ee ie
am eras

se ton neh nw: fe hdg aa
AOS we BE Qe elps germ, ; cme

Pah hekietied aa ak ek ee ne ee ee ae ene n

Po ae OVanetewmatinns wing econo nae 4 bee

oe ieipued Late ao te
ca “ o
OE OMe! IS Ne ek Pe OS en eme

AN ate
Mme Reem 6 & Sw ~ . ¢ \ iP

quibimapharadhedt ied ee
a OR te

Ye orliesiry re

FPR ere ve oem nt it

hat Soe Bare tig SE ve may,
vanes a ‘ " ‘ - : eee ee eine PRA IS ee ~
Pere ne ewe ot owe es, te . i a ’ ~ ideas a
7 ; / h\plintite leds bei te ee
peices ean se :
Rwiestan staan acer See prey ee
NV ey I eer : “ - a yi . teat qy Per oane wy -
naseagrtted . : J 4 sop liaise iment ee penance
=e 7 + an 4° ‘es sles : a hs : : ulin tee Med ~~ a . inca

Ree + Metheny, oy

A de As me balay aire

heteee a ee a
oe ay adie are

5a SP Nite 0 2 Se cin Oe ere tata balked he ee nee

Tice teglthee wn eae ito

me oe ow _ ‘ oe NO eines ty ay, ae

OD 4 we ete S ve adetalintin- tet

en en 7 nseniiitnadaatiiinncaten ae ee

i

ees
- enero pre agen
i deditietenen ae
. - 1 r8 rep pe oe
~ ale ee el ee - -
< SES :

+ Ne an
Rte we Kw CR we KO Oe ee
Oem saie n= wate e Neb yp lee

ee Sn ew
Pan ee

r ote ee
- ein)
RS Oe ay eee —
raatank acl Rien — sab liaiediene ee ee
: mingininedinindes

on a
Oe at eal Ne eR

FOS) OE wy

~~ 9S Nore arene «

~ wy ey

+ wee
i at ee

“es ete re

ee

5 ee eee

Sicdienincadee ee
a oe

— Oe oy nee

7 = Oe ee ee Pe ow
ee nt nh Ba atl ie nn ee By itd
OE Se gr a et me - a + he =!

j

se oie

a mi

Te

Digitized by the Internet Archive
in 2010 with funding from
University of Toronto

http://www.archive.org/details/bulletinofmuseu19harv

?

Ls ee a ie

,

rn

,

o >

«
*

ny cy

via) Mla if)

“yee pagltif fi
Os aaa
~ TAL a

Bee), ,

/ / > de
cams ~
.
P !

BULLETIN

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

HARVARD COLLEGE, IN CAMBRIDGE.

VOL. XIX.

CAMBRIDGE, MASS., U.S.A.»
1890.

UNIVERSITY PRESS:
JOHN WILSON AND SON, CAMBRIDGE, U.S.A.

615333 _
Ai Sat

CONT EATS.

No. 1.— Contributions from the Zodlogical Laboratory. XV. Studies on
Lepidosteus. By E. L. Marx. (9 Plates.) February, 1890

No. 2. — Contributions from the Zodlogical Laboratory. XVI. On the Egg
Membranes and Micropyle of some Osseous Fishes. By C. H. E1géeNMANN.
(8 Plates.) March, 1890

No. 3.— Report on the Results of Dredging by the United States Coast
Survey Steamer “Blake.” XXXII. Report on the Nudibranchs. By R.
eto kiatess)- March: 1800"... e653 ne Oe te i

No. 4.—A Third Supplement to the Fifth Volume of the Terrestrial Air-
Breathing Mollusks of the United States and adjacent Territories. By
W.G. Binney. (11 Plates.) May, 1890

PAGE

129

155

188

No. 1. — Studies on Lepidosteus. Part I. By E. L. Marx!

ConTENTS.
PAGE PAGE
I. Introduction. . .. . . . 1/| B. Historical and Critical Review of
II. Habits of the Young Fishes . 5 the Literature on the Primary
Ill. The Respiratory Function of Egg Membranes and the Mi-
the Air-Bladder. .... 18 cropylein Fishes... . . 54
Vo cmppyology..) . wo. ee QT a; Cyclostomata .. «Vs, sod
1. Egg Membranes. .. . 27 Pere ACnI jy ee hen “OO
me Observations. °. . . 28 e (Ganoidel 60). e) ei.) os 2 Ge
a. Zona Radiata and Vil- Ce eOINOP: fe de sea) | 9s OO
lous Layer. . . . 28 ex@beleostel a) cts). (a GF
& Diicropyle.. . . + -48 1. Zona Radiata and Vil-
eeGranulosa . . «+s 4d lous: Layer «., 22:1 5-3- 68
d. Origin of the Zona Ra- 2. Capsular Membrane . 94
diata and the Villous 3. Micropyle and Plug . 102
aver ys £8 4. Micropylar Cell. . . 110
Pert). ss & » 115)| Bibliography .. «<4 5<%s « 5 120
Peete ws sw Sl. CCU ALD | Explanation of Figures... . . . 128

I. Introduction.

I BECAME deeply interested in the embryology of Lepidosteus through
reading the paper on that subject published by Mr. Agassiz in 1878, and
determined to avail myself of the first opportunity of following up the
study. I desired particularly to pursue the development of the early
stages. A little later I was further incited to this by the brief account
of it which Balfour gave in the second volume of his Comparative Em-
bryology (1881).

I had already formed plans for going to Black Lake, in the vicinity
of Ogdensburg, N. Y., in the spring of 1882, for the purpose of getting
material for the contemplated study, when I learned from Mr. Agassiz
that Balfour had in hand an extensive paper on the subject. Mr. Agas-
siz also informed me that he still had left a part of the material from
which Balfour had been supplied, and he kindly placed this at my

1 Contributions from the Zodlogical Laboratory of the Museum of Comparative

Zoology at Harvard College, No. XV.
VOL. XIX. — NO. l. 1

2 BULLETIN OF THE

disposal. J thought it would be desirable, nevertheless, to procure an
additional supply of eggs and embryos, and especially to endeavor to rear
the young beyond the stages already in my possession.

On account of my duties in College it was impossible for me to leave
Cambridge until nearly the middle of June, — almost a month after the
usual time of spawning. Nevertheless, owing to the extreme backward-
ness of the season, I hoped that I might be able to procure some mate-
rial, and was confirmed in this by correspondence with Mr. J. H. Perry,
through whom I learned that up to a day or two before the time I had
fixed upon for setting out the gar-pike had not spawned.

I arrived at Black Lake on the evening of June 13th. The weather
had meantime grown warm, and the fish had already spawned, but I was
able to secure some eggs which were not very far advanced: in develop-
ment. By anumber of processes I killed and preserved at short inter-
valssets of embryos which presumably belonged to the same lot of
spawn. The eggs which were collected from different localities were
kept in separate earthen-ware dishes and supplied with fresh water every
twelve hours. In this way the embryos were easily kept alive until they
hatched. Then they soon attached themselves by means of their pecu-
liar! maxillary disks to the sides of the dishes, and near the surface of
the water, where they clung with a tenacity truly surprising.

A number of eggs were preserved — principally by means of Kleinen-
berg’s picro-sulphuric mixture —at intervals of a few hours, beginning
on the afternoon of June 14th and extending through several succeeding
days. This method was controlled at intervals by preservations made in
alcohol, in chromic acid, in osmic acid followed by potassic bichromate,
and in the last named reagent alone.

Besides the large number of embryos which were preserved at Black
Lake, I took away with me (June 20) many more — upwards of a hun-
dred — which had recently hatched. These living fishes were carried in
a narrow-necked tin pail, to the sides of which they adhered very firmly.

Instead of returning directly to Cambridge I took the gar-pike with
me to my summer residence on the south shore of Lake Erie, about
forty miles southwest of Buffalo, N.Y. Although this journey extended
over more than three hundred miles the fishes survived it well. The
greater part of it, however, was rendered comparatively easy, since it
was made by steamboat instead of railway. It is fortunate for such an
undertaking that these fishes have so large a yolk-sac, since it obviates
the necessity of procuring food for many days after hatching.

The question of being able to raise: them beyond the stages already

MUSEUM OF COMPARATIVE ZOOLOGY. 3

secured by Mr. Agassiz so evidently depended on finding a suitable
food for them, that I spared no pains to accomplish this end. Many
kinds of meat and fish were minced and fed to them, but none of these
was acceptable. The minced liver, which Mr. Agassiz used with success
at first, was likewise refused. Fragments of meat were suspended in the
water by fine threads, but neither when moved about nor when left per-
fectly quiet did they seem to attract attention. Great numbers of water-
fleas (Cladocera) were put in the water with the young fishes, but the
latter made no attempts to catch them. It was not until after many
fruitless trials that organisms were found which were seized with such
eagerness, and so persistently, as to leave no doubt that they were the
natural food of the young gar-pike. These were the larve of the common
mosquito. They constituted the exclusive diet of the young fishes until
the latter became large enough to catch and swailow minute “mud
minnows ” (Fundulus), on which they subsequently fed as long as they
were kept alive.

When it was once ascertained that the young fishes would take mos-
quito larvee, there was no longer any serious question about the feasibil-
ity of rearing them, nor was it doubtful that these larvee formed their
natural food, for the shallow and quiet waters at the margins of Black
Lake and along the creeks which feed it abound in mosquitos. It
was by means of this diet that the fishes were kept in a thriving con-
dition during the stages immediately following the absorption of the
yolk.

From the 20th of June until the Ist of July specimens were, with a
few exceptions, killed every twenty-four hours; and from the Ist of July
until the beginning of August, usually at intervals of about forty-eight
hours.

By the 3d of August there remained besides those which had not
been preserved only about a dozen living fishes. On that date I started
for Newport, R. I., travelling by rail to New York. These remaining
fishes were carried by hand in a tin pail suspended by a spring; but
owing to the difficulty of carrying in the pail a sufficient number of
mosquito larve, and more particularly to the impossibility of properly
renewing the water, about half of them succumbed to the unfavorable
conditions of railway travel and were put into alcohol. One more died
on the way from New York to Newport; but the remaining ones, hay-
ing been fed on larve after my arrival at Newport, appeared to thrive.
At the end of a month they were taken to Cambridge, where they were
put into a large glass jar and supplied with running hydrant water.

4 BULLETIN OF THE

Here they were also kept on the diet of mosquito larve until from one
cause or another they had all died.

My failure to secure early stages of the eggs in the spring of 1882
made me desirous of repeating the attempt at a more seasonable time
the following year. With this object in view I left Cambridge for
Ogdensburg, May 18, 1883.

Judging from my previous experience that it would be difficult to
procure fertilized eggs in sufficient quantities without great labor, if they
were to be individually detached from the rocks, I procured several yards
of thin muslin of a color resembling the stones in the lake. I planned
sinking this and loading it with small stones in the water on some of
the “points” most frequented by the gar-pikes at spawning time. I
hoped in that way to secure a large number of eggs firmly attached to
the cloth, which I could then remove to a box suitably provided with
wire nettings to allow the necessary circulation of the water. Had it
proved successful, this device would have enabled me to have under
control the eggs thus acquired, and would have allowed me a degree of
certainty as to the age of the preserved material not otherwise easily
attainable.

Unfortunately for my plans, the weather proved to be in several ways
very unpropitious. A long period of cold and rain delayed the spawn-
ing to a time much later than common, and when at length, a few days
after the lst of June, the weather and water became warm enough to
impel to the act of spawning, such high winds prevailed that it was
impossible to watch the movements of the fishes, and the most of them
had spawned before the water became quiet enough to allow one to dis-
cover their places of rendezvous. Some of the localities which they had
visited with the greatest constancy during the past years were appar-
ently deserted. Moreover, the cloths, which had of necessity been an-
chored near the shore, were either set free by the dashing of the waves,
or rolled into ropes which presented a comparatively small surface for
the reception of ova. |

The limited time in the latter part of May left at my command, after
completing some work which I was compelled to take with me, was util-
ized in studying the ovarian ova of females captured before the spawn-
ing period arrived, and in some attempts at artificial fertilization. I
then succeeded in getting a fairly satisfactory knowledge of the interest-
ing structure of the egg membranes and of the micropyle, but I did not
learn the peculiar relation of the latter to the granulosa until a few

MUSEUM OF COMPARATIVE ZOOLOGY. 5

months later, when I had made microscopic sections of the whole ovarian
egg, including membrane and granulosa.

As the outcome of this journey I secured a number of series of eggs,
beginning with the early stages of segmentation, from which I prepared
at intervals and by various methods a considerable quantity of material.
I was also able to bring to Cambridge about two hundred young fishes
just hatched, Some of these were kept alive until September 26, 1886,
—nearly three and a half years.

The fishes brought to Cambridge were put into running hydrant
water and fed on mosquito larvee for several weeks, — until about the
1st of August, — when they were large enough to swallow small ‘‘ mud
minnows” of nearly their own size. These were gradually substituted
for the larve, and those fishes which were large enough to avail them-
selves of this kind of diet grew much more rapidly than their mates.

I have also received some eggs from Mr. Perry since my last trip to
Black Lake, and although his attempts at fertilization did not prove to
be more successful than my own, I stili hope to secure before long the
early stages which are needed to fill the gaps in my material.

II. Habits of the Young Fishes.

The habits of young gar-pike have already been quite fully described
by A. Agassiz (’78*) and Wilder (’76, ’77), so that I shall not have
much to add to what has been previously published.

When first hatched the fish is so small in comparison with the size
of the yolk-sac that it swims only with the greatest difficulty, and its
movements are anything but graceful. It is so disinclined to swim, that,
were it left alone in water sufficiently pure to meet its requirements, I
have no doubt it would not move from the point of its first attachment
for many hours, or even days. When hatched in confinement the young
fishes always swim nearly up to the surface of the water and attach them-
selves to the sides of the dish. When there are a large number of them
they may attach themselves to floating objects. Frequently the super-
ficial film of the water — aided possibly by secretions from the oral disk
— serves to support an individual in the middle of the dish. Sometimes
half a dozen or more individuals form in a cluster, and appear to hang
suspended simply from the surface of the water. It is evident that they
are not merely floating in the vertical position, because in such cases
the surface of the water in their vicinity is always more or less depressed,
and upon the slightest touch the fishes begin at once to sink slowly ; if

6 BULLETIN OF THE

the water has not been too much disturbed, they make no motion while
sinking until they have nearly reached the bottom. Before they actu-
ally touch the bottom of the dish they appear to recognize their prox-
imity to it, and then begin to make vigorous efforts to swim up to the
surface again. ‘This is apparently a very laborious undertaking, and, if
they fail to attach themselves at once, they again begin to sink slowly ;
they seldom attach themselves at the bottom, — especially if the water
has remained for some time unchanged, — but always as near the sur-
face of the water as possible. If there are too many to be accommodated
in a single row, those last to come crowd in between the tails of those
already attached, thus forming a second row ; but if there are still others,
they usually attach themselves to other fishes rather than take a lower
position on the sides of the dish. During the period of yolk absorption
they hang pendent and nearly motionless, except for the respiratory
movements; those which hang from the surface of the water are verti-
cal, and any sudden motion in the water shows that their bodies are
quite limp. When the absorption of the yolk is well advanced, the
flexibility of the body is shown in a striking way by the snake-like
motions which the animal slowly executes while remaining attached.
The disinclination to swim lasts about as long as the yolk-sac per-
sists. With the gradual disappearance of the latter, the fishes show an
increasing tendency to swim about. When at length the sac is nearly
absorbed they rest in quite another way. They float near the surface,
taking a horizontal position, and remaining perfectly straight and mo-
tionless until disturbed; whereupon, by vigorous strokes of the tail,
they swim away with remarkable celerity. In transferring the fishes
from one dish to another I was accustomed to use a small spoon, but
after the absorption of the yolk-sac I found it exceedingly difficult to
capture them in that way, so rapid were their movements. The stage
at which the fishes begin to swim and float is reached in eight or ten
days after hatching. Even at this early age locomotion is accomplished
by two distinct methods. The rapid motions are executed by vigorous

strokes of the whole caudal region. A slower, gliding motion is main- .

tained by means of the very rapid vibrations of the extreme end of the
tail, which are so characteristic of the caudal filament at a later stage,
and by the still more rapid motions of the pectoral fins. Not only do
the pectorals vibrate when the tip of the tail is motionless, and vice versa,
but either of the pectorals may be in rapid motion while the other is at
rest. This second method of locomotion is apparently very serviceable
to the fish, in allowing it to approach its prey unobserved.

MUSEUM OF COMPARATIVE ZOOLOGY. 7

At a later stage of development, when the upper lobe of the tail is re-
duced to a caudal filament, this gliding motion is accomplished, princi-
pally at least, by the action of the pectorals. When the fish is advancing,
these fins are directed obliquely backward ; but when, as often happens,
the motion becomes retrogressive, they are directed more nearly at right
angles to the body. The motion of the fins is so rapid, that I have been
unable to determine by observation if, as is probable, the direction of
the stroke is reversed in the two cases. Not only the direction of the
long axis of the fin, but also the inclination of its transverse axis to the
horizon, is conspicuously changed at such times.

The vibrating movement of the caudal filament perhaps assists the
forward motion of the fish, but it cannot be considered essential to it,
since the filament often remains motionless while the animal is gliding
by means of the pectorals. The amplitude of the vibrations made by
the filament is not great in any case, — about 15°,—and the terminal
half alone is vibratory. When in motion the direction of its axis
is usually continuous with that of the spinal column, although it may
droop more or less while in motion, and is quite liable to do so when at
rest; it then presents an even curvature, as seen from the side, and
often inclines a little either to the right or left.

When the fish is stemming a current, or, in swimming, is directing
the head downward, the caudal filament is kept in rapid vibration ; it
then takes a dorsal turn, and the curvature is rather abrupt at its base.
The whole curvature may amount to enough to make the extreme tip
of the filament perpendicular to the axis of the body, but usually it is
much less.

During the night of August 6-7, 1883, one of the individuals hatched
in June of that year escaped from the tank, and was found in the‘morn-
ing lying in only sufficient water to keep the body moist. Upon being
returned to the tank, though still able to swim, it showed evident signs
of weakness. The body was considerably arched, just as it has been fig-
ured for somewhat younger fishes by A. Agassiz (’78*, Plate IV. Fig. 39,
and Plate V.). I think this case suggests an explanation of the peculiar
curved shape exhibited by the fishes reared by Mr. Agassiz. I had
already in the previous year imagined that the arched condition was not
common, for all my fishes were quite straight, at least so long as they were
well nourished. The curved condition of the escaped fish was apparently
due to muscular weakness ; the curvature was also accompanied by a
slight distortion from the sagittal plane. Inasmuch as it subsequently
regained its normal condition and became straight, I have no doubt that

8 . BULLETIN OF THE

the curved condition is abnormal, and its appearance in the fishes raised
by Mr. Agassiz was probably owing to insufficient nutrition. The con-
trol of the pectorals in the case of this escaped fish was not lost, although
somewhat impaired. The fish was also able at once to execute vigorous
though ill-planned motions with the tail, and was therefore able to swim
rapidly, but in a reckless manner. The control of the caudal filament,
however, was entirely lost; there was not the least trace of the rapid
vibratory movement, even when the pectorals were in active motion. By
the following day the use of the caudal filament had been partially re-
gained; but its vibrations were rather feeble, and were resumed only
after long intervals of repose.

The temperature of the water greatly affects the power of locomotion,
and very cold water may even produce fatal results in a comparatively
short time. In the summer of 1882 I placed a dozen or more fishes in
the cold water of a spring; within twelve hours half of them were dead.
The first signs of an abnormal condition are shown by uncorrelated
movements, reckless swimming, and the inability to keep the dorso-
ventral axis vertical. The appearance is as though the centre of gravity
were located above the centre of volume, and the fish gradually became
incapable of remaining in its normal position of unstable equilibrium. In
swimming the body rolls from side to side. After a time the fish sinks
to the bottom of the vessel, and can regain the surface only with consid-
erable exertion ; it cannot remain at the surface in a motionless condi-
tion. At length the body usually becomes curved sidewise. In this
condition the fish may remain for many hours, or even several days.
Restoration to fresh water of the ordinary temperature seems to have
only a temporary effect, or none. The only way to afford relief is to
place the fishes in direct sunlight until the water becomes warm, when,
if not already too much affected, they will gradually recover.

The movements of the eyes are principally in the horizontal plane.
When one eye is directed obliquely forward, the other looks obliquely
backward at about the same angle, so that the axes of the two eyes
are kept approximately parallel.

The manner in which young gar-pike capture and swallow their prey
is interesting, and serves to show why it is so difficult to get them to
feed on anything except living objects. When very young, as previously
stated, they will not feed on anything but mosquito larve. The fish
always approaches the larva by a slow, even motion, resulting from the
vibration of the pectorals and the tip of the tail, until the prey is about
opposite the middle of the “bill”; then with a quick lateral motion of

MUSEUM OF COMPARATIVE ZOOLOGY. 9

the head the larva is snapped up. Occasionally it is bitten through,
and, whether struggling or not, is allowed to escape; but the reason of
this I did not discover. Usually the larva is not snapped at until by
some slight movement it shows signs of being alive, and then only a
single snap is made. If unsuccessful, the fish remains quite motionless,
_ or glides slowly away by means of the same motion of the pectoral fins.
Fishes, especially when less than an inch in length and when well fed,
‘rarely make a lunge forward toward their prey when they snap at it ;
the motion is usually only a sudden sidewise bend in the neck region.
So cautious are the fishes not to snap at dead larve, that, after they
have advanced so that the insect is opposite the middle of the bill, if it
does not move, they begin to swim obliquely forward, pushing the larva
before them without allowing it to glide backward along the bill. In
this way a larva may be pushed about several inches before it makes
the fatal movement which betrays its condition ; or it may, if it remain
entirely passive, escape altogether, since the fish, failing to discover evi-
dences of life, leaves it and glides off in search of other food.

The number of insects caught by a single gar-pike is undoubtedly
large. I was accustomed to feed the fishes twice a day. In July, 1883,
I had the curiosity to watch the largest of them, then a month old and
about 25 inches (62 mm.) long, during its feeding. In the course of
ten minutes it caught twenty-one large mosquito larve and made nine
ineffectual attempts at seizure. But the voracity of young fishes was
still more forcibly exhibited when, early in August, they were given
small minnows (Fundulus) for food ; for then they did not hesitate to
catch the minnows even when the latter were considerably larger than
themselves ; but they succeeded in swallowing only those which were
of their own diameter or smaller. Since the minnows have much
thicker heads than the gar-pikes, the total weight of the former was
doubtless always somewhat less than that of the latter.

The minnows were caught in the same way as the mosquito larva,
by a sudden sidewise motion of the head; but being too large to be
swallowed at a single gulp, they were at first impaled on the sharp
teeth, and then by a series of deliberate movements put in a position
' to be swallowed. Almost without exception, the minnows were swal-
lowed head first, irrespective of the region of the body first seized. If
caught near its tail, as usually happens, then the prey is moved be-
tween the jaws — to which it lies crosswise — by successive shiftings,
until it is held near the base of its head. A few movements usually
suffice to make it take a direction parallel to the jaws, and the head

10 BULLETIN OF THE

end of the prey is thus brought to lie in the gar-pike’s throat, which is
often greatly distended by it. The movements by which the shifting of
the minnow is accomplished are rather complicated, and require a nice
correlation to be successful. During the process of transferring it cross-
wise between the jaws, the latter have to be opened quickly, and this
motion is instantly followed by a quick lateral motion of the whole head
in the proper direction, This lateral motion is accompanied by two
others ; one a forward thrust of the whole body, and the other a de-
pression of the floor of the mouth. To accomplish the first movement
there is a preparatory curving of the post-anal portion of the body, the
sudden straightening of which at the instant the jaws are loosened gives
the necessary forward impetus, and helps to prevent the escape of the
prey; this is further guarded against by the second motion, — the de-
pression of the floor of the mouth. The gill covers being in contact
with the sides of the body, this latter motion produces a tendency to a
vacuum in the mouth, which can be satisfied only by a sudden influx of
water between the jaws. The current thus produced of course has the
effect of carrying with it any movable object in the mouth or its imme-
diate vicinity, and of impeding the escape of the prey until the jaws are
again closed upon it. While the hold upon the prey is gradually shifted
from its tail region to its head region, the part of the jaws which holds
it is also not the same as at first. By the time the fish has been fully
shifted laterally, it will be very near the base of the jaws, for at each
loosening of the latter they have been thrust forward a little by the
motion of the whole body. But that is not all, for when the prey has
been shifted back to near the base of the jaws, — as one can see better
in the case of the larger gar-pikes, — each subsequent movement of the
latter causes the prey to take a slightly different direction, so that it is
finally swung around until it is parallel with the jaws instead of cross-
wise to them. Iam not entirely certain how this is effected ; but it is
all accomplished when the fish is held in the jaws near their base. It

. is possible that the teeth of one ramus of the jaw are not loosened quite

as promptly as those of the other, and that as a consequence they act
for an instant as a sort of pivot for the rotation of the prey. But that
both sides of the jaw are ultimately set free at each motion is probable,
from the great liability of the prey to escape at this very critical step in
the process. Or it may be that the lower jaw is moved slightly toward
one side as the jaws are being opened, thus giving a swing to the prey
which changes its axis slightly before it is again caught by the closing
jaws. Whatever the means by which it is effected, it is certain that

MUSEUM OF COMPARATIVE ZOOLOGY. it

this changing of the direction of the prey is all accomplished at the time
of the renewals of the grasp.

After the fish has been thus nearly oriented, its head end is introduced
into the throat by a single forward lunge on the part of the gar-pike.
These forward movements are continued for some time, the jaws being
slightly opened at each lunge. When the head is well introduced into
the gullet, the jaws are no longer suddenly opened and closed, but re-
main more or less gaping, while the prey slowly disappears, doubtless
being drawn on by peristaltic motions of the gullet. There are usually
slight motions of the jaws during this latter process, — a slow opening
and partial closing of them. The movement of the lower jaw is some-
what unsteady, almost tremulous, a peculiarity which is also seen, al-
though not so distinctly, at other times than when feeding.

The advantage of a great divaricability of the rami of the lower jaw
and of the fulness of the skin connecting them is at once apparent
when a comparatively large fish is half swallowed, for then the thin
membranous floor of the mouth is greatly distended and the rami
pushed far apart. A side view of the head of the fish then resembles
somewhat the appearance of the throat of a feeding pelican.

Gar-pikes sometimes take food which they apparently discover to be
objectionable only after they have partially swallowed it. In such cases,
it is ejected from the throat with a sudden jet of water.

They do not hesitate to snap at each other when kept in confinement,
as I have many times observed, and as the mutilated condition of the
fins, and especially of the tail, makes very evident. I think it may be
inferred that they are not altogether free from danger from their own
kind when living in their natural haunts, for they always show a remark-
able sensitiveness to being touched in the region of the tail. The caudal
filament especially is so sensitive that the slightest touch from a foreign
body causes the fish to dart away with utmost speed, whereas one may
touch any other part of the body with comparatively little effect. The
young fishes become easily accustomed to the touch of the hands, and
may even be lifted altogether out of the water without offering any
resistance, provided it be done gently and without any quick motion.
But none of them ever become so tame as to allow the slightest contact
with the caudal filament without immediate efforts to escape. It is
almost invariably the tail end which is snapped at by their mates,
though I have a few times seen two individuals with interlocked jaws
carry on a short contest without fatal results to either. I have known
of only one case in which a gar-pike swallowed one of its mates. Near

pO BULLETIN OF THE

the end of July, 1883, I found an individual in the process of swallow-
ing a somewhat smaller mate. The bill of the victim and part of its
head were still protruding from the distended jaws of the captor; so
that this individual was swallowed tail first, contrary to the more usual
method.

The movements of the gill covers vary in rapidity at different times,
but they are always executed with considerable promptness. Their
adduction is quickly followed by their abduction, but the interval that
follows before another adduction is usually rather prolonged ; it is the

more variable element. It may be so short as to make the abduction .

and adduction separated by equal intervals, or it may be prolonged. to
several seconds. These respiratory movements seldom exceed sixty per
minute, and may diminish during the torpor of winter to scarcely more
than one a minute.

The emission of bubbles of gas, which begins soon after the young
fishes detach themselves from their fixed supports, at first takes place
through the gills of one side. It is usually preceded by a forward
lurch of the body, accompanied by a slight rolling to one side and the
elevation of the gill covers. The bubble usually emerges before the fish
regains its normal position, and consequently comes through the gill
slits of the side which happens to be uppermost. Occasionally two
smaller bubbles escape from beneath the gill covers, one from each side.
When the fishes have become much older, the amount of gas is so great
that the bubbles often escape not only from beneath both gill covers,
but also through the mouth opening, and the rolling of the body does
not always accompany the escape of gas.

In the earlier stages of their growth the gar-pikes remain most of the
time very near the surface of the water in a horizontal position. In
such cases the only premonitory symptoms of the escape of gas are the
motions just described ; but as they grow older they gradually habituate

themselves to lying in deeper water, and then they almost invariably

ascend to the surface before emitting gas. The ascent is nearly always
accomplished by a slow forward and obliquely upward motion, the body
being at an angle of about 45° with the horizon. The motion is usually
deliberate, and at a uniform rate. After reaching the surface the body
is allowed to assume the horizontal position before any effort is made to
expel the gas. This motion of ascent is so characteristic, that after
studying their habits one may predict with tolerable certainty whether
a given fish is about to emit gas.

I believe the slight rolling of the body to one side is for the purpose

MUSEUM OF COMPARATIVE ZOOLOGY. 13

of bringing the slit in the roof of the throat, through which the gas
must be forced, into a position more favorable for its escape. If the
opening through which the gas is obliged to pass is directed downward,
its expulsion will require greater effort than if the opening is directed
sidewise. An advantage depending on the same physical properties of
the gas may perhaps also explain the universal habit of coming to the
surface of the water to disengage the bubble. At least, the pressure of
the water to be overcome in forcing out a bubble when the fish is at a
considerable depth must be greater than when it is near the surface.

The escape of gas, which may be several times repeated during the
process of swallowing a large fish, shows clearly enough that the bubbles
are not simply air taken in at the mouth to be immediately discharged
through the gill openings. The repeated emission of gas from the same
fish, without the possibility of any fresh air having been taken in through
the mouth, even led me to conjecture at one time that air was never
taken in through the mouth. At least, it was certain that the young
fishes often discharged gas without lifting any portion of the body out
of the water.

The rate at which gas escapes from the gill openings is extremely
variable, depending on the temperature of the water, the recency of
feeding, etc. Perhaps the following observations give a fair idea of
the rate during a warm summer day. In the course of ten minutes
(August 6) eight fishes together emitted forty bubbles, or an average of
one in two minutes for each fish. On August 17, a single fish, 62 mm.
long, came to the surface six times in ten minutes, and caused bubbles
to escape from the gills.

In the case of older fishes, snapping at the prey is frequently accom-
panied by an escape of gas, which is apparently involuntary, the sudden
motion of the head and the opening of the jaws being sufficient to cause
the escape of a few bubbles.

The nature of the gas will be considered in the following section.

III. The Respiratory Function of the Air-Bladder.

Lepidosteus, as well as some other fishes, has the habit of coming to
the surface of the water and emitting through the gill slits or the oral
opening bubbles of gas. This habit has attracted the attention of all
who have had the opportunity of examining the fish while living.

Poey (55, p. 136) observed that, when placed in a basin of water,
‘every five or eight minutes he [ Lepidosteus| would come to the sur-

14 BULLETIN OF THE

face to swallow a mouthful of air, returning downwards immediately.
One second after, half a dozen air-bubbles, some quite large, escaped by
the opening of the branchiz. The air,” he adds, “‘ remains in the blad-
der one second, sometimes one and a half, and this time is probably
sufficient for the absorption, digestion, and expulsion of the inspired air.
Besides, it is certain that, the animal not attempting to swim, the
bladder was not used in augmenting or diminishing the density of the
body, as most fish do, in order to ascend and descend in water.” Poey
further strengthened his opinion that “some sort of pulmonary respira-
tion existed in the Lepidosteus” by dissections and injections of the
aorta, which showed the great vascularity of the bladder (pp. 134-136).

Louis Agassiz (’57), in exhibiting before the Boston Society of Natu-
ral History some young living gar-pikes, called attention to the fact that
this fish is “ remarkable for the large quantity of air which escapes from
its mouth. The source of this Prof. Agassiz had not been able satis-
factorily to determine. At certain times it approaches the surface of
the water, and seems to take in air, but he could not think that so
large a quantity as is seen adhering in the form of bubbles to the sides
of the gills could have been swallowed, nor could he suppose that it
could be secreted from the gills themselves.”

More recently Wilder (’76 and 777) also has studied the young of this
fish. He says (76, pp. 151-153): “ Very often these young indi-
viduals of Z. osseus, and more frequently the adults of the smaller
species (L. platystomus), would protrude the snouts from the water in
the respiratory act ; but the length of the Jaws made it impossible to
determine whether this was intentional, and for the purpose of inhaling
as well as of exhaling the air.” He inclined to the opinion that air is
taken in as well as given out, because the fishes uniformly approached
the surface, whereas, if exhalation were alone sought, that “ could be as
well accomplished at any depth.” As the result of some experiments in
restraining the motions of Amza, he says: ‘There seems no doubt that

’ with Amza there is a true inspiration as well as expiration of the air.

The same may be considered probable, though not yet proved, with
Lepidosteus.”

So far as I know, these are the only published accounts of this pe-
culiar habit in the case of Lepidosteus. While Agassiz maintained a
conservative attitude regarding the question of the source and nature of
the bubbles, Poey and Wilder expressed the conviction that it was air
which had served the ordinary purposes of respiration, and both of
them supported their belief with arguments.

MUSEUM OF COMPARATIVE ZOOLOGY. 15

Close attention to the movements of young fishes kept in confine-
ment led me at first to doubt the accuracy of this conclusion. The
result claimed by Moreau (’63 and ’63*, p. 820) — that, after the arti-
ficial removal of gas from the air-bladder of fishes generally, the gas
is regularly renewed, even when they are restrained from coming to the
surface — also served to confirm my doubts upon this point. Usually
it was only the tip of the bill which protruded from the water, and it
seemed incredible that such a movement could be sufficient for the
acquisition of even a limited volume of air. In addition, it has often
occurred that several times in succession a fish has been observed to
come near the surface and emit bubbles of gas without any portion of the
head region breaking the surface of the water.

This seemed conclusive upon at least one point: coming near the
surface could not be solely for the purpose of acquiring air. I have
given elsewhere (pp. 12, 13) what appeared to me to be the probable ex-
planation of the upward movement of the fish previous to emitting gas.

The possibility that fresh air was not always — perhaps not generally
—acquired, led me to reflect further about the possible functions of
the air-bladder. Starting with the assumption, that no organ arises
absolutely de novo, and that the air-bladder is the result of a differen-
tiation in the alimentary tube, what, in a phylogenetic sense, was likely
to have been its first function? Did it arise as a hydrostatic apparatus
to be ultimately diverted to the service of respiration for the higher
vertebrates, or was the hydrostatic function already a superimposed
one? Might it not be that the original function was purely one of
excretion ? Perhaps in so primitive a fish as Lepidosteus this original.
function was still the dominant one.

These reflections led me to think it more than ever desirable to
subject the gas emitted by the gar-pike to chemical analysis, — really
the only method of arriving at definite and satisfactory conclusions.
I therefore determined to undertake an analysis as soon as the fishes
had attained a sufficient size to make the amount of gas given off in the
course of a few hours voluminous enough for easy experimentation.

Accordingly, in December, 1883, when the fishes remaining were six
or eight inches in length, some preliminary experiments were begun.
The analysis was attempted by using a bent tube of about 10 mm.
calibre, roughly graduated to 5 mm.; but the result showed that satis-
factory conclusions were attainable only by employing more suitable
apparatus. Through the kindness of Professor Cooke the mercurial
bath and other apparatus for gas analysis at the Chemical Laboratory

16 BULLETIN OF THE

of Harvard College were placed at my disposal, and all the subsequent
work of analysis was carried on in that Laboratory. I am also indebted
to Professor Cooke for valuable suggestions during the progress of the
work. To Prof. A. V. E. Young, then a private student in the Chem-
ical Laboratory, I am under especial obligation, since he carried on the
manipulations and measurements of the gases. From his skill and
previous experience in gas analysis the results are entitled to more
consideration than if the measurements had been conducted by one
less familiar with such manipulations. |

It was my aim in these experiments to ascertain, without sacrificing
the fishes, the composition of the gas in the air-bladder, or at least of
that which was emitted in the form of bubbles, presumably from the air-
bladder. No attempt was made to ascertain the results of branchial
respiration.

The collection of the gas, even from fishes 20 cm. long, is a tedious
process. An inverted glass funnel large enough to allow the fish to
swim freely inside, was first employed ; but the inclined sides of the
funnel not allowing the fish to assume a horizontal position except
when deep in the water, appeared to interfere with the emission of
the gas. After ascending to the apex of the funnel, the fish would
make violent efforts, apparently for the purpose of attaining the surface
of the water, but only occasionally would it emit gas under these cir-
cumstances, the emission usually
taking place only after it had be-
gun to descend. If freed from
the restraint of the funnel, the fish
invariably came at once to the
surface outside the funnel, and
emitted the customary bubbles.

Afterwards a nearly flat-bot-
tomed glass dish, only a trifle
| IW smaller than the one containing
ggo EM xqwurriituing) *S the fish, was inverted over the

latter (see Figure 1); but this
apparatus was only slightly more
successful than the one at first employed. The fishes came to the under
surface of the inverted vessel, and after long intervals some of them
emitted small portions of gas, but this was too small an amount to
be satisfactorily analyzed. Another method — the one finally employed,
although a tedious process — will be mentioned further on.

FIGURE 1.

MUSEUM OF COMPARATIVE ZOOLOGY. ag

In my preliminary analyses I had found that a very thin stratum of
air under the inverted vessel was sufficient to allow the ordinary emis-
sion of gas. The first experiment in which the gas was accurately
analyzed was therefore conducted upon fishes confined under the in-
verted dish, beneath which was a bubble of air of known volume,
which had been deprived of ali its carbon dioxide. The details of the
experiment were as follows.

Experiment 1.— Dec. 13, 1883. At 11:45 a.m. six gar-pikes, varying
in length from 13 to 21 cm., were removed from running water (+ 12° C.)
to the experimental jar, containing recently distilled aerated water at
about the temperature of the room (+20°C.), over which the glass
vessel, with a bubble of air (deprived of carbon dioxide) measuring
66.71 c.c.1 had been previously inverted. The experiment was con-
tinued until 3:15 p. m. (34 hours) without renewing the water. The
fishes were then taken out and the gas collected. In the final transfer
over the mercurial bath a portion of the gas was lost, so that its total
volume could not be ascertained. The portion remaining was collected
over mercury, and found to measure 19.27 c¢.c. (reduced). Treated with
potassic hydrate, the volume was diminished to 19.07 cc. Absorbed,
.20 ¢.c. = 1.04% = carbon dioxide. This volume (19.07) transferred
to water and treated with pyrogallate of potash measured over water
16.27 c.c. Absorbed, 2.80 ¢.c. = 14.5% = oxygen.

Experiment 2. — Dec. 14, 12:30 p.m. Six gar-pikes (the same as in Ex-
periment 1) were removed from water (+11° C.) to the experimental jar
containing water at +21° C. The inverted vessel contained 65.20 c.c.
of air deprived of its carbon dioxide. The experiment was continued
until 4 p. m. (34 hours). During the last hour all the fishes remained
at the top of the water, and became somewhat swollen in appearance.
One of the six did not recover from the effects of the abnormal
conditions. The gas, re-collected and measured, was found to have

1 The eudiometer employed was graduated to fifths of a cubic centimeter.
All the volumes given are the volumes reduced to 0° C. and 1 m. pressure.

_? Ina preliminary experiment it was found that under nearly the same condi-
tions there had been a slight increase in volume. The conditions were as follows.
Dec. 3, 1883, 12:30 p.m. Four fishes (17-20 cm.) were transferred from cold water
into an experimental jar (Fig. 1, p. 16) containing water at +20° C. and a bubble
of ordinary air measuring 53c.c. At 2:30 p.m. five other somewhat smaller fishes
were added. At 4:30 p.m. the experiment was discontinued, and it was found that
the bubble had increased to about 56.5 c.c., the increase being nearly 8.5 c.c.

VOL. XIX. — NO. 1. 2

18 BULLETIN OF THE

diminished nearly 18.3% in volume ; it measured only 53.29 ¢.c. This
diminution, however, is evidently not to be accounted for as the result
of actual absorption by the fishes, but rather as the result of their over-
distention with gas swallowed, — probably in the vain endeavor to com-
pensate themselves for the altered condition of the atmosphere. The
recovered gas (53.29 c.c.) was divided into two portions (A, B), and
treated separately.

A = 43.20 c.c. (reduced) measured over water. This was treated
with potassic hydrate solution, but the volume remained unchanged.
Therefore carbon dioxide = zero. This volume (43.20 cc.) was then
treated with pyrogallate, which reduced it to 38.88 c.c. The diminu-
tion, 4.32 cc., = 10% = oxygen. :

B. The average of three readings made this volume = 10.07 c.c.
Treated with pyrogallate, it was reduced to 9.12 ¢.c., the reduction in
volume, 0.95 c.c., being equivalent to 9.44+ % = oxygen.

Experiment 3. — The gas for this experiment was secured by collecting
the bubbles as they were given off by the fishes. A number of the gar-
pikés were placed in an uncovered experimental jar containing recently
distilled water at +20° C. It was found that by using a small glass fun-
nel, — the one employed was about 65 mm. in diameter, — held near the
surface of the water, the fishes were not prevented from the ordinary
movements accompanying the emission of bubbles, as they were when
confined under a larger funnel. The funnel was held so that the tip of
the snout, but not the gill region, projected beyond its edge. The
otherwise very slow process of collection is somewhat hastened by em-
ploying a number of fishes in the same jar. As they rise one after
another, at comparatively short intervals, the collector, anticipating
their movements, holds the funnel in the proper position, and seldom
fails to secure the desired bubbles. But even in this manner the
amount to be collected in the course of a few hours is not large. In
this experiment the (reduced) volume of gas collected in about two
hours was 9.54 c.c. Treated with potassic hydrate, the volume was
reduced to 9.38 c.c. The difference = 0.16 cc. = 1.7-% = carbon
dioxide. Treated with pyrogallate, the volume was further reduced to
8.20 cc. The diminution, 1.18 c.c., == 12.4—% = oxygen.

For convenience of comparison the results of measurements may be
tabulated as follows :—

MUSEUM OF COMPARATIVE ZOOLOGY. 19

Volumes in c.c. reduced to 0° C. and

1m. pressure. ae as cy re cent
sos After Potas- Aft on Dioxide. xygen.
Original. | sic Hedrata Bir icaliate.
I. Mixed air, 19.27 19.07 16.27 1.04 14.5+
it ip 53.29
A, 43.20 43.20 38.88 0. 10.
B, 10.07 —- 9.12 0. 9.4
III. Respired air, 9.54 9.38 8.20 ie 12.4—

In considering the significance of these results, it is to be kept in mind
that there are two principal problems involved. I have not attempted
to find out whether zztrogen is consumed or produced in this process ; but
simply to ascertain the changes effected in the atmosphere respired as
far ag regards (1) the oxygen and (2) the carbon dioxide. The difficulty
in drawing at once a satisfactory conclusion from the analyses rests pri-
marily on the fact that the emitted gas could not be collected under
conditions which allowed it to be assumed that at the time of analysis
it was in the same condition which it presents in the air-bladder of the
fish. For previous to analysis the gas had bubbled through the water
in which the fishes were living, and had remained exposed to a limited
portion of its surface. The reduction of the proportion of oxygen con-
tained in the atmospheric air is so great, — between one third and
one half, —that the influence of the water upon the composition of the
emitted gas in this respect would certainly have been too insignificant
to modify essentially the result. With the carbon dioxide, however, the
ease is different. The coefficient of absorption in water for the latter
is immensely greater than for oxygen, and the total amount of carbon
dioxide in comparison with the volume of the water to which it was
exposed is so small (never having exceeded 1.7% of the volume of gas
collected) as to make immediate deductions from the analyses of little
value.

In a series of very carefully conducted experiments upon the intestinal
respiration of Cobztis fossilis, Baumert (’53) arrived at the conclusion
that the gas from the intestine, in bubbling through the water and
during its subsequent exposure to that liquid, was not essentially altered
either by the water or the gases contained in it (p. 48). The condi-
tions under which his collections of gas. were made’ were so like

1 Baumert (53, p. 39) employed five or six fishes (Cobitis fossilis) of medium size
which were kept in a vessel containing about twelve litres of water from the river
Oder. The gas was collected by means of a large inverted funnel terminating
in a small-necked receiver, and usually about two hours were required for its
accumulation. ‘Three analyses, made at intervals of three days without any

20 BULLETIN OF THE

those which I employed, that it seems justifiable to make use of his |
conclusions in the present case. ‘The ingenious device employed by
Baumert (’53, p. 46), in verifying the accuracy of his previous results,
to suppress the branchial respiration, and thus secure the effects of
intestinal respiration alone, could not, from the nature of the difference
between the modes of respiration, be made available in the case of
Lepidosteus.

The means (viz. protracted boiling) which Baumert employed for
extracting the absorbed gases contained in the water of experimenta-
tion were as complete as the chemical methods at his time permitted.
Since then, however, Grehant (69) has shown that simply boiling gas-
impregnated water, although sufficient to eliminate all the oxygen and
nitrogen, does not remove all the carbon dioxide, but that a combina-
tion of the mercurial pump with the method of boiling is capable of
accomplishing this result. By this process he has demonstrated that

renewal of the water, gave in 100 volumes of gas respectively 1.77, 0.47, and 0.13
volumes of carbon dioxide, and 10.46, 13.71, and 11.92 of oxygen. It will be seen,
therefore, that not only the conditions under which the collections were made have
been in both cases very similar, but also that the composition of the gas as deter-
mined by analysis was nearly the same in Lepidosteus that it was in Cobitis. Since
the ratio of the gases in the two cases was practically the same, there is no reason
to suppose that the water absorbed a greater proportion of gas in one instance than
in the other.

As regards the temperature of the water at which the experiments were made,
although not definitely stated by Baumert, it is safe to infer, from the temperature
at which his numerous experiments on branchial respiration were made, that it was
considerably lower than the temperature of the water in which the Lepidostei were
placed, so that the tendency to an absorption of carbon dioxide in the latter case
would certainly have been /ess than in Baumert’s experiments.

One of the plans adopted by Baumert to test the influence of the water and its
contained gases upon the emitted gas — whether the latter were either deprived of
any of its oxygen or carbon dioxide, or received accessions while exposed to the
water — was (1) to prepare artificial mixtures of atmospheric air and carbon diox-
ide in different proportions (4:1 and 6:1) and cause them to bubble through water
in which fishes were living, and after collection to allow these gases to remain
exposed to the water, as in the experiment with the natural gas; (2) in a similar
way, to cause atmospheric air to pass through water in which the fishes had been
living for several days. In the second of these experiments, the atmosphere
remained absolutely unaffected, showing not only that it did not lose oxygen, but
that it also did not acquire carbon dioxide. In the first of these methods, howeyer,
while the proportion of nitrogen and oxygen remained practically the same, the
carbon dioxide was diminished by absorption to the extent of about one half its
original volume. To this evidence of considerable absorption, it seems to me,
Baumert does not allow proper weight, when, by a more satisfactory method, he
subsequently finds what he believes to be the true composition of the emitted gas.

MUSEUM OF COMPARATIVE ZOOLOGY. 21

the water of the river Seine contains oxygen and nitrogen in the pro-
portions determined by Humboldt et Provengal (’09), but that, owing
to the defective means of extracting the carbon dioxide employed by
them, they secured only one fortieth of the volume of that gas which
was actually present (Grehant, ’69, p. 379).

Such being the case, it is evident that reliance cannot be placed on
Baumert’s conclusions, for his opinion — that there was no perceptible
absorption of gases under the conditions of experimentation previously
detailed — rested ultimately upon his experiment with intestinal respi-
ration in distilled water, and his ability subsequently to extract the
whole of the gas absorbed by the water during the experimentation.

Being unwilling to sacrifice the limited number of young fishes in my
possession for the purpose of securing the contents of the air-bladder,
and not having the opportunity of getting the gas from the air-bladder
of adults, I determined to employ a method which seemed likely to
furnish positive evidence of the presence of carbon dioxide, if any were
really eliminated, even though it would not give a rigid quantitative test.

The plan was, to place the fishes in a vessel of water having a limited
air-chamber, which could be rapidly swept of its contents by a con-
tinuous flow of pure air through it, and to lead the air, thus drawn
from the experimental jar, through baryta water, which would indicate
the presence of even a small amount of carbon dioxide by the formation
of a precipitate.

Accordingly, a glass jar (Figure 2, 3) of about 20 litres’ capacity, pro-
vided with a thick ground-glass cover having a central neck and orifice
for a stopper, was selected. The cover was larded te secure a close
joint, and in addition the edge was sealed with melted paraffine, and
the jar nearly filled with recently boiled distilled water. This was
siphoned into the jar with as much precaution as possible in order to
prevent exposure to the atmosphere. To make the water habitable for
the fishes, it was of course necessary to aerate it, and it was at the same
time important not to introduce in this process any carbon dioxide. To
accomplish this successfully was found to be a task much more labori-
ous than was at first anticipated. The apparatus shown in Figure 2
was that which finally proved satisfactory.

A Bunsen pump (P) attached to the hydrant provided the suction
required, and the system of tubes and chambers through which the air
was drawn was arranged as follows. Four glass combustion tubes (kK)
each about 2 meters long and 2 cm. in diameter were loosely filled with
fragments of potassic hydrate, and joined by short pieces of rubber

BULLETIN OF THE

22

‘G FANT

S
Tv ICT

FOV On) We IO OE CIE ER RTL

MUSEUM OF COMPARATIVE ZOOLOGY. Bs

tubing, as indicated in the figure. After traversing these tubes, the air
was compelled to bubble through two sets of potash bulbs (x’), and
thence by a glass tube was led through the stopper of the experimental
jar (s). The glass tube was carried to within about 10 cm. of the bottom
of the jar, and, after describing a U-shaped bend, terminated in a plati-
num “rose” (R). Thence bubbling through the water, the air was led
from the air-chamber of the jar by a glass tube to the baryta bulb (s).
To guard against the effect of an accidental diminution of pressure in
the hydrant water, and a consequent reflow of water into the baryta
bulb, the latter was not directly connected with the pump, but an ordi-
nary Wolff’s bottle (w) was interposed.

The preliminary test imposed upon the apparatus was, that it should
run twenty-four hours with no fishes in the water, but otherwise under
the same conditions that would be required when the fishes were intro-
duced, without giving a perceptible precipitate in the baryta bulb. It
required some time and considerable attention to details before this was
attained. All the connections were made by means of close-fitting rub-
ber tubing, which was made as limited as possible, narrow glass tubing
being used wherever practicable, and all the joints were in addition
sealed with melted paraffine. While the pump was in operation, the
water in the jar was of course under diminished pressure,! the diminu-
tion being equivalent to a column of water equalling the distance be-
tween the surface of the water in the jar and the lower end of the bent
glass tube to which the ‘‘rose” was attached. Evidently the removal
of the stopper from the neck of the jar (through which it was proposed
to introduce the fishes) under such circumstances would allow the en-
trance of considerable impure air, which might give a precipitate as
soon as the pumping was resumed. To obviate this source of error, the
tubing of the apparatus was clamped at x, and by means of a hand
bellows air was injected through the potash system until equilibrium
was restored in the chamber of the jar, this being indicated when the
water rose inside the bent tube to the same height as outside. In the
preliminary test, this was done in the same manner as subsequently
when the fishes were introduced, and the stopper was also removed for
the length of time which experience had shown would be necessary for
the introduction of the fishes. Thus, as nearly as possible the same
conditions were observed in the preliminary and the final experiments.

1 It is to be observed that, so far as this diminution of pressure affects the
absorption of gases by the water, it can only act favorably in preventing such
absorption.

eee a

24 BULLETIN OF THE

When the preliminary experiment had proved that the air which for
twenty-four hours was being pumped through the apparatus did not
contain enough carbon dioxide to cause any trace of a precipitate in the
baryta bulb, then the tubing was clamped at x. Equilibrium was re-
stored by forcing in air as before, four fishes were introduced through the
neck in the cover to the jar, and the pumping was resumed. It was kept
up for six hours, at the end of which time not the slightest trace of a
precipitate had been formed, although the fishes had been emitting bub-
bles at the surface of the water in the ordinary manner during the whole
time. The baryta water was subsequently tested to make sure that no
error had been made in the matter of its sensitiveness to carbon dioxide.

The immediate inference from this experiment is, that the gas elimi-
nated by the fishes contained no trace of carbon dioxide; but here, again,
the possibility of an absorption of eliminated carbon dioxide by the water
cannot be rigidly excluded.’ It seems highly improbable, however, that
with an air-chamber which required at most only a few minutes to be
swept of its contents, there could have been a sufficiently prolonged
exposure of the gas to allow the absorption of its carbon dioxide. It
might be urged that this, being but a single experiment, can be entitled to
little weight, — that, even if in this one instance there was no elimina-
tion of carbon dioxide, it is to be considered simply as an exceptional in-
dividual case. It is, however, to be borne in mind that there were four
fishes under experimentation, and that in the case of the analysis which
failed to show the presence of carbon dioxide there were szx.

As the result, then, of these various experiments to determine
whether carbon dioxide is eliminated, it must be concluded for the
present, until opportunity is presented for a careful analysis of the
gas taken from the bladder without exposure to water, that the gas
of the air-bladder of Lepidosteus is not likely to contain more than two per
cent of carbon dioxide at the most, with a strong probability that m many
cases it does not contain an appreciable amount of that gas.”

1 If this experiment were ever repeated, it is perhaps desirable that control
experiments should be carried on at the same time. In the control experiments
atmospheric air mixed with known volumes of carbon dioxide, varying from zero
to 2% of the total volume, should be made to bubble near the surface of the water.
By observing the frequency and estimating the volume of the gas emitted by the
fishes in a given time during the actual experiment, practically the same conditions
could be observed in the control experiments. The introduction of the carbon
dioxide mixture could easily be accomplished by means of a third glass tube,
piercing the rubber stopper of the jar and dipping just below the surface of the
water.

2 The recent analyses made by Jobert (’77 and ’78*) on the gases taken from

MUSEUM OF COMPARATIVE ZOOLOGY. 25

But if little or no carbon dioxide is eliminated by the air-bladder,
what is the purpose of this aerial respiration? The answer must be
evident from the amount of oxygen which invariably disappears. The
emitted gas contains only from two thirds to one half the amount of
oxygen found in an equal volume of atmospheric air. I believe it 1s
therefore definitely and satisfactorily settled by these experiments, that
the air-bladder respiration in Lepidosteus is subservient to the oxygenation
of the blood.

There arise two more questions, which I believe to be intimately
related to each other. If the air-bladder does not provide for the
elimination of a larger amount of carbon dioxide than is stated, how is
this product disposed of? And, secondly, What is the relation between
branchial and aerial respiration in Lepidosteus, and how has it been
brought about ?

The most probable method of elimination of the carbon dioxide is by
means of the gills. To determine the question definitely would require
the complete separation of the effects of branchial respiration from those
accomplished by other means; but it seems to me at present extremely
doubtful if any method can be devised which will enable one to do this
in the case of Lepidosteus. With Cobitis, or any other fish with intes-
tinal respiration in which the gas is emitted through the anus, the solu-
tion of the problem would be comparatively easy.

But there is already some evidence that the gills in such cases effect
an increased amount of elimination. At least the comparisons which
Baumert instituted between Tinca and Cyprinus on the one hand, and
Cobitis on the other, indicate clearly that there is a more rapid elimi-
nation of carbon dioxide through the gills in the latter case than in the
former.

If, then, it can ultimately be shown that in cases of aerial respiration
the gills accomplish most of the elimination of carbon dioxide, and that
the air-bladder is serviceable principally as a means of securing addi-
tional oxygen, something will have been accomplished toward appreciat-
ing the natural conditions which must have led to the imposition of this

the air-bladder and from the intestines of several Brazilian fishes gave the
following results : —

I. Fishes with intestinal respiration,
a, emitting gas from the anus, Callichthys asper C. & Val. : 1.5-3.8%, carb. diox.
b, emitting gas from the mouth and gill slits, Hypostomos : 1.5-2.8% carb. diox.
Il. Fishes with air-bladder respiration, Krythrinus tenvatus

» 129 Aer :
Spix, or Hrythrinus brasiliensis Spix, Spates Hom

26 BULLETIN OF THE

function upon the alimentary canal, or one of its dependencies. The
water in which such fishes lived may have been at times incapable of
furnishing the necessary amount of oxygen, but sufficiently serviceable
as a means of removing, through the gills, the carbon dioxide. The
solubility of carbon dioxide in water, as compared with that of oxygen,
favors such an explanation.

I have not succeeded in finding many analyses of the gases held
in solution by water which seem capable of throwing light on this
question.

A. Morren (41, pp. 471, 478-480, and 744, p. 12), who conducted
numerous experiments to ascertain the effect of light and of green or-
ganisms on the composition of the gases dissolved in water, has recorded
some interesting facts which seem to me to bear upon the problem. He
ascertained that when the per cent of oxygen in the gas extracted (by
boiling) from the water fell below 18, 19, or even 20, the fish contained
in his experimental reservoir began to languish, and many of them died.
He also deduced from his experiments these conclusions: that water
which [is stagnant or] flows slowly over a slimy bottom is subject to
conditions which serve to explain why it may be habitually less oxy-
genated than water which runs rapidly over a sandy bottom, and why
it undergoes greater variations in the composition of its dissolved
gases.

Morren also found that the oxygen in the gas contained in the waters
of the river Marne fell as low as 18 per cent on the 18th of June, 1835,
when there was a remarkable mortality among the fishes in the river,
and he ascribed this mortality to the want of oxygen in the water.

If such gas is incapable of supporting the life of fishes, it might occur
under certain circumstances that the proportion of oxygen would be
considerably below the normal 32 per cent, and still be far from pro-
ducing asphyxia. Under such conditions, which presumably happen
more often in the stagnant water of shaded swamps than elsewhere,
fishes which could avail themselves of the oxygen in the atmosphere
would be able to survive when others could not. They might still
employ their gills for the elimination of carbon dioxide into the water,
— 100 volumes of which can absorb about 120 volumes of the gas, —
but would have to depend largely upon the atmosphere for their supply
of oxygen.

One might conclude, then, that in the transfer of the respiratory
function from the gills to the prospective lungs the two components of
the respiratory process were separated from each other, and that the

\

MUSEUM OF COMPARATIVE ZOOLOGY. 27

oxygenating function was the one first to be transferred ; that, so long
as the animal lived in the water, the gills were, under all ordinary cir-
cumstances, the customary channel for the elimination of carbon dioxide ;
and that finally, during the passage from water to land, this function was
also imposed on the vesicular organ of the intestinal tract, which thus
became a lung in the fullest sense of the word.

IV. Embryology.

The interesting accounts of the ontogeny of Lepidosteus given by
A. Agassiz and by Balfour and Parker have put us in possession of
important data concerning vertebrate development. It is hoped that
a more extended study in this field will serve to answer some of the
questions left unsettled by them, and will afford additional results of
general value in discussions concerning the phylogeny of vertebrates.

There would be some advantages in beginning the subject of the
development of Lepidosteus with an account of the formation and
growth of the ova. Although I have considerable material for a study
of odgenesis, it has seemed to me better, on the whole, to defer a con-
sideration of the topic until the close of that part of these studies which
deals with organogeny, and to begin the description with the ovarian
ovum as it exists at the time of oviposition. But I have deviated from
the plan in the first part of the subject, —the egg membranes, — in
order to give a more complete exposition of the structures which
envelope the mature ovum.

1. Egg Membranes.

The only account of the structure of the egg membranes of Lepi-
dosteus is that given by Balfour and Parker (81, p. 112, and ’82,
p- 362 and foot-note).? It is as follows: “They [the ova of Lepi-

1 Ryder (’85, p. 146) has since given the following brief account: “In the ova of
Ganoids, Amia and Lepidosteus, the zona is composed, in the first instance at least,
of short, parallel, elastic fibers disposed in a plane vertical to that of the mem-
brane, these fibers being fused at their ends or just below the inner and outer
surfaces of the membrane. Sections through the egg membrane of Lepidosteus
seem to indicate the same condition of things as in Amia, in fact Dr. E. L. Mark
of Cambridge, Mass. has kindly shown me drawings which show the fibers of the
zona of the former isolated in the same condition as I have been able to separate
those forming the egg membrane of the latter.”

It will be seen by the following description that I do not agree with all that

28 BULLETIN OF THE

dosteus] have a double investment consisting (1) of an. outer covering
formed of elongated, highly refractive bodies, somewhat pyriform at
their outer ends (Plate 21, fig. 17, f. e.), which are probably metamor-
phosed follicular cells, and (2) of an inner membrane, divided into two
zones, viz.: an outer and thicker zone, which is radially striated, and
constitutes the zona radiata (z.7r.), and an inner and narrow homo-
geneous zone (z.7.').” In a foot-note the authors state, in addition, that
‘‘the ripe ova in the ovary have an investment of pyriform bodies
similar to those of the just laid ova,” but that their attempts to ascer-
tain the nature of these peculiar pyriform bodies proved futile on
account of the bad state of preservation of the material at. their
command.

A. OBSERVATIONS,

The observations which I have made on the egg membranes of Lepi-
dosteus will be followed by an historico-critical account of what is
known about the structure of egg membranes in other fishes.

I have examined the membranes in fresh eggs, as well as in those
which have been treated with various reagents, and have been able to
carry my investigation somewhat further than Balfour and Parker. It
will appear in the course of the following account to what extent my
results agree with theirs, and in what they differ.

a. Zona Radiata and Villous Layer.

Omitting for the present the modifications in the micropylar region
of the membranes, I will consider first the structure of the envelopes in
the recently laid egg which has not been subjected to the action of re-
agents, and subsequently will describe what is to be gained by the study
of sections made from eggs that have been hardened and stained. The
differences between the membranes of recently laid eggs and those of
mature ovarian eggs are so unimportant that it will not be necessary to
give a separate account of each. The nature of the differences, when
such exist, will be pointed out in the course of the description. When
the ripe eggs are artificially removed from the female by “ stripping,”
they have at first irregular, more or less polyhedral forms, due to mutual
pressure in the ovary, and the membrane investing each is in a pliable
condition. This state is retained for a long time, provided the eggs do

Ryder here states. He seems to have entirely overlooked the existence of two
membranes, and gives such an account of his “ zona” as to make me believe that
he has had in view what I have described as the villous layer.

MUSEUM OF COMPARATIVE ZOOLOGY. 29

not come in contact with water; but when immersed in water they
soon exchange the flaccid for a more rigid condition, like the eggs of
many other fishes. Whether the egg is in water or in air, its surface is
excessively sticky, as Mr. S. Garman? has already accurately observed.
The eggs adhere equally well to polished and to roughened surfaces.
When let fall directly from the female into 95 per cent alcohol, with a
view to ascertaining if there was any special superimposed layer of viscid
substance such as that described by Kupffer (’78%, p. 178) for the
herring, the eggs have furnished no evidence of the existence of any
such continuous film, nor of any covering additional to that which is
distinguishable in the mature ovarian ovum, except small amber-colored
bodies mentioned later. |

When laid the egg of Lepidosteus is enclosed in a single membranous
envelope about 50-60 w thick (Plate I. Figs. 5,11). This membrane is,
however, composed of two distinct but firmly united layers. The outer
layer, which embraces from one fourth to one third of the total thickness
of the membrane,’ I shall call the wllows layer (st. vil.) ; the inner, the
zona radiata (z.r.). The former is the outer covering of “elongated,
highly refractive bodies,” described by Balfour and Parker ; the latter
undoubtedly embraces both the zones —z.r. and z.r.!— described by
those authors. It will be convenient to consider the two layers together
rather than separately.

Examined in the fresh condition, the outer surface of the egg envelope
is of a faintly yellowish or brownish tint, which is in part due to the
presence of small ovoid bodies of variable size, of an amber color and a
waxy appearance (Plate VIII. Fig. 5), which are scattered over the surface.
I am unable to say how or where these bodies are formed, but possibly
they result from the disintegration of the granulosa. Aside from these
bodies the surface presents a roughened or shagreen-like appearance,
which is found upon microscopic study to be due to slightly rounded
prominences of nearly uniform diameter which are separated from each
other by regular nearly straight lines, so that a view perpendicularly
upon the surface (Plate I. Fig. 1) presents a field divided by these lines
into small polygonal (four- to six-sided) nearly equal areas. The aver-
age size of the areas increases slightly as one approaches the vegetative
pole of the egg. When the envelope has been removed and torn

1 See A. Agassiz, ’784, p. 66.

2 Measurements of the fresh membrane of an ovarian egg left twelve hours in
glycerine gave a total thickness of 68 4, of which 50 uw represented the zona and 18 u
the villous layer. After the addition of weak hydrochloric acid the latter increased
to twice its original thickness (36 u).

30 BULLETIN OF THE

into pieces, it is readily seen to be composed of the two layers men-
tioned, which for the most part remain firmly united. Along the torn
edges, however, it often happens that the lines of rupture in the two
layers do not coincide, so that for a considerable area there is a separa-
tion of one layer from the other (Plate I. Fig. 5). Such regions are the
most satisfactory ones for the separate study of the two structures.
Both layers are translucent; to the outer belongs the brownish tint
seen in surface views, while the inner is slightly opalescent. When seen
from the edge or in optical section, or, better still, when cut with a
razor into thin radial sections, both exhibit radial striations, which are
much closer to each other in the inner layer than in the outer (Plate I.
Figs. 4, 5, 11). .
Aside from certain exceptions which will be considered later, the fine ra-
diate markings of the inner layer, or zona, appear as nearly straight par-
allel lines, which are traceable through the whole thickness of the layer,
but which become gradually less prominent toward the deep surface.
The markings of the outer or villous layer, on the contrary, are less
uniform ; they traverse the whole thickness of the outer layer, but are
most clearly defined near the periphery, their deeper portions being
more irregular and confused, and often exhibiting a tendency to a zig-
zag course. They indicate the boundaries of highly refractive prismatic
bodies, of which the layer is composed, and which seen endwise produce
the appearance of polygonal areas already alluded to. When viewed
from its deep surface (compare Plate I. Figs. 2, 7), the villous layer
has a somewhat ragged appearance ; it also exhibits polygonal areas, but
they are less regular and less clearly defined than those seen from the
external surface. When the egg membrane has been left for some time
in water, or, better still, in a mixture of water and glycerine to which a
trace of hydrochloric acid has been added, the prismatic elements which
compose the villous layer undergo a remarkable change, during which
the cause of the peculiar zigzag appearance of their boundaries is made
evident. After a time the free rounded ends of some of the prisms
appear to protrude above the neighboring ones (Plate I. Fig. 4), thus
giving the surface a less even contour than it had at first. On com-
paring the conditions before and after the application of acid, it at once
becomes apparent that the layer has increased in thickness, At its free
edges the prisms become more or less detached from each other, and it
then is possible to appreciate their real form.
There are recognizable at least three distinct regions in each prismatic
villus, and each may be roughly compared to a stalk of grain, with its

MUSEUM OF COMPARATIVE ZOOLOGY. aL

head, shaft, and root. These names may be applied not inappropriately
to the three regions of a villus. The peripheral portion or head (cap.),
embracing one fourth, or sometimes as much as one third, of the original
thickness of the layer, is distinctly prismatic and highly refractive ; its
sides are parallel, and it is little affected by the acid, so that, although it
increases very slightly in size, it still retains to some extent its angular
form. Its free end is always more or less rounded (Plate I. Figs. 4, 9, 2).
Following this terminal head, and marked off from it by a slight con-
striction, comes the stalk (Fig. 9, 2, pd.), a long, also prismatic fibre,
which is less highly refractive than the head, and is so crowded as to be
folded back and forth, thus giving to it the appearance of a spiral spring.
In fact, many of the fibres are coiled into a tolerably regular spiral, but
the majority are simply folded irregularly, evidently being accommo-
dated to the space most available.

Through the action of acid these stalks begin to swell, and some of
them — since they are affected more promptly than others — cause an
earlier protrusion of the correspondiyg prismatic heads (Plate I. Fig. 4).
The increase in the thickness of the layer — which soon reaches twice
its original dimensions — is due almost entirely to the swelling of these
deeper portions of the villi. When isolated, they may in some cases be
elongated to ten or twelve times the length of the coils which they at
first formed (a, 6, Fig. 9, Plate I.). A portion of this elongation is due
simply to the unfolding of the compressed stalks ; but ultimately in
proportion as it elongates the stalk becomes more attenuated. It often
happens in this process that different portions of the stalk are at first
unequally affected. Usually it is the deeper portion which is first to
uncoil and to become attenuated (n, Fig. 9). When fully extended the
stalk is slightly tapering, being narrowest at a little distance from its
basal or root end, and although generally quite uniform in calibre, it
occasionally exhibits varicosities. In many cases the isolated villi (Fig.
9, g, h) appear as though temporarily prevented from straightening out
because of delicate longitudinal structures of a band-like appearance.
The aspect of the stalk is then remarkably similar to the pouched con-
dition of the mammalian colon, to the longitudinal muscles of which
these band-like structures correspond. The apparent ‘bagging ” is
usually all in one direction, namely, toward the attached end of the
villus (hk, Fig. 9). The basal end or root (rz.) appears to terminate
regularly in a number (3-9) of tapering root-like diverging prolonga-
tions (Fig. 9, f, g, 7, &), which are often apparently connected with each
other by membranous expansions of the basal portion of the villus.

32 BULLETIN OF THE

These roots serve to fasten the villous layer very firmly to the zona
radiata, in a manner to be explained in connection with the account of
that layer.

The much finer radial markings of the zona radiata (z.r.) are entirely
different in character from those of the villous layer; seen from either
surface with a moderately high power they appear as punctations, dark
or light according to the focusing (Figs. 3, 8, Plate I.), evenly scattered
over the surface, and yet so arranged as to give the whole area a very
characteristic appearance. Although rather evenly distributed, they are
arranged in groups or systems. One may trace over a considerable area
a series of dots placed at the intersections of a system of imaginary equi-
distant lines crossing each other at right angles ; near by may be other
series, in which the systems of imaginary lines cross at angles varying
widely from that of 90°; in still other series, the lines are arcs of circles ;
the circles may vary somewhat in size, but the arcs are never to be
traced for more than a few degrees. These different systems abut upon
each other in the most fortuitous manner, and the intervening spaces are
filled with dots so evenly arranged as not to interfere with a fairly uni-
form distribution over the whole surface (compare Plate III. Fig. 5).
Higher powers show that the punctations are circular in outline, of very
nearly equal diameters (0.5 w at their outer ends), and placed at inter-
vals averaging about 1.5 ». My notes of May 24 and 25, 1883, make
the intervals between the pore-canals, as determined by measurements
on the shell of an ovarian egg that had lain in glycerine over night,
only 2 p, less than half the value given above ; but I believe that the
larger distance fairly represents the average condition.

Thin tangential sections show by focusing that these markings are
due to minute canals (pore-canals), which are ordinarily hollow, or at
least contain a substance that is less refractive than the common homo-
geneous mass of the matrix which they traverse. I have not seen any
evidence of a differentiation in the optical properties of the walls of these
pore canals which would allow one to speak of them as tubules.

Weak hydrochloric acid causes the zona to swell slightly, and ulti-
mately renders the pore-canals less conspicuous or entirely invisible.
There are certain of them which do not fade away, however, even after
treatment with acid, and which at length become the only visible struc-
tures in what otherwise appears as a homogeneous layer. (Fig. 10.
Compare also Figs. 4, 6.) These canals very generally have a spiral
course, and are noticeably broader at the outer surface of the zona than
elsewhere ; they taper gradually toward the inner surface of the layer,

MUSEUM OF COMPARATIVE ZOOLOGY. ay

but seldom reach it; most of them are traceable only a short distance
from the outer surface. They owe their prominence to the fact that
they are filled with a highly refractive substance having the form of a
corkscrew. When the villous layer has been torn from the zona, this
substance appears to terminate exteriorly in a ragged, broken end, which
in some instances is drawn out into a tapering appendage (Fig. 10, «).
There can be no doubt —according to evidence to be gained from
ovarian eggs — that this substance is continuous with that of the pris-
matic columns of the villous layer, of which they are in reality the roots.
The roughened appearance of the inner surface of the separated villous
layer is largely due both to the lacerated ends of these roots and to the
fact that many of them are wholly withdrawn from the pore-canals when
the two layers are torn asunder. This relationship of the layers also
explains why it is so difficult to separate them over even a limited area.
Sections of stained eggs, both radial and tangential, give instructive
views of the egg membranes. In radial sections the difference between
the villous layer and the zona becomes at once apparent from the deeper
stain which the former takes on. The pore-canals are also usually more
distinct than in the fresh egg, although the effect of certain acid pre-
servative reagents (Perenyi’s fluid, picrosulphuric mixture) is such as to
obscure the radial markings of the zona. In the villous layer a still
more striking contrast is produced between the heads and the stalks of
the villi, since the former almost invariably take a much deeper stain |
than the latter. Especially is this true when stained in picrocarmine,
by which the heads are colored a deep carmine while the stalks and roots
remain unstained or take a greenish-yellow hue from the action of the
picric acid (compare Plate IV. Fig. 1). In borax carmine both portions
are usually stained, and almost invariably the head much deeper than the
rest of the villus ; but it has occasionally happened that the heads were
less deeply colored, and presented a slightly yellowish tint (Plate IX.
Fig, 2). I am unable to account for the difference, unless possibly a
prolonged decoloration in hydrochloric acid is the cause of the feeble
stain of the head ends. In all these stained radial sections it is to be
seen that the transition from the head to the stalk, although not marked
by a sharply defined line, is nevertheless abrupt. Owing to this, and
the fact that the stalks, as well as the heads, are of nearly uniform
lengths, radial sections of well stained specimens always exhibit the
stalks in the form of a broad band or zone, sharply marked upon both
edges, — more deeply stained than the zona radiata, but less deeply

than the narrower well defined band which is made up of the heads of
VOL. xIx.— No. 1. 3

— +s ee

34 _ BULLETIN OF THE

the villi. Each of the heads has its external free surface more or less
rounded and not quite smooth, its sides nearly parallel and straight,
and its ill-defined deep face also tolerably straight. Along the last it
is distinguishable from the stalk, with which it is continuous, by its
greater refractive power as well as deeper color, and by a slight dimi-
nution in the size of the stalk. The last distinction becomes more
conspicuous the more the stalk is elongated. The differences between
head and stalk are emphasized by the fact that the villi have a greater
tendency to rupture along this line of union than elsewhere (Plate II.
Fig. 1, and Plate III. Fig.1). The outlines of the free end and the sides
of the head are sharp, and in thin sections, especially such as cut the
heads crosswise, the margins seem to be limited by a narrow double-
bordered dark band (compare Plate III. Fig. 3), as though the head were
invested in a thick deeply staining membrane. Since I have never been
able to find evidence of the separation of any membranous structure
from the surface of the head, I am disposed to believe that the appear-
ance simply results from a differentiation of the cortical portion of the
head, which otherwise appears perfectly homogeneous. In some cases
this cap-like cortical part seems to exert a restraining influence on the
swelling of the central portion ; at least I interpret in that sense certain
conditions of the heads frequently met with. In such cases their sides
are not strictly parallel, especially when the villi stand in an isolated
position (Plate II. Fig. 1). The head, instead of being marked off from
the stalk by a constriction or shoulder of the ordinary form, has its
outline gradually broadened or flaring as it approaches the peripheral
end of the stalk, and its cap-like sheath appears to end abruptly with
edges which are slightly everted ; the connecting portion of the stalk is
as broad as, or even broader than, the basal end of the head, so that the
direction of the resulting shoulder is just the reverse of that commonly
seen. The most natural explanation of this appearance which occurs to
me is, that the free edge of the cap-like sheath is distended, and even
sometimes everted, by the swelling which takes place in the region
where head and stalk are continuous, and that the sheath in all probabil-
ity acts as a restraining investment in preventing any great distention
in the rest of the head.

The stalks, in radial sections of eggs which have been subjected for
some time to the action of water before hardening, have the appearance
of comparatively slender columns, which are often slightly sinuous, but
in general nearly parallel. They taper at first quite rapidly for a short
distance from the head, and then only very gradually toward the basal

MUSEUM OF COMPARATIVE ZOOLOGY. 35

or root end. The spaces between the stalks are much greater than those
between the heads; while the latter sometimes remain — even after the
prolonged action of water—éin a continuous layer, the stalks often
appear to stand individually isolated. It is more common, however, to
find, as the result of the swelling, that both the heads and the stalks are
arranged in groups or patches, — better shown in tangential sections.
Even when the heads are not thus separated, the stalks may be gathered
into clusters which leave in radial sections broad lenticular spaces be-
tween them (Plate III. Fig. 1).

\ The stalk gradually diminishes in size to near its zonal end, where it
enlarges rather promptly into a sort of conical foot, which exhibits dark
longitudinal or radiating markings continuous with the dark outer ends
of corresponding pore-canals in the zona. In some cases the foot is
split into two or three strands, between which there is then left a space
that in radial sections is triangular, with its base resting on the zona
and its more acute angle rising into the stalk. The roots proper embrace
only the portions of the villi still occupying the pore-canals of the zona.
In some cases they are to be recognized as occupying every pore-canal,
in others some of the canals appear to be destitute of villous contents.
The roots are highly refractive, like the stalk, and seem to stain even
more deeply than the latter. They are always broadest at the outer
end, and taper until they are exceedingly fine threads. They seldom
reach more than a tenth or an eighth of the way through the zona,
although longer and larger roots are met with at intervals. They
always appear more tortuous — zigzag, or spiral — than the pore-canals
which do not contain roots, and are at times so irregular in form as to
have caused great distortions in the canals (Plate IX. Fig. 2). Their
finest. tips, however, always appear continuous with the much more
famtly marked pore-canals. I cannot doubt, therefore, that they are
accommodated by simple enlargements of the pore-canals. The great
regularity in their distribution, too, allows no other interpretation than
that the position of the roots is practically determined by that of the
pore-canals.

Tangential sections of stained eggs (Plate III. Figs. 2-5) afford the most
satisfactory evidence of the shape and grouping of both heads and stalks,
and is the only safe means of controlling the views of the foot region
gained by radial sections. The heads are at first close set, leaving only
the finest narrow lines, with here and there an irregular opening where
the prisms incompletely match (Plate ITI. Fig. 2) ; their cross sections are
angular and range from variously proportioned triangles to six- or seven-

———_-— ge

36 BULLETIN OF THE

sided polygons. After the prolonged action of water they become less

angular, and begin to separate along irregular lines, so as to leave the
heads arranged —as already indicated —in patches, which vary con-
siderably in size but are for the most part of a characteristic polygonal
outline, with borders which are necessarily jagged owing to the nature
of the lines of separation, for the latter never split a prism, but simply
separate adjacent ones. The heads may vary in diameter in the ratio
of one to two.

The dark border already alluded to is best seen in thin tangential
sections (Plate III. Fig. 3), and is readily distinguishable on all the
heads when well stained and cut sufficiently thin. The line of separa-
tion between prisms is not always distinguishable, but whether this is
due to actual contact or not it is difficult to say, since the least obliquity
in the section is sufficient to obscure so faint a marking.

The stalks also are found upon cross section to be prismatic, even
after the process of swelling has completely isolated them (Plate III.
Fig. 4). They are also arranged in groups which correspond fairly to
those of the heads, but the spaces between them are much greater.
Occasionally sections of stalks are to be seen, even from the middle of
the stalk-zone, the central part of which has not been stained (Plate
IIL. Fig. 4, a, «). Careful examination shows that such stalks are really
hollow, the boundary of the colorless area being sharply defined. I have
never seen vacuoles in the middle region of any of the stalks examined
in radial section ; besides, these cavities can often be traced continuously
on successive tangential sections toward the foot. They are, moreover,
increasingly frequent as one approaches the zonal attachment of the
stalks. The consideration of all these facts makes me quite sure that
many of the stalks, at least in their basal halves, are really hollow
prisms, although I have never been perfectly certain that I have seen
this condition in radial sections. One may, however, as before stated,
readily see on radial sections that the expanded foot of the stalk is often
apparently split into diverging roots, and that there is an intervening
unstained region. The prolongation of this space, which is triangular
in side view, forms, I believe, the cavity of the stalks in question. A1-
though the prisms appear sharply marked in cross sections, there is very
generally a trace of a filmy substance projecting here and there from
their edges in the form of faintly marked threads, which sometimes end in-
definitely in the inter-prismatic spaces, but at other times appear loosely
to connect neighboring stalks (Plate III. Fig. 4, 8). This substance
seems to stain less deeply than the stalks themselves, but it is exceed-

OO

MUSEUM OF COMPARATIVE ZOOLOGY. ae

ingly difficult to decide whether the faintness of color is due to a specific
difference of substance, or is simply the result of the tenuity of the film
itself. In the former case, one would perhaps be justified in concluding
that there was an inter-prismatic substance which served the purpose of
a cement to hold the stalks together. The peculiar longitudinal band-
like structures noticed during the elongation of the stalks (Plate I. Fig. 9,
g, h, 1) are possibly to be referred to the same substance.’ But on the
second assumption these shreds of faintly stained substance could be
hardly more than the lacerated edges of the stalks themselves. I con-
sider the latter the more probable explanation.

Owing to the spherical form of the egg, tangential sections are circular
in outline, and in a given section the centre represents the deepest part.
When the centre of such a section is occupied by the superficial part
of the zona radiata, the periphery is formed by a circular band of the
villous layer, the deeper portions of which are nearer the centre of the
section.

A segment from that portion of the band which cuts through the
bases of the villi, their roots, and the superficial portion of the zona, is
shown in Plate III. Fig. 5. Proceeding from the outer (in the figure
upper) toward the central portion of the section, one observes that the
cross sections of the villi increase somewhat in size, that the stalks
which embrace cavities become more numerous, and that the outlines
of the stalks become more and more star-shaped, and then irregular,
and that finally they break up into detached spots, which a little farther
along become smaller and smaller until they cannot be distinguished in
size from the pore-canals.

Since the sections of the membrane are successively increasing in
diameter, the deep face of each will pass through a broader portion of
the zona than the upper face will. If for the purpose of examination
the section be inverted, so that the deep face is uppermost, the relation
of parts can be much more easily and satisfactorily studied than if it be
viewed from the upper face only, because the zona offers less impediment
to vision than the thick-set columns of the villous layer. Attentive
focusing shows conclusively on such preparations that the rays or
branching roots of the prismatic columns lead each to a pore-canal, and
it becomes possible in many cases to note the exact number of pore-

1 Dilute hydrochloric acid causes the distance between the heads of the villi to
increase. This would be readily explainable as the result of the swelling of an
inter-villous substance, could the existence of such a substance be satisfactorily
established.

38 BULLETIN OF THE

canals in which a given stalk takes root. The substance of the roots

and of their rib-like extensions up the stalk appears to be more deeply

stained than that of the expanded foot of the stalk ; but this is perhaps
only an appearance due to the fact that they are considerably thicker
than the membranous portion which connects them.

As the successive sections pass through deeper and deeper portions of
the zona radiata, the calibre of the pore-canals grows very gradually
finer, and those which are plugged with deeply stained villous roots
become less numerous, but otherwise there is no essential difference
in the appearance of the sections. The characteristic arrangement of
the pore-canals previously described is visible here, and may be made
out more easily than on the fresh egg-shell, provided the sections are
made perpendicular to the canals and are sufficiently thin.

In radial sections from eggs that have been hardened and stained, the
zona is usually of a uniform faint tint (Plate IJI. Fig. 1), but often
there is a very gradual deepening in the intensity of the color in passing
from the outer to the inner boundary of the layer (Plate II. Fig. 1). In
a few instances this deeper stain seems to extend toward the outer sur-
face of the zona in flame-like jets (Plate IV. Fig. 1). The outer boun-
dary of the zona, although appearing slightly irregular, owing to the
variable lengths of the root-like prolongations of the villous layer, is in
reality fairly even and sharply marked (Plate II. Figs. 1, 7, 8, Plate
III. Fig. 1, and Plate IV. Fig. 1). The inner boundary is still more °
precisely defined, and appears as a fine continuous line, which sharply
separates the zona from the peripheral layer of the yolk. Nowhere is
there any evidence of a gradual transition from the yolk to the mem-
brane. Occasionally, when the section is not exactly perpendicular to
the inner surface of the zona, this boundary appears double, but careful
focusing in such cases always shows this to be an optical illusion. In
a few instances I have seen a similar appearance which was not thus
explainable. For a considerable distance a layer of nearly uniform
thickness appeared to intervene between the zona radiata and the yolk
(Plate II. Fig. 7). But the line which separated this from the rest of
the zona was never to be made out for more than a small portion of the
circumference of a section, for it either terminated abruptly, or, gradu.
ally approaching the inner boundary of the zona, became confluent with
it. Its inconstancy and its want of continuity are together sufficient to
show that the layer in question is not entitled to be considered a dis-
tinct membrane, nor even a differentiated portion of the zona radiata.
I may add, that I have never seen a section of this kind in which it was

‘

MUSEUM OF COMPARATIVE ZOOLOGY. 39

not possible to discover in some part of the layer evidences of pore-
canals continuous with those of the remaining portion of the zona
radiata. I am therefore convinced that the zona radiata is a single
homogeneous layer which is in direct contact with the surface of the yolk,
and is traversed by pore-canals which reach from the yolk to its outer
surface.

When radial sections of the zona are broken, they occasionally show a
tendency to rupture in lines concentric with the surface of the egg, but
this is so rarely the case as hardly to be characteristic. The fracture is
usually irregular, and not dependent on any structural feature ; even
the pore-canals do not appear to have much influence on the direction
of the line of separation. The nature of these canals can be more readily
studied on sections of hardened specimens than on the fresh shell. Their
proximity to each other is not so readily determined from radial sections
as by means of the tangential sections already described. The same
general features which were mentioned in describing their appearance
on the fresh egg are usually visible with even greater clearness on those
which have been hardened. The distinctness of the pore-canals varies,
however, considerably in different specimens, depending undoubtedly
upon the refractive power of the mounting medium, which penetrates
the canals, as compared with that of the matrix of the zona itself.
Upon the most favorable preparations the canals can be easily traced
from end to end, so straight is their general course. At the periphery
of the zona they are uniformly somewhat broader than at its deep sur-
face; but they taper so gradually as to make the difference in calibre,
even at their two ends, trifling. In the case of almost every canal a
slightly spiral course is noticeable near the outer end, whether it be
plugged with the root of a villus or not; and throughout the whole
length there is usually the faintest trace of a wavy or zigzag course.
Aside from this, however, the canals are remarkably straight and paral-
lel. There are no enlargements or irregularities in the calibre, save
those which appear to result from the distention of the canal with the
substance of the villous roots already described.

There still remain to be considered some peculiarities of the villous
layer, which either result from particular methods of treatment, or have
not been observed sufficiently often to allow one to consider them
characteristic features.

Of those dyes which I have used, acetic acid carmine gives the sharp-
est differential staining for the heads of the villi. While the stalks
and roots remain comparatively pale, the heads (Plate II. Fig. 2) take a

40 BULLETIN OF THE

deep rose tint, and the transition from the substance of the head to
that of the stalk is rather abrupt. It happened that many of the villi
from the shell of a mature egg, that was let fall into ninety per cent
alcohol without contact with water, and was afterward stained for twelve
hours in acetic acid carmine, exhibited a very peculiar appearance at
the free surface of their heads. At or very near the middle of this
surface the dark border, so characteristic of the heads of the villi, seems
to be interrupted, and there projects from the free end of the head a
short conical or longer finger-like process. This issues from the head,
apparently through a circumscribed opening in the cortical layer, and
may assume a variety of forms, several of which are shown in Plate II.
Fig. 2, a-m. This peculiarity is interesting, as showing that there is
a region of least resistance in the cortical layer near the apex of each
head, which allows the protrusion of a part of the substance of the head
when it is subjected to the swelling influence of the acetic acid; but
whether this fact is capable of throwing any light on the source of this
villous layer, or the method of its formation, I greatly doubt.

There is often to be seen in radial sections of the villous layer a
strong tendency for the villi to fuse (Plate II. Fig. 2, 7). This is
especially true of the region of the stalks, although it is also to be
observed among the heads. Since this tendency seems to be much
greater in some cases than in others, I am induced to believe that it is
due to the influence of the reagent with which the egg was hardened,
and sometimes perhaps is dependent on the length of time the egg has
been in the water before hardening.

In a few cases — especially in certain nearly mature ovarian eggs
which were hardened in chromic acid —I have seen peculiar markings
in the villi, which at first led me to think they might be traversed by
spaces analogous to the pore-canals of the zona. They were first noticed
on tangential sections, and appeared there like minute circular holes in
the segments of the prismatic villi (Plate II. Figs. 4, 5). Focusing
showed that their contents were much less refractive than the substance
of the villi, and they were consequently very sharply defined. But the
notion that they were optical sections of tubes, like pore-canals, was at
once corrected upon finding isolated villi, which had fallen out of the
layer and were seen sidewise. When the thickness of the section is
about equal to the diameter of the villi, it is difficult, if not impossible,
to decide whether the isolated angular blocks are seen endwise or side-
wise; but by selecting the thicker sections, where the length of the
villous segments is greater than their diameter, the difficulty is avoided,

‘

MUSEUM OF COMPARATIVE ZOOLOGY. 41

and it at once becomes evident that the spaces supposed to be canals
are for the most part minute spheroidal cavities or vacuoles. Usually
there is only a single vacuole in a villus, although occasionally two are
to be seen in the same cross section (Plate II. Fig. 4). Not all of the
villi contain these cavities. Taking into the account their abundance. on
successive sections from the villous layer, I should estimate that not
more than one half or three quarters of them present this feature. The
proportion to be seen upon a single section is, of course, much less than
this. They are most abundant in the stalks, but occasionally one is
also seen in the head. Upon the egg, where they were found most
abundantly, they were rather more numerous in the micropylar region
than at the opposite pole. In the latter region there were, however,
sometimes as many as three or four in one villus, although the size (0.5
to 1 ») was the same as at the micropylar pole. I have in a few cases
observed that the vacuoles were elongated, and then they were always
of uniform calibre and were curved. Occasionally (Plate II. Fig. 4) such
a tubular vacuole appears to communicate at one end with the inter-
villous spaces. Concerning the nature of the contents of these vacuoles,
I can only say that they do not stain, and do not appear differently
from what one would expect if they were cavities simply filled with the
mounting medium. :

The differences between the membranes in mature ovarian eggs and
those recently deposited are principally the result of the swelling of the
layers by the water, and do not require any further explanation.

The foregoing account of the zona radiata in Lepidosteus contains
descriptions of two features which appear to me to bear directly on the
condition of the zona radiata of fishes in general.

First. The proof that the striate appearance of the zona is due to
pore-canals, although very generally assented to by the most competent
observers, especially in recent years, has nevertheless hitherto rested
upon comparatively slight evidence. That this evidence has been
meagre depends upon the excessive minuteness of the structures in
question. It is not to be overlooked, in the first place, that the tubu-
lar nature of the pore-canals in the case of the perch, as originally de-
scribed by Johannes Miiller (’54) for what he called the “ Eikapsel,”
has not the slightest bearing upon the nature of the pore-canals of the
zona radiata, since the egg capsule of Miiller is a structure entirely
different from the zona. I cannot, however, avoid the conviction that
his opinion as to the tubulated condition of that capsule has had con-

42 BULLETIN OF THE

siderable influence in effecting the general acceptance of similar conclu-

sions as to the nature of the radiate ‘markings of the zona. Neither the

evidence produced by Miller, —the possibility of pressing yolk globules
through the “ pore-canals ” of the capsule, — nor the vacuolated condi-
tion described by Ransom (’68, p. 455), can have any direct bearing on
this question.

Leuckart (755, p. 258) appears to have been the first to assert with
the utmost positiveness that the radial striations of the zona were due
to pore-canals; and although he nowhere states the exact nature of
the evidence which convinced him, we are doubtless at liberty to infer
that it was, in part at least, the kind of evidence which he elsewhere
(755, p. 106, foot-note) makes use of; namely, the now well understood
differences in optical effects produced by elevations and by depressions
of surfaces. Kolliker (58, p. 83) soon furnished additional evidence,
derived partly from the study of thin sections of the zona in the trout,
but more especially, as it appears to me, from the fact that maceration
in fresh water causes the middle region of these supposed pore-canals to
be converted into vacuoles. Aside from the arrangement of the dot-like
appearances as seen from the surface of the zona, which has been very
generally recognized, and the features emphasized by Leuckart and
Kolliker, I am not aware that any additional evidence in proof of the
nature of the pore-canals has yet been produced. If, then, the facts
warrant the description I have given of the zona in Lepidosteus, the
evidence that it is a canaliculation which produces the radial striate
markings in the zona radiata of fishes’ eggs has received an additional
confirmation.

Secondly. Although Miiller (as well as more recent observers) has
shown that the pore-canals in the outer envelope, or capsule, in the
case of the perch may have a spiral course, no one has hitherto observed
a similar feature in the case of the canals of the true zona radiata.
The natural injection of these canals in Lepidosteus with a substance
continuous with that which constitutes the villous layer, renders it
comparatively easy to establish the spiral course of the canals in that
fish ; and this makes probable the inference, that certain irregularities
in the direction of these canals, shown by other observers to exist in
the case of other fishes, may in reality be referable to the same
spiral condition, which, from the minuteness of the canals, has not
been recognized.

MUSEUM OF COMPARATIVE ZOOLOGY. 43

b. Micropyle.

The micropyle was apparently overlooked by Balfour and Parker,
since it is not mentioned by them; nor has it been mentioned, I believe,
by any one else, although it occupies a region which is so conspicuously
marked that, having once seen it, one could readily find it with the aid
of a simple lens. Except in eggs that have lain for some time in water,
the region of the micropyle appears, when seen under a hand lens, like
a minute hole in the shell; in surface views with a higher power it looks
like a deep circular pit (Plate IV. Figs. 3, 4) sunk in the egg membrane.
Its diameter is five or six hundredths of a millimeter. Its outline is
nearly always circular, and it has a clearly cut edge. In a few cases a
cross section of the pit has proved to be oval instead of circular, occasion-
ally with one diameter of the oval more than twice as long as the other
(Plate VII. Fig. 4). A similar appearance, though not so marked, is often
produced, even when the pit is really circular, if the plane of the section
is oblique to its axis. Sometimes the pit is partly filled by a whitish, ap-
parently spheroidal body (Plate IV. Fig. 4). When the egg is so viewed
that this depression lies in the equator, the profile of the egg in its vicinity
may be slightly modified, and show a low conical elevation, at the apex of
which the pit is located. This is not commonly the case, however, for
usually there is nothing in the profile to denote the position of the pit.
In eggs nearly mature, and in those which have been recently laid, its
place can be easily found by its relation to the lighter colored animal
pole of the egg. It is invariably located over some part of the germinal
area, and usually precisely over its centre (Plate IV. Fig. 3).

The real nature of this pit and its relations to the two layers of the
eg membrane and to the yolk can be studied on optical, but still better
on actual sections. For a general survey radial sections are most in-
structive, but for the elucidation of some questions sections tangential
to the egg at the animal pole are more valuable.

In strictly radial sections through the region of the micropyle, it is
to be seen that the surface of the egg is deeply depressed. The form of
the depression varies somewhat in different eggs, from that of a funnel,
i.e. with sloping walls (Plate IV. Figs. 1, 5), to that in which the walls
are for some distance almost parallel (Plate V. Fig. 2). This depression
results from an infolding of both layers of the egg membrane ; it forms,
however, only an approach to the true micropyle, or micropylar canal,
the latter being a minute passage through both layers which begins at
the bottom of the depression.

SS

44. BULLETIN OF THE

The funnel, as I shall call that portion of the egg membrane which
forms the walls of the depression, involves a modification of both the
zona radiata and the villous layer. Both are affected in two ways, in
thickness and in direction.

The villous layer begins to grow thinner at some distance from the
edge of the funnel. Sometimes it retains its normal thickness to within
a distance equal to the diameter of the funnel ; at other times it begins
to grow thinner at three or four times that distance from the pit. Its
diminution in thickness is quite gradual and very nearly uniform until
it reaches a minimum at the micropylar canal. The stalks of the villi
are shortened more than the heads, in comparison with their appearance
on other parts of the capsule, and the boundaries between them grad-
ually become less distinct. The diameter of the villi also decreases
considerably. Near the bottom of the funnel they become very short,
but frequently it is evident that, instead of constantly diminishing in
diameter, they may even increase as compared with other regions of
the funnel (Plate VI. Figs. 5-8).

In all parts of the funnel the villi retain a direction perpendicular to
the outer surface of the zona. In the lateral wall, and especially near
the bottom of it, they are slightly wedge-shaped or conical, the head
ends being narrower than the root ends. They thus accommodate them-
selves to the diminished space at their disposal (Plate VI. Figs. 6-8).

The zona radiata (Plate IV. Fig. 1) likewise begins to diminish in
thickness at some distance from the micropylar canal, and continues to
do so until it reaches the canal ; but it does not, like the villous layer,
grow thinner at a uniform rate. Its thickness decreases very slowly to
within a short distance of the region where the membranes begin to bend
inward to form the funnel, and then it suddenly narrows to one third its
normal dimension, after which it again decreases more slowly until it
reaches the micropylar canal. The pore-canals are not perceptibly finer
nor more closely set in the vicinity of the micropyle than elsewhere. They
retain in most regions a rectilinear course perpendicular to the surface
of the zona, but at the region of most rapid reduction in the thickness
of the latter, and for a little distance on either side of it, their course
is curved, the concave side facing the micropylar canal.

The change in the direction of the two layers of the egg-shell results
in the formation of an external depression, which is considerably deeper
than the total thickness of the shell, so that, even with a great diminu-
tion in the thickness of the latter, its inner surface projects into the
yolk as a conical elevation, which is nearly as high as the thickness of

‘

MUSEUM OF COMPARATIVE ZOOLOGY. 45

the shcll. In this deflection of the membranes, the zona radiata seems
to bend more abruptly than the villous layer; this, however, is due
principally to the fact that the region of greatest curvature is also the
region of most rapid change in the thickness of the zona. From this it
results that the inner contour of the zona is much more abruptly curved
than the outer, in some cases appearing almost angular. As a further
consequence of this, the conical elevation appears to arise abruptly from
the inner surface of the membrane ; its apex is rounded, and in the
ovarian egg its surface is everywhere in contact with the yolk. An
inquiry as to whether this infolding is the result of a process of absorp-
tion, or is due to a peculiar local modification of the activities which
produce the membrane, will best be deferred until I have given a de-
scription of the layer of cells which immediately invests the ovarian
ovum.

The micropyle proper, or the micropylar canal (Plate I. Fig. 11%,
Plate IV. Fig. 1, Plate VI. Figs. 3, 4), is straight and of uniform calibre.
It begins at the centre of the bottom of the funnel, and passes through
both villous and zonal layers of the egg membrane ; it is about 8 » long.
Its cross section is circular and about 2 4 indiameter. There is no flare
to the canal, either at the external or internal end, so far as I have been
able to observe. I am unable to say whether the diameter which I have
given is that which the micropyle possesses at the moment the egg is
laid. From measurements of spermatozoa allowed to dry upon the slide
(Plate VII. Fig. 3), one would imagine that the calibre of the micro-
pylar canal must be at least 3, that being the diameter of the heads of
spermatozoa thus treated ; but according to measurements made upon
living spermatic cells the heads are only about 1.8 uw in diameter, so
that I think 2 is probably the normal average calibre of the canal.
Still, [ have sections in which its diameter is 2.5 uw, and in the case of
some fresh membranes it was only 1.5 ~ in diameter. The narrowness
in the latter case I attribute to the swelling of the zona when exposed
to water and glycerine, in which the membranes were examined.

ce. Granulosa.

Nearly mature ovarian eggs are closely enveloped by an uninterrupted
cell layer, which is everywhere in contact with the outer surface of the
villi, Over the greater part of the egg this layer — the follicular epithe-
lium or granulosa — is composed of thin, flat polygonal cells, arranged in
a sheet only one cell thick. In surface views the granulosa cells (Plate V.
Fig. 4) appear of fairly uniform size, — 15-20 uw in diameter, — are

46 BULLETIN OF THE

slightly granular, stain feebly, and exhibit each a single large (5-10 ,)
nucleus, with an even outline and a circular or oval form. When seen
in profile, —as in radial sections of the egg with its membrane and
granulosa (Plate V. Fig. 3),-— a majority of the cells are observed to be
very thin, and their nuclei flattened; but there is occasionally a cell
whose nucleus is not so much flattened, and which therefore protrudes
beyond the general surface of the granulosa. Radial sections of the
ovum with its granulosa are further instructive in showing the rela-
tions of the cells to the heads of the villi. ach granulosa cell cor-
responds in size to from four to eight villi, but there ts no constancy in the
position of the cells or their nuclei in reference to the underlying villi.
Nothing intervenes, however, between the cells and the villi except
occasional artificial spaces. Externally the granulosa is limited by a
thin, homogeneous delicate membrane, the membrana propria (th. fol.)
of the theca follicult.

This is the condition which obtains over all parts of the egg except
in the vicinity of the micropylar funnel. Elsewhere the granulosa re-
tains great uniformity of thickness. At a considerable distance from
the micropyle its cells begin to elongate so that the granulosa grows
thicker ; as the cells approach more and more the condition of columnar
epithelium they become inclined, their onter ends being directed to-
ward the axis of the micopyle (Plate VII. Fig.1). They still continue
to form a layer only a single cell deep until they reach the vicinity of
the rapid declivity in the wall of the funnel. Here the cells, having
now attained an elongated columnar form, become superposed, and jill
completely the micropylar funnel. With a single exception the cells com-
posing this mass are fairly similar to each other. They are considerably
elongated, irregularly columnar or spindle-shaped, and contain each a
single oval nucleus about 10 by 8 or 9 in diameter. The cells
themselves vary from 15 to 40 in length, and are about 10m in
diameter. When the hardened egg, with its membranes, is removed
from the follicle, it often happens that this conical plug of granulosa
cells is left with the rest of the granulosa in the follicle. But even
when the majority of the granulosa cells of the plug are thus removed
from the funnel, there is usually left behind a single one which is unlike
the others. It occupies the bottom of the funnel, which it completely
fills, and is much larger than any other of the granulosa cells (m py. el.
Plate IV. Figs. 1, 4, 5, Plate V. Fig. 2, Plate VII. Fig. 2).

When I first became aware of the existence of such a cell it was from
the study of radial sections of a recently deposited egg in which a

MUSEUM OF COMPARATIVE ZOOLOGY. 47

“maturation spindle” was visible near the micropylar pole (Plate IV.
Fig. 1). As there were no other granulosa cells left attached to the
egg, the first impulse was to regard this as one of the “polar cells”
formed by the ovum during maturation. This seemed the more prob-
able on account of the undoubted existence of a maturation spindle.
A serious obstacle to this view was the great size of the cell as com-
pared with the narrow mycropylar canal. Even the elongated condition
of the cell would hardly warrant the assumption that it had passed
through so narrow an orifice. The examination of suitable sections
from ovarian ova (Plate IV. Fig. 5, Plate V. Fig. 2, Plate VII. Fig. 2)
soon showed that this interpretation was inadmissible, and made it as
certain as one could expect, without having traced it from its origin, that
the cell in question was a specially modified granulosa cell. It may be
appropriately called the mzcropylar cell, for, whatever may be its function,
the morphological fact remains that it occupies the micropylar funnel,
and lies directly over the micropylar canal. I have not been able to dis-
cover that its substance extends into the canal, but the number of favor-
able cases which I have examined is not enough to allow me to say that
such a condition is improbable. So far as | know, nothing of this kind
has been found in the case of any of the osseous fishes, unless the figure
given by Hoffmann (’81, Taf. I. Fig. 20) for Leuciscus is capable of
being thus interpreted.» Hoffmann himself has evidently not considered
the condition of the granulosa in the region of the micropyle sufficiently
important to give it any attention in the text, but there is not the least
doubt in my mind that the accumulation of granulosa cells which he
has figured is the equivalent of the granulosa plug in Lepidosteus. I
am inclined to believe, moreover, that Hoffmann has overlooked a real
difference between the cells in this region, and that an equivalent of the
micropylar cell of Lepidosteus will be found in Leuciscus, and perhaps
in many other of the osseous fishes, especially in those where there is a
large micropylar funnel. In fact the three cells which in Hoffmann’s
figure (Plate I. Fig. 20) seem to occupy the funnel, are all slightly larger
than the remaining granulosa cells, and one of them —the deepest —
fairly represents in its position the micropylar cell. Since all the cells
have a somewhat diagrammatic appearance, it is not too much to expect
that a more careful examination would show a difference between them.

1 Since this account was written, Owsjannikow and Cunningham have both
found similar conditions in other fishes. A review of their articles will be found
at the end of the historical section of the present paper, pp. 104-110.

48 BULLETIN OF THE

d. Origin of the Zona Radiata and the Villous Layer.

The youngest ovarian eggs in which either of the egg membranes has
been observed were about 430 » in diameter, and the ovaries to which
they belonged were preserved just before the period of spawning began.
Sections of such an egg are shown in Plate VIII. Figs. 1 and 2. Tan-
gential sections (Fig. 2) show that the egg is enveloped in a layer of
polygonal granulosa cells whose boundaries are exceedingly faint, and
whose nuclei have very irregular outlines, being lobed or deeply incised,
in some cases almost to complete division. The nuclei contain one, and
frequently two small nucleoli, but otherwise appear homogeneous, and
are uniformly stained. Upon focusing just below this layer of granulosa
cells, one sees the surface of the yolk covered with innumerable fine,
close-set points, which are evenly distributed.

Radial sections (Plate VIII. Fig. 1) supplement the surface views, and
show that the granulosa cells are relatively thin, and easily separable
from the underlying structures. Their protoplasm is finely granular,
and their boundaries are not distinguishable; neither do their deep
surfaces appear to be defined by any membrane. Their nuclei are
considerably flattened, and irregular in outline. ,

Immediately beneath the granulosa the surface of the yolk exhibits
fine radial, nearly parallel markings, which are close together and very
short. They are so intimately joined to the yolk that they seem to form
an integral part of it, and nowhere show the least tendency to become
detached from it. With high powers one can recognize avery thin corti-
cal portion of the yolk (membrane ?), with which they seem to be contin-
uous. It is very difficult to ascertain the distance between the markings,
but about 21 of them may be counted in the space of 17 pw, so that the
average distance is not far from 0.8 p. The length of each is about 0.5 p.

It would not be easy to determine from this stage alone whether the
markings indicate the beginnings of the formation of the zona radiata or
the villous layer. But even in this early condition the punctate mark-
ings of tangential sections appear brighter rather than darker when one
focuses high, so that the inference must be that they are due to minute
bodies which are more highly refractive than the surrounding substance.

This conclusion is abundantly confirmed by the study of somewhat
larger ova. These bodies seem to increase in length with considerable

1 In Fig. 2 (Plate VIII.) these punctations appear much too scattered in the
middle of the area which shows them. They are better represented toward the
margin of the area.

MUSEUM OF COMPARATIVE ZOOLOGY. 49

rapidity, for when the egg has attained a diameter of about 600 p
(Plate VIII. Fig. 3, Plate IX. Figs. 4, 5) they. may have reached the
leneth of 3-3.5 uw. In this stage the layer when seen from the surface
presents an appearance (lower half of Fig. 3, Plate VIII.) which so
closely resembles that of the zona radiata in the mature egg, that one is
involuntarily led to believe that it is the zona. Even the peculiar ar-
rangement of the markings in curved lines recalls the appearance of the
zona when seen in a similar position. Notwithstanding the striking
resemblance, there cannot be the slightest doubt that this layer is not
the zona radiata. In radial sections it is difficult to distinguish between
a layer composed of a homogeneous matrix pierced with minute parallel
canals, and one composed of parallel rod-like structures, but in surface
views this is much easier. Careful focusing shows the same optical
properties as were observed in the earlier stage, and with much greater
distinctness. The staining, too, is such as is to be observed in the villous
layer rather than in the zona; for the highly refractive bodies take the
deeper stain, the intervening substance having the paler color of the yolk.
But the last possibility of doubt concerning the nature of this layer is dis-
pelled by the appearance presented when the elements which compose it
are separated from each other. It frequently happens in mounting thin
sections that portions of the layer are detached, and even resolved into
their constituent elements. In such cases clusters of two or three rod-
like bodies, and even single ones, can be found in such proximity to the
layer as to leave no doubt that they are elements detached from it.
They have the same length and thickness as the markings of the layer ;
they are highly refractive and deeply stained. They can in no way
correspond to anything that is observed in the zona radiata, but do
resemble in several particulars the villi of older eggs.

From all this evdence I am certain that the layer which is first to
make its appearance between the yolk and the follicular epithelium is the
villous layer.

In this stage, too, the union of the layer with the yolk is much more
intimate than its relation to the granulosa. The latter is often sepa-
rated from the layer, the yolk never.

The cells of the follicular epithelium (Plate VIII. Fig. 3) have be-
come somewhat smaller than in the previous stage, but their nuclei
retain the same dimensions and the same lobed appearance which they
had during the earlier stage. As a consequence, the nuclei are closer
together. It will be seen that in the stage figured on Plate VIII.
Fig. 3, the diameter of a single average-sized granulosa cell corresponds

VOL. XIX. —NO. l. 4

50 BULLETIN OF THE

to the distance occupied by about a dozen of the villi. This tact will be
of some interest later, when a comparison is made with the conditions in
the mature egg.

In the sections figured on Plate LX. the villous layer has become still
thicker and the villi are correspondingly elongated ; they are also some-
what farther apart, as well as thicker.1 The thickness of the individual
villi is really greater than that of the spaces intervening between them,
but the appearance as seen under the microscope is represented with
tolerable accuracy in the figures. The villi of the egg shown in Fig. 3
(Plate IX.) have attained a little greater length than those of the other
eggs figured, but the egg itself was probably somewhat smaller than the
one shown in Plate IX. Figs. 1 and 5. Iam not entirely certain of this,
because the egg was incomplete, the yolk having all disappeared except
a portion directly underneath the villous layer.

A more advanced condition in the development of the ovum and its
membranes is to be seen in Plate VII. Fig. 5. The evidence that this
egg is more advanced than those last described is found in its slightly
greater size (nearly 0.7 mm.), and also in the increased size and elon-
gated condition of the yolk bodies which already occupy all parts of the
egg except a peripheral layer. The villous layer has here attained a
thickness of 5.5 », or about one third its thickness in the mature egg,
but the individual villi have not changed perceptibly from the condition
in the previous stage, except in regard to length.

I have no stages between this condition and that which the eggs pre-
sent at maturity, but already enough of the egg membranes has been
formed to allow several conclusions as to the method of their produc-
tion.

It is to be observed, that immediately before spawning there is no
structure, even in the latest of the stages here described, which can be
considered the zona radiata ; neither are there at this time any stages
older than the one last described, except the mature ova. It seems to
me, therefore, perfectly safe to infer that the ZONA RADIATA ts developed
after a large part, if not the whole, of the villous layer has been produced,
and that it is wholly formed during the twelve months immediately preced-
ing the spawning. From its late production and its position inside the
villous layer, as well as its intimate relations to the yolk, it is further to

1 In Fig. 5 (Plate IX.) they are represented a little too far apart and not quite
thick enough, whereas in Fig. 3 (Plate IX.) they have been represented too close
together. The granulosa cells in Fig. 5 are too sharply defined, especially on the
side toward the villous layer.

MUSEUM OF COMPARATIVE ZOOLOGY. 51

be inferred that the zona radiata is excluswely the product of the yolk. It
is also probable, from the evidence of stratification sometimes seen in
the completed structure, that the zona is produced in successive layers.
If such is the case, it follows that portions of the zona nearer the yolk
are formed after those which have a more peripheral position.

The question as to the source of the villous layer is not so easily an-
swered. The fundamental difference between it and the zona radiata at
once suggests for it a different origin. If the latter arises from the yolk,
the former might be produced by the follicular epithelium. This view
would seem to receive confirmation from the peculiar way in which the
roots of the villi in the mature egg penetrate the pore-canals of the zona
radiata. I have no doubt that this condition would be regarded by
many observers as a welcome confirmation of the theory that the pore-
canals are primarily for the purpose of transmitting nutritive material
to the growing egg. Such observers might look upon the villi as secre-
tions from the granulosa, which, owing to slight physical and chemical
changes, had not passed through the pore-canals as nutriment, but re-
mained partly outside the zona to subserve other functions. This view
might be further supported by the fact that during the formation of the
villi the inner surfaces of the granulosa cells are not sharply marked off
by membranes from the underlying structures.

Nevertheless, it seems to me that the arguments which may be
adduced to support the opposite view, — that the villous layer is the
product of the secretive activity of the ovum itself, — greatly outweigh
these considerations.

During the early stages of their formation the villi are so intimately
related to the ovum that they appear to be rods imbedded in its sub-
stance, and at no time during its formation is the villous layer separable
from the yolk. If the latter is by any means removed from the mem-
brane, there is always a superficial portion of the ovum which remains
attached to its inner surface. The separation of the granulosa cells
from the membrane during this period, on the contrary, is quite com-
mon. What might otherwise be a serious obstacle to considering the
villi the product of the ovum,—the presence of a zona between the
two, —is entirely nullified by the fact, previously established, that the
villous layer is produced before the zona radiata.

Whatever renders improbable the formation of the villi from the fol-
licular epithelium is, of course, favorable to the opposite view. If the
villi were products of the-epithelium, one would expect some constancy
in the numerical relations between the two, but this is certainly wanting.

52 BULLETIN OF THE

I have made some measurements and comparisons between eggs half
a millimeter in diameter and those having a diameter of about two milli-
meters, which indicate that the number of the villi remains constant
during the period of growth from the smaller to the larger size.

In an egg 0.5 mm. in diameter there occur about 30 villi in a space
of 35 w; i. e. the villi are about 1.15 « from centre to centre. In an
egg 2 mm. in diameter from the same ovary, treated in the same manner
and cut at the same time, the villi are 4.5 » from centre to centre (com-
pare Plate V. Figs. 3, 4). Allowing for the growth of the smaller egg,
which at the larger size would have a diameter four times as great as at
first, it is evident that the interval for a villus would be four times 1.15 p,
or 4.6 », which agrees fairly well with the space (4.5 u) actually occupied
by a villus in the larger egg measured. There are also other reasons
for believing that the villi do not increase in number after the egg has
reached a diameter of half a millimeter. If new villi were interpolated,
one would reasonably expect to find the younger ones shorter than the
older ones; but at no stage which I have seen is there any marked
difference in their lengths.?

In the larger eggs measured (2 mm.), the nuclei of the granulosa were
on the average about 14 » apart, from centre to centre; i. e. there were
about three villi to the diameter of each cell. But in eggs about half a
millimeter in diameter (compare Plate VIII. Fig. 3, and Plate IX. Fig. 5)
it is to be seen that from six to fourteen villi correspond to the diameter
of a single granulosa cell. If there has been no change in the num-
ber of villi, it follows that the granulosa cells must have increased in
number at least fourfold between the half-millimeter stage and the two-
millimeter stage. It is for this reason I contend that there is no con-
stancy in the numerical relations of villi and granulosa cells, and that
consequently it is improbable that the former are the product of the
latter.

1 It is evident that there has been a corresponding increase in the diameter of the
individual villi during the growth of the ovum, for in the mature condition they
form a continuous layer, with little or no intervening substance.

Ransom (’67) has claimed that the pore-canals of the zona radiata increase in
number during the growth of that membrane. If one were to disagree with me,
and to regard the markings which first appear at the surface of the ovum as the
incipient zona instead of the villous layer, he would be compelled to adopt Ransom’s
view, for the intervals between the markings on eggs half a millimeter in diameter
(1.15 ~) would become, unless there were interpolations, 4.6 4 apart when the eggs
had increased to two millimeters in diameter. In order to reduce the intervals to

the condition actually found in the zona of the mature egg (1.4 «), the number of
pore-canals would have to be increased more than threefold!

Es

=m.

LK ee

ation, +

MUSEUM OF COMPARATIVE ZOOLOGY. 53

At first thought one might regard the modifications of the villous layer
in the micropylar region as the direct result of an alteration in the secre-
tive powers of the granulosa cells situated at that place; but it seems
to me that the thickness of the layer ought, on this assumption, to be
greater than elsewhere, since the granulosa cells are here more numerous
and larger. Besides, the corresponding diminution in the thickness of
zona radiata could not be thus accounted for, but must be assumed to
be the result of diminished secreting activity on the part of the ovum in
this region. Hence the same explanation would certainly be more rea-
sonable in the case of the villous layer. This is a point which seems to
me of considerable importance; the diminished activity of this region
which is shown during the formation of the zona was already manifest
during the formation of the villi.’

From these several considerations, I believe there can be little ques-
tion that the VILLOUS LAYER of the egg membranes in Lepidosteus is also
the product of the ovum itself rather than of the follicular epithelium
surrounding it.

If this conclusion is established, it follows that the parts of the villi
first to be produced are those which are most superficial. I believe that
this is confirmed by the fact that the forming villi are readily stained in
carmine. It is probable that, even in the latest stage of the immature
eggs (0.7 mm.) which I have seen, not much, if anything, more than the
heads of the villi have been produced. The length and the highly re-
fractive condition of the villi at this stage, and the fact that they are
not at all folded, all point to this conclusion.

There still remains much to be done in following out the exact course
of the development of the membranes in Lepidosteus, — especially in
determining when the formation of the zona begins in relation to the
completion of the villous layer, —but I think that the main features of
the process as outlined above will not be disproved by subsequent
study.

I have no explanation to offer of the apparently sudden change in the
nature of the secretions from the ovum which is registered in the pro-
duction of structures so dissimilar as the zona and villous layer are ;
but it is possible that some light may be thrown on this question when
the period of the transition has been carefully worked out.

1 This is an evidence of the polar differentiation of ova (which exhibits itself in
many other phenomena) to which attention has not hitherto been called.

54 BULLETIN OF THE

-B. HistoricaL AND CriticAL REVIEW OF THE LITERATURE ON THE
Primary Eaa MEMBRANES! AND THE MICROPYLE IN FISHES.?

It is possible that the eggs of fishes may present as many as four
essentially distinct kinds of enveloping membranes before separation
from the ovary. The innermost of these, if it exists, may be considered
a true witelline membrane, the equivalent of the cell membrane in general.
I have made no observations concerning it, and shall have little to say
regarding the conflicting testimony as to its existence. The second, pro-
ceeding from the yolk outward, is radially striate, and I shall call it, as
in the preceding description, zona radiata. Although this is totally dif-
ferent in structure from the next membrane, there are several reasons
why it will be best to consider both at the same time. This third
membrane [ shall call, as previously, the vellous layer. The fourth and
outermost, when it exists, is formed exclusively from the granulosa cells,
and may be called by the name first given to it by Johannes Miiller, —
capsular membrane.®

a. Cyclostomata.

The eggs of the myxinoids are enveloped in a “horny capsule,” which
was first described by Thomson (’59, pp. 50, 51) for Myaine glutinosa.
He evidently considered it the equivalent of the egg cases of selachians.
Since the latter are formed in the oviduct, they cannot be considered

1 I use the expression primary egg membranes in the sense in which it has been
employed by Ludwig (’74 p. 197), i. e. for all membranes which are the product of
either the ovum itself or the follicular epithelium surrounding it.

2 Owing to delays in publishing my studies I have been able to extend this
review, and to bring it down so as to include papers which have appeared since
my own account was written.

8 J have the less hesitancy in adopting this name because Miiller (54, p. 189) —
notwithstanding some misconceptions as to its real nature in the perch — gave the
following concise, and, in my opinion, still perfectly applicable definition: “ Kine
von dem Eifollikel, Ovisac eines Wirbelthiers erzeugte Eihille scheint von der
Eischale anderer Eier unterschieden werden zu miissen als capsulare Evhiille, oder
Hicapsel.”” When subsequent observers, —as for example His (’73), — ignoring the
true explanation of Miiller’s investigations given by Leuckart (’55, pp. 257-260),
transfer the name Eicapsel to the zona radiata, one is compelled to protest that
that was not the structure described by Miiller under the name of “ Eicapsel,” and
that no one has yet brought forward satisfactory evidence that the zona is “ pro-
duced by the egg follicle,” as Miiller’s definition demands. It therefore seems to
me that it is better, for the sake of avoiding confusion, to drop entirely the name
capsule — whether egg capsule or “cartilage capsule” (His)—as a designation
for the zona radiata.

——

~

MUSEUM OF COMPARATIVE ZOOLOGY. 5d

as primary egg-membranes; but Steenstrup (’63) subsequently showed
that the egg of Myxine glutinosa possesses this covering before it leaves
the ovary, from which it follows that the “ horny capsule ” is really a
primary membrane.

THomson’s (759) account is brief: “I have found that in the Myxine
glutinosa the globular yolk is enclosed in a horny capsule of similar
consistence and structure [to that of the oviparous cartilaginous fishes],
but of a simple elongated ellipsoidal shape, and in place of four terminal
angular tubes, a number of trumpet-shaped tubular processes projecting
from the middle of the two ends, which probably serve the same pur-
poses as the differently shaped appendages of the ova of the shark and
skate.”

Sreenstrup (’63, pp. 233-238, Figs. a-h) also saw the horny egg-
shell and the peculiar projections from its ends. He says (p. 236):
“Tn the last received individuals the eggs now had not only the same
considerable size [as some large eggs previously described] and more
oval-elliptical form, but besides they were surrounded with a somewhat
firmer, almost horn-like egg-shell, which was furnished at the ends with
a large number of slightly curved or S-shaped horn-threads. Each
horn-thread ends in a head-shaped portion with three or four projecting
spines or hooks, and has thereby some resemblance to a ship’s anchor.
The threads recall — even though somewhat remotely —the horn-threads
projecting from the eggs of the rays and sharks, much as the shell itself
recalls the firm capsule of these cartilaginous fishes. The accompanying
figures exhibit both the appearance of the capsules (f, g) and the man-
ner in which they hang in the mesovarium (h** and h***), together
with eggs of the same appearance as ¢, d, e (Fig. h*), and with a large
number of only slightly developed eggs (0, 0, in Fig. h).”

In the two eggs with horny shell figured by Steenstrup, the shell has
been represented as though it were composed of two parts separated by
a sharp continuous line; the egg appears cut through near one pole by
a plane perpendicular to its long axis. The appearance recalls that seen
in the egg-shells of certain trematodes, where one end serves as a lid
which opens to allow the larva to escape; but whether the author re-
garded this as a similar provision for the escape of the young hag, or as
an accidental condition, is not stated in the text.

Witnetm Mouser (’75, pp. 114-117, Taf. V. Figs. 14, 15) appears to
regard the “Testa” of Myxine glutinosa — which I suppose to be the
same as the “horny capsule ” of Allen Thomson — as resulting from the
secretions [metamorphosis ?] of a layer of [granulosa] cells, which imme-

56 BULLETIN OF THE

diately invest the ovum. He does not expressly state this, but it seems
to me he leaves one to draw such an inference. He says that the
ovarian egg when 0.6 mm. in diameter is surrounded by a single layer
of very flat polygonal cells, outside of which is a thick layer of fibrous
connective tissue,’ and that when the eggs have attained a length of
18 mm. and a thickness of 6 mm. there are two connective-tissue
envelopes; an outer thinner, a continuation of the mesovarium, and
an inner, which at the ends of the egg is thickened (0.4 mm.) and
vascular. At its inner surface the inner membrane is condensed into
a lustrous membrana propria 2 thick, and is firmly attached to the
underlying “ Testa.”” In contact with the inner surface of this mem-
brana propria is a layer of cells. In the middle of the egg the cells are
cubical, but they become more and more cylindrical towards its poles,
where the layer becomes three or four cells deep.

I believe there can be no question that this layer of cells inside the
membrana propria represents the granulosa; but it seems as though
Miiller must have overlooked the egg membrane, if one existed at that
stage, and must have taken the granulosa to be in some way the equiva-
lent of it. Perhaps, assuming that the granulosa cells secreted the
membrane, his idea was that the granulosa ought itself to be considered
as a part of the “Testa,” for he afterwards (p. 126) mentioned, in. the
case of Petromyzon Planer, “a very thin folded egg membrane which
exhibited a polygonal pattern when seen from the surface.” Moreover,
he says, with regard to two deposited eggs of Myxine which he examined,
that there was no trace of either inner or outer connective-tissue en-
velope, and from this fact concludes that they must have undergone
complete regressive metamorphosis, similar to that which the enamel
organ of the teeth suffers after the completion of the enamel.

W. Miiller is the only person who has seen anything of a mzcropylar
apparatus in the myxinoids. ‘Exactly in the middle of the white pole
of the egg,” he says (p. 115), “this cell layer [granulosa] exhibits a
conical infolding 0.1 mm. deep and 0.06 mm. broad, which contains a
funnel-shaped opening, the micropyle, which is directed straight toward
the underlying nucleus and the protoplasm surrounding it.” This is
the whole of his description ; and from it I infer that he has seen that
portion of the granulosa which occupies the micropylar funnel, but that
the micropylar canal— which is a passage through a membrane, not an
involution of a cell layer —has not been seen by him. If the condition
in Myxine is at all comparable with that in Lepidosteus, it is certain

MUSEUM OF COMPARATIVE ZOOLOGY. 57

that Miiller has seen the equivalent of what I have called the micro-
pylar plug of granulosa cells, and it is therefore probable that he was
the first person to observe that peculiar structure in any fish-like animal.
If he were less positive in his assertion that the infolding contained an
opening, I should question if the cells took the form of a hollow funnel ;
even as it is, I doubt if the membrana propria is infolded.?

The first account of the membranes in Petromyzon Planert was by
Max Scuutrze (56, pp. 1-5). When taken from the body, the eggs had
besides the yolk membrane a firm “ Hischalenhaut,” or “chorion,” which
was surrounded with a scarcely discernible thin layer of gelatinous sub-
stance, which was quickly swollen, when it came in contact with water,
to a thickness of not more than a quarter of a line. It was delicate and
fugitive, and was easily removable from the firm underlying membrane.
In the course of eight days it mostly disappeared, being dissolved in the
water ; it was not an “albuminous layer,” but was rather to be compared
to the gelatinous mass uniting frogs’ eggs; its chemical composition was
not known.

The firm ‘ Eischalenhaut,”’ which closely enveloped the ege, was a
clear membrane about 0.0015!” (probably should be 0.015’, or about
0.03 mm.) thick, which had a tendency after being torn to roll in at
the edges. it appeared very finely punctate when viewed from either
the inner or the outer surface. Schultze was inclined to regard the
punctations as due to very fine canals traversing the membrane, but
on account of the delicacy of the object he could not reach a perfectly
satisfactory conclusion on this point. For this finely punctate mem-
brane and that found in bony: fishes, the author would use the name
chorion rather than vitelline membrane, for a true vitelline membrane
(or egg-cell membrane) exists inside the punctate structure.

OwssANnNIKowW (’70, p. 184) says that the gelatinous layer of the outer
egg membrane is very little developed, so that the fertilized eggs are
only feebly attached to the objects on which they fall, the least current
carrying them away.

CaLBERLA’S (’78, pp. 438-441) account is in some particulars more
extended than that of Schultze. The eggs, he says, instead of being
round, are slightly ellipsoidal. The membrane (zona) consists of two
layers, which are not, however, sharply separated from each other. The
outer is highly refractive, rough externally owing to all sorts of eleva-
tions and tooth-like structures (Zacken) ; the inner is much thinner

1 For a review of more recent work on Myxine, see pp. 91-93, 107-110.

58 BULLETIN OF THE

and translucent. With low powers the outer appears as though made

up of concentric layers, but with higher powers it is seen to be a

homogeneous substance traversed by fine radial canals which are con-
tinuous with those passing through the inner layer. At the outer sur-
face each of these canals opens out at the base of one of the elevations
(Zacken). Calberla regards this whole layer as a secretion from the
peripheral layer of the yolk. The proof of it he finds in the conditions
of the membranes in nearly ripe and in over ripe eggs. On the former,
the boundary between the two layers is sharper and the inner layer is
much thicker than on mature eggs; whereas on the latter all distinction
between inner and outer layer has disappeared.

As soon as the egg comes in contact with the water, the tooth-like pro-
jections on the surface of the egg membrane (zona) quickly swell, in con-
sequence of which the whole egg appears as if surrounded with a delicate
area of hyaline substance. This may well be the cause, he adds, of the
stickiness of the surface of the egg.

It seems to me that there is considerable reason for believing that
these external projections described by Calberla correspond to the villi.
of Lepidosteus, both in function and in position. An examination of his
figures (Taf. XX VII.) lends support to this view. I believe also that,
when the genesis of the membrane has been studied, it will be found
that these ‘‘Zacken ” are formed before the zona itself. It is true that

Kuprrer UND BENeEcKE (’78, pp. 9, 10) find the conditions somewhat
different from those recorded by Calberla. They claim that the envelope
of the egg consists in both P. Planeri and P. fluviatilis of a double mem-
brane (Eihaut), and of a continuous covering of gelatinous material which
is replaced at the watch-glass-like elevation of the membrane by a struc-
ture known as A. Miiller’s ‘‘Flocke.” The inner membrane — which
they figure as being much thicker than the outer — contains closely set
pore-canals, but these they assert positively are not continued into the
outer layer. The difference in structure between the two membranes is
demonstrable by means of 0.5 per cent hydrochloric acid. The outer
membrane swells more in water than the inner, but not quite uniformly.
It appears here and there as though it were restrained by a filament of
less easily-swelling substance. And this, they say, is probably the cause
of “Calberla’s unzutreffende Angabe, dass diese Rindenschicht mit
allerlei Erhebungen und Zacken besetzt sei, an deren Basis Poren-
caniiJe miindeten.”

But even if Calberla’s description is not quite satisfactory, it is evident
that this outer envelope is not homogeneous, and that the toothed appear-

MUSEUM OF COMPARATIVE ZOOLOGY. 59

ance which he has figured must have had a basis in optically different
portions of that envelope. According as the imbibition of water has
proceeded less or more, this marking might be more or less conspicuous.
From a comparison of the figures by Calberla with those by the last
mentioned authors, I should think that Calberla’s outer layer of the
zona by no means corresponded with the outer layer of Kupffer und
Benecke, and that the latter, being very thin, had been overlooked by
Calberla.

_ The micropyle of Petromyzon, though sought for by Schultze (56)
and A. Miiller (64) was not found by them.

Owssannikow (’70%, p. 184), who discovered it, says that it is very
small, but that it remains visible for several days after fertilization. In
mature eggs it occupies a position over the eccentric nucleus.

CaLBerLA (’78, pp. 439, 440) has given a careful description of the
micropyle, which, he says, agrees in all essential particulars with that of
osseous fishes. His account is substantially as follows. At one pole
of the elongated egg its membrane is thickened, and bulges out, much
as though a shallow watch-glass — with shorter radius of curvature than
the rest of the egg membrane — had been set into one end of the mem-
brane. Radial sections which pass through the centre of the elevated
portion of the membrane show that in the middle of it there is a very
flat saucer-shaped depression, the centre of which is further depressed
into a funnel. From the narrow end of the funnel a canal is continued
through the membrane, and opens on its inner surface with a slight
flaring. A little below its middle the canal exhibits a spindle-shaped
enlargement, which is shown in Calberla’s Taf. XX VII. Figs. 2 and 3.

The views held by Kuprrer unp BeENECcKE (78, pp. 9-15) regarding
the nature of the micropyle are not easily summarized. They are based
on close observations of the deportment of the egg and spermatozoa at
the time of fertilization, but do not appear to have been corroborated
by sections of the egg membranes.

In the region of the watch-glass segment of the membrane described
by Calberla, the mucilaginous envelope outside the membranes is want-
ing, and in its place is a hyaline dome (A. Miiller’s ‘‘ Flocke ”) composed
of a substance which, unlike the mucilaginous layer, is permeable for
spermatozoa. Usually only one spermatozo6én passes through the inner
and outer egg membranes and reaches the yolk; but the place of its
passage is by no means always the centre of the watch-glass area. It
was such only six times out of fifty. The passage may occur even near

60 BULLETIN OF THE

the margin of this area. Neither is it always the spermatozoén that
first reaches the outer membrane, after having traversed the “ Flocke,”
which passes through.

The statement that the egg membrane is not alone permeable at a
single spot would lead one to suppose that the authors were ready to
deny the existence of a micropyle. They do not, however, directly assert
its absence, although they were unable to find anything of it on the un-
fertilized egg. But as soon as the spermatozoén has passed through the
membrane, a small circular spot may be seen from the surface; this is
due to a shallow depression in the surface of the znner layer of the mem-
brane, the outer layer never showing any passage through it. The authors
hint at the possibility of a chemical action on the part of the spermato-
zoon resulting in a loosening of the two layers and a partial solution of
them, and endeavor to make that view harmonize with the conclusion
that the micropyle “is the remnant of an opening in the inner layer of
the egg membrane, which exists during the stay of the egg in the fol-
licle, corresponding to the condition which Herr von Jhering recently
established in the case of the eggs of the mussels.” The outer layer
would be formed, they imagine, afterwards, and would cover over this
opening, leaving a remnant of it recognizable on the inner membrane.

‘The micropyle, therefore, is not an open passage, as it appeared from
Calberla’s description and drawings, but only a permeable place.”

b. Selachia.

What Ludwig wrote in 1874 concerning odgenesis in the selachians,
that it had been studied by only a very few investigators, was equally
true of the primary egg membranes of the group. Lupwie (74, p. 145)
himself, although he studied the development of the ova, had nothing to
add to what was already known about the egg membranes, and since him
there have been only two writers who have dealt with the subject,
Schultz and Balfour.

Leypie (’52, pp. 87, 88) speaks incidentally of a vitelline membrane, and
a thin albuminous layer surrounding it, inthe case of Raja batis. The latter
probably corresponds to one of the membranes seen by later observers.

GrcEeNBAUR (’61, p. 518) recognized the existence of a homogeneous
egg membrane on eggs of Raja from 1!’ to 2!!! (2-4 mm.) in diameter ; its
external contour was delicate, but internally it was sharply limited. In
the case of Acanthias there was only this one membrane to be observed ;
it attained a thickness of 0.08!” (175 yp) on eggs 4/’/-5!"’ (9-11 mm.) in
diameter.

MUSEUM OF COMPARATIVE ZOOLOGY. 61

Gegenbaur considers it probable that this membrane is produced by
the follicular epithelium, but is evidently not certain of it. He says:
“Es liegen hier wohl bei den Selachiern andere Verhaltnisse vor als bei
den Vogeln und Reptilien, und eine Dotterhaut, wie sie dort von Seite
des Dotters durch Umwandlung seiner peripherischen Schichte zu Stande
kam, kommt hier wohl nicht vor, sondern der Dotter bleibt auf dem
friiheren Stadium der Differenzirung bestehen, dagegen bildet sich eine
Hiille von aussen her, wozu wabrscheinlich die Zellen des Follikelepithels
das Material abscheiden, wenn man den Vorgang der Bildung jener
Membran nicht auf die Oberfliche des Dotters selbst verlegen will.”

Schultz and Balfour disagree in their conclusions as to the origin of
the fugitive membranes which envelop the ovarian eggs of selachians.
Schultz ascribes their formation to the follicular cells ; Balfour, to the
ovum itself.

Scnuttz (’75, pp. 574-576) claims that in Torpedo oculata the fol-
licular epithelium is composed of two kinds of cells: genuine granulosa
cells, derived from the germinal epithelium of the ovary, and, alternating
with them, lymphoid cells, which are derived from the stroma of the
sexual organ. ‘The cells of this follicular epithelium, especially the
lymphoid cells, are merged at their deeper ends into a homogeneous
cuticular layer (Fig. 8), and there form a structure having the morpho-
logical value of a chorion.” This homogeneous layer at no time has a
morphological relation to the egg protoplasm, but retains the closest con-
nection with the follicular cells. On objects subjected to pressure the
outer margin of the homogeneous layer appears jagged like a wood-saw,}
the remnants of the lymphoid cells corresponding to the teeth, in the
intervals between which the granulosa cells are lodged. The latter are
also attached, he says, to the homogeneous layer by means of proto-
plasmic processes, and even appear to fuse with it, but do not show any
differentiation within the substance of the layer. It is not possible
even with the highest powers to demonstrate any such structural pecu-
liarities (radial striation, pore-like perforations) as are met with in the
egg membranes of most classes of animals, even in Raja batis itself.

“Finally, when the egg cell has reached maturity and the follicle
approaches the stage of rupturing, the lymphoid cells together with the
homogeneous layer are converted into connective issue, in the inter-
stices of which the granulosa cells persist, although the latter finally
undergo fatty degeneration. Only at a single place, corresponding to
the whole extent of the germinal disk, do the follicular cells and the

1 “ Gleichsam hohlsageformig [hohlzsigeformig ?] gezackt.”

62 BULLETIN OF THE

homogeneous layer persist unchanged up to the bursting of the follicle.
It is from this part that those granulosa cells come which are occasion-
ally encountered on the escaped [egg] and within the empty follicle.”

On eggs of Acanthias, Scymnus, and Mustelus, Schultz found a
homogeneous layer joined with the follicular layer, and inside the latter
a zona radiata, the inner margin of which was sharply defined against
the yolk. “The pores of this cuticular zona were traversed by proto-
plasmic processes, which stretched from the homogeneous layer to the
ege protoplasm and fused with the latter.”

The author concludes that, so far as his own observations reach, there
are to be distinguished in selachians the four “following conditions
of the follicular epithelium”: (a) simple epithelium (embryonic stage
of selachians); (0) epithelium with homogeneous basal margin (Tor-
pedo); (c) epithelium with homogeneous perforate basal margin (Raja) ;
(d) epithelium with broad homogeneous and narrower perforate basal
margin (Squalide).

Batrour (’78", pp. 402, 403) has confirmed the existence and subse-
quent disappearance of two membranes—an outer homogeneous, and
an inner striate—in one of the Squalidz, Scyllium ; but he believes
that they are produced by the ovum, not by the follicular epithelium,
and that they are absorbed, not converted into connective tissue. Two
similar membranes are also found in Raja, and are believed by Balfour
to be common probably to all sharks. The homogeneous membrane is
formed before the striate one. In Scyllium “the [homogeneous] mem-
brane would seem indeed to be formed in some instances even before the
ovum has a definite investment of follicular cells.” Consequently it is
called a witelline membrane. In ova 0.12 mm. in diameter it is not thick
enough to be accurately measured ; in those of 0.5 mm. diameter it has
a thickness of 2 », and there may also be observed inside it faint indi-
cations of the differentiation of the outermost layer of the vitellus into
the perforate or radially striate membrane of Schultz. The latter Bal-
four does not hesitate to call a zona radiata.

In ova 1 mm. in diameter the zona has increased in thickness (to 4 )
and “is always very sharply separated from the vitelline membrane, but
appears to be more or less continuous on its inner border with the body
of the ovum, at the expense of which it no doubt grows in thickness,”
In larger eggs both membranes increase in thickness, especially the zona,
which now becomes marked off from the yolk. ‘In many specimens it
appears to be formed of a number of small columns as described by
Gegenbaur [for the alligator] and others.”

MUSEUM OF COMPARATIVE ZOOLOGY. 63

The size of the ova at the time of the maximum development of the
membranes is not stated; but after this stage is reached both mem-
branes gradually atrophy. “The zona is first to disappear, and the
vitelline membrane next becomes gradually thinner. Finally, when the
egg is nearly ripe, the follicular epithelium is separated from the yolk
by an immeasurably thin membrane, —the remnant of the vitelline
membrane,” which is no longer visible when the egg becomes detached
from the ovary. Both vitelline membrane and zona are found in Raja,
_ but in a much less developed condition than in Scyllium, and the zona
is developed at a much later period than in that species.

If the account given by Balfour is correct, — and his description
seems to be both more complete and more accurate than that of
Schultz, — then there is an interesting parallelism between the pri-
mary egg membranes in selachians and those of Lepidosteus and the
bony fishes. Not that the villous layer in Lepidosteus is structurally
comparable with that which Balfour calls in sharks a vitelline mem-
brane, but genetically they are alike. They are the membranes which are
first to be produced, —i. e. before the zona radiata, — and they are in
both instances the product of the ovum, not of the follicular epithelium. I
shall take occasion later to refer to the theoretical importance of this
discovery by Balfour.

The ultimate disappearance of both membranes renders the formation
of a micropyle superfluous. It would be interesting to learn, however,
whether there is at any time a trace of such a structure.

ce. Ganoirdei.

The only ganoids besides Lepidosteus whose egg membranes have
been described are Amia and Acipenser. I have elsewhere (p. 27)
quoted Ryder’s account of the membranes in the case of Amia.

From the accounts given by Kolliker, by Kowalevsky, Owsjannikow
und Wagner, and by Salensky, I believe there must be considerable
similarity between the conditions of the egg membranes in Acipenser
and Lepidosteus.

KO Lurker (57, p. 197) says that the porous membranes in the case
of the sturgeon form ‘three layers; two inner, darker, thinner, closely
porous, and an outer pale, thicker layer, apparently with fewer pores.
This outer layer, which is also softer than the others, shows its outer
surface divided into small polygonal areas, which appear to correspond
to the epithelial cells of the egg capsule [follicle]. These cells are ex-
tremely delicate and pale, but yet seen from the surface they show a

64 BULLETIN OF THE

fine punctation. It as the appearance, therefore, as though these cells
secreted the porous layers ; however, concerning this, as well as concern-
ing the corresponding parts of the eggs of other animals, only careful
studies made on eggs of all ages can give an answer, wherefore I abstain
for the present from any opinion.”

According to KowaLevsky, OwsJANNIKOW UND Waaner (’708, p. 172),
the outer membrane in Acipenser is thick, shagreen-like, and possesses
numerous very fine canals. When the ripe eggs fall from the oviduct,
this membrane sticks to objects ; with a certain amount of skill it may
be rather easily detached from the egg. The inner membrane is much
finer [has finer canals ?], transparent, and very firm.

SALENSKY (’81, pp. 234-236) applies the name chorion to the outer of
the two layers composing the “thick capsule’ which envelopes the ripe
egg of Acipenser ; the inner he calls vitelline membrane. At first the
two are so intimately joined to each other that it is difficult to separate
them ; but after the egg has been deposited for some time, the chorion
is easily detached from the vitellme membrane, and may be removed
from the whole egg. From the study of sections of stained eggs the
author determined that the stickiness and the roughness of the surface
were due, not to the chorion, but to two special cell layers which in-
vest it.

“ As to the origin of these two membranes, there is no doubt that they
are derived from the two layers of cells which constitute the epithelial
wall of the ovarian follicle.”

“The chorion,” he adds, “is probably a product of the secretion of
the membrana granulosa of the follicle; when the latter ruptures, the
epithelial cells remain adherent to the chorion and are expelled with
the egg, and are again met with slightly modified at the surface of the
deposited egg.”

“The examination of microscopic sections of the egg shows that this
envelope, which, as has been said, is divided into two after deposit, —
into chorion and vitelline membrane, — presents three distinct layers.
The external and the internal have about the same thickness; the mid-
dle one is thinner. In separating the chorion from the vitelline mem-
brane one may convince himself of the fact that the chorion is composed
of two layers, the vitelline membrane of only one.”

From Mayzel’s abstract (Salensky, ’79, p. 220) I learn further, that
the outer layer is stained deeply by hematoxylin, the two remaining
layers not at all, and also that all three layers are radially and finely
striate. In the figure of the membranes given by Salensky (’78°,

‘

MUSEUM OF COMPARATIVE ZOOLOGY. 65

Tab. I. Fig. 8 B) the inner membrane is only slightly thicker than the
middle one, and both present a lighter appearance than the outer
one.

I believe it is probable that the outer layer will be found to corre-
spond closely to the villous layer of the gar-pike. There can be no
doubt that the inner layer is the zona radiata, and I am inclined to
regard the middle layer simply as the differentiated outer half of the
zona; but the question can be answered satisfactorily only after renewed
investigations which give more particular attention to this point. The
principal reasons for my conclusion regarding the middle layer are, that
it evidently resembles the middle layer more than the outer, especially
in its capability of being stained, and that differences between the inner
and outer portions of the zona have been observed in the case of other
fishes. I know of no case, I admit, in which the outer half of the zona
may be easily removed with the villous layer, so that it still is possible
that the middle layer in Acipenser corresponds to the stalk region of the
villi in Lepidosteus.

I do not clearly understand what the author means by saying that
the two membranes are derived from the two layers of cells which con-
stitute the epithelial wall of the follicle. It is true he claims that there
are two distinct cell layers, which, according to his figures of early
stages (’78°, Tab. I. Figs. 5-7), are ‘ granulosa”? — next to the yolk —
and ‘‘follicular epithelium ’’ —immediately outside the latter; but he
gives no figure showing both these at advanced stages of development.
I doubt their existence. But even if there were two separate epithelial
layers, I fail to understand how both could share in the production of
membranes which lie wholly inside the inner cell layer.

The eggs of Acipenser are altogether uniqne so far as regards the con-
dition of the membrane in the micropylar region.

The earliest observer of the micropyle was KéxuixKer (’57, p. 197),
who incidentally remarks that there is a single micropyle in the stur-
geon’s egg; but subsequent observers have claimed that there is a group
of several macropyles.

KowaLevsky, OwsJANNIKOW UND WacneR (70%, p. 172) state that
at one pole of the egg there are seven micropylar openings, — one in the
centre, with the other six arranged in a circle around the first.

SALENSKY (’81, pp. 235, 236, Planche XV. Fig. 1 A) gives a more com-
plete account of the micropylar region, which is illustrated by figures.
He says: ‘‘ At the germinative pole of the egg there is found a micropyle.

VOL. XIX.— No. 1. 5

a esses sss ssssessessssssssssss sss estan sess as

66 BULLETIN OF THE

The orifices of the micropyle are so small on hardened eggs that it is
probable that they are narrowed by the action of the reagents, which
also cause a retraction of the capsule of the egg. For this reason it is
very difficult to find these orifices.”

The micropylar apparatus is composed of several (5 to 13) orifices,
and Salensky states that, although he has examined a great number of
eggs, he has never found two which were identical either in the number
or distribution of the orifices. ‘‘ Each orifice consists of a quite small
pit (fossette) having the form of a crater; it is surrounded by small,
very slender cylinders.”

Salensky has given only surface views of the micropylar region, but it
has occurred to me that his ‘‘ very slender cylinders” may be villi which
surround the crater-like depression. They would appear to have a radial
arrangement about the crater as a centre, if the villous layer were seen in
optical section at a plane a little below the outer margin of the crater.

d. Dipnor,

The conclusions reached by Brepparp (’86") relative to the ova of
Lepidosiren possibly rest on too limited material to receive immediate
acceptance. So far as regards the egg membranes, the story is certainly
far from being satisfactorily completed. In the youngest eggs there was
no trace of any membrane; but as development proceeded, a delicate
homogeneous membrane encircled the ovum. This was from analogy
thought to be the product of the egg protoplasm, even though it was
more firmly adherent to the follicular epithelium than to the ovum,
and it is called vitelline membrane. In more mature ova there was
underneath this a much thicker radially striate membrane, probably
corresponding to the zona radiata of other vertebrates, which in places
seemed to pass gradually into the substance of the protoplasm. This
membrane (zona radiata) began to disappear with the first steps in the
formation of yolk. During the period of yolk formation the vitelline
membrane became thicker, and also radially and coarsely striate. The
author believes that there was a stage which succeeded this, during
which there was no membrane of any kind, and that at this time an
immense number of follicular cells migrated into the yolk. But in
addition there followed still another stage, when the ovum was en-
tirely occupied by yolk, —in which the epithelium was separated from
the contents of the ovum by an extremely delicate homogeneous mem-
brane, which either corresponded (in some cases) to the persistent vitel-
line membrane, or (in other cases) was a new formation; but even in

‘

MUSEUM OF COMPARATIVE ZOOLOGY. 67

the latter case the author maintains that it is homologous with the
vitelline membrane !

The fact that the layer called by Beddard wtelline membrane becomes
radially and coarsely striate, suggests comparison with the villous layer
of Lepidosteus. As in the ganoid, so in Lepidosiren this is the layer
which is produced first.

e. Teleoster.

The numerous descriptions which have been given of the egg mem-
branes in different osseous fishes show that there is not uniformity either
in their number or structure. Besides the wall of the follicle with its
epithelium, the granulosa, there is perhaps only one investment of the
egg which is universally present, the zona radiata, and even this may
be wanting, or at least may wholly disappear in the case of certain
viviparous fishes. I believe it is certain that the homologue of the zona
radiata is invariably present in oviparous fishes, and I am likewise
of opinion, notwithstanding the inability of some observers to discover
the presence of pore-canals in the eggs of certain fishes, that the canals
are also always present.

In many cases there is a membrane intervening between the zona and
the granulosa, sometimes thin and homogeneous, at other times of a
more complicated nature. It may even (villous layer) resemble some-
what the zona itself, although in reality very different from it in struc-
ture. Whether the thin homogeneous layer is homologous with the
more complicated villous layer cannot as yet be definitely answered, but
Balfour’s account of the origin of a similar structureless membrane in
Elasmobranchs makes me incline to the belief that it is.

In addition, the cells of the granulosa undergo in some instances a
remarkable metamorphosis, accompanied probably by a process of secre-
tion, and thus furnish still another primary envelope to the ovum.

As to the existence of membranous structures z¢mside the zona radiata,
there is much less certainty. The presence of a structureless wtelline
membrane has been maintained with more or less confidence by Vogt,
Aubert, Thomson, Lereboullet, K6lliker, Eimer, and others, and more
recently by Owsjannikow and Scharff. Its existence has also been
denied by eminent authority.

The question as to the presence and nature of a so called zonoid layer
seems, especially in the light of Brock’s recent contributions, to demand
a more cxtended and thorough investigation. In the case of some of
the eggs studied by Eigenmann (’90), structural conditions have been

68 BULLETIN OF THE

observed which can hardly bear any other interpretation than that of a
striate zone of substance inside the zona radiata proper; but in other
cases (Esox, Amiurus) a somewhat similar though perhaps not identical
appearance is due to a retraction of the vitellus from the zona, which
leaves strands of vitelline substance stretching across the space thus
produced.

1. Zona Radiata and Villous Layer.

JOHANNES M@xuer (54) has often been credited with having discov-
ered the fine radial canals which traverse the zona radiata, and give to it
its most characteristic appearance. This is a mistake which has arisen
from Miiller’s misunderstanding the relation of the peculiar membranes
of the perch to those of other fishes. What he described as “ pore-
canals” in Perca belonged to a much thicker membrane than the zona.
This membrane lies outside the latter, and is a result of the activity
aud metamorphosis of the granulosa cells. Miiller, it is true, supposed
this to be the equivalent of the ‘‘shell membrane” previously described
by Vogt, and therefore imagined that he had been able to demonstrate
on a more favorable object what Vogt had claimed on grounds of analogy
rather than on satisfactory proof.

In Voer’s studies on Coregonus palea he (’42, pp. 1, 8-10, 27, 28)
claimed the presence of two membranes. The inner one — being thin,
transparent, and without apparent texture—he called wtelline mem-
brane; the outer one he called a shell membrane, and homologized it
with the “membrane coquilliére ” of birds’ eggs. This outer membrane
presented the appearance of shagreen, which seemed to result from a
quantity of small opaque points uniformly distributed over its surface.
Treated with hydrochloric acid, the points became more transparent,
and then resembled minute warts. Valentine called Vogt’s attention to
the resemblance between this structure and that of the carapace of the
cray-fish, where he had found that a similar effect was due to perpen-
dicular tubes, filled with lime, traversing a membrane composed of regu-
larly polyhedral cells. Vogt, admitting that the “shell membrane” was
too thin to allow the attainment of exact results relative to the nature
of the “points,” nevertheless claimed that the position, behavior, and
reticulate appearance of the latter warranted one in supposing that the
structure was analogous to that of the carapace of the cray-fish. Thus
it appears, he continues, that the shell membrane is formed by the union
of flattened cells, which are arranged around the primitive egg only
toward the epoch of its maturity ; the presence of these minute tubes,

CO

MUSEUM OF COMPARATIVE ZOOLOGY. 69

which traverse the membrane, would in his opinion sufficiently explain
the absorption of the water into the interior of the shell membrane.
(Compare also Vogt, ’42, pp. 27, 28.)

From this summary it is evident that Vogt had under consideration

appearances which were due to the pore-canals of the zona radiata, and
that he moreover believed in the presence of canals, but it cannot be
claimed that he demonstrated their existence. Unlike his predecessors,
he rightly claims that the shell membrane (in Salmo umbla) originates
in the ovary.
« In a paper written in 1845, but not published till many years later,
Voet ert PappENHEIM (’59, pp. 357, 361, 362) also maintain that the
shell membrane of the eggs of fishes, which is uniform and elastic, is
constituted by the fusion of a cell layer formed wm the ovary, and there-
fore not to be compared with shell membranes which are produced in
the oviduct. They made the mistake of insisting that “this cell layer
is not to be confounded with another epithelial layer which one finds in
the ovisacs of the youngest ovules, and which is composed of large ex-
tremely pale cells which subsequently disappear and give place to this
second layer.” |

MeckeL von Hemspacn (’52, p. 421, Taf. XV. Fig. 1) saw and fig-
ured a radial structure of the “zona pellucida” in the case of Cy-
prinus auratus, after treating the membrane with acetic acid and
crushing it, but he expresses no opinion as to the real nature of the
striation.

Levcxart (53, pp. 796, 797), who probably had not yet seen Meckel’s
paper, was evidently not impressed with the explanation which Vogt
adopted to explain the appearance of the outer membrane in Coregonus ;
for he says simply, “That which characterizes the eggs of teleosts is
the possession of a special firm egg-shell (chorion), which is already
formed in the follicle around the primitive yolk membrane, and generally
presents a delicate markmg resulting from regularly grouped granules
or points.”

During the five years beginning with 1853 there appeared a large
number of papers dealing with the egg membranes of fishes, and the
subject was brought to a temporary conclusion by the thorough work
of Kolliker.

AUBERT (’53, pp. 94, 95, Taf. VI. Fig. 1) was the first to figure well
the appearances due to the pore-canals of the zona radiata. In addition
to “a very finely granular, but otherwise structureless skin, which en-
velops the yolk,” and which I believe must have been in his opinion the

70 BULLETIN OF THE

vitelline membrane,! Aubert describes the “shell” of the egg of Esox as
a transparent thin membrane, furnished with fine points, which closely
envelops the yolk. These points exhibit a great regularity of arrange-
ment, being placed at the intersection of symmetrical curved lines.
When the shell has lain some time in water, it separates in many places
into two membranes, of which the outer is very thin, finely granular,
and irregularly elevated, while the inner is thicker, uniform, and upon
sections exhibits fine radially placed streaks.

I believe it is certain that the radial streaks described were due to
the pore-canals, although the author does not fully commit himself te
that view. ‘The spermatozoa are so large,” he says, “that it would be
difficult for them to pass through the ‘points’ of the shell, in case the
latter are regarded as the lumina of fine canals.”

Also LerEBouLLet (54, pp. 240, 242, 245, 249) wrote concerning the
pike: “The ripe egg is surrounded by two membranes : the external is
pierced by microscopic tubes, which serve for the absorption of water,
and consequently for the respiration of the egg; the internal, applied
to the vitellus, is a simple, extremely thin and amorphous protecting
envelope.” He also saw the pore-canals in the perch, and argued that
the expulsion of albuminous globules from the fertilized egg proved
the absence of a vitelline membrane at that time, and that. it went to
confirm the opinion of those who regarded the chorion as produced by
the primitive vitelline membrane, which was itself detached from the
vitellus.

Although J. MUuuer (’54) contributed much to the knowledge of the
egg membranes, especially of the perch, and was also the first to appre-
ciate the importance of the difference in origin between the egg-shell in
birds and what he called the “egg capsule” in fishes, he did not fully
comprehend the structure of what he called the ‘‘ Dotterhaut ” (zona).

1 I cannot agree with His (’73, p. 2, foot-note, compare also p. 14) in his criti-
cism of Aubert when he says: “ Der Name Dotterhaut, welchen die friilieren
Schriftsteller fiir eine besondere den Potter unmittelbar umhiillende structurlose
Membran gebraucht haben, wird von H. Aubert (Beitrage, ete., 94) auf die Eikapsel
angewendet. Er spricht nimlich bein Hecht-Ei von einer Trennung der Dotter-
haut [!] in zwei Schichten, eine dussere diinne, fein granulirte und eine innere,
dicke, mit radiiren Streifen. Eine Begriindung seiner abweichenden Bezeich-
nungsweise giebt er nicht.”

Aubert says distinctly enough that it is the “ Schale” which is divided into two
membranes; and although he nowhere employs the word “ Dotterhaut,” there seems to
me no doubt that he has Dotterhaut in mind when he says: “ Der Dotter wird

von einer sehr fein kornigen, sonst structurlosen Haut iiberzogen.” His may have
been misled by the statement that “ Die Schale etc. den Dotter eng umgibt.”

MUUSEM OF COMPARATIVE ZOOLOGY. 71

He claimed for Cyprinus erythrophthalmus, Perca fluviatilis,and Acerina
vulgaris a velvety appearance of the external surface of this membrane
“as if beset with tufts”? (Zotten!), which he ascribed to ‘“‘ very small cy-
lindrical projections or rods with rounded ends,” — prolongations of the
vitelline membrane itself. Here again Miiller unfortunately confounded
the unlike conditions of different eggs. While his conclusions have been
confirmed in the case of Cyprinus erythrophthalmus, I know of only one
author whose observations give any evidence of the presence of such a
“pile” or velvety structure outside the zona in either of the other fishes
mentioned. Hoffmann (81, pp. 19, 20, Taf. I. Figs. 9, 10), it is true, not
only figures an external layer of the zona in a nearly ripe egg of Perca,
which is thinner and much more sparsely striate than its inner layer, but
he also describes and figures an October egg as possessing outside the still
thin zona a layer of minute, close-set tubercular projections which “ fully
correspond to the Zéttchen of the Cyprinoids.” I believe that Hoff-
mann has in some unaccountable way fallen into an error in this matter.
At least, no other observer has seen any trace of the structure which he
describes, and an examination of the eggs of our American perch in
October reveals nothing of the kind. I think the condition in Acerina
is probably similar to that in Perca,—certainly no one has shown the
presence of a “ Zéttchen” layer. That being the case, what was it
that Miiller mistook in Perca and Acerina for ‘ Zéttchen”? A state-
ment by Owsjannikow (’85, p. 18) makes me believe that Miiller may
have had under view the branching deep ends of the tubular structures
which according to Owsjannikow traverse the gelatinous envelope, and
which are left sticking in the pore-canals of the zona when the two
layers are artificially separated. It may be, however, that Miiller saw
the more conspicuously — bué not, as Hoffmann makes it, more sparsely
—striated outer portion of the zona itself, in which event it might be

1 Miller was not the first to see the appearance presented by these “ Zotten,”
nor even to suggest the name. Von Baer (’85, p. 7) had, twenty years before,
seen the same thing in species of carp. “Die dussere Eihaut ist aber nicht
ganz formlos und gleichartig in sich. Sie enthalt in den Karpfenarten, die ich zu
untersuchen Gelegenheit hatte, keine [should be kleine] dunkleren Vorragungen, die
ihr bei starker Vergrisserung ein zottiges Ansehen gaben.” It is evident that Von
Baer confounded the radial “ tubes ” in the outer envelope of the perch egg with
these villi of the carps, for he adds: “Im Barsche ist diese Hiille noch sehr viel
dicker und man sieht, dass die dunklern Flecken, die hier lang und schmal sind, in
der dussersten Schicht sich befinden.” One will find little occasion for surprise at
this parallelism, in view of the fact that some of the most recent observers, with
the best modern appliances at command, have arrived at a similar, though I believe
erroneous conclusion.

72 BULLETIN OF THE

fairly claimed that he had seen the pore-canals ; but even in that case
it remains perfectly evident that he did not understand at all the
structure of the zona radiata.

When, a few months later, Remax (’54) announced the discovery of
radial striations traversing the whole thickness of the zona pellucida in
the case of the rabbit’s egg, and attempted to determine the cause of
the appearance by a comparison with the conditions found in the egg
membrane of a fish (presumably Gobio fluviatilis), Miiller (54%, p. 256)
was unable to agree with him in the conclusion that the appearance was
probably due to ‘an alternation of canals and cylinders,” but en-
deavored to show that it was “merely an optical expression of the sum-
mation and partial superposition of the images of the rods [Zapfen] as
seen when viewing the vitelline membrane in profile.” The images of
the overlying rods which fall in one line would, in his opinion, cause
the striations to appear much longer than the individual rods really
were, and thus make the lines appear to traverse the whole thickness
of the membrane.

Ransom (’56) maintained that there was present at an early stage
in the growth of the ovum of Gasterosteus a very thin membrane, hav-
ing a finely and regularly dotted structure ; but he does not appear to
have realized as yet that the dots were evidence of pore-canals. He also
discovered that in older eggs the part of the membrane immediately
surrounding the micropylar depression exhibited a number of cup-
shaped pediculated bodies scattered over its surface. These have since
been claimed by Kolliker (’58, p. 81) and subsequent authors to be the
localized equivalents of the “ Zapfen” layer discovered by Miller.

In the same category must also be placed the remarkable filamentous
structures discovered by Harcxen (55) on the eggs of several of the
Scomberesocidee. Although Haeckel described the filaments as lying
inside the finely punctate vitelline membrane (zona radiata) and having
no connection either with it or the yolk, there can be no doubt, as
Kolliker (58, p. 81) first showed, that they are really outside the zona.
I have not yet had the opportunity of examining the eggs of any of the
Scomberesocidze, but conjecture that the sheath which envelops the bases
of their filaments may be only a part of a membrane external to the
true zona radiata, and comparable with that which Eigenmann (90)
has found in Fundulus.

Leuckart (°55, pp. 257-264), who in an appendix to his celebrated
paper on the micropyle of insects’ eggs deals, at least incidentally, with
the structure of the egg membranes in fishes, was the first to perceive

MUSEUM OF COMPARATIVE ZOOLOGY. ie

the true significance of Miiller’s discoveries in the perch. He retained the
name chorion for the zona radiata, and from a study of the trout was
fully convinced that the punctate appearance is due to “ delicate tubules
or canals which traverse the membrane perpendicularly, without open-
ing, however, at its inner surface.” The latter part of this statement
has not been confirmed by subsequent observers. Leuckart, however,
gave an excellent description of the structure of the zona radiata in the
perch, for he not only recognized that it was composed of an outer thin-
ner, firmer membrane, and an inner thick layer of viscid sarcode-like
substance, but he also saw that the two layers were so intimately joined
to each other that the canals were continued through both. While I
prefer to regard these two layers as substantially a unit, basing my con-
clusions on a variability in the apparent independence of the outer layer
and on the continuity of the pore-canals through both, I recognize that
this is a minor point, and that already Leuckart was in possession of the
important facts of structure. It was the presence of this “chorion” in
addition to Miiller’s capsule with its coarser pore-canals which convinced
Leuckart that the latter could not be considered the equivalent of the
radially striate membrane in other fishes, such as the salmon and trout.
I believe that Leuckart was less fortunate when he concluded that there
was in Esox a layer immediately outside the “ chorion”? which was homol-
ogous with the Miillerian “capsule” of the perch; for in my opinion
there can be no doubt that the layer in question is the same as that in
which Eigenmann has found the pore-canals to be continuous with those
of the deeper portion of the zona. Even Leuckart describes the canals
as straight, not spiral as in Perca. In my judgment, therefore, this layer
corresponds to the thin outer layer of the zona seen by Leuckart in the
perch, rather than to the capsular layer described by Miiller.

Although Volume V. (Supplementary Volume) of Todd’s Cyclopedia of
Anatomy and Physiology was not issued until 1859, the article “Ovum”
by Auten Tomson was published much earlier, and in two parts, ac-
cording to Gegenbaur (’61, p. 495). The first part (pp. 1-80) appeared
in 1852, and the second part (pp. [81]-[142]), which contains the por-
tion devoted to osseous fishes, in 1855.

Tomson (’59, pp. [99], [100], [103]), besides giving a very brief
summary of previous work on the subject, treated at some length the
structure of the zona radiata, basing his conclusions partly on the work
of Ransom and partly on his own studies. On the strength of Ransom’s
work he claimed that “the structure [Eikapsel] described by Miiller in
the perch was peculiar to that fish, and belonged only to an outer cover-

74 BULLETIN OF THE

ing superadded to the surface of the dotted membrane, which last re-

sembles in all respects that of other fishes.” ‘This outer covering,” he
adds, “ appears to be of cellular origin; and Dr. Ransom thinks it may
be due to the separation of the tunica granulosa along with the ovum.”
Thomson was able “to perceive the circle or lumen of the tubes” in
the dotted membrane by using a high magnifying power, and also
thought he could distinguish a hexagonal marking of the intervals be-
tween the pores (which he figures) in the salmon; he also pointed out
that the size of these pores was only about one third of that of the tubes
in the perch as described by Miller. The author was less successful,
when, in explaining Fig. 67*, he said: “A granular or dotted appear-
ance of these [granulosa] cells seems to indicate their conversion into
the dotted membrane, which is probably formed in successive layers from
the exterior. . . . The ovum (p. [103]) receives its firm porous mem-
brane [zona] while within the ovarian capsule, but only in the latter
part of the time of its formation.” The origin of the membrane he
is inclined to connect with the interior of the ovarian follicle; “but
whether by exudation from it, or by amalgamation of the innermost layer
of epithelial cells of the follicle,’ he has not been able to determine.
The latter he believes the more probable, and that the membrane is the
true vitelline membrane. I am entirely unable to comprehend how
Thomson could have reconciled the two statements in the last sentence,
for surely the vitelline membrane was not, even at that time, regarded
as the product of anything but the ovum itself.

On ripe fish-eggs ReicHert (’56, p. 89) was able to distinguish two
membranes, both of which were formed within the follicle. The inner
was the punctate membrane, but owing to the fineness of the markings
it was impossible to determine whether they were the result of elevations
or depressions of the surface. Reichert was unable to adopt without
reserve the conclusion that the dotted appearance is due to radial canals,
even though such an explanation was suggested by his finding the mark-
ings on the inner as well as the outer surface of the membrane in the
case of Cyprinus carpio. He evidently reposes great confidence in
Miiller’s explanation of the striation as an optical illusion, which in his
opinion accounts for the appearances figured by Aubert. He is also
uncertain in regard to the existence of a vitelline membrane, so that a
positive conclusion as to the nature of the zona radiata was not reached.
The smallest eggs possess a transparent homogeneous membrane without
punctations, and too thin to be measured, which may be regarded as a
vitelline membrane. With an increase in the size of the egg this mem-

MUSEUM OF COMPARATIVE ZOOLOGY. 15

brane becomes thicker, and in Acerina punctate markings become visible
on its outer surface when it is only about 2.8 » thick. There is never
any appearance which allows the supposition that the thickening is pro-
duced by secretions from the epithelium of the follicle. Since the punc-
tate appearance becomes visible only after the thickening of the original
membrane, it is to be concluded, says Reichert, that the punctate mem-
brane of the ripe egg is not the original vitelline membrane, but a
secondary egg-membrane, which, however, has been formed by the depo-
sition of thickened layers (‘‘ Verdickungschichten’’) of the egg outside
the vitelline membrane. But what has meantime become of the vitelline
membrane is not stated.

Reichert distinguished on the majority of the eggs which he studied a
second membrane outside the punctate one. In Esox it was clear, homo-
geneous, and viscid. In cyprinoids it had the velvety appearance already
described by Miiller, from whom the author differs in regarding this
membrane as not a part of the punctate one. His reasons for this con-
clusion are, that it is as sharply marked from the punctate membrane
as is the capsular membrane in the perch and may even be detached
after treatment with chromic or nitric acids, that the rod-layer is not so
firm as the punctate, and finally that the rods are much fewer than the
punctations of the same egg. ‘The rods are set in a clear, homogeneous
layer, only their rounded ends protruding.

Reichert was evidently influenced by his belief in the probability that
the capsular membrane in the perch owed its origin to the membrana
granulosa, and consequently he left unsettled the question of the origin
of this villous layer, although its intimate adhesion to the punctate
membrane indicated a common origin with the latter.

Kouuiker (’57, p. 197) found in 1856 that the porous membrane in
the case of sturgeons’ eggs presented conditions which favored the view
that it was formed by the cells of the follicle, but he discreetly abstained
from forming an ultimate judgment before eggs in all stages of develop-
ment had been studied. Such studies he found the opportunity of be-
ginning during a sojourn at Nice a few months later, and he continued
them on fresh-water fishes during the beginning of the following year;
the results were published in 1858.

KOLLIKER’s (58, pp. 80-93) observations embraced a large number
of fishes, and —— what is of more importance —he also studied eggs in
several stages of development. With Reichert, he recognized two cap-
sular egg membranes, which he called “ Dotterhaut” (zona radiata) and
“Gallerthiille” (the latter in Perca), but he dissented from Reichert’s

\

76 BULLETIN OF THE

explanation of the striate appearance of the former, and demonstrated

_ by means of thin sections that it was due to the presence of poré-canals.

Kolliker also claimed that there was an outer thin, resistant layer of
the porous vitelline membrane in the case of all fishes, —such as
Leuckart alone had recognized in Perca, — and that this layer might
retain the striate appearance even when the rest of the membrane had
been made pale and apparently homogeneous by the use of reagents.
It is in this outer layer that the viscid and fatty-looking “ Zéttchen”
are rooted with their slightly enlarged bases. Kolliker therefore re-
gards the “ Zéttchen” layer as belonging to the “ Dotterhaut,” rather
than as a separate layer, and compares its elements to the peculiar
appendages discovered by Haeckel in the Scomberesocidee. To these
he ‘adds a description of peculiar mushroom-shaped appendages of the
vitelline membrane in Gasterosteus and Cottus gobio. There is, how-
ever, one difference between the “ Zottchen” and the mushroom-shaped
bodies; in caustic potash the former are greatly swollen and become
pale, whereas the latter are made to shrink somewhat and to become
darker.

In reference to the origin of this membrane and its appendages,
Kdlliker gives the first positive information which we have, for Rei-
chert’s conclusions were at best only theoretical and tentative. Both
in the case of Gasterosteus and Cobitus barbatula he showed that the
villous structures made their appearance before the zona radiata. In
the case of the first-mentioned species, the mushroom-shaped bodies
were distinguishable as minute wart-like points resting on the outer
surface of a membrane so thin that it presented only a single contour.
As these wart-like structures continue to occupy the outer surface of
this vitelline membrane while the latter increases in thickness, it is not
to be doubted, he says, that the increase is due to deposits upon the
inner surface of the membrane. The warts continue to increase in size,
while the membrane becomes still thicker and shows radial markings.

The first appearance of the villi in Cobitis is to be seen in eggs
0.08/’ [0.08/24 = 175 »] in diameter, where they appear as deposits or
outgrowths on the external surface of a thin membrane (Reichert’s primi-
tive vitelline membrane) ; at first they are low and narrow, but they
gradually increase in length, and also, though more slowly, in breadth.
It is only when the villi have attained their full length that the porous
layer begins to be formed by deposits on the zxner surface of the thin
membrane, but this proceeds with such energy that the porous layer
soon exceeds in thickness the villous. At the same time the villi in-

MUSEUM OF COMPARATIVE ZOOLOGY. 17

crease in breadth, though not in length, and the thin membrane persists
as the outer layer of the porous membrane, which in the ripe egg bears
the villi. Upon the first formation of the villi the appearance of a sur-
face view of the membrane so closely resembles that of a porous vitelline
membrane (zona radiata) with fine close-set pores, that one must follow
the whole course of development, and convince himself of the late ap-
pearance of the porous layer before he can be certain that the fine points
are due to the villi.

Kdlliker believed that the formation of this peculiar “ Dotterhaut”
(zona radiata) of fishes could be easily understood as one of the so
called secondary cell secretions, if there were on the inner side of it
another membrane, which latter would then be regarded as the origi-
nal cell membrane of the egg. Although he found some evidence of
the existence of such a thin structureless membrane in Cobitis, he
was unwilling to give much importance to that fact, but inclined to a
belief in its existence, rather upon the theoretical ground that it would
offer a satisfactory explanation of the radial pores as being the equiva-
lents of pores in cuticular structures.

This assumption (p. 104) that the fine pore-canals are to be explained
as resulting from the presence of fine openings in the cell membrane
may be satisfactory enough for those cases in which a definite cell
membrane is demonstrable previous to the appearance of the cuticular
secretion ; but it seems to me superfluous to assume the universal exist-
ence of a cell membrane in order to explain the conditions. Where no
cell membrane exists, the same phenomenon may take place, and is no
more difficult of explanation than in the case where there is a cell
membrane with the supposed structure; for the latter must in its turn
be explained as the result of localized activity on the part of the cell:
protoplasm in secreting its membrane. So, in the end, it comes to one
and the same thing, whether we assume the presence of a cell membrane
or not: the explanation must rest on the ability of protoplasm to
localize its activities; but further than that we are at present unable
to advance. Why or just how protoplasm is able to effect such a histo-
logical division of labor is still unexplained.

The important paper by GucrnBaur (’61) on the structure and devel-
opment of the vertebrate ovum adds nothing to our knowledge of the
zona radiata in bony fishes, but is valuable for the way in which it
illuminates the subject of the vitelline membrane.

The only articles of much importance during this decade were one by
Buchholz and two by Ransom.

78 BULLETIN OF THE

Ratuke’s (61) posthumous work on the development of vertebrates
evidently treats the subject from the standpoint of thirty years before,
when little was known about the matter ; and

LEREBOULLET (761, pp. 120, 121, 123) does not greatly add to the
knowledge of the zona radiata when he says that the chorion of the
trout egg is thin and very soft at the moment the egg is laid, and does
not present the resistance and elasticity which it acquires after it has
remained for some time in the water. For the absorption of water and
the passage of gases necessary for the respiration of the egg and em-
bryo, the chorion is pierced, he says, by an infinite number of excessively
narrow, short parallel tubes, which give a striate appearance to perpen-
dicular sections of the shell.

OwENn (’66, pp. 593-595), although following the accounts by Ran-
som and Thomson, fails to recognize one of the important points estab-
lished by the latter author, for he does not distinguish between the
“ectosac”’ in the perch, and that of salmon and other fishes. More-
over, this “ ectosac”’ (evidently the zona radiata) “‘is composed of close-
set series of hollow columns.” (!) As Ransom (’67, p. 3) has since
pointed out, Owen also erroneously states, possibly under the influence
of Rathke’s exposition, that the villi are formed after the ova escape into
the cavity of the ovary. >

Bucuuouz (63, pp. 71-81, and 63%, pp. 367-372) was the first to
describe a very peculiar appearance of the egg membranes in Osmerus
eperlanus. In addition to the porous membrane (zona radiata), which
continues to invest the egg after it is laid, there is a second one external
to the first, and like it traversed by similar pore-canals. Buchholz states
that these canals are much more readily recognized to be pore-canals in
the case of this fish than in that of other fishes (p. 73). When the egg
has lain for some time in water, the outer envelope, or at least a portion
of it, is found to be attached to the inner membrane around the circum-
ference of the micropylar canal, whence it depends as a loose funnel-
shaped frill with its originally inner surface now directed outward. The
pore-canals which traverse these two membranes, instead of being cylin-
drical, are funnel-shaped, the wider end being directed outward. By
treating the fresh ovarian egg with acetic acid the owter membrane,
which at first lies closely in contact with the inner, is made to swell
up with irregular foldings until it becomes entirely separated from the
inner ; but a striate appearance, which is visible for a moment, becomes
quickly obliterated by the action of the acid. If, however, the eggs
are first treated for twenty-four hours with very dilute chromic acid,

MUSEUM OF COMPARATIVE ZOOLOGY. 79

and then with acetic acid, the radiate structure remains easily distin-
guishable in both membranes.

With regard to the formation of the two membranes, Buchholz argues
that the outer is the older, since one often finds, in the earlier stages of
their formation, that the inner is thinner than the outer, whereas subse-
quently they are both of equal thickness. The increase in the thickness
of the inner membrane was observed between the middle of February
and the middle of April, —the spawning time,— and meanwhile the
outer membrane was found to be thinner over about one third of the
surface. It is to be assumed, according to the author, that this attenu-
ated portion finally disappears altogether, since the persistent portion
which remains attached in the region of the micropyle is too small to
have completely enveloped the egg. Even nearly up to the time of
maturity there is no fusion between the two membranes, which must,
therefore, take place rather late.

The homology of the outer membrane in Osmerus is not at once
evident from this account by Buchholz If one were to accept un-
questioned his description, it would be most natural to regard it simply
as a detached portion of the zona radiata, for he maintains that the two
are identical in structure. There are, however, two other possibilities ;
it may be homologous either with the villous layer of Lepidosteus, or with
the capsular envelope of the perch. I regret that I have not yet found
the opportunity to acquire from personal study additional evidence in
favor of one or the other of these explanations; but there are two or
three things connected with the account given which incline me to
believe that the outer membrane is the equivalent of the villous layer.
The very fact of its becoming detached from the deeper layer and
thrown into folds after the egg has lain in the water suggests a similar
though less striking feature of the villous layer in Lepidosteus ; and al-
though there is no evidence that any such eversion of the membrane
takes place in the latter case, or that it eve: becomes regularly attenu-
ated on one side, as in the case of Osmerus, still I can imagine that a
similar condition might be artificially produced in Lepidosteus, so far at
least as regards the peeling off and eversion of the outer covering, and
it is possible that a slightly different physical condition of the villi
would cause them to adhere to each other so persistently as to allow the
attenuation of the whole membrane on one side of the egg without
separating the individual elements. Since Buchholz asserts that the
pore-canals are more readily distinguished as such in this fish than in
other instances, I infer that the markings which he observed must have

ae

80 BULLETIN OF THE

been coarser than is usual. If his attention was mainly directed to the
‘outer layer, and if, as I imagine, this is a villous layer, the reason of his
statement would be obvious. Besides, he mentions that the canals are
not cylindrical, but funnel-shaped, the wider end outermost ; this, too,
though suggestive of the capsular membrane, would be entirely com-
patible with the idea that he had under view club-shaped villi. And
finally, the argument to show that the inner layer is produced after the
formation of the outer is exactly applicable in the case of the villous
layer, as the observations of Kolliker and my own conclusively show.

Ransom’s (’67) account of the “yelk-sac ” (zona radiata) in Gas-
terosteus is principally interesting to me from his asserting that the
pore-canals as well as the villi increase in number during the erowth of
the egg, and from his consequent conclusions as to the method in which
membranes are formed. The “yelk-sac,” he says, is formed in very
young ova (s}5", or 125 p, in diameter), in which it is easily recognized
by the button-shaped villi attached to the outer surface surrounding the
micropyle. The finely dotted structure is first discoverable in eggs 715!
(180 ») in diameter, and it is the same in character in these as in the
ripest eggs. The membrane is composed of very fine concentrically
arranged lamin, each of which is marked by dots of equal size, so
arranged as to mark (in surface view) the angles of equal-sized lozenge-
shaped spaces, and corresponding in position in the successive laminze so
as to form vertically placed lines or striz. In eggs .01 inch in diam-
eter there were about 24,000 dots to the [linear] inch, and when the
egg had attained .06 inch there were 11,000 to the inch; the distance
between dots being scarcely more than doubled, while the diameter of
the egg had been multiplied about six times. From this the author
argues that there must have been an increase in the number of the dots
during the growth of the sac, and therefore that the membrane does
not increase by apposition of layers either from the inside or outside,
either by the hardening of an exudation or by the conversion of the
substance of the yolk into that of the yolk-sac. “It grows in some
way by interstitial molecular deposit.” A similar increase in the num-
ber of the button-shaped villi was also observed to occur during the
development of the ova.

I do not recall that any one has corroborated or disproved these ob-
servations, or the deductions made from them; but I have shown that
in the case of Lepidosteus there does not seem to be sufficient evidence
to prove that there is any increase in the number of the will. I believe
that a careful investigation of the question in the case of the pore-canals

MUSEUM OF COMPARATIVE ZOOLOGY. 81

would well repay one who should undertake it, for it could not be without
influence upon theories on the method of the growth of membranes.

In his more extended paper Ransom (68) gives an account of the
egg membranes of a number of fishes, but more particularly of Gaste-
rosteus, Esox, and Perca. To what he had already stated about the
“yelk-sac ” of Gasterosteus he adds (pp. 440, 444, 448) that it is diffi-
cult, if not impossible, to determine the precise period at which it is
formed. In eggs g},/! (31.5 yw) in diameter it is not found, but is prob-
ably indicated by the smooth, hard outline which the yolk shows on its
surface. ‘It is separable in eggs 5$,y/’ in diameter, and may be seen in
the fluids on the slide as a homogeneous-looking collapsed sac.” With
a power of 500 diameters the dots appear round, and with one of 3,000
they are but obscurely hexagonal ; they are the same distance apart on
the inner surface as on the outer. Besides these minute regular dots,
there are larger and darker ones of a stellate form, which the author
suggests may in some way be connected with the interstitial growth of
the membrane. They occur irregularly at intervals of about 3,455’, and
act like bodies of low refractive power ; “‘at the cut edge they may be
seen to pass radially about two thirds into the substance of the yelk-sac,
gradually coming to a point and ceasing.” I am not aware that any
other observer has confirmed this appearance, which I imagine may be
due to the presence of protoplasmic prolongations of the yolk into some
of the pore-canals, just as in Lepidosteus the substance of the villi is
traceable in many cases for some distance into the zona, although from
the opposite direction.

‘There are no facts known to me,” says Ransom, “to point out
whether the pabulum for the growth of this membrane is derived di-
rectly from the currents passing inwards, or from the material elaborated
in the egg and passing out of it, or from both sources indifferently.”

Watpeyer (’70, pp. 80, 81, 83), however, did not experience any such
uncertainty concerning the source of the zona radiata, for while he con-
tinued to call it the “‘ Dotterhaut ” and could find no vitelline membrane
inside it, he was very explicit in stating that it was a cuticular formation
produced by the follicular epithelium, and that the pore-canals were occu-
pied by delicate protoplasmic filaments which were in direct connection
with the epithelial cells of the follicle on the one side, and with the finely
granular yolk substance on the other. In his general conclusions con-
cerning the eggs of vertebrates he says: ‘The complete homology of the
zona pellucida [mammals] with the vitelline membrane of other verte-
brates can... no longer be denied. The vitelline membrane is cer-

VOL, X1x.— No. 1. 6

ee ee ere

ae ee

82 BULLETIN OF THE

tainly a structure which does not belong to the primordial egg, but is
deposited upon it from without.”

Like Waldeyer, Ermer (’72°, pp. 417-428, Taf. XVIII. Figs. 9-13)
regards the zona as a cuticular product, but, unlike him, he maintains that
it is produced (precisely as in reptiles) by the egg, not by the granulosa.
But the homology between the egg membranes in these two groups goes,
in his opinion, still further. The delicate membrane described by Au-
bert and others as external to the zona in Esox and other fishes is to
be regarded as a chorion, — which in reptiles is produced by the fol-
licular epithelium. The trumpet-shaped structures of the outer mem-
brane (Kikapsel) in the perch are formed from granulosa cells. They
are the homologues of the trumpet-like structures which the author
described for Coluber and placed in the category of ‘beaker cells.”
The villi which repose on the outer surface of the zona are, as held
by Reichert, to be regarded as belonging to the second membrane, and
not in accordance with Kolliker’s views as the outer layer of the zona
itself. In Eimer’s opinion they are simply yolk substance which has
emerged from the egg through the pores of the zona. Eimer even main-
tains that he has observed the protrusion and subsequent. disappearance
of such villi.

I have already (p. 54) given my reasons for dissenting from the use
which His (’73, pp. 1-3) makes of the word “ Hikapsel” to signify the
zona radiata. According to His, the “capsule” in Salmo salar is from 33
to 35 » in thickness, the pore-canals are straight, only 1.5 to 2 w apart,
and not funnel-shaped at either end. In Esox it is 16 to 17 mw thick,
and has fine parallel (concentric) as well as radial markings (p. 13).

His (’73, pp. 17, 35, 36, et passim) also discovered that in numerous
cases the young eggs possessed a peripheral layer of clear substance
which exhibited fine radial markings after treatment with certain acids.
This did not appear to be constantly present, nor necessarily of uniform
thickness. He called it the zonozd layer, and thought there was proba-
bly a physiological connection between it and the porous egg capsule,
but what the nature of that connection was, remained to be ascertained.
Nothing in either figures or text allows one to draw conclusions as
to the relative fineness of the striations in the zonoid layer and the
“capsule.”

In 1878 appeared three papers which dealt with the structure and
development of the egg membranes of bony fishes.

A pupil of Waldeyer, Kotussyixow (’78, pp. 402, 403, 407-409),
states that in both Perca and Gobio the ‘‘ Dotterhaut ” consists of two

MUSEUM OF COMPARATIVE ZOOLOGY. 83

membranes, which are not easily separable either in the fresh state or after
treatment with osmic acid; but on thin sections of eggs hardened in the
latter reagent or potassic bichromate the two are sharply marked. This
condition is held to indicate a gradual growth of the striate Dotter-
haut, “und zwar ihrer dusseren Stibchenschicht,” which is the youngest
formation, the ener layer being the older. Both layers consist of rods
radially placed side by side, those of the inner layer being much finer,
longer, and nearer together than the outer. It is evident, I believe,
that the inner layer referred to must be the zona radiata, and the outer
the “ Kikapsel”’ of Miiller. Fine granules, which are colored black by
the osraic acid and are regarded as yolk granules, are found in the fine
processes between the rods, as well as in the follicular epithelium itself.
When the ovary hardened in alcohol is stained in hematoxylin, the
inner surface of the follicular epithelium resembles ciliated epithelium,
since the rods rest like cilia on the epithelium. On sections of a very
young follicle (4.65 ~) the membrane, composed of rods, can be seen on
the inner surface of the follicular epithelium. It is still thin and not
easily detached, and the rods are fine and close together. When the
follicle has reached a diameter of 465 », the membrane is much thicker,
and composed of two layers. It is then to be seen that in some places
the rods of the outer layer are continued into the inner layer, being
consequently longer. In this case the rods gradually diminish in thick-
ness as one passes from the follicular epithelium toward the yolk sub-
stance, but they are always sharply limited from the latter. Both layers,
the author concludes, are cuticular formations of the follicular epithelium,
and in no case is the inner layer to be regarded as a special membrane
produced by the egg itself.

One of the best contributions recently made to this subject is that of
Brock (’78, pp. 547-559), who gives, besides a condensed summary of
previous observations, his own valuable results. Aside from the capsular
membrane of the perch, which he calls “‘ Gallertkapsel,’”’ and the related
structures of other fishes, the author is able to recognize only one egg
membrane, which, to avoid prejudging its genetic relation, he calls, from
its most evident morphological peculiarity, the zona radiata, reserving
the term membrana vitellina for egg envelopes, which are the equivalent
of other cedd membranes and therefore differentiations of the [egg] proto-
plasm. The outer lamella of the zona described by Reichert and by
Kolliker he finds with varying distinctness in different cases. Being in
some instances unable to discover it at all, he doubts its constancy. It
may often be demonstrated by the use of acetic acid when not visible

a SE ATE YN ATE er te ret
a

84 BULLETIN OF THE

in the fresh condition. In the perch it has a striate appearance, and
is much more coarsely marked than the true zona,’ but in Serranus
hepatus it is homogeneous. ,

The elongated club-shaped villi resting upon the outer surface of the
zona were found in many cyprinoids, and also in Osmerus. They are
not, as asserted by Eimer, expressed droplets of yolk substance ; they
are secondary appendages of the zona, which “have nothing whatever
to do with either follicular epithelium or yolk.” Brock finds the zonoid
layer of His well developed in Alburnus lucidus, Salmo fario, and Perca
fluviatilis. He is inclined to regard it as a general structural condition,
often overlooked because distinctly shown only at certain periods in the
development of the egg. It is often divisible into two layers, of which
the inner remains homogeneous. When the zona radiata is removed
from the yolk, the “zonoid” remains attached to the latter, to which it
must therefore belong. Its striations are intermediate in fineness be-
tween those of the villous layer and those of the zona. Notwithstanding
certain objections, Brock regards the follicular epithelium as the princi-
pal, if not the exclusive, source of nutrition and growth for the yolk,
which are accomplished by means of cell processes sent by the granulosa
through the zona into the yolk. The evidence, aside from that which
Waldeyer produced for other groups, is to be found, says Brock, in the
fact that, when the granulosa is separated from the zona by a secondary
membrane (Perca, Serranus), it sends processes through the latter which
are traceable up to the zona.

According to the opinion which I have formed, however, in regard to
these so called processes, they are not outgrowths from the granulosa
cells; on the contrary, the cells, retaining their original contact with
the zona, are by the accumulation of the capsular secretion greatly
attenuated.

In regard to the order in which the different membranes make their
appearance, Brock comes to views diametrically opposite to those ex-
pressed by Kolliker. Leaving out of consideration the capsular mem-
brane, which, as Brock rightly states, is late in being formed, and the
outer lamella of the zona, concerning the origin of which he could dis-

1 It is not possible to say with certainty from his figure (Fig. 7, f) whether
Brock regarded the striations of the outer lamella as less numerous than those of
the rest of the zona. They are represented as broader than the latter, and as
thickest at theiryperipheral ends, which agrees with Eigenmann’s observations ;
but they certainly are not represented as continuous with the striations of the inner
portion, and in this I believe that Brock is in error.

MUSEUM OF COMPARATIVE ZOOLOGY. 85°

cover nothing, he asserts that he saw in all cases with the greatest
distinctness that the zona radiata first appeared, and that when it
had attained a certain thickness then for the first time the villi and
the zonoid layer made their appearance, almost simultaneously. Un-
fortunately, Brock has given no details concerning the proof of this
assertion, either in figures or description. I am tolerably certain, not-
withstanding the very positive way in which he maintains his conclu-
sion, that Kolliker was right, and that he is wrong, for I cannot believe
_that in so fundamental a matter there is such a difference between
fishes as would be implied by admitting both views to be correct.

That the differences of opinion which the egg membranes have given
rise to are not exclusively due to the study of different fishes, is clearly
seen from the results reached by Kupffer, and soon after by Hoffmann,
in the study of the herring’s egg. Kuprrer (’78*, pp. 177, 178) found
the yolk to be closely invested by an egg capsule 6 to 8 w in thick-
ness, and the latter to be covered by a layer of viscid semifiuid sub-
stance, which was found to be of nearly uniform thickness if the eggs
were dropped into alcohol without contact with water. In water it soon
becomes solid. The capsule consists of two firmly united layers, the
inner one being finely striate radially, and alone equivalent to the po-
rous capsule (zona radiata) of other eggs. The striation is due to pore-
canals. These do not traverse the outer layer, which has concentric
striations. A boiling ten per cent solution of potash dissolves the po-
rous layer quickly, but does not affect the outer layer of the capsule,
which is believed to be the same as that described for Esox by Aubert,
and fer other fishes by Kolliker. This difference in the two layers
Kupffer regards as favoring Eimer’s view, that one is produced from
the egg and the other from the follicular cells.

On the other hand, Horrmann (’81, pp. 15-33), who has given the
subject of egg membranes in fishes the most extensive treatment of any
recent writer, differs materially from Kupffer in his account of the her-
ring, although he. offers a partial explanation of their differences in
saying that Kupffer examined only fully mature eggs, and such as had
been in contact with sea-water. Hoffmann makes the total thickness of
the membranes in not quite mature eggs to be 32.5 u. The outer layer,
10 w thick, is not separable from the inner, although a dark line marks
the boundary. Both layers are traversed by numerous pore-canals,
which, to judge from his figures, appear to be finer (not necessarily
nearer together) in the inner than in the outer layer ; but whether they
are continuous it is impossible to say on account of the sharp line sepa-

86 BULLETIN OF THE

rating the layers. The inner layer is composed of two parts, the outer

part being often striated concentrically. The membrane of the fully
ripe egg is different. Before contact with water the pore-canals of the
outer layer are not so easily distinguishable as in immature eggs, and in
this layer a great number of small lustrous spherules are now visible.
In eggs which have been in contact with sea-water the outer layer is
raised up from the inner, forming a viscid sheet 10-12 » thick, which
causes the adhesion of the eggs. The outer portion of the inner layer
now exhibits a tendency to split into concentric layers, which obscures
the radial pores, although they are still visible in the deeper part of
this portion.

It appears to me probable that the two portions of the inner 1a
represent the whole of Kupffer’s capsule, and that Hoffmann’s outer
layer is the equivalent of Kupffer’s viscid semifluid substance. In view
of the striate appearance of the outer layer described by Hoffmann, and
the greater coarseness of the striations as compared with the inner layer,
and also in view of its viscid nature, I am strongly inclined to believe
the outer layer will ultimately be shown to be equivalent to the villous
layer in other bony fishes. :

Two layers resembling those of the herring were also found in Creni-
labrus. Greater interest attaches, however, to the account of Leuciscus
rutilus, in which, as the author says, one again finds the two layers of
the zona radiata. But in the next sentence he shows that he does not
distinguish sharply between zona and villous layer: ‘‘The outer layer
here forms the well known Zéttchenschicht.” The villi are club-shaped,
close set, and clothe the zona as a uniform layer, with the exception of
the place where the micropyle is situated. In the extent of their distri-
bution, therefore, the villi in Leuciscus differ from those of Lepidosteus,
— which, though shorter, are not wanting in the periphery of the micro-
pyle, —and also present a condition which is the complement of that
exhibited by Heliasis, Gobius, etc., in which villous structures are
restricted to the region of the micropyle.

Hoffmann arrives at these general conclusions : In adhesive eggs the
zona radiata consists of two layers, of which the outer effects the ad-
hesion. ‘The latter may form a part of the whole zona, or may exist in
the form of villi over the whole surface, or of long filaments limited to the
micropylar region. “But whatever form this outer layer may assume,
it always has a like origin with the rest of the zona; it is nothing else
than a part of the zona itself, which sooner or later undergoes peculiar
metamorphoses.” Hoffmann recognizes the difficulty of determining how

‘

MUSEUM OF COMPARATIVE ZOOLOGY. 87

the zona radiata arises, whether as a product of the yolk or of the granu-
losa cells. An examination of maturing eggs shows him that its periph-
eral layers are always the most distinct. It is as though new layers
were being deposited from within, and this leads to the conclusion that
the zona radiata represents a true vitelline membrane.

Although Hoffmann seems to me to come very near the true solution
of the problem, this presentation of the matter appears altogether un-
satisfactory, because he insists that what he calls the zona is practically
a unit in structure, and fails to recognize a fundamental difference be-
tween the outer villous layer and the “true zona radiata,” as he terms
the inner portion of his zona. From this last conclusion one would be
led to infer that the zona (including both layers) was the result of a con-
tinuous process, and that there could not be any radical structural differ-
ence between villous and porous layer. But to my mind a common origin
from the yolk by no means implies identity of structure, nor the contin-
uous operation of the same formative process.

Ryper (’81°, 82°, 83) has contributed a good deal to our knowledge
of the occurrence and functions of those modified forms of villi first
described by Haeckel, but I think he cannot have given their structure
very close attention or he would not have said (83, p. 195) that “they
are apparently composed of the same tough material as that which enters
into the formation of the egg-membrane itself.” He (’81°, pp. 137, 138)
regards it as probable that: the egg membrane (zona radiata) is secreted
from the cellular walls of the follicle.

Since writing the above summary of and comments on the observations
of authors there have appeared a number of papers, some of them of con-
siderable importance, which I was unable to utilize in forming my opin-
ions of the nature of the membranes in Lepidosteus and other fishes.

Stockman’s (’83) account of the egg capsule in Salmo describes the
appearances of the pore-canals under a Reichert 34 homogeneous im-
mersion lens. In sections the limits of the pore-canals appear toothed,
owing to the presence of minute folds which have for the most part a
direction tangential to the capsule. These extend to the two ends of the
canal, and consequently its mouth appears angular rather than circular.
The substance of the capsule is beset with minute spaces which commu-
nicate with the pore-canals between the tooth-like projections, and are
believed to have a function in the transmission of nutritive material to
the egg.

Ryprr (’84, pp. 3, 4, 11, Plate I. Fig. 5) states that the cod’s egg is

88 : BULLETIN OF THE

“covered by a vitelline membrane which is not porous or enveloped in
adhesive material. It is thin, very transparent, and laminated, as has
been stated by Sars, and at one point is perforated by a single minute
opening, the micropyle.” Ryder was unable to discover the lamination
until after the action of osmic acid, and was uncertain whether it was
a natural condition or the result of the action of the acid. “The cod’s
egg,” he says, “is without the zona radiata found enclosing the egg
proper of the shad, whitefish, and sculpin, and, inasmuch as it is unques-
tionably true that a micropyle perforates the zona in a number of these
cases, it does not appear that sufficient grounds exist for the declaration
that a micropyle perforates the zona radiata alone, in the face of the fact
that the vitelline membrane only is perforated in this one instance.” I
have no doubt that the membrane in question possesses pore-canals, and
that it is therefore a true zona radiata. I can confirm Eigenmanfi'$
observations in this particular, and believe that Ryder himself would
have come to the same conclusion had he observed the membrane under
the same favorable conditions.

In an extensive paper on the eggs of bony fishes Owssannrkow (’85)
describes the egg membranes of a number of the more common fresh-
water forms. The most important part of his paper deals with the cap-
sular membrane in Perca and the equivalent structure in other fishes;
but the consideration of that part will best be deferred until I come to
review the other papers which deal with that subject. I may here add,
however, that he does not recognize the existence of a villous layer
outside the zona, but regards the structure which immediately envelops
the zona in Osmerus and other fishes as the equivalent of the capsular
membrane of the perch. His description of a thin transparent mem-
brane (membrana vitellina) inside the zona radiata in Salmo trutta is
materially affected by the subsequent statement that it is not found in
other cases (Lota, e. g.), and by the admission that it may have been
an artificial product.

The pore-canals of the zona are often more deeply stained than the
substance of the matrix, and by treatment with certain reagents minute
points can be seen in the canals when highly magnified. In Lota vul-
garis the zona is very thin, and the pore-canals in patches do not pene-
trate to its inner surface. It is generally stratified, the strata being
laid down successively and all being perforated. The zona might, in his
opinion, better be called perforata than radiata. Concerning its devel-
opment in Gasterosteus the author says that the first trace of it is
seen to be a very thin membrane without any pores. These appear when

MUSEUM OF COMPARATIVE ZOOLOGY. 89

the zona has become thicker, but they are much finer than in the mature
egg. After describing the condition in Acerina vulgaris (p. 18), the
following statement is made as embodying the author’s idea of what
takes place in the formation of the zona. The granulosa cells secrete a
substance (Zwischensubstanz), which surrounds the egg, one layer upon
another. ‘The pores in this substance arise by the growth into it of the
points of the granulosa cells, or plasmatic processes from them. The
way in which the author describes the mushroom-shaped villi which
surround the micropyle in Gasterosteus shakes one’s confidence in this
part of his work. It is very probable, he says, that they are derivatives
of the granulosa cells. After admitting that they have been described
in a masterly way by Kolliker, he gives an interpretation of them that
must appear to that author very remarkable. They are nothing less
than individual cells (!), each with a nucleus that is stained red in
carmine while the cell protoplasm resists for a time the action of the
dye. From the base of the cell emerges a thread which can be traced
into the zona. Very young eggs possess these appendages, as Kolliker
maintained, but they are much smaller, and consist essentially of a
nucleus which is attached by a thread and surrounded by a thin film
of protoplasm.

If everything that is stainable is to be called either nucleus or cell
protoplasm, then the current notions of what constitutes a cell will
have to be abandoned ; that will give room, it is true, for the admission
into this category not only of the villi and the yolk cells of His, but also,
I fear, of many other structures as well.

The author’s views concerning the villous layer in other cases appear
from his account of Coregonus. The granulosa cells send processes into
the zona radiata. When they are removed from the immature egg, the
zona appears to be covered with small more or less pointed “ Zottchen,”’
which are therefore to be regarded as processes of the granulosa. Out-
side the zona radiata, Owsjannikow finds in many cases a thin viscid
layer, which he suggests is derived from the oviduct in the case of
S. trutta, but looks in Lota as though it resulted from the fusion of
endothelial [ granulosa ?] cells.

SoieeR (85, pp. 330-332) is evidently inclined to bring the villous
layer into connection with the presence of intracapsular corpuscles found
in the perivitelline fluid. Without committing himself unreservedly to
the views of Eimer, — that the villi are simply exuded drops of vitelline
substance, —he confirms the statement that they are not of uniform
length or shape in the case of Leuciscus rutilus, and says that Eimer

Ee ii

90 BULLETIN OF THE

must certainly have the credit of having especially emphasized the fact
that after a certain epoch there exists a contrivance which prevents
the further entrance of water into the intracapsular space.

I believe a comparison of the conditions in Lepidosteus with the early
account by Kolliker will convince the author that the villi have no con-
nection whatever with the interesting conditions of the perivitelline
space which he has discussed.

Ryper (86, pp. 18, 23, 30, 35, 36, and ’87) has recently noted the
existence of a zona radiata in several species of fishes, but without
having given the structure special attention. In some cases the eggs
when laid are covered with an adhesive material, the source of which
is not alluded to. In the case of Ictalurus albidus (white catfish),
there is an interesting condition of the egg membranes which he (’86,
p-. 47, and ’87, p. 535) describes as follows: “The egg membrane is
double, that is, there is a thin inner membrane representing the zona
radiata, external to the latter and supported on columnar processes of
itself which rest upon the inner membrane ; there is a second one com-
posed entirely of a highly elastic adhesive substance. The columns
supporting the outer elastic layer rest on the zona and cause the outer
layer to be separated very distinctly from the inner one. . . . This pe-
culiar double egg membrane, with a well defined space between its inner
and outer layers, is highly characteristic, and bears no resemblance to
the thick, simple zona investing the egg of Alurichthys, nor has any-
thing resembling it ever been described, as far as I am aware, in the
ova of any other Teleostean.” Eigenmann (’90) has attempted to show
by comparisons that the whole of this double membrane is probably a
true zona radiata, and that the columns are protoplasmic substance
which occupied the pore-canals before the separation of the two por-
tions of the zona; but it seems to me more probable, from the “ highly
elastic adhesive’’ condition of the substance of the outer membrane,
that it corresponds to the villous layer in other fishes. There can be
no certainty, however, as to the real homology of the outer membrane
until it has been subjected to a more careful study with especial
reference to the time and manner of its production.

CunnincHam (’86) has recently rediscovered what Buchholz found out
upwards of twenty years ago about the peculiar egg membranes of
Osmerus. It is unfortunate that Cunningham overlooked the valuable
work of Buchholz, and the more surprising since he refers to Owsjanni-
kow’s paper, — which I should suppose he must have consulted, — in
which Buchholz is cited. Cunningham gives figures and an account

—_—

‘

MUSEUM OF COMPARATIVE ZOOLOGY. 91

which practically confirms that of Buchholz as to the eversion of the
outer membrane. According to Cunningham it serves as the so called
suspensory filament by which the deposited eggs are attached. He calls
-the outer membrane the external zona, and agrees with Owsjannikow
that it is traversed by pores which are larger and (according to his
figure, about four times) farther apart than those of the internal zona.
He makes no mention of the outer membrane differing in any other way
from the inner, except that it is somewhat thinner.

Scnarrr (’87, ’87*) regards Beddard’s zona radiata’ in Lepidosiren as
a “zonoid ” layer, which he also finds present in Trigla gurnardus, where
it has a thickness of 25 pw, while the true zona is only about 8 yw thick.
“ Both layers are striped, i. e. provided with minute radial pores,” which
are apparently continuous through both layers. The zona is firm,
granular, and stains deeply; the zonoid is semifluid, usually devoid of ,
granules, and stains only slightly. In ripe ova the latter disappears
entirely. The zona is formed before the zonoid layer. In Blennius
pholis there is no zonoid layer. Scharff “has no doubt that the egg
membranes originate from the yolk,” although, as far as I can under-
stand, he advances no new arguments to prove the fact.

Hennecuy (88, p. 419) says that, notwithstanding the use of high
powers, he has been unable to find in the trout the three layers of
Owsjannikow or the denticulations described by Stockman.

A recent paper by CunnincHam (’86%) deals with the interesting
question of the structure and origin of the egg membrane in Myzxine
glutinosa. Even when the eggs have become 11 mm. long there is no
trace of any membrane between the yolk and the single layer of granu-
losa cells, which latter project irregularly into the surface of the former,
Sections of an egg 16 mm. in length showed that the granulosa had
become several cells deep, though not arranged in regular layers, and
that there was beneath this, and in direct contact with the yolk, a thin
homogeneous membrane.

It certainly is an interesting fact, to which Cunningham calls atten-
tion, that the epithelium is thicker at the poles than at the equator of
the elongated ovum, and that the thickness of the membrane varies
[directly] with that of the epithelium. W. Miiller also observed the
same fact.

In eggs 20 mm. long the follicular cells are elongated, and form prac-
tically a single-cell layer, with the nuclei at the ends nearest the yolk.

1 Compare Beddard (’86*), and the summary of his paper given on pp. 66, 67.

92 BULLETIN OF THE

The membrane at the poles presents thickenings which are shown by
the later stages to be the beginnings of the peculiar polar filaments of
the egg membrane. These thickenings project into corresponding de-
pressions of the granulosa, the cells of which are so thinned at these
places as to be barely discernible. When the egg has increased slightly
in size (21 mm. long, 7 mm. in breadth), these thickenings affect the
external appearance of the follicle. They have the form of finger-shaped
processes covered with a single layer of granulosa cells, and project
into the connective-tissue wall of the follicle. The statement that in
the thinnest places (i. e. over the tips of the projections) the granulosa
is only .02 mm. thick, while its thickness at the exact pole of the egg is
.5 mm., is not borne out by the figure, and I therefore suppose that it is
a typographical error for.2 mm. In either event, however, the granu-
losa does not regain its normal thickness over the ends of the rapidly
growing filamentous projections of the membrane, and this fact may
have significance in determining the source of the membrane. The
author says that the membrane, if prepared in chromic acid, appears
homogeneous even when highly magnified ; but, if hardened in Perenyi’s
fluid, striz perpendicular to the surface are to be seen with a high power.
The strize are not represented as rigidly straight and parallel ; they may
even branch, and are often moniliform. They have been traced to the
outer surface of the membrane, where the author believes they are con-
tinuous with fibrils of the epithelium. He is also convinced that the
membrane is homologous with the zona radiata of teleosts. I cannot
agree with him, however, in the statement, “ When there are two zone
radiate, as in Perca fluviatilis and Osmerus eperlanus, according to Ows-
jannikow, these seem to be simply parts of one membrane differentiated
in physical properties, but essentially similar in structure.”

Cunningham believes it more probable that the zona is produced by
the follicular epithelium than by the outermost layer of the yolk, and
in the following manner: ‘‘ The deeper part of the elongated epithelium
cells is gradually changed into the zona radiata, the substance of the
cells being partly transformed into the substance of the membrane,
while threads of protoplasm, at more or less regular intervals, remain
unchanged, and thus give rise to the pores of the membrane.”

Against this view I would urge that a metamorphosis of the epithelial
cells, especially if prolongations into the membrane occur at intervals,
would be likely to result in the closest union between the membrane
already formed and its generatrix, whereas it is exactly along this line
that the artificial separation, which the author notes in all his eggs,

i

MUSEUM OF COMPARATIVE ZOOLOGY. 93

takes place. Moreover, a diminished thickness in the epithelial cells
corresponding to the most rapidly increasing part of the membrane
(the filaments), not at the close of the period of their formation, but at its
beginning, is the reverse of favorable to his “ hypothesis.”? And, finally,
I would suggest that the zona is everywhere in contact with the sub-
stance of the ovum, and that the increased thickness of the membrane
at the poles may be due to the accumulation there of a greater pro-
portion of the active protoplasm than is found at the surface elsewhere.
Perhaps it might be urged against this view that the explanation would
be only partial, —that, while it might account for an increased thick-
ness of the membrane at the formative pole, it would leave the con-
dition at the opposite pole unaccounted for, and therefore could not fully
satisfy the needs of the case.

I have recently sectioned the eggs of a rather poorly preserved speci-
men of Myxine australis, with a view to getting additional light on this
question. Although the eggs had attained a considerable size, — 22 mm.
long by 6 mm. in diameter, — still there was as yet no indication of the
filamentous projections; in fact, 1 could not trace a membrane continu-
ously around the egg. At the formative pole there was unquestionably a
membrane about 3.5 u thick ; it was faintly marked like the zona radiata
of teleosts, and it presented a deep micropylar infolding, with a cellular
epithelial plug. Nevertheless this membrane gradually grew thinner
in passing from the formative pole, until it could no longer be recognized.
It had about the same extent as the protoplasmic cap. At this stage
there was no more accumulation of protoplasm at the opposite pole
than at the sides of the egg; but there also was no more evidence of a
zona radiata at this pole than at the sides of the egg. That is all I
can offer at present in reply to the possible objection which I have
suggested. If it could be shown that the zona is developed at the
nutritive pole of the egg without the presence of an accumulation of
protoplasm there, and that the granulosa is more highly developed there
than on the sides of the egg, I should admit that a strong case would be
made against the view I defend.

1 Tt is true the author has offered an explanation of this; viz. that the filaments
are formed from the cells at the sides of the process, where the epithelium is very
thick, and that they are pushed up by the growth at the base. But I should
imagine it would be difficult to explain how secretions from Jateral cells could do
anything more than increase the diameter of the process.

94 BULLETIN OF THE

2. Capsular Membrane.

As I have already pointed out, the capsular membrane, since it was
first described by Miiller in the perch under the erroneous supposition
that it was the same as the zona radiata of the salmon, has often been
confounded with that membrane. In looking over the literature on the
egg membranes of fishes, after I had worked out the structure of the
villous layer in Lepidosteus, I was forcibly impressed by the resemblance
of that layer to the descriptions that had been given of the outer mem-
brane in the perch, and at first thought they might be‘ homologous
structures. It was particularly the account given by Ransom (’68,
p. 455, Plate XVI. Figs. 30, 31) of the root-like prolongations of the
tubules in the capsular membrane which suggested comparison. It
therefore seemed necessary to examine carefully all that had been
written on the egg capsule of the perch. The result has not confirmed
my first supposition.

Mutter (54) himself gave an excellent account of the structure of
the capsule, and accurately formulated the most interesting question
concerning its morphological significance. He described the egg enve-
lope (capsular membrane) as about 0.11 mm. thick; its outer surface
as covered with six-sided facets, which average 19 » in diameter. Each
facet contained in its centre an open funnel, which was continued into
a vertical tubule as long as the thickness of the capsule, and from 2.2 p
to 4.7 » in diameter. In fineness these were comparable to dentinal
tubules. They terminated on the inner surface of the capsule in fun-
nel-shaped enlargements, just as they did on the outer surface. Upon
eggs that had been boiled or hardened in chromic acid, it was possible
to see that the tubules had a spiral course, but they also appeared
narrower (1.1 ») than in the fresh state. The tubules were filled with
a thickish (albuminous?) mass, which in the fresh egg was clear, without
deposits, and under pressure projected from the funnel like a rounded
stopper or cylinder, but appeared to be coagulated by boiling and treat-
ment with chromic acid. When one compressed the fresh egg to burst-
ing, the oily substance of the yolk might be pressed into and through
the tubules; thus was effected a delicate injection which might greatly
distend them. Between the tubules, however, there was nothing pressed
out, which proved that on its deeper surface between the tubules the
capsule was closed. In the inter-tubular portions of the membrane,
after the eggs are hardened, there were to be recognized, besides a
gelatinous nearly invisible material, exceedingly delicate projections or

MUSEUM OF COMPARATIVE ZOOLOGY. 95

filaments placed alternately and running across between adjacent tu-
bules. These were thickest next to the tubule, and rapidly tapered to
very fine threads. In his opinion, the method of the formation of the
tubules might be made out during the winter. “ Zhe question is, whether
each of the tubes arises from a single cell, which becomes open, or whether
the tubes are originally inter-cellular, and whether ther walls result from
the remnants of several cells in contact with each other.”

A similar condition is maintained by Miiller for Acerina vulgaris, but
the membrane in this case was much thinner and the tubules conse-
quently shorter.

LEREBOULLET (54, pp. 242, 246) also discovered independently, per-
haps even before Miller,’ that there were in the perch what he called
hollow closely interlaced piliform appendages (also called stiff curved
filaments), which traversed the whole thickness of the shell, and to which
he attributed the agglutination of the eggs into a network. He also saw
besides these the much finer pore-canals.

In regard to the chemical nature of the capsular membrane, it was
maintained by Von Barr (’35) and by Levoxarr (55, p. 260) that it
was an albuminoid substance. KOLLIKER (58) called it gelatinous, but
His (’73, p. 15) proved that it at least closely resembled chondrin, and
consequently claimed the right to call it a cartilage capsule.

ReIcuERT (’56, p. 93) was not able to add much to Miiller’s account of
the structure of the capsular membrane. Concerning its origin he was
at first inclined to believe that it resulted from [a metamorphosis of!]
cells (“aus Zellen hervorgegangen ”), and therefore to regard it as a pro-
duct of the membrana granulosa. This conclusion was strengthened by
finding the granulosa composed of cylindrical cells in the case of Esox,
and that when this membrane appeared in the perch the granulosa cells
had disappeared ; but subsequently, finding that the follicular cells in
the perch were round, and not finding any transitional stages from the
epithelium to the membrane, he was compelled to leave the question
unsettled. Reichert was certainly looking in the right direction, and
evidently very near to a fair settlement of the question.

It remained for KétumKer (’58, p. 90) to confirm this supposition of
Reichert. He found that in February the capsular membrane had a
thickness of 45 » to 75. The tubules, he says, are formed by the
outgrowth of the epithelial cells of the follicle, so that the jelly which
joins them can only be a substance secreted by these cells. These so called
tubules were after all not hollow structures (“noch keine deutlichen

1 See Kolliker (’58), p. 81, foot-note.

96 BULLETIN OF THE

Hohlgebilde”’), but apparently solid pale processes of the epithelial cells,
on which the anastomosing filaments found by Miiller were visible.
Kolliker did not doubt, however, that they were from the beginning
hollow cell processes, but they still contained at the time of their forma-
tion cell contents, and only subsequently became clear. Their indepen-
dent nature was shown by the fact that in chromic preparations they
could be drawn out from the jelly without losing their union with the
[rest of the] epithelial cells. As long as the eggs remained in the fol-
licle the epithelial cells probably continued in union with the tubules;
but at the liberation of the eggs the cell bodies probably fell off, with
the exception of the walls, which were continuous with the tubules, and
then constituted the hexagonal facets of Miiller. Kdélliker was able to
produce a similar effect by artificially separating the cells from the
capsule.

Ransom (’68, p. 455, Plate XVI.), who does not seem to have been
acquainted with the papers of either Reichert or Kélliker, compares the
capsular membrane in consistence with fresh fibrine. ‘The striz look
like tubes, have a distinct double contour for each wall (Fig. 28), but are
filled with a vacuolating material, and do not seem to convey anything
into or out of the egg.” The outer surface was thrown into folds which
radiated from the ends of the “tubes,” but the hexagonal markings seen
by Miiller could not be made out. The tubes, instead of being funnel-
shaped, at their inner terminations divided into root-like processes, and
were in some way intimately adherent to the outside of the dotted yolk-
sac (zona). The clear matrix was elastic and concentrically laminated.
‘The appearance described by Miiller, of oil granules passing through
the tubes, may possibly have been due to vacuolation in them.” Experi-
ments with colored fluids to ascertain if there were any absorption of
fluids along the “tubes” always gave negative results: the cleavage
went on, the yolk-sac was dyed throughout, the clear matrix more than
the tubes, the germinal mass not at all. Either, therefore, the tubes
did not subserve imbibition at all, he contends, or in a much smaller
degree than the clear matrix.

Watpeyer’s (’70, p. 81) conclusion about the origin of the zona from
the granulosa appears to me to have resulted, in part at least, from the
fact that he was unable to discover any essential difference between it
and the capsular membrane of the perch. He says the latter does not
differ from the former in the principle of its structure. He rightly adds,
that “here [capsule] it is to be seen with the greatest distinctness that
the filaments are connected with the subsequently somewhat degenerated

i

MUSEUM OF COMPARATIVE ZOOLOGY. 97

remnants of the follicular cells.” He further adds, that occasionally
it appeared as though there were between these two membranes a thin
flat expanse of granular protoplasm in which the filaments terminated.

His (73, pp. 14, 15), who examined perch eggs in April, also confirms
the opinion of Kélliker, and gives a figure to show the relation of the
radial processes to follicle cells. The radial streaks consist, he says, of
a turbid substance, which stains in osmic acid, and is continuous with
conical nucleated bodies which form a continuous layer between the fol-
licle wall and the outer surface of the capsule. KdOlliker is therefore
right in considering the layer as “granulosa,” and the capsule as its
product.

Brock (’78, p. 556) gives the following clear, and I believe correct,
account of the capsular membrane of the perch. The follicular cells,
which at first are in close contact with the young egg, are raised up from
the zona radiata by the developing gelatinous layer, and with the advan-
cing growth of that layer are drawn out on the side toward the egg into
long processes which can be followed up to the zona. In older eggs
these follicular cells, separated by considerable intervals, (an indication
that their multiplication soon stops,) lie in shallow depressions of the
gelatinous capsule, and with their lower pointed ends continuous with
the processes. These appear to end at the zona with conical enlarge-
ments, but the author will not affirm that this is a constant feature.
Brock also maintains that Serranus hepatus has a very similar gelatinous
capsule. The follicular cells, however, are very peculiar. They form a
network of thin flat cells, which are in contact with each other only by
means ef lateral processes, while perpendicular processes, which are
sometimes branched and exceedingly fine, can be traced through the
jelly to the zona radiata. Concerning the development of the capsule,
nothing is known.

So far as regards the capsular membrane of the perch, Horrmann (’81,
pp. 19, 20, 27-29) comes to totally different conclusions from Brock, and
expresses views which seem to me untenable. In October, ovarian eggs
from 600 to 700 » in diameter possess a membrane 5 yu thick, which is
composed of two layers of nearly equal thickness. “The inner is the
true zona radiata ; the outer is composed of very numerous, small knob-
like projections, which stand very close together and correspond exactly
to the villi of the cyprinoids. On the free surface of the conical villi lie
the granulosa cells.”

It cannot be denied that the figure cited (. c., Taf. I. Fig. 9) corrob-
orates the description given. But there is one fact which I should im-

VOL, XIX.—NO. I. 7

98 BULLETIN OF THE

agine would have caused the author more concern than it seems to have
done. If his drawing accurately reproduces the conditions, there must
have been about four tumes as many villi as there were granulosa cells.
That in itself alone might not be of any significance to the author, espe-
cially as he disclaims any genetic connection between the granulosa and
the underlying capsular membrane, but it does seem as though it should
have received some explanation in view of the ultimate relation (Taf. I.
Fig. 10) which Hoffmann admits to exist between the radial fibres of the
capsular membrane and the cells of the granulosa. This is what he says
about the later (February) stage of the egg: The zona itself is seen
to consist of two layers, the inner much thicker than the outer. From
the latter there arise with small triangular bases long peculiar fibres
only 1 » in thickness, which are stained in osmic acid precisely like the
outer layer of the zona from which they arise. The outer ends of these
fibres are thickened, even more than their inner ends, and form a con-
tinuous layer, between which and the zona radiata the fibres themselves
are stretched like so many columns. ‘Over the proximal’ ends of these
fibres the granulosa cells are arranged in such a way that a granulosa cell
fits into nacu thickened end (Taf. I. Fig. 10).”

If after the [ovarian] eggs have lain in water a short time they are
transferred to osmic acid, it becomes very easy, says Hoffmann, to iso-
late both this layer (formed by the expanded ends of the fibres) and the
granulosa in the form of large shreds (“ Lappen”). From such prepa-
rations one can readily convince himself, he says, that the expanded
ends of the fibres form a continuous sheet, and are not processes of the
granulosa cells, as one is at first inclined to assume, and that the jibres
are, as the examination of the early stages proves, only the greatly grown-
out conical villi. |

That which seems to me to need explanation is, Why is the numeri-
cal relation between the villi and the granulosa cells so different at dif-
ferent stages in the growth of the egg, and why does this relation become
such an invariable one in the later stages of development ?

It might be answered, in reply to the first question, that the granu-
losa cells undergo rapid multiplication, and that cell division occur-
ring twice for all granulosa cells between October and February would
explain the altered relations; but is it not more reasonable to suppose
that, through some unexplainable accident, Hoffmann has been led to
attribute eggs to a perch which were taken in October from some other

1 Jt is not clear to me in what sense the author can use the word “ proximal” of
the ends of the fibres which are directed away from the centre of the egg.

‘

MUSEUM OF COMPARATIVE ZOOLOGY. 99

fish possessing a villous layer, than to ignore the evident constancy be-
tween villi and granulosa cells, and to assume an extensive multipli-
cation of the latter in eggs during the period of their growth from a
diameter of .7 mm. to .75 mm. ?

The latest paper dealing with the capsular membrane of Perca is that
of OwssannIKow (’85, pp. 3-8), who reaffirms Miiller’s claim that the
“tubules” are hollow structures, and corroborates Ransom’s discovery
that their inner ends are divided into branches which penetrate the
pore-canals of the zona radiata.

On gold or silver preparations of ovarian eggs, one finds the granu-
losa cells ' bounded by broad lines of a precipitate, so that there must
be present a large amount of intercellular substance. The cells them-
selves lie, as previous authors have shown, over the beginnings of the
corkscrew-shaped canals. These “beginnings,” in the fully developed
ege, are not at all cells, and have no nuclei; they are little funnels, with
the narrow end sunk as it were into the jelly. The finely granular
substance lies more compactly in the bottoms of the funnels; it is
scantier on the margin, and in many places extends beyond the rim,
of the funnel. This tissue (granulosa cells) often presents the appear-
ance of stellate cells joined together by numerous processes, and sepa-
rated from one another by abundant intercellular substance. The
more closely packed molecules at the bottom of the funnel had given
occasion, he says, to the assumption that there was a nucleus there.

1 In his account of the Graafian follicle, not always readily harmonized with his
figures, and sometimes obscure, he claims the presence of a greater number of epi-
thelial (or endothelial, for he recognizes no difference between the two) cell layers
than have usually been admitted. Thus, if I rightly understand the explanation
of Fig. 4, Taf. I., there are in Perca, e. g., two epithelial cell layers between the
vascular layer and the capsular membrane, — an outer layer of flat endothelium and
a deeper layer of cylindrical granulosa cells; but in the text (pp. 4-8) he speaks of
only two cell layers, an inner and an outer granulosa (!), which are separated by
the vascular layer. From his description of the latter as the source of the new
eggs, there can be no doubt that it is the “germinal epithelium” of authors. I
can reconcile this apparent contradiction between text and figures only on the
assumption that Figure 4 and its explanation belong to a period in Owsjannikow’s
studies when he was not as yet convinced of the error of taking “ the granular accu-
mulation in the bottom of the funnels of the ‘tortuous canals’ for a cell nucleus.”
That he ultimately supposed that to be an error appears from his description, and
the statement (p. 5), “ Die am Grunde des Kelches dichter an einander gelagerten
Moleceln gaben die Veranlassung dort einen Kern anzunehmen.”

In the case of Osmerus, moreover, he recognizes the existence of two layers of
large flat endothelial cells in addition to the granulosa. (Compare his explanation
of Fig. 8, Taf. I.)

100 BULLETIN OF THE

From the bottom of the funnel there proceeds to the zona radiata a
strongly and evenly twisted canal, which breaks off easily at the bottom
of the funnel. Owsjannikow cannot agree with Kélliker, who maintained
that in February these were solid fibres, because already in Novem-
ber and December he finds them hollow. The superficial layer of
the gelatinous mass, as well as that part which immediately surrounds
the canals, appears to be more compact than the rest of it, and the
vicinage of the funnel is more deeply stained by aniline red or gold
chloride than the other parts. The granules which occupy the canals
or the funnels never enter into the gelatinous substance when the canals
are ruptured, but escape into spaces which surround the canals. The
inner end of the canal does not terminate in a pointed manner, as
figured by His, but is often enlarged into a funnel, and sometimes
divided into two or three fibres,— in one case into so many that
it looked like a brush. On one occasion these branches were traced
through the zona. These processes are to be seen only in stained speci-
mens (gold chloride followed by aniline blue), because, having the same
refractive power as the substance of the zona, they are otherwise undis-
tinguishable. The fine molecules which lie on the inner surface of the
zona were found to be deeply stained, and the author concludes that the
dye must have penetrated through the spiral canals. The function of
these canals must consist in the transportation of nutritive material to
the yolk. They arise out of the granulosa cells, are similar to those
seen by Eimer in the adder, and are not processes of the zona radiata, as
affirmed by Hoffmann. The lateral processes from the canals were also
seen by Owsjannikow, but he has for them another interpretation. The
matrix (Zwischensubstanz) appears to lie in layers parallel to the surface.
Upon its being swollen by any fluid, narrow jfissures are formed be-
tween these layers, which join the canals and appear as processes from
them.

Besides the difficulty of trying to comprehend how fissures could
arise as a result of the swelling of a gelatinous mass, — it would seem
that the reverse process ought to be more favorable to their appearance,
— the sufficient answer to this last claim is, that the transverse processes
are more deeply stained than the remaining portions of the matrix,
which could hardly be the case if they were simply fissures.

It was in the hope of ascertaining something more about the inter-
esting capsular membrane in the perch, that I advised Mr. Eigen-
mann (’90) to include that fish in his studies on the development of

MUSEUM OF COMPARATIVE ZOOLOGY. 101

the micropyle. I believe it will be seen from his results that there is
still very good reason for maintaining that the tubular or columnar
structures of the capsular membrane, which have been the objects of
so much study, are derived from the granulosa cells, one from each cell,
and that the process by which the capsular membrane is formed is
neither simply a cell secretion nor exclusively a cell metamorphosis.
Although Eigenmann has not succeeded in getting stages which show
clearly all the steps in the formation of the capsule in Perca, he has
shown that there exist conditions in the later stages of the development
of the egg in Ksox (Kigenmann, ’90, Plate III. Fig. 37) which seem to
me of considerable importance in interpreting the conditions in Perca.
In Esox the cells become elongated, and the central (axial) portion
retains the granular and stainable properties of unmodified cell proto-
plasm. This axial portion is not cylindrical, but conical; its base is
directed outward and contains the nucleus. The peripheral portion
—which is more and more abundant as one approaches the zona —
is more homogeneous than the axial part, and reacts with dyes in a
different way. ‘he cell boundaries have been previously lost. The
boundary between these two constituents of the cell is not at first

sharp, so that this phase of the process may perhaps be regarded as
‘ one of metamorphosis rather than of secretion.

I believe that Perca must pass through some such stage as this during
the earlier part of the process which produces the capsular membrane.
I imagine that the distinction between the axial and the peripheral por-
tions of the cell becomes more and more sharply defined as the thicken-
ing of the capsule goes on. Meanwhile the axial portion does not long
retain the indifferent condition, but is metamorphosed, especially at its
periphery, into a highly refracting substance, so that there is reason for
regarding the structure as tabular. _ This metamorphosis advances till it
has practically obliterated the cell, even though a nucleus with a small
amount of enveloping protoplasm may still be made out at its distal
end in very late stages of ovarian growth. At any time before this, and
after the distinction between a funnel part and a tubular part has arisen,
the less modified distal portion of the cell may doubtless be easily sepa-
rated from the secreted gelatinous substance and also from the meta-
morphosed cell process. Such at least is the view which I have formed,
after a comparison of the granulosa in Perca and in Esox.

It would certainly be remarkable if the perch were the only represent-
ative of bony fishes in which such a process took place. I believe that
there are a few cases already known which may prove upon renewed

102 BULLETIN OF THE

inquiry to be essentially the same as the perch. The similarity to the
perch shown by Acerina vulgaris was recognized by Miiller (54, p. 189).
‘In Acerina,” he states, ‘the egg membrane has the same structure [as
in Perea], only it is much thinner, and consequently the tubules are only
short, not longer than the breadth of the areas.” Ransom (’68, pp. 453,
454) also says of this species that its ‘ yelk-sac has an outer layer or
‘Kikapsel.’” But he adds, that “the outer layer appears to be contin-
uous with, and similar in structure to, the yelk-sac proper.” However,
Owsjannikow (’85, pp. 17, 18, Taf. I. Fig. 13) has given an account of
it which points still more strongly to the resemblance claimed by Miiller.
That part of his account is especially significant in which he states that
in some preparations the follicle cells have the form of very narrow
cylindrical epithelium ; the broad end of the cell is directed outward,
the pointed end inward toward the zona. The cells, he adds, lie in a
transparent non-staining layer, similar to that in which the spiral canals
(in the perch) are located. Finally, a third condition is described in
which the cell form is lost. The structure begins with a broad short
funnel, and passes at once into a narrow, straight hollow fibre which
imbeds itself in the zona radiata.

The peculiar follicular layer described by Scharff (’88, p. 69, Plate V.
Fig. 15) in the interesting egg of the shanny (Blennius pholis) also
appears to have begun to undergo a modification in the same direction
that leads in the perch to the formation of a capsular membrane. The
substance which Scharff calls “interstitial” is, I believe, morphologically
the same as the gelatinous secretion of the follicular cells in Perca.

3. Micropyle and Micropylar Plug.

Since DoyrreE (’50) discovered, in 1850, an aperture leading through
the membrane of the egg of Syngnathus, and gave to it the name of micro-
pyle, there has been a good deal of attention given to that structure.

Independently of each other, and probably without knowledge of
Doyére’s discovery, Ransom (’56) in England and Bruc# (’55*) in Ger-
many rediscovered, in 1854, this structure in fish eggs, and both applied
the name which had meantime become current through Miiller’s (51 and
’54”) discovery of a similar canal in the egg membrane of Holothuria, to
which he also gave the name of micropyle, borrowed from the usage of
botanists.

Ransom in fact succeeded in observing in the egg of the stickleback
the passage of spermatozoa through this opening; Bruch was less fortu-
nate with Salmo fario and S. salar, although he made special effort to

a ee ee ee

—
=

— ee

|
|
|

MUSEUM OF COMPARATIVE ZOOLOGY. 103

discover if such a passage took place, and particularly emphasized the
fact that the orifice of the micropyle is of exactly the same size as the
spermatozoon.

These discoveries were soon (1855) confirmed by Leuckart (’55,
pp. 257-264) and Reicuert (56, pp. 83, 84, 98-104, Taf. IV. Figs. 1-4)
on the Continent, and by Tomson (’59, p. [100]-[104]) in England.
Their observations established the fact of the existence of the micropyle
in numerous fishes, and under several modifications of form. Ransom
had given a fairly accurate account of the structure of the micropylar
region, but Reichert especially insisted upon the differences between an
invaginated portion of the membrane and a passage through the latter
in the case of cyprinoids. He distinguished three regions,—an ap-
proach (Eingang), a fundus, and a neck or cylindrical canal, the length
of which was diminished from what it would otherwise have been by a
reduction in the thickness of the membrane to one third its normal
dimension.

It was, however, the interest in fertilization stimulated by Newport’s
researches on the impregnation of the ovum in Amphibia, and by Keber’s
paper, “ De introitu spermatozoorum in ovula,” etc., Kénigsberg, 1853,
that gave paramount importance to these discoveries, and attracted
general attention to them.

Perhaps it is not surprising, in view of this fact, that the ovarian
egg was less studied, and that the relation of the micropyle to the gran-
ulosa cells in its vicinity was not especially examined; and yet those
observers who concerned themselves with the questions relating to the
origin of the different egg envelopes must have been very near to inquir-
ing what share the granulosa had in producing so remarkable a modifi-
cation in the egg membranes. KOLLIKER (’58, p. 92), although only
incidentally making observations on the micropyle, established the fact
of its existence in a large number of cases. He regarded it simply as an
enlarged radial pore of the secondary vitelline membrane (zona), which
might be produced, he thought, by a process of resorption.

Several subsequent writers have concerned themselves only with ques-
tions relating to the form and position of the micropyle, and its probable
function, Thus BucuHouz (63, p. 72) compared the micropyle in Os-
merus eperlanus to a crater the floor of which is closed except for a mi-
nute canal in the middle of it, which traverses the thickness of the wall ;
Ransom (68), besides adding some unimportant details to his earlier
description of Gasterosteus, described briefly the micropyle in a large
number of other (fresh-water) fishes, claiming that it always terminated

104 BULLETIN OF THE

in a small elevation lying directly over the germ, and concluded, as the
result of experiments (pp. 459-462) made on Gasterosteus pungitius,
that ‘the function of the micropyle is to admit the spermatozoids to the
surface of the yolk”; and His (73, pp. 3, 4) described with some detail

the structure of the micropyle in Salmo salar and S$. fario, in both of
which he recognized a shallow depression (“ Mulde’”’) surrounding the

crater, which in S. fario terminated in a deep funnel, and this in the
canal. He also showed that only one spermatozoon at a time can pass
the micropylar canal, which terminates somewhat eccentrically over the
germinal disk.

HorrMann (’81, pp. 833-36) has confirmed for a large number of fishes
the observations that the inner end of the micropyle terminates in a
papillary elevation of the zona radiata, and that in the ovarian egg it
lies directly over the germ. From a comparison of the dimensions of
the spermatozoa and the calibre of the canal, he also draws the conclu-
sion that not more than one spermatozodn can traverse the micropyle
at atime. In nearly all the micropyles figured by Hoffmann the canal
is a tubular passage without any special enlargement; but in the case
of the herring’s egg — which is the one most carefully described — the
outer half of the passage is enlarged into a conspicuous bulbous cavity
(Taf. I. Fig. 19), which, so far as I recall, has been seen in only one
other instance, that of Petromyzon as figured by Calberla. But the
greatest interest attaches to the conditions figured for Leuciscus rutilus
(Taf. I. Fig. 20). In this case there exists a distinct plug of granulosa
cells occupying the depression in the egg membrane at the micropylar
region. Since the author does not mention the fact in the text, it is
probable that he attached to it no importance. I think, however, it is
the first clear proof published of the existence of a specialization in the
granulosa of the micropylar region in any teleost. On account of the
rather diagrammatic rendering of the granulosa cells, it is not possible
to be very confident about the existence of a specialized micropylar cell,
- but the fact that a single cell forms the apex of the plug favors that
view, and I shall be surprised if such a structure is not hereafter demon-
strated in this European fish.

Of the more recent writers on the micropyle, Owssannikow (785,
pp- 11-13, Taf. I. Figs. 5-7) describes for Osmerus eperlanus a micro-
pylar apparatus composed of two portions, an external and an in-
ternal, corresponding respectively to the two membranes which envelop
the egg, —the “external zona radiata” (which corresponds, in his
opinion, to the outer [capsular] layer in Perca) and the ‘internal zona

‘

= ————e

MUSEUM OF COMPARATIVE ZOOLOGY. 105

radiata.” The apparatus has the form of a crater-like depression, about
as described by Buchholz and Ransom. Where the membranes take the
direction of the crater, they form folds with the pointed ends directed '
inward. But of more interest are his statements, that the zona radiata
externa takes the greater share in the formation of the crater, and that
“‘ other tissues, especially the endotheleum and granulosa cells, participate in
the same.”” Owsjannikow was thus, I believe, the first person after W.
Miiller (75) to call attention to an intimate relation between granulosa
and micropyle.t But there are some elements of uncertainty about his
descriptions and figures that seem to baffle every attempt to reduce
them to harmony. The most perplexing thing about his description is
the use of the term “ endothelium,” which is at first used for Osmerus
in the following connection (p. 10): ‘Die Graafschen Follikel der |
Osmeruseier bestehen aus Endothel, Gefissen, Bindegewebe und Folli-
kelzellen.” In the description of other eggs (Perca) the word “ Endo-
thel” is also used as though applied to cells which lie outside the
vascular layer, and even as though including the germinal epithelium
of other authors (p. 4). Unless an endothelium having a very different
position from that previously described by him is meant, when he
says that it participates (as well as the granulosa) in the micropylar
structure, I believe that the author has fallen into some error; for I
am of opinion that neither the connective-tissue layer with its blood-
vessels nor the germinal epithelium shares directly in the formation of
the micropylar apparatus. Neither do I believe that there exists inside
the connective tissue any layer of cells except the granulosa. Moreover,
I do not think that any layer of endothelium, either inside or outside
the vascular layer, has been figured as sharing in the formation of the
micropyle. It does not help matters in the least to add that the author
discountenances (p. 4) any attempt to draw a distinction between epr-
thelium and endothelium ; for after saying that ‘‘Endothel urd Granu-
losazellen” share in this formation, he proceeds with a description
which certainly allows the assumption that there is only a single layer
of cells involved, to which, however, he gives successively the names of
granulosa, endothelium, and epithelium.

1 Although published several years ago, Owsjannikow’s studies were not made
until some time after I had demonstrated the conditions in Lepidosteus which have
been described above. His paper, as well as the more recent one of Cunningham,
has therefore had no influence in determining the course or the results of my studies
on Lepidosteus, nor did it influence me to suggest a comparative study on the eggs

of bony fishes, such as Mr. Eigenmann has undertaken; for I had already proposed
that question to one of my students before either paper was published.

106 BULLETIN OF THE

On “teased ” preparations Owsjannikow often finds the zona deprived
of its cellular covering, but ordinarily the detached cells are to be found
in the form of a continuous membrane, on which a conical projection is
to be seen. ‘The form of the projection corresponds exactly to that of
the external micropyle ; it is hollow; the cells at the entrance to the
micropyle have the form of a crown, and become smaller and smaller
toward the bottom of the crater. I am not sure that I fully understand
the figure (Taf. I. Fig. 6) which the author refers to in this connection,
but it appears to me to be a view of an egg from the animal pole; the
granulosa cells of the crater, having been detached, are seen partly in
side view, but somewhat obliquely, as a conical structure, and the pore-
canals of the external zona of the crater are visible where the granulosa
cells have been lifted. In the middle of the figure is an optical section
of that portion of the zona which forms the internal projection, and in its
centre the micropylar canal. There is apparently a single layer of cells,
and this I take to be the granulosa. It is to be regretted that the
author has not furnished us with a strictly radial section through the
micropyle and the accompanying structures on a sufficiently large scale
to enable one to determine what becomes of the membrana propria of
the theca folliculi in the region of the micropyle. One would infer, from
the statement that this granulosa cone was hollow, that the theca must
follow the course of the crater; but if it does, it must be different from
all other known cases. Not even in Perca does the membrana propria
suffer any deflection or infolding due to the participation of the gran-
ulosa cells in the formation of a micropylar structure. I have also been
considerably perplexed by Owsjannikow’s Figure 5 (Taf. 1.). At first I
took it to represent a strictly diametric section of the egg and its mem-
branes. With that understanding of it, I imagined that the large oval
body just above the micropyle might possibly represent a single micro-
pylar cell in some way loosened from its natural position; but more
careful study leads me to believe that this is a figure representing in
part an optical section, in part a surface view, and that the oval struc-
ture presents an oblique view of the external entrance to the funnel-
shaped cavity of the crater, still lined with granulosa cells, while all the
rest of the figure represents a view of the egg as it would be seen in
optical section. Owsjannikow states that the inner micropyle can be
regarded as a somewhat enlarged canal of the zona, and claims that it
subserves the nutrition and growth of the egg; for he has traced from
the inner end of the canal a row of granular bodies which were con-
tinued in the yolk as a fine thread, which at last disappeared. This row

MUSEUM OF COMPARATIVE ZOOLOGY. 107

of granules and its thread-like continuation the author regards as a pro-
duct of the granulosa cells. But at present there are no data, he says,
either teleological or phylogenetic, which can explain the remarkable
structure of the external micropyle.

Ryper (86, p. 30, Plate VII. Fig. 35), who has demonstrated the
existence of the micropyle in the eggs of several fishes, speaks of the
passage through the capsular membrane in Perca Americana as “a
wider pore-canal which leads to the micropyle.” It is evident from
Ligenmann’s (90) account and from his figures that this statement is
inaccurate.

Cunnincuam (’86%, pp. 59, 61-63, ete., Figs. 2-4, 12) has shown much
more satisfactorily than W. Miiller (’75) did, that in Myxine glutinosa
the follicular epithelium plays an important part in the formation of the
micropyle. Of an egg 16 mm. in length he says: “ At the exact pole of
the egg there is a differentiated portion of epithelium, where a prolife-
ration of the latter has taken place. This portion is composed of poly-
gonal cells, which are little or not at all elongated, and towards the egg
it runs out into a thin cylindrical process which penetrates the next layer
[zona radiata] as shown at ¢. p. [Fig. 2]... . This process penetrates
the vitelline membrane [zona radiata], occupying a tubular cavity in the
latter, and passing through it to form a hemispherical projection on its in-
ner surface. ... This cellular projection is covered by a thin membrane
continuous with the vitelline membrane, and is not in immediate con-
tact with the germinal disk. . . . There is thus at one pole of the nearly
ripe ovum a tubular canal extending through the chorion [zona radiata],
but. not open internally, filled up by a cylinder of cells projecting from
the follicular epithelium. ... It is evident,” the author adds, “ that this
aperture is to form the mcropyle in the ripe ovum... . It is probable
that careful investigation would show that in all Teleosteans whose ova
possess a micropyle that structure is produced by a projection of cells
from the follicular epithelium.” Cunningham also believes it “at least
possible that in all vertebrates the micropyle will be found on investiga-
tion to be produced in the same way as in Myxine, namely, by the
growth of a cellular process from the follicular epithelium towards the
vitellus while the vitelline membrane is being formed.”

In an egg 20 mm. in length “ the proliferation and differentiation of
cells at the pole in the follicular epithelium have disappeared, but the
cylinder of cells, though reduced in size, still remains in the micropyle,
and is evidently destined to keep the latter open until the maturation
of the ovum is complete.” This egg was from material obtained in

108 BULLETIN OF THE

December. Older eggs, obtained at the end of January, although only
slightly larger than the December egg (21 mm. long), presented a very
different appearance at the micropylar region. Cunningham says of the
latter: “ The micropyle is somewhat narrower, and the cells present in
it at previous stages have disappeared almost completely, only a little
débris remaining. The micropyle seemed also in these ova to be open
internally, though of this point I am not absolutely certain. If there
is a membrane closing the inner end, it is an extremely thin one.”
Cunningham has shown conclusively, I believe, that the granulosa
has much to do with the modifications of the egg membrane in the
micropylar region, but there are several particulars concerning which
his description and figures leave me in a rather unsettled state of mind.
In the first place, the author does not seem to distinguish with sufficient
sharpness between a funnel-like region, which may be partly the re-
sult of an infolding of the membrane, and a passage through the mem-
brane, which I have called the micropylar canal. It seems to me possible
that his uncertainty as to whether the micropyle is closed at its inner
end at a late stage may be due to this fact. The disproportion between
the calibre of the “ micropyle ” and the size of the spermatozoa? is not
alluded to, but at once suggests to me that the structure in question
may be the equivalent of the micropylar funnel only. I should be quite
certain that it was so, were it not that the drawing of the latest stage
(Fig. 12) — which is not sufficiently explained —admitted two inter-
prétations. In this figure the whole passage is divided into two por-
tions of about equal length, but of very unlike calibre. The inner half
is a narrow canal with parallel walls, about one third of a millimeter in
diameter (actual size about 10»); the outer half is abruptly widened
to 6 or 8 mm. in diameter, and gradually increases toward the gran-
ulosa to 10 mm. (nearly 300 p actual size). There are no granulosa
cells, however, in either portion of the passage. The almost flat-bot-
tomed outer half of the passage would appear to be the equivalent of
-the micropylar funnel in bony fishes, and I should certainly have so
regarded it if Cunningham had not evidently considered the inner nar-
rower portion as a part of the “micropyle” of previous stages from
which the cells had disappeared, leaving “ only a little débris.” If the
author is right in this assumption, that the narrow part of the ap-

1 In its narrowest place the “ micropyle ” of the author’s Figure 3 is represented
as 5 mm. in diameter in the drawing, which, being magnified 280 diameters, makes
its actual diameter about 18 uw, whereas the actual diameter of the head of a sper-
matozoon (Fig. 14) is not over 7 u.

‘

MUSEUM OF COMPARATIVE ZOOLOGY. 109

paratus was once occupied by granulosa cells, then either the true
micropylar canal has not been seen in Myxine, or it is formed in a
different manner from that of bony fishes, for in no case is it occupied
in the latter by a plug of cells from the follicular epithelium. The
former alternative is, in my opinion, the more probable. This con-
clusion I base partly on the appearance of Cunningham’s Figure 3, and
partly on the conditions presented by sections of an ovarian ovum of
Myxine australis, which I have studied since reading Cunningham’s
paper. In his Figure 3, the plug of granulosa cells which is sunk
into the yolk is completely enveloped in a uniform layer not more than
2 thick, which not only separates its deep end from the yolk, but
also its sides from the membrane called by him vitelline membrane, or
zona radiata. What the significance of the part of this thin mem-
brane lining the chimney-like elevation of the zona radiata may be, I
cannot say, unless it is reflected at the upper edge of the chimney to
form the outermost layer of the zona; but the portion which separates —
the yolk from the granulosa plug I regard as the equivalent of the first
formed portion of the zona radiata, and believe that the true micropylar
canal will be found in the form of a minute passage through that mem-
brane. My principal reason for this opinion rests on the condition of
this region in the egg of M. australis. This egg was about 22 mm.
long and 6 mm. in diameter. It was enveloped by a follicular epithelium
composed at the animal pole of a single layer of cells, averaging about
10 w thick. Over the region of the flattened germinative vesicle the
granulosa was gradually thickened to about 25 uw, and from the middle
of the thickening a solid plug of cells about 35 » long and 25 wp in di-
ameter projected into the yolk. The membrana propria of the follicular
theca passed over the micropylar region without being at all infolded, so
that the total thickness of the granulosa, measured from the apex of
the plug, was about 60 uw. The cells of the plug were not well pre-
served, but appeared to be of about the same relative size as in Cun-
ningham’s Figure 3, — i.e. of one half or one third the diameter of the
plug. There was no enlargement at the apex of the plug as seen in
his figure. Between the granulosa and the yolk, and in contact with
both, was a highly refractive thin (3.5 4) membrane, which at first
appeared homogeneous, but in which I believe I have detected at
intervals radial markings. This membrane became thinner at some
distance from the pole.

The whole apparatus had such a striking resemblance to the micropy-
lar apparatus in Lepidosteus that I cannot doubt that the granulosa

110 BULLETIN OF THE

plug represents the same structure in Lepidosteus, and that conse-
quently the micropylar canal is to be found at the bottom of the
funnel-shaped infolding produced by the plug.

4. Micropylar Cell.

Although no other cases are yet known in which a single cell of
the follicular micropyle-plug is so evidently differentiated from its
neighbors, as in Lepidosteus, still it is clear from the results of Eigen-
mann’s (’90) studies that the existence of such a specialized cell is not
uncommon.

The most interesting question relative to the micropylar cell is that
of its physiological signification.

That it sustains some intimate and constant relation to the micropyle
itself can scarcely be doubted. Perhaps its primary function is fulfilled
by serving as a source of passive resistance to the forming membranes
in the region of the micropyle—a kind of mould for them — during
the process of their formation, therefore a mechanical device for pro-
ducing a micropylar funnel. In that case it would doubtless often be
more than the single micropylar cell on which devolved the function; it
would be rather that plug-like accumulation of granulosa cells, with the
micropylar cell at its apex, which attains such an extensive development
in Lepidosteus and some other cases.

But the very fact that one cell is generally, if not always, differen-
tiated more than the rest, suggests that the function referred to may
not be the only one, — perhaps, indeed, not the primary one. There
are two other possible functions which are naturally suggested in this
connection.

Concerning the first of these it may be said that it still remains to be
shown to what extent the micropylar canal and its funnel are the result
of an exclusively progressive process of development ; whether, in other
words, any part of this structure is produced by a process of resorption.
Such a process would not be without a parallel. At least Leuckart (55,
pp- 108, 216, 247) has asserted in most positive terms that the micropyle
of certain insects’ eggs is not to be found in the chorion from the very
beginning of its formation, but that it arises after the production of the
chorion by means of resorption.’

1 More recently, it is true, Korschelt (’84, pp. 421, 422) has shown that the
micropyle in Meconema varians is formed by a single cell, and he apparently leaves
no room for a process of resorption, since he says: “‘ Die Entstehung des Canals ist
wohl so zu denken, dass die Zellen schon friihzeitig einen Fortsatz ausstrecken, der

‘

MUSEUM OF COMPARATIVE ZOOLOGY. 111

The fact that Eigenmann (’90) has been unable to discover a micro-
pyle in the earlier stages of the formation of the egg membranes in fishes
may also point in the same direction. I would not wish, however, to
place too much weight upon such negative evidence ; it requires exten-
sive and indeed the most exhaustive examinations to make such testi-
mony satisfactory. Especially am I compelled to this reserve, in view
of the fact that Cunningham (’86*) has found the micropylar apparatus
well developed at a relatively early stage in the formation of the zona
in the case of Myxine glutinosa ; but it will be observed that he says
(p. 61) : “This cellular projection zs covered by a thin membrane continu-
ous with the vitelline membrane, and is not in immediate contact with the
germinal disk.” An actual opening does not exist, therefore, at the time
of which he speaks. Although I have found in M. australis at an ap-
parently early stage in the development of the zona a deep infolding of
that membrane, as described above, still I have not satisfied myself in
the single specimen sectioned that there is an ordfice through the mem-
brane at this stage of development. But on this observation I would
not care to speculate were it not confirmed by Cunningham’s studies
on more extensive and I presume histologically more favorable material,
for I know how easily one may be deceived as to the existence of so
minute a structure as that with which we have to deal.

The facts which I have given above do have a certain weight with me
as rendering it possible that in Myxine at least the micropylar canal, in
the strict sense of the word, is not present until near the maturity of
the ovum, and that consequently it may be the function of the cell
nearest to the bottom of the funnel—the micropylar cell — to absorb
so much of the already formed membrane as is necessary to allow the

anfangs nur kurz ist, spater mit dem Dickerwerden des Chorions und dem entspre-
chenden Zuriickweichen der Epithelschicht aber linger und linger wird.”

In his final paper, Korschelt (’87, p. [43] 223) suggests a method of reconciling
Leuckart’s views with his own observations: ‘Die Beobachtung Leuckart’s, nach
welcher die Mikropylkanale sich nicht von Anfang an am Chorion finden sollen,
liesse sich vielleicht mit unseren Befunden in Uebersinstimmung bringen. Wir
erkannten in mehreren Fallen, dass die Masse des in der Bildung begriffenen
Chorions eine durchaus weiche und plastische sein muss. Deshalb wire es leicht
denkbar, dass die Zellfortsitze, welche die Kaniile entstehen lassen, nicht von
Anfang der Bildung an vorhanden wiren, sondern erst spater in die weiche Masse
des jungen Chorions hinein entsendet wiirden. Dann wiirde es ein Stadium geben,
in welchem das Chorion eine ununterbrochene Oberfliche besasse. Seine Ansicht,
das die Mikropylkanale an dem bereits gebildeten Chorion durch Resorption
entstehen sollen, diirfte Leuckart wohl aufgegeben haben, sobald er die Entste-
hungsweise der Porenkaniale der Eischale kennen lernte.”

112 BULLETIN OF THE

passage of the spermatozoén.? I realize that this is a speculation on
very narrow foundations ; for even if it could be shown to result from
absorption, it might be the protoplasm of the ovum, not the granulosa
cell, which accomplished the work. There are, however, still more seri-
ous objections to this view, which, though not disproving it, render it
very doubtful. ‘The fact that in general pore-canals and orifices in cu-
ticular secretions are the results, not of resorption, but of the previous
existence of protoplasmic projections, makes it probable, without positive
proof to the contrary, that the same would be true in this case.

The serious, and indeed, as it seems to me, insurmountable objection
to considering the whole micropylar funnel as the result of absorption, is
the condition which it exhibits in Lepidosteus, where there is a very
gradual diminution in the thickness, not only of the zona, but also of
the villous layer. It is not probable that any process of absorption
could result in diminishing the thickness of both layers so evenly with-
out affecting their mutual relations, unless perchance it should be
imagined that the zona was absorbed through the agency of the yolk,
and the villous layer by means of the granulosa cells. But even that
assumption would not help the case very much, for it would still have
to be explained why the shorter villi retained all the parts of the longer
ones, and in nearly che same proportions.

While, then, the conditions clearly preclude the possibility of looking
upon the micropylar apparatus in general as resulting from a process of
absorption, it by no means follows that the micropylar canal may not be
produced by such activity.

The other purpose which it has occurred to me the micropylar cell
may subserve, is to facilitate the penetration of spermatozoa. Not pre-
cisely that a minute drop of slimy substance, resulting from the degen-
eration of this cell and covering the orifice of the narrow canal, would
offer less resistance than water, but that its presence might prevent the
occlusion of the orifice by the accidental introduction of impenetrable
substances without itself offering any serious obstacle to the free en-
trance of the fertilizing element.

If one were to attempt a phylogenetic explanation of the micropylar
funnel and canal in bony fishes, he might reason somewhat as follows,

1 In Cunningham’s Figure 3, the cells of the granulosa plug which form the
layer nearest the yolk — four of them being cut in the section figured —are all
larger than the remaining cells of the plug, but I am unable to say that any one of
them is the largest of all.

MUSEUM OF COMPARATIVE ZOOLOGY. ES

The development of a persistent egg membrane impervious to sperma
tozoa would evidently be possible only with the concomitant production
of one or more orifices ; for without such provision no egg could be fer-
tilized, and the transmission of such tendeucies would clearly be impos-
sible. That would necessitate the development of the micropyle by
what I should call an exclusively progresswe method. It would not
imply any regressive or resorbent process. How, then, could one find
reason for claiming any such process? I believe it would only be neces-
sary to assume that the zona radiata in its original development subserved
some useful end during the development of the ovum, in order to form an
idea of the possible course of events which has led to the present con- |
dition. Imagine eggs at oviposition provided with a zona radiata which
remained or had become penetrable by spermatozoa; such eggs would be
in the most favorable condition for fertilization, but on account of the
condition of the zona they would be poorly protected against external
agencies. If, under these conditions, a portion of the zona radiata in
some eggs should become more resistant, even to the point of preventing
the entrance of spermatozoa, the eggs thus modified would be better
protected from injurious external influences than those which remained
in the original condition, and yet they would be almost as readily fertil-
ized as the latter, provided some portion of the zona remained, as at first,
penetrable. In short, the advantages of such a changed condition would
be greater than the disadvantages, and consequently in the long run the
more favorable condition would predominate. Evidently the optimum
protection would have been reached when a region no larger than that
absolutely necessary to admit a single spermatozodn had been left for
that purpose. But this process of restriction in the area accessible to
the spermatozoa may easily have been accompanied by another process,
which may have begun as early as the former. The zona was assumed
to remain, or become at the time of oviposition, penetrable to spermatozoa.
It seems to me entirely reasonable that a process tending to modify a
portion of the zona and make it more readily penetrable should be set
up in the ovary, and that such eggs in the matter of fertilization alone
would have some slight advantage over eggs less easily penetrable. If
now these two tendencies were operative at the same time,—the one
serving to soften a part of the zona, in order to make it the more readily
penetrable, the other to harden another portion, to make the egg less
subject to adverse environment, —the former would become localized
by the encroachment of the latter until at length there would be only

a limited area in which the process of softening went on; but this might
VOL. XIX.—No. 1. 8

114 BULLETIN OF THE

be — as we must believe often happens in the animal economy — corre-
spondingly intensified, until an activity which in the beginning resulted
in only a feeble modification in the condition of the zona ultimately ter-
minated in its complete liquefaction and absorption.

A single argument, which it seems to me may have some value, in
support of this hypothesis, is to be drawn from the condition of the zona
in Elasmobranch fishes. So far as I now see, the complete disappear-
ance of the zona at the maturity of the egg would be entirely in har-
mony with the hypothesis. The condition there at any rate seems to
me to favor the assumption previously made, that the zona originally had
a function distinct from that which it now appears to possess in protect-
ing the embryo after fertilization. For if not, it must be an inheritance
from ancestors which, like bony fishes, had their embryos thus protected.
There are probably few who would defend the idea that the Elasmo-
branchs are descended from bony fishes, and the evidence of common
ancestors with eggs thus protected still remains to be found.

If, as the Russian naturalists assert, there are several micropylar open-
ings in the egg-shell of sturgeons, it may be that those fishes present a
condition which is intermediate between an extensive region of penetra-
bility and the extreme restriction which now prevails in bony fishes.

The funnel portion of the micropylar region is certainly the less essen-
tial and least constant part of the structure. It may reasonably be
considered, I believe, a secondary condition, and the explanation of its
development might lie either in the fact that it served, in a passive way,
to direct the motion of a greater number of spermatozoa toward the
actual orifice, or, possibly, that it served to preserve from accidental
removal the protective products of the degenerated micropylar cell.

Thus the micropylar canal might be regarded as the result of two to
a certain extent antagonistic tendencies, — the fittest solution of a prob-
lem requiring the fulfilment of two conditions. The micropylar funnel
could obviously be regarded as a partial compensation for the diminished
opportunities for fertilization caused by a restriction of the area available
for that purpose, and might have arisen simultaneously with the restric-
tion, or only after the latter had attained its present maximum.

SS eS eee ee

MUSEUM OF COMPARATIVE ZOOLOGY. 115

Summary.

1. The first food of young gar-pikes after the absorption of the yolk
is mosquito larvee ; later, they feed on small fishes.

2. Very cold water is injurious to young gars.

3. The arched form which the body sometimes exhibits is probably
always the result of an abnormal condition.

4. The acts of catching and swallowing the prey are complicated.
Fishes are usually swallowed head first.

5. The young gar-pike lives at the surface of the water after the
absorption of the yolk-sac.

6. The emission of bubbles of gas through the gill slits is accompanied
by a rolling of the body to one side and a forward movement.

7. Analyses of the bubbles of gas emitted by young fishes, and also
of a limited amount of atmospheric air which had been used by the
fishes in respiration, showed a reduction of the oxygen to 10-15 per
cent, and the presence of only a small amount of carbon dioxide, 0-1.7
per cent.

8. Air which had been previously deprived of its carbon dioxide gave
no evidence of containing even a trace of that gas, after having been
respired. Consult the text for the conditions of experimentation.

9. It is probable from these experiments that the air-bladder respira-
tion serves to oxygenate the blood, but that the elimination of carbon
dioxide is effected in some other manner. It is possible that the two
elements of the respiratory function of higher vertebrates were succes-
sively transferred to the air-bladder, that of oxygenating the blood being
the first to be transferred.

10. There are two egg membranes in Lepidosteus, —a zona radiata
and a villous layer, —and they are intimately joined together. Both are
radially striate, the zona finely, the villous layer coarsely.

11. Balfour and Parker were mistaken in stating that there was a
homogeneous non-striate membrane inside the striate zona, and also in
supposing that the pyriform bodies (villi) of the outer covering were
metamorphosed follicular cells.

12. The outer layer is not as thick as the zona, and is made up of
radially compressed and folded columnar structures, the villi.

116 BULLETIN OF THE

13. Each villus is composed of three parts, — head, stalk, and roots.
The roots project into the pore-canals of the zona.

14, The zona presents the usual structure of that membrane in bony
fishes. The pore-canals are slightly enlarged and spiral at their periph-
eral ends.

15. The egg of Lepidosteus has a single micropyle, but it has been
overlooked by previous observers.

16. The micropylar apparatus embraces a funnel and a canal. The
funnel results from an infolding and a reduction in thickness of both the
villous layer and the zona radiata. The micropylar canal is a narrow pas-
sage of uniform calibre (2 ») and circular cross section, through the thin-
nest part of the two membranes, namely, at the bottom of the funnel.

17. The granulosa of the mature ovarian egg consists of a single layer
of polygonal cells, except in the region of the micropylar funnel, where it
forms a plug of cells completely filling the funnel.

18. The membrana propria of the theca folliculi is not infolded in the
region of the micropylar funnel.

19. A single large granulosa cell, the “micropylar cell,” forms the
apex of the plug, and occupies the bottom of the funnel.

20. The villous layer is formed before the zona radiata. It first
appears at the surface of the yolk in eggs about 0.4 mm. in diameter,
and reaches one third its maximum thickness a year before the egg ma-
tures. It is produced by the ovum, not by the granulosa cells.

21. The number of the villi is not increased during the growth of the
villous layer.

22. The zona radiata is likewise the product of the ovum, and its
formation requires less than twelve months.

23. The name “capsule ” should not be used for the zona radiata; it
ought to be restricted to membranes outside the zona, which, like that
of the perch, are the product of the granulosa.

24. An egg membrane comparable structurally and genetically with
the zona radiata of bony fishes is to be found in representatives of all
the groups of fishes except Amphioxus. It is fugitive in selachians and
Lepidosiren, and probably in viviparous teleosts. The zona is produced
by the ovum, not by the follicular cells, and is traversed in all cases by

MUSEUM OF COMPARATIVE ZOOLOGY. 117

pore-canals, which rarely (Myxine?) branch. J. Miller is wrongly cred-
ited with their discovery.

25. An egg membrane genetically, but not always structurally, com-
parable with the willows layer of Lepidosteus, is found in several other
cases: possibly in Petromyzon, probably in selachians and Lepidosiren,
and certainly in several teleosts. This membrane is also produced by the
ovum, and earlver than the zona.

There is some reason to believe that it exists in the herring and the
smelt (Osmerus). Hoffmann is probably in error in attributing the
presence of villi to the eggs of Perca in October, and Owsjannikow cer-
tainly is in asserting that the mushroom-shaped villi in Gasterosteus are
individual cells.

26. The capsular membrane is produced, as originally defined by
J. Miller, by the follicle or follicular cells. It has often been confounded
with the zona, and also with the villous layer.

27. Althongh the “tubules” in Perca have been described as possess-
ing root-like prolongations which penetrate the pore-canals of the zona,
they are genetically unlike the villi found on other eggs, being produced
by the granulosa cells alone. Hoffmann’s statement to the contrary
rests on insufficient evidence.

28. A comparison of the condition of the granulosa in Blennius pholis
and Esox with that in Perca, is believed to shed some light on the prob-
able steps by which the capsular membrane is produced.

29. The most important paper on the origin of the zona and the
villous layer I believe to be that of Kolliker, whose conclusions I have
confirmed in the case of Lepidosteus. All authors who have maintained
that either zona radiata or villous layer is the product of the granulosa
I believe to be in error ; and in particular I maintain that the reasons
assigned by Cunningham to prove that the zona in Myxine is produced
by follicular epithelium are not adequate to establish his proposition.

30. It is doubtful whether the micropylar canal has been seen in
Myxine. What W. Miiller called the micropyle was the micropylar
funnel, and possibly that which Cunningham describes as the micropyle
does not embrace the canal.

31. What I have called the micropylar plug of granulosa cells was
first seen by W. Miiller in Myxine ; later it was figured, but apparently
not appreciated, by Hoffmann in Leuciscus; it was described as hollow
by Owsjannikow in the case of Osmerus, who recognized its relation to

118 BULLETIN OF THE

the micropylar funnel ; and finally, it was described more fully for
Myxine by Cunningham. Eigenmann has found the same structure in
a number of bony fishes.

32. The micropylar cell has never before been recognized. Eigen-
mann has now found it in a number of osseous fishes, — Perca, Pygos-
teus, Esox, etc. I believe it may fairly be assumed to exist in the
greater number of those fish eggs which possess a micropyle, and that it
has an important function in connection with the formation of the micro-
pyle or the fertilization of the egg. )

33. I have made the following suggestions as to the possible func-
tions and history of the micropylar apparatus: The micropyle, being
evidently a provision for the fertilization of the ovum, may have its
present structure as the result of two to some extent conflicting tenden-
cies ; one induced by the advantages of protection to the egg, the other
by the necessity of some provision for the penetration of the fertilizing
element. But the best protection is not compatible with penetrability
of the membrane at all points. Any reduction in the extent of the
penetrable surface would be favorable to protection. An optimum con-
dition would be reached when the penetrable area is reduced to a mini-
mum, and that is the diameter of the head of a spermatozo6n.

The funnel may be a partial compensation for such reduction.

The micropylar plug may mechanically determine the presence and
form of a funnel.

The micropylar cell may serve to form the canal by resorption, or to
prevent the occlusion of the canal by less penetrable matter at the time
of oviposition,

CAMBRIDGE, April 7, 1889.

MUSEUM OF COMPARATIVE ZOOLOGY. 119

Postscript.

Since the completion of the present paper, I have received from the
author, Dr. J. Beard, a copy of his “Preliminary Notice, On the
Early Development of Lepidosteus osseus.” This paper was presented to
the Royal Society of London, April 20, 1889, and was printed in the
Proceedings of the Society, Vol. XLVI. pp. 108-118, May 16, 1889.

Dr. Beard’s material was procured by him in the spring of 1888 from
the same place as that which supplied Mr. Agassiz and myself, — Black
Lake, New York.

In this preliminary notice the author does not devote much attention
to the egg membranes. What he says about the imner egg membrane
coincides with the views which I: have expressed. He says that it “is
not composed of two layers either in Lepidosteus or in the sturgeon: It
is a simple zona radiata, the strize reaching to the innermost portions of
the membrane. The division into two layers, sometimes seen, is the
optical effect of thick sections.”

I cannot agree with his conclusions regarding the external layer, and
am confident that his final paper will not contain proof of the accuracy
of his statements. He says: “The pyriform bodies are certainly modi-
fied cells, each with the remains of a nucleus at its outer end. These
modified cells have degenerated into a sort of glue, which causes the ex-
cessive stickiness of the newly laid eggs. . . . In the ovarian egg these
‘pyriform bodies’ are probably nutritive cells to the ovum, for their
outer ends near the nuclei contain a number of minute yolk particles.”

As far as regards the external layer, it is difficult to conceive how
our views, whether morphological or physiological, could have been
more divergent.

120 BULLETIN OF THE

BIBLIOGRAPHY.

Agassiz, Alexander.

'78*. The Development of Lepidosteus. (Presented Oct. 8, 1878.) Pro-
ceed. Amer. Acad. Arts and Sci., Vol. XIV. No. 4, pp. 65-76, Pis. L-V.
1879.

Agassiz, Louis.

42. Sce Voet, Cart, '42.

’57. Living Specimens of young Gar-pikes from Lake Ontario. Proceed.
Bost. Soe. Nat. Hist., Vol. VI. (1859), pp. 47, 48. Jan., 1857.

Aubert, Hermann.

53. Beitrage zur Entwickelungsgeschichte der Fische. Zeitschr. f. wiss.

Zool., Bd. V. Heft 1, pp. 94-102, Taf. VI. 16 Aug., 1853.
Baer, Karl Ernst von.

'35. Untersuchungen iber die Entwickelungsgeschichte der Fische, nebst
einem Anhange iiber die Schwimmblase. 52 pp., 1 Taf. u. mehreren
Holzschn. im Texte. Leipzig: Vogel. 1835.

Balfour, Francis M.

78, On the Structure and the Development of the Vertebrate Ovary.
Quart. Jour. Micr. Sci., Vol. XVIII. No. 72, pp. 383-438, Pls. XVII-
KIX, Oct. 1878:

’81. A Treatise on Comparative Embryology. Vol. II., xi + 655 + xxii pp.
429 figs. London: Macmillan & Co. 1881.

Balfour, F. M., and W. N. Parker.

’81. On the Structure and Development of Lepidosteus. (Rec'd Nov. 24,
1881.— Abstract.) Proceed. Roy. Soc. London, Vol. XXXIIT. No. 217,
pp. 112-119. 8 Dec., 1881.

’82. On the Structure and Development of Lepidosteus. (Ree’d Nov. 24.
— Read Dec. 8, 1881.) Philos. Trans. Roy. Soc. London, 1882, Pt. II.
pp. 359-442, Pls. XXI.-XXIX.

Baumert, Friedr. Moritz.

'53, Chemische Untersuchungen iiber die Respiration des Schlammpeizgers
(Cobitis fossilis). Annalen der Chemie u. Pharmacie (Liebig), Bd.
LXXXVIII. pp. 1-56. 1853.

Beddard, Frank E.

86%. Observations on the Ovarian Ovum of Lepidosiren (Protopterus).
Proceed. Zool. Soc. London, for 1886, Pt. III. pp. 272-292, Pls. XXVIII,
XXIX. 1 Oct., 1886.

MUSEUM OF COMPARATIVE ZOOLOGY. 121

Brock, J.

'78. Beitrage zur Anatomie und Histologie der Geschlechtsorgane der Kno-
chenfisehe. Morph. Jahrb., Bd. [V. Heft 4, pp. 505-572, Taf. XX VIIL.,
XXIX. 1878.

Bruch, C.

’55. Ueber die Befruchtung des thierischen Hies u. wber die histologische
Deutung desselben. Mainz. 1855.

55°. Ueber die Mikropyle der Fische. Zeitschr. f. wiss. Zool., Bd. VII.
Hefte 1 u. 2, pp. 172-175, Taf. IX B. 20 May, 1855.

Buchholz, Reinhold.

’63. Ueber die Mikropyle von Osmerus eperlanus. Arch. f. Anat., Physiol.
u. wiss. Med., Jahrg. 1863, pp. 71-81, Taf. IIT A., Figs. 1-4. 1863.

63%. Nachtragliche Bemerkungen iiber die Mikropyle von Osmierus eper-
lanus. Arch. f. Anat., Physiol. u. wiss. Med., Jahrg. 1863, pp. 367-372,
Taf. VIII A. 1863.

Calberla, Ernst.

78. Der Befruchtungsvorgang beim Ei von Petromyzon Planeri. Zeitschr.
f. wiss. Zool., Bd. XXX. Heft 3, pp. 437-486, Taf. XX VII.-XXIX.
7 March, 1878.

Cunningham, J. T.

’86. On the Mode of Attachment of the Ovum of Osmerus eperlanus. Pro-
ceed. Zodl. Soc. London, for 1886, Pt. III. pp. 292-295, Pl. XXX. (Read
May 4.) 1886.

"86°. On the Structure and Development of the Reproductive Elements in
Myxine glutinosa, L. Quart. Jour. Micr. Sci., Vol. XX VII. Pt. I. pp.
49-76, Pls. VI., VII. Aug., 1886.

Doyére, M. P. L. N.

’50. Note sur ’@uf du Loligo media et sur celui du Syngnathe. L’In-

stitut, Tom. XVIII. p. 12. 1850.
Eigenmann, Carl H.

90. On the Egg Membranes and Micropyle of some Osseous Fishes. Bull.

Mus. Comp. Zodl., Vol. XIX. No. 2, pp. 133-154, Pls. I-III. 1890.
Eimer, Th.

72°. Untersuchungen iiber die Hier der Reptilien. Arch. f. mikr. Anat.,

Bd. VIII. pp. 216-243, 397-434, Taf. XI., XII, XVIII. 1872.
Gegenbaur, Carl.

61. Ueber den Bau und die Entwickelung der Wirbelthiereier mit partielle
Dottertheilung. Arch. f. Anat., Physiol. u. wiss. Med., Jahrg. 1861, pp.
491-529, Taf. XI. 1861.

Gréhant, [Nestor].

69. Recherches physiologiques sur la Respiration des Poissons. Ann. Sci.

Nat., sér. 5, Zool., Tom. XII. pp. 371-382. 1869.
Haeckel, Ernst.

’55. Ueber die Hier der Scomberesoces. Arch. f. Anat., Physiol. u. wiss.

Med., Jahrg. 1855, pp. 28-31, Taf. IV., V. 1855.

122 BULLETIN OF THE

Henneguy, Felix.
_’88. Recherches sur le Développement des Poissons osseux. Embryoge-
nie de la Truite. Jour. de l’Anat. Physiol. norm. et path., Tom. XXIV.
No. 5, pp. 413-502; No. 6, pp. 525-617, Pls. XVIII-XXI. 16 Dec.,
1888, and 20 Feb., 1889.
His, Wilhelm.
'73. Untersuchungen iiber das Ei und die Entwickelung bei Knochenfischen.
Leipzig: F. C. W. Vogel. 1873. 4+ 54 pp., 4 Taf., 4to.
Hoffmann, C. K.
’81. Zur Ontogenie der Knochenfische. Verhandl. d. koninkl. Akad. v.
Wetenschappen, Amsterdam, Deel XXI., 164 pp., 7 Taf. 1881.
Humboldt, A. von, et Provencal.
See PROVENGAL UND HumBo pt.
Jobert.
'77._ Recherches pour servir a |’ Histoire de la Respiration chez les Poissons.
Ann. Sci. Nat., sér. 6, Zool., Tom. V. No. 8, 4 pp. 1877.
'78*. Recherches Anatomiques et Physiologiques pour servir a |’ Histoire de
la Respiration chez les Poissons. Ann. Sci. Nat., sér. 6, Zool., Tom. VII.
No. 5,7 pp. dvAug, 1878:

Kolliker, Albert von.

'57. Nachweis von Porenkanalchen in den Epidermiszellen von Ammocetes
durch Professor Leuckart in Giessen nebst allgemeinen Bemerkungen tber
Porenkanale in Zellmembranen. Verhandl. physical.-med. Gesellschaft in
Wirzburg, Bd. VII. pp. 193-198. 1857.

’58. Untersuchungen zur vergleichenden Gewebelehre, angestellt in Nizza im
Herbste 1856. Verhandl. physical.-med. Gesellschaft in Wurzburg, Bd.
VIII. pp. 1-128, Taf. I-III. 1858.

Kolessnikow, N.
'78. Ueber die Hientwickelung bei Batrachiern und Knochenfischen. Arch.

f. mikr. Anat., Bd. XV. Heft 3, pp. 8382-414, Taf. XXV. 23 Aug., 1878.

Korschelt, Eugen.
'84. Ueber die Bildung des Chorions und der Micropylen bei den Insecten-

eiern. Zool. Anzeiger, Jahrg. VII. Nos. 172, 173, pp. 894-898, 420-425.
21 July, 4 Aug., 1884.

87. Zur Bildung der Eihillen, der Mikropylen und Chorionanhange bei den
Insekten. (Hing. 10 Aug., 1885.) Nova Acta Leop.-Carol. Deutsch.
Akad. der Naturforscher, Bd. LI. No. 8, pp. 181-252, Taf. XXXV.-
AXXIX, Halle. 1367:

Kowalevsky, A., Ph. Owsjannikow und N. Wagner.

'70. Die Futwickelungsgeschichte der Store. Vorlaufige Mittheilung. Bull.
Acad. Imp. Sci. St. Pétersbourg, Tom. XIV. col. 317-325, 7 figg. 1870.

"70%, Also in Mélanges Biol. tirés du Bull. Acad. Imp. Sci. St. Pétersbourg,
Tom. VII. (1869-70), pp. 171-188. 1870.

MUSEUM OF COMPARATIVE ZOOLOGY. 123

Kupffer, Carl.

'78*. Die Entwicklung des Herings im Hi. Jahresber. der Commission zur
wissensch. Untersuchung der deutsch. Meere in Kiel fir die Jahre 1874
-76, Jahrg. 4-6, pp. 175-226, 4 Taf. Berlin. 1878.

Kupffer, C., und B. Benecke.

78. Der Vorgang der Befruchtung am Ei der Neunaugen. 24 pp., 1 Taf.

Kénigsberg: Hartung. [1878.
Lereboullet, Auguste.

54. Résumé d’un Travail d’Embryologie comparée sur le Développement du
Brochet, de la Perche et de Ecrevisse. Ann. Sci. Nat., sér. 4, Zool.,
Tom. I. pp. 237-289. 1854.

61. Recherches d’Embryologie comparée sur le Développement de la Truite,
du Lézard, et du Limnée. Premiére Partie. HEmbryologie de la Truite com-
mune (Salar Ausoni, Val.; S. fario, L. Bl.). Ann. Sci. Nat., sér. 4, Zool.,
Tom. XVI. pp. 113-196, Pl. II., III. 1861.

Leuckart, Rudolph.

’53. [Article] Zeugung. Wagner’s Handworterbuch der Physiologie u. s. w.,
Bd. IV. pp. 707-1000. 1853.

’55. Ueber die Mikropyle und den feinern Bau der Schalenhaut bei den
Insecteneiern. Zugleich ein Beitrag zur Lehre von der Befruchtung.
Arch. f, Anat., Physiol. u. wiss. Med., Jahrg. 1855, pp. 90-264, Taf.
VIL-XI. 1855.

Leydig, Franz.

’52. Beitrage zur mikroskopischen Anatomie und Entwicklungsgeschichte
der Rochen und Haie. Leipzig: Engelmann. 1852. iv. + 127 pp., 4
Taf.

Ludwig, Hubert.

"74. Ueber die Hibildung im Thierreiche. Arbeiten aus d. zoolog.-zoot. In-

stitut in Wirzburg, Bd. I. pp. 287-510, Taf. XIII.-XV. 1874.
Meckel von Hemsbach, Heinrich.

’52. Die Bildung der fir partielle Furchung bestimmten Eier der Vogel, im
Vergleich mit dem Graafschen Follikel der Decidua des Menschen. Zeitschr.
f. wiss. Zool., Bd. III. Heft 4, pp. 420-434, Taf. XV. 1852.

Moreau, Armand.

63. Sur Air de Ja Vessie natatoire des Poissons. Comptes Rendus de
Paris, Tom. LVII. No. 1, pp. 37-39. 6 July, 1868.

63". Sur l’Air de la Vessie natatoire des Poissons. Comptes Rendus de
Paris, Tom. LVII. No. 20, pp. 816-820. 16 Nov., 1863.

Morren, Auguste.

'41, Recherches sur |’Influence qu’exercent et la Lumiére et la Substance
organique de Couleur verte souvent contenue dans |’Eau stagnante, sur le
Qualité et la Quantité des Gaz que celle-ci peut contenir. Ann. de Chim. et
de Phys., sér. 3, Tom. I. pp. 456-489. 1841.

'44. Recherches sur les Gaz que |’Eau de Mer peut tenir en Dissolution en

124 BULLETIN OF THE

différente Moments de la Journée, et dans les Saisons diverses de l’ Année,
Ann. de Chim. et de Phys., sér. 3, Tom. XII. pp. 1-56. 1844.
Morris, Charles.
’85. On the Air-bladder of Fishes. Proceed. Acad. Nat. Sci., Philadelphia,
1885, Part II. pp. 124-135. 1885.
Miiller, August.
’64. Ueber die Befruchtungs-Erscheinungen im Eie der Neunaugen. Schrif-
ten d. k. phys.-dkonom. Gesellsch. zu Koénigsberg, Jahrg. V. pp. 109-119,
Taf. 1V. 1864.
Miiller, Johannes.
51. [See also Jonannes Miuimr, ’54>.] Monatsberichte d. Akad. d. Wis-
senschaften Berlin. 28 April, 1851.
"54. Ueber zahlreiche Porencanale in der Hicapsel der Fische. Arch. f. Anat.,
Physiol. u. wiss. Med., Jahrg. 1854, pp. 186-190, Taf. VIII. Figs. 4-7.
) 1854.
54°, Anmerkung des Herausgabers [oz R. Remak, Ueber LEihiillen und
Spermatozoen]. Arch. f. Anat., Physiol. u. wiss. Med., Jahrg. 1854,
p. 256.
"54>, Ueber den Canal in den Hiern der Holothurien. Arch. f. Anat., Physiol.
u. wiss. Med., Jahrg. 1854, pp. 60-68.
Miiller, Wilhelm.
'75. Ueber das Urogenitalsystem des Amphioxus und der Cyclostomen.
Jena. Zeitschr., Bd. IX. Heft 1, pp. 94-129, Taf. IV., V. 30 June, 1875.
Owen, Richard.
66. On the Anatomy of Vertebrates. Vol. I. Fishes and Reptiles. Lon-
don: Longmans, Green, & Co. 1866. 42 + 650 pp., 452 figures.
Owsjannikow, Ph.
"70%. Die Entwickelungsgeschichte der Flussneunaugen (Petromyzon fluvi-
atilis). Vorlaufige Mittheilung. Mélang. Biolog. tirés du Bull. de PAcad.
Imp. Sci. St. Pétersbourg, Tom. VII. (1869-71), pp. 184-189. 1870.
’85. Studien iiber das Ei, hauptsachlich bei Knochenfischen. Mém. Acad.
Imp. Sci. St. Pétersbourg, sér. 7, Tom. XX XIII. No. 4, 54 pp., 3 Taf.
1885.

Poey, Felipe.
’55. Observations on different Points of the Natural History of the Island of

Cuba, with reference to the Ichthyology of the United States. Ann. Lye.
Nat. Hist. New York, Vol. VI., 1858, pp. 183-187. Oct., 1855.
Provencal et A. von Humboldt.
'09. Recherches sur la Respiration des Poissons. Mém. de Phys. et de Chim.
de la Soc. d’Arcueil, Tom. IT., Paris, pp. 859-404. 1809.
09%. Also in Journ. de Phys., Tom. LXIX. pp. 261-286. 1809.
Provencal und A. von Humboldt.
‘11. Untersuchungen iiber die Respiration der Fische. Jour. fiir Chemie u.
Physik (Schweigger), Bd. I. Heft 1, pp. 86-121. 6 Feb., 1811. German
translation of ProvencaL ET A. von Humsotpr, ’09, dy Dr. Sigwart.

MUSEUM OF COMPARATIVE ZOOLOGY. 12d

Ransom, W. H.
56. On the Impregnation of the Ovum of the Stickleback. Proceed. Roy.
Soe. London, Vol. VII. pp. 168-172. 1856.
67. On the Structure and Growth of the ovarian Ovum in Gasterosteus lei-
urus. Quart. Jour. Micr. Sci., n. ser., Vol. VII. pp. 1-4, Pl. I. Jan.,
1867.
68. Observations on the Ovum of Osseous Fishes. Philos. Trans. Roy. Soe.
London, Vol. CLVII. Pt. II. pp. 431-502, Pls. XV.-XVIII. 1868.
Rathke, Heinrich.
’61. Entwickelungsgeschichte der Wirbelthiere. Mit einem Vorwort von A.
Kolliker. Leipzig: W. Engelmann. 1861. 9 + 201 pp., 8vo.
Reichert, Karl Bogislaus.
’56. Ueber die Micropyle der Fischeier und iber einen bisher unbekannten,
eigenthiimlichen Bau des Nahrungsdotters reifer und unbefruchteter
Fischeier (Hecht). Arch. f. Anat., Physiol. u. wiss. Med., Jahrg. 1856,
pp. 83-124, 141, 142, Taf. IL., III., und IV. Figg. 1-4. 1856.
Remak, Robert.
'54. Ueber Hihillen und Spermatozoen. Arch. f. Anat., Physiol. u. wiss.
Med., Jahrg. 1854, pp. 252-255. 1854.
Ryder, John A.
’81°. Notes on the Development, Spinning Habits, and Structure of the Four-
spined Stickleback (Apeltes quadracus). Bull. U. 8. Fish Commiss.,
Vol. I. pp. 24-29. [1881] 1882.
Ble. Development of the Spanish Mackerel (Cybium maculatum). Bull.
U. 8. Fish Commiss., Vol. I. pp. 135-172, 4 pls. [1881] 1882.
’82°. Development of the Silver Gar (Belone longirostris), with Observations
on the Genesis of the Blood in Embryo Fishes, and a Comparison of Fish
Ova with those of other Vertebrates. Bull. U. S. Fish. Commiss., Vol. I.
pp. 283-301, Pls. XIX.-XXI. May 2 and 19, 1882.
’83. On the Thread-bearing Eggs of the Silversides (Menidia). Bull. U. 8.
Fish Commiss., Vol. ILI. pp. 198-196. 1883.
’'84. A Contribution to the Embryography of Osseous Fishes, with special,
Reference to the Development of the Cod (Gadus morrhua). Ann. Report
U. 8. Commissioner of Fish and Fisheries for 1882, XVIT. pp. 455-605,
Pls, 1.-XII.
84%, Also separate, with title-page and cover. 149 pp., 12 pls. Washington:
Government Printing Office. 1884.
’85. On the Development of Viviparous Osseous Fishes. Proceed. U. S.
National Museum, Vol. VIII. Nos. 8-10, pp. 128-155. Pls. VI.-XI.
25 May, 188).
86. On the Development of Osseous Fishes, including Marine and Fresh-
Water Forms. Extracted from Ann. Report U. 8. Commissioner of Fish
and Fisheries for 1885 pp. [1]-[116], Pls. I-XXX. 1886.
87. [Same as RypzER, ’86.] Ann. Report U. 8. Commissioner of Fish and
Fisheries for 1885, pp. 484-604, Pls. 1-XXX. 1887.

126 BULLETIN OF THE

Salensky, W.

'77. Entwickelungsgeschichte des Sterlets. Vorlaufige Mittheilung. (Rus-
sian.) In den Beilagen zu den Protokollen d. 84 u. 89 Sitzung. d. Ge-
sellsch. d. Naturforscher an d. Kaiserl. Universitat in Kasan. 34 pp. 1877.
(Cited from Hofmann u. Schwalbe, Jahresb. f. 1878, Bd. VII. Entwicke-
lungsgeschichte, p. 213.)

78. Entwickelungsgeschichte des Sterlets. Vorlaufige Mittheilung. (Con-
tinuation of W. Sauensxy, '77.) 5. Postembryonale Entwickelung.
(Russian.) Jn den Beilagen zu den Protokollen d. 97 Sitzung d. Gesellsch.
d Naturforscher an d. Kaiserl. Universitat in Kasan. 21 pp. 1878.
(Cited from Hofmann u. Schwalbe, Jahresb. f. 1878, Bd. VII. Entwicke-
lungsgeschichte, p. 213.)

'78*. Zur Embryologie der Ganoiden. JI. Befruchtung u. Furchung des
Sterlet-Hies. IL. Entwickelungsgeschichte des Skelets beim Sterlet.
Zool. Anzeiger, Jalrg. 1. Nos. 11-18, pp. 2438-245, 266-269, 288-291.
Nov. 4,411, 25, 1878.

'78>. History of the Development of the Sterlet (Acipenser ruthenus).
Pt. I. Embryonic Development. (Russian.) Mém. Soc. Naturalists
Imp. Univ. Kasan, Tom. VII. No. 3, pp., 1-226, Pls. 1-IX. 1878.

"79. Abstr. of W. Satensky, '77, 78, and '78*, by Mayzel ix Hofmann u.
Schwalbe, Jahresb. f. 1878, Bd. VII. Entwickelungsgeschichte, pp. 213,
218-225. 1879.

’80. History of the Development of the Sterlet (Acipenser ruthenus).
Part. II. Post-embryonic Development and Development of the Organs.
(Russian.) Mém. Soc. Naturalists Imp. Univ. Kasan, Tom. X. Part II.
pp. 227-545, Pls. X.—XIX. Kasan, 1880. (Continuation of W. Sa-
LENSKY, '78°.)

’81. Recherches sur le Développement du Sterlet (Acipenser ruthenus).
Arch. de Biol., Tom. II. fase. 2, pp. 233-341, Pls. XV.-XVIII. 1881.

Scharff, Robert.

’'87. On the Intra-ovarian Egg of some Osseous Fishes. (Ree’d Nov. 17,
1886. — Abstract.) Proceed. Roy. Soc. London, Vol. XIV. No. 249,
pp. 447-449. 1887.

’87*%, On the Intra-ovarian Egg of some Osseous Fishes. Quart. Jour.
Mier. Sci., Vol. XXVIII. pp. 53-74, Pl. V. Aug., 1887.

Schultz, Alexander.

'75. Zur Entwickelungsgeschichte des Selachiereis. Arch. f. mikr. Anat,

Bd. XI. pp. 569-582, Taf. XXXV. 1875.
Schultze, Max. Sigmund.

'56. Die Entwickelungsgeschichte von Petromyzon Planeri. Eine von der
hollandischen Societaét der Wissenschaften zu Haarlem i. J. 1856 gekrénte
Preisschrift. 50 pp., 8 Taf. 4to. [1856 ?]

Solger, Bernhard.
’85. Dottertropfen in der intracapsularen Flissigkeit von Fischeiern. Arch.

f. mikr. Anat., Bd. XX VI. Heft 2, pp. 321-334, Taf. XII. Dec. 14, 1885.

it a ee

MUSEUM OF COMPARATIVE ZOOLOGY. - 57

Steenstrup, J.

’63. En lagttagelse af Mg med hornagtige Aghylstre hos Slimaalen (Mysxine
glutinosa, Linn.) med Hensyn til det om denne Fisk udsatte Prisspergs-
maal. (An Observation on Eggs with horn-like Egg-case, in the Slime-
Kel, Myxine, etc.) Oversigt o. d. kgl. danske Videnskabernes Selskabs
Forhandlinger i Aaret 1863, pp. 233-239. [1864 ?]

Stockman, Ralph.

’83. Die aussere Hikapsel der Forelle. Mittheil. a. d. Embryol. Institut

Wien, Bd. II. Heft 3, pp. 195-199. 1883.
Thomson, Allen.
’'59. [Article] Ovum iz The Cyclopedia of Anat. and Physiol., edited by
Robert B. Todd, Vol. V. (Suppl. Vol.), 1859, pp. 1-80 and [81}{142].
Note. — Part I., pp. 1-80, was issued in 1852; Part IL., pp. [81}{142], in
1855.
Vogt, Carl.

'42. Hmbryologie des Salmones. Neuchatel. 1842. 6 + 328 pp., 8vo.
Avec Atlas, fol. obl. de 7 pls.

Being Tome I. of Li. Agassiz, Histoire Naturelle des Poissons d’Hau douce de
? Europe Centrale.
Vogt, Carl, et S. Pappenheim.

’'59. Recherches sur |’Anatomie comparée des Organes de la Génération chez
les Animaux Vertébres. (Déposé dans les Archives de l’Acad. le 30 Dec.,
1845.) Ann. Sci. Nat., sér. 4, Zool, Tom. XI. pp. 331-369, Pl. XTIT. ;
Tom. XII. pp. 100-131, Pls. IL., III. 1859.

Waldeyer, Wilhelm.

"70. Hierstock und Ei. Ein Beitrag zur Anatomie u. Entwickelungsgeschichte
der Sexualorgane. Leipzig: W. Engelmann. 1870. 8 +174 pp., 6 Taf.
8vo.

Wilder, Burt G.

"76. Notes on the North American Ganoids, Amia, Lepidosteus, Acipenser,
and Polyodon. Proceed. Amer. Assoc. Adv. Sci., Vol. XXIV B, Detroit
Meeting, pp. 151-196, Pls. 1-III. 1876.

'77. Gar-Pikes, Old and Young. Popular Sci. Monthly, Vol. XI. Nos. 61,
62, pp. 1-12, 186-195, 10 figures. May and June, 1877.

or ou). ao Ree
3 | or AY ee ee eae eos a dad agi i
a rr ers ihe skin. nutoh, aaacnlaet Te fag haetor ight a See:
om pic PL wet Tap etd tenth LW ia yn ele jie wh).
2 ble ets latinas 5a un the chin de Ae OF Come aero mt

ic Fs ne ay iy iT yy. GO phe f ut ; Se esr) uy j
a | E Wookie fii lta Asi Iaay Wat iene
-" | Rg ic ea 1 thi

! \
\ oP ael ee ei te tt 7, ty | Maemtegs Sh faba “i?

J ; é

as iy Po or an oak cue '- “ Pa

tir i iy ee Meo Ver he Ae) my
: P 1 ; i Hi ~ L

. ‘ 1
7 ; : "a ‘ ie ¥ |
_, |
ve

‘ é, ,
. A 2 A fi 4 Lf ys Ne P :
tay - .
‘i a a is 4 }

det Tieiet ue
as oF ; 4 ey ee, gar! aia Pm ‘> Wh 6) he
7 4 ; ‘
ys | | im }
7 ; age:
. . ; P Phe } Ke) an 7 t2 phd Wy eh Aiyte Fy
s , ‘ ' as : ; T r ‘ ee
= ’ riko AED Fob Ainl i Vd ee) A a ee

“4 % -j

7 y u 4 bal a. e 1
y

ee i

7) i ;
vA ' i j od dat ; A
re ‘ » 1 , i

re

oi } ’ 1 vf ’ } , vat. ’ tt ee
Ri ,

Hi |

a) ; = T
$ 4 " epr ee . pak .,

* SP Da ; ' a uy nie
Pr F a * ‘ae viet ler ey ey “94 eee

lie ()

A :

‘EXPLANATION OF FIGURES.

, )

figures were drawn with the aid of the camera lucida, and were made
from preparations of Lepidosteus osseus.

MARK, — Lepidosteus,

ABBREVIATIONS.
cap. Head of villus. pd. Stalk of villus.
c.-t. cp. Connective-tissue corpuscle. rz. Root of villus.
Jus. mat. Maturation spindle. st. vil. Villous layer of egg membrane.
gran. Granulosa. th. fol. Membrana propria of theca folliculi.
m py. can. Micropylar canal. vac. Vacuole.
m py. cl. Micropylar cell. vit. Vitellus.
nl. Nucleus. vs.g. Germinative vesicle.
nl. gran. Nucleus of granulosa cell. z.r. Zona radiata.

PLATE I.

All the figures of this plate were made from material that had not been hardened,
and all the figures except Figs. 7 and 11 are magnified 472 diameters.

Fig. 1. A surface view of a small portion of the villous layer of egg membrane.

‘* 2. The appearance presented by the same layer when the region near the
boundary between it and the zona radiata is in focus. Some of the
roots of the villi are seen between the stalks.

« §. The zona radiata when the focusing is a little below its outer surface.
A few pore-canals are occupied by roots of villi and appear darker.

« 4. A portion of a radial section after being treated with weak hydrochloric
acid. Two of the villi much more elongated than the others.

« 5. A-radial section of a fresh egg-shell, showing the relative thickness of
the zona radiata and the villous layer.

“ 6. A portion of the same with the villous layer removed, but leaving its
roots in the spiral pore-canals, Examined in glycerine.

“7. Portions of the villous layer removed from the zona radiata and much
swollen in water. The roots appear like a fine fringe. X 145.

«“ 8. The appearance of the pore-canals after treatment with hydrochloric acid.
The most of them, especially toward the margin of the figure, should
have been drawn larger but faint. A few are conspicuous from the
presence of roots of villi.

“ 9. atoh andj ton, isolated villi in various stages of elongation after imbi-
bition of very dilute hydrochloric acid; 7, after soaking in water
only.

“ 10. A fragment of the zona radiata deprived of the villous layer and treated
with weak hydrochloric acid until all the pore-canals except those
containing villous roots had disappeared. The zona, having become
soft, was partly crushed, so that the roots were seen obliquely, the
ends toward the top of the plate being the ones torn from the stalk.

“ 11, Optical radial section of the micropylar region of a fresh egg, the wall of
the membrane beyond the micropyle being projected on the same
plane. X 145.

“ 114, Optical cross-section of the same, at plane a of Fig. 11.

“ 11%. Optical cross-section of the same at plane d of Fig. 11.

leah opi |

MARK —LEPIDOSTEUS.

TUM

UAT

a

\

B. Meisel, lith.

E.L.Mark, del.

ae!

>a

= <

A inne: Beis ya

+ a —

net a

a, 4 ange at sinktty ie ‘a : siti
ee does oh, shopeg wedi. WIT a) e ¥ =i
via pac ven a Wa i per re) th 'ndaes "Ty ro we
wl siharcTiuone wt @ >, eae ire . Hy
whpanyt sa ¥ age a rane b eekebo 47 he) Pad

iy: Ai wna? apn Ne aah , ie ely ured a Ws R
A all Papel tee eae

Pel -
a at eh ial

i" Ai irer ve BAGS A ai Satan ilies. io Snel

AD |) ‘ a, f “- es é

en oe & ; ; } a

-

x
<™
a

y s ’ 5 H 4
el

y Wal ye : . + : His, bh he 4
wits eta) ait wih Tae, /Oyo le sre Blake Aes my wh deals JA

‘ POS ast ORS MCE Miidieiede Pncbow's i Yeas iby WR gl
| ’ vai Merihete WR rol ALAS iiie oie. Wile oa
Pm eedadd ol hpninkl . ~duy!od

Fete,

i oy

I
Magy AER cl fw'e bd

Dare rre is Terre # ai: ry i. ty hhh) Sie ii yay f

a Waheed miget '} iu b
abraleersi ts Md pars th 2 “os ur? weer ik — oviat 3h} Phys at ra A ee t

s y el ey ran hin ewe. ahi e, Myleene iieelt By) & Ms law.

“ e ain ee Wea ’ ‘oll al) ‘ ) iy ie
ung uit waives | i duke ‘ Cire ts C0 Viaw basal al’! beni
ay, ‘aioe site filiiw ye bd
é eee ATP TL Y Ey, ta a ee biahaty qnyeneest ‘5 ‘
ne, f k y A
‘aa F 5 . i Baie ‘ail y tea . ‘ Vy
“ ; PE :
eal] i “ys en IS sa? STP iN sé tf) ij iy ry, Winns Maoh P ’
l+/ Y ' ‘ y ; Lit 4 ?
“ par ve ei ei] ns gla ‘4 Ta | Mies eer hy TT a i
4 ie i | t
vil; Mirae 2H ibe. yp eae ;
¥4 ual Pa — 4 JALVIV AN vi Psi Aha! Ray) > -¥ Hy ve Vea
a: Gir) "1 rid
Vv it ‘ { ¥tgs ash 5 aay Be 2% 8 Wie ee
4 } id Aen 1 74 ey’ ir. wd Le
1 a TEM ogh ie a BAY LE Meh te file Ky ty 1 ae '
5 : lie s
: a é 4, _— q
ie ‘ es yan tea PARE hele SAT ATT ve A ay i’ "
; 7 1, a iy pec hs ot p 2 K 4
, Wiiaest! rt } wit A 0 ; thy Hey tt wea hy Y Tee hi }
i = 4 : ais ; =
. i ‘AG Fe ¥ Pa
Pe MR Meicts MHeltite AY boats Hore ° r
bie " held y y i ‘Vasall i phish 4 ah % : Wey iy miStiomy 4 ie bite ‘ta. shit : 1 ae
f 4 ’ :
o- ry

’, WwiAe an ; a 7 a3 4
h ?

WP sricas

Mark. — Lepidosteus.

ABBREVIATIONS.
cap. Head of villus. pd. Stalk of villus.
c.-t. cp. | Connective-tissue corpuscle. rx. Root of villus.
Jus. mat. Maturation spindle. st. vil. Villous layer of egg membrane.
gran. Granulosa. th. fol. Membrana propria of theca folliculi.
m py. can. Micropylar canal. vac. Vacuole.
m py. cl. Micropylar cell. vit. Vitellus.
nl. Nucleus. vs. g. Germinative vesicle.
nl. gran. Nucleus of granulosa cell. z.r. Zona radiata.

PLATE II.

All the figures of this plate are magnified 515 diameters.

Fig. 1. Portion ofa radial section through the shell of a deposited egg preserved
in 5 per cent potassic bichromate. Stained in carminic acid dis-
solved in 80 per cent alcohol. Mounted in benzole-damar.

“ 2,38. Isolation preparations of villi from a mature egg pressed from female.
The egg was let fall into 90 per cent alcohol, afterwards soaked in
water, and then stained in acetic-acid carmine. Examined in glyce-
rine. The heads were of a much brighter rose-color than is shown in
the lithographic print. In Fig. 3,c, the villi are seen edgewise; in
a, b, and d, flatwise.

“ 4-6. Sections through villi of a mature ovarian egg which was preserved in
0.25 per cent chromic acid forty-eight hours, washed in water six
hours, and further hardened in grades of alcohol. Stained in alco-
holic borax-carmine (Grenacher). In Figs. 4, 5, the ends of the villi
are seen; in Fig. 6, the sides.

a 4. Sections of stalks from the pole opposite the micropyle.

“ 5,6. Sections of stalks from near the micropyle.

* 7. Radial section of an egg preserved in Perenyi’s fluid (44 hours) followed
by alcohol and stained in alcoholic borax-carmine, showing an inner
portion of the zona radiata partly detached from the outer portion.
It is not a separate membrane.

~ 8. Radial section through zona and villous layer of an egg preserved in 5 per
cent potassic bichromate, stained in carminic acid in 80 per cent alco-
hol, and mounted in benzole-damar.

Pie

B. Meisel, litk

me
pe

ae He
oxy Bog. 4% -ae
onl

s

MARK —LEPIDOSTEUS.
' E.L.Mark, del.

ns ee Se

» rr? ‘
° “ns
0
wf
‘
"
4
> ‘
> ‘
on fe
& ‘
2
} “
.
*
;
t
.
ca]
; f
¢
7 is
7
me °
i
‘ a ‘
;
le
.
oC
1
i
’
)
=
er
»
j -
”
A
ia
.
-
be 4
-
» =~ 7 ®
*

he Bi ot Re
rh Ries Mii thy 1 my ‘ i iy ‘_ wan
i eu Cea ih “J Nine). avai

Mark. — Lepidosteus.

cap.
c.-t. Cp.
Sus. mat.
gran.

m py. can.
m py. cl.
nl.

nl. gran.

ABBREVIATIONS.
Head of villus. pd. Stalk of villus.
Connective-tissue corpuscle. rz. Root of villus.
Maturation spindle. st. vil. Villous layer of egg membrane.
Granulosa. th. fol. Membrana propria of theca folliculi.
Micropy lar canal. vac. Vacuole.
Micropylar cell. vit. Vitellus.
Nucleus. vs. g. Germinative vesicle.
Nucleus of granulosa cell. z.r. Zona radiata.

PLATE IIL.

All the figures of this plate were drawn from the shell of an egg preserved in
cold corrosive sublimate (4 hours) followed by alcohol, stained in Kleinenberg’s
hematoxylin, sectioned in paraffine, and mounted in benzole-damar. All except
Fig. 38 magnified 515 diameters.

Fig. 1.
<< 9
tf 3.
Ci9 4,
Oe:

Radial section, the heads of some of the villi broken off.

Tangential section through the heads, at A of Fig. 1.

Similar section of four heads more highly magnified to show the deeply
stained peripheral portion. X 750.

Tangential section through the middle region (B of Fig. 1) of the stalk.

Section parallel to preceding through the region of the roots of the
villi (C of Fig. 1). The lower portion of the figure cuts through a
deeper part of the membrane (zona) than the upper portion does. The
middle portion shows the branching roots of the villi as they enter
the pore-canals.

iganetiy

MARK —LEPIDOSTEUS.

(nd
Vv
> hey
¢
B. Meisel, lith

“e

*\ Pagh Gee 4
abtee a aes
e

ae
oie
4
mm
;

sy
eres

ot
s

4
CJ

,

%
E..L.Mark, del

Ir

=

=

"* ' ‘
rein \ WES
*
i Eee DENI NM ae Verano), ROAR OF Wi Bi
4
' oh Ae by
! i y

es

e 4 petite a Ne Renal ) .
Pasian) ivsa Sn
es Peabiitiys pati oniitnk” Saby a

ne iy aE fps) 2 nt a4

nee Haken dh eras

ponve Saris é iiwke eins Siig beg Re
"| ES ib ‘tw AE eral i aren oe:
1 LL iitaye WY Dari: a): swe aiid: " re Pt ze
er

vt

tiation. Wied lk). hw aan et, ae *

es Bats

i sor WE He 1a ti Ege roby vi Say Riliing Hithast 4 saat
MWA BP patie & O0ip, Faw al yegeeriied
Wap te gi) GE eh), i ail sone) tesa®
4 Riise ,tRD Gh aU Ladi ed Van (mend va

5 an : rab igh 4 AON Da ved by sintish (arane ye
i aa ah PN. GF Lovrcnt Gl scale ERs ticki wily Relves 9: rs
Sy =e ( Ce hin erie ak oe evil’ hiitdi eat ee a hyivee a 7 : L ut ;
. a ra Aid 10), py rary} ny ith pe 4%, uf! Sait rn aid Dee ie 7 ; é ei f
ERD Lenin Ga ebot hyuly’: Cran: Ves Mat Dh Pies iano
( Bebiineiieiy ngod | itd Tihaier ty abi 4) Fie arony
Eat eileen POP W ON a Wie ane A Pinta E WA sc ae peel
aa ot dali mich erected vy nga tyes ded dere: ate Wits 7
a : pon drondy x ane oul} at Silos ays
rer rt Ercageil (hws Ai rea i wale ae) Upurea prith { i) > | <
bg Swat ie Ly whcngy Hvis ah Pye tnd uy bec eae oF,
bin li aand a5 if \ eh shi ape eR hd tp ‘ ea r P
yi be BreoL ") C. a pMmOULN €t{ erie eee sre Te itd y ,
Rete) MMM MEET Ri Rbk eral ye 4 fd Wee
| ie ti Fi MU es ul
1 s
f _ >
’ Py i
- i? “
acnt ( ny
. " t a i
rh bie ae
Pa 50h ;
5) id wus
NM ae nae '
yah by i wu .< . 3 ’
ST Ae 4
Es! vi Th The ra i

Mark. — Lepidosteus,

cap.

c.-t. cp.
Sus. mat.
gran.

m py. can.
m py. el.
nl.

nl. gran.

“ oF

ABBREVIATIONS.
Head of villus. pd. Stalk of villus.
Connective-tissue corpuscle. rz. Root of villus.
Maturation spindle. st. vil. Villous layer of egg membrane.
Granulosa. th. fol. Membrana propria of theca folliculi.
Micropylar canal. vac. Wacuole.
Micropylar cell. vit. Vitellus.
Nucleus. vs. g. Germinative vesicle.
Nucleus of granulosa cell. z.r. Zona radiata.

PLATE IV.

Radial section through the micropyle and micropylar funnel, showing the
micropylar cell and a portion of the maturation spindle of an egg
“stripped” from the fish, preserved in 0.5 per cent chromic acid
(5 hours) followed by washing in water, and hardened in alcohol.
Stained in picrocarmine. X 516.

The second section preceding that shown in Figure 1, and passing nearly
through the middle of the maturation spindle. X 515.

View of the animal pole of an egg preserved in Merkel’s fluid. The
germinal disk was rather more than half as broad as the diameter of
the egg, and its outline should have been represented more distinctly
by the lithographer; it was of a yellowish color, but much lighter
than the rest of the egg. The micropylar funnel is seen exactly over
the centre of the disk. XX about 10.

View of the micropylar funnel and contained micropylar cell of the egg,
a section of which is shown in Figure 1. X 158.

Radial section through the micropylar region and germinative vesicle of
an ovarian egg preserved in alcohol. A portion of the granulosa still
adheres to the outer surface of the villous layer. Stained in alcoholic
borax-carmine. X 158. ;

eee

AACE

B. Meisel, lith

>;

wer fir IVI 4

3 ‘ os: ;

vor A xi Ben

a f

See Se ae " high bo haerh

Wy: o.. Si oa he nin BN han!

oe a Eee : Psi pardayit: nae
Mod Sulla nad i Witiys Pit ae, tea Se
urns ta aL; Yaests ee are
(an fo, RANE RL Yt Ora) me :

He etnies es!) ; hiralany

Fabian arn ‘ ere phi mabyiainy’ oo a he

a Frei fc

Wie Aven hye ni ber Ap iory Ape) ict Taltiaal
yy: Weaevu 4) hoi prt yO CRM jails th tt cnt’
Ane ae) Al Cape ght inhi Wily anne
{ met “.¢ Tes f t eA) Daa Chive aan es Be | Fa ’ ase
la West us ia Ve Mabe | iy wae ss Oe ee he at

2 A eh yithada is Pt foe ‘ph vite Senate AAO Worst na)

M 7 .
aR xe ne’

BOY paiy CS ce tei, Eadie

Annie are TR Blan) ihe | <a) ¥ dtabliy reat ry

jubeat Weal apat niente 2H femalpty Lies 5a
Pek y. Hie) Sea M4 a. ONES ey Phim Lica de I ae

6 we

Mark, — Lepidosteus.

ABBREVIATIONS.
cap. Head of villus. pd. Stalk of villus.
ct. cp. Connective-tissue corpuscle. rz. Root of villus.
Jus. mat. Maturation spindle. st. vil. Villous layer of egg membrane.
gran. Granulosa. th. fol. Membrana propria of theca folliculi.
m py. can. Micropylar canal. vac. WVacuole.
m py. cl. Micropylar cell. vit. Vitellus.
nl. Nucleus. vs. g. Germinative vesicle.
nl. gran. Nucleus of granulosa cell. z.r. Zona radiata.

PLATE V.

Fig. 1. Radial section through germinative vesicle and micropylar funnel. Ow-

&e

ing to a distortion of the section, the curvature of the part of the
membrane shown is less than it should be.- The finely granular and
vacuolated portion of the yolk (vac.) beneath the germinative vesicle
is in the centre of the egg. The egg was from an ovary which was
hardened in 0.25 per cent chromic acid and stained in alcoholic borax-
carmine. X 72.

2. The section following that shown in Figure 1, more highly magnified.
xX 335.

3. Radial section through villi and granulosa of an ovarian egg preserved in
alcohol and stained in alcoholic borax-carmine. XX 515.

4. Surface view of a portion of the granulosa from the same egg as that of
Figure 8. X 516.

e!

MaRK—LEPIDOSTEUS. | | Pay

at

>.

B. Meisel, lith.

daft 4 hy S30 Bh fy, .

’

mgottiy. . Ajests | Ne ee
Ve We aii aye Dare tae WAN PAN Sofi 0
ay BPM WiIATY i say bs
i } ce
iene Meo a,
i ua dd awit Aas 9

" ra "

> ,
i irri anos ® aay,

SG yneed Ry Vides | oO ew iM ey

Hii fp i hie

g
|

oe wih ete tek

‘a Thy i Uivten:
x.) wy ‘Pole "

Mark, — Lepidosteus.

ABBREVIATIONS.

cap. Head of villus.

c.-t.cp. Connective-tissue corpuscle.
Jus. mat. Maturation spindle.

gran. Granulosa.

m py. can. Micropylar canal.

m py. cl. Micropylar cell.

nl. Nucleus.
nl. gran. Nucleus of granulosa cell.

pd. Stalk of villus.

rz. Root of villus.

st. vil. Villous layer of egg membrane.

th. fol. Membrana propria of theca folliculi.
vac. Wacuole.
vit. Vitellus.

vs. g. Germinative vesicle.

zr. Zona radiata.

PLATE Vi

All figures on this plate are magnified 515 diameters.
Fig. 1. Radial section through the micropylar canal, somewhat oblique to its

axis.

The egg was preserved in 5 per cent potassic bichromate and

stained in carminic acid dissolved in 80 per cent alcohol.

“ 2-4,

Three successive tangential sections through the bottom of the micro-

pylar funnel and the micropylar canal of an egg stripped from the
fish preserved in 90 per cent alcohol, and stained in alcoholic borax-

carmine.

“ 5-8.

The zona radiata is closely enveloped by the yolk.
Tangential sections through the deeper portions of the micropylar funnel

of an ovarian egg hardened in 0.25 per cent chromic acid and stained

in alcoholic borax-carmine.

In Figures 5 and 6 the sections pass

through the deep portion of the zona radiata which is not infolded to
form the funnel, but in Figures 7 and 8 only that portion of the zona
is cut which projects as a conical elevation into the substance of the
yolk. Only alternate sections were drawn.

Put

4 i] Ny
WY y yt ji
. iy | Lid
LVF fh hy 4
i ee ,
Ps i 4! VW
al aly
"|, i
Gy fi

\

a

e-77t V. CAML,

a,”

: : B. Meisel. lith
E.L Mark, del.

.
)
.
&
.
'
a ¢
>
u
g
=
‘
:
{
Sc tts,
i -
‘
{
= i
-
a f
>
iy
Ab
j
'
’
2
.
*
.
=! 5 -
'
i t.

A aeRO el

BMOT OKT: oie

Pag

EW te
PA)? ue

jae’ Fh
mah Betal em ba ts
w Depo ey aren praly, a, nq
Rita, 0) 8
vs ‘
2 TAN
Morey BYE whit) 7¢
f Mab un mates

oid oil
ony. Wx) ie

a- -)

Mee) Oho! ea
ey) fi!
on Sit

ss Na sued a1

A ane Tea ic be

eke

a ae "Let dort he
Ae Olde hE
ean am 4 have

od wie eA ay)

ee ables heat)
“i a ie Ca bar ys peumigerty,)
Dah ye HOMD ty,

ga th eed

Miser tals ld

tae HG

“ ul j
fer Se UAT b sc Tye PAN

Cy Ce AY sed)
Pea Y NT tay een

mart

MARK. — Lepidosteus.

cap.
C.-t.cp.
Sus. mat.
gran.

m py. can.
m py. cl.
nl.

nl. gran.

Fig. 1.

ABBREVIATIONS.
Head of villus. pd. Stalk of villus.
Connective-tissue corpuscle. rz. Root of villus.
Maturation spindle. st. vil. Villous layer of egg membrane.
Granulosa. th. fol. Membrana propria of theca folliculi.
Micropylar canal. vac. Vacuole.
Micropylar cell. vit. Vitellus.
Nucleus. vs. g. Germinative vesicle.
Nucleus of granulosa cell. z.r. Zona radiata.

PLATE VIL.

Radial section through the granulosa plug which fills the micropylar fun-
nel. From an ovarian egg preserved in 0.25 per cent chromic acid,
and stained in picrocarminate of ammonia. X 515.

Micropylar cell and outlines of the egg membranes in the region of the
micropylar funnel, from an ovarian egg preserved in alcohol. Radial
section. X 515.

Four spermatozoa, dried on the slide. X 472.

Micropylar funnel ; optical cross-section as seen from the yolk side of the
egg membranes; showing the oval form of the funnel which is
sometimes met with. > 515.

Section of an ovarian egg through the germinative vesicle. Only one
membrane besides the granulosa present; it is the villous layer.
Preserved in 0.25 per cent chromic acid (48 hours). Stained in
alcoholic borax-carmine. X 158.

E.L.Mark, del.

we

SS

Ie

“

B.Meisel, lith

Mann det

a) AR UR een Apap 9

PT saatii Meat ;
( Nudity Ne ent v1 Rp Ra anat hh Pic ne
histo Pre WhO? Ss ee Se eat Ris
He Yr, ae teen Whey | Py ie ae wer!
vi | yO ee we Mr des tut

given) v- -4F y' ied Yu!

se die tale” Bide Lh ; Ml

Se mie a aa ky CAVA LA

yn ‘ 4
bi dendan Wee mY Ao a ee ae fea

ye KS (20gpaj

tally pi (PR se
A ae dae = £
‘ with: arlbiws ¢ dhe # BARIE : eh ol Ly
Aga Sa il) J ea fio ne y 2 Mf mPa
Potion sky distin ve. hes. ie Et leek a
vty Aad oh chetys ely PUA IReCLS aie Yel ai ye eas
bss Ww Mya eld ly ert TO ee att \ NAyEIe
50 ety CU ie) babi '
py # ea) es Nha
Gas AR SHge
.

hile CU ee DF) iat rte a) |i
yh ets Rt We a Wty bay)
Beak? y rainy, ‘ he ha S asin!
Dial Mevliiania vith Wo vd
"pha Bed! 40 hin espe

Hott: | aa Bk

4

Lira fl ir

Mark. — Lepidosteus.

ABBREVIATIONS.
cap. Head of villus. pd. Stalk of villus.
c.t.cp. Connective-tissue corpuscle. rz. Root of villus.
fus. mat. Maturation spindle. st. vil. Villous layer of egg membrane.
gran. Granulosa. th. fol. Membrana propria of theca folliculi.
m py. can. Micropylar canal. vac. Vacuity.
m py. cl. Micropylar cell. vit. Vitellus.
nl. Nucleus. vs. g. Germinative vesicle.
nl. gran. Nucleus of granulosa cell. z.r. Zona radiata.

PLATE VIII.

Figures | and 2 are from sections of an ovarian egg about 0.4 mm. in diameter
which was hardened in chromic acid. X 510.

Pig. o,
‘“ 2.
‘cc 3.
“<< 4,
“e 5.

Part of the peripheral portion of a radial section in which the earliest ob-

served trace of the villous layer has made its appearance. The
membrana propria of the theca and the follicular epithelium are arti-
ficially separated from the yolk and villous projections.

Tangential section from the same egg. The section embraces connective-

tissue cells of the stroma, as well as follicular epithelium, and has also
cut off a portion of the periphery of the yolk, with its villous pro-
jections, which last give it a dotted appearance. The nuclei of the
epithelium are often lobed. vac. indicates vacuities evidently due
to depressions in the surface of the yolk, not to vacuoles in its
substance.

Portion of a section which, owing to the wrinkled condition of the surface

of the egg, affords a surface view of the granulosa, as well as a radial
section and surface view of the villous layer. Some of the detached
villi are seen at one side. The nuclei of the granulosa cells still have
irregular lobed forms. Chromic acid preparation of an egg about
0.6 mm. in diameter. X 510.

View of the villi as seen from the surface of an egg after it has lain for

some time in water. 472.

Amber-colored bodies found at the outer surface of the villous layer of

the egg membrane. X 472.

S
:
ae :
: a
2
=
i
f
|
1
rd —
ci :
eI ¢
=
4
fx}

.

e4)) WPatidden at hn

we

A, ha
re a
iabiby } ap . } :
ane De) Sew ii- 4
Paitin Oe
Aisle ac WO ue i,

at eye 4
Vy fi
era 2) ALS ny il)

LD Me

bitte
;
pills

ie TVR A

a

ioe

i ee

ed TU il hh

‘

"th a Ait Ae tried4 Lo

‘ ae ralegay) if ests MU DD Eos Wire vey Tk |
i a ae } i.
y i? 4

Pes lagiia’) $4

4
Ne AP s4 ea rl -

Pete te

4 / i
LB a Pup

4

any lis if

kl ee
OYA

Aja

Snes

pat,
‘a Ao ein]

riding UR |
pila kA | Yetr i
be ?
i \ ms,"
Pay te al io 7) ab

ad

ni Fy yy

oni

b

ee) ae : lates di Tie ane

yet

hed hs

AY,

(Al - ‘ A

Nanas
Hive i

Wied

|

eV

hala aus vi pias

Mark. — Lepidosteus.

ABBREVIATIONS.
cap. Head of villus. pd. Stalk of villus.
ct. cp. | Connective-tissue corpuscle. rz. Root of villus.
Sus. mat. Maturation spindle, st. vil. Villous layer of egg membrane.
gran. Granulosa. th. fol. Membrana propria of theca folliculi.
m py. can. Micropylar canal. vac. Vacuole.
m py.cl. Micropylar cell. vit. Vitellus.
nl. Nucleus. vs. g. Germinative vesicle.
nl. gran. Nucleus of granulosa cell. z.r. Zona radiata.

PLATE IX.

Fig. 1. Section of an ovarian egg about 0.6mm. in diameter through the ger-
minative vesicle. The villous layer is at all points in contact with
the yolk; but it is separated from the granulosa at intervals. The
egg was hardened in 0.25 per cent chromic acid and stained in alco-
holic borax-carmine. X 168.

2. Portion of a radial section through a mature ovarian ovum, hardened in
0.25 per cent chromic acid, showing the penetration of the roots of
the villi into the pore-canals of the zona radiata. X 516.

« 3,4. Radial sections of ovarian eggs preserved in alcohol, showing stages
in the formation of the villous layer. The eggs were somewhat more
than 0.5 mm. in diameter, and were stained in alcoholic borax-car-
mine. X 510.

« 5. A portion of Figure 1 enlarged. The outlines of the granulosa cells,

especially on the side toward the villi, are much too sharp. X 515.

Pie

a

MARK —LEPIDOSTEUS.

B. Meisel, lith

E.L.Mark, del.

No. 2.— On the Logg Membranes and Micropyle of some Osseous
Fishes. By Caru H. EIGENMANN.!

At the suggestion of Dr. E. L. Mark, I undertook the study of the
development of the micropyle and egg membranes in some of the bony
fishes.

The eggs of the following species were examined: Amiurus catus,
Tachisurus sp. (?), Catostomus teres, Notemigonus chrysoleucus, Caras-
sius auratus, Clupea vernalis, Alosa sapidissima, Fundulus heteroclitus,
F’. diaphanus, Apeltes quadracus, Pygosteus pungitius, Lepomis mega-
lotis, Morone americana, Esox reticulatus, Anguilla anguilla rostrata,
Cyclogaster ? lineatus, Gadus morrhua, and Hippoglossoides platessoides.
In many of these species the eggs were not in a condition favorable
for tracing the development of the micropyle or even the membranes.
My account will be confined to the eggs of Amiurus catus, Notemi-
gonus chrysoleucus, Clupea vernalis, Fundulus heteroclitus, Pygosteus
pungitius, Perca americana, Morone americana, Esox reticulatus, and
Cyclogaster lineatus.

I am indebted to Dr. Mark for the use of his manuscript abstracts
of the papers on egg membranes published before 1881.

It has long been known that fish ova are provided with a membrane,
the zona radiata. The eggs of certain fishes have, in addition to and
outside of the zona radiata, a second membrane which bears in some
cases long filaments, in others short processes which serve to attach the
eggs to foreign bodies.

Fundulus heterochitus and F. diaphanus.

The fact that the eggs of some fishes are provided with long filaments
was first noted by Haeckel (’55). He found them on the eggs of many
species of Scomberesocide, but mistook their position, describing them
as thin fibres lying inside the egg membrane (zona radiata). A con-
nection of the fibres with cells could not be traced.

1 Contributions from the Zodlogical Laboratory of the Museum of Comparative
Zodlogy at Harvard College, under the Direction of E. L. Mark. — No. XVI.

2 Liparis of authors.
VOL. XIX. — NO. 2, 9

130 BULLETIN OF THE

KGlliker (’58) corrected the mistake made by Haeckel as to the posi-
tion of the filaments.

Hoffmann (’81) found filaments on the eggs of Heliasis, Gobius, and
Blennius.

Ryder (’82°) described the filaments of the eggs of Belone longirostris,
and, in passing, mentioned the probability of their existence in the eggs
of Mugil albula. He afterwards (’86*) found them on the eggs of Fun-
dulus heteroclitus, and has also (’83) shown that the eggs of Menidia
(Chirostoma) notata are provided with four of them.

I have examined eggs of Fundulus heteroclitus at intervals of about
two weeks from October, 1887, till June, 1888. The eggs undergo
scarcely any change between October and April. A series taken between
April 1st and June 1st shows all the stages covered by the longer period.
The filaments can best be studied in fresh material. They appear in
the form of hyaline threads, which are more highly refractive than any
other part of the egg membrane. In an ovary of October 27 there were
filament-bearing eggs in three stages of development.

In the smallest eggs— about 0.16 mm. in diameter —in which
filaments can be seen (Plate I. Fig. 1) they appear as hyaline dots, or as
conical bodies with rounded bases, uniformly distributed over the entire
surface. They either lie wholly below the granulosa, or the tips of the
longer ones may lie in between the granulosa cells (Fig. 8). In this
stage the diameter of the threads is much greater than the thickness of
the membrane, which can scarcely be distinguished in sections. I was
not able to discover sheaths enveloping the filaments such as Haeckel
describes for the Scomberesocide. In other slightly larger eggs belong-
ing to this same stage of development (Figs. 2-4, 6) the filaments are no
longer conical, but appear in the fresh condition as short, curved threads
equally blunt at both ends.

In the second stage, the eggs being intermediate in size between those
just mentioned and the largest, the threads (Fig. 5) are of about the
same thickness as those of the earlier stage, but they are much longer,
and taper near the free end. They do not seem to be closer together
than in the smaller eggs. The filaments are bent in a more or less
regular manner, first to one side and then to another. On stained sec-
tions it was to be seen that the threads usually follow the margins of the
granulosa cells, and that they are correspondingly curved (Fig. 6).

On the largest eggs — about 0.4 mm. in diameter — the filaments are
much longer, and cover about as much of the surface of the egg as they
leave exposed. They are so long and so tortuous that it is almost

MUSEUM OF COMPARATIVE ZOOLOGY. 131

impossible to follow a single filament throughout its whole length. It
often happens (Fig. 10) that several filaments are parallel to each other
for a considerable distance. In sections the filaments are found to lie
in between the bases of the granulosa cells, and also to rise between
these cells nearly to their outer surfaces.

In the ripe ovarian eggs the basal ends of the filaments pass directly
through the granulosa layer, and the greater part of the filament thus
comes to lie between the outer portions of the granulosa cells, or even
quite outside of them (Fig. 9). The regularity of their windings can-
not be seen as well as in eggs of the third size. The filaments are of
varying lengths, but most of them are several times as long as the
diameter of the egg. The distances between filaments are not materially
altered during the growth of the egg; but since the surface of the egg
increases during its development to many times the size which it had
when the filaments first appeared, the total number of the latter must
also be greatly increased. The earliest stages in the formation of new
filaments would be difficult to find after the egg has reached its second
stage, because they would be hidden by the larger filaments.

In ripe eggs forced from the ovary, the filaments extend out from the
egg for some distance, and then form a network, several filaments deep,
over the whole surface (Fig. 11).

Concerning the origin of the filaments it may be said that they do
not have any connection with the granulosa cells at any stage of their
growth (Figs. 3, 4, 6, 8). In tangential sections it is seen that they
arise at places corresponding to the boundaries between two or three
cells. In a ripe egg examined in the fresh state under pressure (Fig. 12)
indistinct processes are seen to radiate from the base of each filament,
forming a stellate figure. In no case, either in fresh specimens or sec-
tions, could the filaments be traced into the substance of the zona radi-
ata. They are outgrowths from a thin membrane which lies outside of
the zona and is formed before the latter, not processes of the zona itself.

When the filaments first make their appearance, the egg membrane,
as stated above, is much thinner than the diameter of a filament, and
the granulosa cells are lens-shaped, barely touching by their mar-
gins (Fig. 8). In the largest eggs found in the ovary of April 2d, the
granulosa was about 8 p» thick, but the egg membrane had only reached
the thickness of 2 uw. That it is radially striate is rather to be in-
ferred than directly seen. In places the outer surface shows slight
elevations at regular distances, which I believe to be prolongations of
granulosa cells sunk into the radial canals (Fig. 13). While the largest

132 BULLETIN OF THE

eggs of April 2d were only about 0.4 mm. in diameter, and therefore
scarcely exceeded in size those of October and November, the largest
ovarian eggs of May 2d measured over 0.8 mm. Between May 1 and
June 1,—by which time the eggs have reached their full size, —the
growth is still more rapid. ‘The egg membrane of early May eggs meas-
ures about 6.5 » in thickness, and has distinct pore-canals.

There exists an exceedingly thin outer membrane overlying the zona
radiata. It was discovered in the examination of fresh ripe eggs, in
which the striation of the zona itself could be seen much better than
in sections of hardened eggs. In one instance, in which the zona of a
fresh ripe egg was ruptured, this overlying membrane was left intact.
It is with this membrane that the bases of the filaments are continuous.

In view of this condition in Fundulus, and of the fact that other pro-
cess-bearing eggs (Cyprinids and Gasterosteide) possess a thin outer
membrane, it would be interesting to re-examine the eggs of the Scom-
beresocide, of Menidia, and of Mugil to find whether they do not also
possess this structure.

The outer surface of the fresh ripe egg of Fundulus heteroclitus shows
a network of lines (Fig. 7). This appearance is doubtless due to the
presence of superficial ridges, which in radial sections have the appear-
ance of minute projections fitting in between the bases of the granulosa
cells (Fig. 9). Where two or more lines meet, there is a thickening.
The whole arrangement bears a superficial resemblance to the appear-
ance presented near the surface of the zona in the perch (Fig. 31). In
the case of the latter, however, the thickenings correspond in position
to filaments, each of which corresponds to the middle of a granulosa
cell, whereas in Fundulus heteroclitus the thickenings correspond in po-
sition to the boundaries between granulosa cells. From the position of
the filaments in Fundulus it is probable that, like the ridges, they are
outgrowths of the outer structureless egg membrane. It is evident from
what has been said that there is a fundamental difference between the
filaments found in Perca and those in Fundulus. In Perca they owe
their origin to the granulosa, and are formed after the zona has nearly
reached its full growth; in Fundulus, on the contrary, they owe their
origin to the activity of the egg itself, and they begin to be formed
before the zona.

Pygosteus pungitius.

After Haeckel had described the long filaments peculiar to the eggs
of the Gobiesocide, Kolliker (58) described external appendages in the

MUSEUM OF COMPARATIVE ZOOLOGY. 135

eggs of Abramis brama, Chondrosteus nasus, Squalinus argenteus, Cobitis
barbatula, Gobio fluviatilis, Cyprinus rufus, and Gasterosteus pungitius,
In all these species he found the appendages inserted in a very thin
membrane, which ultimately lies just outside the zona radiata and which
makes its appearance before the latter.

The most important paper on Pygosteus is that of Ransom (’67). He
studied Gasterosteus pungitius and G. leiurus, and found that the eggs
of the two species do not differ greatly. He says that in the oviduct the
eggs are surrounded by a viscid layer, and that the zona radiata lies be-
low this layer. The zona is in contact with the yolk except in ripe eggs,
in which a thin homogeneous membrane covers the yolk and follows the
constrictions at the time of cleavage. The micropyle and the dotted
appearance of the egg membrane were first made out in eggs 74,5! thick,
and in eggs 53,’ in diameter the membrane could be separated from the
yolk. The button-shaped processes can be made out in eggs somewhat
less than 74,5!’ (0.17 mm.) in diameter. They are attached to the outer
surface of the yolk-sac by a bright, highly refractive point. In the case
of the smallest ova there are on an average seventy buttons, in that of
the largest two hundred and seven. ‘They serve to attach the egg to
foreign substances. tansom describes and figures the micropyle.

Owsjannikow (’85) found that in ovarian eggs the granulosa cells cover
the micropyle. In fully grown eggs only a single membrane is present,
while in the younger ones the zona seems differentiated into two layers,
owing to the fact that the zona is laid down by successive additions.
The pores do not appear till the membrane has attained considerable
thickness, and they are then much finer than in the ripe egg. The
mushroom-shaped processes are maintained by him to be cells that pos-
sess nuclei which are colored red with carmine. From the base of the
process a thread can be traced into the zona radiata. In young eggs
the processes consist principally of a nucleus attached to a filament. He
does not believe that they are derived from the zona, but thinks they
come from the granulosa ; why he thinks so is not stated. Inside the
zona he has found the zonoid layer of His.

I have examined ovarian eggs of fishes taken in November, December,
and April. A few days after the spawning, in early April, the ovaries
contain a considerable number of eggs (about 0.55 mm. in diameter) in
which the formation of the yolk is well advanced. ‘These are evidently
destined to be laid before the recurrence of the next annual spawning
season, for they are much larger than any of the ovarian eggs found in
December. These eggs show no signs of degeneration, and their pres-

134 BULLETIN OF THE

ence can therefore hardly be explained as due to their failure to pass off
with the first lot of eggs laid ; nor can they be eggs which properly belong
to the first set of spawn, as their size in comparison with that of the
mature eggs (1.1 mm.) sufficiently proves. Therefore I believe that, as
Ransom has inferred, these fishes deposit eggs more than once during
the season.

The ovaries are most available for study after the first set of eggs are
deposited. As in the case of Fundulus heteroclitus, all stages of growth
are shown in ovaries taken during a period extending from one or two
months before till a short time after spawning, —the months of March
and April.

In eggs 0.15 mm. (Plate I. Fig. 14) or more in diameter there are
two membranes,—an outer more highly refractive, and an inner stri-
ated one. In many sections the two are artificially separated (Fig. 15*).
In ripe eggs the outer membrane had either entirely disappeared, or its
structure had become so much like that of the true zona that the two
could not be distinguished from each other. Their total thickness is
from 15 tol8 p. In many sections of ripe eggs an outer layer, much
thicker than the outer layer seen in the earlier stages of development,
was in places separated from the rest of the zona. If it represents the
outer membrane of the earlier stage, then the latter must undergo a
great change in its later development, for it is now much thicker, and
is traversed by the same pore-canals as the deeper portion.

The rivet-shaped processes which are found in the region of the
micropyle are inserted, as Kélliker says, in the thin membrane which
lies outside the zona, and which is formed before the latter makes its
appearance. They take a much deeper stain than the thin membrane,
but I have seen nothing which would warrant one in claiming that
they contain each a nucleus. The smallest egg in which these processes
could be seen had a diameter of about 0.14 mm.; only a single thin,
structureless membrane was to be made out in this stage. The largest
eggs examined had a diameter of about 1 mm.

When the processes make their appearance, the granulosa is so thin
that it is difficult to determine from surface views whether they lie
above or below it ; but radial sections show that they lie below. There
is no such constant relation between the processes and individual cells
of the granulosa as to suggest the origin of the former from the latter ;
but at a later stage the heads of the rivets occupy nearly the same
plane as the nuclei of the granulosa cells (Fig. 16), and therefore appear
to have an intimate connection with the granulosa cells. When the

MUSEUM OF COMPARATIVE ZOOLOGY. 135

granulosa is torn from the egg membranes, as, owing to the shrinkage of
the egg, it frequently is, the processes no longer show the same sharp
outer margins. Their edges are often frayed, and are not stained as
deeply as when the granulosa and the membranes are in their normal
relations to each other. With the separation of the granulosa the thin
outer membrane is sometimes torn (Fig. 15°); and whether torn or
not, it is often separated from the inner membrane. This may be due
to the fact that the processes are from the beginning adhesive, and
have thus acquired an intimate secondary relation to the cells of the
granulosa. In such sections it can be clearly seen that the rivet-shaped
processes are joined to the outer membrane and not to the zona, though
their bases have projected into the zona for a greater or less distance.
When the granulosa is torn from the egg membranes, the processes
always, even in the smallest eggs, remain attached to the membranes
rather than to the granulosa. I have been able to find neither the
nuclear structure within nor the prolongations from these processes
which Owsjannikow has described.

I have not succeeded in finding the micropyle in eggs that were much
less than 0.4 mm. in diameter; in such the zona has an average thick-
ness of about 5. The portion immediately surrounding the micro-
pyle shows a considerahle local thickening. Owing to the variation in
the thickness of the zona in different regions of the same egg, and to the
inconstancy of the position of the micropyle in relation to this varia-
tion, it sometimes happens that the zona at the micropylar region has
already reached a thickness of 10 or 11 yu.

It is a noticeable fact, that at this earlier stage the micropyles of
nearly all the eggs were cut radially when the sections were made
in planes perpendicular to the axis of the ovary. Furthermore, the
micropyles uniformly lie in the half of the egg opposite the side of
attachment.

In the vicinity of the micropyle the zona becomes thickened by the
elevation of its outer surface, the deeper surface undergoing no change of
direction. At a distance of about 10 w on either side of the micropylar
canal it attains its greatest thickness, and then its outer contour curves
inward until it becomes continuous with the wall of the micropylar
canal. The inner end of the canal is sometimes slightly enlarged
(Figs. 19-22).

At this stage the pore canals of the zona radiata do not seem to be
modified in direction in the region of the micropyle; they are all
radially arranged. The outer membrane could not be distinguished in

136 BULLETIN OF THE

this region ; it probably is entirely wanting in the area immediately sur-
rounding the micropyle. The granulosa cells are two or three layers
deep in the vicinity of the micropyle, and a single cell larger than the
others is always to be found directly above the canal. It usually sends
a prolongation into the canal itself (igs. 18, 21).

In eggs about to be laid, the greatest thickness of the zona in the
vicinity of the micropyle is approximately 24 yw, and the thickening in
this region is not so conspicuous as at the earlier stage. The zona bends
inward slightly, so that its inner surface no longer forms a simple
curve. The micropylar passage through the zona presents three re-
gions: a shallow funnel-shaped depression, which occupies the outer
third of the layer ; a narrower tubular portion, which is a prolonga-
tion of the bottom of the funnel, and is rounded at its lower end ;
and finally a very narrow canal, which traverses the inner sixth or
eighth of the zona, and opens at the apex of the low elevation of the
inner surface (Fig. 18).

The outer or funnel-shaped portion is wholly filled even at this
advanced stage by the single large micropylar cell which was seen at
the earlier stage (Figs. 18, 21).

Perca.

The egg of the perch has been a favorite subject for study. Almost
every writer on teleostean ova has examined it. Von Baer (’35, pp. 6, 7)
first described it as having a double membrane, the outer portion being
traversed by long narrow dark spots (“‘ dunklern Flecken ”).

Miiller (’54) gives a fuller account. He separates the membrane into
an inner, the zona radiata, and an outer, the capsule. The outer sur-
face of the zona is described as being covered with exceedingly small
cylindrical projections. These are doubtless nothing but the elevations
between the pore-canals, which are rather wide on the outer half of the
zona. The capsule is radially traversed by small spiral tubes, which
are enlarged and funnel-shaped at both ends. ‘Transverse filaments are
sometimes seen between these radial tubes. On applying pressure, yolk
granules were forced into the spiral tubes, but in no case was any yolk
matter forced between the tubules ; from which he concludes that the
capsule must be closed between them.

Kdlliker (58) discusses the origin of the “tubules.” He considers
them to be outgrowths from the follicular cells, and the substance between
them as a secretion from those cells. He denies the statement made by
Miiller, that they are hollow, but has seen the anastomosing filaments

MUSEUM OF COMPARATIVE ZOOLOGY. 137

described by him. The tubules are independent of their jelly matrix,
and in chromic acid preparations they can be separated from the latter.
When the eggs are deposited, the granulosa cells probably fall off, leav-
| ing shallow depressions having polygonal outlines, from the centres of
which “tubules ”’ arose.

Ransom (’68) described the canals passing through the outer portion
as having a double contour for each wall, and as filled with material
containing vacuoles; but they do not seem to him to convey anything
either fluid or solid into or out of the egg. This outer layer is separable
only by tearing it from the yolk-sac (zona), and does not leave a dis-
tinct outline. The tubes divide at their inner terminations into branch-
like roots, and adhere closely to the zona radiata. The internal ends are
not expanded as Miiller described, and it is rarely that filaments pass
from one to the other. He supposes that the granules seen by Miiller
were only vacuoles. The eggs when deposited are arranged in the form
of hollow tubes with the micropyles all turned to the inside.

His (73) mentions having seen the micropyle, but neither figures nor
describes it.

Brock (’78) describes the zonoid layer, and finds its striations inter-
mediate in fineness between those of the villous layer and those of the
zona. Judging by his drawing of Alburnus lucidus there are about three
striations in the zonoid layer to four in the zona. The latter, he says,
makes its appearance before the villous layer.

Hoffmann (’81) finds that in October the zona and the villous layer
are of equal thickness. The latter is said to be composed of numerous
small projections which correspond exactly to the villi of the Cyprinoids.
At the free ends of the villi lie the granulosa cells. In February the
zona is differentiated into two layers, of which the inner is four times as
thick as the outer. There arise from the outer layer long fibres with
triangular bases and with their distal ends expanded to form a continu-
ous layer on which the granulosa cells rest. Each filament corresponds
to, but is not a process of, a granulosa cell.

Owsjannikow (’85) recognizes the usual divisions of the egg membrane.
The contents of the distal ends of the filaments are granular, which has
given rise to the belief that they are nuclei. The filaments end ex-
ternally in funnel-shaped enlargements described by Miiller. He suc-
ceeded in forcing granular matter from the yolk into their deep ends.
The latter divide and enter the pores of the zona, through the whole
thickness of which they can be traced. He states (p. 7) that, contrary
to Hoffmann’s belief, the filaments are derived from the granulosa. In

138 BULLETIN OF THE

a subsequent part of his paper (pp. 29-31), where he gives an account of
the development of the ovarian egg, his statements seem to be conflict-
ing as to the relation of the spiral canals to the granulosa cells, but at
the end he repeats that the canals are outgrowths of cells as stated by
Kolliker, The interstitial matter (Zwischensubstanz) is arranged in
lamellze which are parallel to the surface of the egg. By the swelling
of the lamelle fissures arise which have the appearance of processes
from the canals.

I have studied the ovarian eggs of Perca killed in October, February,
and May. It is probable that the formation of the egg membranes is
less advanced in the American species of this latitude than in the Euro-
pean species at a corresponding season. ,

Contrary to Hoffmann’s statement that in October the capsular layer
and zona are of equal thickness, not a trace of the capsular layer, dis-
tinct from the granulosa, could be found at this time of the year. The
zona is well developed, and is differentiated into two layers of about
equal thickness. The outer layer is radially striate, while the inner
appears to be structureless. The granulosa cells lie immediately in
contact with the zona radiata (Fig. 23, Plate III.). I have not been
able to find the micropyle in October eggs.

In February the zona remains practically as it was in October, but
vacuoles — which may be caused by the method of treatment — are to
be seen in the inner portion (Fig. 25, Plate III.). They are much flat-
tened radiaily, and thus suggest an approach to a stratified condition of
this portion of the zona. The radial striations of the outer half of the
zona are more strongly marked than at the earlier stage, and much
fainter striations may also be seen traversing the inner half. The lat-
ter, though less distinct, are just as numerous as, and continuous with,
those of the outer half. At this date the capsular layer is already well
developed, but it has attained only half the thickness which it has in
May.

Up to the month of May the thickness of the zona radiata has not
changed, but the pore-canals can now be more readily traced passing
entirely through it. They still remain much more evident in the outer
than in the inner half of the zona. This is due to the greater calibre
of the canals, not to their being farther apart in the outer half.

The different descriptions of the capsular layer are in part due to the
fact that it presents different conditions according to varying circum-
stances. The radially arranged spiral structures traversing this layer
arise as funnel-shaped tubules, one beneath each cell of the granulosa.

OO

MUSEUM OF COMPARATIVE ZOOLOGY. 139

In the early stages of their development the tubules have a more or
less spiral course, while in the later stages they become more nearly
straight. In February eggs (Fig. 25, Plate III.) their inner ends are
slightly expanded, and terminate in a thin structureless film overlying
the zona. In radial sections of eggs taken in May, they often appear
triangular at the base, and their contents divide into branches which
enter the pores of the zona. The “filaments ” connecting the canals
are sometimes much more abundant than at others. In the vicinity of
the micropyle one finds on tangential sections (Fig. 31, Plate II.) that
the tubules at or near their bases are joined to each other by what
appear like slender filaments, but these may be the cut edges of nearly
perpendicular membranes. This results in the production of an irregu-
lar network with meshes of variable size and shape, at the angles of
which the spiral tubules are located.

The micropyle was seen in eggs taken in February and in May.
Immediately surrounding it, the zona radiata is thickened by a slight
elevation of its internal surface (Fig. 24, Plate III.). The micropyle con-
sists of a funnel-shaped opening in the zona with the wide end directed
outward. In some cases the inner end of the canal also flares slightly.
In a February egg in which the micropylar region was somewhat distorted
(Fig. 26, Plate II.) the micropyle seems to have been composed of two
regions, separated from each other by a distinct shoulder, the inner
end of the outer portion being much wider than the outer end of the
inner portion. The granulosa cells and their tubules are greatly crowded
above this region (Fig. 24, Plate IIJ.). At some distance on either
side of the micropyle it is to be seen that the outer funnel-shaped ends
of the canals begin to be more elongated than in other parts of the egg,
and continue to increase in length up to the micropyle. The nuclei of
the granulosa cells, which are situated near the bottom of the funnel-
shaped expansions, also become more and more elongated as one ap-
proaches the centre of the micropylar region, and at the same time they
come closer to the zona radiata. The effect of this is to produce in radial
sections through the micropyle the appearance of an immense funnel-
shaped depression in the whole capsular layer (Fig. 24). But the ap-
pearance is misleading ; there is no such broad depression ; the granulosa
cells of this region extend outward beyond their nuclei until they reach
the theca folliculi at the same level that the neighboring cells do. The
thickness of the capsular layer is therefore not changed in the vicinity
of the micropyle, and the theca folliculi does not bend inward, but
stretches over this region with a uniform curvature. The granulosa

140 BULLETIN OF THE

cells stain more deeply than the inter-tubular substance of the capsular
layer. This peculiarity is very serviceable when one is searching for
the micropyle. Notwithstanding the absence of a broad depression,
there is a narrow irregular canal left in the centre between the modified
granulosa cells, which can best be seen upon sections tangential to this
part of the egg. (Figs. 27-32, Plate Il. Compare Explanation of Fig-
ures.) The appearance is similar to what one might imagine would
result if the central cell of this region had dropped out of its original
place. That such a cell has not wholly disappeared, but has simply
lost its peripheral connection with the wall of the theca, is rendered
probable by the presence of a peculiar cell at the bottom of this canal.
Directly over the micropyle, in contact with the zona and filling more
or less completely its micropylar depression, lies a single cell of large
size. Its nucleus is more nearly spherical than the nuclei of the other
cells, and it is not stained as deeply as they are. (Fig. 24, Plate III. ;
and Figs. 26, 31, Plate II.) There can be no doubt that it is a
peculiarly modified granulosa cell.

Morone americana.

The egg membrane of the white perch has never been described, but
Ryder (82) has described the micropyle.

There is only a single membrane, the zona radiata, but it is composed
of two distinct layers, both of which are traversed by pore canals. ‘The
eges examined were taken from fishes caught in February, April, and
May. In February the ovary contained eggs in four stages of develop-
ment; in the older stages there are well developed membranes. Eggs
of 0.16 mm. in diameter have a single homogeneous membrane 1.2 p
thick. When they have reached a diameter of 0.28 mm. the zona is
composed of two layers (Fig. 33, Plate II.), a very thin inner and a
thicker outer one; together they measure 39 p in thickness. By the
time the eggs have reached a diameter of 0.40 mm. (Fig. 34, Plate II.)
the total thickness of the membrane is more than doubled ; that of the
outer layer is 49 » and that of the inner 39 yp. The outer layer is
formed first and takes a deeper stain. It does not increase much in
thickness after the appearance of the inner layer, and in the older eggs
it contains vacuoles. The inner is at first apparently homogeneous, but
with its great increase in thickness there appear in it the radial stria-
tions characteristic of the zona. The granulosa cells are small and low,
and have flattened nuclei situated in the middle of the cell.

MUSEUM OF COMPARATIVE ZOOLOGY. 141

Esox reticulatus.

The egg membrane of Esox was first described by Aubert (53). He
says the shell of the egg is a thin, transparent punctate membrane, which
closely envelops the yolk and in sections exhibits radially placed streaks.
After lying in water some time, an outer very thin granular membrane
makes its appearance.

Lereboullet (54) describes two membranes, the outer of which is
pierced by microscopic tubes. ‘The inner is a simple extremely thin and
amorphous envelope, which has no homologue in the perch.

Reichert (’56, p. 94) states that the membrane discovered by Aubert
surrounding the zona radiata is to be found on all eggs of this species,
but that it is in the fresh condition entirely homogeneous.

Kdélliker (58, pp. 84 and 85) maintains the existence of a thin outer,
resistant layer in all fish eggs, and was able to isolate it in fresh eggs of
Esox.

Ransom (’68) says that in Esox the egg membrane is similar to that
of Gasterosteus ; he also, as I think erroneously, supposes the thin
outer membrane to be homologous with the “ Eikapsel” of the perch.
He figures the micropyle. |

Finally, His (73) described for the zona radiata concentric as well as
radial strize.

The eggs examined by me were taken from the ovary in February.
Leaving out of consideration the smallest eggs, 0.063 mm. and less in
diameter, which have no membrane except the granulosa, the ovary
contained eggs in three stages of development, respectively about 0.50,
1.00, and 1.50 mm. in diameter. In eggs of the first stage the zona
radiata is about 3 » thick and very faintly striate. There is no evidence
of its being differentiated into concentric layers. At the micropyle (Fig.
35, Plate III.) it reaches a thickness of 7». Very generally the yolk is
more or less retracted from the zona by the action of the hardening re-
agents, so that a narrow space, which varies a good deal in thickness over
different parts of the egg, is left between the two structures. Spanning
this interval are numerous fine threads, which have the appearance of
being prolongations of the substance of the yolk continued into pore-
canals of the zona. This is a condition which remains at subsequent
stages, and will therefore be discussed further on. The granulosa cells
are still thin, and their nuclei much flattened.

In the second stage (Fig. 36, Plate III.) the zona has a total thickness
of 11 or 12 p, and is distinctly differentiated into two layers, the outer of

142 BULLETIN OF THE

which is only about one fifth as thick as the inner. The latter is faintly
stained, and distinctly striate radially ; the outer is deeply stained, and
striations are usually not to be seen in it, but on favorable sections, espe-
cially such as are very thin, the striations may frequently be made out
to pass continuously through the whole thickness of both layers. Upon
this point there is not the least doubt, so that it is certain the outer layer
in question is truly a part of the zona, and I have been unable to find in
ovarian eggs any membrane intervening between this and the granulosa
cells. In sections of the micropylar region, the inner portion of the zona
radiata exhibits vacuoles elongated in the direction of the pore-canals.
In this region the latter are not strictly radial, but converge towards the
outer end of the micropylar canal. Inside the zona there is a region to be
seen which bears some resemblance to a membrane with coarser (more
distant) striations than those of the zona. It varies in thickness on dif-
ferent parts of the egg, and corresponds, I believe, to the sub-zonal space
seen in the eggs of the first stage ; but it may represent the zonoid
layer of His.

The membranes of eggs of the third or oldest stage (Fig. 37) differ
somewhat from the conditions just described. The vacuoles of the zona
radiata, found in the second stage near the micropyle only, are here
found over all portions of the egg; they are always most numerous near
the inner surface, and are not found at all in the outer fifth of the
membrane. They are more or less regularly arranged in series parallel
with the surface of the zona. Kdlliker (’58, p. 84) attributed the pres-
ence of such vacuoles in the pore-canals to the effect of fresh water on
the zona.

The granulosa cells in the second and third stages have nearly spheri-
cal nuclei, which lie at their distal ends (Fig. 37, Plate III.). Below
the nuclei, tapering columns of granular protoplasm extend to the zona
radiata. These columns are separated by less deeply stained tracts of
substance, but the boundaries of the columns are not sharply marked.
The appearance is as though the columnar cells were being gradually
metamorphosed into an intercellular substance. This condition is evi-
dently an approach to that found in Perca.

The micropyle was found in eggs of both the first and second stages.
In the first stage (Fig. 35, Plate III.) the zona is twice as thick around
the micropyle as in other regions. This thickening results in a consid-
erable elevation of the inner surface of the zona, the outer surface being
only very slightly changed. The micropyle is a wide canal, the outer
third of which tapers rapidly and is continuous with the inner two

MUSEUM OF COMPARATIVE ZOOLOGY. 1438

thirds, which taper only slightly from without inward. The micropylar
canal is partially filled with a plug of substance which appears to be con-
tinuous with the yolk. The granulosa cells overlying the micropyle do
not appear different in size from those which envelop the rest of the egg,
but a single cell is sometimes seen to overlie the micropyle in addition
to the regular layer of granulosa cells. In the second stage (Fig. 36,
Plate III.) the micropylar canal is narrower than in the first; it no
longer tapers gradually from the outside inward, but is slightly nar-
rowed at two points, one near the outside and one at its deep end. By
the retraction of the yolk from immediate contact with the zona near
the micropylar canal in the case of one of the eggs, a space was formed
through which could be traced a cord of substance continuous with that
which occupied the canal itself. The portion of the substance which trav-
ersed this space was funnel-shaped, with the wide end next the yolk. The
thickness of the zona does not now differ so greatly in different regions as
at the first stage. At some distance from the micropyle in the egg last
mentioned (Fig. 36), the inner surface of the zona was raised rather
abruptly ; nearer the micropyle it was slightly depressed, but the mar-
gin of the canal was raised in the form of a Jow cone, which thus occu-
pied the centre of a very shallow inverted crater, the rim of which was
formed by the outer circular elevation. Above the micropyle in the
granulosa was a large spheroidal space nearly filled with a granular mass
somewhat denser than the yolk. The mass was slightly contracted, leav-
ing a narrow space at its periphery. I am in doubt whether to regard it
as a cell or not, since no nucleus could be detected. On both sides of this
granular mass there were several highly refractive homogeneous bodies
(Fig. 36, 2). It is however doubtful if they have any significance in
relation to the micropyle. The granulosa cells at this stage are tall
and have elongated nuclei, which are broad at the exterior end, and
taper towards the egg membranes.

Notemigonus chrysoleucus.

The ovary of this species contained ova in four stages of development
on May 9th. In all but the smallest eggs the zona radiata was present.
The largest had a diameter of 0.6 mm., and the zona varied gradually
from a thickness of 2 » on one side of the egg to that of 4 » on the
opposite side. The pore-canals are very fine, being almost invisible in
balsam preparations.

The micropyle was observed in only a single case; it was found in
the middle of the thickest portion of the membrane, which is exactly in

144 BULLETIN OF THE

the middle of the attached side of the egg. The direction of the inner
surface of the zona was not altered in the vicinity of the micropyle
(Fig. 38, Plate II.), but its outer surface exhibited a broad circular
depression, by which the thickness of the zona was diminished about one
half. The micropyle proper at the centre of the depression appeared
as a narrow canal of uniform calibre. Between the zona and the yolk
there was a narrow space, probably formed by the contraction of the
yolk ; beneath the micropylar region, this space was abruptly enlarged
into a hemispherical depression. Across this space the radial strands of
protoplasm characteristic of almosteall the spaces between the zona and
the yolk were plainly visible.

The granulosa, which over all other parts of the egg is composed, as
usual, of a single layer of cells, is thickened in the region of the micro-
pyle. As the direction of the long axes of their oval nuclei show, the
cells near the margin of the micropylar depression in the zona have
their peripheral ends inclined toward the axis of the micropyle. The
cells which fill up the depression have larger and more elongated nuclei,
and the obliquity of the latter has become so great that the depression
appears to be filled with a granulosa layer two or three cells deep. It
would seem that in this case the single micropylar cell found in other
eggs was represented by a number of enlarged granulosa cells.

Clupea vernalis.

The chief interest in the egg membranes of this species centres in the
presence of a thin, highly refractive, structureless membrane overlying
the zona radiata of eggs in an advanced stage of development (Fig. 39,
Plate III.). This outer membrane is intimately connected with the gran-
ulosa cells, so that it usually retains its connection with the granulosa
when the latter is artificially separated from the zona. In all such cases
slender striations extend from it to the zona radiata. The appearance of
these markings is such as to show clearly that they are prolongations of
the substance of the outer membrane, and there can be little doubt that
the projections penetrate the pore canals of the zona radiata, from which
they are partially withdrawn by the artificial separation of the two mem-
branes. This structural condition suggests an explanation for simi-
lar appearances below the zona radiata in Esox and other fishes, and
between the zona radiata and an inner layer in Amiurus (Fig. 45) and
Ictalurus. It will be more fully discussed later.

i i

MUSEUM OF COMPARATIVE ZOOLOGY. 145

Cyclogaster lineatus (Liparis lineatus Auct.).

The ovaries of this species contain ripe eggs in May, the time at
which I examined them. The largest eggs were about 0.63 mm. in
diameter, the membrane averaged about 0.043 mm. in thickness. The
zona radiata seems to be filled with small spaces connected by the much
finer radial canals (Fig. 40, Plate III.), the spaces causing the latter to
appear moniliform. Near the inner and outer margins of the zona the
canals are.simple tubes, as in most other fishes.

The eggs next in size are 0.25 to 0.30 mm. in diameter ; their zona is
always only half as thick on one side as it is on the opposite, the change
in thickness being nowhere abrupt. In eggs of this stage the zona is
traversed by simple pore-canals, which are indistinct near its outer sur-
face. In some cases (Fig. 41, Plate III.) the transition from the inner
to the outer portion is so abrupt that the zona appears to be composed
of two layers of unequal thickness, —an outer, thinner, more nearly
homogeneous and unstained, and an inner which is thicker, more dis-
tinctly striate, and usually faintly stained.

The micropyle (Figs. 42-44, Plate III.) was observed only in eggs
about 0.16 mm. in diameter. As in the case of Pygosteus it seems to’lie
in a plane perpendicular to the long axis of the ovary. The micropyle is
a long narrow tube, with parallel sides, in a local thickening of the zona.
This increase in the thickness of the zona affects the outline of the inner
surface more than that of the outer, and is entirely independent of the
above mentioned gradual change in thickness between opposite poles of
the egg. It is produced principally by additions to the inner surface of
the zona. The outer surface is slightly elevated at a little distance from
the micropyle, but is abruptly depressed immediately over it. The reg-
ularity in the arrangement of the nuclei of the granulosa cells is dis-
turbed in the immediate vicinity of the micropyle, where the whole layer
is slightly thickened. Usually an enlarged single nucleus lies immedi-
ately above the micropylar canal (Fig. 44).

On the Number of Egg Membranes.

The views held concerning the number of egg membranes in teleosts
have been many and various. Authors have generally been agreed about
the presence of a membrane perforate with radial canals, the zona ra-
diata ; but doubts have been raised by Ryder whether this membrane
is always present. He (’82°) found no striations in the egg membrane of

VOL. XIX.—NO. 2. 10

146 BULLETIN OF THE

Belone longirostris or (’84, p. 457) the cod, and states (’85, p. 145) that |

the eggs of Gambusia patruelis do not possess any membrane. I have
found striations in the membrane of the fresh cod egg. It may be stated
here that the striations of the zona sometimes show plainest in fresh
eggs, sometimes not until reagents have been applied. Haeckel says, for
forms related to Belone longirostris, that the membrane is structureless,
but that it is covered with minute black dots. These dots were doubt-
less pore-canals seen from the surface. The zona radiata of Osmerus
eperlanus was found by Buchholz (’63) to consist of an inner and outer
portion, joined together in the micropylar region only. On deposition
of the eggs the outer membrane is turned wrong side out, and serves
to attach the eggs to foreign substances. These conditions have been
redescribed by Cunningham (’86). Hoffmann (’81) found that the zona
is differentiated into two layers in all adhesive eggs, the outer portion
being ultimately transformed into a viscid mass.

Ryder (’86) describes a peculiar arrangement of the egg membranes
of Ictalurus albidus. He says: “'The egg-membrane is double, that is,
there is a thin inner membrane representing the zona radiata, external to
the latter and supported on columnar processes of itself, which rest upon
the inner membrane ; there is a second one composed entirely of a highly
elastic adhesive substance. The columns supporting the outer elastic
layer rest on the zona, and cause the outer layer to separate very dis-
tinctly from the inner one.” I have found similar conditions in Amiurus
catus (Plate II. Fig. 45), but am inclined to think that the two mem-
branes represent the outer and inner portions of the zona radiata; for
the outer shows the striations peculiar to the zona, and the columnar
layer is of varying thickness, The inner membrane, being closely asso-
ciated with the yolk, would cling to it when the yolk contracts; the
protoplasm in the pore-canals being partially withdrawn would give rise
to these columnar processes. Where the two membranes were separated
for a considerable distance, the columnar structure was destroyed. Simi-
lar conditions obtain in the eggs of Clupea vernalis, but in this case the
columnar structures lie between the zona radiata and a thin owter homo-
geneous layer which is in contact with the granulosa. There cannot be
the least doubt concerning the meaning of the columnar layer in Clupea
vernalis, for the two membranes lie directly in contact in some parts of
the egg. The peculiar structures in Ictalurus and Amiurus doubtless
have an origin and meaning similar to that of Clupea.

The eggs of all the species of fishes examined by me possess a perfo-
rate zona radiata. The radial strize could never be made out on the

—————————e ee

MUSEUM OF COMPARATIVE ZOOLOGY. 147

first appearance of the egg membrane. The absence of striz in these
younger eggs may be accounted for by assuming, as Reichert has sug-
gested, that the zona radiata is a later growth, and that the imperforate
membrane of younger eggs is a different structure, or that during the
earlier stages the material composing the membrane is less dense, al-
lowing the food material to have ready access to the yolk. Granting
for the moment that the zona grows by apposition of layers from within,
the latter view is the more probable, because in the perch the inner
portion of the zona is not perforate even after the outer is distinctly so,
and in most cases the pore-canals are much more distinct and wider in
the outer than in the inner portion of the zona. The meaning of the
pore-canals, in the intra-ovarian egg at least, needs little discussion. In
most of the sections prepared, where the granulosa cells are slightly
raised from the zona radiata, processes of the granulosa cells can be seen
to enter the pore-canals.

Various membranes have been described for different fishes as over-
lying the zona radiata. The peculiar capsular layer of the perch has
been seen by all authors who have examined the eggs of this fish. It
was first described by Von Baer (’35).

Rusconi (’36) describes a thin membrane overlying the ovarian egos
of Cyprinus. Aubert (’53) saw an outer membrane on eggs of Esox
which had lain in water some time. Kdlliker (’58) succeeded in isolat-
ing this membrane in the case of Esox.

Reichert (56) discovered that whenever processes are present, as
in many cyprinoids, they are set in a thin outer membrane. KOl-
liker (58) confirmed this statement, and added that this outer mem-
brane is developed before the zona radiata. Reichert also found that
the membrane of the smallest membrane-bearing ovarian eggs is not
striate, and concluded that the zona radiata must be a secondary for-
mation.

Vogt (42) was the first to describe a membrane within the zona
radiata. He found that in the eggs of Coregonus palea and Salmo
umbla this membrane cannot be readily seen until after the eggs have
been in water for some time, and that it passes (p. 29) gradually into
the germ. Ransom (’56) found a similar structure in eggs of Gaste-
rosteus pungitius, in which this inner membrane takes part in cleavage.
Eimer (’72*) claims to have isolated this vitelline membrane, which he
saw in trout, pike, white-fish, and perch. Oellacher (72) also succeeded
in separating it in the brook trout. I believe that the structures de-
scribed by Vogt, Ransom, Eimer, and Oellacher are, as others have

148 BULLETIN OF THE

maintained, not to be considered as vitelline membranes, but as the
superficial part of the protoplasm of the egg.

His (’73) found that the cortical layer of the yolk in many ovarian
eggs is more finely granular than the rest of the yolk, and that it is
radially striate. This outer portion of the yolk he called the zonoid
layer. Many others have seen similar structures. According to the
accounts of some authors the zonoid layer is found only in eggs which
are not mature, and even then it is not always present.

The condition of the egg membranes in Amiurus and in certain stages
of Esox has suggested the idea that similar appearances may in some
cases have given rise to a belief in the existence of a zonoid layer when
there really was none. A partial withdrawal of the egg protoplasm
occupying the pore-canals produces an appearance which at first sight
suggests the presence of a striate membrane internal to the zona; in
fact, I at first supposed it to be a distinct membrane, and was the more
easily misled because in some cases it seems to be of nearly uniform
thickness. However, more careful study showed that it was not a mem-
brane, and that the appearance was due to fine threads of highly refrac-
tive substance stretching across a space between the inner surface of the
zona and the yolk. There are two things especially which make it im-
possible for me to believe that this is a normal condition: the great
variability in the thickness of the supposed membrane in different parts
of the same egg, and the fact that the radial striations are due to a sub-
stance which is more highly refractive than the substance, if any, filling
the intervening spaces. If, on applying reagents, there is great contrac-
tion of the yolk, either it is torn from the protoplasm in the pore-canals,
or the protoplasm contained in the pore-canals is suddenly withdrawn
from them and distorted; in either case, there would be no appearance
of a zonoid layer. If, however, the protoplasm should not be withdrawn
from all the pores, but should in the case of many remain stretched
across the space between the zona and the yolk, as might no doubt fre-
quently happen, we should find the supposed zonoid layer more coarsely
striate than the zona, a condition described by recent authors. Such
an origin of the zonoid layer might also explain its absence in ripe eggs.
After the egg has attained its full size, the connection of the yolk sub-
stance with the canals would naturally be less intimate than at an earlier
stage, and then a contraction of the yolk would not be accompanied by
the stretching of any filaments across the space thus produced.

Scharff (87 and ’87*) has recently described, within the zona radiata
in young eggs of Trigla, a zonoid layer, which subsequently disappears.

MUSEUM OF COMPARATIVE ZOOLOGY. 149

The eggs examined by me may be divided as follows : —

I. Eggs with a single membrane, the zona radiata.
a. Zona radiata a single layer of uniform structure. Notemigonus
chrysoleucus, Carassius auratus.
aa. Zona radiata differentiated into an inner and outer layer. Morone
americana, Esox reticulatus, Cyclogaster lineatus, Amiurus catus.

II. Eggs with a zona radiata and a thin homogeneous outer layer.
6. Outer membrane without appendages. Clupea vernalis.
6b. Outer membrane bearing filiform appendages. Fundulus hetero-
clitus, F. diaphanus.
66b. Outer membrane .with short appendages. Pygosteus pungitius.

III. Eggs with a zona and a thick outer layer produced, by a secretion
from and metamorphosis of the granulosa cells. Perca americana.

Origin of the Egg Membranes.

Concerning the origin of the different egz membranes of fishes several
views have been held.

Vogt (42) and Vogt and Pappenheim (’59) maintained that the zona
radiata is formed by the compression of a layer of cells surrounding the
egg; Reichert (’56), Kolliker (58), Gegenbaur (’61), and Eimer (’72°),
that it is derived from the yolk ; Thomson (’59) and Waldeyer (’70), that
it is derived from the follicular epithelium ; Ransom (67) argued that it
cannot grow by apposition of layers from within or without, and that it
must grow by interstitial deposition of material. Whether this material
comes from ingoing or outgoing currents, he was unable to determine.

I think that the zona is undoubtedly derived from the yolk. Kélliker
found that in all the filament-bearing eggs studied by him the zona
radiata was formed after the filament-bearing membrane. I have found
the same to be true in Fundulus. In the case of Morone the outer layer
of the zona does not become much thicker after the inner layer has begun
to be formed, whereas the latter continues to grow rapidly.. In the case
of Cyclogaster lineatus, where the outer layer of the zona shows columnar
structures, these do not bear any definite numerical relation to the over-
lying cells of the granulosa. The outer portion of the zona is almost
always more uniform in its structure, and stains deeper, than the inner
portion.

Reichert (56) and Kélliker (’58) are inclined to believe that the cap-
sular layer of the perch is derived from the granulosa cells, an opinion

150 BULLETIN OF THE

with which Hoffmann (’81) does not agree. It certainly does not make
its appearance till after the zona is well developed ; if it were derived
from the yolk, its substance would first have to traverse the zona radiata.
How the nourishment for the egg could pass into the latter through the
pore-canals, and the formative substance of the villous layer at the same
time pass out through them, is scarcely conceivable. Moreover, at the
distal end of each of the villi lies the nucleus of a granulosa cell, there
being as many villi as there are cells, a fact which proves beyond a
doubt the intimate relation of the two structures.

The membrane just external to the zona in Clupea vernalis may be
considered homologous to that in Gasterosteus, Fundulus, and many
Cyprinoids, even though it does not in Clupea bear appendages as it
does in Gasterosteus. From the development of the appendages in
Fundulus and Gasterosteus it is evident that this membrane has no
connection with the granulosa cells. In these cases each of the appen-
dages does not correspond to a single cell as in the perch, nor to any
definite number of cells. If Reichert is correct in saying that the
homogeneous membrane found in young eggs is a different structure
from the zona radiata, the membrane under consideration may perhaps
be looked upon as the primitive membrane described by him. It is
certain that it appears before the zona, and I am inclined to think that
it is derived from the yolk.

CAMBRIDGE, December, 1888.

MUSEUM OF COMPARATIVE ZOOLOGY. 151

BIBLIOGRAPHY.

Aubert, Hermann.
53. Beitrage zur Entwickelungsgeschichte der Fische. Zeitschr. f. wiss.
Zool., Bd. V. Heft 1, pp. 94-102, Taf. VI. 16 Aug., 1853.

Baer, Karl Ernst von.
35. Untersuchungen tber die Entwickelungsgeschichte der Fische, nebst
einem Anhange tiber die Schwimmblase. 52 pp., 1 Taf. u. mehreren
Holzschn. im Texte. Leipzig: Vogel. 1835.

Brock, J.
'78. Beitrage zur Anatomie und Histologie der Geschlechtsorgane der Kno-
chenfische. Morph. Jahrb., Bd. IV. Heft 4, pp. 505-572, Taf. XX VIIL.,
BAX. 1878.

Buchholz, Reinhold.
’63. Ueber die Mikropyle von Osmerus eperlanus. Arch. f. Anat., Physiol.
u. wiss. Med., Jahrg. 1863, pp. 71-81, Taf. III A., Figs. 1-4. 18638.
63%. Nachtragliche Bemerkungen iiber die Mikropyle von Osmerus eper-
lanus. Arch. f. Anat., Physiol. u. wiss. Med., Jahrg. 1863, pp. 367-372,
Taf. VIII A. 1863.

Cunningham, J. T.
’86. On the Mode of Attachment of the Ovum of Osmerus eperlanus. Pro-
ceed. Zool. Soc. London, for 1886, Pt. III. pp. 292-295, Pl. XXX. (Read
May 4.) 1886.

Eimer, Th.
'72". Untersuchungen iiber die Hier der Reptilien. Arch. f. mikr. Anat.,
Bd. VIII. pp. 216-243, 397-434, Taf. XI., XII., XVIII. 1872.

Gegenbaur, Carl.
61. Ueber den Bau und die Entwickelung der Wirbelthiereier mit partielle
Dottertheilung. Arch. f. Anat., Physiol. u. wiss. Med., Jahrg. 1861, pp.
491-529, Taf. XI. 1861.

Haeckel, Ernst.
’55. Ueber die Hier der Scomberesoces. Arch. f. Anat., Physiol. u. wiss.
Med., Jahrg. 1855, pp. 23-31, Taf. IV., V. 1855.

152 BULLETIN OF THE

His, Wilhelm.
'73. Untersuchungen tber das Ei und die Entwickelung bei Knochenfischen.
Leipzig: F.C. W. Vogel. 1873. 4 + 54 pp., 4 Taf., 4to.
Hoffmann, C. K.
’81. Zur Ontogenie der Knochenfische, Verhandl. d. koninkl. Akad. v.
Wetenschappen, Amsterdam, Deel XXI., 164 pp., 7 Taf. 1881.
Kolliker, Albert von. .
'58. Untersuchungen zur vergleichenden Gewebelehre, angestellt in Nizza im
Herbste 1856. Verhandl. physical.-med. Gesellschaft in Wirzburg, Bd.
VIII. pp. 1-128, Taf. J.-III. 1858.
Lereboullet, Auguste.
54, Résumé d’un Travail d’Embryologie comparée sur le Développement du
Brochet, de la Perche et de PEcrevisse. Ann. Sci. Nat., sér. 4, Zool.,
Tom. I. pp. 237-289. 1854.
Miiller, Johannes.
’54. Ueber zahlreiche Porencanile in der Hicapsel der Fische. Arch. f. Anat.,
Physiol. u. wiss. Med., Jahrg. 1854, pp. 186-190, Taf. VIII. Figs. 4-7.
1854.

Oellacher, J.
'72. Beitrage zur Entwickelungsgeschichte der Knochenfische nach Beo-
bachtungen am Bachforelleneie. Zeitschr. f. wiss. Zool., Bd. XXII. Heft 4,
pp- 378-421, Taf. XXXIT., XXXIII. 20 Sept., 1872.

Owsjannikow, Ph.
85. Studien tiber das Hi, hauptsachlich bei Knochenfischen. Mém. Acad.
Imp. Sci. St. Pétersbourg, sér. 7, Tom. XXXIII. No. 4, 54 pp., 3 Taf.
1885.
Ransom, W. H.
56. On the Impregnation of the Ovum of the Stickleback. Proceed. Roy.
Soc. London, Vol. VII. pp. 168-172. 1856.
’'67. On the Structure and Growth of the ovarian Ovum in Gasterosteus lei-
urus. Quart. Jour. Micr. Sci., n. ser., Vol. VII. pp. 1-4, Pl. 1. Jan.,
1867.
’68. Observations on the Ovum of Osseous Fishes. Philos. Trans. Roy. Soc.
London, Vol. CLVII. Pt. II. pp. 431-502, Pls. XV.-XVIII. 1868.

Reichert, Karl Bogislaus.

56. Ueber die Micropyle der Fischeier und iiber einen bisher unbekannten,
eigenthiimlichen Bau des Nahrungsdotters reifer und unbefruchteter
Fischeier (Hecht). Arch. f. Anat., Physiol. u. wiss. Med., Jahrg. 1856,
pp. 83-124, 141, 142, Taf. IL, TII., und IV. Figg. 1-4. 1856.

Rusconi, R.

’36. Ueber die Metamorphosen des Hies der Fische vor der Bildung des
Embryos. Arch. f. Anat., Physiol. u. wiss. Med., Jahrg. 1836, pp. 278-
288.

MUSEUM OF COMPARATIVE ZOOLOGY. 153

Ryder, John A.
’81e. Development of the Spanish Mackerel (Cybium maculatum). Bull.
U. S. Fish Commiss., Vol. L. pp. 185-172, 4 pls. [1881] 1882.

’82. The Micropyle of the Egg of the White Perch. Bull. U. 8S. Fish
Commiss., Vol. I., p. 282. May 2, 1882.

82°. Development of the Silver Gar (Belone longirostris), with Observations
on the Genesis of the Blood in Embryo Fishes, and a Comparison of Fish
Ova with those of other Vertebrates. Bull. U. 8. Fish. Commiss., Vol. I-
pp. 283-301, Pls. XIX.-XXI. May 2 and 19, 1882.

’83. On the Thread-bearing Eggs of the Silversides (Menidia). Bull. U.S.
Fish Commiss., Vol. III. pp. 1938-196. 1883.

’84. A Contribution to the Embryography of Osseous Fishes, with special
Reference to the Development of the Cod (Gadus morrhua). Ann. Report
U. S. Commissioner of Fish and Fisheries for 1882, X VIL. pp. 455-605,
Pls. I.-XII.

"84°. Also separate, with title-page and cover. 149 pp., 12 pls. Washington:
Government Printing Office. 1884.

’85. On the Development of Viviparous Osseous Fishes. Proceed. U. S.
National Museum, Vol. VIII. Nos. 8-10, pp. 128-155. Pls. VI.-XI.
25 May, 1885.

’86. On the Development of Osseous Fishes, including Marine and Fresh-
Water Forms. Extracted from Ann. Report U. 8. Commissioner of Fish
and Fisheries for 1885. pp. [1]-[116], Pls. l-XXX. 1886.

"86%, The Development of Fundulus heteroclitus. American Naturalist,
Vol. XX. p. 824. Sept., 1886.

’87. [Same as RyprR, ’86.| Ann. Report U. S. Commissioner of Fish and
Fisheries for 1885, pp. 484-604, Pls. 1-XXX. 1887.

Scharff, Robert.
’87. On the Intra-ovarian Egg of some Osseous Fishes. (Rec’d Nov. 17,

1886.— Abstract.) Proceed. Roy. Soc. London, Vol. XIV. No. 249,
pp. 447-449. 1887.

"87%. On the Intra-ovarian Egg of some Osseous Fishes. Quart. Jour.
Mier. Sci., Vol. XXVIII. pp. 538-74, Pl. V. Aug., 1887.
Thomson, Allen.
59. [Article] Ovum iz The Cyclopedia of Anat. and Physiol., edited by
Robert B. Todd, Vol. V. (Suppl. Vol.), 1859, pp. 1-80 and [81]}-[142].
Note. —Part I., pp. 1-80, was issued in 1852; Part IL., pp. [81]-[142], in
1855.
Vogt, Carl.
’42. Hmbryologie des Salmones. Neuchatel. 1842. 6 + 328 pp., 8vo.
Avec Atlas, fol. obl. de 7 pls.

Being Tome I. of Lu. Agassiz, Histoire Naturelle des Poissons d’Eau douce de
? Europe Centrale.

154 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

Vogt, Carl, et S. Pappenheim.

- '59. Recherches sur l’Anatomie comparée des Organes de la Génération chez
les Animaux Vertébres. (Déposé dans les Archives de l’Acad. le 30 Dec.,
1845.) Ann. Sci. Nat., sér. 4, Zool., Tom. XI. pp. 331-369, Pl. XTIT. ;
‘Tom. XII. pp. 100-131, Pls, IL, Ill, 1859.

Waldeyer, Wilhelm.

'70. Lierstock und Ki. Ein Beitrag zur Anatomie u. Entwickelungsgeschichte

der Sexualorgane. Leipzig: W. Engelmann. 1870. 8+ 174 pp., 6 Taf.
S8vo.

EXPLANATION OF FIGURES.

ABBREVIATIONS.

cp. Blood corpuscles. prj. tcl. Intercellular ridges.

Jil. Filaments of Fundulus. spa. Space below micropyle.

fil. vt. Filaments of vitellus. tbl. Tubules of the capsular mem-
gran. Granulosa. brane in Perca.

m py. Micropyle. the. fol. Theca folliculi.

m py. cl. Micropylar cell. vac. Vacuole.

nl. gran. Nucleus of granulosa cell. yk. Yolk.

nl. m py. Nucleus of micropylar cell. Zi Zona radiata.

po. can. Pore-canals of the zona radiata. z. r.’ Zona radiata externa.

pre. Rivet-shaped processes of zona. z.r.” Zona radiata interna.

All the figures were made with the aid of the camera lucida, and all except
Figs. 1, 2, 5, 7, 11, and 12 from preparations mounted in benzole-balsam. Figs. 39-
41 were drawn by Dr. Mark, and the others by the author.

E1ig=NMANN. — Egg Membranes.

a.

nS,

Pe aberes

1G,
ey

PEATE I;

Figures 1-18 are of Fundulus heteroclitus.

Surface view of one of the smallest filament-bearing eggs of October 27.
Diameter of egg, 0.16 mm. Examined fresh. X 750.

Surface view of another egg of the same size and date, with somewhat
larger filaments. Examined fresh. X 425.

Tangential section of an ovarian egg 0.15 mm. in diameter. The ovary
was hardened, December 28, in Flemming’s chromic-osmic-acetic mix-
ture, and subsequently stained with hematoxylin. X< 425. The section
is seen from its inner surface.

Tangential section of an egg from the same ovary with longer filaments.
x 425. This section is also seen from its inner surface.
Surface view of an egg of October 27, about 0.25 mm. in diameter.

Examined fresh. X 425.

Tangential section of an egg 0.25 mm. in diameter, from the same ovary
as Fig. 3. xX 750.

Surface view of a ripe (June) egg from which the granulosa cells had been
removed, showing the network of ridges between their bases. Exam-
ined fresh. XX 750.

Radial section of the egg represented in Fig. 3. Transsections of fila-
ments are seen at fil. X 425.

Radial section of an egg of May 2, about 0.8 mm. in diameter. Preserved
in Perenyi’s fluid, and stained with picrocarminate of lithium. X 750.

Tangential section of an egg of December 23, about 0.4 mm. in diameter.
From the ovary mentioned under Fig. 38. X 425.

Radial optical section of a ripe egg shortly after being forced from the
ovary (June 1). Examined fresh. X 60.

Base of one of the filaments of the ripe egg. Examined fresh under pres-
sure. X 700.

Radial section of an ovarian egg of November 23. Preserved in Perenyi’s
fluid, and stained with Grenacher’s alcoholic borax-carmine. X 750.

Figures 14-22 are of Pygosteus pungitius, all except Figure 20 being of eggs
from a single ovary, which was cut transversely.

Radial section through an ovarian egg 0.15 mm. in diameter. The ovary
was preserved in Perenyi’s fluid, April 18, shortly after spawning, and
subsequently stained in picrocarminate of lithium. X 750.

Tangential section near the micropyle. %X 112.

Radial section of an egg 0.37 mm. in diameter. X 750.

Radial section of an egg 0.38 mm. in diameter. X 750.

Radial section of an egg 0.87 mm. in diameter. X 750.

“ 18-22. Radial sections through the micropyles of eggs, about 0.4 mm. in diam-

eter. In Figs. 20 and 21 the micropyle is cut obliquely. X 750.

“« 20 is from an ovary hardened April 4, i. e. some time before spawning.

—s. .. @

rt

E1GENMANN-EGG MEMBRANES. Ppl

the. fol.

—~

tj Py eee ; = ees: —s
eee
tobe le ve aa, teed \ \ ar “x Yo “nd Mh py,
: Se j : / | } ‘ \ .
BP Ce Sen es | | en Sena
¢ > a) aS | ! | me py.

——— ee ZT. 7 — °: a
SN Nett Hy, ey ; nN
oe = NE en ET Qed
ia So, grands.) >) ( ee sion Gp etd Cy
ere gs is, OO OG
sa aa 22 rs See) eh) SO
‘ ee Sa . XK ase, a NN es
2 seis — Prt Aes ae | ———.
——— = — - _— — y Sonn AS

f roar a ie oe —— ia ‘a ee ~~
\ / a 4 po . x
\ / = == a., is f td. m. py.
iy SS ar. f :
“UN 77 Se anf

m7 py. 7 a: 7 ST Ss Si. ee ea — °C ir 2 Sa cee ————— a

iF pv
HE, del. B. Meisel, lith

y H

ee

te

a in inal a he

i
mh as is bl f ‘
th . 3) = : 1
‘ 7 ».

é
> As i ‘ ruta
aoe y
Le) ere ad onga® 2
i SS
h ’ “ai
ru AQ =

: pe ie.
he waersy avy) M. ’ Ff : = a

nt 4 i, b dl ide iting A : 7
ae oa as
, watcher wit
‘a ie Wap. iv pe
- 4 a
: ¥
“<a 4 Tay Tye4 r Fix 4 Py we y

. iy
}
P “ é 3 : = ’
& . etry
* &
. M : a vv +
—< os oS , pe
bo » 1 ome Ge lee in; a A 4 a
, ' ; te
ae deere dure bated Gaye M
= ' Ss ,
: i ES erst whe Es at a
os - i a
Meinl: alee ar ia
no . i y Ne
BN OQ Re by iran dh SRR ey | =
a es i as ie SY Pears i - .
es

- a - lh any Pema Wel HH iy } d ; .
‘hs aa BELL 74h Myatt. Sie <<. ') a ad é nt *
= maid nae Basis 5 otal r ; ‘in

ri oF Te Wek lite whic <2) I, i
A's ‘ide Wis, (eiOed ; ~ 7“
Mksts has byes : ; n 7

> RTS 1; 24 Witty: 98 le ae
were ‘2 Fe ry i ee

SSS en gee Pei
eg Sr Wea ‘eagles ie Sure ie '
) |? aD oP it

Vitae we it ; \
Vaid wo 7 '
DR Pe oes th

hort uae A ery Ey Lima aay 3» Teeth) | id un 2
RE ir.) 1, aL oy

Hie ath 0 iy" f.” shave a We ker ve i bale 2, pe fa SAEs
Be. ut, et 14 cist lel ails RANE Pa 2

SO Se GK fuiher sci

a‘ .

EIGENMANN. — Egg Membranes.

‘ ABBREVIATIONS.

cp. Blood corpuscles. pr j.vcl. Intercellular ridges.
Jil. Filaments of Fundulus. spa. Space below micropyle.
Jil. vt. Filaments of vitellus. tbl. Tubules of the capsular mem-
gran. —- Granulosa. brane in Perca.
m py. Micropyle. the. fol. Theca folliculi.
m py. cl. Micropylar cell. vac. Vacuole.
nl. gran. Nucleus of granulosa cell. yk. Yolk.
nl. m py. Nucleus of micropylar cell. Za Tn Zona radiata.
po. can. Pore-canals of the zona radiata. z. r.’ Zona radiata externa.
pr... Rivet-shaped processes of zona. z.7r.” Zona radiata interna.
PLATE II.

Figures 23, 24, and 25 are on Plate III. ; Figures 26-32 are from Perca.

Fig. 26. Radial section through the micropyle of an ovarian egg from Perca killed

in February. X 750.

26%. Section through the micropyle (?) of an egg of Perca. X 750.

27-32. A series of sections tangential to the surface of an egg of Perca 1mm.
in diameter at a point somewhat to the right of the micropyle. The
portion of the sections to the right of the micropylar region lies deeper
than the portion to the left. In

27, the cells lying to the right of the region of the micropyle are crowded, and
have a curved band-like arrangement. The cells which contain dark
points and lie farther to the right are from the central portion of the
section, and are therefore cut across deeper than the others. In the
second section,

28, only the deeper, filamentous prolongations of these cells are seen. In

28, 29, a median pit can be traced through the centre of the column of cells
which occupies the micropylar region. In

30 the nucleus of the enlarged micropylar cell (n/. m py.) is seen. In

31 is seen the enlarged mouth of the mycropyle (m py.) and a few cells which
lie somewhat higher and to the left of it.

82 is a section through the zona radiata and micropylar canal.

This egg was preserved February 15 in Perenyi’s fluid, and was stained
with Czoker’s alum-cochineal. X 425.

33. Radial section of an ovarian egg of Morone americana 0.28 mm. in diam-
eter. Hardened in chromic acid February 25, and stained with picro-
carminate of lithium. %X 780.

34. Radial section of a larger egg (0.4 mm. in diameter) of A/orone americana
from the same series of sections as that of Fig. 83. 760.

38. Radial section through the micropyle of an egg of Notemigonus chryso-
leucus 0.63 mm. in diameter. Ovary of May 5, killed in Perenyi’s
fluid, and stained with picrocarminate of lithium. X 750.

45. Radial section through the egg of Amiurus catus. Ovary of May 9, killed
in Perenyi’s fluid, and stained with picrocarminate of lithium. X 750.
The radial markings have been accidentally omitted and fil. vl. placed
for fil. vt.

———————

Go

ae ey

, B. Meisel, lith
CHE, del.

*

Th HABE pooh
Sd ae
, cyeper 70 i *

F Ae

a is 7 (0 «@
' 7 7 ew, Ve
RR” ie
“ay Le en -
Ain it b's

iat ew Siwess i--

gs ee ot ee RT et fay
., tlegtadeald Oe cdl A oe.) ee

= ae Mihi , 7

EIGENMANN. — Egg Membranes.

m py.

m py. el.
~ nl. gran.
nl. m py.

po. can.

pre.

25.

30.

40.

41.

ABBREVIATIONS.
Blood corpuscles. pr j.vcel. Intercellular ridges.
Filaments of Fundulus. spa. Space below micropyle.
Filaments of vitellus. tbl. Tubules of the capsular mem-
Granulosa. brane in Perca.
Micropyle. the. fol. Theca folliculi.
Micropylar cell. vac. Vacuole.
Nucleus of granulosa cell. yk. Yolk.
Nucleus of micropylar cell. aie Zona radiata,
Pore-canals of the zona radiata. z. r.’ Zona radiata externa.

7

Rivet-shaped processes of zona. <z. r.’ Zona radiata interna.

PLATE III.

Radial section of an egg of Perca in October, 0.5 mm. in diameter. The
ovary was hardened in 0.25 per cent chromic acid, and subsequently
stained with Czoker’s alum-cochineal. X 750.

Radial section through the micropyle of an egg of Perca. The ovary was
preserved in Perenyi’s fluid, May 9, and stained in carminate of
lithium. > 425. The definite line at the outer margin of the zona
radiata should have been omitted.

Radial section of an egg of Perca, 0.9 mm. in diameter. From an ovary
hardened in February. X 425.

Figures 26-54 are on Plate II.

Radial section through the micropyle of an egg of Hsox, 0.47 mm. in
diameter. Ovary of February 23 killed in chromic-osmic-acetic mix-
ture, and stained with picrocarminate of lithium. X 750,

Radial section through the micropyle of an egg of Hsor, 0.94 mm. in
diameter, from the same series represented in Fig. 35. X 750.

Radial section of an egg of Esox, 1.5 mm. in diameter, from the same
series. X 750.

Radial section of an egg of Clupea vernalis, 0.54 mm. in diameter. Pre-
served in Perenyi’s fluid, and stained with picrocarminate of lithium.
x 515.

Radial section through the egg of Cyclogaster liparis, 0.7 mm. in diameter.
Ovary of April 26 preserved in Perenyi’s fluid, and stained with picro-
carminate of lithium. X 5165.

Radial section through the egg of Cyclogaster liparis, from an ovary of
May 7 preserved in Perenyi’s fluid, and stained with picrocarminate
of lithium. X 515.

42,43, and 44. Radial sections through the micropyles of three eggs of Cyclo-

gaster, about 0.25 mm. in diameter. Ovary of May7 preserved in
Perenyi’s fluid. X 515.

_ ELGENMANN-EGG MEMBRANES

the fol... <a

i!

__f ee

B. Meisel, lith

Vv

No. 3. — Report on the Results of Dredging, under the Supervision
of ALEXANDER AGASSIZ, in the Gulf of Mexico (1877-78), and in
the Caribbean Sea (1879-80), by the U. S. Coast Survey Steamer
“ Blake,” Lieut.-COMMANDER C. D. SIGSBEE, U. 8S. N., and Com-
MANDER J. R. BARTLETT, U. 8. N., Commanding.

[Published by Permission of Cartite P. Paterson and J. E. Hinearp, Superin-
dents of the U. S. Coast and Geodetic Survey.]

A XXIE.

Report on the Nudibranchs. By Rup. Bereu.

Wahrend dieser Expedition wurden nur ganz wenige Formen von
Nudibranchien gefischt, aber fast alle neu und darunter noch dazu ein
Paar ziemlich interessante. Diese Formen waren die folgenden : —

. Tethys leporina, L. var.

. Chromodoris scabriuscula, Bgh., n. sp.
Chromodoris punctilucens, Bgh., n. sp.
Chromodoris sycilla, Bgh., n. sp.

. Phlegmodoris? anceps, Bgh., n. sp.

. Nembrotha gratiosa, Bgh., n. sp.

. Phyllidiopsis papilligera, Bgh., n. sp.

SIO OP wD =

Fam. TETHYMELIBIDA.

TETHYS, L.

R. Bergh, Malacolog Untersuch. (Semper, Philipp. I. ii.), Heft IX., 1875, pp. 346-
362, Taf. XLV. Fig. 19-26, Taf. XLVI. Fig. 1-22, Taf. XLVII. Fig. 1, 2.
H. v. Hering, Tethys. Ein Beitrag zur Phylogenie der Gastropoden. Morpholog.
Jahrb., II., 1876, pp. 27-62, Taf. IT.
R. Bergh, Notizen iib. Tethys leporina. Jahrb. d. deutschen Malakolog. Ges., IV.
4, 1877, pp. 335-339.
R. Bergh, Beitr. zur Kenntn. d. Aeolidiaden. VII. Verh. d. k. k. zool. bot. Ges.
in Wien, XXXII., 1882, pp. 67-68.
VOL, XIX.—NO. 3.

156 BULLETIN OF THE

H. de Lacaze-Duthiers, Sur le Phenicurus. Comptes Rend., Cl. I., 1885, pp. 30-35.

RK. Bergh, Sur la Nature du Phenicure. Arch. de Zool., 2 ser., III., 1886,
pp. 73-76. d

H. de Lacaze-Duthiers, Contrib. & l’Hist.du Pheenicure. Arch. de zool., 2 ser., IV.
1887, pp. 77-108, Pl. IV.

List, Zur Kenntn. d. Driisen im Fuss von Tethys fimbriata, L. Arb. aus dem zoolog.
Institut zu Graz, I. 6, 1887, pp. 287-305.

>

Diese merkliche aberrante Nudibranchien-Gruppe ist erst durch die
zwei ersten der obengenannten Arbeiten niher bekannt worden und
diese Kenntniss ist nicht ohne wesentlichen Einfluss auf das Studium
der ganzen Gruppe gewesen.

Die untenstehende Untersuchung hat wesentlich nur dadurch Inter-
esse, dass sie das Vorkommen von einer 7Jethys, der altbekannten oder
einer neuen Form, im westlichen Theile des atlantischen Meeres nachweisst.

T. leporina, L. var.

Tafel I. Fig. 1-3.

Hab. M. atlant. occ. (Dominica).

Von dieser Form wurde ein Individuum in der Nahe von Dominica aus
einer Tiefe von 138 Faden hinaufgefischt.

Das in Alkohol ganz schlecht bewahrte, verdrehte, theilweise erhartete und
Papillenlose Individuum hatte eine Ldnge von 4.3 cm., von welchen die volle
Halfte auf dem Segel kamen, der Querdurchmesser des letzteren 3 cm.; die
Hohe der Rhinophorscheide 7 mm., der Keule 2 mm.; die Lange der Rand-
faden bis 10 mm.; die Lange des Mundrohres 4 bei einem Durchmesser am
Grunde von 3.5 mm.; die Breite des Riickens bis 13 mm.; die Hohe des
Korpers bis 10 mm.; die Lange des Fusses 2.5 bei einer Breite bis fast 2 cm.,
der Vorderrand 7 mm. frei vortretend. — Die Farbe der Aussenseite des colos-
salen Segels ist gelblichweiss wegen dichtgedrangter ganz feiner gelblichweisser
Piincktchen, die gegen den Rand hin zu unregelmassigen Fleckchen fast zu-
sammenfliessen. Die Unterseite des Segels ist hinten kohlenschwarz so wie
auch das grosse Mundrohr (aussen und innen), wird dann in der mittleren
Strecke mehr braungrau, gegen den Rand hin schwarzlich und (theilweise
fleckig) schwarz; die Randfaden des Segels meistens gelblichweiss, der Boden,
auf dem sie sitzen, aber schwarz. Die Scheide der Rhinophorien schwarz,
mit grossen gelblichweissen Flecken; die Keule am Grunde schwarz, sonst
weisslichgelb. Der Riicken und die Korperseiten fast von der Farbe der
Oberseite des Segels, aber mehr gelblich und im Genicke so wie in der Ge-
gend des Riickenrandes starke, grosse, kohlenschwarze Flecken. Die Riicken-
papillen fehlten ganz; die Kiemen (am Grunde der Papillenfacetten) weisslich,
Die obere Seite des Fusses ringsum wie die Korperseiten gefarbt ; der Fuss-
rand weisslichgelb ; die Fusssohle braungrau.

MUSEUM OF COMPARATIVE ZOOLOGY. tov

Der grosse Segel wesentlich wie bei der typischen Tethys des Mittelmeeres;
an der Innenseite der Randparthie stehen die Randfdden in meistens 4-6 (8)
sehr undeutlich geschiedenen Reihen; die aussersten sind ganz klein, die
innersten von bedeutender Lange; die Dorsalen Currhen des Segelrandes kamen
in gewohnlicher etwas sparsamer Menge vor und von einer Hohe bis 2mm.
Das starke Mundrohr am Vorderende (Fig. 1 a) in gewohnlicher Weise ge-
kluftet; der gahnende Aussenmund bis an den Rand und bis in die Tiefe, bis
an die schniirlochartige Pharynxoffnung mit starken Héckerchen, nur ausnahms-
weise reihegeordnet, besetzt. Im Genick, dicht an der Gegend des hinter-
sten Theils des Segels, ziemlich weit von einander stehend, die zusammenge-
driickten, oben etwas breiteren Rhinophorien, deren vorderer Theil oben eine
Vertiefung mit umgeschlagenem Rande tragt, in welcher sich die zurickge-
bogene Keule fand; diese letztere etwas abgeplattet, mit 11 breiten Blattern. —
Die Korperform wie in der typischen Tethys, der Récken nur vielleicht etwas
breiter. Am gerundeten Riickenrande, wie es schien, 7 rundliche Papillen-
facetten gewohnlicher Art, die Papillen selbst aber fehlend (wie so oft bei
Exemplaren von Tethys) ; dicht neben jeder Facette zwei Kiemenbischel, ein
vorderer kleinerer, ein hinterer grésserer; die Kiemenbiischel wie gewohnlich.
Vor der zweiten rechten Papillenfacette die etwas hervorragende Anal-Pro-
tuberanz, neben derselben die Nierenpore. Die Kérperseiten vorne ziemlich
hoch ; aus der Genitaloffnung ragte ein Theil des Penis etwa 2 mm. hervor.
Der grosse Fuss ganz wie bei der typischen Tethys; eine mediane Langsfurche
fehlte nicht hinten an der Sohle.

Die Eingeweidemasse an die Kéorperwande durch Bindesubstanz geheftet.

Das weisslichgelbe Centralnervensystem zeigte die Hauptganglien von ein-
ander viel deutlicher geschieden, als ich es sonst bei Tethyden gesehen
habe, nur zwischen den beiden pleuralen Ganglien war die Grenze undeut-
lich. Die buccalen (vorderen Eingeweide-) Ganglien zwischen dem hinteren
Theile der Speicheldriisen liegend (Fig. 1d), oval, durch eine Commissur
verbunden, die langer als der Querdurchmesser des Ganglions war ; oberhalb
der Wurzel des nach vorne gehenden Nerven fanden sich mehrere Nervencellen
eingelagert (Ganglion gastro-oesophagale). Der Riechknoten am Grunde der
Keule des Rhinophors. Kleine (sympathische) Ganglien kamen an und
zwischen den Eingeweiden zerstreut vor, besonders im Gebiete des Geni-
talsystems.

Die kleinen schwarzen Augen an der Oberflache der Gehirnknoten nach
aussen fast sessil, oval, von 0.12 mm. grésstem Diam., mit gelber Linse, reich-
lichem schwarzem Pigmente und ziemlich grossen Retinazellen. Die Ohr-
blasen als kalkweisse Punkten aussen an der oberen Seite der cerebralen Gang-
lien neben den pleuralen gelagert, kugelformig, ganz kurzgestielt (Fig. 3),
etwas kleiner als die grossen Nervenzellen, von 0.16 mm. Diam., mit zahl-
reichen runden und ovalen Otokonien von einem gréssten Durchmesser von
0.016-0.02 mm. Die Haut mit Driisencellen und Drischen iiberall reichlichst
ausgestattet.

Die Pharynx6ffnung unten am Grunde der Mundrohre in die Speiserdhre

158 BULLETIN OF THE

iibergehend ; die etwas langer als die Mundréhre war; das vordere Ende (Fig.
1 b) derselben aussen schwarzlich, dann ringartig gelblichweiss (Fig. 1c), dann
wieder und in der iibrigen Strecke schwarzlich. Die Innenseite vorne schwarz,
mit etwa 15 starken Lingsfalten, die sich vorne in den Pharynx hinein fort-
setzen, hinten an dem erwahnten, nicht ganz schmalen, fast farblosen Ringe
plotzlich anhalten; im vorderen Theile der folgenden Strecke kamen wieder
etwa 15 starke Falten vor ; diese Falten waren von einer schwach gelblichen
Cuticula iiberzogen, die ganz fein und zierlich gefaltet war. In dem hinteren
Theile der den Schlundkopf reprasentirenden (Fig. 1 6) vorderen Strecke der
Speiserdhre miindet jederseits die langgestreckte, feinknotige (Fig. 1 dd, 2)
gelblichweisse Speicheldriise ein; der Ausfiihrungsgang ganz (Fig. 2a) kurz. —
Der eigentliche aussen schmutzig schwarzblaue Magen 5.5 mm. lang, oval, von
3.5 mm. Durchmesser, von den gelblichen vorderen Lebern mit Ausnahme
der Mitte der Riickenseite (und des Hinterendes) bedeckt (Fig. 1c). Ge-
offnet zeigt der Magen feine Langsfalten der Innenseite ; etwas nach vorne
findet sich rechts die Oeffnung des Gallenganges der rechten Nebenleber ;
schrag gegentiber die Oeffnung fiir die mit einander verbundenen linke
Neben- und Hauptleber. Hinten und rechts setzt sich der Magen in den
Darm fort (Fig. 19); die schwarze Farbe hort plotzlich und scharf am
Pylorus auf. Der Pylorustheil des Darmes ist gelblichweiss, und hier 6ffnet
sich, dicht neben dem Pylorus, wie durch ein Schnirloch der sogenannte
zweite Magen. Dieses ziemlich grosse Diverticulum (Fig. 1f) ist gelblichweiss,
fast kugelf6rmig, von 3 mm. Durchmesser ; die Innenseite mit einem feinen
pennaten Faltensystem. Der (Fig. 1 g) Darm erst nach unten und hinten, dann
hinaufsteigend, kurz, ziemlich weit, nur in der letzten Strecke enger; aussen
mit Ausnahme der ersten Strecke schwarzlich; die Innenseite schwarz, mit
feinerer Langsfalten und einer starkeren, die von der Oeffnung des Diverticu-
lums anzufangen scheint. — Der Magen und der Darm von Nahrung vollge-
stopft ; dieselbe bestand aus Massen von kleinen niederen Crustaceen (Cope-
poden, Ostracoden) und Stiicken von kleinen Decapoden, mit Bruchstiicken
von kleineren Gasteropod-Schalen und Sandkoérnern vermischt.

Die rechte Nebenleber, wie erwahnt (Fig. 1 hi), den rechten Theil des
Magens mit einem dicken gelblichen Lager einhiillend ; von derselben geht
(wenigstens) ein Zweig an die erste rechte Papille (und wahrscheinlich an
den (Fig. 1) Rhinophorstiel) ab ; diese Lebermasse offnet sich durch einen
ganz kurzen Gallengang in den Magen. Die linke, der vorigen ganz ahnliche,
Nebenleber, den linken Theil des Magens (Fig. 1 4/) einhillend, sich nach
hinten etwas verlangerend und sich mit (Fig. 1 m) dem Ausfiihrungsgange ©
der Hauptleber vereinigend; auch von dieser Leber geht ein Zweig an die
Gegend der Facette der ersten Papille ab; diese Leber 6ffnet sich links in
den Magen. pie Hauptleber viel grosser als die vorigen, an Lange etwa 1.8 cm.
betragend bef einer Breite vorne von 11 und einer Hohe von fast 9.5 mm.; §
das Vorderende schief nach rechts- hinten- unten abgestutzt und (wegen der
vordern Genitalmasse) vertieft; das Hinterende gerundet ; nur central am
Vorderende trat die graubraune Farbe der Leber hervor, sonst war sie von

MUSEUM OF COMPARATIVE ZOOLOGY. 159

der gelblichen Zwitterdriise gedeckt; das Organ bestand aus Lappen von ver-
zweigten Lappchen, deren Ausfihrungsgange sich allmahlich traubenartig
vereinigen und nach und nach den central verlaufenden Hauptgallengang
bilden, welche links am Vorderende frei hervortritt (Fig. 1m) und sich mit
der linken Nebenleber vereinigt. An den Seitentheilen des Riickens der
hinteren Eingeweidenmasse durchbrechen mehrere Leberzweige das Zwitter-
driisenlager und steigen an die Papillenfacetten auf.

Das Pericardium und das Herz wie gewohnlich. — Die Nvere mit ihrer baum-
artigen Veranstelung von schodnen Kolben und Rohren den grossten Theil
der hinteren Eingeweidemasse tberziehend und die Langsfurche derselben aus-
kleidend; in der Auskleidung von jenen viele horngelbe und braungelbe rund-
liche Concremente von einem Durchmesser von meistens 0,025-0.035 mm.
Der Ureter wie gewohnlich; in denselben offnet sich der Pericardialtrichter,
der kurz- birnformig war, von 1mm. Lange, gelblichweiss, mit stark durch-
schimmernden Langsfalten; der Gang kurz, fast ohne Vegetationen der
Innenseite.

Die gelbliche, die Leber mit Ausnahme des grossten Theils ihres Vorderen-
des uberziehende Zwitterdriise wie gewohnlich ; in den Lappchen entwickelte
Zoospermien. Der rechts am Vorderende der hinteren Eingeweidemasse ent-
springende Zwitterdrisengang an die Hinterseite der vorderen Genitalmasse iiber-
tretend. Diese letztere 9mm. lang bei einer Hohe von 7 und einer Dicke
von 5mm.; am oberen Rande vorne die Prostata, hinter derselben der Knauel
der Windungen der Ampulle des Zwitterdriissenganges, unter dem letzteren
die Samenblase; die Hauptmasse ist von der Schleimdriise gebildet. Die gelb-
liche Ampulle durchgehends von beilaufig 0.5 mm. Diam.; aufgerollt hinter
der Prostata einen Knauel bildend, der ein wenig kleiner als die Prostata war ;
ausgerollt mass dieselbe 2.5 cm. Der aus der Theilungsstelle der Ampulle
ausgehende Samenleiter etwa doppelt so lang wie der Durchmesser der Pros-
tata. Diese letztere gelblich, fast kugelférmig, von 4mm. Diam., mit einem
kleinen Nabel der hinteren und einer tiefen Kluft der Vorderseite, aus welcher
die Fortsetzung des Samenleiters hervortretet; die Oberflache fein kérnig, der
Bau ganz der gewohnliche. Die aus der tiefen Kluft vortretende Fortsetz-
ung des Samenleiters graulich, ziemlich diinn, etwa doppelt so lang wie der
Durchmesser der Prostata, sich durch den Penis bis an seine Spitze windend.
Der halb hervorgestreckte Penis gelblich, lang, kegelformig; der gewéhn-
liche Nebensack konnte nicht gefunden werden. Der Eiergang geschlangelt
an den Schleimdrisengang gehend, ausgestreckt beiliufig 1.5 cm. messend,
etwa so dick wie die Ampulle. Die gelbliche, sich in das Vestibulum geni-
tale neben dem Schleimdriisengange éffnende Samendblase birnférmig, von
etwa 5 mm. Lange bei einem Durchmesser von etwa 2.3 mm., von Samen
strotzend; der Ausfiihrungsgang fast ebenso lang, mit starken Langefalten
der Innenseite. Die Schleimdriise gross, kalkweisslich ; die ‘Eiweissdriise
gelblich ; der Schleimdriisengang mit der gewodhnlichen starken Doppel-
falte.

160 BULLETIN OF THE

Ob diese Form nun eine (locale) Varietit der bisher nur im Mittelmeere
und bei den canarischen Inseln gefundenen Tethys leporina darstellt oder eine
eigene Art, muss vorlaufig hingestellt werden. Das Erste ist wohl das wahr-
scheinlichste, obgleich die schwarze Farbe der Verdauungshole und Abweich-
ungen im Genitalsystem wohl auch die letzte Annahme ermoglichten, Die so
trage und nie schwimmende Staurodoris verrucosa kommt doch auch im west-
lichsten Theile des atlantischen Meeres (unweit von Rio Janeiro) vor (Staurod.
Januarii, Bgh.).}

Fam. DORIDIDZ] CRYPTOBRANCHIATZ:.

CHROMODORIS, Ald. et Hauc.

Vgl. R. Bergh, Report on the Nudibranchiata. Challenger Exped., Zool., X., 1884,
pp- 64-78.

Vgl. R. Bergh, Malakolog. Unters., Heft XV. 2, 1884, pp. 64-78, pp. 347-350; Heft
XVI. 2, 1889, pp. 831-837.

Die fast immer schlanken und meistens lebhaft gefarbten Chromodoriden
haben einfach gefiederte Kiemenblitter, starke Lippenplatten, und die Rhachis-
parthie der Radula tragt hochstens nur Verdickungen, aber keine Zahnplatten.?
Die Aphelodoriden,? die sonst sehr ahnlich sind, unterscheiden sich durch
mehrfach gefiederte Kiemenblatter und durch Fehlen von Lippenplatten.

Die Gattung ist bisher nur aus den warmeren (Mittelmeere) und den tropi-
calen Meeresgegenden bekannt. Sie scheint die Artenreichste Gruppe von
Doriden zu sein; sie wird hier wieder durch mehrere neue Arten bereichert.

1. Chr. scabriuscula, Bgh., n. sp.
Tafel i. Fig. 11-19.

Hab. M. atlant. occidentale.

Von dieser form wurden am 24° 44’ Lon. und 83° 26’ Lat. (d. h. in der
Nahe von Straits of Florida) aus einer Tiefe von 37 Faden 3 Individuen
gefischt, die fast vollstandig von derselben Grdésse und Formverhaltnissen
waren,

1 Vg). Thering, Zur Kenntn. d. Nudibranchien d. brasilianischen Kiiste. Jahrb.
f. d. Malacolog. Ges., XTIII., 1886, pp. 230-233.

2 Nur die Chr. Scabriuscula, B. macht hier eine Ausnahme.

3 R. Bergh, Neue Chromodoriden. Malakolog. Bl. N. F. 1., 1879, pp. 107-113.

R. Bergh, On the Nudibr. Gaster. Molls. of the North Pacific Oc. (Dall, Explor.
of Alaska), II., 1880, Pl. VIII. (XVL.), Figs. 12-18.

MUSEUM OF COMPARATIVE ZOOLOGY. 161

Die in Alkohol gut bewahrten Individuum hatten eine Ldnge von 12 bei
einer Breite bis 6 und einer Hohe bis 3.5 mm., die Lange des Fusses 10 bei
einer Breite bis 2.6 mm.; die Breite des Mantelgebrames 1.5 mm.; die Lange
der Tentakel 0.6 mm.; die Hohe der (zuriickgezogenen) Rhinophorien 1.8, der
(zurickgezogenen) Kieme 1.5 mm.— Die Farbe war durchgehends gelblich-
weiss, die Keule der Rhinophorien und die Kieme mehr gelblich.

Die Form war langlich-oval, etwas niedergedriickt; die Riickenseite etwas
gewolbt, iiberall bis an den Rand mit ziemlich zahlreichen kleinsten conischen
Hockerchen besetzt, die am Mantelgebrime zahlreicher waren; die weit nach
vorn stehenden Rhinophorlocher, und die weit nach hinten stehende Kiemen-
spalte schniirlochartig zusammengezogen; die Keule der Rhinophorien stark,
mit beilaufig 20 nicht diinnen Blattern; die Kieme aus 9, einem vorderen und
jederseits 4, nach hinten an Grosse allmahlich abnehmenden, einfach-pinnaten
Blattern gebildet; die Analpapille niedrig. Der Kopf klein; die Tentakel
kurz-cylindrisch, am Ende gleichsam eingestilpt. Die Unterseite des Man-
telgebrames eben, mit durchschimmernden, gegen den Rand senkrecht gehen-
den Spikelziigen. Der Fuss langgestrekt, mit parallelen Seitenrandern; der
Vorderrand mit Furche und gerundeten Ecken; der Schwanz 2.2 mm. lang,
etwas zugespitzt. Die Genitalpapille mit zwei Oeffnungen neben einander.

Die Eingeweide schimmerten nirgends durch, waren an der Korperwand
angeheftet.

Das Centralnervensystem zeigte die cerebro-pleuralen Ganglien kurz-nieren-
formig, die nach unten stehenden pedalen Ganglien grdosser als die pleuralen;
die gemeinschaftliche Commissur ziemlich kurz, kaum so lang wie der Quer-
durchmesser des Fussknoten. Die ganz kurzstieligen Ganglia olfactoria un-

gewohnlich gross, fast halb so gross wie die Ganglia cerebralia; die buccalen:

und gastro-oesophagalen Ganglien wie gewohnlich. -

Die ganz kurzstieligen Augen ziemlich gross, mit schwarzem Pigment. Die
Ohrblasen kleiner als die Augen; mit Otokonien gewohnlicher Art prall ge-
fullt, unter denen ein kugelférmiger, der doppelt so gross wie die anderen
war. In den Blattern der Rhinophorien zahlreiche, auf den Rand senkrecht
stehende, gelbliche, harte Spikeln von einem Durchmesser bis 0.03 mm. Die
Riickenhaut im Ganzen und besonders die Hockerchen derselben mit ahnlichen
Spikeln stark ausgestattet.

Die Mundrohre stark, 1.5 mm. lang, wie gewohnlich. Der kurze Schlund-
kopf 1.6 mm. lang ; hinten an der Unterseite trat die Raspelscheide hervor.
Die horngelbe ringartige Bewaffnung der Lippenscheibe unten viel breiter als
oben, aus den gewohnlichen, bis beilaufig 0.027 mm. langen, an der Spitze
geklufteten (Fig. 11, 12), dicht zusammengedrangten Hakchen zusammenge-
setzt. Die Zunge von gewohnlicher breiter Form mit tiefer Kluft ; in der
hellgelben Raspel 58 Zahnplattenreihen, weiter nach hinten in der starken
Scheide 46 entwickelte und etwa 4 jiingere Reihen; die Gesammtzahl dersel-
ben somit 108. In der Raspel jederseits 25 Platten, und die Anzahl weiter
nach hinten kaum 30 iibersteigend. Die Zahnplatten schwach gelblich ; die
Breite der medianen Platten 0.01, die ersten lateralen 0.016 mm.; die Hohe

VOL. XIX. — NO. 3. 11

162 BULLETIN OF THE

der aussersten Platten meistens 0.028, die Hohe der Seitenplatten bis 0.04 mm.
Es kamen wirkliche mediane, am Rande gezahnte Platten vor (Fig. 13 a, 14).
Die innerste laterale Platte (Fig. 13) mit 8-10 Dentikeln des ausseren und 4-5
des inneren Hakenrandes ; an den tbrigen Seitenplatten fanden sich, wie ge-
wohnlich, nur Zahnchen am Aussenrande, aber in sehr variabler Menge, mit-
unter 5-6, mitunter nur 2-3 (Fig. 15-17); an den aussersten (Fig. 18)
Platten war der Grundtheil kirzer, und unterhalb der Hakenspitze fanden
sich nur 2-3 Zahnchen. — Die langen und weisslichen Speicheldriisen wie
gewohnlich.

Die Speiserdhre etwa so lang wie der Schlundkopf; der Magen wie gewohn-
lich; der Darm vor der Mitte der hinteren Eingeweidemasse hervortretend,
sein Knie in gewohnlicher Weise bildend, und in gewohnlicher Weise ver-
laufend, gelb. — Die hintere Eingeweidemasse (Leber) 6.5 mm. lang bei einer
Hohe und Breite von 3.2 und 3.5 mm., vorne sehr schief abgestutzt und hin-
ten gerundet, (gelblich-) weiss. Die Gallenblase langgestreckt-birnformig,
weisslich, links am Pylorus erscheinend.

Das Pericardium mit dem Herzen, die weisslichen Blutdriisen, die Niere
und der Pericardialtrichter wie gewohnlich.

In den gelblichen Lappen der Zwitterdriise grosse Eierzellen. — Die vordere
Genitalmasse gross, etwa 4mm. lang, von ovaler Form, planconvex, gelblich ;
am Vorderende die starken Windungen des Samenleiters. Die Ampulle des
Zwitterdriisenganges weisslich, geschlangelt. Der Samenleiter lang; der weiss-
liche prostatische Theil kirzer als der gelbliche muskelose (Fig. 19 a); die
kurzkegelformige, ziemlich dicke glans penis am Boden des (Fig. 19)) rau-
migen Praeputiums kaum vortretend. Die Spermatotheke kurz-birnformig,
‘die Spermatocyste wurstformig und kleiner; der vaginale Gang lang, nach
vorne weiter, mit einer starken gelben Cuticula ausgefuttert; die Vagina fast
so lang wie das Praeputium, doppelt so dick wie der vaginale Gang, von einer
diinneren Cuticula ausgekleidet. Die Schleimdrise gross; die Eiweissdriise
etwas mehr gelblich.

Diese Form unterscheidet sich von den allermeisten Chromodoriden durch
die harten Héckerchen des Riickens und damit durch die ziemlich starke Ent-
wickelung der cutanen Spikeln, so. wie besonders durch wirkliche mediane
(rhachidiale) Zahnplatten. Auch die Auskleidung des vaginalen Ganges ist
eigentbimlich.

2. Chr. punctilucens, Bgh., n. sp.

Tafel I. Fig. 4-10.

Hab. M. atlant. occid.
Ein einziges Individuum wurde aus einer Tiefe von 37 Faden auf 24° 44’
Lon. und 83° 26’ Lat. (d. h. in der Nahe der Straits of Florida) gefischt.

MUSEUM OF COMPARATIVE ZOOLOGY. 163

Das in Alkohol bewahrte Individuum hatte eine Lénge von 3.5 bei einer
Breite von 1.6 und einer Hohe von 1.5 cm.; die Breite des Mantelgebrames
2 (vorne) bis 4.5 mm. ; die Hohe der (zuriickgezogenen) Rhinophorien 3, der
(zuriickgezogenen) Kieme 5 mm.; die Lange des Fusses 3 bei einer Breite
bis 1 cm. — Die Farbe der obern Seite war durchgehends olivenbraungrau
mit ziemlich zahlreich zerstreuten gelben und weissen Punkten, die oft eine
schwarze oder schwarzliche Areola zeigten; am Mantelrande ein schmales,
schwarzes, seiner Lange nach durch eine weisslichgelbe oder gelbe Linie ge-
theiltes Band ; die Unterseite des Mantelgebrames von der Grundfarbe des
Riickens oder mehr grau, hier und da schwarzfleckig ; der Rand der Rhino-
phorlécher so wie der Kiemenspalte schwarz mit gelben Punkten und Bruch-
stiicken von gelben Linien ; die Keule der Rhinophorien schwarz, am Vor-
derrande und gegen die Spitze gelblich; die Kiemenblatter schwarzlich, die
Rhachisparthien, die Spitze und theilweise die Rander der Blatter gelbfleckig ;
die Analpapille schwarz mit gelblichem Rande. Die Korperseiten von der
Farbe des Riickens, die gelben und weissen Punkte kommen aber sehr sparsam
vor. Die Tentakel mit gelber Spitze; der Aussenmund schwarz. Die Fuss-
sohle graulich; das Fussgebrame oben von der Farbe der Korperseiten, aber
mit starken schwarzen Flecken ; der Fussrand gelb, hier und da mit schwarzen
Fleckchen ; am Schwanzricken zerstreute gelbe Punkte.

Im Aeusseren simulirte diese Form (in Alkohol bewahrt) ganz eine Dorvopse,
nur war der Mund wie bei den Doriden, und die Tentakel kurz kegelférmig (an
der Spitze, wie bei so vielen Chromodoriden, gleichsam halb eingesttilpt). Die
Form langlich, die Consistenz weich. Der Ricken etwas gewolbt, das Mantel-
gebrame ziemlich breit, wellenformig gebogen, an der Unterseite wie der ganze
Riicken eben. Die Rhinophorlécher fast *glatrandig; die Keule der Rhino-
phorien kraftig, mit beilaufig 30 breiten Blattern. Die Kiemenspalte quer-
oval, fein rundzackig. Die Kieme jederseits aus 7 einfach pinnaten Blattern
gebildet, denen sich hinten eine Spirale von 13 Blattern anschliest ; diese letz-
teren etwas schmachtiger und unbedeutend niedriger als die andern, die alle
fast von gleicher Grosse waren. Hinten zwischen den Spiralen die cylindrische,
oben abgestutzte, etwa 3mm. hohe Analpapille; rechts und vorne neben der-
selben die Nierenpore. Die Kéorperseiten ziemlich hoch; die Genitalpapille
wie gewohnlich. Der Fuss vorne gerundet-abgestutzt, mit feiner Randfurche ;
das Fussgebrame nicht schmal; der Schwanz stark, nicht kurz.

Das gelbe Centralnervensystem von den Blutdriisen bedeckt, in reichliche, fest
anhangende Bindesubstanz gehillt; die Ganglien ziemlich dick. Die zwei
Abtheilungen der cerebro-pleuralen Ganglien sehr ausgepragt; die pedalen
ausserhalb und unterhalb der vorigen liegend; die pleuralen grésser als die
cerebralen, die pedalen wieder grésser als die pleuralen ; die gemeinschaftliche
Commissur weit, doppelt so lang wie der Querdurchmesser des Centralnerven-
systems. Die Riechknoten, die buccalen und die gastro-oesophagalen Ganglien
wie gewohnlich.

Die Augen fast sessil, mit schwarzem Pigment. Die Ohrblasen so gross wie
die Augen, mit Otokonien gewohnlicher Art prall gefillt. In den Blattern

164 BULLETIN OF THE

der Keule der Rhinophorien keine Spikel. In der Haut des Riickens kamen
erhartete Zellen sparsam vor.

Die Mundrohre sehr stark, etwa 6 mm. lang bei einem Diam. hinten
von 6mm.; aussen gelblich, innen vorne schwarz und hinten gelblich; die
starken Retractoren wie gewOhnlich. — Der sehr kraftige Schlundkopf 5.5 mm.
lang bei einer Breite von 4.5 und einer Hohe von 4.5 mm.; das abgeplattete
Hinterende stark schrage; von der Unterseite ragt die starke (1.1 mm.
in Diam. haltende) Raspelscheide 3mm. nach oben und links empor. Die
runde Lippenscheibe von 4mm. Diam., von der schon dunkel ambergelben
Lippenplatte (Fig. 4) iberzogen, welche oben schmaler, unten (von vorn nach
hinten) viel breiter (bis 2.6 mm.) war, unten continuirlich, oben durch einen
ganz schmalen Zwischenraum in zwei Halften geschieden. Die Lippenplatte
in gewohnlicher Weise von dicht zusammengedrangten gelblichen Stabchen
gebildet, welche (in gerader Linie gemessen) eine Lange bis zu fast 0.06 mm.
erreichten, gebogen und in der Spitze gekluftet (Fig. 5, 6) waren. Die
Zunge breit, abgeplattet, mit breiter Kluft; in der gelblichen Raspel 60
Zahnplatten, weiter nach hinten und in der ziemlich langen Raspelscheide 98
entwickelte und 12 jiingere Reihen ; die Gesammtzah] derselben somit 170.
Die vordersten 16-18 Reihen sehr incomplet. In den hintersten Reihen der
Zunge fanden sich jederseits bis 53 Seitenzahnplatten, und die Anzahl stieg
kaum wesentlich weiter nach hinten. Die Zahnplatten gelblich; die Hohe
der aussersten Platten 0.04-0.05 mm. betragend, allmahlig stieg die Hohe der
Platten bis zu etwa 0.1 mm.; die Lange der medianen (Fig. 7 a) Verdickungen
meistens 0.035mm. Die Zahnplatten von der gewohnlichsten Hakenform;
an den Aussersten ist der Korper in gewoéhnlicher Weise reducirt, und die
Platten mehr aufrecht. Die innerste (Fig. 7 bb) Zahnplatte an beiden Ran-
dern des Hakens gezahnelt; alle die wbrigen (Fig. 8, 9) nur am ausseren
Rande mit 6-10 feinen Dentikeln ; die 5-7 aussersten (Fig. 10) ohne
Dentikel.

Die weissen Speicheldriisen sehr langgestreckt, vorne etwas dicker, sich bis
an die Unterseite der hinteren Eingeweidemasse hinab erstreckend.

Die Speiserdhre kaum langer wie der Schlundkopf bei einem Durchmesser
von beilaufig 1mm. Der in die hintere Eingeweidemasse eingeschlossene
Magen rundlich, nicht klein. Der Darm vor der Mitte der oberen Seite die
hintere Eingeweidemasse durchbrechend, in gewohnlicher Weise verlaufend
und sein Knie bildend; im ganzen 7 cm. lang bei einem Durchmesser von
1.5-2 mm. — Der Inhalt der Verdauungshohle war ganz unbestimmbare thieri-
sche Masse, worin Stiicke von Zahnplattenreihen des Thieres selbst.

Die hintere Eingeweidemasse (Leber) 2 cm. lang, bei einer Hohe und Breite
von 1.2 cm.; nach hinten zugespitzt; die vordere Halfte der rechten Seite
(durch die vordere Genitalmasse) stark abgeplattet ; die Substanz gelb. Die
Gallenblase links neben dem Pylorus, sackformig, von 4mm. Lange, graulich.

Das Pericardiwm gross, queroval, von 8 mm. kurzestem Diam. Die gelbe
Herzkammer von 3.5mm. Lange. Die Blutdriisen in den Randern etwas lap-
pig, graugelb, abgeplattet; die vordere gestreckt-herzformig mit der Spitze nach

Pie

MUSEUM OF COMPARATIVE ZOOLOGY. 165

vorn, 6mm. lang; die hintere breit, querliegend, 7 mm. breit. — Die Ver-
breitung der Niere tber die hintere Eingeweidemasse sehr schon, die /Urin-
kammer weit; der Pericardialtrichter stark, birnformig, 2mm. lang.

Die Zwitterdrise mit einem 0.5-1 mm. dicken, mehr gelben Lager den
eréssten Theil der Leber tiberziehend; in ihren Lappchen grosse Eizellen. —
Die (sehr stark erhartete) vordere Genitalmasse gross, planconvex, 14mm. lang
bei einer Breite von 7 und einer Hohe von 11 mm. Am Vorderende die
ziemlich dicke, geschlangelte, opak-weissliche Ampulle des Zwitterdriisen-
ganges. Der Samenleiter lang, gewunden, der gelbliche prostatische Theil
kirzer als der muskuldse; die glans penis kegelformig. Die Spermatotheke
von ovaler Form, von 3.5mm. Lange; die Spermatocyste wurstformig, gebo-
gen, ein wenig langer. Die Schleimdrise graulichgelb und kalkweiss, die
Eiweissdriise braunlich; der ausserste Theil des Schleimdriisenganges schwarz.

Dieses Thier reprasentiert gewiss eine neue Art. Unter den wenigen bis-
her! bekannten Arten aus dem westlichen atlantischen Ocean (Chr. Moerchii,
B.; Chr. gonatophora, B.) giebt es keine zu welcher sie hingefihrt werden
konnte, und eben so wenig kann sie mit irgend einer der vielen im Mittel-
meere vorkommenden identificirt werden.

3. Chr. sycilla, Bgh., n. sp.
Tafel III. Fig. 5-13.

Hab. M. atlant. occ. (Sin. Mexicanum).

Von dieser Form hat die Blake-Expedition 16 Meile gegen Nord von den
Jolbos-Inseln (an der Kiiste von Yucatan) ein einziges Exemplar gefischt, aus
einer Tiefe von etwa 14 Faden.

Das in Alkohol gut bewahrte, nur etwas zusammengezogene Indivi-
duum hatte eine Ldénge von 2.5 bei einer Breite von 1 und einer Hohe
von 1.4 Cm.; die Hohe der (zuriickgezogenen) Rhinophorien fast 4, der
(zurickgezogenen) Kieme 4.25 mm.; die Breite der Fusssohle 4.5 mm. —
Die Grundfarbe des Korpers war ein sehr schénes und lebhaftes Dunkel-
blau. Diese Farbe war am Ricken wie an den Seiten von zahlreichen, kalk-
weissen, diinnen, oft zerstiickelten Langslinien durchzogen; die Stiickchen
mitunter an dem einen oder anderen Ende kurz- schlingen- oder osenformig
oder mit einem kurzen Seitenaste; zwischen den Linien kamen noch hier und
da einzelne rundliche oder ovale Fleckchen vor. Am Riicken fanden sich
etwa 9-10 solche Linien vor, an den Korperseiten 5-6. Der Mantelrand
(Fig. 13) so wie der Fussrand mit einer ganz ahnlichen, ebenso unterbroche-
nen, kalkweissen Linie geziert. Der Rand der Rhinophorlécher weiss ; die
Rhinophorien schmutzigblau. In dem theilweise weissen Rande der runden
Kiemenspalte endigt die grésste Zahl der weissen Riickenlinien; die Kiemen-

1 Vgl. die von mir vor einigen Jahren gelieferte Liste in Challenger Exped.,
Zool., X., 1884, pp. 65-72.

166 BULLETIN OF THE

blatter sehr schén blau; ihre an der Aussenseite ziemlich breite Rhachis weiss
gerandert, der schmale innere Rhachisrand mitunter auch weiss. Die Fuss-
sohle schmutzig gelblich.

Die Formverhiltnisse wie bei den meisten Chromodoriden; das Stirnge-
brame, der Mantelrand und das Schwanzsegel schmal. Ringsum an der Un-
terseite des Mantelgebrames fanden sich grossere und kleinere, durchsichtig-
gelbliche, kugelformige, sessile, ungleichgrosse Blasen (Fig. 13 aa) von einem
Durchmesser von beilaiifig 0.3-2 mm; die grossten kamen am Schwanzsegel
vor (Fig. 13 aa); jede zeigte am Scheitel eine meistens schon unter der Loupe
sehr deutliche Oeffnung. Oberhalb und ausserhalb des Aussenmundes jeder-
seits ein gleichsam eingestiilpter Tentakel. Die etwas zusammengedrickte
Keule der Rhinophorien mit etwa 40-45 breiten Blattern. Die Kieme weit
nach hinten stehend, mit 12 schénen Blattern, von welchen das hinterste Paar
kleiner, die wbrigen fast gleichgross. Im Centrum des Kiemenkreises die
niedrige (oben weisse) Analpapille, rechts und vorn neben derselben die Nie-
renpore. Der Fuss wie gewohnlich ziemlich schmal.

Das Peritonaeum farblos oder hier und da blaulich.

Die das Centralnervensystem eng einhillende starke Bindesubstanzcapsel mit
der Unterseite der vorderen und mit dem Vorderende der hinteren Blutdriise
innig verwachsen. Die Ganglien an der Unterseite der ganzen Ganglienmasse
deutlich geschieden. Die cerebro-pleuralen Ganglien langlich-nierenformig,
die cerebrale grésser als die pleurale Abtheilung; die rundlichen pedalen Gang-
lien etwas grosser als die cerebralen. Die grosse gemeinschaftliche Commissur
ziemlich weit, doppelt so lang wie der Querdurchmesser des Centralnerven-
systems. Die proximalen und distalen Riechknoten wie gewohnlich. Die
buccalen Ganglien oval, fast unmittelbar mit einander verbunden; die gastro-
oesophagalen sehr kurzstielig, etwa + der vorigen betragend.

Die: Augen mit schwarzem Pigment und schwach gelblicher Linse, durch
einen kurzen N. opticus mit dem kleinen Gangl. opticum verbunden. Die
Ohrblasen wie gewohnlich, mit zahlreichen Otokonien gewohnlicher Art. In
den diinnen und breiten Blattern der Keule der Rhinophorien kamen zerstreute
erhartete Zellen, aber keine Spikel vor.

Die Mundrohre aussen blaugrau, innen gelblichweiss, kurz und weit; der
Diam. und die Lange etwa 5 mm. betragend. — Der Schlundkopf sehr stark,
6 mm. lang bei einer Breite von 5 und einer Hohe von 4.75 mm., von gewohn-
lichen Formverhaltnissen ; die 3.5 mm. lange, starke Raspelscheide langs des
Hinterendes des Schlundkopfes hinaufgekrimmt; die Lippenscheibe gross,
gewolbt, mit sehr starker, griinlich- olivenfarbiger Lippenplatte. Diese letztere
einen etwa 3mm. breiten Ring bildend oder eigentlich zwei Halbringe, die in
der Mittellinie oben und unten durch ein schmaleres Zwischenstick vereinigt
sind. Die Platte in gewohnlicher Weise von ganz dicht gedrangten Stabchen
mit gebogenem hakenartigem Kopf gebildet (Fig. 5) ; sie erreichten eine Hohe
bis zu beilaufig 0.04 mm.; die Stabchen der erwahnten Zwischenstiicke ganz
klein. Die Zunge breit ; in der griinlich- olivenfarbigen Raspel 39 Zahnplat-
tenreihen, weiter nach hinten kamen dazu 41 entwickelte und 4 jiingere Reihen,

MUSEUM OF COMPARATIVE ZOOLOGY. 167

die Gesammtzahl derselben somit 84. Die 8 vordersten Reihen mehr oder
weniger incomplet. Die hintersten Reihen der Zunge enthielten (jederseits)
etwa 290 Zahnplatten, und die Anzahl stieg kaum wesentlich weiter nach
hinten. Die Zahnplatten schwach gelblich mit etwas grinlichem Anfluge;
die Seitenzahnplatten erreichten eine Hohe bis zu 0.14 mm., die der aussersten
betrug etwa 0.06-0.08 mm. Die Rhachisparthie sehr schmal, meistens mit
einer seichten medianen Langsfalte. Die Zahnplatten von der allergewohn-
lichsten Form (Fig. 6-9); die Haken gabelig, der obere Ast langcr und mehr
gebogen als der untere ; unterhalb dieses letzteren eine Andeutung von feinen
Rundzacken, die nach aussen in den Reihen besonders etwas deutlicher wur-
den und selbst in feine Dentikelbildungen tbergehen kénnen (Fig. 6). Die
(meistens) 5-6 aussersten Platten sind von etwas abweichender Form (Fig.
10, 11), zeigen den Haken reducirt und mit gerundetem Ende; die 1-2 aller-
aussersten haben keine Auskerbung oben (Fig. 12).

Die Spercheldriise sehr lang, sich uber die Unterseite der vorderen Genital-
masse erstreckend, kalkweiss, dinn; vorne etwa 1.25 mm. breit, in der hin-
teren Halfte kaum halb so dick ; die ganz kurzen Ausfihrungsgange in die
Wurzel der Speiserdhre einmiindend.

Die Speiseréire diinn, etwa 14mm. lang (bei einem Durchmesser von
0.8 mm.), ganz unten am Vorderende der hinteren Eingeweidemasse eintre-
tend und sich.in die weite Leber- Magenhéhle 6ffnend. Der Darm die Leber
vor der Mitte ihrer oberen Seite durchbrechend, vorwarts gehend ; sein Knie
uber die vordere Genitalmasse legend und dann nach hinten verlaufend; die
Lange des Darmes im Ganzen etwa 5 cm. betragend, bei einem wechselnden
Durchmesser von 1.5-4mm. Der weissliche Inhalt des Darmes (und der
Leberhohle) war unbestimmbare thierische Masse, mit langen und spitzen
Spikeln vermischt.

Die hintere Eingeweidemasse (Leber) war 15 mm. lang bei einer Hohe von
12 und einer Breite von 9 mm. (stark zusammengezogen), hinten gerundet,
vorne schief abgestutzt; ihre Substanz hell gelblichgrau. Die Gallenblase
horizontal an der linken Seite des Pylorus legend, 4mm. lang bei einem
Durchmesser von 1 mm., gelblichweiss.

Das Pericardium blaugrau. Das Herz wie gewohnlich. Die grinlich-gelb-
grauen blutdrusen an der oberen Seite mit hell griinlichblanem Ueberzuge, die
vordere kleiner, 3mm. breit bei einer Linge von 2.5 mm.; die hintere grésser,
gerundet-dreieckig, die Spitze nach hinten kehrend, 5.5 mm. breit bei einer
Lange von 3.5 und einer Dicke von 0.8 mm. — Die Niere wie gewohnlich ;
das pericardio-renale Organ birnformig, 1.8 mm. lang.

Die gelbliche Zwitterdriise als ein diinnes Lager die Leber fast vollstandig
iiberziehend ; in den Lappchen der Driise kamen reife Zoospermien vor. —
Die vordere Genitalmasse 8 mm. lang bei einer Héhe von 6 und einer Breite
von 4mm.; die dunkelblauen Hauptausfiihrungsginge noch 4 mm. lang; das
Hinterende der Masse wird zum gréssten Theile von der grossen Samenblase
gebildet, die aber oben und an der Ausseren (rechten) Seite von den Win-
dungen des Samenleiters gedeckt wird. Die Ampulle des Zwitterdriisen-

168 BULLETIN OF THE

ganges opak-gelblichweiss, wurstartig, etwas zusammengebogen, ausgestreckt
an Linge 6 mm. bei einem Durchmesser von beilaufig 0.75 mm. messend.
Der lange, viele langere und kiirzere Windungen machende, weissliche prosta-
tische Theil des Samenleiters ausgestreckt etwa 5-6 cm. lang bei einem fast
durchgehenden Diam. von 0.5 mm.; der mehr gelbliche muskulése Theil nur
beilaufig 12 mm. lang und etwas dinner. Der letztere geht in den sich nach
und nach verdickenden, am Ende blauen Penis iiber, der eine Lange von 4.5
bei einem Diam. (vorne) bis zu 1.5 mm. hatte ; nur der unterste Theil des-
selben ist hohl (Praeputium), auch an der Innenseite blau, am Boden der
Hohle die gewohnliche, wenig vortretende papillare glans. Die Spermatotheke
gross, kugelfoérmig, von 5 mm. Diam., die Ausfiihrungsgange nicht lang; die
Spermatocyste aneeatiye 2.5 mm. eae: ziemlich kurzstielig. Die Schleim-

und Hiweissdriise kaum die Halfte der ganzen Genitalmasse betragend, 5.2 mm.
lang bei einer Hohe von 4.8 und einer Breite von 3 mm., gelblichweiss und
weiss ; der weite Schleimdriisengang aussen und innen ee

Ringsum die Gegend der Cardia, an die Leber (Niere?) angeheftet, fanden
sich vier, 1.5-2 mm. lange Individuen eines mit dem Dzstoma glauci} wenig-
stens ganz nahe verwandten Thieres.

Man kennt jetzt eine kleine Reihe von Chromodoriden (Chr. runcinata, pan-
tharella, sannio (Fig. 15), picturata, camoena, elegans (Fig. 16), glauca, cali-
forniensis (Fig. 14), Marenzelleri, gonatophora, sycilla (Fig. 13)), mit eigen-
thiimlichen blasenartigen Driisenbildungen am Mantelgebrdme, wozu jetzt auch
die hier untersuchte Form gehort. — Sie scheint von den schon bekannten
Chromodoriden specifisch verschieden.

PHLEGMODORIS, Bea.
R. Bergh, Malacolog. Unters., Heft XIII., 1878, pp. 593-597.

Corpus molle quasi subgelatinosum, dorso tuberculoso. Tentacula pro ma-
jore parte affixa, applanata. Branchia e foliolis tripinnatis paucis formata.
Podarium sat latum, sulco marginali anteriori non profundo, labio superiore
capite affixo.

Armatura labialis nulla, Radula rhachide nuda, pleuris multidentatis ;
dentes intimi forma simpliciori, reliqui hamati. — Penis inermis.

Die Phlegmodoriden sind von weicher Kérperbeschaffenheit, der Riicken mit
Knoten und Knétchen bedeckt. Die Tentakel etwas applanirt, zum grossten
Theile angeheftet. Die (retractile) Kieme aus wenigen (5) tripinnaten Federn

1 Vgl. meinen: Report on the Nudibranchiata. Challenger Exped., Zool., X.,
1884, p. 18, Pl. X., Figs. 5-17.

MUSEUM OF COMPARATIVE ZOOLOGY. 169

gebildet. Der Fuss ziemlich breit, mit nicht tiefer vorderer Randfurche, die
obere Lippe derselben an die Seiten des Kopfes angeheftet. — Keine Luppen-
platte. Die Raspel ohne Mittelzahnplatten ; die Seitenzahnplatten ziemlich
zahlreich, die innersten von einfacherer Form, die anderen hakenformig. —
Der Penis unbewaffnet.

Die Phlegmodoriden gehoren den tropischen, hauptsachlich den indischen
Meeresgegenden. .
Phi. mephitica, Bgh. M. philippin.
Phi. areolata (Ald. et Hanc.). M. indic.
Phi. sponyiosa (Kelaart). M. indie.
Phl.? anceps, Bgh. M. mexican.

a HE

Phlegmod.? anceps, Bgh., n. sp.
Tafel I. Fig. 20-26. Tafel II. Fig. 6.

Hab. M. atlant. occ.

Von dieser Form lag ein einziges, in Alkohol mittelmassig conservirtes
Individuum vor, an der Long. 89° 16’ und Lat. 23° 13’ (d. h. im mexicanischen
Golfe) aus einer Tiefe von 84 Faden gefischt.

Die Lénge des Individuums betrug 10 mm. bei einer Breite bis 5 und einer
Hohe bis 2 mm. ; die Linge des Fusses 7 bei einer Breite bis 2.2mm.; die
Breite des Mantelgebrames 2 mm.; die Hohe der Rhinophorscheide 0.8, des
Kiemenhiigels 1 mm. — Die Farbe war durchgehends hell schmutzig gelblich,
am Riicken mit dunklen erhabenen Punkten (Hockerchen). Die Consistenz
des Korpers ziemlich weich.

Die Form langlich-oval, abgeplattet, mit breitem und ziemlich dinnem Man-
telgebrame. Der Ricken mit Andeutung von einem medianen und jederseits
einem, der Grenze des eigentlichen Riickens folgenden, lateralen Kamme; der
Ricken wbrigens tiberall mit zerstreuten spitzen Hockerchen bedeckt, die, beson-
ders am Mantelgebrame, durch Auslaufer oft mit einander verbunden waren;
am medianen Kamm so wie an den hohen Rhinophorscheiden (Fig. 20), und
am hohen Kiemenhiigel waren die Hoékerchen hoéher und dichter stehend,
besonders am Rande von jenen und diesem. Die Keule der Rhinophorien
beilaufig so hoch wie die Rhinophorscheide, mit etwa 25 diinnen Blattern;
die Kieme aus 5, bis 1.2 mm. hohen, einfach- hier und da doppelt- gefiederten
Blattern gebildet, von denen die 3 vorderen héher ; die Analpapille niedrig.
Die Unterseite des Mantelgebrames eben. Die Kérperseiten ganz niedrig;
die Genitalpapille wie gewéhnlich. Der Fuss nicht schmal, vorn gerundet und
mit Randfurche; die obere Lippe stark vorspringend, in der Mitte ausge-
randet; der Schwanz nicht ganz kurz. Die Tentakel fingerformig.

Das Centralnervensystem abgeplattet ; die cerebro-pleuralen Ganglien ziemlich
rundlich, die Grenze zwischen den zwei Abtheilungen derselben wenig aus-
gepragt; die pedalen Ganglien rundlich, grosser als die pleuralen, ausserhalb
derselben liegend. Die proximalen Riechknoten fast sessil, ziemlich gross;
die einander fast beriihrenden buccalen und die gastro-oesophagalen Ganglien

170 BULLETIN OF THE

wie gewohnlich; die kugelformigen sessilen Ganglia optica kleiner als die
Augen.

Die Augen ziemlich gross, fast sessil, mit reichlichem schwarzem Pigment.
Die Ohrblasen etwas kleiner als die Augen, von beilaufig 0.08 mm. Diam., von
Otokonien gewohnlicher Art strotzend, die einen Durchmesser bis 0.009 mm.
erreichten. In den Blattern der Rhitophorien lange, aber nicht stark erhirtete,
auf dem Blattrand senkrecht und schiefstehende Spikel. In der Réickenhaut
sehr zahlreiche, lange, mehr oder weniger erhirtete Spikel, die auch, und zum
Theile biindelweise, in den Hoéckerchen vorkommen, hier aber weniger er-
hartet und meistens mit den Spitzen am Scheitel der (Fig. 6, Fig. 21) Hécker-
chen hervorragend; eben derselben Art war das Verhaltniss an den Rhinophor-
scheiden und am Kiemenhigel.

Die aussen weisslich, innen gelbliche (Fig. 22) Mundréhre stark, etwa
1.5 mm. lang ; hinten scheinen mehrere drisenartige Korper einzumiinden
(Fig. 22). — Der kriftige Schlundkopf etwa so lang wie die Mundrdhre, hinten
an der Unterseite trat die Raspelscheide als eine dicke Papille hervor; die
kraftige, rundliche, gelblichgraue Lippenscheibe zeigte sich von einer starken
gelben Cuticula iberzogen. Die Zunge breit und etwas abgeplattet; in der
breiten gelben Raspel 7 Zahnplattenreihen, von denen die erste sehr incom-
plet ; weiter nach hinten 8 entwickelte und zwei jiingere Reihen; die Ge-
sammtzahl derselben somit 17. Die Rhachis ziemlich breit, nackt; von late-
ralen Platten jederseits 17-18 hinten an der Zunge und weiter nach hinten
19-20. Die Zahnplatten horngelb. Die Linge der 4 innersten betrug mei-
stens 0.06—0.08-0.1-0.11 mm.; die Hohe des Hakens der Platten wbrigens bis
0.11 steigend, die der diussersten nur 0.04-0.06 mm. betragend. Die innersten
(Fig. 23, 24) 4 Platten sind wenig gebogen, schlanker, mehr aufrecht; danach
entwickelt sich schnell die durch die Reihe bleibende Form (Fig. 25), die
allergewohnlichste Hakenform; die diusserste oder die zwei aussersten Platten
mit verkiirztem Kéorper, mehr aufrecht stehend (Fig. 26 aa); die dusserste
schlanker als die nichst stehenden.

Die weisslichen Speicheldriisen langgestreckt.

Die Speiserohre beilaufig so lang wie der Schlundkopf, ziemlich weit. Der
1.5 mm. lange, freie Magen und der Darm wie gewohnlich. Die Verdauungs-
hohle leer. — Die hintere Eingeweidemasse (Leber) kurz-kegelférmig, vorne
schief abgestutzt, hinten gerundet, schmutzig-weisslich.

Das Pericardium mit dem Herzen wie gewohnlich; ebenso die abgeplatte-
ten, griulich-weisslichen Blutdriisen.

Die Zwitterdriise schien den gréssten Theil der Leber zu iiberziehen, kaum
etwas heller als diese ; in den Liippchen Massen von Zoospermien. — Die vor-
dere Genitalmasse beilaiufig 1.5 mm. lang, etwas zusammengedrickt; die Am-
pulle des Zwitterdriisenganges ziemlich dick, wurstformig, gebogen, ausge-
streckt ein wenig langer als die Genitalmasse, opak gelblichweiss. Der
Samenleiter nicht lang, der kurze Penis schien unbewaffnet. Die Sperma-
totheke kugelformig ; die Spermatocyste sackformig, gebogen, etwas kleiner.
Die den gréssten Theil der Genitalmasse bildende Schleimdriise weisslich,
die Eiweissdriise gelblich.

ee eee

Se ee ee TSEC

MUSEUM OF COMPARATIVE ZOOLOGY. 171

Ob diese Form nun wirklich zur Gattung Phlegmodoris gehort, ist sehr
zweifelhaft. Diese Thierform zeigt wie die letztere Gattung die inneren
Seitenzahnplatten von einfacherer Form, hat auch eigenthumliche drisen-
artige Koérper hinten am Mundrohre, so wie stark vortretende Rhinophor-
scheiden. Die Kieme ist hier aber nicht tripinnat wie bei den Plegmodo-
riden, und das Vorderende des Fusses scheint von anderer Beschaffenheit.

Fam. DORIDIDA] PHANEROBRANCHIATSAs.

NEMBROTHA, Beu.

R. Bergh, Malacolog. Unters., Heft XI., 1877, pp. 450-461.
R. Bergh, Beitr. zu einer Monogr. d. Polyceraden, I]. Verh. d. k. k. zool. bot.
Ges. in Wien, XXX., 1880, pp. 658-663; III. Ib., XX XIIL., 1883, pp. 164, 165.

Corpus limaciforme, fere laeve; tentacula breviora, lobiformia ; rhinophoria
retractilia clavo perfoliato ; branchia paucifoliata, foliolis bi- vel tripinnatis;
podarium angustius. .

Armatura labialis inconspicua vel nulla. Radula sat angusta; rhachis den-
tibus depressis subquadratis vel arcuatis; pleurae dente laterali majori falci-
formi singulo et dentibus externis depressis compluribus.

Glandula hermaphrodisiaca hepate connata ; prostata discreta nulla; glans
penis armata.

In den Formverhialtnissen stehen diese Thiere den Trevelyanen sehr nahe,
zeigen auch den Korper Limax-artig, eben, und den Fussrand von den Ko6rper-
seiten fast nicht oder nur wenig vortretend. Die Tentakel sind auch kurz,
lappenformig; die Rhinophorien retractil, mit durchblatterter Keule. Die
(nicht retractile) Kieme auch an etwa der Mitte der Linge des Riickens
stehend, aber aus wenigen (3-5) Federn gebildet.— An der Lippenscheibe keine
Bewaffnung oder eine ganz schwache (N. nigerrima). Die Zungenbewaffnung
gewissermassen an die der Polyceren erinnernd. An der Rhachis kommen
(im Gegensatze zu der nackten Rhachis der Trevelyanen) subquadratische
oder bogenformige, niedergedriickte Mittelzahnplatten vor; neben der Mittel-
zahnplatte eine grosse unregelmassig sichelformige Seitenzahnplatte ; die dus-
seren Platten niedergedriickt, ohne entwickelten Haken. Die Zwitterdriise
ist (4m Gegensatze zu dem Verhiltnisse der Trevelyanen) von der Leber
nicht gesondert. Der Penis ist in gewohnlicher Weise mit Hakenreihen
bewaffnet.

Die Nembrothen sind bisher nur aus den tropischen Meeresgegenden be-
kannt und zwar fast nur aus dem philippinischen und dem Stillen Meere.

172 BULLETIN OF THE

Der kleinen Reihe von Arten wird die untenstehende neue aus dem mexi-
canischen Golfe hinzufiigen sein.

N. nigerrima, Bgh. M. philippin., pacific.
N. Kubaryana, Bgh. WM. pacific.

N. gracilis, Bgh. M. philippin.

N. cristata, Bgh. M. philippin.

N. morosa, Bgh. M. philippin.

N. diaphana, Bgh. M. philippin.

N. gratiosa, Bgh., n. sp. M. mexican.
N.? Edwardsi (Angas). M. pacific.

OE so eee ee

N. gratiosa, Bgh. n. sp.
Tafel II. Fig. 1-5. Tafel III. Fig. 1-4.

Hab. Sinum Mexicanum.

Es fand sich nur ein einziges Individuum vor, an der Breite von 24° 26!
und Lange von 83° 16’ aus einer Tiefe von beilaufig 36 Faden gefischt.

Das in Alcohol bewahrte Individuum hatte eine Lange von 22 bei einer
Hohe von 6 und einer Dicke von 4 mm., die Hohe der Kieme noch 4 mm.
betragend; die Hohe der Rhinophorien 2.5, des Schwanzkammes so wie
der Rhinophorkimme 1.5 mm.; die Breite des Fusses 2.6mm. — Die Farbe
des Thieres wird im Leben prachtvoll gewesen sein; die Grundfarbe des
Korpers war jetzt hell gelblich, am Ricken wie an den Korperseiten mit
zahlreich zerstreuten, runden und ovalen, gringrauen und graugriinen Flecken
von einem Diam. von meistens 0.6-0.8 mm. ; die Rhinophorkamme an ihrem
Grunde aussen von einer Linie von ahnlicher Farbe eingefasst, ihr oberer
Rand schwarzblau, ebenso der Stirn; die Rhinophorien schwarzblau oben,
gelb unten; der Rand der Becherartigen Tentakel schwarzblau, ebenso der
Scheitel und der Grund der Hocker des Schwanzkammes und des Fussran-
des oben, die Rhachis-Parthien der Kiemenblatter hell gelblich, das Laub
schwarzblau ; die Fusssohle gelb.

Das Thier war von etwas mehr zusammengedrickter Form und langer als
andere bekannte Nembrothen. Die Tentakel wegen einer sich ihrer Lange
nach erstreckenden Furche fast ohrenférmig, am ausseren Ende etwas gelost
(Fig. 2a). Zwischen den Tentakeln der rundliche Aussenmund. Oberhalb
des Mundes tritt der ziemlich schmale, im Vorderrande ein wenig ausgekerbte
Stirn etwa 1.5 mm. hervor. Hinter dem Stirne erhebt sich jederseits ein
starker Rhinophorkamm (Fig. 1 a) mit gebogenem, ebenem Rande ; innen am
Grunde des Kammes die rundliche Oeffnung der Rhinophorhohle, der Rand
derselben hinten mit einem vortretenden Zipfel (Fig. 1c); die Rhinophorien
kurzstielig, ihre Keule mit etwa 35-40 Blattern. Der Ricken schmal, gerun-
det in die K6rperseiten iibergehend ; ein wenig vor seiner Mitte stand die
Kieme, von drei doppelt-fiederigen Blattern gebildet, von denen das hin-

:
|
|
,

MUSEUM OF COMPARATIVE ZOOLOGY. To

derste an seinem Grunde noch ein kleines Blatt trug. Dicht hinter der Kieme
die wenig vortretende Analpapille, an ihrem Grunde rechts die feine Nieren-
pore. Die Mitte des Schwanzes (des hinter der Kieme liegenden Ko6rpertheils)
trug (in einer Lange von 5mm.) einen Kamm, der sich in mehrere, groéssere
und kleinere, zusammengedrickte, oben gerundete Hocker erhebt. Die Kor-
perseiten ziemlich hoch; die (zusammengezogene) Genitaloffnung in der Mitte
zwischen dem Hinterrande des Rhinophorkammes und der Kieme, etwas nach
oben liegend. Der Fuss wie gewohnlich schmal ; der Vorderrand mit tiefer
(Fig. 26) Furche ; das Fussgebrime schmal.

Die Eingeweide schimmerten am Vorderk6érper undeutlich (weisslich) durch.
— Das Peritonaeum farblos. Die Eingeweidehdhle sich nur bis etwa dicht
hinter der Gegend der Analpapille erstreckend.

Das Centralnervensystem in eine diinne Bindesubstanzhille eingeschlossen ;
die Ganglien ziemlich dick. Die cerebro-pleuralen Ganglien je ein fast 8-
Zahl-ahnliche Masse bildend ; die beiden Abtheilungen derselben fast gleich-
gross; die rundlichen, von vorne nach hinten nur ein wenig zusammenge-
drickten, pedalen Ganglien etwas grésser als die pleuralen ; die gemeinschaft-
liche Commissur ziemlich kurz, nur noch ein halbes Mal so lang wie der
Querdurchmesser des plenralen Ganglions. Die proximalen Riechknoten fast
sessil, zwiebelformig; die distalen ein wenig grosser, kugelformig. Die buc-
calen Ganglien abgeplattet-rundlich, fast unmittelbar mit einander verbunden,
etwa so gross wie die proximalen Riechknoten ; gastro-oesophagale Ganglien
wurden nicht gesehen.

Die Augen kurzstielig, mit schwarzem Pigment, hellgelblicher Linse. Die
Ohrblasen etwas kleiner, mit runden und ovalen Otokonien gewohnlicher Art
gefillt. Die Blatter der Keule der Rhinophorien ohne Spikel. In der inter-
stitvellen Bindesubstanz kamen erhartete Zellen nur sparsam vor.

Die Mundrohre kurz und weit, an Lange und in Durchmesser 1.5 mm.
messend. — Der Schlundkopf von gewohnlicher Form, 2.6 mm. lang bei einer
Hohe und Breite von beiliufig 2mm.; vom hintersten Theil der Unterseite
ragt die Raspelscheide 0.75 mm. hinab ; die Lippenscheibe ziemlich gross, nur
von einer, besonders im Innenmunde und oben, ziemlich starken gelblichen
Cuticula tiberzogen. Die Zunge stark, etwas abgeplattet. In der hell horn-
gelben, in der Randparthie (wegen der Aussenplatten) braungelben Raspel 10
Zahnplattenreihen ; weiter nach hinten fanden sich deren 4 entwickelte und
2 jungere ; die Gesammtzahl der Reihen somit 16. Die vorderste Reihe war
auf die mediane Platte und die letzte Aussenplatte reducirt. Die Reihen sonst
an jeder Seite der medianen eine laterale und drei Aussenplatten enthaltend.
Die medianen und die Aussenplatten stark horngelb, die lateralen fast farblos.
Die Breite der vordersten medianen Platten 0.24, der hintersten 0.29 mm. ;
die Lange der lateralen Platten hinten an der Zunge 0.56, die Linge der Aus-
senplatten von innen nach aussen meistens 0.2-0.18-0.14 mm. Die medianen
Platten (Taf. II. Fig. 3) flach, mehr breit als lang ; der Vorderrand nicht um-
gebogen, convex, nicht oder kaum in der Mittellinie ausgerandet, mit etwas
vortretendem gerundetem Ecken; der Hinterrand mit dem Vorderrande parallel,

174 BULLETIN OF THE

etwas diinner als dieser (Fig. 3a); die Seitenrander fast gerade, mit einander
parallel. Die lateralen Platten (Taf. Il. Fig. 1 aa, 2) gross, unregelmassig,
sichelformig oder eigentlich gleichsam unregelmassige, ein wenig zusammenge-
bogene, zum Theil am Ricken ausgehohlte, in den Randern theilweise verdickte
und oben kurz-gekluftete (Fig. 1, 2) Blatter bildend; von dem Doppelthaken
der Spitze ist der untere Theil der kleinste. Von den drei Aussenplatten, die
alle vorne breiter waren, war die innerste fast doppelt so gross wie die folgende,
subquadratisch, mit einem ziemlich starken, nach innen gerichteten Kamm
(Fig. 1 6b). Die folgende Platte war ziemlich convex (Fig. 1 cc, 3aa), mit
Andeutung einer Lingsleiste. Die. dusserste Platte (Fig. 1 dd, 3 bb) war auch
convex, nicht halb so gross wie die vorige.

Die gelblichweissen, nicht recht dicken Speicheldriisen begleiteten den iiber
den Schlundkopf verlaufenden Theil der Speiserdhre; die Ausfiihrungsgange
kurz.

Die Speiseréhre etwa 3.5 mm. lang, vorne weiter, hinten schmaler, sich oben
am Vorderende der hinteren Eingeweidemasse in die Leberhohle (den Magen)
offnend. Der Darm aus der letzteren an der linken Seite der Cardia aus-
gehend; in seinem Verlaufe erst links, dann quer, dann rechts und nach
hinten gehend, mehrere grosse Biegungen machend; ausgestreckt beilaufig
16mm. messend bei einem Durchmesser von 1-1.5 mm., in seiner ganzen
Lange (wegen seines Inhalts) kalkweiss. Der Inhalt des Darmes und der
weiten Leberhohle war thierische Masse, theilweise von Bryozoen herrihrend,
und parenchymatose-pflanzliche. ,

Die hintere Eingeweidemasse (Leber) 11 mm. lang bei einer Hohe und
Breite von 4; sie war fast cylindrisch, hinten gerundet, vorne schief nach unten
und vorne abgestutzt; ihre Farbe war aussen schwirzlichgrau, dieselbe aber
zim grossten Theil von der Zwitterdriise verdeckt; die Substanz der Leber
und die Wand der weiten Hohle schwarz oder schwarzbraun.

Das Herz wie gewohnlich. Die Blutdriise gelblichweiss, queroval, ziemlich
abgeplattet, hinter dem Centralnervensystem liegend und etwa so breit wie
dieses. — Die Niere wie gewohnlich, der Nverentrichter birnformig, etwa
0.55 mm. lang, mit etwa 10 Hauptfalten.

Die gelbliche Zwitterdriise mit einem fast einfachen Lager von dicht-
stehenden meistens an einander stossenden Lappchen (Taf. II. Fig. 4), die
Leber fast iiberall iiberziehend. In den Ovarialfollikeln der Lappchen grosse
Eierzellen, in der nicht sehr abgeplatteten Testicularplatte keine reife Zoo-
spermien. Der diinne weissliche Zwitterdriisengang frei an der rechten Seite
der Cardia vortretend und langs der Speiserdhre an die vordere Genital-
masse verlaufend. Diese letztere, etwa 2.5 mm. lang bei einer Breite und
Dicke von 2.2 mm.; die Ausfiihrungsgange noch 1.6 mm. vortretend; das
Vorderende der Masse wird von der Schlinge der Ampulle des Zwitterdrisen-
ganges gebildet ; hinten an der oberen Seite liegt die grosse Samenblase, und
an der ausseren (rechten) Seite schlangelte sich der Samenleiter. Die er-
wihnte Ampulle wurstférmig, stark zusammengebogen, ausgestreckt 3 mm.
lang bei einem Durchmesser von 0.8. Der stark geschlangelte prostatische

MUSEUM OF COMPARATIVE ZOOLOGY. 175

Theil des Samenleiters etwa 5 mm. lang; der muskulose beilaufig 4mm. lang,
eine grosse Schlinge bildend, unten endigte derselbe als eine kleine Glans am
Boden des etwas dickeren, etwa 0.7mm. langen Praeputiums. In fast dem
unteren Viertel des musculdsen Samenleiters findet sich eine sich bis in die
Glans fortsetzende Bewaffnung. Dieselbe besteht aus etwa 10-12 Quincunx-
Reihen von kleinen gelblichen Haken, die eine Hohe bis zu beilaufig 0.02 mm.
erreichen (Fig. 5). Die Spermatotheke (Taf. Il. Fig. 4a) kugelformig, von
etwa 1mm. Durchmesser. Die (von dem Samenleiter verdeckte Spermato-
cyste ein wenig kleiner, auch (Fig. 4d) kugelfOrmig; ihr Ausfiihrungsgang
etwas langer als die Blase, in den uterinen Ausfiihrungsgang der Spermato-
theke (Fig. 4c) titbergehend, Die Schleim- und Eiweissdriise (wie alle die
ibrigen der vorderen Genitalmasse gehorenden Organe) weiss und gelblich-
weiss. Das Vestibulum genitale mit starken Langefalten.

Diese unzweifelhaft neue Form der Gattung Nembrotha scheint der N. dia-
phana am Nachsten zu stehen, unterscheidet sich aber schon im Aeusseren
deutlich genug durch die starken Rhinophorkamme und durch die ganz ver-
schiedene Farbenzeichnung, noch dazu durch die etwas verschiedene Beschaf-
fenheit der Raspel.

Fam. PHYLLIDIADA.

PHYLLIDIOPSIS, Beu.

R. Bergh, Neue Beitr. zur Kenntn. d. Phyllidiaden. Verh. d. k. k. zool. bot. Ges.
in Wien, XXV., 1875, pp. 661, 670-673, Taf. X VI. Fig. 11-15.

R. Bergh, Malacolog. Unters. (Semper, Philipp. Il. ii.), Heft XVI., 2, 1889, pp.
859, 866-867, Taf. LXXXIV. Fig. 23-27.

Dorsum ut in Phyllidiis propriis. Apertura analis dorsalis.
Tubus oralis ut in Doriopsidibus ; glandula ptyalina discreta (?).

Die Phyllidiopsen bilden gewissermassen ein interessantes Zwischenglied
zwischen den Phyllidien und den Doriopsen. Im Ganzen sehen sie den achten
Phyllidien ahnlich aus und haben dieselbe Lage der Analoffnung. Die Ten-
takel sind sehr klein und wie bei den Doriopsen ihrer ganzen Lange nach
angeheftet. Die Mundrohre ist wie bei den Doriopsen ; es scheint, auch wie
bei den Doriopsen, eine gesonderte Mundspeicheldriise (Gland. ptyalina) vor-
zukommen.

Die Gruppe ist, wie andere Phyllidiaden, nur aus den tropischen Meeres-
gegenden bekannt, und umfasst bisher nur die unterstehenden Arten.

1. Ph. cardinalis, Bgh. M. pacific. (Ins. Tonga).
2. Ph. striata, Bgh. M. africano-indic. (Maurit.).
3. Ph. papilligera, Bgh., n. sp. M. mexicanum.

176 BULLETIN OF THE

Phyllidiopsis papilligera, Bgh., n. sp.
Tafel II. Fig. 7-14.

Hab. M. mexicanum.

Von der Form lag nur ein einziges Individuum vor, aus einer Tiefe von
101 Faden an 25° 33’ Br. und 84° 21’ L. (d. h. im mexicanischen Golfe) hinauf
gefischt,

Das in Alkohol bewahrte Individuum hatte eine ZLénge von 12 bei einer
Breite bis 11 und einer Hohe bis 4.5 mm.; die Breite des Mantelgebrames
3mm., die Hohe der (zuriickgezogenen) Rhinophorien 1.5 mm.; die Lange des
Fusses 7.5 bei einer Breite bis 6mm. — Die Grundfarbe des Riickens weiss-
lich, an derselben viele runde und ovale, grosse und kleine, sammetschwarze
(bis 2.5 mm. breiten) Flecken, die meistens Papillen tragen, welche theil-
weise auch schwarz sind; an der weisslichen Unterseite des Mantelgebrames
schimmerten die schwarzen Riickenflecke durch; die ibrige Unterseite (gelb-
lich-) weisslich. Die Rhinophorien und der Aussenmund gelblich.

Die Form fast rundlich, etwas gewolbt (Fig. 7), mit breitem dinnem Mantel-
gebrame. Die Consistenz des Thieres nicht hart, nicht recht steif. Der Ricken
eben, aber mit ziemlich zahlreichen, bis etwa 1.6mm. hohen, zusammenge-
drickten, mehr oder weniger, besonders an der einen (meistens vorderen)
Seite, schwarzfarbigen Papillen bedeckt. Die Rhinophoroffnungen (Fig. 7 a)
ziemlich weit von einander liegend, die starke Rhinophorkeule mit etwa 20-25
Blattern. Die Analpore median hinten am Riicken (Fig. 7b). Der innerste
Theil des Mantelgebrames ist dicht mit quergehenden, meistens an der Mitte
hoheren, bis 1.5 langen diinnen Blattern bedeckt ; hinten begegnen sich die
Blatterreihen tiber den Schwanzgrund, vorn erstrecken sie sich bis an den Aus-
senmund; die Anzahl der Blatter jederseits 45-50. Keine Spur von Tentakeln
wurde gesehen; der Aussenmund fand sich als eine starke durchbohrte Papille
vor dem Vorderrande des Fusses. Die Genitalpore an gewohnlicher Stelle
der niedrigen (rechten) Koérperseite. Der Fuss gross, breit, vorne abgestutzt-
gerundet und mit Randfurche, das Fussgebrame nicht schmal, der Schwanz
nicht kurz.

Das Centralnervensystem (Fig. 9) zeigte die cerebro-pleuralen Ganglien nie-
renformig, schrage gegen einander liegend, nach vorne convergirend (Fig.
9 ab); die pedalen Ganglien an der Unterseite der pleuralen liegend, grosser
als diese, rundlich (Fig. 9 cc); die gemeinschaftliche Commissur doppelt, diinn
(Fig. 9d). Die proximalen Riechknoten fast sessil, zwiebelformig (Fig. 9);
die distalen kugelformig. Die buccalen Ganglien (Fig. 13) an gewohnlicher
Stelle, rundlich, einander berithrend.

Die Augen fast sessil, von 0.1 mm. Diam., mit reichlichem schwarzem Pig-
ment (Fig. 9). Die Ohrblasen weit von den vorigen an der Unterseite (Fig. 9)
der Gehirnknoten liegend, von beilaufig 0.06 mm. Diam. ; etwa 50-100 ovalen
Otokonien von einem Durchmesser bis 0.013 mm. enthaltend, unter denen ein
grésserer rundlicher (Fig. 12). In den Blattern der Keule der Rhinophorien,

MUSEUM OF COMPARATIVE ZOOLOGY. LTE

wie gewobnlich, dinne, mehr oder weniger erhartete, kirzere und langere
Spikel, die letzteren zum grossen Theile auf dem freien Rande senkrecht
stehend (Fig. 10). In der Haut des Riickens eine Unmasse von grosseren
und kleineren Spikeln und Bindel von solchen, welche auch unter der Loupe
schon durchschimmerten (Fig. 7); im Mantelgebrame waren dieselben zum
grossen Theile senkrecht und schrag (Fig. 11) gegen den Rand geordnet; sonst
lagen sie mehr ungeordnet. Die Spikel waren von den gewohnlichen bei
diesen Thieren vorkommenden Formverhaltnissen (Fig. 10, 11), meistens stark
erhartet, oft glasartig; von einem Durchmesser bis 0.16 mm., von sehr wech-
selnder Lange, die oft bis iber 0.4mm. stieg. Bindel von ahnlichen Spikel
stiegen in die Papillen bis an ihre Spitze auf (Fig. 8). In der interstitiellen
Bindesubstanz kamen iiberall Massen von grésseren und kleineren meistens
stark erharteten Spikel vor, so wie verkalkte Klumpen und Kugeln.

Durch den Aussenmund war das Ende des Mundrohres etwas hervorge-
stilpt; unter jenem fand sich die Oeffnung der Mundrohrendrise (Fig. 13 q).
Die gelblichweisse Mundrohre (Fig. 13a, 14a) weit, nicht kurz, 2 mm. lang,
hinten mit (Fig. 13) einer kreisartigen Einschnurung ; die Innenseite (Fig.
14a) mit Langsfalten; in das vertiefte Hinterende derselben senkt sich der
gelbliche Schlundkopf, der am Boden der Mundrohrenhohle stark vorspringt
(Fig. 140). Dieser Schlundkopf (Fig. 13 b) von gewohnlicher Form, fast
cylindrisch, von starker gelblicher Cuticula an der Innenseite iberzogen, etwa
2mm. lang; am etwas engeren Hinterende des Schlundkopfes (Fig. 13 c) die
buccalen Ganglien. Hinter den letzteren finden sich (Fig. 13 d) die gewohn-
lichen, hier fast kugelformigen eigentlichen (hinteren) Speicheldrisen (G1.
salivales) (Fig. 13d). Es kommt aber jederseits (?) noch eine langliche, et-
was lappige, weissliche vordere Speicheldrise (Gl. saliv. access.) vor, (Fig. 13 f),
die neben dem Schlundkopf das Hinterende der Mundrohre durchbohrt; sein
Hinterende geht in einen bindegewebigen Strang iber. Unter dem Schlund-
kopf liegt die lappige, weissliche Mundrohrendrise (Gl. ptyalina), welche in
einen starken Ausfiihrungsgang tbergeht, die sich hier nicht in die Mund-
rohre, sondern unmittelbar unter dem Aussenmunde 6ffnete (Fig. 13 gq).

Das Hinterende des Schlundkopfs geht etwas enger in die gestreckt-schlauch-
formige Speiserohre (Fig. 13 ¢) iber, welche ein wenig kirzer als der Schlund-
kopf ist und die obere Seite der hinteren Eingeweidemasse durchbohrt. Die
in dieser letzteren eingeschlossene Magenhohle nicht weit. Der Darm die
Eingeweidemasse am Anfang des letzten Drittels durchbrechend und in ge-
wohnlicher Weise verlaufend. — Die Verdauungshohle war leer.

Die hintere Eingeweidemasse (Leber) 5.5 mm. lang bei einer Breite von
4 und einer Hohe von 3mm., vorne schrag abgestutzt, hinten gerundet ; die
Substanz gelblichweiss.

Das querliegende Pericardium ziemlich gross; das Herz wie gewohnlich.
Die Blutdrise gerundet-viereckig, graulichweiss. Die Nvere in gewohnlicher
Weise die obere Seite der hinteren Eingeweidemasse iiberziehend; die Urin-
kammer wie gewohnlich.

Die Zwitterdriise durch mehr gelbliche Farbe von der Leber hier und da

VOL. XIX.—No 3 12

178 BULLETIN OF THE

unterscheidbar; in den Lappchen Eierzellen und reife Zoospermien. — Die
vordere Genitalmasse gerundet-viereckig, beilaufig 3mm. lang. Die weissliche
Ampulle des Zwitterdriissenganges wurstformig gebogen. Die Samenblasen
weisslich; die Spermatotheke kugelf6rmig, die Spermatocyste eiformig. Der
Samenleiter nicht lang; das Dasein einer Penis-Bewaffnung konnte nich nach-
gewiesen werden. Die Schleimdriise weisslich, die Eiweissdriise mehr gelb.

Bisher war keine am Riicken Papillen-tragende Form von Phyllidiaden
bekannt worden. Diese nimmt in dieser Beziehung eine ahnliche Stellung
unter den Phyllidiaden wie die Echinodoris? unter den Doriden ein.

1 R. Bergh, Neue Nacktschnecken der Siidsee, II. Journ. d. Mus. Godeffroy,
Heft VI., 1874, pp. 19-22, Taf. III. Fig. 4-20.

Fig.

Fig.
Fig.

Fig.
Fig.
Fig.
Fig.

Fig.
Fig.
Fig.

Fig.
Fig.
Fig.

Fig.
Fig.
Fig.

ce oe

Dd;
12.

13.

14.

MUSEUM OF COMPARATIVE ZOOLOGY. 179

TAFEL-ERKLARUNG.

TAFEL I.

Tethys leporina (L.).

Verdauungssystem; a das (an der Unterseite gekluftete) Mundrohr;
6 vorderer, c hinterer Theil der Speiserohre; dd die in den ganz rudi-
mentaren Schlundkopf einmiindenden Speicheldriisen, zwischen den
Hinterenden derselben die buccalen Ganglien; e Hinterende des (ersten)
Magens, f zweiter Magen, g Darm; h Zweig der rechten Nebenleber
in das Rhinophor, i in die vorderste rechte Papille; k/ linke Nebenleber
mit ihren Zweigen, m Hauptausfiihrungsgang der Hauptleber.

Speicheldriise (linke), mit Cam. gezeichnet ( Vergr. 55).

Otocyste, mit Cam. gezeichnet (Vergr. 350); a Stiel.

Chromodoris punctilucens, Bgh.

Lippenscheibe mit Mundoffnung und Lippenplatte.

Stiick der Lippenplatte.

Grosste Elemente derselben.

Von der Rhachisparthie der Raspel ; a rhachidiale Verdickung, bb innerste
Seitenzahnplatte.

Zahnplatte aus dem inneren Drittel einer Reihe.

Eine der grossten Platten.

Aeusserer Theil zweier Zahnplattenreihen ; a aiusserste Platte der Reihen.

Fig. 5-10 mit Cam. gezeichnet (Vergr. 350).

Chromodoris scabriuscula, Bgh.

Elemente der Lippenplatte, von vorne.
Aehnliche, von der Seite.

Stiick der Raspel; a mediane Platte.

Fig. 11-13 mit Cam. gezeichnet (Vergr. 350).
Mediane Platte, von oben.

15,16. Zahnplatten vom inneren Drittel einer Reihe.

17.

Eine der grossten Platten.

180

Fig. 18.

Fig. 19.

Fig. 20.
Fig. 21.
Fig. 22.
Fig. 23.
Fig. 24.
Fig. 25,
Fig. 26.

Fig.
Fig.
Fig.
Fig.
Fig.

Hig, 6.

Fig. 7.

Fig. 8.
Fig. 9.

Fig. 10.
mig; 11:
Fig. 12.
Fig. 13.

Fig. 14.

Seer ee

BULLETIN OF THE

Aeusserste Platte einer Reihe.

Fig. 14-18 mit Cam. gezeichnet (Vergr. 750).

a muskuloser Theil des Samengangs, 6 Praeputium mit zuriickgezogener
Glans; mit Cam. gezeichnet ( Vergr. 55).

Phlegmodoris ? anceps, Bgh.

Rhinophorscheide, a Grund; mit Cam. gezeichnet (Vergr. 55).

Hockerchen des Riickens, mit Cam. gezeichnet (Vergr. 200).

Mundroéhre, aa Retractoren, 6 Driisen am Hinterrande des Mundrohres.

Innerster Theil einer Zahnplattenreihe; a erste Platte.

Aehnlicher von zwei Reihen, ab erste Platte derselben.

Eine der grossten Platten.

Aeusserster Theil zweier Zahnplattenreihen mit 8 und 9 Platten, aa
ausserste.

Fig. 23-26 mit Cam. gezeichnet (Vergr. 350).

2 Meu!) 0) ied O Coe aa

Nembrotha gratiosa, Bgh.

Rhinophorkamm, in 6 den Riicken iibergehend, c Rhinophoroffnung.
a Tentakel, 5 Vorderrand des Fusses.

Mediane Zahnplatte, mit Cam. gezeichnet (Vergr. 200), a Hinterrand.
Lappchen der Zwitterdriise.

Haken der Penis-Bewaffnung, mit Cam. gezeichnet ( Vergr. 750).

Phlegmodoris% anceps, Bgh.

Stiick der Riickenhaut, vom Rande gesehen; mit Cam. gezeichnet (Vergr.
200).

Phyilidiopsis papilligera, Bgh.

Das Thier, von der Riickenseite; a Gegend der Rhinophor-Oeffnungen,
b Gegend der Analpore.

Eine der kleineren Riickenpapillen.

Das Centralnervensystem, mit Cam. gezeichnet (Vergr. 55); ab cerebro-
pleurale, cc pedale Ganglien; d gemeinschaftliche Commissur.

Rand eines Rhinophor-Blattes, mit Cam. gezeichnet (Vergr. 350).

Vom Mantelrande; mit Cam. gezeichnet (Vergr. 55).

Otocyste, mit Cam. gezeichnet (Vergr. 350).

a Mundrihre, b Schlundkopf, c buccale Ganglien, d Speicheldriisen (Gl.
salivales), e Speiserhre, f Accessorische Speicheldriisen (GI.
access.), gg Ausfiihrungsgang der Mundrohrendriise (Gl. ptyalina).

a gedffnete Mundrohre, b Vorderende des Schlundkopfs.

Fig.

Fig.
Fig.

Fig.

’ Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.

Fig.

Fig.

Fig.

Fig.

14.

15.

16.

MUSEUM OF COMPARATIVE ZOOLOGY. 181

. TAFEL III.
Nembrotha gratiosa, Bgh.

Zwei Reihen von pleuralen Zahnplatten (linker Seite) von oben; aaa
laterale Platten ; 66 innerste, cc mittlere, dd dusserste Aussenplatte.

Laterale Platte von der Riickenseite.

Aeusserster Theil dreier Zahnplattenreihen mit den je zwei dussersten
Aussenplatten.

Fig. 1-3 mit Cam. gezeichnet (Vergr. 200).

a Spermatotheke, 6 vaginaler und ¢ uteriner Gang; d Spermatocyste,
e Diverticulum des Ausfiihrungsganges der Spermatocyste.

Chromodoris sycilla, Bgh.

Elemente der Lippenplatte.

Zwei Zahnplatten aus der Mitte einer Reihe der Zunge.

Zahnplatte vom inneren Zehntel einer Reihe.

Zwauzigste Zahnplatte von aussen ab.

Hine der innersten Seitenplatten.

Vierte Zahnplatte, von aussen ab.

Dritte Zahnplatte, von aussen ab.

Die drei 4ussersten Zahnplatten, von innen; a dusserste.

Fig. 5-12 mit Cam. gezeichnet (Vergr. 350).

Hinterende des Korpers, von oben (Mantelgebrame), mit den weissen
Flecken; aa Driisenbildungen der Unterseite des Mantelgebrimes.

Chromod. Californiensis, Bgh.

(Vgl. R. Bergh, On the Nudibr. Gaster. Moll. of the North-Pacific Ocean.
IL, 1880, Pl. XIV. Fig. 5. Scientific Results of the Explor. of Alaska,
Wel. J Art, vi.; 2;) |

Hinterende des Mantelrandes, von der Unterseite, mit 6 Driisen beutel ;
a Fuss.

Chromod. sannio, Bgh.

(Vgl. R. Bergh, Malacolog. Unters. (Semper, Philipp. II. ii.), Heft X VII.,
1890, Taf. LX XXVII. Fig. 1.)

Hinterende des Mantelrandes, von der Unterseite, mit 4 grossen Driisen-
beuteln; a Fuss.

Chromod. elegans (Cantr.).

(Vgl. R. Bergh, Untersuch. d. Chromod. elegans und villafranca. Mala-
kozool. Blatter, XX V., 1878, Taf. I. Fig. 4.)
Driisenbeutel von der Unterseite des Mantelgebrimes.

oa

Blake Mollusca , Nudtibranchiata’

Levendal

ttbhranchiata’

‘ue

Blake 1 elusca’, lV

a. -_—
le
~~ At - - .

Blake. Mollusca; Nudibranchtata-

No. 4.— A Third Supplement to the Fifth Volume of the Terres-
trial Atr-Breathing Mollusks of the United States and adjacent
Territories. By W. G. BINNEy.!

As promised in the Second Supplement, the Eastern Province Species
are here given, with addenda to those of the other Provinces. My pur-
pose is to bring the subject down to this date. The “ Manual of Amer-
ican Land Shells,” published subsequently to Vol. V., must also be used
in connection with the present paper. I have added figures of many
species to replace those of Volume V.

Buriincton, New JERSEY, January 1, 1890.

SPECIES OF THE NORTHERN REGION.

It must be borne in mind that the Universally Distributed Species are also
found here. They are: —

Patula striatella, Anruony.

Microphysa pygmea, Drape.

Placed in this genus on account of the similarity of its jaw and lingual
dentition to those of other species of Microphysa. See 2d Suppl., p. 35.

Helicodiscus lineatus, Say.
Vallonia pulchella, Mix.

Pupa muscorum, Liyvy.
See below, p. 186, for vars. bigranata and Lundstrom.
It may readily be doubted whether this species is not rather confined to the
Northern Region.

1 The Terrestrial Air-Breathing Mollusks of the United States and the adjacent
Territories of North America, described and illustrated by Amos Binney. Edited
by A. A. Gould. Boston, Little and Brown, Vols. I., II., 1851; III., 1857. Vol.
IV., by W. G. Binney, New York, B. Westermann, 1859 (from Boston Journ. Nat.
Hist.). Vol. V., forming Bull. Mus. Comp. Zodl., Vol. IV., 1878. Supplement to
same, in same, Vol. IX. No. 8, 1883. Second Supplement, in same, Vol. XIII.
No, 2, 1886.

VOL. XIX.—NO. 4.

184 BULLETIN OF THE

Zonites nitidus, Mu...
arboreus, Say.
indentatus, Say.
See Suppl., p. 139.
Zonites minusculus, Biny.
Dall thus describes a var. Alachuana (Pr. U.S. Nat. Mus., 1855, 270): —

A form of it which, at first sight, looks different from minuscula is rather larger
than usual, and above shows no differences. On the base in the type the junction
of the inner lip with the body whorl takes place, following the course of the whorl,
inward from the middle line of the base of the whorl and generally about the inner
third. This gives a peculiarly thimble-shaped umbilicus. In the variety under
consideration, the above mentioned junction takes place outside of the middle line,
or even at the outer third, while the aperture is a little dilated. The result of this
is to show a much larger portion of the base of the penultimate whorl, and to alter
the facies of the umbilicus. For this form, found in Alachua County, Florida, I
would suggest the varietal name Alachuana.

Zonites viridulus, Mxe.
milium, Morse.
fulvus, Drape.

These will not be repeated in the lists of the various Regions into which the
Province may be divided. (See Vol. V., p. 17.)

The following are Northern Region Species: —
Vitrina limpida, Gp.
Angelicee, Brecx.
Vitrina exilis, More ter.
The distinction between the Eastern, Central, and Pacific Provinces not being
marked in these high latitudes, this species is given here. It might, perhaps,
with Patula pauper and Pupa borealis, rather be considered a species of the

Pacific Province.
Zonites Fabricii, Brecx.

Binneyanus, Morse.
ferreus, Morse.

Zonites exiguus, STIMPSON.
Plate ITI. Fig. 4.

The figures are copies of original drawings of Dr. Stimpson.
Zonites multidentatus, Brnney.
See Suppl., p. 144.

Acanthinula harpa, Say.
Patula asteriscus, Morse.

MUSEUM OF COMPARATIVE ZOOLOGY. 185

Patula pauper, Goutp.
See remarks under Vitrina extlis, above.

Pupa Blandi, Morse.
borealis, More tet.

See remark under Vitrina exilis.
The figure was drawn by me from a specimen collected at the
original locality.

Pupa decora, Goutp.

a e- ae Pupa borealis
Hoppii, Mouzer. enlarged. ”

Vertigo Gouldi, Binney.
Bollesiana, Morse.
A variety Arthuri, from Dakota, is mentioned by Von Martens, Gesell. Nat.
Freunde zu Berlin, 21 Nov., 1882, p. 140.
Very near, if not identical with, V. maliwm.
Vertigo simplex, Goutp.
ventricosa, Morse.

Very near, if not identical with, V. Gouldi.
Ferussacia subcylindrica, Liyn.
In the mountains of McDonnel Co., North Carolina, a colony of this species

was found by Mr. Hemphill. He found no colony of Vitrina, which might
be expected to exist at those high elevations.

Succinea Haydeni, W. G. B
Verrilli, Bianp.
Gronlandica, Brecx.
Higginsi, Buanp.
Totteniana, Lea.

Dr. Westerlund, in the “ Land- och Sottvatten-Mollusker ” of the Vega Ex-
pedition, quoted in the Manual of American Land Shells, pp. 473, 474, also
catalogues from Arctic America the following species: —

Limax hyperboreus, Wesrtrertunp. (See below, p. 205.)
Pupa arctica, Watt.
columella, Brenz.
Succinea chrysis, WestERLunp. (See p. 186.)
turgida, WEsTERLUND.
annexa, WEsTERLUND. (See p. 186.)
Vallonia Asiatica, Nevin.
Pupa edentula, Drapr. ?
signata, Ms.
Vertigo Bollesiana, var. Arthuri.

186 BULLETIN OF THE

Pupa muscorum, var. bigranata, Ross.
muscorum, var. Lundstromi, WrEsTER.LUND.
columella, Benz., var. Gredleri, Crxssin.
Krausseana, Re1nu.

Of the above, descriptions and figures are given of two only, Succinea chrysts
and S. annexa, which are copied here.

Succinea chrysis, WesTerLunp.
(Figures copied on my Plate I. Fig. 14.) .

Testa oblongo-ovata, solida, irregulariter transversim striata vel seepe costu-
lato-plicata, colore varia, seepissime spira pallidiore, apice rubro, anfractu ultimo
antice rotuntiore, subviolaceo-rufescente, postice pallidiore, ubique strigis trans-
versis numerosis albidis; spira elevata, acuta, anfr. 34, convexi, ultimus deorsum
lente attenuatus, penultimus subtus tumidulus, antepenultimus transyersalis,
extus depressus, sutura forte excisa, anfr. ultimo minutissimo; sutura perim-
pressa, apertura ovata, intus aureo-micans, pariete arcuatula, obliqua; peristoma
obscure marginatum, marginibus equaliter arcuatis (exteriore superne ad inser-
tionem forte curvato), in pariete callo tenuissimo albido conjunctis.
Long. 114, diam. 74, ap. 74 mm. 1.,5 mm. d.; long. 13, diam. 74, ap.
long. 9, diam. 74 mm.; long. 10, diam. 6, ap. long. 64, diam. 5 mm.

Asia: America, Port Clarence, Alaska.
etna I figure also a specimen from St. Michael’s, Alaska (Dall), which
chrysis. has usually been referred to a form of S. lineata.

Succinea annexa.
(Figures copied on my Plate I. Fig. 15.)

Testa elongato-ovata, fragilis, intus rugas incrementales fuscas (in spec. max.)
validas et extus abruptas dense striata, anfr. penultimo dense distincte spira-
liter lineata, anfr. ultimo transversim irregulariter alternatim rufo- et albido-
strigata ; sutura impressa; spira exserta, apice mamillata; afr. 4, ultimus
convexus, penultimus tumidus, antepenultimus altus, exitus convexus, sutura
tenui a precedente sejunctus, summus (subtus visus) globosus; apertura ovata,
pariete obliqua, columella arcuata, marginibus linea tenui alba junctis. Long.
11, diam. 8, apert. long. 8, diam. 6 mm.; long. 10, diam. 64, apert. long. 6,
diam. 44 mm.

Fort Clarence, Alaska.

INTERIOR REGION SPECIES.

Macrocylis concava, Say.

Zonites capnodes, W. G. B.
fuliginosus, GrirriTu.
friabilis, W. G. B.

MUSEUM OF COMPARATIVE ZOOLOGY. 187

Zonites levigatus, PFEIFFER.
Rugeli, W. G. B.

See Suppl., p. 138.
Zonites demissus, Binney.

The variety acerrus has been found near Fort Gibson, Indian Territory,
by Mr. Simpson.

Zonites ligerus, Say.

A variety Stonet is thus described by Mr. Pilsbry: “From Mr. Witmer
Stone I have received a form of Z. ligerus differing from the type in having a
concave, broadly excavated base, with comparatively wide umbilicus, collected
by him in New Castle Co., Del. The axis in the type is barely perforated ;
but in this form it is a millimeter or more wide, and the base around it
broadly concave.” (Nautilus, ILI. 4, p. 46, Aug., 1889.)

Zonites intertextus, Binney.
subplanus, Binney.

See Suppl., p. 139.
Zonites inornatus, Say.

sculptilis, Bianp.

Elliotti, Reprrep.

limatulus, Warp.
capsella, Gouxp.

Lawe, W.G. B.

See Suppl., p. 142, Plate II., Fig. E. The name is suggested for the shell
figured by me in Vol. V. (Fig. 44) as Z. placentula.
Zonites placentula, SaHuTtrLeworra.
See Suppl., p. 142. i

Zonites Wheatleyi, Brann.
See Suppl., p.141. Clingham’s Peak, N. C. (Hemphill).
Zonites petrophilus, Buianp.
Habersham Co., Ga.; Clarkesville, N. C. (Hemphill). See Suppl., p. 140.
Zonites Sterkii, Datu.

Shell minute, thin, yellowish translucent, brilliant, lines of growth hardly
noticeable, spire depressed, four-whorled ; whorls rounded,
base flattened, somewhat excavated about the centre, which is
imperforate; aperture wide, hardly oblique, not very high,
semilunate, sharp-edged, the upper part of the columella
slightly reflected ; upper surface of the whorls roundish,
though the spire as a whole is depressed. Greater diam. 1.1, — Zonites Sterkii,
height 0.52 mm. ae

188 BULLETIN OF THE

New Philadelphia, Ohio. Collected on a grassy slope, inclining to the north-
ward, and covered. with grass, moss, and small bushes, and so far has not been
found elsewhere. Clearly not young of a Pupilla or Zonites. It is probably
one of the smallest species known, and remarkable for its imperforate umbilicus.
The above forms a portion of the description by Dall of Hyalina Sterhii,
from Proc. U.S. Nat. Mus., XI., p. 214, Figs. 1,2, 3, 1888, The figure given
by me is drawn from an authentic specimen.

Zonites gularis, Say.
‘suppressus, Say.

cuspidatus, Lewis.

See Suppl., p. 148.

Miss Law thus wrote from Philadelphia, Tenn., of this species: ‘ Unlike
gularis it seems to be a rare shell, and I find it only by scraping off the
surface of the ground in the vicinity of damp mossy rocks. Its habits are
more like placentula than gularis. Neither Miss Clara Bacome nor I ever mis-
take one for a gularis, even before picking it up ; the thickened yellow splotch
near the lip, and the thinner spot behind, showing the dark animal through it,
as well as its more globular form, particularly on the base, make it look very
different when alive.” :

Zonites lasmodon, Putters.
Plate III. Fig. 5.
Enlarged drawings by Miss Lawson are given of this species.
Zonites macilentus, Suuttt.
See Suppl., p. 143.
Zonites significans, Bianp.
See Suppl, p. 144.
Zonites Andrewsi, W. G.B.
See Suppl, p. 144.
Zonites internus, Say. 3
Vitrinizonites latissimus, Lewis.

See Suppl., p. 145; for other localities, see Man. of Am. Land Sh., p. 231,
Also in Washington Co., N. C., and in Watauga Co. at Banner’s Elk (Hemp-
hill).

Limax campestris, Binney.

Limax montanus, castaneus, occidentalis, hyperboreus, and Hemphilli are

probably identical with this.

Tebennophorus Caroliniensis, Bosc.
Tebennophorus dorsalis, Binney.
Tebennophorus Wetherbyi, W. G.'B.

See Plate VI. Fig. F.

MUSEUM OF COMPARATIVE ZOOLOGY. 189

Tebennophorus Hemphilli, W. G. B.
Plate VI. Fig. H.

See Man. of Amer. Land Sh., p. 247.

The animal is long, narrow, cylindrical, with pointed tail. Its color is
black. The jaw is strongly arched, with median projection, and four or five
ribs converging to the centre, all crowded on the middle third, the outer thirds
being ribless. The lingual membrane has 24—14~—1—-14-24 teeth, all of same
types as figured by Morse for that of 7’. dorsalis. Length of largest individ-

ual contracted in spirit 25 mm.
The penis sac is long, cylindrical, receiving retractor muscle and vas deferens

at its summit.
Patula solitaria, Say.

alternata, Say.
Cumberlandiana, Lza.
perspectiva, Say.

Bryanti, Harper.
See Suppl., p. 147.

Helicodiscus fimbriatus, Weruersy.

See Suppl., p. 148.

A curious form, wanting the epidermal fringe and most of the revolving
ridges, was found in great numbers near Fort Gibson, Indian Territory, by Mr.
C. T. Simpson. The same form has been found by Mr. Hemphill on Salmon
River, Idaho. He proposes for it the name Salmonacea.

Strobila labyrinthica, Say.
A form from Venezuela, without the cost, is noticed by Dall as var. Morsei
(U.S. Nat. Mus. Proc., 1855, p. 263).

Polygyra leporina, Goutp.
Hazardi, Brann.
Troostiana, Lexa.
fastigans, Say.
Stenotrema spinosum, Lea.
labiosum, Goutp.
Kdgarianum, Lea.
Kdvardsi, Buann.
barbigerum, Reprietp.
stenotremum, Ferussac.
hirsutum, Say.
A widely separated locality is the bank of the Yaqui River, near Guaymas

(Palmer).
Stenotrema maxillatum, Goutp.
monodon, Rackert.

Triodopsis palliata, Say.

190 BULLETIN OF THE

Triodopsis obstricta, Say.
appressa, Say.

It is quoted by Von Martens from the banks of the Columbia River, but
from drawings and description of the single specimen found by Kraus, kindly
sent me by Dr. Von Martens, it appears that the species was confounded with
flattened forms of Mullani or devius. |

Triodopsis inflecta, Say.

A depauperated form of this species is about being described and figured as
T. edentula by Mr. F. A. Sampson.

Triodopsis Rugeli, Saurrieworru.
' tridentata, Say.

The deformed specimen figured is one of appressa, not of this species.

Triodopsis fallax, Say.
introferens, Buanp.

Van Nostrandi, B :
Also, Jacksonville, Florida. ae

Mesodon major, Binney.

On Plate I. Fig. 2, I have figured the dentition of an individual of this spe-
cies differing from that figured in Vol. V. Plate VIII. Fig. G, by wanting the
side cusps and cutting points of the central and
first lateral teeth. The individual from which the
lingual was extracted is labelled B in the collec-
tion given by me to the United States National
Museum. Fig. 3 gives an outer lateral of the
same membrane, on which the side cusp and cut-
ting point are present, Fig. 1 gives a central tooth
with side cusps and cutting points from the mem-
brane of the specimen labelled A.

The figures show a larger range of variation in
the dentition of individuals of the same species than would have been antici-
pated. (See also M. Andrewsi.)

Mesodon albolabris, Say.
Andrewsi, W. G. B.

In the Manual of American Land Shells, p. 302, I have described and fig-
ured specimens of a larger form of this species, which would be called major
by most collectors, but which has the genitalia and lingual dentition of An-
drewst. (See figure above.)

The penis sac of Andrewst was described by me as constricted in the middle.
Further study has convinced me that it is rather twisted than constricted.
On Plate I. Fig. 4, I give a figure of the genitalia to show this; and in Fig. 5,
the penis sac of still another individual. :

Mesodon major.

MUSEUM OF COMPARATIVE ZOOLOGY. 191

In studying the lingual membrane of many individuals of M. Andrewsi, I
have found some variation. I give here notes on membranes of specimens
labelled as specified in the Binney collection in the United States National
Museum.

AA. 60-1-60 teeth, with about 14 laterals on each side.

N. 51-1-51 teeth, with 11 laterals ; some extreme marginals have decid-
edly multifid cusps.

Q, from Hayesville, N. C., has also about 11 laterals.

V has 9 laterals, 60-1-60 teeth.

M. 60-1-60 teeth, with about 14 laterals. Some outer laterals have side
cusps : one is figured on Plate I. Fig. 12.

G has same count as M; no side cusps to outer laterals.

N has 64—-1-64 teeth, with 14 laterals. The extreme laterals have side
cusps.

L has 61-1-61 teeth, with 11 laterals ; no side cusps on outer laterals.

J same. 64—1-64 teeth, with 14 laterals.

B. 60-1-60 teeth, with 16 laterals, none with side cusps.

F. All laterals, even first, have decided side cusps (see Plate I. Fig. 10)
and cutting points: and marginals also (Fig. 11). 50-1-50 teeth, with 15
laterals.

K. 53-1-53 teeth, with 14 laterals.

I. 50-1-50 teeth, outer laterals with side cusps.

O. 68-1-68 teeth, with 14 laterals.

As remarked above, most collectors will refer this large form of Andrewsi to
major. It differs from that species as hitherto understood very decidedly
in its lingual dentition and genitalia. In its shell, also, the species differs
from the generally known major in so marked a manner, that from it alone I
could say, before examination, what were the characters of the dentition and
genitalia of every specimen collected by Mr. Hemphill in the mountains of
North Carolina. One of the puzzling questions to be left to future solution is
the limitation of albolabris, major, and Andrewst. It must be studied from the
lingual dentition and genitalia, as well as from the shell. The student must
also consider whether the Helix major of the Boston Journal and of the Ter-
restrial Mollusks are the same species.

Practically, the simplest way of treating specimens in collections is to refer
to a variety of albolabris all forms more resembling that species than they do
the major of the Terrestrial Mollusks, and to call major all specimens most
nearly conforming to the figure and description of that species in Terrestrial
Mollusks of U.S., Vols. IT. and III. In the former category would be placed
the major of the Boston Journal; in the latter, the large forms I have referred
to Andrewst in Manual of American Land Shells, such, for instance, as are fig-
ured in Fig. 3223, repeated here, ante, page 190. This variety of albolabris
and this major, as above identified, would be found to differ widely in den-
tition and genitalia, the former in these respects resembling albolabris, the

192 BULLETIN OF THE

latter Andrewst. The latter species must also be recognized as subject to
variation, rendering it in some cases difficult to separate from major, —never
from the large variety of albolabris.

The original specimen of major of the Terrestrial Mollusks was included in
the collection given by Mr. J. S. Phillips to the Philadelphia Academy of
Sciences. The points in which it differs from the large form of albolabris are
pointed out in Terrestrial Mollusks, Vol. II. p. 98.

Mesodon multilineatus, Say.
Pennsylvanicus, GREEN.
Mitchellianus, Lea.
elevatus, Say.
Clarki, Lea.
Christyi, Buanp.
exoletus, Binney.
Wheatleyi, Briann.
dentiferus, Binney.
In aspecimen collected by Mr. Hemphill, at Banner’s Elk, N. C., I found the
retractor muscle of the penis sac near its junction with the vas deferens, not
at half the length of the latter. There was no constriction to the penis sac.

Mesodon Wetherbyi, Buanp.
thyroides, Say.
clausus, Say.
Downieanus, Bianp.
Lawe,: Lewis.
profundus, Say.
Sayi, Binney.

Pupa pentodon, Say.

The enlarged view of the aperture gives on the left P. Tappanzana, on the
right P. curvidens. |

Under the name of Pupilla Floridana, Mr. Dall has described what I con-
sider as a form of this species in Proc. U. 8. Nat. Mus., 1885, p. 251, Plate
SVL Aig, 1A.

Shell greenish spermaceti-white ; when living, the tissues of the animal show
with pale salmon-color through the shell in the apical whorls; surface smooth or
lightly striated, with a tendency to retain dirt upon itself; form subcylindrical, with
a rather obtuse apex, the last whorl forming nearly half the shell; suture evident ;
whorls five, neatly rounded; aperture longer than wide; lip white, thin, reflected ;
teeth about nine, of which there are generally three larger than the rest, their tips
nearly meeting, and their bases mutually nearly equidistant; one is on the pillar,
one on the body whorl, and one on the anterior margin; on either side of the latter
are two generally subequal much smaller denticles. Lon. 1.60, lat. 0.75 mm.

MUSEUM OF COMPARATIVE ZOOLOGY. 193

Habitat. — Under loose oak bark, oak hamak, Archer, Alachua County, Florida,
April, 1885, W. H. Dall, sixteen specimens.

This is one of our smallest species, and is related to P. pentodon and P. pellucida.
It is about half the size of the former and much more slender. Its teeth recall
those of P. curvidens, Gould, in their arrangement, but the shell is more cylindrical
and smaller than it is in P. pellucida (servilis) as figured by Gould. The teeth are
more numerous than in the latter shell, and set, as in P. pentodon, in one series ;
not, as in pellucida, partly deeper in the throat.

I describe this with some hesitation, for the condition in which the Pupide and
Vertigos of North America are is most unsatisfactory, and offers an excellent field
to some careful student who shall be able to examine and figure large series of
authentic specimens. Still, as there is absolutely no other form with which I feel
able to unite this one, it is better to give it a name than to leave it erroneously
with some other species.

The above description is copied from that of Dall, while the figure, Plate
XVII. Fig. 11, is copied in my Plate III. Fig. 2. I have seen no specimen of it.

Pupa fallax, Say.
armifera, Say.
contracta, Say.

Pupa Holzingeri, Srerxt.

Shell narrowly perforated, turrited-cylindrical, vitreous (or whitish), very
minutely striate, shining; apex rather pointed; whorls 5, regularly increasing,
well rounded, especially the upper ones, the last somewhat narrowed and a
little ascending towards the aperture, compressed at the base but not carinated,
at some distance from the outer margin provided with an oblique, rather prom-
inent, acute crest corresponding in direction to the lines of growth, extending
from the base to the suture, formed by a whitish callosity; behind the crest the
whorl is flattened, and corresponding to the lower palatal lamella, impressed;
aperture lateral, scarcely oblique, relatively small, inverted subovate, with a
slight sinus at the upper part of the outer wall, margins approximated ; peri-
stome moderately reflected ; lamellae 6; one parietal, rather long, very high, in
its middle part curved outward, towards the aperture bifurcated, the outer
branch reaching the parietal wall ; one columellar, longitudinal, rather high,
its upper end turning in nearly a right angle towards the aperture, but not
reaching the margin; basal exactly at the base, short, high, dentiform; 3 in
the outer wall, viz.: the lower palatal long, ending in the callus, highest at
about its middle ; the upper short, rather high on the callus ; above the upper,
one supra-palatal, quite small, dentiform, nearer the margin.

Length 1.7 mm., diam. 0.8 mm. = .068 X .032 inch.

As already stated, our species ranges beside P. armifera and P. contracta,
Say, standing nearer the latter. Yet it is different from this species by the
shape of the aperture, the wanting callus! connecting the margins on the

1 In many specimens of P. contracta so strongly developed that the peristome is
rendered continuous.
VOL. XIX. — NO. 4. 13

194 BULLETIN OF THE

body whorl, by the longer crest behind the aperture, which in contracta disap-
pears in about the middle of the (height of the) whorl, and by the wanting
constriction, especially in the columellar wall, not to speak of the size and
shape of the whole shell. The lamella also show some marked differences,
such as the presence of a high basal, the shorter columella not reaching the
base, but with relatively larger horizontal part, the bifurcation of the parietal
and the presence of a supra-palatal, the last just as it is in P. armifera.

It must be added here that the specimen first obtained from Minnesota in
several respects differs from those found in Illinois and
Iowa, which I consider as typical ; by its size which is one
third smaller, by the basal lamella developed in a peculiar
way, being rather longer at the truncated top than at its
foot, and by the stronger, thicker palatal lamelle. Yet, as
there was only one specimen, it was liable to be an individ-

ml Y ual peculiarity, — even then of interest. Should, however,
We more specimens be found with the same configuration, they

2a /
‘A \\

would represent a distinct and well characterized variety ;
Pupa Holzingeri, possibly it is a peculiar northern form.
enlarged. New Philadelphia, Ohio, June, 1889.

The above is a description by Dr. V. Sterki1 of a Pupa received by him
from Winona, Minn., and Northern Illinois. He kindly furnished me the
above figure.

Pupa rupicola, Say.
corticaria, Say.
Vertigo milium, Goutp.
ovata, Say.
Succinea retusa, Lea.
ovalis, Say.
avara, Say.
aurea, Lema.
obliqua, Say.

SOUTHERN REGION SPECIES.

Glandina Vanuxemensis, La. °
truncata, Say.
bullata, Gou.n.
decussata, PFEIFFER.
Texasiana, PFEIFFER.

Lingual membrane as usual in the genus. Teeth 35-1-35. Central small,
narrow, with a single blunt rounded cutting point. See Plate IX. Fig. G.

1 The Nautilus, Vol. III., No. 4, p. 37, August, 1889.

ar

MUSEUM OF COMPARATIVE ZOOLOGY. 195

Zonites caducus, PFEirrer.
cerinoideus, ANTHONY.

Gundlachi, Preirrer.
Found also in Texas, at Hidalgo, by Dr. Singley.

Zonites Singleyanus, Pirssry.

Shell minute, broadly umbilicate, planorboid, the spire scarcely perceptibly ex-
serted; subtranslucent, waxen white, shining, smooth, under a strong lens seen to
be slightly wrinkled by growth-lines; whorls three, rather rap-
idly increasing, separated by well impressed sutures, convex, the
apex rather large; body whorl depressed, slightly descending,
indented below around the umbilicus; aperture small, semilunar,
oblique; peristome simple, acute. Umbilicus nearly one third
the diameter of the shell, wide, showing all the whorls.

Alt. 1, diam. 2 mm.

New Braunfels, Comal Co., Texas.

Allied to Z. minusculus, but much more depressed, more shin-
ing, smoother, smaller, with broader umbilicus and a complete
whorl less than mznusculus.

This species, one of the most distinct of the smaller forms of
Hyalina, was communicated to me by Mr. J. A. Singley, in whose
honor it is named. I have also found a few specimens among the
shells collected by myself in Central Texas, during the winter of 1885-86. With
Z. Singleyanus at New Braunfels are found quantities of Z. minusculus. The latter
species exhibits some variation, being often more depressed than more northern
specimens. This depressed form has been noticed in Mexico by Strebel, who pro-
poses for Z. minusculus the new generic title of Chanomphalus, which is, of course,
completely synonymous with Pseudohyalina, Morse, 1864, and this, again, is not dif-
ferent enough from Hyalina to warrant the erection of a new genus or subgenus.
There is some variation in the width of the umbilicus in Texan specimens of Z. mi-
nusculus, but I have not seen specimens with it so wide as Dr. Dall indicates for
his var. Alachuana from Florida. JH. elegantulus, Pfr., is about the size and form of
my Zonites Singleyanus, but it is a strongly sculptured species.

The above description was published by Pilsbry, Proc. Phil. Acad., N. S.,
1889, p. 84, Plate XVII. Figs. 6, 7,8. A specimen kindly furnished me by
Dr. Singley for the purpose is drawn in my figure.

Zonites Singleyanus,
enlarged.

Zonites Dallianus, Simpson.

Shell minute, depressed, narrowly umbilicated, fragile, pale straw-
colored, somewhat shining; under a lens seen to be marked with
delicate growth-lines above, smoother beneath. Spire a little con-
vex; apex subacute; sutures scarcely impressed. Whorls three
and one half, scarcely convex, the last wide. Aperture oblong-
lunate, oblique, upper and lower margins sub-parallel, slightly con-

verging; peristome acute. Alt. 14, diam. maj. 3, min. 2} mm. Zonites
West Florida, at Shaw’s Point, Manatee Co., and Little Sarasota pecan
Bay.

196 BULLETIN OF THE

Differs from Z. arboreus, Say, in the smaller spire and wider last whorl; fewer
whorls ; differently shaped aperture. It is about half the size of Z. arboreus, and
‘the sculpture is the same as in that species. The Helix Ottonis of Pfeiffer, of
which specimens from Cuba and Hayti are before me, has no special relationship
to this species, but is undoubtedly a synonym of Z. arboreus, as Pfeiffer himself
concluded. HH. Ottonis differs from arboreus in nothing but the lighter color; the
form and dimensions are precisely as in arboreus. (See Pfr. in Wiegm. Archiv fiir
Naturgeschichte, 1840, p. 251; the species was never described in the Monographia
Heliceorum. ) ‘

The aperture in Z. Daillianus is less lunate than in Z. arboreus, embracing less of
the penultimate whorl; seen from beneath, the greater portion of the aperture lies
outside of the periphery of the penultimate whorl; whilst in Z. arboreus the reverse
is the case. The much smaller size of Dallianus also separates it from Z. arboreus.

This species was sent me under the above name by Mr. Charles T. Simpson, the
well known student of Floridian shells. The same form I find in the museum of
the Academy, collected by Mr. Henry Hemphill.

The above descriptién was published by Mr. Pilsbry in Proc. Phil. Acad.,
N.S., 1889, p. 83, Plate III. Figs. 9,10, 11. A specimen kindly furnished
me for the purpose by Mr. Pilsbry is also figured above.

Microphysa incrustata, Pory.
vortex, PFEIFFER.

All the specimens received from West Florida collected by Mr. Hemphill,
and from East Florida by Mr..G. W. Webster, are heavily incrusted with dirt.

Microphysa (?) dioscoricola, C. B. Apams.

Shell minute, subperforate, conic globose, thin, very delicately striate, horn-
colored ; spire elevated, obtuse ; whorls 3-34, convex, the last
medially subimpressed; aperture lunately rounded; peristome
simple, acute, the columellar margin subvertically descend-
ing, very slightly reflected, diam. greater 13, lesser 12, height
14 mm. (Pfr.). .

This species is placed by Von Martens (Die Heliceen, p. 73
in Conulus, a subgenus of Hyalina, with fulvus, Gundlachi, and

Microphysa others. Mr. Dall tells us (Nautilus, ITI. 25) that it belongs to

eae ak Microconus. This last is synonymous with Microphysa, a sub-

genus of Zonites, according to Tryon, Syst. Conch., IIT. 24.

Mr. Dall says also that the species was originally described from Jamaica by

Adams, and subsequently from Trinidad by Guppy as ceca. In its jaw and

lingual dentition it seems to agree with most of the other species of Microphysa
which I have examined. I retain it, therefore, in that genus.

The species seems widely distributed in Florida. St. Augustine; Blue
Spring, St. John’s River ; Lake Worth to Hawk’s Park along the east coast;
Hilo River, emptying into Mosquito Inlet, east coast, not Hillsborough River,
emptying into Tampa Bay, as stated by Dall. The specimens examined by me

MUSEUM OF COMPARATIVE ZOOLOGY. 197

were collected by G. W. Webster at Hawk’s Park, “widely distributed in dry
places, where other species are not found.” Also at Hidalgo, Texas (Singley).

The shell is figured on preceding page.

The jaw (Plate III. Fig. 6) is high, strongly arched, with acuminated ends ;
it is very thin, membranous, light horn-colored and transparent ; there are
numerous — some fifteen on each side the median line — narrow, delicate ribs,
running obliquely to this line, denticulating either margin; on the upper
median portion the ribs meet before reaching the lower margin, leaving upper,
median, triangular plates as in Orthalicus. The jaw is quite such as I have
described and figured for Macroceramus in Terr. Moll., V. 384. It also resem-
bles that of Microphysa turbiniformis (Ann. N, Y. Acad. Sci., III., Plate XV.
Fig. C), excepting that the latter wants the upper median triangular plates.
A greatly magnified view of the central portion of the jaw is given.

The lingual membrane is long and narrow. Owing to its small size, it was
very difficult to determine the shape of any but the lateral teeth. Three of
these last are figured on Plate II., Fig. 5, drawn by camera lucida. They have
wide, square bases of attachment, bearing, as usual, two cusps, both stout and
blunt, and bearing short, stout cutting points ; the centrals appear of the same

_ shape and tricuspid, but I failed to distinguish them clearly enough to draw

by camera; the laterals are separated, low, wide, quadrate, with long irregu-
larly serrated cusp. I failed also to distinguish these clearly enough to draw
by camera. I have represented them in the figure as they appeared to me.
The laterals seera like the teeth of Pupa, the marginals much like those of
Cronella subcylindrica. 'The dentition is somewhat similar to what I have fig-

‘ured of vortex on page 356 of the Manual of American Land Shells. There

are about 15-1-15 teeth, with six perfect laterals on each side the median line.

Mr. Dall says of this species that the shell is much smaller than that of
granum, olive-greenish, with a silky lustre and few inflated whorls, the first
of which is usually finely punctate. The suture is very deep, and the umbili-
cus is proportionally larger than in granum.

The figure of the dentition of an undetermined species found by Dr. W. M.
Gabb, in Costa Rica, published by me in the Annals of the New York Acad-
emy of Science, Vol. III. p. 261, Plate XI. Fig. G, is said by Mr. Pilsbry to
represent that of this species, — he having identified the shell from which the
lingual was extracted to be H. ceca, Guppy.

Hemitrochus varians, MENEE.
Strobila Hubbardi, Brown.

Polygyra auriculata, Say.

Dall (U. S. Nat. Mus. Proc., 1855, p. 263) thus characterizes a variety
microforis :—

This form is quite well marked, and when fully adult shows as a rule little vari-
ation from the form figured by the Binneys, and generally regarded as typical. <A
quite uniformly characterized variety was found, however, by me at Johnson’s

198 BULLETIN OF THE

Sink, Alachua County, Florida, where it was abundant. Some twenty specimens
were picked up in a few moments during a hurried visit made with other ends in
view, and a quart could easily have been gathered in half an hour. This form is
distinguished by its generally smaller size (max. diam. 12.0, min. diam. 10.0, alt.
6.0mm.) as compared with the type (15.0, 12.0, and 7.9 mm.), and by being more
closely rolled, thus having not only an actually smaller umbilicus, but one in
which one third less of the preceding whorl is visible. The specimens were uni-
form in this, and in all other respects were like the typical auriculata.

Polygira uvulifera, Suurrteworru.
auriformis, BLanp.
Postelliana, Bianp.
espiloca, Bianp.
avara, Say.
ventrosula, PrerrreEr.
Hindsi, Preirrer.
Texasiana, Moricanp.
triodontoides, Buanp.
Mooreana, W. G. B.
hippocrepis, PFEIFFER.

Through the kindness of Mr. Singerly, I have the opportunity of examining

the jaw and lingual membrane.
Jaw long, low, ends blunt; anterior surface with over 14 ribs denticulating

either margin.

Lingual membrane long and narrow (Plate III. Fig. 8, a, 6). Centrals tri-
cuspid, laterals bicuspid, marginals low, wide, irregularly denticulate. Teeth
30-1-30, the ninth lateral having its inner cutting point bifid.

Polygyra Jacksoni, Buanp.

A form was found abundantly near Fort Gibson, Indian Territory, Mr.
C. T. Simpson, who thus describes it in Proc. U. S. Nat. Mus., 1888, p. 449.
Instead of the bicrural tooth on the body whorl, at the aperture there is a heavy
elevated deltoid callus, which is joined to the upper and lower margins of the peri-
stome, and which occupies about the same area as the tooth in the type. The num-
ber of whorls is 5; greater diam. 7, lesser 6, height 3mm. In examining several
hundred specimens, I have found none which approach the type, and I would
therefore propose for it the varietal name of de/toidea.
Polygyra oppilata, Moricanp.
Dorfeuilleana, Lea.
Ariadne, PFrirFrer.
septemvolva, Say.
cereolus, Mun.re.pr.
Carpenteriana, Bianp.

MUSEUM OF COMPARATIVE ZOOLOGY. 199

Polygyra Febigeri, Bianp.
pustula, Férussac.
pustuloides, Buanp.

Triodopsis Hopetonensis, SHuTTLEWoRTH.
Levettei, Buanp.

See 2d Suppl. This species may perhaps be considered one of the Central
Province. A variety, however, approaches very nearly the Indian Territory
shell lately described as Mesodon Kiowaensis. This variety is toothless. It is.
smooth, like Levette, and has six whorls.

Triodopsis vultuosa, GouLp.
Copei, WETHERBY.

See 2d Suppl. To the synonymy add Triodopsis Cragini,
Call, Bull. Washburne Coll. Library, I., No. 7, p. 202, Fig. 5,
Dec., 1888, Topeka, Kansas. I have seen an authentic speci-
men, given by Mr. Call to the National Museum. It is figured
here.

Mesodon Romeri, Preirrer.

divestus, GouLp.

The typical form has few separated, very stout ribs ; a va- Triodopsis Cragini,
riety from Eufala, Indian Territory, sent me by Mr. C. T. a
Simpson, has numerous fine ribs and revolving microscopic lines. One in-
dividual is 24 mm. in greater diameter.

Mesodon jejunus, Say.
See Manual of American Land Shells, p. 390.

Mesodon Kiowaensis, Simpson.

Shell umbilicated, orbicularly depressed, solid, dark brown in color ; whorls
5, with rather coarse strie, and fine revolving impressed lines,
which are much more conspicuous on the last whorl. Suture
deeply impressed, leaving the whorls well rounded ; aperture
oblique, somewhat transversely rounded, forming fully three
fourths of a circle ; peristome thick and solid, whitish or pur-
plish, evenly reflected, with a slight constriction behind it ;
umbilicus moderate, deep, exhibiting but little more than one
of the whorls. Greater diam. 15, lesser 13, height 7 mm.

Kiowa Station, about thirty specimens, mostly dead. Lime-
stone Gap, two dead specimens. Another badly bleached shell
was obtained not far from Eufaula (Indian Territory).

Jaw with 9 ribs; teeth with fewer laterals than Sayii, and the inner cusp is
bifid on the marginals, while in Sayii it is entire (Simpson).

Mesodon
Kiowaensis.

g

200 BULLETIN OF THE A

The foregoing description is copied from the Proceedings of the U. S.
National Museum, 1888, p. 449, while the figure is drawn from a specimen
kindly furnished by Mr. Simpson.

The shell appears to me to be a toothless form of some T'riodopsis, rather than
a Mesodon (see above, under Triodopsis Levettet). It also resembles nearly
some of the toothless forms of Triodopsis Mullan.

Acanthinula granum, StTrReBet and PFEIFFER.

Shell small, umbilicated, thin, scarcely shining, light horn-colored, with
rib-like striz of growth, crossed obliquely with rib-like
folds, in fresh specimens hirsute or with punctate epi-
dermis. Whorls 44, four of them broad, rounded, regu-
larly increasing in size, rapidly in elevation, the last
descending, impressed at the umbilicus. Peristome
; simple, broadly reflected at its columellar margin, par-
gos a tially covering the deep umbilicus, within with whitish,
light thickening. Greater diam. 2.8, lesser 2.6, height

2.8 mm.}; of aperture, height 1.2, breadth 1 mm. (Strebel and Pfeiffer.)

Acanthinula granum, SvREBEL and PFEIFFER, Beitrag zur Kennt. der F. Mex. L.
und S. W. Conch., IV., 1880, p. 31, Plate IV. Fig. 15, not Plate [X., as quoted
in text.

A Mexican species, found also in Florida; Archer, Alachua Co.; Evans
Plantation, Rogers River; Lake Worth (Dall).

Mr. Dall says the shell, when perfect, is nearly the size of labyrinthica, very
thin, reddish brown, with very deep sutures and a rather deep, small tubular
umbilicus. It is covered with beautiful deep oblique epidermal ridges, which
are easily lost, and do not agree with the lines of growth.

The figure is drawn from a specimen kindly furnished by G. W. Webster.

Dorcasia Berlandieriana, Moricanp.
griseola, PFEIFFER.

Bulimulus patriarcha, W. G. B.
alternatus, Say.

I am assured by Dr. Singerly and Mr. Simpson that the form known as
alternatus does not always have a dark aperture, and the intermingling of the
forms leads an observer on the spot to believe alternatus, Schiedeanus, Moore-
anus, and dealbatus varieties of one and the same species. They were so treated
by my father in Vol. II. i

Bulimulus Schiedeanus, Preirrer,
var. Mooreanus, W. G. B.
dealbatus, Say.

*

MUSEUM OF COMPARATIVE ZOOLOGY. 201

Bulimulus serperastrus, Say.
multilineatus, Say.
Dormani, W. G. B.

Bulimulus Floridanus, Preirrer.

I have already in Terr. Moll., IV., Plate LX XIX. Fig. 3, figured the front
view of the typical specimen in Mr. Cumings’s collection, drawn by Mr. G. B.
Sowerby. The back view is now offered (Plate III. Fig. 7), received from the
same source.

A comparison of the front view of Mr. Sowerby’s drawing referred to above,
with the figure of Bulimulus Hemphilli (Plate III. Fig. 9), recently received
from Mr. George W. Webster, will lead one to believe the
two to be identical. Iso suggested in Manual of American
Land Shells (p. 408), when treating the variegated shell fig-
ured in Fig. 449 of that work, here repeated. There appear
to be two varieties of coloring, one corresponding to Pfeiffer’s
description, and one to Sowerby’s figure. te

I give the description of B. Hemphilli in full, though I be- _—_ Bulimulus
lieve it to be identical with Floridanus. rears:

Shell imperforate, very thin, transparent, amber-colored and marked by coarse
lines of growth; body whorl with six revolving and slightly interrupted brownish
red bands, the lower two being close together and upon the rounded base, spire
obtuse, whorls five, slightly convex, the body whorl constituting two thirds of the
entire length of the shell. Suture slight, base uniformly and gracefully rounded.
Aperture direct and oval, peristome thin. Length, 19 mm.; diameter, 8 mm.
Hab. both coasts of South Florida.

Remarks. Mr. Henry Hemphill, of San Diego, Cal., first found a few dead
and badly preserved specimens of this shell in 1884, at Marco, west coast of
Florida. These Mr. Binney thought identical with b. Floridanus, Pf. (See Manual
of American Land Shells, 1885.) Numerous specimens collected during the past
summer by the author and Mr. G. W. Webster and son, prove beyond a doubt
that this is not identical with the shell figured and described on page 407 of Mr.
Binney’s Manual. The 5. Hempii/li is more ventricose, not angular at base, im-
perforate, differs in color, and in fact there is a general difference.

Mr. Berlin H. Wright describes the above species in the West American
Scientist, San Diego, April, 1889, p. 8. He found also a variety of uniform
light brown or russet color, bandless, which I have figured on Plate III. Fig. 9.
This form had a jaw and lingual membrane the same as in B. Marielinus

and Dormant

Bulimulus Marielinus, Pory.
Cylindrella Poeyana, D’Orzreny.
jejuna, Say.

202 BULLETIN OF THE

Macroceramus pontificus, Gout.

I give here, for comparison, a figure of the true M. Kienert, from a type in
Dr. Pfeiffer’s collection, from Honduras.

Macroceramus Kieneri.

Macroceramus Gossei,

Macroceramus Gossei, Preirrer.
The figure given represents the species.

Pupa variolosa, Goutp.
modica, Gou.p.
pellucida, Prreirrer.

Strophia incana, Bryyey.

Holospira Romeri, Preirrer.

Pfeiffer says “allied to Goldfussz, but from all species easily recognized by

the basal carina of the last whorl, and its singular twist, which at first sight
gives a sinistral appearance to the shell.”

Holospira Goldfussi, Menxe.
Stenogyra octonoides, D’Ors.
subula, PFreirrer.
gracillima, Pre1rrer.
Ceecilianella acicula, Mu.izr.
Liguus fasciatus, MU..ieEr.

See p. 435 of Manual of American Land Shells for still another variety of
coloring of this species.

Orthalicus undatus, BrucuikrRe.

Succinea Concordialis, Gou.p.
luteola, Gouxp.
effusa, SHUTTLEWORTH.
Salleana, PFEIFFER.
campestris, Say.

Veronicella Floridana, Binney.

MUSEUM OF COMPARATIVE ZOOLOGY. 203

Onchidium Floridanum, Dat.

“ See Plate VI. Figs. B, C, for a drawing of an original specimen, enlarged
three times.

To Mr. Hemphill is due the credit of adding this genus to the fauna of Eastern
North America. ‘The specimens arrived as this paper is going through the press,
and a detailed description must be deferred. The following notes, however, will
indicate its external characters : —

When living, the creature is of a uniform slaty blue, the under parts bluish
white, with a greenish tinge to the veil. The surface appears beautifully smooth
and velvety without dorsal tubercles ; just within the slaty margin of the mantle
is a single row of about (in all) one hundred whitish elongated tubercles. When
crawling, it is of an oval shape, about an inch long, and two tentacles extend for-
ward beyond the mantle margin, resembling the oculiferous ones of Vaginulus
Floridanus. In spirits the surface is still smooth, but numerous circular hardly
elevated domelets cover the back, each appearing to contain one of the dorsal eyes
described by Semper. The tentacles are entirely retracted ; a narrow veil, with
lightly escalloped edge, precedes the head ; the muzzle is not prominent, is in-
dented in the middle, and puckered at the edges. The foot is about one third
wider than the mantle at each side of it. There is nojaw. The penis resembles
that of Siphonaria in form and position. The animal exudes very little mucus. It
was found on rocks between tides associated with Chiton piceus. Fifteen speci-
mens were found at Knight’s Key by Hemphill.

Onchidium indolens of Couthouy (Rio) and O. armadillo of Morch differ from the
above in coloring. The latter, described from-St. Thomas, has a very different
dorsal surface. No others are known from East America. It would seem as if
the small Northern species, possessing a jaw like O. boreale, Dall, and O. Celticum,
Cuvier, might appropriately be separated from the agnathous tropical forms as a
subgenus, for which the name of Onchidella might be revived in a restricted sense.

The above description is by Dall (Proc. U.S. Nat. Mus., 1885, p. 288).
Specimens received by him have the lingual dentition of the genus. (See my
Plate III. Fig. 10, where a central tooth and adjacent lateral are given.) There
are numerous rows of over 97-1-97 teeth.

The following are to be added to the species treated in the Second Sup-
plement.

PACIFIC PROVINCE SPECIES.

Microphysa Stearnsi, Buianp.
Lansingi, Buanp.

It must be borne in mind that the other species of Microphysa examined by
me have quadrate marginal teeth, while Stearnst and Lansingi have the acule-
ate marginal teeth of the Vitrinine. Thus they can hardly be classed in
Microphysa. The name Pristina has been suggested by Ancey (Conchologists’

204 BULLETIN OF THE

Exchange, I. 5, p. 20, Nov., 1886). As a substitute for this preoccupied-name,
Mr. Pilsbry suggests Anceyia. (See same, I. 6, p. 26, Dec., 1886.) Mr.
Ancey’s description is: —

Pristina, Anc. (nov. subg. Hyaline). Testa parvula, imperforata, cornea, nitens,
multispirata; spira depresse conica. Apertura interdum lamellis radiantibus sub-
serratis in palato sitis insignis.

Geographical Distribution: Western and Arctic North America.

Types: Hyalina Stearnsi, Bland, and Lansingi, Bland.

Mr. W. G. Binney put these species, but with doubt, in Microphysa, while other
authors consider them as Hyaline; they differ from the latter by anatomic fea-
tures, and from the former by the form of the shell. Altogether I am inclined to
place the group in Hyalina, as a series nearly allied to Conulopolita, Boettger
(type, C. Radder, Boettg.); I am confident the presence or absence of internal
laminez or tooth-like processes within the aperture of Helices are not generic char-
acters ; in some instances they are either present or absent in closely allied species.
I established this fact when at work (Le Naturaliste, 1882) on the New Caledonian
forms, and I now repeat this as my opinion in regard to Pristina and Gastrodonta.
In the latter the teeth are frequently absorbed by the animal when growing larger.

Macrocyclis Duranti, Newc.

To the synonymy add : —

Selenites calatura, Mazycx, Proc. U.S. Nat. Mus., 1886, p. 460, with figures of that
form and of typical Duranti. Also, Proc. Elliott Soc., Feb., 1886, p. 114, same
figures.

Mr. Mazyck’s description and figures are repeated here : —

Shell’ small, depressed, brownish horn-color, with very coarse, rough, crowded,
sub-equidistant, irregular ribs, which are obsolete at the apex;
whorls four, rounded, somewhat inflated below, gradually in-
creasing, the last not descending at the aperture; suture im-
pressed ; umbilicus wide, clearly exhibiting all of the volutions ;
aperture almost circular, slightly oblique; peristome simple, its
ends approaching and joined by a very thin, transparent, whitish
callus, through which the ribs are distinctly seen. Greater
diameter, 4 mm.; height, 1? mm.

Santa Barbara, California, Dr. L. G. Yates. Hayward’s, Ala-
meda County, California, W. H. Dall, U. S. National Museum.
Macrocyclis Duranti, Newcomb’s description of this little shell (17. Duranti) is as

var. celata, follows: —
enlarged. > : 5
Shell depressed, discoidal, pale corneous, under the lens
minutely striated, opaque, broadly and perspectively umbilicated ; whorls 4, the last
shelving but not descending (at the aperture); suture linear; aperture rounded,
lunate, lip simple, the external and internal approaching.

“ Habitat. — Santa Barbara Island.”

Mr. Binney’s description, which is repeated in each of his works above named,
differs in this important particular. For Newcomb’s “ Under the lens minutely
striated,” he substitutes the contradictory words “ with very coarse, rough striz.”

MUSEUM OF COMPARATIVE ZOOLOGY. 205

In a note written in answer to an inquiry addressed to him regarding this singular
discrepancy, he says, “ My description and figure are from an individual, not from
the species. I am absolutely sure my specimen was one of the original find.”
His figure, drawn by Morse, rather represents a comparatively smooth, semi-
transparent shell.

Limax hyperboreus.

See Manual of Amer. Land Shells, p. 473. I have figured on Plate VIII.
Fig. F, an individual from British Columbia. Here I give the dentition.
Jaw arched, smooth, with blunt median projection. Lingual

membrane with 42-1-42 teeth ; centrals tricuspid ; laterals bi- ‘om
cuspid, 12 in number on each side; marginals about 30 on each |i \
side, aculeate, simple, without bifurcation or side spur. A 1 |

The figure shows a central tooth with its adjacent lateral, and
three extreme marginals.

Limaz montanus, L. castaneus, L. occidentalis, and L. campes- ae
tris all have side spurs to their marginal teeth. Otherwise, their Q\
dentition shows no specific distinction from that of hyperboreus.

Until the genitalia of the last is shown to vary, I am inclined Limax hyper-
to believe all four to be one and the same species. boreus.

Limax Hemphilli.

Mr. Henry Hemphill has sent me in spirits from Julian City, California, a
small, slender, smooth, dark species of Limaz, 20 mm. long in its contracted
state. It does not outwardly resemble Limax agrestis, nor does it seem prob-
able that that species would have been accidentally introduced from the Eastern
cities! The dentition, however, agrees with that of agrestis by its having the
peculiar side spur to the larger cutting point of all the lateral teeth. I venture
to propose a specific name for it, in hopes of having an opportunity later to
fix its specific position by an examination of the genitalia. It is figured on
Plate VIII. Fig. E.

The jaw is as usual in the genus.

There are 50-1-50 teeth to the lingual membrane, of which ten on each
side are laterals. Centrals tricuspid; laterals bicuspid, the larger cutting point
having a well developed side cutting point on its inner side; the laterals have
also an inner, slightly developed, horizontal side cusp, bearing a small, stout
cutting point (see Plate I. Fig. 13); marginals simple, without side spur.

The figure on Plate II. Fig. 3, shows one central with its adjacent laterals,
an outer lateral, and several extreme marginals.

A specimen, apparently of the same species, from British Columbia, has
53-1-53 teeth, of which 13 on each side are laterals.

I have the same species, with similar dentition, from San Tomas, Lower

California (Hemphill).

1 Tt is, however, found in San Francisco.

206 BULLETIN OF THE

Limax Hewstoni, J. G. Cooper.

On Plate II. Fig. 4, will be found a better figure of the dentition of this
species than is given in Terr, Moll., V. It will be seen that the inner side
cusp of the lateral teeth is quite distinct from the side spur found in Limaz
Hemphilli and agrestis. (See line third of p. 223.)

I have figured (Plate VIII. Figs. D and I) individuals received from Dr.
Cooper, drawn by Mr. Theo. D. A. Cockerell.

Limax campestris, var. occidentalis.

The specimen figured on Plate VIII. Fig. H, was kindly furnished by Dr.
Cooper. I have already expressed my belief in the identity of this with the
Eastern form.

Arion foliolatus, Gouxp.

It is with the greatest pleasure that I announce the rediscovery by Mr.
Henry Hemphill of this species, which has hitherto escaped all search by
recent collectors. It has till now been known to us only by the description
and figure of the specimen collected by the Wilkes Exploring Expedition,
almost fifty years ago, and given in Vols. II. and III. of Terrestrial Mollusks.
A single individual was found in December, 1889, at Olympia, Washington,
and sent to me living by Mr. Hemphill. It can thus be described. (See
Fig. A of Plate VITI.)

Animal in motion fully extended over 100 millimeters. Color a reddish
fawn, darkest on the upper surface of the body, mantle, top of head, and eye-
peduncles, gradually shaded off to a dirty white on the edge of the animal,
side of foot, back of neck, and lower edge of mantle, and with a similar light
line down the centre of back ; foot dirty white, without any distinct locomo-
tive disk ; edge of foot with numerous perpendicular fuscous lines, alternating
broad and narrow ; mantle minutely tuberculated, showing the form of the
internal aggregated particles of lime, the substitute of a shell plate, reddish
fawn color with a central longitudinal interrupted darker band and a circular
marginal similar band, broken in front, where it is replaced by small, irregu-
larly disposed dots of same color; these dots occur also in the submarginal
band of light color. Body reticulated with darker colored lines, running
almost longitudinally, scarcely obliquely, toward the end of the tail, and con-
nected by obliquely transverse lines of similar color, the areas included in
the meshes of this network covered with crowded tubercles, as in Prophysaon
Andersoni, shown in Plate IX. Figs. I,J. Tail cut off by the animal. (See
page 207.)

What appears to be the same species, or a very nearly allied one, was found
by Mr. Hemphill at Gray’s Harbor, Washington, on the banks of the Chehalis
River, near its mouth. This form is figured on Plate VIII. Fig. C. When
extended fully, it is 70 millimeters long. It is more slender and more pointed

MUSEUM OF COMPARATIVE ZOOLOGY. 207

at the tail than the large form. The body is a bright yellow, with bluish
black reticulations. The edge of the foot and the foot itself are almost black;
shield irregularly mottled with fuscous ; the body also is irregularly mottled
with fuscous, and has one broad fuscous band down the centre of the back,
spreading as it joins the mantle, with a narrower band on each side of the
body. The other characters, external and internal, are given below. This
smaller form loses its colors on being placed in spirits, becoming a uniform
dull slate color.

The large Olympia form is surely Arion foliolatus, Gould, agreeing perfectly
with his description in Vol. II., and with his figure in Vol. III., excepting
that the latter is colored with a deeper red.

Mr. Hemphill writes of it: “I have to record a peculiar habit that is quite
remarkable for this class of animals. When I found the specimen, I noticed
a constriction about one third of the distance between the end of the tail and
the mantle. I placed the specimen in a box with wet moss and leaves, where
it remained for twenty-four hours. When I opened the box to examine the
specimen, I found I had two specimens instead of one. Upon examination of
both I found my large slug had cut off his own tail at the place where I no-
ticed the constriction, and I was further surprised to find the severed tail piece
possessed as much vitality as the other part of the animal. The ends of both
parts at the point of separation were drawn in as if they were undergoing a
healing process. On account of the vitality of the tail piece, I felt greatly
interested to know if a head would be produced from it, and that thus it would
become a separate and distinct individual.” The animal on reaching me still
plainly showed the point of separation from its tail. (See Plate VIII. Fig. A.)
The tail piece was in an advanced stage of decomposition. I noticed the con-
striction towards the tail in one of five individuals of Prophysaon coruleum
from Olympia. (See page 209.) Another individual of the same lot had a
truncated tail, having undergone the operation. The edges of the cut were
drawn in like the fingers of a glove.

The tail of the Arion folvolatus having been cut off, I was unable to verify
the presence of a caudal pore from this individual. On the only living one
of the lot from Gray’s Harbor, the pore was distinctly visible, and is figured
on Plate VIII. Fig. C. Usually, it seemed more ‘a conspicuous pit” than a
longitudinal slit, as in Zonites. At one time I distinctly saw a bubble of
mucus exuding from it. It opened and shut, and is still plainly visible on the
same individual, which I have preserved in alcohol and added to the Binney
Collection of American Land Shells in the National Museum at Washington.
Another individual from Seattle plainly shows the pore.

Five specimens of the Gray’s Harbor lot had, concealed in the mantle, a
group of particles of white limy matter which it was impossible to remove as
one shell plate. In the large Olympia individual these irregularly disposed
particles of lime, of unequal size, seemed attached to a transparent membranous
plate. With care, I removed this entire, and figure it. It is suboctagonal in
shape (Plate VIII. Fig. B). Under the microscope it appears that the par-

208 BULLETIN OF THE

ticles of lime do not cover the whole plate ; at many points they are widely
_ separated. This aggregation of separate particles is the distinctive character
of the subgenus Prolepis, to which A. foliolatus belongs.4

The genitalia of the large individual from Olympia is figured on Plate IX.
Fig. D. The ovary is tongue-shaped, white, very long and narrow; the oviduct
is greatly convoluted ; the testicle is black in several groups of coeca; the
vagina is very broad, square at the top with the terminus of the oviduct, and
the duct of the genital bladder entering it side by side ; the genital bladder is
small, oval, on a short narrow duct; the penis sac is of a shining white color,
apparently without retractor muscle; it is short, very stout, blunt at the upper
end where the extremely long vas deferens enters, and gradually narrowing to
the lower end. There are no accessory organs. The external orifice of the
generative organs is behind the right tentacle.

The form from Gray’s Harbor (Plate IX. Fig. H) has its generative system
very much the same as described above. The ovary is much shorter and
tipped with brown, and is less tongue-shaped. The penis sac tapers to its
upper end. The vagina is not squarely truncated above. The system much
more nearly resembles that of Prophysaon Andersoni (see Terr. Moll, V.) than
that of the Olympia folvolatus.

, The jaw of both forms is very low, wide, slightly arcuate, with ends attenu-
ated and both surfaces closely covered with stout, broad separated ribs, whose
ends squarely denticulate either margin. There are about 16 of these ribs in
one specimen from Gray’s Harbor, and over 20 in that of the true foliolatus from
Olympia (see Plate IX. Fig. B). The lingual membrane in each form is long
and narrow, composed of numerous longitudinal rows of about 50-1-50 teeth,
of which about 16 on each side in the true foliolatus (Plate [X. Fig. C), and
19 in the other form, may be called laterals. Centrals tricuspid, laterals
bicuspid, marginals with one long inner stout cutting point, and one outer
short side cutting point. The figure shows a central tooth with its adjacent
first lateral, and four extreme marginals.

I have figured both the true foliolatus from Olympia (Plate VIII. Fig. A)
and the smaller form from Gray’s Harbor (Plate VIII. Fig. C) of natural size.
Should the latter prove a distinct species or variety, I would suggest for it the
name of Hemphilli, in honor of the discoverer of it and the long lost foliolatus.

Prophysaon Hemphilli.
See Plate VII. Fig. D, drawn by Cockerell from the living animal.

Prophysaon Andersoni, J. G. Cooper.

Figure 1 of Plate III. was drawn from a specimen received from Dr. Cooper.
It represents the true Andersoni, distinguished by a light dorsal band, and by
genitalia such as I have described for P. Hemphilli. The same form, also re-

1 Mr. Theo. D. A. Cockerell, finding the slug not to be a true Arion, is about to
suggest for it the generic name of Phenacarion.

MUSEUM OF COMPARATIVE ZOOLOGY. 209

ceived from Dr. Cooper, is drawn by Mr. Cockerell on Plate VII. Fig.C. Mr.
Cockerell has shown me that I have confounded with it another species, which
he proposes to call P. fascoatum. See next species.

Prophysaon fasciatum, CocKkERELL.

This species is described by Mr. Cockerell as distinct from Andersoni, with
which I have formerly confounded it. (2d Suppl. to Vol. V., p. 42.) It has
a dark band on each side of the body, running from the mouth to the foot.
To this must be referred the descriptions of animal, dentition, jaw, and geni-
talia formerly published by me as of Andersonz.

I am indebted to Mr. Theo. D. A. Cockerell for a figure and description of
this species. The former is given on Plate VII. Fig. A, while the latter is
given here in the words of Mr. Cockerell, whose name must consequently be
associated with it as authority : —

Length (in alcohol), 19mm. Mantle black, with indistinct pale subdorsal bands, —
an effect due to the excessive development of the three dark bands of the mantle.
Body with a blackish dorsal band, commencing broadly behind the mantle and
tapering to tail, and blackish subdorsal bands. No pale dorsal line. Reticulations
on body squarer, smaller, more regular, and more subdivided than in P. Andersont,
Cooper. Penis sac tapering, slender. Testicle large. Jaw ribbed.

Prophysaon coeruleum, CocKERELL.
Plate VIII. Fig. I, J.

In the Nautilus, 1890, p. 112, it is thus described : —

Length (in alcohol), 224 mm.; in motion, 43 mm. Body and mantle clear blue-
gray, paler at sides, sole white. Mantle finely granulated, broad, without mark-
ings. Lengthof mantle, 7 mm.; breadth,5 mm. Respiratory orifice, 24 mm. from
anterior border. Body subcylindrical, tapering, pointed. (In one specimen eaten
off at the end.) Distance from posterior end of mantle to end of body, 102 mm.

The reticulations take the form of longitudinal equidistant lines, occasionally
joined by transverse lines, or coalescing. Sole not differentiated into tracts. Jaw
pale, strongly ribbed. Liver white.

Mr. Binney sends me colored drawings of the living animal; the neck is long
and white, or very pale. Mr. Binney has examined the jaw and lingual, and finds
them as usual in the genus.

Several specimens were sent from Olympia, Washington, by Mr. Hemphill to
Mr. Binney.

P. ceruleum is an exceedingly distinct species, distinguished at once by its color
and the character of its reticulations.

Prophysaon cceruleum, var. dubium, n. var., COoCKERELL.

Length (in alcohol), 8mm. Length of mantle,4 mm. Distance from posterior
end of mantle to end of body, 3 mm. Mantle broad, with four bands composed
of coalesced black marbling, very irregular in shape, and running together anteri-
orly. Body dark, tapering. Sole pale, its edges gray. Liver white.

VOL. XIX. — NO. 4. 14

210 BULLETIN OF THE

With the P. cwruleum is a small dark slug, probably a variety of it, but differing
as described above. It will easily be distinguished by its blackish color and the
peculiar markings on the mantle.

Prophysaon Pacificum, CockERELu.
Plate VII. Figs. B, E, F, H.

Mr. Theo. D. A. Cockerell gives the following in the Nautilus of February,
1890, pp. 111-113 :—

Length (in alcohol), 173 mm. Body and mantle ochrey brown, head and neck
gray. Mantle granulated, rather broad, with a black band on each side not reach-
ing the anterior border ; these bands are farthest (24mm.) apart near the respira-
tory orifice, from which point they converge posteriorly, and anteriorly by the
bending of the band on the right side. Length of mantle, 73 mm. ; breadth, 4 mm.
Respiratory orifice 84 mm. from anterior border. Body cylindrical, rounded and
very blunt at end, not conspicuously tapering. Distance from posterior end of
mantle to end of body, 8mm. Body dark grayish-ochre above, with an indistinct
pale dorsal line; sides paler. Reticulation distinct, with indistinct “ foliations.”
Sole somewhat transversely wrinkled, but not differentiated into tracts.

Jaw dark, strongly curved, blunt at ends, with about ten well-marked ribs (Plate
VII. Fig. F).. Lingual membrane with about 35-1-35 teeth; centrals tricuspid, the
side cusps very small, laterals bicuspid, marginals with a large sharp straight inner
point and a small outer one. Compared with P. humile the centrals are slightly
shorter and broader. Liver dark gray-brown.

Found by Mr. H. F. Wickham under logs in ditches by the roadside and damp
places at Victoria, Vancouver Island, 1889.

This is a very distinct species, easily recognized by its color, the absence of dark
bands on the body, the pale dorsal line, and the blunt posterior extremity.

Prophysaon flavum, CocKERELL.
Plate VII. Fig. K.
From the Nautilus, 1890, p. 111:—

- Length (in alcohol), 25mm. Body and mantle dull ochreous, head and neck
ochreous. Mantle tuberculate-granulose, grayish ochre, pale at edges, and with
black marbling or spots in place of the bands of P. Pacificum. Length of mantle,
11 mm.; breadth, 54 mm. Respiratory orifice 5 mm. from anterior border. Body
cylindrical, hardly tapering, and blunt at end. Distance from posterior end of
mantle to end of body, 14 mm. Body dark grayish-ochre above, with a pale
ochreous dorsal line not reaching much more than half its length ; sides paler.
Reticulations distinct, “foliated.” Sole with well marked transverse lines or
grooves, those of either side meeting in a longitudinal median groove, which
divides the foot into two portions. Liver pale grayish.

Uniform tawny, as is Limax flavus. It stretches itself out in a worm-like shape
unlike other species. Internal shell plate, jaw, and tongue as in Andersont.

Gray’s Harbor, Washington. (Hemphill, 1889.)

This is probably a variety of P. Paczficum.

MUSEUM OF COMPARATIVE ZOOLOGY. yal

Prophysaon humile, Cockers tt.
Plate VII. Figs. F, G, L, M.
From Nautilus, 1890, p. 112.

Length (in alcohol), 163 mm. Body above and mantle smoke-color, obscured
by bands. Mantle wrinkled, and having a broad dorsal and two lateral blackish
bands, reducing the ground-color to two obscure pale subdorsal bands. Length of
mantle, 7 mm.; breadth, 54 mm. _ Respiratory orifice 23 mm. from anterior border.
Body subcylindrical, somewhat tapering, rather blunt at end. Distance from pos-
terior end of mantle to end of body, 8mm. Back with a blackish band reaching a
little more than half its, length, and lateral darker blackish bands reaching its
whole length. Reticulations distinct, ‘‘ foliated.” Sole strongly transversely striate-
grooved, but not differentiated into tracts.

Jaw pale, strongly striate, moderately curved, not ribbed. (See Fig. F.) Lingual
membrane long and narrow. Teeth about 80-1-35. Centrals tricuspid, laterals
bicuspid, marginals with a large inner point, and one (sometimes two) small outer
points. Liver pale chocolate.

Found by Mr. H. F. Wickham under the bark of rotten logs in the woods around
Lake Cceur d’Alene, Idaho, 1889.

In its reticulations, and general external characters, this species resembles
P. Andersoni, of which it is possibly a variety.

Hemphillia glandulosa.

(See also p. 216.)

From Olympia and Gray’s Harbor, Washington, Mr. Hemphill sent me liv-
ing specimens of this species, both young and mature. Several of the young
had the horn-shaped process to the tail noticed in the original description of
the genus. The shell in these young individuals is very slightly attached, ap-
parently simply by having its posterior margin lightly covered by the mantle.
It often becomes detached. In these young, the mantle is proportionally
smaller, and the neck much longer. I have figured an enlarged view of a
young individual, Plate IV. Fig. D.

Ariolimax! Columbianus, Goutp.

Found also by Mr. Hemphill on Santa Cruz Island.

Plate VI. Fig. A, represents the mottled variety, found recently b7 Mr.
Hemphill in the State of Washington. Mr. Cockerell suggests for it the vari-
etal name maculatus. This form shares with the type the peculiar penis sac
(Fig. G) distinguishing it from the next species.

Ariolimax Californicus, Coorrr.

See Plate V. Fig. E, for the animal in motion, and a portion of the genital
system (Fig. H), showing variation from that of A. Columbianus.

1 The name should be Arionilimarx.

Al be BULLETIN OF THE

Ariolimax Andersoni.
See Plate V. Fig. F, showing the typical specimen in spirits restored.

Ariolimax Hemphilli.
Plate V. Fig. B, G.

A variety maculatus, Cockerell, is figured in B. The Figure G is drawn
from a typical specimen, with the tail, the pore, and the locomotive disk.

Ariolimax niger, J. G. Coorrr.

Plate V. Fig, A, gives a lighter-colored form ; Fig. I, the typical form; Figs.
C and D, the caudal pore.

Triodopsis inflecta, Say.

This has erroneously been quoted from the Pacific Province, at the mouth
of Columbia River. It is difficult to decide what species Middendorff had in
view. His words are thus translated :—

Let it not be objected that Helix clausa up to this time has not. been discovered
west of the Rocky Mountains. The Northwest Coast of America is almost wholly
unexplored conchologically, and I do not doubt that H. clausa will be there
found, just as I can now assert with reference to //. planorboides. Even the
American authors know this hitherto only from the Ohio and Missouri. Its dis-
tribution nevertheless appears to extend over the whole of North America, since I
have received a great number of specimens of the same through Mr. , from
Sitka, whereby it becomes incorporated with our Russian Fauna. Southwards it
extends to the west coast of America, at least to Upper California, where they
were likewise collected by Mr. It appears to have undergone no altera-
tion whatsoever, and presents in Sitka a considerable size, as the ordinary repre-
sentations show (up to 22, etc.). Moreover, Binney in the Boston Journal, IIL,
Plate XIV., has them copied equally large.

Polygyra Roperi, Pitssry.

Shell umbilicated, plane above, slightly inflated below, shining, pellucid,
light horn-color, with delicate wrinkles of growth ; spire flattened ; whorls 54,
scarcely rounded, very regularly increasing, the last flattened’
above, abruptly deflected at the aperture, deeply constricted
behind the peristome ; aperture transversely lunar, gaping,
much contracted, tridentate ; peristome thickened, broad,
white, gradually thinning and scarcely reflected at its edge,
and not extending beyond the surface of the whorl, its ends
approached, joined by a light callus, on which is a heavy

Polygyra Roperi, : ; %
enlarged. white callus bearing a stout, white, broad, blunt, transverse

tooth, slightly curving inward, its basal margin with an erect
conical, short tooth, separated by a small circular sinus from another rather
more deeply seated similar tooth on its upper margin. Umbilicus broad,

MUSEUM OF COMPARATIVE ZOOLOGY. 215

showing the volutions clearly. Greater diameter, 9 mm. ; lesser, 7 mm. ;
height, 24 mm.
Helix (Triodopsis) Roperi, Pitspry. The Nautilus, Vol. III. No. 2, June, 1889,
p. 14.

Redding, Shasta Co., California, in drift of the Sacramento River, three
dead shells were collected by Mr. Edward W. Roper, of Chelsea, Mass.

The above description is drawn from one of the original specimens, kindly
lent me by Mr. Roper, while another in the collection of the Academy of Natu-
ral Sciences of Philadelphia, from which Mr. Pilsbry drew his description, is
figured above. The third specimen was given by Mr. Roper to Mr. Henry E.
Dore of Portland, Oregon.

Never having seen a specimen of P. Harfordiana, I cannot say if this species
is identical with it. At least, it must be nearly allied.

Aglaja fidelis, Gray.

New figures of several forms of this species are given. Plate X. Fig. A
represents the black elevated form approaching infumata. Its sculpturing is
given in Fig. B. The small, black, elevated form is given in Fig. C, with its
sculpturing in D ; the small, depressed form, in E.

Aglaja infumata, Goutp.

Plate X. Fig. F, gives an enlarged view of the hirsute surface.

Arionta arrosa, GouLp.
Plate XI. Fig. A gives this species. A form of arrosa nearly approaching
A. exarata is given in Fig. B, its sculpturing in Fig, C.
Arionta exarata, PFEIFFER.

The typical form and its sculpturing are given on Plate XI. Figs. D and E.

Arionta Mormonum, PFEIFFeEr.

The typical form is given on Plate XI. Fig. F. The fF
variety (Vol. V. p. 141) approaching Aglaja Hillebrandi,
is given in Figs. G and H ; the sculpturing of the same
form, on Plate X. Fig. G. The genitalia of this form
are the same as of the type. if

Arionta sequoicola, J. G. Coorrr.

A figure of the sculpturing of this species is here pijargea seuipuuctid of
given, greatly enlarged. Arionta sequoicola.

214 BULLETIN OF THE

Arionta Californiensis, Lea.

I give here new figures of two forms of this species, Arionta Diabloensis and
the depressed variety of A. Bridgesi, the former drawn from a shell received
from Dr. Cooper.

so ~
Arionta Diabloensis. Arionta Bridgesi, depressed.

Onchidella Carpenteri, Datu.
An alcoholic specimen received from Mr. Dall is figured on Plate VI.
Figs. D, E, enlarged twice.
Veronicella olivacea, Stearns.

I have failed to receive Californian specimens. That figured on Plate IX.
Figs. E, F, is one of the original lot from Folvon, Central America.

CENTRAL PROVINCE SPECIES.

Limax montanus, INGERSOLL.

A specimen is figured on Plate VIII. Fig. G.
The species is surely identical with L. campestris.

Patula solitaria, Say.

Mr. Hemphill found this species very abundant at Old Mission, Coeur
d’Alene, Idaho. There was an albino variety, a depressed form, and one very
much more elevated than that which I figured in the Second Supplement,
Plate I. Fig. 10.

MUSEUM OF COMPARATIVE ZOOLOGY. 215

Patula strigosa.

Among the shells recently collected by Mr. Hemp-
hill at Old Mission, Coeur d’Alene, Idaho, was a
marked variety of this species, for which Mr. Hemp-
hill suggests the name subcarinata. The specimens
vary greatly in elevation of the spire, and in the num-
ber and disposition of the revolving bands, often ae ay,
quite wanting. All have a very heavy shell, the body ~~ . Boetean rth
whorl of which has an obsolete carina which is well
marked at the aperture, modifying the peristome very decidedly. See the
figure.

In examining the genitalia I find the base of the duct of the genital bladder
greatly swollen along a fifth of the total length of the duct.

On the banks of the Salmon River, Idaho, Mr. Hemphill found a form like

var. Gouldi, but distinctly carinated. None of the Utah
ea individuals of this form are so characterized.

AKIN . Another form of strigosa from the same locality is very
Xs ae) large, flat, with a transversely oval aperture, the ends of the

. peristome so nearly approached as almost to touch, and
Seals tthidioe. vat je often joined by a heavy callus, which forms a continuous
galis, Hemphill. rim around the aperture. Mr. Hemphill has called this

SRT

var. jugalis.
Microphysa pygmea. CW
Found by Mr. Hemphill at Old Mission, Coeur d’Alene, Na)
Idaho. abi

Microphysa Ingersolli, BLanp.
A better figure of this species is here given.
Triodopsis Hemphilli. \ ie
&

Mr. Tryon has suggested the name binom- Aidtopliyaa Inger-
inata for this species, though Hemphillc is _ solli, enlarged.
not preoccupied in Triodopsis.

Triodopsis Sanburni.

: 4 : The cut is drawn from one of the original specimens.
Triodopsis Sanburni, :

enlarged.
Mesodon ptychophorus.
At Old Mission, Coeur d’Alene, Idaho, Mr. Hemphill found

a form of this species characterized by a heavy, dead white
shell with scarcely any trace of ribs or wrinkles of growth
which are usually so characteristic of the species. On the ,,

f esodon ptycho-
banks of the Salmon River he found a small form, the lesser _ phorus, var.
diameter of which is only 12 mm. See figure.

216 BULLETIN OF THE

Triodopsis Harfordiana.

Ancey suggests commutanda, and Tryon Salmonensis, as a substitute for the
name Harfordiana. I retain the last name, it not being preoccupied in the
genus Triodopsis.

Prophysaon Andersoni ?

Specimens collected by Mr. Hemphill at Old Mission, Coeur d’Alene, Idaho,
appear to agree with specimens of this species received from Dr. Cooper. The
jaw is low, wide, slightly arcuate, with over 12 broad, stout ribs, denticulating
either margin. The lingual membrane is given in Plate II. Fig. 2. The
central and lateral teeth are slender and graceful. The latter have, apparently,
a second inner cutting point, as is found in Limaz agrestis. I have so figured
it, hoping to draw attention to it, and thus settle the question of its being
there.

Hemphillia.
Plate IV.

From Old Mission, Ceeur d’Alene, Idaho, Mr. Henry Hemphill has sent me
fine large specimens of Hemphillia alive. From these I am able to give the
outward characteristics of the animal in drawings by Mr. Arthur F. Gray.

The animals are larger and much lighter in color than those originally found
at Astoria. They do not while in motion differ from other slugs, though my
former figure of the animal in spirits shows a very great difference, owing to
the contraction being resisted by the internal shell. The rear end of the mantle
seems swollen and blunt, separated from the back, however, and thus alone

does there seem to me any difference in its appearance from Limaxz, whose _

mantle lies flat upon the back. The slit in the mantle is sometimes open,
sometimes closed, and the slit seems to extend quite to the rear of the mantle.
There is a profuse flow of mucus from over the slit. There seem on the man-
tle to be little protuberances, rather than the elongated reticulation of the
rest of the animal. The caudal pore opens and shuts, and exudes mucus in
bubbles sometimes, which occasionally form a solid lump of mucus on the tail.
The horn-like process of the tail so prominent in the first specimens from
Astoria — contracted in alecohol—does not exist in these living specimens,
though occasionally there is a kind of hump above the pore. (See Plate IV.
Fig. D.)

Mr. Hemphill writes: “‘ Hemphillia has a peculiar habit when removed from
its resting place of switching its tail, so to speak, quite rapidly, —a habit I
never noticed in any of our other slugs. I find them hibernating in old
rotten logs.”

The viscera are enclosed under the mantle.

Mr. Gray in drawing the animal called my attention thus to the characters
of the outward markings of the slug :—

i

MUSEUM OF COMPARATIVE ZOOLOGY. 217

“ You are right in saying that the slit in the mantle extends to the back margin.
The central pit seems flooded with mucus at all times, but does not change its
form; the slit, however, seems to widen and show a little ridge on either margin
when the animal is fully expanded. The little tubercles, or small pimples as it
were, seem to cover the posterior portion of the mantle, while the elongated tuber-
cles seem to cover the anterior half, though these at times disappear and the ante-
rior portion runs into folds, which break up the surface, and starting from the
margin of the mantle run to its centre in parallel lines like miniature waves.
They move steadily inward from both margins, disappearing before reaching the
little mucous pit in the centre of the mantle, little wavelets rising at the margins
and keeping up a constant rhythmic motion toward the centre.”

The jaw of this specimen has about 25 ribs, denticulating either margin. It
is low, wide, slightly arcuate, with slightly attenuated ends. (See Plate IX.
Fig. A.)

The lingual membrane is as described and figured by me in Vol. V.; there
are, however, in this form, 57-1-57 teeth, with some eleven true laterals.

The genitalia I have figured in Plate III. Fig.3. It agrees with my figures
in Vol. V. of the genitalia of the original specimens, excepting that the penis
sac, as represented there in Plate XII. Fig. K, is here doubled on itself.

Pupa hordeacea, Gass.

An authentic specimen of this species is figured in the Second Supplement,
Plate III. Fig. 10, referred by mistake to P. Arizonensis in the explanation of

Plate IIT.
Pupa Arizonensis, Gass.

The reference to hebes in Second Supplement should be Fig. 12, not Fig. 10.

LOCALLY INTRODUCED SPECIES.

Tachea nemoralis, Linn.

Fine large specimens of this species have been sent me by Prof. James H.
Morrison, found by him living during the last three years at Lexington, Vir-
ginia. They form part, no doubt, of a colony descended from living individ-
uals introduced from Europe around plants.

Zonites cellarius, MULLER.

Also at San Francisco (Cooper).

Limax maximus, Linn.

Also at New Braunfels, Texas (Singerly).

A drawing of the lingual dentition on Plate II. Fig. 1, shows the cutting
points of central and lateral teeth to be trifid. This is not shown in my figure
in Vol. V.

218 BULLETIN OF THE

SINcE the foregoing was written, the following species have been de-
scribed ; —

Zonites selenitoides, Pitspry.

This species is similar in form and general appearance to Z. minusculus, Binn.,
though decidedly larger. The umbilicus is broad, as in the latter species. The
shell is thin, light yellowish horn-color, almost white. Surface
shining, covered with close strong oblique rib-stria, like Putula
striatella; these striz, while generally regular, sometimes bifurcate,
or separate to give room for another to be inter-
calated. The spire is flatter than minusculus, nearly
plane. The earlier 12 to 2 whorls are smooth, pol-
ished, not striate; the sutures are well impressed.
There are 34 whorls in all, convex, gradually widen-
ing, the last proportionately wider than in Z. minus-
culus. Aperture slightly oblique, lunate, narrower
than in Z. minusculus, its margins thin, acute, scarcely converging, the columellar
shortly subreflexed.

Alt. 1.2 mm., diam. 3 mm.

The specimens were presented to me by Mr. W. G. Binney, who, regarding them
as new, kindly permitted me to describe them. They were gathered by Hemphill,
prince of collectors! at Mariposa Big Trees, California. The name selenitoides is
given because of a certain resemblance to the little Selenites Duranti of Southern
California.

Sculpturing, en-
larged.

Zonites selenitoi-
des, enlarged.

The above description was published by Pilsbry in Proceedings of Academy
of Natural Sciences of Philadelphia, 1889, p. 413, Plate XII. Figs. 13-15.
I give a figure of the original specimen, and of its sculpturing.

Zonites Simpsoni, Pirssry.

This species belongs to that group of /yalina comprising capsella, Gld., Lawe,
W. G. Binn., and placentula, Shutt.,— species with narrow umbilicus, numerous
closely coiled narrow whorls, and without a callus or thickening within the base
of the last whorl. Z. Simpsoni differs from placentula in its much smaller size,
nearly straight, instead of arcuate, basal lip, seen from beneath, proportionately
wider last whorl, and the more trigonal, wider aperture. With Z. Lawe I need not
compare it, as that species is much larger and more elevated. Z. capsella is about
the same size, color, and texture as Simpsoni, but has a narrow umbilicus and
very much narrower aperture, narrowly semilunar instead of trigonal in outline.
Z. Simpsoni has 5 whorls. Alt. 2, diam. maj. 45, min. 4 mm.

The specimens before me were collected by Mr. C. T. Simpson, at Limestone
Gap, Indian Territory. The trigonal form of the aperture is so peculiar that the
species may be separated from Z. capsella at a glance. My comparisons were made
with specimens of capsel/a received from Gould, and placentula from W. G. Binney.
The figures are Gamera lucida drawings.

From Proc. Acad. Nat. Sci. Phila., 1889, p. 412, Plate XII. Figs. 8-10.

MUSEUM OF COMPARATIVE ZOOLOGY. 219

Pupa calamitosa, Pitssry.

Shell minute, cylindrical, very blunt at apex, chestnut-colored; whorls 44, the
first one and a half smooth, the following regularly costulate striate, the costulz
separated by spaces wider than themselves ; last whorl abruptly turning forward,
rounded beneath, encircled by a slight central constriction or furrow; aperture
about one third the total length of shell, rounded, truncated above, contracted
within; peristome thin, expanded, without crest or callous thickening behind;
columellar margin rather dilated; parietal wall bearing two entering lamelle, one
arising near the termination of the outer lip, the other more deep seated, elevated,
entering less obliquely ; columella with a strong white deep-seated obliquely enter-
ing fold; outer lip with two short white lamelle. ;

Alt. 1.70, diam. 0.80 mm.

Two trays of this tiny species are before me. One received from Henry Hemp-
hill, collected near the mouth of San Tomas River, Lower California, the other
collected by Orcutt near San Diego, California. Most specimens show the widen-
ing inward of the outer lip shown in the figure. Several specimens have only one
lamella on the outer lip, and are rather larger than the typical form described,
measuring 1.90 mm. alt. The second parietal lamella is usually much larger than
the first, but in one or two specimens before me this is not the case. The umbili-
cal rimation terminates in a tiny depression, perhaps minutely perforated at the
axis. The formula of denticles or folds (according to Dr. Sterki’s scheme!)
AA BD E or AA BE. The species is of a decidedly different type from any
known American Pupa. P. hordacea, Californica, and Rowelli, abundant Western
forms, belong in quite diverse groups; the first being allied to P. corticaria and
pellucida, the last two grouping with P. decora, Rowelli, and corpulenta,

From the Pupe of the Mexican fauna, leucodon, pellucida, and chordata, the pres-
ent species is quite distinct in every respect.

The inward continuation of the parietal and columellar folds is shown in Figure
17. They are white, regularly veined with darker, like polished plates of agate.

From Proc. Acad. Nat. Sci. Phila., 1889, p. 411, Plate XII. Figs. 16, 17.

Mr. Hemphill sends me the following descriptions, which must be fully
credited to him ; —

Helix tudiculata, var. Binneyi.

This beautiful variety belongs to the globosely depressed forms of H. tudicu-
lata, Binn. It is of a uniform greenish yellow color, without blotches or
markings, except a very faint trace of a band at the periphery. H. tudiculata
is very variable in form, size, and sculpture, and with the umbilicus either
open or closed, but it is very constant in its dark chestnut-color in Southern
California. North of Merced County, however, it becomes a shade lighter,
and passes towards the light, thin form of H. arrosa, which I regard as the

1 See Proc. U. S. Nat. Mus., 1888, p. 369. I have repeated the letter presenting
the parietal fold, as the two seem to be of equal importance.

220 BULLETIN OF THE

progenitor of tudiculata, arrosa in turn having evolved -from its northern
neighbor, H. Townsendiana, Lea, and Townsendiana from the form we now
call H. ptychophorus, Brown, found in Eastern Oregon and Idaho.

Habitat. Mountains of San Diego County, California. Only one specimen
found.

Helicodiscus fimbriatus, var. Salmonensis.

This variety varies from the Eastern or typical forms in the absence of the
revolving lines; otherwise the shells are alike.

Habitat. Banks of Salmon River, Idaho, Old Mission, Idaho, and Oakland,
California.

Helix Kelletti, var. albida.

This is a beautiful clear white translucent variety, with no markings or
stains of any kind. It is quite thin and frail, and a trifle smaller than the
average size of Kelletti. |

Habitat. Santa Catalina Island, California. Two specimens only found
by me.

Helix Kelletti, var. castanea.

Among the numerous patterns of coloring assumed by H. Kelletti, none are
more conspicuous than this well marked variety. The body whorl is of a
deep shiny chestnut-color above the periphery, and becomes lighter as it fol-
lows the whorls of the spire to the apex. The band at the periphery is quite
variable in the different speeimens; it is generally light, and well defined
above, but below it is irregular and spreads over the base of the shell more
or less.

Habitat. Santa Catalina Island, California. This variety is not rare.

Patula strigosa, var. Buttonii.

Shell umbilicated, elevated, or moderately depressed, nearly white, some-
times stained with light chocolate; whorls five, convex, with numerous oblique
strize ; suture impressed, aperture circular ; peristome thickened, not reflected,
darker than the body of the shell; extremities nearly approached and joined
by a callus; with or without a basal tooth ; tooth when present very variable,
generally consisting of a single tubercle ; in some specimens it is nearly or
quite square, as high as long; in other specimens it is long and bifid.

Diameter of the largest specimen, 7 inch; height, $ inch. Diameter of the
smallest specimen, 4 inch ; height, 3 inch.

Habitat. Box Elder Co., Utah.

I dedicate this interesting form of strigosa to my friend, Mr. O. Button, of
Oakland, California.

MUSEUM OF COMPARATIVE ZOOLOGY. 221

Selenites Duranti, var. Catalinensis.

Shell widely umbilicate, depressed, white, transparent when fresh ; whorls
4, flattened above and below, with fine oblique striz; spire planulate ; aper-
ture transversely rounded ; peristome simple, acute ; extremities approached
and joined by a very thin callus in fully matured specimens.

Greatest diameter, } inch ; height, ~ inch.

Habitat. Santa Catalina Island, California.

My little shell differs from the typical Duranti in its greater size, smoother
surface, broader umbilicus in specimens of the same size, but principally in its
transparent shining surface. It is larger than the largest Duranti that I have
seen, but not so large as the costate variety of that species described by Mr.
Mazyck as distinct under the name of S. cwlata, which I have in my possession.
My specimen of that species is larger than his measurements.

I can add the following to his locality : Los Angeles and San Diego, Cali-
fornia, Point Abunda, and banks of San Tomas River, Lower California; thus
giving it a range of about two hundred miles up and down the coast. I have
collected the typical S. Duranti at the following places: Etna Springs, Napa
Co., Healdsburg, Sonoma Co,, Bolinas and San Rafael, Marin Co., Oakland,
Alameda Co., Santa Cruz, Monterey, Santa Barbara Island, Santa Catalina
Island, and San Clemente Island, a range of over one hundred miles north
and south. It is confined to the Coast Range as far as we know at present.

Fig,
Fig.

Fig.
Vig.
Fig.
Fig.
Fig.
Fig.
Fig.

Fig.

BULLETIN OF THE

EXPLANATION OF THE PLATES.

PLATE I

Central tooth of lingual membrane of Mesodon major, the specimen la-=
belled A (see p. 190).
Central tooth, two adjoining lateral teeth, and two marginal teeth of lin-
gual membrane of MJesodon major, the specimen labelled B (see p. 190).
Same: an outer lateral tooth bearing a side cusp and cutting point (see
p. 190).
Mesodon Andrewsi: the genitalia.
ov. oviduct.
g.b. genital bladder.
d.g.b. duct of same.
v.d. vas deferens.

te retractor muscle of penis sac.
p.s. penis sac.

or. common orifice.

Dp: prostate gland.

Penis sac of another specimen of same.

Lingual dentition of same, from specimen labelled E. Two central teeth,
with an adjoining lateral tooth.

Same: marginal teeth.

Same: extreme marginal teeth.

Same: first lateral tooth of specimen labelled F (see p. 191).

Same: marginal tooth (see p. 191).

Same: specimen labelled M (see p. 191), an outer lateral tooth.

The fourth lateral tooth of Limax Hemphilli (see p. 205).

Succinea chrysis, Westerlund, copied from the ‘ Vega Expedition,” Plate
III. Fig. 10.

Succinea annexa, Westerlund, copied from the same, Fig. 11.

PLATE II.

Lingual dentition of : —

Bigs. i.

Fig.

2.

Limar maximus. A central tooth with two adjacent laterals; an outer
lateral; two marginals, the left hand one the last.

Prophysaon (see p. 216). A central tooth with its adjacent lateral tooth ;
an outer lateral tooth; an extreme marginal tooth.

MUSEUM OF COMPARATIVE ZOOLOGY. 993

Fig. 3. Limax Hemphilli. A central tooth with two adjacent laterals; an outer
lateral tooth; two outer marginal teeth.
Fig. 4. Limax Hewston. A central tooth with adjacent lateral on either side ;
incorrectly numbered on the plate; two extreme marginals.
Fig. 5. Microphysa dioscoricola (see p. 196).
PLATE III.
Fig. 1. Prophysaon Andersoni, J. G. C., received from Dr. Cooper.
Fig. 2. Pupilla Floridana, Dall, from original figure.
Fig. 3. Genitalia of Hemphillia, from Old Mission, Coeur d’Alene, Idaho (see
p. 217):—
t. testicle.
ep. epididymis.
ov. ovary.
ovid. oviduct.
pr prostate.
g.b. genital bladder.
d.g. b. duct of same.
v.d. vas deferens.
r retractor muscle of penis.
p. Ss. penis sac.
or. common orifice.
Fig. 4. Helix exigua, from an original drawing by Dr. Stimpson.
Fig. 5. Zonites lasmodon, Phillips, enlarged. Drawn by Miss Helen E. Lawson.
Fig. 6. Central portion of jaw of Wicrophysa dioscoricola, greatly enlarged.
Fig. 7. Bulimus Floridanus (see p. 201). Drawn from original specimen in Mr.
Cumings’s collection, by G. B. Sowerby.
Fig. 8. Lingual dentition of Polygyra hippocrepis.
a. central and two lateral teeth.
b. marginal teeth.
Fig. 9. Bulimus Hemphilli.
Fig. 10. Dentition of Onchidium Floridanum.
PLATE IV.
Fig. D was drawn by W. G. Binney, the other figures by Arthur F. Gray: all
from life.
Fig. A. Hemphillia glandulosa, twice the natural size.
Fig. B. The same; animal in motion, natural size; the slit on the mantle par-
tially open.
Fig. C. The same; partially contracted and at rest.
Fig. D. The same; the very young animal.
Fig. E. The same; dorsal view of posterior portion of the animal, twice the nat-
ural size; pore closed.
Fig. F. The same; lateral view, pore closed.

Py: BULLETIN OF THE

Fig. G. The same; dorsal view, pore open.
a. mucus beads exuding.
b. slit widely opened, the walls or lips rolled out.
c. Mucus accumulations.

Fig. H. The same; lateral view, pore open.

Fig. I. The same as last.

big. J. The same; the internal shell plate.

PLATE V.

Figs. F, H, drawn by W. G. Binney; A, C, D, by Arthur F. Gray; B, E, G, I,
by T. D. A. Cockerell, of West Cliff, Custer Co., Colorado: all from life.
Fig. A. Ariolimax niger, fully extended.
Fig. B. Arvolimax Hemphilli, var. maculatus, Cockerell; animal contracted in alcohol.
Fig. C. Ariolimax niger; the caudal mucus pore, twice the natural size, dorsal
view, the pore open.
a. mucus exuding.
b. b. ridges each side of slit or channel.
ce. mucus channel or pore.
d. little channels conducting mucus from back of animal into channel c.
Fig. D. The same; posterior view.
Fig. E. Ariolimax Californicus, in motion, natural size.
Fig. F. Ariolimax Andersoni, restored from an alcoholic specimen.
Fig.G. Ariolimar Hemphilli, in motion, with end of tail and pore.
Fig. H. Portion of genitalia of E.
p. s. the penis sac.
Jj. — the flagellum.
r. the retractor muscle.
v. d. the vas deferens.
Fig. I. Ariolumax niger, partially extended.

PLATE VI.

Figures B, C, D, E, H, were drawn by A. H. Baldwin, the last from life, the
others from specimens preserved in spirits; Figures F, G, by W. G. Binney, from
life; A, from life, by Arthur F. Gray.

Fig. A. Ariolimax Columbianus, var. maculatus, Cockerell, natural size; from a
specimen collected by Mr. Hemphill.
Fig. B, C. Onchidium Floridanum, three times natural size ; from type.
Fig. D, E. Onchidella Carpenteri, twice natural size.
Fig. F. Tebennophorus Wetherby’; from type.
Fig. G. Portion of genitalia of A.
p. s. the penis sac.
r. the retractor of same.
v. d. tle vas deferens.
Fig. H. Tebennophorus Hemphilli; from the type.

MUSEUM OF COMPARATIVE ZOOLOGY. 225

PLATE VII.

All the figures drawn by T. D. A. Cockerell, excepting I, which was drawn by
Miss Annie Roberts.

Fig. A. Prophysaon fasciatum.

Fig. B.
Fig. C.
Fig. D.
Fig. E.
Fig. F.
Fig. G.
Fig. H.
Fig. IL.

Fig. J.
Fig. K.
Fig. L.
Fig. M.

Pacificum.
Andersoni.

a Hemphilli.

ss pacificum, jaw.

humile, jaw.
“« the animal contracted in spirits, and the surface.
Pacificum ; the same views as last.
ceruleum.
“6
flavum.

humile.
6é

PLATE VIII.

Figure C was drawn by F. W. Earl, from life; A, from life, by W. G. Binney;
B, D, G, I, from life, by T. D. A. Cockerell; E, F, H, were restored by Mr. Cock-
erell from specimen in spirits

Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.

HO OOD p

Phenacarion foliolatus, natural size ; the tail eaten off.
Internal shell of A.

The same, var. Hemphilli, natural size.

Limax Hewstoni ; in motion and at rest.

Hemphilli ; same views as last, and surface.
hyperboreus ; same views as last.

montanus ; same views.

occidentalis ; same views.

Hewstoni ; a larger individual.

PLATE. LX:

Figures A, B, C, D, G, H, were drawn by W. G. Binney; E, F, by T. D. A.
Cockerell; I, J, by Arthur F. Gray.

Fig. A.
Fig. B.
Fig. C.

Fig. D.

Jaw of Hemphillia glandulosa.
Jaw of Phenacarion foliolatus.
Lingual membrane of same; one central tooth, with its adjacent lateral

and three extreme marginals.

Genitalia of same; one half of natural size.

ov. ovary.

ovid. oviduct.

t. _— testicle.

g. 6. genital bladder.
p. s. penis sac.

v. d. vas deferens.

226 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

Fig. E, F. Veronicella olivucea, from one of original lot from Folvon.

Fig. G. Lingual membrane of Glandina decussata.
THis. TH. Genitalia of Phenacarion foliolatus, var. Hemphilli; same references as in
D; one half of natural size.
Fig. I. Prophysaon Andersoni ; surface magnified sixteen times.

a.a. a. reticulations of the body.
b. b. foliolated spaces between reticulations.

Gs lower edge of the body.
d. locomotive disk.
Fig. J. The same, magnified eight diameters ; upper surface; same references

as the last.

PLATE X.

Drawn by A. H. Baldwin, Smithsonian Institution.

Fig. A. Aglaja fidelis ; the large, elevated black variety.
Fig. B. Sculpturing of same.
Fig. C. The same; small, black, elevated form.
Fig. D. Sculpturing of last.
Fig. E. The same; small, depressed form.
Fig. F, Aglaja infumata ; sculpturing.
Fig. G. Arionta Mormonum ; sculpturing of the form figured on Plate XI. Figs. G, H.
PLATE XI.
Drawn by A. H. Baldwin.
Fig Arionta arrosa.
Fig Variety of last, approaching A. exarata.
Fig Sculpturing of last.

Arionta exarata; type.
Sculpturing of last.
Arionta Mormonum.
H. Variety of last, connecting with Hillebrand:.

es
QanoOw>

“

ee a

"

ie, | i |
“a :

Sr, 3

Pte

Need DBs
os ie!
4

z eo ~~, eR i ee = ee

ae

d Suppl to Terr Moll. Plate. II

.

,
!

Binney

Buhords

| ae
’ »
f ‘ tl
5 °

Binney; 34 Suppl. to Terr Moll. Plate. II

~ ot ee Ce ve
Ket
ae

Binney; Suppl. to Terr. Moll.

Plate 1V-

.

Suppl. to Terr. Moll. Plate v.

Binney ;

Binney Suppl. to Terr. Moll. Plate VI.

— ———SSaSSS —o
= —————— = SSS

Binney Suppl. to Terr. Moll, Plate VIE

“

Binney ; Suppl. to Terr. Moll. Plate VIII

ee ee

Binney ; Suppl. to Terr. Moll. Piate IX.

— -_ _ . “wy

7, \’
_. %
i =
:
-
al
ft
.
G »
A
{
———
'
i
o
y
i
i
»
-
a
5
| ~ 2 a
fe a :
a o ons S « :

Lie
Ze y
my

Binney; Suppl. to Terr. Moll. Plate X.

Suppl. to Terr. Moll. Plate XI.

>

Binney

BULLETIN

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT

HARVARD COLLEGE, IN CAMBRIDGE.

VOL. XX.

CAMBRIDGE, MASS., U.S.A.
1890-1891.

(!

UNIVERSITY PRESS: iby.
' JOHN WILSON AND SON, CAMBRIDGE, U.S. A.

CONTENTS.

PAGE
No. 1.— Contributions from the Zodlogical Laboratory. XVII. The His-

tology and Development of the Eye in the Lobster. By G. H. ParxKer.
Ce RO Ce ek! kg I ee 7

No. 2.— On the Rate of Growth of Corals. By A. Acassiz. (4 Plates.)
CEE ics. fa US Rae eal 2 ee ae GE

No. 3.— Preliminary Account of the Fossil Mammals from the White
River and Loup Fork Formations. Part Il. Carnivora and Artiodactyla,
by W. B. Scott. Perissodactyla, by H. F. Osborn. (3 Pisteg ie. aa BB

No. 4.— Contributions from the Zoological Laboratory. XIX. Cristatella:
the Origin and Development of the Individual in the Colony. By C.
eeremmmeort. (ti Plates.) 0.05. 2 2. oe me we ey ew es LOE

No. 5.— Contributions from the Zodlogical Laboratory. XX. The Eyes
in Blind Crayfishes. By G. H. Parxer. (1 Plate.) November, 1890 . 153

No. 6.— Notice of Calamocrinus Diomedz, a New Stalked Crinoid from
the Galapagos, dredged by the U. S. Fish Commission Steamer “ Alba-
meee ty AS Acassiz. December, 1890... >. b.. 6. wk. 186

No. 7.— Contributions from the Zodlogical Laboratory. XXI. The Origin
and Development of the Central Nervous System in Limax maximus.
By Annize P. Hencuman. (10 Plates.) December, 1890 ..... . 169

No. 8.— Contributions from the Zodlogical Laboratory. XXII. The Parietal
Eye in some Lizards from the Western United States. By W. E. Ritter.
Pe re Pannen POOP i 8.4 RRO el ce 4). 209

,

, Oy

ie mn fe wi

No. 1.— The Histology and Development of the Eye in the Lobster.
By G. H. ParKer.?

TABLE OF CONTENTS.

PAGE PAGE

MeaMInPOURION 5 2k ke 1 6. Accessory Pigment-cells . . 25

CO a re 3 7. Innervation of Retina . . . 26

Oi Easiology .. ee 41 TIT, Development a9 ..0.).0..5 «) » 8h

1. Corneal Pe paces. ee 6 1. Planprthe Bye .-4 <..« » 3b

eee ww el elf CO 2. Optic Nerve .. 43

3. Distal Retinule . .. . ~15 3. Differentiation of Omiautdia 45

4. Intercellular Spaces of ha 4. Types of Ommatidia .. . 56

Rec el iet ve pvirn GeO | LV 9 A DHOMEMORE Ste cilgim nulina ree

5. Proximal Retinule . . . . 20] V. Explanation of Figures . . . 60
INTRODUCTION.

THroueH the kindness of Mr. Alexander Agassiz it was my privilege
to spend the greater part of the summer of 1887 at the Newport Marine
Laboratory. During the preceding winter I had been interested in the
structure of the eyes in Arthropods, especially in the inversion of the
retina in Arachnoids and my instructor, Dr. E. L. Mark, had called my
attention to the importance of ascertaining whether the retina in the
compound eyes of Crustaceans was inverted or not. At about this time
Kingsley (’86*) published his preliminary account of the development of
the compound eye of Crangon, and claimed that in this crustacean, as
in spiders, the retina was inverted. For reasons which I shall mention
in the course of this paper, Kingsley’s account did not seem fully sat-
isfactory to me, and consequently I decided to study for myself the
development of the eye in a crustacean. My visit to the Newport
Laboratory offered an excellent opportunity to collect embryological
material for such a study. During August and September spawning
lobsters were easily obtained, and I therefore determined to study
the eye in the lobster, Homarus americanus, Edwards. A series of
lobsters’ eggs were collected, and before leaving Newport my observa-

1 Contributions from the Zodlogical Laboratory of the Museum of Comparative

Zodlogy, under the direction of E. L. Mark, No. XVII.
VOL, xx. — NO. l. 1

2 BULLETIN OF THE

tions had been carried far enough to satisfy me that the retina in the
lobster was a simple ectodermic thickening. On returning to Cam-
bridge from Newport, the study of the lobster’s eye was continued in
the Embryological Laboratory at Harvard College, under the direction
of Dr. Mark. Here I completed the observations on the development
of the eye, and studied its histology. In the fall of 1888 a brief pre-
liminary account of the results which are now presented in full was
published in “ The Proceedings of the American Academy of Arts and
Sciences,” Vol. XXIV. pp. 24, 25.

In procuring at Newport the necessary stages in the development of
the lobster I proceeded as follows.

Female lobsters with eggs were obtained from the fishermen, and
kept in floating latticed boxes which were anchored in the small cove
beside the Laboratory. A few eggs were taken daily from each lobster.
The reagents which I employed in killing the eggs were Kleinenberg’s
picro-sulphuric acid, Perenyi’s fluid, a saturated aqueous solution of
corrosive sublimate, and hot water. The eggs which were prepared with
corrosive sublimate were rendered almost useless by the subsequent
formation of a fine precipitate. Those which were killed in Kleinen-
berg’s picro-sulphuric acid and in Perenyi’s fluid gave fair results ;
the latter reagent left the yolk in good condition for cutting. The
best results, however, were obtained by the use of hot water. Eggs
which had been prepared in this way could be easily shelled, and the
embryos could be readily dissected from the yolk. The separation of
the embryo from the yolk proved to be a great advantage, and obviated
the necessity of cutting the yolk, a tedious process in an egg as large
as the lobster’s.

In the following account of the development of the lobster’s eye, the
stages which it is necessary to describe are taken from different sets of
eggs. These sets were from different lobsters, consequently I cannot
state with exactness their relative ages. I shall therefore characterize
them by their most evident structural peculiarities. Beginning with the
earliest stage and proceeding to the later ones, I have lettered them
A, B,C, D, E, and F. Set A is in the stage of the “egg-nauplius”; in
this set the characteristic three pairs of appendages are easily distin-
guishable. In set B the thoracic appendages have begun to form.
This stage corresponds very closely to what Reichenbach (’86, Plate III.
Fig. 11) has designated in the crayfish as stage H. In stage C the
first trace of pigment in the retina is visible. Stage D is from the
same series of eggs as stage C, but is seven days older than C. In both

MUSEUM OF COMPARATIVE ZOOLOGY. 3

stages C and D, the abdomen of the embryo is recurved, and reaches
forward covering the space between the optic lobes. Stage E corre-
sponds to the time of hatching. Stage F is represented by a young
lobster one inch in length.

The younger stages which follow the hatching of the lobster are
obtained with considerable difficulty, and I am under obligations to
several of my friends for material which covers this period. For some
lobsters in the “ Schizopod” stage I am indebted to Mr. Sho Watase.
Mr. H. H. Field and Mr. Carl H. Eigenmann kindly collected for me
some young lobsters one inch in length. From Mr. F. L. Washburn I
received the eyes of several half-grown lobsters, six to eight inches in
length. The material which I used in studying the histology of the
eye in the adult was very kindly supplied to me by A. T. Nicker-
son and Company, of Charlestown, Mass.

Methods.

The methods of staining, embedding, etc., which I have employed,
are those known to all students of modern histology. In one case,
the staining of nerve-fibres, | have used a method which I accidentally
discovered while experimenting with Weigert’s hematoxylin. ?

In employing this method it is necessary to stain the sections on the
slide. ‘The way in which I have stained sections on the slide has already
been described (’87, p. 175). Further experience has shown, however,
that the successful employment of this method necessitates a careful
observance of certain precautions. These I have not sufficiently em-
phasized in my former account, and I therefore redescribe the method,
calling especial attention to the precautions. The method consists in a
cautious use of Schiallibaum’s fixative. The fixative which I have em-
ployed is composed of clove oil three parts and Squibb’s flexible collo-
dion one part. The mixture before being used should be allowed to
stand for about a week. After several months it may become ineffective.
When working, I usually employ the fixative frequently enough to fol-
low its changes, and at the first signs of failure I make a new mixture.
If for any reason I have not used the fixative for some time, I test it
with a few waste sections before employing it with valuable material.
In using it a moderate amount is applied to the slide, and the sections
in paraffine are placed on it. The slide and its sections are now sub-
jected to a temperature of 58° C. for fifteen minutes. It is important
to observe carefully both the length of time during which the slide is
heated and the temperature to which it is raised. At the end of fifteen

4 BULLETIN OF THE

minutes, the slide, while warm, is thoroughly washed with flowing tur-
pentine. This can be applied conveniently from a small wash-bottle. AI
of the paraffine should be removed from the slide before it becomes
cool, otherwise on cooling some paraffine may solidify. This is liable to
loosen the film of collodion. The wash of turpentine should be contin-
ued not only till the paraffine is thoroughly removed, but till the slide is
cool. Then, and not till then, can the turpentine be safely replaced by
alcohol, first 95%, then 70%, 50%, and 35%, and finally it can be im-
mersed in water. After once having got the slide with its sections into
water, the subsequent treatment with alcohol and water seems to have
no effect in loosening the sections, although the film of collodion will
dissolve easily in ether. I have very generally employed this method
of staining for two years, and as it obviates the difficulties which arise
from maceration or partial penetration of dyes, I use it in preference to
staining zm toto. I have lost very few sections by it, and such accidents
as I have had were due, I believe, to a neglect of some of the precau-
tions which have been mentioned.

The method of staining nerve-fibres which I have employed consists
of a modified use of Weigert’s hematoxylin. The tissue which was
stained by this method was for the most part killed in hot water,
although I have also successfully stained nerve-fibres which were killed
in chromic acid and Kleinenberg’s picro-sulphuric acid. Sections of the
optic nerve which had been mounted on the slide and carried into water
were treated for about half a minute with an aqueous solution of
potassic hydrate ~5%. They were then thoroughly rinsed in distilled
water and transferred to Weigert’s hematoxylin. Here they remained
for about three hours at a temperature of 50° C. They were then
rinsed again in distilled water, carried through the grades of alcohol,
and after being dehydrated with alcohol of about 99%, they were cleared
in turpentine and mounted in benzole balsam. ach nerve-fibre when
so treated had a distinct blue-gray outline. The sections do not over-
stain even when they are kept in the dye for a prolonged period, and
there is of course no subsequent decoloring. This method yields fair
results when applied to nerves from any part of the lobster’s body, but
it is especially successful in treating that portion of the optic nerve
which intervenes between the retina and the optic ganglion.

THe Histouoey.

The two movable eye-stalks of the lobster are situated one on either |
side of the rostrum, at the angle which that structure makes with the

MUSEUM OF COMPARATIVE ZOOLOGY. 5

anterior edge of the carapace. The form of the eye-stalk approaches
that of a short cylinder terminated by a hemisphere. The cylindrical
part of the stalk resembles the general surface of the body in that it is
covered with a firm, calcified cuticula. Excepting a portion of the
surface next the rostrum, the whole of the hemispherical part during
life is black, and covered with a flexible cuticula. The black area de-
fines the position of the retina. That portion of the hemispherical
surface which is not black, and which faces the rostrum, is covered with
a peninsula-shaped piece of inflexible cuticula. A broad isthmus of
the same kind of cuticula connects this with the shell of the cylindrical
part. The absence of the retina from the peninsula-shaped. portion of
the hemisphere is due in all probability to the fact that the field of
vision for this part of the hemisphere is cut off by the rostrum. The
remainder of the hemisphere, that part on which the retina is devel-
oped, faces away from the lobster’s body, and its field of vision is not
permanently obstructed by any part of the animal.

A section perpendicular to the surface, and cutting the eye-stalk in a
region where the cylindrical and hemispherical parts unite, is shown in
Figure 26. The thick, calcified cuticula of the cylindrical part is indi-
cated at cta. On the inner surface of this cuticula is a thin hypoder-
mis (4 d.). The hypodermis is bounded on its inner face by a basement
membrane (md.). The cuticula of the hemispherical part (crn.) is thin
and flexible. It can be designated by the name corneal cuticula.
(Compare Patten, ’86, p. 544.) Resting on the deep face of the corneal
cuticula is the thick cellular layer, named by Lankester and Bourne the
ommateum (omm’.). The proximal face of the ommateum is limited by
a basement membrane, which is continuous with that bounding the
corresponding face of the undifferentiated hypodermis. The ommateum
is continuous with the hypodermis, and in fact can be regarded as a
thickening of that layer. Carriére (85, p. 169) has already pointed
out in the eye of Astacus a similar relation between the hypodermis
and ommateum, and he believes that this relation holds good for all
Decapods.

On inspecting the external face of the corneal cuticula, one finds it
divided into an immense number of square facets, one of which is shown
in Figure 2. Although as a rule the outline of the facet is square, it is
- not invariably so ; for on the margin of the retinal area close to where
the ommateum passes over into the undifferentiated hypodermis, the
outline often becomes somewhat irregular, and more frequently presents
the form of a hexagon than of a square (Fig. 59). The number of facets
‘in each eye of an adult lobster is about 13,500.

6 BULLETIN OF THE

In the ommateum the cells are arranged in specialized groups or
ommatidia. There is a single ommatidium under each. corneal facet,
consequently in any given eye the number of ommatidia equals the
number of facets. The cellular composition of each ommatidium is
best understood from a comparison of longitudinal and transverse sec-
tions. Figure 1 represents a longitudinal section through an ommatid-
ium. The thick lamellated layer (ern.) at the distal end is the corneal
cuticula. Directly below this is a thin layer of cells, the corneal hypo-
dermis (crn. hd.). Following on the corneal hypodermis are the cone-
cells (cl. con.). They are very long, and extend from the corneal hypodermis
inward till their proximal ends disappear in the deep part of the retina.
In reality they terminate upon the basement membrane. Their distal
ends in the region of the crystalline cones are surrounded by pigment-
cells, to which I give the name distal retinule (rtn’. dst.). These, like
the cone-cells, extend to the deeper part of the retina. Here the proxi-
mal retinule and accessory pigment-cells occur. The proximal retinule
are elongated cells (rtn’. px.), and contain black pigment. They sur-
round the rhabdomes (rhb.). The accessory pigment-cells are irregular
cells, which fill the space between the deep ends of the proximal reti-
nule. They contain a pigment which is whitish by reflected and yellow-
ish by transmitted light. Their nuclei are shown at nl. pig., Figure 1.
The last two kinds of pigment-cells described rest upon the basement
membrane (mb.); below this membrane the fibres of the optic nerve
can be seen (n. fbr.).

From this description it will be seen that the ommateum lies between
the corneal cuticula and the basement membrane, and is composed of
the following kinds of elements: cells of the corneal hypodermis, cone-
cells, distal retinule, proximal retinule, and accessory pigment-cells.
The numbers and positions of these cells are best made out from trans-
verse sections. The several kinds of cells will be discussed in the order
named.

The Corneal Hypodermes.

That the corneal cuticula in Decapods is separated from the cone-cells
by an intervening layer of cells is a view which has been held only by
recent investigators. Grenacher (’79, p. 123), in his account of the eyes
in Decapods makes no mention of such a layer, and leaves one to con-
clude that the cone-cells abut against the cuticula. Claus (’86, p. 57)
suspected the presence of a corneal bypodermis in Decapods, Schizopods,
and.Stomatopods, but his search for it was in vain.

|
|
|

a

~~.

MUSEUM OF COMPARATIVE ZOOLOGY. a

The view that the cuticula and cone-cells are in contact, is strongly
contrasted with that, maintained by Patten ('86, pp. 626, 642). Ac-
cording to this writer, the corneal cuticula is due to the activity of a
layer of cells, the corneal hypodermis, which intervenes between the
cuticula and the cone-cells. Patten has identified the corneal hypo-
dermis in the following genera of Decapods: Penzeus, Paleemon, Pagurus,
and Galathea. It has also been described by Kingsley (’86, p. 863) in
the eyes of Crangon, and by Herrick (’86, p. 43) in the eyes of Alpheus.
Carriére (’89, p. 225) has recorded it in the eye of Astacus, and there is
now good reason for believing that a corneal hypodermis exists in the
_ eyes of all Decapods.

Patten’s statement (86, pp. 665, 666) that the corneal hypodermis
“has been invariably overlooked by Grenacher,” and Kingsley’s asser-
tion (86, p. 863) that the existence of the corneal hypodermis “was
utterly ignored by Grenacher,” are perhaps a trifle too strong. It seems
much more probable that Grenacher confused the nuclei of the cone-
cells and corneal hypodermis. He evidently never saw both kinds of
nuclei in the eye of the same Decapod. In some cases he may have
described the nuclei of the cone-cells, in other cases those of the corneal
hypodermis. In both instances what he described he took to be the
nuclei of the cone-cells. In the eye of Mysis, I believe that he (’79,
p. 118) described the nuclei of both the cone-cells and corneal hypoder-
mis, although in this case he was of the opinion that both sets of nuclei
belonged to the cone-cells. Where only one set is figured, it is difficult to
decide whether he has given the nuclei of the cone-cells or of the corneal
hypodermis. So far as I am aware, there are always in each ommatid»
ium of a Decapod two hypodermal nuclei, and fowr nuclei in the cone-
cells. This numerical relation is sufficient to distinguish the groups of
nuclei, but it can only be employed satisfactorily where transverse sec-
tions at the proper niveau are given. Unfortunately, in the Decapods,
Grenacher did not figure any such sections, and it is therefore difficult
to decide in particular cases which kind of nuclei he has described.

In the lobster a well differentiated corneal hypodermis has already
been pointed out (Fig. 1, crn. hd.). In transverse sections this presents
the appearance of squares of granular protoplasm (Fig. 3). Each square
contains two nuclei, and is bounded by a membrane. <A narrow space
filled with granular substance separates the membranes of adjacent
squares. From the longitudinal section (Fig. 1) it will be seen that
these squares are relatively thin, so that their proportions are somewhat
like those of square tiles. The outer face of the tile is flat; its inner

8 BULLETIN OF THE

face is hollowed, however, so that its centre is the thinnest part. In a
- few cases the corneal hypodermis has appeared as cubical blocks, rather
than as tiles. This thickened condition probably indicates an increased
functional activity, and the more frequently occurring tile-like condition
may correspond to a quiescent stage.

The two nuclei contained in each square are placed some distance
apart, and on one of the diagonals of the square (Fig. 3). Their long
axes are approximately parallel to the other diagonal. In a given eye all
the squares agree in having the nuclei on parallel diagonals. The pres-
ence of two nuclei in a square indicates that the square consists of two
cells. Any membrane separating the two cells must necessarily pass
between the two nuclei, but all attempts to discover such a membrane
have failed. However, for a reason which will be given shortly, I be-
lieve that the protoplasm of the hypodermal square is divided by the
diagonal which lies between the nuclei. In the centre of each square
several oval or round outlines are usually visible (Fig. 3). These are
vesicular bodies which occur in the distal ends of the cone-cells, and
which can be seen through the very thin corneal hypodermis.

The corneal cuticula is the result of the activity of the corneal hypo-
dermis. Viewed from the surface, the cuticula is divided by narrow
bands into square facets (Fig. 2). Each facet is external to a hypoder-
mal square. The proximal and distal faces of each facet, as can be seen
in the transverse section (Fig. 1), are very nearly flat, the proximal
face only being a trifle convex. This convexity, however, is so slight
that one cannot attribute to the facet the character of a lens.

When a piece of corneal cuticula is cleaned by treating it with potas-
sic hydrate, and is then examined in water, the markings which are
visible with difficulty in preparations mounted in balsam are easily seen.
Each facet in addition to its narrow limiting bands has a faintly marked

diagonal band which divides the square into two equal triangles (Fig. 2).

In the different facets of a given eye the diagonal bands are parallel.
Newton (’73, p. 327, Plate XVI. Fig. 3), in describing the structure of
the eye in the lobster, states that each facet is crossed by two diagonals
at right angles to each other. This statement I cannot confirm, for,
although I have searched with care, I have never succeeded in finding
more than a single diagonal in each facet. In the middle of the diagonal
there is an irregular hazy patch. This at times has a distinctly marked
cross in it. When the cross is present, one of its axes lies in the diago-
nal band, the other extends at right angles to the band (Fig. 2).
Whether all of these markings extend through the substance of the

——_-

MUSEUM OF COMPARATIVE ZOOLOGY. 9

cuticula, or whether they are confined to its surface, is difficult to say.
The production of the cuticula is such a uniform process that one would
naturally expect to find that the marking extended through it, for the
successive layers would be similarly marked, and thus bands would be
established extending from its deep to its superficial face. Concerning
the vertical extension of the bands between the facets there is no ques-
tion, for in transverse sections of the cuticula (Fig. 1, ~) they reappear
in their proper positions, and extend from one surface to the other.
Owing to the roughness of the cut face, they are much less readily de-
tected in sections than when viewed from the outer surface of the cutic-
ula (compare Figs. 1 and 2). The diagonal band and its central spot
have not been observed in transverse sections, even when the plane of
section is in the most advantageous position for demonstrating these
structures. Notwithstanding their apparent absence, both may be pres-
ent, although indiscernible. For even in the superficial view, when the
outline of the facet was so readily visible, the diagonal band was only
faintly seen. In transverse sections, where the distinct boundary of the
facet is visible with difficulty, one should not expect to see the much
fainter diagonal. On comparing the diagonal band and the boundary
of the facet by focusing through the corneal cuticula, I was unable to
distinguish a greater vertical extension in the one than in the other.
Since it has been shown that the boundary of the facet extends through

the cuticula, this observation supports the conclusion that the diagonal-

band also extends through it.

Patten (’86, pp. 626, 627) has described in the facet of Penzus a
band which has many resemblances to the diagonal band in the lobster.
It is not diagonal, however, but transverse, and divides the square facet
into two equal rectangles, in which the sides are in the proportion of
one to two. I have already given my reason for believing that the
diagonal band in the cornea of the lobster extends through the sub-
stance of the cuticula. Patten states that the transverse band in
Penzeus is only a superficial structure, and says (86, p. 627) that in
cleaning the cuticula “ when the treatment with caustic potash has been
carried to excess, all markings disappear except the contours of the
facets.” I have subjected the corneal cuticula of a lobster to a boiling
solution of potassic hydrate (75%) for a quarter of an hour, and, al-
though the potash completely cleaned the cuticula, the outlines of the
facets, the diagonal band, and its spot were as readily visible after this
treatment as before. A second and third trial with the same piece of
cuticula did not noticeably effect the markings. In this respect, then,

10 BULLETIN OF THE

the diagonal band in the lobster is materially different from the trans-
verse band in Penzus, and I conclude that in the cornea of the lobster
the limiting and diagonal bands are essentially similar in that they both
extend through the cuticula.

In all probability the bands between the facets were produced during
the secretion of the cuticula by the interference of the partitions which
separate the hypodermal squares. If this be true, it is probable that
the diagonal bands represent a like interference. It is important to
notice that the diagonal band in the cuticula corresponds to the imagi-
nary diagonal which lies between the nuclei of each hypodermal square,
never to the diagonal which crosses the nuclei (compare Figs. 2 and 3).
This diagonal then corresponds to the position in which one would look
for a membrane between the pair of hypodermal cells; and although
such a structure has not been observed, the diagonal band in the cornea
is a strong indication of its presence.

Admitting this to be the significance of the diagonal band, it is but
natural to expect that, if deeper cells touch the cuticula, they would pass
outward between the hypodermal cells. The fact that the hazy patch
which lies in the middle of the facet is always on the diagonal band,
and directly external to the distal tips of the cone-cells, leads to the
belief that this patch marks the place where the cone-cells pass between
the cells of the hypodermis and touch the cuticula. I am not of opin-
jon that the patch is produced by the secretion of the cone-cells, al-
though I have no evidence that the cone-cells cannot produce cuticula
at their distal tips. It seems to me more probable that they have given
rise to the patch by a series of interrupted interferences with the ac-
tivity of the corneal hypodermis. If such be the case, a distinct cross
might be produced when the area of interference was definitely circum-
scribed. When the area was not so sharply bounded, a hazy patch with
indistinct outlines might be the result.

From the facts which have been presented, I conclude that each hypo-
dermal square consists of two flattened cells, triangular in outline, and
very intimately applied on their longest sides. °

The Cone-cells.

One of the most important questions in the anatomy of the cells of
the crystalline cones (retinophoree) concerns the relation which these
cells bear to the rhabdome. Max Schultze was the first to maintain
(67, p. 407) that the cone-cells and rhabdomes were separate struc-
tures. Grenacher’s researches lead to the same conclusion. As an oppo-

MUSEUM OF COMPARATIVE ZOOLOGY. 11

nent of this view, Patten (’86, p. 670) has claimed that the cone-cells
and rhabdome were continuous, and in fact that the rhabdome of the
compound eye was only an enlargement of the proximal end of the cone-
cell. Kingsley (’86, p. 863) in his description of the eye in Crangon
supported Patten’s view.

Of those authors who maintain the separateness of the cone-cells and
rhabdome, no one, I believe, has given a fully satisfactory account of the
way in which the proximal ends of the cone-cells terminate. Grenacher,
in describing the eye in Palemon said (’79, p. 123): “ Die fein ausge-
zogene Spitze dieser Pyramide [the cone-cells] durchsetzt, bevor sie in
contact mit der Retinula tritt, zuerst eine in Form eines Hohlcylinders
sie umhiillende Pigmentmasse um sich dann in das Vorderende der
Retinula eine Strecke weit einzusenken.” A more detailed account was
given by Schultze, who, after stating (’68, p. 10) that in some crusta-
ceans the cone-cells appeared to terminate a little in front of the distal
end of the rhabdome, said that in the crayfish “geht der Krystallkegel
nach unten in vier Spitzen aus, welche sich aus den vier Kanten der
Oberflache entwickeln und das obere Ende des nervésen ebenfalls vier-
kantigen Sehstabes umschliessen. Die vier Spitzen legen sich dabei an
die Kanten des letzteren an und laufen als lange feine Faden auf der
Oberfliche des Sehstabes herab, diesen umklammernd und mit ihm ober-
flachlich verbunden aber durch Maceration isolirbar. Gegen das Ende
spitzen sie sich fein zu und verlieren sich auf der Oberfliche des Kérpers,
den sie umfassen.” This account is the most complete of any that I
have seen, and yet that Schultze was not fully satisfied that he had seen
the proximal termination of the cone-cells is probable from the fact that
he says the fibres are Jost on the surface of the rhabdome.

The relation of the rhabdome to the cone-cells, and the way in which
these cells terminate in the lobster, is as follows. As in other Deca-
pods, each ommatidium in the eye of the lobster contains four crystalline
cone-cells. ‘Together these cells form an elongated pyramid, with its
base next the corneal hypodermis and its apex on the basement mem-
brane (Fig. 1, cl. con.). At the distal end of the ommatidium, in the
region which corresponds to the base of the pyramid, the four cells
are closely applied to each other. This condition is maintained till
the deeper part of the ommatidium is reached. Here the four cells,
reduced to fibres, separate and end independently on the basement
membrane.

A transverse section of the distal ends of the cone-cells is shown in
Figure 4. On the external faces of each group of four cells there is a

12 BULLETIN OF THE

distinct bounding membrane (mb. pi ph.). This can be called the periph-
_ eral membrane. The four cells in each group are separated one from
another by delicate membranes (md. 7 cl.), which often show undoubted
continuity with the peripheral membrane. These membranes, since they
lie between the cone-cells, can be called the intercellular membranes.
The distal end of each cell contains coarsely granular protoplasm and a
nucleus (Fig. 4, 2/. con.). The nuclei usually lie in the external angles
of the cells, and do not readily take up coloring matter. The terminal
granular protoplasm of the four cells forms a distal cap (Fig. 1, cap.).
This cap fills the concavity on the proximal face of the corneal hypo-
dermis, and its central distal tip probably passes between the pair of
hypodermal cells and touches the cuticula. The spot or cross which is
thus probably produced in the centre of each facet has already been
described.

Below the cap of granular protoplasm is the crystalline cone. The
firm peripheral membrane of the cap is continued over the cone and
proximal part of the cone-cell. The distal end of the cone in cross-
section is a square with rounded angles. At this end there is no indi-
cation of a division of the cone into four segments corresponding to the
four cells. A transverse section midway the length of the cone (Fig. 5)
shows no features essentially different from those of the section across
the distal end. The proximal end of the cone in cross-section is nearly
circular (Fig. 6). On the sides of this end of the cone one often notices
small re-entrant angles (Fig. 6, x). The peripheral membrane dips into
these. The angles are usually four in number, never more, and occupy
positions which indicate the planes of separation between the four cone-
cells. In some cases delicate membranes originating from the angles di-
vide the substance of the cone into its four constituents (Fig. 6, mb. 7 cl.).
These membranes correspond in position to the intercellular membranes
at the distal end of the cone-cells. The substance of the cone is very
finely granular. The four constituents of each cone terminate very
nearly at one level. In passing in a proximal direction through a series
of sections, the substance of the cone is last seen as a thickening which
flanks the cell membranes, especially the intercellular membranes.
(Compare Figs. 1 and 6.)

Below the cones the outlines of the four cone-cells are well marked
by both peripheral and intercellular membranes (Fig. 7). The inter-
cellular membranes are continuous with those seen in the proximal ends
of some cones. In this region the cells contain coarsely granular proto-
plasm. In passing from the deep ends of the cones to the proximal

—

MUSEUM OF COMPARATIVE ZOOLOGY. ig

retinule, the most striking difference noticed in the cone-cells 1s a dimi-
nution in their diameter. (Compare Figs. 1, 6, and 8.) On a level with
the distal ends of the proximal retinule, the groups of cone-cells still
retain their four-parted character (Fig. 9, cl. con.). Each group is easily
traced between the retinule till it approaches the distal end of the
rhabdome. The change which here takes place is represented in Fig-
ure 12. Of the four ommatidia which are shown in this section, the one
indicated at a is cut slightly above the rhabdome. In the case of om-
matidia 6, c, and d, the plane of section passes through the end of the
rhabdome. In each of these three, it will be noticed that the rhabdome
is surrounded by four bodies, which correspond in position to the four
cone-cells. The four ommatidia which are drawn in Figure 12 are in
no way exceptional, but represent a very usual condition. Many such
cases have been examined, and whenever the tip of the rhabdome was
in the section, it was invariably surrounded by the four bodies previously
mentioned. When, on the other hand, the plane of section did not pass
through the rhabdome, only the four cone-cells were present. The
round bodies at the sides of the rhabdome can be traced from section
to section, and I therefore believe them to be fibres. Moreover, there
is no break observable in their continuity with the cone-cells, and I
therefore further believe that they are the fibrous prolongations of the
cone-cells. They have one peculiarity which is worthy of comment.
As the cone-cell passes over into the fibre, a considerable diminution in
its diameter takes place. This is accomplished at the distal end of the
rhabdome, and within a space equal to the thickness of one or at most
two sections (7.5-15 »). Occasionally there is to be seen a group, in
which one or two cells have been reduced to fibres, and the remaining
ones are as large in transverse section as an individual in ommatid-
ium a (Fig. 12). The conclusions which are arrived at from the study of
sections are confirmed by isolation-preparations. Figure 28 represents
a portion of an isolated group of four cone-cells from a single omma-
tidium. In the distal part of the specimen the four cells are intimately
bound together, but at the proximal end they appear as four separate
fibres. As in the transverse section (Fig. 12), the continuity of the
fibrous and thicker portion of the cone-cell, and the rapid reduction of
the cone-cell to form the fibre, are plainly seen. The cone-cells are usually
somewhat separated before they reach the rhabdome. It can scarcely
be said that they touch the rhabdome, although this is the region in
which they are nearest to it. As the fibres pass into the deeper part of
the retina they are found to lie nearer the periphery of the ommatid-

14 BULLETIN OF THE

ium. They lie between the proximal retinule, but at some distance
from the rhabdome (Fig. 14, el. con.). They still retain, however, the
same relative positions in the ommatidium. ‘Their peripheral location
is maintained (Fig. 17) till they are very close to the basement
membrane.

The changes which the fibres undergo as they approach the basement
membrane is shown in Figure 19. In this section the plane of cutting
was slightly oblique to the basement membrane. Of the three omma-
tidia which are here represented, those indicated at ¢ and d are cut at
about the same level, and very close to the basement membrane. Om-
matidium 0 is cut farther from the membrane than either c ord. The
fibres of the cone-cells are closer to each other in ¢ and d than in 0.
As this condition is generally true in other sections, it follows that, as
we approach the basement membrane, the fibres of the cone-cells con-
verge. The convergence is also shown in Figure 21. Of the omma-
tidia here figured, 6 is cut farthest from the basement membrane. In
it the four cone-cells (cl. con.) can be seen, and between them a dot
which represents the proximal end of the rhabdome (rhb.). Nearer the
basement membrane is ommatidium a, in which the four cone-cells can
be recognized, and to one side a fibrous area. The rhabdome does not
extend as deep as this. The fibrous area represents a region in which
the plane of section passes through an elevation on the distal face of
the basement membrane. Ommatidium d is still nearer the membrane.
The cone-cells are here brought more closely together, and are sur-
rounded by the fibrous substance of the elevation. The form of the
elevation is now seen to be that of a cross. At # is shown a basal section
of the cross-shaped elevation surrounded by four large openings through
the basement membrane. The cone-cells are no longer visible at this
level, and I therefore believe that without penetrating the basement
membrane they terminate in these elevations. This belief is further
supported by the fact that in transverse sections of the basement mem-
brane the cone-cells distinctly end in the substance of these elevations
(Fig. 29, cl. con.).

The facts obtained from a study of the lobster’s eye support the claim
made by Schultze and Grenacher, that the cone-cells and rhabdomes are
separate structures. In the case of the crayfish, Schultze, moreover,
saw the prolongations of the cone-cells, and traced them into the deeper
part of the retina. It is probable that in the crayfish, as in the lobster,
the fibrous ends of the cone-cells terminate in the basement membrane,
but this Schultze did not see. Such an omission is by no means sur-

MUSEUM OF COMPARATIVE ZOOLOGY. 15

prising ; for when we reflect upon the methods at his command, it is
remarkable what success he had in tracing the course of the fibres, and
in demonstrating the relation of the cone-cells and rhabdome.

Patten has advanced the view that the cone-cells are provided with
an axial nerve-fibre, and that the cone itself is the true perceptive ele-
ment. I shall defer the consideration of this topic till I describe the
innervation of the retina.

The Distal Retinule.

Surrounding each crystalline cone are two pigment-cells, the distal
retinule. These cells not only surround the cone, but extend as fibres
into the proximal part of the retina (Fig. 1, rtn!. dst.).

The relation which the distal retinule sustain to the cone can be
studied most readily in transverse sections. In a section passing
through the distal end of the cone (Fig. 4, ommatidium a), it will be
observed that of the four lateral faces which the cone presents, two, the
lower and left-hand ones, are covered by a single retinula (rén’. dst.).
The retinula is thickest at the lower left-hand angle of the cone, and
becomes thinner the farther it extends on the two adjacent faces. At
the more distant edges of these two faces the retinula terminates.
Thus the retinula is composed of a central portion and two blade-like
extensions. Hach blade covers one face of the cone. The second
retinula is essentially like the one just described, but lies at the upper
right-hand angle of the cone and covers its upper and right-hand
faces. In this way the four faces of the cone are sheathed by a pair of
retinule.

On inspecting the arrangement of the retinule in adjoining omma-
tidia (Fig. 4), it is evident that they are so placed that the thick end
of each blade-like portion is opposite the thin end of the blade of a
neighboring retinula, and that, in passing along the space between the
cones, as one retinula becomes thicker the other becomes thinner. The
delicate membranes which separate the blades consequently extend
obliquely across the spaces between the cones (Fig. 4). In the space
which is left between the angles of four adjoining cones the mem-
branes of the retinule are very much thickened. That this is a
thickening in the membrane of the retinula, and not due to substance
produced by the cone-cells, seems probable for two reasons. First, the
membrane is often somewhat thickened in regions between two retinule,
and where the cone-cells could not well touch it. Such thickenings are
directly continuous with the larger ones already mentioned (Fig. 4).

16 BULLETIN OF THE

Secondly, in isolation-preparations the thickenings always remain at-
tached to the retinule ; the cones, on the other hand, are covered with
membranes of uniform thickness. The thickening in the membrane
is not characteristic of the whole length of the retinula, but is pecu-
liar to the region corresponding in level to the distal end of the cone
(Fig. 1).

The foregoing description applies to the structure of the distal
retinulz as seen in the plane of Figure 4. This plane passes through
the outer ends of the cones. In other regions the retinule present
somewhat different conditions. The relation of the retinule to the
hypodermal squares is shown in a section (Fig. 3) which is slightly
more superficial than that just described. The two retinule which
were located at the two angles of the cones here occupy the corre- '
sponding angles of the hypodermal squares. They do not, however,
entirely cover the four lateral faces of the square, as they did those of
the cone, but from the angles at which they are located they extend
over half of each of the adjoining faces. It follows from this that
together they flank only one half of the lateral exposure of the square.
The blades are now no longer wedge-shaped in transverse section, nor
do they overlap neighboring blades, but each one stretches completely
across the space in which it lies. Of the lateral surface of the square.
that half which is not sheathed by its own pair of retinule is covered
by the arms of four adjacent retinule. Consequently, six retinule in
all touch each hypodermal square. Two of these belong to the omma-
tidium which is represented by the square; four belong to adjoining
ommatidia. The relation of these will be readily seen by referring to
Figure 3.

In passing from the plane in which each cone is surrounded by its
own pair of retinule to the one in which the corresponding hypodermal
square is surrounded by six retinule, the blades of the two retinule
proper to the cone undergo a gradual narrowing ; so that, instead of
each blade covering the whole of one face of the cone, it covers less and
less, and eventually sheathes only one half of the corresponding face of
the hypodermal square. As the blades of the retinulee become narrower,
they expose the surface of the cone, but this is still kept covered by the
retinule of adjoining ommatidia. In any ommatidium there are four
blades which become narrow, consequently there are four regions in
which adjacent retinule touch the cones; and as there is a separate
retinula for each region, it follows that four additional retinule here
come in contact with the cone.

MUSEUM OF COMPARATIVE ZOOLOGY. 1

The distal retinulz touch the cuticula along the band which marks
the boundaries of the facets. That the retinule contribute to the
formation of the cuticula is very improbable, although I believe that
it is largely through their interference that the outlines of the facets
are produced.

The lateral surfaces of each cone are completely enclosed by retinulz ;
the pair of retinulz belonging to the cone play the principal part.
That portion of each retinula which encloses the proximal two thirds of
the cone is densely pigmented (Fig. 1). In transverse sections through
the pigmented region one can see that each cone is completely sur-
rounded by a pigment band. Ona level with the middle of the cone
each retinula contains a nucleus (Figs. 1 and 5, nl. dst.). This is im-
bedded in pigment. The membranes of the retinulz are less distinct in
the pigmented region than near the distal end of the cone. The only
membrane which was observed (Fig. 5) was one which corresponds to
the thickened membrane shown in Figure 4. At the proximal end of
the cone, the retinule rapidly contract till they are reduced to fibres
(Figs. 6 and 7, rin’. dst.). The pigment is present for only a slight
distance below this level. The fibres of the retinule are grouped in
pairs, and in this relation extend to the proximal part of the retina. It
is noticeable that the two fibres which constitute a pair are derived, not
from a single ommatidium, but from two adjacent ommatidia. These
fibres when seen in longitudinal sections were probably mistaken by
Newton (’73, pp. 328, 329) for an investing membrane. At least, in
all attempts to demonstrate the existence of such a membrane I have
failed, and there is so strong a resemblance between the fibres of the
distal retinule and the structure which Newton figured (’73, Plate
XVII. Fig. 15) as the cut edge of an investing membrane, that I am
inclined to think them identical. Transverse sections from the proper
region would have settled the question whether these bodies were fibres
or membranes, but unfortunately Newton has not figured any such
sections.

The pair of fibres in passing from the basal ends of the cones to the
proximal retinule retain the same relative position, and are only slightly
reduced in diameter. (Compare Figs. 7 and 8, rtn’. dst.) Deeper than
this, they are still identifiable, and can be distinguished from the fibres
of the cone-cells by their slightly greater diameter, and by the fact
that they are always in pairs. They lie in the space between four
ommatidia. (Compare Figs. 12, 15, and 18.) Till within a very short
distance of the basement membrane they maintain the condition shown

VOL. xx. — NO. l. 2

18 BULLETIN OF THE

in Figure 18, rtn!. dst. Beyond this I have not been able to trace
_ them with certainty. The groups are no longer observable, and it is
probable that the fibres have separated. I know that in this region the
other cells suffer a very considerable rearrangement ; and such being
the case, it wonld be a very difficult matter to identify single fibres,
especially fibres as small as these are. I have not found any satisfactory
method of staining the fibres so as to distinguish them, as in the case of
the fibrous ends of the cone-cells. I can therefore claim to have traced
the fibres only to within about 20 » from the basement membrane.

As I have already mentioned, the distal face of the basement mem-
brane has cross-shaped thickenings on it. In the angles which the arms
of the cross make with each other, the basement membrane is perforated.
There are consequently around each cross four openings through the
membrane. Hach opening, however, lies between two crosses, so that in
reality only one half of each opening belongs to a given cross, or, if one
counts whole openings only, half of the four openings, i. e. two openings,
belong to each cross. The crosses correspond in number and position
to ommatidia, hence there are also two openings for each ommatidium.
In each opening, beside three or four large fibres which will be described
later, one finds a single small fibre (Fig. 21, rtn’. dst.). That this fibre
represents the continuation of the fibrous end of a distal retinula seems
probable, for two reasons. First, the diameters of this fibre and of the
fibrous part of the distal retinulee are so nearly the same as to be undis-
tinguishable. Secondly, the number of fibres which pass through the
basement membrane, two for each ommatidium, agrees with the number
of distal retinule in each ommatidium. I therefore believe that the
small fibres which are seen in the openings through the basement mem-
brane are the proximal continuations of the fibres of the distal retinule.
If this explanation be true, then it is only natural to expect that, asa
pair of fibres approaches the basement, the individual fibres should
separate, one passing through each opening. As I have already ex-
plained, the fibres, while separated, could be identified only with great
difficulty.

If the fibres pass through the basement membrane, as I believe they
do, they terminate only a short distance below it. For at about 15 yp be-
low the membrane all of the fibres are of nearly the same size, i. e. some-
what larger than the large fibres which pass through the membrane
(Fig. 22). In this region, then, the smaller fibres have either increased
to the size of the larger ones, or diminished till no trace of them is left.
The fibres here are in groups, however, and these are directly continu-

MUSEUM OF COMPARATIVE ZOOLOGY. 19

ous with the groups which pass through the membrane. Lach group of
large fibres as it passes through the membrane consists of either three
or four individual fibres. If the smaller fibres disappear, the groups
below the membrane should consist of three or four fibres also; if, on
the other hand, the smaller fibres increase in size, the deeper groups
should consist of four or five fibres. By either method of change there
would be groups of four fibres, so that it is the groups of three or five
fibres which will be decisive. Asa matter of fact, the fibres are very
commonly in groups of three, and not in groups of five ; consequently I
conclude that the smaller fibres dwindle out a short distance below the
basement membrane.

The distal retinule have not been identified in many Decapods.
Carriere (’85, p. 169) has described them in the eye of Astacus. In
Penzus, Patten (86, p. 634) has observed four cells which belong to
the pigmented collar of the retinophora, ‘Two of these, the inner ones,
evidently correspond to the distal retinule of the lobster. They sur-
round the cones. The other two, the outer ones, appear to have no
homologue in the lobster’s eye.

The Intercellular Spaces of the Retina.

In the region of the retina which lies between the proximal ends of
the cones and the distal border of the deeper band of pigment, the
groups of cone-cells and the pairs of distal retinulze are separated by
considerable intervening space (Fig. 1, spa. ¢ cl.). This space is filled
with a fluid which contains a very small amount of albuminoid sub-
stance. Patten (86, Plate 31, Fig. 73, x) has figured a similar fluid-
filled space in Penzeus. On the application of heat this albuminoid
substance in the lobster coagulates and forms larger or smaller vesicular
bodies, which vary much in size. They are usually loosely attached to
the cone-cells and the fibres of the distal retinule. They readily take
up coloring matter. They have never been observed in fresh retinas
when teased in normal salt solution, nor in maceration-preparations.
It was probably these bodies which Newton (’73, p. 329, Fig. 15, c’)
described as the nuclei on the investing membrane.

In addition to the albuminoid substance which I have described, one
occasionally meets with a thin layer of homogeneous material which
lies slightly in front of the rounded ends of the proximal retinule.
This forms a dividing membrane which separates the retina into a prox-
imal and distal portion. The membrane is of course pierced in many
places. There is an opening in it for each pair of distal retinule, and

20 BULLETIN OF THE

each group of cone-cells. In many cases the membrane has not been
observed. It was noticed by Newton (73, p. 328), and as it is non-
cellular it is probably a feeble representative of what Herrick (’86, p. 44)
has described as a ‘chitinous” framework in the deeper part of the
retina in Alpheus.

The Proximal Retinule.

The proximal retinule are pigment-cells which closely invest the
rhabdome. With the brownish accessory pigment-cells they constitute
the proximal band of brownish black pigment on the distal side of the
basement-membrane (Fig. 26, pig. px.). In some cases they appear to
terminate distally in rounded knobs, each of which contains a nucleus
(Fig. 1, rtn’. px.). In other instances, and these are of frequent occur-
rence, their distal ends, in addition to having a swollen nucleated part,
are prolonged into delicate fibres (Fig. 30). These fibres when present
extend toward the outer surface of the retina, and are applied, not to
the cone-cells, but to the fibrous portion of the distal retinule. The
fibres have been traced only a short distance beyond the rounded ends
of the cells from which they originate. As the region into which they
extend is one readily studied in both sections and maceration prepa-
rations, and as these methods of study have given no evidence of fibres
other than that of the very short ones already mentioned, it seems fair
to conclude that the distal retinulze terminate as fine fibres a short dis-
tance in front of their nucleated portions.

In transverse sections the distal retinule first clearly appear in the
plane represented in Figure 9. Here each group of four cone-cells is
surrounded by a circle of seven retinule. The section from which this
figure was drawn is in a slightly oblique plane. In moving from right
to left, one passes into deeper and deeper regions. In the more super-
ficial part of the section, the right-hand half, each retinula contains a
nucleus, which is surrounded by a small amount of pigmented cell-sub-
stance. In the deeper part of the section, the left-hand half, the plane
is below the region of most of the nuclei, and one sees the seven reti-
nul densely filled with pigment. In the next section (Fig. 10), the
retinule are broader in transverse section. In their expansion they
have so far encroached on the space which they surround that it is
only large enough to allow the passage of the four cone-cells. The con-
traction of the space within the circle of retinule takes place almost
in one plane, as can be seen in the longitudinal section (Fig. 1). In the
plane of Figure 10, the retinule show a tendency to group themselves.

—

MUSEUM OF COMPARATIVE ZOOLOGY. 2h

One can recognize an odd, usually larger retinula, which occupies the
lower right-hand corner of each group. The remaining six retinule
are disposed in pairs. In Figure 11, which represents a plane of section
deeper than that shown in Figure 10, the retinule, although somewhat
reduced in thickness, present nevertheless the same method of group-
ing as was pointed out in Figure 10. In this plane one also notices
next the odd retinula, a nucleus. This is remarkably constant in its
occurrence, both as to position and as to the fact that there is always a
single nucleus. When compared with the nuclei of the surrounding retin-
ul, it is found to resemble them very closely. The nuclei of the seven
retinule are characterized by their sharply marked oval outlines, and
by the possession of one or two very distinct nucleoli (Fig. 30, nl. pz.).
In both of these respects the single nucleus agrees so closely with the
nuclei of the retinule, that, were it not for its somewhat smaller size
and deeper position, it could not be distinguished from them. The
regularity with which it occurs, and its structural peculiarities, incline
me to believe that it represents a reduced retinula in which pigment
has never been developed. This belief is further supported by the fact,
that the additional nucleus is always found next the larger retinula,
which from its great size seems to have replaced a second cell. It is
therefore probable that each ommatidium of the lobster’s eye possesses
eight proximal retinule rather than seven, and that one of these is
rudimentary.

Below this additional nucleus, the proximal retinule pass around the
rhabdome. In this region they are deeply pigmented, and so completely
envelop the rhabdome that I am of opinion that no appreciable amount
of light gains access to it except through the cone-cells. The cone-cells,
it will be remembered, extend through the central region of each group
of proximal retinulz, until they almost impinge on the distal end of
the rhabdome. Thus, by excluding the retinule, they form a trans-
parent shaft, which leads to the distal tip of the rhabdome, and by
which that structure can receive light. The rhabdome in transverse
section has a four-sided outline. Three sides of the rhabdome are occu-
pied each by a pair of retinule. These pairs are composed of the same
couples which were previously noticed. (Compare Figs. 10 and 14.)
The fourth side of the rhabdome is occupied by the seventh or odd
retinula. Thus, again, it is noticeable that this retinula occupies a
position where, if perfect symmetry were shown, we should expect two
retinule. The rhabdome in transverse section is broadest midway of
its length. In this position the retinule are small, as if closely pressed

22 BULLETIN OF THE

against its sides (Fig. 14). Below its middle transverse plane the
rhabdome becomes gradually smaller and smaller, till finally it termi-
nates about 15 uw from the basement membrane. As the rhabdome con-
tracts in size, the retinulz enlarge. (Compare Figs. 14 and 17.)

As I have already mentioned, the retinule are definitely arranged
around the rhabdome, and this arrangement persists nearly to its
proximal termination ; but between the end of the rhabdome and the
basement membrane the retinule rearrange themselves. This re-
arrangement of the retinule is a step preparatory to their passing
through the apertures in the basement membrane, the general struc-
ture of which has already been described. It will be remembered that
under each ommatidium the distal face of the membrane presents a
cross-shaped thickening, and that in each of the four angles which the
arms of the cross make with each other there is an opening (Fig. 21).
The openings are oval in outline, especially on the distal face of the
membrane. One end of a given oval lies in the angle of the cross,
and the crosses are so close to each other that the other end of the
same oval lies in the angle of a neighboring cross. Each opening then
lies in the angles of two adjoining crosses, and through it pass two
groups of retinule, one from each of the two ommatidia to which the
crosses correspond.

The four groups into which the retinule of a single ommatidium
are divided pass one through each of the four surrounding apertures.
Three of the groups consist of pairs of retinulze; the fourth group is
represented by only a single retinula. Although these groups agree
numerically with the groups of retinule, which, as I have already
shown, surround the rhabdome, they are not composed of the same
individual retinule.

For convenience of comparison, numbers can be assigned to the dif
ferent retinule. This has been done in ommatidium c (Fig. 15), where
the large odd retinula is numbered 1, and the remaining retinule, pro-
ceeding ina circle to the left, are numbered 2 to 7. On this plan of
numbering, the four groups of retinule which have been already indi-
cated as surrounding the rhabdome are composed as follows. What
may be called the first group is formed of retinule 2 and 3, the second
group contains retinule 4 and 5, the third retinule 6 and 7, and the
fourth retinula 1. It will be observed (Fig. 15) that the seven retinule
are also divided into four other groups by the fibres of the cone-cells.
Three of these groups are composed of pairs of retinule, the individuals
of which lie nearest the angles of the rhabdome. In Figure 15, omma-

-. ————— es

MUSEUM OF COMPARATIVE ZOOLOGY. 23

tidium c, these taree groups are represented by retinule 1 and 2, 3 and
4,and 5 and 6. The fourth group is composed of the single retinula
number 7. |

The four groups thus defined are identical with the ones which pass
through the four openings in the basement membrane. In F igure 21
the four openings which surround the cross («) can be designated from
their positions as the upper, lower, right-hand, and left-hand openings.
The upper and lower openings each present the transverse section of
four large fibres. The right- and left-hand openings are each occupied
by three large fibres. The source of these fibres can be ascertained by
comparing Figures 20 and 21. In ommatidium ¢ (Fig. 20) retinule
5 and 6 unite with retinule 1 and 2 of ommatidium d, and thus consti-
tute the four fibres which pass through the upper aperture (Fig. 21).
In a similar way, retinule 1 and 2 of ommatidium c unite with 5 and 6
of an ommatidium which lies below ¢, and pass as four fibres through
the lower opening (Fig. 21). Retinule 3 and 4 of ommatidium ¢
(Fig. 20) unite with retinula 7 of an ommatidium to the left of c, and
emerge as three fibres through the left-hand opening (Fig. 21). Retin-
ule 7 of c, and 3 and 4 of 6 (Fig. 20), unite and form the three large
fibres of the right-hand opening (Fig. 21). This plan of distribution is
repeated in each ommatidium, and thus brings about the groups of
three or four fibres which occur in each opening through the basement
membrane. The groups of fibres are distinguishable for only a very
short distance below the basement membrane. The individual fibres of
each group soon separate, and in the deeper part of the optic nerve they
never again present this grouping. The description of the termination
of the fibres will be deferred to a later part of this paper.

The relation of the rhabdome to the cone-cells in Homarus has al-
ready been described. That they are separate structures, as Schultze
and Grenacher have asserted, I believe there can be no doubt. I have
seen nothing which favors the view held by Patten, namely, that the
rhabdome is an enlargement in the proximal part of the cone-cells. In
Homarus the rhabdome has the general form of a spindle. In trans-
verse section, however, it is not circular, but square. Its four sides are
thrown into ridges, the crests of which extend across the sides at right
angles to its longest axis. The inner face of each proximal retinula is
thrown into corresponding undulations. The rhabdome and retinula
are so adjusted to each other that a crest on one fits into a furrow
on the other. (For a similar condition in Penzeus, compare Patten,
°86, Fig. 72.) The retinule and rhabdome are thus intimately bound

24 BULLETIN OF THE

together. On comparing opposite faces of the rhabdome, it will be seen
that the crests of one side are in the same horizontal plane as those of
the other ; but on comparing adjoining faces, it will be observed that the
crests of one correspond to the furrows of the other.

In the rhabdome of the lobster I have not found a complicated system
of plates, such as Patten describes in Penzus. The substance of the
rhabdome in the fresh condition is apparently homogeneous, but in har-
dened preparations it is finely granular and stratified. The strata are
at right angles to the long axis of the rhabdome, and the rhabdomes
often break transversely. The stratified condition of the rhabdome, and
the close relation which the proximal retinule bear to it, support the
conclusion that the rhabdome is the product of the proximal retinule.

In the structure of the rhabdome there is one peculiarity which,
although I cannot explain it, requires some comment. In transverse sec-
tions the square area of the rhabdome is often divided into four smaller
squares by two intersecting lines (Figs. 13 and 43). I made this obser-
vation before I had studied the relation of the distal tip of the rhab-
dome to the cone-cells, and I concluded then, that, if Patten was correct
in believing that the cone-cells and rhabdome were continuous, these
four divisions of the rhabdome must correspond to the four cone-cells.
I recognized the fact, however, that, if such was the case, the group of
cone-cells in the region of the rhabdome was turned through an angle of
45° as compared with its position at the surface of the retina (contrast
Figs. 4 and 13). After having satisfied myself that the cone-cells and
rhabdome were separate structures, I was forced to the conclusion, that
the four segments of the rhabdome were independent of the four cone-
cells. That there can be no question of the independence of these two
structures is shown by the condition of the cone-cells and rhabdome
in Mysis. In all Decapods, so far as I am aware, the cone is formed of
four cone-cells, and the rhabdome has four segments; in Mysis, how-
ever, the cone is formed of only two cells, although the rhabdome has
four segments.!

I can offer no explanation of the cross lines which occur in the rhab-
dome. As Grenacher (’79, p. 124) has observed, one might at first take
them for the outlines of such parts of the rhabdome as were produced by
individual retinule. There are, however, seven retinule, and only four
segments. Not only do the numbers disagree, but the position of the
lines in the rhabdome is difficult to explain. If the lines are related
to the retinule, it would be natural to expect that they would coincide

1 My attention was called to this fact by my friend, Mr. H. H. Field.

MUSEUM OF COMPARATIVE ZOOLOGY. 25

with the junction of two retinule. Asa matter of fact, three lines do
come between retinulz, but the fourth one abuts against the middle of
the large odd retinula (Fig. 34). This relation of the cross-lines and
retinule persists from the earliest stage in the production of the rhab-
dome, and although the lines are doubtless formed during this process
they show a strange independence of the retinule.

An exactly similar relation between the cross-lines and retinule has
been described by Grenacher (79, p. 124) in Palemon. I cannot
agree with Grenacher in believing that the four lines are due to the
fact that only four retinule are concerned in the production of the
rhabdome. The relations of the retinule and rhabdome are the same in
the young lobster (Fig. 58) in which the rhabdome is being produced
as in the adult. This fact was not known to Grenacher. It shows, I
believe, that the rhabdome is the product of the surrounding seven reti-
nule, and that the problematic lines have some other significance than
that of indicating regions of production.

The Accessory Pigment-cells.

These cells occupy the open space at the base of the ommatidia.
They are characterized by possessing a pigment which, as I have before
stated, is brownish by transmitted and whitish by reflected light. The
cells are bounded proximally by the basement membrane, and their
distal ends rarely reach beyond the middle of the rhabdomes (Fig. 1).
They are extremely irregular in form, and seem to fit themselves’ to a
cavity of almost any shape. Their function seems to be that of filling
what would be otherwise an unoccupied space, as though to lend solid-
ity to the tissue in the base of the retina. (Compare Figs. 13 and 16.)
Their nuclei are irregular in form and size (Fig. 18, n/. pig.). Judging
from the number of nuclei, two or three cells are associated with each
ommatidium. The number, however, is variable. The physical prop-
erties of the pigment which these cells contain are very characteristic.

Streaks of this pigment, and even whole cells, are to be met with in

the open space on the proximal side of the basement membrane.

Cells similar to those which I have called accessory pigment-cells have
been described by Carriére (’85, p. 169) in Astacus, and by Patten (’86,
p- 636) in Peneus. It is highly probable that the yellow pigment
which Grenacher (’79, Fig. 114) figured in the base of the -retina of
Mysis represents accessory pigment-cells as well as the dark pigment
which he described (’79, p. 124, Fig. 117) in Paleemon.

ect bates

—o > ee

Pitas, Se 2

—

we 6s

- “hes <Seeaeeet oC oe

26 BULLETIN OF THE

The Innervation of the Retina.

The study of the termination of the nerves in the retina is of partic-
ular importance, since it affords a means of identifying the perceptive
elements. Wherever these elements may be, the ultimate branches of
the nerve-fibres must unquestionably lead to them; hence the impor-
tance of discovering the termination of the nerve-fibres.

Students who have investigated the compound eyes of Arthropods
have held two opinions as to the position in which the nerve-fibres ter-
minate. One school has maintained that the fibres terminate in the
crystalline cones, and that therefore these bodies are the perceptive
elements. The other school has endeavored to show that the. fibres
end in the region of the rhabdome, and that for this reason the rhab-
dome is the perceptive element. I shall not attempt to give an his-
torical account of this subject, but only call attention to the fact, that
of late years the majority of writers have expressed the opinion that the
rhabdome is the perceptive element, and that the cone is merely dioptric
in function. This conclusion has been recently criticised by Patten,
who believes the cone to be the perceptive body.

Many of Patten’s statements are based upon observations which were
possible only by his methods of investigation. On this account, as well
as for the reason that his paper is the most important recent contribu-
tion to the study of nerve-termination in compound eyes, I shall not
refer to the older publications, but limit myself to what he has pre-
sented in his article on “‘ Eyes of Molluscs and Arthropods,” and to such
papers as have appeared since the publication of that work.

When comparing Patten’s statements on nerve-termination with
those of other recent investigators, I was inclined to believe, since
Patten used new and probably better methods of study, that his results
were more trustworthy. The contrast between his views and those
which are more generally accepted is so striking, however, that in begin-
ning a study of the nerve-terminations the first step to be taken was
necessarily one of confirmation. I was unable to obtain the same
species of Crustaceans as Patten had used, but I believe that I was safe
in assuming that the difference which may exist between the innerva-
tion of the retina in Penzeus and Homarus could not be a fundamental
one, and that the more important feature which had been demonstrated
in one genus could be shown in the other. I consequently prepared
sections of the retina of Homarus according to the methods which Patten
had recommended, and although I was careful in both preparation and

MUSEUM OF COMPARATIVE ZOOLOGY. 27

examination of these sections, I was entirely unsuccessful in discovering
any trace of nerve-fibres such as Patten had described. I therefore
resolved to try his methods of maceration. ‘This I did, following closely
his directions as to the strength of solutions, length of time during
which reagents should be employed, etc., but my results were again
negative. Still adhering to the reagents which he had recommended,
especially chromic and sulphuric acids, I varied the time during which
the eyes were treated, hoping thereby to obtain a combination more
favorable for Homarus. The separation of the elements of the retina
was often very successful, but I never saw in any of my preparations
systems of nerve-fibres which resembled those figured by Patten. In
almost all cases the isolated parts of the retina presented many delicate
fibrous projections. These projections might be interpreted as shreds
of broken nerve-fibres, although in no case did they show a systematic
arrangement. Moreover, they were found on all kinds of tissue. For
these two reasons, I believe that they were not broken nerve-fibres, but
simply shreds of tissue. The substance of the cones was finely granular,
and was never penetrated, so far as I could discover, by any fibres. My
results from both sections and isolation preparations were invariably
negative ; and as my observations had been made upon somewhat over
sixty lobsters’ eyes, I concluded that in Homarus there was no evidence
in favor of the method of nerve-termination which Patten had described.
As I have previously mentioned, it is highly improbable that the
methods of innervation in the retinas of Penzeus and Homarus are fun-
damentally different, and since I have found in the retina of Homarus
no confirmation of Patten’s views, I am of opinion that he must be mis-
taken as to the method of nerve-termination in Penzeus. Many of
Patten’s figures of the individual nerve-fibres in Penzus (86, Plate 31,
Figs. 69, 70, 71) resemble so closely the fibres which I have seen in all
of my isolation preparations, and which, for reasons already given, I am
persuaded are not nervous, that I am forced to believe that Patten has
mistaken for nerve-fibres shreds of non-nervous tissue.

My criticism of Patten’s results refers only to those which he obtained
from a study of the Crustacea. No one, so far as I am aware, has fully
confirmed his views concerning the nerve-terminations in this group.
Kingsley (86, p. 864) claims to have seen the axial nerve-fibre in
Crangon, but he was unable to trace its finer ramifications in the cone.
He states (’87, p. 57), however, that the method of preparation which
he employed was not intended especially for the fibres, and that there-
fore it is not surprising that they were not identifiable. Herrick

28 BULLETIN OF THE

(89, p. 168) could not distinguish fibres in the cones of Alpheus, al-
though, like Kingsley, he admits that his failure may be due to the
method which he used. Relative to nerve-terminations, the evidence
which the work of these two investigators presents can scarcely be called
critical, and I therefore hold to my former conclusion, namely, that the
fibres which according to Patten represent the nerve-terminations are
in reality not nerve-fibres at all.

After having reached this conclusion, I was naturally led to look
elsewhere for the true nerve-fibres. It occurred to me that, in order to
be certain that I was dealing with nerve-fibres, it was safer to begin
studying them in regions where their identity was beyond question. I
therefore examined maceration-preparations of parts of the larger nerves
from the lobster’s body. These nerves were readily resolved into a
number of fibres, which in transverse section were enormous when
compared with such fibres as the axial nerve-fibre figured by Patten
(’86, Plate 31, Figs. 72, 74, 108). I then studied in a similar way the
optic nerve. The fibres in this nerve were smaller than those from
the other nerves which I had macerated, but they were much larger
than those figured by Patten. Each fibre (Fig. 36) possessed a distinct
sheath, and its contents were marked by lines which extended parallel
to its long axis. These lines I interpreted as indications of fibrillee
which composed the fibre. In addition to the large fibres, the optic
nerve, as well as the other nerves, showed, when macerated, the fibrous
shreds which I have previously mentioned. They were very insignifi-
cant in amount, forming, I should judge, not more than a fraction of one
per cent of the whole optic nerve, and I was never able to trace them
as continuous fibres for any considerable distance. I believe that here,
as in the retina, they arose from an artificial tearing of the tissue.

At about this time I happened to find the modification of Weigert’s
method of straining nerve-fibres, which I have described in the Intro-
duction. As soon as I was aware of the results which could be obtained
by this method, I applied it to a series of transverse sections of the
optic nerve. The series extended from the retina to the optic ganglion,
and demonstrated conclusively, I think, that.the proximal ends of the
seven proximal retinule, after passing through the basement membrane,
as I have described, continued inward as fibres, and finally passed into
the substance of the optic ganglion. Figures 21, 22, 23, 24, and 25
illustrate steps in the passage from the retina to the ganglion. In Fig-
ure 21 the groups of three or four proximal retinule are seen as they pass
through the basement membrane. Figure 22 is taken at a level imme-

7
f

MUSEUM OF COMPARATIVE ZOOLOGY. 29

diately below the basement membrane. Here two groups of four fibres,
and one of three, can be distinguished. Below this level the groups of
three and four fibres are no longer to be recognized. The fibres diminish
rapidly in diameter as they leave the retina. Figure 23 represents a
transverse section of the fibres at one fourth the distance from the base-
ment membrane to the distal face of the ganglion. Figure 24 represents
a similar section midway between retina and ganglion. Figure 25 repre-
sents the fibres as they enter the ganglion. The same kind of fibres
have been identified in the substance of the ganglion. These fibres, I
believe, are the true fibres of the optic nerve, and, as I have shown,
they connect the seven proximal retinule of each ommatidium with the
optic ganglion.

The termination of the nerve-fibres in the proximal retinule is so
directly opposed to the method of termination which Patten has de-
scribed, that, before making a final statement based on a study of the
lobster only, it seemed prudent to seek confirmation in other species.
This I have done, and I can now state that the termination of the nerve-
fibres in the retinule has been demonstrated in Eupagurus, Cambarus,
and Gammarus. From this I conclude that it is highly probably that
in the compound eyes of all Crustacea the nerve-fibres terminate in the
retinule.

This method of nerve-termination, namely, the direct continuity of
the nerve-fibre and the perceptive cell, has also been demonstrated by
Watase (’89, pp. 34-36) in the eyes of Limulus.

There are some interesting phases in the transition from the nerve-
fibre to the retinula. On examining the transverse sections of nerve-
fibres at a level slightly below the basement membrane, one observes
that it is in the form of a transparerit cylinder the periphery of which
has scattered over it a few pigment granules. As the fibre passes
through the basement membrane the pigment increases in quantity, and
when the retinula is reached the nerve-fibre is represented by a trans-
parent axis in its centre (Fig. 19, ax. .). The nervous axis is transpar-
ent, because it contains no pigment granules, while the peripheral portion
of the retinula is densely pigmented. In this way the nerve-fibre proper

is continuous as a transparent axial shaft from below the basement

membrane to a level in the retina, which corresponds to the middle of
the rhabdome. From this level to its distal end the retinula is com-
pletely filled with pigment, no trace of a transparent central axis being
visible. (Compare Figs. 19 and 14 with Figs. 11, 10, and 9.) The trans-
parent nervous axis of each retinula terminates, as I have said before,

30 BULLETIN OF THE

on a level with the middle of the rhabdome. When it is cut very close
to its distal end, the nervous axis is seen to lie almost next the rhab-
dome. I have never seen a case, however, in which the axis and rhab-
dome were not separated by a line of pigment granules. In transverse
sections which have been thoroughly depigmented, the distal half of
each rhabdome is surrounded by a great number of bodies which resem-
ble coarse granules (Fig. 32, for!.). These bodies are limited almost
entirely to the distal half of the rhabdome. At first I supposed that they
were the colorless skeletons of pigment granules, but they were easily
distinguished from the latter by their larger size and sharper outline.
It then occurred to me that, since these bodies were found only about
that portion of the rhabdome which was distal to the nervous axis, they
might therefore be the cut ends of the finer fibrille into which that
axis had been resolved. If such was the case, these bodies were in re-
ality fibres, not granules. In order to determine whether they were
fibres or granules, I examined oblique sections of the rhabdome. Fig-
ure 33 is taken from one of these sections. The granular body in the
figure is the rhabdome, and the transverse bands are the strata in its
substance. The projection of the long axis of the rhabdome in this
figure would be a vertical line. The lower end of the figure is proximal,
the upper end distal. Owing to the fact that the rhabdome is cut
obliquely, what is seen at its proximal end belongs on one side of it,

_ and what is seen at its distal end belongs on the opposite side. The —

first proximal dark band in the substance of the rhabdome is very
nearly midway between its ends. At the proximal end of the rhabdome
three distinct fibre-like bodies are seen. ‘These are in reality longitu-
dinal sections of the greatly compressed retinule, which have been
noticed in transverse sections. (Compare Figs. 15 and 18.) At the
distal end of the rhabdome, in place of the three retinule, there are a
great number of fine fibres. The fibres occur only around the distal
half of the rhabdome, and I believe that in transverse sections these
fibres are represented by the small round bodies previously described.
From the distribution of these small fibres and their relation to the
distal end of the nervous axis, I am of opinion that they represent the
fibrillze of the nerve-fibre.

The entrance of the fibres of the optic nerve into the proximal reti-
nulz is of itself strong evidence that the rhabdome, not the cone, is the
perceptive body. This conclusion is further supported by the ultimate
distribution of the optic fibrille. If it be admitted that the rhab-
dome is immediately concerned in the perception of light, it is only

SL Se ee

MUSEUM OF COMPARATIVE ZOOLOGY. ok

natural to expect that this structure would be accessible to the light.
As I have already shown, the cone-cells form a transparent axis, which
leads directly to the rhabdome, and through which light could readily
reach that structure. Once having penetrated to the rhabdome, it is
probable that, either in the substance of the rhabdome itself, or in the
superficial layer of pigment which immediately surrounds the rhabdome
and in which the fibrille are, the light is transformed into that kind of
energy which is transmitted by nerve-fibres.

THE DEVELOPMENT.

In discussing the development of the eyes in Arthropods, the more
recent investigators have given much attention to the general plan of
these organs. One of the objects of their researches has been the
reduction of the eyes of these animals to a single structural type. If
such a type were found to exist, it would very probably reproduce the
essential structural feature which the eye of the ancestral Arthropod
possessed. The desirability of ascertaing if there is a common type for
the eyes of all Arthropods is evident, for upon the nature of the answer
to this question must depend to some extent the conclusions concern-
ing the phylogenetic relationship of the group, and its different classes,
Possibly in the phylogeny of two classes of Arthropods the eyes may
have originated independently. Of course organs independently devel-
oped could not be homologous, and they might be so differently con-
structed that it would be impossible to reduce them to a common type.
Notwithstanding the difficulty in homologizing the eyes of one class of
Arthropods with those of another, the homologies among the members
of a single class are much more readily determined, and many impor-
tant comparisons can be safely made. It would, therefore, seem more
prudent to limit investigations to the eyes of a single class until they
were well understood, rather than institute comparisons between tbe
eyes of different classes, where, from the limitations of our knowledge,
such comparisons must be more or less hazardous.

The Plan of the Eye.

The work of different investigators has already suggested several
structural types for the compound eyes of Arthropods, but the differ-
ences which some of these types present are of such a fundamental
character, that to accept one is to reject another. It was with the hope
of gaining confirmation for some one type that I was led to study the

‘

32 BULLETIN OF THE

development of the eye in a Crustacean, and, as I have previously ex-
plained, the species most available for such a study was the lobster.

The results which have been presented in the more recent papers on
the embryology of the compound eyes require a brief notice before the
» development of the eye in the lobster is described. What is said in this
connection is purely introductory ; no criticism of the views of different
authors will be made until after the development of the lobster’s eye
has been described.

The writers who have thus far published accounts of the development
of compound eyes can be grouped under four heads, depending upon the
type of eye which their researches indicate. The first type is that
represented by the eye of Peripatus. Patten (’86, p. 688, ’87, p. 211)
is of the opinion that the compound eyes of Hexapods, as well as Crus-
taceans, are constructed upon this plan. Each eye should then consist
of a closed vesicle which was produced by an involution of the hypo-
dermis. The eye would be composed of three layers, which in the
order of their positions are as follows: first, the superficial hypodermis ;
second, the outer wall of the vesicle; and third, the inner wall of the
vesicle. Patten is of opinion that in the eye of an adult individual
these three layers are modified in the following way: the superficial
hypodermis becomes the corneal hypodermis; the outer wall of the
vesicle is so far reduced as to be inconspicuous, and the inner wall gives
rise to the retina. The retina includes the crystalline cones and the
pedicels (rhabdomes). Patten supports these conclusions mainly from
theoretical grounds, but he believes that he has found evidence of the
existence of this type in the development of the compound eyes of
Vespa, Blatta, and the Phryganids.

The second structural type which I shall mention has been advocated
by Kingsley (87, p. 51) in his description of the development of the
compound eye of Crangon.’ In this type the eye results from a vesicu-
lar infolding, as in that which was proposed by Patten, but it differs
from the latter in the fate which is ascribed to the two walls of the
vesicle. According to Kingsley, the outer wall of the vesicle is not
reduced, but gives rise to the retina. In it are developed the crystal-
line cones and the pedicels (rhabdomes). The inner wall of the vesicle,
instead of forming the retina, as Patten believed, is converted into a part
of the optic ganglion.

The third structural type is that which is presented by Reichenbach
(’86, pp. 85-96) in his account of the development of the crayfish.

1 See the note on page 41.

MUSEUM OF COMPARATIVE ZOOLOGY. =]

Here a hypodermal involution takes place, and a vesicle is produced,
but the cavity of the vesicle is soon obliterated. ‘The mass of cells
which results from the fusion of the walls of the vesicle now divides into
an outer and an inner layer. ‘These two layers do not necessarily cor-
respond to the outer and inner walls of the original vesicle. The super-
ficial hypodermis and outer layer of the infolded mass fuse, and give
rise to the retina. The crystalline cones are developed in that part of
the retina which is derived from the superficial hypodermis ; the rhab-
domes probably originate in that part which is derived from the outer
layer of infolded cells; the inner layer of cells becomes ganglionic.

Bobretsky’s (73) account of the development of the compound eyes
in Astacus and Palemon agrees in its essential features with the de-
scription which Reichenbach has given for the eyes in the crayfish.
Bobretsky, however, does not describe an involution in the optic disks.
As Reichenbach suggests, the fact that Bobretsky did not have the oppor-
tunity of studying very early stages may explain his failure to observe
the involution. In other respects, the accounts are essentially alike, and
there is little doubt that the plan of eye which Bobretsky’s researches
indicate is the same as that suggested by Reichenbach’s studies.

Each of the three structural types which have thus far been described
are dependent upon the formation of a hypodermal vesicle. The fourth
type is simpler than any of the three preceding ones, in that a vesicle is
not necessarily produced, the retina being supposed to originate as a
simple thickening in the superficial hypodermis. This type has been
advocated by Herrick (’86, p. 43) in his account of the development
of Alpheus. The researches of Nusbaum (’87, pp. 171-186) on the
development of Mysis indicate the same type. Neither in Grobben’s
(79) account of the development of the eyes in Moina, nor in Claus’s
(86, pp. 307-324) description of those in Branchipus and Artemia is
any mention made of an involution. These authors might, therefore,
be cited as favoring the fourth type of eye, although it is to be observed
that the special question of the vesicular origin of the eye is not dis-
cussed by them.

The advocates of the fourth type find support, not only in the embry-
ology of the Crustacea, but also in that of the Hexapods. According to
Carriere (’85, pp. 181-186), the compound eyes of some Hymenoptera
and Lepidoptera develop as simple thickenings of the hypodermis.

Of the four types which have been mentioned, the one with which
the development of the lobster’s eye accords will be seen from the
following description.

VOL. XX.—NO. 1. 3

34 BULLETIN OF THE

The first traces of the eyes in the development of the lobster are the
optic disks. These disks lie one on either side of the median plane, and
are for a considerable time the most conspicuous structures in the an-
terior region of the embryo. In the early stages of development the
disks face ventrally, but as the head of the animal becomes differen-
tiated they come to face almost in the opposite direction, i. e. dorsally.
At stage A (Fig. 37) the optic disk, so called, is oval in outline rather
than disk-shaped; its longer axis is transverse to the principal axis of
the embryo. From that portion of each disk which is near the median
plane a band of tissue extends posteriorly, and connects the disk with
the ventral plate of the embryo. The disk and the band with which it
is connected to the ventral plate are distinguished from the surrounding
tissue by their greater number of nuclei. The disks comprise the tissue
from which both retina and optic ganglia develop.

A section passing through an optic disk in a plane perpendicular to
the longitudinal axis of the embryo is shown in Figure 38. In this
case the section is from the right optic disk. The left side of the sec-
tion is farthest from the median plane; the right side is near that
plane. Since at this stage the disk faces ventrally, and since the ventral
edge of the section is uppermost in the figure, it is the posterior face of
the section which is presented. To the right and left of the disk one
can see the undifferentiated ectoderm with its occasional nuclei. This
ectoderm is directly continuous with the tissue of the disk; in fact, the
disk is only a local thickening in the ventral ectodermic layer of the
embryo. It is due to its greater thickness that the disk contains more
nuclei than the surrounding ectoderm. ‘The deep face of the ectoderm,
both of that which is undifferentiated and that which forms the disk, is
limited by a delicate but distinct basement membrane (Fig. 38, mbd.).

Some idea of the method of growth in the optic disk can be gained
from a study of its nuclei. It will be noticed that the nuclei in the
deeper part of the disk are rather small and irregularly grouped when
compared with those which are found next its outer surface. These
superficial nuclei form an almost regular series, which extends from one
margin of the disk to the other. They are usually also characterized by
having their long axes perpendicular to the surface of the disk. When
they increase by division, their planes of separation are in most cases
either parallel or perpendicular to the outer surface of the disk. When
the plane of separation is perpendicular to the surface of the disk, the
tendency for such divisions is to increase the diameter of the disk.
When, on the other hand, the plane of separation is parallel to the

= MUSEUM OF COMPARATIVE ZOOLOGY. 30

surface of the disk, the tendency is to thicken it. In order to determine
the distribution of these two methods of division, the planes of sepa-
ration in the superficial nuclei of four disks were carefully observed.
Each disk can be divided into halves by a plane parallel to the sagittal
plane of the embryo. That half of the disk which lies near the median
plane can be called the proximal half; that which is farther from the
median plane, the distal half. Among the superficial nuclei of the distal
halves of the four disks which were examined there were seen thirteen
cases of division. In all of these cases the tendency was to broaden the
disk. There was no case where the division of the nucleus would have
thickened it. In the proximal halves of the disk there were six cases
of division observed ; five of these tended to thicken the disk, and one
would have broadened it. It is therefore apparent that in the super-
ficial layer of each disk the proximal half is becoming thicker, while the
distal half is becoming broader.

The deep nuclei of each disk, those lying below the row of superficial
nuclei, divide in different planes. Among these nuclei in the four disks
which were examined, seven instances of division were noticed. Five of
these were in planes which would have thickened the disk ; the remain-
ing two would have broadened it. This part of the disk consequently
shows a tendency both to become broader and thicker. Of these two
tendencies, that which would thicken the disk is the stronger.

The method of growth which has been described for different parts of
the optic disk can be easily distinguished only in its earlier stages. The
subsequent changes which affect the structure of the disk render it
rather difficult to follow in detail the growth of the disk as a whole; but
in general the broadening of the distal superficial region, the thickening
of the proximal superficial part, and the broadening and thickening of
the deeper parts are continued.

The most important changes in the differentiation of the optic disks
are the following: first, the separation of the retina and optic ganglion
by the formation of the basement membrane; second, the production of
the optic nerve ; and third, the differentiation of the ommatidia. These
three changes will be consideréd in the order named.

The first step in the differentiation of the retina and optic ganglion
occurs at stage B. Figure 39 represents a section from the optic disk of
an embryo of this stage. The plane of section in this case corresponds
to that of Figure 38. The chief difference observable between the disk
at stages A and B 1s its greater thickness in the older embryo. Not only
has the disk thickened, but it has spread laterally, so that the small

36 BULLETIN OF THE

angle which is seen on the outer margin of the disk in stage A (Fig.
38, x), and which indicates a tendency on the part of the disk to grow
over the adjacent ectoderm, is represented in stage B by a very acute
angle (Fig. 39, x), while the disk itself forms an actual fold, which
covers a portion of the undifferentiated ectoderm.

The separation of the retina and its ganglion is accomplished by the
production of a basement membrane. This is gradually developed in
the substance of the disk between the regions of the retina and the optic
ganglion. In order to distinguish the newly formed membrane from
the original basement membrane which bounds the under surface of the
disk, I shall speak of the former as the intercepting membrane. The in-
tercepting membrane (Fig. 39, mb. 2 cpt.) takes its origin from the base-
ment membrane on a line a little within the lateral edge of the disk.
From this position it extends at this stage as a delicate lamella for a
short distance through the tissue of the disk. The direction of its course
is approximately parallel to the outer surface of the disk, and it divides
the distal portion of the disk into two masses, one of which is superficial,
the other deep. ‘The superficial mass is the first portion of the retina
to be differentiated, and, as I have previously stated, the growth in this
region is chiefly lateral. The deep mass is the beginning of the optic
ganglion. The intercepting membrane does not extend so far as to
separate the superficial portion of the disk from the deeper part in the
proximal half. This condition is what one might expect, since in the
proximal region of the disk the superficial nuclei divide in such planes
as to thicken the disk, and a membrane which would separate the deep
and superficial parts would be a source of interference in the process of
thickening.

The intercepting membrane is not an involution of the original base-
ment membrane, but is produced by the activity of the ectodermic cells
between which it lies. There is no reason for doubting, I believe, that
at this early stage (B) it is strictly an ectodermic secretion. At this
stage it is fan-shaped, the handle of the fan being formed by that part of
the membrane which is in contact with the original basement membrane.
The plane of the fan is parallel with the outer surface of the disk. In
sections which are either anterior or posterior, but parallel to that shown
in Figure 39, portions of the intercepting membrane are often visible,
and may be apparently unconnected with the basement membrane.
This is due to the fact that the plane of section cuts the anterior or
posterior edge of the fan without including the handle.

The chief difference between the intercepting membranes at stages

MUSEUM OF COMPARATIVE ZOOLOGY. 37

B and C is the greater extension of the membrane at stage C. (Com-
pare Figs. 39 and 41.) The retina is now much more completely
separated from the ganglion than formerly. The superficial and deep
portions of the proximal part of the disk are, however, still continuous.

In stage D (Fig. 45) an important step is taken in the development
of the membrane. It splits at this stage into two layers, one of which
adheres to the retina, and the other to the optic ganglion. The retinal
and ganglionic layers of the membrane have probably been distinct from
the time of their formation; but until stage D is reached, the double
nature of the membrane is not apparent. Occasionally in stage C one
notices at the point where the intercepting and basement membranes
unite a re-entrant angle (Fig. +l, 2). This in some cases extends a
short distance between retina and ganglion, and doubtless represents the
first step in the separation of the components which form the intercept-
ing membrane.

In stage E (Fig. 46) the membrane has completely severed the super-
ficial from the deep ectoderm, and the separation of its retinal and gan-
glionic constituents extends over a broader area than in the previous
stage.

The subsequent changes in the intercepting membrane consist in a
complete separation of its retinal and ganglionic portions. This separa-
tion is effected by the withdrawal of the optic ganglion from the super-
ficial ectoderm. In the adult, the ganglionic membrane remains relatively
thin, but the retinal membrane becomes much thickened. This mem-
brane, which has already been described as the basement membrane of
the adult retina, is not uniformly thickened, but presents local eleva-
tions, each of which is in the form of a cross. The four apertures which
pierce the membrane in the angles of the cross-shaped elevations, and
the relation which adjoining crosses and apertures bear to one another,
have already been described. The proximal face of the basement mem-
brane is nearly flat (Fig. 29). The cross-shaped elevations occur on its
distal face. The substance of the membrane is apparently homogeneous,
and contains no traces of cells or nuclei. The fact that its substance is
alike throughout favors the idea that it has been derived from a single
source. From stage E to the adult condition a few mesodermic cells
have been noticed next. its proximal face (Fig. 1, cl. ms d.). These cells
are not intimately attached to it, and I am of opinion that they con-
tribute little or nothing to its composition.

From the foregoing account I draw the following conclusions concern-
ing the growth of the basement membrane of the eye. In its earliest

38 BULLETIN OF THE

stages the basement membrane is strictly ectodermic in origin. “In the
adult condition it is much thicker than in its early stages, and the
greater part of its substance has probably come from the ectoderm, al-
though mesodermic cells rest against its proximal face, and possibly may
contribute to its formaticn.

From the description which I have given, it is evident that the optic
disks are thickened regions in the superficial ectoderm, and that these
disks are cut by an intercepting membrane into two parts, one deep,
the other superficial. The deep part is converted into the optic gan-
glion ; the superficial part becomes the retina. So far, then, as the
development of the retina in the lobster is concerned, it supports the
view that the compound eye in Crustaceans is developed from a simple
thickening of the ectoderm.

In describing the formation of the retina and optic ganglion in the
lobster, | have made no mention of an involution. Both Reichenbach
and Kingsley have described an infolding in the formation of the eyes, —
the former in Astacus, the latter in Crangon, —and it is therefore only
natural to look for a similar condition in the eyes of Homarus. Any
evidence of an involution in the production of the lobster’s eyes is to
be sought in the early stages of development. I regret that in the very
early stages my material is deficient, and I have not grounds enough to
warrant the statement that no involution occurs. All that I can state
is, that in all stages which I have examined I have not been able to find
any evidence of an involution. The youngest individual which I have
studied was one in which the optic disk was about two thirds as thick as
that represented in Figure 38. Excepting its thinness and the smaller
number of its nuclei, it presented essentially the same appearance as the
one which is figured. It will be noticed that near the centre of the
disk in Figure 38 there is a space devoid of nuclei. It occurred to me
that such a space might represent the last traces of an involution, and I
therefore plotted carefully the nuclei in five pairs of disks, some of which
were less mature than the disk shown in Figure 38. The result of the
plotting was that the light space which is seen in Figure 38 proved to
be an individual peculiarity, and I did not find in the arrangement of
the nuclei in the other disks any evidence of an involution.

The plotting of the nuclei, however, brought to light the method by
which the disks increased in size. This has already been described, and
offers, I believe, an explanation of the fact that in some cases, as for
instance in the crayfish, the formation of the eye is attended with an
involution, while in other instances, as in Alpheus, no involution is
present.

MUSEUM OF COMPARATIVE ZOOLOGY. 39

It will be remembered that the superficial layer of nuclei in the optic
disk of the lobster was divided into two regions. The one farther from
the median plane has been called the distal region ; the one nearer, the
proximal region. The broadening of the distal region produced the
retina. By a proliferation of its cells the proximal region resulted in
the formation of the optic ganglion. It is my opinion that this prolifer-
ation of cells represents what is produced in the case of some Crusta-
ceans by an involution, and that either an involution or proliferation, or
possibly a combination of both processes, occurs in the eyes of all Crus-
taceans. Whichever process characterizes the development of a given
eye, it must be borne in mind that the involution or proliferation is
connected with the formation of the ganglion only, and takes no part in
the production of the retina. The latter is a simple thickening in the
ectoderm ; the optic ganglion is developed either as a proliferation or
involution of the ectoderm which lies close beside the retina. The
results at which various investigators have arrived, different as they
may at first appear, can be harmonized, I believe, by this interpreta-
tion of the origin of the retina and optic ganglion.

In his account of the development of the eye in the crayfish, Reichen-
bach (’86, p. 85) has described an ectodermic involution, which occurs
nearly in the centre of the optic disk. That portion of the disk which
is farther from the median plane than the region of involution gives
rise to the retina, so that the region of involution in the crayfish
occupies a position which corresponds to the region of proliferation in
the lobster. Not only do the two regions correspond, but the masses
of tissue which are developed from each have certain peculiarities com-
mon to both animals. In the case of the crayfish the mass of tissue
which results from the involution becomes divided into an outer and an
inner wall. These two walls are separated from each other by a band
of nuclei, which are larger and lighter in color than the surrounding
nuclei. In the lobster the ganglionic tissue which arises by proliferation
is divided into an outer and an inner part. The separation is effected
by a band of nuclei, which in position and structure resemble the band
figured by Reichenbach. (Compare Reichenbach, ’86, Plate XII. Fig.
174, and Plate III. Fig. 41 of this paper.) The similarity presented by
the bands of nuclei in the lobster and crayfish supports the conclusion
that the involution in the crayfish and the proliferation in the lobster
are homologous structures.

An objection to this comparison might be raised on the ground that,
according to Reichenbach’s statement (’86, p. 93), the involution in the

40 BULLETIN OF THE

case of the crayfish gave rise to a mass of tissue which formed on the
_ one hand the deeper part of the retina, and on the other a portion of
the optic ganglion, while the cells which arise by proliferation from this
region in the lobster produce only a part of the optic ganglion. But,
as Patten (87, pp. 208, 209) has shown, Reichenbach himself was not
certain that any part of the infolded ectoderm contributed to the for-
mation of the retina. Reichenbach found it difficult to locate exactly
the position which the developing rhabdomes occupied. On page 92 in
his account he describes a part of the outer wall (Ausserwand) of the in-
folded ectoderm, and states his belief that in it the rhabdomes develop ;
in fact, he describes certain red bodies which he says are without doubt
the rhabdomes themselves. On page 96 he admits that the layer in
which he supposed the rhabdomes originated may be a layer of nerve-
fibres. Granting this interpretation, it is no longer possible to consider
the previously described red bodies as rhabdomes. Reichenbach does not
make this last statement, but his description implies it when he states
that, although the region of the red bodies may not be the region of the
rhabdomes, yet the rhabdomes doubtless originate in a somewhat more
superficial part of the outer wall. Apparently he has not identified the
rhabdomes in their new position ; at least, he makes no such statement
in his text or description of plates. Since he has also admitted that the
red bodies may not be rhabdomes, I cannot see that he has positively

identified any structure as a rhabdome. Such being the case, it is diffi-

cult to understand on what grounds he can maintain the assertion that
the rhabdomes develop in the outer wall. If this assertion cannot be
defended, then it is possible that they may develop in the superficial
hypodermis. This would be analogous to the condition presented in the
lobster.

If the rhabdome in the eye of the craytish is developed, as I believe
it is, in the superficial hypodermis, and not in the outer wall of the in-
folded hypodermis, the objection which was suggested in homologizing
the involution in the eye of the crayfish with the proliferation of cells
in the eye of the lobster has no weight.

In the development of the eye of Crangon, according to Kingsley’s
(86*) account, there is also an optic invagination. If what I have at-
tempted to show in regard to the eyes of the crayfish be true, then this
invagination in Crangon should be connected with the formation of the
ganglion only. This, as I have already stated, is not the view held by
Kingsley, for he maintains that the outer wall of the invaginated pocket
gives rise to the retina, and only the inner wall is concerned in the pro-

MUSEUM OF COMPARATIVE ZOOLOGY. 41

duction of the ganglion. This is fundamentally different from the con-
dition found in the Jobster. The difference, however, is due, I believe, to
the fact, that in describing the later stages of development Kingsley has
pointed out a cavity which he believed to be the cavity of the invagi-
nation, but which in reality is not. The cavity which he has marked oc
in his Figures 3, 4, and 5, is unquestionably the cavity of involution,
but the space marked oc in his other figures is, in my opinion, a part of
the body cavity.

My reason for this belief is as follows. The cavity of an involution
such as is found in the anterior median eyes of spiders or the median eyes
of scorpions is, when it has lost its connection with the exterior, a closed
-ectodermic sac. That such a cavity should be occupied by migrating
mesodermic cells seems to me extremely questionable, for in order to
enter the cavity it would be necessary for the cells to penetrate one
wall of the vesicle. This of course is not impossible, but it is not borne
out by analogy with the Arachnoids. The cavity which Kingsley has
marked oc in Figure 7 is occupied by several mesodermic nuclei, and
what is more important, perhaps, is that it is apparently connected with
other cavities in the embryo. These cavities also contain mesodermic
tissue. Excepting Figures 3, 4, and 5, the cavities marked oc I believe
to be homologous, and I further believe that they represent, not the
cavity of involution, but the space which intervenes between the infolded
pocket and the superficial ectoderm. This is a part of the embryonic
body cavity, and is of course readily accessible to mesodermic cells.
The fate of the real cavity of involution is not so easily discovered.
Probably, as in the case of the crayfish, it is obliterated in the mass of
tissue from which the retinal ganglia arise.!

If the interpretation which I have suggested in the preceding para-
graph be admitted, the development of the eye in Crangon is essentially
the same as in the Jobster and crayfish. The proliferation in the optic
disks of the lobster is represented by an involution in the disks of
Crangon. The cavity of the involution disappears in Crangon, and its
walls give rise to ganglionic tissue. The part of the superficial ecto-

1 Since this paper was written, I have received a copy of the third part of
Kingsley’s studies on the development of Crangon. As the following quotation
will show, Kingsley (’89, p. 20) has materially changed his views as to the forma-
tion of the retina: “I may say here that I am inclined to believe that I fell into
error in my account of the development of the Compound Eye of Crangon, and
that the invagination or inpushing which I there described as giving rise to the
ommatidial layer of the eye, in reality gives rise to the ganglion of the eye which
in the adult is contained within the ophthalmic stalk.”

42 BULLETIN OF THE

derm which is immediately external to the ganglionic involution thickens
and produces the retina.

~The mode of development which Patten has suggested for the com-
pound eye of Arthropods has not received any support, so far as I am
aware, from the embryology of the Crustacean eye. ‘The only observa-
tions which go to confirm Patten’s opinion are those of his own on
Vespa, Blatta, and the Phryganids. Possibly the compound eyes of
Crustaceans may be developed upon a different plan from those of Hexa-
pods. Certainly the evidence which Patten has given for the Peripatus-
like type of the compound eye of Hexapods has not been found in any
of the Crustacea. According to Patten’s view of the origin of the com-
pound eyes, the corneal hypodermis should arise on the sides of the
optic area, and spread over the retina until the latter is entirely cov-
ered. This constitutes the closing of the shallow vesicle. When I
come to describe the differentiation of the ommatidia, I believe it can be
shown beyond a doubt that in the lobster the corneal hypodermis arises,
not by any lateral growth from the edge of the optic area, but by a
simple process of delamination. The cells of the corneal hypodermis are
the differentiated superficial cells of that thickening in the hypodermis
which produces the retina. Thus the plan which Patten has suggested
for the compound eyes in Arthropods is not supported by the evidence
derived from the development of the lobster. Patten himself (87,
p- 202) admits that in Vespa he did not see the closure of the hypo-
dermis over the retina, and that in Blatta and the Phryganids (87,
pp. 208 and 211) the process is very obscure.

The researches of Carriere point to a type of compound eyes for the
Hexapods which is similar to that exhibited by the lobster. It is possi-
ble that the vesicular origin of the compound eyes of Hexapods, owing
to their obscure method of formation, may have been overlooked by
Carriére, but I am inclined to believe, after a consideration of both sides
of the question, that the evidence favors the simpler method of origin,
and therefore that the compound eyes of Hexapods as well as Crusta-
ceans arise as simple hypodermal thickenings.

In Mysis, according to Nusbaum (’87, pp. 171-185), the develop-
ment of the eye follows essentially the same course as that which I have
described in the lobster. The optic disks are formed and the ganglionic
and retinal portions are differentiated in the same manner in both cases.
The development of the retina is slightly complicated in Mysis by the
fact that, instead of lying in one plane, or very nearly so, the retina is
strongly folded on itself, so as to give the appearance of two lamelle, an

MUSEUM OF COMPARATIVE ZOOLOGY. 43

internal and an external one. The retinal elements are differentiated
earliest in the internal lamella. As the eye becomes more distinctly sep-
arated from the body, the internal and external lamelle are unfolded so
that they are no longer distinguishable as separate parts. The retina. is
developed from the thickened layer of hypodermis. So far as Nus-
baum’s observations extended, the ganglion is produced without an
involution.

The development of the compound eye in Alpheus has been stud-
ied by Herrick (86, 88, and ’89). The course of development is
almost identical with that of the lobster.‘ The optic disks after they
thicken are cut by a basement membrane into a ganglionic and a reti-
nal portion. There is no involution connected with the formation of
.the ‘eye.

In the introduction to the development of the lobster’s eye, mention
was made of four structural types which the work of different investi-
gators indicated as possible plans for the compound eyes of Crustacea.
I have given reasons for excluding three of these. The type of eye
which Patten has advocated is unsupported by the embryology of the
Crustacea. Reichenbach and Kingsley misinterpreted structures in the
eyes which they studied, and were consequently led to erroneous con-
clusions. If the interpretations which I put on the work of these two
investigators be admitted, all studies on the development of the com-
pound eyes of Crustacea point to one conclusion, namely, that in these
eyes the retina originates as a thickened layer of hypodermis, and is not
modified by any form of involution. The involution when present is
connected with the formation of the optic ganglion only. In the pro-
duction of the ganglion, the involution can be replaced by a proliferation
of the cells.

The Optic Nerve.

The development of the optic nerve? is intimately connected with the
formation of the intercepting membrane. Before the formation of this

1 I have had an opportunity of examining Dr. Herrick’s unpublished plates on
the development of Alpheus, and Dr. Herrick has kindly looked over my figures
of the lobster. The correspondence between the method of growth in the eye of
Alpheus and Homarus is certainly very close. The few differences that were
noticed were such as might be expected between different species. I take this
opportunity of thanking Dr. Herrick for his kindness in extending to me the use of
his plates.

2 The optic tracts of a lobster consist of four principal parts. The first of these
is the retina, from which nerve-fibres lead to the optic ganglion. These three

)

|
|

44 BULLETIN OF THE

membrane the retina and ganglion is one continuous mass of cells
(Fig. 38). When the intercepting membrane is formed, the retina and
ganglion are apparently separated (Fig. 39). I am of opinion, however,
that this separation is only apparent, and that in reality the two struc-
tures are still in connection. At least in this stage and in stage C
(Fig. 41) nuclei are frequently found lying directly across the mem-
brane. These nuclei present so normal an appearance, and their oc-
currence is so frequent, that I cannot believe that their position is due
to accidental displacement. My only way of accounting for the place
which they occupy is by supposing that the basement membrane is
perforated where they are found. The membrane was probably pro-—
duced in the form of a net, through the meshes of which the retina and
ganglion retained their original connection. Either this is true, or it
must be admitted that the retina and ganglion were first connected,
then separated, and finally reconnected ; a supposition which seems to
me unnecessary as well as improbable. From stage C to the adult con-
dition the retina and ganglion are unquestionably connected by nerve-
fibres. If the conclusion arrived at concerning the origin of the optic
nerve is correct, it follows that the optic nerve cannot be properly
described as an outgrowth either from the retina or from the ganglion,
but it must be considered as a remnant of the original connection which
existed between retina and ganglion. Patten (’87, p. 196) has described
essentially the same method of origin for the optic nerve of Vespa. The
formation of the optic nerve from tissue which represents the original
connection of a portion of the optic ganglion with the superficial ecto-
derm is doubtless a reproduction of the method by which that nerve
arose phylogenetically.

The fact that the fibres of the optic nerve are from the outset attached
to the proximal ends of the retinule is of significance in determining
the plan of the eye. Of the four structural types of compound eyes
which have been suggested, that which Kingsley has presented involves
the inversion of the middle or retinal layer. Mark (87, pp. 91, 92)
has shown that when the retina is inverted, as in the anterior median
eyes of spiders, the nerve-fibres are at first attached to the morphologi-
cally deep ends of the retinal cells, and that the attachment afterwards
migrates toward the other ends of these cells. From the fact that in

parts, retina, nerve, and ganglion, are contained in the eye-stalk. The fourth part
is a second bundle of nerve-fibres which connect the optic and cephalic ganglia.
In using the term optic nerve I refer to that collection of fibres which unites the
retina and optic ganglion. 4

MUSEUM OF COMPARATIVE ZOOLOGY. 45

Crustaceans the nerve-fibres are always attached to the proximal ends of
the retinule, it can be argued that the retina in this group has never
been inverted, but retains its original position, and that any structural
plan which involves the inversion of the retina is therefore probably
wrong.

The Differentiation of the Ommatidia.

In the development of the compound eye in the lobster, the deposi-
tion of pigment and the differentiation of the ommatidia take place
at about the same time. These changes occur at stage C. (Compare
Figs. 39, 41, and 42.) At this stage (Fig. 41) it will be observed that the
retinal layer is thickest at the lateral margin of the disk (the extreme
left in Fig. 41). The retina becomes thinner as one proceeds from the
margin toward the median plane (from left to right in the figure).
The thickest part of the retina, it will be recalled, was the part first to
be separated from the ganglion by the intercepting membrane. As it
was the first part of the retina to be separated, so it is the first part in
which the ommatidia are differentiated and pigment is deposited.

The first steps in the differentiation of the ommatidia are seen in
Figure 42. Here the nuclei in the thicker part of the retina have sep-
arated into two bands, one distal (y), and one proximal (x). The distal
band, as I shail presently show, can be further separated into a super-
ficial and deep layer. ‘These two layers are close to the outer surface
of the retina, and approximately parallel with it.

The arrangement of the nuclei which make up the distal band is most
easily observed when the retina is viewed from the surface. The group-
ing of these nuclei, from the time when the ommatidia are differentiated
to the adult condition, is so characteristic that the fate of the individual
nuclei can be easily traced. For the sake of simplicity I shall therefore
designate the different nuclei, from their first appearance in groups, by
the names of the cells in which they are ultimately found, although it
is to be borne in mind that in the early stages the cell walls are not as
yet differentiated.

Viewed from the surface, the superficial layer of the distal band of
nuclei presents the appearance which is shown in Figure 43. In this
layer there are two kinds of nuclei, one elongated, the other roundish.
The elongated ones are always in pairs. They ultimately become the
nuclei of the corneal hypodermis. These hypodermal nuclei arise from
among the superficial nuclei of the distal band, and do not originate, as
Patten believed (’86, p. 645), from a fold which ‘grows over the retinal

46 BULLETIN OF THE

area. Of course the number of pairs of hypodermal nuclei equals the
number of ommatidia. In the earliest stages in which the ommatidia
have been seen there were always three or four pairs of hypodermal
nuclei, so that it is probable that the first step in differentiation is the
simultaneous production of three or four ommatidia.

The second kind of nuclei in the superficial layer are roundish and
arranged in circles of six around each pair of hypodermal nuclei. The
circles are so combined that each nucleus plays a part in three circles ;
and as there is one circle for each ommatidium, it follows that only one
third of each nucleus belongs to a given ommatidium, or, if one esti-
mates the nuclei as units, only one third of each circle of six nuclei, or
two nuclei, belong to an ommatidium. These nuclei represent the cells
which in the adult have been called the distal retinule, and of which
there was a single pair to each ommatidium. |

The deep layer in the distal band of nuclei lies directly below the
superticial layer (Fig. 42). The nuclei which constitute this layer are
all of one kind, and are arranged in groups of four (Fig. 44). They
represent the cells of the crystalline cones. The centre of a given group
of cone-nuclei is directly below the centre of a pair of hypodermal nuclei.
At this stage the cone-cells can be observed as elongated pyramids,
which lie with their bases in the region of their four nuclei, and their
apices extending into the deeper part of the retina (Fig. 42).

The nuclei of the proximal band show at this stage no special arrange-
ment. They migrate to the deeper part of the retina, and there undergo
further change. Between them and the basement membrane the pig-
ment is deposited.

The most noticeable changes which the retina as a whole now under-
goes are two. First, it thickens until it is throughout nearly as thick
as in the region where the ommatidia were first differentiated. (Com-
pare Figs. 45 and 46.) Second, the number of ommatidia greatly in-
creases. These two changes, the thickening of the retina and the
production of new ommatidia, go hand in hand and spread, over the
general surface of the eye, from the region in which the first ommatidia
appear. It is worthy of notice that the new ommatidia are constructed
from the undifferentiated cells which immediately surround the area of
ommatidia already formed. Cells once incorporated in a given omma-
tidium never in any way contribute to the formation of other ommatidia.
Moreover, in the differentiation of an ommatidium no cells are left
between it and the neighboring ommatidia, so that ommatidia once ad-
joining each other remain so. In other words, new ommatidia are not

MUSEUM OF COMPARATIVE ZOOLOGY. 47

produced between old ones, but only on the edge of the ommatidial
area.

The changes in the ommatidia themselves can be studied most readily
in longitudinal and transverse sections of the retina. Figure 47 is an
enlarged drawing of that portion of the retina which is marked by a
bracket in Figure 46. It will be noticed that in this stage, E, the nuclei
are limited for the most part to the distal half of the retina. The
middle of the retina is occupied by a band of pigment, which gradually
fades as it approaches the basement membrane. The space between the
basement membrane and the ganglion is relatively narrow, and is occu-
pied chiefly by nerve-fibres. Returning now to the nuclei in the distal
half of the retina, it is to be observed that the general arrangement
which was pointed out in the last stage also persists here. The nuclei
of the distal retinule are still in circles of six (compare Figs. 47 and
48, nl. dst.), and are very close to the external surface of the eye. In
the centre of each circle is seen a round pink body, the tip of the cone-
cells (Fig. 48 con.). Directly below the level of the nuclei in the distal
retinule are the pairs of nuclei belonging to the corneal hypodermis
(Figs. 47 and 49, nl. crn.). Each pair of corneal nuclei surrounds the
distal end of the cone.

Below this level one meets the four nuclei of the cone-cells (Figs. 47
and 50, nl. con.). From the side view (Fig. 47) it will be seen that the
groups of cone-cells are spindle-shaped in outline, and have their nuclei
arranged in a transverse plane at the thickest part of the spindle. The
nuclei which in stage C formed the proximal band are scattered in this
stage between the deep ends of the cones (Fig. 47, nl. px.). They are
not definitely arranged. In order to estimate the number of nuclei in the
proximal band for each ommatidium, I counted these nuclei as seen in a
series of tangential sections. In the outermost section in which the prox- °
imal nuclei occur there were six nuclei around each cone. These nuclei,
however, were arranged in circles similar to the circles of the corneal
nuclei; consequently for each ommatidium there were only two nuclei
in each circle of six. The remaining nuclei were all embraced in the
two succeeding deeper sections. In each of these two sections there
were about nine nuclei around each cone. The arrangement of these
nuclei was extremely irregular, and it was consequently difficult to
estimate the number of nuclei which belonged to one ommatidium.
Most of the nuclei were situated between three cones, therefore, about
one third of the nuclei around a given cone can be considered as belong-
ing to the ommatidium represented by that cone. As there were in

48 BULLETIN OF THE

these two sections about eighteen nuclei around each cone, it follows
_ that one third of this number, or six, represents approximately the num-
ber of nuclei for each ommatidium. If then the deeper sections con-
tain six nuclei for each ommatidium and the outermost section two, the
total number of proximal nuclei for each ommatidium must be eight.
I do not mean to imply that this estimate can be insisted upon as abso-
lutely invariable ; but I wish to show that, as these nuclei represent the
proximal retinule in an adult lobster’s eye, and as there are eight
such retinule and about eight of these nuclei to an ommatidium, the
change which the eight embryonic cells undergo in becoming adult
retinule is chiefly that of arrangement, and certainly does not involve
any considerable increase in numbers. :

At stage E, which was the one last described, the young lobster was
about to escape from the egg-shell. The next stage, F, is that of a
lobster about one inch in length. At this stage the optic lobes are
represented by optic stalks, and the distal rounded end of each stalk is
occupied by the retina. Figure 51 represents a longitudinal section of
a single ommatidium from this stage. The distal end of the ommatid-
ium is covered with a well developed corneal cuticula. The cuticula
is marked out into hexagonal corneal facets (Fig. 52). The facets of
course indicate the arrangement of the ommatidia. This arrangement
at the first differentiation of ommatidia was such that hexagonal and
not square facets would have resulted if a cuticula had been then
produced.

Directly below each corneal facet is a pair of crescentic nuclei, those
of the corneal hypodermis (Fig. 53, x. crn.). ‘These nuclei have in all
preceding stages shown a tendency to become elongated and crescentic
in outline, but it is in this stage that this peculiarity reaches its highest
development. Between each pair of hypodermal nuclei the distal end
of the cone-cells is usually seen (Fig. 53, con.). The four cone-cells
with their nuclei occur immediately below the corneal hypodermis
(Figs. 51 and 54, nl. con.). The cone itself is already formed in part,
and lies below the nuclei of the cone-cells. Proximally the four cone-
cells can be traced to near the middle of the retina; here they are no
longer distinguishable. The proximal end of the cone itself terminates
as four processes, one on the outer lateral wall of each cone-cell (Fig. 56).
Surrounding the cone about midway on its length are the nuclei of the
distal retinule (Figs. 51 and 55, nl. dst.). In transverse sections of
stage F, these nuclei show the same grouping in circles of six as they
showed in previous stages. The outlines of the retinule are not visible.

MUSEUM OF COMPARATIVE ZOOLOGY. 49

The substance in which the nuclei lie contains a few granules of pig-
ment. Below the proximal end of the cone the nuclei of the proximal
part of the retina can be seen. These are not so definitely arranged
that they can be counted. They occur on several planes. Fortunately,
their cells in this stage have very distinct outlines, and a short distance
below the nuclei one can see with perfect clearness the seven retinule
which surround each rhabdome (Fig. 57, rtn!. px.). Occasionally, as is
seen in Figure 57, the nucleus of the retinula occupies a position in its
cell as deep as the plane of this section. The proximal retinule contain
a few pigment granules. On account of the great similarity of the
groups of proximal retinule, I have not been able to plot and super-
impose sections with certainty ; and as the nuclei of the proximal retin-
ulz are placed at different levels I have not succeeded in identifying
an eighth nucleus, which, it will be remembered, was pointed out in
the histology of the adult eye as probably representing a degenerate
retinula.

At this stage the first trace of the rhabdome appears (Fig. 57, rhd.).
It is a cylindrical thickening in the centre of each group of proximal
retinule, and extends from a short distance below the crystalline cone
very nearly to the basement membrane. From its earliest appearance
it is divided into four segments, which bear the same relation to the sur-
rounding retinulz as they do in the adult. (Compare Figs. 58 and 34.) —
Nothing has been observed in the development of the rhabdome which
indicates the significance of the four lines by which the segments of the
rhabdome are separated.

Owing to a lack of distinctness in the tissue near the basement
membrane, I have been unable to identify the individual retinule in
that region. What I have observed is, that fibres pass from the retina
through the basement membrane and into the optic ganglion. Pre-
sumably these fibres are from the proximal ends of the retinule, and are
grouped as I have described in the retina of mature lobsters.

At this stage the only observation which I have made bearing on the
question of nerve termination is as follows. In each proximal retinula
near the rhabdome there are one or two fibres which extend nearly the
whole length of the retinula. In transverse sections they of course
appear as dots (Fig. 58, fbr’.), and might be mistaken for the remains
of pigment granules were it not fom their sharper outlines and the regu-
larity of their arrangement. I am of opinion that they are the first indi-
cations of nerve-fibrille, which, as I have pointed out in the section on
Histology, lie in the adult eye next the rhabdome.

VOL. XX. — NO. l. 4

50 BULLETIN OF THE

The cells which I have thus far described in the differentiation of the
ommatidia are unquestionably ectodermal in origin. In stage F certain
nuclei appear which may have another origin. It will be recalled that
in stage C (Fig. 41), although the intercepting membrane is well devel-
oped, the retina and optic ganglion are still closely applied to each
other. In stage D (Fig. 45) the retina and ganglion have separated
enough to form an intervening space of considerable extent. In stage
E (Fig. 46) this space not unfrequently contains several nuclei. These
are smaller than the ectodermic nuclei in the retina, and of about the
size of those in the ganglion. Their chromatine differs from that of
both the retinal and ganglionic nuclei, in that it has the form of very
distinct particles which give the nucleus a decidedly granular appearance.
Moreover, these nuclei, which I believe to be mesodermic in origin, are
variable in shape, whereas the different kinds of ectodermic nuclei pos-
sess characteristic forms (Fig. 47). In stage F the space between the
retina and ganglion also contains a few mesodermic nuclei, and similar
ones are noticeable in the base of the retina (Fig. 51, nl. pig.). The
latter are different from the nuclei of the proximal retinule (Fig. 51,
nl. px.), and resemble so closely the nuclei which in the adult have been
described as belonging to the accessory pigment cells, that I believe
them to be the nuclei of those cells."

From stage F to that of the fully grown lobster the changes in the
retina are, with one exception, rather insignificant. The parts of the
retina increase considerably in size, especially at the distal end of
the ommatidium, and additional pigment is deposited in the distal and
proximal retinule. The only change of importance which the eye
undergoes before full maturity is reached is a rearrangement of the
ommatidia. It will be remembered that up to stage F the ommatidia
were so arranged in relation to each other that the resulting corneal
facets were hexagonal in outline. In the adult lobster the facets are
square. The change from the six- to the four-sided facet is effected,
I believe, by a partial slipping of rows of ommatidia on each other.
Imagine, in Figure 55, a row of ommatidia extending in the direction
of the arrow and on either side, and parallel to this another row. Only
one ommatidium in each of these two lateral rows is given in the figure.

1 When the preliminary notice of this paper was written, I was somewhat in
doubt as to the origin of the accessory pigment-cells. I had not then had the op-
portunity of studying stage F, and I was of opinion that these cells were probably
of ectodermic origin. I believe that the evidence which is now at hand indicates
that they are cells derived from the mesoderm.

an

MUSEUM OF COMPARATIVE ZOOLOGY. | Rea

Suppose the individual ommatidia of the three rows to be arranged in
reference to one another as the four in Figure 55; i. e. the ommatidium
of one row covering the open spaces between two ommatidia in the
adjoining row. Under these conditions hexagonal facets would be pro-
duced. But now imagine the middle row to move in the direction of
the arrow through the distance of half the thickness of an ommatidium.
After the completion of this movement, the ommatidia of the middle
row are directly opposite the ommatidia of the adjacent rows. With
this arrangement the ommatidia can be grouped in square blocks of
four, nine, etc. This is the grouping which obtains in the adult retina,
and the facets resulting from it are square in outline. In the more
primitive arrangement, that with hexagonal facets, the ommatidia can
also be grouped in blocks of four, nine, etc. These blocks, however,
are never square in outline, but lozenge-shaped (Fig. 55).

The process of rearranging the ommatidia is accomplished at a period
in the growth of the young lobster much later than that at which the
nerve-fibres have arranged themselves in relation to the openings in
the basement membrane. Such being the case, one would expect to
find in the deeper part of the adult retina traces of the more primitive
arrangement. This is seen, I believe, in the lozenge-shaped outline
which groups of four ommatidia present in the deeper part of the retina.
A good instance of this is to be seen in Figure 14, although it is ap-
parent in almost all of the transverse sections which include the proxi-
mal retinule. One might say that the ommatidia were rooted to the
basement membrane when the hexagonal system prevailed, and that the
rearrangement fully affected only their free distal ends.

The movement suggested as a means of rearranging the ommatidia
not only explains the new position of the ommatidia, but also accounts
for the situation in which the nuclei of the distal retinule occur. In
Figure 55 each ommatidium is surrounded by a circle of six nuclei.
These belong to the distal retinule. Instead of describing them as
being in circles of six, it might be said that there is one nucleus at each
corner of every cone. Thus, cone x (Fig. 55) has nuclei 1, 2, 3, and 4
at its four corners, and cone y has in a corresponding way nuclei 5, 6, 7,
and 8. If the cones were to move as I have already described, and
were to carry their surrounding nuclei with them, the result would be
that nucleus 5 would come to lie next to nucleus 2, and 7 next to 4, and
so on. In other words, a pair of nuclei would occupy each space be-
tween the adjoining angles of four adjacent cones. This is the position
which the nuclei of the distal retinulee occupy in the retina of an adult
lobster (Fig. 5).

52 BULLETIN OF THE

Several investigators have already described the development of the
ommatidia in the compound eyes of the higher Crustaceans. The differ-
ent accounts disagree especially in two particulars ; first, as to the source
of the different structures in the retina, and, secondly, as to the number
and arrangement of the cells in an ommatidium. In describing the
development of the retina, I have already discussed the first difference,
and need here only recall my conclusion; namely, that the retina, includ-
ing the corneal hypodermis, crystalline cones, retinule, and rhabdome,
originates as a simple hypodermal thickening, and that no part of it is
derived from the deeper ectoderm, which becomes the central nervous
system. As to the number of cells which constitute an ommatidium, it
will be recalled that in the lobster there are at least sixteen in each
ommatidium ; two in the corneal hypodermis, four cone-cells, two distal
retinulz, eight proximal retinule one of which was rudimentary, and a
small, but variable number of accessory pigment-cells. The last are
probably mesodermic in origin; all of the others are derived from the
ectoderm.

The several accounts of the corneal hypodermis given by various
authors differ principally in the number of cells which are said to be
found in each ommatidium. Reichenbach (’86, p. 91) and Nusbaum
(’87, p. 179) state respectively that in Astacus and Mysis there are
four hypodermal cells in each ommatidium. Nusbaum’s statement
is further supported by Grenacher (’79, p. 118), who describes four cells
under each facet in Mysis. Herrick (’89, p. 167) has found two hypo-
dermal cells in the ommatidium of Alpheus. Patten states that the
number of corneal cells in the ommatidia of all Decapods which he has
examined is two.

All the direct evidence that I have seen points to the conclusion that
the ommatidia of the Decapods possess two cells in the corneal hypo-
dermis. Reichenbach’s observation directly opposes this view. I have
not had the opportunity of examining the same species as Reichenbach
did, but I have studied a representative of the fresh-water crayfishes,
Cambarus Bartoni, and there is no question that in the ommatidium
of this species only two corneal cells are present. The nuclei of these
two cells lie in the angles of the hypodermal squares, each one directly
above a nucleus of a cone-cell. When viewed from the surface, it is
difficult to say whether there are two or four hypodermal nuclei, be-
cause the four nuclei of the cone lie so near to the hypodermis and
resemble its nuclei so closely. It seems to me possible that in
his surface view Reichenbach (86, Fig. 226) may have drawn with

,
;
:
J
i

=
—

MUSEUM OF COMPARATIVE ZOOLOGY. 53

the hypodermal nuclei those of the cone-cells, thus giving four instead
of two.

The four nuclei in the corneal hypodermis of each ommatidium, as
described by Nusbaum in the case of Mysis, had already been seen by
Grenacher, who further stated that the nuclei were arranged in pairs at
two levels. Grenacher does not describe nuclei in the two segments of
the crystalline cone. In the retina of Mysis stenolepis, the four nuclei
which were described by Grenacher are easily identified, but I cannot
agree with Grenacher’s statement (’79, p. 118) that all four belong to
the same category. The more superficial pair unquestionably belong to
the corneal hypodermis, but the deeper pair, I feel confident, are the
nuclei of the crystalline cone. The conclusion to be drawn from this
interpretation of the work of Reichenbach, Nusbaum, and Grenacher is,
that in Decapods and Schizopods each ommatidium pussesses two cells
in the corneal hypodermis and only two.

Concerning the number of cone-cells in the ommatidium of Decapods,
different writers very generally agree. Reichenbach, Kingsley, Herrick,
and Patten all state that there are four cone-cells in the Crustaceans
which they have studied.

In the cone-cells the number four, according to Grenacher, is charac-
teristic not only of Decapods, but, excepting the Schizopods, it is the
distinguishing feature of the Podophthalmata. In Mysis, Grenacher
states that the cone has two, instead of four segments. I have studied
the eyes of Mysis stenolepis, Smith, and my observations confirm this
statement. Notwithstanding Grenacher’s assertion, Nusbaum (’87,
p- 179) claims that the nuclei of the cone-cells in Mysis are grouped
in fours. Iam confident that there are only two nuclei in the adult
cone, and having seen no evidence of a suppression of two nuclei, I
must consequently side with Grenacher in his belief that the cone of
Mysis is composed of only two cells.

The differentiation of the retinulz into distal and proximal groups is
more complete in the lobster than in the majority of other Decapods
studied. As a result of this incomplete differentiation, it is often
impossible to get exact statements from the descriptions of different
authors concerning the number and position of the retinule.

Reichenbach (’86, p. 93) maintains that in Astacus, after the four
cone-cells, and as he believes the four hypodermal cells, were differ-
entiated, the remaining cells became pigment-cells (retinule). Essen-
tially the same account is given by Nusbaum (’87, p. 180) for Mysis.
Kingsley (’87, p. 53) states that, in addition to the corneal hypodermis,

54 BULLETIN OF THE

each ommatidium in Crangon at first consists of four vertical series of
nuclei, each series containing five nuclei. This would give a total of
twenty cells for each ommatidium. After deducting four, the number
of cone-cells, from twenty, the original number of cells, there remain
sixteen cells to be accounted for, presumably as pigment-cells. I have
examined the eye of an adult Crangon vulgaris, and I find in it, as in
the lobster’s eye, two distal retinule and seven proximal retinule, This
can scarcely be reconciled with Kingsley’s account, unless one admits
an extensive suppression of cells. Such a suppression seems to me
scarcely probable, and I am therefore inclined to believe that there
has been some error in Kingsley’s method of counting. Possibly
the series of nuclei were so placed that they were shared by neigh-
boring ommatidia, and did not all belong to one ommatidium. This,
however, could only be settled by re-examining the early stages of
Crangon.

Herrick (’86, p. 44) describes seven retinule in the ommatidium of
Alpheus, and makes the statement that they do not possess nuclei. He
then describes some undifferentiated ectodermic cells, the nuclei of
which can be seen in the space between the cones. As this is the
position which in the young lobster is occupied by the nuclei of the
proximal retinule, and as Herrick has not identified any nuclei for
the proximal retinule in Alpheus, I am inclined to regard the deeper
nuclei of this group as belonging to these cells. The more superficial of
the nuclei described by Herrick are apparently arranged in circles of
six around each cone. (Compare Herrick, ’86, Figs. 1 and 2.) As in the
early stages of the lobster this arrangement was characteristic of the
nuclei in the distal retinule, it is possible that these superficial nuclei
in Alpheus may represent the distal retinule.

Reichenbach (’86, p. 92), Kingsley (’87, p. 52), Nusbaum (’87, p. 179),
and Herrick (’89, p. 167) describe an ingrowth of mesodermic tissue
between the retina and ganglion, and in Mysis, according to Nusbaum
(’87, p. 180), these cells, as in the lobster, give rise to what I have
called the accessory pigment-cells.

From the investigations which have been summarized in the preced-
ing pages, it is difficult to draw any general conclusion concerning the
number of retinule in the ommatidia of the higher Crustacea. Reichen-
bach and Nusbaum make no statement as to the number of these cells
in Astacus and Mysis. Kingsley’s enumeration of them in Crangon
seems to be erroneous. Admitting the undifferentiated ectodermic
nuclei in Alpheus to be the nuclei of the retinule, Herrick’s statement

MUSEUM OF COMPARATIVE ZOOLOGY. : BH

that there are seven retinule in the ommatidia of this Crustacean coin-
cides fairly with the results obtained from the lobster.

From the histological evidence of the adult, on the other hand, writers
are generally agreed that the ommatidia of Decapods possess seven
proximal retinule. It is probable, however, that this statement requires
some qualification. It will be recalled that, in describing the proximal
retinule of the lobster, I referred to an additional nucleus which ap-
parently represented an eighth rudimentary retinula. I have identified
this eighth nucleus in the ommatidia of Cambarus, where, as in the
lobster, it lies in a plane different from that of the other seven nuclei.
If the eighth nucleus should be present in approximately the same
plane as the other nuclei, it could be identified only with great difficulty.
It is my belief that it often occurs in this position, and probably
for this reason it has generally escaped the attention of investigators, —
for I am of opinion that it is present in the ommatidia of all Decapods.
When, therefore, the statement is made that the ommatidia of a certain
Decapod contains seven proximal retinulse, the probabilities are that
the ommatidium in reality contains eight proximal retinule, one of
which is rudimentary.

Concerning the Schizopods, Grenacher states (’79, p. 119) that in
Mysis there are more than four proximal retinule, but how many more
he is not certain. In Mysis stenolepis my own observations have shown
me that there are certainly seven pigmented proximal retinule. This
number agrees with the number of functional retinulz in Decapods, and
my opinion is that in Mysis, as in Decapods, an eighth rudimentary
proximal retinula may be expected.

The presence of distal retinule in the ommatidia of Decapods seems
to have generally escaped attention. Patten and Carriere, however,
describe these cells in Penzus and Astacus, and the fact that they are
easily recognized in the eyes of Homarus, Cambarus, and Eupagurus
inclines me to the belief that they form a constant element in the om-
matidia of all Decapods. They have also been seen in the retina of
Mysis. In the eyes of this genus their nuclei occupy the position indi-
tated by d in Grenacher’s Figure .12. It is of interest to observe that
their permanent position in Mysis is an early and transitory one in the
lobster. (Compare Grenacher, ’79, Fig. 112 d, with Figs. 48 and 55 of
this paper.) In both the Decapods and Mysis the number of distal
retinule is two.

56 BULLETIN OF THE

The Types of Ommatidia.

With the conclusions arrived at in the foregoing account as a basis,
an ommatidium can be constructed which will serve as a type for the
ommatidia of all Decapods. Omitting the accessory pigment-cells, this
typical ommatidium would be composed of sixteen cells as follows:
two cells in the corneal hypodermis, four cone-cells, two distal retinule,
and eight proximal retinule. In a similar way a typical ommatidium
can be constructed for the Schizopods. This type would differ from
that proposed for the Decapods, in that its crystalline cone would be
formed of two, instead of four cells.

The ommatidia of the Schizopods and Decapods, as the development
of the lobster shows, are closely related. The arrangement of the om-
matidia by which hexagonal facets are produced is permanent in Mysis,
and temporary in the lobster. The distal retinule are grouped in the
adult Mysis in a way which is reproduced only in the early stages of
the lobster. In Mysis the outline of the nuclei in the corneal hypoder-
mis, aS seen in sections tangential to the retina, is strikingly crescen-
tic. This form is temporarily assumed by the corresponding nuclei in
the young lobster (Fig. 53). Thus it is evident that the ommatidia of
the lobster pass through a stage in which they closely resemble the
permanent condition of the ommatidia in Mysis. For this reason, [
believe that the ommatidium of Mysis represents a type ancestral to
that of the lobster.

The only important difference which exists between the ommatidium
of Mysis and that of the lobster in its early stages is that in the latter
the cone consists of four cells, while in Mysis it is composed of only two.
If one admits that the ommatidium of Mysis is the forerunner of that
of the lobster, this difference in the number of cone-cells can easily be
explained on the supposition that the cells which form the cone in
Decapods divided once more than those in Schizopods. If these two
types of ommatidia are thus related, it is only natural to expect that
this process of cell-division may connect the ommatidium of Mysis
with that of some lower Crustacean.

The ommatidia in the eyes of Amphipods are constructed upon a
type which seems thus connected with that of Mysis. In Gammarus,
for instance, I have found that the ommatidia possess an undifferen-
tiated corneal hypodermis, a crystalline cone of two cells, and five retin-
ule. If one desires to convert this type into that of Mysis, or through
Mysis into that of the Decapods, the necessary change merely involves

MUSEUM OF COMPARATIVE ZOOLOGY. o7

the division and differentiation of cells. The two cone-cells in the
simpler type would remain unaffected in Mysis; in the Decapod they
would divide once, thus producing a cone of four cells. By division,
the five retinule of the Amphipod would become the ten retinule of
the higher type. These ten retinule are differentiated into two sets,
eight proximal retinule which surround the rhabdome, seven of them
possessing nerve-fibres, and two distal retinule which envelop the cone,
but have no nervous connections. If the ten retinule of the higher
type were developed from the five retinule of the lower type, it follows,
since all of the five retinule possessed nerve-fibres, that those of the
ten retinule which do not possess nerve-fibres must be considered as
degenerate. The degenerate cells of the higher type of ommatidium
are consequently the eighth proximal retinula and the two distal retin-
ule. The structural condition of these three cells favors this view.
The eighth proximal retinula is identifiable only through its nucleus.
Each distal retinula, as I have previously described, possesses a proxi-
mal fibre. This fibre passes through the basement membrane with the
nerve-fibres of the proximal retinule, but it is very much smaller than
these fibres, and terminates without reaching the optic ganglion. This
condition is easily interpreted as one of degeneracy.

The process by which distal and proximal retinule were Tacbabky differ-
entiated from unmodified retinule is easily suggested. A characteristic
distinction between the ommatidia of the higher and lower Crustacea,
as, for instance, between Gammarus and Homarus, is that in the former
the cone and rhabdome are very close to each other, whereas in the
latter the cone proper and rhabdome are separated by considerable
space. Without attempting to assign physiological reasons, the state-
ment may be made that it seems necessary that the sides of the cone
should be sheathed with pigment. In the ommatidia of Gammarus this
is accomplished by the distal ends of the five retinule ; these reach be-
yond the rhabdome and envelop the cone. They thus form a funnel, in
the large central cavity of which the cone rests, while the neck of the
funnel is occupied by the rhabdome. Imagine now the division of
these five retinulze into ten, and the separation of the cone from the
rhabdome. It is easy to understand how two of the ten retinule may
retain their connection with the cone, while the remaining eight adhere
to the rhabdome. The two retinule connected with the cone would lose
their nervous function and become simple pigment-cells. The retinule
which remain attached to the rhabdome would continue as perceptive
cells. The separation of the rhabdome and cone not only offers an

58 BULLETIN OF THE

explanation of the way in which the distal and proximal retinule have
arisen, but it also accords well with the fact that the proximal retinule
end distally in fine fibres which stretch toward the cone, and are ap-
plied to the fibres of the distal retinulee.

If what has been said of the growth of ommatidia be true, the course
whieh the development of these structures takes in the case of the
lobster must be somewhat different from that by which they arose
phylogenetically. In the lobster the division of the nuclei is entirely
completed before the ommatidia are differentiated. Consequently, after
differentiation has occurred, no further cell-division ensues among the
elements of an ommatidium. ‘This fact, however, is by no means a
serious objection to the view which I have expressed, for of the two
processes, cell-division and the differentiation of ommatidia, it is only
necessary to imagine that in successive generations the differentiation
was retarded until a stage was reached in which all cell-division was
accomplished before the differentiation of ommatidia began. It is of
interest to observe, however, that in the lobster the planes of division
among the nuclei, which eventually enter into the formation of omma-
tidia, correspond in direction to the plane in which the nuclei of the
ommatidium in Gammarus would divide, should this ommatidium be
converted into that of the lobster. Thus the plane of nuclear division
which was so characteristic of the distal superficial part of each optic
disk in the lobster may have a phylogenetic significance.

In how far the ommatidia of all the Crustacea can be brought into
relation by a process of cell-division such as I have outlined, and what
constitutes the simplest form of an ommatidium, are questions which
require careful and extensive comparative study, and which I am not
able to discuss here. Suffice it to say, that between the ommatidia of
the higher and lower Crustacea there is reason to conclude that such a
relation as I have pointed out probably exists.

CAMBRIDGE, October 1, 1889.

MUSEUM OF COMPARATIVE ZOOLOGY. 59

BIBLIOGRAPHY.

Bobretsky, N.
"73. Development of Astacus and Palemon. Kiew, 1873. (The substance

of this paper, so far-as it refers to the eye, is known to me through the
abstract in Balfour’s Comparative Embryology, Vol. II. pp. 397, 398.
London, 1881.)

Carriére, J.
’'85. Die Sehorgane der Thiere vergleichend-anatomisch dargestellt. Miin-

chen u. Leipzig: R. Oldenbourg. 1885. 6+ 205 pp., 147 Abbildg. u.
1 Taf.

’°89. Bau und Entwicklung des Auges der zebnfiissigen Crustaceen und der
Arachnoiden. Biologisches Centralblatt, Bd. IX. No. 8, pp. 225-234.
15 June, 1889.

Claus, C.
’'86. Untersuchungen iiber die Organisation und Entwickelung von Branchi-

pus und Artemia nebst vergleichenden Bemerkungen uber andere Phyl-
lopoden. Arbeit. Zool. Inst. Wien, Tom. VI. Heft 3, pp. 267-871,
12 pls. 1886.

Grenacher, H.
'79. Untersuchungen iiber das Sehorgan der Arthropoden, insbesondere der

Spinnen, Insecten und Crustaceen. G6ttingen: Vandenhock u. Ruprecht.
1879. 8 +188 pp.; 11 Taf.

Grobben C.
"79. Die Entwickelungsgeschichte der Moina rectirostris. Arbeit. Zool. Inst.

Wien, Tom. II. Heft 2, pp. 2038-269, 7 pls. 1879.
Herrick, F. H.
’86. Notes onthe Embryology of Alpheus and other Crustacea and on the
Development of the Compound Eye. Johns Hopkins Univ. Circulars,
Vol. VI. No. 54, pp. 42-44. Dec., 1886.
’88. Development of Alpheus. Johns Hopkins Univ. Circulars, Vol. VII.
No. 63, pp. 36, 37. Feb., 1888.
’89. The Development of the Compound Eye of Alpheus. Zool. Anzeiger,
Jahrg. 11, No. 308, pp. 164-169. 25 March, 1889.
Kingsley, J. S.
’'86. The Arthropod Eye. Amer. Naturalist, Vol. XX. No. 10, pp. 862-867.
Oct., 1886.
‘862. The Development of the Compound Eye of Crangon. Zool. Anzeiger,
Jahrg. 9, No. 234, pp. 597-600. 11 October, 1886.

60 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

Kingsley, J. S.
'87. The Development of the Compound Eye of Crangon. Journ. of Mor-
phology, Vol. I. pp. 49-67, 1 pl. 1887.
'89. The Development of Crangon vulgaris. Third Paper, with Plates I.,
IL., III. Bull. Essex Inst., Vol. XXI., pp. 1-42. 1889.
Mark, E. L.
’'87. Simple Eyes in Arthropods. Bull. Mus. Comp. Zodl. at Harvard Col-
lege, Vol. XIII. No. 3, pp. 49-105, 5 pls. Feb., 1887.
Newton, E. T.
'73. The Structure of the Eye of the Lobster. Quart. Jour. of Micr. Sci.,
Vol. XIII., New Series, pp. 325-343. 1873.
Nusbaum, J.
'87. L’Embryologie de Mysis chameleo (Thompson). Arch. Zool. exper.,
sér. 2, Tom. V. pp. 123-208. 1887. 7

Parker, G. H.
’'87. The Eyes in Scorpions. Bull. Mus. Comp. Zodl. at Harvard College,

Vol. XIII. No. 6, pp. 173-208, 4 pls. Dec., 1887.
’'88. <A Preliminary Account of the Development and Histology of the Eyes
in the Lobster. Proc. Amer. Acad. Arts and Sci., Vol. XXIV. pp. 24, 25.

Nore. — Copies of this paper were distributed in November, 1888, although the volume in which it
appeared was not issued till December, 1889.

Patten, W.

’86. Eyes of Molluscs and Arthropods. Mittheilungen a. d. zool. Station
zu Neapel, Bd. VI. Heft 4, pp. 542-756, Pls. 28-32. 1886.

’°87. Studies on the Eyes of Arthropods. 1. Development of the Eyes of
Vespa, with Observations on the Ocelli of some Insects. Journ. of Mor-
phology, Vol. I. pp. 198-227, 1 pl. 1887.

Reichenbach, H.
86. Studien zur Entwicklungsgeschichte des Flusskrebses. Abhandl.
Senckenb. Naturf. Ges., Bd. 14, Heft 1, pp. 1-137. 1886.
Schultze, M.
‘67. Ueber die Endorgane des Sehnerven im Auge der Gliederthiere. Arch.
f. mikr. Anat., Bd. III. pp. 404-408. 1867.

’'68. Untersuchungen iiber die zusammengesetzten Augen der Krebse und

Insecten. Bonn: Max Cohen & Sohn. 32 pp., 2 Taf. 1868.
Watase, Sho.

89. On the Structure and Development of the Eyes of the Limulus. Johns

Hopkins Univ. Circulars, Vol. VIII. No. 70, pp. 84-37. March, 1889.

PARKER. — Lobster Eye.

EXPLANATION OF FIGURES.

All the figures were drawn with the aid of an Abbé camera. Unless otherwise
stated, the figures are from specimens stained with Grenacher’s alcoholic borax-
carmine and mounted in benzol-balsam. Where sections have been depigmented
the reagent employed was an aqueous solution of potassic hydrate +% (see Par-
ker, *87, p. 175). Figs. 1 to 36 refer to the histology of the adult lobster’s eye.
Figs. 37 to 59 inclusive deal with the development of the eye.

PLATE I.

All figures on this plate illustrate the histology of the lobster’s eye.

Fig. 1. A longitudinal section of an ommatidium. This is a composite figure, its
parts having been drawn to one scale from various sections, and after-
wards combined. The distal retinula on the right side of the cone
contains its natural pigment; that on the left side has been depig-
mented. The letter x indicates the position of the band-which limits
the corneal facet. The numbers to the right of the figure correspond
to the numbers of the following figures of transverse sections, and
mark the level at which the latter were taken. xX 105.

The remaining figures on this plate (Nos. 2-25) are transverse sections
either of ommatidia or bundles of nerve-fibres. In each case the sections were
studied and drawn from their distal faces. Figs. 2 to 20 are magnified 375
diameters ; Figs. 21 to 25, 575 diameters.

“ 2. A corneal facet. This specimen was cleaned in a boiling solution of
potassic hydrate and studied in water.

“ 3. Four pairs of cells from the corneal hypodermis. Around the lower right-
hand pair of cells the outlines of the six surrounding distal retinule are
indicated in part by dotted lines.

“ 4. Four groups of cone-cells. The ommatidia to which the cone-cells be-
long are indicated by the letters a, b,c, andd. These letters are used
to indicate corresponding ommatidia in deeper sections,

“ 5 to 20 represent transverse sections through ommatidia in the various regions
indicated as follows : —

“ 5. The middle region of four cones. Completely depigmented.

The proximal ends of four cones. The re-entrant angle on the surface is

indicated at x. Completely depigmented.
7. Slightly below the proximal ends of the cones.
8. Beyond the distal ends of the proximal retinule.
“ 9. The distal ends of the proximal retinule.
0
1,

“10. The thick distal portion of the proximal retinule.

“ 11, The proximal retinule immediately above the distal termination of the
rhabdome.

“ 12. At the distal ends of the rhabdomes. Completely depigmented.

Parker. — Lobster Eye.

Fig. 13.

The rhabdome between its distal end and middle. In this section the ac-
cessory pigment-cells of each ommatidium are apparently separated
by an intervening space from those of neighboring ommatidia. This
space is probably the result of shrinkage and subsequent rupture.

In the same plane as that shown in Fig. 13. Partially depigmented.

The middle of the rhabdome. The distal retinule of ommatidium c have
been numbered. Completely depigmented.

The proximal end of the rhabdome. The accessory pigment-cells in this
section have probably been ruptured, as in Fig. 13,

In the same plane as that shown in Fig. 16. Partially depigmented.

The proximal tip of the rhabdome. Completely depigmented.

The proximal retinule close to the basement membrane. Partially de-
pigmented.

A diagram of Fig. 19. The proximal retinule of the different ommatidia
have been numbered to correspond with those of ommatidium c in
Fig. 15. The retinule of ommatidium « are tinted pink.

Transverse section of nerve-fibres as they pass through the basement
membrane. The line y z shows the plane of section for Fig. 29; z
indicates the cross-shaped thickening in the basement membrane.
Completely depigmented and stained in Weigert’s hematoxylin.
(See page 4.) xX 575.

“ 22, 23, 24, and 25 represent transverse sections of the fibres in the optic nerve.

Weigert’s hematoxylin. xX 575. The planes at which these sections
‘were taken are as follows: —

“ 922. Directly below the basement membrane. The fibres are in groups of
threes and fours.
« 23. At one fourth the distance from the basement membrane to the optic
ganglion.
“24. At half the distance between the membrane and ganglion.
“ 25. At the surface of the ganglion.
ABBREVIATIONS.

az.n. Nervous axis of retinula. mb. pi ph. Peripheral membrane.
cap. Protoplasmic cap ofcone- n. fbr. Nerve-fibre.
cl. con. Cone-cell. |cell. nl. con. Nucleus of cone-cell.
cl.ms d. Mesodermic cell. nl. crn. as corneal hypodermis.
con. Cone. nl. dst. a distal retinula.
crn. Corneal cuticula. nl. pig. 2 accessory pigment-cell.
crn. h d. Corneal hypodermis. nl. px. . proximal retinula.
cta. Cuticula. omm'. Ommateum.
enc. Brain. pig. dst. Distal band of pigment.
fir'. Fibrillz. pig. px. Proximal band of pigment.
gn. opt. Optic ganglion. r. Retina.
hd. Hypodermis. rhb. Rhabdome.
mb. Basement membrane. rin’. dst. Distal retinula.
mb. i cpt. Intercepting membrane. rtn'. pr. Proximal retinula.
mb.icl. Intercellular membrane. spa.ic/. Intercellular space of retina.

The other abbreviations which occur on the plates are explained in the descrip-
tion of the figures with which they are found.

. —_— as

DARKER - LOBSTER EYE

rin’ dst

a. }

mb. cl. iG \. ee) ;

GHP del

rape ra SE ae

2d rtatdst.

indern

rin px.

mb...

al pig

Pat easton
clmsd.-»-
nfor

“I

MASE concen

clon

clcon
8 4

cl.con. . 2

PUN SE

clon &

g ~~

rlivdst

B, Meisel, lith

ik
»
*, “
* J ‘
’
A i Z ra
Oye Ae
' pen 5,
Lb
7.
me :
i
.
=)
ane
f
’ x
i
i
=
y i
{
i
f
4
=a
f
\
'
if
y
iy
3
: e :
i
|
*
fa
>
‘
‘
- % 1:
e ihe
he A v2 ~S '
oA ~< Bi ; d iva
. . lag Rae ae 91

soa wile a
ka Leese rs as

Fest
a Fai ik 7
wh = + i
A) <ot oing hs a,
= seit
A Re bya
Saas 3

— TT) eg
ae Paes

a) ee
é A SEEK
Mike CAM
— wpe mt Siphy 7,

PARKER. — Lobster Eye.

PLATE II.

Figs. 26 to 36 deal with the histology of the lobster’s eye ; Figs. 37 to 39, with its

development.

Fig. 26. Vertical section through that portion of the eye-stalk where the transi-

eet.

“ 28.

bier

fo.

Sol.

jou

‘C30:

88:

“ 39.

tion from the undifferentiated hypodermis to the ommateum is accom-
plished. The open space z is due to shrinkage. X 81.

A group of the four cone-cells of an ommatidium and the attached cor-
neal hypodermis. Isolated and studied in Muller’s fluid. x 366.
Proximal end of a group of four cone-cells where they separate as fibres
to pass around the rhabdome. Isolated and studied in Miuller’s fluid.

x 365.

Transverse section of the basement membrane. The distal face is upper-
most; the proximal face below. The section is taken in a plane
which would be represented in Fig. 21 by the line z y. x 575.

A rhabdome and its seven surrounding proximal retinule. At the distal
end the free tips of the seven retinule can be seen. At the proximal
end the rhabdome and the four groups of retinule which pass through
the basement membrane are visible. Isolated and studied in chromic
acid 25%. X 200.

An individual proximal retinula. Isolated and studied in chromic acid
Bo% xX 200.

Transverse section of a rhabdome and its surrounding cells. The plane
of section is about half-way between the middle and distal end of the
rhabdome. Completely depigmented with potassic hydrate. x 460.

Oblique section of a rhabdome from the same series of sections as Fig. 32.
The upper end of the figure is distal; the lower, proximal. x 460.

Transverse section of a rhabdome and its surrounding retinule. The
nervous axis of each retinula is distinctly stained. Completely de-
pigmented ; stained with Kleinenberg’s alum-hematoxylin. x 460.

Longitudinal section of a bundle of nerve-fibres extending through the
basement membrane toward the optic ganglion. Depigmented ;
Weigert’s hematoxylin. xX 575.

Optic nerve-fibre. Isolated and studied in Miiller’s fluid. x 575.

Superficial view of a left optic lobe. Enough of the right lobe is drawn
to indicate the position of the median plane (x y). «x is anterior; y is
posterior. Stage A (see page 2). X 280.

Posterior face of a section from a right optic disk cut transversely to the
longitudinal axis of the embryo (see page 34). At xis an angle formed
by the growth of the retinal cells over the undifferentiated ectoderm.
Stage A. x 280.

A section from the optic disk of an embryo in stage B. The plane of
cutting corresponds to that in Fig. 38. In this figure x indicates, as
in the preceding one, the angle between the undifferentiated ectoderm
and the growing retina. X 280.

Tdi lhl =

PARKER —- LOBSTER EYE.

|

i@ Po) \
Boast)
ge Ge )\

\

“O

@ Sere!
oe
S Og

G.H.P del.

‘
i '

= ’ *

A a
: me] +0
ea Pees

AAMAS Ageia

my
ANG
Pamhies.

ge: vs

’ ek

Pau rit

PARKER — Lobster Eye.

Fig. 40.

“é

PLATE III.

All figures on this plate illustrate the development of the lobster’s eye.

41.

42.

45.

44,

A section through the left optic lobe and left half of the supra-cesophageal
ganglion. ‘The plane of section is tangential to that part of the sur-
face of the egg on which the embryo rests. The position of the
median plane is indicated at xy. The surfaces tinted with deeper pink
in the figure represent areas containing nuclei in the specimen; those
in lighter pink, areas in which no nuclei were present. The optic lobe
is divided into two parts by a band of large, faintly colored nuclei,
which, with the smaller surrounding nuclei, are shown in the figure.
To the right of the nuclei the broad tinted marginal area represents
the retina, r. The remainder of the optic lobe gives rise to the optic
ganglion. Stage C (see page 2). X 280.

Posterior aspect of a transverse section of aright optic lobe. The plane
of section corresponds to that in Fig. 88; x is the angle which indi-
cates the separation of the retinal and ganglionic constituents of the
intercepting membrane. Stage C. x 280.

This figure is taken from a region which corresponds to the left-hand por-
tion of Fig. 41. Although from the same set of eggs the embryo from
which Fig. 42 was drawn was somewhat more advanced than that
from which Fig. 41 was taken. At x the proximal band of retinal
nuclei can be seen; at y the distal band is shown. Stage C. x 460.

The superficial layer from the distal band of retinal nuclei; seen from
the external surface of the retina. Stage C. x 460.

The deep layer of the distal band of nuclei. These are seen in opti-
cal section somewhat within the outer face of the retina. Stage C.
x 460.

A transverse section of an optic lobe from a lobster at stage D. The
plane of section corresponds to that of Fig. 88. As in Fig. 40, the
deeply tinted areas were nucleated; the lighter areas were without
nuclei. X 148.

¥

t

De a

PARKER ~- LOBSTER EYE.

mb.topt.

g
"e
tn
a
=
©

y =
i
° re °
; f -
ve 4 — t y
- 5
A i 7 s \ "
4 1 = > ar - =
. & : ;
= = ’
i “Y é
3 : Z
- - on :
a J
= ¢ A
‘= 4 * me
a : =
, _ F -
- i
=r
y a
= .
f 1
’
¥ us 2 1 | r
és 1 3
. q ;
a i a
— a =z - es i ' °
> we ‘ me a.
Pr , 1 ' of
= f { J t'
¥, r
F <
® a .
« » +
) ir x Z a i ‘
| - » o -
7 - i 7 , Ff
> 4 i x i

a

: tphit i (iw
A

—

oy i ae “
F wi eat

4

a | 4 (9 *~

Parker. — Lobster Eye.

PLATE IV.

All figures on this plate, except Fig. 59, illustrate the development of the

Fig. 46.

al

“* 48.

“* 49.

** 60.

lobster’s eye.

A transverse section of an optic lobe at stage E (see page 2). The plane
of section and the method of coloring the figure are the same as in
Fig. 45. x 146.

An enlarged drawing of that portion of the retina which is in brackets in
Fig 46. Stage E. x 460.

A view of the external surface of the retina. The distal ends of four
ommatidia are seen. Stage E. X 460.

A transverse section of four ommatidia in the region of the hypodermal
nuclei. (Compare Fig. 47.) Stage E. x 460.

A transverse section of four ommatidia in the plane which the nuclei of
the cone-cells occupy. Stage E. xX 460.

Longitudinal section of a single ommatidium. Stage F. 460.

Four corneal facets seen from the external surface. Stage F. x 460.

to 58 represent transverse sections of four ommatidia at Stage F. The
numbers on the left side of Fig. 51 indicate the heights at which these
sections were taken, and correspond to the numbers of the following
figures. In Figs. 53 to 58 the magnification is 460.

A transverse section in the region of the corneal hypodermis.

A transverse section through the region in which the nuclei of the cone-
cells occur.

A transverse section in the same plane as the nuclei of the distal
retinule.

A transverse section of the proximal ends of two cones.

A transverse section through the rhabdomes and proximal retinule.

A transverse section of arhabdome from Fig. 57. Fig. 58 was drawn
with a higher magnification than Fig. 57 in order to show the relation
of the proximal retinulz to the segments of the rhabdome. X 640.

A corneal facet from near the periphery of the retina in an adult lobster.
The hexagonal outline is noteworthy. This specimen was cleaned in
boiling potassic hydrate and examined in water. X 280.

aN PL. IV.
PARKER —- LOBSTER EYE. | | | |

-

Al dst.

Bice = nb.con:

he 1@:.....rLdst.

Bot ae 2. a 65
AL pg ©

B. Meisel, lith.

No. 2.— On the Rate of Growth of Corals. By ALEXANDER
AGASSIZ.

WE know as yet comparatively little regarding the rate of growth of
corals under different conditions. Dana has given, in his “Corals and
Coral Islands,” * a résumé of our knowledge on the subject, so that it
is only necessary for me here to refer the reader to his account of the
statements of Darwin, Stuchbury, Duchassaing, Verrill, and others, re-
lating to this subject. .

The specimens figured in this communication have been kindly sent
me by Lieut. J. F. Moser, commanding the U. 8. Coast and Geodetic
Survey steamer “Bache.” They were all taken (as stated by Mr.
Hellings, the cable manager) off the cable laid between Havana and
Key West, in June, 1888, from a portion of the cable repaired in the
summer of 1881; so that the growth is about seven years. Lieutenant
Moser writes: ‘ Taken from the shore end of the International Cable ;
the specimens were taken between the triangular buoys and the outer
reef, the shore end being that portion between Key West and the outer
reef.” The Coast Survey maps indicate a depth of from six to seven
fathoms, and this portion of the cable is most favorably situated as re-
gards food supply, being directly in the track of the main flow of the
tide as it sweeps in and out from the outer reef into Key West Harbor,
and over the flats to the northward.

Some of the specimens belong to different species from those of which
the rate of growth was already known.

Orbicella annularis (Plates I. and Il.) shows a much greater increase
in the thickness of coral formed than the case mentioned by Verrill,
where the thickness formed in sixty-four years was not more than
about eight inches. The specimens sent by Lieutenant Moser grew
to a thickness of two and a half inches in about seven years.

* Coral and Coral Islands, by James D. Dana. Third edition. New York,
1890. (Pp. 125, 253, 418.)
VOL. xX. — NO. 2.

62 BULLETIN OF THE

The Manicina areolata (Plate III.) shows also a very rapid rate of
—inerease. This corresponds to the rate of growth of allied genera
(Meandrina labyrinthica) observed by Pourtales at Fort Jefferson,
Tortugas.

The Jsophyllia dipsacea (Plate IV.) shows a still more rapid increase.

Of course, we are unable to state that these corals began to grow the
first season the cable was laid; but, judging from the favorable locality
in which the corals were found, it is not probable that more than a few
months passed before some of the swarms of pelagic coral embryos which
must have floated past the cable found a place of attachment.

The specimens have all been figured of the natural size.

The figures all show, with the exception of those of Manicina, the
size of the cable to which the corals were attached.

CAMBRIDGE, August, 1890.

MUSEUM OF COMPARATIVE ZOOLOGY. 63

EXPLANATION OF THE PLATES.

PLATE I.

Orbicella annularis Dana (natural size).

The thickness of the coral at the edge of the mass varies from $ to } of an inch.
The greatest height of the mass above the cable is 2} inches.

PLATE II.

Orbicella annularis Dana (natural size).

The thickness of this specimen is very much less than that of Plate I. It varies
at the edge of the mass from } toi of aninch. The greatest height above
the cable is 2} inches.

PLATE IIL. >

Mancina areolata, Ehrenb. ;

1. Seen in profile. The thickness above the cable is one inch.
2. Same, seen from above.
Both figures natural size.

PLATE IV.

Isophyllia dipsacea Ag. (natural size).
The greatest thickness is 24 inches

~

t Li fale
Th FM Ly? ae Vi ea ae a Ha BAG aie rah saath

gt NR ui pages
Mera

PRATER: I.

‘DUDG

8

IaVTONNY

Vv

TIAOIAIG

ee Sith oan al die aa

nie

*

PLATE II.

pupg SIuvt

ONNV WTTHOILAC

—

a

May ae
hs
1
“
'
'

PLATE IIL

Y

Ehrenb

MANICINA AREOLATA

ae J _

7 = : a
ee See =a
3 Sr y a) 5
@ “eel A Tes }
t = ~ =

PLATE IV.

by vaovsaia

VITTAHAOS]

&

No. 3.— Preliminary Account of the Fossil Mammals from the
White River and Loup Fork Formations, contained in the Museum
of Comparative Zoology. Part Il. The Carnivora and Artiodac-
tyla by W. B. Scott. The Perissodactyla by HENRY FAIRFIELD
OSBORN.

THis paper, the second upon the Fossil Mammals of the Museum of
Comparative Zodlogy, is a continuation of the one published by the
writers? in August, 1887, upon the White River Mammalia, and in-
cludes a number of additions to and corrections of the results there
described. It is, however, especially devoted to a consideration of the
upper Miocene or Loup Fork mammals collected in Nebraska by Messrs.
Garman and Clifford, and in Kansas by Mr. Sternberg. The specimens
from these different localities exhibit a considerable range of specific
variation.

The Loup Fork species here described have for the most part been
long established, but these collections add much to our knowledge, and
enable us to determine very fully the structure of forms which have
been known hitherto only from fragments. Of such new observations
we may mention: (1) the determination of the foot structure of Meryco-
choerus ; (2) of Blastomeryx; (3) the restoration of Cosorysx ; (4) discovery
of the mandible of Alurodon hycenordes ; (5) the discovery of an exceed-
ingly large feline animal ; (6) observations upon the molars of the equine
series ; (7) the manus and pes of Aceratherium; (8) the skeletal char-
acters and restoration of Aphelops fossiger ; (9) the homologies of the
elements of the molar teeth in the rhinoceroses; (10) the brain char-
acters of Aphelops and Mesohippus ; (11) the discovery of a Loup Fork
species of Chalicotherium.

We have again to express our thanks to Dr. F. C. Hill, Curator of
the Geological Museum at Princeton, for his skilful excavation and
mending of the specimens, and to Mr. R. Weber for the very accurate
series of drawings which accompany this paper.

GEOLOGICAL MusEuM, PRINCETON, N. J., July 8, 1890.
1 The authors, as initiated in their Memoir upon the Uinta Mammalia, have

divided the subjects for their present and future joint papers.
VOL, XX.—NO. 3. 5

66 BULLETIN OF THE

CARNIVORA.

CANIDA.

ZELURODON, LeIpy.
(Syn. Epicyon, Leidy. Canis, Leidy, in part. Palhyena, Schlosser.)

The dogs of this genus are the most abundant of the Loup Fork Canida, and,
as their relations and systematic position have been very generally misunder-
stood, it will be well to describe them in some detail. The special peculiarity
of the genus is to be found in the development of a large anterior basal lobe on
the superior sectorial, as in the cats. The postero-internal cone (metaconid)
of the lower sectorial is much reduced, and in some species almost disappears.
The talon of this tooth is rather short, and consists of an internal and external
cone or tubercle, being of the basin-like character. The premolars are remark-
ably heavy, and possess well developed basal conules. There are four well
marked species of this genus, of which the best known is

Atlurodon szevus, Lerpy (Cope).
(Syn. Canis sevus, Leidy. -dlurodon ferox, Leidy. -dlurodon sevus, Cope.)

This species is characterized by the very small size of the internal cusp of
the upper sectorial, and by the nearly
~ : straight and slender mandible; the in-
cisors are rather smail, and the first
upper molar is very large and subtri-
angular in shape. The skull as fig-
ured by Cope (American Naturalist,
XVII.) presents a rather short, nar-
row muzzle, and is in general quite
bear-like in appearance. Notwith-
standing its peculiarities of dentition,
this animal is an unmistakable dog,
and the structure of the skull, ver-
tebra, limbs, and feet is character-
istically cynoid. The metapodials are, however, somewhat less elongated
proportionally than in existing dogs.

‘“
ie u
Ss sg 2

Ficure 1.— “lurodon sevus, fragment of
right superior maxillary X 4.

4@lurodon Haydeni, Lripy.
(Syn. Canis Haydeni, Leidy. Epicyon Haydent, Leidy.)

This species is very large, and is remarkable for the short, massive mandible
and the strong upward curvature of the posterior portion of the alveolus, so
that the inferior tubercular molars may almost be said to be inserted in the
ascending ramus. In Dr. Leidy’s type of the species (Ext. Mam. Fauna, Dak.
and Neb., Plate I. fig. 10) the third lower molar is inserted by two fangs, and in

MUSEUM OF COMPARATIVE ZOOLOGY. 67

the Cambridge specimen by only one; but this does not appear to be a constant
character. The postero-internal cusp (metaconid) of the lower sectorial is
reduced to a rudiment, and the talon is much shortened antero-posteriorly.
Pm. 1 and 2 are relatively quite small, while pm. 3 and 4 are quite high and
massive.

f@lurodon Wheelerianus, Corr.
(Syn. Canis Wheelerianus, Cope. -d!lurodon Wheelerianus, Cope.)

This species is nearly as large as the preceding one, but differs from it (1) in
the much less strongly curved alveolar region, and (2) in the very large size of
the external upper incisor, which at the base is nearly as large as the canine.
The species is represented in the collection by the facial region of a very old in-
dividual, the teeth of which are worn down to mere stumps. The face appears
to be proportionately longer than in AZ. sevus, the orbit lying somewhat farther
back; it is also very deep, and encloses an unusually large nasal chamber.

AMlurodon hyzenoides, Core.

FIGURE 2. —Mandibles of 4lurodon X 3. A. 4. hyenoides; B. 4. Haydeni;
C. 4. sevus.

This, the smallest species of the genus, has been described by Cope from the
superior dentition: “ The second and third premolars are robust and somewhat
swollen at the inner base. Each has a short heel, but no median posterior lobe.
The principal lobe is robust, in the third [pre]molar as wide as long at the base.

68 BULLETIN OF THE

The internal anterior lobe of the superior incisor [sectorial] is very large, and
its apex is distinct from the inner side of the rest ot the tooth. It is relatively
larger than in Crocuta brunnea. . . . The first true molar is somewhat wider
near the inner extremity of the crown than at the external extremity.’’ (Bul-
letin U.S. Geological and Geographical Survey of the Territories, Vol. VI.
p. 388.)

The Cambridge collection contains a mandible which should almost certainly
be referred to this species. It is proportionately short, stout, and of nearly
uniform depth, not tapering anteriorly as in 4. sevus; the symphysis is very
obliquely placed and the chin abruptly rounded, giving the jaw a somewhat
cat-like appearance. The first and second premolars are small, and the latter
is implanted by a fang which is but imperfectly divided into two; the third
and fourth premolars are low but strong, and differ from the corresponding
teeth of the other species in the presence of a small anterior basal cusp. The
sectorial is large, and has a well developed metaconid ; the talon is obliquely
worn upon its outer side, showing a different mode of opposition of the teeth
from that which obtains in 4!. sevus. The incisors are very closely crowded
together, and the median one is pushed very far back out of the line of the
other two.

MEASUREMENTS.
m. m.
Length, inferior molar series 032 Blade of sectorial (ant. post.) .013
af «premolar series .031 Talon sf as 006
“ sectorial (m. 1) 019

? Ailurodon ursinus, Core.
(Syn. Canis ursinus, Cope.)

Some large specimens agree best with the figures and descriptions of the
Canis ursinus, but they are so damaged as to render any final reference of them
impossible. Indeed, it is by no means clear that the species here named can
be regarded as distinct.

The systematic position of lurodon has been somewhat disputed. Leidy
placed it provisionally among the Felide (op. cit., pp. 68 and 367). Cope,
though referring it to the Canide, has regarded it as the forerunner of the
hyenas. “I nevertheless suspect that this genus is the ancestor of the Hye-
nide, through the intermediate forms Ictitherium and Hyenictis.” (American
Naturalist, Vol. XVII. p. 244) Professor Cope has, however, informed us
that he does not attach much importance to this view. Schlosser has adopted
the same opinion, but believes that 4%. sevus should be generically separated
from the other species. “ Der Cands seevus, Leidy, wird von Cope zur Gattung
Ablurodon gestellt, indess offenbar ohne hinreichenden Grund, denn sowohl der
Schadelbau, als auch die Beschaffenheit der einzelnen Zahne, namentlich des
oberen Pr, 1 sprechen sehr fiir die Zugehérigkeit zu den echten Caniden, wah-

4
,

— ee

MUSEUM OF COMPARATIVE ZOOLOGY. 69

rend die beiden iibrigen Hlurodon-Arten sich héchst wahrscheinlich als Vor-
laufer der Hyanen erweisen werden.” (Beitr. z. Palaont. Oesterr. Ungarns,
Bd. VIII. p. 252.)

In these statements Schlosser has been misled by the fact that the specimen
of Atlurodon sevus which was figured by Cope is very old, and the teeth so
much worn down that the anterior lobe of the upper sectorial is hardly distin-
guishable. The specimens before us demonstrate clearly that Cope’s reference
of the species is correct, and that A’. Wheelervanus and hyenoides cannot be
generically distinguished from it. The only characters of 4lurodon which in
any way resemble those of the hyznas are (1) the massive premolars, (2) the
presence of an anterior basal cusp on the upper sectorial, and (3) the reduction
(in some species) of the postero-internal cusp of the lower sectorial. These
resemblances are obviously merely analogical, and are of far less importance
than the characters of the skull and limbs, which are distinctively cynoid.
These animals are genuine dogs, if somewhat peculiarly modified, and to regard
them as ancestors of the hyenas is to ignore the close connection between the
latter and the viverrines, besides being improbable on geographical grounds.

CANIS.
? Canis vafer, Lerpy.

This small alopecoid is represented in the collection by a mandible with
broken teeth and some other fragments. It agrees almost exactly with Leidy’s
type (op. cit., Plate I. fig. 11), except that the diastema between the canine and
pm. 1 is shorter. The small size of the sectorial places the species in the mi-
crodont division of the alopecoid series. M. 2 is very elongate antero- posteriorly,
and m. 3 is implanted by two fangs. The mandible is very slender, much

Ficure 3.—? Canis vafer x 3, A. First superior molar; B. Mandible.

curved, and non-lobate. The first upper molar is nearly quadrate in shape, the
metaconule being almost as large as the protocone, and placed upon nearly the
same antero-posterior line. The cingulum is internally very greatly enlarged
and thickened, and is disposed symmetrically around the inner side of the
crown, instead of being confined to the postero-internal angle, as is usual
among the recent Canide.

70 BULLETIN OF THE

The head of the radius is less transversely extended and more discoidal than
in recent dogs, apparently indicating the retention in some degree of the power
-of supination.

FELIDA.
FELIS.

? Felis maxima, sp. nov.

This species is founded upon a well preserved humerus from the Loup Fork
of Kansas. The chief peculiarity of the specimen is its great size, which very
much exceeds that of any living

ae m feline. In construction it closely

é 7 A “ resembles the humerus of the
Fh + \ lion, with some minor differences.
i 1} | The external tuberosity rises high
“sii? = above the head, and is somewhat

less rugose; the deltoid ridge is
exceedingly broad and massive,
and descends far down upon the
shaft; the outer condyle for the
capitellum of the radius is less
decidedly convex; the internal
epicondyle is very prominent and
massive, and is surmounted by a
large epicondylar foramen. ‘The
presence of this epicondylar fora-
men shows that the specimen be-
fore us cannot be referred to
Smilodon, for the humerus of S.
necator figured by Cope (Ameri-
\ can Naturalist, Vol. XIV. p. 857)
a \ has no such foramen. The su-
) pinator ridge is somewhat broken,
but it appears to have been pro-
F _ portionately less robust than in
IGURE 4.— Humerus of ? Felis maxima x 4; é :
internal and anterior views. the lion. The following table
will exhibit the great size of this
specimen. The measurements of the humerus of Smilodon are taken from
Cope’s figure.

MEASUREMENTS.
Smilodon necator. Felis leo. F.? maxima.
m. m. m.
Humerus length . . . wna 92> abe 313 429
as width of distal aaa | eG .O87 .054 072

ee antero-posterior diameter gieimal a a -088 118

MUSEUM OF COMPARATIVE ZOOLOGY. 71

Another cat, perhaps a smaller individual of the same species, is represented
by a phalanx of the median row, which is of the characteristically asymmetrical
shape, so as to allow the retraction of the claws. It agrees best in shape with
the median phalanx of the fourth posterior digit of the lion, but is much larger,
measuring 37 mm. in length, and the proximal end is 20 mm. wide; in the
lion these dimensions are 27 and 14 mm.

A third very large feline is indicated by the proximal end of a radius from
the Loup Fork of Nebraska: it agrees closely in shape and size with that oi
the lion.

Still another cat is represented by the third and fourth
metatarsals from the same horizon and locality. The
mode of interlocking, the shape and character of the prox-
imal articular surfaces, are very cat-like, but the bones are
short and massive, showing strikingly different propor-
tions from those to be observed in the recent forms.

MEASUREMENTS.
Felis leo. ?
m. m.,
Metatarsal Ill., length . . .. . 119 .089
a “ width proximal end . .021 022
s By ERE (A iho old 093
a “¢ width proximal end . .015 018

These specimens show that the number of cats occur-
ring in the Loup Fork formation is much more consider-
able than has hitherto been supposed. Unfortunately,
however, these remains are not associated with teeth, so
that they cannot be referred to their proper genera and

species. FraureE 5.— Third
and fourth metatarsals
of unknown feline x 4.

? Pseudalurus intrepidus, Lerpy.

This species is doubtfully indicated by a humerus, lacking the proximal end,
which is distinctly feline in character, but remarkable for the very weak de-
velopment of the supinator ridge.

MUSTELID A.

Carnivora of this family are not certainly known to occur in any American
formation older than the Loup Fork, and they are very rare even in that forma-
tion. The mustelines are represented in the collection by only a fragment of a
lower jaw supporting pm. 4. In the absence of the molars, it is impossible to
determine to what genus this specimen should be referred ; but it would appear
to agree best with the Mustela parviloba of Cope.

(p- BULLETIN OF THE

ARTIODACTYLA.
OREODONTIDA.

MERYCHYUS, Leipy.

Merychyus elegans, Lerpy.

The genus Merychyus is abundant in the Loup Fork, but has been known
hitherto chiefly from the dentition. The Garman collection contains some
portions of the skeleton, which are therefore of great interest. These speci-
mens show that the genus has departed but little from the type of the fam-
ily, Oreodon, but present, nevertheless, some important approximations to the
ruminants.

The ulna and radius show no tendency to coalesce, and the former has the
shaft considerably more reduced than in Oreodon. The radius differs in many
ways from the ordinary oreodont type; the groove for the intertrochlear ridge
of the humerus is narrower, the inner flange of the head smaller and less ob-
lique, the outer larger and more concave, and the upward projection from the
anterior edge much better developed, almost as in a true ruminant. The shaft
is broader and more flattened, the walls much thinner, and the medullary cav-
ity larger. The distal end is less expanded and thickened, and the tendinal
sulcus barely indicated. The facets for the scaphoid and lunar are very dis-
tinctly separated ; the former is shaped much as in Oreodon, but more deeply
incised and more obliquely placed; between the two is a very deep notch,
which penetrates from the posterior side through nearly half the thickness of
the radius. This notch is indicated in Oreodon, but is not nearly so deep. The
only bone of the manus which is preserved is the magnum, the shape of which,
however, shows that it has moved entirely beneath the scaphoid, and has a
deeply concave facet upon its ulnar side which embraces the side of the lunar
almost in a semicircle. No facet for the second metacarpal is to be seen upon
the radial side of the magnum, whence it follows that the third metacarpal was
in contact with the trapezoid, and that an adaptive reduction of the manus had
commenced, which, except in Merycocherus, is unknown in other oreodonts.

Of the tibia only the distal end is preserved, and this portion differs but little
from that of Oreodon; the astragalar facets are somewhat more deeply grooved
and of more unequal size, and the fibular surface is deeper, as if the distal end
of the fibula had commenced to wedge itself between the tibia and the calca-
neum. The pes is higher and more slender than in Oreodon, but shows few
important changes. The middle and external cuneiforms are united, but the
limits of the two elements are plainly shown by the step cut in the distal sur-
face. Metatarsal II. occupies the whole of the distal surface of the mesocunei-
form, and abuts against the side of the ectocuneiform, while metatarsal III. is
confined to the latter alone. Merychyus thus presents the curious condition of
an adaptively reduced manus and an inadaptively reduced pes. The metatar-

MUSEUM OF COMPARATIVE ZOOLOGY. 73

sals are relatively very long and slender, more so than in any other member of
the entire family, though far from reaching the elongation seen in the true
ruminants; the lateral digits are especially slender, though proportionately as
long as in Oreodon. The phalanges are likewise long and slender, but the un-
guals are still plainly of the true oreodont pattern.

Another specimen of this species is the skull of a very young animal with
the milk dentition, which shows some interesting differences from that of Oreo-
don. In the latter genus, as in the Tragulina and the older selenodonts gener-
ally, the third upper milk molar, 4.8, is of a triangular shape, having only the
posterior crescents developed, with the anterior portion elongated and tren-
chant, while Merychyus agrees with the true ruminants in the fact that this
tooth is like a permanent molar, consisting of four crescents,

Of all known oreodonts Merychyus is perhaps the one which most closely ap-
proximates the true ruminant type. This is apparent in the elongated and
more or less prismatic crowns of the true molars, in the increased size and com-
plexity of pm. 2, in the character of the milk dentition, in the structure of the
long bones of the skeleton with their large medullary cavities and thin walls,
as well as in the adaptive method of reduction assumed by the manus,

MERYCOCHG:RUS.

‘Merycochcerus cenopus, Scort.

The type of this species is the specimen consisting of a beautifully preserved
manus and pes contained in the Garman collection from the Loup Fork of
Nebraska, which unfortunately are not associated with teeth. It is therefore
possible that they may belong to some already described species, though they
do not agree well in size with any of them.

The foot structure of this genus has been briefly noticed by Cope, who states
that the feet are tetradactyle, and that “the os magnum is entirely beneath the
scaphoid, and there is a distinct trapezium. The posterior foot is constituted as
in Hucrotaphus.” (Proc, Am. Ass. Adv. Sci., 1884, p. 484.) The manus in Mery-
cocherus agrees much more closely with that of Merychyus than with that of any
other genus of the family. The carpus is higher in proportion to its breadth
than in Oreodon, and very much higher as compared with the height of the meta-
carpus, thus giving the manus very different proportions in the two genera.
When the individual carpal bones are compared, we find many differences of de-
tail. The scaphoid has lost its cuboidal shape and become higher, narrower, and
deeper (antero-posteriorly); the proximal surface has a convex anterior ridge
which is very oblique, rising to a high point on the ulnar, and dying away on
the radial side. The distal surface is anteriorly much narrower than in (reo-
don, broadening however behind; the magnum facet is much larger, and the
trapezoidal smaller and more lateral. The trapezium appears not to have been
in contact with the scaphoid. The Junar is very peculiar, especially in the great
downward prolongation of the beak-like process, which is wedged in between
the unciform and magnum, and almost reaches the metacarpals. The proximal

74 BULLETIN OF THE

surface has not the simply convex shape seen in Oreodon, but rises high towards
the ulnar side, and is much depressed on the radial. The distal surface is-occu-
pied by the long concave and obliquely placed facet for the unciform, that for
the magnum being altogether fateral, This is the culmination of a tendency
already noticeable in Protoreodon of the Uinta formation, the earliest known
member of the family, and more marked in Oreodon, namely, the movement of
the magnum away from the lunar and under the scaphoid. In Merycocherus
(and Merychyus) the lunar does not rest upon the magnum at all, touching it
only laterally. The cuneiform is much like that of Oreodon. The pisiform
is very different from that seen in the earlier genera of the family, and shows a
tendency to assume the form characteristic of the pigs, though relatively much
larger than in those animals. Compared with that of Oreodon, it is shorter,
heavier, and especially much more expanded at the free end. No trapezium
is preserved in connection with this specimen, and as no facets for it are clearly
distinguishable on the other carpals, it may not have been developed. The
trapezoid is very different from that of Oreodon, in being very much higher, nar-
rower, and deeper; the facet for the scaphoid is oblique and almost as much
posterior as superior; behind, the bone is drawn out into a projecting process,
not abruptly truncated by the facet for the trapezium, as it is in Sus, The sig-
nificant characters of the trapezoid are shown by the distal surface, which is
constituted as in the pigs, having a large facet for mc. II. and a small one for
me. III.; in the pig the two facets are of nearly the same size. The magnum
is very peculiar; as Cope has shown, it lies entirely beneath the scaphoid and
internal to the lunar; its proximal surface is occupied by a large, slizhtly con-
vex facet for the scaphoid, very different in shape from the same facet in Oreo-
don, as it lacks the abruptly rounded posterior rising ; the ulnar side is even
more deeply concave than in Merychyus, encircling the convex lunar. The
unciform differs but little from that of Oreodon, except that the proportions of
the proximal facets have changed, that for the lunar being considerably the
larger.

The metacarpals are relatively much shorter and broader than in Oreodon,
the lateral digits are somewhat reduced, though not very much, while the me-
dian ones have greatly increased in thickness. In proportions the metacarpus
is quite like that of Sus, though as in all the oreodonts the keels of the distal
trochleze are confined to the palmar surface. Mc. II. is short, stout, and com-
pressed ; it articulates by a narrow surface with the trapezoid, but is excluded
from the magnum. Me. III. is very suilline in appearance, but its proximal
end is not much extended transversely ; on each side of the magnum surface is
a facet for the trapezoid and unciform, the latter considerably the larger, while
in the pig they are of nearly equal size. Mc. IV. is of about the same breadth
and thickness as me. III.; its proximal end is transverse, as in the pig, not
oblique, as in Oreodon. Me. V. is not preserved in the specimen, but the facet
for it on the unciform shows that its head was flatter than in Oreodon, and that
it did not rise so much upon the external side of the unciform.

The phalanges resemble those of Oreodon, except for their greater stoutness,

MUSEUM OF COMPARATIVE ZOOLOGY. 75

and, as compared with the metacarpus, their greater length, for the three pha-
langes of the fourth digit are together as long as the metacarpal. The unguals
are of the same general shape in the two genera, but broader, more depressed,
and with the ends less pointed in Merycocherus.

The pes in the species of Merycocherus from the John Day and Deep River
beds differs less from that of Oreodon than does the manus, but the species be-
fore us appears to show an important departure from that type. The astragalus
is shorter, broader, and more massive than in Oreodon, and the distal trochlea
has a broader surface for the cuboid. The calcaneum is not preserved. The
cuboid is low and broad; the surface for the astragalus is broader than that for .
the calcaneum, reversing the proportions seen in Oreodon; the calcaneal surface
is also of a different shape, as it does not project outwards, and its external mar-
gin is straight, not rounded; unlike the pigs, this facet is not notched on its
outer margin. The astragalar surface is not so deeply concave as in the White
River genera, and another difference lies in the presence of a broad shallow
groove which separates the articular surface into anterior and posterior portions.
The peroneal sulcus is shallow. The distal end is almost entirely taken up by
the large facet for mt. 1V., that for mt. V. being very small and more lateral
than distal; in Oreodon it is entirely distal. The navicular does not. differ
sufficiently from that of the earlier genera to require description. The ento-
cuneiform is relatively large, and in general resembles that of Oreodon, but has
a larger bearing upon mt. II. The ento- and meso-cuneiforms are missing, but
they were doubtless ankylosed together as in all the other members of the
family.

As a whole, the tarsus has changed in an opposite sense to the changes in
the carpus, having become lower and broader, while the carpus has become
narrower and remarkably high.

The metatarsus is suilline in generai appearance ; the median digits are short
and massive, while the laterals are reduced, especially in length, being not only
proportionally but absolutely shorter than in Oreodon Culbertsoni. Mt. II. has
an exceedingly small surface for the mesocuneiform, but the head is not oblique
asin Sus. Mt. III. has a minute facet upon the tibial side of the head, which
appears to encroach upon the mesocuneiform; and if this is the case, we have
here the beginnings of an adaptive reduction of the pes, which is not known to
occur in any other member of the family. Except for its heavier propor-
tions, mt. IV. is like that of Oreodon; mt. V. has a smaller, more concave and
obliquely placed facet for the cuboid than in the latter genus.

Merycocherus and Merychyus thus agree with each other, and differ from
other oreodonts in which the foot structure is known in the adaptive reduction
of the manus, and it is interesting to note that this adaptive method has been
independently assumed in several distinct lines of artiodactyles, e. g. the true
ruminants, the pigs, and the camels. A study of the oreodonts shows that they
are not closely connected with any existing artiodactyles, and it is difficult to
see how the same result could be so often reached independently, unless it be
the effect of the similar mechanical conditions to which the extremities are
subjected.

76 BULLETIN OF THE

MEASUREMENTS.

Merycocherus cenopus. Merychyus elegans.
m.

Garpus, height. 5. iiss. acne (OSU
Shes ROMORORGED 4 1'5°1)\5) Lec ts a Bg tO
Lunar, height)... . ¢. MMMEAR oy BORD
ss breadth proximal a Petal) tts) OLA
Metacarpal IT., length... .; #2) 8) 049

oe ‘¢ breadth proximal a Shho: te oS
e TPT, Temigthi ys ese a ae si > O64
a “breadth proximal a teat a,
eg PW, (demote Sie. eememeie, 8 oe «Ooo,
Phalanges of IV. digit, length. 2)... . 057 m.
Astragalus, eighty: |5:4hehsss\\ «nee pe yao .027
breadth) c)hea\s tah)-4s)- aed ieee 014
Metatarsal TL, length). «V0 Ytten @ ox GOD0 .056
a TT yy 88. ujondat Chose et he sa .067
< “breadth proximalend . . .015 .009

ee EV, \deni@th. 2), 3) ky iaaeaneteu 0G
GP Re ey en ered oT

SUID.
DICOTYLES.

Several species of peccaries have been described from the Loup Fork beds.
The Clifford collection contains two jaw
fragments, apparently of different species.
One of these differs from existing species
in the fact that the last lower premolar
is of much simpler construction than the
Fieure 6.— Fragment os mandible of molars, and more perfect specimens would

tact ean probably show that this represents a dis-
tinct genus; but it would be premature to propose a name for it in the absence
of more complete material.

GHLOCIDA.
BLASTOMERYX, Core.

This genus of true ruminants is abundantly represented in the collection.
The type species, B. gemmifer, Cope, is from the Loup Fork, and differs from
the closely allied Cosoryx chiefly in the brachyodont dentition. The later
described species from the John Day formation not improbably belong to Pa-
lwomeryx, from which Blastomeryz is distinguished by the absence of the char-
acteristic fold on the lower molars, and the greater narrowness and compression
of the molar crowns.

——— > ee. ee

MUSEUM OF COMPARATIVE ZOOLOGY. ye

THE SKULL.

A little of the superior wall of the cranium is preserved in one of the speci-
mens, which probably belonged to a young animal, as the horn is a mere rudi-
ment. The frontals are extended back of the orbits and form a considerable
part of the cranium, but they are shorter than in Carvacus. Distinct though
not prominent ridges converge from the back of the orbits, and probably unite
behind in a sagittal crest, though as this part of the cranium is broken away
the existence of a sagittal crest cannot be certainly atlirmed. If present at all,
it must have been a mere indication. The orbits are large, and have sharp su-
perior borders. In Cosoryx and Antilocapra the horn arises directly over the
orbit, and the same is probably true of the John Day species of DBlastomeryz ;
but in the Loup Fork species of the latter genus the base of the antler has
shifted its position somewhat, so as to spring from the posterior portion of the
orbit, and it is also directed obliquely backwards, which apparently is the be-
ginning of a process which results in the position of the pedicels observed in
Cariacus. So much of the frontals as is preserved shows no trace of any sinuses,
only the ordinary diploetic structure of the cranial bones. The bases of the
antlers are much farther apart than in the deer, and are not connected by any
intervening ridge. The coronal suture is nearly straight. As usual in rumi-
nants, the parietals have coalesced into a single large bone, which clearly makes
up most of the roof of the cranium. In the anterior portion the supra-orbital
ridges are carried over from the frontals and converge to a point. Only the
anterior part of the parietal is preserved. The inferior surface of this, and of
the frontals as well, is deeply channelled by the winding and coniplex cerebral
convolutions. Of the dentition we possess only a superior molar, and several
inferior molars of a smaller species. The upper molar is very cervine in struc-
ture. The crown is brachyodont, and nearly as broad as long, while in Cosoryx
it is strikingly narrow as well as hypsodont. The valleys are deeper and the
external crescents more flattened than in Paleomeryx, while the internal cres-
cents are somewhat simpler. The cingulum has almost disappeared, but a small
basal pillar occurs between the inner lobes, as in many deer. The lower mo-
lars have low and rather narrow crowns; the valleys are shallow, and disappear
after a comparatively short time of attrition. The cingula are but faintly indi-
cated. As compared with the molars of Carzacus, those of Blastomeryx are
simpler in the uncomplicated inner crescents of the upper teeth and the shal-
lower valleys.

THE SKELETON.

The scapula is much like that of Cosoryx, though with some difference, the
neck is more contracted, the coracoid more prominent, and the acromion more
overhanging. The glenoid cavity is small, nearly circular in shape, and quite
deep ; the anterior or coracoid border is thin and curved, the glenoid border
much thickened and nearly straight. The spine is not very high and divides
the blade into unequal fossee, the postscapular being much the larger, and the
acromion overhangs the neck, but does not nearly reach the margin of the
glenoid cavity.

78 BULLETIN OF THE

The humerus. The proximal end of this bone is not preserved in any of the
specimens ; the shaft is rather short and slender, and shows a distinct sigmoid
curve; an indistinctly marked deltoid ridge runs for some distance down the
‘shaft. The distal end is moderately expanded and thoroughly cervine in ap-
pearance ; the inner condyle is much the wider, and the intercondylar ridge is
sharp and prominent. The anconeal fossa is deep and narrow, but does not
perforate the bone. A moderate internal tuberosity forms a downward projec-
tion at the postero-internal angle. The ridges for muscular attachment are but
feebly developed. .

The radius is entirely distinct from the ulna, no co-ossification between the
two occurring at any portion of their length. The proximal end is much ex-
panded, as this bone carries nearly the entire weight of the fore limb, and covers
the whole of the distal end of the humerus, the ulna being confined to the pos-
terior aspect. The groove for the intercondylar ridge is deep, and emarginates
the anterior and posterior edges. Two small facets for the proximal end of the
ulna occur on the posterior side, the inner one very small, the outer larger and
quite deeply concave. The shaft is long, slender, and considerably flattened,
forming in section a transversely directed oval. The distal end is expanded and
thickened, and is deeply grooved in front by the tendinal sulcus. On the ex-
ternal side there is a strong and roughened extension, which fits into a corre-
sponding depression in the side of the ulna. The facets for the carpus are
separated by a strongly defined ridge, and are placed very obliquely to the axis
of the bone. - That for the scaphoid is deeply concave in front and as markedly
convex behind, and is continued well up on the posterior side of the bone.
This portion has the greatest antero-posterior diameter. The lunar facet is
smaller and less deeply incised. External to it is a small oblique surface which
articulates with the cuneiform.

The ulna is much reduced, though still retaining its independence. The ole-
cranon is rather short and much compressed, though of considerable fore and
aft depth. As the radius has usurped the entire distal end of the humeral
trochlea, the sigmoid notch of the ulna is shallow, and the internal radial facet
has become minute, though the external one forms quite a protuberance. The
shaft is exceedingly slender and compressed ; for most of its length hardly more
than a thread of bone. The distal end is somewhat expanded, though very
small, and deeply excavated on the inner side for the protuberance of the ra-
dius. The cuneiform facet is saddle-shaped, and sends down a well marked
process on the outer side.

The pelvis, so far as it is preserved, is much like that of Cosoryx, though the
ilium has a longer neck, and was appprently even less everted than in that
genus. The ischium is rather short.

Little of the femur is preserved. The head is small and compressed, and
rises little above the ridge connecting it with the great trochanter. The
rotular trochlea is very broad, with rounded and somewhat more prominent in-
ternal edge; the outer edge is lower and sharper. The condyles are rather
small and widely separated ; above the inner one there is a small but distinct
plantaris rugosity. |

——_— SS ae

MUSEUM OF COMPARATIVE ZOOLOGY. 79

The ézbza has a large trihedral head, with large external and small internal
surfaces for the femoral condyles, and prominent bifid spine. The cnemial crest
is very well developed, and just posterior to it on the external side is a very
deep tendinal sulcus. The shaft is quite long and stout, with oval section and
broad distal end. ‘The surfaces for the astragalus are deeply inciséd, and the
external one is somewhat the larger. The tongue is broad and thick, corre-
- sponding to the breadth of the groove in the astragalus. The internal malle-
olus is very long, and forms a tongue-like projection from the antero-internal
corner.

The fibula is as completely reduced as in any ruminant. The proximal: end
is ankylosed with the tibia, where it forms a short sharp process. The distal
end is not represented in any of the specimens, but from the structure of the
tibia it is plain that it was a small nodule wedged in between the distal end of
the tibia and the fibular process on the caleaneum. Between the two distal
fibular facets of the tibia is a groove for the reception of the rudimentary shaft.

The carpus is like that of recent deer; the bones of the proximal row are
high and narrow, those of the distal row low and broad. The scaphoid is deep
antero-posteriorly, and broader in front than behind,
where it is much narrowed by the great lateral exten-
sion of the lunar; the proximal surface is directed
very obliquely backwards and inwards, and is deeply
incised so as to form a very firm interlocking joint
with the radius; this facet is divided into a strongly
convex anterior portion, and an as strongly concave
posterior portion. The lunar is curiously shaped ; it
is broadest in front and behind, and contracted in the
middle ; the anterior surface is transverse, the poste-
rior very oblique. The radial surface is directed ob- |
liquely inwards parallel to that of the scaphoid. The fioure 7. —Carpus of
distal surface is divided nearly equally between the Blastomeryx, nat. size,
magnum and the uncitorm, which meet at a very open Dee Age
angle, and the “beak” is barely indicated. The inner face of the lunar is
nearly vertical, the outer very oblique, as the upper part of the bone is consid-
erably wider than the lower, corresponding to the reduction in width of the
proximal portion of the cuneiform. The latter bone has a broad distal surface
and small saddle-shaped facet for the ulna, which is prolonged well down upon
the postero-external surface. The cuneiform rises above the level of the lunar,
and presents a small oblique surface, which articulates with the radius. The
pisiform facet is high and very narrow. The trapezium is not preserved in any
of the specimens, but its presence is demonstrated by a small facet on the pos-
tero-internal angle of the trapezoid. It was obviously very small and did not
reach the scaphoid. The trapezoid and magnum have coalesced ; the proximal
surface of the compound bone is mostly occupied by the scaphoid, the facet for
which is low and concave in front, rising behind to a low broad convexity; the
distal surface is nearly flat. The unciform is narrower and higher than the

80 BULLETIN OF THE

trapezo-magnum ; its proximal surface is divided obliquely into facets for the
lunar and cuneiform. The unciform projects below the level of the trapezo-
magnum, and so presents upon the radial side a small facet for the corresponding
projection of me. III.

The metacarpus consists of the IIT. and IV. metacarpals, which have coalesced
to form a cannon-bone, and the II. and IV., which are very slender, styliform
bones, perhaps interrupted in the middle of the
shaft. The cannon-bone is more slender than
the anterior one of Cosoryx figured by Cope
(Am. Nat., 1881, p. 547). The proximal sur-
face is unequally divided, considerably the
larger part belonging to me. III., which pro-
jects above the level of me. IV., and so comes
into contact with the unciform. This arrange-
ment occurs also in Dremotherium (see Gaudry,
Enchainements, Fig. 142), and to a much less
degree in Antilocapra. The shaft of the can-
non-bone is broader and flatter proximally, be-
coming narrower and more rounded distally,
and the distal trochleze are completely encircled
by sharp keels, as in existing ruminants. On
the posterior side of the proximal end are two
small facets, probably for the heads of the lat-
eral digits. At all events, much of the shafts
of the metacarpals II. and V. were preserved
in the shape of very slender and compressed
splint bones,

The phalanges are long and slender, and so
asymmetrical as to produce a decided conver-
gence of the toes. The proximal ends of the
first row are deeply grooved for the keels of the
metapodials, but are not emarginated in front
as are those of the recent Pecora. The ungual
phalanges differ from those of the deer and an-
Bicone (8) — Marius “and pes of telopes only in their greater slenderness. The

Blastomeryx % 2. phalanges of the lateral digits are of about the
same proportionate size as in existing Cervide.

The tarsus is also cervine in character, and differs little from that of Pale-
omeryx (Cervus) Flourensianus as figured by Fraas (Fauna von Steinheim,
Taf. VIII. fig. 24). The astragalus is high, narrow, and deeply grooved, and
the distal end shows hardly more than an indication of the ridge which passes
between the navicular and cuboid. The calcaneum is long and much compressed,
though with considerable depth, antero-posteriorly in the lower third (in Pale-
omeryx the caleaneum is thicker and more rounded); the cuboidal facet is
narrow, pointed in front and quite concave from before backwards; the sus-

MUSEUM OF COMPARATIVE ZOOLOGY. 81

tentaculum is short, but broad and thick, and the fibular facet is high, but short
and narrow. As compared with the caleaneum of recent deer, that of Blasto-
meryx has less antero-posterior diameter below, and a somewhat less thickened
tuberosity at the free end, quite unimportant differences. The cuboid and
navicular are firmly co-ossified, as in Palwomeryx and all existing Pecora ; these
bones are quite low, and the navicular rises but little in front to fit the distal
groove of the astragalus, but on the postero-internal side it sends up a strong
and high process, which makes the astragalar facet very deeply concave; dis-
tally the navicular shows two facets, a large one for the compound cuneiform,
and a much smaller one for the entocuneiform. On the distal surface of the
euboid, besides the large facet for mt. IV., is seen a minute, oblique infero-
lateral one, obviously for a rudimentary fifth digit. The only other tarsal
bone preserved in the specimens is the compound cuneiform, which is rather
low, narrow, and deep; in front it is nearly on a level with the cuboid, but be-
hind descends somewhat below it, and thus affords a lateral attachment to
mt. IV. The presence of a distinct entocuneiform is demonstrated by the
facets for it upon the navicular and cannon-bone.

The metatarsus presents some features of much interest. Rosenberg (Zeitschr.
f. wiss. Zool., Bd. XXIII.) has shown that in the sheep embryo there are at
one stage four complete metatarsals ; he states, however, that the lateral ones
are ultimately absorbed. In Blastomeryx, as in Amphitragulus, and probably
all existing Pecora and Tylopoda, there are at least three elements which enter
into the formation of the poste- |
rior cannon-bone ; viz. mt. III.
and IV., and the proximal por-
tion of mt. II. The latter, though
ankylosed with mt. III., shows
its limits distinctly; it has a
small facet for the entocunei-
form, and ends below in a point.
Mt. V. was obviously present, as

upon the postero-external side of - B és

the cannon-bone there is a shal- FIGURE 9.—Proximal end of posterior cannon-
bones X %, internal view; A. Amphitragulus;
low groove, and upon the euboid, B. Blastomeryx; C. Antilocapra.

as already stated, there is a small
facet for the head of it. An examination of the metatarsus of a modern rumi-
nant seems to show that the portion articulating with the entocuneiform is the
head of mt. II.; whether mt. V. be also present is more difficult to decide, but
in existing forms there is no portion which can be identified with it, while in
Blastomeryx, though undoubtedly present, it does not coalesce with the cannon-
hone. There is nothing in any of the specimens to indicate that any portion
of the distal ends of the lateral metatarsals were retained, though doubtless
phalanges were preserved, as in the deer.
In this brief, but fairly comprehensive review of the osteology of Blastomerys,
we have seen nothing which can be opposed to the view expressed by Professor
VOL, XX. — NO. 3. 6

82 BULLETIN OF THE

Cope, that this genus should be placed in the ancestral line of the distinctively
American deer. Alces, Tarandus, and Cervus are really immigrants from the
Old World, and do not belong in this category; but the truly American types,
of which Cariacus is the chief example, have a peculiar skull structure, first
pointed out by Garrod, which seems to show that the American deer were sep-
arated from those of the Old World at a comparatively early date, though it is
very questionable whether both series could have independently acquired the
extraordinary peculiarity of the deciduous antler.

COSORYX, Lrrpy.

Cosoryx fureatus, Lrrpy.

This very interesting animal is represented in the collections of Garman and
Sternberg by several specimens, which enable us to add materially to the de-
scriptions hitherto published. These descriptions are so brief that the relation-
ships of this genus have been very generally misunderstood. Schlosser says :
“Tn Nordamerika finden sich im oberen Tertiar zwar Geweihe von Hirschen,
Kiefer derselben sind indessen noch nicht mit Sicherheit ermittelt, wenigstens
nahern sich die Gebisse des Dicrocerus Cope, Merycodus Leidy, Cosoryx Marsh,
zweifellos eher den Antilopen, besonders dem lebenden nordamerikanischen
Genus Antilocapra, als den Hirschen. Sie sind zugleich viel einfacher gebaut
und schliessen sich namenthch die Marken sehr bald, was bei den Hirschen erst
in einem ziemlich spiten Stadium der Abkauung auftritt. Das Gleiche diirfte
wohl auch der Fall sein bei Blastomeryx Cope == Cosoryx gemmifer, trotzdem
Cope denselben als Stammvater von Cervus und Cariacus betrachtet. Die von
Marsh behauptete Existenz von Seitenzehen bei Cosoryx diirfte wohl mit Recht
bezweifelt werden; die Hand hat Cope abgebildet und zeigt dieselbe keine Spur
von etwaigen Griffeln. Wenn auch die systematische Stellung dieser Formen
noch nicht vollig klar gelegt erscheint, so konnen wir doch mit grosser Wahr-
scheinlichkeit annehmen, dass wir hier einen eigenen Seitenzweig der Rumi-
nantier vor uns haben, als dessen letzter Rest die merkwirdige nordamerika-
nische Gabelantilope zu betrachten ist. Die Verastelung des Geweihes ist
bisweilen fast so stark wie bei den echten Hirschen. Wahrscheinlich war es
von Hornmasse iiberzogen — den verwachsenen Haaren des Bastgeweihes. Fir
diese Annahme spricht die auffallende Glatte der von Cope und Leidy abge-
bildete Geweihfragmente.” (Morph. Jahrb., Bd. XII. p. 70.)

As we shall see, some of these inferences are probably quite correct, others
are equally probably misleading. This group of closely allied species is not
confined to the “upper Tertiary” or Loup Fork beds, but appears first in the
lower middle Miocene of Oregon in the John Day beds, where the genus Blas-
tomeryx is abundantly represented by some large species. Now Blastomeryz is,
so far as we can at present determine, almost identical with the type variously
named in Europe Palewomeryx and Dremothertwm, about the only difference of
importance being the absence of the characteristic “Paleomeryx fold” on the

MUSEUM OF COMPARATIVE ZOOLOGY. 83

lower molars. Cosoryx is very closely allied to Blastomeryx, and is distin-
guished from it chiefly by the much more hypodont molars. The bones of
the various Loup Fork species of this genera cannot be distinguished apart in
the absence of associated teeth, and it is quite probable that the John Day
species of Blastomeryx will prove to belong to a different genus from the Loup
Fork species.

THE DENTITION.

One undoubted specimen of Cosoryx contained in the Cambridge collection
consists of a fragment of the superior maxillary containing one molar, the lower
jaw with first and third molars, an antler, the sacrum, all the lumbar and the
five posterior dorsal vertebre in unbroken succession, the scapula, humerus»
pelvis, and posterior cannon-bone. The resemblance of these bones to Antilo-
capra is very striking, and fully justifies what Schlosser has said with regard
to the relationships of the two genera. The second upper molar is not much
extended in the antero-posterior direction, and has a fairly high crown, though
not hypsodont to the same degree as in the prong-buck; the median fold of
enamel on the external wall, or, more properly speaking, the projecting anterior
horn of the postero-external crescent is less strongly developed than in the re-
cent form, and the corresponding horn of the anterior crescent hardly projects
at all. The valleys are shorter and wider than in Antélocapra, and though the
tooth is in an advanced state of wear, they are still quite deep, in contrast to
what occurs in the lower molars. The lower incisors and canines are all
broken away, but from the alveoli and remaining fangs it may be seen that they
were of the ordinary ruminant pattern, probably not very long; they decrease
in size from the median incisor outwards, and the canine is the smallest, of the
series. The premolars, three in number, are represented only by their alveoli,
which shows them to have been very small. The most anterior is implanted
by a single root, the others by two. Leidy’s figure (Merycodus necatus) shows
them to possess considerable complication, but they are less molariform and
more trenchant than in Antilocapra. The true molars are more truly hypso-
dont than in the upper jaw ; the first is very small, but the third resembles that
of the modern genus exceedingly closely.

The same may be said as to the form of the mandible itself; the horizontal
ramus is very long, compressed, and rather shallow, and with an extremely long
diastema between the canine and premolar 3; the ramus is less rounded on the
external side than in Antilocapra, and in that genus there is no such descent of
the upper margin in front of the premolars as occurs in Cosoryx. The sym-
physis is short (much shorter than in C. trilateralis, Cope) and much contracted,
and on a level with its posterior edge is a large single mental foramen. The
antler is branched like the one figured by Leidy with the name of Cervus Warreni,
but with a much longer beam, and the tines meeting at a more open angle.
The beam is longer and the tines shorter than in any of the antlers figured by
Cope, except, perhaps, the imperfect specimen named Cosoryx (Dicrocerus) teres
(Wheeler, Pl. LX XXII. fig. 6). The antler is composed of dense bone, with a

84 BULLETIN OF THE

smooth and here and there furrowed surface, a texture which, as Schlosser has
remarked, is very different from that of a deer antler. The burr is very large
and prominent, but a vertical section shows that the beam passes into the
pedicel without any perceptible break or change in the tissue. In Cosoryx the
burr is very variable, as may be seen from Cope’s figures. In this collection
are some specimens without any burr, others with a single burr, and some with
two or three. They can hardly be regarded as an evidence that the antler was
deciduous.

THE SKELETON.

The vertebra, so far as they are preserved, resemble very much those of Anti-
locapra. Owing to the fact that only the posterior part of the column is pre-
served in the specimen, it will be most convenient to describe them from behind
forwards. The only caudal represented is like the second of the prong-buck,
but a little more complete, and clearly shows that the tail was short, as may
also be inferred from the sacrum. This caudal is short and narrow, especially
in front, with short wide transverse processes near the posterior end. There are
a pair of rudimentary prezygapophyses, and an exceedingly minute neural canal,
which will just allow the passage of a needle, and a corresponding neural spine.
In the prong-buck the second caudal has neither canal nor spine, and the trans-
verse processes are wider.

The sacrum consists of four completely ankylosed vertebre. The first has a
broad depressed centrum, well developed prezygapophyses, and much enlarged
pleurapophyses, which occupy most of the sacral surfaces of the ilia. The spine
is coalesced with the others into a high and arched ridge. In the prong-buck
the spines are more distinct. The other sacrals have expanded pleurapophyses,
but only the second has any contact with the ilium. The centra decrease rap-
idly in size from the first posteriorly, and that of the last is exceedingly depressed
and thin. The whole sacrum is quite strongly arched from before backwards.
The lumbar region is quite long, and consists of six vertebre, which are slen-
derly constructed; the centra are anteriorly comparatively narrow and trihe-
dral in section, posteriorly they are broader and more depressed. The spines
are low and comparatively broad, and are inclined well forward, with concave
anterior borders. The transverse processes on the first lumbar are short, de-
pressed, but comparatively broad ; these processes lengthen as we pass backwards,
but are very slender as compared with those of Antzlocapra, and the neural
spines are lower than in that genus. The zygapophyses are of the interlocking
cylindrical type usual among artiodactyles, and there are no metapophyses.
We may infer with considerable confidence that the number of dorsal vertebrae
was thirteen; on this assumption, the most anterior dorsal of this specimen is
the ninth. In this the centrum is short and trihedral in section, with the infe-
rior border sharp and arched from before backwards ; the spine is rather short,
and directed very obliquely backwards; the transverse processes are short and
slender, and have well marked facets for the tubercles of the ribs; the prezyga-
pophyses are flat and placed on the pedicels of the neural arch, and, separated

ee EEE eee

— ee ee ee ee eee

MUSEUM OF COMPARATIVE ZOOLOGY. 85

from them by a short interval, arises a pair of small metapophyses. The tenth
is the anticlinal vertebra; the spine is at first very oblique, but curves, and
in its upper portion is vertical. In other respects this vertebra is like its
predecessor. On the eleventh the spine is directed slightly forwards, but the
end is rounded like that of the anterior dorsals; the metapophyses have ap-
proached the median line so as to touch the post-zygapophyses of the tenth,
while the post-zygapophyses of the eleventh have assumed the cylindrical shape
found in the lumbar region. The twelfth and thirteenth vertebree are much
like lumbars in their construction, and are distinctly longer than the three an-
tecedent vertebre ; the spines have the nearly straight thickened free ends seen
in the lumbars, and the metapophyses have disappeared. The transverse pro-
cesses, however, are very short, though they still retain the rib-facets, even on
the thirteenth.

The ribs, so far as can be judged from the fragments, are narrow and very
slender. Of course this may be true only of the posterior part of the series.

The scapula is characteristically ruminant. The glenoid cavity is nearly round
and quite shallow, the coracoid process is prominent, recurved and thickened at
the end ; the neck is very long and much contracted, the borders sloping away
from it very gradually; the coracoid border is thin and rounded at the edge, it
curves gently forwards and upwards from the neck; the glenoid border is very
much thickened and somewhat overhanging, from the neck it is nearly straight,
and forms a right angle with the very thin suprascapular border. The spine
rises abruptly from the neck into the high acromion ; the latter overhangs very
slightly, in sharp contrast to the condition found in Anttlocapra. The spine
divides the blade into unequal fossz, the prescapular being much the smaller,
as is ordinarily the case among the ruminants. Except for the nearly straight
inferior edge of the spine, and the consequent lack of an overhanging acromion,
this scapula very closely resembles that of the prong-buck.

The humerus has a broad and flattened head, which projects but little beyond
the shaft. The external tuberosity is large, and curves over the deep bicipital
groove ; the internal tuberosity very small; both are much less developed than
the corresponding processes in Antilocapra. Proximally the shaft is broad and
compressed, below it is rounded and slender. No ridges for muscular attach-
ment are more than very faintly indicated. The distal end is broken away, but
in all probability it was like that of Blastomeryx described above.

The pelvis is also entirely ruminant in character. The ilium has a short, deep,
and much compressed neck, expanding into a curved and strongly everted plate,
which projects a considerable distance in front of the sacral attachment. The
ilium is somewhat trihedral in section, the median rounded ridge of the plate
being more prominent, and the expansion itself smaller than in the prong-buck.
The ischium is very long ; above the acetabulum its superior border shows the
convexity so usual in the recent ruminants, though in a less marked degree.
The tuberosity of the ischium is very long and prominent, and directed straight
outwards; behind the tuberosity the ischium is prolonged further than in the
prong-buck. The cannon-bone belonging to this specimen is broken, and its

86 BULLETIN OF THE

proximal end obviously diseased, so that it does not merit description; the only
fact of importance which it shows is the comparative slenderness of the bone.

So far as the material will enable us to judge, the feet of Cosoryx differ in no
important respect from those of blastomeryx, aud the same statement applies to
the long bones of the limbs.

RESTORATION OF COSORYX FURCATUS.

(See Plate I.)

This drawing is made from the specimen already described, completed vy
fragments of others, while the feet are drawn trom blastomeryx; the cervical
vertebrae are represented only by the axis, the others being conjectural, as are
also the anterior dorsals. The skull is taken chiefly from that of the closely
allied European genus, Pa’@omeryx, and from specimens of the large Cosoryx
teres, Cope, belonging to the Smithsonian Institution. The fortunate associa-
tion of the mandible in the same specimen with the vertebre, pelvis, scapula,
ete., gives a very useful standard as to the length and character of the skull,
position of the molars, etc. It may be assumed with some confidence that the
drawing gives a fairly accurate representation of the animal. Marsh’s account
of the feet of Cosoryx shows that they were constructed much like those of Blas-
tomeryx. In general appearance Cosoryz seems to have had the same light,
graceful build as Antilocapra, but with a very different skull and deer-like
antlers. The proportions of the limbs also differ somewhat, the hinder cannon-
bone being considerably longer than the fore, while in the prong-buck they are
of nearly the same length. Cosoryx was a much smaller animal, the bones are
all more slender than in Antilocapra, and the carpal and tarsal bones are much
higher and narrower proportionately.

The view held by Cope that Cosoryx is the ancestor of Antilocapra is very
probably the true one. So far as the dentition, the vertebra, and the limbs are
concerned, the differences between the two genera are only such as might be
expected to occur between a Miocene and a recent ruminant. A distinction of
some importance, however, consists in the character of the horns. In Cosorya
they are branched, but probably not deciduous antlers; in Antilocapra, a core
with a horny sheath, which, however, differs strikingly from the horn of the
typical Cavicornia. But the unique branched horn of Antilocapra not improb-
ably indicates, as has been suggested by Cope, a remnant of a former branching
of the bony core itself, and so this difference does not preclude a genetic con-
nection between the two forms. In Cosoryx the antler was almost certainly
covered with skin; its smooth surface, as Schlosser points out, shows that it
could not have been naked, as in the true deer.

Both Blastomeryx and Cosoryz are probably to be derived from the species re-
ferred to the former genus which occur in the John Day beds, but there is no
form yet known in the White River which could have given rise to these John
Day ruminants. The latter are most probably descended from some Palao-
meryx of the Old World, which migrated to this continent. The very close con-

MUSEUM OF COMPARATIVE ZOOLOGY. 87

nection between these American genera and the Amphitragulus, Dremothervum,
etc. of St. Gérand le Puy is obvious from the most superficial comparison.

The collection contains specimens probably indicative of other species of
Cosoryx, some of them much larger than C. furcatus; but in the absence of asso-
ciated teeth, it is not possible to refer them to their proper categories.

PERISSODACTYLA.
ANCHITHERIID A.

MESOHIPPUS, MarsH.

THe BRAIN.

Mesohippus had a large and well convoluted brain. The length and breadth
indicate that it weighed about one third as much as the brain of the recent
horse, while if we estimate the body weights of the fossil and recent animals by
the relative size of the humeri, the brain of the Miocene species was proportion-
ally heavier. The cerebrum
of the horse is, however,
much more highly convo-
luted, and the frontal lobes
are relatively broader. The
Mesohippus brain is distin-
guished in a marked manner
by the longitudinal direction
of the parietal and occipital
sulci, and by the deep trans-
verse frontal sulci, as con-
trasted with the oblique sulci
of all recent ungulates. In
fact, in this respect it bears a
marked general resemblance
to the brain type of recent
Carnivora, and conforms with
the higher Ungulata of the
Eocene.

On either side of the lon-
gitudinal fissure is a long deep fissure forking anteriorly and marking off the
median gyrus, m, of the parieto-occipital region. Parallel with this is a short
fissure, which separates the two medilateral gyri, ml, mi’. The third fissure
extends to the posterior transverse, and thus entirely separates the supersylvian
gyrus, ss, from the medilateral. The fourth fissure is shallower. There are
three transverse frontal fissures (FR. 1, 2,3) which divide this lobe into three
gyri; the median fissure extends almost to the longitudinal fissure, and sug-

FicurE 10.— Brain of Mesohippus Bairdii X %. From
above, and from side.

88 BULLETIN OF THE

gests the crucial sulcus of the Carnivora. The sylvian fissure is very shallow.
‘The temporo-sphenoidal lobe is very prominent, and is divided into three gyri
(s,m, 7) by two sulci. Beneath the third frontal gyrus is a vertical sulcus, par-
allel with the sylvian.

The cerebellum has a large central lobe with transverse simple furrows.

THE DENTITION.

There are a few new points to be noted in regard to the teeth of Mesohippus,
which bear upon the dentition of the horses in general, and are clearly shown in

FicurE 11.— Superior and inferior molars of Mesohippus Bairdii x 4.

a series of unworn crowns of the upper and lower jaws. Scott has already
pointed out that the incisors in this genus are simple, there being no indication
of the infolding of the enamel, such as is seen in
Anchitherium aurelianense. In some of the John
Day species of Anchitherium the enamel is not in-
folded, as observed in the lower jaw of a specimen
referred to A. equiceps, Cope.

The upper molars of Mesohippus clearly show the
first step in the formation of the posterior pillar, pp,
which is so conspicuous a feature in Anchithervum,
in the posterior valley. This can also be observed
in a still simpler stage in a specimen of Anchilophus
from the French Phosphorites. Step by step with
the development of this cusp appears the posterior
pillar, p, in the lower molars, behind the entoconid ;
this accessory cusp can be traced back to the teeth
of Epihippus. When it finally unites with the en-
Nene 12.— Superior mo- toconid, in Hipparion, it forms the posterior twin

ar of Anchitherium longi- Neds é :

criste X 1, Superior and cusp (0, 6, Riitimeyer), which is analogous to the

Ha view. Cope col- anterior pair formed by the union of the metaconid

and anterior pillar, a (a, a, Riitimeyer).

Thus the transition from the Mesohippus to the Anchitherium molars is very
gradual, as shown in the accompanying figures. By tracing back the rise of

Jva he 7ve

EOD

|
‘

aa ee

nn

Ly
4
A

MUSEUM OF COMPARATIVE ZOOLOGY. 89

the eleven elements which compose the upper Equus molar, we find that six
belong to the primitive sextubercular bunodont crown. Two elements of the
ectoloph, the antervor pillar and median pillar, rise from the simple primitive
basal cingulum of the Hyracothervwm molar; the same mode of development,
we have just seen, is true of the posterior pillar. The eleventh elemeut, the
fold of the postero-external angle of the crown, p, is not prominent until we
reach Equus. The term “posterior pillar” is taken from Lydekker ; the other
terms, ‘‘median” and “ anterior,” are applied to parts which have an analogous
origin from the basal cingulum. The remaining coronal cusps are readily iden-
tified with their homologues in the primitive tritubercular molar.

? Anchitherium parvulus, Marsu.
(Syn. Equus parvulus, Marsh.)

Among the Loup Fork specimens collected by Clifford are found two lower
molars, m, and mz, which are almost identical in size with those of Mesohippus
Baird. The crown of m, measures: antero-posterior, .011 m.; transverse,
009 m. Unlike the Mesohippus molars, there is no external cingulum. The
“posterior pillar” has the same degree of development as in Anchitherium. The
fangs are separate. ‘There is no trace of cement. Marsh has described a di-
minutive horse (Hquus parvulus), estimated at two feet in height, from the
same beds, and it is highly probable that these teeth belong to this species. The
generic reference is of course very uncertain. The brachydont crowns point
either to Merychippus or Anchitherium, but the stage of development of the
coronal pattern approximates most closely that in the latter genus, being a little
more advanced than in Mesohippus.

RHINOCERIDA.

ACERATHERIUM.

THe MANUS AND Pks.

The characteristics of the pes of Hyracodon from the lower White River beds
have been fully enumerated by us.! They are principally as follows: cuboid
not supporting astragalus anteriorly ; lateral digits reduced and not spreading ;
ectocuneiform not articulating laterally with mts. II. We may subsequently
find that the feet of the later species of Hyracodon varied in some of these re-
spects, although this is not probable, owing to the fixity of foot-types once
established. We have, however, no present means of distinguishing between
the Metamynodon and Aceratherium foot-bones.

On page 169 of the first Bulletin a high, rather slender tarsus was described,

1 See Scott, E. M. Museum Bulletin, No. 3, May, 1883, p. 19. Also, Osborn, Mam-
malia of the Uinta Formation, May, 1889, Part IV. “Evolution of the Ungulate
Foot,” p. 549.

90) BULLETIN OF THE

which probably belongs to the Aceratheriwm of the lower beds. It differs:
widely in its proportions from other specimens found in this collection, which
belong either to the Aceratheriwm of the higher beds, or to Metamynodon. The
best preserved specimen of this second type
(marked a®) is comparatively short and broad,
with spreading digits and rugose surfaces for
muscular attachment (Figure 13). The pro-
portions of the metapodials to the tarsals: are
similar to those in Ceratorhinus. The caleaneum
has a powerful tuber ; the ectal astragalar facet is
very convex ; the sustentaculum is narrow, and
its oval facet is continuous with the inferior;
the cuboidal facet is nearly horizontal. About
one fifth of the astragalus rests upon the cuboid.
The relations of the cuboid, navicular, and ecto-
cuneiform repeat those observed in Rhinocerus.
The mesocuneiform is very short, giving mts. II.
a wide articulation with the ectocuneiform. The
metatarsals are powerful, the lateral pair having
approximately the same length as in R. indicus.
This type of foot is related directly to that of
Aphelops. :

The manus and pes of a third specimen (marked
a®) show several interesting differences. In the
pes, the metatarsals are of the same proportions,
but the calcaneo-cuboidal facet is oblique and
narrow, resembling that in Hyracodon, and the sustentaculum is very small.
The remains of the carpus show that the species to which this specimen belonged
had a greatly reduced fifth digit, constituting a functionally tridactyl manus.
The evidence for this is in the greatly reduced lunar-magnum facet, which is
invariably characteristic of tridactylism.?

It may be noted here that among the carpals of Titanotherium there is a well
preserved lunar, which has its magnum facet much reduced anteriorly, so there
is little question that we shall yet discover a tridactyle species of the genus.

Fiagure 13 —Right pes of Ace-
ratherium X #.

THE Rartnoceros Monars.

The peculiarities of the molars of Aphelops will be made more clear by a few
observations upon the molars of the rhinoceroses in general. The three main
crests of the lophodont crown may now be distinguished in part by terms which
express their homologies with the elements of the sextubercular superior and
quadritubercular inferior molars of the primitive ungulate, Phenacodus. In the
upper molars, the outer crest is formed by the union of the primitive paracone

1 See Osborn, Mammalia of the Uinta Formation, p. 567. It is possible that these.
feet belong to Metamynodon.

MUSEUM OF COMPARATIVE ZOOLOGY. 91

and metacone, to which is joined the anterior pillar (see Mesohippus, p. 88); it
may be called the ectoloph. As the anterior crest is formed by the union of the
protocone, protoconule, and paracone, it may be termed the protoloph. The
posterior crest, which unites the primitive metacone, the metaconule, and the
hypocone, may be termed the metaloph. The
outer surface of the ectoloph in the primitive
molar of the Rhinoceros is marked by three
vertical ridges corresponding to its three prim-
itive component elements, me, pa, ap; one or
all of these disappear in the flattening of the
surface. It will be observed that nothing cor-
responding to the ‘median pillar’ of the su-
perior molar of the horse is developed. In the
lower molars (the paraconid disappearing), the
union of the metaconid and protoconid forms
the anterior crest, or metalophid, while the
hypoconid and entoconid unite to form the

)
Wg

hypolophid. FicurE 14.—Superior molar of
: Rhy ‘os (sp. 1 4 r
The secondary enamel folds, which are de- vaaeer ale Amit) Sa ever

veloped from the three crests, bear a most in-

teresting analogy to those observed in the horse series, beginning with Proto-
hippus; they are outgrowths of the same regions of the crown and subserve
the same purpose. ‘They are moreover of like value in phylogeny.. The useful
descriptive terms introduced by Busk, Flower, and Lydekker, should be adopted
in part.!_ These secondary elements consist, first, of three folds projecting into
the median valley, one from the ectoloph, the crista; one from the protoloph,
the crochet; one from the metaloph, the anticrochet. Secondly, the ecto-
loph unites with the posterior cingulum and metaloph. Thus the anterior and
posterior valleys may be cut off by the union of these folds into from one to
three ‘fossettes,’ precisely analogous to the ‘lakes’ in the horse molar, except
that they are not filled with cement.

The accompanying diagram is taken from a fossil molar figured by De Blain-
ville. (Osteogr. Gen. Rhin, Plate XIII.) It is remarkable in exhibiting all
the primary and secondary elements, for they are very rarely combined in a
single tooth. Similar accessory folds are frequently developed in the lower
molars.

1 The terms ‘ protoloph’ and ‘ metaloph’ are, however, substituted for ‘anterior
collis’ and ‘posterior collis’ of Lydekker. The term ‘anterior pillar’ = ‘first
costa,’ and ‘ paracone’ = ‘second costa.’ The mode of evolution of the ‘pillar’
must have been similar to that in the horses, where Lydekker has proposed this
term for the ‘posterior pillar.’ It is very appropriate, because the pillars in their
earliest development can be shown to rise independently from the cingulum (see
Mesohippus, p. 88), and not as folds of the main elements of the crown, as we should
infer from their fully developed stage.

92 BULLETIN OF THE

APHELOPS, Corn.

The generic characters of Aphelops have been given by Cope as follows. Den-
tition, I. 274, C. $, P. #98, M. 3; post-glenoid and post-tympanic processes
in contact but not co-ossified ; digits, 3-3; nasals hornless. To these charac-
ters may be added: magnum not supporting lunar anteriorly ; absence of the
‘crista’ and invariable presence of the more or less strongly developed ‘ crochet ’
and ‘anticrochet’ in the superior molars.

The specific nomenclature of A phelops is in confesion. The type of 4. (Rhi-
noceros) crassus, Leidy,} is a last upper molar, which is closely similar to that of
A. megalodus ; the characters of the milk molar associated with this type cannot
be used in definition? The penultimate upper molar, the type of A. meridi-
anus, Leidy,? corresponds in the development of the two ‘crochets’ to the same
tooth in A. fossiger, Cope, but the posterior ‘fossette’ is not enclosed by the strong
cingulum as in the latter species. A. (Aceratheriwm) acutum, Marsh, is identical
with A. fossiger. A. malacorhinus, Cope, resembles A. meridianus in the open
posterior fossette and the development of the ‘crochets.’ It is impossible, how-
ever, to clear up this synonymy without bringing the original types together
for comparison. General characteristics of all these types are the invariable
development of the ‘crochet,’ absence of the ‘crista,’ usual development of the
‘auticrochet.’ The specific names proposed by Cope are here adopted because
they are established upon a very complete knowledge of the skull as well as
of the teeth.

Aphelops fossiger, Core.

Dentition: I. 4,C. 9, P. 4, M. 3. First premolar simple, conical, sometimes
absent; nasals not overhanging premaxillaries; foramen lacerum medium
confluent with foramen ovale ; occiput broad and low ; limbs short and bulky ;
molars with well developed ‘ crochet’ and ‘anticrochet.’

In the figure given by Marsh (Am. Journ. Sci., Oct., 1887, p. 3) and by Cope
(Ain. Nat., Dec., 1879, p. 771 ¢), the third and fourth premolars have both the
‘crochet’ and ‘anticrochet.? There is some ground for the supposition that the
skull here described belongs to a different species, since the ‘ anticrochet’ is not
developed in the premolars. This reference is therefore provisional.

This is apparently the only species which is represented in this collection.
All the specimens are from Kansas, and include several skulls and well pre-
served bones from all parts of the skeleton, enabling us to give a complete
description and restoration of the animal.

1 See Ext. Mamm. Fauna, Dak., p. 228.

2 Cope has nevertheless employed the ‘criste’ developed in this milk molar in
his definition of A. crassus. “On the Extinct Species of Rhinoceriide of North
America,” etc., Buil. U. S. Geol. Survey, Vol. V. No. 2, p. 257.

in — 7

MUSEUM OF COMPARATIVE ZOOLOGY. 93

THE BRAIN.

One of the most interesting features of Aphelops is the very large size of the
brain. The walls of the cranium are solid. There are no vacuities or air-cells
in the diploé of the mid-region of the brain-case, such as attain from 1 to 14
inches in thickness in Ceratorhinus. Thus the brain is relatively much larger
WIIK,

My
: ny

Ficurer 15.— Brain of Aphelops fossiger X 4. Lateral view of intracranial cast.

than that of the recent rhinoceros, and presents a marked advance upon that of
Aceratherium oceidentale. The bulk of the fore- and mid-brain, or the divisions
in front of the cerebellum, is approximately as follows :—

Acerathervwm, 420 ¢.c. Aphelops, 1240 ¢.c. Ceratorhinus, 720 c.c.

The bulk of the entire brain is: Aphelops, 1470 c.c. Ceratorhinus, 850 c.c.
The relative body weight of the two animals can be roughly estimated from a
comparison of the femora as Aphelops 4, Ceratorhinus 3. It thus appears
that the steady brain growth of the ungulates during the Eocene and early
Miocene periods reached its
highest point in some fami-
lies of the later Miocene,
and was followed by a de-
generation.

The cerebellum in Aphe-
lops is small and partly over-
hung by the hemispheres.
The lateral view of the
hemispheres shows a very
marked predominance of
transverse sulci, which ra-
diate from the vertical syl-
vian fissure, S, so that in the basal view of the frontal lobes the fissures are
antero-posterior. The dorsal surface of the cast is somewhat imperfect, giving
an incomplete reproduction of the parietal and occipital regions. The superior

FicurE 16 — Brain of Ceratorhinus Sumatrensis x }.
Lateral view of cast.

94 BULLETIN OF THE

sulci of the frontal lobe are directed obliquely backwards to the longitudinal
fissure, thus reversing the direction observed in the recent ungulates.

THE SKULL AND DENTITION.

The skull (Plate ITI.) is broad in relation to its length, owing to the shorten-
ing of the ant-orbital region and the recession of the nasals. The maxillaries
spread very widely for the powerful series of molars, while the premazillaries
are slender. The orbit is placed above the first molar. The nasals are com-
pressed anteriorly, and extend only so far as to overhang the premaxillary suture.
A marked feature of the skull is that the upper surface is in a nearly straight
line from the supra-occipital ridge to the tip of the nasals, while in A. megalodus
it is concave. The orbit is very slightly overhung by the supra-orbital process.
The zygomatic arch is deep vertically, but compressed laterally. The post-
glenoid process is deep and narrow ; it has contact with the post-tympanic of
variable length. The remarkable feature of the post-tympanic is its extension
into a broad flat plate behind the auditory meatus. The occiput is broad and
low, and does not overhang the condyles; it is deeply cleft in the median line.
On the base of the skull, the foramina rotundum and spheno-orbitale are con-
fluent, as observed by Cope. The foramen ovale is either confluent with or
separated by a slender ridge of bone from the foramen lacerum medium.

The molars and premolars are remarkable for the extreme flattening of the
outer surface of the ectoloph, all trace of the three vertical ridges having dis-
appeared. The first premolar is a simple
conical tooth implanted by a single fang ; it
is apparently inconstantly developed, for
Marsh makes no mention of it in his de-
scription of A. (acutwm) fossiger. The inner
angles of the protoloph and metaloph unite
by the ‘crochet’ in pm? and pm? to enclose
the median valley, as in Acerathertwm. The
fourth premolar resembles the molars except
in the non-development of the ‘anticrochet.’
The true molars are characterized as follows :
Figure 17.—First superior molar of })y the constriction of the inner portion of

Aphelops fossiger X 3. ‘
the protoloph into a separate column; by
the strong development of the ‘crochet, which in m? and m? unites early with
the metaloph to enclose the anterior ‘fossette’; by the development of the
‘anticrochet’ at the inner angle of the metaloph and ectoloph; by the com-
plete enclosure of a posterior ‘ fossette’ in the first and second molars.

The inferior molars are of the simple rhinoceros pattern, there being no trace
of accessory folds. The first premolar is missing; the second is separated by a
rather narrow diastema from the large lateral tooth. Between the pair of large
semi-procumbent caniniform teeth are two small incisors.

The lower jaws are very massive, with a strongly arched lower border. The
condyles are broad and elevated. The posterior border is broad, but not rugose.

ee

MUSEUM OF COMPARATIVE ZOOLOGY. 95

THE SKELETON.
(Plate IIT.)

Vertebre. — The atlas resembles that of R. wnicornis, with extremely broad
transverse processes. A well preserved axis has a low tuberosity representing
the spine; there is some doubt whether this is the normal adult condition, al-
though the absence of the spine would accord with the low occiput and hornless
nasals, The cervicals 3-6 have deeply opisthoccelous centra, rather high and
narrow in proportion, with powerful zygapophysial processes.. The inferior
lamellz of the transverse processes project downwards and forwards, and- ex-
pand very slightly at the tip; the width of this lamella increases somewhat in
C. 6; the superior lamelle project opposite the vertebrarterial canal. The sixth
and seventh cervicals apparently have slender elevated spines, in the remainder
the spines are low or tuberous. The centrum of C.7 is subcircular in front
and broad posteriorly.

The dorsals are represented by a number of vertebree in the mid-region. The
centra are laterally compressed with distinct keels; the zygapophysial facets
are very small and horizontal; the metapophyses are well developed. The
length of the spines in the anterior dorsal region was apparently as in L. javanus.
No lumbars are found in this collection.

Fore limb. — The scapula is very short and heavy. The general outline is
triangular; the glenoid border is concave; the coracoid border is convex ; the
superior border rises to a point above the spine; the upper third of the spine
shows a very stout recurved acromial process.

The humerus is remarkably short and heavy, and is distinguished by the un-

usually elevated position of the deltoid ridge, which is much higher upon the
shaft than in the recent rhinoceroses. The tuberosities are heavy and sessile ;
the external condyle is unusually prominent. The wlra has a deep, powerful
olecranon process and stout trihedral shaft, which is suddenly compressed in-
feriorly for the cuneiform articulation. The proximal and distal faces of the
radius are subequal ; the shaft is very slightly arched and closely united with
that of the ulna, giving this segment a very massive appearance.

The structure of the manus is in keeping with the short and heavy upper
segments; it is broader and more powerful than in any of the recent rhinoce-
roses. The three short, widely spreading digits are faced by rugose areas for
the attachment of powerful muscles. Mtc. III. is much the largest; the lat-
eral metacarpals, II. and IV., are short and directed outwards; the phalanges
are short and wide, especially the distal series. As in all tridactyle forms the
carpal displacement is extreme; the scaphoid covers the whole upper surface of
the magnum anteriorly ; the lunar is rather small, and rests anteriorly wholly
upon the unciform; posteriorly the pivotal process of the magnum supports
the lunar; the cuneiform is high and narrow. The trapezium is missing in
both the carpal series before us, but is indicated by the usual facets upon mte. IT.
and the trapezoid. The magnum is broad and quadrilateral. The unciform
has an unusually wide mtc. III. facet, and is vertically compressed.

96 BULLETIN OF THE

Hind limb. — There is a complete left innominate bone, which gives all the
characters of the pelvis. The upper surface of the ilium, unlike that of Cerato-
_rhinus, is nearly flat, The supra-iliac border is evenly arched, and, as the

ischial and acetabular borders are of approximately the same length, the ilium
is unusually symmetrical. The ischium and pubis are in a plane perpendicular
to that of the ilium; the pubic symphysis is short; the obturator foramen is
an elongate oval. The tuber-ischil is not very prominent. The border extend-
ing from the tuber to the symphysis is evenly rounded.

The femur is relatively longer and more slender than the humerus, having
the form and proportions observed in Ceratorhinus. The great trochanter stands
out widely ; below this the shaft is of a broad flattened section ; the lesser tro-
chanter presents a long low ridge; the third trochanter is only half as promi-
nent as in the recent rhinoceros, and is not recurved. The tibia is characterized
by a marked asymmetry of the tuberosity ; the internal malleolus is not promi-
nent ; the popliteal space is deeply excavated; the astragalar facets are shallow.
The fibula is of the same proportions as in the recent rhinoceros.

The tursus is unusually short and spreading. The astragalo-tibial facet is
flattened laterally, and shows little fore and aft play; the ectal and sustentacn-
lar facets are either confluent or slightly separate; the inferior is distinct and
separate ; the cuboidal facet is extremely broad. The cuboid is shallow, with
subequal calcaneal and astragalar facets; posteriorly it articulates with both
the navicular and ectocuneiform, anteriorly with the latter only; it has a very
deep posterior hook. The presence of the entocuneiform is indicated by the
articular facets for it. The mesocuneiform is narrow and deep. The ectocu-
neiform is very broad; this bone and the navicular have the same proportions
as in the rhinoceros. The middle digit is much the largest of the three, and
Mts. IIT. has a considerable cuboidal facet.

The following measurements are made from specimens which belong to dif-
ferent individuals, a, b, c, etc.; they therefore cannot be used in estimating the
exact proportions of the different parts. The proportions have, however, been
very carefully determined in the accompanying restoration of the skeleton.

MEASUREMENTS.
Skull.
m.

Spec. s. Total length, sagittal crest toend of nasals . .. . . . . .490
os Breadth, outside zygomatic arches . 2. . «© s+» ts) «@ WG)» toOU
oe Depth, penultimate molar to top of cranium. . . . . . « .2385
4 Occiput, diameter of, transverse, .268m.; vertical. . . - . .198
2 From occiput to anterior end of orbit . . . . a eee

i Antero-posterior, diameter Te SS angie series ane 115 T.,
it) a) any", Ps A . 3h opeen aa
& Diameter first mele, Bi isterise 057 m., mansveess oi) Oe
6 seconc me .068 < are) ns Ee

‘5 “ third “ Pe 058 . . Swe ae

MUSEUM OF COMPARATIVE ZOOLOGY. 97

m. m.

Spec. s. Diameter fourth premolar, antero-posterior .045; transverse . .065
6 3 third “ SS .035 a + (e050

se Ly second 2 O28. Hue “ a yi OSB

* + first 4s . 017 . Hie SOT

¢ Lower jaw, length, angle to front of canine . . . . . . . .470

a a depth, tip of coronoid to inferior border . . . . .295

Vertebre.

Spec. h. Atlas, greatest width, .356m.; greatest depth . . . . . . .100
Spec. pp. Axis, greatest width, .18m.; length ofcentrum . . . . . .090
“¢ = ** depth, spine to base of centrum, estimated . . .140
Spec. p. Fifth cervical centrum, antero-posterior .074 m., vertical .068 m., .
transverse .076 m.
Spec. 0. Twelfth dorsal centrum, antero-posterior .075 m., vertical
.055 m., transverse .058 m.

Appendicular Skeleton.

Spec. c. Scapula, vertical diameter, approx., .295 m.; glenoid ee ant.
ot rr - . 900
= Humerus, length of, 308 nt: brenatte iene and Habe Pay 5 5
Spec. a. Radius, length, .285 m.; breadth, proximal, .093 m.; distal . .098
‘ Ulna, greatest length, .36 m.; sigmoid facet to cuneiform facet .295

‘“ Carpus, greatest transverse diameter, .130 m,; ditto vertical . .057
of mar Prt. vreau. . JOO mt. ; lenotr: Soy SEO
a ule 3B - » « 043 i: 5 en ee oe OO
- a ks qj eee Ot a ede a AY vargh es

Spec. e. Left innominate bone, diameter, antero-posterior . . . . . .495
Length of pubis, .185 m.; of ischium, .20 m.; ofilium . . . .340
Spec. f. Femur, length of, .46 m.; diameter, head and great trochanter. .165

spec. g- Tibia, length of, .37 m.; width, proximal. . . . . a PEO
Spec. q bas 7. Tarsus, tuber éidfets to distal facet of mts. III., apptDR. - .220
Beye raniewemse (hiaImOver eho eg tt. Le) OS

. Second metatarsal, length 09,0. 2 . 1k. » 088

RESTORATION. (See Plate II.)

The restoration of Aphelops fossiger confirms Cope’s statement that the pro-
portions of the animal were rather those of the hippopotamus than the rhi-
noceros. The body was long, the chest deep, the limbs and feet short and
massive, and supplied with powerful muscles. The skeleton is about 9 feet
long and 4 feet 6 inches high. Thus Aphelops presented a wide contrast to its
tall, comparatively slender predecessor, Aceratherium, of the lower Miocene.
The increase in brain capacity shows that its nervous organization kept pace
with its general muscular and skeletal development. We may infer that the
extinction of Aphelops was due to climatic changes, rather than to any defects
in its internal organization, because the brain, teeth, and feet are, in themselves,
as adaptive as in any of the present persisting types.

VOL. XX. — NO. 38. ii

98 BULLETIN OF THE

CoMPARISON WITH ACERATHERIUM AND RHINOCERUS.

There is nothing, however, which precludes the supposition that the Ameri-
can lower and upper Miocene Aceratheria are genetically related.

All portions of the skeleton of A. occidentale are now known to us, excepting
the scapula, pelvis, and dorso-lumbar vertebrae; they indicate an animal in the
same stage of skeletal evolution as the recent tapir; the proportions are practi-
cally similar; the displacement of the carpals and tarsals is in a corresponding
stage. ‘The mode of progression was also probably similar, for all the articular
facets and protuberances for muscular attachment present innumerable points
of resemblance. Cope! first pointed out the tapir resemblances in A ceratherium,
especially in the separation of the foramina spheno-orbitale and rotundum ovale
and foramen lacerum medium; the separation of the post-glenoid and post-
tympanic; and the form of the femur. We have shown that this resemblance
apples to the carpus? and tarsus ; it is also true of the humerus and forearm,
and of the atlas and axis. The remaining cervicals are widely different ; it is
probable, also, that the pelvis and scapula were different. This is of course
simply an instance of functional and structural parallelism. It follows that an
enumeration of the differences between the recent tapir and rhinoceros would
also embrace the majority of the features which distinguish Aceratherium from
Aphelops, for the latter is in most respects a fully developed rhinoceros.

Thus, if the descent from Acerathertum to Aphelops took place, it was accom-
panied by wide-spread modifications of the skeleton. In Aphelops megalodus we
find a probable transition species. Its proportions are more intermediate. The
narrow elevated occiput, the less degree of separation of the foramina of the
skull, the lophodont character of the first upper premolar, the small develop-
ment of the ‘anticrochet’ in the superior molars, — these characters all point
towards Acerathervum.

A. fossiger is a highly modified form, with its broad occiput, simple first pre-
molar, and confluent cranial foramina. In many respects the modifications it
exhibits are simply steps towards the recent rhinoceros type; for example, its
tridactylism, the extreme displacement of the podials, and the characters of the
spinal column. But there are many points in which Aphelops differs from the
recent rhinoceroses; namely, the sub-triangular shape of the scapula, the very
elevated position and sessile character of the deltoid ridge of the humerus, the
spreading manus, the oval obturator foramen, and the comparatively feeble de-
velopment of the third trochanter. The marked peculiarity of the upper molars
is the development of both the ‘crochet’ and ‘anticrochet,’ and absence of the
‘crista.?_ This combination is very distinctive, since all the living rhinoceroses
present combinations of the ‘anticrochet’ and ‘crista.’? The molars of Aphelops

1 Bull. U. S. Geol. Surv., Vol. V. No. 2, p. 285. Also, “On Extinct American
Rhinoceroses and their Allies,” Am. Nat., Dec., 1879, p. 771 c.

2 Osborn, “ Evolution of the Ungulate Foot,” Mem. Uinta Mamm., p. 550.

8 See Flower, “On some Craniai and Dental Characters of the Existing Species
of Rhinoceroses,” Proc. Zool. Soc., 1876.

’

MUSEUM OF COMPARATIVE ZOOLOGY. 99

resemble in this respect those of R. tichorhinus. Briefly stated, in all living
forms the protoloph is simple, and the accessory folds are developed, first from
the metaloph, then from the ectoloph; while in the known extinct American
forms the ectoloph is simple, and the protoloph develops a fold to which a fold
of the metaloph is sometimes superadded.

In view of these facts, together with the numerous divergences in the skele-
ton, there is strong corroboration for the opinion advanced? by Scott in 1883,
that Aphelops should not be regarded as ancestral to any of the recent foreign
species, but rather as the last known of an extinct American series. The ques-
tion is still an open one whether its distribution was confined to this continent.

CHALICOTHERIOIDIA.’

CHALICOTHERIUM, Kavp.

Specimens of this genus are rare in American formations, and have not as yet
been reported from the Loup Fork. Marsh* has mentioned the occurrence of
it in the John Day Miocene of Oregon, and in view of the discoveries of Forsyth
Major and Filhol, it is altogether probable that the foot-bones from that for-
mation, which Marsh has referred to the Edentata under the names Moropus
distans and M. senex,* belong to the same genus. A third species of the same
genus is announced by Marsh® from the Loup Fork, M. elatus, which is prob-
ably represented in the Garman collection from the Loup Fork of Nebraska.

Chalicotherium elatum? Marsa.
(Syn. Moropus elatus, Marsh.)

The specimen is a portion of a right superior maxillary containing the third
and fourth premolars and the first molar. The premolars have a flattened ecto-
loph connected by two convergent crests, with a large internal cone which is
cleft at the summit; the base of this cone is surrounded by a strong internal
cingulum. The ectoloph is worn by two symmetrical incisions alternating with
the transverse crests in the third premolar, but in the fourth these incisions are
asymmetrical. The first molar is partly of the Titanotherium type, with its

1 ¥. M. Museum Bulletin, No. 8, 1883, p. 17.

2 Gill, Arr. of the Fam. of Mammals, Smithsonian Mise. Coll., No. 230, p. 271.
This order was properly defined by Gill, but was erroneously placed among the
Artiodactyla, owing to the reduced condition of the superior incisors. Filhol’s forth-
coming memoir upon the Mammals of Sansan will probably enable us to determine
its phylogenetic relations.

8 American Journal of Science and Arts, 8d Series, Vol. XIV. p. 362.

* Ibid., pp. 249, 250.

5 Thid., pp. 250, 251.

100 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

protocone isolated, but the hypocone, as in all known Chalicotherioids, is united
with the metacone by a low ridge (metaloph).

Figure 18. — Superior premolars and first molar of Chalicotherium elatum X %.

The available figures and descriptions are so imperfect that the relation-
ships of this species to those of the Old World cannot be definitely made out.
It is, however, decidedly smaller than that which occurs at Pikermi (Ancy-
lothervum).

MEASUREMENTS.
m. m.
Third premolar, antero-posterior diameter, .024; transverse, .025.
Fourth —“ es or s02Di- “ .028.

First molar, a se 036 ; +6 033.

PLATE I,

RESTORATION Or THE SKELETON OF COSORYX FURCATUS.

sixth natural size.

One

‘OZIS [BANJVU Y}J[IM} UO

‘MADISSOU SdOTAHAY AO NOLATANS AHL AO NOILVUOLSAY

‘TT ALVId

PLATE III.

SKULL OF APHELOPS FOSSIGER.

One sixth natural size.

No. 4.—Cristatella: the Origin and Development of the Indi-
vidual in the Colony. By C. B. Davenport.!

ConTENTS.

Page Page
i,anfowattion. ....« » » 101 5. Origin of the Central
II. Architecture of the Colony . 108 Nervous System. 127

III. Origin of the Individual . . 106
A. Observations. — Origin of
jhe Bud. 2... 107
2. The Alimentary Tract 111
8. The Central Nervous
System. . 3. . HS
4. The Kamptoderm . 115
5. The Funiculus and
Miiscles)) 0... . 1S
6. The Body-wall . . 117
7. The Radial Partitions 119
B. Comparative and Theo-
retical Review of the
Observations on the
Origin of the Indi-
Viddal s.r). oy f. 120
1. Origin of the Polypide 121
2. Interrelation of the
Individuals in the
Colony.) 22. DE
3. Origin of the Layers 123
4. Origin of the Alimen-
fry Pract. < . .. 127

6. Origin of the Funicu-
lus and Muscles’. 128
IV. Organogeny.—Development of

i. The Ring Canal . .-129
2. The Lophophore . . 1380
3. The Tentacles. . . 135
4. The —Lophophoric
Nerves’ .f ows). Wo
5. The Epistome . . . 1388
6. The Alimentary Tract 139

7. The Funiculus and
Muscles?t,, . ~~. 141

8. Origin and Develop-
ment of the Parieto-
vaginal Muscles. . 148

9. Disintegration of the
Neck of the Polyp-

GG i ate at eee

10. Development of the
Body-wall. . . . 144
SaUMMpaeVe ce ho ws ise ete Sen LEO
Bibliograpliyc< <.) os...) «-«' i.” » 145
Explanation of Figures . . . . 182

I. Introduction.

At the suggestion of Dr. E. L. Mark, I began, in the spring of
1889, the study of fresh-water Bryozoa. While at the Laboratory of the

United States Fish Commission, at Woods Holl, Mass., where, through -

the kindness of Mr. A. Agassiz, I had the opportunity of spending the

1 Contributions from the Zodlogical Laboratory of the Museum of Comparative
Zoology, under the direction of E. L. Mark, No. XIX.

VOL. XX. — NO. 4.

102 BULLETIN OF THE

following summer, I gathered most of the material for this study. I
found an excellent place for collection in Fresh Pond, Falmouth, where
_ Fredericella and Plumatella were also gathered. Upon my first visit to
this pond (July 5th), I found at its outlet Cristatella exceedingly abun-
dant on the leaves of the pond-lilies. A month later, the same locality
yielded very few specimens ; but about September 5th I found them
plentiful again, and at the same time noticed the phenomenon described
by Kraepelin and by Braem, — that some of the statoblasts of Pluma-
tella had already hatched. Colonies of from five to twenty individuals
were observed with the two halves of the statoblast still adhering to
their bases. A few colonies of Cristatella were also gathered in the
latter part of August from Trinity Lake, New York.

The material collected was killed with a variety of reagents. Cold
corrosive sublimate gave the best results. In staining, I always found
Czoker’s cochineal the most satisfactory dye for the study of the
embryonic cells of the bud.

As Haddon (’83, pp. 539-546) has reviewed the most important part
of the bibliography of budding in Phylactolemata which had been
published at the time of his writing, I shall be relieved from giving
here any extended historical account of the earlier researches. The
contributions of Nitsche (75) and Hatschek (’77) are well known.
Reinhard has published a preliminary article (’80*, ’80°) on this subject
in the Zoologischer Anzeiger; but his two more important papers
(82 and ’88) I have unfortunately not seen. Braem’s (’88, ’89%, and
°89°) three preliminary papers concerning budding in fresh-water Bry-
ozoa correct some erroneous statements of Nitsche, and support Hat-
schek’s view of the origin of the polypide. The results at which I
have arrived concerning this last problem are similar to those of
Braem, but his work has apparently been done chiefly on Alcyonella,
mine on Cristatella. Finally, I believe there will be found in this paper
something new on the organogeny, which Braem does not seem to have
especially studied, and which may be of general morphological impor-
tance. For these reasons, it has seemed to me desirable that I should
publish my observations and conclusions, and I am the more inclined
to do so because our views are not in all points the same.

In the matter of nomenclature, my studies have not led me to a final
conclusion as to the homologies of the axes of the individual, and there-
fore I fall back by preference on non-committal terms. The individual
is bilaterally symmetrical. Parts nearer the mouth end of a line joining
mouth and anus (i. e. nearer the margin of the colony) will be desig-

MUSEUM OF COMPARATIVE ZOOLOGY. 103

nated “anterior” or “oral”; parts nearer the anal end, “ posterior” or
“anal.” ‘To parts nearer the roof of the colony will be applied the
term “superior,” or ‘“tectal”; to those nearer the sole, “inferior.”
Parts situated at either side of the sagittal plane of the individual are
“lateral,” and either right or left, —the individual facing the margin
of the colony. In naming organs, I have preferably used the terms
employed by Kraepelin (’87). I adopt the term polypide simply be-
cause it is a convenient name for a number of organs closely united
anatomically, and arising from a common source embryologically.

II. Architecture of the Colony.

The colony of Cristatella, as is well known, consists of a closed sac,
which is greatly elongated in old specimens, and has a flattened base or
“sole,” and a convex roof. The wall of this sac is known as the wall
of the colony or cystiderm (Kraepelin). Suspended from the dorsal
wall, and hanging in the common cavity of the colony, which may be
culled the cwnocel, are to be seen numerous polypides in different stages
of development. A more careful observation shows that the polypides
lying nearest the median plane of the colony are the largest and oldest,
those nearest the margin, conversely, smallest and youngest (Plate I.
Fig. 1). All young colonies of Cristatella have been derived from one
of two sources, eggs or statoblasts. According to Nitsche (’72, p. 469,
Fig. 1), there are two polypides of the same age first developed in the
cystid, which is a product of a fertilized ovum, and regarding these he
fully agrees with Metschnikoff’s (’71, p. 508) statement, ‘Die beiden
Zooiden entwickeln sich wie gewohnliche Knospen.”

Nitsche (’75, pp. 351, 352) observed that in Alcyonella the primary
polypides are placed with their oral sides turned from each other, and
that the younger buds arise-in the prolongation of the sagittal plane of
the older polypides, and from that part of the cystid lying between the
cesophagus of the older buds and the margin of the colony.

As Braem (’89°, pp. 676-678) has shown, there is but one primary
bud in the statoblast embryo. ‘The younger buds formed in the stato-
blast arise on the oral side of the primary bud.

In Cristatella, says Braem (’88, p. 508), the newly hatched stato-
blast embryo already exhibits to the right and left of the adult primary
polypide two nearly complete daughter individuals of unlike age, which
are generally followed by two other sisters in the same relative posi-
tions, and a fifth in the median plane,—oral with respect to the

104 BULLETIN OF THE

mother bud. These buds may produce new ones until the whole
colony has attained the size of a pea; then young buds arise anal-
- wards of the primary polypide, and as the margin of the colony is pro-
truded on each side of this point, the colony becomes heart-shaped.
The two upper lobes of the heart are regions of great reproductive
activity ; they separate from each other, and thus transform the heart-
shaped colony into an elongated one. Through the heaping together
of buds effected by this process, a misproportion between the area
(Flichenraum) and the circumference of the colony results, and the
buds, which lie in longitudinal rows, soon come to be crowded. After
this, they each give rise to only two daughter buds, a lateral and a
younger median one.

To these observations of Braem I have little to add. I have figured
(Plate X. Fig. 88) a young colony of Cristatella, containing about thirty
polypides. This was taken in the latter part of July, and is probably
an egg colony. My reasons for thinking so are, that the statoblasts of
the preceding year form colonies in the early spring; that statoblasts
of any year have never been seen, like those of Alcyonella, to hatch in
the fall; and that there are, occupying the centre, two polypides of
very nearly equal size and development, and probably therefore of
nearly equal age. Surrounding these are eight younger individuals,
nearly equal to each other in size, and these are in turn followed by two
generations, of thirteen and seven individuals respectively, —the last
generation evidently being as yet incomplete.

As Kraepelin (’87, pp. 38, 139, 167) clearly states, the Cristatella
colony is comparable with those of Pectinatella, Plumatella, etc., and may
be derived from them by imagining a condensation of those branching
colonies. The radial partitions seen in Figure 88, di sep. r., Plate X., are
thus homologous with the lateral walls of the branches of a Plumatella
colony ; and just as in the latter, so here young individuals arise near
the tips of the branches, and the older individuals degenerate. As in
Plumatella, young individuals are produced not only distad of older, but
also laterad, thus founding new branches, so in Cristatella we find young
buds having the same positions. These facts will be better appreciated
by a reference to Figure 1, which shows a portion of the margin of a
mature colony. It is here clearly seen, (1) that, as has long been known,
the youngest individuals are placed nearest to the margin, and that
therefore, as one passes towards the centre, one encounters successively
older and older individuals; and (2) that, as Kraepelin (’87, Fig. 134) has
already figured, the older individuals are arranged in a quincunx fashion.

MUSEUM OF COMPARATIVE ZOOLOGY. 105

The bit of the margin figured may be regarded as typical, not only on
account of its symmetry, but also because of the fact that the youngest
individuals are placed at the normal distance from the margin. Al-

Figure A.

though I have seen these conditions
repeated in enough instances to assure
me of their normal nature, yet, owing
to a crowding of polypides, both among
themselves and to the margin of the
colony, and also to the consequent dis-
placement of polypides, the appear-
ances which I am about to describe are
often obscured.

First, the interrelations of the indi-
viduals included within compartments
1-8 are exactly repeated in compart-
ments 9-16. The same repetition
holds true for the remainder of this
side of the colony. On the opposite
side, the number varies from six to
eight. At the ends of the colony,
owing to crowding of individuals, it is
difficult to count with accuracy. Since
all individuals are derived from pre-
ceding ones, the conclusion seems rea-
sonable that the inhabitants of these
eight branches were derived from a
common ancestor. It is interesting
that from each of these ancestors the
same number of branches and an
almost equal number of individuals
are produced, and that the correspond-
ing individuals in each of these fam-
ilies, e. g. Figure A, 4, 5 and 12, 13,
and 7, 8 and 15, 16, are similar in

position, and of the same stage of development.

Secondly, most individuals figured have given rise to two individuals ;
some, on the contrary, to but one. Of the two individuals produced,
one (the older) passes into a second (new) compartment, and so forms
a new branch. The younger, however, remains in the ancestral com-
partment, and thus continues the ancestral branch. See, e. g., individual

106 BULLETIN OF THE

4, 5, of Figure A. The buds which give rise to new compartments may
be called lateral buds, in accordance with Braem’s terminology; those
which prolong the ancestral branch, median buds. Where only one
individual arises, it is a median bud. These conclusions regarding the
relationship of buds are based solely upon the length of the radial par-
titions, the inner extremities of which correspond to the angle formed
by two branches in branching genera like Plumatella.

Thirdly, while the lateral buds, Figure A, 4, 5, and 12, 13, give rise
directly to new buds, median buds of the same or younger age, 6, 14, have
moved to a considerable distance from their mother buds before giving
rise to new individuals. The effect of this is, that the median bud
comes to lie, not alongside of the lateral bud, but in a quincunx position
relatively to it.

Fourthly, lateral buds (branches) may arise from either side of the
budding individual. Although most of the branching in the part of
the colony figured in the cut is to the right, yet the youngest lateral
buds are being given off to the left. So in compartments 4, 6, 7, 12,
the funiculus indicating the point- where the median bud will arise.

To recapitulate : The descendants of common ancestors are arranged
similarly in the same region of the colony ; a lateral and a median bud
may arise from a single individual, the first forming a new branch, the
latter continuing the ancestral one; median buds migrate towards the
margin before producing new buds; and new branches are formed on
either side of the ancestral branches.

III. Origin of the Individual.

Two essentially different views of the origin of the polypide in the
adult colony of Phylactolemata have been maintained within recent
years. The first is that advanced by Nitsche (’75, pp. 349, 352, 353),
and adopted by Reinhard? (’80%, p. 211, ’80°, p. 235). According to
these authors, the outer of the two layers of the colony-wall gives rise,
either by a typical or a potential invagination, to the inner cell layer of
the bud, — the layer from which the lining of the alimentary tract and
the nervous system both arise, —and pushes before it the inner layer

1 Reinhard says in his preliminary article, “Meiner Meinung nach entwickelt
sich die Knospe in Folge einer Verdickung des Ectoderms, in welche dann die
Zellen des Entoderms eindringen,” but Brandt’s abstract of the paper read by Rein-
hard before the Zodlogical Section of the Russian Association, places entoderm
for ectoderm, and vice versa, —a rendering more in accordance with Reinhard’s
statements in the context.

MUSEUM OF COMPARATIVE ZOOLOGY. 107

of the colony-wall, which thus becomes the outer layer of the bud.
Hence the buds arise independently of each other.

The second view is that advanced by Hatschek (77, pp. 538, 539,
Fig. 3). He asserted that in Cristatella “Die Schichten der jiingeren
Kfospe stammen von denen der nachst dlteren direct ab.” Finally,
Braem (’88, p. 505) agrees essentially with Hatschek, and believes that
a typical double bud, although it does not always appear, is the funda-
mental condition. His preliminary account clearly shows that precisely
the same condition of affairs, except in so far as modified by the less
metamorphosed condition of the ectoderm, exists in Alcyonella as in
Cristatella.

A. OBSERVATIONS.

1. Origin of the Bud. — The result of my own work has been to lead
me to a conclusion differing from both of these two views, but more like
the second than the first. By my view, as well as by Braem’s, Nitsche’s
two types of single and “double” buds are united into one. I would not
say, with Hatschek, that the two layers of the younger bud arise directly
from those of the next older, but that each of the corresponding layers
of the younger and next older buds arises from the same mass of indiffer-
ent embryonic tissue. In some cases, each of the layers of the daughter
polypide does arise from the corresponding layers of the very young
mother bud. In other cases each of the two layers out of which the two
layers of the older bud were constructed contributes cells to form the
corresponding layers of the younger bud, but the cells thus contributed
have never formed any essential part of the older bud. All gradations
between these two types occur. Tor convenience’ sake, we may always
call the older polypide the mother ; the younger polypide, the daughter.
Figure 3 (Plate I.) shows a well advanced bud (Stage VIII.) which con-
sists of two layers of cells, an inner, 2., composed of a high columnar
epithelium arranged about a narrow lumen; and an outer, ex., of more
cubical cells. In a region (I) on the bud which is near the attachment
of its oral face to the body-wall there is a marked evagination of the
contour, caused in part by a thickening of the outer layer, and in part
by a slight increase in the diameter of the inner. This thickening o
the wall is the first indication of the formation of a younger. bud, which
is to arise at this place. Figures 22, II., 16, VI. (Plate III.), and 11,
VI. (Plate II.) show later stages of buds originating in the same manner
as that of Figure 3. The mother bud has grown larger, as has also its
lumen. The outline in its upper oral region has become much folded as

108 BULLETIN OF THE

a result of cell proliferation, and a deep pocket has been formed lined by
a layer of cells which are still a part of the inner layer of the mother
bud. The outer layer of the latter has also been protruded by the ac-
tivity of the inner layer, and its cells go to form the outer layer of the
young bud. Still another point is to be observed. The centre of the
young bud has moved away from the centre of the neck of the mother
bud, and thus the former lies nearer to the margin of the colony than the
latter. Figure 17, VII. (Plate III.) shows a still more advanced stage
in the development of the bud, in which it is sharply separated from
its parent, but its inner and its outer layers are still in direct continuity
with the inner and outer layers respectively of the mother.

I have selected this series from the many which might have been
chosen to show the origin of the polypide, because it is an intermediate
type between two extremes, and because by it the other cases receive an
easy explanation. All cases of budding, however, seem to conform to
this general law: the greater the difference in age between the youngest
and the next older bud, the greater the distance between the points at
which they begin to develop. Thus the typical case of a “double bud”’
is that in which two buds appear to arise at the same time. They origi-
nate, as Nitsche observed, from a common mass of cells. <A case of two
buds, one only slightly younger than the other, is seen in Figure 5. By
comparing with Figure 3, in which the older polypide is older than VIL.,
Figure 5, the difference between the younger buds will be apparent. On
the other hand, Figure 4 illustrates a comparatively late formation of
the younger bud. The older bud had attained a stage corresponding to
Figure 18 (Plate IIJ.), but the younger bud is not older than that seen in
Figure 22, II. Just as in the latter case the two layers of the older bud
went respectively into those of the younger, so in the present case a
direct continuity can be traced between the cells of the inner and outer
layers respectively of the younger and older buds. The evidence that
the cells composing the inner layer of the young bud have not arisen di-
rectly from the ectoderm is derived not only from the continuity of both
cell layers of the two buds, but from the presence of the apparently un-
modified ectodermic cells lying above the inner layer of the young bud,
and sharply marked off from it. Figure 6 shows a later stage of this
same type, in which the layers of the young bud are seen well formed, but
still very sharply separated from the overlying ectodermal cells. These
series afford an interpretation of the extreme type of budding shown in
Figure 2, which is not uncommon. The mother polypide has reached a
stage corresponding to Figure 18, Plate III. To the left of the neck of

MUSEUM OF COMPARATIVE ZOOLOGY. 109

the polypide and towards the margin is the funiculus, fun. Between it
and the neck of the polypide the coelomic epithelium is thickened and its
cell boundaries have become evident. Directly above this region, and im-
mediately above the muscularis, is a row of cells, which stain deeply and
show other evidences of being embryonic. These are directly continuous
with the neck cells of the older polypide, exactly as was the case with
the cells of the inner layer in Figures 4 and 6 (Plate I.). In fact, they are
in every way comparable with these. Figures 8 and 9 (Plate II.) show
slightly later stages. The funiculus has moved farther from the parent
bud, the future outer layer (ex.) has become thicker, and its cells are
columnar and sharply marked off from each other. The inner layer (2.)
of the new individual is represented by a thicker, stolon-like mass of cells,
which is in direct continuation with the innér layer of the mother bud,
from which it was doubtless derived.

A stage which, on account of the greater distance of the funiculus
from the older polypide, I believe to be slightly more advanced than
Figure 8, is shown in Figure 9. In the section drawn, the inner cell
mass (?.) exhibits few nuclei, but they are more numerous in adjacent
sections. The band of protoplasm connecting this young bud with the
mother is perceptibly smaller than in the preceding stage. The cells of
the inner layer form a mass sharply marked off from the ectoderm ;
those of the outer layer are greatly thickened, as in the last stage.

A peculiar thickening of what I regard as Nitsche’s “ Stiitzlamella ”
takes place between the young bud and the mother polypide. This is
shown in Figures 9 and 10 at mu. It is not stained by Czoker’s cochi-
neal, and the circular muscle fibres, here cut transversely, are very con-
spicuous in the midst of it. As I have noticed this appearance only in
the cases of young buds which have originated like those of Figures 9 and
10, and of others of about their age, (and it is in these buds and at this
age that migration from the older polypides takes place,) I believe that
there is some connection between this condition of the muscularis and
the disturbance which such a migration must cause.

I have already (page 106) referred to the fact that in some cases me-
dian buds are found far removed from the mother polypide, although
in an early stage of development. Stage IV. is the youngest in which
such buds have been found.

The cells of the mass destined to form the inner layer of the bud
multiply rapidly after they have reached a proper position, and there is
considerable protrusion of the coelomic epithelium into the body cavity.
The fact, that during its extensive migration the bud increased only

110 BULLETIN OF THE

slightly, whereas it now begins to develop rapidly, leads to the presump-
tion that it has ceased to migrate, and has come to a state of comparative
rest, relative to the surrounding ectodermal cells.

One of the first indications of further development is seen in the
arrangement of the cells of the inner layer, which is such that their
nuclei come to lie near the surface of a hemisphere whose convex side is
turned toward the coenocel. The beginning of this process is seen in
Figure 10 (Plate IT.), and, further progressed, in Figure 11. Figure 14
exhibits a still later stage in the development of the polypide. In the
lower portion of the two-layered sac of this figure a separation of the
cells (lu. gm.) has begun. ‘This is the first indication of the atrium.

In all cases there exists at this stage a condition of the ectoderm like
that shown in Figures 5 and'14. The absence of ectodermal cells directly
over the bud may be accounted for by supposing that they have come to
lie upon, and form part of, the neck of the polypide. While it would be
impossible to deny that they maght migrate through the cells composing
the neck of the polypide, and thus come to form the nervous elements,
a careful study of the successive stages figured will not show the slight-
est evidence of any such migration, nor is it a preort probable, from what
ts known of the action of epithelium the world over, that such a migra-
tion would occur.

According to the description of Nitsche (75, p. 353), there is a
lumen in the bud of Alcyonella (where the ectoderm is much less meta-
morphosed than in Cristatella), which is always in direct communication
with the outer world, the bud having been formed by a typical invagi-
nation. Braem (’88, pp. 506, 507), however, states that he has never
seen in Alcyonella this communication of the bud cavity with the outer
world. In the much more obscure process of polypide development
in Gymnolemata the lumen first appears after the cells of that mass
from which the bud is to arise have arranged themselves in two con-
centric layers. In Endoprocta, according to Nitsche (75, p. 374) and
Seeliger (’89, p. 179), the lumen arises by a virtual or actual invagi-

1 Since writing the above paragraph I have cut some sections of Plumatella
in which this process is much clearer, owing to the absence of secreted bodies in
the ectoderm. Instead of a few ectodermal cells dropping down upon the upper
part of the neck of the polypide, as is the case in Cristatella, there is a cup-shaped
invagination of the ectoderm, which is quite deep, and thus gives rise to an elongated
“neck.” That none of these ectodermal cells go to form any part of the polypide
proper is certain in Plumatella. But it is also true, that ectodermal cells are thus
incorporated into the neck of the polypide, and probably into the stolon which pro-
ceeds from it.

MUSEUM OF COMPARATIVE ZOOLOGY. 11%

nation and remains always in communication with the surrounding
medium. In Cristatella the lumen is formed in the bud at the time
when its diameter perpendicular to the roof of the colony slightly ex-
ceeds that parallel to it. As Figure 14 (Plate II.) shows, this cavity
(lu. gm.) first makes its appearance in the distal part of the central
mass of cells. There are always cells lying above the lumen, and thus
cutting off the ectoderm from contact with it. The two layers from
which, according to my view, all of the cells of the adult polypide are
derived, are now completely established ; and the cavity has already
appeared which, by enlargement, out-pocketing, and the concrescence of
its walls, gives rise to the atrium, and the lumina of the alimentary
tract and supra-cesophageal ganglion.

The bud elongates, and often at this time, preparatory to giving rise to
a new bud from its upper marginal angle, becomes bent or curved, the
concavity always being next the daughter bud (see Figs. 3, 5, 11, and
22). By this change in form the bud becomes bilaterally symmetrical.

2. Origin of the Alimentary Tract. — The first organ derived from the
two-layered sac is the alimentary tract. Nitsche (75, p. 356) described
the process of its formation ina very clear manner ; but I believe he is
in error. The original lumen of the bud represents, he says, the atrium
and the lumen of the alimentary tract. The part lying nearest the
attached end of the bud gives rise to the former; the latter is derived
from the lower part of the lumen. These two regions become separated
by the invagination of the two layers of the bud along a furrow on each
side of the bud; just as though the walls of a two-layered hollow rubber
ball were pressed together by a finger of each hand acting at opposite
sides until the points of the fingers should be separated by the four
layers of the ball only. By this process, mouth, anus, and the entire
gut, would of course be formed at one time. Reinhard (’80%, p. 212)
appears to agree with Nitsche as to the method of origin of the alimen-
tary tract. Braem (’89, pp. 677, 678) describes and figures diagram-
matically this process in the statoblast of Cristatella. In the median
plane of the bud there is an out-pocketing from the anal side of the
atrium which involves both layers of the bud ; it assumes the form of a
comma, its blind end curving forward to meet the blind end of a lesser
evagination of the oral part of the atrium, — the cesophagus. The blind
ends of these two pockets meet, and by the breaking through of the
intervening tissue their lumina freely communicate with each other,
' thus completing the alimentary tract. He adds: “ Auch bei den Poly-
piden der fertigen Colonie der Darm durch Verwachsung eines analen

112 BULLETIN OF THE

und eines spater auftretenden oralen Schlauches, an deren Bildung
freilich das dussere Knospenblatt nur secundiar sich betheiligt, zu Stande
kommt.”

My own observations are nearly in accord with the statements of
Braem, as opposed to Nitsche’s. The older bud of Figure 11 (Plate II.)
shows the first indication of the lumen of the posterior part of the ali-
mentary tract near the attached portion of the bud. The cells of the
inner layer are multiplying, and the lumen of the bud is broader here
than elsewhere. The position of a daughter bud, VI. (on the oral side of
the one under consideration), sufficiently indicates that the point marked
rt. is in the region of the future anal opening. Figures 12 and 13 (Plate
II.) show further stages in this same process. The lumen of the intestine
is formed, not by constricting off a part of the original lumen of the bud,
but by the rearrangement of cells at the progressing blind end of the
pocket, which gradually moves towards the distal part of the larger or
bud cavity. It is important to establish the fact that the alimentary
tract is formed in the inner layer of the bud, and that its cells alone line
the digestive cavity. Figures 20 and 21 (Plate III.) represent two suc-
cessive sections out of five which pass through the inner layer; namely,
the second and the third counting from the attached to the free end of
the bud. The sections were cut at right angles to the plane of Figure
11 nearly along the lines 20 and 21 respectively. It will be seen that
the inner layer alone is implicated in the lining of the alimentary pocket
at this early age. They also show clearly the incorrectness of the state-
ments of Nitsche on this point. Figures 24, 25, and 26 (Plate IV.) are
three sections cut through a bud of about the age of that represented in
Figure 13 (Plate I1.), and at right angles to the plane of the latter, and
in the direction of the lines 24, 25, and 26 respectively. A section cut
beyond the end of the intestine, in Figure 13, is not represented. It
shows that the lumen of the alimentary tract is absent at this plane. A
comparison of Figures 20 (Plate III.) and 26 (Plate IV.) shows that the
lumen of the bud, dw. gm. (which we may call the atrium from its resem-
blance to a space having the same relations in Entoprocta), has increased
in volume owing to a growth of the lateral walls. On account of the
more rapid elongation of the anal than of the oral side, the axis of the
alimentary tract comes to take a horizontal position, as shown in Fig-
ures 17 and 18, Plate III. (Compare also Figs. 27-29, Plate IV.) The
blind end of the digestive sac comes very close to the blind end of
another pocket formed on the oral side, the cesophagus, and soon the
two communicate directly. At the same time, the inner cell-layer of the

|
;
'
‘

a

MUSEUM OF COMPARATIVE ZOOLOGY. 113

middle portion of the alimentary tract has been quite cut off from that
of the atrium by a constriction, the beginning of which is seen at ex.,
Figure 24 (Plate IV.) and in a later stage at ex., Figure 28. The cells
of the outer layer are next pushed into the place of constriction and re-
move the alimentary tract at this point still further from the atrium, as
is shown in Figures 18 (Plate III.) and 28, ex. (Plate IV.). The error
of Nitsche is explainable on the ground that he believed the stage of
Figure 18 to be the earliest in the development of the alimentary tract.
3. Origin of the Central Nervous System. — Metschnikoff (71, p. 508)
first clearly recognized that the supra-cesophageal ganglion of Phylacto-
lemata is derived from the inner cell-layer of the bud, — the same
layer which gives rise to the inner lining of the alimentary tract.
Nitsche (75, pp. 359, 360) described and figured in an insufficient and
not wholly accurate way the process of the formation of this organ.
According to my observations, the central nervous system arises directly
over the middle of the horizontally placed alimentary tract in the posi-
tion marked gn. in Figure 18, Plate III. (compare also Figs. 17 and 28,
pam. gn.). The process by which the ganglion with its internal cavity
(Plate VIII. Fig. 73, lu. gn.) is formed will be more easily understood if
the reading of the text be accompanied by reference to the following sec-
tions. Figures 17, 18, 19, Plate III., and Figure 73, Plate VIII., show
successive stages in sagittal section. Figures 27-29, Plate IV., from a
single individual, are vertical right-and-left sections, the positions of
which are indicated by the lines 27, 28, 29 of Figure 17. Figures 30-32
are similar sections from an older individual (see lines 30, 31, and 32,
numbered at the lower border of Fig. 18), and Figures 33-38 are
from a still older polypide (compare lines 33-38, Fig. 19). By a study
of these sections, it is seen that the cells forming the floor of the brain,
pam. gn., are derived from the inner layer of the bud, and indeed from
the very region of the layer which furnished cells to line the alimentary
tract (Plate II. Fig. 13, Plate IV. Figs. 25 and 24, ga.), and therefore
that the layer of cells forming the floor of the ganglion is directly con-
tinuous posteriorly through the anal opening (Plate II. Fig. 13, an.)
with the wall of the rectum, and anteriorly with the lining of the
cesophagus. ‘The first marked differentiation of this region is effected
by the sinking of the centre of the floor of the neural tract (Fig. 18, gn.),
thus forming a shallow pit, which opens directly into the atrium above.
The closure of the walls of the ganglion above must now be considered.
Concerning this process, Nitsche says: “ Die Rinder dieser Einstiilpung
[my ‘shallow pit’] wachsen nun wie die Rinder der Medullarrinne
VOL. XX.— NO. 4. 8

114 BULLETIN OF THE

eines Wirbelthierembryos gegen einander, und wie in letzterem Falle
eine hohle Réhre von der dorsalen Leibeswand des Thieres abgeschniirt
wird, so wird hier eine hohle Blase von der Wand des Polypids abge-
schnirt. In unserem Falle ist aber die Wandung an der diese Ab-
schniirung vor sich geht, zweischichtig.” The two layers referred to were
those of the median walls of a pair of invaginations of the latero-anal
sides of the wall of the atrium, —the beginnings of the lophophoric
arms (br. loph., Figs. 37, 38, Plate IV., and Figs. 61, 62, Plate VII.).

The process of closure is in reality somewhat different from Nitsche’s
conception of it. The axes of the pockets which go to form the lopho-
phoric arms are, at first, directed inward, upward, and slightly oral-
ward (Plate I. Fig. 7, br. loph.). By means of these invaginations the
cell layers lining the atrium on opposite sides are brought into contact
at a point between the rectum and the ganglionic pit (Plate V. Fig. 43,
loph.’). This approximation of the walls may, perhaps, better be said
to be a continuation upward of the process by which the alimentary
tract was cut off from the atrium (after the lumen of the former was
formed), and by which cells of the outer layer of the bud came to
intrude themselves between these two regions (Plate IV. Fig. 35, ex.) ;
for the lateral furrows, by the formation of which this act is performed,
are, on each side, continuous with the lophophoric pockets, and above
end blindly in them. By the approximation and fusion of the inner
layers of the atrium several things are accomplished. The posterior
wall of the brain is formed (Plate IV. Fig. 39, loph.’), the anus is car-
ried farther up (compare Plate III. Fig. 19, and Plate VIII. Fig. 73, an.),
and by a continuation of the constricting process the cavities of the
lophophore on opposite sides of the polypide are brought into communi-
cation between the ganglion and the rectum at a point opposite the
letters Zu. gm. in Figure 63 (Plate VII.), whereas they formerly com-
municated only outside the alimentary tract.

Oralward from the lophophoric pockets there is a thickening of the
inner layer above the floor (pam. gn.) of the ganglion on each side
(Plate IV. Figs. 28 and 31). Later, each of these thickenings becomes
a fold involving the inner layer of the bud only (Plate IV. Fig. 35).
The upper and lower halves of this pair of folds respectively fuse in the
sagittal plane, the last point at which the union occurs being near the
cesophagus (Plate III. Fig. 19). Anteriorly the rim of the shallow brain-
pit rises up as a third fold, and the ganglion becomes a sac whose mouth
is bounded by the edges of the folds, the advance of which causes it to
become more and more constricted. These folds are the pair of folds

MUSEUM OF COMPARATIVE ZOOLOGY. 115

above the cavity of the ganglion, and the one between the cavity of the
ganglion and the csophagus. The outer layers of these three folds
respectively fuse immediately behind the cesophagus; the inner layers
are constricted off, but without closing the neck of the sac. Conse-
quently the neck of the ganglionic sac, instead of opening into the
atrium, now abuts upon the inner cell-layer at the angle between the
floor of the atrium and the esophagus. The lower layers of the hori-
zontal folds thus become the upper wall of the ganglion (Fig. 35, ¢ct.
gn.); the upper layers form the new floor of the atrium (Fig. 73,
pam. atr.), which lies between the lophophore arms, is continuous with
its median walls, and passes over into the walls of the alimentary tract
both in front and behind, The outer layer of the young bud only
secondarily makes its way in between the upper and lower layers of ’
these folds. It ultimately takes the form of a double layer embracing
a space, which is the epistomic canal. (Plate VIII. Fig. 73, lu. gn.,
Plate V. Fig. 52, lu. gn., can. e stm.)

4. Origin of the Kamptoderm. — While the alimentary tract, lopho-
phore arms, and nervous system are being marked out in the lower por-
tion of the bud, these organs become farther removed from the wall of
the colony by an enlargement of the atrium to meet the demands of
the augmenting volume of the lophophore. Pari passu with this en-
largement of the atrium, its walls diminish in thickness (compare kmp.
drm., Fig. 73, Plate VIII., with Fig. 18, Plate III.). This is rather the
result of a failure of the cells to multiply in proportion as the area of
the wall increases, than of a decrease in the number of cells already
formed. Both the inner and outer cell-layers of the bud take part in
the formation of this wall, as is evident from the figures. The wall of
the atrium was called ‘‘tentacular sheath” by Allman (56, p. 12) and
Nitsche, but Kraepelin (’87, p. 19) employs the name “kamptoderm ”
for this structure. I prefer this term to “tentacular sheath,” and have
employed it both on account of the reasons given by him and because it
may be easily inflected, whereas “tentacular sheath” may not. The
kamptoderm, then, is formed of the upper portion of the bud, and both
of its cell-layers are concerned in its formation and persist in the adult.

5. Origin of the Funiculus and Muscles. — Nitsche (’75, pp. 353, 354)
did not see the origin of the funiculus, but states that it suddenly occurs
lying close along the oral side of the bud, to which one end is at-
tached. Its proximal end is fastened, he says, to the inner layer of the
colony-wall, and by the growth of the latter between the funiculus and
the neck of the bud this end retreats from the young polypide. Braem

116 BULLETIN OF THE

(88, p. 533) asserts that the funiculus arises as a longitudinal ridge on
the outer layer of the oral wall of the young polypide at the time of
the formation of the alimentary tract, and that the cells of this ridge
are cut off from the bud to form the funicular cord. Soon after this,
embryonic cells from the inner layer of the young polypide penetrate
into the midst of the cord through its proximal end, and thus lay the
foundation of the statoblast.

Concerning the origin of the muscles, Nitsche (’75, p. 354) states that
they are simple elements of the outer cell-layer of the bud, which were
originally situated in the angle of attachment of the bud to the inner
layer of the colony-wall, and that by the growth of this wall they be-
come drawn out into spindle-shaped cells.

I have decided to treat of these two organs together, since their ori-
gin and development are curiously similar. According to my belief,
both arise, in part at least, from the inner cell-layer of the colony-
wall. At a stage slightly earlier than that of the first appearance of the
fully formed funiculus (Plate II. Fig. 11, cl. fun.), I have always found
a disturbed condition of the coelomic epithelium. This is particularly
noticeable on that side of the young lateral bud upon which the median
bud is about to arise. In some cases I have seen the cells of this
layer taking on all the characters of wandering cells, as seen at cl. fun.,
Figure 22, Plate III, where some have already begun to group themselves
into a funiculus-like cord. At Figure 57, cl. fun., Plate VI., the funiculus
is seen lying close to the oral wall of the polypide. That it has not
arisen in precisely the manner described by Braem is probable from this
figure alone, for the proximal end of the funiculus is not yet connected
with the wall of the colony. If my view is correct, this connection
arises only secondarily (Fig. 2, fun.). I am, however, inclined to be-
lieve that the distal end of the funiculus arises in a different way from
the proximal, and in the manner described by Braem. My evidence
for this is, that I have twice seen at this point cells in the act of
dividing so as to contribute daughter cells to the funiculus. Figure 53,
Plate VI., shows the condition of the distal end of the funiculus, fun.,
which passes, without any line of demarcation, into the outer layer of
the bud; this layer is normally one cell thick, but in the region of
funicular formation it is two cells thick, The proximal end of the
funiculus is, at this stage, attached to the celomic epithelium of the
roof of the colony, tct. That an attachment should occur in this man-
ner, and become quite intimate, is not strange, considering the origin
of the funiculus from amoeboid cells, and the fact that, even at a late

~

MUSEUM OF COMPARATIVE ZOOLOGY. LVF

stage of development, this character is still retained by much of its
tissue. (See, for example, Fig. 77, fun., Plate TX.)

The great retractor and rotator muscles have, I believe, like the funic-
ulus, a double origin. ‘They arise from the outer layer of the bud, on
the one hand, and from the celomic epithelium on the other. The
first indication of the differentiation of the muscle cells consists in a
disturbance in the upper lateral edge of the outer layer of the bud at
about the stage of Figure 17, Plate III. This is shown in dorso-ventral
sections through this region (cl. mu., Figs. 24, 26, Plate [V.). Later, the
disturbance becomes more marked, and cells having a semi-amceboid
character appear to be proliferated (cl. mu., Fig. 33, Plate IV.), and to
migrate from the bud towards the ccelomic epithelium. During this
process the cells of the latter layer also are active, and some of them,
elongating, reach towards the young polypide, as seems to be clearly
shown at cl. mus., Figure 54, Plate VI. It is significant that, since each
of the two upper lateral edges of the bud lies near a radial partition, the
muscles also are always formed in close proximity to one (di? sep. r., Fig. 54,
Plate VI.; Fig. 30, Plate IV.). It will thus be observed that, both in the
case of the funiculus and of the muscles, the end which is attached to
the wall of the colony arises at a point which is remote from that of its
attachment to the adult. The migration to the later positions will be
treated of farther on. (See page 141.)

6. Origin of the Body-wall.— As already shown (page 104), the
body-wall of the individual of a Cristatella colony includes not only
the endocyst of authors, — the roof and the sole, — but also the radial
partitions.

Braem (’88, pp. 506, 507) concludes “dass die polypoide Knospen-
anlage . . . nicht allein das Polypid nebst den Tochterknospen liefert,
sondern dass auch die zugehérigen Cystide aus ihr und zwar aus ihrem
Halstheil entwickelt werden.” I believe that a portion of the “ cystid,”
or body-wall, is thus formed in Cristatella, but not the whole.

If one compares the relations of the polypide to its danghter bud in
Figures 3 (Plate I.) and 17 (Plate ITI.), and reflects that later the daugh-
ter bud is to be found still farther from the mother bud, he is forced to
one or the other of two conclusions: either the young bud is pushed
from the mother by a proliferation of cells from the neck of the polyp-
ide without causing an increase in the length of the body-wall itself,
or there is an actual increase in the length of the body-wall, produced
either by the proliferation of cells already existing in it, or by the addi-
tion and subsequent proliferation of cells from the neck of the mother

\

118 BULLETIN OF THE

polypide ; and this increase in length, occurring between the polypide
and bud, carries the two apart. Unfortunately, I am unable to state
definitely how this migration of the young bud away from the mother is
effected. If the ectoderm increases in length between the two buds by
the proliferation of cells already existing in it, that fact ought to be
evinced by a distorted condition of the old cell-walls of the highly meta-
morphosed cells of the ectoderm. For, since most of the active proto-
plasm is at the base of the ectoderm, its area will increase faster than
will the area of the surface of the ectoderm; and the latter will either
rupture or stretch, or else the ectoderm will become concave on its
outer side. An application of these criteria to sections of the body-wall
in the budding region leads to the conclusion that the ectoderm of
Cristatella increases here very slightly, if at all, by a proliferation of
cells already existing in it. A search for cell division in this region
has yielded the same negative results. There can be no doubt that
cells are added to the ectoderm from the neck of the polypide. The
process takes place, however, after the daughter bud is well established
at some distance from the mother bud. The proliferation of these cells
ruptures the old cell-walls of the ectoderm, and increases the area of
the body-wall. I shall have occasion to speak of this process more
fully in treating of the later period to which it belongs.

There remains, then, the conclusion, that the cells which go to form
the inner layer of the young bud are pushed from the neck of the next
older bud by a proliferation of cells in the stolon-like mass, without
causing an increase in the area of the body-wall itself. Moreover, I
have seen cell proliferation in the stolon-like mass. Another series of
facts will lead us to this same conclusion.

Though the body-wall does not increase by cell proliferation between
buds, it does so, I believe, at the margin of the colony. This, it is
true, cannot be directly observed with ease, since the multiplication of
cells, which tends to increase the breadth of the colony, must also occur
at the margin, and one cannot be certain what dimension of the colony
wall will be augmented by any given case of nuclear division. My
belief rests on the following evidence. (1) In the same adult colony
the distance of the youngest bud from the margin is not the same
in all regions. This is not what we should expect if the distance of
the youngest polypides from the margin remained unchanged during the
growth of the colony. (2) There is a gradual increase in the amount
of metamorphosis exhibited by the cells as one passes from the margin
towards the middle of the roof. Figure 60 (Plate VI.) shows a rather

MUSEUM OF COMPARATIVE ZOOLOGY. 119

marked example of a very common, although not universal, condition
of the lateral margin of the colony. The epithelium of the margin is
composed of columnar cells, which are higher (54 ») than those of the
roof (48 »), and also of a less average diameter (8.4 ) than the latter
(18.2 »). Moreover, the cells are very little metamorphosed. In pass-
ing towards the roof (¢ct.), the cells are seen to become more and more
metamorphosed, the secreted bodies (cp. sec.) becoming relatively larger.
Figure 55 represents the margin in a more metamorphosed state than
Figure 60. Although this condition of things is not incompatible
with the idea of a passive margin, it strongly suggests that this region
is one of proliferation, by which cells are added to the roof, and thus
the distance from the youngest polypide to the margin is virtually
increased. This conclusion receives a very important confirmation
from the study of the origin of the radial partitions, the treatment of
which must be deferred for the moment. Although new cells are being
added to the roof at the margin, yet the distance from the youngest
polypide to the margin is not greater in old than in young colonies.
How, then, is the approximate constancy of this distance maintained ?
Evidently it can only be by the process (which I have already shown
must take place) of migration of some of the young buds at. the base
of the ectoderm, particularly in the case of median buds. The ten-
dency of the migration of young buds towards the margin is to diminish
the distance between the front of the budding region and the margin
of the colony. The tendency of cell proliferation at the margin is to
increase that distance. The actual distance is the resultant of these
two opposing factors, and may be less or greater in different parts of
the same colony, according as the one or the other is the more active.
If we assume, further, that the cells added to the roof and sole from
the margin plus those derived from the necks of the polypides are
equal in amount to those lost by the degeneration of individuals in the
middle of the colony, we have a sufficient explanation of the fact,
observed long ago, that the adult colony of Cristatella maintains a
nearly constant width.

7. Origin of the Radial Partitions. — I know of nothing on this sub-
ject by any previous author. The radial partitions consist of a muscu-
laris covered on both faces by a very thin epithclial layer (Plate X.
Fig. 95,1). The muscle fibres of the muscularis arise from the already
formed longitudinal muscles of the wall of the colony at the region of
transition from the sole to the roof (Plate VI. Fig. 55, mu.). As the
muscle fibres move into the ccenoceel, they carry before them the ccelo-

120 BULLETIN OF THE

mic epithelium of the region from which they arise. It is owing to
this method of origin that the epithelium comes to clothe both faces
of the partition. ‘The process by which the muscle fibres move into
the ccenoceel appears to be this. The end of a fibre nearest the roof
becomes fixed to a certain part of the muscularis of the roof, and is
left behind with it when the margin is carried outward (potentially) as
the result of cell proliferation. Thus from a nearly horizontal position
the fibres attain a direction at first oblique, and then perpendicular to
the sole. In some instances the upper ends of the fibres move through
an arc of more than ninety degrees, so that they are ultimately directed
upward and inward, i. e. towards the centre of the colony. (Compare
mu, mu’, mu, Fig, 55, Plate VI.). This process is also indicated in
two horizontal sections (Plate X. Figs. 95 and 96), the former being
nearer the sole than the latter. This is a region of active budding, and
in consequence new compartments or branches are being rapidly formed.
The numbers 3, 4, 5, and 6 (Figs. 95, 96) show the positions of young
partitions, which are shorter above than below, owing to the oblique
position assumed by the innermost muscle fibres of the partition. The
oblique position is due to the fact already demonstrated (Fig. 55,
Plate VI.), that the tectal end of the muscle of the partition first
appears at the margin nearer the sole than the roof. At 2 (Fig. 96)
there is apparently an interesting case of the formation of a new par-
tition by the detachment of certain fibres from the muscularis of an
old one. The fibres, moving away laterally, take with them a covering
of ccelomic epithelium. Near the sole this process has progressed far-
ther than it has nearer the roof, so that in Fig. 95 the detachment
appears complete, whereas in Fig. 96 the union is still visible. This
method of formation is intelligible when one considers that the muscu-
laris of the partition often contains more than a single layer of muscle
fibres. Thus, in Figure 87, mu., there are two or three layers of fibres
in the section. Figure 86 represents a section cut vertically and at
right angles to a partition near its union with the marginal wall of the
colony, and shows three fibres of the longitudinal or inner layer of
the muscularis lying side by side in the partition.

B.— CoMPARATIVE AND THEORETICAL REVIEW OF THE OBSERVATIONS ON
THE ORIGIN OF THE INDIVIDUAL.

What bearing have the facts here adduced on those given for other
groups of Bryozoa, and what is their probable significance in relation
to the general problem of non-sexual reproduction ?

MUSEUM OF COMPARATIVE ZOOLOGY. 121

1. Origin of the Polypide. — Lateral budding (as distinguished from
linear budding, such as occurs in Turbellaria, Cheetopods, c&c.) may
be roughly classified under two types, in one of which the young indi-
vidual arises dzrectly from the body-wall of the parent, as in Hydra.
In the other, the young arise, one after the other, from a mass of
embryonic material derived from a parent individual, — from a stolon,
as in Salpa. In the group of Bryozoa both of these methods seem to
be present. In such a form as Paludicella (Allman, ’56, pp. 35, 36,
Korotneff, ’75, p. 369) we have an example of the direct type; in Pedi-
cellina we have a stoloniferous genus. Also in the marine Ectoprocta
examples of both types appear to occur (e. g. Flustra, Hypophorella).
To which of these classes does budding in Cristatella belong? It seems
to me that we have here an instructive example of a transitional condi-
tion. The young polypide of Figure 3 arises directly from the mother
polypide, and may represent a case of the first class. Is the type of
Figure 2 a representative of the stoloniferous class? It seems to me
that it partakes of the essentials of that class, although, as I have

shown, it may be united by intermediate stages with the first class. —

I understand a stolon, in its morphological sense, to signify a mass of
embryonic cells derived from a parent individual, and capable of repro-
ducing non-sexually one or more daughter individuals at some distance
from that parent. The condition shown in Figure 15, Plate II., in which
the embryonic cells of the two layers represent the stolon, may fairly
be said to answer to this definition. The mass of cells (III.) represents,
then, the distal end of the stolon. But the stolon does not end here,
although its further progress towards the margin is delayed. Not all
of its cells go to form the polypide which arises at this place. On the
contrary, some of them remain in the ‘neck ” of the new polypide, in
an indifferent histological condition, and later give rise, either directly,
or by the intervention of a typical stolon, or by both, to one or two
new buds. Those cells of the neck which do not thus pass over into
new buds for the most part degenerate (page 144). According to this
view, the neck of the polypide is to be regarded as at first essentially
a portion of the stolon.

2. Interrelation of the Individuals in the Colony. — The interrelation
of individuals in the colony in Cheilostomata has been most carefully
investigated from a morphological standpoint by Nitsche (’71, pages 35,
36), who showed that, in opposition to Smitt’s theory, each new indi-
vidual arose from a single preceding one, and that the latter, in order
to increase the breadth of the colony, might give rise to two individuals

a2 BULLETIN OF THE

instead of one. Reichert (69, p. 311, Fig. 28, Plate VI.) has shown
that in Zodbotryon (one of the Ctenostomata) “an der Mantelfliche,
und zwar einseitig, inseriren die Bryozoenkipfe mit Alternation in par-
allelen, wie es scheint, langgezogenen spiralig verlaufenden Reihen an-
geordnet.” Nitsche (75, p. 370) states that the buds in Loxosoma
arise from the mother alternately on opposite sides, and that the
younger the bud, the nearer it is to the foot of the parent individual.

Both Hatscheck (’77, pp. 517, 518, Fig. 33, Plate XXIX.) and Seeli-
ger (’89, p. 176) show that in Pedicellina young individuals are devel-
oped in the plane of the older ones, and are successively formed at the
growing tip of the stolon, towards which the cesophageal side of all
individuals is turned. ‘This relation is the same as that which we have
found in Cristatella. In Cheilostomata, however, it is apparently the
anus which is turned towards the budding margin. °

Thus, throughout the group of Bryozoa, we find that the position which
young buds assume in relation to older individuals is very definite.

I am inclined to believe that the radial partitions of Cristatella sep-
arate the morphological equivalents of the isolated branches of such a
form as Plumatella punctata (see Kraepelin, ’87, Taf. V. Figs. 124, 125).
The type of budding which gives rise to the series of median buds may,
then, be represented, as seen from the side, by

es

F 8 z Figure B. The margin (*) will then represent

on , that portion of the body-wall of the youngest
- individual, which will give rise to a part of
TGURE B,

the body-wall of the next younger individual.
The process by which the body-wall of the individual of Cristatella is
formed is therefore, in my opinion, different from that which Braem
describes in the case of Alcyonella, for he maintains that in Alcyonella
the proper body-wall of an individual arises later than its polypide. In
fact, the tip of the branch of Alcyonella is somewhat different from
that of Cristatella. In the former, it is occupied by the polypide of a
budding individual; in the latter, a part of the body-wall of the bud-
ding individual is pushed out beyond the polypide. In the former, the
foundations of the daughter polypide are pushed out upon the body-
wall of the mother, and begin to form their own proper body-wall; in
the latter, the young bud migrates away into the modified part of the
body-wall of the mother, which forms the extremity of the branch, and
which now becomes a part of the body-wall of the daughter polypide.
This distal part of the body-wall grows independently of the polypide
by interstitial growth, and thus differs from any part of the body-wall

MUSEUM OF COMPARATIVE ZOOLOGY. 123

of the individual of Alcyonella, for all of it, according to Braem, is de-
rived from the neck of its own polypide. This last method of origin
of the body-wall I believe to be also present in Cristatella, as well as in
Alcyonella, as I shall have occasion to show later (page 144).

In Paludicella, according to both Allman (’56, pp. 35, 36) and Korot-
neff (’75, p. 369), the formation of the body-wall of the new individual
is begun before the appearance of the polypide. In Cheilostomata, as
both Nitsche (71, p. 22) and Vigelius (84, p. 75) have shown, and in
Ctenostomata, as demonstrated by Ehlers (76, pp. 91, 92), the “ zow-
cium” arises before the polypide takes on its definite form.

3. Origin of the Layers. — Although later researches have only con-
firmed the conclusion arrived at long ago, that in Tunicates cells from
all three germinal layers of the parent pass over into the bud, the
facts in Bryozoa have seemed not to favor the view of the fundamen-
tal nature of this process. To be sure, Hatschek (’77, pp. 517-524)
believed that he had found evidence of a condition in the budding
of Pedicellina exactly comparable with that in the budding of Tuni-
cates ; but the more recent studies of Harmer (’86, p. 255) and Seeliger
(89) have failed to confirm his results, if they have not es
explained the source of his error.

What is the relation of the condition I have described in Cristatella
to the question of the transmission of a part of each germinal layer to —
the bud, and in how far do the conditions here agree with our present
knowledge of the budding process in other groups of Bryozoa? AlI-
though my results accord with Hatschek’s in this, that the youngest
and next older buds are intimately related, that the corresponding
layers in each are derived from the same cell layers, and that the inner
layer of the bud is not derived directly from the overlying ectoderm,
they do not strengthen the idea of the fundamental importance of his
doctrine, ‘‘ Die Schichten der jiingeren Knospe stammen von denen der
nichst alteren direct ab.’ Moreover, they afford no evidence of the
uccuracy of his conclusion, that the inner layer of the bud is com-
posed of entoderm ; indeed, since this inner layer does not give rise to
the alimentary tract alone, as he supposed, but to the nervous system
also, the facts in Cristatella tend to weaken his hypothesis. In order to
determine finally just what the origin of the stolon from which the
inner layer arises is, it will be necessary to study the origin of the first-
formed polypides. This I have not yet been able to do. Our present
knowledge on the subject is still in an unsatisfactory state.

Allman (’56, pp. 33, 34) has described and figured some stages in the

124 BULLETIN OF THE

development of the egg, but without referring to gastrulation, or the
layers involved in the first polypide.

Metschnikoff (71, p. 508) and Nitsche (’75, p. 349) maintain that
the outer layer of the embryonic “ cystid” goes to form the inner layer
of the primitive polypides, and that its inner layer forms the outer
layer of the polypides.

Reinhard (’80%, pp. 208-212) is more explicit concerning the early
stages than preceding authors. Apparently the egg segments regularly,
and undergoes embolic invagination, The blastopore closes. There is
a circular groove in the anterior part of the embryo (Barrois’s mantle
cavity), and from the cap or “hood” which the mantle cavity sur-
rounds, the wall of the ‘‘cystid”” or colony-wall is subsequently formed.
The embryo is already composed of three layers, ‘an outer, the tunica
muscularis, and the entoderm.” All three layers of the “ hood ” share
in the formation of the polypides, but the fate of each layer is not
clearly described.

Haddon (’83, p. 543) suggests that the gastrula is to be regarded as
one in which the alimentary tract is retarded in development, and that
the enlarged coelomic diverticula, such as occur in Sagitta, etc., line
nearly the whole of the so-called archenteron. From the small mass of
true entoderm at the pole opposite the blastopore the alimentary tract
arises. This suggestion, unfortunately, has no positive facts for its
support, and could be of service only upon the assumption that the
alimentary tract of the first polypide is formed from the inner layer of
the ‘“cystid”; but this assumption is contrary to the observation of all
who have written on this subject.

Kraepelin (86, p. 601) has also observed the “ gastrulation,” but he
believes that it is to be interpreted as the precocious formation of an
enteroccel, in which case the invagination to form the first polypide is
to be regarded as the true gastrulation, the inner layer of the cystid as
mesoderm, and the inner layer of the bud as entoderm.

By far the most satisfactory and complete account of the embryology
of fresh-water Bryozoa is that of Korotneff, ’89. The genera studied
were Alcyonella and Cristatella. Since the development takes place
inside of an ocecium, the use. of the section method is necessary for the |
elucidation of the details of the embryological processes. Apparently
the egg segments regularly and forms a blastula. Loose cells are given
off from the inner surface at one pole of this blastula. These arrange
themselves in an epithelium, lying immediately inside of the ectoderm,
over a part only of its inner surface ; so that while the upper two-thirds

MUSEUM OF COMPARATIVE ZOOLOGY. 125

of the embryo has two layers, the lower third is one-layered. The
cavity of the lower third contains some scattered cells, which, the au-
thor hints, may be representatives of the mesoderm, while the cavity in
which they lie may represent an enterocel. The author regards the in-
ner layer of the upper two-thirds as true entoderm. The method of its
formation recalls that of the entoderm of some Coelenterata, as demon-
strated by Metschnikoff. There is no epithelial invagination, such as
Kraepelin maintained, and therefore the cavity which the inner layer
lines cannot be regarded, says Korotneff, as an enteroceel. Later, the
entire embryo becomes two-layered by an extension of the inner layer.
The two polypides arise from two distinct invaginations of the double-
layered wall.

Unfortunately, Korotneff does not demonstrate by figures the method
of origin of the alimentary tracts of the first polypides; but there is
little reason to doubt that it is essentially like that in other buds. If
it is admitted that the inner layer is entoderm, as Korotneff maintains,
then the entoderm takes no part in forming the digestive epithelium ;
but the latter is derived solely from ectoderm.

In his discussion of the theoretical bearing of his results (p. 404), the
author seems to maintain that the polypide is to be regarded neither as
an individual (Nitsche’s view), nor, on the other hand, as an assemblage
of organs homologous with organs of the same name in other groups ;
but rather as a new structure, developed upon the cystid, to aid in its
nutrition. :

In criticism of Korotneff’s view, that the loose cells given off from
one pole of the blastula are entoderm, I may point out that this process
bears quite as much resemblance to the process of ‘mesenchyme ”
formation (as described by Korschelt for the Echinoids), as it does to
the origin of the entoderm in some Ceelenterates. Compare Figs. 13 E
and 182, in Korschelt und Heider’s Lehrbuch der Vergleichenden Ent-
wicklungsgeschichte.

Braem (’89°, pp. 676, 677) has shown that the primary polypide of
the statoblast arises from the cell layers of the statoblast, exactly as the
primary polypide of the egg embryo does from those of the “ cystid,”
and the alimentary tract is formed as in buds of Cristatella.

To sum up: The outer layer of the colony-wall is ectodermal in ori-
gin; the inner layer arises by an embolic (?) invagination of the blas-
tula, and would therefore appear to be entoderm, although the possibility
of its being homologous with the mesoderm in other forms is perhaps
not excluded. The first polypides so arise that their inner layers are

126 BULLETIN OF THE

formed by an invagination of the outer layer of the colony-wall, and
their outer layer from the inner layer of that wall.

In Lndoprocta, Seeliger (’90, pp. 176-187) has shown decisively that
the inner layer of the bud is derived solely from the ectoderm, and that
this inner layer gives rise to the digestive epithelium of the alimentary
tract, to the nervous tissue of the brain, and to the outer layer of the
tentacles. Here mesenchymatous cells, representing undoubtedly meso-
dermal tissue, come secondarily to surround the polypide as a loose
outer tissue. In Loxosoma the same is probably true.

The conditions of budding in G'ymnolemata are more difficult to un-
derstand. In Paludicella the bud seems to arise as in Phylactoleemata
(Allman and Korotneff). The same is probably true for Aleyonidium
(Haddon, ’83, p. 523, Plate XX XVIII. Fig. 23). In the Cheilostomata,
however, the fact of the great development of a loose mesenchyme-like
tissue obscures the process, and makes it difficult of interpretation.
This tissue, which is known under three probably homologous terms, —
‘“‘Funiculargewebe,” Nitsche, “ Parenchymgewebe” in part, Vigelius,
and ‘‘ Endosare,” Joliet, — is to be considered as representing the funicu-
lar and coelomic tissues of Phylactolemata. The most careful observa-
tions on the origin of this tissue are those of Joliet (77, pp. 249, 250,
and *86, pp. 39, 40) and Vigelius (84, p. 76). Both authors assert
that this tissue is derived from cells given off from an epithelium at the
distal end of the budding individual. Vigelius (84, pp. 19, 79) believes
that this epithelium is ectodermal, and that it is the sole rudiment of
this layer; but Ostroumoff (85, p. 291) and Pergens (89, p. 505)
have shown that the ectoderm persists and secretes in its cells the cal-
careous.ectocyst. It seems more probable, however, that the “ funicu-
lar tissue” arises from the inner layer of the body-wall (Nitsche, ’71,
p. 37, Plate III. Fig. 5, c.), and is the equivalent of the coelomic epithe-
lium of Cristatella. The fact that many of these mesenchymatous cells
conglomerate in the formation of the polypide sufficiently accounts for
the origin of its outer layer of cells. The origin of the inner layer is
problematical, if, as is asserted to be the case by several authors, the
bud is not formed in the region of the body-wall.

It will be premature to speculate upon the significance of the facts
of budding in the Ectoprocta until we shall have gained a more com-
plete knowledge of the ontogeny of the group, and of the relationship
of the Cheilostomatous to the Phylactolematous type through compara-
tive agamogenetic studies. It may appear in the end, that, under cer-
tain circumstances, undifferentiated embryonic tissue, derived from a

MUSEUM OF COMPARATIVE ZOOLOGY. 127

certain germ layer, can assume the task of building organs in budded
individuals similar to those derived from a different layer in the sex-
ually produced individual.

Whatever may be the truth of the conclusions reached by Haddon
(’83, pp. 548, 549, 552) and by Joliet (’86, pp. 54-56), that the nervous
system and the alimentary tract arise from two distinct layers, or kinds
of cells, in the species studied by them (and their evidence is certainly
not conclusive even for these), their attempts (Haddon, ’83, p. 540,
Joliet, 86, p. 57) to apply their results to the Phylactolemata are not
justified by the observations which are here presented, nor by those
which have been made upon most Gymnolzmata and Endoprocta.

4. Origin of the Alimentary Tract.— There is a curious difference
between the Endoprocta and the Ectoprocta in the development of the
organs of digestion. Seeliger (89, pp. 182-184) has shown for Pedi-
cellina, that the cesophagus and stomach arise as an evagination of
the oral wall of the young bud, which secondarily becomes connected
with the proctodeum. Haddon (83, pp. 517, 518) has shown for
Flustra, Barrois (’86, pp. 73-86) for Lepralia, Braem (89°, pp. 677, 678)
for the statoblast polypides of Cristatella, and the present paper for the
polypides in the adult Cristatella, that the cesophagus only is formed
on the oral side, the stomach arising with the rectum on the anal side
of the atrium. In all cases the esophagus is formed first (Plate IT.
Fig. 13). A comparison of my Figure 18 with Figure 41, Plate XXX.,
of Hatschek (77), shows a striking resemblance between the two. The
form of the alimentary tract and the depression to form the ganglion
are practically ideutical; and were the tentacles to arise directly from
the immature lophophore arm (br. loph., Fig. 18), and from the circum-
oral fold which has already appeared, it would be difficult to decide
whether the anus opened outside or inside the circlet of tentacles, —
whether, at this stage, the Cristatella polypide were ectoproct or
endoproct.

5. Origin of the Central Nervous System. — The only observations on
the origin of the brain in Bryozoa relate to Phylactolemata and Endo-
procta. In buds of Pedicellina, the ganglion is formed, according to
Hatschek (77, p. 520), as an invagination of the floor of the atrium,
which later becomes cut off as a hollow sac. Harmer (85, pp. 274, 275)
has studied the origin of the ganglion in the bud in Loxosoma. He
states that it is derived from the floor of the vestibular [atrial] cavity,
and (apparently on purely theoretical grounds) that this latter is ecto-
dermic. ‘In a longitudinal section through a fairly advanced bud

128 BULLETIN OF THE

(Fig. 15) it is seen that a narrow slit-like diverticulum of the vestibule
passes behind the epistome. This diverticulum, which remains in very
‘much the same condition throughout life, does not give rise in toto to
the ganglion, which is merely formed by a differentiation of some of its
ectodermic cells.” Harmer further doubts Hatschek’s account of the
formation of the ganglion in Pedicellina, and believes that the lumen of
Hatschek’s hollow sac is in reality the commencement of the fibrous
tissue which occupies the centre of the ganglion in the adult, and
which in optical sections might easily be mistaken for an empty space.
“Similarly,” he continues, “‘ Nitsche has described the ganglion of Alcyo-
nella as originating as a diverticulum from the tentacle sheath. I regard
it as probable that the explanation which I have suggested for Pedicel-
lina will hold also for Aleyonella.” The conditions which every student
of the embryology of Phylactolemata has stated since Metschnikoff’s
paper in 1871, and which my own results reaffirm, do not warrant Har-
mer’s conclusions. The nerve fibres are very evident in the adult ganglion
of Cristatella, and in addition to them there is a cavity, ontogenetically
derived from the atrium, which, as Saefftigen (’88, p. 96) has also
shown for Phylactolemata, contains no histological elements (Plate V.
Fig. 52).

6. Origin of the Funiculus and Muscles. — The origin of the so-called
funicular tissue in Gymnolemata has been described already (page 126).
This same tissue also gives rise, according to Vigelius (84, pp. 34, 35)
and others, to the retractor muscles of the polypide. As I have already
shown (pages 115-117, Figs. 22, 54), in writing of the origin of these
tissues in Cristatella, the coelomic epithelium gives off cells, some of
which take on an amoeboid appearance, and, uniting together, form that
end of the funiculus which is attached to’the colony-wall. Other cells
from the coelomic epithelium pass directly to the adjacent outer layer of
the bud, to form the nascent retractor and rotator muscles. Both of these
organs are, however, formed in part from cells composing the outer
layer of the bud, —itself closely related ontogenetically to the ccelomic
epithelium.

These facts would seem to confirm the conclusion which the similar
relation of the two layers would suggest, namely, that the coelomic
epithelinm of Phylactolemata is the homologue of the “ endosare ” of
Gymnolzmata.

MUSEUM OF COMPARATIVE ZOOLOGY. 129

IV. Organogeny.
1. Development of the Ring Canal. — Nitsche (’75, p. 358) describes

the ring canal as a furrow arising from the opening of each of the lopho-
phoric pockets, and running towards the oral side of the bud. In a later
stage, both layers become deeply implicated in this furrow, and the
ring canal is completed by a growing together of the edges of the
furrow.

Braem (’89°, p. 679) merely states that he cannot fully agree with
Nitsche’s description of the formation of the ring canal.

As a result of my own studies on this subject, I have reached the
conclusion that the circumoral branch of the ring canal makes its first
appearance in the median plane in the oral region at about the time
that the depressions of the lophophoric pockets are first indicated. The
formation of both organs is preceded by a preliminary thickening of the
inner layer of the bud (Plate IV. Fig. 26, b7. doph., and Plate III. Fig.
17, can. crc.). It is only later, after the lophophoric pockets have at-
tained considerable depth, that the groove of the incipient “ring canal ”
appears continuously on the side of the polypide, extending from the
pre-oral region to the lophophoric pockets (Plate IV. Figs. 33, 35, 37,
can. cre.).

As indicated in the successive stages of Figures 18 and 19, Plate IIT.,
the thickening of the inner layer anterior to the mouth is followed by a
fold at this point involving both layers. The fold is deepest in the pre-
oral part of the median plane, and becomes shallower as it proceeds pos-
teriorly. Finally, the outer-layer cells of the lips of the fold approach
each other and fuse, thus forming a true canal (Plate IV. Fig. 33, can.
crc.). Kraepelin (’87, p. 57, Figs. 72, 73, gb.) asserts that this canal
does not communicate at its neural ends with the ccenoccel, but that it
is always closed by a strong “ Querbriicke ” connecting the “ Kampto-
derm” with the alimentary tract. By making sections of the colony
parallel to the sole, dozens of individuals are cut through the entire
length of the circumoral ring canal. Although I have examined many
individuals cut in this way, I have never succeeded in finding in Crista-
tella this closing “ Querbriicke”; but in both young and old specimens,
sections nearly corresponding to Kraepelin’s Figure 72 show a perfectly
uninterrupted semicircular space surrounding the cesophagus, and open-
ing freely into the coenoccel on each side of the brain (Plate IX. Fig. 78,
can. cre.). JI must therefore conclude that in Cristatella the fluids of

the cavities of the circumoral branch of the ring canal, and therefore
VOL Xx —NO. 4. 9

130 BULLETIN OF THE

of the tentacles also, are in free communication with the fluids of the
common body cavity. As Figure 51, Plate V., shows, the posterior ends
of the ring canals open into a pair of cavities which are the bases of the
lophophoric pockets, and by a comparison of Figures 61-63, can. ere., can.
cre.’, Plate VII., it will become apparent that they each become confluent
with a furrow which passes up the lophophore arm, and from which the
outer lophophoric row of tentacles is developed. Further, by a com-
parison of can. erc.”, can. cre”, in Figures 61-63, Plate VII. (dextro-
sinistral vertical sections), and Figure 50, Plate V. (horizontal section,
compare also Fig. 52, a sagittal section), it will be seen that from
the tip of the lophophoric arm a groove (can. erc.’’) passes down upon
the side opposite to the ascending groove (can. erc.’), and, reaching the
base, turns abruptly anteriorly (can. ere.”, Fig. 50), and finally, in
a later stage, becomes confluent with its fellow of the opposite side
in the median plane just behind the epistome and above the brain.
It would be quite unnecessary for me to give figures showing the
course of this supraganglionic canal (cf. Fig. 52, Plate V.). It has
long been recognized, and is shown in Kraepelin’s (87) Figure 66,
Taf. I]. This is probably what Verworn (’87, pp. 114, 115, Figs. 20a,
20 b, Taf. XII.) has described as a “segmental organ.” Braem (’89?,
p. 679) has given to it the name ‘‘Gabelkanal.” The “ Ringkanal” of
Nitsche is, then, to my mind, merely the circumoral portion of a groove
which is elsewhere unclosed to form a proper canal and which lies at
the base of all tentacles. My reason for avoiding another term for the
unenclosed portion of the ‘‘canal” is, that I regard the whole as mor-
phologically equivalent to the ring canal of Gymmolemata, which is
said to be closed throughout.

2. Development of the Lophophore.— The early stages in the forma-
tion of this organ are well known, both from the descriptions of Nitsche
(75, pp. 857, 358) and the earlier ones of Allman and others.

I have already (page 114) shown how the cavities of the lophophoric
pockets become confluent between the rectum and ganglion, and how
their opposed walls, formerly passing over into each other through the
floor of the brain, are now anteriorly continnous by means of the new
floor of the atrium, and posteriorly are fused together.

The union of the inner layers of the two opposed walls of the lopho-
phore arms (Plate V. Fig. 44, loph.’) continues, however, for some dis-
tance above the floor of the atrium, up to within a short distance of the
tips of the young arms (Plate VII. Figs. 61, 62, loph.’). As the arms
grow longer, the relative extent of their free and fused portions remains

MUSEUM OF COMPARATIVE ZOOLOGY. Lat

approximately the same. The free ends of the arms are shown in
Figure 99, Plate XI., just above loph’. The polypide figured here is
only slightly older than that of Figure 77, Plate IX. The connection
between the two arms is not one of contact merely, for in the region
of fusion one can count roughly three layers of nuclei, whereas each of
the two free portions of the same cell layer contains but one layer
of nuclei (Fig. 99).

Before the atrial opening is formed, a separation of the two arms
begins to take place. This process commences at the base of the arms,
und proceeds upward as the tentacles of the inner row successively
reach a certain stage of development. As the work of separation pro-
gresses, the cells of the connecting band lose their capacity for becoming
stained and appear vacuolated. The vacuoles increase in size until the
connection between the arms is reduced to a series of fine threads
(Plate VILL. Fig. 75, loph.'), which are probably sundered when the ten-
tacles of the inner row (can. cre.”, Fig. 76) bend at right angles to their
former position to become parallel to those of the outer row. In at-
tempting to find an explanation of this process, it must first be ascer-
tained how the arms of the lophophore grow in length. One is perhaps
inclined to think of a terminal growth, but this does not take place.
So far as I can judge from an examination of many longitudinal sec-
tions of the arms, cell proliferation goes on throughout the whole length
of the arm, and with nearly equal rapidity in all parts. The distance
between the centres of the terminal tentacles is about the same as in
the case of the more fully developed proximal ones, but they are closer
together in the young arm than in the adult one. This being the case,
there ought to be as many (incipient) tentacles in the young as in the
adult, and I find that to be, so far as I can determine, very nearly or
exactly the case. |

The horseshoe-shaped lophophore being characteristic of the Phy-
lactolemata, a study of its development is important, since it may be
expected to throw light on the phylogeny of the group. We have in
Cristatella, Plumatella, and Fredericella, a series in which the arms of
the lophophore are shorter and shorter, in correspondence with other
changes, by which is effected a gradual transition to the Gymnolemata,
which have a circular lophophore. In Gymnolzmata, the ring canal
lies at the base of all tentacles in the adult. The anus lies outside the
circle of this canal. The brain lies within the lumen of the canal.

Nitsche (71, pp. 43-45) has given the best description extant of the

| ey BULLETIN OF THE

development of the lophophore in Gymnolemata. At a very early
stage, the rudiments of the tentacles, he says, are seen lying in a
- U-shaped line, surrounding the mouth in front, but unclosed behind.
The same is true for Paludicella (Korotneff, ’75, p. 371). The post-
oral tentacles make their appearance at the posterior free ends of the
row of tentacles. They are bent slightly downward, so as to be con-
cealed by the tentacles above. At a later stage, the tentacles lying
next to the anus gradually come to lie nearer to the anal side of the
mouth opening, the nearly parallel lateral rows lose their compressed
appearance, and a circular basin is formed whose walls are constituted
by the corona of tentacles.

In Pedicellina (Hatschek, ’77, pp. 520, 521) the tentacles arise as
five pairs of papilla-like processes in the upper part of the atrium.
Two additional pairs are formed later nearer the anal opening. In the
adult (Nitsche, 769, p. 21) the tentacles are arranged with bilateral
symmetry, and so that the plane of symmetry passes through two inter-
tentacular spaces, which are thus the only unpaired spaces; they are
also much broader than the others.

One might be inclined to ask by what modification of the condition
of the tentacles in Endoprocta we may suppose the condition in Kcto-
procta to have arisen, but the question is not a fair one. I have
already (page 127) shown that the young bud of Cristatella has many
points of similarity to a well advanced Endoproct. This similarity
leads me to the conclusion that the common ancestor of the Endo-
procta and Phylactolemata more nearly resembled the former than the
latter group. But the Endoprocta are not that common ancestor ;
rather they are themselves more or less modified descendants of it.
The proper inquiry is, To what ancestral relation between tentacles and
anal opening does a comparison of the ontogeny of Endoprocta and
Ectoprocta point, and by what modifications of that ancestral type may
the two divergent types of the present be derived? Eliminating for a
moment the evidently coenogenetic character of the lophcphore arm, an
early stage of either Endoprocta or Ectoprocta reveals a U-shaped band
from which tentacles are to arise. This band completely encircles the
mouth, and passes posteriorly as far as the anus. This is the condition
of the Endoproct bud, with only five of its seven pairs of tentacles
formed ; it is also the condition of the Cristatella bud of Stage XIII.
(compare Figs. 19, 44). Starting from this common condition, that of
the adult Endoproct, on the one hand, was attained by the addition of
two pairs of tentacles posteriorly, thus nearly completing the circlet

MUSEUM OF COMPARATIVE ZOOLOGY. 133

bebind the anus. The condition of the adult Ectoproct, on the other
hand, was reached by the curving oralwards, and the meeting of the
free ends of the rows of tentacles between the mouth and anus, thus
shutting the anus outside of their circle. In evidence of this latter
assertion, I submit the following comparative statement.

As Nitsche has shown for Gymnolemata, the tentacles on the ring
canal are first arranged in two rows, placed bilaterally, and meeting
in front, but not behind. Later the hindermost of the tentacles move
forward and toward the median plane, thus completing the circlet of
tentacles at a point behind the mouth, but in front of the anus. I be-
lieve the circumoral ring canal plus the early invaginations of the lopho-
phoric arms in Phylactolemata to be homologous with the ring canal
of Gymnolemata in its early stage ; like the latter, it is closed in front,
but has two free ends behind. The difference lies in the greater devel-
opment of the posterior ends of the canal, which latter have become
thrown into a vertical fold to afford space for more tentacles. At this
stage of development it would be difficult to say whether the anus
opened within or without the corona of tentacles. As in Gymnolemata
the circle is completed by a movement inward of the posterior tentacles,
so in Phylactolemata the corona of tentacles is completed in front of the
anus by the two anterior processes, can. cre.’”, Figure 50 (cf. Fig. 44),
of the lophophore arm, which come to unite just behind the epistome,
Figures 52, 81, can. cre.” The lumen of this process of the lophophore
arm thus forms that portion of the ring canal which, as I shall show
directly, is the morphological equivalent of the most posterior portion
of the ring canal in Gymnolemata. The tentacles which arise froin
this portion of the ring canal are ontogenetically, and therefore phylo-
genetically, the youngest. As in Gymnolemata, so here the moving for-
ward of the most posterior tentacles obliterates the basin-like floor of
the atrium, such as we see in Endoprocta, and leaves the anal opening
far outside the circlet of tentacles. |

The answer to the question, How may the horseshoe-shaped tentac-
ular corona of Phylactolemata be homologized with circular ones? is
involved in the answer to the preceding query. Nitsche (’75, p. 357)
believed the lophophoric arms to be “primary tentacles,” and the
tentacles borne on them to be secondary tentacles, ‘‘ Gar nicht ohne
Weiteres mit den Tentakeln der Infundibulata von Gervais zu verglei-
chen.” The only evidence which he offers in support of his theory is
the fact that the tentacles on the lophophore arm arise later than the
arm itself.

134 BULLETIN OF THE

The tentacles of Phylactolemata may be distributed into two groups.
The first includes those which arise from the circumoral branch of the
ring canal. The ring canal, from which they spring, begins to be formed
at nearly the same time as the lophophoric arms. These tentacles are
undoubtedly homologous with those of the same region in Endoprocta
and Gymnolemata. The second group of tentacles includes those which
are borne upon the lophophore arms and upon the supraganglionic ring
canal. Are these comparable with the posterior tentacles of Gymnole-
mata? I believe they are, and for the following reasons. Nitsche’s
reason for supposing that they are not is unsatisfactory, since, if we re-
gard the lophophore arms as mere upward folds of the wall of the ring
canal, we should expect to have the tentacles arise later than the arms.
The fact that the tentacles of the lophophore arm arise much later than
those of the circumoral region is what we should expect, since the pos-
terior tentacles arise later than the circumoral ones in both Endoprocta
and Gymnolemata,—a criticism which Hatschek (’77, p. 541) has
already applied. In direct support of my belief are the facts, (1) that
the ring canal is continuous along two sides of the lophophore arms,

which would be the case if they were mere upward folds of the wall of
the ring canal; (2) the structure of the tentacles is the same as that
of the oral ones, and the relation of their intertentacular sept to the
ring canal of the arms is the same as that of the septee of the oral ten-
tacles to their ring canal, as Kraepelin (’87, pp. 55, 56) has shown. If
both circumoral and lophophoric tentacles find their homologues in
Gymnolemata, we have only to conceive of an elongation of the postero-
lateral angles of the lophophore of Gymnolemata, after the forward
movement of the posterior tentacles, to effect the condition which is
found in Phylactolemata.

The significance of the fusion of the lophophore arms is difficult to
determine. I had thought it might be possible to find a phylogenetic
explanation for it, by regarding the unfused tips of the arms in Crista-
tella as homologous with the short arms of Fredericella. In studying
Plumatella, however, where the length of the lophophore arnis is inter-
mediate between that of Cristatella and Fredericella, I have been able
to find no trace of this fusion. It does exist, however, in Pectinatella.
I have had no material of Lophopus, upon which it is important to study
this point. The evidence so far seems to indicate that this fusion of the
arms during the period of their development is a secondarily acquired
adaptation to some condition concerning the nature of which I am
ignorant. |

MUSEUM OF COMPARATIVE ZOOLOGY. 135

3. Development of the Tentacles. — Nitsche (75, p. 359) observed that
both layers of the bud went to form the tentacle in Phylactolemata,
and that the inner layer was derived from the outer layer of the polyp-
ide; the outer, on the contrary, form the inner cell layer. He states,
moreover, as already mentioned, that the oral tentacles arise first, then
those of the outer row of the lophophore arms, of which the basal are
fully formed before the terminal ones. The tentacles of the inner row,
he says, are formed last, and in Alcyonella are yet lacking when the
polypide is first evaginated.

My own observations confirm in general those of Nitsche. The long-
est tentacles in a polypide of about the age of that shown in Figure 77,
Plate IX., are those arising from the region of transition from the circum-
oral ring canal (can. cre.) to the outer lophophoric ring canal (can. cre.’).
The tentacles lying near the median plane, and in front of the mouth,
are somewhat shorter than these (75 4:52). The tentacles situated
near the proximal extremity of the inner lophophoric ring canal (can.
erc.”) are still shorter (50 »). Those situated at the tips of the lopho- ,
phore arms are at this stage about 30 w in length. The tentacles behind
the mouth, arising from the supraganglionic part of the ring canal (can.
erc.'"), are shortest of all at this stage (15 p).

The two layers which, as we have seen, go to form the upper wall of
the ring canal in all its parts, are the ones which give rise to the ten-
tacles. In Figure 74, ta.’, Plate VIII. (compare Fig. 51, éa.’), young oral
tentacles are cut transversely at different heights. The circumoral part
of the ring canal is seen at a point (can. crc.) near which it opens into
the cavity of the lophophore arm. The plane of the section passes ob-
liquely upward and anteriorly from this point. The most posterior ten-
-tacle in the lower part of the figure is cut at the base. The calibre of
the canal (including its walls) is evidently much enlarged at this point.
The enlargements of the canal at the base of the tentacles are seen also
in Figure 78, can. cre., Plate IX. The more anterior tentacles in Figure 74
show the two layers well marked, but as yet enclosing no lumen. Since
the tentacles arise from the ring canal at intervals only, the ring canal
is a tube (or groove) whose lumen is alternately constricted and ex-
panded laterally as well as vertically. The lumen is, indeed, often so
small between the tentacles that the ring canal appears divided into
separate chambers by a series of transverse septa, which, however, are
always penetrated by an opening (Fig. 78, can. crc.). Figures 73 and 77,
ta.’, show, in longitudinal section, successive stages in the development
of the oral tentacles. The formation of tentacles begins by a rapid cell

136 BULLETIN OF THE

proliferation at intervals in the upper wall of the ring canal; thus a
_ projection is formed at each of these points, which constantly elongates
to form the tentacle. Figures 70 and 69 (Plate VII.) are longitudinal
sections of two later stages in the development of tentacles. The inner
layer, ex. (Fig. 70), becomes gradually thinner as the tentacle grows
older, and its cells finally become thread-like (Fig. 69, ex.).

Figure 81 (Plate [X.) shows the arrangement of the tentacles about the
mouth and over the ganglion in a young polypide. The supraganglionic
part of the ring canal is cut tangentially just behind the epistome (can.
erc."), J have often noticed that, in polypides of about the age of that
of Figure 77, or older, certain of the nuclei seen in a cross section of a
tentacle stain more deeply than the others. These nuclei are usually
two or three in number on each of the lateral surfaces of the tentacles.
They are evident in Figure 81. I do not know what this difference in
staining properties signifies. Vigelius (’84, p. 38, Fig. 23) describes
and figures a condition of the nuclei in Flustra, as seen on cross-section,
which is similar to that just described. The deeply staining nuclei in
Flustra lie on the inner face of the tentacle, are larger than the others,
and belong to cells which possess no cilia.

Nitsche (71, p. 43) described the development of the tentacles in
Flustra as though they were derived exclusively from the inner layers of
the bud; but Repiachoff (’75%, pp. 138, 139, ’75”, p. 152) showed that in
Cheilostomes both cell layers of the bud took part in their formation,
and he figures an early stage which is quite similar to my Figure 70.

4. Development of the Lophophoric Nerves. — It has long been known
that a large nerve passes along the middle of the upper wall of each
lophophore arm, connecting proximally with the corresponding side of
the ganglion. No observations have been made, so far as I know, upon
the origin of this organ. Evidently there are, @ priorz, two possibilities.
Either (1) the lophophoric nerve is formed by a direct outgrowth of
the ganglion, or (2) it arises in place from the inner layer of the bud,
which, since it here forms the outer layer of the lophophoric pocket, is
the same as that from which the ganglion itself is constructed. By a
careful study of this nerve in many stages of development, and from sec-
tions in different directions, I have come to the conclusion that it arises
as an outgrowth of the walls of the ganglion, and that it penetrates
between the outer and inner layers of the arm.

The facts which have led me to this conclusion are these. First, dur-
ing the formation of the brain, soon after its lumen is cut off from its
connection with the atrium, its cells begin to divide rapidly (Plate V.

MUSEUM OF COMPARATIVE ZOOLOGY. 137

Fig. 51, Plate VII. Figs. 63, 68); but that the new cells so formed do not
all remain in the brain is indicated by the fact that the brain does not in-
crease very rapidly in size. (Compare Plate IIL Fig. 19, and Plate IX.
Fig. 77.) This rapid cell division would be inexplicable upon the as-
sumption of an origin 2 situ. Secondly, at an early stage the lopho-
phoric nerve is already seen extending from the brain to the adjacent
inner layer, with which it remains in contact. A longitudinal section
through the middle of this nerve shows a prolongation of the lumen of
the brain extending into it, so that its upper wall passes directly into the
upper wall of the brain, and its lower wall into the corresponding part
of the central organ (Plate VII. Fig. 68, lu. gn., n. loph.). The proxi-
mal part of the lophophoric nerve is thus to be regarded as a pocket of
the brain. The existing condition is not what we should expect if a
cord of cells derived from the outer layer of the lophophoric arm had
secondarily fused with the brain. Thirdly, I have never found any good
evidence that cells were being given off from the outer layer of the arm
at its tip to form the nerve, where we should look for such a process, if
anywhere ; on the contrary, the nerve is quite sharply marked off from
the outer layer at this point, as will be seen by reference to Figures 64
67 (Plate VII.). These figures represent successive transverse sections
from a young lophophore arm of about the stage of development of that
shown in Figure 71. Figures 65-67 were drawn from one arm in about
the position indicated by the lines 65-67 in Figure 71. Figure 64 was
drawn from the opposite arm of the same individual, and in about the
region of Figure 65. In Figures 64 and 65 there is a small space be-
tween the nerve (7. Joph.) and the overlying cells of the inner layer (2.).
This may be due to shrinkage, but in any event it indicates a complete
independence between the two cell masses which it separates. Over the
nerve the cells of the layer ¢ are shorter than elsewhere. This might
be considered as an indication that the cells had recently divided in
order to give up cells to the nerve, which, on this assumption, would be
formed a situ. Three appearances, however, indicate that the cells of
the layer « have been rather subjected to crowding at this point, as
though by a mass of cells forcing their way between them and the layer
ex., and gradually increasing in volume. (a.) The surface of the layer ¢.
is raised above the general level directly above the nerve. (b.) The cells
of the layer 2. are somewhat broader over the nerve than elsewhere, and
the nuclei are shorter, but thicker. These are the conditions which we
should expect in an epithelium subjected to pressure by the intrusion
of a mass of cells at its base, for in volume the crowded cells compare

138 BULLETIN OF THE

fairly with their neighbors, whereas, if they had by division given rise
to nerve cells, they should all be smaller. (c.) In Figure 67, which is a
section immediately in front of the advancing tip of the nerve, the po-
sition corresponding to that opposite the nerve in the preceding sections
is indicated by an asterisk (*). The nuclei are here crowded together,
indicating pressure. Fourthly, there is a considerable difference in size
between the nuclei of the cells of the layer ¢. of the lophopbore and the
nerve cells. This is not what one would expect upon the assumption of
the formation of the nerve directly from the overlying cells. Fifthly, a
longitudinal section through the young lophophoric nerve (Plate VII.
Fig. 71) shows a more active cell division in it than in the walls of the
arm (compare Fig. 64, 2. loph.), and a crowding together of nuclei of
the outer layer of the arm, 2, at its distal end, rather than a passage of
nuclei into the nerve.

The conclusion to which I have arrived from considering these facts
is that the peripheral nervous system in Phylactolemata arises from the
brain as an outgrowth of ws walls.

5. Development of the Epistome.— The epistome was regarded by
Lankester at one time (74, p. 80) as homologous with the foot of
Mollusca, and on another occasion (’85, p. 434) as representing the
preoral lobe of Annelids, —a view for which Caldwell (83) first pro-
duced evidence from comparative embryology. In view of such diver-
gent opinions, and of the occurrence of an organ which is possibly its
homologue, in quite aberrant genera, such as Phoronis, Rhabdopleura,
etc., a careful consideration of its origin and development is desirable.

After the ganglion is fully formed, its oral face remains in contact
in front with the posterior wall of the cesophagus (Plate V. Fig. 52,
Plate IX. Fig. 77), and on each side with the outer wall of the lopho-
phoric pockets by means of the lophophoric nerves (Plate VII. Fig. 63,
n. loph.). The outer layer of the bud penetrates between the gan-
glion and rectum, but not between the ganglion and the cesophagus
(Fig. 51,*). This layer also comes to lie between the floor of the
atrium above, the ganglion below, and the lophophoric nerves on either
side, having made its way in from behind as a double cell-layer enclosing
a flat cavity (Plate V. Fig. 52, Plate VI. Fig. 56, Plate VIII. Fig. 74,
can. e stm.). My description of the process by which the inner layer
comes to envelop the ganglion above and behind differs considerably
from Nitsche’s, already quoted (page 114). As the ganglion becomes
farther removed from the floor of the atrium, the cavity above it (can. e
stm.) enlarges, and the two lateral walls of this canal, each composed of

MUSEUM OF COMPARATIVE ZOOLOGY. 139

two layers of cells, both belonging to the outer layer of the bud, form the
‘‘Verbindungsstrang des Ganglions mit dem Lophoderm” of Kraepe-
lin (87, p. 63, Taf. IT. Fig. 59, vs.). (See Plate V. Fig. 51, Plate VI.
Fig. 56, and Plate IX. Fig. 80,*.) This canal is the only one by which
communication between the body cavity and the cavity of the epistome
can occur. It may be called the epistomic canal (Plate V. Fig. 52,
Plate VIII. Fig. 72, can. e stm.).

The epistome proper arises at the point where the epistomic canal
ends blindly, above and in front of the brain (Plate VIII. Fig. 73, Plate
IX. Fig. 77, e stm.) ; it is a pocket, the outer wall of which is contin-
uous on its under surface with the cesophageal epithelium, and on its
upper surface with the floor of the atrium. The growth of this organ
is disproportionately great after the first evagination of the polypide.
That part of its wall which is turned towards the alimentary tract is
then much thicker than the remaining part ; it forms the posterior wall
of the pharynx (Plate VIII. Fig. 72, e stm. ; compare Plate IX. Fig. 81).
Is the epistome innervated by fibres from the brain, as maintained by
Hyatt (68, pp. 41-43)? I have not succeeded in finding such fibres,
and the conditions of the formation of the epistome, cut off as it is from
the brain at every point, make such a connection improbable.

Allman (756, Fig. 8, Plate XI.) and Korotneff (75, p. 371) have
shown for Paludicella, and Nitsche (’71, p. 44) has shown for Flustra,
that an epistome-like fold occurs at an early stage of development, but
is absent in the adult. Such an organ has been described by Allman
(756, p. 56) and other observers in Pedicellina, and it is still more
prominent in Loxosoma, in which the relation of the epistome to the
body cavity is similar to that in the Phylactolemata.

The constant occurrence of this organ in the development of Bryozoa,
and its presence in so many aberrant genera which seem to be some-
what allied to this group, can only be interpreted, it seems to me, as
signifying that it is an ancient and morphologically important organ.
The manner of its development in Cristatella seems to throw very
little light, however, upon its significance; it arises rather late, and
does not become of any considerable size until the atrial opening is
made. ;

6. Development of the Alimentary Tract. — The later development and
histological differentiation of the alimentary tract have not been hereto-
fore carefully studied.

At the stage at which we left the alimentary tract (Plate III. Fig. 19)
only two parts were clearly differentiated, the cesophagus and the intes-

140 BULLETIN OF THE

tine. Inthe next stage shown (Plate VIII. Fig. 73), further changes are
seen to have taken place. The most prominent is the down-folding of
the lower wall of the intestine at its middle region to form the ccecum.
Even at this early stage histological differentiation of the cells of this
region has occurred to such an extent that the lumen of the coecum is
nearly obliterated by the great elongation of some of the cells lining it.
This condition of affairs will be understood by studying the cross sec-
tion of the cecum at a later stage, as shown in Figure 94, Plate X.
The cavity of the rectum has also enlarged, and its cells have taken
on the regular columnar appearance which exists in the adult.

At a still later stage (Plate IX. Fig. 77), the position of the cardiac
and pyloric valves, separating respectively the cesophagus (@.) from the
stomach (ga.), and the cecum (ce.) from the rectum (77.), is clearly in-
dicated. The blind sac is still further elongated and well differentiated
from both stomach and rectum. In order to attain the adult condition
(Plate VIII. Fig. 72), the oral portion of the alimentary tract has merely
to become divided, by a difference in the character of its cells, into
pharynx (phx.) and cesophagus (@.), the stomach (ga.) to increase in
diameter, and the blind sac (cw.) to elongate. The anus (an.) finally
comes to lie at the apex of a small cone, or sphincter valve.

The histological changes which the cells of the different parts of the
alimentary tract undergo are considerable, and will be treated of in
order, beginning with the

Gsophagus. — At a stage a little later than Figure 77, the esophagus,
as is shown in Figure 84, Plate X., has a small diameter relative to that
of the rest of the alimentary tract (cf. Plate VIII. Fig. 72, @.), and its
inner lining is composed of high columnar epithelium, like that of the
oral groove. The shape of the cells is not greatly different in the adult ;
but they become vacuolated, and since these vacuoles lie near the base
of the cells, and either nearer to or farther from the lumen than the
nuclei, the latter acquire that irregular arrangement referred to by
Verworn (87, pp. 111 and 112).

Stomach. — Figure 93 (Plate X.) represents a section across the stomach
immediately below the cardiac valve, from the same individual as that
from which Figure 84 was taken. The proximal ends of all cells stain,
more deeply than the distal ends, but the cells are all alike as far as re-
gards receptivity to stains. Already, in certain regions, the cells are
higher or lower than the average, and have even begun to group them-
selves as typical ridge- and furrow-cells. Figure 82 is a section through
the same region as Figure 93, but from an adult individual. The ridge-

= —— =

MUSEUM OF COMPARATIVE ZOOLOGY. 141

cells are distinguishable from those of the furrows by their greater
height, their weaker attraction for dyes, and their vacuolated and gran-
ular appearance. Moreover, the cell boundaries of this epithelium are
gradually lost. Kraepelin (’87, p. 51) has argued that the elongated
cells are the true digestive cells, and that the deeply dyed cells of the
furrows are, functionally, liver cells.

Cecum. — Figure 94 is from a cross section of the coecum at the
stage of Figures 84 and 93. The cells are more differentiated here than
at any other part of the alimentary tract. They stain uniformly, however,
except for a narrow light zone next to the lumen, and all reach to the
muscularis. The digestive cells are swollen at their free ends; the liver
cells, on the contrary, are thickest at the base. Figure 83 is from a sec-
tion of the proximal part of the ccecum of an adult. The changes which
the cells have undergone are of a similar character to those experienced
by the gastric epithelium, only there has been an exaggeration in this
region of the features shown by the stomach. Figure 85 represents a Sgc-
tion near the blind end of the coecum of an adult. The diameter of the
tube is smaller here than in the section last described, but the inner
epithelium is thrown into still higher ridges and more profound furrows.
Nearly all of the cells, however, seem to extend to the muscularis. The
“liver” cells do not extend so far towards the blind end of the coecum as
this region. The cytoplasm is not at all stained. Evidently, here the
process of digestion reaches a maximum. The circular muscles of the
muscularis are striped, and are developed here to an extraordinary de-
gree, and the coelomic epithelium is greatly thickened, another evidence,

it seems to me, of the intimate relation of this layer to the muscularis.

The number of ridges is not constant in different parts of the alimentary
tract of the same individual, and varies somewhat for the same region
in different individuals. In sections corresponding in position to Figure
83, I have, however, usually found six ridges.

7. Development of the Funiculus and Muscles, — It has already (page
117) been pointed out that the fixed ends of both the funiculus and
muscles originate at a great distance from their position in the adult.
Thus the funiculus originates upon the oral face of a young bud. As
this bud grows older, the fixed end of its funiculus becomes gradually
farther and farther removed from its neck towards the margin, until
finally the funiculus is inserted upon the colony-wall at the margin, or
even upon the sole. So the retractor and rotator muscles arise together
on each side of the polypide and in the angle formed by the colony-wall
and the radial partitions. Later (Plate V. Figs. 44, 45, mz. ret. + rot.)

142 BULLETIN OF THE

they are found on the partitions immediately below the colony-wall.
Still later (Plate VI. Fig. 59, mw. rot., mu. ret.) we see them on the lower
portion of the partition, and finally (Fig. 56, mw. rot., mu. ret.) they are
found attached to the sole, at some distance, it may be, from the
radial partition.

The question arises at once, How do these changes of position take
place? Examination shows that the union between the cclomic epi-
thelium and the cells of that portion of the funiculus which is attached
to the roof is very slight after the funiculus has passed to some distance
from the mother polypide. Although occasionally I have seen the cells
of the fixed end closely applied to the coelomic epithelium, the only
connection between the two is usually effected by means of amoeboid
cells (Plate V. Figs. 46-48, cl. mz.). On cross sections of the fixed end
of the funiculus these cells (Fig. 49, cl. mz.) are seen to surround it as
a loose layer, and in longitudinal sections some of the ameeboid cells
arg seen to be connected with the coelomic epithelium. It is difficult to
determine the origin of these cells, but they have the position and
character of the cells of which the funiculus was exclusively composed
before the entrance into it of the ectodermal plug described by Braem.
The only explanation of the migration of the funiculus which occurs to
me has been suggested by the facts given above; namely, that the
“ migratory cells,’ by which the funiculus is attached to the coelomic
epithelium, change their position, carrying with them the funiculus.
Remembering that the ccenoccel is filled with a fluid in which the
funiculus floats, and that by the growth of the funiculus it is elongated
in proportion as the distance from its origin to the coecum increases, this
hypothesis does not seem improbable, although its truth can hardly be
tested by the study of preserved material. When the funiculus has
reached its permanent position its attachment to the coelomic epithe-
lium is more intimate. Meanwhile the end attached to the polypide
has become more and more attenuated (Plate IX. Fig. 77, fun.), until,
in the adult, I have usually been unable to discover any attachment.
In any case, it must certainly be broken when the polypide begins
to degenerate.

The migration downward of the ends of the muscles which are
attached to the partition is even more difficult of explanation. During
this migration their point of origin seems to be in the muscularis of the
partition itself. The fixed point of the muscle in the adult is probably
in the muscularis of the sole, since I have traced muscle fibres through
the coelomic epithelium, and to the muscularis (Plate VI. Fig. 58, mu.

MUSEUM OF COMPARATIVE ZOOLOGY. 143

ret.). The insertion is in the muscularis of the polypide (Fig. 56), but
I have not been able to determine the precise relation between the
muscle fibres of the great coelomic muscles and those of the muscu-
laris. A comparison of Figures 44, 59, and 56 shows quite plainly that
both the retractor and the rotator muscles originate from a common
mass of muscle cells, and become distinct from one another by a
separation of their points of attachment to the polypide. The re-
tractor muscles (mz. ret.) are attached to the cesophagus immediately
below the ganglion (Plate IX. Fig. 78); the rotator muscles (mw. rot.),
on the contrary, to the lateral walls of the opening leading from the
ccenoceel (cen.) to the cavity of the lophophore arms. These two re-
gions are near to each other in the young polypide, but become con-
stantly more widely separated with the growth of the lophophore.
Compare Figure 78 with Figures 74 (Plate VIII.) and 51 (Plate V.),
which are younger stages, cut somewhat above the level of Figure 78,
and more than twice as highly magnified.

I have been able to obtain in thick sections various stages in the
development of the muscle fibres, some of which are shown in Figures 89
to 92 (Plate X.). In the earlier stages, all parts of the muscle cell stain
uniformly in cochineal. Later, the cell body becomes differentiated into
two portions, easily distinguishable by their different receptivity to the
dye. The more retractile portion becomes greatly elongated, highly
refractive, and incapable of being stained. A mass of indifferent pro-
toplasm, including the nucleus, still remains stainable (Fig. 90). The
undifferentiated portion continues to diminish relatively to the whole
mass of the cell, which has greatly increased in size, until little remains
but the nucleus, placed on one side of the muscle fibre (Figs. 91, 92).
Figure 92 is one of the retractor muscle fibres, in a partly contracted
state. ‘The end placed uppermost in the figure was that which abutted
upon the muscularis of the cesophagus. Its more intimate relation to
the muscularis could not be traced.

8. Origin and Development of the Parieto-vaginal Muscles. — These
consist of two sets, the lower, or posterior, and the upper, or anterior.
The posterior arise earlier. At about the time when the neck of the
polypide begins to disintegrate in order that the polypide may become
extrusible, a disturbance is seen in the cells of the outer layer of the
kamptoderm immediately below the neck of the polypide, and in the
coelomic epithelium opposite to them (Plate XI. Fig. 97, mw. inf.). As
a result, several cells of each layer become organically connected with
those of the opposite layer, and give rise to muscle cells. A later stage of

144 BULLETIN OF THE

such a process is seen at Figure 98. By the time the atrial opening is
established these cells have become plainly muscular (Plate IX. Fig. 79).
Farther up in the angle of attachment of the kamptoderm to the roof
of the colony, the coelomic epithelium and the outer layer of the bud are
both seen to be somewhat disturbed (Fig. 97, mu. su.). At different
points, a single one of these cells reaches across, and later becomes
differentiated into a genuine muscle cell (lig. 99, mu. su.). Of these
there may be three rows.

9. Disintegration of the Neck of the Polypide.—The neck of the polyp-
ide, having fulilled its function as the most important part of the stolon,
must now give way to allow of the extrusion of the nearly developed pol-
ypide. The first indication of this process is the formation within the
cells of the neck of a secreted substance (cp. sec.'), apparently like the se-
creted bodies of the ectoderm. This metamorphosis first involves the outer
and middle cells of the neck only (Plate XI. Fig. 97, cev. pyd.). Later
(Plate IX. Fig. 77, of. atv.) a depression occurs in the ectoderm. This is
due, I believe, to a cessation of cell proliferation at the centre, although
it remains active at the edges of the neck. The depression gradually
deepens until the atrium is closed by a thin layer of cells only (Fig. 98).
The cells of the side of the neck do not disintegrate, but go to form the
“ Randwulst”’ of Kraepelin (’87, p.40). The cells of this region remain
unmetamorphosed. Only a thin layer of cells now stands between the
polypide aud the outside world. This ruptures, as is shown in Figure
99, and by the relaxation of the muscularis, which is thickened about
the atrial opening into a sphincter (Fig. 98, spht.), the polypide is ready
to expand itself.

10. Development of the Body-wall.— As already stated (page 117),
Braem believes that the whole body-wall in Alcyonella is derived from
the neck of the young polypide, after it has begun to give rise to
daughter polypides; and I have given my reasons for believing that in
Cristatella a portion of it at least is derived from the margin.

In addition to this, cells are undoubtedly added to the body-wall, as
Braem states, after the time of origin of the buds. Particularly after the
formation of the median bud, the neck appears to continue to furnish cells
to the ectoderm. Figure 73,* Plate VIII., shows such a mass of cells.
Later stages show that these cells secrete a gelatinous substance within
their protoplasm (cp. sec.', Plate XI. Figs. 97, 98) ; they gradually in-
crease in width and height from the neck outward (Figs. 97-99), and at
the same time become more and more completely metamorphosed. ‘The
result of the addition of these cells from the neck of the polypide is to

MUSEUM OF COMPARATIVE ZOOLOGY. 145

carry the body-wall at the region of the atrial opening to a considerable
height above the level of that portion of the roof lying between polyp-
ides. (Compare Fig. 73, Piate VILI.; Figs. 98 and 99, Plate XI.) This
method of origin of the body-wall is of much less importance in Crista-
tella than in Alcyonella, since the extent of the proper body-wall about
the atrial opening is much less in the former than in the latter case.

The development of the gelatinous bodies deserves further attention.
Kraepelin (’87, p. 24) concluded, from a study of the condition in a
statoblast embryo, that they are formed by a metamorphosis of the cell
protoplasm, beginning at the outer end of the cylindrical cell, and
finally involving, in some cases, the entire cell, together with its nucleus.
Some appearances which I have noticed in the ectoderm of Cristatella
lead me to conclude that the origin is not always so simple as Kraepelin
describes. Figure 79, Plate IX., shows at cp. ser. a number of small gelati-
nous masses occurring at various regions in the protoplasm. Such an
appearance is quite common, and must be interpreted, it seems to me,
as the formation of the gelatinous balls by an intra-cellular metamor-
phosis of the cytoplasm. The balls, flowing together, produce the larger
masses. ‘The metamorphosed matter from several cells may also fuse
into one mass (Plate VI. Fig. 55, cp. sec.). The final result. of this pro-
cess of cell metamorphosis in the ectoderm is a frame-work of old cell
walls, having a thin layer of protoplasm and nuclei at its base, and in-
closing the great gelatinous balls. Such a condition exists near the
centre of the colony between adult polypides, and is shown in Figure 100,
Plate XI.

Summary.

1. Most individuals give rise to two buds, of which one forms a new
branch, the other continues the ancestral branch.

2. The median buds migrate away from the parent polypide to a con-
siderable distance before giving rise to new buds.

3. The descendants of equal age from common ancestors are arranged
similarly in the same region of the colony.

4. New branches are formed upon either side of ancestral branches.

5. The greater the difference in age between the youngest and the
next older bud, the greater the distance between the points at which
they begin to develop.

6. In typical “double buds,” both polypides arise from a common
mass of cells at the same time. From the neck of old polypides a stolon-

VOL. XX. — NO. 4. 10

146 BULLETIN OF THE

like process of cells is given off to form median buds. Between these
two extreme types, intermediate conditions occur.

7. The alimentary tract is formed by two out-pocketings of the lumen
of the bud in the median plane, one forming the csophagus, the other
the rectum and stomach. The blind ends of these two pockets fuse, and
thus form a continuous lumen.

8. The central nervous system arises as a shallow pit in the floor of
the atrium ; the pit becomes closed over by a fold of the inner layer
only of the polypide, which thus forms a sac, the walls of which become
the ganglion.

9. The kamptoderm arises by the transformation of the columnar
epithelium of the two layers of the wall of the atrium into pavement
epithelium.

10. The funiculus arises from ameeboid cells derived from the coelomic
epithelium.

11. The retractor and rotator muscles arise together from the ccelomic
epithelium of both body-wall and bud, and in the angles formed by the
radial partitions and the body-wall.

12. The wall of the colony grows by cell proliferation at its margin.

13. The radial partitions arise as follows: certain muscles of the
muscularis at the margin of the colony leave the latter, and are carried
into the ccenoceel, taking with them a covering of ccelomic epithelium.

14. Budding in Cristatella presents conditions transitional between
direct and stoloniferous budding.

15. Throughout the group of Bryozoa, the youngest and next older
buds are intimately related, and the place of the origin of the younger
buds relatively to the older is determined by a definite law.

16. Cristatella differs from Alcyonella in possessing a region of the
colony-wall,—the tip of the branch, — which grows independently of
the polypides.

17. Each of the layers of the younger bud arises from a part of the same
cell mass as that which gave rise to the corresponding layer of the next
older bud.

18. The digestive epithelium and the nervous tissue are both derived
from one and the same layer of cells, the inner layer of the bud.

19. The alimentary tract of Cristatella at an early stage is similar to
that of a young Endoproct.

MUSEUM OF COMPARATIVE ZOOLOGY. 147

20. Harmer’s conclusion, that the ganglion of Phylactolamata arises
exactly as in Endoprocta, is not confirmed.

21. The “ring canal” lies at the base of all tentacles.

22. The circumoral region of the ring canal in Cristatella is in free
communication with the coenoccel in all stages of development ; and not
closed, as maintained by Kraepelin.

23. The two arms of the lophophore arise independently of each
other. Their adjacent surfaces undergo a secondary fusion, which per-
sists until the inner row of tentacles is about to be formed on the lopho-
phore. The two arms then become entirely separate.

24. The ancestor of Bryozoa probably possessed a U-shaped row of
tentacles, encircling the mouth in front, and ending freely behind near
the anus.

25. The tentacles near the mouth are phylogenetically the oldest.

26. Both layers of the bud are involved in the formation of the
tentacles.

27. The lophophoric nerves arise as outgrowths of the central ganglion,
which make their way into the lophophore arms. :

28. The epistome arises as a fold continuous with the wall of the
cesophagus below and the floor of the atrium above, and it communicates
with the ccenoceel by means of the epistomic canal.

29. The coecum of the alimentary tract, which occurs only in Ecto-
procta, is produced relatively late in the ontogeny by an ont-pocketing
of the lower wall of the alimentary tract at the free end of the polypide.

30. The funiculus migrates (probably with the aid of ameeboid cells)
from the roof of the colony to the margin, or even to the sole.

31. The “ origins” of the retractor and rotator muscles migrate along
the radial partitions from roof to sole. The separation of the two mus-
cles takes place secondarily as their points of insertion separate.

32. The parieto-vaginal muscles arise from the coelomic epithelium of
the body-wall and polypide. )

33. The disintegration of the neck of the polypide is begun by a meta-
morphosis of the protoplasm of its cells. The metamorphosed cells
break away, leaving the atrial opening.

34. The part of the body-wall lying around the atrial opening arises
by proliferation of cells derived from the neck of the polypide.

35. The ectodermal cells become metamorphosed by an intercellular
secretion of small ‘Gallertballen,” which fuse to form the larger ones.
Often the contents of more than one cell fuse into a single large mass.

CaMBRIDGE, June, 1890.

148 BULLETIN OF THE

BIBLIOGRAPHY.

Allman, G. J.
56. A Monograph of the Fresh-water Polyzoa, including all the known Spe-

cies, both British and Foreign. London: Ray Society, 1856, pp. i—viii.,
1-119, 11 Pls.
Barrois, J.
’86. Mémoire sur la Métamorphose de quelques Bryozoaires. Ann. des
Sciences Naturelles. Zoologie. 7° série, Tom. I. No. 1, pp. 1-94,
Pis., L=-LV.» 1886:
Braem, F.
’'88. Untersuchungen iber die Bryozoen des siissen Wassers. Zool. Anzei-
ger, XI. Jahrg., No. 288, pp. 503-509; No. 289, pp. 533-539. 17 Sept.,
1 Oct., 1888.
89%. Ueber die Statoblastenbildung bei Plumatella. Zool. Anzeiger, XII.
Jahrg., No. 299, pp. 64, 65. 4 Feb., 1889.
89>. Die Entwicklung der Bryozoencolonie im keimenden Statoblasten.
(Vorlaufige Mitth.) Zool. Anzeiger, XII. Jahrg., No. 324, pp. 675-679.
30 Dec., 1889.
‘Caldwell, W. H.
’83. Preliminary Note on the Structure, Development, and Affinities of
Phoronis. Proc. Roy. Soe. Lond., Vol. XXXIV. pp. 371-383. 1883.
Ehlers, E.
"76. Hypophorella expansa. Ein Beitrag zur Kenntniss der minirenden
Bryozoen. Abhand. d. kénigl. Gesellsch. d. Wiss. zu Gottingen, Bd. XXI.
pp. 1-156, Taf. 1-V. 1876.
Haddon, A. C.
83. On Budding in Polyzoa. Quart. Jour. of Mier. Sci., Vol. XXIII.
pp. 516-555, Pls. XXXVII. and XXXVIII. Oct., 1888.

Harmer, S. F.
’'85. On the Structure and Development of Loxosoma. Quart. Jour. of
Mier. Sci., Vol. XXV. pp. 261-337, Pls. XIX.-XXI. April, 1885.
’'86. On the Life-History of Pedicellina. Quart. Jour. of Mier. Sei., Vol.
XXVII. No. 101, pp. 239-264, Pls. XXT., XXII. Oct., 1886.
Hatschek, B.
77. Embryonalentwicklung und Knospung der Pedicellina echinata. Zeitschr.
f. wiss. Zool., Bd. XXIX. Heft 4, pp. 502-549, Taf. XXVIII. -XXX.,

u. 4 Holzsch. 18 Oct., 1877.

MUSEUM OF COMPARATIVE ZOOLOGY. 149

Hyatt, A.

'68. Observations on Polyzoa. Suborder Phylactolemata. Salem [Mass. ].
Printed separately from Proc. Essex Inst., Vols. IV. and V., 1866-68,
pp. i-iv., 1-103. 9 Pls.

Joliet, L.

"77. Contributions a |’ Histoire Naturelle des Bryozoaires des Cotes de France.
Arch. de Zool. Expér., Tom. VI. No. 2, pp. 193-3804, Pls. VI—-XIII. 1877.

’86. Recherches sur la Blastogénése. Arch. de Zool. Exper., 2° série, Tom.
IV. No. 1, pp. 37-72, Pls. IL, III. 1886.

Korotneff, A. A.

"74. Jlouxopaniz PatupiceLLa. Bull. Roy. Soc. Friends of Nat. Hist.
Moseau, Vol. X. Pt. 2, pp. 45-50, Pls. XIL., XIII. 1874. [Russian.]

"75. [Abstract of Korotneff, °74, by Hoyrr, in Hofmann u. Schwalbe’s
Jahresber. Anat. u. Phys. f. 1874, Bd. ILI. Abth. 2, pp. 869-372. 1875.]

'89. Sur la Question du Développement des Bryozoaires d’Hau douce. Mé-
moires de la Société des Naturalistes de Kiew, Tom. X. Liv. 2, pp. 393-
410, Tab. V., VI. 21 Oct., 1889. [Russian.]

Kraepelin, K.

’'86. Ueber die Phylogenie und Ontogenie des Siisswasserbryozoen. Biol.
Centralblatt, Bd. VI. Nr. 19, pp. 599-602. 1 Dec., 1886.

'87. Die Deutschen Siisswasser-Bryozoen. Eime Monographie. I. Anato-
misch-systematischer Teil. Abhandl. der Naturwiss. Verein in Hamburg,
mo X., 168 pp., 7 Taf. 1887.

Lankester, E. R.
| "74. Remarks on the Affinities of Rhabdopleura. Quart. Jour. Mic. Sci.,
; Vol. XIV. pp. 77-81, with woodeut. 1874.
: 85. [Article.] Polyzoa. Encyclopedia Britannica, Ninth Edition, Vol.
XIX. pp. 429-441. 1885.
Metschnikoff, E.

'71. Beitrage zur Entwickelungsgeschichte einiger niederen Thiere. 6. Al-
eyonella. Bull. de ?Acad. Imp. Sci. de St. Pétersbourg, Tom. XV.
pp. 507, 508. 1871.

Nitsche, H.

69. Beitrage zur Kenntniss der Bryozoen. I. Beobachtungen iiber die Hnt-
wicklungsgeschichte einiger chilostomen Bryozoen. Zeitschr. f. wiss.
Zool., Bd. XX. Heft 1, pp. 1-86, Taf. I-III. 1 Dec., 1869.

‘71. LBeitrage zur Kenntniss der Bryozoen. III. Ueber die Anatomie und
Entwicklungsgeschichte von Flustra Membranacea. IV. Ueber die
Morphologie der. Bryozoen. Zeitschr. f. wiss. Zool., Bd. XXI. Heft 4,
pp. 416-498, Taf. XXXV.-KXXVIII. and 4 Holzschn. 20 Nov., 1871.
[Also separate, pp. 1-83.]

'72. Betrachtungen iiber die Entwicklungsgeschichte und Morphologie
der Bryozoen. Zeitschr. f. wiss. Zool., Bd. XXII. Heft 4, pp. 467-472,
2 Holzschn. 20 Sept., 1872.

150 BULLETIN OF THE

‘75. Beitrage zur Kenntniss der Bryozoen. V. Ueber die Knospung der
Bryozoen. A. Ueber die Knospung der Polypide der phylactolemen
Susswasserbryozoen. B. Ueber den Bau und die Knospung von Loxo-
soma Keferstemii Claparéde. C. Allgemeine Betrachtungen. Zeitschr.
f. wiss. Zool., Bd. XXV., Supplementband, Heft 3, pp. 343-402, Taf.
XXIV.-XXVI. 22 Dec., 1875.

Ostroumoff, A. ;

’85. Remarques relatives aux Recherches de Mr. Vigelius sur des Bryozo-
aires. Zool. Anzeiger, VIII. Jahrg., No. 195, pp. 290, 291. 18 Mai,
1885.

’86>. Contributions & ’Etude zoologique et morphologique des Bryozoaires
du Golfe de Sébastopol. Arch. Slaves de Biologie, Tom. II. pp. 8-25,
184-190, 329-355, 5 Pls. 1886.

Pergens, E.

’°89. Untersuchungen an Seebryozoen. Zool. Anzeiger, XII. Jahrg., No.

317, pp. 504-510. 30 Sept., 1889.

Reichert, K. B.

70. Vergleichende anatomische Untersuchungen iiber Zoobotryon pellucidus
(Ehrenberg). Abhandlungen der kéniglichen Akademie der Wissenschaften
zu Berlin, aus dem Jahre 1869, II., pp. 233-338, Taf. I.-VI., Berlin,
1870.

Reinhard, W. W.

‘807. Zur Kenntniss des Sisswasser-Bryozoen. Zool. Anzeiger, III. Jahrg.,
No. 54, pp. 208-212. 3 May, 1880.

’80>. Embryologische Untersuchungen an Alcyonella fungosa und Cristatella
mucedo. Verhandl. d. Zool. Sect. VI. Vers. Russ. Naturf. Abstract by
Branpt, A., in Zool. Anzeiger, III. Jahrg., No. 55, pp. 234, 235.
10 May, 1880.

82. “Skizze des Baues und der Entwickelung der Stisswasser-Bryozoen.”
Charkow, 1882. 7 Taf. [Russian. |

’88. “Skizze des Baues und der Entwickelung der Siisswasser Bryozoen.”
Arb. Naturf. Gesellsch. Charkow, Bd. XV. pp. 207-310, 7 Taf. 1888.
[ Russian. ]

Repiachoff, W.

759, Zur Entwickelungsgeschichte der Tendra zostericola. Zeitschr. f. wiss.
Zool., Bd. XXV. Heft 2, pp. 129-142, Taf. VIT-IX. 1 Marz, 1875.

’'75>. Zur Naturgeschichte der Chilostomen Seebryozoen. Zeitschr. f. wiss.
Zool., Bd. XX VI. pp. 189-160, Taf. VI.-IX. 8 Dec., 1875.

Saefftigen, A.

’88. Das Nervensystem der phylactolemen Sisswasser-Bryozoen. (Vor-
laufige Mittheilung.) Zool. Anzeiger, XI. Jahrg., No. 272, pp. 96-99.
20 Feb., 1888.

Seeliger, O.
'89. Die ungeschlechtliche Vermehrung der endoprokten Bryozoen. Zeit-

MUSEUM OF COMPARATIVE ZOOLOGY. 151

schr. f. wiss. Zool., Bd. XLIX. Heft 1, pp. 168-208. Taf. IX. u. X., 6
Holzschn. 138 Dec., 1889.
Verworn, M.

’87. LBeitrage zur Kenntnis der Sisswassserbryozoen. Zeitschr. f. wiss.

Zool., Bd. XLVI. Heft 1, pp. 99-130, Taf. XII. u. XII]. 25 Nov., 1887.
Vigelius, W. J.

’84. Die Bryozoen, gesammelt wahrend 3. u. 4. Polarfahrt des ‘‘ Willem
Barents” in den Jahren 1880 und 1881. Bijdragen tot de Dierkunde.
Uitgegeven door het Genootschap Natura Artis Magistra, te Amsterdam,
1l¢ Aflevering, 104 pp. 8 Taf. 1884.

ee

a ,

EXPLANATION OF FIGURES. |

All figures were drawn with the aid of a camera lucida from preparations of
Cristatella mucedo.

An.

an.

atr.

br. loph.
can. cre.

can. crc.’

47

can. crc.

4st

can. Cre.

can. e stm.

cav. loph.

cev. pyd.
cl, fun.
cl. mt.
cl. mus.
cee.

cen.

cp. sec.

cta.

di sep.
di sep. r.
ec.

e stm.

e t. cal.
ex.

fun.

ga.

gn.

:
kmp. drm.
loph.’

lu. gm.
lu. gn.

ABBREVIATIONS.

Anal side of polypide.

Anus.

Atrium.

Lophophore arm.

Ring canal, circumoral
part.

Ring canal, outer lopho-
phoric part.

Ring canal, inner lopho-
phoric part.

Ring canal, supra-gan-
glionic part.

Epistomic canal.

Cavity of lophophore
arm.

Neck of polypide.

Young cells of funiculus.

Migratory cells.

Young muscle cells.

Coecum.

Ceenoceel.

Secreted bodies of ecto-
derm.

Cuticula.

Intertentacular septum.

Radial septum of colony.

Ectoderm. -

Epistome.

Celomic epithelium.

Outer layer of bud.

Funiculus.

Stomach.

Ganglion.

Inner layer of bud.

Kamptoderm.

Place of union of arms
of lophophore.

Lumen of the bud.

Lumen of the ganglion.

mu. inf.

Mu. SU.

or.
pam. atr.

pam. gn.
phe.

Muscularis.

Inferior parieto-vaginal
muscles.

Longitudinal muscle fi-
bre of muscularis.

Retractor muscle of
polypide.

Rotator muscle of pol-
ypide.

Superior parieto-vaginal
muscles.

Transverse (circular)
muscle fibre of mus-
cularis.

Lophophoric nerve.

‘Nucleus of muscle fibre-

CHsophagus.

Atrial opening.
Ovum.

Oral side of polypide.
Mouth.

Floor of atrium.
Floor of ganglion.
Pharynx.

pyd. {i., ii., &c.] Polypide.

pyd. fill.
pyd. ma.
rt.

sol.
spht.
sul. or.
ta.

tct.

tet. gn.
vac.
vlv. cr.

vlv. py.

Daughter polypide.
Mother polypide.
Rectum.

Sole.

Sphincter.

Oral groove.
Tentacle.

Oral tentacle.
Roof of colony.
Roof of ganglion.
Vacuole.

Cardiac valve.
Pyloric valve.

ae

“ .

‘ a 1 :

be)

DAVENPORT. — Cristatella.

Fig. 1.
“cc ae
“ 3.
«é 4.
pe es
SS eLeG:
fe the

PLATE I.

A portion of the lateral rim of a colony. An optical section taken just
below the roof of the colony, showing the arrangement of polypides.
xX 72.

Origin of the stolon (I.) from the neck of a mother polypide of about
Stage XII. (Fig. 18). Sagittal section of mother polypide. The
margin of the colony is to the left. x 890.

Earliest stage in the origin of a bud from a young mother polypide.
Sagittal section. Margin to left. x 390.

Origin of a bud from a mother polypide of about the age of that of Fig. 3.
Sagittal section. The margin of the colony is to the right of figure.
x 890.

Sagittal section of a double bud. Margin of colony to the left. x 390.

Later stage in bud formation of same type as Fig. 4. Sagittal section.
x 890.

A part of the right side of a polypide of a stage of development interme-
diate between those of Figs. 19 and 73. Seen from the sagittal plane.
The cut surface lies to the right of the sagittal plane, and passes
through the orifice of the right lophophore arm. The alimentary tract
thus lies immediately above the plane of the paper. X 150.

_ DAVENPORT. — CRISTATELLA.

G

@a
——
“ut et cel,

li cain ct i i te ie ee

B Meisel, lith. Boston

i
«

«

ig
i
‘
’
‘
.

rr
i a
a

a»

re yy,
eh Gs bh
Ba at:

/ SP Ale a ; :
Ve awe Pe), UL oe MS ie

CUD sey Toy) Wh

DAVENPORT. — Cristatella.

PLATE II.

All figures are magnified 390 diameters, and are from sagittal sections.

. 8. Stage II. in the same series as Fig. 2. The funiculus, fim., has moved

farther from the mother polypide. Margin to left.
9. Stage 1V. The inner layer, 7., of the bud is definitely formed, and the
external layer is greatly thickened. Margin to left of figure.

10. Stage V. The cells, 7., have arranged themselves in a layer, and begin
to form an invagination. Margin to right.

11. Stage VIII. The first indications of the alimentary tract appear as a
depression in the inner layer, rt. The funiculus, cl. fun., has begun
to form, as is indicated by a disturbance of the celomic epithelium.
Daughter bud forms Stage VI. in a series beginning with L, Fig. 3.
Margin to left.

12,13. Successive stages in the formation of the alimentary tract.

14. Stage VI. The two cell-layers are now definitely formed, and a lumen
has begun to appear in the inner. Margin to right.

15. Stage III. in the stoloniferous type of budding. Stolon has elongated
greatly, and active cell division is taking place at its distal (i. e. mar-
ginal) end.

DAVENPORT. - GRISTATELLA,

®& See pn hl
pyding PB e0

CoN

009 COS

Pe

ene aa ‘ Ps
j \\
rare

Dk

C.B.D. del, B Meisel lith.Boston

»
i 2
- ’ 7
J
- 4
‘ ’
‘
’
=
> *
7 ‘
4
se “N ,
j ii?
"
’ ‘
. F
7) .
’ '
Ly .
al
rm
F
. i)
——.
’
'
i
: ’
‘
7
a
‘
=
‘
F
hen. }
a 3
3 4
FE
*
,
= ~
- 4
\
’
s
‘
‘
i
Fi
@
i
¥
’
.
‘ Fi aS
:

eae 4 7 Cy Lee .
Le ie Sa | fe He,

DAVENPORT. —Cristatella.,

Fig. 16.

olan

PLATE III.
All figures are magnified 390 diameters.

A later stage (VI.) in the direct type of bud formation. The mother polyp-
ide is cut to one side of its sagittal plane, and shows the invagination
of the lophophore arm (br. oph.). The funiculus appears as scattered
cells about both buds.

A still later stage in the same series as Fig. 16. The daughter bud ( VII.)
has alumen. In the mother polypide (XI.) the atrium has enlarged
by the inshoving of the lophophore arms. The cesophageal and rectal
invaginations are not yet continuous, and the ring canal (can. erc.) has
begun to appear oralward in the sagittal plane. Sagittal section. |

Stage XII. Alimentary tract nearly complete. Beginning of the forma-
tion of the ganglion. One of the lophophore arms is cut tangentially.
Sagittal section.

Stage XIII. Ganglion closing. The lophophore arm cut tangentially.
Sagittal section.

and 21. The positions and directions of the planes of these sections are

shown by their projections on a sagittal section (Fig. 11, lines 20, 21)
of an individual of the same age. To show non-participation of the
outer layer in the first stage in formation of the alimentary tract.

Karly stage in direct bud-formation. Origin of funiculus, c/. fun. Sagit-
tal. Margin to right.

The position and direction of the plane of this section are shown by its
projection on a sagittal section (Fig. 19, line 23) of an individual of
the same age. This figure shows the folds of the inner layer at the
mouth of the ganglionic sac.

9
fi >loph.
{

“a 6, i , Saat d
Soho 90890-36”” ,.

Pan.gr-
EX.

i

ane 7 rp A
AO) NAAGY ©

C.B.D. del. B Meisel, lith Boston.

>

p

y

\
, aan)

ie
i ; - SPD tep ante. at bid ? ) as
‘ i fs ei'y AW i " 5 Ler | aad a | ‘ ‘af ai’, A enol F ini OP To ~ ’ : 4 i} iw _ A :

DAVENPORT. — Cristatella.

PLATE IV.

All figures, except Fig. 39, are vertical right-and-left sections, and all are magnified

390 diameters.

Figs. 24-26. Three sections from a series passing from the oral to the aboral face

“ce

“cc

“cc

27-29.

30-82.

33-38,

39.

of a polypide of about Stage X., and cutting it in the planes indi-
cated by the lines 24-26, Fig. 13.

Three sections of a series cut from a polypide of Stage XI. The
planes of section are indicated in the lines 27-29, Fig. 17.

Three sections; whose positions are indicated by the lines 30-82, Fig.
18, cut from a polypide of Stage XII.

Six sections cut from a pclypide of Stage XIII. in the directions indi-
cated in Fig. 19 by the lines 33-38.

A horizontal section of a polypide somewhat older than that repre-
sented in Fig. 18, and passing nearly in the direction of the line 43.

— |

DAVENPORT. — CRISTATELLA. i

QQTOMDEFR ,

SAS

KEE a \
6 Hj is ig! LEN ost CLLIN en
0) 9) Co, G vig Ws sty

CANCE

Nae 662. a on
LOELY/ WAS Oa OES
9 lun. py, (al ®aoeliso\
ute oh! | \Oj :
ANGLE’ a ae 9 ge
SUELO ED Nees?
. ~— :

CRD del. GM. BMeisel,lith. Boston.

i=

Fee Se

a

MA
ACh. Nine

DAVENPORT. — Cristatella.

PLATE V.

Figs. 40-45 are four horizontal sections of a polypide of Stage XII. passing in the
direction indicated by the lines 40-45, Fig. 18. x 390. .

“ 44, 45 are horizontal sections of a polypide of Stage XIII. The direction of
the cutting planes is indicated by the lines 44, 45, Fig. 19. x 390.

“ 46-48. Sections through the migrating end of the funiculus, showing its rela-
tion to the coelomic epithelium of the roof. The ectoderm is not
shown. The arrow indicates direction of motion. X 390.

“ 49. Transverse section through the funiculus, showing the loose migratory
cells. X 390.

“ 60,51. Horizontal sections of a polypide slightly younger than Stage XIV.,
Fig 75. Of these two sections, Fig. 50 is nearer the roof of the colony,
and immediately above the ganglion. Fig. 51 is the second section
below, and passes through the middle of the ganglion. X 390.

“ 62. Sagittal section of the region about the brain of a polypide somewhat
older than that shown in Fig. 77. This figure is reversed relatively
to Fig. 77. x 600.

| DAVENPORT. — CRISTATELLA. ead

| © - boph
© 09 oem im
i
\

oO ~
an (o) Ors

-. x PN
a S334%>

fo!

oO
0)
g
oO
O7 fl |

00! fo

COC!

6
ro) /9
@O\kc
@

=

Oo?
78
Q

yO,

¥

(e

3°09 % 99
= ~
"24

Oy
©
STA
\e
(sy

Wc

e 6
a!16
]

B.Meisel ith Boston

J
.
.*
‘
oa
Uy
;
7

iy

A

A ry
mir ee

7,4 7] . ye 8

DAVENPORT. — Cristatella.

ee

c

“

PLATE. VI.

Young funiculus, showing its connection with polypide. x 390.

Origin of muscles. The section, passes diagonally across a partition at
the left, di sep. r., and cuts the polypide tangentially at the right.
x 390.

Section including a radial portion, showing the position of the muscles in
the partition near the margin of the colony. 890.

Section through the retractor and rotator muscles of a polypide of about
the age of that shown in Fig. 77. X 890.

Young funiculus, whose upper end is free from the coelomic epithelium of
the roof of the colony. x 390.

Section through the sole, showing the relation between the muscle cells
and the muscularis of the sole. X 600.

Section across a radial partition, and both rotator and retractor muscles
which are migrating from the roof to the sule. 390.

Section at right angles to the wall of the colony, showing the elongated
and unmetamorphosed cells of the margin. X 390.

pamatr ey

Cane stom,
AY loph,

(a)
Oh

ea
x = et.coel.
sol

’ CBD. del. = B Meisel lith.Boston.

+

my

a

Che ian [Ot 0 t ; : ivf ff
i}, ah wera yay ; ers ey
7 > fe * a i
Lins eal ea 1 | ray mer Y
a ‘
»
ee b |
SPR yes £ aa j
‘ : ee AR |
j
P, be hh } nM e
F aj
ue i” ee ite
, .
= ) ?
A ‘ ‘PTY
| ‘ | ‘
b- 9 or,
;
, }
: :
iv
oe ‘’ . - >
j iva
; / 4 ,

. '
i
1 a wh Cc 4
]
.
7 ad ; 4 '
iA
a
‘ i i =
v4
- eal
~3 i ; :
‘
1) + Ve
i “
‘ ty
f
4 r
' ws
te
‘
7]
‘ _
\ ~
*
7
t = << ‘
.
i
5 :
‘
{
‘
’
% ,
.
.
y
.

DAVENPORT. — Cristatella.

PLATE VII.

Figs. 61-63. Three vertical right-and-left sections of same polypide passing from
posterior end anteriorly. About Stage XIV. (Fig. 78, Plate VIII.)
x 890.

“ 61. Section through lophophore arms, showing their fusion, /oph.’, and the
position of the ring canal, can. erc.', can. cre.”

* 62. Section just posterior to anal opening, showing openings af lophophoric
pockets.

« 63. Section through ganglion, showing early stage in formation of lopho-
phorie nerve, parts of the ring canal, and young tentacles.

“ 64. Cross section of lophophore arm, near termination of young nerve, at
place marked 64, Fig. 71. 1000.

“ 65-67. Three successive sections through end of lophophore nerve in re-
gions marked 65, 66, and 67, Fig. 71. These figures are from the
same individual as Fig. 64, but from the opposite lophophore arm.
x 1000.

“ 68. Vertical right-and-left section through ganglion of an individual slightly
younger than Fig. 63, showing origin of cornua by outgrowth of the
walls of the ganglion, with an extension of the lumen of the latter.
x 600.

“ 69,70. Longitudinal sections of two stages in the development of a tentacle,
Fig. 70 being the younger. X 390.

“ 71. Section through ganglion and growing lophophore nerve. Stage XIV.
x 490.

Pr Vil.

DAVENPORT. — GRISTATELLA.

p.drmn.

Am,

{

Vo

/
7
‘
C
*
. 4 \) .
©
e ,
S

B Meisel, lith.Boston

*-

7 /

see on eS eaaal

Davenport. — Cristatella.

75.

76.

PLATE VIII.

Sagittal section of an adult polypide. The lophophore has been omitted.
Outlines with camera lucida. Nuclei put in free hand. X 175.

Sagittal section of bud. Stage XIV. The margin of colony to left.
* Ectodermal cells derived from neck of polypide. X 590.

Nearly horizontal section of a bud a little older than that shown in Fig.
73. The plane of section passes obliquely upward and forward. The
tentacles are cut at different heights. X 390.

Transverse section of lophophore arms before separation. The connecting
band, loph.’, is reduced to threads. The polypide has already evagi-
nated. The section figured is the seventh from the distal end of the
arms, — about 404 distant. x 390.

Transverse section of lophophore arms immediately after separation. The
tentacles arising from can. crc.” were previously fused. X 390.

DAVENPORT. — GRISTATELLA.

AE
@O 86
Sous

B Meisel lith.Baston,

ie a, a i wt An iis
ere! wr . CU AEN a

. ark re

Bin mie 8 bie

uit»

ci
ah pn Bat j roe o a ib sah

Pherae Pw yi ie Fi

a,
Aart fade 4a bi Wi :
calm ar? Ay ae, hata ert)

een, 5 | :
aw i sh
apa OAD et yeliad oy
Las yer
; (‘Kyo

DAVENPORT. — Cristatella.

Fig. 77.

ee ioF

= Ato.

neu:

- Oe.

PLATE IX.

Sagittal section through a polypide, of which the atrial opening (of. atr.)
has already begun to form. X 390.

Horizontal section through the circumoral part of the ring canal, can. crc.,
showing its free communication with the ccenoceel (cen.). Adult..
x 175. |

Vertical section through the roof of the colony (to the left) and the
kamptoderm (to the right), showing their connection by the inferior
parieto-vaginal muscles (mu. inf.) at an early stage of their develop-
ment. %X 600.

Horizontal section in position marked 80, Fig. 72, Plate VIIL, showing
epistomic canal, can. e stm., and supra-ganglionic part of ring canal,
can. cre’ X $90.

Section cutting lophophore at base of tentacles. The arm of the right
side only is shown entire. Stage of Fig. 77. Xx 176.

DAVENPORT. — CRISTATELLA.

a
= ’
————_ ‘i
2) ,
“ale ,
'

CPSC...

ae

1 C iS)

atl a)

ee ‘;
| si[s°'e fk» ae BN @ ee) < se
rlagsie fo SARS Ae 000
Vos Sls sei’ oe AVES
ee off Tios'| sSvaNs .
y NG oes cele \
° y

CBD. del.

B Meisel lith Bos ton

7 aa
' § if -
i : : %
\,
:
)
.
{ ~ Pl S
. 7 A
= yh
.
’
s
‘
= 1
i & i ‘
a
)
\
t
\
ac 7 pen
. '
: - a
?
hen
~ beste
» a7 it
i
5 '
i
‘
‘ cs
~ »
i}
-
>
‘ {
i
‘
.
-
is {
-
mu
[
= t
:
a

of be 7 ‘ne re ae
, Dd my iad |
; Lorery ant

a

DAVENPORT. — Cristatella.

PLATE X.

Fig. 82. Transverse section of stomach of adult polypide. x 3890. Compare with

Fig. 93.

83. Transverse section of proximal part of coecum of same individual as that
of Fig. 82. x 390. Cf. Fig. 94.

84. Transverse section of esophagus of a polypide whose atrial opening is
just formed. xX 390.

85. Transverse section of the cecum of an adult polypide near its distal
extremity. 590.

86. Vertical section across a radial partition at its junction with colony-wall.
x 600.

87. Horizontal section of radial partition at its junction with colony-wall.
x 600.

88. Small colony of Cristatella, drawn from transparent object, showing pol-
ypides in optical section at different focal planes. X circa 40.

89-92. Muscle fibres in successive stages of development. From thick sec-
tions. ~ 390.

93. Transverse section of stomach of the same polypide'as that from which
Fig. 84 was taken; representing, therefore, a considerably younger
stage than Fig. 82. Xx 390.

94. Transverse section of coecum of the same polypide as that from which
Figs. 84 and 93 were taken, cut in a region nearly corresponding to
the position of that shown in Fig. 83. 890.

95, 96. Two horizontal sections of a part of the margin of a small colony
in which radial partitions are being rapidly formed in correspondence
with rapid budding. Fig. 95 lies near the sole; Fig. 96, near the roof.
The same figures refer to the same partition. X 300.

a

DAVENPORT. — CRISTATELLA.

7 Al

i

. rs =< = S : 5
CBD dei. B Meisel, lith Boston.

-_
-

4 >

ix
we F

al ri) >) nts ka “il vin
oT cy ly Aa ee a eee | ee

DAVENPORT. — Cristatella.

PLATE XI.

Figs. 97-99. Vertical sections, showing three successive stages in the degeneration
of the roof to form the atrial opening, of. atr., and development of the
parieto-vaginal muscles. X 390.

“ 100. Late stage in the development of the ectoderm, showing its extreme :
modification between adult polypides. x 390.

.

/

DAVENPORT. — CRISTATELLA.

Pi. Xl.

B Meisel, lth. Boston.

No. 5.— The Hyes in Blind Crayfishes. By G. H. PARKER!

Tw the fall of 1888 Mr. Samuel Garman placed at my disposal several
erayfishes ? which had been collected by Miss Ruth Hoppin in the caves
of Jasper County, Missouri. The specimens were given to me with the
suggestion that I should ascertain the extent to which their eyes had
degenerated, for, judging from external appearances, these organs had
become as rudimentary as the eyes of the blind crayfish, Cambarus
pellucidus, Tellk., from Mammoth Cave. In order to establish compari-
sons it was desirable to study the eyes in C. pellucidus, and for this
purpose specimens of this species were kindly furnished me from the
collections in the Museum of Comparative Zodlogy. These specimens,
as well as those collected by Miss Hoppin, were preserved in strong
alcohol. My study of this material was carried on in the Zodlogical
Laboratory of the Museum, under the direction of Dr. E. L. Mark.

Notwithstanding the general interest which zodlogists have shown in
the blind crayfishes there have been very few publications on the minute
structure of the eyes of these animals. The earliest contribution to this
subject was from Newport, who, in discussing the ocelli of Anthophora-
bia, incidentally described the structure of the eye in Cambarus pellu-
cidus. According to Newport’s account (’55, p. 164), the eyes in this
species would seem to be only partially degenerated, for although the
retinal region is not pigmented, the corneal cuticula is nevertheless
divided into irregular facets, or ‘‘ corneales,” as they are termed, “‘ and
the structure [hypodermis] behind these into chambers to which a small
but distinct optic nerve is given.”

The second investigator who studied the eyes of blind crayfishes was
Leydig (’83, pp. 36 and 37). The material which was accessible to him
was unfortunately so poorly preserved that it was of little value for his-
tological purposes. He nevertheless satisfied himself that the cuticula
in the corneal region was not facetted. He also quoted from an abstract

! Contributions from the Zodlogical Laboratory of the Museum of Comparative
Zoology, under the direction of E. L. Mark, No. XX.

2 These crayfishes had previously been submitted to Dr. Walter Faxon for
determination. They have since been described by him as a new species, under
the name of Cambarus setosus, an account of which will be found in Mr. Garman’'s
recent paper (’89, p. 237) on “‘Cave Animals from Southwestern Missouri.”

VOL. XX. — NO. 5.

154 BULLETIN OF THE

of Newport’s paper, to the effect that the eye is ‘ohne Hornhaut, Pig-
ment und Nervenstiabe.” The phrase “ohne Hornhaut ” means, I be-
lieve, that a facetted cornea is not present ; at least this seems to be the
interpretation placed on it by Leydig, for the quotation is shortly fol-
lowed by this sentence : ‘‘ Dort wo man eine gefelderte Cornea zu suchen
hatte —am Gipfel des Kegels — zeigt sich die Haut von der gewohn-
lichen Beschaffenheit.” There was greater reason for Leydig’s regret
that he could not consult Newport’s original paper than Leydig himself
appreciated ; for, although he probably had no reason to consider the
abstract incorrect, if his quotation from it is exact, it differs at least in
one respect from Newport’s account. Newport described the cornea as
facetted ; Leydig’s quotation from the abstract states that it was not
facetted. I have been unable to discover where this abstract was pub-
lished, but, since Leydig quotes directly from it, the probabilities are
that the discrepancy between his quotation and Newport’s actual state-
ment is to be attributed to an error in the abstract. Aside from this
difficulty, it must be borne in mind that Leydig and Newport in their
observations on the cornea by no means agree ; for while Newport really
describes the cornea as facetted, Leydig states from his own observa-
tions that it is without facets. According to Leydig, then, the eye of
C. pellucidus is more completely degenerated than the observations
of Newport would lead one to suppose.

The latest account of the eyes in blind crayfishes forms a part of
Packard’s paper on “‘ The Cave Fauna of North America” (’88, pp. 110
to 113). Newport and Leydig studied C. pellucidus ; Packard had the
opportunity of studying not only this species, but also C. hamulatus,
Cope and Packard, from Tennessee. In both species according to Pack-
ard the cornea was without facets, and the hypodermis was not thick-
ened in the retinal region, but an optic nerve and ganglion were present.

The results obtained by Packard thus confirm those given by Leydig.

From this brief historical review it will be observed that one of the
principal questions concerning the eyes of blind crayfishes deals with
the extent of their degeneration. This change has not only affected the
finer structure of the retina, but it has also altered the shape of the
optic stalk. I shall therefore begin with a description of the external
form of the stalks.

The optic stalks of blind crayfishes are not only proportionally smaller
than those of crayfishes which possess functional eyes, but they have in
the two cases characteristically different shapes. In crayfishes with

MUSEUM OF COMPARATIVE ZOOLOGY. 155

fully developed eyes the stalk is terminated distally by a hemispherical
enlargement ; in the blind crayfishes it ends as a blunt cone. This
cone-shaped outline is especially characteristic of C. pellucidus (Fig. 2).
It will be observed that in this species the optic nerve (x. opt.) termi-
nates in the hypodermis immediately below the blunt apex of the cone.
In C. setosus (Fig. 1) the termination of the optic nerve is also at the
apex of a blunt cone. In this case, however, the axis of the cone does
not coincide with the axis of the stalk, as it does in C. pellucidus, but
the two axes meet each other at an angle of about forty-five degrees,
and in such directions that the conical protuberance at the distal end of
the stalk is directed forward and outward from the median plane of the
animal. The protuberance is rather more blunt in C. setosus than in
C. pellucidus (compare the regions marked r. in Figs. 1 and 2).

Through the kindness of Dr. Walter Faxon I was enabled to examine
two specimens of C. hamulatus. In this species the stalks also termi-
nate in blunt cones. They are not so pointed as in C. pellucidus, but
approach the more rounded form of C. setosus.

The three species, C. pellucidus, C. hamulatus, and C. setosus, are the
only blind crayfishes thus far known in North America, and, as they
agree in having a conical termination to the optic stalks, a peculiarity
not observable in crayfishes with functional eyes, it may be concluded
that the conical form is characteristic of the stalks iu blind crayfishes.
Unquestionably, this conical shape is coupled with the degenerate con-
dition of the retina. .

In describing the finer anatomy of the eye it will be more convenient
to begin with the condition found in C, setosus. Figure 1 is drawn from
a longitudinal horizontal section of the optic stalk in this species. The
plane of section passes through the region where the optic nerve and
hypodermis are in contact. This region (Fig. 1, 7.) corresponds to the
retina of other crayfishes. The optic stalk is covered with a cuticula
(Fig. 1, ¢t.), which is of wneform thickness and which resembles the
cuticula of the rest of the body. In this respect the stalk differs from
that of decapods with well developed eyes, for in these, although much
of the stalk is covered with ordinary cuticula, the retinal region is pro-
vided with a thin flexible cuticula. This has been named by Patten the
corneal cuticula ; it cannot be said to be differentiated in C. setosus. In
optic stalks with functional retinas the corneal cuticula is usually
facetted, but in C. setosus no indication of facets is discoverable.

The undifferentiated condition of the cuticula leads one to antici-
pate a simple condition in its matrix, the hypodermis. The latter is a

156 BULLETIN OF THE

continuous layer of cells (Fig. 1, 4d.) with its distal face applied to the
cuticula and its proximal face bounded by a fine but distinct basement
membrane (mb.). The layer is throughout very nearly uniform in thick-
ness ; at least it is not thicker in the region of the retina than at many
other places, and the slight variations in its thickness are not in signifi-
cant regions. The only feature of the retinal hypodermis which would
suggest that it was unlike the rest is the somewhat closer crowding of
its cells. This manifests itself in the arrangement of the nuclei in two
or three irregular rows, instead of a single one. In other respects the
nuclei of the retinal region and the surrounding hypodermis are essen-
tially similar.

The optic nerve (Fig. 1, 2. opt.) consists of a poorly defined bundle
of nerve-fibres which extend from the optic ganglion to the hypodermis.
The nerve-fibres are doubtless intimately connected with the cells in the
hypodermis, for the basement membrane is interrupted where the nerve
and hypodermis are in contact. It is probable that the basement mem-
brane is reflected from the hypodermis to the optic nerve, although I
have not been able to observe this with clearness.

Recent investigations support the conclusion that the retina in the
crustacea is derived from the hypodermis. In C. setosus that portion
of the hypodermis from which the retina would be derived is scarcely
distinguishable from other parts of the same layer. The retina in this
species, therefore, has so completely degenerated that it has at last
returned to the condition of almost undifferentiated hypodermis.

That the optic nerve still retains its connection with the retinal area
is, on the whole, not so significant a condition as one might at first sup-
pose. It is probable that the optic nerve arises in this species as it
does in the lobster. I have elsewhere (Parker, ’90, p. 43) attempted to
show that in the lobster it is not an outgrowth from either the optic
ganglion or the retina, but that, as the ganglion was differentiated from
the hypodermis, the optic nerve remained as a primitive connection be-
tween these two structures. So long, then, as an optic ganglion should
be differentiated one might expect an accompanying optic nerve; but
the nerve would be present as a passive connection between hypodermis
and ganglion, rather than as a structure which had retained that posi-
tion by virtue of its continued functional importance.

The foregoing account of the eye in C. setosus is based upon obser-
vations on three individuals of this species: Two of these measured,
from the tip of the rostrum to the end of the telson, 6 cm.; the third,
4.2 cm. In the three individuals the eyes presented essentially the

MUSEUM OF COMPARATIVE ZOOLOGY. 157

same condition. Figure 1 is taken from one of the larger individuals.
In this specimen the cuticula was somewhat thinner and the hypoder-
mis rather thicker than in the other two. This I believe was due
to the fact that the animal had recently moulted.

So far, then, as the eye of C. setosus is concerned, although the optic
ganglion and optic nerve are present, the retina has undergone a com-
plete degeneration, and is now represented by a layer of undifferentiated
hypodermal cells.

The eyes of Cambarus pellucidus present a somewhat different condi-
tion from that described in C. setosus. A longitudinal horizontal sec-
tion of the optic stalk of C. pellucidus is shown in Figure 2. The outer
surface of the stalk is covered with a cuticula (c¢.) of uniform thickness,
and there is no indication of facets. Excepting at the apex of the
stalk, the hypodermis (Ad.) is composed of a remarkably uniform layer
of cells. As in C. setosus, it is bounded on its deep face by a deli-
cate basement membrane (mb.). Both an optic ganglion (gn. opt.) and
nerve (z. opt.) are present, the latter being connected with the hypo-
dermis. In all these respects C. pellucidus resembles C. setosus, but
when the retinal part of the hypodermis in the two species is compared
a striking difference can be seen. The retinal hypodermis in C. se-
tosus (Fig. 1, 7.) is, as we have seen, substantially like the remaining
hypodermis of the optic stalk. The retinal hypodermis in C. pelluci-
dus (Fig. 2, 7.) is much thicker than the hypodermis of the stalk. With
this thickened region of the hypodermis the optic nerve is connected,
and there is no question, therefore, that this thickening represents the
rudimentary retina. Omitting minor details, the form of the thick-
ening is that of a plano-convex lens, the curved surface of which is
applied to the concave inner face of the cuticula at the distal end of the
stalk. The optic nerve is attached to the central part of the flat face
of the thickening.

When the retinal thickening is carefully studied by means of radial
sections, one can see that it differs from the neighboring hypodermis
not only in thickness, but also in the fact that it contains two kinds of
substance: a protoplasmic material uniform with that of the rest of the
hypodermis, and a number of relatively large granular masses (Fig. 3,
con.). These granular masses contain two, three, four, or sometimes five
nuclei, and nuclei are also to be found scattered through the undiffer-
entiated protoplasmic substance. The nuclei in the granular masses
are slightly smaller than those in the surrounding portion of the hypo-

158 BULLETIN OF THE

dermis ; they are, moreover, round in outline, while the other nuclei are
usually somewhat elongated. The same features can be observed in
tangential sections (ig. 6). Here, however, the outlines of the larger
nuclei no longer appear oval, since these nuclei are now cut in a
plane at right angles with their elongated axes. The nuclei in the
hypodermis which adjoins the retinal thickening resemble the larger
oval nuclei of the thickening. Nowhere in the adjoining hypodermis
have the granular masses with their smaller nuclei been observed. It
is therefore clear, that in C. pellucidus the retinal hypodermis is dis-
tinguished from the neighboring hypodermis, not only by its greater
thickness, but also by the fact that it is composed of two kinds of sub-
stance, each with its special form of nucleus. Since the protoplasmic
material of the retinal region contains nuclei which resemble those of
the surrounding hypodermis, it is probable that this material represents
hypodermis which has remained unmodified after the differentiation
of the granular bodies. As shown in Figure 3, the granular bodies are for
the most part limited to the deeper portion of the retinal thickening, and
the oval nuclei occupy the more superficial part. If these oval nuclei
represent undifferentiated hypodermal cells, it is only natural that they
should occupy a superficial position, for it is there that the function of
such cells, namely, the secretion of cuticula, could be most advanta-
geously carried on. In tangential sections of the retinal thickening,
both the nuclei of the undifferentiated hypodermis and the outlines of
the cells to which they belong are distinguishable (Fig. 5). These cells
when compared with those from the hypodermis of the sides of the stalk
(Fig. 4) are seen to be much smaller than the latter. Like those from
the sides of the stalk, however, they present no definite grouping.
This accords with the fact that the cuticula presented no special mark-
ings, such as facets, etc., for such markings could of course result only
from some special grouping of the secreting cells.

It is difficult to say what the granular bodies with their contained
nuclei are. Doubtless they represent some element in the retina of the
functional eye reduced by degeneration to this form. The ommatidium
or structural unit in the retina of a crayfish consists of five kinds of
cells. , These are as follows: first, two cells in the corneal hypodermis,
lying next the cuticula; second, four cone-cells directly below the
corneal hypodermis; third, two pigment-cells, the distal retinule,
flanking the cone-cells ; fourth, seven pigment-cells, the proximal reti-
nul, surrounding the rhabdome ; fifth, a few yellowish accessory pig-
ment-cells limited to the base of the retina. Excepting the accessory

= aes

MUSEUM OF COMPARATIVE ZOOLOGY. 159

pigment-cells, all the cells in an ommatidium are ectodermic in origin ;
the accessory pigment-cells are probably derived from the mesoderm.
Of these five kinds of cells, the granular bodies probably do not repre-
sent the accessory pigment-cells, for in fully developed eyes the latter
lie on both the distal and proximal sides of the basement membrane,
whereas the granular bodies are found only on the distal side of that
structure. The granular bodies, then, more likely represent one of the
four remaining elements, all of which naturally occur only on the distal
side of the membrane. It is not probable that the granular bodies
represent the cells of the corneal hypodermis, for these produce the cu-
ticula of the retinal region, and if they have any represeutatives, those
representatives must be the distal layer of unmodified hypodermal cells
already indicated in the retinal thickening. The position of the granular
bodies, therefore, precludes their representing corneal hypodermis. If
then the granular bodies are not accessory pigment-cells nor corneal
hypodermis, they must be either distal or proximal retinule or cone-
cells. In a previous paper I have given reasons for considering the
proximal and distal retinule as both originating from a common group
of cells, the retinule. These are essentially sensory in function, as con-
trasted with the cone-cells, which are merely dioptric. The question
then narrows itself to this: Are the granular masses clusters of dioptric
cone-cells or sensory retinule ?

In determining to which of these two groups of cells the granular
masses belong, the relation which the latter sustain to the fibres of
the optic nerve would doubtless be of great importance, for the nerve
fibres in fully developed eyes are known to terminate in the retinule,
not in the cone-cells. Unfortunately, the histological condition of my
material was such as to preclude the possibility of determining this
question.

The fact that each granular mass contains several nuclei clearly indi-
cates that it consists of several cells. The number of cells in each
mass, Judging from the number of nuclei, varies from one to about five,
the more usual number being three or four. When one compares the
condition of intimate fusion which the cells of each mass present with
the normal condition of the retinule and cone-cells, the masses must
certainly be admitted to resemble more closely the cone-cells. More-
over, the number of cells in each mass, although variable, is nearer to
that of the closely united cone-cells than to that of the retinule. Not
only do the number of cells involved and the intimacy of their fusion
favor the idea that each mass represents a degenerate cone, but the

160 BULLETIN OF THE

eranular substance of the mass also closely resembles the granular ma-
terial of a cone. For these reasons it seems probable that the granular
nucleated masses in the retinal region of C. pellucidus are the degen-
erate representatives of the cones in normal eyes.

The fact that, of all the ectodermic elements of the retina, only the
eranular nucleated masses continue to be differentiated, throws them
into strong contrast with the surrounding structures. The retention of
these masses may mean that on account of their extreme differentiation
they have had time to respond only incompletely to the influence of
degeneration ; or it may imply that phylogenetically they were among
the earliest retinal structures differentiated. Admitting them to be
degenerated cone-cells and merely dioptric in function, one can scarcely
conceive how they could have been differentiated before the sensory
cells which they serve. But even if they cannot be regarded as more
primitive structures than retinule, their retention still may be signifi-
cant, as an indication that the ommatidia of primitive crustaceans con-
tained cone-cells as well as retinule.

Former studies have led me to believe that the difference in the
ommatidia of various crustaceans could be explained on the assump-
tion that the number of elements has been gradually increased from
lower to higher forms by cell-division. The simplest conceivable rep-
resentative of an ommatidium in the crustacea might then be a sin-
gle cell. This would be of course a sensory cell; by its division,
the more complicated ommatidia might subsequently be derived from it.
In such an event, the cone-cells must be modified sensory cells; but
the fact that these cells persist in so rudimentary a retina as that of
C. pellucidus points rather to the conclusion, that they are probably
almost as old, phylogenetically, as the retinule themselves, and that
primitive ommatidia consisted of at least two kinds of cells, sensory
cells or retinule, and cone-cells, derived not from degenerated sensory
cells, but from the undifferentiated hypodermis.

As I have already shown, the results which Newport, Leydig, and
Packard arrived at are not always in agreement. ‘This might be ex-
plained by the fact that the organ under consideration is a degenerated
one, and consequently subject to considerable individual variation. This
supposition, however, is not supported by anything I have observed.
The preceding account of the eye in C. pellucidus is based upon the
examination of three individuals. These were respectively 6.5 cm.,
5.6 cm., and 4.4 cm. long. Figure 2 was drawn from the optic stalk of
the shortest individual. In all essential features the eyes of the two

MUSEUM OF COMPARATIVE ZOOLOGY. 161

other crayfishes presented the same condition as that shown in Figure 2.
In the specimen 5.6 cm. in length, the granular bodies were less dis-
tinct than in the other two, but they were nevertheless recognizable,
and the retinal thickening was as pronounced in this as in either of the
other specimens. The fact that these three individuals show so little
variation leads me to believe that the condition of the eye in the blind
crayfish is not so variable as I at first supposed it would be. The same
constancy is also true of C. setosus. Hence it seems to me improbable
that the differences between Newport’s observation and those of the
later investigators are due to individual variations in the specimens
studied. The fact that Newport’s work was done before the develop-
ment of present methods of research offers, I believe, a more natural
explanation of some of his results, than the supposition of individual
variations. That the methods of his time were imperfect is evident
from the fact that Newport himself seems to have overlooked the gan-
glion of the optic stalk, a structure readily discoverable by means of serial
sections. (Compare Newport’s Figure 13 [’55, p.-102] with Figure 2 in
this paper.) Leydig’s observations, so far as they extend, are fully con-
firmed by my own. Packard’s account differs from mine in only one par-
ticular, but that is of considerable importance ; he states that there is
no retinal thickening in the two species studied by him. This difference
may possibly be due to individual variations in the crayfishes. Unfor-
tunately, Packard does not state the number of specimens which he
examined, and consequently one is uncertain how much weight to give
to his general statements.

The conclusions to be drawn from the foregoing account may be
summarized as follows. In both species of crayfishes studied, the optic
ganglion and nerve are present, and the latter terminates in some way
not discoverable in the hypodermis of the retinal region. In C. setosus
this region is represented only by undifferentiated hypodermis, com-
posed of somewhat crowded cells, while in C. pellucidus it has the form
of a lenticular thickening of the hypodermis, in which there exist multi-
nuclear granulated bodies. These I have endeavored to show are
degenerated clusters of cone-cells. If Packard’s observations are correct,
the retina in C. pellucidus may be reduced in some individuals as much
as it is in C. setosus, which I have studied, but my own examinations
do not render this view probable.

CAMBRIDGE, February 24, 1890.

qa

162° BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

BIBLIOGRAPHY.

Leydig, F.
’83. Untersuchungen zur Anatomie und Histologie der Thiere. Bonn,
Emil Strauss, 1883. 174 pp., 8 Taf. :
Newport, G.
55. On the Ocelliin the Genus Anthophorabia. Trans. Linn. Soc., Lon-
don, Vol. XXI. pp. 161-165, Tab. X., Figs. 10 to 15 incl. Read,
April 19, 1853.
Packard, A. S.
’88. The Cave Fauna of North America, with Remarks on the Anatomy of
the Brain and Origin of the Blind Species. Mem. Nat. Acad. Sci., Vol.
IV. Pt. 1, pp. 1-156, 27 Pls. Read, Nov. 9, 1886.
Parker, G. H.
90. The Histology and Development of the Eye in the Lobster. Bull. Mus.
Comp. Zoél. at Harvard Coll., Vol. XX. No. 1, pp. 1-60, 4 Pls. 1890.
Garman, S.
’89. Cave Animals from Southwestern Missouri. Bull. Mus. Comp. Zodl.
at Harvard Coll., Vol. XVII. No. 6, pp. 225-240, 2 Pls. Dec., 1889.

PARKER, — Blind Cray fishes,

EXPLANATION OF FIGURES.

ABBREVIATIONS.
con. cone. mb. basement membrane.
ct. cuticula. nl. con. nucleus of cone-cell.
gn. opt. optic ganglion. nl. hd. nucleus of hypede
hd. hy podermis. n. opt. optic nerve.
r. retina.

The specimens from which the following figures were taken were killed and
preserved in strong alcohol, and stained in Czocher’s alum-cochineal. The cray-
fish from the optic stalk of which Figure 1 was drawn was 6 cm. long. That from
which the remaining figures were made was 4.4 cm. long.

Pies di:

A longitudinal horizontal section through the right optic stalk of Cam-
barus setosus, Faxon. The histological detail is given in the hy-
podermis only. The optic ganglion and the optic nerve are tinted.
Between these structures and the hypodermis the space is filled with
a loose connective tissue. X 65.

A longitudinal horizontal section through the right optic stalk of Cam-
barus pellucidus, Tellk. This drawing was made in the same manner
as Figure 1. xX 65.

An enlarged drawing from the distal end of the section which immediately
follows that from which Figure 2is taken. This figure shows the details
in the retinal enlargement of the hypodermis. The space between
this enlargement and the cuticula was artificially produced. Xx 275.

Tangential section of the hypodermis from the side of an optic stalk of
Cambarus pellucidus. X 275.

Tangential section of the superficial portion of the retinal thickening in
the eye of Cambarus pellucidus. X 275.

Tangential section of the deep portion in the retinal thickening of the eye
of Cambarus pellucidus. This section is taken from the same series
as the one from which Figure 5 was drawn. X 2785.

BMeisei,lith. Boston.

a

No. 6.— Notice of Calamocrinus Diomede, a new Stalked Crinoid
Jrom the Galapagos, dredged by the U. S. Fish Commission
Steamer “Albatross,” LIEUT.-COMMANDER Z. L. TANNER, U. S. N.,
commanding. By ALEXANDER AGASSIZ.

[Published by Permission of Marsuatt McDona xp, U. S. Fish Commissioner. ]

In 1887, Professor G. Brown Goode, Acting U. 8. Fish Commissioner,
was kind enough to invite me to join the ‘‘ Albatross” at Panama, and
to take part in the dredging operations to be carried on between that
port and the Galapagos Islands.

I always hoped to have the opportunity of comparing, at some time,
the deep-water fauna of the Pacific side of the Isthmus of Panama with
that of the Caribbean, and to see how far the parallelism which has
been traced between the littoral fauna of the two sides was carried out
with the deep-water fauna. Unfortunately, I was unable to avail my-
self of this exceptional opportunity, although Colonel McDonald, the
U.S. Fish Commissioner, detained the “ Albatross ” at Panama to allow
me to join her at the last moment.

To have thoroughly dredged the line from Panama to the Galapagos
would have been to collect material for the solution of many an inter-
esting problem in the geographical distribution of marine animals, to
say nothing of the rich harvest likely to have been gathered, when
dredging in a district so prolific as that of the Bay of Panama, in
shallower waters; and if the haul made at Station No. 2818, off In-
defatigable Island, is at all a measure of what we may obtain in the
way of novelties, the naturalist who is the first to run that line may
be prepared for remarkable discoveries.

In addition to the Stalked Crinoids collected by the “ Albatross,”
which the Fish Commissioner has kindly placed at my disposal for
study, he has also intrusted to me the Kchini collected by the “ Alba-
tross”” on her voyage from the east coast of the United States to San
Francisco. The route she followed was about the same as that taken
by the “ Hassler,” and the material collected differed but little from
the collection made by the latter vessel. The Echini were more nu-

voL xx —wNo. 6.

166 BULLETIN OF THE

merous; but with the exception of the young stages of a few species,
and additional data regarding the geographical distribution of many
species, there were no novelties brought to light. I shall take another
occasion to publish a final report on the Echini.

The ‘ Albatross” dredged on her voyage from New York to San
Francisco, off Indefatigable Island, one of the Galapagos, at a depth of
392 fathoms, three imperfect specimens of a most interesting Stalked
Crinoid. At the first glance, it might readily pass for a living repre-
sentative of the fossil Apiocrinus ; but on closer examination we found
that it revealed some features which ally it with Millericrinus, and
others with Hyocrinus:and Rhizocrinus. It soon became apparent that
we were dealing with a new type, combining structural features of all
the genera above named. It has, like Hyocrinus and Rhizocrinus, only
five arms; they are, however, not simple, but send off from the main
stem of the arm three branches to one side and two to the other.

As in Hyocrinus, the first radials are high, the second radials much
narrower than the first. The system of interradial plates is highly
developed, as in Apiocrinus and Millericrinus, six rows of solid polygonal
imperforate plates being closely joined together, and uniting the arms
into a stiff calyx as far as the sixth or seventh radial, and to the third
or fourth joints of the first and second pinnules. These two pinnules
are on the fourth and fifth radials; the third pinnule is on the sixth
radial ; and they are all below the first axillary, which is the eighth ra-
dial, and which gives rise to the first branch from the main stem. The
second and fifth, sixth, or seventh radials have syzygies.

The imperforated interradials are followed by smaller, somewhat
thinner and perforated perisomatic plates, which extend to the promi-
nent lateral plates of the food groove. The interradial calycinal plates
extend along the arms for a considerable distance beyond the first
branch.

The ventral surface extends nearly horizontally from the mouth to
the level of the seventh radial, and this plane may be considered the
greatest width of the cup, the interradial spaces arching very slightly
toward the mouth, at the junction of the imperforate interradial plates
with the perforated perisomatic plates.

The solid imperforate interradial plates extend over the prominent
anal proboscis. The oral plates at the interradial angles of the food
groove are small, but easily distinguished from the adjacent lateral
and covering plates. They are separated from the so called calyx in-
terradials by three or four rows of perforated perisomatic plates, except

|

MUSEUM OF COMPARATIVE ZOOLOGY. 167

on the anal interradii. The stem was somewhat curved at the upper
extremity, the terminal joints expanding slightly to form a continuation
of the outline of the cup of the base of the calyx. The stem tapered
very gradually, and in its general appearance recalled that of Apiocrinus,
expanding again towards the base, the root of which, however, was not
obtained by the “Albatross.” The stem is cylindrical, without cirri. In
the upper third the joints are alternately ribbed transversely, or even
ornamented near the base of the calyx with more or less prominent
tubercles, as in Millericrinus. The uppermost joint is convex, and in
the space left vacant between it and the central part of the basal ring
a small lobed delicately reticulated pentagonal disk was found resting
upon the upper face of the “article basal” of De Loriol. This is prob-
ably a modified anchylosed infrabasal ring, which may or may not be
resorbed in older stages of the genus.

There are five distinct basals in one of the specimens ; in the second
their sutures can fairly be distinguished, while in the third they were
completely anchylosed, much as they so frequently are in Rhizocrinus.
As in Hyocrinus, the basals are about half the height of the first radials ;
the second radials cut deeply into the first radials.

The stem of this crinoid must have attained a length of from 26 to
27 inches; the height of the calyx to the interradials is 7% of an inch ;
its diameter at the inner base of the second radials is }4 of an inch, at
the height of the third joint of the second pinnule 1 inch, at the level of
the proximal face of the radials 2 of an inch, and at the level of the su-
ture of the basals with the uppermost joint } of an inch; and the length
of the arms is probably about 8 inches. |

I propose to name this crinoid Calamocrinus Diomede, after the ves-
sel which discovered it. I have to thank Colonel Marshall McDonald,
the U. S. Fish Commissioner, for the opportunity of studying this
crinoid. With his consent, a detailed account of Calamocrinus will
be published in the Museum Memoirs as soon as the plates can be
prepared.

CAMBRIDGE, November 28, 1890.

No. 7. — The Origin and Development of the Central Nervous System
in Limax maximus. By ANNIE P. HeNcuMaN.!

For several years the origin of the central nervous system in Mollusks,
both as to method and time of appearance, has been a matter of contro-
versy. It has been of especial importance to determine from which of
the embryonic layers its parts arise, and to ascertain if its development
throws any light on the relations of Mollusks to other important groups
of the animal kingdom, particularly Worms.

Since the observations of the earlier writers, down to about 1874, were
carried on without the aid of sections, their conclusions do not merit
that degree of confidence which is te be accorded those who have availed
themselves of this means of study.

Most of the later authors agree that the central nervous system arises
from the ectoderm, either by an invagination, or by a simple local thick-
ening which later becomes detached. However, Bobretzky (’76, pp. 162-
169), — the first to use sections, — while conceding that in Fusus there
are invaginations of the ectoderm to form the sense organs, concludes
that the supra-cesophageal and pedal ganglia arise from the mesoderm,
and Butschli (77, pp. 227, 228) is inclined to believe that the same is
true in Paludina vivipara.

Von Jhering (74, p. 321) claims for Helix, and both Lankester (’74,
pp- 382, 383) and Wolfson (’80, pp. 95, 96) for Lymneeus stagnalis,
that the central nervous system arises simply from a thickening of the
ectoderm.

Fol (’80, p. 664) has since pointed out, however, that Lankester’s con-
clusions are based on an erroneous interpretation of cells (“nuchal cells”),
which he believes are not at all nervous in their nature. They are the
same cells which Wolfson has called the embryonic brain ; but Wolfson’s
opinion, previously stated, has reference to the definite nervous system,
not to this so-called embryonic brain.

Haddon (’82, pp. 368-370) believes that he has seen the rudiments
of the cerebral and pedal ganglia of Nudibranchs in the form of thicken-

1 Contributions from the Zodlogical Laboratory of the Museum of Comparative
Zoology, under the direction of E. L. Mark, No. XXI.
VOL. XX.—NO. 7.

1'70 BULLETIN OF THE

ings of the ectoderm, and he has made sections of Purpura lapillus’ and
Murex erinaceus which show, as he maintains, that similar rudiments
are also formed in them by proliferation from thickenings of the
ectoderm.

Kowalesky (’83, pp. 23-26) shows for Chiton Polii that the lateral and
pedal nerve trunks are formed simply as thickenings of the ectoderm.

Rabl (’75, pp. 206-208) maintained that the supra-cesophageal gan-
glia and the sense organs in Lymnzus, Physa, Ancylus, and Planorbis
were formed by an invagination of the ectoderm, and that the pedal
ganglia were produced by delamination from the same germinal layer.
He has more recently (’83, pp. 57, 58) expressed doubt as to the man-
ner in which the pedal ganglia arise in Bythinia tentaculata, because he
has seen them so connected to the ectoderm of the dorsal wall of the
foot by means of cells as to indicate that they arise by proliferation
from that region.

Sarasin (’82, pp. 45-48), who has also recently studied Bythinia ten-
taculata, and who is the only author that has hitherto followed the de-
velopment of the entire nervous system in a Gastropod, contends that
the whole of it arises from ectodermic thickenings, without any invagina-
tion even for the supra-cesophageal ganglia. He also believes that the
pedal ganglia arise from the dorsal wall of the foot.

Fol (80, pp. 165-169) admits no invagination for the central nervous
system in aquatic pulmonates, and he even inclines to the opinion that
it may be derived from the mesoderm, which, however, has itself origi-
nated from the ectoderm. He considers it an unimportant question, and
therefore one which it is useless to discuss, whether the nervous system
arises from ectoderm or mesoderm. If the mesoderm were derived from
the entoderm, then it would be an important question. He believes that
the supra-cesophageal ganglia of land pulmonates (pp. 192-195) originate
by invaginations of the ectoderm, while the pedal ganglia arise from the
mesoderm of the foot.

The latest investigations are those of Salensky (86, pp. 655-759) on
the development of Vermetus, one of the Prosobranchs. He concludes
that the cerebral ganglia are formed by two invaginations of the ecto-
derm, while the pedal ganglia arise by proliferation from the ventral and
lateral walls of the foot on each side of the median depression which runs
along its ventral face. These ganglia arise separately, and later become
connected with each other by a commissure, and with the cerebral gan-
glia by connectives, both of which are outgrowths from the ganglia
(pp. 694, 695).

MUSEUM OF COMPARATIVE ZOOLOGY. 17

Thus we find that, of the authors cited, Bobretzky, Biitschli, and
—as far at least as regards the aquatic pulmonates — Fol consider the
central nervous system as originating from the mesoderm. Rabl is a
little doubtful as to its mesodermic origin in Bythinia, Rabl, Fol, and
Salensky are the only investigators who consider any portion of the cen-
tral nervous system as arising by invagination, and then only in certain
Gastropods.*

The following observations were made upon embryos of Limax maxi-
mus obtained from adults kept in captivity. Under favorable circum-
stances, they lay abundantly during the latter part of September, and
through October and November. After numerous trials, the best method
found was to keep about twenty-five or thirty in a large tin pail, the
cover being perforated with small holes. Instead of using moss to se-
cure the necessary moisture, the slugs were fed upon lettuce or cabbage ;
the latter is the better of the two. This food affords at the same time
sufficient protection against desiccation, a suitable retreat for the slugs,
and a place where they may lay the eggs. It should be changed every
other day, — every day if the weather is warm, — and the pail should be
washed thoroughly each time. One of the advantages of using a tin
vessel is the ease with which it may be kept clean. Cabbage will keep
longer than lettuce, and the slugs lay more abundantly when fed upon
it. The eggs were generally found in the morning, sometimes at night,
in bunches of from thirty to forty. They are more abundant at first
than after the slugs have been kept some time in confinement; it is
therefore better to obtain at intervals fresh supplies of small numbers
of slugs than to procure a larger number at one time. As soon as
found, the eggs were removed to a watch-glass containing water; this
was placed in a tumbler already about half filled with moss or moistened
paper, having a perforated tin cover. The eggs must not be allowed
to become dry. For a few days they should be carefully examined under
a microscope, every twenty-four hours or oftener, and all those which fail
to develop should be removed at once. In the course of a few days these
can be readily detected with the naked eye by reason of the greater
opacity of the eggs, and the presence of a whitish spot in them due to
the disintegration of the embryo.

1 The brothers Sarasin, in later researches in Ceylon (’87, pp. 59-69) on a spe-
cies of very large Helix, find that there are two invaginations of the ectoderm on
each side of the head to form the cerebral ganglia, and Kowalesky (’83*) had found

several years before that there were in Dentalium two deep invaginations, one on
each side.

72 BULLETIN OF THE

A very large per cent of eggs kept in this way remain in good condi-
tion until hatching, which, in a moderately warm room, occurs between
the twenty-second and twenty-seventh day.

The best reagents for killing embryos were found to be either chromic
acid, 0.33%, or Perenyi’s fluid. The chromic material when well stained
with alcoholic borax-carmine shows the differentiation of nerve cells and
nuclei excellently, but it is more difficult to stain sufficiently chromic
material than such as has been preserved in Perenyi’s fluid. The latter
may be stained with alcoholic borax-carmine or picrocarminate of lithium,
Good results for the study of cell division have also been obtained by
staining with Czoker’s cochineal. The picrocarminate of lithium is par-
ticularly valuable in the older stages, because it brings out the nerve
fibres, the latter being stained yellow, while the ganglionic cells are
colored red.

To obtain the embryos in an uninjured condition, it is advisable, in
using the chromic-acid method, to remove only the outer envelope
before killing. The egg may be held between the thumb and forefinger
of the left hand, while with a finely pointed stick, somewhat like a
wooden toothpick, the outer membrane is gently punctured ; the probe
should be run under the membrane a little way, to make a larger open-
ing, and the egg carefully pressed with the thumb and forefinger,
whereupon the albumen, containing the embryo and surrounded by the
inner membrane, will come out in a perfect condition. This may be
dropped at once into water, if several are to be treated together, for it
is more convenient to put them all into the chromic acid at the same
time. When all have been shelled, they should be put into 0.33%
chromic acid for two or three minutes only, simply to kill the embryo
without hardening the albumen. Then they should be transferred to a
watch-glass of water, to which a few drops of the acid have been added.
While in this fluid, the inner membrane may be removed with needles.
To accomplish this, it is advisable, in the very young stages, to make as
large an opening in the membrane as possible, and then with a needle
gently to press the embryo out, even if the albumen adheres to it, for
the albumen becomes slightly coagulated in the weak acid, and then
can easily be washed off. In the stages from the tenth to the sixteenth
day, the large size of the pulsating sacs of both head and foot regions
makes it extremely difficult to extract the embryos uninjured; great
care must therefore be taken, and no pressure used, While employing
one of the needles to hold the membrane, the other should be forced
through the membrane, which may then be ruptured and turned back

MUSEUM OF COMPARATIVE ZOOLOGY. 173

over the embryo, being drawn off like the finger of a glove. In the
older stages, not much care is necessary, because the embryos bear with-
out injury considerable handling, and there is so little albumen left that
their position is not readily changed while the membranes are being re-
moved. When freed from the membranes and as much of the albumen
as possible, the embryos are to be returned with a large-mouthed pipette
to the chromic acid (0.33%), where they may be left for an hour or
two; after washing in running water for two or three hours, they may
be carried up to 70% alcohol by adding to the water, drop by drop,
35% alcohol; then 50% alcohol, etc. This dehydration must be made
very carefully, to avoid shrinkage. The embryos are extremely deli-
cate, and must be handled with great care through every step of the
process.

In using Perenyi’s mixture, it is best to free the embryos while lung
from the surrounding membranes and the albumen, removing the inner
membrane under clear water. When set free, they should be trans-
ferred at once with a pipette into a dish of Perenyi’s mixture, where
they may remain from two to three minutes. They are then to be washed
thoroughly in distilled water at least five minutes, put into a 5¢@ aq. sol.
of alum for thirty minutes, washed again in water, and finally car-
ried through the grades of alcohol as in the chromic method. It is
necessary to remove the embryo while living, because otherwise the
albumen becomes in this reagent like a jelly, and cannot be removed
without injury to the embryo. Material designed to be sectioned must
not be left in alcohol longer than a month, since the albumen in the
nutritive sac gradually becomes too hard to be cut, especially if pre-
pared in Perenyi’s mixture. The stages from the tenth to the sixteenth
day can still be used, even if they have been thus overhardened, by re-
moving the nutritive sac; but in the younger stages this is apt to
destroy the embryo, and in the older ones — much of the albumen hay-
ing been swallowed — its removal is still more certain to have the same
effect. Attempts subsequently to soften the albumen by prolonged
treatment with weak acetic acid proved to be only partially successful.
If the embryos are to be kept at all, they should be left unstained ;
but the safest way is to carry them through to embedding as soon as
possible.

They can be stained whole; but to do this successfully, they must
be carried gradually through successively weaker grades of alcohol until
a grade corresponding to the stain is reached. It is advisable to make
the necessary steps from the stain to the parafine as quickly as possible.

174 BULLETIN OF THE

Staining with picrocarminate of lithium has the advantage of saving
time, since it acts rapidly,—the older specimens requiring only one or
two hours, the younger from half an hour to an hour. A few grains of
picric acid may be added to the dehydrating alcohols which follow the
stain, in order to prevent the total extraction of the picric acid, and the
consequent disappearance of the yellow color from the nerve fibres. If
the object is too deeply stained, the differentiation of nerve tissue does
not show well; the nerve fibres ought to be yellow, the surrounding
nuclei pinkish red with a yellow tinge, and all the other tissue pinkish
red. As this and Czoker’s cochineal are both aqueous dyes, the chromic
material is apt to macerate in them; neither does it stain so well in
them as in alcoholic borax-carmine.

The chloroform method of embedding in parafine was used exclusively.
When the embryo has been transferred by the well known method to a
vial containing chioroform, the vial should be placed uncorked on the
water-bath at 55° to 60° C. Pendent spoons in the large cups are not
very serviceable, as the least jar sends the objects off, and it is almost
impossible to recover them from the bottom of the cup without injury.
It-is better to have ready on the bath an empty warm glass dish, —a
common salt-cellar is very good; also one filled with parafine which
melts at about 52° C. The embryos are to be left in the chloroform
only as long a time as is necessary for them to sink, and are then to be
transferred with the chloroform to the empty glass dish. The transfer
is best made by means of a warm pipette, if the embryos are small.
Cold soft parafine is then added, a small piece at a time, until the
chloroform has so thoroughly evaporated as to leave no trace of its odor.
After remaining for fifteen minutes in the soft parafine, the embryo is
to be transferred to the “harder” parafine (52° C.), where it should
remain from fifteen to thirty minutes. It is important to handle the
object with great care, and to carry it through the period of heating as
quickly as may be; the latter is necessary, because the embryos are
very apt to become brittle if subjected to the heat too long. They
should be embedded within an hour or an hour and a half from the
time they are first put upon the bath in the chloroform. It is espe-
cially dangerous to allow the parafine to harden about the embryo be-
fore the latter is finally embedded, because upon the remelting of the
parafine the object is almost certain to fall into fragments, owing to its
great delicacy.

The embedding, especially for the younger stages, must be done un-
deralens. It is most convenient to use a dissecting microscope, the

————— ee SCO

MUSEUM OF COMPARATIVE ZOOLOGY. 175

stage of which should be kept warm. I have found that parafine which
melts between 50° and 52° C. is better for embedding than that which
is harder, for the latter is liable in hardening to cause the embryo to
crack.

Sections from 10 to 15 w thick, and in the oldest stages even thicker,
are better than very thin ones.

The central nervous system of Limax consists of four pairs of gan-
glia, —namely, cerebral, pedal, pleural, and visceral,—-together with
one abdominal ganglion. To these more central ganglia are joined in
addition a pair of buccal ganglia, and one mantle or olfactory ganglion.

To summarize briefly in advance my conclusions: The ganglia arise
separately. The components of three of the five pairs are joined to-
gether later by commissures. Secondarily-produced connectives? also
serve to join the cerebral ganglia to the pedal, the pleural, and the
buccal ; the pleural to the pedal and the visceral; and the visceral to
the abdominal. The growth of the ganglia is rapid; they are well
formed, and in their ultimate positions by the sixteenth day. The prin-
cipal changes from that time until hatching, eight or nine days later,
are increase in size, and modifications of the histological conditions.
According to my observations, all the ganglia, with the possible exception
of the pleural, are derived directly from the ectoderm, — the cerebral in
part from invaginations, the others exclusively by cell proliferation with-
out invagination. The cerebral ganglia are formed by extensive invagi-
nations, one on each side of the head region, just below and behind the
base of the ocular tentacles. During the invagination a rapid cell
proliferation takes place at the deep end of the invaginated portion of
the ectoderm, and also at a region of the ectoderm corresponding to
the depression between the labial tentacles and the upper lips. The
lateral halves of the cerebral mass arise as two separate structures, —
each from a double origin, — which are only secondarily joined. This
union is the result of outgrowths from each of the ganglia which,
uniting, form the cerebral commissure. The invaginations begin a
little later than the proliferation of cells which gives rise to the pedal
ganglia, and they remain open as narrow tubes until towards the period
of hatching, or even later. In one instance they have been found in
this condition as late as eight days after hatching. The cerebral com-

1 In accordance with the usage introduced by Lacaze-Duthiers, the term com-

missure is employed for the nerve fibres joining the components of a pair of
ganglia, and connective for those between ganglia on the same side of the body.

176 BULLETIN OF THE

missure is formed a little earlier than the commissural fibres joining
the pedal ganglia. The latter are connected by two distinct commis-
sures, the anterior of which is formed earlier than the posterior. The
visceral ganglia precede a little in their development the pleural,
abdominal, buccal, and mantle ganglia. The buccal ganglia make their
appearance at about the same time as the pleural, and undergo almost
no change in position.

The nervous system in Limax maximus makes its appearance on the
sexth or seventh day after the ege is laid. At this time the foot is a
conical projection, less than half as long as the diameter of the more
or less spherical remaining portion of the embryo, and its pulsating sac
is very small. It is a stage which is only slightly older than that rep-
resented by Fol in his Planche 17-18, Fig. 7. The ocular tentacles
are now distinguishable as small elevations of the head region, near
the beginning of the primitive nephridial organs, but the labial tenta-
cles are barely to be made out. The radula sac is a nearly spherical
outfolding of the floor of the oral sinus ; its fundus is composed of only
a single layer of cells, but the part of the sac which is continuous with
the wall of the cesophagus is more than a single cell deep ; the lumen
of the oesophagus is traceable close up to the yolk, where it ends
blindly. Both the cesophagus and the radula sac are covered with a
continuous layer of somewhat flattened mesodermic cells. ‘The shell
gland has the form of a large thin-walled sac containing concretions.

When this condition has been reached, the head region (Plate I.
Fig. 2) exhibits no sign of cerebral invaginations, nor have I been able
to find regions of cell proliferation or thickenings in the ectoderm which
were referable with certainty to the cerebral ganglia.

So far as I have been able to make out, the first contribution to the
formation of the pedal ganglia occurs in the form of small clusters of
cells, which are still imbedded-in the ectoderm of the ventral wall of the
foot (Plate I. Fig. 5), from which they are subsequently detached.
Each of these clusters has a spheroidal or more ridge-like form, and con-
tains from four to eight cells. The boundaries of the cells are not
sharply marked, but the whole cluster is limited by a definite outline
separating it from the rest of the ectoderm. Each cell contains a nu-
cleus, which is large, but less deeply stained than those of the ectoderm,
and each nucleus has a large nucleolus, which is very deeply stained
(Plate I. Fig. 1).

The region in which this proliferation takes place is definitely located,

4
4
a. oe.

— SS — -

MUSEUM OF COMPARATIVE ZOOLOGY. dW ya

for it lies in the same transverse plane in which the otocysts (Plate I.
Figs. 3, 4) are situated, and it is found at a region in that plane inter-
mediate between the lateral border of the foot and its middle line, but
considerably nearer the former. The proliferating cells project into the
cavity of the foot, and ultimately are separated from the ectoderm.

Although cells which closely resemble these are found in groups in
other parts of the body wall, their nuclei do not become as large as
those of the cells destined to form the ganglia. Moreover, the prolifer-
ations are constant and most abundant in the regions where the differ.
ent ganglia of the nervous system take their origin. Besides, in these
cases there is generally a sinking in of the surface of the ectoderm in
the same region.

Somewhat later than at the stage described, usually on the seventh
day, the external conditions still remain nearly the same, the ocular
tentacles being perhaps a little more prominent, and the concretions in
the shell gland more numerous.

The cells of the primitive entoderm, which surround the yolk, form a
striking feature of the condition at this stage. These entoderm cells
are very large. vacuolated, and only slightly stainable. They contain
large ovoid nuclei, which are crowded to one margin of the cells by the
nutritive contents accumulated in the cells. Each nucleus contains one
large deeply stained nucleolus, and a network of chromatic substance
(Plate I. Fig. 2). The ectoderm, except over the nutritive sac, consists
of elongated cells, whose nuclei are so arranged as to give the appear-
ance of two or more layers. The ganglionic cells at this time closely
resemble the mesodermic cells, and this makes it difficult to distinguish
between the two (Plate I. Fig. 7).

The internal ends of the primitive nephridial organs are situated one
on each side of the head, immediately above and back of the ocular tenta-
cles. These organs pass at first forwards and upwards, then in an arch
backwards over the nutritive sac, and finally downward and forward.
Their external openings are far back in the lateral walls of the body,
behind the head region. The organs are readily distinguished in sec-
tions by their large slightly stained cells, which are arranged in a single
layer around an oval lumen. The large nuclei contain each a single
deeply stained nucleolus (Plate I. Figs. 2 and 6). The primitive en-
toderm and the nephridial organs retain this histological condition
throughout the embryonic stages.

The cerebral invaginations at first appear as shallow depressions in

VOL. XX.— NO, 7. 12

178 BULLETIN OF THE

the ectoderm at the base of the ocular tentacles, at a point immediately
below the nephridial organs. At the same time that the infolding takes
place, cells, whose nuclei are larger than those of the mesodermic cells,
are being proliferated from the deep surface of the invaginating portion
of the ectoderm (Plate I. Fig. 6).

In the region of the ventral wall of the foot referred to in the stage
previously described, there are in the ectoderm of each side of the body
two groups of ganglionic cells (prf. pd., Plate I. Fig. 7), one behind the
other. These cells project into the cavity of the foot, and reach nearly
to another small group of cells situated not far from the ventral wall.
The cells of the latter group (there is one group on each side of the
body) have nuclei similar to those of the cells still connected with the
ectoderm. Each group lies in the position subsequently occupied by the
pedal ganglion of its side of the body, and is undoubtedly the beginning
of that ganglion, for the cells in the ventral wall of the foot continue to
be proliferated during several days, and are found in some individuals
to be in direct continuity with the ganglia after the latter have at-
tained considerable size. In the individuals shown in Figures 7 and 9
(Plate I.), the right otocyst (Fig. 7) is seen as a closed vesicle, which is
not yet wholly detached from the ectoderm. The otocysts undoubtedly
vary in regard to the time of their detachment, as will be seen by a
glance at the left otocyst of the same individual, which has entirely
lost its connection with the ectoderm (Tig. 9).

All the other ganglia, with the exception of the one near the olfactory
organ and the buccal ganglia, arise by cell proliferation from ectoderm
which lies between the foot and the head region, either at or a little
above the posterior angle formed by the body wall with the dorsal sur-
face of the foot, or along a depression which runs forward from this
point. This angle marks the posterior limit of a furrow which passes
obliquely forward and downward, partially separating the head and
visceral mass from the foot. This depression will be designated as
the pleural groove. Of the remaining ganglia, only the visceral have
begun to be formed at this time. The cells destined to form these
ganglia are situated immediately above the angle produced by the
pleural groove (Plate I. Figs. 8 and 9). Some of those of the left
ganglion are wholly detached from the ectoderm, but those of the
right (Fig. 7) are still continuous with the ectoderm, though pro-
jecting into the body cavity. The cells have large, round, faintly
stainable nuclei, each containing one large nucleolus, which takes a
deep stain.

MUSEUM OF COMPARATIVE ZOOLOGY. 179

Twenty-four hours later, about the eighth day, the pulsating sac of the
foot has become still larger, and the oral sinus has extended backward and
downward as a very narrow tubula» passage, — the cesophagus, — which
follows the surface of the nutritive sac for some distance, and subse-
quently opens into it. The peculiar ciliated cells of great size and spongy
appearance, which occupy a linear tract along the middle of the roof of
the mouth and cesophagus, are at this time very prominent (Plate I.
Fig. 2, loph. cil.). These cells form what Fol (’80, pp. 190, 191) has
called the “ciliated ridge.” ‘They persist until after the completion of
the nervous system. The ingrowth of the ectoderm to form the rectum
is now composed of a compact group of small cells, which shows a small
lumen in its central portion, but is still closed at both ends.

The cerebral ganglia remain in nearly the same condition as that last
described. About twelve hours later, between the eighth and ninth days,

Ficury A. — The right face of a section parallel to the sagittal plane from an embryo of the
ninth day xX 220.

gn. pd. Pedal ganglion ocy.s. Left otocyst.

the two cerebral invaginations have become deeper, and the two groups
of cells which form the main portions of the corresponding ganglia con-
tain a greater number of cells. (Plate IT. Fig. 15.)

The pedal ganglia are also now composed of many more cells than in
the previous stage. Each ganglion is usually pear-shaped, and tapers to-
wards the posterior end of the foot. They both continue to receive ac-
cessions from the ectoderm (Figure A), and at the same time are rapidly
increasing in size by division of the cells already in position. The nuclei
are larger and more easily distinguished than in the previous stage from
those of the mesodermic cells, the latter being more spindle-shaped than

180 BULLETIN OF THE

before. The cells of the mesoderm form a continuous layer along the
inner surface of the ectoderm, except where cell proliferation is taking
place (Figure A).

As yet nothing is to be seen of the pleural ganglia.

The visceral ganglia have increased in size (Plate I. Figs. 10, 11, 12);
they are still connected with the ectoderm (Figs. 10, 12), although a few
cells with large nuclei have become detached from it (Fig. 11). The
ganglion and the otocyst of the same side of the body lie in nearly the
same sagittal plane. Each ganglion is situated just above the angle
caused by the pleural groove. The right visceral ganglion (Fig. 10) is
somewhat farther forward and more dorsal than the left (Figs. 11, 12).

About in the median plane of the body, and above the angle made by
the pleural groove, are the cells which form the abdominal ganglion
(Plate I. Fig. 13). The greater part of them are still embedded in the
ectoderm. Although in some regions they project into the body cavity,
they are nowhere wholly separated from the ectoderm. The abdominal
ganglion seems to be at first more intimately connected with the left
visceral ganglion than with the right, but a connective is formed with
both of them a little later, and the abdominal ganglion thus appears to
occupy the place of a direct commissure between the two visceral ganglia.
As development proceeds, the abdominal ganglion becomes closely fused
with both the visceral ganglia.

Quite an advance in external conditions is made by the nenth day.
But individuals of the same age vary so much in the degree of develop-
ment attained by both their external and internal organs, that the age
assigned can be taken only as an approximation to the average condition
at the time indicated.

The tentacles appear as protuberances, the labial tentacles being much
smaller than the ocular; the shell gland contains more concretions, the
mantle is larger and bends backward over the dorsal surface of the foot.
The radula sac makes its appearance and extends backward into the foot,
where it ends blindly immediately back of the pedal ganglia. In trans-
verse sections it appears flattened dorso-ventrally ; its lumen is oval, and
the ectoderm lining it is more than one cell deep.

The cerebral invaginations (Plate IT. Figs. 15, 19, Plate III. Figs. 25,
26) are much deeper, the infolding ectoderm is greatly thickened, and
the incipient ganglia receive accessions from ectodermic depressions be-
tween the rudiments of the upper lips and the labial tentacles (Plate II.
Fig. 21). The cerebral commissure (Fig. 21) is also being formed, the

MUSEUM OF COMPARATIVE ZOOLOGY. 181

cells of the median portion of each ganglion growing out to meet the
corresponding cells from the opposite ganglion, The commissure at this
stage is composed of a small number of cells, which are very much elon-
gated. The fibres resulting from their elongation already make a con-
tinuous bridge from one ganglion to the other.

The pedal ganglia (Plate II. Fig. 20, Plate III. Fig. 27, Plate V. Fig.
60) consist of two small groups of cells, situated about midway be-
tween the sole of the foot and the posterior end of the radula sac. They
are a little below and behind the pleural groove and the otocysts, and
they are farther from each other than from the lateral wall of the foot.
There is a slight indication of a commissure (Plate III. Fig. 27) joining
their anterior portions to each other. The commissure is formed in the
same manner as the cerebral commissure, the individual cells composing
it being spindle-shaped, with their nuclei somewhat elongated in the
direction of the fibres.

The otocysts (Plate II. Fig. 20, Plate III. Fig. 27, Plate V. Fig. 60)
are on a level with the lower margin of the radula sac, and are nearer the
pedal ganglia than in the preceding stage.

On each side of the body above the pleural groove is a group of a
few cells, which are in all probability the first indications of the pleural
ganglia (Plate II. Figs. 14 and 20). The centre of each cluster is seen
on cross sections (Fig. 20) to be nearly on a level with the lumen of the
radula sac. The cells at this stage are very small, and so loosely associ-
ated that it is difficult to distinguish them from mesodermic cells. I
have not satisfactory evidence of their origin directly from the ectoderm,
for, although I have found them at times very near to the ectoderm
(Fig. 20), I have never found them at any stage continuous with it. On
the other hand, [I have not seen conditions which would warrant the
conclusion that the ganglia were the result of outgrowths from either of
the pre-existing ganglia.

A little before the ninth day the cells detached from the ectoderm
to form the visceral ganglia (Plate II. Figs. 17, 18) increase rapidly in
size, and the diameter of their nuclei often becomes four or five times as
great as that of the ectodermic nuclei. The ganglia consist of elongated
groups of such cells, still attached to the ectoderm above the pleural
groove (Figs. 16, 18). The want of symmetry in the positions of the
right and left ganglia is more conspicuous than in the preceding stage,
the ganglion of the right side being considerably more dorsal and far-
ther back than that of the left side (Plate IT. Fig. 23, Plate V. Fig. 60).
Owing to the infolding of the ectoderm on the right side of the body to

182 BULLETIN OF THE

form the respiratory chamber (Plate II. Figs. 16, 24), the region from
which the ganglionic cells arise is now located on the ventral and median
walls of the infolding. The ganglia have also grown forward, and lie
between the nephridial organs and the nutritive sac (Plate II. Fig. 23).
In an individual cut crosswise, the posterior portion of the ganglia is
found to be two or three sections back of the otocysts. Both the
ganglia may be traced through five or six sections.

A little behind the visceral ganglia, and to the left of the median
plane of the body, are the prominent cells of the abdominal ganglion
(Plate II. Figs. 24, ab.).

All of these ganglia still consist of groups of loosely associated cells.
Later they become more compact, and are surrounded by connective-
tissue cells. ’ |

The buccal ganglia (Plate II. Fig. 22), first seen with certainty at
this stage, arise, one on each side of the radula sac, at the angle between
it and the cesophagus. It is to be seen from cross sections that the cell
proliferations from which they spring take place from the dorsal wall
of the neck of the sac, where its lumen begins to be separated from that
of the esophagus. This is also their permanent position; they are later
joined together by a commissure, which results from outgrowths of the
cells composing the two ganglia.

On the tenth day the external appearance of the embryo remains
nearly the same as before, with the exception that there is an increase
in the size of the embryo, and especially of its pulsating sacs. The sac
of the radula has become more elongated, and the anal opening (Plate
III. Fig. 31, an.) is formed. :

The cerebral invaginations still appear, in sections parallel to the
sagittal plane (Plate III. Figs. 28 and 29), as shallow depressions. The
number of cells in each ganglionic group (Plate IV. Fig. 58, Plate V.
Fig. 63) has increased perceptibly. At the same time the groups have
extended backward, and show indications of the cerebro-pleural connect-
ives. In specimens cut in the sagittal plane, the cerebral commissure
cut crosswise may be seen above the oral opening (Plate III. Fig. 30).

The pedal ganglia (Plate IV. Figs. 54, 58, Plate V. Fig. 63) have in-
creased in size. Their anterior borders now reach as far forward as the
plane of the pleural groove, and they extend backward into the foot
much farther than before. In cross sections (Plate IV. Fig. 54) they
appear as rounded groups of cells, which are far apart and not yet very
compact ; they still continue to receive accessions by the proliferation of

MUSEUM OF COMPARATIVE ZOOLOGY. 183

ectodermic cells from the walls of the foot (Plate IV. Fig. 57, 58, pr/.).
The first decided evidence of a pedal commissure makes its appearance
during this stage. It consists (Fig. 54) of a few very much elongated
nerve cells, which stretch across from one ganglion to the other a little
posterior to the region of the otocysts. The commissure may be traced
on about half a dozen successive sections, or for a distance of some 50
or 60 ». From its position it evidently is the beginning of the anterior
commissure. ‘The thickness (10 ») of a single section contains only
three or four cells, the nuclei of which have the chromatic substance
so concentrated into a single nucleolus as to make the nuclei appear
clearer than those of the surrounding connective-tissue cells. There
is at present no trace of a posterior commissure. ‘The otocysts are now
nearer the ganglia (Plate IV. Fig. 58, Plate V. Fig. 63) than at any
previous stage.

The pleural ganglia (Plate V. Fig. 63) are still inconspicuous, being
composed of only a few scattered cells, which lie nearly dorsal to the
otocysts, about midway between the visceral and the cerebral ganglia
of the same side of the body. Many of the cells are elongated in the
direction of the ganglia between which they are located, and appear to
form the beginning of a connective between them.

The visceral ganglia (Plate IV. Figs. 58 and 59, usc.) are still con-
nected with the ectoderm, but project more prominently from the wall
of the body, and extend forward more than before. The right (Plate IV.
Fig. 59) is larger, and still lies more dorsal, than the left (Plate V.
Fig. 63). The cells which compose the ganglia are numerous and
large, and the nuclei of those which form the centre of the ganglion are
conspicuously larger than those at the periphery. In cross sections of
a stage possibly a little less developed than the one last described, the
ganglia (Plate IV. Figs. 53, 56, 57, 55) lie, one on each side of the
body, immediately above the pleural groove, a little below and inside
the external orifices of the primitive nephridial organs. On the right
side of the body the ectoderm which constitutes the anterior wall of
the infolding to form the mantle chamber is seen in sagittal sections
(Plate 1V. Fig. 58) to be much thicker in the region adjoining the pleu-
ral groove than in that which forms the deeper portion of the infold-
ing. The transition from the thick to the thin ectoderm is very abrupt,
and is marked by a pocket-like depression. The right visceral ganglion
is situated at the side and in front of this depression. Some of the cells
in the anterior portion of this ganglion (Plate IV. Fig. 56) are traceable
toward the median plane of the body. The left visceral ganglion

184 BULLETIN OF THE

(Figs. 56, 57) is not yet as large as the right, and it consists of fewer
cells.

The position of the connective between the visceral and pleural gan-
glia (Plate V. Fig. 63) is indicated by the presence of spindle-shaped
cells with fibrous projections. The connective is at this time long, and
the cells and fibres composing it are only joined to one another loosely.

As the abdominal ganglion increases in size, it extends more toward
the right side of the body (Plate V. Fig. 61), and the connective be-

Figure B. — The left face of a section parallel to the sagittal plane from an embryo of the
eleventh day. X 73.

ab.-vsc. s. Left abdominal-visceral connective. pd. Pedal ganglion.
ced. s. Left cerebral ganglion. pes. Foot. ;
ench, Shell gland. sac. rad. Radula sac.
0 cy. Otocyst. ta. Ocular tentacle.

tween it and the right visceral ganglion, which is hardly perceptible at
this stage, is much shorter than that to the left visceral ganglion.

The buccal ganglia remain in the same condition as in the preceding
stage (Plate II. Fig. 22).

By the eleventh day the embryo has increased greatly in size (Figure B) ;
the tentacles are prominent, and the pulsating sac of the foot is very
large. A narrow slit-like infolding of the ectoderm (compare Plate VIII.
Fig. 101, gl. pd.) has arisen in the median plane of the body at the an-
terior end of the foot, into which it extends backward a short distance.
It is the beginning of the foot gland. The salivary glands also make

MUSEUM OF COMPARATIVE ZOOLOGY. 185

their appearance during this stage as a pair of evaginations of the
lateral walls of the esophagus, immediately above its communication
with the radula sac, and a little in front of the buccal ganglia (Plate
VI. Figs. 77-80).

The cerebral invaginations still open broadly at the sides of the head
(Plate III. Figs. 32-34, and Figure C). They are, however, quite deep,
and in a series of sagittal sections the depression becomes deeper and
deeper as one approaches the median plane, and at the same time the
orifice which leads to the depression becomes narrower and narrower,

-- gud. plu.

Fiaure C. — The posterior face of a transverse section from an embryo of the eleventh day. x 73.

ab. Abdominal ganglion. mt. Mantle.
cav. mt. Mantle cavity rad. Radula sac.
ench, Shell gland. sul. plu. Pleural groove,

iw. ceb. dz. Right cerebral invagination.

until it is almost slit-like (Figs. 32-40). The deep ends of the invagi-
nation are turned a little towards the median plane. These invagi-
nated portions of the brain are composed of small, closely packed cells,
whose nuclei stain deeply. The proliferated portions of the cerebral
ganglia, which are deeper than the sacs (Plate V. Fig. 64, Plate VI.
Figs. 70, 71), extend toward each other in the median plane, and back-
ward and downward toward the pedal ganglia (Fig. 71). They have
now become differentiated into a fibrous central part (Fig. 71), in which

186 BULLETIN OF THE

are lodged the larger scattered cells with their very large nuclei,-and a
peripheral part, where the cells are crowded together and the nuclei
are smaller (Fig. 70). They are loosely enveloped by spindle-shaped,
very much elongated, connective-tissue cells (Fig. 71). Immediately
above the oral cavity is the cerebral commissure (Plate VI. Fig. 80°).
It can be traced from one side of the brain to the other, and its cross
section appears as a very small round patch of fibrous substance, sur-
rounded on the dorsal side by a layer of flat cells.

The cerebro-pedal connectives are indicated (Fig. 71) by a few cells
extending from the ventral-posterior ends of the cerebral ganglia to the
anterior ends of the pedal, a little in front of the cerebro-pleural con-
nectives (Fig. 70). The latter extend from the posterior ends of the
cerebral to the anterior ends of the pleural ganglia, thus diverging
somewhat from the cerebro-pedal connectives. There are found in the
ganglia many cells which are in different stages of division. It is owing
to this cell division that the ganglia increase rapidly in size, especially
after they are wholly cut off from the ectoderm ; cells in the commis-
sures and connectives are also found in process of dividing in planes
perpendicular to the direction of their fibres.

The principal change in the pedal ganglia (Plate VI. Fig. 71) is due
to an increase in size, particularly in the antero-posterior direction. The
central portion of these ganglia has the same fibrous appearance as that
described for the cerebral ganglia, and the pedal nerves can be traced for
a considerable distance toward the tip of the foot (compare Figure E,
page 191). The anterior commissure (Plate III. Fig. 44, Plate VI.
Fig. 74) is now somewhat shorter than in the previous stage, and con-
sists of a greater number of cells. Cell proliferation is still taking
place from the ectoderm of the ventral wall of the foot (Plate VI. Fig.
71), and the ganglia continue to receive accessions from these sources.
More highly magnified views of the regions of proliferation are given in
Plate VI. Figs. 72 and 73.

The pleural ganglia (Plate VI. Fig. 70) are now easily recognized.
Each ganglion is formed of a triangular group of cells, occupying a posi-
tion immediately above and anterior to that part of the pleural groove
which is nearest to the otocyst. The cells composing the ganglion are
fewer than those of any of the other pairs of ganglia, but resemble them
in their histological conditions; they are only loosely connected, and
their fibres are elongated in the directions of the three connectives. At
this stage the ganglia are not closely enveloped in connective tissue.

The pleuro-visceral connectives are well developed, especially the left

MUSEUM OF COMPARATIVE ZOOLOGY. 187

one (Fig. 70) ; the right one is much longer and more attenuated, since
the right visceral ganglion is farther from the pleural than the left vis-
ceral. The ganglia are most distinctly seen in specimens cut in a
sagittal direction.

The visceral ganglia (Plate V. Figs. 67-69, Plate VI. Fig. 70) are
much larger and more elongated in the direction of the pleural ganglia —
i, e. downward, forward, and outward — than they were during the pre-
vious stage. They are still connected with the ectoderm at their pos-
terior dorsal ends, while the opposite ends are much drawn out toward
the pleural ganglia (Figs. 69, 70). The right visceral ganglion (Figs.
67-69) is larger than the left, and its longest axis has a dorso-ventral
direction (Fig. 68). The fibrous prolongations continue into the pleuro-
visceral connectives (Fig. 71).

The abdominal ganglion (Plate III. Figs. 43, 44, 46,47, Plate VI.
Figs. 75, 76), although still connected with the ectoderm, is also larger,
and projects more into the body cavity than on the tenth day. A large
portion of it still lies to the left of the median plane of the body (Plate
VI. Figs. 75, 76), and the connective to the left visceral is well devel-
oped (Plate III. Figs. 41, 42, Plate V. Fig. 68); that to the right is
less complete (Plate III. Figs. 45, 51).

The buccal ganglia (Plate V. Fig. 62, Plate VI. Fig. 77) are now
very distinct ; the dorsal wall of the radula sac still contributes to their
increase in size.

Cell proliferation takes place from the ectoderm bordering the en-
trance to the respiratory cavity. A few cells, which probably form the
olfactory ganglion, are seen at this stage to be separating from the ecto-
derm in this region.

For the next twenty-four to thirty-six hours (twelfth and thirteenth
days) the external appearance of the embryo remains nearly the same
as on the eleventh day. In the living embryo the larval heart may be
seen pulsating, and the foot gland extends somewhat farther towards the
posterior extremity of the foot.

The cerebral invaginations appear simply as long narrow sacs filled
with a coagulated substance; the inner ends of these sacs have grown
upward as well as backward (Plate VII. Fig. 94). The proliferated
portions of the cerebral ganglia (Fig. 94) are much larger, and have
now assumed more nearly their ultimate positions (Plate [II. Figs. 48,
49; Plate VII. Figs. 81, 82, 94). The central portion of each has
become more fibrous (Fig. 81).

188 BULLETIN OF THE

The connectives, both to the pedal (Plate VII. Fig. 81) and to the
pleural ganglia (Plate III. Fig. 48), are well developed, and are both
thicker and shorter than in the stage last described.

The pedal ganglia do not differ materially from the condition de-
scribed for the eleventh day. The anterior end has increased in diam-
eter, and has grown a little farther forward (Plate III. Fig. 50, Plate
Vila Bigs.8h901):

Both commissures are now present; the anterior (Fig. 92) is a little
behind the otocysts (compare Fig. 92 with Fig. 91), and the posterior
(Fig. 90) is directly above the blind end of the foot gland, and about
0.2 mm. back of the anterior commissure.

The pleural ganglia (Plate III. Fig. 48, Plate VII. Figs. 82, 83, 88)
are very near the cerebral ganglia, as may readily be seen in sagittal
sections (Figs. 48, 82), and the fibrous connectives to the other ganglia
are plainly to be distinguished. The ganglia have become more com-
pact and rounded, and occupy a position nearer the middle plane of the
body (Figs. 86, 88).

The visceral ganglia (Plate III. Fig. 49; Plate VII. Figs. 83, 84,
86-89), although they have increased greatly in size, are still connected
with the ectoderm which forms the anterior wall of the mantle chamber
(Figs. 88, 89).

They have also moved inward and forward. The right ganglion
(Figs. 49, 83, 87-89) is especially well developed, and much farther
forward than in the previous stage. Its axis is prolonged into a nerve,
which runs upward and backward, probably to the olfactory ganglion
(Figs. 84, 87).

The connective from the right visceral to the abdominal ganglion
passes backward and inward (Plate VII. Figs. 83, 84). Where the
connective leaves the visceral ganglion (Fig. 83), the nuclei of the
ganglionic cells are very large, and the fibres are very much elongated
in the direction of the connective.

In specimens cut crosswise the nerve which forms the dorsal prolon-
gation of the axis of the visceral ganglion is found far forward, in front
of the anterior face of the abdominal ganglion; it passes upward and
inward (Plate VII. Figs. 87, 88), and is connected with the ectoderm
that forms the wall of the small infolding from the respiratory cavity
(Fig. 88) referred to in the account of the tenth day. This region is at
the same level as that with which the abdominal ganglion is connected
farther back (Plate VII. Fig. 93). The ectodermic cells to which this
nerve is distributed form the lining to an irregular infolding from the

MUSEUM OF COMPARATIVE ZOOLOGY. 189

median face of the respiratory cavity, and the lumen of the infolding
connects by a narrow orifice with the respiratory chamber (Fig. 88,
can. wa.): I believe this is the organ first described by Lacaze-
Duthiers.

A little farther forward the right visceral ganglion sends to the right
side of the body a nerve (Plate VII. Fig. 89 ».), which passes between
the wall of the mantle chamber and the primitive sexual duct, probably
to be distributed to the right half of the mantle.

At this time the greater portion of the abdominal ganglion (Plate VII.
Figs. 81, 82, 85, 86, 93) lies on the right side of the median plane, al-
though it is joined to the left visceral by a large and prominent connect-
ive (Plate ITI. Figs. 50, 52, Plate V. Figs. 65, 66, Plate VII. Fig. 93).
Since the visceral ganglia have grown inward and forward, the ab-
dominal ganglion now occupies a position considerably posterior to
them (Plate VII. Figs. 83, 86); it lies above the right side of the radula
sac. Its posterior dorsal margin is still continuous with the ectoderm
of the wall of the respiratory cavity (Fig. 93), but farther forward it is
entirely separated from the ectoderm (Fig. 85), and is surrounded by a
layer of connective-tissue cells. All the other ganglia are similarly en-
veloped in connective tissue except where they are continuous with the
ectoderm.

The connective to the left visceral ganglion (Plate VII. Fig. 93) passes
downward, forward, and outward to the left side above the radula sac.

The buccal ganglia (Plate VII. Fig. 81) are larger than on the tenth
day, but are closely applied, as before, to the walls of the radula sac.
Their commissure (Plate V. Fig. 65) is embraced in the angle between
the cesophagus and the neck of the radula sac, and in sagittal sections
presents a circular outline.

On the fourteenth day the upper lips as well as both pairs of tentacles
are very prominent, and the foot gland has grown backward still farther
into the foot (Plate VIII. Fig. 102). The salivary glands have now be-
come elongated into tubular organs with a circular lumen and _ thick
walls consisting of a single layer of epithelial cells (Plate VIII. Fig. 106).
They reach a little farther back than the buccal commissure ; in passing
forward they lie on either side of the cesophagus, about on a level with
its lower border. They pass along the dorsal side of the buccal ganglia,
and then suddenly bend downward to open into the cesophagus.

The cerebral invaginations (Plate VILI. Fig. 96) present the same gen-
eral appearance as in the stage last described, but the lumen of the sacs

190 BULLETIN OF THE

is smaller (Plate X. Figs. 121, 126) ; in cross sections (Fig. D) it ap-
pears oval. The walls are thick, being composed of spindle-shaped cells
arranged perpendicularly to the axis of the sac and so crowded that the
nuclei are three or four deep.

The proliferated portion of the cerebral ganglia (Plate IX. Fig. 114) re-
tains its pear-shaped condition, but is shorter and thicker. A ventral and

FicureE D. — Posterior face of a transverse section from an embryo of the fourteenth day. X 190.

cnch. Shell gland. nph. Nephridial organ.

Lie Heart. @. (&sophagus.

gn. ceb. Cerebral ganglion. pi. cr. Pericardium,

hp. Liver. plu. Pleural ganglion.

hp. dz. Right lobe of liver. sac. rad. Radula sac,

in, Intestine. ta.dz. Right ocular tentacle.
iv. ceb. Cerebral invagination. ta’. Labial tentacle.

mt. Mantle. VSC. Visceral ganglion.

median portion of each ganglion forms a small rounded lobe (Figure E).
These lobes are near the bases of the upper lips, and in sagittal sections
appear almost completely separated from the larger part of the ganglia
by ingrowths of connective tissue. It is from these lobes that the pedal
connectives arise. The connectives to the pleural ganglia emerge from

MUSEUM OF COMPARATIVE ZOOLOGY. 191

the larger portion of the ganglion; they are thicker and shorter than
the cerebro-pedal connectives, from which they are separated by only a
narrow space.

The cerebral commissure is much shorter than before (Plate X.
Fig. 126), but it has not increased much in thickness (Plate VIII.
Fig. 101). In sagittal sections it is seen to be composed of a central
portion made up of nerve fibres cut crosswise and a peripheral layer of
nuclei; but the nuclei are wanting on the face of the commissure which
is in contact with the dorsal wall of the cesophagus.

Fiaure E. — The left surface of a section parallel to the sagittal plane from an embryo of the
fourteenth day. 73.

art. pd. Pedal artery. plu.-pd. Pleuro-pedal connective.
ceb, dx. Right cerebral ganglion. plu.-vsc Pleuro-visceral connective.
ceb.-pd. Cerebro-pedal connective. pd.dzx. Right pedal ganglion.

cnch, Shell gland. rt. Rectum.

lab. Lip. ta. Ocular tentacle.

Nn. Nerve. ta’. Labial tentacle

0 cy. Otocyst. USC. Visceral ganglion.

plu. Pleural ganglion.

The pedal ganglia (Plate VIII. Figs. 97-100, Plate IX. Figs. 114,
118, 119) lie between the radula sac, which is above, and the foot gland
which is below them. They are nearer together than on the twelfth
day, and their anterior ends are more rounded (Fig. 114). Their pos-

192 BULLETIN OF THE

terior ends are elongated and continued as two large nerves far back. into
the foot. In specimens cut crosswise these nerves appear as rounded
patches of fibres, situated one on each side of the body, above the plane
of the foot gland and about midway between it and the lateral walls of
the foot. Each is surrounded by a layer of connective-tissue cells. As
one approaches the pedal ganglia in passing from behind forward, the
nerves increase in size and lie nearer to each other. In the region of
the posterior commissure (Plate IX. Fig. 119) the ganglia are nearly as
broad as in the region of the anterior commissure (Fig. 118), but they
are not much more than half as thick in the dorso-ventral direction. In
front of the posterior commissure they are separated by a narrow space,
which is wider behind than in front, where it is terminated by the ante-
rior commissure. The commissures are both well developed (Plate VIII.
Figs. 101, 102, Plate IX. Figs. 118, 119), and owing to the approxima-
tion of the ganglia have become shorter than in the last stage. The
nuclei in the region of the posterior commissure (Fig. 119) are of nearly
uniform size; but in front of it each ganglion (Figs. 114, 118) contains
a fibrous central portion immediately surrounded by the greatly enlarged
nuclei of cells which form the most of the fibrous substance.

The pleural ganglia (Plate VIII. Fig. 106, Plate IX. Figs, 114, 116,
Plate X. Figs. 123, 125, and Fig. E) have increased considerably in size,
and are more compact. They have moved downward and inward ; and
each now lies in contact with the posterior face of the corresponding
cerebral mass (Plate IX. Fig. 114), and below and in front of the ven-
tral portion of the corresponding visceral ganglion (Figs. 106, 123,
125). They are much smaller than either the cerebral or visceral
ganglia. The nuclei of their central cells are, as in the pedal ganglia,
much enlarged.

The visceral ganglia (Plate VII. Fig. 95, Plate IX. Fig. 114, Plate
X. Figs. 123, 125) are now entirely detached from the ectoderm, and
have moved downward, forward, and inward.

The left ganglion (Plate VIII. Fig. 106, Plate X. Fig. 125) is smaller
than the right, and more closely connected with the left pleural (Fig.
125) than in the previous stage. Its dorsal surface is slightly above
the level of the dorsal wall of the radula sac, and its connective with
the abdominal ganglion (Plate VIII. Fig. 104) is much broader than
before. The right visceral ganglion (Plate VII. Fig. 95, Plate VIII.
Figs. 102 and 106, Plate IX. Fig. 114, Plate X. Fig. 123) is much
larger than in the last stage ; it is also closely connected with the right
pleural ganglion (Fig. 123). It extends dorsally much farther than the

MUSEUM OF COMPARATIVE ZOOLOGY. 193

left visceral, and also nearer to the median plane (Plate VII. Fig. 95,
Plate VIII. Fig. 102, Plate IX. Fig. 120). It is in contact with the
lower surface of the right end of the abdominal ganglion (Plate VII.
Fig. 95).

The abdominal ganglion (Plate VII. Fig. 95, Plate VIII. Figs. 101,
102, 104, Plate IX. Figs. 115-117, Plate X. Fig. 123) is entirely un-
connected with the ectoderm, and has moved forward, so that there is a
considerable space between it and the pleural groove, but its posterior
face extends farther back than that of the right visceral ganglion (Plate
VIII. Fig. 102). The greater portion of the ganglion is now situated
on the right side of the body, immediately above and to the right of the
radula sac (Plate VIII. Fig. 104, Plate IX. Figs. 115-117). It is
elongated, and its chief axis is directed obliquely across the body, the
right end being considerably higher and a little farther back than the
left end. In passing downward and forward to the left side of the body,
it lies between the cesophagus aud the posterior part of the radula sac.
Its left end is prolonged into a connective, which passes forward and
outward to join the left visceral ganglion (Plate VIII. Fig. 104, Plate X.
Fig. 123). A large nerve, which passes upward and backward to be dis-
tributed to the viscera, emerges from the most dorsal portion of the
abdominal ganglion on the right side of the body (Plate VIII. Fig. 104,
Plate IX. Fig. 117). The histological condition of the abdominal gan-
glion is similar to that of the previously described ganglia of this stage.
The fibrous portion, as well as the enlarged cells and nuclei, are espe-
cially prominent in the portion of the ganglion which lies to the right. of
the median plane of the body (Plate IX. Fig 117).

The buccal ganglia (Plate VII. Fig. 95, Plate VIII. Figs. 102, 106,
Plate IX. Fig. 120, Plate X. Fig. 121) have become larger, and with
their commissure (Plate VIII. Fig. 101, Plate IX. Fig. 120) stretch
across the dorsal wall of the neck of the radula sac, to which they are
still closely united. The nuclei immediately surrounding the central
fibrous portion of the ganglion are already slightly enlarged, though the
cells are not so far advanced in their histological differentiation as are
those of the other ganglia. A single pair of connectives passes obliquely
forward, downward, and outward, to join the buccal to the cerebral
ganglia (Plate X. Fig. 121).

By the sexteenth and seventeenth days, besides a general increase in size
of the external organs, the foot gland extends backward much farther
VOL. XX. — NO. 7. 13

194 BULLETIN OF THE

than the pedal ganglia (Plate VIII. Fig. 107), and the viscera lie rather
more to the left side of the body (Figure G).

The central nervous system (Figure F) now consists of five well devel-
oped pairs of ganglia and an azygos ganglion (Figure G). The cerebral
ganglia with their commissure form the dorsal portion of three nerve
rings, the remainder of which are completed respectively, (1) by the
cerebro-pedal connectives, the pedal ganglia, and their commissures ;
(2) by the cerebro-pleural connectives, the pleural ganglia, the pleuro-

wee
SC

|

Fiaure F. — Posterior face of a transverse section from an embryo of the sixteenth day. X 70.

art. ce. Cephalic artery. a@. Csophagus.

art. pd. Pedal artery. plu. Pleural ganglion.

com. ceb. Cerebral commissure. plu.-pd. Pleuro-pedal connective.
dt. sx. pr, Primary sexual duct. pd Pedal ganglion.

gl. sal. Salivary gland. sac. rad. Radula sac.

gl. pd. Pedal gland. ta. dx. Right ocular tentacle.
n. pd. Pedal nerve. ta. s. Left ocular tentacle.

n. ta. Tentacular nerve.

visceral connectives, the visceral ganglia, the viscero-abdominal connect-
ives, and the abdominal ganglion ; (3) by the cerebro-buccal connectives,
the buccal ganglia, and their commissure. The first and second rings
are further joined to each other by means of the pleuro-pedal connect-
ives. Each of these three rings encircles the cesophagus. The posterior
end of the radula sac in the earlier stages, up to the present one, is
usually found to occupy a position above the pedal ganglia and their

MUSEUM OF COMPARATIVE ZOOLOGY. 195

commissures ; but with a greater concentration of the nervous ganglia
toward one another, the sac is forced to occupy a position below the
pedal ganglia and their commissures. The relations of the different
ganglia to each other is even more definite than before, and can be more
readily understood from transverse sections than from sagittal ones.
The peripheral nerves from the cerebral, pedal, visceral, and abdominal
ganglia are well developed; the principal changes from this time until
hatching are histological.

The cerebral invaginations have become narrow and shorter, but are
still open to the exterior (Plate X. Fig. 124, w.). The deeper portion
of the invagination, that in contact with the proliferated portion of
the cerebral ganglion, has become a solid and rounded mass (Plate X.
Fig. 122, lob. lat.), which is intimately connected with the ganglion
by means of fibrous outgrowths from its ganglionic cells. It is com-
posed of small deeply stained cells, which have undergone no such
histological change as those which compose the proliferated portion of
the brain. It forms a lobe on the antero-lateral face of each cerebral
ganglion (Plate X. Figs. 122, 124, 127). From this time forward the
principal change in the cerebral sacs consists in the gradual obliteration
of the lumen of the invagination. This is usually completed somewhat
later in the embryonic life; but, as previously stated, the sacs have in
one instance at least been found open several days after hatching. Be-
sides this, there is no other connection now remaining between the
ectoderm and any of the ganglia, except such as is effected by means
of the peripheral nerves.

The median proliferated portions of the cerebral ganglia now extend
dorsally farther than in the last stage, and their commissure is much
shorter (Plate VIII. Fig. 105, and Fig. F).

The pedal ganglia (Plate VIII. Figs. 1037, 109-113) have moved
forward, and are broadly in contact with the pleural ganglia. They
have become more compact, and rather more triangular in shape, than
before. From the ventral portion of each ganglion emerge four or
five large nerves, which terminate in the ventral wall of the foot ; from
the dorso-lateral region two nerves are given off to the lateral walls, and
the antero-ventral part of each ganglion tapers off into a stout nerve
running forward to the anterior wall. The connectives with the cerebral
ganglia are well developed (Plate VIII. Figs. 103, 107).

The pleural ganglia (Plate VIII. Figs. 103*, 111-113) are nearer to
the median plane than previously. The ventral posterior face of each is
closely joined to the corresponding pedal ganglion (Figs. 103%, 112),

196 BULLETIN OF THE

the dorsal median face to the visceral ganglion (Figs. 103*, 112, 113),
and the anterior face to the cerebral ganglia (Fig. 107). No nerves
arise from the pleural ganglia.

The visceral ganglia (Plate VIII. Figs. 103%, 110-113) have also
moved nearer to the median plane. The left ganglion is directly below

SPN >
AIT Sra e.
as y

i: SS “SN Be a
glsa wei ue =

eee

Ficure G. — Posterior face of a transverse section from an embryo of the seventeenth day. > 60.

ab. Abdominal ganglion. n. pd. Pedal] nerve.

art. ce. Cephalic artery. mph. Nephridial organ.

art. pd. Pedal artery. nph. dx Right nephridial organ.

dt. sz. pr. Primary sexual duct. 0 cy. Otocyst.

gl. sal. Salivary gland. pd. s. Left pedal ganglion.

gl pd. Pedal gland. pt. cr. Pericardium.

hp. Liver. ret. ta. Retractor muscle of right ocular tentacle,

mt. Mantle. ret. ta. s. Retractor muscle of left ocular tentacle.

nN. Nerve. USC. S. Left visceral ganglion.
the cesophagus (Fig. 111 and Figure G), since the latter occupies a posi- i

tion more to the left side of the body than before. The right visceral |
ganglion still remains larger, and extends farther dorsally than the
left (Figs. 1037, 111, 112). It is nearer the median plane than in the

MUSEUM OF COMPARATIVE ZOOLOGY. 197

stage last described (Figure D., page 190); it lies in front and only a
little to the right of the abdominal ganglion (Fig. 111).

The abdominal ganglion (Plate VIII. Figs. 109-111) is less elongated
than in the last stage (Fig. 104). It is wedge-shaped, and appears as
though crowded in between the two visceral ganglia from behind and
above. It is so intimately connected with these ganglia that it almost
appears to form a part of them (Fig. 111). But the presence, between
the ganglionic masses, of connective-tissue cells, which reach nearly to
the connectives, enables one to make out with some certainty the extent
of each of the three ganglia. Since the planes which separate them are
oblique to the transverse planes of the body, these boundaries are not
always readily seen in cross sections. The right and left visceral gan-
glia have no direct commissural nerve fibres uniting them; they are
joined only by such fibres as pass through the abdominal ganglion.

The buccal ganglia (Plate VIII. Fig. 108, Plate X. Fig. 124) are
now entirely separated from the dorsal wall of the radula sac, from
which they arose, and are surrounded by a layer of connective-tissue
cells. The differentiation of their ganglionic cells is well advanced.

Summary.

1. In Limax maximus the whole of the central nervous system arises
directly from the ectoderm.

2. The cerebral ganglia originate in part as a pair of true invagina-
tions, one on each side of the body in front of the pleural groove and
behind and below the bases of the ocular tentacles. In the course of
their development, the neck of each invagination becomes a long, nar-
row tube-like structure, which remains open throughout the period of
embryonic life. The main part of the cerebral ganglia is formed from
cells which are detached at an early period from the deep ends of these
cerebral invaginations, or from neighboring ectoderm ; the portions
which persist as the walls of the infoldings finally form distinct lateral
lobes of the brain.

3. Ail the other ganglia originate by cell proliferation from the ecto-
derm without invagination.

4. The ganglia arise separately, and, with the exception of the ab-
dominal and mantle ganglia, in pairs, one on each side of the body.
Their connection with each other is the result of a secondary process in
the development, — the outgrowth of nerve fibres.

In advanced stages, the central nervous system consists of five pairs

198 BULLETIN OF THE

of ganglia and an azygos ganglion. Together these form three complete
rings surrounding the esophagus.

The relative positions of the ganglia are best appreciated from cross
sections. In passing from behind forward, they are encountered in the
following order: (1) the pair of pedal ganglia, which lie under the
radula sac, and are joined to each other by an anterior and a posterior
commissure ; (2) one abdominal ganglion a little to the right of the
median plane; (3) a pair of visceral ganglia occupying the posterior
angle formed by the outgrowth of the radula sac from the cesophagus.
They are separated by the abdominal ganglion, from which connectives
pass to them; (4) a pair of pleural ganglia, not joined by a com-
missure, and not giving off nerves. They are united by means of con-
nectives to the pedal, visceral, and cerebral ganglia of the same side ;
(5) a pair of cerebral ganglia, with their supra-cesophageal commissure
and connectives to the pleural, pedal, and buccal ganglia; (6) a pair of
buccal ganglia, with a commissure under the cesophagus posterior to its
connection with the sac of the radula.

The mantle ganglion lies far back, and is joined to the abdominal
ganglion by a large nerve.

It seems as if there could be no doubt that the infolding of the ecto-
derm of the anterior wall of the respiratory cavity on the right side of
the body gives rise to the special-sense organ discovered by Lacaze-
Duthiers (’72, pp. 483-494). It corresponds in its position and its
connection with the right visceral ganglion to his description of the
adult, and also to Fol’s description (80, pp. 166-168) of the origin and
position of that organ in the aquatic pulmonates.

As is well known, Limax belongs to that group of Gastropods in which
all the nerve centres, except the cerebral and buccal, lie on the ventral
side of the intestinal tube; not to the group in which the connection
between the right pleural and right visceral ganglia passes above the
cesophagus, and in which that of the left lies below it. Limax, there-
fore, is not directly referable to Von Jhering’s group of Chiastoneura,
although the want of symmetry in the position of its ganglia does not
allow one to say that it is orthoneuric.

The Gastropod in which the details of the origin and fate of the
nervous centres have been most carefully studied is Bithynia, a chias-
toneuric form, in which Sarasin has found that the abdominal ganglion
is joined to the right visceral ganglion only, and is located at the fundus
of the gill cavity. The relation is different from that found in Limax

MUSEUM OF COMPARATIVE ZOOLOGY. 199

maximus, where the abdominal ganglion is intimately fused with the
right visceral, and is also in close connection with the left visceral
ganglion.

As was to have been anticipated, the abdominal ganglion of Limax
corresponds more nearly in position to that in Lymneus and other
fresh-water pulmonates, as described for the adult by Lacaze-Duthiers
(72, pp. 437-500).

Of the authors who have studied the origin and development of the
cerebral ganglia in Mollusks, Fol (’80, pp. 168, 169, 193-195) is the
only one who has pursued his investigations on Limax maximus,
He says (80, p. 193): “Vers lepoque de la fermeture de la vésicule
oculaire, se montrent deux autres enfoncements de l’ectoderme. L’un
des deux, assez vaste et situé 4 la base du tentacule, 4 son bord intérieur,
est l’origine du ganglion cerebroide ; je le décriai plus loin. L’autre
enfoncement, plus petit, est situé au-dessous de ce dernier, 4 la base du
pied, et méne a la constitution de la vésicle auditive.”

As to the method by which the cerebral ganglia originate, this agrees
in part with that which I have found ; but as to the time of origin, my
investigations lead me to a different conclusion. The otocysts are pres-
ent as small groups of cells (Plate I. Fig. 4), and the cellular elements
which go to form the beginning of the pedal ganglia are also being pro-
liferated (Fig. 3), before there is a trace of the invaginations which go
to form the cerebral ganglia (Fig. 2).

A little later the otocysts assume the form of closed vesicles, uncon-
nected with the ectoderm (Plate I. Fig. 9), while the cerebral invagi-
nations are now seen as shallow pits (Fig. 6). Therefore, in Limax
maximus the formation of both pedal ganglia and otocysts precedes
that of the cerebral invaginations.

Sarasin (’82, pp. 1-68) maintains that in Bythinia tentaculata there
are no invaginations to form the cerebral ganglia. They arise as thick-
enings of the ectoderm, one on each side of the body, which he calls
die Sinnesplatte.

In the recent researches of the Sarasin brothers (’87, pp. 600-602,
*88, pp. 59-69) on Helix Waltoni, of Ceylon, it is asserted that each uf
the cerebral ganglia is at first represented by a group of cells derived
from the part of the ectoderm called ‘Sinnesplatte’”’ before there is
any invagination. There are two groups of these cells, one on each side
of the body. Somewhat later two infoldings arise from each Sinnes-
platte, one above the other. These infoldings become long, narrow

200 BULLETIN OF THE

“cerebral tubes,” the deep ends of which are enlarged (’88, Fig. 24).
From their inner ends a rapid cell proliferation takes place, the prod-
ucts of which join the cerebral cells already in position. The invagi-
nated portions later form the “accessory lobes” of the brain. At a
late stage only one pair of tubes remains open to the exterior, and the
openings to these are closed before the end of embryonic life. The
Sarasins (’88, p. 61) do not know the precise time at which they are
closed, but are certain that the openings do not persist. They express
their belief that the cerebral tubes are homologous with the organs of
smell in Annelids, which, according to Kleinenberg’s studies on Lopado-
rhynchus, also originate as invaginations of the Sinnesplatte, and by
cell proliferation furnish a part of the material for the brain.

Prior to any knowledge of the investigations on Helix by the Sarasins,
I found very similar conditions in Limax maximus. In this case, how-
ever, there is but one invagination of the ectoderm on each side of the
body. It corresponds in position to those described in Helix, being
perhaps the equivalent of the upper or larger invagination in that
species.

The invaginations in Limax have the form of shallow pits before any
other ganglionic cells are to be seen. The cell proliferation, which re-
sults in the production of the main portion of the ganglia, takes place
during their ingrowth. Possibly the proliferation from the depression
between labial tentacle and upper lip represents what was originally
a true invagination, and corresponds to the lower of the two invagina-
tions described by the brothers Sarasin. In Limax maximus the ex-
ternal openings persist until a late stage, and occasionally even after
hatching. Here, also, the invaginations form a lobe of the brain, exactly
as in the case of Helix (Sarasin, ’87, p. 601).

Two well developed “ Seitenorgane” were found by the Sarasins
(88, p. 54) in Helix Waltoni, situated near each other in the “ sense-
plate” ; and they think (p. 60) that these may correspond in position
to the cerebral tubes of later stages.

The groups of cells embedded in the ectoderm, from which, in my
opinion, the greater part of the nervous system in Limax maximus
takes its origin, resemble both in the arrangement of the cells and their
histological condition the ‘“Seitenorgane” described by the brothers
Sarasin (’88, pp. 53-57). But I have never observed bristles, or other
terminal structures, projecting toward the outer world. Moreover, in
Limax unmodified ectodermic cells usually lie between these groups of
large cells and the outer surface of the body.

MUSEUM OF COMPARATIVE ZOOLOGY. 201

The Sarasins (’88, p. 57) consider these clusters of cells homologous
with the “taste-buds ” and “lateral organs” of vertebrates, and say that
they are to be found in and at the margin of the Sinnesplatten, and
along the sole of the foot,— more rarely on the sides of the foot. I
think these organs are probably the same as those which I have seen in
Limax, and to which I attribute simply the function of contributing to
the formation of the ganglia.

Salensky (’86, pp. 685-690) describes the cerebral ganglia of Verme-
tus as arising from a pair of ectodermic thickenings, which early show
pocket-like invaginations, and become deeper and narrower. From the
inner ends of these invaginations are formed the main portion of the
ganglia. The latter are united to each other by a very small commis-
sure, composed of fibrous prolongations of the ganglionic cells surrounded
by other nerve cells. .

The principal difference between the method of development in Ver-
metus and that in Limax maximus consists in the fact that the detach-
ment of the deep portion of the invaginations to form the ganglia in
Vermetus is not effected until the invaginations have reached their ulti-
mate size, whereas in Limax the detachment of cells from the invagi-
nated area begins as early as does the invagination, and accompanies it
during the whole of its formation.

Kowalesky (’83*, pp. 1-54) found in Dentalium two deep invagina-
tions, which he calls the “sincipital tubes,” one on each side of the
head region, a little ventral to the middle of the velar area. From the
posterior deep ends of these sacs the cerebral ganglia are subsequently
formed ; but he is uncertain whether all the cells concerned in the in-
volution share in the formation of the ganglia. If his Figure 65 is
compared with Figures 27 and 33 A in Salensky’s paper, the close resem-
blance in the method of origin of the cerebral ganglia in the two types
becomes apparent.

Fol (80, pp. 169, 170) asserts that the pedal ganglia of the aquatic
pulmonates appear as condensations in an already formed mesoderm,
and that they are nearer the pharynx than the ectoderm when they
begin to be discernible. ‘‘One may therefore say,” he adds, “that these
ganglia arise from the mesoderm without prejudging the unsettled ques-
tion, viz. from which of the primordial layers arises the mesoderm
which forms them.” Of the pedal ganglia of the terrestrial pulmonates,
he says that they are diffentiated en leew et place in the midst of the
mesodermic tissues of the foot.

202 BULLETIN OF THE

With this I cannot agree, although I admit that at the time when
the groups of cells which form the ganglia begin to be proliferated from
the ectoderm, it is extremely difficult to distinguish them from the
mesodermic elements (Plate I. Fig. 5). It is to be observed, however,
that Fol considers it an unimportant distinction, whether the ganglia
are formed from groups of mesodermic cells which have themselves
recently originated from the ectoderm, or by a proliferation of cells
directly from the ectoderm.

I am unable to reconcile the account of the development of the pedal
ganglia in Bithynia given by P. Sarasin (’82, pp. 47-49), with the
conditions seen in Limax; nor can I think it probable that any consid-
erable difference exists between nearly related mollusks in regard to
the place whence the ganglionic cells arise. Sarasin maintains that in
Bithynia the pedal ganglia arise from a single median thackening of the
ectoderm of the dorsal wall of the foot, in the region where that wall
bends over to become continuous with the posterior wall of the visceral
sac. Anteriorly, in the region of the oral invagination, this median
band of cells forks, and each branch becomes joined to the correspond-
ing cerebro-pleural cell mass by a slender cord of cells. Subsequently,
the posterior unique portion of the proliferated cell mass is completely
divided into lateral branches by a separation which progresses from in
front backward. It seems to me that, according to this account, both
the pedal ganglia must be regarded as arising from a common mass of
cells, and that they are not from the beginning wholly separate, as I
maintain for Limax.

The relative positions. of pedal ganglia and otocysts present, to my
mind, a serious objection to Sarasin’s view, which may not have seemed
so important to him on account of his uncertainty about the origin of
the otocysts. I believe it is sufficiently evident that the otocysts do
not arise, as Sarasin thinks probable, from the cerebro-pleural prolifera-
tions, but independently, and from the dorso-lateral wall of the foot in
the region of the “ pleural groove.” They ultimately lie immediately
dorsal to the corresponding pedal ganglia. If Sarasin’s view as to the
origin of the pedal ganglia as a median dorsal proliferation were correct,
the ganglia would have to migrate to a lower plane than that occupied
by the otocysts. But there is no evidence either in Limax or the figures
given of Bithynia which would confirm such a supposition. As further
corroboration of my opinion that the pedal ganglia arise from the ventral
and lateral walls of the foot, I would cite the conclusion reached by
Salensky (’86, pp. 691, 692) for Vermetus. He has shown that the

MUSEUM OF COMPARATIVE ZOOLOGY. 203

pedal ganglia originate from the ventral wall of the foot, in a region
and by a method corresponding to that seen in Limax maximus, as will
be seen by comparing his Figures 21 C to 23 with my Plate I. Figs. 5
and 7, and Plate IV. Fig. 57. The only important difference between
Vermetus and Limax lies in the fact that, in the case of the former, the
cells forming the ganglion remain from the beginning a more compact
mass than they do in the latter.

No one except Lacaze-Duthiers (’72, pp. 456, 457) has mentioned
the existence of more than a single pedal commissure. He maintains
that there are in Lymnzus as many as three. After speaking of the
cerebral ganglia as being connected by one commissure, he goes on to
say (p. 456), “ Au contraire les ganglions pédieux ont trois commissures
réelles.”” He seems, however, uncertain as to whether the most posterior
ought to be considered a true commissure: “ La troisitme commissure
mérite-t-elle bien ce nom? elle est constante dans les Pulmonés et se
présente sous la forme d’un petit nerf gréle transversal naissant 4 peu
pres a la hauteur du troisieme nerf pédieux inférieur ; elle donne vers
son milieu naissance 4 un filet nerveux trés-délié, impair médian que
l’on suit dans les tissus de la fosse pédieuse sans trop pouvoir définir et
limiter exactment son réle.” (p. 457.) His investigations were made
exclusively upon the adult.

In Limax maximus two commissures are certainly distinguishable
during a greater part of the embryonic life; no trace of a third has
been seen. ‘The adult has not been studied.

None of these authors, with the exception of Sarasin, say anything
conclusive concerning the origin of the remaining ganglia, although
Salensky (86, p. 697) speaks as if the pleural ganglia of Vermetus
originated in the cerebro-visceral connectives, which are shown in his
Figures 31B to 31F.

Sarasin asserts (’82, pp. 46, 47) that in Bithynia the pleural ganglia
originate as part of the “Sinnesplatte,” from which the cerebral ganglia
arise, and that these ganglia, cerebral and pleural, are so closely fused
with each other in the later stages of development as to form on either
side of the body a single mass.

I believe that they arise in Limax maximus by cell proliferations
from the lateral walls of the body, behind the cerebral ganglia, and just
above the pleural groove ; they are closely connected (not fused) with
the cerebral ganglia only in late stages.

Sarasin (’82, pp. 50-52) says that the visceral ganglia in Bithynia

204 BULLETIN OF THE

arise by cell proliferation from the dorsal margin of the ventral wall of
the head or trunk region, above that which I have called the pleural
groove. Further, that the right visceral (or supra-intestinal) ganglion
is connected by a nerve fibre to the olfactory ganglion under the gill
cavity. Farther back than the visceral ganglia he finds a median pro-
liferation of cells lying at the ventral margin of the gill cavity, from
which the abdominal ganglia arise. He asserts that there are two ab-
dominal ganglia, — one connected with the supra-intestinal ganglion,
the other with the sub-intestinal ganglion.

In Limax maximus the visceral ganglia and the abdominal ganglion
arise by the same method as that described by Sarasin; but the former
are produced from the lateral walls of the head region, above the pleural
groove, one on each side of the body. The right ganglion in later stages
is more dorsal than the left. It appears to be formed in part from the
inner wall of the respiratory cavity, to which it remains connected by a
nerve. It is in this region that is developed an organ which I believe
to be the olfactory organ of Lacaze-Duthiers.

There is only one abdominal ganglion ; this takes its origin a little to
the left of the median line of the body, from the anterior margin of the
body wall immediately above the pleural groove.

Sarasin (’82) is the only author who gives attention to the origin of
the buccal ganglia. He describes them as arising in exactly the same
manner, and in the same situation in relation to the walls of the radula
sac and the cesophagus, that they do in the case of Limax maximus.

CAMBRIDGE, November, 1889.

MUSEUM OF COMPARATIVE ZOOLOGY. 205

BIBLIOGRAPHY.

Bobretsky, N.
76. Studien tber die embryonale Entwicklung der Gasteropoden. Arch. f.
- mikr, Anat., Bd. XIII. pp. 95-169, Taf. VIII.-XIII. 1876.
Biitschli, O. ;
'77. Entwicklungsgeschichtliche Beitrage. Zeitschr. f. wiss. Zool., Bd. X XIX.
pp. 216-254, Taf. XV.-XVIII. 1877.
Fol, H.
80. Développement des Gasteropods Pulmonés. Arch. de Zool. exp. et
gén., Tom. VIII. pp. 103-232, Pls. IX.—XVIII. 1877.
Jhering, H. von.
"75. Ueber die Entwicklungsgeschichte von Helix. Zugleich ein Beitrag zur
vergleichenden Anatomie und Phylogenie der Pulmonaten. Jena. Zeitschr.,
Bd. IX. Heft 3, pp. 299-388, Taf. XVIII. 1875.
'77. Vergleichende Anatomie des Nervensystems und Phylogenie der Mol-
lusken. Folio, x + 290 pp., 8 Taf., 16 Holzschnitte. Leipzig, 1877.
Jourdain, S.
'85. Sur le Systéme nerveux des Embryons des Limaciens et sur les Relations
de ’Otocyste avec ce Systeme. Compt. Rend. Acad. Sci. Paris, Tom. C.
pp. 383-385.
Kowalesky, A.
’83. Embryogénie du Chiton Polii (Phillippi), avec quelques Remarques sur
le Développement des autres Chitons. (Odessa, Dec., 1882.) Ann.
_Musée Hist. Nat. Marseille, Zool., Tom. I., Mém. No. 5, 46 pp., 8 Pls.
1883.
'83*, Etude sur l’Embryogénie du Dentale. Ann. Musée Hist. Nat. Mar-
seille, Zool., Tom. I., Mém. No. 7, 54 pp., 8 Pls. 1883.
Lacaze-Duthiers, H. de.
60. Mémoire sur l’Anatomie et l’Embryogénie des Vermets. 2° Partie.
Ann. Sci. Nat., 4 série, Tom. XIII. pp. 266-296, Pls. VIT-IX. 1860.
'72. Du Systeme nerveux des Mollusques Gastéropodes pulmonés aquatiques
et d’un nouvel Organ d’Innervation.. Arch. de Zool. exp. et gén., Tom. I.
pp. 437-500, Pls. XVIIT-XX. 1872.
’85. Le Systéme nerveux et les Formes embryonales du Gardinia Garnotii.
Comp. Rend. Acad. Sci. Paris, Tom. C. pp. 146-151.
Langerhans, P.
'73. Zur Entwickelung der Gastropoda Opisthobranchia. Zeitschr. f. wiss.
Zool., Bd. XXIV. pp. 171-179, Taf. VIII. 1873.

206 BULLETIN OF THE

Lankester, E. R. :

73. Summary of Zodlogical Observations, ete. Ann. Mag. Nat. Hist.,
4th series, Vol. XI. pp. 85-87. 1873.

'74. Observations on the Development of the Pond Snail (Lymneus stag-
nalis), and on the early Stages of other Mollusca. Quart. Jour. Mier. Sci.,
Vol. XIV. pp. 3865-391, Pls. XVI., XVII. 1874.

Rabl, C.

'75. Die Ontogenie der Siisswasser-Pulmonaten. Jena. Zeitschr., Bd. IX.
Heft 2, pp. 195-240, Taf. VII-IX. July, 1875.

"79. Ueber die Entwicklung der Tellerschnecke. (Wien. Mitte Feb., 1879.)
Morph. Jahrbuch, Bd. V. Heft 4, pp. 562-655, Taf. XXXII.-XXXVIIL.,
7 Holzschnitte. 1879.

’83. Beitrage zur Entwickelungsgeschichte der Prosobranchier. Sitzungsb.
der K. Acad. der Wissensch. Wien, Math.-naturw. Cl., Bd. LXXXVII.
Abtheil. ITI. Heft 1, pp. 45-61, Taf. I., IT. 1883.

Salensky, W.

'72. Beitrage zur Entwicklungsgeschichte der Prosobranchier. Zeitschr. f.
wiss. Zool., Bd. XXII. pp. 428-454, Pls. XXXV.-XXXVIII. 1872.

'86. Etudes sur le Développement du Vermet. Archives de Biologie,
Tom. VI. pp. 655-759, Pls. XXV.-XXXII. 1886.

Sarasin, P. B.

’82. Entwicklungsgeschichte der Bithynia tentaculata. Arbeit. a. d. Zool.-

Zoot. Inst. Wiirzburg, Bd. VI. Heft 1, pp. 1-68, Taf. 1—-VII. 1882.
Sarasin, P. und F.

87. Aus der Entwicklungsgeschichte der Helix Waltoni Reeve. Zoolo-
gischer Anzeiger, Jahrg. X., No. 265, pp. 599-602. 21 Nov., 1887.

88. Aus der Entwicklungsgeschichte der Helix Waltoni Reeve. Ergeb-
nisse naturwissenschaftlicher Forschungen auf Ceylon in den Jahren
1884-1886, Bd. I. Heft 2, pp. 35-69, Taf. VI-VIII. 1888.

Wolfson, W.

'80. Die embryonale Entwickelung des Lymneus stagnalis. Bull. Acad.
Sci. St. Pétersbourg, Tom. XXVI. No. 1, pp. 79-99, 10 Figs. 12 Mars,
1880.

MUSEUM OF COMPARATIVE ZOOLOGY. 207

EXPLANATION OF FIGURES.

All the figures were drawn with the aid of the camera lucida, and were made from
preparations of Limax maximus.

INDEX TO STAGES.
The Roman numerals indicate Plates. The Arabic numerals, Figures; those

which are enclosed in a parenthesis belong to the same specimen. Skeleton num-
bers on the plates refer to the number of the section in its series.

6th day. (+), (+).

ith “ (= =) (=) (a) (sam a) z rs

eh (|.

14, 16, 19
1 oe a Vv

Sth “ Luge | ) eye
20-24? 25-27/? \60

10th “ Ut. Ty. V. )
28-31? 58, 59 61, 637°

: III. V. VI. salah a
itt * CS I ea oa |,
(sar 51’ 62, 64, 67? = 68, 69? So

12th “ Mee) ¥,) EVIE ae) ( Ill. ae ( VI. ) VII.
48, 49’ 65, 66? 81, 82/7 \50, 52? 85-94/? \74-76, 8028/7 \g3-84/°

14th « ea VIII. Hs) ( VIII. 1X: >. 2 \?
95 ? 96, 101, 102? 1147? \97-100, 104, 106? 115-120? 121, 123, 125, 126

16th ** Sra ), =).

7th “ ( VIII. XS
103, 1034, 105, 108-113? 122, 127/°

208

The right side of the animal is indicated by the letters dr., the left side by s.

BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

ABBREVIATIONS.

These

letters are usually affixed to one or more of the abbreviations used to designate organs.
The skeleton figures rmmediately under the number of a figure on the plate indicate the

number of the section in the series to which the figure belongs.

Stages” (p. 207).

ab.
ab.-vse.
an.

buc.

cav. mt.
ceb.-buc.
ceb. dx.
ceb. s.
ceb.-pd.
ceb.-plu.
ench,
com. da.
com. buc.
com. ceb.
com. pd.
com. pd. a.
com. pd. p.
dt. sx. pr.
dx.

en.

gl. pd.

gl. sal.
gn.

tv. ceb.

Abdominal ganglion.

Abdomino-visceral connective.

Anus.

Buccal ganglion.

Mantle cavity.
Cerebro-buccal connective.
Right cerebral ganglion.
Left cerebral ganglion.
Cerebro-pedal connective.
Cerebro-pleural connective.
Shell gland.

Anterior pedal commissure.
Buccal commissure.
Cerebral commissure.
Pedal commissure.
Anterior pedal commissure.

Posterior pedal commissure.

Primary sexual duct.
Right.

Entoderm.

Pedal gland.

Salivary gland.
Ganglion.

Cerebral invagination.

lab.
lns.

Consult also “‘ Index to

Upper lip.
Lens.

lob. lat. Lateral lobe of brain.

loph. cil. Ciliated ridge.

mt. Mantle.

n. Nerve.

nph- Nephridial organ (primitive kid-
oc. Eye. [ney.)
ocy. Otocyst.

@. (Esophagus.

pd. Pedal ganglion.

pes. Foot.

plu. Pleural ganglion.

plu.pd. Pleuro-pedal connective.
plu.-vsc. Pleuro-visceral connective.
pr f- Cell proliferation.

rad. Radula sae.

ret. ta. Retractor muscle of tentacle.
8. Left.

sul. plu. Pleural groove.

ta. Ocular tentacle.

ta.’ Labial tentacle.

USC. Visceral ganglion.

vsc.-plu. Viscero-pleural connective.

ds pA od Mla Bia
we Pea,
oe

a ee Kis Ho

bray: tae

igi Hoe boven PRR Hit iF wre Wi TA ‘ |
Patton en red’ Te mye ie . |

| if :
Ay od ¢ ; |
sn eles a eran aii

bi :
| J

GM oN ea te tifigt? ge; Aa) ,;
a ae Cer ee: Ate ne “hid wreath: ; Wat
4) ean fr ‘anianpiliane on vs: Bs : 1 wh
eh hay ots mA wy a aght Lb) '
v at tae de aie fry at AAs Jabot hae dc, .
ew aettoscbe nite sion
Pert ar oly, ae tn ett ria eoniied taht, gh aie Aa
alii Ce ee eee od RT bets Mh Seder é
- sdoinit ee on Auta baat é hak wired" ae Anew) &
steer rita Wikis - esilaya Peay tN eee jy vitct ae "
» bedony wait mid cd ai: cath ‘RY CRE 1 A ogih wa¥ las ro oe
arr ore RAD tot Are Cty
PARTY RIE. Lo Hig okt a bat rset et ee
tami io oa O30" wooly Aine i ae
om Wek ites agairye 2 We viene ep hie Bi. A nee
A a pipet aa) Oth aT ae aia dst ee ; ? Is
a TO OMA) th Lote fay ox rigs nt ; ae
ty itr GME Kyi wii weer ys asaling Bee Baty Pada. yh | east
; Sp Aa ie only Ce oy penny Pp oly LA eis See p
am OF ett Dt AA oe rela wish A nha
o ” vn pate ihe deta Bae pce 9 jaw Fiesty ty ty, ant yar cenele
dink BA hehe yr ah .ehe neice BEER :
Sod ed Rav hity iS ua SMe, ah aT Biv facing | ‘
rep thre Ipiea ys deaths TA pr aretianny)® eehe Pes froliguer diately ity
silt dae Wy tet al eg tS) ae es
“ot itt mail bogelavirtt (itd ie iepaet’

caste

t
\ =
»
se *
y ie
— _
*. — oe
~
‘
i
"
j ri
are: t
=
*
.
. . t
‘
— ‘ }
= i
iM
{
{
{

HENCHMAN, — Nervous System of Limax.

PLATE I.

All the figures of this plate were made from material killed in Perenyi’s fluid, and

Fig. 1.

«ec

“cc

2.

all except Fig. 1 are magnified 250 diameters.

A small portion of Fig. 5 more highly magnified to show the cell prolifer-
ation for the right pedal ganglion.

Posterior face of a transverse section from an individual about six days
old. The section passes anterior to the “ pleural groove,” and through
the region where the cerebral invaginations subsequently arise; the
left side is cut a little anterior to the right. Stained in alcoholic
borax-carmine.

A section from the same individual posterior to the pleural groove in the
region of the cell proliferation for the pedal ganglia.

A section from the same, still farther back.

Transverse section from an embryo a few hours older than the preceding,
in the region of the proliferation to form the pedal ganglion. Stained
in alcoholic borax-carmine.

6-9. The left surface of sections parallel to the sagittal plane from an em-

bryo of the seventh day. Figs. 6, 8, and 9 represent respectively the
11th, 16th, and 18th sections of the series, and are from the left half
of the embryo. Fig. 7 1s from the right half, and passes through
the right otocyst. Stained in Czoker’s cochineal.

10-13 exhibit the right surface of sections from another individual (between the

seventh and eighth days) cut parallel to the sagittal plane, the anterior
portion a little in advance of the posterior. Fig. 10 is a section passing
through the proliferation forming the right visceral ganglion. Figs 11
and 12 are two successive sections passing through the left visceral
ganglion; the latter also passes through the left otocyst. Fig. 13
shows the region of the forming abdominal ganglion. Stained in picro-
carminate of lithium. In both these individuals the left ganglia and
otocysts are more developed than the right.

Ds ae ai Ba

' = : wil, . BMeisel,lith. Beston.

"a

.
*
‘
'
-
|

_

Ra ees.
, my Y : fart
es ae a tt pee are woe
ra

a i athe oh fe Canis 1; mis

ye
q es ‘
oe iets tian ‘Aj mg? tal

MBhwnses ‘WIR VA) C8 f ‘Ty
bh atin! ei) cn ila emis

Belin HH Lat inc:
Ty th ae
M4 Tay Wi ii ee re
; Lewes Al trait) iwahaiie vy

+ Ww thet, fe City eae ed
ae 10 OAT ry gee
| a ps id .

i] <1 yl? ‘eps
© 7 ee Rite ry ras, Gy
av Acad ty lr

+ eds y j Pie
he ore ky

oye \ ; “4 ‘

, baa TRU we

.

Politi 208 5 0

HENCHMAN. — Nervous System of Limax.

Fig.

PLATE UU.

All the figures of this plate are magnified 250 diameters.

14. A section parallel to the sagittal plane from an individual of the eighth
day. It passes through the cerebral and pleural ganglia of the left
side of the body, and also shows four cells of the left otocyst poste-
rior to the pleural groove. The material was killed in 0.88% chromic
acid, and stained in alcoholic borax-carmine.

15. The left surface of a section cut parallel to the sagittal plane from an
embryo of the seventh day (but more advanced than in Figs. 6-9), pass-
ing through the cerebral invagination and a group of cells belonging
to the proliferated portion of the cerebral ganglion of the right side.
The material was treated as in that of Fig. 14.

16. A section from the same individual as Fig. 14, passing through the cell
proliferation to form the visceral ganglion of the right side.

17 and 18 are from the same individual as Fig. 15.

17. A section passing through the visceral ganglion and the external opening
of the nephridial organ of the right side. |

18. The second section nearer the median plane than Fig. 17, showing the
cell proliferation to form the right visceral ganglion.

19. From the same individual as Fig. 14, showing the cerebral invagination
and proliferation of the right side, and also a cross section of the
primitive kidney. :

20-24. The anterior surfaces of transverse sections from an embryo of the
ninth day. Material killed in Perenyi’s fluid, and stained in alcoholic
borax-carmine.

20. Portion of a section which passes through the proliferation of cells form-
ing the pleural ganglion (dorsal to the pleural groove), and through
the pedal ganglion and the otocyst of the left side.

21. The 37th section, which passes through the cerebral commissure and
shows the proliferation of cerebral cells on the left side.

22. The 51st section, which passes through the buccal ganglion of the
right side.

23. The 69th section, showing the unsymmetrical position of the visceral
ganglia and a cross section of the right nephridial organ.

24. The 75th section of the series, passing through the abdominal ganglion
and the invagination to form the mantle cavity. It is in the region
where the anterior portion of the embryo is bent backward over the
foot by the nutritive sac; the foot is not represented in the figure.

B Meisel, Jith. Boston.

teal £
AG

> Cavan

\

=
by
ra)
=
ae
-
icp)
g
S
a
a
i
S)
a

ef

er

HeEncuMaN. — Nervous System of Limax.

PLATE III.

Figs. 25-27 are from the same individual (ninth day) as Figs. 20-24 on Plate II.

ii

They are magnified 83 diameters.

25. This section passes through the left cerebral ganglion, and the cerebral
invagination of the right side.

26 shows, in addition to the cerebral invagination, that of the right eye (at
the left of the figure).

27. Section passing through the pedal ganglia and their anterior commissure.
It shows a cross section of the cesophagus, the radula sac, and the right
otocyst.

28-31.. The left surface of sections parallel to the sagittal plane from an em-
bryo of the tenth day. (Figs. 58, 59, Plate IV., and Figs. 61, 638, Plate
V., also belong to this series.) ‘This individual was killed in Perenyi’s
fluid, and stained in picro-carminate of lithium. X 100.

28 and 29 are successive sections passing through the invaginations of the cere-
bral ganglion and the eye of the left side.

30 shows a cross section of the cerebral commissure.

31. The second section nearer the median plane than Fig. 30, showing the
cerebral commissure, the position of the right visceral ganglion, and
the anus already open to the exterior.

82-47, 51. The left surface of sections cut parallel to the sagittal plane, from
an embryo of the eleventh day, magnified 100 diameters. Kilied in
Perenyi’s fluid, stained in Czoker’s cochineal. (Sections shown in
Plate V. Figs. 62, 64, 67, and Plate VI. Figs. 70-73, also belong to
this series.)

32-38. Seven successive sections passing through the invagination for the
cerebral ganglia of the left side.

39. The second section nearer the median plane than Fig. 38, showing the
inner sac-like end of the invagination.

40. The next section, showing the blind end of the invagination and the
proliferated portion of the ganglion.

41 and 42. Successive sections (32d and 33d of the series) passing through the
left pedal ganglion and the connective between the abdominal and
left visceral ganglia.

43. Section cutting the abdominal ganglion crosswise.

44. The 35th section shows, in addition to the abdominal ganglion, a cross
section of the anterior pedal commissure. (The 36th, 87th, 39th, and
40th sections of the series are shown in Figs. 47, 46, 45, and dl,
respectively.)

45. Section passing through the connective between the abdominal ganglion
and the right visceral ganglion.

46 shows the abdominal ganglion still connected with the ectoderm.

47. The next section to that shown in Fig. 44, passing through the abdominal
ganglion.

48,49. The left surface of sections cut parallel to the sagittal plane from an

individual of about the twelfth day, magnified 100 diameters. Killed in
(See obverse.)

PLATE III. (continued.)

0.383% chromic acid, and stained in alcoholic borax-carmine. (Sec-
tions shown in Plate V. Figs. 65, 66, and Plate VII. Figs. 81, 82, also
belong to this series.) P

Fig. 48. A section passing through the cerebral and pleural ganglia of the right

se

-

‘

ce

side, and also through the abdominal ganglion.

49 shows the proliferated portion of the cerebral ganglion and the visceral

ganglion of the right side.

51. (See explanation of Figs. 82-47.) The section following that shown in

50, 52.

Fig. 45. It passes through the connective between the abdominal
ganglion and the visceral ganglion of the right side.
The posterior surface of transverse sections of an embryo at the same
stage of development (twelfth day) as that represented in Figs. 48 and
49. ‘The ventral portion and the right side cut a little in advance of
the dorsal portion and the left side, magnified 100 diameters. Killed
in 0.33% chromic acid, stained in alcoholic borax-carmine.

Figs. 85-94, Plate VII., continue this series. The following shows
the sequence of the sections : —

Section 77, 101, 103, 109, 110, 112, 118, 117, 125, 126, 128, 146.

Figure 90, 92, 91, 93, 50, 52, 85, 86, 88, 87, 89, 94.

50. This section passes through the left pedal ganglion and the abdominal

ganglion where the latter is still attached to the ectoderm. The
mantle cavity open to the exterior.

52. The second section in front of Fig. 50, showing that in this region the ab-

dominal ganglion is free from the ectoderm.

WSS? \“e\)
4

i @ = [; \ Sa 2

‘We ee \F y
by & & therse.de pe

| oe

B.Meisel lith.Baston.

OM ; ,
‘ ; Si.
' in A
‘ .
4 J
x : ed,
rn ¢ Sgt ’ 2
wr. Ni uA ad
f i 3 alee 7 .
1 REDE INTE ATI, 02-5. | git
bs ‘ieee fe r
te dal ¥ i) T
J ad ‘ ’ 4
: i 4 4
4 af ; LA
ay } F b 4
: ' Bry" 7h, - 4
. vi = | . \ ’ 4 :
iis ti : - = Ah by ‘a ; Wi; ;
, ie 5 ip 7 at f Wiens ON a aes Se ° a
7 ; iW mt f j Poy BAG Ue 5 a M
Dany | nd Gr yt Ys : ‘py. ee i » i ne
ee eat on ; vue mn a, el
ya : Dp, be Leet ae i Mes
; 1 rs . rr * : ee é
’ a Fh uf i :
a win Later
1 f ih
i]
Mi

eer Vi) since an

. BN | i
rlasolall Beit DyMira nie wale Pie aA at

ahi Pre PijGHians OV lane/oue qed ie tid dna 4 ‘

wh Sil Seatrigpany's. « Ra VN eae oink thi's eB Ae fm . ith

Cee boitea: "i Jot Tia ote A yh a ‘ nh r
1a repeat eet teshh we : th

ivi ses write Aloe 4) feiss yee sr PRAT a ‘thy ala trvide ie in b : . if: J

F Mn Su ney Palbrcstivon fe alg ja geeletiqu AenProty ois if a ni J

PEMA DOME tin wre zits edi ST sik Snirinans > oR

Be tis Rae,

) ovina PAR n6 Oe le rent rte: woe Piguidin ‘ait i
DORN More ld Powda by, SY. emery beds, “8

On LAW the Ud ie gt night Cie Ania nvig. yh! we ; 7
a Hwide Point BP tak Aj rao hrtwle oy lieth: ati
Pe aay we te Mee jee} ool (One (atogey ty Jagd TPs ;

wi yarity f intnatiy 4 og a eee MMe “
hy, sates, Pty a y Wo te Gib oh AvitGa Se aidigh las

RENAE MMR S18 ul Ni Tie Lomaye eds Ee) a
iin vine adit ai Wa faa ol bata ih ate eri rest PGS

i Sai iy GET Ch RN tat a, UPON PET Hadid ied: 2 Ba dit 7, os
ci ania shun MEL WMLY Hage ice aes Me he Tose Saar. er
ai Ws i, fe ye i # i ay iad a) = Ar tw tad aN st gy Yee ;
an at rn eag en Dees ols wR ip iA HINA v9}

: ; td wo aa ves ts we evry mye vis. tt Depiecnrtyn } neh einen = i
ils 2. : Vw en

2 , i)
sl wise Py J - ‘uta if t fie REN rye Ont Wi AP ‘ce hk f Vissi sé Oe: s Week

we | Pad Oy ahi Sos la «ll bhi’ Pel) 34

Ty a
.
t .
7 i ) .
Powe
g ie
ee 2 Sali ~
— ot
’
al ’
.
be °
‘
1
. ¥ . ,
=
1
‘ W
‘ ? '
-
é *
F a
i 2 .
¥
4
4 °

MENCHMAN. — Nervous System of Limax,

PLATE IV.
All figures of this plate magnified 250 diameters.

Figs. 53, 56, 57, and 55 are four successive sections from the same embryo. This
was seven days old, but much more advanced than the embryos represented
in Figs. 6-18, 15, 17, and 18. Killed in 0. Lig chromic acid, stained in
alcoholic borax-carmine.

“ 63. Posterior face of a transverse section passing through the visceral ueore
the external opening of the nephridial organ of the right side, and the
opening into the mantle chamber or respiratory cavity. (Compare
Figs. 56 and 55.)

“54. The posterior surface of a transverse section passing through the anterior
pedal commissure. Embryo of about the same stage of development
as the preceding, and prepared in the same way as that.

“ 55. Portion of the third section following that shown in Fig. 53; it passes
through the pedal and visceral ganglia of the right side.

‘“« 56 shows both the visceral ganglia.

“ 57. Portion of section showing the left pedal ganglion and cell proliferation
from the ventral wall of the foot; the visceral ganglion and the ex-
ternal opening of the nephridial organ of the left side are also seen.

“ 58, 59. Consult explanation of Figs. 28-31, Plate III.

“ 68. Sagittal section passing through the proliferated portion of the right
cerebral, pedal, and visceral ganglia and right otocyst. Shows cell
proliferation from the ventral wall of the foot, and that the visceral
ganglion is attached to the ectoderm at the margin of the mantle
cavity.

“ 59. A small portion of the section following Fig. 58 to show the right visce-
ral ganglion and its attachment to the ectoderm.

8

eg
an a
SES.
on

S

ene

ed

GO
©)

Wer

»\
is
in De

é B Meisel lth. Boston

)
: ~ :
|

———

3 a
i iy

yee

HENCHMAN, — Nervous System of Limax.

Fig. 60.

61.

62.

63.

64.

PLATE V.
All figures magnified 250 diameters.

The left surface of a section cut parallel to the sagittal plane from an
embryo of the ninth day. The section passes through a few cells of
the cerebral, the pedal, and the visceral ganglia of the right side of
the body, and also shows a section of the right otocyst. Killed in
Perenyi’s fluid, and stained in alcoholic borax-carmine.

(Consult explanation of Figs. 28-51, Plate III.) Sagittal section (the 22d)
to show the abdominal ganglion, which lies embedded in the ectoderm
anterior to the pleural groove.

(Consult explanation of Figs. 32-47, Plate III.) The 32d section of the
series, showing a transverse section of the cerebral commissure and a
portion of the left buccal ganglion.

(Consult explanation of Figs. 28-31, Plate III.) The 15th section of the
series ; it passes through the left cerebral, the pedal, the pleural, and
the visceral ganglia, and the wall of the left otocyst.

(Consult explanation of Fig. 62.) The 24th section; it shows the inter-
nal end of the cerebral invagination and the cell proliferation to form
the larger part of the brain of the left side.

65, 66. Compare explanation of Figs. 48, 49, Plate ITI.

65.

66.

67.

68.

69.

The 91st section of the series, showing transverse sections of the buccal
commissure and the connective between the abdominal and the left
visceral ganglia.

The 102d section of the series; it passes through the abdominal ganglion.

(Consult explanation of Fig. 62.) A section through the right visceral
ganglion.

The posterior surface of a transverse section from an embryo of the
eleventh day. It shows the right visceral and the abdominal ganglia.
Killed in Perenyi’s fluid, stained in Czoker’s cochineal.

The second section anterior to Fig. 68, passing through the visceral
ganglion of the right side.

(Additional sections from this specimen are shown in Figs. 77-80,
Plate VI.)

SHMAN.- NERVOUS SYSTEM OF LIMAX.

7

l itpig
| eee 4

B.Meisel.lith,Baston.

=D

-
°

{ Lees : ; U
ry au ina dad eet fos ES,
' } a aa : ; J

uy 1 ) rr eS |

B71 ad OME bovis o stn gee tt tbh SHAME Yi liy pitt AG. pantry Be ald
s Mrsihinynihy (08 Pepi gbin orty So Loa EE ral Nissho the

hs a 7
ae ee b= ingadgyadinenut thee
ans AME RND!6 Noe Cuchi Landieras what casts eer Oe
tet TS rikiy ai Tele Dita hive Hy-utd nas ‘ u |
‘ in Salatse at) iid Seek Ost egy be hase avy olinad B iyi
ee et a oR Fed MT hse |e ee Mi
PNR AE UR GBI Ciery Lecittet a ware) dee’ Only Eyptey avril. ta
, +e Ow Silo rely
Re MOD Vict oilt Lo Ine Tinaed olf to ny ing A
ID 3s EH Gay Lsbay outset wetemilita lio self
Peeled oe to hie tity! ih nso PP RY YTD ee inte
PAO Ie Porawbnarl 3) wealiey velsptes Poe Ni ea)
, tah Behe ie sha a ist hoo REMY Sete to’

meuiiitat: ; “pir eh thd TE Le ; : F 4 7 a)
Matos At heotage

mR REA mts ie meiy is BRAT Fini sive , i ae cele eT Bt
TMG Thi tee { Opa: it i ria

‘

ee: OMe TAS (1 To Sréit.o2 HE Ga UL ot OY banal
MW Whivce!. he mit Lo Beis ots . | ;
MOR wilroiunt ” tS PD .ni'l On othanlgnd obe eke
#9 ons OF Migwer ott wo oanu Sede tle, :
Vie) Gee

| ‘wt ts pieces st) ATT Wo nother
Gre an 46 nies By Bb. 64 beat Nady ' wld

Mahe,
By A)

Vee

M j
i

HENCHMAN. — Nervous System of Limax.

PLATE: Vi

All the figures of this plate were made from material killed in Perenyi’s fluid, and

all except Figs. 72, 73, and 77 are magnified 250 diameters.

. 70-738. Consult explanation of Figs. 32-47, Plate III.

70 shows the left cerebral, pleural, and visceral ganglia in the region of the
cerebro-pleural and pleuro-visceral connectives. .

71. Section passing through the otocyst and the cerebral, pedal, and visceral
ganglia of the left side. The visceral ganglion is connected with the
ectoderm, and the pleuro-visceral connective is much more elongated
than on the ninth day.

72. A portion of the ventral wall of the foot from the 82d section, to show
the cell proliferation for the pedal ganglion. X 665.

73. Same as Fig. 72, but from the right side of the body.

74-76, 80%. The posterior surface of transverse sections from an individual
of the twelfth day. The right side cut slightly in advance of the lett.
Stained in picro-carminate of lithium.

74. The 81st section, which passes through the otocysts and both pedal
ganglia in the region of their anterior commissure.

75 and 76. The 59th and 60th sections of the same series, showing a part of
the left side of the embryo in the region of the abdominal ganglion.

77-80. (See explanation of Fig. 68, Plate V.) Posterior faces of four suc:
cessive transverse sections through the mouth of the left salivary
duct where it connects with the esophagus. Fig. 77 shows also the
buccal ganglia.

80%. (See explanation of Figs. 74-76.) The 107th section of the series. It
shows the cerebral commissure, a few cells of the right cerebral
ganglion, and the right eye.

IENCHMAN.—NERVOUS SYSTEM OF LIMAX.
iP. ve

—

See oo ae) lorhont.
Haag” _ J
“NSS! Ss nie Yi

107

B Meisel, lith. Boston

. '
, f =
¢ » I
F = i i
¢ t P =
is
j
* = 4 j =
x
q 5 = j 5
=." he —_
: : =
ek
Fee - sd
= ‘ -
.
e
‘
‘ .
’
Ge
,
r of
i
\
i
(
la .
‘ =. =
-
‘
U if
’
. o ie
\ y
oo ow . . '
. ’ =

Hencuman. — Nervous System of Limax.

PLATE VIL

All the figures of this plate, except Fig. 95, were made from material killed in
0.33% chromic acid, and stained in alcoholic borax-carmine. All figures are
magnified 250 diameters.

Figs. 81-82. (See explanation of Figs. 48, 49, Plate III.)

“ 81. The 114th section of the series; it shows the right cerebral and pedal
ganglia, together with the cerebro-pedal connective, the right buccal
ganglion, and the abdominal ganglion.

“ 82. The 126th section of the series. It passes through a portion of the right
cerebral ganglion, the right pleural and the abdominal ganglia, from
the last of which a nerve runs dorsalward.

« 83,84. The left surface of two sections (the 96th and 99th) parallel to the

sagittal plane from an embryo of the twelfth day.

88. The section shows a very small portion of the cerebral ganglion, the
pleural and visceral ganglia of the right side in the region of the
pleuro-visceral connective, and the abdominal ganglion together with
its connective with the right visceral ganglion.

84 shows the right visceral and the abdominal ganglia; also, a portion of the
connective between the abdominal and the visceral ganglia, and a large
nerve extending dorsalward from the latter.

«“ 85-94. (See explanation of Figs. 50, 52, Plate III.) Transverse sections,

twelfth day.

* 85. This section (113th) shows a portion of the abdominal ganglion sepa-
rated from the ectoderm.

‘© 86 (117th section) shows a small portion of the abdominal ganglion, as
well as the pleural and visceral ganglia of the right side, together
with their connective.

“« 87. (126th section; compare also 125th section, Fig. 88.) A portion of the
right visceral ganglion and the large nerve running dorsalward from
its left dorsal margin are shown.

“ 88 (125th section) shows both visceral ganglia — the left one still con-
nected with the ectoderm — and the left pleural ganglion, together
with the pleuro-visceral connective.

‘“ 89. The 128th section, which touches the right visceral ganglion, and a large
nerve running dorsalward between mantle cavity and sexual duct
from the right dorsal margin of the ganglion.

“ 90. The 77th section; it passes through both pedal ganglia and their poste-
rior commissure, which is directly above the blind end of the pedal
gland, the tip of which is cut.

“ 91. The 103d section, which shows the right pedal ganglion and otocyst.

“ 92. The 101st section, which passes through the pedal ganglia and their an-
terior commissure, above which is the radula sac, and below which a
blood-vessel and the pedal gland are to be seen.

(See obverse )

PLATE VII. (continued.)

Fig. 93. The 109th section; this passes through the abdominal ganglion at the
place of its connection with the ectoderm lining the median wall of
the mantle cavity. It also shows a portion of the connective from
the abdominal ganglion to the left visceral ganglion.

“ 94, The 146th section of the series; it shows a cross section of the neck of the
cerebral invagination, and a small portion of the cerebral ganglion of
the left side.

“ 95. The left surface of the 54th section, from an individual of the fourteenth
day. The section passes through the right visceral and the abdomi- .
nal ganglia, showing their close connection with each other at this
time, and it also cuts the right buccal ganglion. The material was
killed in Perenyi’s fluid, and stained in picro-carminate of lithium.

The following figures are drawn from sections of the same series as
Fig. 95: Plate VIII. Figs. 96, 101, 102; Plate IX. Fig. 114. The
sequence of sections is this :—

Section 52, 54, 56, 66, 82.
Figure 101, 95, 102, 114, 96.

HENCHMAN.- NERVOUS SYSTEM OF LIMAX.

in)

a

~

WeNcuMAN. — Nervous System of Limax.

PLATE VIII.

Fig. 96. (See explanation of Fig. 95, Plate VII.) The 82d section; it passes

ce

ce

through the ocular tentacle, showing in section the cerebral invagi-
nation and ganglion of the right side. » 100.

97-100. Posterior faces of transverse sections, the right side a little in ad-
vance of the left, from an embryo of the fourteenth day. Killed in
Perenyi’s fluid, stained in picro-carminate of lithium.— Figs. 104,
106; also Plate IX. Figs. 115-120, and Plate X. Figs. 121, 128, 125,
126, belong to the same series as Figs. 97-100. The sequence of
sections is : —

Section 81, 87, 92, 98, 102, 104, 109, 111, 113, 113, 115, 121,
®igure, 97,119; 98, 99, 118,. 100,115, 116, 104, 117, 123, 120,
1235. 122,497,139.
106, 125, 121, 126.

97. The 81st section ; it passes through the pedal ganglia, a few sections be-
hind the posterior commissure. X 100.

98 and 99. The 92d and 98th sections; they pass through the pedal ganglia
between the two commissures. X 100.

100. The 104th section, two sections in front of the anterior commissure. It
shows the right otocyst in addition to the radula sac and pedal gland.
x 100.

101, 102. (See explanation of Fig. 95, Plate VII.)

101. The 52d section of the series, showing the abdominal ganglion, and a
cross section of the cerebral, buccal, and both pedal commissures.
x 100.

102. The 56th section; it passes through the visceral and buccal ganglia of
the right side, and a portion of the abdominal ganglion. It shows
the cerebral and pedal commissures, as well as a sagittal section of the
foot gland. X 100.

103, 103. The posterior surfaces of transverse sections from an embryo ot
the seventeenth day. 0.83% chromic acid; alcoholic borax-carmine. —
Additional sections from this series are shown in Figs. 105, 108-113,
and Plate X. Figs. 122, 127. The sequence of sections is indicated by
the following : —

Section 111, 176, 180, 186, 187, 194, 211, 212, 217, 226, 225.
Figure 108%, 109, 110, 111, 112, 113, 108, 105, 122, 108, 127.

103. The section passes through the cerebral ganglia in the region of the
cerebro-pedal connective. The wall of the body is not represented,
but merely the ganglia, together with the esophagus, the ducts of the
salivary glands, the radula sac, and the right ocular tentacle. x 140.

103%. This section (111th) passes through the pedal, pleural, and visceral
ganglia in the region of the pleuro-pedal connective. A large nerve
passes from the dorsal margin of each pedal ganglion to the lateral
wail of the body. X 83.

(See obverse. )

Fig. 104.

66

105.

106.

107.

PLATE VIII. (continued.)

(See explanation of Figs. 97-100.) The 118th section; it passes through
the pleural ganglia and the abdominal ganglion. A large nerve con-
nects the abdominal ganglion with a pocket-like infolding from the
wall of the mantle cavity. X 100. Figure 117 (Plate IX.) shows a
portion of this section more highly magnified.

(See explanation of Figs. 103, 103%.) The 212th section; it passes
through the cerebral commissure. X 100.

(See explanation of Figs. 97-100.) This section, the 122d, passes
through the right cerebral invagination, the right cerebral and buccal
ganglia with their connective, and the right visceral ganglion. It also
shows the left pleural and visceral ganglia with their connective.
x 100.

Is a combination of two successive sections cut parallel to the sagittal
plane from an embryo of the stxteenth day. It shows the cerebral,
pedal, and pleural ganglia, with their connectives, and the otocyst of
the left side. The position of the pedal gland is shown by dotted
lines. Perenyi’s fluid; picro-carminate of lithium. 100.

108-113. See explanation of Figs. 103, 1033.

108.

109.

110.

111.

112.
115.

The 225th section of the series shows the cerebral invaginations. ‘The
right one (an enlarged view of which is seen in Fig. 127) is cut length-
wise, it being still open to the exterior; the left one transversely.
The cerebral ganglia and their lateral lobes, and the buccal ganglia
with their commissure crossing between the radula sac below and the
cesophagus above, are also shown. %X 88.

The 176th section, which passes through the pedal ganglia and the ab-
dominal ganglion. X 83.

The 180th section; it shows, in addition to the organs seen in Fig. 109, a
smali portion of the left visceral ganglion and the otocysts.

The 186th section ; it passes through the pedal ganglia, a portion of the
pleural ganglion of the right side, both visceral ganglia, and the ab-
dominal ganglion. It also shows the connective from the abdominal
to the right visceral ganglion, and a stout nerve arising from the
latter. X 83.

The 187th section shows the pedal, pleural, and visceral ganglia. X 83.

The 194th section. This shows, in addition to the ganglia, the nerve
which arises from the left visceral ganglion. X 83.

AN.- NERVOUS SYSTEM OF LIMAX.
4

96.

cele

nupa
O-pa p-

com buc. ~ _.
: rf

B Meisel, lith.Boston.

Pbk ve oY avy ee
id dlonhsbh bie Soot
Nie? wi)
; .

Mey Dar + ¥ a Lava By

See Lele CaP hci Visk wily “et!

Bie iAP) / 1)

HER wt;
lied aa i ve

a le! ,

e ok

es i pate wie ot} ine)

wid MY 0
wok gay i a)

wis stn vit

Org hp

PV OV an
an wit ¥ anh,

(tea fia &

pall PP ba) Le as Aide Ad
Ay) CMLEME: ANE Ne we
ke alg Mri 5’ Dep ke Ve oD its whi aie ibe
. i ' F hea
«A Re ee ity
OPT, 5 hig4 eihi8 na a , ts I ;
isan “ : hl ict y \4 Eas
de Whi, LA dae Sea iin Bed ae
’ 2) : ‘
—> 4) : a i
, 4 +S )
Tid Ae Dh dae ga A coe Ae TUL an | ae
PHOERY. Peal | (ane Ta “as al
Mi :
re ia caiy! i La a
at tT
iow Wey j i “a i
=
ij = AC. A
f ! ~ * Pal
.
{
J
a
—
3
: i) , f

“y
‘ j

5

.

PAA BOA

aM

ane wy lif. iis ae se) ‘ae ule
oye vl heynne! nd ae Cony

at Pra wn

a mS

fe ny
1 af
ey |
Mi P 1
Py
1)
av ¥
iP!
i
Ms
ie i
i ith
; j ¥;
‘
Vy f
ne ete
vit im.
wi ny
am t
=e wy
i : ty
hon ~
'
' ;
; y
, ‘
‘ ar
he
cy
\
ey
4
s
¥
AL P
Ro
‘E ra
Ey ‘
:
J
-
\
}
‘
< :
j ag

HencuMan. — Nervous System of Limax.

PLATE IX.

All the figures of this plate are magnified 250 diameters, and were made from
material killed in Perenyi’s fluid, and stained in picro-carminate of lithium.

Fig. 114.

(See explanation of Fig. 95, Plate VII.) The 66th section; it passes
through the cerebral, the pedal, the pleural, and the visceral ganglia
of the right side in the plane of the cerebro-pedal and cerebro-pleural
connectives. It also shows the right otocyst.

“115-120. See explanation of Figs. 97-100, Plate VIII.

REL:

o6.

me tie

«120.

The 109th section; it shows a portion of the abdominal ganglion at the
right of the radula sac.

The 111th section ; it passes through the abdominal and right pleural
ganglia.

The 118th section; it shows the abdominal ganglion where it passes
above the radula sac, and a portion of the right pleural ganglion.
(Compare Fig. 104, Plate VIII.)

The 102d section; it passes through the pedal ganglia in‘the plane of
their anterior commissure. It also shows the right otocyst.

The 87th section. The pedal ganglia in the plane of their posterior
commissure.

The 121st section, which passes through the visceral and buccal ganglia
of the right side, and shows a portion of the buccal commissure.

A
HENC HMAN.—NERVOUS SYSTEM OF LIMAX. PLIX.

<I

B Meisel, ith. Baston.

‘hone

‘ '
a a

he)
boven

ee) if
v7

3

a ave

Rep

we hid)

|
"
>

:

Fal i mir | 7%

te
6
‘hig }
% ad 5
he +e wa ieee
i
'
¢
.
. i
, a
‘
iY
~e rl
wk, Dea
¥ a a
¢ {2004 Fee Maas ,
pie ae rare Nis
&\ ‘MA >,
. >
7 j } are a q
sald 4 fay yee PAY Wh fiat}
L nar 4 We
" P - ie a ty |
.
, «
1) ‘ | ‘Tae if wey \
y (7) 3 - "Aes ‘) ei "4
i F
Ye ely eee —19y
‘ r
Zo
fr ak i) A 4)
4 ,
>
7 4
4 ,
+ “a : } Lis 3 iA |
¢ en al ora. ate hy
of ‘ ’
, . i T) “eb! j ‘lL
F ‘ > .
oe vt bai
’
it ‘
\ ie ¥
i a
' ‘s a
’ ; Wig »
wr j j
: a
a
, \F
‘
5 ‘
—
=
j
‘
‘
’
,
s]
>
)
i
‘ :
—

¥

bear af

~~

7. ap» ce

a

a

*
>)

2
-_

\
ae 6 Ons

ee

HENCHMAN. — Nervous System of Limax

PLATE X.

Fig. 121. (See explanation of Figs. 97-100, Plate VIII.) The 127th section; it

passes through the left cerebral invagination, also through the left
cerebral and buccal ganglia, and their connective. X 237.

122. (See explanation of Figs. 103, 1039, Plate VIII.) The 217th section of
the series. It passes through the right cerebral ganglion and its lat-
eral lobe; it also shows the ocular tentacle and the wall of the right
eye, transverse sections of the cesophagus, salivary glands, radula sac,
and primary sexual duct. X 237.

125. (See explanation of Figs. 97-109, Plate VIII.) This section passes
through the right pleural and visceral ganglia, a small portion of the
left pleural and visceral ganglia, and the abdominal ganglion, which
lies between the cesophagus and radula sac. X 250.

124, ~The posterior face of a transverse section from an embryo of the siz-
teenth day. The left side is cut a liitle in advance of the right. The
section passes through the cerebral invagination, — still open to the
exterior, — the cerebral ganglion of the left side and its lateral lobe,
and the left buccal ganglion. It also shows in cross section the duct
of the salivary gland, and a small portion of the wall of the radula
sac. X 237. Perenyi’s fluid; picro-carminate of lithium.

125, 126. See explanation of Figs. 97-100, Plate VIII.

125. The 122d section; it shows the left pleural and visceral ganglia, with
the pleuro-visceral connective, and a small portion of the left cerebral
ganglion. X 250.

126. The 189th section; it shows in section the left cerebral invagination,
the cerebral ganglion, and the cerebral commissure. X 250.

127. (See explanation of Figs. 103, 103°, Plate VIII.) The 225th section; it
shows the cerebral invagination, and the right lateral lobe of the
brain. (Compare Plate VIII. Fig. 108.) x 287.

HENCHMAN.-NERVOUS SYSTEM OF LIMAX.

No. 8. — The Parietal Eye in some Lizards from the Western United
States. By W.E. RITTER

WirH a single though notable exception, the numerous authors who
have written on the parietal organ in vertebrates since the papers of
de Graaf: (’86* and 86”) and Spencer (’86 and ’87) appeared, have
agreed that the structure is, or at least was in ancestral vertebrates,
an eye. This belief is based entirely on the structure of. the organ, no
physiological experiments or observations on the habits of the animals
possessing it having yet been produced in proof of its function.

Leydig (89) alone, in a recent preliminary paper on the subject, has
denied its optical nature, and has assigned to it an entirely different
function; though in a second preliminary, still more recent (90), he
expresses his denial with considerably less confidence. He rejects the
eye hypothesis, however, on the same grounds that have led others
to adopt it; namely, on the grounds of its structure, and especially
of its relation to the brain.

He believes that what is generally held to be an optic nerve is in fact
merely a string of connective tissue. .

Among those who believe the organ is or has been an eye, there are
important differences of opinion as to its present value. By Ahlborn
(784), de Graaf, Spencer, and several other more recent writers, it is
believed to be degenerate and entirely functionless in all living verte-
brates. Rabl-Riickard (’86) has expressed the opinion that the organ
may still be of use in furnishing its possessors with a more delicate
means of detecting differences of temperature than exists elsewhere on
the body. Béraneck (’87) believes that, while the structure is probably
of an optical nature in some vertebrates, it has become so secondarily ;
that the primitive function of the epiphysis, common to the brains of
all vertebrates, was something entirely unknown to us now, though not
concerned with vision; but that in the Cyclostomes, the Amphibians,
and the Reptiles it has taken on, secondarily, the function and form
of an eye.

1 Contributions from the Zodlogical Laboratory of the Museum of Comparative
Zoology, under the direction of E. L. Mark, No. XXII.
VOL, Xx. — NO. 8. 14

210 BULLETIN OF THE

But even were it established beyond question that the organ is a
degenerate eye, there would still remain several quite distinct and very
interesting problems to be solved. The most fundamental of these is
probably that of its homology. Much has been written on this ques-
tion by the various recent authors, but even less unanimity of opinion
has been reached here than on the question of its structure and fune-
tion. The question why the organ has remained so well developed in a
few systematically widely separated groups of vertebrates, while in all
others the process of degeneration has gone so far as to leave but a
mere trace of the proximal portion of the epiphysis, has not been
much discussed. It is not my purpose in the present paper to enter
upon a discussion of the theoretical questions involved, and they are
here adverted to merely to point out the need —as indicated by their
importance and the discordance of the opinions now held with regard
to them — of a larger body of facts on the subject than we yet possess.
For the present, I confine myself to a presentation of the facts observed,
and my interpretation of them as bearing upon some of the minor
- conclusions reached by other writers, hoping to be able to pursue the
subject further in the near future, when situated in a region where an
abundance and a variety of material, adult and embryonic, can be ob-
tained, and where observations on the habits of the animals can be
made.

The present work was undertaken at the suggestion of Prof. E. L.
Mark. I wish here to acknowledge my indebtedness to Mr. G. H.
Parker, of the Museum of Comparative Zodlogy; to Mr. J. J. Rivers,
Curator of the Museum at the University of California; and to Mr. T.
C. Palmer, of the United States Department of Agriculture, Wash-
ington, for material used ; and also to Mr. S. Garman, of this Museum,
for assistance in determining the species studied.

A word as to technique. For studying the structure of the retina it
is very desirable to remove the great quantity of pigment that inva-
riably obscures the histological elements in this region. Neither nitric
nor hydrochloric acid, nor the alkalies, have any visible effect on this
pigment, but the desired result was reached by the use of chlorine gas.
The mounted, unstained sections were covered by a film of ninety per
cent alcohol, and placed in a tight glass chamber, in which was also
confined a small vessel containing a mixture of potassium chlorate and
hydrochloric acid for generating the gas. By being careful that the
slide on which the sections were mounted occupied a perfectly horizon-
tal position, and was so placed that the film of alcohol could not be

Pi a

——

MUSEUM OF COMPARATIVE ZOOLOGY. yaal

drawn off by capillary attraction, the film soon became saturated with
the gas, and did not need renewing. From forty-five minutes to an
hour, depending on the quantity of pigment, was sufficient time in
which to accomplish the work. Considerable difficulty was found in
removing the chlorine from the sections. As it had thoroughly pen-
etrated the tissue, simple washing, even though prolonged, did not
- wholly remove it; but by washing carefully, and then leaving the whole
slide immersed in ninety per cent alcohol for twelve or fourteen hours,
the gas was entirely removed. A good quality of Schallibaum’s fixative
held the sections perfectly through all this and the subsequent staining.

For decalcifying and hardening the tissues I have found Perenyi’s
fluid more satisfactory than anything else tried, the two processes being
accomplished at the same time by this reagent.

Of the several species of lizards which I have studied I shall describe
the structure in only three, namely, Phrynosoma Douglassii, P. coronata,
and Uta Stansburiana, these being the only ones that have presented
anything new or of special interest.

Phrynosoma Douglassii.

1. External Appearance. — Concerning the external appearance of the
organ little need be said, since it differs in no essential particular from
what has been amply described and illustrated in numerous other liz-
ards. The scale marking the position of the eye is quite conspicuous,
especially in very young individuals, where it is of a rather lighter color
and larger size, relatively, than in the adult. In old individuals the
great development of the surrounding scales and tubercles renders it
somewhat less noticeable than it otherwise would be, but it is always
readily distinguished, not only by its median position, but also bv the
absence of pigment and by its translucent appearance.

2. The Parvetal Vesicle. — Figure 1, drawn from a sagittal section
through the dorsal wall of the head, shows the form of the vesicle and its
position within the parietal foramen and with reference to the external
and internal surfaces of the wall. It lies within the parietal foramen,
though extending somewhat above the dorsal surface of the parietal
bone, firmly embedded in connective tissue, so that when the wall of the
head is separated from the brain the vesicle always goes with the former.
The tissues composing the dorsal wall of the head are, excepting the
corneous layer of the skin, quite different immediately over the vesicle
from those of the surrounding regions. The epidermal layer of the skin

212 BULLETIN OF THE ‘

elsewhere sends down irregular cone-shaped masses, which penetrate
and become lost in the underlying connective tissue, thus firmly uniting
the two layers. Over the vesicle, however, these processes are wholly
wanting, the under surface of the epithelial layer being even, and
sharply limited from the connective tissue. These processes are espe-
cially well developed immediately beyond the margin of the disk of
the vesicle, where they carry the cells of the epidermal layer (e’drm.)
considerably deeper than their general, level. ‘The connective tissue
between the vesicle and the epidermal layer is composed of fibres con-
siderably finer and looser than those found in other places, and, further-
more, the fibres are here disposed at various angles to the surface of the
skin, whereas elsewhere they are approximately parallel to this surface
(con’t. tis.'). Pigment, which is found in great abundance in the skin in
all other regions of the body, is always entirely absent here. It will
thus be noticed that each of the tissues over the vesicle is considerably
more penetrable to light than are the corresponding ones elsewhere.
The connective-tissue fibres immediately around the vesicle are arranged
concentrically to its surface, and are, especially in the proximal two-
thirds of their extent, considerably finer and closer than elsewhere.
A kind of capsule for the vesicle is thus formed, and it is this alone
which separates it from the cranial cavity. The fibres of a string of
tissue extending from the distal end of the epiphysis can be traced,
though with some uncertainty, to this capsule, but I find no indication
of their passing through it, or even entering it, though I have given
special attention to this point. 3

The internal surface of the cranial wall in the region of the vesicle
presents a depression, which is much less marked, however, than a cor-
responding one in P. coronata, to be referred to hereafter: Junning
through the connective tissue at the bottom of this depression, and
hence. near the deep surface of the vesicle, are found a number of blood-
vessels of considerable size and well filled with blood corpuscles (va. sng.).
The vesicle itself is elliptical in sagittal section, the major axis, 258 p
long in the specimen figured, having the direction of the long axis of
the head. In transverse section it is slightly elongated dorso-ventrally,
and measures in this axis 171 p. .

The cavity in sagittal section shows a triangular outline, the base of
the triangle being on the dorsal or lens side. From this outline in the
sagittal section the form gradually changes to that of an ellipse in
the last sections on each side that cut the cavity; so that the form of
the cavity is approximately that of a broad, flat cone, the base directed

MUSEUM OF COMPARATIVE ZOOLOGY. 213

outward and the apex inward. The base of the cone is slightly con-
cave, corresponding to the convexity of the inner surface of the lens.

The wall of the vesicle is very distinctly differentiated into lens (/ns.)
and retinal (rtn.) portions, the latter forming about two thirds of the
whole. The lens is slightly biconvex, the two convexities being very
nearly equal. The line of demarcation between the lens and the retina is
a sharp one, though the two portions are plainly continuous. The cells
composing the lens are large and distinct in outline, each one extending
entirely through its thickness (Plate II. Fig. 5, el. dms.). Their nuclei
are large, easily stainable, and somewhat granular ; they are uniformly
situated near the internal ends of the cells. The lens is entirely with-
out pigment.

Figure 5 represents a highly magnified portion of a longitudinal
vertical section of the vesicle taken from near the median plane. In
the retinal portion six regions or zones may be distinguished. Passing
from the external surface toward the cavity, we find (1) a basement
membrane (mb. ba. ex.). This is very thin, but uniform in thickness,
and is of a structureless nature. From many points on this membrane
fine processes radiate into the connective tissue enveloping the vesicle
(Plate I. Fig. 3, pre. 7.). These processes do not appear to be of a mus-
cular nature, but rather the same in structure as the basement membrane
from which they arise. (2) A zone containing a few scattered nuclei
(nl.'), and fine-grained sparsely but evenly distributed pigment (pzg.).
No cell boundaries can be made out in this zone. ‘The nuclei, few in
number, form a single layer, and are situated near the basement mem-
brane. They are very nearly round, exhibiting no tendency to elongate
in the radii of the vesicle. Areas in their centres, which are somewhat
raore deeply stained than the rest of the nuclei, and which are probably
nucleoli, are to be seen. (3) A zone (z./’) in which are distinguish-
able neither cells, nuclei, nor pigment; only a uniform, fine-granular,
slightly stainable substance, of much the same nature, apparently, as
the cell substance in those regions of the retinal portion in which cell
boundaries can be distinguished. Whether or not this zone repre-
sents the centrally directed ends of a layer of cells, the nuclei of which
are the ones found in zone 2, I am unable to say, but it probably
does. (4,5) The next two zones are distinguished from each other
only by the difference in the elements composing them, no distinguish-
able line of separation existing between the two. The most obvious
difference between the constituent elements of these two regions is in
the shape of the nuclei, those in zone 4 being approximately spherical

214 BULLETIN OF TUE

(nl.!'), while those in zone 5 are much elongated in the radii of the ves-
icle (n/.!’). The suggestion at once comes that this difference is due
solely to the crowding together of the cells nearest the internal surface
of the retina, and hence that the two zones should in reality be re-
garded as but one. If, however, the difference in shape of the nuclei
were the result solely of such crowding, we should find a complete
gradation from the spherical to the elongated form in passing from
without inward ; but such a gradation is not found in fact. Further-
more, on close examination with high powers, it is found that the nuclei
differ in structure as wéll as in form. An irregular stellated area can
be detected in the centres of some of the spherical ones which does not
exist in the elongated ones; also, the entire substance of the former
is slightly more granular than that of the latter. In the fifth zone
cell boundaries (though not well shown in the figure) can be quite dis-
tinctly traced to the internal basement membrane ; but how the cells
of the fourth and fifth zones are related I have been unable to deter-

mine, since cell boundaries in the fourth zone cannot be traced. (6) The

last layer may be designated as an internal basement membrane (mod.
ba. 2.), though it differs somewhat in structure from the external base-
ment membrane, being of a granular nature. It extends over the
surface of the lens, as well as over the retina, and is rather more com-
pact in the former than in the latter region. Projecting into the cavity
of the vesicle from the retinal portion are found certain structures con-
cerning the nature of which Iam not quite sure, but believe them to
be secretions from the cells of the fifth zone. They are in general
elongated, and pointed at their free ends, though their outlines are
ragged and indefinite. They always stain most deeply at their internal
free ends. In many cases, as at *, they are seen to be continuous with
the cells of the fifth zone through the internal basement membrane.
These structures may correspond to what de Graaf has described and
figured as existing on the internal surface of the retina’of Anguis, and
has called “Staafjeslaag,” but which Spencer and others believe to
be merely a coagulum from the fluid that probably filled the cavity in
the recent state. It is, however, scarcely possible to account for the
structures here under consideration in this way, as is to be seen from
my description and figures of them; furthermore, a.coagulum (cog.)
does exist in addition to these.

Within the substance of the retina (Fig. 5, va. rtn.) are found a num-
ber of cavities varying in diameter, as measured in the plane of the
sections, from 5.5 w to 22. The sections of these cavities are never

MUSEUM OF COMPARATIVE ZOOLOGY. 215

quite circular, but are never much elongated. In many, though not in
all, an exceedingly thin endothelial lining can be seen, and in a few in-
stances blood corpuscles are found in the cavities (Plate I. Fig. 4, en’th.
va. and cp. sng.). Although none of these cavities were found to extend
through more than four or five sections, each 7.5 » in thickness, and
although in no instance was it possible satisfactorily to trace a connec-
tion between them and the blood-vessels lying outside the vesicle, it still
seems quite certain that they form a network of fine blood-vessels rami-
fying through the substance of the retina. Owing to the fact that in some
instances no lining membrane to these cavities can be found, and that
their outlines are not sharply marked, the possibility of their having been
artificially produced by the removal of pigment masses suggests itself ;
but the definiteness of the outline of many others and their endothelial
lining membranes, in which much-flattened nuclei are found, strips this
conjecture of its plausibility. If these are really blood-vessels, it might
appear that some of them would be seen cut longitudinally ; and while
it is true that in many cases focusing shows the cut walls to be very
oblique to the plane of the section, still no sure instance of a vessel
cut lengthwise has been seen. When, however, one considers the
exceeding delicacy of the endothelial lining, and the fact that no differ-
ential staining takes place, it does not seem impossible that such sec-
tions may exist, and yet escape detection. These cavities have no
regularity of arrangement, but are for the most part confined to zones
2, 3, and 4. In no instance has one been seen confluent with the
cavity of the vesicle.

These may possibly correspond to what Owsjannikow mentions as
having been seen by him in Chameleon vulgaris. He says: “Am hin-
tern Rande der Retina findet sich an einigen Schnitten das Lumen
eines Rohrs, von dem nicht mit Bestimmtheit gesagt werden kann, ob
es einem Blutgefiasse oder einem anderen Gewebe angehért.” (Owsjan-
nikow, ’88, p. 16.)

3. The Epiphysis. — Figure 9 (Plate III.) represents a sagittal sec-
tion of the epiphysis, and so much of the brain as is necessary to show
the relation of the former to the latter. The entire structure, or, more
properly, the combination of structures that must be considered at this
time, presents the form of a curved cylinder, one end of which is pro-
duced into a cone, while the other end has a hopper-shaped excava-
tion. In keeping with the usual method of designation, I shall call the
whole structure the epiphysis, though, as the sequel will show, it is
doubtful if this is justifiable. The excavated end is proximal, the

216 BULLETIN OF THE

excavation being the continuation of the cavity of the third ventricle
into the epiphysis. The conical end, then, is distal, and rises somewhat
above the level of the cerebral hemispheres. The curved axis forms
very nearly a segment of the circumference of a circle, and is directed
upward and forward from its point of origin from the brain. Continu-
ing anteriorly from the apex of the cone is a string of connective tissue
(con’t. tes.), which passes to the region of the parietal vesicle, and in the
distal portion of its course comes close in contact with the dura mater
of the brain. The axis of the cylinder, if we consider it as continued
to the anterior termination of this connective-tissue string, describes
very nearly a semicircumference. The most anterior point in the con-
nection of the epiphysis with the brain is at the junction of the cere-
brum with the optic thalamus, somewhat anterior and dorsal to the
superior commissure (com. su.). For a short distance above its connec-
tion with the brain in this anterior part, the epithelial nature of the
epiphysial wall is less distinct than at a higher level, where the wall
becomes thicker, and is composed of a single layer of more or less cuboid
nucleated cells, which stain readily in borax carmine or hematoxylin
(Plate IIT. Figs. 8, 9, e’th.). Also at this level the wall becomes thrown
into a highly complicated system of folds; and it is this folded epithe-
. lium, containing within its folds great quantities of blood corpuscles,
that forms a large bulk of the whole epipbysis (Figs. 8 and 9, e’th.
and cp. sng.).

In the section represented in Figure 9 no connection exists between
the epithelium of the posterior portion of the epiphysis and the brain,
and it is doubtful if such connection exists here in any of the sections
of this specimen; at any rate, if it does exist, it is exceedingly thin
and limited in extent. There is, however, an undoubted connection in
this region in P. coronata, which will be described later ; but even in
this latter species the posterior wall of the epiphysis is much less de-
veloped than the anterior wall. The exceedingly thin epithelium that
forms the posterior wall in P. Douglassii would, as is evident from its
position and from comparison with P. coronata (Plate IV. Figs. 11 and 12),
form a connection with the brain roof had not a separation taken place,
either artificially or as a result of degeneration. This wall is closely
applied to the anterior, concave side of the blood sinus to be presently
deserjbed, and at a considerable distance above the brain is continuous
with the anterior wall of the epiphysis. The space included by these
walls is the hopper-shaped excavation in the proximal end of the cyl-
inder already mentioned, — an extension of the cavity of the third

»—.4

MUSEUM OF COMPARATIVE ZOOLOGY. 217

ventricle (vzt.?) into the epiphysis. Intimately connected with the dis-
tal end of the portion of the epiphysis thus far described is found a
vesicle (eph. vs.), the thick walls of which are composed of columnar
epithelium, and thus differ markedly from the folded epithelium of the
anterior wall previously described. ‘This vesicle is much flattened
antero-posteriorly, its longest axis lying very nearly in the axis of the
cylinder to which the epiphysis as a whole has been compared. That
the structure here described is a separate vesicle, and that its cavity
is not continuous with the cavity already described as a continuation
of the third ventricle, admit of easy and satisfactory demonstration,
not only in this particular instance, but also in all other individuals
both of this species and of P. coronata of which sections have been
made. In passing through the entire series of sections, it is easily seen
not only that the two cavities nowhere approach more nearly to conflu-
ence than in the one represented in the figure, but also that the walls
of the vesicle and those of the more proximal part of the epiphysis with
which they are in relation are clearly distinct. The separateness of
these two structures will appear more clearly when we come to consider
the same parts in P. coronata. Passing upward and forward from the
distal end of this vesicle is to be seen a bundle of connective-tissue
tibres which becomes blended with the string of connective tissue already
described as running from the apex of the cone to the region of the parietal
vesicle. There is no indication that the epithelial wall of the epiphysial
vesicle, as it may be called, passes into this string.

Covering the whole postero-dorsal convex side of the portion of the
epiphysis thus far described, and even extending considerably beyond
its distal extremity, is an immense blood sinus fulty distended with
blood corpuscles (Fig. 9, sn. sng., and Fig. 8, cp. sng.).

Phrynosoma coronata.

1. General Description. — Figure 2 (Plate I.) represents a transverse
section of the dorsal wall of the head, passing through the middle of the
parietal eye of P. coronata. The description of the external appearance
and of the vesicle and its surrounding structures given for P. Douglassii
requires modification in only a few points to become applicable to this
species. The depression mentioned as existing on the internal surface of
the wall of the brain-case immediately under the vesicle in P. Doug-
lassii becomes in this species a deep pit. To correspond with this
pit the external surface of the wall immediately over the vesicle forms

218 BULLETIN OF THE

a low, broad cone, a condition which gives quite a different general ap-
pearance to the sections in the two species. In P, coronata the vesicle
is situated somewhat nearer the external surface of the cranial wall than
in P. Douglassii; and the intervening connective tissue differs less,
both as regards the fineness and direction of its fibres, from the adjacent
tissues, than in the case of P. Douglassii. The vesicle, with its con-
. nective-tissue capsule, protrudes into the bottom of the pit considerably.
The pit is bridged over by the dura mater of the brain, and thus a
chamber is formed in which a great quantity of blood corpuscles is
found (cp. sng.). It will be remembered that no such blood sinus in
this region exists in P. Douglassii, but that numerous blood-vessels do
occur here. In P. coronata, however, the sinus replaces the vessels.

2. The Parietal Vesicle. — With regard to the vesicle itself, the only
points in which it differs very essentially from that found in P. Douglassii
are the absence of the cavities in the retina regarded as blood-vessels, and
the far less perfect development of the structures projecting from the
internal surface of the retina into the cavity of the vesicle. The latter
difference I am inclined to think due to the probably somewhat. greater
degree of degeneration of the retinal cells which secrete these struc-
tures. That this portion of the retina is more degenerated in P. coro-
nata may be supposed from the fact that we find here considerably
more pigment than in the corresponding region in P. Douglassii. How-
ever, too much stress must not be laid on the greater or less quantity
of pigment, since the quantity is quite variable even within the same
species. In one individual of this species pigment was found, though
in small quantity, in the lens.

3. The Epiphysis. — Although this structure does not differ in any
essential particular from what we have already seen in the preceding
species, the fact that several of the points which go to make the study
of the epiphysis of much interest are here well brought out, has made it
seem best to describe and illustrate the organ in detail. Figures 10, 11,
and 12 (Plate IV.) present vertical longitudinal sections from the same
animal at different planes to the left of the median plane, Figure 12
being very nearly median, and Figure 10 farthest removed from it. It
should here be said, however, that the sections are not quite vertical ;
so that, while the epiphysial vesicle is situated more to the left than to
the right side of the sagittal plane, yet it is less so than would be
inferred from the way in which it appears in the figures. The form
of the epiphysis, as a whole, is nearly the same as that found in P.
Douglassii, and it is composed of the same parts ; — namely, a proximal

MUSEUM OF COMPARATIVE ZOOLOGY. 219

part with an anterior much-folded epithelial wall, and a posterior not
folded and thinner epithelial wall; an epiphysial vesicle ; a blood sinus ;
and a string of connective tissue extending from the distal end of the
_ vesicle and blood sinus to the region of the parietal vesicle. In the
anterior wall of the proximal portion the folding extends down some-
what nearer to the brain than is the case in P. Douglassii, and just at
its junction with the brain a large blood-vessel is found filled with
blood corpuscles (Fig. 12, cp. sng.). As already said in describing the
posterior wall in P. Douglassii, the connection (opposite the letters
ont.*) with the brain is here complete and very evident, though the roof
of the third ventricle (¢ct. th/. opt.) appears in the section to constitute
a part of this wall.

The cells composing the walls of the proximal part are about two or
three deep, but not arranged in layers. They are small, distinctly nu-
cleated, and the nuclei are apparently perfectly round. They stain
readily. On the outer surface of this wall is found, throughout most
of its extent, a very thin layer of tissue, the cells of which are much
flattened. This layer becomes continued from the apex of the epiphysis
as the connective-tissue string (con’t. tis.) already mentioned as passing
to the region of the eye; another portion of it also becomes continuous
with the pia mater of the brain.

Figure 10 represents a section through the longest portion of the
epiphysial vesicle. In this plane the proximal portion of the epiphysis
has not yet appeared, and is not found till we pass to a section in
which the long axis of the vesicle has become considerably shortened.
In the wall of the vesicle three zones or layers are found. The external
one is similar to— in fact, on the posterior surface is continuous with
—the thin external layer mentioned in the proximal portion. The
second zone, comprising more than half of the entire thickness of the
wall, is composed of cells apparently of the same nature as those
described as forming the chief portion of the wall of the proximal part ;
but the layer is considerably thicker here than there, and on the whole
rather more compact (e’th., Figs. 10 and 11). The third and most
internal zone is a deeply pigmented one (pig.). This pigment is so
dense that when destroyed no distinguishable structure remains. In
the presence of this pigment the species now under consideration differs
entirely from P. Douglassii, where no pigment in this region is found.
Again, however, attention is called to the fact that great importance
cannot be attached to the presence or absence of pigment. Figure 11
shows the relation between this vesicle and the proximal portion of the

220 BULLETIN OF THE

epiphysis. In this section it will be seen that a distinct line of demar-
cation exists between the true epithelial portions of the two walls
where they come in contact. This distinctness is maintained through-
out the entire series of sections. When the median section is reached,
the vesicle has entirely disappeared. From the distal end of the vesicle
the connective-tissue string extends forward to the region of the eye, as
in the case of the proximal portion (con’t. tzs.). The blood sinus (Fig.
12) does not, in this species, come in contact with the epiphysial vesicle,
but occupies the same position on the proximal part as in the case of
P. Douglassii. It is much smaller in P. coronata, but in other respects
is of the same nature. Whether or not this epiphysial vesicle may
be homologized with the secondary vesicle in Petromyzon (Ahlborn,
83, Beard, ’89, Owsjannikow, ’88, Wiedersheim, ’80) can be profitably
discussed only after its development has been studied. So far as the
condition in the adult is concerned, there is little to indicate such a
homology.

I mention here an observation which may be of significance in con-
nection with this complicated structure of the epiphysis. In both:spe-
cies and in all the individuals of Phrynosoma of which I have made
sections favorable for exhibiting the entire dorsal surface of the brain,
I have noticed that the pia mater appears to form a junction with the
connective-tissue string described as passing from the distal extrem-
ity of the epiphysis to the region of the parietal eye, and also that it
is thrown into several folds on the dorsal surface of the cerebellum.
The membrane where folded is considerably thicker than elsewhere,
contains within its folds numerous blood-vessels, and is composed of a
single layer of cells very regular and distinct in outline and of a de-
cidedly epitheloid appearance. The condition reminds one strongly
of the folded portion of the wall of the epiphysis.

Uta Stansburiana.

As I have had but two specimens of this species, both preserved in
alcohol, and hence not in the best histological condition, my study of
it has been less satisfactory than that of the species of Phrynosoma.
A few points, however, have been observed which are of some interest. ;
but these can be presented without entering into a detailed descrip-
tion of the structure. Figure 6 (Plate II.) represents a portion of a
sagittal section through the dorsal wall of the head and the parietal
vesicle. The parietal foramen, too broad to be embraced in the figure,

ee

MUSEUM OF COMPARATIVE ZOOLOGY. PPA

is much larger here than in Phrynosoma, and the vesicle can scarcely
be said to be embedded in the connective tissue of the brain roof,
us in the case of Phrynosoma, but rather is suspended from the under
side of the wall in a connective-tissue capsule.

The most striking features about this vesicle, as seen in the section,
are its dorso-ventral flattening, and the entire separation of the lens
from the retina. The lens, a well defined structure, composed of much
elongated, almost fibrous, non-stainable cells, has its margins widely
separated from the retina, and the intervening space is occupied by a
uniformly fine granular substance (cog.?), which also occupies the nar-
row space corresponding to what would be the cavity of the vesicle,
were the lens and retina continuous at the margins of the former. The
retina shows no structure beyond two deeply pigmented layers, cor-
responding to its external and internal surfaces, connected at short but
irregular intervals by pillars of pigment, between which are seen a few
scattered nuclei. This distinct separation of the margins of the lens
from the retina is the only undoubted case of the kind, so far as I know,
that has been seen, and if normal may be of significance in connection
with the theory of the origin of the eye recently advanced by Beard
(89). I am, however, inclined to believe, notwithstanding the fact
that the condition here found is apparently confirmed by the sections
of my second specimen of this species, that the separation is in reality
due to the extreme differentiation of the two structures, by means of
which the. connection between them was weakened, and then to artifi-
cial rupture by the flattening of the vesicle. The point certainly needs
confirmation in more carefully preserved specimens.

I was unable to study the epiphysis in the material which I had, but
no trace of anything like a nerve or even like a connective-tissue string
extending from the parietal vesicle could be detected, nor were there
any indications of blood-vessels or sinuses corresponding with those
existing in Phrynosoma found here.

Conclusions.

The general bearing of the facts here presented I discuss at present
only in connection with the question of the function, past and _pres-
ent, of the parietal organ. I concur in the opinion held by most of
the persons who have written on the subject, that the organ is a de-
generate eye, although my observations furnish, perhaps, no evidence
in addition to what has been presented by former writers, in support

222 BULLETIN OF THE

of the belief. From the morphologist’s point of view, the evidence
that would remove all doubt as to the correctness of this opinion would
be that the vesicle regarded as the eyeball should be composed of ele-
ments essentially similar to elements found somewhere in organs known
to perform the mechanical part in the act of vision ; and, second, that
this vesicle should be connected with the brain by a nerve comparable
with the optic nerve of some known functional eye. .I think no one
familiar with the structure of the vesicle as it exists in many Lacertilia
and in Petromyzon, will refuse to accept as satisfactory the evidence
on the first point. The evidence on the second point is less conclusive.
In many cases where the vesicle is well developed, as in Phrynosoma, it
is certain that nothing which can be justly compared to an optic nerve
exists. Spencer (’86 and ’87) and several succeeding writers have held
it as beyond doubt that in several species, notably of the genera La-
certa, [latteria, and Varanus, there is a nervous connection between the
brain and vesicle. Leydig (’89), however, in his preliminary, based on
his study of Lacerta ocellata, Varanus elegans, and other forms, says
‘‘der von Spencer beschriebene Nerv ist kein Nerv sondern das strang-
artig ausgehende Ende der Zirbel.” Lacerta ocellata is one of the
forms in which Spencer ascribes, with least question, a nervous nature
to the structure under consideration; but apparently Leydig has not
examined either of the species of Varanus, viz. gigantea and Bengalen-
sis, which Spencer studied ; while, on the other hand, V. elegans, Leydig’s
species, is not mentioned by Spencer as having been studied by him.
This denial zn toto of the existence of the nerve as described by Spen-
cer, Leydig practically repeats in his most recent contribution to the
subject (Leydig, ’90), and adds, as further confirmation of his opinion,
that he has studied Hatteria (he does not tell us what species) and finds
that here also the so-called nerve is of the nature of connective tissue.
He also comes to the conclusion in this communication, that, while from
the structure of the vesicle alone the organ must at least be put among
the sense organs, it is yet “as good as impossible to do so while it is
recognized that in the parietal structure of all the animals investigated
by me not one contains a nerve, for we must hold fast to the proposition
that for the equipment of a sense organ the peripheral end of a nerve is
necessary.” It appears to me, however, that we are not compelled to
relinquish the belief that the organ was originally an eye, even though
we accept Leydig’s statement, as against Spencer’s and others, regarding
the nature of the supposed nerve in the cases which both have exam-
ined; or even should it appear that in no case does the nervous con-
nection now exist.

ee a te eS ee

MUSEUM OF COMPARATIVE ZOOLOGY. 225

It seems to me that Leydig has not given suflicient prominence to
the possibility, not to say great probability, that the nervous connection
has been lost by the modification and degeneration which the whole
structure has certainly undergone; and especially must we hesitate in
rejecting this explanation, when we remember that by so doing we are
compelled to seek another. To be obliged to ascribe a function other
than that of vision to a structure entirely like an organ of vision in
most of its essential parts, and differing widely from one in no essential
point, is requiring us to accept a conclusion that would throw suspicion
on all our morphological reasoning. Should it be shown conclusively
that the vesicle never has, in any vertebrate, either in the adult or dur-
ing its ontogeny, nervous connection with the brain, then we should be
obliged to abandon the optical explanation of its origin, and turn to
the exceedingly difficult task of finding another. But until such
knowledge is at hand, it seems to me we must suppose that the organ
was produced as an eye, that in some way entirely unknown to us it
lost its optical function, and that, in the consequent modification and
degeneration, the optic nerve degenerated more rapidly in some cases
than did the optic vesicle; and that in this way the separation which
we now find took place.?

In .previous discussions of the nature and function of the parietal
organ, I believe sufficient attention has not been given to the structure
and development of the epiphysis and its relation to the parietal ves-
icle, and especially its relation to the so-called choroid plexus. I have
designated the entire structure found in connection with the roof of the
thalamencephalon as the epiphysis; but, as already said, I have consid-
erable doubt as to the wisdom of so doing. For the sake of precision it
would seem best that the term epiphysis should be limited to the
structure which arises as an evagination from this portion of the brain.
Certain it is that the large blood sinus which I have described as a part
of the epiphysis in Phrynosoma cannot be regarded as forming an essen-
tial portion of the structure, and I think it quite possible that what I
have called the epiphysial vesicle is not a portion of the epiphysis, should

1 Concerning the nervous connection between the eye and the epiphysis in
Anguis fragilis, Strahl and Martin say (88, p. 154), “Der Nerv der nach hinten
am Vorderrand der Epiphyse scheinbar verschwindet, tritt von unten her in das
Auge ein.” Francotte (’88, p. 782) also describes essentially the same condition
in this species. But such a condition would be so anomalous that C. K. Hoffmann
(88, p. 1991), notwithstanding the agreement of these independent statements,
has, 1t seems to me with reason, expressed doubt as to the trustworthiness of the
observations.

224 BULLETIN OF THE

the term be limited as I have suggested that it ought to be. The dis-
tinctness of the epiphysial vesicle from the proximal portion of the epi-
physis in the adult Phrynosoma is without exception, so far as my
observations have gone; and if it is regarded as having been derived
from the epiphysis, then we have two vesicles instead of one that have
arisen in this way, and the difficulty of explaining the nature and
function of the whole structure is correspondingly increased.

In his recent paper, Leydig (90) has expressed the belief that there
are two forms of parietal organs. He says: ‘¢ From the posterior por-
tion of the embryonic thalamencephalon (Zwischenhirn), especially in
Lacerta agilis, two thick-walled vesicles (Blasen) bud out just in the
middle line, lying one behind the other and springing from a common
root (einem Wurzelpunkte). The anterior vesicle gives rise to the
parietal organ, and the posterior one constitutes the epiphysis (Zirbel).”
It is only, he says, from the anterior of these two vesicles (Blasen) that
a vesicle (Blase) becomes cut off, and attains an eye-like character; the
posterior one ends in the expanded blind terminal portion of the epi-
physial thread (Zirbelfaden). But Selenka (90) informs us, in a still
more recent communication, that, after studying the development of the
brain in a large number of reptiles and other vertebrates, he is unable
te confirm Leydig’s statement as to the origin of the parietal eye. He
does find, however, in ali cases, an evagination from the dorsal wall of the
fore brain very similar to the one that forms the epiphysis from the roof
of the thalamencephalon; also that the two structures elongate para
passu, the epiphysis becoming directed upward and forward, while the
anterior evagination, which he calls the “ paraphysis,”’ becomes directed
upward and backward. After the parietal vesicle is cut off from the
epiphysis, the distal end of the paraphysis grows in between the vesicle
and the end of the epiphysis from which it was detached, and the vesicle
comes to lie on the paraphysis as on a pillow.

The relation of the two structures in the adult he does not know.
C. K. Hoffmann (’85) has also described an evagination from the roof
of the brain at the place of transition from the fore brain to the thala-
mus, which he calls the ependyma,— the beginning of the choroid
plexus, —and he says that in the grown animal “it comes to take a
not inconsiderable part in the formation of the epiphysis.” Although
there is nothing in the brief papers of either Leydig or Selenka to indi-
cate whether or not the additional more anterior evagination seen by
them is the same as that described by Hoffmann, yet, since all have
studied the same forms, viz. of the genus Lacerta, it seems quite prob-

MUSEUM OF COMPARATIVE ZOOLOGY. 235

able that they have all observed the same structure. Whether or not
any portion of the epiphysis as I have found. it in Phrynosoma cor-
responds to the paraphysis of Selenka, or the ependyma of Hofimann,
can of course be determined only by studying the development of this
portion of the brain. |

Bearing in mind the highly vascular condition of all parts of the
parietal organ, the numerous large blood-vessels surrounding the vesicle
in P. Douglassii, and the great sinus in the same region in P. coronata,
the sinuses of the epiphysis in both species, as well as the great quan-
tity of blood contained in the much folded anterior wall of the epi-
physis, it seems to me impossible to escape the belief that, in this genus
at least, the organ must have some physiological significance. Leydig
(89) has expressed the opinion that it belongs primarily to the lymph
system. From what has already been said, it is evident that I cannot
accept this conclusion ; but it does appear to me highly probable that
the structure has become secondarily of such a character. From the
numerous instances of change of function in the animal organism to
which attention has been directed by Dohrn (’75), Kleinenberg (’86),
Lankester (80), Weismann (’86), and others, there are certainly no
a priori objections to such a view, and it seems to afford more nearly a
satisfactory explanation of the present condition of the organ than does
any other. .

CamBRipGE, August 15, 1890.

VOL. Xx. — NO. 8. 15

bo
bo
op)

BULLETIN OF THE

BIBLIOGRAPHY.

Ahlborn, F.
’°83. Untersuchungen uber das Gehirn der Petromyzonten. Zeitschr. f.
wiss. Zool., Bd. XX XIX. pp. 191-294, Taf. XIII.—X VII.
’84. Ueber dic Bedeutung der Zirbeldriise. Zeitschr. f. wiss. Zool., Bd. XL.
pp. 331-337, 1 Taf.
Beard, J.
’87. The Parietal Kye in Fishes. Nature, Vol. XXXVI. pp. 246-248 and
pp. 340, 341.
’°89. Morphological Studies. —I. The Parietal Eye of the Cyclostome
Fishes. Quart. Jour. of Micr. Sci., Vol. XXIX. pp. 55-73, Pls. VI.
and VII.
Béraneck, Ed.
’87. Ueber das Parietalauge der Reptilien. Jenaische Zeitschr., Bd. XXI.
pp. 374-410, Taf. XXII., XXIII.
Carriére, Justus.
’85. Die Sehorgane der Thiere vergleichend-anatomisch dargestellt. Min-
chen. 205 pp. }
’89. Neuere Untersuchungen iiber das Parietalorgan. Biolog. Centralbl.,
Bd. IX. pp. 1386-149.
Cattie, J. Th.
’82. Recherches sur la glande pinéale (epiphysis cerebri) des Plagiostomes,
des Ganoides et des Téléostéens. Arch. de Biologie, Tom. ILI. pp. 101-
194, Pls. IV.-VI.
Dohrn, A. '
'75. Der Ursprung der Wirbelthiere und das Princip des Functionswech-
sels. Leipzig. 87 pp.
Ehlers, E.
'78. Die Epiphyse am Gehirn der Plagiostomen. Zeitschr. f. wiss. Zool.,
Bd. XXX. (Supplement), pp. 607-684, Taf. XXV., XX VI.
Francotte, P.
’88. Recherches sur le développement de l’épiphyse. Arch. de Biologie,
Tom VIII. pp. 757-821.
Gotte, A
'75. Die Entwicklungsgeschichte der Unke. Leipzig. 956 pp., 22 Taf.
Graaf, H. W. de
’86*. Zur Anatomie und Entwickelungsgeschichte der Epiphyse bei Am-
phibien und Reptilien. Zool. Anzeiger, Bd. [X. pp. 191-194.

MUSEUM OF COMPARATIVE ZOOLOGY. ZT

—’86». Bijdrage tot de Kennis van den Bouw en de Ontwikkeling der Epi-
physe bij Amphibien en Reptilien. Leyden. 4to, 61 pp., 4 pls.
Hoffmann, C. K.

85. Weitere Untersuchungen zur Entwickelung der Reptilien. Morph.
Jahrb., Bd. XI. pp. 176-218, Taf. X.-XII.

88. Hpiphyse und Parietalauge. Bronn’s Klassen und Ordnungen des
Thier-Reichs, Bd. VI. Abth. III., Reptilien, pp. 1981-1993.

Hanitsch, P.

’88. On the Pineal Eye of the wines and Adult of Axguis fragilis. Proc.

Liverpool Biolog. Soc., Vol. III. pp. 87-95, Pl. I.
Herdman, W. A.

'86. Recent Discoveries in Connection with the Pineal and Pituitary Bodies
of the Brain. Proce. Liverpool Biolog. Soc., Vol. I. pp. 18-25, Pls. I.
and II.

Julin, Ch.

’'87. De la signification morphologique de l’épiphyse (glande pinéale) des
vertébrés (avec trois planches). Bull. Scientifique du Nord de la France
et de la Belgique, 2° série, 10° année, pp. 81-142, Pl. TTT:

Kleinenberg, N.

’'86. Die Entstehung des Annelids aus der Larve von Load’
Nebst Bemerkungen tber die Entwicklung anderer Polychaeten. Zeitschr.
f. wiss. Zool., Bd. XLIV. pp. 1-227, Taf. 1-XVI. (Ueber die Entwicke-
lung durch Substitution von Organen.)

Kolliker, A. von.

’87. Ueber das Zirbel- oder Scheitelauge. Sitzungsb. d. Wirtzburger phys.-

med. Gesellsch., Sitzung vom 5. Marz, pp. 51, 52.

Lankester, E. Ray.
’°80. Degeneration. A Chapter in Darwinism. London. Nature Series,
76 pp.
Leydig, Fr. von.
'72. Die in Deutschland lebenden Arten der Saurier. Tiibingen. 4to, 262
pp-, 12 Taf.
’89. Das Parietalorgan der Reptilien und Amphibien kein Sinneswerkzeug.
Biolog. Centralbl., Bd. VIII. No. 23, pp. 707-718
90. Das Parietalorgan. Biolog. Centralbl., Bd. X. No. 9, pp. 278-285.

McKay, W. J.
’89. The Development and Structure of the Pineal Eye in Hinulia and
Grammatophera. Proc. Linn. Soc. N. 8. Wales, 2d ser., Vol. III. Pt. 2,
pp- 876-889, Pls. XXII.-X XIV.
Owsjannikow, Ph.
88. Ueber das dritte Auge bei Petromyzon fluviatilis, nebst einigen Be-
merkungen itber dasselbe Organ bei anderen Thieren. Mém. de ]’Acad.
imper. de St.-Pétersbourg, 7° série, Tom. XXXVI. No. 9, 21 pp., 1 Taf.

228 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

Ostroumoff, A. von.

87. On the Question concerning the third Eye of Vertebrates. Beilage zu
den Protocollen der Naturforsch. an der kaiserl. Universitat zu Kasan.
13 pp. (CRussian.)

Peytoureau, S. A.

’87. Le glande pinéale et le troisitme ceil des vertébrés. These, avee 42

figs. Paris. Doin, 68 pp.
Rabl-Riickard.

82. Zur Deutung und Entwickelung des Gehirns der Knochenfische. Arch.
f. Anat. u. Phys., Jahrg. 1882, Anat. Abth., pp. 207-320.

86. Zur Deutung der Zirbeldriise (Epiphysis). Zool. Anzeiger, Bd. IX.
pp. 405-407.

Selenka, E.

"90. Das Stirnorgan der Wirbelthiere. Biolog. Centralbl., Bd. X. Nr. 11,

pp. 823-3826.
Spencer, W. Baldwin.

86. The Parietal Hye of Hatteria. Nature, Vol. XXXIV. pp. 33-35.

"87. On the Presence and Structure of the Pineal Eye in Lacertilia. Quart.
Jour. of Micr. Sci., Vol. XX VII. pp. 165-238, Pls. XIV.—XX.

Stieda, L.

’65. Ueber den Bau des Haut des Frosches (Rana temporaria). Arch. f.
Anat. Phys. u. wiss. Med., Jahrg. 1865, pp. 52-66, 1 Taf.

'75. Ueber den Bau des Centralnervensystems des Axolotl und der Schild-
krote. Zeitschr. f. wiss. Zool., Bd. XX V. pp. 361-406, Taf. XXV., XX VI.

Strahl, H. :

84. Das Leydig’sche Organ bei Eidechsen. Sitzungsb. d. Gesellsch. zur

Beford. d. gesam. Naturwiss. zu Marburg, Nr. 3, 1884, pp. 81-83.
Strahl, H., und Martin, E.

‘88. Die Entwickelungsgeschichte des Parietalauges bei Anguis fragilis und
Lacerta vivipara. Arch. f. Anat. u. Phys., Jahrg. 1888, Anat. Abth.,
pp- 146-163, Taf. X. -

Weismann, A.

87. Ueber den Riickschritt in der Natur. Ber. d. Gesellsch. zu Freiburg,

Bd. IL. pp. 1-380.
Wiedersheim, R. von.

80. Das Gehirn von Ammoccetes und Petromyzon Planeri. Jenaische
Zeitschr., Bd. XIV. pp. 1-23, Taf. I.

’86. Ueber das Parietalauge der Saurier. Anat. Anzeiger, Bd. I., pp. 148,
149.

EXPLANATION OF FIGURES.

All the figures are camera drawings excepting where otherwise indicated in the
explanations, :

Rrirer, — Parietal Kye.

ABBREVIATIONS.

cav. e’phy. Cavity of the epiphysis. mac. opt. Spot marking the position

cbl. Cerebellum. of the parietal organ.

ceb. Cerebrum. mb. ba, ex. External basement mem-

chs. opt. Optic chiasm. brane. [brane.

cl. 2. Cells of zone 5 of the retina. mb. ba.i. Internal basement mem-
cl. Ins. Cells of the lens. nl. Nucleus.

coy. Coagulum. nl’. Nuclei of zone 2 of retina.

com. a. Anterior commissure. nl”, Nuclei of zone 4 of retina.

com. p. Posterior commissure. rial Nuclei of zone 5 of retina.

com. su. Superior commissure. os par. Parietal bone.

con’t. tis. | Connective tissue. rig. Pigment.

cp. sng. Biood corpuscles. pre. T. Processes radiating from the

e’drm. Ectoderm. external basement mem-

en’th. va. Endothelium of retinal blood- brane.

vessels. rtn. Retina.
eph.vs. Epithelium of the epiphysial sn. sng. _—_ Blood sinus.

vesicle. tct. thl. opt. Roof of the optic thalamus.
e’th. Epithelium. thi. opt. Optic thalamus.

la, trm. Lamina terminalis. va. rtn. Retinal blood-vessels.

Ins. Lens. vnt.3 Third ventricle of brain.

lob. opt. | Optic lobes. vs. Epiphysial vesicle.

m. SCUu. Scale of the parietal eye. zg. Second zone of retina.

PLATE).

Fig. 1. Left face of a section through the dorsal wall of the head of Phrynosoma
Douglassii in the sagittal plane, and consequently passing through the
middle of the parietal organ. Diagrammatic in unimportant details.
x 140.

«2. Transverse section through the dorsal wall of the head and middle of
the parietal organ of P. coronata. Diagrammatic in unimportant de-
tails. > 140.

«3. Section of a small portion of the deep wall of the parietal organ and the
enveloping connective-tissue capsule, to show the processes radiating
from the external basement membrane. X 1060.

« 4. A transverse section of one of the retinal vessels, in which a blood

corpuscle is seen. XX 1060.

Fie PO MS Oe OO” Os par:

B Meisel, lith.Boston.

a.

‘

or "5 Cnr AE:

A yes
oy ak da Et
San homes. eee
Gory bet ere IRE

Wie iNS 71%

\ ?
‘i ‘wy
ry a they
Mee 2
i ie
5 <a
4 > j
2 v2 ; v rte
ch - % . ve Thad ATW Pata
ree “ah rn ’ ‘ f
ja Res a) UE er eee ke Pat
P cy Ae ee sf oy OE ry
jes } Ais ; . Vi os At oil | fy SAA 7 Dy a ons
wy, MS , a 94 " : ( a A : y
: SUBS A asc) Teen, «alae
“Sh a x Se) ake a dh a i N,
aa ? “uy e ‘ oe he) 4 &ayy 2] i e ¥
, ‘ is, Utes e '
r . i 2
at i :
it 7 Hn i 4 i J
Ss ¢ .
‘%? j a \
: Opes iy ra oo . 1 i
—_— dh i a ane | A

¥ 5 F ip nl r ? 4 ¢
F r, . ee ie } wa ‘ |

5 —

A : [ duet b a> a) i : hier
, 3 } ( ; rt be, Mite oe :
f 1 ; St ed rk” OT eae) Veen
a Pal us _ *
i air fe : '
he = s, h, ; ie “i a \e ‘ -

D esd

: wry :
er 71> : fy. . a FY he i G Ve , » 4a hier 5 4?

aA © Le
iF] f
me ‘
if?
, - ‘ * ee 7 far * ces
id 7 y y
uy
Ng ;
rey | oe Ng -
y
t .
ray
pn
—
ry ie -
.
Y id 4
Ls
4 it
‘
4
{ fia
* at
~

Ritrer. — Parietal Eve

ABBREVIATIONS.

cav e’phy. Cavity of the epiphysis. mac. opt. Spot marking the position
cbl. Cerebellum. . of the parietal organ.
ceb. Cerebrum. mb. ba. ex. External basement mem-
chs. opt. | Optic chiasm. brane. [brane.
hyd: Cells of zone 5 of the retina. mb. ba.2. Internal basement mem-
cl. Ins. Cells of the lens. nl. Nucleus.
cog. Coagulum. nl’. Nuclei of zone 2 of retina.
com. a. Anterior commissure. nl”. Nuclei of zone 4 of retina.
com. p. Posterior commissure. ae aor Nuclei of zone 5 of retina.
com. su. | Superior commissure. os par. Parietal bone.
con't. tis. Connective tissue. py. Pigment.
cp. sng. Blood corpuscles. pre. r. Processes radiating from the
e’drm. Ectoderm. external basement mem-
en’th. va. Endothelium of retinal blood- brane.

vessels. rtn. Retina.
eph- vs. Epithelium of the epiphysial sn. sng. Blood sinus.

vesicle. , tct. thl. opt. Roof of the optic thalamus.
eth. Epithelium. thi. opt. Optic thalamus.
la. trm. Lamina terminalis. va. rtn. Retinal blood-vessels.
(ns. Lens. ont.3 Third ventricle of brain.’
lob. opt. Optic lobes. vs. Epiphysial vesicle.
m. scu. Scale of the parietal eye. at Second zone of retina.

* Processes secreted from the inner surface of the retina.

PLATE II.

Fig. 5. A portion of a section near the median plane, through the same eye as
that represented in Figure 1, more highly magnified. X 570.

“6. Sagittal section of the dorsal wall of the head, with the parietal organ
of Uta Stansburiana. Diagrammatic in unimportant details. X 312.

“7. External view of the parietal eye and surrounding structures in Uta
Stansburiana. X 8.

t

R TER. — PARIETAL EYE. | Pell

edrm.
conttis.

es én.

BMeisel,lith Boston.

Os —

Piet.

a it

Gos Y » Pir bh a

sis Gi.

jit mii: Padi

Age LT rey,
neat !

aS Poaidiey tieks Ans “ae!
en AIFS She Peis

pie re ifs ok Gish :

a

Rat aay ‘anna’ :
range 2 hone

:, aaah
ee 2.9 ia

/ i Ponce sal sented
sh via gllnede ne

ict cane oe te te
a yee De lait ga ‘a ie we

VT plone Sit a ole a

= Weed Stat

vineea: vay wo inde, cbt He

pens heath
oj IS

vi oe du ee ae af
Py RAR aa: sail Behe"

1 Wea rae Sad ty is wis ote ari»
ity pul Peal fh tae pealaaedseis j

a é Se ican oy aed a a ;

+s
i=

=o thas “ha eae so seas ys

Rirrer, — Parietal Eye.

ABBREVIATIONS.

cav. e’phy. Cavity of the epiphysis. mac. opt. Spot marking the position
cbl. Cerebellum. of the parietal organ.
ceb. Cerebrum. mb. ba. ex. External basement mem-
chs. opt. | Optic chiasm. brane. - [brane.
el. tv. Cells of zone 5 of the retina. mb. la.i. Internal basement mem-
cl. Ins. Cells of the lens. nl. Nucleus.
cog. Coagulum. nl’. Nuclei of zone 2 of retina.
com. a. Anterior commissure. nl”. Nuclei of zone 4 of retina.
com. p. Posterior commissure. Te Nuclei of zone 5 of retina.
com. su. Superior commissure. os pur. _- Parietal bone.
con’t. tis. Connective tissue. pg: Pigment.
cp. sng. Blood corpuscles. pre. r. Processes radiating from the
e’drm. Ectoderm. external basement mem-
en’th. va. Endothelium of retinal blood- brane.

vessels. rtn. Retina.
eph. vs. Epithelium of the epiphysial sn. sng. Blood sinus.

vesicle. ict. thi. opt. Roof of the optic thalamus.
eth. Epithelium. thl. opt. Optic thalamus.
la. trm. Lamina terminalis. va. rin, Retinal blood-vessels.
Ins. Lens. ont. Third ventricle of brain.
lob. opt. | Optic lobes. US. Epiphysial vesicle.
m. scUu. Scale of the parietal eye. es Second zone of retina.

PLATE III.

Fig. 8. Left face of a sagittal section through a portion of the epiphysis, a short
distance above its connection with the brain in P. Douglassvi. It is in
’ part diagrammatic, though the outlines of the figure as a whole, and
of most of the foldings of the epithelium, were drawn with the
camera. From the same individual as Figure 1. X 312.

“ 9 Similar view of a sagittal section from the same individual, to show the
relation of the epiphysis to the brain and the blood sinus. X 80.

~ RITTER.— PARIETAL EYE.

Pu JIL

see oe — Sa

a

C

BMeisel,lith. Boston.

WER del.

he een 2 horas

bar

aren tae a a ee
o ue meses wile ty “se

1

eda
sites is ae ty eae
a ae S|) ina ea
Tae ihe: tsi," Lee Seats a
ag ane ear

nists rae

ge atetiaies © i i y
Foti a" G8 ada i 2 wig = ss ot Bt ’ ea
STIG PRPEE ENG |) Sry $13 Bees »vaet a ae Ai wag ya ay : ie
ie Festa vA) pe frags RAR hae coe peskpacc vg

ee Sh ee Begs, os, seh ot a

a . ? o
4 P x 0
7 } ‘ yer
Pa
4 —— —_

{ 7 en

PR ’

F
™
’
;
ati
———.
; } Me : ip Th ia .

My Oe aan ‘at, ba wee | WO pds ed , ON i

‘ " j is 5 ant ; ’ :
i See hy eds Ti maf rea ea Ul Fe a mY aa alas Ps, , peat forged Smee

Rirrer. — Parietal Eye.

ABBREVIATIONS.
cav. e’phy. Cavity of the epiphysis. mac. opt. Spot marking the position
cbl. Cerebellum. ” of the parietal organ.
ceb. Cerebrum. mb. ba. ex. External basement mem-
chs, opt. Optic chiasm. brane. [ brane.
ets Cells of zone 5 of the retina. mb. ba.7. Internal basement mem-
cl. Ins. Cells of the lens. nl. Nucleus.
co7. Coagulum. a Nuclei of zone 2 of retina.
com. a. Anterior commissure. ae Nuclei of zone 4 of retina.
com. p. Posterior commissure. nl”. Nuclei of zone 5 of retina.
com. su. Superior commissure. os par. Parietal bone.
con’t. tts. Connective tissue. pg: Pigment.
cp. sng. Blood corpuscles. pre. r. Processes radiating from the
e’drm. Ectoderm. external basement mem-
en’th. va. Endothelium of retinal blood- brane.
vessels. rtn. ' Retina.
eph. vs. Epithelium of the epiphysial sn. sng. Blood sinus.
vesicle. tct. thi. opt. Roof of the optic thalamus.
eth. Epithelium. thl. opt. Optic thalamus.
la. trm. Lamina terminalis. va. rtn. Retinal blood-vessels.
Ins Lens vnt.8 Third ventricle of brain.
lob. opt. Optic lobes. vs, Epiphysial vesicle.
Mm. SCU. Scale of the parietal eye. ai Second zone of retina.
PLATE IV.

Figs. 10, 11, 12. The left faces of three sections of P. coronata, parallel with the
sagittal plane, — Figure 12 nearly median, Figures 10 and 11 to the
left of it. Figure 10, farther to the left of the median plane than
Figure 11, passes through the longest part of the epiphysial vesicle.
Figure 12 is more highly magnified, to show the histological structure.
Figs. 10 and 11 x 40. Fig. 12 x 90.

W.ER del.

ize

on

roe

B Meisel. lith. Boston.

t

Dy

i. Paty, aay rm
nie", ; Oo aaa d
Poy A uy Lae My heath NG ,

a
F f
tty
H oN
; i /
7 }
Z i"
Ps y ff ae
} fe th
Tem ws
A iv 4 ”"
a 7
a!
e ‘
ek eal
: ' ree
: 7 ; Ay
” il,
i itr 5
wail t)
i \ | yi . iy
iV
NN aE
\ . f o ‘
’ th) ;
‘ t 1 us
iy : ’ i
oh Ny a
cre oe
: mr 4
ble
ia rh
ths ,

:
i
' J t )
Wal) i 7 h
ALU | Mak: Tae, ; } n ‘ Lian
i y il

[ ve i,
wah A ay \h Hi
Mba tinesy: Sans) di)

a hy aa

aa Ps, Goo & eee WS : » ws Pais fy - Tas

QL Harvard University. Museum
4 of Comparative Zoology

H3 Bulletin

v.19—20

Biological

& Medical

Serials

PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET

oc a

UNIVERSITY OF TORONTO LIBRARY

HANDBOUND
AT THE

we.

Se

UNIVERSITY OF
TORONTO PRESS

Pn Se ne NaCl a aA TR lm giana
POL Oy pe el hetteidnnd Veetealins idee ee
. le ene, ee eee 2 + ne,

7s ha 7 7 , 7 . - as 4s SARS Sou I I hme ~ Wit meng
j 7 sa8 i on eee ee oe ne ee) ee re Maplin ete me ek jeter my mace
00. ade e misten DoR es tee em I, ed Ra eeenipansiny Mile PN i me A EA ate OER Ue ony cee ee
ite? : 7 eee = ae ee errs a Ae! Oenummumpanpene ne ee men ete ered Lr ieee — pape
canned apapipeathey a cag hapa epigrtpiaioca apisa- tS See ci : pb een ah. gine ge es ee
spine tear eee " > 7 "E57 pea Path
= eon are ee RI tes wishib_srce_agn soe ne Nene a ge eR ee ORTON a oe tae cag SO A tye
pg mareg ce aetna - neon wee Tome ns ae mn be =e. ee oe _ Hie apie =.
_ - “2 b= -_- -~— - ~ se + fe oon epi aha a mane a a s : oie oP i Pa anche ¥ = 7. se es
ay Soa - -a+ *~ ¢ pao atin tie aes “ we a
— 2 Bo eae ae SS Ses mpl :
J pe a ein oie =
ot tenet Ses pile het : pera ie oe eg nS ne ee NE GE A rt satin A ge
ds Ree a dk + hse - ue aap,
Fone eS ao sages esas El a Noe
7 a ~~ = 7 - _- sanyo a
= F? Es aan = apes <*> 7 —~ ~~ ~~
Sumas Recs pe ita ae > ae ie
a oe ene Waste ee ei een et a ela eee Ot ee
SMe ala ae st aa ~ SAE Aves :
fe RS ot eee Bin ae eS alee ae my Ke 7 den Geek cee oe
a (Stir A Fab Sh Nap ag ce sm - eens > oe amie rece
Sn i es oe i eee ae an fa a a Sane SN SIS IE natal - FS An ETO TPM teen 0 My ane ee ae a
eater ie ie ee oe etree et we ee ee =: a oe ‘ ST me aan oe me te eal ls
ee A SP SS ee ee Sr GS Orme pw eles po
- ee Aes as SRR him Rg a LS enw aS 8 oy See eae ee te age Wie we we ee nT a OR ee ee Wihe aae a ages bi ee eee -
phic eee ee =- * an er er tee ne ene Se,
[— So a ea ek eee Se Ser MUM. - TESS UR SS Se SE tet 2 ee
m Sloe ae eae a ne A ere ee ge Ea ee eo pia 7 Fk WNT He Seer me Ge aoe bn ge
Rebels ayer ones ee chi en oe aw a SN of ee een im ene oe
Fe 6 pele, @ age ae. She iy oe oo ee =~ ~
x aes peck out Se ae? PS OE SS me er me ey = Dad “~ i tian
he imc me eran, BRI See ee ee a ee See me gry ~+ Ms re cama Te ee! aha oo osmg A aon den mea ee ee we
coms win eco ater eee OL ee pees a a a lee mame a se oe
wee Ores ee ee A ee me we ong ee ay eS = =. “ar wo ne fees Se Onde yng wins aApaiptt oN ee oo oo. nee a mee
We Fm eee mS oe ee, ¥ — OE A AARNE PAI Gs 8 a eny aber in Senn ah nha todamerickal ee oe a
= ie eh ty a mn oP = - -s Sree sham ene agen Steere ee . —erenes L,
OS OS See ae a nin eee ie eats, 1 eel atie nie ae -- cm + 2 Dale nuiiewne ome wa od wre pi
ey ae ee omer ewe ae 4 ~~ mes Se epee M0 A 08 be om amen guree ee
i. a Sh a - ~~ = eee 2 iN ne se ae Hiden emer again mpinpertecites att
Sg teeta ci . F Nihal ee ear pe alpen eae hn ESD iat ten acta ee aeidnepenh doles ee
ons ncaa HEE: yy Hla Ss ak a eraoe a mae? on a A? St ne = PLSD em ie eee ee Pt mp see Wee,
.- ie ai ne . a = se . * etn kage «on ae 2? the, 2 gh
ome eR eee Re ee Ce Senne a a Se ON Am ss, ee ares Fe AO eats gute mum por a
a PA EEE Se ee me uipeadtie tz er EET One me ee we SOFT EGR ePs we ane,
Se ee ee aie plage ~ ae eels AO to een nw
2.6 pe ap wan tee mate = wy a Re St ~~ "~~ =e ER. perpegae mpd ee
OS Are elas i eK AES Re ee eo etal! ge = A ie : FORM oe waa enemun Senge npn
Oe ee ee ee ee Fa i see ne OL eet - - - Sia ~ a reitiped at oui eid Meni ee ee ee
OE PI OA ET a ay ey Sa a gs o ee ee NN eS te Bak oa co idecienien iy leben tiedn sete te
re oe lee inte meer = > ~ ree a Oe + tems ee — an tay ee ae pon pes oon mri nape on
FOR le eye te pe ap tee Re en tn ee Ee ne ee SO A Nee os ~ - ot et ae taal AG tee Ai amg galne ab suaageemee on
WRN eH Le ree eat tS mee eee ~ s Iti Ame Hw adnate oe silliniethinthath semen eae Le ee >
oo mee nee BOE Vile NS Ee a ey a, eS a te ea a = aon be A tet) tite authenticate
Mm ape em ere te mt Se ere ew « mor gent papaya tras a <—l a mony oe Se a a nen rater eoahnnet orn aoe teeny
id ne ee ee a7 2 es ~“ > Stl haggis. 0 J * " ~ PL OOOO eat ne ae mt seme ag
jee ne hs 88 we et ee oe te “ nee - a Nt BO trp mst, open PORE + wi nea te ee ee
om way em ol fw eee eee ew eee - Piipendinsetentmeein he Ra lg a celigeenge i gthl~ore-ae wi~- oma, On ment nee
IE Gm AE SO OO me ome eI oe SU SO eng ee “s LO PE te ih mens On nae eae rom . tyne ne a ee ee nO weg ne magerewngee ncn tens Potala
a” sla We cab Ponce ok Fw Se wee me ee ft rmwe tie NS OT MIS oe OE ea oem tere prc CD AA eS tlc PPCM RM whe Vg ot © moninn wen as naeretadtaten (Aeterna Gem onpe seaea oreo an
Mice, = ows a st be Oe el ee eee neta A led eh eine OTe TE FE NNO CON Ale enipineetinm. aplen e a OS LE Oe nt nk A neat Delian te ee ne nag a eo eo ner mew amnar gee een ee mma
mS er mnt ee em eee ee ee a ee we Crewe Se NOONE Ret he & Miata, a. oh SON dean ent ~~ mel eames elaine at ete tate ee se a erromn beter na mega ene aon nnaparwnen aaa eae arg snot
np Ng A on gas a ae a ast wae ahem noe ae afr ne ae Ga eh ne EN pele me anaes open an ek ee OOS nrenabetior syns onee-co renee ee
snciuanliicees Te” eons on, ene = wee Pe ms noe MO RM ec ee Oe epee a ent Rae WP mee lovastatin Ee Doe Ie tape eaip ome ae eT Opera ee plete et
a ae 4 - es a ee sopcapiny SSS eo a mi erefotegrome Anode ee a ee ene Ae OO nh a eh Rr ie mone ew ne OR et mee aie tcptabees gudemaeet aaeeiie a en
2 ttie mewe = se were ee 7 omer + ee oe... ~~e- ay wecayee > Oa am oF ee neaichassetnaiinindiete aa ae <n ee em em gk cow SoS ad te
weet nine ep a one ear a ye ee ee re a eal eens On Tre ot Sua ns Staenaee 9 et a net ons bt pe tee crt Meroe ns cana gos er eteeovertewcmoa se
a eh an My dgem a eee eee ET eee ee ee ee, LO COD A A hn at nn nn potinde NE i ne rea irene ype, ass Sapineypseeedyanlinm ner iepapes cate ee ee Ie orb Aine Oe AsRi APO adie
Op Ae ro eet A wm Verein e Mane ee ne Oe ne, nent fnslaiaebeied. sagan de car otie Sn a a RN CAND uty ps tmies oactee priate ekore ne < nt convenes n coulandratydeemmaen =
OE Ny = pm pens Epa ne we te ee pene ey eta fet in nae Atle be NORE PPG TO sen os epee TE Sot ioe Mehner mee Penpatin parensesealogeare oO rocapee ene pera nmrge agen
SR ORE am a dt Me peepee oe CIR Cael yee fas re, Aen prt ares bm ah Menaigen man's magma ee 7S ioanliagligieiaceeeencme te ea (OP OAS ne Ney 8 oh rma ;
OF Oe mee eee NO i OE Rae the cin emeny oe rene cee ee a rare etree cine ma sparymaianeeipatene naa nere J
PO Bt ee een y aE Rese, bese ieaticuliteeererces oo alee ee ee Fee . te, _ r
OE OSL OR te. ~ age ayes agree ner ee

etn rate
tee 6 Nin Sree AF 08 an!

Ov ep OR weer n nae Mehdphaeaio tassios ae
J ve as irthecahetninins nteit-omn enn ee too
PLE ET QUOT Re ie ee i.

let cence ie ee ae

Seer oe e-LN Setind ON Negi om Oe te. ae vada tpdivewkteac ee

EPS GO Rated tent @ = ee em pe ani aroes Lobe Neprnnyt omg aaarcenmmm ane een

ae PS Fe were es cron ae veo 1 am pine tree ettineeiiee mi iaen marta hon int pn Cornainnnbeptemas,

baptist atl nde ee a ee Siete olined Naas EE rao neretnn Snare vernon semencten iron
Oe Se eee a eae many trial gee .

ws nner abena sue gnaeate tet

treomrens oi eee en

bye en soteod bap encom aban ao seor ta ite eens .

= “J ee tat arin eaten tetas mercer ree eeey) 2+ oh oh ererapadiqineta reMatemahns entre, neers! “ wo alen bart wmenetinion ge ry elie oeqeeseruee ne hnefe ei ot=ashet

. : "Oe eran Nag, enema TE ote 8 oe, nora - Caen oe taty ee tA we ohn fm Loaeatentnalbociaall iele>grtiuneediuen en

ree i mer hae an ero aah moma patoaonialon SPP anime np iensn epee main poe ree eh gewcinngan act, etl eaten saethctongio tn toe te om et we OE alc on ft
- + aovenall msm Retin tnairnsips, alee hom mmeger ana nt . Mt eerynm os oes : ee FO ARO wre we lipntee

ib ittitichiinah dnbumetien hee SemningmeiN Mrmr ton aaas Suton cantata tages, epee writ

(te aoe oe
- Laperepadatign etienel eet rac hekantterens
Paral tenene al —

- 1° © @ aera Gr elevedg etre nee
nb-btsuinpket aiid et eee are

2 SM ere Se

te ewer me Rad Ot'9 Oe te,
ee 8 ae wate 0K = ye ob