BULLETIN

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT

HARVARD COLLEGE, IN CAMBRIDGE.

VOL. XXIV.

CAMBRIDGE, MASS., U.S. A.
1893.

Reprinted with the permission of the original publisher
KRAUS REPRINT CORPORATION
New York
1967

Printed in U.S.A.

CONTENTS.

No. 1. — Contributions from the Zodlogical Laboratory. XXXII. On Urna-
tella gracilis. By C. B. Davenport. (6 Plates.) January, 1893

No. 2. — Contributions from the Zoólogical Laboratory. XXXIV. Note on
Carotids and the Ductus Botalli of the Alligator. By C. B. DAVENPORT.
(1 Plate.) January, 1893

No. 8. — Contributions from the Zodlogical Laboratory. XXXV. On the
Eyes, the Integumentary Sense Papille, and the Integument of the San
Diego Blind Fish (Typhlogobius californiensis, Steindachner). By W. E.
Rrrrer. (4 Plates.) April, 1893

No. 4. — Reports on the Dredging Operations off the West Coast of Central
America to the Galapagos, ete, by the U. S. Fish Commission Steamer
“ Albatross.” IV. Vorlüuflger Bericht über die erbeuteten Holothurien.
Von HunrnT Lupwia. June, 1893

No. 5. — Contributions from the Zoölogieal Laboratory. XXXVI. The
Development of the Scales of Lepidosteus. By W. S. Nickerson. (4
Plates.) July, 1893

No. 6. — Contributions from the Zoölogical Laboratory. XXXVII. Studies
in Morphogenesis. I. On the Development of the Cerata in ZBolis. By
C. B. Davenport. (2 Plates.) July, 1893 .

No. 7.— Reports on the Dredging Operations off the West Coast of Central
America to the Galapagos, etc., by the U. 5. Fish Commission Steamer
« Albatross.” VI. Preliminary Descriptions of New Species of Crustacea.

By WarrER Faxon, August, 1803

PAGE

105

115

141

140

No. 1. — On Urnatella gracilis. By C. B. DAVENPORT.!

CONTENTS.
PAGE PAGE
E Introdubtlon «610v. 1 4. Histology of the Buds . . 24
Ts sAMBtOMY Po en a 8 5. Formation of new Stocks . 24
LG SIE 6 ne m é 6. Formation of the Individual 28
AID, OBS WAS UE yo 9 7. Regeneration . . . . . 2
III. Non-sexual Reproduction . . 16] IV. Affinities of Urnatella . . . 30
1. Architecture of the Stock . 16| V. Affinities of the Bryozoa . . 30
2, Segmentation of the Stalk . 19| VI Summary ...... . 35
3. Orientation of the Individ- VIL, Literature cited... vo ees 00
Wg) O Dis ES ADE ya Inter c v as US

I. INTRODUCTION.

In November, 1851, the late Dr. Leidy published, in the Proceedings
of the Academy of Natural Sciences of Philadelphia, a notice of “ what
I suspect to be the conweium of a new genus of Polyzoa; although I
have never been able to detect the polypides if such exist.” He gives
three figures of the “comacium,” consisting of a series of urn-shaped
segments placed so that the base of one segment rests on the mouth of
its proximal neighbor, To this new genus he gives the name Urna-
tella, with a genus diagnosis, which three years later (54, p. 191), after
discovery of the polypide, was revised to read as follows: “ Joenoecium
consisting of a series of segments up to eighteen in number, and forming
free, semi-crect, curved stems, attached only by the base of the lowest
segment. Segments excepting the three last ones simple, urniform ; the
antepenultimate and the penultimate oblong, with simple or compound
branches of the same form ; the last segment or active polyp is campanu-
late, and is supplied with cylindrical, ciliated arms arranged in a circle
around the mouth.” He finds that several stems arise from a common
base. There are fourteen tentacles,

In 1856 Allman (pp. 117-119) referred to the discovery of Leidy, and
published a pencil sketch of Urnatella furnished by Leidy. This is the

1 Contributions from the Zoölogical Laboratory of the Museum of Comparative
Zoölogy, under the direction of E. L, Mark, No. XXXIII.

VOL. XXIV. — NO. 1. 1

2 BULLETIN OF THE

first published drawing of the calyx and young branches. Allman was
inclined to refer this genus to the Pedicellinide.

In 1870 Leidy recorded further observations on Urnatella, and in 1884
he published in a thin quarto, with a single plate, everything which he
had found out about its anatomy and physiology. The main points of
these papers will be brought out in connection with the organs to which
they reter.

Finally, Mr. Edward Potts, in a note to one of the editors of the
American Naturalist (April, 1891), states that he has succeeded [as did
Leidy] in obtaining in the spring rejuvenated heads from the headless
stems of Urnatella gathered the preceding fall.

Various writers have called attention to the imperfections in our
knowledge of this aberrant form. Ehlers (’90) has made many sugges-
tions concerning the anatomy of Urnatella, which, being purely hypo-
thetical, require to be settled by observation. Leidy himself was impressed
with the importance of a better knowledge, and he had intended, he says
(84, p. 6), to make a thorough investigation of it. “Other occupations,
and the want of a ready supply of the necessary material, have prevented
my intention, and I am now led to communicate what I have learned of
the animal with the view that some of my younger countrymen and
co-laborers, under more favorable circumstances, may be induced to do
what I had hoped and wished to do.” To supplement the work of Pro-
fessor Leidy is the object of the present paper, which has been largely
inspired by his.

In 1884 Leidy remarked upon the absence of Urnatella from its former
haunts. Apparently nothing had been seen of it since that time, until, in
1889, I opened correspondence with Mr. Edward Potts of Philadelphia
upon the subject, In the summer of 1890 Mr. Potts and I thoroughly
examined the waters of the Schuylkill River, both above and below the
"airmount dam, but without finding any trace of Urnatella. In Septem-
ber, 1891, Mr. Potts found many stocks in the bed of the temporarily
emptied Schuylkill Canal, below Flat Rock dam, and kindly forwarded
some of these, living, to me at Cambridge. In July, 1892, Mr. Potts
and I re-examined in vain the Schuylkill River at Fairmount dam, and
finally, on July 4, made dredgings in the Schuylkill at Flat Rock dam,
near Shawmont Station, Pennsylvania. We found no trace of Urnatella
in the quiet waters above the dam, but in the turbulent waters im-
mediately below the overflow almost every stone brought from the bot-
tom bore stocks, and some were almost completely covered on one face

with luxuriant growths, One barrel-hoop dredged from the mud was

MUSEUM OF COMPARATIVE ZOOLOGY. 3

completely covered with headless stems. As many stones as we could
carry were packed in moist paper or in jars, and transported safely
several miles. A few days later, stones with living stocks of Urnatella
and parts of the barrel-hoop with headless stems were transported in a
pail of water to Cambridge. Upon arrival at Cambridge it was found
that nearly all the heads had dropped off. The stocks were kept in
running water for three months, during which time those which had lost
their calyces did not regenerate them; whereas the stocks on the barrel-
hoop did regenerate their heads, but did not bud very luxuriantly.

I cannot close this introduction without expressing thanks to my
friend Mr. Potts, who, filled with love for the science, has unselfishly
supplied me with material, hospitably received me into his household
during my stay in Pennsylvania, and worked with me beyond his
strength in our dredging operations on the Schuylkill in order to further `
a study of this animal.

The later studies included in this paper were made in the Embryo-
logical Laboratory, under the direction of Dr. E. L. Mark.

I propose to describe first the anatomy of the adult, and, in the second
place, the non-sexual reproduction and regeneration, so far as I have
been able to determine them. In each section of the paper I shall give
first my own observations on Urnatella ; secondly, compare them with
those of others on Urnatella or other Endoprocta ; and, finally, give my
conclusions upon or interpretation of any difficult points.

TI. ANATOMY.

For the purposes of description we must distinguish the stalk and
the calyx.

1. The Stalk.

The stalk is an elongated, filamentous structure, composed of a
number of segments, lying one beyond the other. The number of such
segments varies from one to several, according to the age of the stalk.
Leidy counted as many as eighteen. I have never counted more.
Looked at with a hand lens, the stem is seen to be composed of alter-
nating translucent and opaque rings, of which the former occupy the
middle zone of each segment ; the latter have an intersegmental position.
The surface of the stalk in the translucent zone is thrown into circular
folds, and small processes appear scattered over its surface, more abun-
dantly on the more distal segments (Plate I. Fig. 6).

OF

BULLETIN THE

An optical longitudinal section of the living stalk (Plate I. Fig. 7)
shows that it consists of a thick cuticula, within which is a mass of ill-
defined tissue. ‘The cuticula is thicker and more opaque in the basal
segments, thinner and more transparent in the distal ones. Each of the
basal segments often contains a mass of globular, highly refractive bodies,
which may be forced out from the segment by crushing the latter (Plate
I. Figs. 4, 5). Isolated by this means they are seen to be of the most
varied size, and to resemble yolk spherules. Besides these granules, one
finds lying near the cuticula in all but the oldest segments a fibrillated
layer consisting of elongated, spindle-shaped, highly granular cells in the
midst of the yolk spherules (Plate I. Fig. 7), and, finally, one or several
flickering bodies reminding one by their motion of the cirri! of flame
cells. In the dark zones optical sections reveal indications of a transverse
septum.

A more accurate insight into the histological structures is gained
from thin sections of the stalk.

The cuticula appears in two conditions, which are not sharply separated,
but pass into each other through intermediate conditions. The one
kind (Plate I. Fig. 4, cta.) occurs in the middle zone of each segment.
It is transparent and flexible, and becomes thicker in passing from the
distal to the proximal segments. It stains in hamatoxylin, and is then
seen to be composed of laminæ, for it appears in section distinctly
striated (Plate I. Fig. 4; Plate II. Fig. 16). The other kind of cuticula
occurs chiefly at the constrictions between segments, but it extends over
the two ends of the segments also, and even over the middle zones of
the segments, as an outer cuticula (Plate II. Fig. 8, cta!.). It forms,
moreover, the septa between adjacent segments. It does not stain in
hematoxylin, is more highly refractive than the first kind, and has a
yellowish color by transmitted light. Where it passes over into the first
mentioned kind of cuticula, it becomes distinctly striated, and gradually
loses its specific characters, gaining those of the first kind, of which it is
probably only a modification. It is the presence of these two kinds of
cuticula that causes the superficial appearance of alternating light and
dark zones on the stem. These two kinds of cuticula do not correspond
to the inner and outer cuticulas which Ehlers (90, pp. 16, 17) describes
for Ascopodaria. Two layers with the characters of those of Ehlers ap-
pear not to exist in Urnatella,

The ectoderm (Plate II. Figs. 11, 10, 16) diminishes in thickness from

1 I employ the word “cirrus” in the sense in which it is used by Bütschli
for Ciliata (cf. Bronn’s Thierreich, Protozoa, p. 1324).

t

MUSEUM OF COMPARATIVE ZOÖLOGY.

the distal end of the stalk, where it is a layer of cuboidal cells, to the
base, where its cells are extremely flattened, so that the layer is hardly
distinguishable. Even near the base of the stalk, however, the ectoderm
is a relatively thick layer on each of the surfaces of the transverse septa.

The axial portion of the stalk presents very interesting histological
conditions. In a young individual whose stalk comprises a single seg-
ment only (Plate II. Fig. 15) the axial portion contains a closely packed
mass of cells, which stain deeply, contain few vacuoles, and are slightly
elongated in the direction of the axis of the stalk. At a later stage one
finds the plasma of some of these cells undergoing a metamorphosis.
Such cells appear filled by a vacuole in which lie numerous highly re-
fractive granules. A little later all of the axial cells have become greatly
elongated in correspondence with the elongation of the stalk. Many of
them appear almost completely vacuolated. A cross section of the stalk
at such a stage of development is shown in Plate II. Figure 9, where the
centre is filled with a network of cell boundaries enclosing clear spaces.
This condition is very striking in longitudinal sections through the re-
gion of the septum, where the long vacuolated cells may be seen extend-
ing from one segment into the other.

According to Ehlers (90, pp. 18-22), the stolon of Ascopodaria shows
a condition of the axial portion similar to that here described; but in
the stalks which support the calyces there is a considerable amount of
intercellular substance, and 1 find the same thing in Pedicellina. In
Urnatella, on the contrary, such intercellular substance appears only in
the older stalks.

The musculature of the adult stalk is situated immediately within the
ectoderm, and is found in all segments running from one septum to the
other. On cross section of the stalk it appears as a circle of radially
elongated, highly refractive bodies, placed close to the ectoderm (Plate
II. Fig. 9, mu). Each elongated body has, at a certain focus, a darker
periphery and a lighter central line. With the highest powers the darker
periphery can be resolved into a single row of bodies, which appear as
bright spots at the highest, as dark spots at the lowest focus (Plate II.
Fig. 16, mu). These I interpret as fibrilloo,

The fibrille develop in the protoplasm of the axial cells. In their
earliest stages one sees them forming a single row on the outer edges of
the more peripheral cells (Plate IT. Fig. 17, mw). Later they are formed
on the two lateral edges also (Fig. 11). Additional lateral (or, in rola-
tion to the stalk, radial) rows of fibrillo are next formed in each cell,

the new rows appearing nearer the centre of the cell than those already

6 BULLETIN OF THE

formed (Fig. 10). Finally, in the adult, the nucleus of the cell, which
has been pushed centripetad by the multiplication of fibrillee, comes to
lie opposite several pairs of radial rows or sheets of fibrille. These
sheets always occur in pairs, the components of the pairs being separated
by a narrow clear space, and the pairs by broader spaces.

As I have said, this musculature occurs throughout the entire stalk up
to the base of the calyx (Plate III. Fig. 19, mw.). I do not understand
on what ground Ehlers states (90, p. 146) that in Urnatella “der
Muskelmantel in den gleichmässig dicken Stengelgliedern nicht deren

ganze Lange erfiillt.”

There is a considerable difference between my description of the struc-
tures of the muscle fibres and that of Ehlers (90, pp. 25-28) for Asco-
podaria. In the latter case, the muscle cells are placed three to five

deep, instead of in one peripheral series, as in Urnatella. Moreover, in
Ascopodaria the fibrille lie in two thick peripheral bands on the sides
of each muscle cell ; whereas in Urnatella several muscle bands (each
consisting of a single row of fibrillae) belong to each cell.

The differences between the two genera are easily explained by assum-
ing an increase in the number of muscle cells in Ascopodaria, so that
they can no longer lie in a single peripheral series, and that the fibrille
bands have become greatly crowded together, so that one band is not
equivalent to a single band of Urnatella, but to all of the bands of Urna-
tella which have arisen on one side of the cell. The stalk musculature
of Ascopodaria is thus more highly developed than that of Urnatella.

Ehlers mentions the resemblance between the muscle fibres of Ascopo-
daria and those of Nematodes, especially the Coclomyaria.

I have referred above to vibratile movements in the living stalk.
Stained sections show, placed more or less abundantly throughout the
stalk, elongated spaces terminating blindly at one end in structures
which must be regarded as flame cells (Plate II. Fig. 12, cl. jim.).
Such a flame cell is spherical in form, stains slightly, possesses a large
nucleus, and gives off into the elongated space cl. tb. the well known
cirrus (Wimperflamme).

The form of the cirrus in all cases examined was conical, the axis
of the cone being usually thrown into a sigmoid curve. Where the
cirrus arises from the cell, it is slightly constricted, then enlarges sud-
denly, forming a thickened ring. Distal of this ring the cirrus tapers
gradually toa fine point. That the cirrus is composed of a number of
agglutinated cilia is indicated by the presence of longitudinal strie. I
believe that the flame cell closes the lumen of the tubule at this point.

MUSEUM OF COMPARATIVE ZOOLOGY, Y

Concerning the space into which the free end of the cirrus projects I
can only conclude, from comparison with other cases where a flame cell
occurs, that this must represent the lumen of an excretory tubule. I
have not, however, been able in any instance to trace an individual
tubule to any considerable distance, or until it opens into any other
organ, One thing seems fairly certain, there is no morphological differ-
ence between those elongated clear spaces into which the cirri project
and the elongated vacuolated cells to which reference has already been
made. Iam therefore inclined to regard it as probable that they are
one and the same thing. The vacuolated cells are developed by the
breaking down of the plasma of the elongated cells of the young stalk.
Before the metamorphosis is completed, one finds cells filled with clear
contents, in which lie scattered granules (Plate II. Fig. 15, cl, tb. left).
Finally, even these seem to disappear (Fig. 15, cl. tb. right). Since
the tubules agree in size with these elongated, later vacuolated cells,
I conclude that their lumina are intracellular. Since the so called
vacuolated cells —many of which at least are, according to my view,
tubules — pass uninterruptedly from one segment to the next through
the central opening of the septa, there may easily exist continuous
tubules running from the basal segments into the calyx.

The only excretory tubules which have been heretofore mentioned as
occurring in Endoprocta are the pair which lie in the calyx and open
into the atrium between mouth and rectum. Farther on, I shall
have occasion to inquire into the probable significance of the facts here
recorded.

In specimens of Urnatella gathered in the fall and examined while
living, I found the basal segments filled with yolk granules (Plate I.
Fig. 4). Even in stocks collected in July, the basal segments contained
a greater or less amount of yolk. In passing from the distal towards
the proximal end of the stalk, one can trace the development of this
yolk, and since this is a matter of some general interest I have paid a
little attention to it.

If we examine one of the middle segments of a stem whose basal
segment is full of large yolk spherules, we shall find reserve stuff form-
ing in some of the elongated cells. It appears as a fine granulation
throughout the whole cell, except immediately surrounding the nucleus
(Plate II. Fig. 13). Not all of the cells appear thus granular, for
others are at this stage vacuolated, Still later (Fig. 14) many of
the fine granules appear to have fused into larger ones, and these are
separated by clear spaces. The cells have enlarged very greatly (those

8 BULLETIN OF THE

in Figs. 13 and 14 being equally magnified), and their boundaries
seem to be at some points discontinuous. Finally, in the basal seg-
ment we find all trace of cell boundaries lost, and the central space
filled by a mass of large and small yolk globules, nuclei, and proto-
plasmic débris (Plate II. Fig. 8).

The series of stages in the formation of yolk, which are shown in suc-
cessive segments of the same stalk, points to the conclusion that yolk is
in this case formed inside of the cell by a fusion of many excessively
small granules, —a conclusion similar to that reached by Stuhlmann
(87, p. 23) for the eggs of a Teleost. Hand in hand with the develop-
ment of yolk there seems to go a diminution in the amount of proto-
plasm, which therefore seems to be broken down in the process. The
details of this process resemble remarkably those of the formation of
the yolk in the statoblast of Phylactolsoma, for an accurate description of
which we are especially indebted to Braem (90, p. 76).

Urnatella is quite unique, so far as I know, among all Endoprocta in
the storage of food material in its stalk. This is doubtless of high
physiological importance, as I shall try to show later.

Septa occur at the constrictions between segments, and separate the
latter from one another. They are composed of a circular fold of ecto-
derm, whose free edge surrounds a small opening, through which spindle-
shaped vacuolated mesenchymatous cells pass. The adult condition is
easily interpreted by reference to the development, an early stage in
which is shown in Figure 24 (Plate IV.). Between the layers of the
ectodermal fold a perforated disc of cuticula — continuous at its outer
edge with the superficial cuticula — becomes laid down. This cutica-
lar disc increases in thickness with increasing age.

The most distal septum, which separates stalk and calyx, is more
complicated than the others. The complication is due to the fact that
mesenchymatous cells have placed themselves in and above the opening
of the septum, and havo flattened themselves out perpendicularly to the
axis of the stalk, while still allowing the vacuolated cells to pass at
their margins into the calyx from the stalk. Thus the transversely
flattened mesenchymatous cells appear to send out horizontal processes
between the tubular cells (Plate IV. Fig. 25).

The flattened cells which lie above the opening of the septum are
seen, in longitudinal section of the stalk, to be arched over the opening.
It results from this that the smaller, lower cells lie partly enclosed by
the larger upper ones (Plate III. Fig. 18).

A septum between stalk and calyx agreeing even in detail with that

MUSEUM OF COMPARATIVE ZOOLOGY. 9

just described for Urnatella has been described and figured by Ehlers
for Ascopodaria. Such a highly complicated septum seems indeed
to be common to the Pedicellinid&, which in this respect appear more
nearly allied to Urnatella than does Loxosoma, in which such septa are
absent.

In Pedicellina Benedeni also, which has a segmented stalk, the seg-
ments are separated from one another, according to Fottinger (87,
pp. 301—303), by perforated septa.

2. The Calyx.

Under this heading I shall treat successively of the body wall (includ-
ing the lip of the atrium, and the tentacles), the atrium, the alimentary
tract, the body cavity, the nephridia, the sexual organs, and the nervous
system.

A good idea of the external form of the expanded calyx may be gained
from Leidy’s figures. My own, having been drawn chiefly from preserved
material, show the polypide for the most part in a retracted condition
(cf. Plate I. Fig. 2). When thus retracted, the atrial opening does not
lie at the apex of the calyx, but is thrown sharply over towards the oral
aspect (Plate III. Fig 18).

The body wall is composed of a single layer of excessively thin epi-
thelium, — the ectoderm, — which has secreted a thin cuticula. This
cuticula is thickened in places, producing papilla, which are irregularly
scattered over the calyx. At the lip of the atrial opening, as the pas-
sage into the atrium in the retracted condition may be called, the ecto-
derm is thicker than elsewhere, and folds back upon itself until it reaches
the base of the lophophore. This backward-reflected portion I shall
call by the name kamptoderm, for although its homology with the kamp-
toderm of Ectoprocta may not be entirely beyond doubt, I cannot see
any important difference between the two structures, either in their adult
relations or in development. As in the Ectoprocta, so here the polypide
is formed in the retracted state, and the atrial opening does not break
through until a late stage is reached.

In Ectoprocta the line of union of the kamptoderm and body wall,
i.e. the lip of the atrial opening, is marked by a thickened ring com-
posed of elongated ectodermal cells, at the base of which lie the fibres of
a sphincter muscle. This organ constitutes the * Randwulst ” of Phy-
lactolemata or the “ Diaphragma” of Gymnolemata. Does an organ
homologous with this occur in Urnatella? Tt is in keeping with the
more primitive organization of the Endoprocta that, although a corre-

10 BULLETIN OF THE

sponding. region exists and sphincter muscles are found here, it has not
become so distinctly differentiated from the rest of the body wall as has
the Randwulst of Phylactolemata. This region may be designated the
lip of the atrium.

When the tentacles are expanded, —a secondary condition, as the
development of the calyx shows, — the lip of the atrium forms a circu-
lar ridge lying at the base of the tentacles (Plate IV. Fig. 26, spht. atr.).
When the tentacles are drawn in tightly, the lip of the atrium becomes
puckered. Two of the folds resulting from this process are shown cut
across in Figure 19 (Plate ILL.) above the middle of the atrium.

The tentacles of Urnatella in three cases in which I counted them on
transverse sections, as well as on the entire animal, numbered twelve ;
in one case, thirteen. In addition to these numbers, Leidy ('84, p. 10)
found sixteen (usually) and fourteen. In the specimen with thirteen
tentacles, the odd one was placed on the anal aspect of the calyx in the
median plane. It appeared shorter than the others. In one case with
twelve tentacles, observed fully expanded, the two tentacles of the anal
aspect lying nearest the median plane appeared shorter than the re-
maining ones. Leidy does not refer to this point, and his figures
afford no satisfactory evidence as to the occurrence of this condition
in his specimens.

The tentacles are each composed of a cylinder of columnar epithelium
surrounding a narrow central region which is filled with mesenchymatous
tissue. In addition, on each of the lateral aspects of the tentacle
there is a muscle, composed of one to three fibres lying side by side
(Plate IV. Fig. 27, mu. ret. ta.). The epithelium is ciliated on the lat-
eral and inner faces of the tentacle.

The atrium is bounded by the tentacular corona on all sides. The
floor of the atrium passes into the mouth in the oral region, and rests
upon the rectum in the aboral region, At the centre there opens into
it an elongated pocket, the cloaca, The lateral angles of the mouth
are prolonged aborally, and form two grooves which open into the atrium
along the lateral margins of the floor (Plate IIT. Fig. 19, sul. atr.).
These, which may be called the atrial grooves, approach each other and
become shallower as they pass aborally upon the atrial wall, until they
disappear in the median line above the rectum. The epithelium lining
the grooves is ciliated.

This “atrial groove ” exists also in Pedicellina echinata according to
Nitsche (69, pp. 21, 22), and, according to Ehlers ('90, pp. 52, 53, 59,
60), in his Ascopodaria macropus also. Its function, as has been fre-

MUSEUM OF COMPARATIVE ZOOLOGY. 11

quently pointed out, must be to carry the particles of food from the bases
of the tentacles towards the mouth.

The mouth is limited on the side towards the anus, and separated
from the cloaca by a fold, which may be designated the inner lip. This
represents the organ often referred to as the epistome, — a term which
implies an homology with the organ of the same name in Phylactole-
mata. This organ is of greatest size in Rhabdopleura and Loxosoma
among Endoprocta, where it is elevated far above the general level of the
floor of the atrium. It is less marked in the Pedicellinideo. In Urna-
tella it is not at all evident, because its upper edge is not higher than
the roof of the rectum which forms the actual floor of the atrium be-
hind. If, on the contrary, the cloaca extended underneath the rectum,
as it does in the Pedicellinideo, causing a great space between it and
the intestine, and thus making the roof of the intestine the floor
of the atrium, the inner lip would appear as a very prominent organ
(of. Figs 18),

On the outer edge of the mouth, and forming the “ outer lip,” is a
prominent horizontal fold of the oral part of the atrium (Fig. 18, loph.
atr.). This fold extends aborally, running parallel with, and forming the
outer wall of the “atrial groove ” (Fig. 19). The fold gradually becomes
less pronounced towards the aboral aspect of the atrium until (in Fig.
18) it forms only a slight swelling of the atrial wall over the rectum.

Alimentary Tract. — As in other Endoprocta, so in Urnatella one can
distinguish four regions in the alimentary tract : oesophagus, stomach,
intestine, and rectum. The wall of the alimentary tract is composed
throughout of a ciliated epithelium, except in the so called hepatic cells
of the upper wall of the stomach. In this exception Urnatella agrees
with other Pedicellinidee, but for Loxosoma Harmer makes no such
exception.

I find a highly refractive basement membrane lying at the base of
the digestive epithelium. This stains deeply in hematoxylin, so that it
can be seen with a low power as a distinct line surrounding the alimen-
tary tract. Such a condition, described by Nitsche (69, p. 19) for
Pedicellina, has been denied by Ehlers (90, p. 72) to exist in Asco-
podaria.

There is an indication of a specialized intestinal (sphincter) muscle
surrounding the opening leading from intestine to rectum, as well as at
the anus. The structure in question, which appears in section (Fig. 18,
spht.) as highly refractive and deeply staining areas on either side of
the opening, seems to lie in the epithelial wall of the alimentary tract.

BULLETIN OF THE

Since elsewhere in the calyx of Urnatella, and in other Bryozoa, muscles
arise from mesodermal tissue, I am inclined to believe that, notwith-
standing their position, they are formed from mesenchymatous cells
which have crowded in between the epithelial cells.

The free surfaces of the epithelial cells are different in the various
regions of the alimentary tract. On the wall of the wsophagus they
run out into tolerably close-set and long cilia, which become longer
towards the deep end of the wsophagus, and project into the stomach at
its cardiac end (Fig. 18, @.). Along the lateral and lower sides of the
stomach one finds longer sparsely distributed cilia, and at their base a
close-set layer of short rods (Stäbchen). Around the pyloric opening
of the stomach there is a ring of elongated cilia. The intestinal epi-
thelium bears chiefly short, stout, and close-set Stäbchen. The cilia of
the rectum are confined to the upper wall, and are of medium length,
and sparse. ‘These are omitted in Figure 18.

The position of the mouth has been already defined. It leads into a
broad esophagus, constantly narrowing as it, descends until it opens into
the stomach at the lower (oral) end of the latter. Its wall consists of
a columnar epithelium, some of whose cells are vacuolated, especially at
the base of the epistome. The inner wall of the csophageo-gastric
opening is provided with a valve-like process, covered, like the opposite
wall, with large cilia.

The pear-shaped stomach is lined below and laterally by a cuboidal
or short columnar epithelium, above by a highly granular, more or less
vacuolated, non-ciliated epithelium, whose free ends are ragged, sending
processes into the lumen of the organ. These constitute the so called
hepatic cells. Their granules stain deeply in hematoxylin. In the
epithelium of the lateral and lower walls of the stomach deeply stain-
ing cells alternate irregularly with less deeply staining ones (Fig. 18).
These cells are also full of small vacuoles.

The stomach is separated from the intestine by a circular ridge,
formed of elongated cells bearing long cilia, — the pyloric valve.

The cells of the conical intestine are cuboidal, and possess a thick
external and internal membrane, which stains deeply in haematoxylin.
The passage into the rectum is restricted.

The epithelium of the rectum is composed of flattened cells. Its
lower wall is closely applied to the upper wall of the intestine, the two
walls being separated by the basement membranes of the two cell layers.

The anus lies, as already stated, on the aboral wall of tho cloaca. At
this point the wall of the rectum becomes continuous with the floor of

MUSEUM OF COMPARATIVE ZOOLOGY. 13

the atrium above and the aboral wall of the cloaca below. The cells
surrounding the anal opening are slightly larger than their neighbors.
Lying apparently in the basement membrane is the anal sphincter
already mentioned.

Under the term body cavity I include the space lying between the
alimentary tract and the outer body wall and atrium. This space,
which is much reduced in the retracted condition of the polypide, con-
tains indifferent mesenchymatous tissue, tubular cells, muscles, the
excretory and sexual organs, and the nervous system.

The body cavity is bounded by no other epithelium than the ecto-
derm of the body wall and atrium, and the entodérm of the alimentary
tract; that is to say, there is no mesoderm.

The indifferent mesenchymatous cells have been seen to best advan-
tage in the end of a regenerating stalk shown in Figure 3. Here the
cells, which were studied while living, could be seen migrating on the
inner surface of the ectoderm, and extending through the central region.
Such wandering cells were filled with highly refractive granules.

The tubular cells of the body cavity lie chiefly at the base of the
calyx, near its attachment to the stalk. In this region also I have seen
in the living animal a flickering ciliate movement, and in the sectioned
animal flame cells.

I have been able to distinguish only two systems of muscles in
the body cavity of Urnatella, — the sphincter of the atrium, and the
muscles of the tentacles.

The sphincter of the atrial opening (Plate III. Fig. 18, Plate IV.
Fig. 26, spht. atr.) is composed of circular fibres lying in the lip of the
atrium. Taken together, the fibres form a folded sheet, U-shaped on
cross section, the convexity of the U being directed upwards. The
function of this muscle is, of course, to constrict the atrial opening, and
thus to protect the tentacles and parts below.

The tentacular muscles consist of the pair to each tentacle already
mentioned. These run from the apex of the tentacle to the base, where
they diverge to the right and to the left, and, after breaking up into
many branches, pass through the ectoderm to become inserted upon the
cuticula of the body wall. Leidy (84, p. 10) saw these tentacular
muscles.

Differentiated muscles do not seem to be abundant in the calyx of
any of the Endoprocta. Tentacular muscles are unknown in other spe-
cies. On the other hand, Ehlers (90, pp. 64, 65) has described two sys-
tems for Ascopodaria which I have not seen in Urnatella, namely, lateral

14 BULLETIN OF THE

wall muscles and transverse muscles, the latter running from the right
to the left wall. The fibres of these muscles also break up into branches
before making their attachments.

LExcretory Organs. — The chief excretory apparatus in Urnatella lies in
the calyx below the ectodermal floor of the atrium (Plate III. Fig. 18,
Plate IV. Fig. 22, nph.). It consists of a pair of tubules which unite
proximally and open by a single pore into an unpaired cavity, which in
turn opens into the atrial chamber at about the centre of its floor.
This unpaired cavity is the one I have proposed to call the cloaca.

From its opening into the cloaca the unpaired tract of the excretory
tubules, which may bé designated efferent duct, runs oralwards and down-
wards, and then divides, the two tubules following the posterior wall of
the wsophagus. Finally, the tubules turn back upon themselves, run-
ning outward and towards the rectum. The whole excretory apparatus
has thus the form of the Greek letter Y.

Kach tubule ends blindly in a flame cell which bears the characteristic
cirrus, exactly similar to that found in the stalk (page 6). Figure 22
(Plate IV.) shows the end of the tubule of the left side. The plane of
this section was such that it cut the posteriorly reflected region of the
excretory tubule of the left side throughout all but the middle of its ex-
tent. At the middle line the tubules of both sides sink below the plane
of the section, so that it is'the efferent duct which is cut at mph. in the
median plane.

One of the youngest individuals in which I have found a nephridium
is that from which the section Figure 29 (Plate IV.) was drawn. The
efferent duct (neph.) appears to be composed of two elongated cells
placed end to end. Running through the midst of these is a poorly
marked lumen, partly filled by a granular substance. This and one or
two other similar cases seem to me to support strongly tho view of tho
intracellular nature of the lumen of the nephridium.

The evidence derived from the adult condition is less satisfactory, but
points to the same conclusion. Thus one finds on cross section of the
tubules that the lumen is not sharply limited like the exterior of the
tubule. In fact, one sometimes finds delicate threads traversing the lu-
men (Plate III. Fig. 21, a-c). In one of these sections two nuclei are
cut across, which in so far militates against my conclusion that tho
lumen runs inside of single cells placed end to end. But I believe these
to be the nuclei of two adjacent overlapping cells.

On account of the evidence just presented, I regard the nephridium
of Urnatella as having an intracellular lumen and ending blindly in a

MUSEUM OF COMPARATIVE ZOOLOGY. 15

flame cell, and am thus brought into complete agreement with the results
obtained by Harmer from Loxosoma, His conclusion (’85, p. 279) that
the Endoproct nephridium is probably to be regarded as a head kidney,
like that of Trochophores, seems to me to be justified, A further discus-
sion of this topic must be deferred till towards the close of this paper.

The nepbridium opens, as already stated, not directly into the vesti-
bule, but into a pocket of it, the cloaca (Plate III. Fig. 18, Plate IV.
Fig. 28). The wall of the cloaca consists of a layer of cuboidal epithe-
lium which is perforated by three openings, —the anus, and the proximal
openings of the efferent duct and of the vas deferens.

The cloaca is an organ which does not occur in any other Endoproct.
Its existence here is due to the unique position of the rectum and anus,
and of the opening to the vas deferens. In Loxosoma the last is very
far removed from the opening of the nephridia. In the Pedicellinides
the two openings are less distant from each other. In the male of
P. Benedeni (Fottinger, ?87, Plate X. Fig. 16), the anus and the open-
ings of the nephridium and vas deferens are quite near together; but in
the fomale the oviduct of that species opens far distant from the other

organs.

5

Sexual Organs. — Curiously enough, I have not found among all the
individuals sectioned any ripe females. All the mature individuals of
the lot collected by me on July 4th from the Schuylkill appear to have
been males. In no individual did I notice any difference in the position
of the ducts, such as obtains between the two sexes in Pedicellina Bene:
deni, — no trace of an incubatory chamber between rectum and the floor
of the vestibule.

In the male organs two parts may be distinguished, — testis and vas
deferens. The testis (Plate IV. Fig. 23) is a paired ovoid body lying be-
tween rectum, intestine, and the floor of the atrium. In the figure given
one observes spermatozoa in various stages of development, the wall of
the sac constituting the germinative epithelium.

The vas deferens (Plate III. Fig. 18, Plate IV. Fig. 28) is an un-
paired U-shaped tube, the concavity of the U being turned oralwards.
The wall consists of a cuboidal epithelium which is ciliated, at least at
the proximal end of the tube.

The vas deferens of Urnatella resembles in form that of P. Benedeni
(Foettinger, '87, Plate X. Fig. 15), which in turn seems to be more com-
plicated than that of Ascopodaria.

Nervous System. — I have been able to make only a superficial study
of this system. The main ganglion (Plate III. Fig. 18, Plate IV. Figs.

VOL. XXIV. — NO. 1l. 2

16 BULLETIN OF THE

22, 29, gn.) lies between vas deferens and nephridium, is elongated trans-
versely, and slightly constricted in the median plane. One can distin-
guish a central'region composed of fine fibres running transversely, and
a cortical region of cuboidal, deeply staining cells with large clear nuclei.
I have not succeeded in tracing any peripheral fibres from this gan-
glionic mass, as Ehlers has done in another Endoproct.

III. | NoN-SEXUAL REPRODUCTION.

1. Architecture of the Stock.

Urnatella forms stocks by budding. As in other Bryozoa, the buds
are normally produced in a very regular manner. TI believe that I have
determined the law in part, although it has exceptions, as the law
of budding in every Bryozoan stock has. As Leidy has stated, several
vertical stalks may arise from the same horizontal plate. A single stalk
may remain unbranched, giving rise to new individuals at its distal end
only, or it may give rise to branches which come off irregularly from a
few segments. Usually only one branch arises from a single segment,
but occasionally two do so (Plate V. Figs. 35, 38).

The length of these branches and the number of their segments decrease
towards the distal end of the main stalk, and the conclusion seems jus-
tified that it is only at the distal end that they are formed. Excepting
for these occasional branches, the lower segments of the stalk are bare of
any branches or polypides. The distal end, however, is usually crowded
with polypides so thickly as to make it difficult to count them or to de-
termine their points of attachment (cf. Plate V. Fig. 30).

A very casual observation, however, shows that the buds from the
main stalk are of two kinds; first, those which have given rise to a
linear series of segments at the distal end of which is a calyx, and these
are what I have called branches; and, secondly, those consisting of a
stolon-like process, from one surface of which arise calyces resting upon
a stalk of usually one segment only. The surface of the stolon upon
which such polypides are placed is a definite one, namely, that which is
turned towards the distal end of the main stalk. The budded branches
and also the stolons are grouped upon the oral side of the adult stalk.

In the clearest cases, in young or not too richly branched stocks, I
find the branch arising on the oral surface of the segment, and between
two stolons, which are therefore more nearly lateral (Plate V. Figs. 37,
38, 40, 42).

MUSEUM OF COMPARATIVE ZOOLOGY. 17

One may therefore say that usually the branches are median and oral,
the stolons lateral, Again, the branches give rise, like the primary stalk,
to two kinds of buds, branches and stolons. The stolons give rise only
to stalks of one segment each, bearing a calyx distally. These calyces
are so placed that their oral surface is directed towards the distal end of
the stolon. I have not found more than two individuals borne upon a
stolon,

I have previously (91, p. 72) tried to show how all buds in the stocks
of Bryozoa are to be referred back to embryonic tissue lying at the tips
or margin. In Endoprocta, however, the extreme tips seem to be oc-
cupied by a polypide, and the embryonic tissue lies in a zone at the base
of the latter. This difference may be regarded, however, as only appar-
ent, and the two conditions harmonized by conceiving the polypide in
Endoprocta retracted into the stalk, below the zone of embryonic tissue,
— the condition realized in Eetoprocta. The distal part of the stalk
will then become terminal, constituting an apical ring of embryonic tissue
surrounding the secondary atrial opening thus produced. From the oral
portion of this ring new buds — branches and stolons — are, as in Pluma-
tella, proliferated; and this process is repeated for each segment. At
(or near) the apices of these incipient branches and stolons lies a mass
of embryonic tissue which gives rise in the one case — branches — to the
stalk, the polypide, and the Anlage of new buds; and in the other — in-
cipient stolons — to the stolon and the Anlage of the individuals which
bud forth from it. The differences between the branch and the stolon
are, however, more apparent than real, as a comparison of the diagrams
Figures 59 and 60 (Plate VI.) will make clear. In one case (Fig. 59)
the mass of embryonic cells in any segment does not grow out far beyond
the youngest individual produced from it; in the other case (Fig. 60)
there is a considerable growth beyond the youngest individual, Thus in
the latter case a long stolon is produced, in the former it remains at a
minimum.

As I have already stated, in most cases, particularly in young vigorous
stocks, one meets with the condition which may fairly be called typical,
in which from one segment three buds — one median (branch) and two
lateral (stolons) — arise. This typical condition may be expressed by
the formula on the next page.

This formula will be understood by reference to the diagrams on Plate
VI., of which it is a symbolic expression. The letters represent in all
cases calyx-bearing individuals, the asterisks gemmiparous tissue. The
capital A stands for the individual which forms the main stem under

18 BULLETIN OF THE

consideration. The ¥% immediately below it represents the embryonic
mass situated at the tip of the stalk, from which all the gemmiparous
tissue distributed throughout the stock has taken its origin. As the
stalk has grown, there have been left behind the embryonic masses repre-
sented by x. Three of these have arisen together, and the process has
been repeated as often as segments have been formed. The separation
of the segments is represented in the formula by a dash. Between every
two dashes the same budding process is repeated.

x
x U ox xm
x

*

The median embryonic masses (x) give rise to the individuals a, b,
and c, and leave with each a dowry of embryonic tissue represented by x.
Each of these individuals a, b, and c, with its embryonic mass x, now
acts exactly like Aé. The lateral embryonic masses act precisely as
do the median ones.

The typical condition expressed by the foregoing formula is not often
perfectly realized, for the simple reason that the stock would thereby be-
come too crowded. The more frequent modifications are the absence of
the median bud, the absence of one or both of the lateral buds (stolons)
and, occasionally, the apparent substitution of & branch for a stolon.

The foregoing formula is one of the most complicated that it is neces-
sary to make for Bryozoa, for nowhere else are the embryonic masses so
abundant relative to the polypides, or, in other words, nowhere else are
so many embryonic masses borne by one individual. It is quite com-
mon among other species to find three embryonic masses arising from
one; but here the formation of triple masses is repeated as often as the
number of segments in the stalk. "This fact raises the question of the
significance of these segments, and leads us to discuss the subject.

MUSEUM OF COMPARATIVE ZOOLOGY. 19

2. Segmentation of the Stalk.

Leidy (84, p. 9) did not fail to call particular attention to the divis-
ion of the stalk of Urnatella into segments, and recognized their suc-
cessive production “through the process of division, very much in the
same manner as in the production of the proglottides of tape-worms
from a scolex.”

Tho phenomena of segmentation must have a special interest on
account of the relation of segmentation and strobilization. I shall
treat under this subject, first, of the phenomena of segmentation in
Urnatella; secondly, of the probable origin and significance of the
process in this genus; and, thirdly, I shall examine critically Leidy’s
comparison.

The transverse septum, which is the first indication of the separation
of the stalk into segments, begins to be formed at about the time that
the lateral buds arise, and immediately below them. It arises, as already
stated (page 8), as a ring-like fold of the ectoderm, whose free edge en-
closes a circular orifice, through which the interiors of the two segments
which are being formed are continuous (Plate IV. Fig. 24). By con-
tinued growth of this fold the diameter of the orifice diminishes, never
wholly closing, but leaving a space for the passage of the tubular cells,
to which reference has already been made (page 7). In this process the
sheath of longitudinal muscles, which originally ran immediately inside
of the ectoderm continuously from the base of the calyx to the last fully
formed septum, becomes discontinuous at the place of the ingrowing
fold, and the separated ends of the muscle fibres become attached to their
respective faces of the septum.

In order to discover the significance of this process of segmentation
we must first study its distribution throughout the Endoprocta. Usually
the calyx is borne upon an unsegmented stalk. Such is the condition,
for example, in Loxosoma, Pedicellina echinata, P. Americana, and
Ascopodaria. In other cases, Gonopodaria (Khlers, 90, p. 146) and
P. Belgica (Fottinger, '87, p. 301), there is a poorly expressed and
highly variable segmented condition. In Arthropodaria (Pedicellina)
Benedeni the stalk is distinctly and regularly segmented, there being
partitions and slight external constrictions. Finally, in Urnatella we
find the segmented condition still more sharply expressed.

There is a suggestive parallelism between the formation of segments
and the production of buds. While with one exception all species with
unsegmented stalks produce no buds, and none have yet been described

20 BULLETIN OF THE

as occurring on Gonopodaria and P. Belgica, budding does occur to a
slight extent on Arthropodaria, and still more profusely on Urnatella.
There is one genus, however, Barentsia (Vigilius, 7844, p. 86), which buds
sparsely and has an unsegmented stalk. Here, however, as in Urnatella
(Plate VI. Fig. 58) and Arthropodaria (1), there is a septum between the
branches and the main stalk.

Finally, in all Pedicellinidee which have been studied carefully there
is a septum between the stalk and calyx.

One may bring these scattered facts of distribution of the septum into
some sort of unity by saying that in the Pedicellinide every calyx and
every lateral branch (where such exists) is partly cut off from communi-
cation with the interior of the mother stalk by septa. This is effected
by the production of these organs at the place of junction of the calyx
and the lateral buds to the stalk, and in some cases also by the partial
division of the stalk itself into compartments by septa. The transverse
septa must have either a morphological or a physiological significance.
They might be regarded as a part of the wall of zocecia, equivalent to
the zoccia of Paludicella, and therefore to be morphologically equiva-
lent to the septa of Paludicella. On this hypothesis each segment of
the stalk of Urnatella would be an (incomplete) individual. The idea of
the individuality of the segments would seem to be sustained by the fact
that each gives rise to similar buds, and that apparently any segment
has the capacity of regenerating the lost calyx, or end of the primary
stalk. On the other hand, dissepiments might be regarded as structures
which had grown across the originally unsegmented stem to fulfil some
need of the organism. Comparative anatomical studies seem to me to
favor the second view. In Loxosoma, and in most marine Pedicellinide,
we find a stalk without dissepiments. In other forms dissepiments
are few and variable in number, in still others they are constant in
occurrence, and from the segments arise buds. Finally, in Urnatella,
the segments are separated by the dissepiments, each becomes more in-
dividualized, and has the capacity of giving rise to buds having the same
arrangement. The differentiation of the separate segments has gone so
far that one can hardly see in the stalk of Urnatella, resembling a string
of beads, the assumed smooth cylindrical stem of its Pedicellina-like
ancestors.

If we seek for an explanation of the dissepiments, I think it is to be
found in the protection of the stock against the influx of water and de-
stroying organisms at the time of the loss of calyx or lateral branches,

which would make regeneration impossible. This is the same hypothe-

MUSEUM OF COMPARATIVE ZOOLOGY. 21
sis which I offered in a former paper (91, p. 40) to account for the
formation of a septum in Ectoprocta.

In favor of this hypothesis are not merely the need of such an appara-
tus on account of the frequent loss of the calyx and the lateral branches
through accident, and the fitness of this mechanism for the function,
but also the existence of the special mechanism of radiate cells, covering
over the opening in the dissepiment between the calyx and stalk, — :
dissepiment which will be most useful in the manner indicated by this
hypothesis, owing to the delicacy of the calyx and its liability to acci-
dent. When the lateral branches or the terminal calyx become de-
tached from the parent stem, we find that the pore in the septum,
remaining behind as a part of the wall of the stalk, has become sealed
by a cuticular plug. So also Ehlers (°90, p. 22) in Ascopodaria. In
this case we can see the utility of the dissepiment, and can infer its
value in those positions where it is not certain, but only possible, that
it may be called into play. My conclusion then is, that the dissepi-
ments have a purely physiological meaning, possessing a protective
function, and that the segments of the stem are only physiological di-
visions of a primitively undivided stalk, which have perhaps no other
significance than that they are parts separated by the dissepiments.!

It follows naturally from the foregeing hypothesis, that the segmen-
tation of the stalk has succeeded, rather than preceded, the condition of
bud formation from the stalk, it being rendered desirable owing to the
greater danger to mutilation to which the stalk is exposed. From this
standpoint we can see why buds should be produced on each segment in
a similar manner. The relative profuseness of budding in Urnatella is
explainable on other grounds.

Examining more closely the relation of this process to the production
of proglottides in a tape-worm, — accepting the view that the production
of proglottides is fundamentally a process of continual regeneration of
lost parts, — there seems to be an important difference in this, that the
growth of the stock of Urnatella is limited, more than ten or twelve seg-
ments being rarely formed, while an indefinite number of proglottides
are produced. The limited growth of the Urnatella stem seems to indi-
sate that the production of segments is not the production of new parts,

1 Freely branching stocks of Hydroids have septa interpolated at the base of
the hydranth, which is peculiarly liable to fall off, and sometimes in the middle of
the stems. The occurrence of such similar structures throughout the two most
profusely branching groups of Metazoa is further evidence for the validity of the
physiological explanation of them which I have offered.

22 BULLETIN OF THE

but the division, progressing towards the calyx, of an originally simple
stalk into a number of parts, — that the growth of the Urnatella stalk
is limited just as that of the Pedicellina stock is limited, and for the
same reason.

3. Orientation of the Individuals.

The orientation of the individuals of the stock seems to be here, as in
Ectoprocta, a very definite one. In Pedicellina and its close allies, as
well as in Urnatella, the young individuals budded from the parent
stalk face the same way as the polypide of the parent stock. Seeliger
(90, p. 571) has pointed out that in all the species of Loxosoma the
orientation of the buds with reference to the parent is definite, but
different from that found in Pedicellina. The orientation of the buds in
Loxosoma (which arise, as in Pedicellina, on the oral side of the mother)
is the reverse of that of the mother. Thus it comes about that the
oral aspect of the buds in the Pedicellinide is turned towhrds the pe-
riphery of the stock, that of Loxosoma towards the centre. This is
the same difference which has been found to prevail between the buds
of Phylactolemata and Gymnolemata, and as in this case, so in Endo-
procta the differences may be harmonized by a different method of ex-
pression. Jn all Endoprocta the oral aspect of the buds is turned towards
the centre of proliferation. For in Loxosoma new individuals are pro-
duced between the next older and the parent stock, the centre of prolif-
eration remaining in the parent stock, while in Pedicellina it migrates
away upon the end of the stolon. The typical condition in Loxosoma is
represented by the formula

a b c d * A;
that of Pedicellina by
* d c b a A,

in which the asterisks represent gemmiparous tissue.

The general statement of the relation of the aspects of the buds to the
proliferating region is the reverse of that which I have given for Ecto-
procta (91, p. 82), which reads, “In both Phylactolemata and Gymno-
lemata, the anal aspect is turned towards the gemmiparous region.”
It is important to note, however, that this difference corresponds to a
difference in respect to the part of the alimentary tract which is formed
by the principal pocket of the atrium, for this in Ectoprocta gives rise
to the posterior part of the elementary tract; in Endoprocta, on the con-
trary, to the anterior part. The differences in the process of budding in

MUSEUM OF COMPARATIVE ZOOLOGY. 23

the two groups can be harmonized, and at the same time the physiologi-
cal nature of the differences indicated, by putting the statement thus : —
In all Bryozoa, the formation of the alimentary tract begins at that end
which is turned towards the gemmiparous region (cf. page 28).

The problem of the difference in the method of development of the
alimentary tract in Ectoprocta and Endoprocta, is the same as that of
the differences in the development of the alimentary tract of Triploblas-
tica in general.

As is well known, the midgut in Triploblastica is produced by an in-
vagination whose mouth — the blastopore — comes to be in some cases
at the anterior part of the tract, in others at its posterior part. This
variation in the method of formation has been explained by the hypo-
thesis that the blastopore represents the opening into the gastro-vascular
cavity of Coelenterates, which is functionally both mouth and anus ; and
that as we find a physiological separation of the opening in many Coelente-
rates, so a morphological separation of the gastrula-opening into mouth
and anus by concrescence of the lips of the blastopore in the mid-oral line,
has occurred in the ontogeny of Triploblastica. In some cases both mouth
and anus arise by this process, in other cases only one organ, the other
arising secondarily, or (preferably) later. The part which arises later
might be regarded as a new formation, or, following Caldwell (’85, p. 23),
as derived from a part of the entoderm which had become separated
from the greater part in the separation of the two extremities of the
elongated lip of the blastopore to permit the placing of mouth and anus
at opposite poles of an elongated animal.

The application of these facts and their explanations to the facts of
the formation of the alimentary tract in the Bryozoan polypide is evi-
dent. Gastrulation takes place not in the act of first invagination of the
inner layer of the bud, but in a secondary invagination from the bottom
of the first formed sac. The blastopore does not lie on the surface of
the body wall, but has been carried below the surface, and its position is
indicated by the plane of separation of alimentary tract and atrium,
where the roof of the gut and the floor of the atrium have been produced*
by concrescence of the lips of the blastopore. The atrium then is in
no way lined by entoderm; it is merely a precociously developed, pro-
tecting pocket of the body wall, which occurs in that region in which
invagination of the entoderm is to take place. The primary atrial open-
ing is not at all the blastopore, as some authors have called it.

1 Compare Seeliger, '89*, pp. 181, 182; so also in Paludicella, see my earlier
paper, ’91, p. 19, and probably in Phylactolemata, cf. Krepelin, 92, p. 83.

24 BULLETIN OF THE

My conception of the relation of the processes of atrium formation and
gastrulation will be more easily understood if we compare the formation
of the polypide with the early stages in the development of the egg of
Sipunculus, as given by Hatschek ('83, pp. 78-81). Here gastrulation
occurs only in the depths of the invagination at the vegetative pole ; the
more superficial part of the cavity isthe trunk amniotic cavity. This
corresponds to the atrium of Bryozoa, and, like it, is lined by ectoderm.

The concrescence of the lips of the blastopore takes place slightly
differently in Endoprocta and Ectoprocta, so that in one case the blastd-
pore persists in the region of the permanent mouth, and the procto-
dieum appears later; whereas in the other case the blastopore persists
in the region of the permanent anus, and the stomadeum arises later.

In all Bryozoa, however, the main ganglion arises in the region of con-
crescence, precisely as the ventral nerve cord is now known to do in most
invertebrates. Thus additional support is given to the idea that the
ganglion of Bryozoa is a subesophageal one.

4. Histology of the Buds.

The histological structure of the branches is exactly like that of the
parent stalk. The minute structure of the stolons deserves a word of
description. The base from which the individuals arise is filled with
mesenchymatous cells, which at the distal, growing end are closely
packed and deeply stainable (Plate VI. Figs. 51, 57, sto.). Some of the
cells are differentiated into muscle fibres, which run up into the stalks
of the individuals that have arisen from the stolon (Fig. 57, mu.). The
epithelium at the distal end of the stolon is composed of cuboidal or
slightly elongated cells. At the base of the stolon the epithelial cells
are greatly elongated and closely packed together (Figs. 51, 57, cl. sec.).

At the point of attachment to the parent stem the stolon is consider-
ably constricted, the epithelium forming an inward fold surrounding a
small opening through which the parenchymatous tissue of the bud and
the mother stalk are continuous (Plate VI. Fig. 58, dé. sep.). Between
the layers of this circular fold is secreted a cuticular disk, perforated at
its centre.

5. Formation of New Stocks.

The development of Urnatella from the egg has never been seen, and
I was naturally anxious to get embryological material. On looking over
my preserved material I found no trace of eggs, although there were

many ripe males. I have no direct knowledge, therefore, as to whether

MUSEUM OF COMPARATIVE ZOOLOGY. 25

I was too early or too late for embryological material, although from
some indirect evidence drawn from what follows I am inclined to think
I was too early.

The youngest stocks found I obtained in large numbers, and they
were of almost exactly the same age. Moreover, they agreed with the
youngest stages found by Leidy and figuted by him (84, Plate I. Figs.
5, 6, 7). Ihave represented some of these in Plate V. Figures 31, 32,
and, enlarged, in Plate II. Figure 15, and Plate V. Figure 46.

I have already (pages 3-6) called attention to some of the characters of
the stalks of these young individuals. It remains to mention the “basal
plate” of such stocks.

Unlike its condition in the adult, the “basal plate” is a relatively
large organ of elongated cylindrical form (Plate IL. Fig. 15). Its outer
epithelium consists of sharply demarked cells, cuboidal above, columnar
below. This epithelium has given rise to a cuticula thin and dense
above, and thicker and less refractive below. ‘The columnar cells are
somewhat less deeply stained than those at the end of the “basal plate”
furthest from the two stalks. Their nuclei lie at their inner ends.
The outer part of the cell contains spherical masses of granules. On the
whole, it stains less deeply than the inner end. Finally, one always
finds particles of dirt closely adherent to the under and lateral aspects
of the basal plate.

These facts I interpret as follows. The columnar cells of the under
side are glandular, and secrete a sticky substance which causes the ad-
herence of the surrounding particles of dirt, and thus serves to anchor the
young stock.! ‘The tissue of the interior of the “ basal plate ” is remark-
able, and difficult of interpretation from sections alone. In such sections
one sees bands running through the middle region and crossing at various
angles. Each of these “bands ” is a nucleated cell, and probably repre-
sents a muscle fibre (cf. Plate V. Fig. 46). One sees also fibres having a
different appearance running radially from the base of the stalk to the
columnar cells of the “basal disc" over quite a long stretch (Plate II.
Fig. 15). In addition to these protoplasmic structures there are long
clear spaces which are bounded by thin membranes and contain occa-

1 Seeliger (’90, p. 573) finds in Pedicellina a glandular differentiated zone on the
base of the stolon, where it comes in contact with foreign bodies. This, he says,
an be traced unbroken through the entire length, especially in young stocks, con-
sisting of few individuals. The foot gland of Loxosoma is an organ fulfilling a
function similar to that of the glandular cells of the stolon, but an homology of the

two organs can hardly be maintained.

26 BULLETIN OF THE

sional nuclei. These are the tubular cells already referred to. Some of
them contain the agglutinated cilia characteristic of flame cells, and they
arise from large cells, which must indicate the beginnings of the excretory
tubules of the stalk already mentioned.

What is the meaning of the fact that no younger stages than these
occur, although such and all older stages are abundant? Have these
young stocks been derived from fertilized eggs, or have they some other
origin ?

I have already referred to the fact that the great mass of the buds of
any Urnatella stock are found at the upper end of the parent stalk. The
lower and middle parts of the stalk possess few buds, although they once
constituted the upper end of the stalk. What has become of the buds
which have been lost? Leidy asked this question, and the facts led him
“to suspect that the branches are spontaneously and habitually detached
from the parent stem, to become elsewhere attached, and thus form new
colonies.” I have evidence that raises the suspicion of Leidy to as near
certainty as can be obtained by use of the morphological method. Zhe
“youngest stocks” are derived from the stolons of the parent stalk, which
habitually become free for the purpose of founding new stocks.

To establish this proposition it will be necessary to show, (1) identity
of structure between old lateral buds and young stocks, and (2) the
scar of attachment of the young stock to the parent stalk (cf. Plate VI.
Fig. 58). Of identity of structure there can be no question. Often it
would have been quite impossible to distinguish between young stocks
and “stolons” which had been violently broken off from the parent
stalk and were lying loose in the bottle, were it not for a single criterion,
namely, the young stocks had dirt adhering to their lower surface. An
application of the second criterion leads to the same positive result. In
series of thin sections of young stocks one can always find at one side
of the median plane the scar of former attachment, which appears as a
thickening of the cuticula into which ectodermal cells may sometimes be
seen penetrating (Plate V Fig. 47). We have here, then, a method of
non-sexual propagation quite similar to that obtaining in Loxosoma, where
the buds habitually drop off, so that this genus is commonly said not to
be stock-producing. This resemblance must be regarded as being purely
a physiological one, and not in the least implying any closer relation-
ship of the two genera.

I have already expressed my belief that the stolons are thrown off reg-
ularly for the purpose of founding new stocks. On this assumption we
can account for the rapid growth of the embryonic tissue giving rise to a

MUSEUM OF COMPARATIVE ZOOLOGY. 27

stolon-like body, which will become a basis of support for the new stock,
the “basal plate.” One can thus account for the thickened cells of
the under side of the stolon, which appear before detachment (Plate VI.
Figs. 51, 57). The question remains, Do the median branches play a
similar rôle to the stolons? I do not think they do, for the reasons,
(1) that, having no basal plate, they are not physiologically fit for form-
ing new stocks; (2) that I have found no new young stocks having
one parent stem with one or two generations of budded individuals, —
the condition of the median branches ; (3) that, on the contrary, one
often finds such median branches persisting on even the lower segments
of the stock, (Plate V. Figs. 35, 39, 40. Compare Leidy, '84, pp. 8, 9;
Plate 1. Fig. 4.) Since the median branches frequently persist as a part
of the parent stock, — they are not produced in the first place on every
segment, — I conceive their function to be the increase of the number
of proliferating points in the stock itself.

Starting with the young stock, one can find all stages of growth up to
the most complicated conditions (Plate V. Figs. 33, 36, 43, 44). Dur-
ing the growth of the stock the basal plate gradually undergoes changes.
The parenchyme becomes filled with yolk globules (Plate V. Fig. 49),
and the cuticula becomes thick and dark.

Concerning the morphological significance of the basal plate a few
words must be said. I regard this as a stolon morphologically equiva-
lent to the stolon of the Pedicellinide. In the latter group, as is well
known, the individuals are budded from the upper side of a repent cy-
lindrical stolon, which constantly produces new buds at the growing end,
and which becomes separated into segments by the formation of trans-
verse dissepiments. There is no such stolon in the adult Urnatella,
which is sharply separated from the Pedicellinidee by this single ,charac-
ter. The presence of a stolon in the young stock indicates a derivation
from an ancestral condition possessing a stolon in the adult.

If, however, the “stolon ” of the young Urnatella stock is homolo-
gous with that of the Pedicellinida, we ought to find it, sometimes at
least, giving rise to more individuals than two, and perhaps becoming
segmented. Both of these conditions are occasionally fulfilled, Leidy
observed that three, four, or even five stems may arise from a com-
mon “basal plate.” I have observed only three with certainty. Two
cases of this are shown in Plate V. Figs. 48 and 49. In the first of
the two cases distinct perforated dissepiments were observed dividing
the stolon (basal plate) into three segments, out of each of which a

single stalk: arose.

BULLETIN OF THE

6. Formation of the Individual.

Recent careful studies on the formation of the individual in other
Endoprocta, especially by Seeliger (’89*), render a detailed study of this
process less necessary. There are a few points concerning the physio-
logical rather more than the morphological features of this process which
I have attended to in this case in order to test certain conclusions which
I had arrived at from the study of the earliest stages of budding in
Ectoprocta.

First, the budding regions are areas of cuboidal cells, with relatively
large nuclei and deeply staining plasma. Such a condition is found in
both the ectoderm and mesoderm of the proliferating region (Plate VI.
Figs. 50, 57). The relative enlargement of the ectodermal cells, and
at the same time a bending of the whole layer outward, give rise to
the first fundament of the new individual. The musculature of the
new individuals is certainly not derived directly from that of the old
stalk, for this takes no part in the outbending. Upon the apex of the
cylindrical protuberance thus formed the polypide is produced. The
details of this process I have not followed.

Secondly, the position and time of origin of the buds arising from the
stalk are very definite. They make their appearance in a zone lying in
the lower part of the segment (Plate VI. Figs. 50, 58-60), and shortly
after the formation of the dissepiment which lies just below.

One of the questions the re-examination of which most interested
me was that of the origin of the alimentary tract, since this is stated to
arise differently in Endoprocta and Ectoproota.

In an optical section of the whole bud (Plate VI. Fig. 53), it could
be seen that the atrium was connected with the young alimentary tract
at the oral end only. The same thing is shown in the series of trans-
verse sections, Figures 54-56, in which the distal (anal) part of the
atrial chamber is not confluent with the rudiment of the alimentary
tract which touches its floor, Figure 54, but the two organs are con-
fluent at the proximal (oral) part, Figure 56. Figure 52 is from an
optical longitudinal section of the bud shown in Figure 53, taken in a
plane perpendicular to that of Figure 53. Here the alimentary tract, st.
(Fig. 52; ga. Fig. 53) is being constricted off from the atrium.

Like the young bud, the growing tip of the stolon possesses an ecto-
derm consisting of large cuboidal cells (Plate VI. Fig. 57). The mesen-
chymatous tissue also consists of thickly crowded, undifferentiated, and
deeply staining cells (Figures 51, 57).

MUSEUM OF COMPARATIVE ZOOLOGY. 29

7. Regeneration.

Like other Endoprocta, Urnatella has the capacity of regenerating
its lost calyces. Leidy (’84, p. 13) had already observed this process,
and Potts and I have (as previously mentioned) seen the same thing.
Figures 3 and 2 (Plate I.) show some of the phenomena of regenera-
tion. The formation of a new terminal calyx seems to be preceded
by the formation of the stalk part of the new terminal individual.
This new formation takes place in one of the segments near the distal
end of the parent stalk. The terminal dissepiment of the segment,
cutting it off from the outside world, is completely closed in its cen-
tre by a cuticular plug. This cuticular dissepiment (Fig. 2, di. sep.
at the left) becomes torn off from the lateral cuticula of the segment
along one edge, sometimes, as in the figure, remaining attached at
the opposite edge. I have given on Plate I. (Fig. 3) a drawing of
such a regenerating stalk before the polypide has budded from its
wall. The outer body wall has evidently taken the initiative in the
process, and mesenchymatous cells have migrated in. The drawing
(Fig. 3) was made from the living animal, and the mesenchymatous
cells could be seen changing form like an amoba. The ameeboid
cells also contained highly refractive granules, which I regard as food
material. These granules were seen moving about in the cells as the
latter changed in form.

One striking feature of the young segment was the presence of two
flickering organs, which I now fully believe to be flame cells with their
tufts of fused cilia, I have elsewhere ('91, p. 39) called attention
to the part played by amooboid cells in Paludicella on the nutrition of
the young buds. I believe that the amoeboid cells here have the same
function.

Leidy ('84, p. 13) has suggested that the segments of Urnatella might
function as statoblasts to preserve the species during the winter. Of
the correctness of this suggestion I have no doubt. The ease with
which regeneration occurs from the old segments, and the fact that in
stalks in which the calyces have been for a long time lost one sees the
cirri of the flame cells still flickering, show that the tissue surrounded
by the thick cuticula of the segments has a great capacity for retaining
life, and, under favorable conditions, for reproducing lost parts. In their
róle as statoblasts, the segments need the yolk stored up in their cells
during the summer.

BULLETIN OF THE

IV. AFFINITIES or URNATELLA.

From the frequent opportunity that I have had successfully to com-
pare the organs of Urnatella, even in detail, with those of Pediceilina
and, allied genera, there remains no doubt in my mind of its close rela-
tionship to those forms. Arthropodaria, especially in its segmentation
of the stem and the associated budding process, seems most closely to
resemble Urnatella. In three points of importance, however, Urnatella
differs from other Pedicellinid®, namely : (1) in the possession of a cloaca
(and absence of a brood-sac ?) ; (2) in the presence of water or excretory
canals in the stem and calyx, — which are so striking that they could
hardly have been overlooked if they occurred in other Pedicellinide;
and (3) in the absence of the stoloniferous type of budding.

As for the last difference, however, I have tried to show that there is
a stolon from which the individuals of the Urnatella stock arise, although
it is small. Whether this rudimentary condition of the stolon is an
ancestral or a degenerate character is doubtful; I have been inclined to
consider it the latter.

In regard to the first difference, I must point out that in the male of
Arthropodaria there is a condition resembling that found in Urnatella,
for in the former genus (Foettinger, ’87, Plate X. Fig. 8) the anus, vas
deferens, and excretory tubule open near together.

The second difference concerns a very important set of organs, and if
they should be shown to be indeed absent in Arthropodaria Benedeni, it
would lead us to conclude that in one respect at least, perhaps owing to
physiological needs, Urnatella has retained a more ancestral condition
than its near allies.

V. AFFINITIES OF THE BRYOZOA.

There are three prevailing views concerning the relationship of the
Bryozoa to other groups. According to one view (most recently and
ably defended by Ehlers) they find their nearest allies in the Gephyrea ;
according to the second view, they have sprung from the lower worms, —
from Rotifer-like ancestors, The third view (that of Hatschek) assumes
that Ectoprocta and Endoprocta are in fact not closely related, but that
the former should be placed near the Gephyrea, the latter close to tho
Rotifera.

The reasons for this difference of opinion are not far to seek. Those
who have begun their studies with the Ectoprocta, particularly with

9

MUSEUM OF COMPARATIVE ZOOLOGY. Ə

their anatomy, have been struck by their resemblance — especially in
the possession of a body cavity and of a tentacular corona — to the
Gephyrea, and particularly to Phoronis. Those, on the contrary, who
have devoted their studies chiefly to the Endoprocta, and especially to the
development of that group, have urged the second view. The third view
seeks to reconcile the two conflicting theories. I favor the second of the
views given above because of certain considerations which follow.

I propose first to show the untenableness of the third view. It would
then be nearly sufficient, in deciding between the two remaining views,
to show that the Endoprocta are the more primitive group of Bryozoa ;
but in addition to this, I shall offer positive evidence of derivation of the
Bryozoa from the lower worms.

The chief argument for the diphyletic origin of Bryozoa rests on
these three important differences between Ectoprocta and Endoprocta :
(1) that of the tentacular corona, which includes within it the anus in
one case, and leaves it outside in the other ; (2) that of the body cavity,
which is absent in one case and present in the other; and (3) that of
the kidney, which is a pronephridium in Endoprocta and (it is alleged)
a metanephridium in Ectoprocta.

I have, in an earlier paper (91*, p. 103), shown that the difference in
relations of the anus to tentacular corona is completely and satisfactorily
explained by the study of the development of the polypide, in which the
closure of the tentacular corona between mouth and anus is effected only
at a relatively late stage.

Concerning the second of these differences, Ehlers (790, pp. 152, 154)
has already well argued that it cannot be so fundamental, since other un-
questionably closely allied groups (e. g. Hirudinea and Cheetopoda) differ
similarly. Moreover, the difference between the “ body cavity ” of Gym-
noliemata and Endoprocta is one of degree, not of kind, for in both cases
we have to do with parenchymatous tissue more or less completely filling
the primary body cavity. The existence of spaces in the midst of the
parenchyme of Gymnoliemata may be accounted for (following Harmer,
'85, p. 64, sce also Lang, ’88, p. 77) on the physiological ground of the
necessity of a space into which the polypides can retract. In Phylac-
toleomata this parenchyme has become, in part, a very definite “ coolomic
epithelium,” although, as I have pointed out (90, p. 128), showing
traces of its parenchymatous origin.

Upon the alleged differences in the kidney, no argument can be based,
simply because the existence of an excretory tubule in Ectoprocta is
very uncertain, being at present not even probable.

VOL, XXIV. — NO. l. 3

92 BULLETIN OF THE

A fourth difference, which I do not remember to have seen mentioned,
concerns the position of the sexual glands. In Endoprocta they arise on
the polypide and are connected with the atrium by a special duct. In
Ectoprocta they arise on some part of the body wall. In all cases, how-
ever, they arise from the mesenchymatous tissue, or its equivalent, the
colomic epithelium. The exact position is not significant, for even
within the group of Phylactolemata we find the testis arising sometimes
on the funiculus (Plumatella), sometimes on the body wall (Cristatella),
and in neither case at the same place with the ova.

On the other hand, the resemblances between Ectoprocta and Endo-
procta are striking: in both a curved alimentary tract, with tentacles
of similar histological structure ; in both, an atrium originating in the
same manner; the central nervous system in both alike in position,
form, structure, and development ; in both the polypide originating in
comparable ways, involving the same problem of the relation of the germ
layers to the organs of the bud. Moreover, there must remain unmen-
tioned many minor resemblances which individually are not very signifi-
cant, but which, occurring together, furnish a most powerful argument to
the mind of one studying the animals themselves.

Hatschek’s view seems to me, therefore, untenable, and it remains to
inquire whether the Bryozoa have been derived, accompanied by degen-
eration, from an Annelidan ancestry, through Gephyrean-like forms,’ or
whether they represent a persisting low type. As we have seen, the
first hypothesis must be relinquished, if it can be shown that the Endo-
procta are the more ancestral type.

In any group of sessile, colonial animals, we should expect the more
ancestral type to retain more distinctly its individuality, to possess in a
more marked degree features belonging to a free, non-colonial life, and
in a less marked degree those belonging to a sessile, colonial one. ‘Thus
in its development the group will lose certain characters and gain cer-
tain others? A well marked individuality accompanied by a simple

1 As is well known, Hatschek (’83, p. 69) has long maintained that Phoronis and
the Sipunculacew have not been derived from Annelids as the Echiurida have, but
from unsegmented ancestors; and this view has become very widely accepted.
Ehlers, however, seems to adhere to the older view. Whatever the truth may be
in regard to this matter, the validity of my argument based on embryology and
comparative anatomy concerning the absence of close relationship between Pho-
ronis and Bryozoa is not affected.

2 Lang ('88) has shown, by an instructive analysis, that in sessile animals loco-
motor and sense organs, the nervous system, and the musculature tend to degener-
ate, and that the tentacles and protective coverings become more important.

9

MUSEUM OF COMPARATIVE ZOOLOGY. 33

method of budding (Loxosoma); a relatively poorly developed, incom-
pletely retractile lophophore; a complicated system of sense organs
and nerves (Loxosoma); sexual and excretory ducts; a typical larval
(trochophore) form, — these distinguish the Endoprocta. On the other
hand, the Ectoprocta are marked by a loss of individuality (existence
of coonoceel, Phylactolemata), by a highly complicated lophophore pro-
vided with means for complete retraction, by absence of a complicated
nervous system (small ganglion of Gymnolemata), by absence (2) of sex-
ual and excretory ducts, and by abbreviated larval life (passed within
the body of the mother).

Stronger than this argument is the fact that in the development of
the tentacular corona and of the alimentary tract — at first without a
cacum — Ectoprocta pass through stages more nearly resembling the
adult Endoprocta condition than their own adult condition does.

These facts seem to me to prove, if morphological principles can be
relied upon, that Endoprocta are nearer the ancestral form of Bryozoa
than Ectoprocta.

Admitting that the Endoprocta are more ancestral than the Ecto-
procta, I cannot conceive how any one can maintain a close relationship
with Phoronis. For the line connecting mouth and anus is in Endo-
procta ventral, while the corresponding line in Phoronis is dorsal, as
Caldwell (83, p. 372) has shown, and the kidney is a metanephridium.
These facts far outweigh, in my opinion, similarities in tentacular corona,
epistome, and bent alimentary tract.

The absence of a true body cavity, and the existence of a water or
excretory system ending in flame cells, point conclusively to an origin of
Bryozoa from the lowest worms. For such an excretory system is found
elsewhere only in Platyhelminthes, Rotifera, and in a modified form
in Nemertines (Bürger, 91). On the other hand, the existence in the
stalk of epithelial (in addition to mesenchymatous) muscles looks like
an advance beyond Rotifera and Platyhelminthes. But it does not fol-
low that such muscles existed in the ancestors of Endoprocta; they
may have been produced by causes similar to those by virtue of which
they occur in Nematodes.

Hatschek (77, p. 528) suggested, and Harmer (85, p. 11, 35) has
since shown, that the ganglion of the Endoprocta is to be regarded as a
suboesophageal ganglion. Zelinka’s (91, p. 337) discovery of a subo-
sophageal ganglion in Rotifers is interesting in this connection, as mak-
ing more probable the assumption necessary for the preceding view, that
the ancestor of Rotifers and Endoprocta possessed such an organ.

BULLETIN OF THE

One cannot refrain from noticing the similarity in the relations of the
“under lip” of Rotifers (Zelinka, 91, Taf. III. Fig. 55) and the epistome
of Bryozoa (cf. also the foot of Mollusks).

Zelinka has also shown (p. 397) that in Callidina at an early stage
two lateral folds appear on the ventral side of tho embryo, enclosing be-
tween them the mouth and under lip. These folds extend along about
one half of the length of the embryo. Later they become relatively
shorter, and finally form the lobes of the wheel organ. They have also
precisely the position of the lophophoric ridges of Ectoprocta at an early
stage, as I have figured it "91, Fig. 25), lying on each side of the mouth
and ganglion. "They have also the same relations as the ridges from which
the gill filaments arise in Lamellibranchs. Lankester’s (74, p. 80) view
of the homology of the tentacles of Bryozoa and gill filaments of Lamelli-
branchs is thus strengthened.

Finally, there is between Rotifers and Endoprocta a striking similarity
in the position of anus and urogenital duets, which in Urnatella, as in
Rotifers, open into a common cloaca on the subosophageal aspect, be-
tween mouth and anus. Such a resemblance is especially striking in
footless genera like Asplanchna (which, however, has no anus, cf. Ma-
sius, "91, fig. 1) and Hertwigia (L. Plate, '85, Fig. 7), — genera resembling
more nearly the ancestral form, since the foot, lying behind the anus,
must be considered as a secondarily produced appendage.

To sum up: The embryological as well as the anatomical evidence
seems to sustain the view that Bryozoa are closely related to Rotifers,
the two groups having sprung from an ancestor which was common to
them and Mollusks also ; that after the Rotifer stem had branched off,
the common Mollusco-Bryozoan stem produced tentacles on the lateral
ridges; that the two groups then soon separated, the Mollusca to un-
dergo an extensive and complicated development, the Bryozoa to remain
at a low level. The chief changes which the Bryozoa have experienced
are (1) the acquiring of a body-cavity through the relative decrease in
amount of the mesenchyme, that which remains forming an epithelium
(Phylactolemata); (2) the loss (2) of the protonephridia and sexual
ducts in Ectoprocta; (3) the loss of the epistome (Gymnolemata) ;
(4) the loss of the preoral ganglion ; (5) the acquiring of a cocum (Be-
toprocta) ; (6) the multiplication of methods of reproduction, by regen-
eration, by budding (without and with stock-formation), by division of
stocks, and by statoblasts.

CAMBRIDGE, September 27, 1892.

c

MUSEUM OF COMPARATIVE ZOOLOGY. 9

SUMMARY.

The segmented stem of Urnatella consists of a two-layered cuticula ;
an outer epithelium (ectoderm) consisting of cells flattened except at
the transverse septa and at the distal end of the stalk; and an axial
portion consisting of elongated cells, many of which are vacuolated, and
surrounding which there is no intercellular substance. (Pages 4, 5.)

The musculature of the stalk consists of radial sheets of fibrille, sev-
eral of which develop in a single cell. (Pages 5, 6.)

Many of the vacuolated cells of the stalk end in flame cells like the
water or excretory tubules of Platyhelminthes. (Pages 6, 7.)

Yolk is developed in the cells at the base of the stalk, first as fine
intercellular granules, which later fuse, this process being accompanied
by cell degeneration. (Pages 7, 8.)

The lip of the atrium contains a sphincter, and resembles in its rela-
tions the “margined thickening” of Ectoprocta. (Pages 9, 10.)

The epithelium of the tentacles encloses a parenchymatous core. A
pair of muscles is present. (Page 15.)

The alimentary tract resembles that of the Pedicellinide, except that
the lower wall of the rectum and the upper wall of the intestine are in
close contact. (Pages 11,212)

The nephridial tubules end blindly in flame cells, and open into a
cloaca, into which open also anus and the vas deferens. (Pages 14,15.)

Two kinds of buds arise from the Urnatella stalk; “ branches,” which
are typically median, and “stolons,” typically lateral. (Page 16.)

The segmentation of the stalk is probably an adaptation to the pro-
cess of budding, which is accompanied by a greater liability of the wall
of the stock to rupture, and therefore by a greater need of separation of
the stalk into compartments. (Page 20.)

In all Endoprocta the oral aspect of the buds is turned towards the
centre of proliferation, and in all Bryozoa the aspect in which that end of
the alimentary tract which arises from the principal outpocketing of the
atrium lies is turned towards the gemmiparous zone. (Pages 22, 23.)

The youngest stocks found consisted of a stolon bearing two indi-
viduals. This has been derived from the “stolons” of the parent stalk,
which habitually become free for the purpose of founding new stocks.
(Page 26.)

Urnatella is, structurally, one of the Pedicellinide and most nearly
resembles Arthropodaria Benedeni. (Page 30.)

The Bryozoa have probably been derived from Rotifer-like ancestors.
(Page 34.)

BULLETIN OF THE

LITERATURE CITED.

Allman, G. J.
56. A Monograph of the Fresh-water Polyzoa. viii + 119 pp., 11 Pls.
London: Ray Society.
Biirger, O.
'91. Die Enden des exkretorischen Apparates bei den Nemertinen. Zeitschr.
f. wiss. Zool., LIII. p. 322.
Caldwell, W. H.
^83. Preliminary Note on the Structure, Development, and Affinities of
Phoronis. Proc. Roy. Soc. Lond., XXXIV. p. 371.
185. Blastopore, Mesoderm, and Metameric Segmentation. Quart. Jour.
Mier. Sci., XXV. p. 15.
Davenport, C. B.
'90. Cristatella: the Origin and Development of the Individual in the Col-
ony. Bull. Mus. Comp. Zoöl. at Harvard College, XX. No. 4, p. 101.
’91,. Observations on Budding in Paludicella and some other Bryozoa. Bull.
Mus. Comp. Zool. at Harvard College, XXII. No. 1, p. 1.
Ehlers, E.
'90. Zur Kenntniss der Pedicellineen. Abhandl. kónigl. Gesellsch, d. Wiss.
zu Gottingen, XXXVI. 200 pp., 5 Taf.
Fettinger, A.
'87. Surl'anatomie des Pédicellines de la cóte d'Ostende. Arch. de Biol.,
VIL m 200;
Harmer, S. F.
'85. On the Structure and Development of Loxosoma. Quart. Jour. Mier.
Sci. XXV. p. 261.
Hatschek, B.
"T7. Embryonalentwicklung und Knospung der Pedicellina echinata. Zeit-
schr. f. wiss. Zool., XXIX. p. 502.
'83. Ueber Entwicklung von Sipunculus nudus. Arbeit. a. d. zool. Inst.
Wien, V. p. 61.
Kraepelin, K.
'92. Die deutschen Süsswasser-Bryozoen. II. Entwickelungsgeschicht-
licher Teil. Abhandl. der naturwiss. Verein in Hamburg, XII. 67 pp»,
5 Taf.
Lang, A.
"88. Ueber den Einfluss der festsitzenden Lebensweise auf die Thiere.
Jena, G. Fischer.

MUSEUM OF COMPARATIVE ZOOLOGY. 37

Lankester, E. R.
»74. Remarks on the Affinities of Rhabdopleura. Quart. Jour. Mier. Sci.,

XIV. p. 77.

Leidy, J.
^51. On some American Polyzoa. Proc. Acad. Nat. Sei. Philad., V. p. 320.

54. On Urnatella gracilis and a new Species of Plumatella. Proc. Acad.
Nat. Sci. Philad., VII. p. 191.
"70. [Note on Urnatella.] Proc. Acad. Nat. Sci. Philad. for 1870, p. 100.
^84. Urnatella gracilis, a Fresh-water Polyzoan. Jour. Acad. Nat. Sci.
Philad., (2.), IX. p. 5.
Masius, J.
91. Contribution à l'étude des Rotateurs. Arch. de Biol., X. p. 651.
Nitsche, H.
69. Beiträge zur Kenntniss der Bryozoen. TI. Ueber die Anatomie von
Pedicellina echinata. Zeitschr. f. wiss. Zool., XX p. 1.
Plate, L. H.
785. Beiträge zur Naturgeschichte der Rotatorien. Jena. Zeitschr, XIX.
mas
Seeliger, O.
'89», Die ungeschlechtliche Vermehrung der endoprokten Bryozoen. Zeit-
schr. f. wiss. Zool., XLIX. p. 168.
^90. Bemerkungen zur Knospenentwicklung der Bryozoen. "Zeitschr. f.
wiss. Zool, L. p. 560.
Stuhlmann, F.
'87. Zur Kenntnis des Ovariums der Aalmutter. Abhandl. naturwiss.
Verein in Hamburg, X. 48 pp.
Vigelius, W. J.
^g4*, Ueber Barentsia bulbosa Hincks. Bijdragen en tot de Dierkunde,
XI. p. 85.
Zelinka, C.
'91. Studien über Räderthiere. III. Zur Entwicklungsgeschichte der Rä-
derthiere nebst Bemerkungen über ihre Anatomie und Biologie. Zeit-
schr, f. wiss. Zool., LIIL. p. 1.

BULLETIN

OF THE

ANALYTICAL INDEX

TO DAVENPORT ’90, '91%, AND THE PRESENT PAPER.

The Roman numerals refer to these three papers in the order of publication, The Arabic numerals
refer to pages. Topics are arranged alphabetically, except under ““Groups,” where the main di-
visions are arranged according to an adopted zoülogical system. Under ** Organography ” letters are
placed after certain page numbers to designate the standpoint from which the organ is treated. The
letters apply to all the page nnmbers between them and the next preceding letter, These letters
have the following significations: A, Anatomy; F, Function; H, Histology; O, Ontogeny ;

P, Phylogeny.

AUTHORS,

Agassiz, A., ii. 2, 47, 101.

Agassiz, L., ii. 83.

Allman, i. 115, 121, 123, 126, 130, 189;
ii. 18, 30, 82, 84, 70; iii. 1.

Altmann, if, 99,

Barrois, J., i. 124, 127; ii, 49, 60, 54, 58-
61, 67, 89, 90, 92, 96.

Beneden, P. J. van, ii. 25, 80, 82, 30, 61.

Braem, i. 102-104, 107, 110-112, 115-117,
128, 125, 127, 129, 130, 142, 144; ii. 12-
14, 16, 17, 22, 29, 30, 89, 62, 67, 68, 71-
75, 81, 87-89, 94-96, 98, 101; iii. 8,

Bürger, iii. 33.

Bütschli, ii, 99; iii. 8.

Caldwell, i. 138; iii. 28, 88.

Chun, 11. 72,.78,.82,

Claparede, ii. 34.

Cori, ii. 108.

Dumortier, ii. 25, 30, 32, 39, 61.

Ehlers, i. 123; ii. 34, 60, 78, 78, 79, 102,
108; tits 2, 4, 0,50, 11:18, 10; 19, 21,

30-32.

Eisig, ii. 29, 86, 37.
Ellis, ii. 54.

Farre, ii. 54, 61.
Fottinger, iii. 9, 15, 19.
Fraisse, ii. 71.

Freese, ii. 29, 94, 60.
Goette, ii. 98, 101.

1 See “ Literature Cited,” p. 36.

Grant, ii. 54.

Haddon, i. 102, 124, 126, 127; ii. 60, 64-
66, 89.

Harmer, i. 123, 127, 128; ii. 50, 51, 90, 91,
104; 111, 11, 15, 81, 88.

Hatschek, i. 102, 107, 122, 128, 127, 128,
132,134; ii. 54, 71, 104 ; iii. 24, 30, 82, 33.

Heider, ii. 82.

Hincks, ii. 41, 46, 49, 59, 77, 79.

Hoyer, ii. 83.

Hyatt, i. 189.

Joliet, i. 126, 127; ii. 34, 54.

Jullien, ii. 67, 70, 89.

Jussieu, ii. 54.

Kennel, ii. 84.

Kent, ii. 75.

Kleinenberg, ii. 95.

Korotneff, i. 121, 123, 124, 126, 132, 139;
ii. 82-84, 61, 67, 89, 94.

Korschelt, i. 125; ii. 82, 99.

Kraepelin, i. 103, 104, 115, 122, 124, 129,
130, 134, 139, 141, 144, 145; ii. 25, 26,
32, 35, 75-78, 89, 103; iii. 28.

Kükenthal, ii. 36.

Lankester, i. 138; iii. 84.

Lang, iii. 31, 82.

Leidy, ii. 41, 83; iii. 1, 2, 4, 10, 13, 16, 19,
25-27, 29.

Maas, ii. 101.

Mark, i. 101; ii. 2, 16; iii. 8.

Masius, iii. 84.

MUSEUM OF COMPARATIVE ZOOLOGY. 39

Metschnikoff, i, 103, 113, 124, 125, wi

ii. 36, 90.

Nitsche, i. 102, 103, 100-117, 121-120, 128-
136, 139; ii. 13, 22, 26, 32, 34, 52, 53,
55, 59-63, 88, 102, 103; iii. 10, 11.

Nussbaum, ii. 71, 72, 98, 100, 101.

Oka, ii. 88, 104.

Ostroumoff, i. 126; ii. 34, 52, 58, 60, 64,
66, 67, 90, 92.

Parker, ii. 29, 87.

Pergens, i. 126; ii. 60.

Plate, iii. 94.

Potts, iii. 2, 3, 29.

Prouho, ii. 55, 57-61, 67, 88, 95.

Reichert, i. 122; ii. 18, 93.

Reinhard, i. 102, 106, 111, 124; ii. 89.

Repiachoff, i. 136; ii. 60, 90, 91, 97.

Roux, li. 71.

Saefftigen, i. 128.

Seeliger, i. 111, 122, 123, 126, 127; ii. 58,
55-61, 88, 96, 100-102, 104; iii. 22, 23,
25, 28.

Semper, ii. 72, 84.

Smitt, i. 121; ii. 34, 50, 62, 72.

Stuhlmann, ii. 99; iii. 8.

Tullberg, ii. 16, 29.

Verrill, ii. 46, 48, 49, 83.

Verworn, i. 130, 140.

Vigelius, i. 123, 126, 128, 136; ii. 34, 53,
60, 62, 67, 90-92; iii. 20.

Wagner, Y. v., ii. 87, 100.

Weismann, ii. 72, 83.

Wilson, H. V., ii. 2, 100

Zacharias, ii. 87.

Zelinka, iii. 33, 34.

Zoja, ii. 87.

BIOLOGY.

Continuity of germ plasma, ii. 12.

Fresh water, effects of, on larval stages,
ii. 89.

Sessility, effects of, on Bryozoa, iii. 32, 33.

Sex, iii. 15.

BUDDING, GENERAL.

Architecture of the stock, i. 103-106, 121,
1225115. 2,9; 41-52, 84-86; iii. 16-18.

quineunx arrangement of individuals,
i. 104; ii. 85.
spiral arrangement, ii. 80.
Cause of cell proliferation in B., ii. 102.
Distribution of embryonie tissue in stock,
11.72, 74; 41, 17, 18,
Dependence of B. upon
economy of space, ii. 86.
embryonic tissue, ii. 72, 74, 97, 100-102.
food supply, i1. 86, 87.
Formation of new stocks, iii. 24-27.
Germ layers, homology of,in B 1.125 127.
Knospungszone, ii. 84.
Lateral B., i. 121.
Laws of B., i. 105, 106; ii. 8, 42-52, 71-
87; iii. 16, 18, 21.
deviations from law, i. 105; ii. 3, 43,
51, 62; 111, 18.
Margin of Corm, i. 104, 105, 118, 119; ii.
4, 53, 55.
Orientation of Individuals, i. 103, 104,
122; ii. 8, 8, 82; iii. 16, 17, 22.
with reference to proliferating centre,
1.122; ii. 3, 82; iii. 22.
“Samknopp " of Smitt, i. 121; ii, 52.
Stolon, i. 121; ii. 70, 71; iii. 27, 30.
Tail end, ii. 84.
Time and place of B., i. 104-106, 108; ii
7, 9, 56, 86; iii. 28.
Tip of stock, ii. 4, 5, 14, 72, 19, 86 ; iii. 17.

BUDDING, SPECIAL.
Annelida, ii. 84, 85.
Bryozoa,
Dichotomy absent in, i. 105; ii. 42, 48,
80, 81, 86.
Homology of branch and stolon in
Urnatella, iii. 17.
Polypide, origin of, i. 106-111, 121-127 ;
ii. 4, 7, 18, 27, 54-56, 72-81; iii. 28.
—— layers of, i. 107-110, 122-125 ; ii.
54, 69, 70, 95, 96; iii. 23.
—— primary, ii. 69, 70, 89, 90.
Unity of development of polypide
throughout B., iii. 28, 24.
Hydroidea, ii. 82, 83.

DEGENERATION, li. 64.

40

DISTRIBUTION.
Cristatella, i. 102; ii. 2.
Fredericella, i. 102.
Gymnolemata, ii. 2,
Paludicella, ii, 2,
Plumatella, i. 102; ii. 2,
Urnatella, iii. 2.

Groups.

Protozoa, ii. 86.

Ciliata, ii. 99; iii. 4.
Gregarinida, ii. 99.

Porifera, ii. 100.

Cnidaria, i. 125; ii. 94; iii. 23.
Hydra, i. 121; ii. 87, 89.
Hydroidea, iii. 21.

Pennaria, ii. 88.

Halistemma, ii. 82, 88, 85.
Triploblastica, iii. 28.
Platyhelminthes, iii. 33.

Turbellaria, i. 121; ii. 87, 89.

Cestoda, iii. 18,

Nemertini, iij. 33.

Nematoda, iii. 6, 33.
Coelomyaria, iii. 6.

Trochophora, iii. 15.

Rotifera, ii. 89; iii. 80, 33, 84.
Asplanchna, iii. 84.
Callidina, iii. 34.
Hertwigia, iii. 34.

Bryozoa, i. 120-123, 139; ii. 38, 41, 54, 61,
66, 71, 72, 86-88, 93, 94, 98, 100, 105;

iii. 1, 12, 16-18, 24, 30-34.
Rhabdopleura, i. 138; iii. 11.

Endoprocta, i. 110, 126, 127, 132-134;
ii. 19, 61, 102-104; iii. 7, 8, 11, 18,

15, 17, 19, 22-24, 28-34,

Loxosoma, i. 126, 127, 189; iii. 9, 11,

15, 20, 22, 25, 33.

Pedicellinide, iii, 2, 9, 11, 15, 20,

27, 30.

Arthropodaria. See P. Benedeni.
Ascopodaria, ii. 102; iii. 6, 9-11,

13, 15, 19, 21.
Barentsia, iii. 20.
Gonopodaria, iii. 19, 20.

Pedicellina, i. 121-123, 127, 128,
182, 139; ii. 105; iii. 22, 25.

BULLETIN OF THE

Bryozoa, Pedicellina, — continued,

P
P
E
P
Urn

. Americana, iii. 19.

. Belgica, iii. 19, 20.

. Benedeni, iii. 9, 15, 19, 20, 30.

. echinata, iii, 10, 11, 19.

atella, iii. 1-16, 19-22, 24, 26, 27,
29, 30, 34.

Ectoprocta, i. 102-104, 127, 132, 138;

ii. 61, 90; iii^ 9, 17, 21-24, 28,
30-34.

Gymnolemata, i. 110, 126-128, 131-

134; ii. 4, 40, 54, 55, 60, 67, 81,
82, 89, 97, 98; iii. 9, 22, 31.

Cheilostomata, i. 121, 123, 126, 136;

Ct

ii. 52, 53, 56, 60, 61, 03, 64.

Bugula, ii. 34, 41, 46, 48-51, 55,
-60, 81, 86, 87, 90-92, 94.

Escharella, ii. 52, 54, 59, 60, 81.

Kucratea, ii. 64.

Flustra, i. 127, 136, 189; ii. 34,

3, 58-60, 64.

Lepralia, i. 127; ii. 34, 47-50, 52-
54, 56, 58, 60, 86.

Membranipora, ii. 29, 34, 59, 60,
64, 90, 91, 96.

Serupocellaria, ii. 34.

Tendra, ii. 60.

enostomata, i. 123; ii. 32, 33, 60,
63, 75, 108,

Alcyonididse, ii. 80.
Alcyonidium, i. 126; ii. 90-98.
Bowerbankia, ii. 34, 40, 60-63,

80, 97.

Cylindrecium, ii. 78.
Farella, ii. 61.
Flustrella, i. 121; ii. 28, 49, 53,

55, 56, 58-60, 64, 91-93,

Hypophorella, i. 121; ii. 84, 60,

78, 79.

Laguncula, ii, 61.
Mimosella, ii. 78.
Paludicella, i. 121, 123, 126, 132,

189; ii. 2, 18, 19, 22, 26, 57-
64, 72, 78, 75, 76; iii. 20, 28.

Stolonifera, ii. 78, 79.
Vesicularia, ii. 79.
Victorella, ii. 75, 76, 78, 79.
Zoobotryon, i. 122,

MUSEUM OF COMPARATIVE ZOOLOGY. 41

Bryozoa, Gymnolemata, — continued.
Cyclostomata, ii. 50, 81, 90, 92, 98.
Crisia, ii. 50, 51, 61, 81, 86, 87.
Tubuliporida, ii. 81.
Phylactolemata, i. 102, 127, 128, 132-
185, 189; ii. 19, 20, 22, 24, 20, 30,
54, 57-68, 67; iii, 8-11, 22, 29,
31-34.
Aleyonella, i. 102, 103, 107, 110,
122-124, 135, 144, 145.
Cristatella, i. 108, 104, 107, 111,
122-129, 131, 132, 184; ii. 2, 22,
62, 67-70, 74, 75, 87, 90, 98;
iii. 32,
Fredericella, i. 131, 134.
Lophopus, ii. 61.
Pectinatella, i. 134; ii. 102, 104.
Plumatella, i. 110, 181; ii. 2, 30,
67-70, 73-76; iii. 17, 32.
Mollusca, i. 188; ii. 16, 29, 89.
Lamellibranchiata, iii. 34.
Annelida, i. 138 ; ii. 29; iii. 82.
Chetopoda, i. 121; ii. 84, 89; iii. 31.
Autolytus, ii. 85.
Chetogaster, ii. 84.
Ctenodrilus, ii. 84.
Nais, ii. 84.
Hirudinea, ii. 38, 89; iii. 31.
Gephyrea, iii. 80-82.
Sipunculida, iii. 24, 32.
Phoronide, i. 138; iii. 81-83.
Sagitta, i. 124.
Arthropoda, ii. 38.
Crustacea, ii. 29.
Astacus, ii. 89.
Tunicata, 1. 123; ii. 96, 98.
Pyrosoma, ii. 97.
Salpa, i. 121.
Vertebrata,
'Teleostei, iii. 8.
Amphibia, ii. 98.

Hisroroov.
Cell,
basement membrane, iii. 11.
cell wall, perforated, ii. 03.
cilia, iii. 12.
gelatinous bodies, i. 144, 145.

Cell, — continued.
granules, ii. 10; iii. 12, 13.
staining properties, ii. 10; iii. 12.
nucleus,
deeply staining, i. 136.
large, ii. 5, 6, 9, 10, 69, 98: iii. 6, 28
Stübchen, iii. 12.
yolk, ii. 34, 36, 87; iii. 7, 8, 29.
Cell sorts,
amoeboid, ii. 7, 38, 116, 117, 142.
chlorogogen [of Annelids], ii. 36.
digesting, i. 141.
elongated, iii. 5.
embryonic, i. 109; 5, 6, 10, 11, 65, 69,
72, 98-102; iii. 24, 28.
— yolk within, ii. 98.
flame, iii. 4, 6, 18, 29.
cirrus, iii. 4, 6.
gemmiparous. See embryonic.
granular, ii, 99; iii. 12, 29.
hepatic, i. 141; iii. 11, 12.
mesenchymatous, i. 126; ii. 10, 28-80;
iii. 10, 12, 18, 29.
muscle, i. 143; ii. 28; iii. 5, 6.
attachment, i. 142, 148; ii. 28, 29;
iii. 18.
cross striped, i. 141; ii. 91.
fibrille, i. 143; ii. 28, 30; iii. 5, 6.
migration of point of attachment, i.
117, 141-148; ii. 14, 16, 28, 29.
nerve, ii. 62.
phagocyte, ii. 36.
reticulated, ii. 6.
secreting, iii. 25.
vacuolated, iii. 5, 7, 8, 12, 18, 26.
Cuticula, ii. 5-7, 11, 14, 15, 18, 29, 31;
iii, 4, 9.
Tissues,
epithelium,
ciliated, iii. 10-12.
columnar, ii. 8, 65.
free surface of E., iii. 12,
pavement, ii. 10
derivation from columnar, ii. 65,
68, 101.
transition to mesenchyme, ii. 54.
variations in thickness, ii. 8, 15; iii.
4, 5, 12.

Hıstorvsıs.
Cellular, i. 181, 144, 145; iii. 5, 7, 8.

MORPHOLOGY, GENERAL.

Bilateral symmetry, i. 111; ii. 9.

Change of form, ii. 24.

Evagination, ii. 10, 11.

Form of body, ii. 2.

Germ layers, relation of, to bud layers,
ii. 88.

Invagination, ii. 8, 9, 102.

Region of growth, ii. 9.

Rotation of axes, ii. 24, 25, 62, 63.

Unequal growth, ii. 25.

MORPHOLOGY, SPECIAL.

Bryozoa, aspects of, i. 102, 103; iii. 33.

ONTOGENY, GENERAL.

Blastopore, ii. 104, 105.
in buds, iii. 23.
concrescence of lips of, in bud, iii

23.

Blastulation, ii. 68.

Body cavity, i. 124.

Cleavage, i. 124; ii. 68.

Entoderm, ii. 89-95.
loss of, through loss of function, ii.

93-95.
Gastrulation, i. 124, 125; ii. 90-92; ii.
23.

Mesenchyme, i. 125.

Mesoderm, i. 125; ii. 68, 91-94.

Oógenesis, ii. 69, 97.

Schadonology, General,

Larval life modified by
fresh water, ii. 89.
intra-uterine life, ii. 89; iii. 33.
—— length of, ii. 98, 94.
Schadonology, Special,
Bryozoa, ii. 89-95.
Cyphonautes, ii. 91-94.
Comparison of Endoproct and Ecto-
proct larvz, ii. 105.
Spermatogenesis, iii. 15.

42 BULLETIN OF THE

ONTOGENY, SPÉCIAL.
Bryozoa,
Relation of blastopore to region of ori-
gin of first polypide, ii. 92.
Relation of blastopore to embryonic
tissue, ii. 97, 105.

ORGANOGRAPHY.
Bryozoa,

Alimentary tract, i. 111-118, 1249, 127°,
139-1410; ii. 19, 57, 58, 62, 640, 1040;
iii. 11-144, 280,

— of larva, ii. 90, 91°.

Anus, i. 1400; ii. 20, 21, 24, 57, 58,
62, 630; iii. 11-134, 30P,

Atrial fold, iii. 114.

Atrial groove, iii. 10, 114F

Atrial opening, i. 1440; ii. 31, 320;
iii. 940,

Atrium, i. 110, 1120; ji. 18-21, 57, 58,
1080; iii. 10, 114, 230,

Avicularia, ii. 64.

Basal plate, iii. 164, 25, 27HFP,
homology with stolon, iii. 27.

Branch, ii. 9, 110; iii. 16, 174, 24H, 27F,

Body cavity, iii. 134, 31P,

Body wall, i. 117-120, 122-1240; ii, 12-
18, 530; iii. 9H,

Brown body, ii. 64-660.

Cardiac valve, i. 1400; ii. 199.

Calyx, iii. 9.

Cirrus. See Tentacle.

Cloaca, iii. 11, 14, 15, 294.

Closing cell of diaphragm, iii, 8H,

Coecum, i. 140, 1410; ii. 19, 20, 639;
iii. 38°,

Coelomic epithelium. See Mesoderm.

Conoceel, i. 103; iii. 83°.

Collare setosum, ii. 320, 103P.

Couronne (of larva), ii. 90.

Communication plate, ii. 32, 330, 38-
40, 76, 78". See also Septum.

Diphragma, ii. 32, 63, 1030P ; iii. 9P.

Ectoderm, ii. 979; iii. 4, 5, 9H,

Endosare, i. 120, 1280P,

Entoderm, ii. 89-95, 970,

Epistome, i. 138, 1390; ii. 610; iii. 114P,
34P,

MUSEUM OF COMPARATIVE ZOOLOGY. 43

Bryozoa, —continued.
Epistomic canal, i. 115, 1390,
Excretory organs, i. 130; ii. 64; iii. 14,
154P, 33P,
efferent duct, iii. 144.
tubule, iii. 6, 7HO, 30P,
intracellular lumen, iii. 7.
Foot gland, iii. 25.
Funiculus, i. 109, 115, 116, 141, 142°;
ii. 300P; iii. 32.
migration, i. 116.
Gabelkanal, i. 130P,
Gemmiparous zone, ii. 4, 6°.
Gizzard, ii. 63H,
Incubatory chamber, iii. 154.
Internal sac (larva), ii. 90, 93.
Intestine, iii. 11, 124H,
Kamptoderm, i. 1150; ii. 2, 21, 22, 25,
31, 56, 57, 59-620; iii. 9AP,
Lip of atrium, iii. 9, 104,
Lophophore, i. 132-134? ; iii. 33F.
Lophophoric arms, i. 1140.
fusion of, i. 130, 131, 134.
Lophophorie ridges, ii. 20-230; iii. 384P.
Mantle cavity (larval), ii. 90.
Mantle fold (larval), ii. 90.
Marginal thickening, i. 1449; ii, 32,
103P; iii. 9, 10P.
Mesoderm, ii. 5-7, 10, 54, 56, 899.
Metanephros, iii. 31, 83°.
Musculature, i. 116; ii. 630,
anal sphincter, ii. 32; iii. 11, 184.
atrial sphincter, iii. 13AF,
intestino-rectal sphincter, iii. 11,
1240,
lateral-wall, iii. 18, 144,
parietal, ii. 28-800,
parieto-vaginal, i. 143, 1449; ii. 2, 13,
28, 630,
pyramidalis. See parieto-vaginal.
retractor, i. 117, 141-1430; ii. 13,
28, 630,
rotator, i. 117, 141-1439.
stalk, iii. 5, 6HO, 190, 254,
transverse, iii. 13, 14^.
Neck of polypide, i. 121, 1449; ii. 18,
319, 614, 63, 650,
Neck of ocecium, ii. 69, 700,

Bryozoa, — continued.
Nephridium, i. 130; ii. 103; iii. 14,
1540,
Nervous system, ii. 26, 610.
circumaesophageal nerve, i. 136, 1380;
ii, 26, 620,
ganglion, i. 113-1150, 127, 128P; ii.
24, 26, 55, 61, 639, 103P; iii. 15,
164H, 24, 32, 33".
gastric nerve, ii. 26, 620,
Esophagus, i. 112, 1400; ii. 19, 20,
264, 58, 62, 689; iii. 124H,
Ocecium, i. 124; if. 689.
Operculum, ii. 644.
Oviduct, iii. 154,
Parenchymatous tissue, i. 126; ii. 91.
Pharynx, i. 1400; ii. 21, 24, 26, 57, 630,
Pronephridium, iii. 31, 34°.
Pyloric valve, i. 1400; iii. 124,
Pyriform organ (larval), ii. 90, 93.
Randwulst. See Marginal thickening.
Rectum, ii. 230; iii. 12H,
Ring canal, I. 129, 130, 138-1359 ; ii.
21, 22, 24, 589.
Rosettenplatte.
plate.
Segment of stalk, iii. 1, 3, 164.
relation of formation to strobiliza-
tion, iii. 19, 21.
Septum, ii. 11, 32, 530; iii. 4, 5, 8, 9^.
intersegmental, iii. 4, 5, 8, 94H, 19,
200P,
subcalycine, ii. 44F ; iii. 8, 94H, 210,
Sexual ducts, iii. 154, 30, 32, 34°.
Sexual organs,
ovary, ii. 689.
testis, iii. 154,
Stalk, iii. 1, 8-944,
Statoblast, i. 111, 1259; ii. 34; iii. 8,
29, 34,
Stolon, ii. 75, 77-799 ; iii. 16, 174, 244,
26, 27FP,
Stomach, i. 140, 1419; ii. 19, 28, 24,
629 ; iii. 11, 1248,
Subosophageal ganglion. See Nervous
system.
Tentacle, i. 182, 135, 1360; ii. 20, 22,
58, 590, 103" ; iii. 1, 10, 814,

See Communication

44 BULLETIN OF THE
Bryozoa, Tentacle, — continued.
sequence of formation, i. 132, 135;
ii. 23, 59, 62.
number, ii. 24, 25, 59 ; iii. 1, 10.
odd tentacle, ii. 25, 26 ; iii. 10.
Yolk spherules, iii. 7, 8 .
Cnidaria,
Septum, iii. 21".
Mollusca,
Foot, iii. 34°.
Gill filament, iii. 34P,
Rotifera,
Anus, iii. 344,
Cloaca, iii. 344,
Foot, iii. 344,
Subesophageal ganglion, iii. 334,
Urogenital ducts, Hi. 844.
Wheel organ, 344.
Tunicata,
Nervous system, ii. 960,

MUSEUM

OF COMPARATIVE ZOOLOGY,

PHYSIOLOGY.
Cellular,
Cuticularization, ii. 33.
Excretion,

Bryozoa, ii. 66.

REGENERATION,

Calyces in Urnatella, iii. 2, 29.

Polypide in Gymnolemata, ii. 64, 65.

Relation to normAl polypide formation,
ii. 65.

RELATIONSHIPS.

Bryozoa and other Metazoa, iii. 31-34.

Ectoprocta and Endoprocta, i. 132, 183
ii. 102-105 ; iii. 31, 32.

Gymnolemata, ii. 75-81.

Phylactolemata, i. 104.

Urnatella, iii. 30.

,

EXPLANATION OF PLATES.

All figures were drawn with the aid of a camera lucida, from preparations of
Urnatella gracilis, unless otherwise stated.

DAVENPORT, — Urnatella gracilis.

PLATE I.

ABBREVIATIONS.
atr. Atrium. di. ba. Basal disk.
cic. Scar of fallen-off bud. di. sep. Transverse septum.
cl. fim. Flame cell. gm. Bud.
cl. mi. Migratory cell. mu. Muscle fibre.
cta. Cuticula. nl. Nucleus.
ems Calyx. pli. ere. Circular folds.
cx. rgn. Regenerating calyx. vt. gran. Yolk granules.

Regenerating stalk of Urnatella, viewed as an opaque object. X 96.

Optical section of the same. X 96.

Optical section of the distal end of a stalk of Urnatella which is probably
about to regenerate, from living animal. X 410.

Optical section of the proximal urn of the stalk of Urnatella. Showing
thick cuticula and yolk granules. x 165.

Contents of proximal urn, obtained by crushing. Moving cilia were visible
Nuclei in some of the yolk granules. » 320.

Stock of Urnatella viewed as opaque object. Three stems arising from
the basal disk. Buds and regenerating terminal calyces. X 36.

Optical section of stalk of Urnatella. Xx 96.

DAVEN PORT - U

U

ANATELLA

pli.ere, etan, ai Y j

cla.

au
ex
j air.
ee m diba.
cd flm
E eta
Du elmi
@ Vl i
\ B.
\ j )
\ |
eu
y
Uta,
ete.
» s
| ete,
= oe M eft. sep.
3 k
à { \
gne. :
> E | Pi
CUP. 1 i if
Y :
b i EGP,
di ba T j

dé bu.

DAVENPORT, — Urnatella gracilis,

Fig. 8.

PLATE II.

ABBREVIATIONS,
atr. Atrium, gn. Ganglion.
cl. film. Flame cell. in. Intestine.
cl. hp. Hepatic cells. ms'drm. Mesoderm.
cl. sec. Secreting cells. mu. Muscle fibre.
cl. tb, Tubular cells. nl. Nucleus.
cta. Cuticula. nph. Nephridium.
cta.! Outer cuticula. æ. G?sophagus.
cta.!! Inner y of. atr. Atrial opening.
di. ba. Basal disk. pa'chy. Parenchyme.
di. sep. Dissepiment. rt. Rectum.
ecdrm. | Ectoderm. vt. or

Yolk granules.

ga. Stomach. vt. gran,

Longitudinal section through proximal part of stalk of Urnatella. Third
stage in the formation of yolk granules. X 610.

Cross section through distal segment of stalk, showing ectoderm and pa-
renchymatous tissue inside. X 405.

Third stage in the development of the stalk muscles. Zeiss, jy oilimmers.,
Oc.2. x 1350.

Second stage in the development of the stalk muscles. X 1350.

Termination of excretory tubule of stalk in flame cell. The position of
the calyx is above. 305.

First stage in development of yolk in stalk. X 610.

Second stage in development of yolk in stalk. X 610.

Median section through young stalk (age of Fig. 31 in Plate V.). x 165.

Fourth stage in the development of the stalk musculature. X 1850.

First stage in the development of the stalk musculature. X 1850.

URNATELLA
n i
re ee j
PR
$a B xs
oy = z : |
, | ; ; ^ „ee ab.
j 3 : EOG
ba | |
‘ ^ | | |
A UN
Mitt í ) | .
| é Š niu ;
ec'drh |
= | ae, 9
DE
pa^ chi ete
mae
H |
etarra a
10
ee drm aes | A |
HH oiron , i :
neut leen
'2 4 :
13
ed fle T
3
vé men. i
| ge
re |
hy |
(sep
4
/¢
cla en
pac
oè
| msde
{ "
yd
16,
ase
olt da
Le |
c6 ron new $
eg. set 3 : j

«

“

DAVENPORT, — Urnatella gracilis,

PLATE III.

ABBREVIATIONS.
an. Anus. mu. Muscle fibre.
atr. Atrium. nph. Nephridium.
clc. Cloaca. æ. Œsophagus.
ol, ela, Closing cell of septum. pd. Stalk.
cl. hp. Hepatic cells. Pl. Rectum.
cta. Cuticula. spht. Sphincter between intestine
di. sep. Dissepiment. and rectum.
ga. Stomach spht. atr. Atrial sphincter.
gn. Ganglion. sul. atr. Atrial groove.
in. Intestine. ta. Tentacle.
kmp'drm. Kamptoderm. va. def. Vas deferens.

loph. atr. Atrial ridge.

Fig. 18. Median section of calyx of Urnatella. Camera drawing made from sev-

eral adjacent sections. 405.
19. Section through plane of line “Fig. 19” on Fig. 18. x 96.

Figs. 20, 20°. Longitudinal and cross sections through chief nephriaium of calyx.

x 610.

21* 21^, 21°. Three sections taken across nephridium of one individual at
different places, to show nature of the lumen and wall. Fig. 21* nearer
the distal end of tubule. Fig. 21°, third section from mouth of tubule,

at point of union of two nephridia into one. X 700.

EN à ;
s kmpdren
spht air
spht
: 1 d
| > an
| da
4 ete
in
Ñ
» >
N
k
| ene.

lophat
| chán.
Y | |
| 3 Fig. 25.
à
ge. |
ae
Ela
s
im "Y
si
rrtt
ud

DAVENPORT, — Urnatella gracilis.

PLATE IV.

ABBREVIATIONS.
an. Anus. mu. Muscle fibre.
atr. Atrium. mu. ret. ta. Retractor muscle of tentacle.
clc. Cloaca. neph. or per
cl. cls. Closing cell. nph. | — re
cl. fim. Flame cell. e. CEsophagus.
cl. tb. Tubular cell. pa'chy. Parenchyme.
cta. Cuticula. rt. Rectum.
di. sep. Dissepiment. spht. atr. Atrial sphincter.
ec'drm. Ectoderm. sp'z. Spermatozoa.
gn. Ganglion. te. Testis.
int. Intestine. va. def. Vas deferens.

Section through calyx at level indicated by line “ Fig. 22” in Fig. 18
x 405.

Section through calyx above the last at level indicated by line “ Fig 23”
in Fig. 18. X 405.

Longitudinal section through young stalk showing development of septum.
x 405.

Transverse section through closing cells at base of calyx. Section passes
through a flame cell at the left. The lower part of the section lies
below the dissepiment, the position of which is indicated by the line
di. sep. X 670.

Longitudinal section through tentacle and atrial sphincter. X 405.

Two adjacent tentacles cut transversely near their bases. X 405.

Part of nearly median section of calyx, passing through cloaca and vas
deferens. X 405.

Median section through nephridium aud adjacent organs of young calyx

to show the development of the nephridium. X 1000.

A dus dus u Re ol ob
M s " t (se
| | ; : * Des
f 74 AP yy E
e e
Ay i \ š ss Spat atr
jan i y ela.
nue ism ute deo i p E
E * ® oq + N
E |
N ‘> S ect.
x eun ES /
Mw m S 1 ie : i
j d ed Ar j d :
ER N \ oy nd» fin,
5 > OY ; Fe Xe
; dd @) e :
ae ^ Lf. S
3 3 dise. *
i ‘ og
«O
PR 25.
fe

Davenport. — Urnatella gracilis.

PLATE V.

ABBREVIATIONS.
An. Anal aspect. Or. Oral aspect.
cic. Scar of fallen bud. or. Mouth.
cl. sec. Secreting cell. pd. Stalk.
di. ba. Basal disk. rt. Rectum.
ec’drm. — Ectoderm. sto. Stolon.
mu. Muscle fibre. vt. Yolk spherules.

Fig. 30. Distal end of a richly branching stock of Urnatella, seen from the right
side. X 56.

Figs. 31-45. Views of Urnatella as solid object (exc. Fig. 33). All x 181.

* 81,32. “ Youngest stage” of Urnatella stock. Cf. Plate II. Fig. 15.

Fig. 33. View of somewhat older stage, as transparent object. Disk adhering to
stolon.

Figs. 34, 35. Young stalks with median buds.

Fig. 86. Entire stock of Urnatella, — three stalks arising from one basal disk.
“ Median buds ” arising.

* 37. Stalk with two lateral buds and one median.

* 88, Stalk with profuse lateral budding on the median buds. Anal view,
branches displaced to the right and left. From segment c' two lateral
buds and one median have arisen; from c, two lateral buds, one of
which has given rise to two individuals, the other to only one.

“ 39 Stalk with branches, — oral view. Remnant of a median branch on seg-
ment F. From c! two lateral (?) buds; from K a stolon with two
individuals, and from L two stolons, the left with two individuals,
the right with one.

Figs. 40,41. Branching stocks with median and lateral buds.

Fig. 42. Typical lateral buds, each with two individuals, A median bud on next
younger segment.

Figs. 43, 44. Right and left views of the same stock. The number of individuals
is indicated in the diagram accompanying the figures (Fig. 444).

Fig. 45. Left view of part of stock. Two stolons and one median branch from seg-
ment B (Fig. 459).

* 46. Optical section of “ youngest stage " of Urnatella stock. Stolon with two
individuals arising therefrom. X 120.

* 47. View of cuticula and ectoderm at end of a stolon in a stock of the
“youngest stage," showing scar of former attachment to parent
stock. X 405.

* 48. Optical section of the basal plate of a stock showing three individuals

arising from it, and its division into three segments. X 52.
Three individuals arising from an older basal plate. X 52.

Or
7
2
Ef.
diba
sio s: ^ id
as |
j x
de
K j
i $ b i
z a i
4 ' |
X f y {
ej i |
» m ài
i |
I A ^
D , 4 N

DAVENPORT, — Urnatella gracilis.

PLATE VI.

ABBREVIATIONS.
atr. Atrium. i. Inner layer of bud.
cl. sec. Secreting cell. mu. Muscle fibre.
em, Calyx. or. Mouth.
di. sep, OF |... A spht. atr. Atrial sphincter.
di. sp. f ete i st. Stomach,
ex. Outer layer of bud. slo. Stolon.
ga. Stomach ta. Tentacle.
gm. Bud.

Fig. 50. Surface view of stem stained in Czokor's cochineal and studied in clove oil,
Showing the deeply stained protoplasm and crowded condition of
nuclei indicating the position of an incipient bud. X 318.
“ bl. Lateral bud drawn as transparent object after staining, showing stolon
and secreting cells (cl. sec.). X 98.
Figs. 52, 53, Two optical sections from different aspects of the same bud stained
in Czokor's cochineal and studied in clove oil. Showing formation of
alimentary tract. X 225.
* 54-56. Three transverse sections across a young bud, showing the develop-
ment of the alimentary tract. Fig. 54 nearest the tip; Fig 56 nearest
the base. x 405.
Fig. 57. Enlarged view of lateral bud showing stolon and secreting cells (cl, sec.).
x 205.
* 58. Stalk of young lateral (?) bud showing method of attachment to adult
stalk. X 285.
Figs. 59, 60. Two diagrams showing relation between the two methods of budding
in Urnatella. The shaded cells indicate probable position of embry-
onic tissue.

ye

No. 2.— Note on the Carotids and the Ductus Botalli of the
Alligator. By ©. B. Davenport.

Tun carotids of the Alligator, as is well known, usually present in
the adult an unsymmetrical condition. From the right aortic root two
branches (a and b) arise close together. Of these, one goes to the left,
the other to the right side of the body. The left vessel (a) gives rise
first to a large artery (1) running on the dorsal side of the oesophagus
to the head ; secondly, to a smaller one (2) running alongside the neck,
sending branches to the oesophagus, trachea, and muscles of the body
wall, and anastomosing behind the occiput with a branch of 1; and,
thirdly, to a vessel (3) going to the fore limbs. Other smaller vessels
we may for present purposes neglect. The right vessel (b) gives rise to
all that the left does excepting the equivalent of 1.

Concerning the homology of these vessels with those of other Reptiles
and of Birds, there has been much difference of opinion. Rathke (57,
p. 91) designated the vessels a and b arteria anonyma. He called 1
art. subvertebralis, and from embryological data he interpreted it (as
defined in the preceding paragraph) as follows. The tight and left com-
mon carotids arise separately, and later fuse along the middle of their
extent, but remain separate at their proximal and distal ends. Secon-
darily, the right proximal root degenerates, and is thus absent in the
adult.

By “common carotid,” Rathke implies an homology with a vessel of
this name in Lizards. The latier is morphologically equivalent. to the
ventral distributing trunk between the third and fourth arterial arches in
Rathke’s well known type scheme of the arterial arches. The unfused
distal ends of the common carotids each branch at the occiput to form
vessels homologous, in his opinion, to the internal and external carotids
of Mammals.

Fritsch (69, p. 705) later maintained that the a, subvertebralis is an
a. carotis primaria, i. e. produced by a drawing out of the a. anonyma,

1 Contributions from the Zoölogical Laboratory of the Museum of Comparative
Zoölogy, under the direction of E. L. Mark, No. XXXIV,
VOL. XXIV. — NO, 2.

BULLETIN OF THE

46

and hence essentially unpaired, — a view which Hoffmann in Bronn’s
Thierreich, Reptilia, unfortunately adopts.

Van Bemmelen (’88, pp. 114, 115) combats the view of Fritsch, and
brings forward additional instances of the persistence of the right proxi-
mal root of the a. subvertebralis in embryos. At the same time he
points out that Rathke’s a. subvertebralis must be regarded as equiva-
lent to the dorsal collecting trunks of the anterior branchial vessels, and
therefore to the aa. carotides interna of Rathke’s general scheme. The
vessels designated in my description above by the numeral 2 — called
by Rathke a. collateralis colli — are, on this hypothesis, to be regarded
as aa. carotides externe (ventral distributing trunks of anterior bran

chial vessels).

Mackay (89, pp. 126-136) more than any other has contributed by
his embryological and comparative anatomical studies to an interpreta-
tion of the carotids of the Crocodilia. The vessels a and 6 are, accord-
ing to him, the parts of the ventral distributing trunks which lie between
the fourth and third arches, and correspond to the common carotids of
Rathke's general scheme. The morphologically paired vessels (1) are the
combined internal, or. better, “dorsal” carotids. The part designated in
Figure 1 by 1* is thus homologous with the third visceral (first branchial)
vessel. The part designated by ]^ + 1% has arisen by fusion of the dor-
sal collecting trunks of the three anterior branchial vessels t' ‘ough a
part of their extent. The vessels marked 2 and 2 (Fig. 1) are external
or “ventral” carotids, — these vessels being represented in Birds also,
where the so called common carotids are in reality “dorsal ” carotids,
not equivalent to the common carotids of Lizards. Mackay’s results,
which thus confirm and extend van Bemmelen’s, seem conclusive, not
only because he has traced the development of the homologous vessels
in Birds, but because he has found one instance — like that of Brandt
(72, p. 307) long ago — in which the dorsal collecting trunk persists
between the third and fourth arches. It is connected with the third
arch near the proximal (posterior) end of the a. subvertebralis, and is
thus far removed from Rathke's so called “a. carotis interna ” of the
head region. The hypothesis that the a. collateralis colli of Crocodiles
is homologous with the a. carotis externa of Lizards, receives additional
support from the fact that, as in Birds, the a. subclavia arises near the
point at which the a. collateralis colli is formed by the division of the
true a. carotis communis (a and 1).

The conclusions of Mackay have been recently confirmed by the re-
searches of Hochstetter (90).

MUSEUM OF COMPARATIVE ZOOLOGY. 47

In view of the interest attaching to this subject, I have thought it
not wholly superflaous to put on record still another example of the

persistence of the right root of the a. subvertebralis, — especially since
so perfect a root from so old a specimen has not been heretofore men-
tioned, and since no figure of this condition at any stage has, to my
knowledge, been heretofore published.

The specimen in question was one which was being used for study in
the class in Vertebrate Anatomy by one of my pupils, Mr. H. O. Marcy.
The Alligator was about 28 cm. long. The arteries had been injected
from the heart, and a very evident branch, Figure 1, 1", was shown, fully
injected, running from the right common carotid, b, to the a. subverte-
bralis 1° + 1%, Thus the whole carotid system had a very symmetrical
aspect. The branch in question (1%), which must be regarded as the
proximal end of the right dorsal carotid, was injected throughout its
whole length. It measured 0.7 mm. in diameter, the corresponding
measurement of the left root being 1.0 mm. The two trunks passed
dorsally, encircling the trachea and oesophagus, and met near the median
plane, dorsal to the oesophagus and immediately beneath the vertebral
column.

I add also figures of two cases of persistent ductus Botalli, completing
the VI. or pulmonary arch. Some trace of this condition appeared upon
most of the individuals examined. In one case shown (Fig. 3), the ih-
jection filled only the dorsal and ventral ends of the ductus, the middle
region remaining as a white cord. In the other case (Fig. 2), the pul-
monary end only was injected.

Finally, in a simple diagrammatic lateral view (Fig. 4), built up on
the system of Rathke’s diagrams, as improved by Boas (’87), I have
indicated the chief neck-arteries of the Crocodilia. In this diagram,
the usually occurring arteries are drawn in heavy lines ; those occurring
less regularly or abnormally, in fine lines ; those absent, in dotted lines.
I have introduced as the II. arch the anastomosis between 1° and 2
(Fig. 1), thus adopting van Bemmelen’s (88, p. 115) suggestion ; but
I must repeat his caution, that this is only a rather probable, but by

no means demonstrated homology.

48 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

LITERATURE CITED.

Bemmelen, J. F. van.

'88. Beiträge zur Kentniss der Halsgegend bei Reptilien. Bijdr. tot de

Dierkunde, N. A. M. Amsterdam, XVI. pp. 99-146, 2 Tab.
Boas, J. E. V.

'87. Ueber die Arterienbogen der Wirbelthiere. Morph. Jahrb., XIII.

p 115, Taf, T.
Brandt, E.

"72. Sur le ductus caroticus du Caiman à museau de brochet (Alligator
lucius sive Mississippiensis). Bull. de PAcad. impér. St. Petersb., Tom.
XVII. pp. 307-309, 1 fig.

Fritsch, G.

'69. Zur vergleichenden Anatomie der Amphibienherzen. Arch, f. Anat.

Physiol u. wiss. Med., Jahrg. 1869, pp. 654-758, 4 Taf.
Hochstetter, F.

'90. Ueber den Ursprung der Arteria subclavia der Vogel. Morph. Jahrb.,

Bd. XVI. pp. 484-493.
Mackay, J. Y.

'89. The Development of the Branchial Arches in Birds, with special refer-
ence to the Origin of the Subclavians and Carotids. Phil. Trans. Roy.
Soc. Lond., Vol. 179 B, pp. 111-140, 4 pls.

Rathke, H.
'57. Untersuchungen über die Aortenwurzeln der Saurier. Denkschr. d.

Wien Akad., Bd. XIII., Abth. 2, pp. 51-142, 6 Taf.

E

2.

DavENPORT,— Carotids of Alligator.

15.

y,

EXPLANATION OF FIGURES

ABBREVIATIONS.

ao. Aorta. ad mu. lar. To muscles of larynx.
art. pul. Pulmonary artery. mu, pet. Pectoral muscles.

dt. Bot. Ductus Botalli. (e, Ussophagus.

ade. To esophagus. u Trachea.

Left arteria anonyma, Rathke = ventral distributing trunk between IV. and
III. arterial arches.

Right ditto.

Arteria subvertebralis colli, Rathke = a. carotis dorsalis, Mackay = third arte-
rial arch and dorsal collecting trunk anterior to III. arch. In this vessel

three regions may be distinguished : —
Left vessel of the posterior paired portion, the fellow of which on the right

side (1%) is usually lost, but in Figure 1 persists.

H1». Fused portion of 1. (The leader should be continued to the vessel cov-
ered by the esophagus in Fig. 1.)

Left vessel of anterior paired (unfused) portion.

Left collateralis colli, Rathke = a. carotis ventralis, Mackay == ventral distrib-
uting trunk anterior to III. arch.

Right ditto.

Left a. subclavia == secondary or persistent subclavian (cf. Mackay).

Anastomosing branch between 1° and 2 (== II. arterial arch ?).

Fig. 1. Ventral aspect of head and neck regions of Alligator Mississippiensis,

28 cm. long. Part of the esophagus and trachea are represented lying
above (ventral to) the median artery.

Figs. 2and 9. Lateral views of the arterial arches to show two cases of persistent

ductus Botalli.

Fig. 4, Diagram, showing in side view the relation of the arterial arches of Croco-

dilia to the type. Heavy lines represent normal vessels; lighter lines,
abnormally occurring vessels; dotted lines, vessels wholly absent in

adult.

dt Bat

No. 3.— On the Eyes, the Integumentary Sense Papille, and the
Integument of the San Diego Blind Fish (Typhlogobius cali-
Jorniensis, Steindachner). By W. E. Rırrer.!

Tun work the results of which are embodied in the present paper
was begun and well advanced at Harvard University, and has been
completed at the University of California.

I wish here to express my warmest appreciation of the many kind-
nesses received at the hands of Prof. E. L. Mark, not only during my
residence in Cambridge as a student under him, but particularly since
leaving there. I have also to thank Prof. C. H. Eigenmann both for
specimens sent me from San Diego while I was working in Cambridge,
and for valuable information and suggestions about collecting the fish
during my visit to San Diego last summer.

Typhlogobius californiensis was first described by Dr. Franz Stein-
dachner. The species is the type of the genus, and thus far is the
only one known. — Steindachner's (79, pp. 142, 143) description of the
eyes is as follows: “Die winzig kleinen, wie Punkte durchschimmer-
enden Augen, liegen hoch am Seitenabfalle des Kopfes gegen Ende des
ersten Viertels der Kopflänge ; ihre Entfernung von einander steht der
Schnauzenlänge nach und beträgt circa 4 der Kopflinge.” According
to this author the genus resembles Crystallogobius, Gill, from which it
differs in its dentition and abortive eyes.

Miss Rosa Smith (81, pp. 19-21), — now Mrs. C. H. Eigenmann, —
unaware that the fish had been described by Steindachner, redescribed
it, making for it, as the Vienna ichthyologist had done, a new genus, the
name of which was derived from the rudimentary condition of its eyes.
Othonops was the generic name given it by Miss Smith, and this term,
signifying as it does “veiled or obscured eyes,” is, so far at least as the
younger individuals are concerned, undoubtedly more nearly true to the
facts, as the sequel will show, than is the name chosen by Steindachner,
Typhlogobius signifying “blind goby.” The specific name chosen by
Miss Smith was eos. She says: “This species is most closely related to

1 Contributions from the Zodlogical Laboratory of the Museum of Comparative
Zoölogy, under the direction of E. L. Mark, No. XXXV.

VOL, XXIV. — NO. 3.

€

02 BULLETIN OF THE

Crystallogobius nilssoni (Düb. & Ker.) Gill (Gobiosoma nilssoni Günther,
Cat. Fishes Brit. Mus., III. 86), a species found on the coast of Norway,
from which it is distinguished generically by the obsolete eyes. . . . The
eyes are large and conspicuous in C. nilssoni.”

With reference to the integument of the head aud its tactile organs,
this author says : “On the under side of the head'the skin (in a pre-
served specimen) lies in irregular folds, which conform generally to the
outlines of the lower jaw, the outer folds reaching the gill openings.
Between the lower lip and these folds there is a series of papilla which
has its origin a short distance behind the corner of the mouth, the series
being slightly separated close behind the symphysis of the lower jaw by
two small, rounded flaps. The papilla number about fourteen on either
side of the flaps. On the superior surface of the snont, extending pos-
teriorly half as far as the termination of the maxillary, the skin is finely
wrinkled, and there is on either side a conspicuous flap, which seems to
conceal a nostril.” The largest specimen examined by Miss Smith was
22 inches in length.

The same author (^90, p. 181) publishes a note made by her at San
Diego, July 3, 1882, on the tenacity of life exhibited by this species, which
is so characteristic of it that I quote the passage nearly entire : “Three
specimens were secured and were placed alive in a two-quart tin pail
along with seaweeds, polyzoa, hydroids, living mollusks, a sea-cumber,
and a number of small fishes and crabs. The living forms in the pail
were so crowded and so short of water that all of the fishes except the
three pink blind fish had died before I reached home, the drive of
twelve miles being over a hilly road for some distance. . . . When re-
turning from La Jolla and other points along the seabeach, I have fre-
quently carried home the tide-pool species alive in this manner, and
invariably the Oligocottus analis, one of the small Cottidee, was more
tenacious of life than any of the other species. At this time, however,
Oligocottus expired with the rest, leaving the blind fish to claim the
honor of being the most hardy of the smaller species of the region.
This species is scaleless and exceedingly slippery. I took one of these
examples from the pail, when, like an eel, it slipped through my fingers
into a barrel of rain water standing near, swimming around in the barrel
several times. I then removed it to a clean shallow dish into which I
had poured half a cupful of sea sand, together with the small amount of
dirty sea water which had covered the medley of animate beings before
mentioned. Typhlogobius, still active, tried to bury itself in the sand,
but the dish was too shallow, and several efforts proved unavailing. . . .

MUSEUM OF COMPARATIVE ZOOLOGY. 53

Tt was still quit» active five hours after it was removed from among the
dead fishes. How much longer it may have been able to survive I do
not know, as I then killed it with alcohol.”

In a paper on the “ Point Loma Blind Fish and its Relatives,” Prof.
C. H. Eigenmann (90, pp. 65-71) has given some very interesting facts
on the habits of the species, and also the only account, so far as I am
aware, of some of the profound structural changes that have been
induced in it by its peculiar way of living.

The fact that the fishes pass their lives under stones in crab holes, or
buried in the sand, must of course have been known by every one who
had collected them ; but as Prof. Eigenmann has had much better op-
portunity to study their habits than has any one*else who has written
on the subject, his account is quite full, and so interesting that 1 repro-
duce a considerable portion of it.

About San Diego the fish has beon found at Point Loma only; it
has been taken, however, at Encenada. Its habitat is consequently, so
far as known, quite limited. The crustacean in the holes of which and
with which it lives is a burrowing carideoid, which has the same pink
color as the fish ; but while the crustacean is found throughout the en-
tire bay region, the fish is its companion only at Point Loma, Another
species of the Gobiide, belonging to the genus Clevelandia, also fre-
quents the holes of the same crustacean along with Typhlogobius. *

“Sometimes the fishes [other than the blind fishes 1] live quite out of
water on the damp gravel and sand under rocks. . . . In the bay the
gobies habitually live out of the holes, into which they descend only
when they are frightened, while at Point Loma they never leave their
subterranean abode, and to this fact we must attribute their present
condition.”

It is not the eyes alone that have undergone modification. The
whole frontal region of the skull has been profoundly changed ; the
scales have entirely disappeared, the color has been reduced, and the
spinous dorsal has been greatly diminished in size. “ The skin, and
especially that of the head, has become highly sensitized.”

1 I find Clevelandia in San Francisco Bay at West Berkeley ; and here it often
enters holes in the mud with a species of Crangon. In this case the holes are not,
I think, dug by the crustacean. The general appearance and actions of the two
animals are so similar that at a little distance it is very easy to mistake the one for
the other. The color of the two is absolutely indistinguishable as they rest at the
bottom of the shallow tide-pools ; and it is so like the dark brown mud of the bot-
tom on which the animals are found that it is with great difficulty that they are

seen when not in motion.

54 BULLETIN OF THE

In a specimen about 0.9 of an inch long the color cells were well
formed, and the membranes of the fins were thin (in the adult the fins
are very thick in proportion to their height). The movements of the
fish at this age were similar to those of the other gobies, and not at all
sluggish, like those of the adult. In the adult, says Eigenmann, “the
color has been reduced.”

The eyes have suffered the greatest change of all. In the small
specimen just mentioned they were quite evident, and apparently func-
tional. “Objects thrust in front of them are always perceived, but the
field of vision is quite limited. With age the skin over the eyes thick-
ens, and they are scarcely evident externally. As far as I could deter-
mine they do not see at this time, and certainly detect their food chiefly,
if not altogether, by the sense of touch.”

“The lens is large in proportion to the size of the eye, which does not
materially differ in size in the smallest and largest specimens exam-
ined. The optic nerve is slender and long as compared with that of any
of the other gobies.” Because of lack of facilities for histological work,
Prof. Eigenmann did not study the minute structure of the eyes. All
his attempts to get material for studying the development were unsuc-
cessful, though artificial fertilization was tried, and many visits were
made to Point Loma in search of eggs and embryos. The spawning
time is June and the latter part of May.

During the last of June and the first of July, 1891, I was able to
spend several days at San Diego, but at that time it was impossible
to get specimens of the blind fish, a thing which I had greatly hoped
to do. From Dr. Eigenmann’s experiences I had thought it quite possi-
ble that at this time I might also be so fortunate as to get embryonic
material of the species. On arriving at San Diego my hopes were at
once annihilated, however, as I found that the fish could be caught only
at the very low tides, and at this time of the year these tides come in
the night. I arranged, however, with two local collectors, Mr. L. C.
agg and Mr. O. N. Sanford, to have all the specimens they could
obtain, as soon as the tides would permit of their being collected, sent
to me at Berkeley. By this means I secured twenty-two specimens, the
most of them apparently fully grown, though two were considerably
smaller than any I had previously studied, and these were all carefully
preserved in Perenyi’s fluid.

Through the kindness of the officers of the Pacific Coast Steamship
Company’s steamer “Pomona,” twelve specimens were sent to me in

July from San Diego alive. Only four of these were living, however,

MUSEUM OF COMPARATIVE ZOOLOGY.

when they reached me, and only one was in perfect health. This one
was kept in an aquarium of about four gallons of water until February
10, 1892. I can fully confirm the statements of other writers concerning
the extreme tenacity of life exhibited by these fishes. During the last
two months that this specimen was kept, the water was not changed nor
aerated in any way, nor was food given to the fish ; and I may say that
it did not, so far as I could determine, take food at all during the time
of its confinement, except the small quantity naturally contained in the
water. Worms were placed in the aquarium, but had to be removed,
untouched so far as I could see, to prevent contamination of the water
by decomposition, When the fish was killed, it was to all appearance
as well and as lively as it had been at any time during its ca ptivity.

Experiments for the purpose of determining whether the eyes in this
individual still performed their proper function were not very satisfac-
tory. Very frequently when the water was suddenly illuminated by
a strong light thrown into the aquarium, standing in a dark room, the
fish was found to be moving about at the bottom with considerably more
than its wonted activity. This activity would continue for only a short
time, when the fish would either move more slowly or would settle down
and become quite still. As the movements were almost always rather
slow and infrequent when the fish could be seen, I am inclined to in-
terpret this behavior as indicating that the fish was sensitive to the light.
However, repeated attempts to produce conditions that would cause it to
choose between light and darkness, if it had the power of such choice,
were without positive results. On the whole, both from these observa-
tions on the living fish, and from the structural conditions to be here-
after described, I am of the opinion that the power of perceiving light
is not wholly lost, even in the adult. The specimen kept alive was
32 mm. long, or about 20 mm. shorter than the largest ones that I have
seen.

The pink color mentioned by all those who have written of this fish is
a quite striking feature in its general appearance. It is not at all due
to pigment in the integument, but to the extreme richness of blood-
vessels situated in the sub-epidermal connective tissue, as will be shown
later, It disappears entirely in preserved specimens, the color becoming
a dull opaque white, particularly in large individuals. In small individ-
uals, however, the color of the dorsal side of the head and body is quite
dark from the presence of brown pigment. The causes to which these
different colors, under different conditions of development and treatment,
are duo, are of considerable significance, and 1 shall speak of them more

THE

BULLETIN OF

fully later on. It is not my purpose to give an account of the general
structure of the fish in this paper. I must, however, call attention to
the far reaching modifications that have been brought about by its
peculiar habits, not in the eyes merely, but in many other structures,
As already mentioned, Dr. Eigenmann has shown that the “ entire
frontal region of the skull has been profoundly changed.” He has also
shown that the fins of the adult are much thicker, more fleshy, and
shorter, in proportion to the size of the body, than in the young. My
observations fully confirm these statements.

In the smallest specimens that I have seen, 19 mm. long, the eyes
are distinctly visible without dissection. In some of the preserved
specimens of this and somewhat larger size, the lens is also clearly seen
in surface views ; while in other specimens it is not so distinct, and in
some is scarcely seen at all, though it is probably always present in all
these younger individuals.

The eyes are situated wholly on the dorsal aspect of the head, and
very near together (Plate I. Figs. 1, 2). Their distance backward from
the tip of the nose is also short as compared with the length of the
fish. Thus in the specimens 19 mm. long this distance was 0.95 mm.,
or one twentieth of the entire length of the fish. In large individuals,
especially while living, the eyes are visible from the surface, but appear
as scarcely more than black specks deeply buried in the tissue. In
many cases they cannot be seen at all in preserved specimens.

The epidermis immediately over the eyes does not differ essentially,
either in the smaller or the larger individuals, from what will be de-
scribed further on as existing in other portions of the dorsum and sides
of the head and body. The mucous cells are present here as else-
where, and they are as numerous and as large as in adjacent regions.
The average thickness of the epidermis is 50 y in the smallest specimens
studied ; 63 in a specimen 60 mm. long, and 76 p in a specimen about
65 mm. long, thus showing a gradual increase in thickness with the
increasing size of the animals. In the smallest specimens the sub-epi-
dermal tissue over the eye is not differentiated into a dermal and sub-
dermal layer. The connective tissue in this region is arranged in several
strands which unite with one another at various angles, thus bounding
wide spaces (Plate III. Fig. 17, spa.). In this specimen (Fig. 17) the
space between the epidermis and the sub-epidermal tissue is quite wide,
and is continuous over the entire eye, and for a considerable distance
beyond. This space may be in part artificial ; but even if so, the con-
nection between the epidermis and the immediately underlying tissue

MUSEUM OF COMPARATIVE ZOOLOGY, 57

must have been exceptionally frail, since separations of this kind are
rare in adjacent regions.

The thin layer of tissue next to the lens — in fact closely applied
to it in many cases (Figs. 5 and 17, crn.) — is part of a layer that en-
velops the entire eye, in many places lying close upon the pigmented
layer of the retina. The portion in the region of the retina is undoubt-
edly the sclera, and will hereafter be designated as such. The portion
in the region of the lens may be regarded as representing all there is of
a cornea excepting its epithelial layer ; but on this subject I shall speak
further presently. The strands of connective tissue that have been
spoken of as intervening between the epidermis and the eye in the
smaller specimens are distinctly fibrous, and contain numerous small,
much flattened connective-tissue nuclei. These strands are directly
continuous with the sub-epidermal connective tissue of the surrounding
regions, and are not largely continuous with the sclera, though in part
they are (Fig. 17 a).

In addition to the small flattened connective-tissue elements in these
bands of connective tissue, a few much larger cells are found (el. con't.).
They have distinct round nuclei, and each nucleus has a nuclear mem-
brane and a nucleolus, The membraneless cell body is drawn out into
one or more processes, usually two or three, which become lost among and
are apparently continuous with the fibres of the connective-tissue strands
in which the cells are situated, They are probably embryonal connect-
ivo-tissue cells concerned in the production of the thick layer of this
tissue that intervenes between the eye and the epidermis in older speci-
mens (Plate IT. Fig. 6). In this older specimen (Fig. 6) a large num-
ber of nuclei aro seen, in part immediately over the eye, and consequently
in the same position as the cells regarded as embryonal connective-tissue
cells in the young specimens ; but they are mostly at one side of the
eye (Fig. 6, con't. tis.), and although some of them are undoubtedly
cells of connective-tissue character, at the same time many of them are
certainly not of this nature, but are probably leucocytes. As shown in
Figure 6, there is over the eye in the large specimens a well defined
dermal layer, drm., which usually remains adherent to the epidermis
when the latter is removed. This layer is nearly structureless, though
fine fibres are not uncommon in it. In the specimen shown in Figure 6,
the entire thickness of the tissues over the eye is about 392 a, of which
103 a is epidermis and 289 p sub-epidermis. About midway between
the epidermis and the eye there is a thin stratum of formed connective

tissue (st. con't,), much denser than the surrounding tissue ; and imme-

58 BULLETIN OF THE

diately beneath the dermal layer is a layer of comparatively coarse fibres
arranged in bundles running more or less nearly parallel with one an-
other, Among these bundles are blood-vessels and numerous cells,
mostly of the kind that I have regarded as leucocytes, The remaining
portion of the tissue in this region is composed of rather fine uniform
fibres, containing very few cellular elements. I have said that the eye-
enveloping portion of the connective-tissue capsule that is immediately
over the lens probably represents the cornea, excepting its epithelial
layer. When, however, we bear in mind the method of development of
the cornea, — as first clearly made out by Kessler (77, pp. 83-94), and
now understood by all embryologists, excepting in so far as this author
believed it to contain, besides its epithelial layer, elements derived
from the ectoderm ; and when we remember, further, that in the cornea
of the normal adult eye, the substantia propria, together with the mem-
brana elastica anterior and the membrane of Descemet, make up its
entire thickness, excepting its conjunctival (i. e. epithelial) layer, the
interesting question arises whether in such an eye as is represented in
Figure 6 the layer ern. should be regarded as representing the whole
cornea, or merely the membrane of Descemet. Should the latter inter-
pretation be adopted, then it would follow that the tissue intervening
between this and the dermal layer would be the substantia propria
greatly thickened, and the dermal layer (drm.) would be the membrana
elastica anterior of the cornea; and what I have called embryonal con-
nective-tissue cells (Fig, 17, cl. con’t.) might then be regarded as corneal
corpuscles. However, I hardly think this the right view of the matter,
since, as already pointed out, the tissue over the eye is mostly con-
tinuous with that of the adjacent regions other than the sclera. Tt
is possible that the strands seen at «, Figure 17, give some support to
such an interpretation. But whatever view may be taken, it seems to
me that we are justified in regarding the conditions here presented as
evidence against Kessler’s statement that a portion of the cornea, be-
sides its epithelial layer, is derived from the ectoderm. This author's
account of the development of the cornea in Triton is in substance as
follows.

The first trace of it to appear is a thin layer of hyaline substance
on the inner surface of the ectoderm over the eye. This appears at a
time when the cavity of the lens vesicle has wholly disappeared, and
the retinal layers have begun to be differentiated. (See the author’s
Figure 60.) This layer is held to be secreted from the ectoderm. The
succeeding steps may best be given in the writers (77, pp. 89, 90) own

o]

MUSEUM OF COMPARATIVE ZOOLOGY. 9

words: “Wenn die zuerst vorhandene hyaline Schicht eine gewisse
Dicke (Figg. 60 und 61) erreicht hat, wird dieselbe vom Hornblatt ab-
gedrängt durch eine zweite an dieses sich anbildende hyaline Schicht ;
in das zwischen beiden Schichten entstehende Interstitium dringt von
der Peripherie her eine einzellige Lage der spindelförmigen Kopfplatten-
elemente, die sich vorher schon in einen spitzen Winkel gegen das
Hornblatt am Rand der Corneaanlage gestellt hatten, ein (Fig. 62);
sobald dieselben von allen Seiten her im Pol der Cornea zusammen-
treffen, ist die erste hyaline Schicht von der unterdess zu der gleichen
Dicke entwiekelten zweiten vollständig gesondert. Ebenso wie die erste
durch die zweite, wird dann die zweite durch eine dritte neu sich bil-
dende Schicht vom Hornblatt und darauf durch eine zweite einwandernde
Lage von Kopfplattenelementen von der dritten Schicht isolirt ; diese
wieder vom Hornblatt durch eine vierte neue Schicht und von letzterer
durch eine dritte Zellenlage u. s. f.” It would thus appear that a very
intimate connection is brought about between the ectoderm itself, the
relatively large portion of the substantia propria derived from it, and
the mesodermal elements of the cornea; and it should be especially
noticed that this process goes on at a comparatively late stage of devel-
opment, — viz. at a time when the retinal layers are being differentiated,
and after the pigmented portions of the eye are well formed ; in short, at a
stage only a very little earlier than that at which development is arrested
in the eye of Typhlogobius. If such a process had ever taken place
here, it seems almost certain that we should see some indications of it
in such a stage as is shown in Figure 17 (Plate TIL). But, on the con-
trary, what we do find is no connection between the epithelium over the
eye and the immediately underlying tissue, or at least almost none, aud
no indication of a hyaline layer on the inner surface of the epithelium.
While, on the other hand, in older specimens (Plate IT. Fig. 6) the epi-
dermis and the sub-epidermal tissue are in close connection, there being
no interruptions or spaces at all, and we have here a well defined nearly
structureless layer closely adherent to the epidermis.

There is considerable individual variation in the size of the eye. In
three specimens, 50 mm., 60 mm., and 63 mm. long, the diameters, meas-
ured parallel to the long axis of the head, were respectively 0.44 mm.,
0.46 mm., and 0.47 mm. ; the diameters transverse to the long axis of the
head in the last two of these were, respectively, 0.39 mm. and 0.47 mm.
In another specimen 63 mm. long, the diameter transverse to the head
was 0.372 mm. ; the diameter parallel to the long axis of the head was
not measured in this specimen. This last measurement was made on the

60 BULLETIN OF THE

section, while the measurements in the case of the first three were made
with the eyes in place in the head, the head having been cleared in
clove oil. It is undoubtedly true, that some of the difference in size
between the last mentioned eye and the first three was due to shrinkage
in it during its passage through the paraffine. I have made numerous
measurements of the same eyes before embedding and after cutting, and
have always found the sections somewhat less in diameter than the
whole eyes, even when all precautions had been taken to prevent shrink.
age. But certainly shrinkage cannot account for the great difference
found here between the eye measured in section, and the one of an in-
dividual of the same length measured in the head. This small eye, I
should add, was the only one that I have found in which the lens was
wholly absent. This eye will be described later on. In a specimen
25 mm. long the axial diameter of the eye was 0.28 mm., and its equa-
torial diameter was 0.45 mm. In a specimen 19 mm. long the diameter
in the long axis of the head was 0,28 mm., and transverse to this it was
0.39mm. It appears from these measurements that the eye does in-
crease somewhat with the increase in size of the animal, though it is
true that, in view of the obvious individual variation in size in specimens
of nearly the same length, not enough of the smaller specimens have
been studied to determine definitely how much this increase amounts to.

A sclerotic coat, well defined from the surrounding connective tissue,
is always present, though in some places its fibres, both singly and in
bundles, leave their concentric course and pass off into the connective
tissue, and thus bring about an intimate connection between the two.
In some places the connective-tissue fibres not belonging to the sclera,
but in its vicinity, are seen to have taken on a concentric direction even
at a considerable distance from the eye, and to have become more
numerous and more closely packed than is the case with the subcuta-
neous connective tissue in general. There is thus brought about a
fusion to some extent of the eye bulb with the tissue in which it is
embedded. This statement applies especially to the eye the section of
which is shown in Figure 6. In most sections numerous flattened con-
nective-tissue cells are found in the sclera; and in all the eyes that I
have sectioned the sclera is cartilaginous in the region corresponding to
the ora serrata of the retina. The cartilaginous layer is usually only
one cell thick, but occasionally it is two or more cells thick (Figs. 5, 6,
and 12). In many specimens the cartilage does not extend entirely
around the eye in the equatorial zone, and in no case have I seen it
extend more than half-way back to the entrance of the optic nerve. No

MUSEUM OF COMPARATIVE ZOOLOGY. 61

indication of ossification of the scleral cartilage, as is common in bony
fishes, has been seen. No trace of an argentea is to be found. All the
pigment is of the same dark brown granular variety, and, when seen
by reflected light, never gives the white, silvery color that is character-
istic of the crystalline material of the argentea.

The choroid is exceedingly rudimentary, and in many specimens I
have been unable to detect its presence at all. In the eye from which
Figure 13 was drawn (Fig. 13, chr.), —a specimen the eyes of which, as
will be seen further on, are better developed in several respects than is
usual in these eyes, — it is more distinctly seen than in any other case
that I have studied. Here the layer of pigment is very thin; it is in-
terrupted at short intervals, and cannot be traced for more than one
third of the distance through which it would normally extend.. Whether
this pigment should be regarded as representing the lamina supracho-
roidea, or as belonging to the choroid proper, it is impossible to say.

In a few instances (Figs. 13, 14, 15, chr. cpl.), a layer of cellular tissue
has been found at the proximal pole of the eye, extending for a variable
distance toward the anterior rim of the optic cup, but rarely reaching it.
This layer is always:closely applied to the outer surface of the pigmented
layer of the retina, and in some sections it seems to be continuous with
the pigment of the choroid. In some places (Fig. 15, chr. cpl.) the tissue
has very much the nature of formed connective tissue, while in other
places (Figs. 13 and 14, chr. cpl.) the cells are spherical or elliptical, with
indications at times of blunt processes, and with distinct nuclei, Where
cells of this kind occur, the layer is somewhat thicker than where the
structure is more characteristically that of connective tissue ; and in
several instances blood corpuscles (Fig. 13, cp. sng.) are found scattered
here and there in these thicker portions of the layer, indicating the
presence of capillaries, I identify this layer as the chorio-capillaris.

A conspicuous structure in all the specimens studied is a thick,
usually short, somewhat lenticular mass of pigment occupying a position
usually at the proximal pole of the eye, at or near the entrance of the
optic nerve, by which it is pierced in some cases. (Figs. 14, 15, etc.,
gl. chr.). This mass is concentric with the surface of the retina, but is
always separated from it by a short though somewhat variable interval.
In some places the cellular layer just described in part occupies this
space, and in some places the thin layer of choroid pigment is seen to
enter the same space. The mass always lies within the sclerotic, and
is always composed entirely of pigment. I have been unable to find
any cellular or other protoplasmic elements in it. This body I interpret

62 BULLETIN OF THE

as representiug the “choroid gland.” The evidence for this is princi-
pally in the position which it occupies, and very little in its structure.
It is true that this body may contuin a small amount of pigment in the
normal eye, as I find to be the case in Clevelandia, yet its characteristic
structure consists, as is well known, in its richness in blood-vessels ; but,
as already said, none of these occur in Typhlogobius. Its position — viz.
at the proximal pole of the eye at or very near the entrance of the optic
nerve, and between the chorio-capillaris and the sclerotic — is, however,
strong evidence in favor of regarding it as the “choroid gland.” This
structure is described in text-books (Wiedersheim, ’86, p. 412) as being
situated in the normal fish eye between the argentea and the pigmented
layer of the choroid. The fact that no argentea is present in the eye of
Typhlogobius weakens somewhat the force of the evidence that I have
used to support the assumption that I have made with reference to the
significance of the pigment mass described. But its relation to the
chorio-capillaris and the pigment layer of the choroid are the same as
that of the “choroid gland”; as is also its relation to the sclera, with
the exception that no argentea is interposed between the two. Of
course it is impossible to say that, were the argentea developed, it would
lie between the sclera and the pigment mass, rather than between the
lgtter and the pigmented layer of the choroid. "We however have as
much reason to suppose it would occupy the former position as the latter.

The pigment layer of the retina is exceedingly thick. In a specimen
about 50 mm. long, the entire thickness of the retina including the pig-
ment layer was 0.108 mm., and that of the pigment layer was 0.07 mm.;
while in a specimen of Clevelandia of about the same length, the entire
thickness of the retina including the pigment layer being 0.13 mm.,
the thickness of the pigment layer was only 0.037 mm.; that is, in
Typhlogobius the thickness of the whole retina is to the thickness of the
pigment layer as 1.5 : 1, while in Clevelandia the corresponding ratio is
3.5:1. In Gasterosteus, I find that about the same proportion holds
as in Clevelandia, whereas in the perch (H. Müller, ’57) the proportion
is at least not less, and in the salmon a year old (Hoffmann, '83) the
proportion is 64- : 1.

The layer is composed wholly of pigment of the dark brown granular
variety. I have searched in vain for cellular elements within it, In
most specimens the pigment is a very uniform mass; but occasionally
one finds an eye in which very distinot and perfectly round nodules of
pigment occur. Some of these are so clear-cut and smooth in outline
that they have the appearance of perfectly round cells wholly trans-

MUSEUM OF COMPARATIVE ZOOLOGY. 63

formed into pigment, though they are considerably larger than any
cells, even blood corpuscles, that are found in the eye (Plate II. Fig. 13,
glb. pig.). They probably merely signify that the. pigment tends to
segregate in such nodules during its formation. On the inner surface
processes of pigment project down among the rods, as in normal eyes,
excepting that in most instances they are relatively much shorter and
less distinct ; in some specimens they scarcely appear at all (Plate II.,
Figs. 5 and 13 prc.). The greater portion of the thickness of the layer
pertains to the region between the bases of these processes, i. e. the
terminals of the rods, and the outer surface of the layer. And it is
hence in this portion that the increase in thickness over that found in
normal eyes has taken place ; for in the latter, this region is relatively
thin. Iam at a loss to know from what source this pigment has come.
As already said, no cellular elements are to be found in the layer, so it
is quite certain that they have completely degenerated into pigment.
I shall return to this question in the comparative part. The layer thins
out rapidly as it approaches the ora serrata, and is frequently thrown
into an equatorially directed fold (Plate IT. Figs. 5 and 13, *), which
may extend entirely around the eye, but more commonly is confined to
one side of it.

Immediately in front of the thinned out region just mentioned, the
pigment thickens again somewhat, to form the pigmented portion of the
iris. This structure, though always present, varies greatly both in form
and size. Ina majority of specimens the pigmented portion constitutes
the entire iris ; and in all cases it forms by far the greater portion of it.
The dense pigment is entirely the same in structure as that in the pig-
ment layer of the retina. In the few cases where a cellular portion is
present, it is in small quantity, and appears to be of the nature of con-
nective-tissue cells and fibres. Neither blood capillaries nor epithelium
have been detected in this region (Figs. 5, 13, and 17, ör.). It fre-
quently happens that the outer surface of the iris is in contact with the
inner surface of the cornea (see figures), and it is thus made to appear
as though the iris has a considerable part in addition to its pigment ;
this, however, is undoubtedly only an appearance caused by the inner
border of the iris having been thrust outward by some artificial means, —
probably by the lens being in most cases moved outward, for this body
is very loosely held in its place. In a few specimens a trace of the
ligamentun pectinatum is present, though in most cases no trace of it
exists. The short blunt processes of pigment occasionally seen project-
ing toward the lens (Plate III. Fig. 17, 8) remind one of the ciliary

64 BULLETIN OF THE

processes, but it is extremely doubtful if they should be so interpreted.
They are situated too far from the ora serrata and too near the free
edge of the iris. They are probably in some way merely incidental to
the extensive pigmentation of the iris.

In no instance are the layers of the retina in Typhlogobius as fully
differentiated as in the normal eyes of fishes. I will first describe them
as they are found in the majority of specimens, beginning with the inner-
most layer, and will afterwards speak of the cases that show deviations
from the common condition.

An internal limiting membrane, distinct from the layer of nerve fibres,
I have been unable to find. Corresponding with the exceedingly rudi-
mentary condition of the optic nerve, the layer of nerve fibres is very
thin, even in the immediate vicinity of its entrance, where, in the normal
retina, it reaches its greatest thickness. In many sections only frag-
ments of it are to be seen; and for considerable areas no traces at all
are found. It is possible that its absence is due to its having been
broken away during the preparation of the sections; but, however that
may be, it is certain that, wherever present, the layer is very thin (Plate
II. Fig. 5, Plate III. Figs. 18 and 21, st. for. opt.). The next layer, viz.
that of the ganglion cells (Figs. 18 and 21, cl. gn.), is always distinct,
and is from one to three or four cells in thickness. As a rule only the
nuclei are distinguishable ; but occasionally the cell bodies can be made
out. Examined under high powers and with careful focusing, some of
these are found to possess one or more processes (Fig. 18, cl. gn). A
nuclear membrane can usually be seen, as can also one to several darkly
stained particles within the nuclei. The nuclei are in general very
nearly spherical, though there is a tendency for them to become ellip-
tical with the longer axis radially directed.

The inner reticular layer (st. rtl. 4.) is well developed in all cases, and
is essentially the same in structure as in the normal eye, though I have
not detected any of the fibres running parallel with the surface of the
retina that are found in normal fish eyes. Within this layer the radial
fibres of Miiller (fbr. Mú.) can usually be made out, though they ap-
pear to be few and indistinct.

The remaining portion of the retina, as far as to the external limiting
membrane, is never fully differentiated into the layers that are found in
the normal eye between this membrane and the inner reticular layer ;
and in many specimens scarcely any indication of a differentiation can
be seen. About the average condition is shown in Figure 18, st. rtn./,
where a layer of nuclei (st. nl. i) about two or three deep may be dis-

MUSEUM OF COMPARATIVE ZOOLOGY. 65

tinguished next to the inner reticular layer. These are slightly larger,
on the average, than are the more superficially situated nuclei (st. nl. ex.),
and they also stain somewhat more deeply. Usually each nucleus has
a centrally placed nucleolus, and a not very distinct nuclear membrane.
These nuclei undoubtedly represent the inner nuclear layer, though
whether the entire layer or only its spongioblasts, it is impossible
to say.

Following this layer there is usually a single layer of nuclei (st. rtl. ex.)
that are considerably smaller and less distinct than those of the layer
last described. Not only are the nuclei themselves here smaller and
less distinct than those of the layers on either side of them, but spaces
(st. tl. ex.’) are seen at intervals in which there are no nuclei at all. It
seems quite certain that the outer reticular layer is represented by
these spaces.

Finally, outside of this layer follows another of nuclei about one or
two deep (st. nl. ex.) that are again somewhat larger and more distinctly
stained than those in the layer last described. No difference between
them and those of the inner nuclear layer can be discovered, excepting,
as said above, that they are slightly smaller, and stain a little less
deeply. They are undoubtedly the nuclei of the rods, i.e. the outer
nuclear layer, though I have been unable to trace a connection between
them and the rods, and it is somewhat surprising that thoy are slightly
greater in diameter than the rods.

The external limiting membrane (mb. lim. ew.) is usually quite
distinct.

The rods are well — probably normally — developed, but I have
searched in vain for cones. In the retina shown in Figure 18, they are
quite variable both in length and diameter, being in a few instances
knobbed at the outer ends (bac. cla.). It is, however, quite possible
that both this appearance and the shortness of some of them are due
to artificial causes, but the variation in diameter could scarcely be so
explained.

With a single exception, to be described more minutely hereafter,
the Jens has been present in all the specimens studied in detail.

It differs in no way in structure from the lens in normal fish eyes that
I have examined; i.e. it appears entirely homogeneous and structure-
less after preservation in alcohol, Perenyi’s fluid, or picro-nitric acid.
It is held in position very loosely, and consequently is easily displaced ;
it is frequently found, in prepared specimens of the eye, pushed entirely
out of its proper situation into the somewhat yielding connective tissue

VOL. XXIV. — NO. 3, 5

66 BULLETIN OF THE

which lies immediately over it. No trace of a suspensory ligament seems
to be present, nor have I been able to find a processus falciformis.

In a single instance, viz. in the eye portions of which are shown in
Figures 13 and 21, and which has already been spoken of as showing
in several respects a higher state of development than any other speci-
men studied, I find close behind the iris, within the cavity of the
eye, a few strands of tissue containing a few small nuclei, which may
possibly be regarded as the hyaloid membrane, though I do not feel at
all certain that this is their nature. In the same specimen a very few
blood corpuscles are also found close behind the lens within the eye
chamber. With this single exception I have been unable to find a trace
of tissue within the chamber.

The optic nerve, although exceedingly slender, is always present, so
far as my observations have gone. A very striking fact in connection
with it is the thick sheath of pigment that surrounds it in its passage
through the retina (Figs 6, 13, and 15, pig.'"). This sheath invests the
nerve very closely, no space existing between it and the latter ; however,
in its course through the retina, a considerable non-pigmented space is
sometimes seen between its outer surfuce and the pigment of the ret-
ina; indeed, in a majority of cases the outer boundary of the sheath
can be traced entirely through the retinal pigment.

These facts make me incline to the opinion that this sheath really
belongs to the nerve, and has arisen by the pigmentation of the outer
portions of it. The fact that in many cases it continues on without
interruption through the cellular portion of the retina, nearly to its
inner surface (Plate IT. Figs. 6 and 13), gives considerable confirmation
to this view. It would seem, however, were this the right interpreta-
tion, that we should find a rather more gradual disappearance of the
pigment in passing, on a section, from the main mass of the wall of
the sheath to the unpigmented portion of the nerve, than we do; but
the inner surface of the sheath is not quite as sharply defined as its
outer surface is.

On account of the position of the eyes, far anterior to the brain, and
near the anterior extremity of the head, the optic nerves are very long.

'The muscles of the eyeballs (Fig. 8) are also very long and slender,
and are probably always present, though I have not been able to detect
them in the sections in all cases ; but in eyes dissected out and cleared
in glycerine or clove oil, or slightly stained in Schneider’s acetic acid
carmine, I have always found them. Figure 8 is a camera drawing
from a glycerine preparation, showing all the muscles excepting the
internal oblique, and also the nerve.

MUSEUM OF COMPARATIVE ZOOLOGY. 67

I will now describe certain eyes that present exceptions to the con-
ditions already described. The first will be the eye that shows the
maximum development; the second, eyes that show the minimum devel-
opment and the maximum degeneration.

The specimen that shows the greatest development has already been
mentioned (page 61) in connection with the choroid and the hyaloid
membrane, and the reader is referred to the statements there made con-
cerning these structures.

The most interesting evidence of unusual development is found in
the greater differentiation of the layers of the retina; and it is in the
greater distinctness of the outer nuclear (st. nl. ex.) and outer reticular
(st. rtl. ew.) layers that the difference chiefly consists (Fig. 21). Whereas
the former is ordinarily, as in Figure 18, only one or two cells thick,
(exclusive of the cells st. rel. ex.), in this instance it is three or four cells
thick ; but more significant than its greater thickness is the fact that
the deepest cells (nl. ba.) are arranged in a regular layer one cell thick,
and closely packed.

What this layer of nuclei represents will be pointed out in the com-
parative part; it may be said here, however, that it probably does not
belong to the outer nuclear layer. Although the externa] reticular
layer even here does not present a well defined boundary either inter-
nally or externally, — particularly not internally, — the irregular areas
which have already been mentioned as being destitute of nuclei in the
average specimens (Fig. 18, st. rtl. ex.) are here considerably more dis-
tinct (Fig. 21, st. rel. ex.), both as to thickness and frequency of occur-
rence. In fact, the layer may be regarded as having the characteristic
structure of the normal eye, excepting that it contains a considerable
number of scattered nuclei, and is without distinct limitation internally.
The rods also seem to be better developed in this specimen than in
others. In many of them what has the appearance of an indistinct
structureless nucleus may be seen occupying the extremity of the fourth
of the rod nearest the external limiting membrane (Plate IV. Fig. 22,
aand b). This one fourth probably represents the inner member of the
rod. In some cases the nuclear-looking structure terminates on the sidé
directed toward the distal end of the rod with a well defined straight line,
but in other cases the whole has a round or elliptical form. This portion
of the rod takes stain slightly, but itis the only portion that does. The
substance of the remaining portion is uniformly opaque-glassy in appear-
ance, excepting that numerous pellucid spots exist in it. . These are con-
siderably more distinct in some rods than in others, and occasionally

68 BULLETIN OF THE

they form a regular row through the whole length of the outer member
(Fig. 22, c). They are mostly confined to the outer members, but in
some rods a single one of unusual distinctness is found in the inner
member,

I now pass to the least developed and most degenerate eyes. Perhaps
the specimen most interesting in this respect is the one shown in
Figure 6. In this the lens is entirely absent, a continuation of the
pigment layer of the retina extending without break entirely over the
space that this structure should occupy. The series of sections of this
eye is complete, and every section is as perfect as the one from which
Figure 6 was drawn; so there can be no possibility that the absence
of the lens is due to artificial causes, or that an error in observation
has been made. I should add, also, that I have an equally complete
series of sections of the other eye of the same individual, and this shows
the same structure in every particular. Over a portion of the outer.
moiety, the pigment is disposed in a double layer (Fig. 7, pig.’ and pig.").
The outer of these layers is more directly continuous, both in thickness
and direction, with the pigment layer of the retina. The inner layer
presents a considerably thickened mass, pig.", irregularly lenticular in
form. From the appearance shown in this figure, one is tempted to
regard this pigment mass as a remnant of the ‘lens, the thinner portion
of the layer at its edges, which connects it with the pigment layer of the
retina, representing the iris. It is very doubtful, however, if such is
the case. With such an interpretation the outer pigment layer (pig.!)
would seem to represent the inner layer of the cornea; and this would
afford an explanation, not obvious otherwise, of the continuity of the
layer over the pupil. But there are several difficulties in the way of
these convenient interpretations. In the first place, the pigment mass
under consideration is found on examining the entire series of sections
to be very irregular in form, and, as shown in Figure 6, to become con-
fluent with the outer pigment layer (pig.!) in some portions of the eye;
in the second place, as is shown in the figures, and has already been
mentioned, the outer layer is directly continuous with the pigment layer
of the retina, which of course it should not be if it belongs to the
cornea; and, in the third place, the inner layer of the cornea is itself
present and not pigmented (Figs. 6 and 7, ern.).

Concerning the condition of the retina in this eye I speak with con-
siderable hesitation, because of a fear that its peculiarities may be due to
artificial causes. I should say, however, that the specimen was hard-
ened in chromic acid, and that all the tissues around the eye are well

MUSEUM OF COMPARATIVE ZOOLOGY. 69

preserved. If the retina is macerated, it is because the preserving fluid
failed to penetrate through the pigment layer in which it is wholly
incased. Consequently, I shall not dwell at length on the subject, nor
insist much on the significance of the structure described. Figure 19 is
a rather highly magnified view of a meridional section of the retina
near the ora serrata. It will be seen that only the layer of ganglion
cells, the inner reticular layer, and undifferentiated layer are here
distinguished. The inner reticular layer does not differ from that
found in ordinary eyes. But with the other layers the case is quite
different. The nuclei are much less closely packed than in other eyes,
the intervening spaces being occupied by a few scattered fibres. The
nuclei appear to be of two quite distinct kinds. One kind (lewey.) is
somewhat larger than the other, stains considerably less deeply, is less
refractive, and shows no trace of a membrane. In the other kind, the
nuclei are smaller, cl., cl,!, cl.!!, stain deeply, and are refractive in such
a way that in focusing through them they take on the three different
appearances shown in the figure; i. e. when seen at a high focus they
appear uniformly dark, as at cl. ; at a deeper focus the appearance is that
of a disk with a dark line at. its circumference, a light yellow zone (rep-
resented in white in the figure) inside of this, and a uniformly dark
spot in the centre, cl.; at a still deeper focus the appearance is that of a
dark ring with a light centre, c." A few of these latter have two or
three longer or shorter processes, but by far the larger number of them
are spherical, or nearly so.

No rods are present in these eyes, though this portion of the retina is
so ragged and formless that it is impossible to say whether they have
ever been developed or not. It is quite possible that the larger bodies,
in which, however, no nuclei are visible, are leucocytes.

Another interesting exceptional case is that presented by the section
of the eye shown in Figure 12. It will at once be seen that the point
of chief interest lies in the double layer of pigment that is situated
behind the lens. within the cavity of the eye. The outermost of these
layers is in close contact with the posterior surface of the lens, while the
innermost is in close contact with the inner surface of the retina. I
cannot, however, satisfy myself that either layer is developed at the ex-
pense of the parts to which they are respectively adjacent. It is true
that tho line of contact between the lens and its pigment layer is not a
sharp one, there evidently being a gradual disappearance of the pigment
here. The lens, however, shows no flattening on this side, as it would
were the pigment layer formed at its expense ; but more than this, the

70 BULLETIN OF THE

pigment layer extends on either side beyond the lens, its outer edge be-
coming on one side (the lower one in the figure) continuous with the
inner layer of pigment, and on the other side with the pigment layer of
the retina. The connection with the outer is shown rather indistinctly
in the section figured, but much more distinctly in some of the other
sections of the series, as is also the fact that at places the iris is entirely
cut off from the retinal pigment. If, then, these layers of pigment have
come neither from the lens nor the retina, the only other structures
in this region that they could come from are the vitreous body and the
hyaloid membrane. In other. respects this eye does not differ appre-
ciably from the usual structure already described, with this exception,
that the undifferentiated layer of the retina shows somewhat less differ-
entiation than usual. I have studied these sections in vain to find
convincing evidence of actual degeneration within the retinal elements
themselves, in addition to the pigmentation.

I now present a table showing some of the results of measurements
that have been made of the retina, lens, and optic nerve : —

Two Speci- | Spec. | Spec. Ion, of
mens,each |50mm. | 60mm, | Cleve-
19mm, long. | long. | long.! landia.

4 2 8 4 5

mm. mm, mm. mm, mm,
Entire retina nans. Pigment A . +. | 096 | .076 | .110 | ‚146 | .186
Pigmentlayer . . . „3° « | 056 | 046) -068 | .078 | .042
Nerve fibre layer. ecs « +... + . + | 0029 .0014¥ .0048| .0058| .05
Ganglion cell layer .004 | .0029| .0072| .0058) .05

Inner reticular layer . + a Er °.0198150088 OTIO a 018 c 02
.0817

Undifferentiated layer. . . +. + . + + + | 017 5.001 | .0208
“ Undifferentiated layer” of Nos. 1, 2, and 8:

Outer nuclear layer . , .0058| .016

Basal nuclei in outer nuclear layer . .0029' .0029

Inner nuclear layer (including c outer retic-

ular layer) . . + " . .0023

Outer reticular layer EN 0018
Tangential cell layer . . . « rau +. + + 0029
Inner nuclear layer . . s . 6 + + 0 0. 029
Diameter of lens . + . .205 A184
"Thickness of optic nerve at entrance to retina .011 | 117
Length of rods . . + + + + + «+ i 082

Perhaps the most important thing that these measurements reveal is
the fact that the eyes are subject to great variation, as well in the pro-

1 The same specimen from which Figures 13 and 21 were drawn.

MUSEUM OF COMPARATIVE ZOOLOGY. 71

portion of their constituent parts as in their size and degree of difler-
entiation in different individuals.

It is quite probable that some of the differences in thickness between
the retinas of different individuals is due to the fact that measurements
have been made on sections not quite meridional in direction. Care has
been taken, however, in each instance to avoid this source of error. But
in the case of the one giving a thickness 0.145 mm. greater than that of
the normal retina in Clevelandia, the sections are so cut that none are
quite meridional. As may be gathered from Figure 13, this eye is so
flattened in its axial direction that the retina is almost disk-shaped ; and
from this fact it was possible so to cut the retina that none of the sections
would be entirely perpendicular to its surface. The sections are so cut
obliquely, though the deviation from the perpendicular is certainly not
sufficient to account for the great difference in thickness that is shown.
But the difference in proportion between corresponding layers in different
retinas cannot be explained, even in part, as due to artificial causes. In
one of the specimens 19 mm. long, the ratio between the inner reticular
layer and the whole retina is 1 : 7, while in the other it is 1: 13.

From the measurements here given alone, it might be concluded that
there is a gradual increase in thickness, and a constantly advancing dif-
ferentiation in the retina, with increase in the size and age of the animals.
Such a conclusion is not warranted, however, when the whole number of
specimens and stages that have been studied by me is considered, though
it must be admitted that, in view of the obviously wide range of in-
dividual variation, the number of specimens examined has not been suf-
ficiently large to justify an unqualified denial that such is the case. All
that can be said with positiveness is, that, notwithstanding the fact that
the thickest and most fully differentiated retina has been found in a
specimen much larger—and therefore presumably older —- than the
smallest studied, yet several still larger individuals have shown retinas
thinner and less differentiated than those of the smallest individuals;
and, further, that in one instance at least one of the smallest individuals
shows in the distinctness of the outer reticular layer as great a degree of
differentiation as any retina examined. It would be a very interesting
and significant thing, if, owing to a retardation in development, differen-
tiation of the retina should continue throughout the entire: life of these
fishes ; since we know quite well that normally the fish eye becomes func-
tional and differentiated at an early period in development (Balfour and
Parker, '82, pp. 371 and 384; Ryder, '84, p: 500 ; Hoffmann, ’83).

It is exceedingly desirable to ascertain what law, if any, controls the

(2 BULLETIN OF THE

variations of functionless organs. Having now before us the facts relating
to the structure of the eye, we may pass to some reflections on their
significance when considered from a comparative and a developmental
point of view.

First of all, I will speak of the pigment layer of the retina. This has
the greater interest since, according to R. Wright (Wiedersheim, ’86,
p. 427), in the retina of the “blind fish Chologaster papilliferus there is
no pigmented epithelium.”

It has already been shown that in Typhlogobius this layer is always
thicker, relatively, than in the normal fish eye, being thicker than the
entire remaining portion of the retina. I am in considerable doubt as to
how this thickenning has taken place. The first explanation that sug-
gested itself to me was that the choroid had become wholly converted
into pigment and fused with the pigment lamella of the retina. How-
ever, the dense and uninterrupted character of the pigment of the layer,
and the evenness of its external surface, at once threw grave doubts in
the way of this explanation, and the more because of the rather meagre
development of the choroid in the normal eye of bony fishes. Then, as
the choroid was found on further study to be present outside of this
layer, the only remaining alternative was to suppose the latter to be
wholly derived from the proximal wall of the primitive optic vesicle ;
i.e. to represent the pigment lamella of the retina. We may possibly
suppose that the proximal wall of the primitive optic vesicle never be-
came thinned out as it does in normally developing eyes; but the fact
that this process takes place very early — in bony fishes, at least, by the
time the differentiation of the retina has begun — is quite a serious ob-
jection to such a supposition. But even if this were the case, it is hardly
possible to believe that this layer was ever as thick as we find the pig-
ment layer in the adult fish to be. We seem forced to suppose that for
some reason the layer has actually increased in thickness concomitantly
with the retardation in the development of the eye, or, it is quite pos-
sible, with the degeneration of this particular part of it.

T would call attention to the comparison of Typhlogobius with Cleve-
landia in this regard. From the figure of the retina of the latter, it will
be seen that the retinal pigment appears in two quite well marked
layers, an outer and an inner, the two being connected at short but
somewhat irregular intervals by crossbeams or processes (Fig. 20, ex.,
t., and m.). From this it seems that the inner extremities of the pro-
cesses of the retinal pigment layer, which in normal eyes, and particu-
larly in many teleostean eyes, project far down among the rods and

MUSEUM OF COMPARATIVE ZOOLOGY. 73
cones, have here become fused together to form a continuous inner
layer, t.

So far as 1 have been able to determine, this condition is peculiar to
Clevelandia, at least to the extent in which it is here seen. The inter-
esting question now arises whether we have here the beginning of a
process that would, under conditions that have brought about the
changes seen in the Typhlogobius retina, ultimately result in a similar
thick, solid retinal pigment layer; this being effected by a still fur-
ther fusion of the cross rods of pigment now seen, As already pointed
out, it is certain, both from Dr. Eigenmann's observations and my own,
that Clevelandia spends some time at least in holes in the ground.*

The only doubt existing concerning the identification of the layers of
the retina is with reference to what I have called the outer reticular
layer (Plate III. Figs. 18 and 21, st. rtl. ex.), and the layer of nuclei
(Fig. 21, nl. ba.) that has been designated by the non-committal term
of “basal nuclei," basal, i. o., with reference to the outer nuclear
layer. On comparing Figure 21 with Figure 20, the section of a Cleve-
landia retina, there will be little doubt of the correspondence of layer
nl. ba. in the two cases; but at the same time the entire absence of
layer ful. (Fig. 20) will be noticed in Figure 21. These two layers
together seem to correspond to W. Müller’s (74, pp. 60 and 61, Taf.
XIII Figs. 4 and 7) layer of tangential fulcrum cells. This layer is
described by this author as being composed in Petromyzon of “zwei
Etagen grosser quadratischer Zellen, zwischen welchen eine Schicht
ganz flacher, in faserartige Ausläufer sich fortsetzender Zellen gelagert
ist." The layer is said, in the same connection, to be subject to much
modification in the different families of fishes, in which alone it is well
developed ; but the Percidw and the Cyprinide are mentioned as teleos-
tean groups in which the layer with both its “ Etagen ” is present. Ac-
cording to this interpretation the external granular layer of M. Schultze,
salled in this paper the external reticular layer and by Krause (76)
the membrana fenestrata, is not present in either Clevelandia or Typhlo-
gobius; and it is instructive to note that Krause does not find this

L I may here add an observation recently made, which indicates that the time
thus passed hidden from the light is not inconsiderable. On some of my visits to
the beach at West Berkeley I have found the fish very numerous in the tide-pools,
while at other times hardly any are seen. Whether their absence is due entirely
to their having gone into the holes Tam not sure ; but however that may be, at such
times I have occasionally found them by digging” Tam not yet able to say whether
their disappearance is in any way correlated with conditions of the weather as

regards sunshine.

74 BULLETIN OF THE

layer in Perca fluviatilis. Its position would be between the external
nuclear layer and the layer nl. ba. (Figs. 20 and 21). The outer of the
two * Etagen" of Müller’s tangential fulcrum cells appears to corre-
spond to Krause’s membrana perforata, and likewise to M. Schultze's
basal plexus, and the inner to Krause’s stratum lacunosum ; this would
make the layers nl. ba. and st. rtl. ex. of the Typhlogobinus retina the
membrana perforata and the stratum lacunosum, respectively. An
objection to this interpretation is possibly presented by Hoffmann’s ac-
count of the development of the Salmon retina. His Figures 10, 11,
and 12 (Taf. V.) show that what he calls the tangential fulcrum cells
become differentiated quite early, certainly as early as the stage of
development represented by the partially developed retina of the adult
Typhlogobius. But judging from the position of this layer in relation
to the inner nuclear layer and the layer that he regards as the outer
granular layer, it would seem that his tangential fulcrum cells corre-
spond to the inner “ Etagen” only of what Müller designates by that
name. But according to my interpretation these cells are not present
in Typhlogobius, unless they be represented by the scattered cells in
layer st. rtl. ex. The chief point to be made in this discussion of the
homologies of the retinal layers is this. In the most differentiated
retina, even though all the layers found ir the normal adult fish eye
may be marked out, the differentiation is much less complete as regards
the zone between the two nuclear layers than it is in the normal eye of
a closely related genus ; while in a majority of individuals development
is arrested at a considerably earlier stage.

In view of the almost universal statement that the rods and cones
are the latest of all the parts of the retina to be developed, it would, I
think, hardly be expected that the rods should be as complete as they
are in these eyes. O. Hertwig ('90, p. 402) says, “Of all parts of the
retina the remarkable rods and cones are the latest developed." Hoff-
mann (’83, p. 68) says, “According to all other authors [Löwe ex-
cepted] they [the external members of the rods and cones] arise latest
of all the retinal elements in the different animals; ond it is likewise so
in bony fishes.” It would certainly seem that the t imony of the eyes
of Typhlogobius is against the absolute cerreotncse se these statements.

The recent papers of Hess (89), Kohl (89), and Schlampp (’91 and ’92)
together with the somewhat older contributions to the same subject by
Leydig, Kadyi, Ciaccio, and others, make possible a detailed comparison
of the eyes of Typhlogobius with those of Proteus anguincus and Talpa
europea,

MUSEUM OF COMPARATIVE ZOOLOGY. 15

On the whole, it appears that the eye of Proteus is more rudimentary
than that of either Typhlogobius or Talpa. The most distinct indica-
tion of this is in the absence of the lens in the adult animal.

With reference to it Schlampp says (92, p. 555) : “Die Linse [in Pro-
teus] wird gleichzeitig mit der Einstülpung der primären Augenblase
angelegt, wächst in den sekundären Augenbecher hinein, wo sie bei der
Larve noch in der Gegend des vorderen Augenpoles zu finden ist. Sie
kommt aber über die zellige Struktur der embryonalen Linse nicht
hinaus, erleidet vielmehr durch Nichtgebrauch alsbald eine Rückbil-
dung, so dass sie bei ganz jungen Thieren an Grösse und Zellmasse
schon bedeutend redueirt ist, im späteren Leben aber resorbirt wird und
spurlos verschwindet.”

The lens is present in Talpa, though it retains its embryonic cellular
structure throughout life, wholly according to Hess (’89, p. 8), partly
at least according to Kohl (’89, p. 385) and others. In this regard, then,
it is more rudimentary than the lens of Typhlogobius. The choroid
is present in Proteus, consisting, according to Kohl (?89, p. 406) “aus
mebreren Zellenlagen mit reichlichem Pigment, das sich stets in zwei
Lagen anordnet, von denen bald die eine, bald die andere die gróssere
Stärke besitzt. Die innere derselben repräsentirt das vielfach (so auch
Hess) schon zur Retina gerechnete Pigmentepithel.” It also contains
blood capillaries according to both Kohl and Hess. As regards the
choroid and the pigment lamella of the retina, it would seem, according
to these statements, that the eye of Typhlogobius, with its exceedingly
rudimentary choroid and greatly thickened pigment lamella, is some-
what more rudimentary — it may be even degenerate —than that of
Proteus; though it must be borne in mind that the choroid is compara-
tively feebly developed in normal teleostean eyes.

In Talpa the choroid reaches a relatively slight development, and has
little pigment, while the pigment layer of the retina is highly developed
(Hess, ’89, pp. 3 and 4). In this regard it more nearly agrees with
Typhlogobius than does Proteus. The iris, considerably thickened with
pigment, the ligamentum pectinatum, ciliary body, and ciliary muscle, are
all present, though reduced, in Talpa, according to Hess.

With regard to the retina of Proteus, Schlampp’s statement in his
summary is as follows: “Die Retina breitet sich, Mangels des central
Glaskörperraumes, nicht flächenhaft aus, sondern wird eine solide
Kugel, welche axial von Sehnerven durchzogen wird. In ihrem histolo-
gischen Baue weicht sie nicht wesentlich von der Netzhaut der Am-
phibien ab, die Endapparate erreichen-aber die endgúltige Form nicht.”

76 BULLETIN OF THE

Also according to Kohl (’89, p. 407), the nerve-fibre layer, the gan-
glion-cell layer, the inner and outer nuclear layers, and the inner re-
ticular layer are present. Regarding the outer reticular layer and the
optic cells he says: “ Zwischen den beiden Kórnerschichten habe ich die
aüssere reticuläre Schicht (Zwischenkörnerschicht) immer durch eine
fortlaufende, oft gar nicht so schmale Spalte repräsentirt gefunden.
. . . Die Sehzellen, die sich mit Picrocarmin meist sehr schön färben
lassen, zeigen ungemein mannigfache Formen: bald ganz flach, bald
nahezu kreisrund. Oefter fand ich vollkommen entwickelte Zäpfchen,
niemals jedoch auch nur annährend stäbchenartige Gebilde. Die Hem-
mung in der Entwicklung ist eben auch hier schon so früh eingetreten,
dass eine ausgesprochene Stäbchen- und Zapfenschicht nicht mehr zur
Ausbildung kommen konnte.”

Of the retina in Talpa, Hess says that the nerve-fibre layer is very
thick near the entrance of the optic nerve, and that the inner reticu-*
lar layer contains cells; he quotes Leydig and Kadyi to the effect that
the. optic cells consist. exclusively of rods, and he adds (89, p. 5),
“Ueber die anderen Retinaschichten ist Besonderes nicht hervorzuhe-
ben.” Kohl (89, p. 384), however, states that “ Zäpfchen sind stets
vorhanden : oft vereinzelt, oft sehr zahlreich und die Stäbchen nahezu
verdrängend. Bei einem Exemplar zeigen die Sehzellen noch jene
Form, die sich bei Embryonen eines gewissen Alters findet, und noch
nicht erkennen lässt, ob die betreffenden Zellen sich zu Stäbchen oder
zu Zäpchen weiter entwickeln werden.” Itthus appears that the three
retinas have reached about the same stage in development; that of
Proteus being probably on the whole the most rudimentary, and that
of Typhlogobius, at any rate as represented by the one shown in
Figure 21, the least so.

As regards the vitreous body, Schlampp finds that, it is entirely absent
in the eye of Proteus ; while Kohl (’89, pp. 406 and 407) finds a struc-
ture which he regards as the hyaloid membrane, or “the membrana
limitans interna, the only representative of the vitreous body in the Pro-
teus eye.” Hess and Kohl both describe the vitreous body as present in
Talpa, and, according to the latter, it contains numerous blood-vessels.
It will be remembered that no trace of this structure has been found in
the eye of Typhlogobius, with possibly a single exception.

All are agreed at present, it appears, that the optic nerve is present in
both Protens and Talpa, though Hess quotes Semper as stating that it is
entirely degenerated in Talpa. I find no account, however, of its ever
having in either of these animals a pigment sheath in its passage through

MUSEUM OF COMPARATIVE ZOOLOGY. Ti
the retina, such as occurs in Typhlogobius. But it is interesting to no-
tice in this connection Kohl’s description of this portion of the nerve in
Proteus. He ('89, p. 408) writes : “ Beim Durchgang des Opticus zeigen
die Zellen der Retina ein eigenthümliches Verhalten. Ihre Kerne werden
sehr langgestreckt und sie ordnen sich um den Nerv in 1-2 dichten Lagen
dergestalt an, dass sie schon. kurz vor dem Eintritt des Opticus in die
aüssere Kórnerschicht und auf der ganzen Strecke, die derselbe sich durch
die Körnerschichten hinzieht, eine Art fester Röhre um ihre bilden."
It is quite possible that the pigment sheath described in the Typhlogo-
bius eye may have been preceded by such a cellular sheath as this ; but
if so, my conjecture that it is derived from the nerve itself, and not from
the surrounding retina, would be, of course, erroneous. It is also worth
mentioning that Berger ('81, p. 262) has described pigmented fibres
arising from the choroid as passing through the optio nerve in some
fishes.

We have not yet sufficient knowledge of the minute structure of the
eye of any of the other blind vertebrates, Myxine and its allies excepted,
to make possible further detailed comparison. With reference to the
eyes of the Myxinideo, it should be said that, from the investigations of
J. Müller (35-41), and, later, W. Müller (74, pp. 7-15), we know that
they aro far more rudimentary than in any other vertebrate whatever,
unless we admit the exceedingly problematical pigment spot at the
anterior end of the nerve cord of Amphioxus to be homologous with the
eye. It is, however, instructive to notice wherein the eyes that we have
been considering may be regarded as passing along the same degenera-
tive road over which the Myxinoid eye has passed, and in what respects
they might seem to be on different roads. The eye of Myxine is buried
in the tissue of the head in much the same way as in the other forms,
excepting that, in addition to the layers of skin and the connective
tissue by which it is covered, there is also a layer of muscle over it, and
it is immediately surrounded by a sort of capsule containing in its sub-
stance much fat. As the foregoing pages have shown, there is no indi-
cation of either the muscle layer over the eye or the fatty layer around
it in Typhlogobius, Proteus, or Talpa; but it is of course entirely beyond
our power to say that there never could be such structures.

Neither lens nor eye muscles, nor anything that can properly be re-
garded as a cornea, sclerotic, or iris, are present in the eye of Myxine.
The primitive optic vesicle never becomes wholly obliterated, and the
retina reaches only a very rudimentary degree of differentiation. Wi
Müller (74, p. 14, and Fig. 3, Taf. XI.) recognizes in it, however, the

78 BULLETIN OF THE

internal limiting membrane; the inner reticular layer (called by him
the neurospongium), and scattered in this the ganglion cells ; the inner
nuclear layer (called by him the ganglion retin); the rudiments of
the rods and cones; and the radial fibres. Krause’s remarks on the eye
of Myxine are interesting. He (86, p. 19) says: “Sein Auge würde
zu den perotischen rückgebildeten, wie das von Proteus anguineus
zu rechnen sein, und man kann die rudimentür entwickelte Retina
deshalb nicht zur Construction phylogenetisches Aufbauten benutzen."
It appears to me that the most interesting fact concerning the Myxi-
noid eye, at least from a comparative point of view, is the entire ab-
sence of pigment in it. I may here say that I have made some sections
of the eye of a member of this family found at Monterey, Cal., and
named by Lockingtou (78, p. 793) Bdellostoma stoutii, and can con-
firm the statements made on this point by all other observers. I have
so far found no trace of pigment in the eye. The proximal layer of
the primitive optic vesicle remains distinctly cellular throughout life,
as always stated, but no pigment appears either in it or in the meso-
dermal tissue immediately surrounding the eye. If, as seems certain
with the rudimentary eyes of the three forms that we have been consid-
ering, an increase of pigment is an incident. to the gradual diminution in
functional importance and structural completeness, I can see no very
satisfactory explanation for the absence of pigment in the Myxinoid eye,
if we are to suppose, as I take it for granted we must, that it too is
the result of arrested development.

Wyman (54, p. 395; '54*, p. 18; see also Putnam, ’72, pp. 18, 19)
has made us acquainted with the eye of Amblyopsis spelæus as far as he
was able to with the methods of morphological investigation of his time.
And it is altogether probable that all he has made known coucerning
this species holds good for Typhlichthys subterraneus, since the two
forms are so nearly alike that systematists are not fully agreed that
they should be considered as separate species.

According to Wyman, the eye of Amblyopsis has “a sclerotic coat,
a choroid coat, a layer resembling the retina, a lens, and a nerve.”
His notes, published by Professor Putnam, give somewhat more of de-
tail as to the structure of these several parts. He says: “ Under the
microscope, with a power of about twenty diameters, the following parts
are satisfactorily made out: . . . 2d, a layer of pigment cells for the
most part of a hexagonal form, and which were most abundant about
the anterior part of the eye; 3d, beneath the pigment a single layer

of colorless cells, larger than a pigment cell, and each cell having a
, E) I e 5 e

MUSEUM OF COMPARATIVE ZOOLOGY. 79
distinct nucleus ; 4th, just in front of the globe a lenticular-shaped,
transparent body, which consisted of an external membrane containing
numerous cells with nuclei.” The pigment layer he regards as repre-
senting the choroid, and the layer of colorless cells within — and it
should be particularly noticed that it is according to both his descrip-
tion and figures a single layer — as representing the retina. It is
very desirable, indeed, that these eyes should be studied anew with
modern methods of preservation aud by means of sections; for, if
Wyman's account of the structure proves to be correct, we have here a
most interesting deviation from the three forms best known and already
compared. It may be supposed that his statement concerning the
cellular condition of the lens is correct ; for this involves merely the ob-
servation that a given structure is composed of cells, while his state-
ment concerning the retina involves an observation as to how the cells
in a given cellular structure are disposed, — two quite different matters,
as every histologist knows. In this particular the lens of Amblyopsis
corresponds, then, to that of Talpa ; but in the latter animal the retina
is fairly well differentiated, and even in Proteus, where the lens is
wholly wanting in the adult, the retina is differentiated to a consider-
able extent. If Wyman is correct in supposing that the retina in Am-
blyopsis is represented by a single layer of cells, then we have a
condition corresponding more nearly to that found in Myxine than in
any other known vertebrate, although even here the retina proper is
far from being a single cell layer, but the eye of this latter form has no
trace of a lens,

Cope (64, p. 232) remarks with regard to the blind Silurid, Gronias
nigrilabris, that in no case has he found anything representing the lens.
Whether a considerable number of specimens were examined with refer-
ence to this point, the author does not state ; but from the general char-
acter of the fish and its eyes, as described, it appears to me quite probable
that, as Packard suggests, further examination will lead to the discovery
that the lens is not entirely absent.

I cannot refrain from saying at this point a few words on the ques-
tion which, in reality, induced me to undertake the study of the eye of
Typhlogobius, viz. the question of the actual degeneration of function-
less organs. There is a belief prevalent among zodlogists, though to
just what extent I am unable to say, that, if a structure undergoes de-
generation in ontogeny it does so in the reverse order of its phylogeny.
It would appear that a degenerating vertebrate eye with its great com-

plexity of organization, this-complexity having been taken on by degrees

80 BULLETIN OF THE

throngh a long course of evolution, would furnish an excellent test of
this belief. The eye here studied throws very little light on the question,
however, — scarcely as much as does that of other known functionless
eyes. But when we consider together the facts presented by the eyes of
Myxine, Typhlogobius, Proteus, and Talpa, and possibly also Ambly-
opsis and Gronias, this much seems quite certain: that the lens disap-
pears before the retina; and that, where degeneration takes place at all
in ontogeny, the lens is affected first and most profoundly, as seen in
Proteus, and probably also exceptionally in Typhlogobius. Supposing
the somewhat doubtful instance of a degenerating retina presented by
the eye shown in Figure 19 to be genuine, we still have reason to be-
lieve that its degeneration has been preceded by that of the lens, since
the latter body is undoubtedly absent in this specimen, There can
scarcely be a doubt, from physiological reasons, that the retina is consid-
erably older, phylogenetically, than the lens, even though it can hardly
be said to be so ontogenetically.

Tun INTEGUMENTARY SENSE ORGANS.

At present I treat this subject no further than pertains to the question
whether the loss of sight in Typhlogobius has been compensated by
an unusual development of the sense of touch, leaving the consideration
of any morphological significance that the sense papillae may have with
the hope that they may be studied developmentally at some future
time.

From the testimony of numerous writers, there is no doubt that
compensations for such loss by the super-development of the other spe-
cial senses, hearing, smell, and touch, are common among animals
both invertebrate and vertebrate. For a discussion of this subject see

’ackard (86, pp. 123-130).

My conclusion with reference to the tactile sense in Typhlogobius
is, that in all probability it not only has not increased, but has actu-
ally diminished par; passu with the diminution of the power of sight.
at least four —

The reasons for this conclusion are that several
genera of the Gobiide closely related to Typhlogobius are as well
provided with tactile papille as is the blind fish, these organs being
considerably more numerous and more widely distributed on different
parts of the body in the other fishes than in Typhlogobius. The
genera to which I refer are Gobius, Gobiodon, Lepidogobius, and Cleve-
landia. The last two I have examined mysclf. The arrangement and

MUSEUM OF COMPARATIVE ZOOLOGY. 81

distribution of the papilla of Typhlogobius are shown, except for a few
scattering ones to be spoken of shortly, in Plate I. Figure 3, pap.

There are two series near the edge of each side of the lower jaw,
running parallel with it. The series of one side do not quite unite,
anteriorly, with those of the opposite side, the interval between their
ends being occupied by a slight prominence in the epidermis. Posteriorly
the series extend beyond the angle of the mouth and turn upward some-
what to terminate about on a level with the mouth opening; the rows
nearer the median line, however, extending slightly farther than the
ones nearer the edges of the jaws (Fig. 3). The papille of the inner
series are considerably larger than those of the outer series, there
being about six of the former to thirteen of the latter. The larger
ones are on the average about 0.08 mm. in diameter, though the size
varies considerably. The papille of the outer series are situated on a
quite prominent ridge, while the inner ones are, on the contrary, in a
shallow furrow. These ridges and furrows are, however, apparently
a part of the longitudinal foldings in the integument that are charac-
teristic of this region of the head, rather than structures expressly for
the accommodation of the sense buds. Another series of papille is
found on each side of the head above the mouth, and having very
nearly the same direction as those below, though inclining toward
the latter somewhat in their course backward. These extend anteri-
orly to near the tips of the fleshy knobs shown at f, Figure 1. The
papilla above the mouth are of about the same size as the smaller ones
on the lower jaw. Still another series is found on each side of the
head on the operculum, extending however at a right angle, or nearly
so, to the series already described (Fig. 3). These are also of the
smaller variety. The number in both this and the upper-jaw series is
more variable that in the lower-jaw series, though the transverse series
never extend far on to the top of the head.

In Lepidogobius and Clevelandia both lowerjaw series are present,
and have precisely the same arrangement and form as in Typhlogobius ;
and in addition papille are numerously present on various parts of
the head and body where they do not occur in Typhlogobins. "Thus
on the head of Lepidogobius there are at least several hundred in addi-
tion to the ones on the lower jaw. On each side of the head, beginning
at a point a little above the mouth and somewhat nearer its angle
than the end of the snout, four rows take their origin and diverge
irregularly. The row nearest the mouth bends downward somewhat as,
in it& backward course, it reaches the angle of the mouth, and it extends,

VOL. XXIV. — NO. 3. 6

82 BULLETIN OF THE

a

as do all the rows, considerably farther back of the angle of the mouth
than it does in front of that point. The uppermost row of the four
runs upward to near the posterior and lower quadrant of the eye, where
it takes a trend more directly
backward, and extends for
& considerable distance back
along the dorsal limit of the
operculum. The other two
rows are situated considerab

nearer the lower than the up-

per TOW, and are nearer each

Head of Lepidogobius, showing the distribution of
the tactile Papille. x 13. other than either is to the

uppermost or the lowermost row. They also run very nearly parallel
with each other. The lower one of these two middle rows contains the
fewest and largest papille of the head, those of the inner mandibular
series excepted. There are about twenty-five papille in the lower
row, nineteen in the next, thirty-five in the third, and fourteen in
the fourth.

Many of the papilla of this species are distinctly excavated on their
summits, and in such a way as to show such an arrangement as is de-
scribed by Solger (80, p. 375) to exist in the lower jaw of Gobius
minutus. The excavations are in the form of grooves, or creases, which
extend entirely across the summit of each papilla, each groove being
somewhat broader in its middle than at the ends. In some of the rows
these grooves are directed lengthwise of the row, while in others they
have a direction crosswise of it. There is some variation in the
direction of the grooves in the papille of the same row, and consider-
ably more in some rows than in others; but the constancy in some of
them is noticeable. In the larger papille the grooves are much more
pronounced than in the smaller ones, in many of these latter the exca-
vation being a pit rather than a groove. In the lower-jaw series of this
species, the grooves of the inner rows extend crosswise to the axis
the head, and those of the outer row lengthwise, thus corresponding
the condition found by Solger in Gobius minutus.

In addition to the four series thus described, there are numbers
scattered on other portions of the head, particularly about the
regions

papillee
tip of the snout and on the opercular apparatus; in these
they are particularly numerous on the subopereulum. Also on each side
of the body, beginning immediately behind the pectoral fins, there are

about thirteen transverse series, containing from five to ten papille

MUSEUM OF COMPARATIVE ZOOLOGY. 83

each ; and still lower down are from five to seven additional transverse
series, extending well down on the ventral surface of the body. It is
possible that these lateral series are derived from a segmentally arranged
type; but if to they have certainly deviated greatly from the typical
arrangement, as they also vary both in the number of series and in the
number of papille in each series. The papille here are considerably
smaller, on the whole, than those of the head. There are, finally, a
number of papille scattered around the bases of the fins, both pectoral
and pelvic.

In Clevelandia there are about twenty-five transverse series on the
sides of the body very uniformly segmentally arranged, being situated
on the inter-myotomic septa. The series contain an average of about
five papille each, though the number varies considerably. I have
worked out the precise arrangement of the series on the septa, — for
not quite all the septa have papilla, — and of the number of papilleo
to each series in a considerable number of specimens, as it has appeared
to me that this may have considerable morphological significance. It is
not necessary for my present purpose, however, to give the results in
detail. It is worthy of mention that the transverse series on the body
of Clevelandia are situated in shallow ditches, the anterior wall of these
being deeper and more abrupt than the posterior. In no case have I
been able to find the papille situated in canals, or in grooves that
approach canals, as is so common in fishes, and is said to be the case in
lobius niger, by Merkel (’80, p. 28). All writers agree, however, that in
the genus Gobius by far the greater portion of the papillee are free on the
surface of the body (F. E. Schultze, '76 ; Merkel, 80 ; Solger, '80). We
know from tho last two of these authors that both the genera Gobius and
Gobiodon have free transverse series of papilla on the sides of the body.
With reference to this subject, Solger (80, p. 378) says: “Bei Gobius
konnte ich *Querreihen von 3-7 Organen’ constatiren ; auf Beziehung
der Organreihen zur Metamerie des Leibes achtete ich damals leider noch
nicht, Auch Gobiodon hat am Rumpfe freie Seitenorgane, die in Quer-
reihen auf Coriumpapillen stehen und höchst wahrscheinlich durchweg
segmental angeordnet sind.” And Merkel (80) shows three of these
series in his Figure 4, Taf. IV.

It being, then, evident that so many of the near relatives of Typhlo-
gobius are provided with sense papilla on the sides of the body, the
question at once arises whether any are found in the corresponding
region on the blind fish. Very naturally it was to the smallest speci-
mens in my possession that I turned to begin the search for them. On

84 BULLETIN OF THE

each of the two individuals 19 mm. long, papillae were found on the
sides of the body, and on one of them a few on the head, besides in the
regions where they occur regularly In one of these the papille on
the right side of the body were distributed in what I regard as repre-
sentatives of nine of the transverse series described in Clevelandia.
The first and second series behind the pectorals were represented by one
papilla each; the third and fourth, by three papille each; the fifth,
sixth, and seventh, by two each ; and the eighth and ninth, by one each.
The series were evidently segmentally arranged, though not all were
on consecutive segments ; thus between the third and fourth series were
two myotomie plates ; between the fourth and fifth, four plates ; between
the fifth and sixth, two plates; between the sixth and seventh, one
plate ; and between the eighth and ninth, two plates.

Figure 25, Plate IV., shows the arrangement of a group of papille
on the right side of the head of this same individual. As seen by the
figure, seven of these papille were much larger than the remaining
ones, and were situated on quite prominent ridges of the skin,

Although the papille have been diligently searched for on the sides
of the body of other specimens, they have been found on the two
small ones only. The question at once arises, Are the papillae absent
from the larger ones because they have degenerated and completely
disappeared during the life of the individual? All the evidence I
have on this point is contained in the facts presented. That the
papille have been found only on the two small specimens examined,
and that they have not been found on any of the numerous large ones,
certainly suggests very strongly an affirmative answer to the question.
It must be said, however, that a considerable percentage of my larger
specimens are not so well preserved but that the papille may possibly
have been present in them and escaped detection, But some of them
are well preserved, and were the papilleo present they would, I am sure,
have been found.

That the sense papilla are less numerous on Typhlogobius than on
several, at least, of its near allies, is evident. The question may now
be asked, Is it not possible that, although there has been no compensa-
tion for the loss of sight by an increase in number of the tactile papilla,
such a compensation has been brought about by a higher development
of the individual papille themselves? So far as structural evidence
is concerned, this is certainly far from probable. Figures 23 end 24
(Plate IV.) show sections of two papille of the inner mandibular series
of Typhlogobius; and for the purpose of comparison a section of a

MUSEUM OF COMPARATIVE ZOOLOGY, 85

papilla from the same region in Lepidogobius is given in Figure 26.
Between Figures 24 and 26 there is considerably less difference than
may often be seen between sections of different papille of the same
animal. Thus the sense cells proper (cl. sns.), which are very regular in
their arrangement in Figure 26, are quite irregular in Figure 24; but
Figure 23 agrees much more closely in this particular with Figure 26
than with Figure 24. Indeed, the difference is due to the position and
direction of the section, both arrangements being found on sections of
one and the same papilla at times. The cuticular spikes, so distinctly
seen on the sense cells in Figure 23, are much less distinet in any of
the Lepidogobius sections examined ; but they are only exceptionally
seen with such clearness in sections of the blind fish papille. It will
be noticed that a considerable space (+t.) exists in Figure 26 between the
sensory cells and the underlying nuclei of the supporting cells, in which
there are no nuclei ; and that such a space does not appear, at least con-
spicnously, in either of the figures from Typhlogobius. This, however,
is not a difference of material significance, since in many sections of
the papille of the blind fish such spaces do exist. In Figure 24, it
will be observed that a blood-vessel, va. sng. (the leader from which
has been misplaced in the engraving), penetrates far into the interior
of the papilla. A similar vessel is present in Figure 26, though it does
not extend quite so far into the base of the papilla, nor have I in any
case found it to do so in this species, though it is true 1 have not
examined as large a number of sections of Lepidogobius as of the blind
fish ; but the difference, if distinctive of the two forms, is nevertheless
insignificant. Neither as regards the mantle cells (cl. mt.), nor the rela-
tion of the papilla to the epidermis, — i. e. its extending entirely through
the thickness of it,— nor the way in which the nerve approaches and
enters the papilla, nor the character of the immediately underlying sub-
epidermal tissue, is there the slightest characteristic difference to be
made out between the two species in such papille as are represented
in Figures 24 and 26.

In only one point a difference may possibly exist between them,
though I have not yet been able fully to satisfy myself of this. By
Figure 23 it will be seen that the papilla is wholly and deeply buried in
the epidermis, only a small pore (po.) communicating with the outer
world. The apparent bridge across the pore near the papilla is prob-
ably a point of contact merely, as adjacent sections show. The whole
appearance is as though the papilla had been withdrawn into the epi-
dermis ; for not only is the latter much thickened immediately around

86 BULLETIN OF THE

the papilla, but the inverted hopper-shaped outline which the inner sur-
face of the epidermis shows immediately under the papilla in most cases
where the latter reaches out freely to the surface (as in Figures 24
and 26), is here entirely obliterated. Another fact that seems to favor
the view that the papilla has been withdrawn, is the very distinct flask-
shaped excavation in the summit of the papilla itself, seen in Figure 23
(fos.), while in the sections represented in the other two figures no such
excavations are present. A natural explanation for this would seem to
be that, on being drawn in, the middle portion of the papilla with the
sense cells had been more depressed than the mantel cells. This mey be
the true explanation, but in.one instance I have found the excavation in
the papilla, even though the papilla itself protrudes through the epider-
mis, even more distinctly than in Figures 24 and 26; yet it should bo
mentioned that in this exceptional instance the papilla is considerably
narrower in proportion to its length than those shown in the figures’
just referred to, or than they usually are. I have searched in vain for
muscle fibres that could bring about such a withdrawal, and have no
other evidence than that presented that it takes place; nor have I
often found the papillæ thus buried, and never in Lepidogobius. Leydig
(79, p. 25) has suggested the probability of the contractility of the
cellular elements of the papilla as the cause of an apparently similar
condition in Acerina cernua.

A word should perhaps be spoken at this point on the possibility
of the loss of sight being compensated by a higher development of the
organs of hearing or smell. This subject lies outside of the purpose of
the present paper, and I have given only superficial attention to it.
The ears examined in dissected specimens mounted in glycerine do not
appear unusually large. The minute structure I have not examined ;
but from this morphological evidence, taken with the fact that all my
efforts to get from my single living specimen responses to sounds of
various kinds were unavailing, I am inclined to believe that the sense
of hearing is not. largely developed.

My sections of the snout show the olfactory epithelium to be very
well developed, though apparently not more so than in other bony
fishes, and certainly not so highly as in some of the long-tailed am-
phibia that I have examined.

What we know about the compensatory development of the tactile
organs in other vertebrates with rudimentary eyes may be summed up
as follows. It is well known from the writings of Tellkampf, Wyman,
Leydig, Putnam, Wright, and others, that the tactile papille are well

MUSEUM OF COMPARATIVE ZOOLOGY. 87

developed on the head and sides of the body of both Amblyopsis and
Typhlichthys. According to Packard (’86, p. 127), Tellkampf regarded
these papilla “as without doubt increasing the tactile sense.” I have
not seen this paper of Tellkampf’s, and do not know whether his mean-
ing would be that the tactile sense is increased as compared with what
it was in the same species before it was deprived of sight, or merely
that it is great as compared with other bony fishes. Leydig also be-
lieves that the tactile organs perform such a compensatory office (783;
see also Wright, '84, p. 272). Packard (’86, pp. 127, 128) gives ex-
tracts from several letters of Dr. John Sloan that are interesting in
this connection. Although the writer does not expressly state his belief
that the sense of touch has been highly developed for the purpose of
compensating the lack of sight, he still gives very convincing evidence
of its extreme acuteness from personal observation on the fishes in their
native surroundings. It should also be noticed that he specially tested
their powers of hearing and the effect of light upon them, and to both
he says they “ manifested total indifference.” — Sloan's observations were
on Amblyopsis. Wyman (’72, p. 19) has described the ear of this spe-
cies as being “largely developed” in all its parts, and Cope (72, p. 410)
found the sense of hearing “evidently very acute.” As to the ques-
tion whether the sense papille in Amblyopsis and Typhlichthys are in
reality developed as a compensation for the loss of sight, the testimony
furnished by Chologaster is of the greatest importance. Although this
genus was discovered and named by L. Agassiz in 1843, its characters
were best made known by Putnam. He (’72, pp. 22, 23) says: “In the
genus Chologaster we have all the family characters as well expressed
as in the blind species, though it differs from Amblyopsis and Typhlich-
thys by the presence of eyes, and the absence of papillary ridges on
the head and body, and by the longer intestine and double the number
of pyloric appendages, as well as by the position of the ovary.”

In 1881, S. A. Forbes (82, p. 3) discovered a fish in Southern Tllinois
which he identified as belonging to the genus Chologaster, but repre-
senting a new species. With reference to the point that we are now
considering, the author writes: “'The most important and interesting
peculiarity of this species indicates a more advanced stage of adap-
tation to a subterranean life than that of its congeners. On all the
surfaces of the head appear short rows of peculiar tubercles. . . .
When thus exposed [by being freed from the adjacent epidermis], they
closely resemble the papille of Amblyopsis in form and size, and are
similarly cupped at the tip." Again (p. 5) he says: “The extraor-

88 BULLETIN OF THE

dinary development, in only a part of the genus, of a special sensory
apparatus peculiarly useful to a fish unable, for any cause, to see,
points the same way, [i e. to the supposition that this genus has a
shorter subterranean history than Amblyopsis,] and gives evidence of a
progressing adaptation of these fishes to their unusual abode. The in-
termediate relation of the sensory tubercles of Chologaster to the much
smaller ones of young fishes and the permanent papille of Amblyopsis,
points out the evident origin of the last through the permanency
and higher evolution of structures evanescent in the young.” This is
probably the clearest case furnished by vertebrates of the loss of sight
being recompensed by a higher development of the tactile sense.

As regards the tactile papilla in the Cuban blind fish (Lucifuga), Put-
nam (272, p. 9), who examined a specimen sent to the Museum of Com-
parative Zoólogy by the discoverer of the fish, Professor Poey, says: “In
the Cuban blind fish we find ciliary appendages on the head and body `
quite distinctly developed, evidently of the same character as those of
Amblyopsis, and answering the purpose of tactile organs. . . . There are
eight of these on the top of the head, . . . and quite a number arranged
in three rows on each side of the body, showing that the tactile sense
is well developed in these fish."

This, so far as I am aware, is all that is known on the subject, and
can be regarded as furnishing nothing more than a probability that
touch papillas have been here developed to compensate the fish for sight-
less eyes. The writer just quoted remarks further, that it is singular
that the barbels on the jaws, so commonly found in the Cod family and
its allies (to the latter of which the Cuban fish belongs), are entirely
wanting. As is well known, Lucifuga is a cave dweller, and consequently
the conditions which have produced its rudimentary eyes are more
similar to those that have produced the corresponding change in Am-
blyopsis than to those that have had the same effect on Typhlogobius.
And this fact may strengthen the probability above referred to; for,
from the difference in conditions of life, Amblyopsis and Lucifuga are in

rj

all probability much more active than Typhlogobius, and this would
make the tactile sense more useful to the first two species than to the
last one.

We will now notice the condition of the blind deep-sea fishes with
reference to the touch papille. The three forms described by Gün-
ther (80), Typhlonus nasus, Aphyonus gelatinosus (p. 548), and
Ipnops murrayi (p. 585), are all without barbels, and, so far as

known, other special tactile structures. The two genera first named
)

MUSEUM OF COMPARATIVE ZOOLOGY, 89

belong to the same family as Lucifuga, and consequently the remarks
made concerning the absence of barbels in the latter will apply in a
measure to these genera, and with reference to Lucifuga receives more
force by the statement of Günther that they “are Brotula organized
for a subterranean life” (p. 547); and in the genus Brotula, which
has eyes, the snout is provided with barbels.

Through the kindness of Mr. C. H. Townsend, naturalist of the
United States Fish Commission Steamer “ Albatross,” I have been
able to examine, though somewhat superficially, a specimen of Ipnops, —
probably the same species as the one above mentioned, — and, so
far as I could discover, Günther's statement that it is “ deprived of
organs of sight and touch” (’87, p. 190) is strictly correct. The same
author makes the following as a general statement on this subject:
“Special organs of touch are not more generally developed in deep-sea
fishes than in the littoral fauna. . . . As such may be considered . . .
the more or less detached rays of the pectoral fin of . . ., and especially
of Bathypterois, which possesses but rudimentary eyes.” (87, p. xxxi.)
And in another connection the same author (p. 722) says: “ Beyond that
depth [two hundred fathoms] small-eyed as well as large-eyed fishes
occur; the former having the want of vision compensated by tentacular
organs of touch, whilst the latter have no such accessory organs.”

I have not been able to find any direct statements concerning tactile
papille on the several species of blind Silurids of South America men-
tioned by Günther (Packard, ’86, p. 107), nor have we any knowledge
that such structures are found on Gronias, the blind representative of

the same family from Pennsylvania.

Tue INTEGUMENT.

I was led to a study of the integument by the question having arisen
as to why the quantity of pigment should have diminished in it, while
under the same conditions of life it had ¿ncreased in the eyes. That
such diminution had taken place in the skin was inferred from the
generally much lighter appearance of the largest preserved specimens
as compared with the smallest. In the latter, the whole dorsal portion
of the body and head is covered with a great number of distinct pig-
ment cells (Fig. 1), while the large specimens never present anything
like so conspicuous a pigmentation ; and in the majority of cases they
appear, on cursory observation, to be almost white.

Closer examination shows, however, that the pigment is in reality

90 BULLETIN OF THE

present in large specimens as well as in small ones, in quantity almost, if
not quite, as great in the one case as in the other; but that it becomes
disguised in the former case, and in a manner that will appear presently.

Figures 9 and 10, Plate II., are from sections of the integument
and immediately underlying tissues of the dorsum of the head of two
individuals, 19 mm. and 72 mm. long respectively. It will be noticed
that the epidermis ‘does not differ essentially in thickness or structure.
In Figure 9 it is about 0.028 mm., while in Figure 10 it is slightly
thinner, though in reality it should be a little thicker, some shrinkage
having taken place here. With the sub-epidermal connective tissue
(con’t. tis.), however, the case is quite different. In Figure 9 its thick-
ness is 0.056 mm., while in Figure 10 it is about 0.025 mm., over
the pigment; and it will be seen that in both the pigment (pig.) is
situated in the deepest stratum of the connective tissue, adjacent, or
very nearly so, to the muscles (mx.).

Within the sub-epidermal layer, in all the larger specimens, there is
found a dense and intricate network of blood-vessels and capillaries.
In general, the vessels of this network appear to run quite uniformly in
one plane, situated about midway between the epidermis and the pig-
ment layer. In many places, however, they will be seen, in sections, to
approach very close to the epidermis, or at least to its basement mem-
brane, which is at times quite distinct. These vessels are shown in
Figure 10 (va. sng.) ; also in Figures 4 and 11 as surface views from
nitric acid glycerine preparations. Their walls are so thin as to be
scarcely distinguishable from the surrounding connective tissue. In-
deed, in parts of many sections their presence can be detected only by
the blood-corpuscles, which are very different in appearance from the
connective-tissue cells, owing to their larger size, more elliptical outline,
distincter nuclei, and slightly yellowish homogeneous non-stainable (in
heematoxyline at least) cell protoplasm. The connective-tissue layer
in which these vessels are situated is somewhat different from cither
the layer above or that below it: its fibres are more closely compacted
together, it contains more cellular elements, and it takes stain rather
more readily than do the adjacent layers. I am inclined to believe
that many of these cellular elements are leucocytes. I may here add
that the blood-vessels shown in Figures 4 and 11 are none of them
of capillary fineness, since in none of them are the blood corpuscles
arranged in a single row, as is characteristic in capillaries. The cap-
illaries are still smaller, and from the method of preparation and

delineation are not shown here.

MUSEUM OF COMPARATIVE ZOOLOGY. 91

It is unquestionably to the presence of this highly developed vascular
network that the pink color of the living fishes is due; and it is un-
doubtedly by this, in part, but mostly by the much thickened sub-
epithelial connective tissue, that the pigment is disguised in the
preserved specimens."

And now as to the reason for this highly vascular condition of the
skin, which is certainly unusual, as I have convinced myself by exam-
ining the integument of several other bony fishes, both by sections and
by the same methods of treatment that were used in preparing the
specimens shown in Figures 4 and ll.

I will consider the several explanations that have suggested them-
selves, in the order in which they have occurred to me,

When I first saw the living specimens, I supposed their pink color to
be due to the fact that the pigment had disappeared from the skin on
account of the constant darkness in which the fishes live; and that, it
having thus become somewhat translucent, no scales whatever being
present, whatever of vascularity there might be in the tissues of the
body wall became visible through the integument, This explanation
lost all its foree, of course, as soon as it was noticed that the pigment
is present in large as well as in small specimens, and that the blood-
vessels are situated between the pigment layer and the epidermis, and
not under the former. I would not be understood to mean by this that
the pigment layer is so dense that it would much obscure the vascular-
layer were it superficial to the latter.

The next hypothesis that presented itself to me was suggested by the
fact mentioned by Dr. Eigenmann, that the crustacean with which the
fishes so constantly live is also of the same pink color, Have we here
a case of protective resemblance? An entirely satisfactory answer to
this question cannot be given until we know more of the habits of the
fish in its native conditions of life, and also of the structure and habits
of the crustacean in company with which it lives. So far, however, as
our present knowledge enables us to see, there are some quite serious
obstacles in the way of this supposition, It is probable that the fish

1 I may add, that on examining several large specimens preserved in alcohol
exclusively, I find that the pigment is very distinctly seen on the whole dorsal
surface, without removing the skin. As the epidermis in these specimens is
quite loose as compared with that of specimens preserved in picro-nitric, picro-
sulphuric, or Perenyi’s fluid, 1 explain the greater distinctness of the pigment
by supposing that in the alcoholic specimens the sub-epithelial connective tissue
has shrunken more by dehydration than it has in the other methods of fixation,

and also more than has the epidermis.

92 BULLETIN OF THE

has adopted its present mode of life as a means of escape from its eno-
mies, so that protective coloration could be of no use, and consequently
natural selection would have no power to establish the color,

It is true that there are recorded a few cases of animals that are
apparently protectively colored, which at the same time depend upon
concealment to escape their enemies ; e. g. the caterpillar of the moth
Mania typica (A. R. Wallace, ’89). But these are certainly very ex-
ceptional ; and if the law of natural selection is to be held as applying
to them at all, we are compelled to assume that either the coloration
must have been produced in some other way than through it, or that
neither the color protection nor the concealment is adequate in itself
to effect the degree of protection necessary for the preservation of the
species. On this supposition, it is possible that natural selection has
been operative in producing the color; but Dr. Eigenmann (90, p. 68)
tells us that Typhlogobius “never leaves its subterranean abode ”: and
the extent to which the eyes are reduced affords very Surong proof in
confirmation of this statement. Again, on physiological grounds, it
would seem that had the color been produced for the mere purpose of
the color alone, it would have been effected by a deposition of pig-
ment, and not by such an enormous increase in the quantity of blood-
vessels and blood; for certainly the former would have been more
economical,

And this brings me to what I believe to be the true explanation of
the condition. I believe it to be for the purpose of cutancous respira-
tion. Says Prof. N. Zuntz (82, p. 114): “Wo auch immer das Blut
mit der Atmosphire oder mit gashaltigem Wasser in Contact kommt,
muss, in derselben Weise wie in der Lunge, ein auf Ausgleich etwaiger
Spannungsänderungen hinzielender Diffusionsstrom der Gase auftreten.”
The conditions for such a diffusion seem to be present here. That
cutaneous respiration takes place as a normal process in many vertebrates,
both terrestrial and aquatic, is generally admitted by physiologists. For
the present purpose I need only to consider in some detail what is known
about the process in some of the aquatic forms.

Spallanzani (1803, pp. 71, 114) was the first to show that the frog by
means of respiration through the skin continues to live for a long time
in air after the extirpation of the lungs. W. F. Edwards (24, pp. 41-
62) confirmed Spallanzani's results, and added the observation that this

1 For a discussion of this question, see the larger works on physiology, and par-
ticularly Milne-Edwards, ’57, pp. 632-635, and Hermann, ’82, pp. 114-117.

MUSEUM OF COMPARATIVE ZOOLOGY. 93

animal would continue to live in water as well as in air, particularly in
flowing water at a low temperature.

Berg (68) investigated the same subject, and, while confirming the
results of his predecessors so far as the fact of cutaneous respiration is
concerned, concluded that the quantity of carbonic acid gas exhaled is
less than that found by Regnault and Reiset. Less attention appears to
have been given to the subject of cutaneous respiration in fishes than to
the same process in amphibians and mammals ; though Spallanzani, and
later Humboldt and Provencal (11, p. 86), found it to occur in these
animals to a slight extent.

Quincaud (’73, p. 1143) found that an eel of 530 grams’ weight ab-
sorbs 0.58 c.c. of oxygen in an hour through the skin,

With this attempted explanation of the color of Typhlogobius the
question at once arises, Is this color peculiar to this fish, or is it com-
mon to all others that live habitually excluded from the light as this one
does? If all the other blind fishes have the same color, and from the
game cause, viz. from the vascularity of the integument, then we should
have to suppose the same explanation to apply to all; and this would
diminish its probability, though of course it would not necessarily
wholly invalidate it. In speaking of the color of blind fishes, Professor
Putnam (’72, p. 8) gives a list of seven partially or wholly sightless gene
era of the family Siluridæ, found in various parts of South America,
Africa, and Asia. Of their color he says: “ All the other members of
this family [Siluridæ] having rudimentary or covered eyes are also dark
colored ; while the blind fish of the Mammoth Cave and of the caves
of Cuba are nearly colorless.” Concerning the color of Gronias nigri-
labris, already mentioned in other connections, Cope (64, p. 232) says:
“The color of the upper surfaces, tail, fins, barbels, and under jaw is
black ; sides varied with dirty yellow, abdomen and thorax yellowish
white.” And this author remarks in the same connection that the
“dark pigment of the skin of this animal comes off upon the hands in
handling it.”

Concerning the color of the several species of the three blind, or nearly
blind, groups of the Gobiide other than Typhlogobius, I gather the fol-
lowing from Günther (61, pp. 133-138).

In the characterization of the group Amblyopina the eyes are spoken
of as “very small, and more or less hidden.” No mention is made in
this connection of the color, though the name Amblyopus roseus (Cuv.
and Val., XII. 164), as applied to the whole genus Amblyopus, is given
ina foot-note. Of the eight species enumerated one is said to have “eyes

94 BULLETIN OF THE

inconspicuous,” color “greenish olive (in spirits)”; the color of another
is “ greenish,” no mention of the eye ; another is “ rose-colored,” no men-
tion of the eye, nor statement as to whether this is the color in life or in
spirits ; a fourth is “ brownish with darker spots, . . . eye small and indis-
tinct” ; another, “eyes invisible,” no mention of color. Of the remaining
three, no mention is made of either the eyes or the color, but for the
name in one species rubicundus is given as a synonym. Of the genus
Trypauchen two species are described, one of which is characterized as
“reddish (during life), brownish (in spirits),” the other as * uniform rose-
colored." No mention is here made of the condition of the eyes, and I
know them to be rudimentary only by the list of blind fishes given by
the same author (Packard, ’86, p. 107).

In the characterization of the genus Trypauchenichthys the eyes are
”

said to be “very small, scarcely visible,” and the only species described

is “ rose-colored (Bl.).” Nothing is given to indicate that these fishes
live particularly excluded from the light. The genus Amblyopus is said
to be “confined to the coasts, estuaries, and fresh waters of the Fast
Indies, extending northward to China and Japan ; one species from the
west coast of South America.” The genus Trypauchen is from the “ Kast
Indian Seas” and the “fresh waters of Borneo,” and Trypauchenich-
thys is from “ rivers of Borneo.”

We are not informed whether the several shades of red here men-
tioned are due to pigmentation; but from the facts that there are
several shades, that in some of the species the color seems to persist in
the alcoholic specimens, and that the fishes come in a category many of
which — particularly of the related genus Eleotris, with eyes normally
developed, inhabiting much the same regions — are of similar shades of
color, it appears probable that such is the case.

Perhaps the most interest attaches to the color of the Mammoth Cave
blind fishes and those of the caves of Cuba; for these are without any
question completely deprived of the influence of light. Cope (’72, p. 410)
speaks of Amblyopsis speleus as swimming “in full sight like white
aquatic ghosts" ; in his original description of Typhlichthys subter-
raneus, Girard (59, p. 63) gives its color as a “uniform dull yellowish
white tint" ; and both these species as well as the Lucifuga are referred
to by Putnam as being “nearly colorless,” as already mentioned.- Also
Jordan and Gilbert (82) describe both Amblyopsis and Typhlichthys
as “ colorless,” and in the same way Günther ('80, p. 618), who re-
gards the two as belonging to the same genus, speaks of the body as
colorless.

MUSEUM OF COMPARATIVE ZOOLOGY. 95

I have dwelt thus at length on this question of color in other blind
fishes because Eigenmann (’90, p. 68) has said with reference to the
color of Typhlogobius that “in its pink color and general appearance
this fish much resembles the blind fishes inhabiting the caves of South-
ern Indiana.” I suppose this to refer to Amblyopsis, as there is not to
my knowledge any other blind fish known from the caves of this region.
Whether Eigenmann’s statement about the color of the Indiana fishes
is to be taken as opposed to those quoted from other writers or not, the
most significant fact for our purpose is that there is certainly no such
degree of vascularity in the integument of Amblyopsis as is found in
Typhlogobius. I have had opportunity to examine a well preserved alco-
holic specimen of this species, obtained by Professor Mark from Professor
Putnam. I prepared fragments of the skin in the same way that had been
employed in studying that of Typhlogobius, and found the blood-vessels
here to be even less abundant than in the integument of the Clevelandia
and Lepidogobius that I have examined.

The most serious objection, I think, to the supposed respiratory func-
tion of the skin lies in the thickness and density of the epidermis, and
the fact that the entire surface is thickly beset with the slime-secreting
cells (see Figs. 9, 10, and 17). I do not believe, however, that the
epidermis here would offer greater resistance to the interchange of gases
than would that of the frog ; certainly, as regards the integumentary
glands and their products, the frog’s skin can hardly be more favorably
constructed for a respiratory function than that of the blind fish. When
we remember the dense cuticular layer that covers the entire surface of
such animals as the earth-worm, where all the respiration must be carried
on through the body wall, this obstacle does not seem so great. More-
over, in Cobitis fossils, where intestinal respiration is well known to take
place to a considerable extent, although it was long supposed that no epi-
thelium was present in the region of the intestine, — in which from the
richness of the blood-vessels the respiration is supposed to be carried
on, — Lorent has shown not only that there is an epithelium present, but
that it consists of two layers, a superficial layer of flat polyhedral cells,
and beneath this a layer of stratified cylindrical epithelial cells, among
which are scattered beaker cells (Wiedersheim, ’86, p. 572).

Of course the ultimate test of my theory must be made by physio-
logical experimentation, and I hope to be able to do this before long. I
annot suppose gill respiration to be to any great extent supplanted by
integumentary respiration, since the gills appear to be normally devel-

oped. It is necessary, then, to suppose that the latter method supple-

96 BULLETIN OF THE

ments the former ; and this may have become necessary from the peculiar
mode of life of the animals. It is quite certain that the water of the
small holes under stones, in which they live, would contain less aerating
oxygen than would that of the open sea; and consequently a greater
absorbing surface would be essential in order to effect a normal aeration
of the blood.

SUMMARY.

The facts observed and the conclusions reached may be summed up
as follows.
The Eyes.
1. In the smallest examples studied the eyes, though very small, are
so distinetly that the

distinctly visible even in preserved specimens,
lens is plainly seen. In the largest examples, on the other hand, they
are so deeply buried in the tissue as to appear even in the living animals
as mere black specks, while in preserved ones they are in many cases
wholly invisible.

2. Neither in small nor in large specimens does the epidermis over the
eye differ in thickness or structure from that of adjacent regions. In
the large individuals the much greater thickness of the tissue here is
brought about by an increase in the sub-epidermal connective tissue, the
growth of which can be seen taking place in the embryonal connective-
tissue cells that are found here.

3. As is the case with rudimentary organs generally, the eye is subject
to great individual variation in size, form, and degree of differentiation.

4. The only parts of the normal teleostean eye no traces of which have
been found are the argentea, the lamina suprachoroidea, the processus
falciformis, the cones of the retina, the vitreous body proper, the lens
capsule, and in one specimen the lens itself.

5. In the parts present the rudimentary-condition of the organ is seen
in the very slight development of the choroid, no cellular elements being
present in this excepting in the chorio-capillaris, and here to a quite
limited extent, the rest of that layer being composed exclusively of pig-
ment; in the fact that the choroid gland is composed entirely of pig-
ment; in the fact that the iris, though of fully the normal thickness,
is almost entirely of pigment, there being on its outer surface in some
specimens a small amount of cellular material, which probably represents
the ligamentum annulare ; in the great proportional thickness of the pig-
ment layer of the retina and the entire absence in it of anything except-

MUSEUM OF COMPARATIVE ZOOLOGY. 97

ing pigment ; in the incomplete differentiation of the layers of the retina,
there being in some individuals scarcely more than a trace of the external
reticular layer separating the two nuclear layers, and there being in no
specimen studied a retina sufficiently developed to enable one to homol-
ogize with certainty the layers marked out ; in the minute size of the
optic nerve, and: the fact that it is ensheathed in a thick layer of pig-
ment for nearly its entire course through the retina ; and, finally, in the
small size of the motores oculi.

6. The surest evidences of actual degeneration are found, first, in the
greatly augmented quantity of pigment in all the portions that are at all
pigmented in the normal eye; and, secondly, in the presence of pigment
in regions where none is found in the normal eye, as in the hyaloid
membrane (Plate IT. Fig. 12, pig.").

No undoubted instances of degenaration through the breaking down and
dissolution of the tissue without the formation of pigment, such as have
been described particularly by Looss, have been found, though in a single
specimen (the one in which no lens is present) a process of this nature
may be taking place.

7. On comparing the eyes of all blind vertebrates that have been
most carefully studied, we find that, in the several degrees of incom-
pleteness of development represented by the different species, all may,
in a general way, be said to be passing along the same degenerative road.
There are apparently, however, a few interesting exceptions to this. The
most marked of these exceptions is found in the entire lack of pigment
in the eyes of the Myxinideo, whereas in all other rudimentary eyes an
increase of this substance over what exists in normal eyes is found.

8. The eyes of blind vertebrates furnish very little evidence on the
question whether structures in undergoing actual degeneration in on-
togeny follow the reverse order of their phylogeny. The little that may
be regarded as bearing on this point is without much doubt of an affirma-
tive character. This is found in the breaking down and resorption of the
lens, — habitually in Proteus, and probably occasionally in Typhlogo-
bius, — possibly in the excess of pigment in the iris and pigment layer
of the retina, and particularly in its occasional presence in the hyaloid
membrane of the Typhlogobius eye, while no evidence of actual degen-
eration in the retina appears in connection with these. The possible case
of a degenerating retina in Typhlogobius is neglected in this considera-
tion, since, as pointed out, the lens being absent in the same eye, it is

immaterial whether it be considered or not,

VOL. XXIV. — NO 3. i

BULLETIN THE

OF

The Integumentary Sense Pupille.

1. These have been considered only so far as pertains to the question
whether they have been developed to compensate the rudimentary con-
dition of the eyes ; and it is concluded that such is not the case.

2. The facts that lead to this conclusion are the presence in several
closely related genera — four at least — of the tactile papille with the
same distribution as those of Typhlogobius, and in addition to this, on
parts of the body where they are not found at all in Typhlogobius, ex-
cepting in the smallest specimens ; and that the papille that are present
in Typhlogobius are not more highly developed than those of correspond-
ing regions in related genera.

3. In comparing the several species of blind fishes with a view to de-
termining under what conditions the tactile sense does become developed
to compensate the loss of sight, it is concluded that, while the greater
activity of the cave blind fishes might explain their more highly developed
tactile papille, this cannot be affirmed as a general law, since other
blind fishes (as some at least of the deep-sea forms and probably also
the blind Silurids) are without tactile papille, while we have no rea-
son to suppose them less active than the cave fishes. It is necessary to
have more knowledge than is yet possessed of the mode of life of the
various blind forms before this question can be fully answered.

The Integument.

1. This structure has been studied with reference to the pigment con-
tained in it, and the pink color of the living fishes.

2. Very nearly if not fully as much pigment is present in the largest
as in the smallest specimens, the lighter color of the former being due
to the obscuration of the pigment by a thickening of the sub-epidermal
tissue between the pigment and the epidermis.

3. The pink color of the living animals is due, in great part at least,
to a highly abnormal development of blood-vessels in the sub-epidermal
portion of the integument.

4. So far as it has been possible to determine, this vascularity of the
skin is unique in this fish.

5. The most probable explanation found of this condition is that it is

for the purpose of cutaneous respiration,

UNIVERSITY OF CALIFORNIA, BERKELEY,
July 4, 1892.

MUSEUM OF COMPARATIVE ZOOLOGY, 99

BIBLIOGRAPHY.

Balfour, F. M.
"81. A Treatise on Comparative Embryology. Vol. II. London.

Balfour and Parker.
'82. On the Structure and Development of Lepidosteus. Phil, Trans.
Roy. Soc. London, Vol. CUXXILL p. 359.

Berg, W.
"68. Untersuchungen über die Hautathmung des Frosches. Dorpat.

Berger, E.

"81. Beiträge zur Anatomie des Fischauges. Zool. Anzeiger, Jahrg, IV.

p. 258.
Ciaccio, G. V.

75. Osservazioni intorno alla membrana del Descemet e al suo endotelio
con una descrizione anatomica dell’ occhio della Talpa europea. Mem.
Accad. Sci. Istituto di Bologna. Ser. 3, Tom. V. p. 501.

Cope, E. D.
'64. On a Blind Silurid from Pennsylvania. Proc. Acad. Nat. Sci. Philad.,
Vol. XVI. p. 231.
"72. On the Wyandotte Cave and its Fauna. Amer. Nat., Vol. VI. p. 406.
Edwards, Milne.

See Milne-Edwards.
Edwards, W. F.

"24. De l'influence des agents p! ysiques sur la vie (pp. 12 et 41-62). Paris.
Eigenmann, Carl H.

'90. The Point Loma Blind Fish and its Relations. Zoe, Vol. T. p. 65.

Eigenmann, Rosa Smith.
"90. Note on Typhlogobius Californiensis. Zoe, Vol. I. p. 181.
Forbes, S. A.
'82. The Blind Cave Fishes and their Allies. Amer. Nat., Vol. XVI.
p. li

Girard, Charles.
"59. Ichthyological Notices [XIIL.]. Proc. Acad. Nat. Sci. Philad., Vol. XI.
p. 63.

106 BULLETIN OF THE

Günther, Albert C. L. G.

’61. Catalogue of the Fishes of the British Museum. Vol. III. London.

'80. An Introduction to the Study of Fishes. Edinburgh.

87. Report on the Deep Sea Fishes. Challenger Reports, Vol. XXII.

“ Ichthyology.” Encyclopædia Britannica, Vol. XII. p. 722.

Hermann, L.
182. Handbuch der Physiologie. Bd. IV. Theil II.
Hertwig, O.

’90. Lehrbuch der Entwicklungsgeschichte des Menschen und der Wirbel-

thiere, Jena.
Hess, C.

"89. Beschreibung des Auges von Talpa europea und Proteus anguincus.

Arch. f. Ophthalmologie, Bd. XXXV. p. 1.
Hoffmann, C. K.
’83. Zur Ontogenie der Knochenfische. Arch. f. mikr. Anat., Bd. XXIII.
p. 45.
Humboldt und Provençal.
See Provençal und Von Humboldt,
Jordan and Gilbert.

82. Synopsis of the Fishes of North America. Bull. No. 16 U. S. Na-

tional Museum, Washington.
Kadyi, H.

78. Ueber das Auge des Maulwurfs (Talpa europea) in vergleichend-ana-
tomischer Hinsicht. (Polish.) Denkschr. Akad. d. Wiss. Krakau, Math.-
naturhist. Section, Bd. IV. p. 124.

Kessler, Leonhard.
77. Zur Entwickelung des Auges der Wirbelthiere. Leipzig.
Kohl, C.

89. Einige Notizen über das Auge von Talpa europa und Proteus angui-

neus. Zool. Auzeiger, Bd. XIL pp. 383, 405.
Krause, W.

"86. Die Retina. II. Die Retina der Fische. Internat. Monatschr. f. Anat.

u. Histol., Bd. III. p. 8.
Leydig, F.

79, Neue Beiträge zur anatomischen Kenntniss der Hautdecke und Haut-
sinnesorgane der Fische. Festschrift der Naturforschenden Gesellschaft m
Halle a. S., p. 129.

'83. Untersuchungen zur Anatomie und Histologie der Thiere. Bonn.

Lockington, W.N.
78. Walks around San Francisco. Amer. Nat., Vol. XII. p. 786.
Merkel, Fr.
"80. Ueber die Endigungen der sensiblen Nerven in der Haut der Wirbel-

thiere. Rostock.

MUSEUM OF COMPARATIVE ZOOLOGY. 101

Milne-Edwards.
'57. Leçons sur la Physiologie, Tom. IL. p. 632. Paris.
Müller, H.
57. Anatomisch-physiologische Untersuchungen über die Retina bei Men-
schen und Wirbelthiere. Zeitschr. f. wiss. Zool., Bd. VIII. p. 1.

Müller, Johannes.
'85-41. Vergleichende Anatomie der Myxinoiden. Abhandl. Akad. Wiss.
zu Berlin. 1834.
Müller, W.
74. Ueber die Stammesentwicklung des Schorganes der Wirbelthiere. Bei-
träge zur Anatomie und Physiologie als Festgabe Carl Ludwig gewidmet.
Heft IT. Leipzig.
Packard, A.S.
786. The Cave Fauna of North America, with Remarks on the Anatomy of
the Brain and Origin of the Blind Species. Mem. National Acad. of
Boin Vol. IV.

Provencal und Von Humboldt.
711. Untersuchungen über die Respiration der Fische. Jour. f. Chem. u.
Physik (Sehweigger), Bd. I. p. 86.
Putnam, F. W.
^72. The Blind Fishes of the Mammoth Cave, and their Allies. Amor. Nat.,
Vol. VI. p. 6.
Quincaud.
73. Expériences relatives à la respiration des poissons. Compt. Rend.
Acad. Sci. Paris, Tom. LXXVI. p. 1141.
Ryder, J. A.
'84. A Contribution to the Embryology of Osscous Fishes, with special
Reference to the Development of the Cod (Gadus morrhua). Ann. Re-
port U. S. Com. Fish and Fisheries for 1882, p. 455.
Schlampp, K. W.
'91. Die Augenlinse des Proteus anguineus. Biolog. Centralb., Bd. XI.
Nr. 2, p. 40.
192. Das Auge des Grottenolmes (Proteus anguineus). Zeitschr. f. wiss.
Zool., Bd. LIII. p. 537.
Schultze, M.
'66. Zur Anatomie und Physiologie der Retina. Arch. f. mikr. Anat.
Bd. IL p. 175.
Schulze, F. E.
76. Ucber die Sinnesorgane der Seitenlinie bei Fischen und Amphibien.
Arch. f. mikr. Anat., Bd. VI. p. 62.
Semper, Carl.
'80. Die natürlichen Existenzbedingungen der Thiere, Erster Theil.

102 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

Smith, Rosa.

"81. Description of a new Gobioid Fish (Othonops cos) from San Diego,
California. Proc. U. S. National Museum, 1881, Vol. IV. p. 19; also,
Smithsonian Miscellaneous Collections, Vo! XXII. p. 19.

'90. See Eigenmann, Rosa Smith, ’90.

Solger, B.

^80. Neue Untersuchusgen zur Anatomie der Seitenorgane der Fische.
III. Die Seitenorgane der Knochenfische. Arch. f. mikr. Anat., Bd.
XVII. p. 364.

Spallanzani.
103. Trois mémoires sur la respiration, trad. par J. Senebier. Geneva,
1803.

Steindachner, F.
"79. Ichthyologische Beiträge (VIIT.). Sitzungsb. Wien. Akad. Math.-
naturw. Cl., Bd. LXXX. Abth. 1, p. 119.
Wallace, A. R.
89. Darwinism. London.
Wiedersheim, Robert.
"86. Lehrbuch der Vergleichenden Anatomie der Wirbelthicre. 3te Aufl.
Jena.
Wilson, H. V.
91, The Embryology of the Sea Bass (Serranus atrarius). Bull. U. S. Fish
Commission, Vol. IX.
Wright, R. R.
84. Leydig's Researches in Anatomy and Histology. Amer. Nat., Vol.
XVIII. p. 272.
Wyman, J.
54, [The Eyes of Amblyopsis speleus.] Proc. Bost. Soc. Nat. Hist,
Vc. IV. (1851-54), p. 395.
154, [On Amblyopsis speleus.] Proc. Bost. Soc. Nat. Hist, Vol. V.
(1854-56), p. 18.
272. The Blind Fishes of the Mammoth Cave, and their Allies. Amer. Nat.,
Vol. VI. p. 6. (See Putnam.)
Zuntz, W.
82, Handbuch der Physiologie (Hermann). Bd. IV. Theil 2, p. 144.

EXPLANATION OF PLATES.

All the figures except 20 and 26 are drawn from Typhlogobius californiensis,
Steind.

bac.
bac. cla.

chr.
chr.-cpl.

Cl COLE

cl. cont,

cl. sns.
con’t. tis.
cp. sng.
ern.

ort.

drm.
ec’drm.
ex.

for.
Jos.
Jul.
gib. pig.

gl. chr.
gl. muc.
n

ir.
lewcy.
lig. ann.

Ins.
m.

mb. lin. ex.

Rods of the retina.

Knobbed or club-shaped
rods.

Choroid.

Chorio-capillaris.

Refractive cells of retina
as they appear at dif-
ferent foci.

Cells in connective tissue.

Ganglion cells.

Ganglion cells showing pro-
cesses.

Mantle cells.

Sense cells.

Connective tissue.

Blood corpuscle.

Cornea.

Cartilaginous
sclera.

Derma.

Ectoderm.

Outermost portion of pig-
ment layer of retina.

Miiller’s fibres.

Flask-shaped pit in papilla.

Tangential fulcrum cells.

Pigment nodules, clusters,
balls, etc.

Choroid gland.

Mucus glands.

Innermost ends of the pro-
cesses of pigment layer
of retina fused together.

Tris.

Leucocy tes in retina.

Non-pigmented elements of
the iris, ligamentum an-
nulare.

Lens.

Middle portion of the pro-
cesses not fused together.

Membrana limitans ex-
terna.

Muscle.

portion of

ABBREVIATIONS.

nl. ba.
n. opt.
n. pap.
ob. ex.
or. ser.
pap.
pig.
pig.
pig”

pig”

pig. rin.
po.

pre.

rt.

rtn.

scl.

set. cl. sns.

spa.

spng-bl.

st, bac. con.

sl. cont.

st. fbr. opt.
st. lac.

st. nl. ex.
st. nl. 1.

st. vtl. ex.
st. rtl. ex.’

st. vtl. 7.

sl. rtn^

It.

va. say.

Basal nuclei.

Optic nerve.

Nerve to papilla.

External oblique muscle.

Ora serrata.

Papilla.

Pigment.

Pigment behind cornea,

Pigment in place of lens
(Fig. 7) and (Fig. 12)
hyaloid membrane (7).

Pigment surrounding optic
nerve.

Pigment of retina.

Pore in epidermis.

Processes of the pigmented
layer of the retina.

Rectus muscles.

Retina.

Sclera.

Bristles of sense cells.

Spaces in the connective
tissue over the eye.

Spongioblasts.

Layer of rods and cones.

Stratum of formed con-
nective tissue.

Optic fibre layer.

Stratum lacunosum.

External nuclear layer.

Internal nuclear layer.

External reticular layer.

Non-nucleated spaces in
st. rtn’, the beginning of
external reticular layer.

Neurospongium (inner mo-
lecular layer) = inner re-
ticular layer (preferable).

Undifferentiated layer of

(See p. 65.)

retina.
Non-nucleated tract be-
tween sense cells and

mantle cells.
Blood-vessels.

RITTER. — Eyes of Typhlogobius calif.

PLATE I.

View of the dorsal side of the head of an individual 19 mm. long, show-
ing the distribution of the pigment cells, the folds of the skin, and the
eyes. X 18,

Similar view of an individual about 55 mm. long. X 21.

An individual of the same size seen from the ventro-lateral side. X 2.

Nitric-acid glycerine preparation of integument from the dorsum of the

head of a specimen 72 mm. long, showing the blood-vessels and the

pigment cells. — Surface view. X 59.

RITTER, — Eyes of Typhlogobius calif,

14.

16.

PLATE II.

Meridional vertical section of an eye from a specimen about 50 mm. long.
X 150. The fold in the pigment layer of the retina is indicated at *,
(See text, p. 63.)

Meridional section of an eye without a lens. Specimen 63 mm. long.
x 88.

The section, not quite meridional, is from the same eye from which
Figure 6 is taken, and is given to show the pigment mass (pig.”) in
the place of the lens. Consequently only the pigmented portion of the
eye is drawn. X 115.

A dissected acetic-acid glycerine preparation to show the eye muscles
and optic nerve. X 60.

Section of the integument from the dorsum of the head perpendicular
to the surface, from an individual 19 mm. long. X 450.

Section similar to the one shown in the preceding figure, but from an indi-
vidual 72 mm. long. X 450.

Preparation similar to the one shown in Figure 4 (Plate 1.), but from the
side of the body. X 63.

Section, not quite meridional, from an individual 60 mm. long. X 115.

The pigmented portion of the retina and choroid, from the same eye.
x 230. The fold in the pigment layer of the retina is indicated at *,
(See text, p. 68.)

A small portion of the pigment layer of the retina, and the choroid gland
with the optic nerve passing through it. Specimen 63 mm. long.
800.

A small portion of the retina, showing the optic nerve passing through it.
Same eye as that shown in the preceding figure. X 800.

The iris and the adjacent parts; same eye as that shown in Figure 5.

x 350,

WR,
wt fBr opt, ert.

ecam.

Pg

tete

SÉ. eom
^
er $
ff
|
|
|
f
ob ex
Ins.
pig à ; ;
va sng
lig. etr
old ^ Fold here

w NC PAG Com des.

ern

mete

gann

rin

pig”

a. apt

Ius

RITTER, — Eyes of Typhlogobius calif.

PLATE III.

Fig. 17. Meridional vertical section of the eye of a specimen 19 mm. long.
X 280. a, connective-tissue strands (see text, pp. 57, 58) ; B, ciliary
processes ¢ (see text, pp. 63, 64).
* 18. Small portion of the retina of the eye shown in Figures 5 and 16. x 980.
“ 19. A small portion of a section of the retina near the ora serrata from the
eye having no lens. X 700.
“ 20. Section of the retina of Clevelandia.
A. small portion of the retina of the eye shown in Figure 13. x 350.

RITTER.- Eyes or TYPHLOGOBJUS CALIF Pe
gà muc. cont. tis eer apa.
d i y cont.
s i JT \ 5 t:
e:
17. >
ih
> i :
(
29. ins
^
e 20
leuy.
strn’

& pig. rtm

mé dime ex
sender |
niba. x ASTON

selec,

nl. th.

Ww net

sprig Boos

SCE

hacelo

s/n opt E. $

i | / , bac,

As " mb dimer.
PE
f Staten. 2/
N P e.
à) Ca M" r ag
QUU t * $ LU rilex dus
AUR RN :
s d x > SL rán
> Ws um í " mò lont ec.
since : 3 sente
à CA uo » | Eds
" ret, ba.
i n c$ sé.rtlar
: ! sintii e sente
e : ANA rtl i
we AA N ,
5 US cl. gre.
dae y

dg st for opt

for Mi. , st. fir opt.

clgn

B Meisel lth. Boston

Fig.

RITTER, — Eyes of Typhlogobius calif.

24.

PLATE IV.

Rods from the same eye as the preceding (Fig. 21); a’, Y, outer mem-
bers; a”, b”, inner members; c, portion of outer member showing the
more transparent round spots. X 720.

Section of a tactile papilla from the median lower-jaw series of a specimen
about 50 mm. long. The section is not quite parallel to the long axis
of the papilla, and this accounts for its appearing not to extend fully
through to the deep surface of the epidetmis. X 720.

Section of another papilla of the same series, same specimen as the pre-
ceding. X 350 about.

Note. — The leader from va. sng. has neither the right direction nor
sufficient length to reach the blood-vessel in the axis of the papilla.
Diagram showing the arrangement and relative size of the papillae found

on the right side of the head of a specimen 19 mm. long.

Section of a papilla of the median lower-jaw series of Lepidogobius.

x 840.

No. 4. — Reports on the Dredging Operations off the West Coast of
Central America to the Galapagos, to the West Coast of Mexico,
and to the Gulf of California, in charge of ALEXANDER AGASSIZ,
carried on by the U. S. Fish Commission Steamer “ Albatross,”
during 1891, Lieut.-Commander Z. L. TANNER, U. S. N., Com-
manding.

[Published by Permission of Marsnauu McDonato, U. S. Fish Commissioner. |

IV.

Vorläufiger Bericht über die erbeuteten Holothurien. Von Hupmrr
Lupwic.

Dim Holothurien-Ausbeute der im Frühling 1891 unter der wissen-
schaftlichen Leitung von A. Agassiz stattgefundenen Fahrten! des
Vereinigten-Staaten-Dampfers Albatross, welche sich auf den östlichen
tropischen Theil des Stillen Oceans erstreckten, ist eine so ungemein
reiche, dass es sich verlohnt einen vorläufigen Bericht über deren Inhalt
zu geben. Die im Manuscript fertig vorliegende ausführliche Abhand-
lung wird von 19 Tafeln begleitet in den Memoirs of the Museum of
Comparative Zoölogy at Harvard College, erscheinen.

Die Sammlung umfasst im Ganzen 46 Arten aus 28 Gattungen. Neu
sind darunter eine Familie, eine Subfamilie, acht Gattungen, 30 Arten
und drei Varietäten.

Von besonderem Interesse dürfte der Nachweis einer Übergangs-
gruppe der Aspidochiroten zu den Elasipoden sowie die Entdeckung
einer seltsamen, dem pelagischen Leben angepassten Form sein.

In systematischer Reihenfolge sind die Untersuchungsergebnisse der
einzelnen Familien die folgenden.

1 Vgl. über diese Fahrten, Bull. Mus. Comp. Zoöl., Vol. XXI. No. 4, 1891, u.
Vol. XXIII. No. 1, 1892,
VOL. XXIV. — NO 4.

BULLETIN

OF THE

I. ASPIDOCHIROTAE.

Ausser einigen längst bekannten littoralen Arten (Holothuria languens
Sel, pardalis Sel., maculata [Br.], marenzellera Ludw., vagabunda Sel.)
wurden zahlreiche Exemplare des interessanten, von Théel beschriebenen
Pseudostichopus mollis aus Tiefen bis zu 1823 Faden erbeutet, deren
Untersuchung verschiedene Ergänzungen zu den Angaben jenes For-
schers gestattet, namentlich aber eine Anderung der Gattungsdiagnose
nothwendig macht. Die veränderte Diagnose lautet :

(19—)20 Fühler; Keine Fühlerampullen ; Steincanal mit der Kör-
perwand verbunden; Bauch abgeflacht; Ambulacralanhánge in Form
ungewöhnlich kleiner, schwer zu bemerkender Füsschen, die mehr oder
weniger deutlich in Längsreihen geordnet sind; Genitalschläuche in
zwei Biischeln (einem rechten und einem linken); After in einer senk-
rechten Furche, ohne Kalkzähne ; Haut ohne Kalkkórper.

Ferner ist die bis jetzt nur durch Théel bekannte Gattung Paelopatides
in mehr als 70 Exemplaren der P. confundens Théel vertreten. Auch
diese F'orm, deren Fühler einen peltato-digitaten Character haben, zeich-
net sich durch den Mangel der Fühlerampullen aus. Ihre nühere Unter-
suchung und Vergleichung lehrt, dass P. agassizii Théel mit confundens
zu verginigen ist und dass die neuerdings von Walsh beschriebene ange-
bliche Elasipoden-Art : Benthodytes gelatinosa in die Gattung Paelopatides
gehórt. Doch muss die Diagnose dieser Gattung dahin erweitert werden,
dass sie auch Arten mit nur 15 Fühlern umfasst. In dem Besitze von
nur 15 Fühlern reiht sich an Paelopatides gelatinosa (Walsh) eine neue
Art: P. suspecta.

Durch den Mangel der Fühlerampullen schliessen sich an die Gat-
tungen Pseudostichopus und Paelopatides drei neue Gattungen an: Sy-
nallactes, Mesites und Meseres, welche mit jenen zu einer besonderen
Unterfamilie der Synallactinae vereinigt werden. Die übrigen Aspidochi-
roten werden zu einer Unterfamilie der Holothurinae zusammengefasst.

Die Gattung Synallactes ist durch zwei Arten: ©. alexandri n. sp.
und S. aenigma n. sp. vertreten und durch folgende Merkmale charac-
terisiert: (18—)20 Fühler; keine Fühlerampullen ; Steincanal mit
der Kórperwand verbunden; Bauch abgeflacht ; Füsschen des Bauches
und Papillen des Rückens in Längsreihen geordnet und auf die Radien
beschränkt; Genitalschläuche in zwei Büscheln (einem rechten und
einem linken): After nicht in einer Längsfurche und ohne Kalkzähne ;
Kalkkörper in der Haut und in den Ambulacralanhängen vorhanden.
Von der neuen Gattung Mesites liegt nur eine Art: M. multipes n. Sp.

MUSEUM OF COMPARATIVE ZOOLOGY. 107

vor. Die Diagnose der Gattung lautet: (18 —) 20 Fühler; keine Füh-
lerampullen; Steincanal an, aber nicht in die Kórperwand tretend ;
Bauch etwas abgeflacht; kleine, gleichartige, zahlreiche Füsschen sind
über den ganzen Körper gleichmissig vertheilt ; Genitalschlàuche nur
in einem (linken) Büschel ; After ventral oder subventral, ohne Auszeich-
nung; Kalkkórper in der Haut und in den Fiisschen vorhanden.

Die Merkmale der neuen Gattung Meseres mit der einen Art: M.

macdonaldi n. sp. sind die folgenden: 15 Fühler; keine Fühlerampullen ;
Steincanal?; Kórper niedergedrückt mit abgeflachter Bauchseite, deren
Rand ringsum von einer einfachen Reihe finer Füsschen wie von einem
Randsaume besetzt ist; ausserdem unregelmássig (1) über den ganzen
Körper vertheilte Füsschen (? Papillen) ; Genitalschláuche in zwei Bü-
scheln (einem rechten und einem linken) ; After ventral, ohne Auszeich-
nung; Haut ohne Kalkkórper.
: Die Unterfamilie der Synallactinae umschliesst demnach Formen,
welche in dem Mangel der Fühlerampullen, im Verhalten des Steincanals
und, wie hier hinzugefügt werden kann, auch in dem Fehlen eines Wun-
dernetzes von den Aspidochiroten zu den Elasipoden hinüberleiten und
so die schon vor einiger Zeit von mir geäusserte Ansicht bestätigen, dass
die Elasipoden Abkómmlinge von Aspidochiroten sind.

II. ELASIPODA.

1) Psychropotinae. Die sechs vorliegenden Arten sind mit Aus-
nahme einer einzigen, der Benthodytes sanguinolenta Théel, siimmtlich
neu. Eine gehórt ebenfalls zur Gattung Benthodytes: D. incerta, Die
vier anderen vertheilen sich auf die Gattungen Psychropotes : Ps. raripes
und Ps. dubiosa und Kuphronides: E. tanneri und E. verrucosa.

Huphronides tannert n. sp. schliesst sich nahe an die typische Art der
Gattung: Æ. depressa Théel an, unterscheidet sich aber durch die
grössere Zahl der Rückenpapillen und durch stärkere Bedornung der
vierarmigen Kalkkórper. Auffallend ist ihr Reichthum an Kalkkórpern
in den inneren Organen, namentlich in der Wand der Genitalorgane,
des Darmes und der Darmblutgefüsse. Kuphronides verrucosa n. sp.
unterscheidet sich von Æ. depressa und #. tanneri schon durch die zahl-
reichen Warzen, mit welchen der Rücken des Rumpfes und Saumes
übersät ist; dazu kommt eine geringere Abflachung des Körpers, eine
geringere Breite des Saumes, eine andere Zahl der Fühler und Rücken-
papillen sowie Form und Grösse der Kalkkórper.

Psychropotes raripes n. sp. ist durch die geringe Zahl (7—8) der freien

108 BULLETIN OF THE

Füsschen in den seitlichen ventralen Radien gekennzeichnet. Die
Diagnose der neuen Art lautet: Körper gestreckt, ohne den Schwanz-
anhang etwa 4$ mal so lang wie breit. .Schwanzanhang nahe am Hin-
terende des Rückens entspringend, mehr als halb so lang wie der Körper
und mit einfacher, abgerundeter Spitze endigend. 18 Fühler, Rücken
jederseits mit fünf bis sieben winzigen Papillen. Bauch nur vorn und
hinten von einem deutlichen Randsaume begrenzt, dagegen an jeder
Seite mit einer Reihe von sieben bis acht grossen Füsschen. Kalkkórper :
vierarmige, nach aussen convexe Kreuze von gedrungener Gestalt mit
sehr kräftiger Bedornung der Aussenseite. Psychropotes dubiosa n. sp.
besitzt wie Ps. loven? Théel nur zehn Fühler, dagegen sind die Füsschen
der Flanken nicht zu einem zusammenhängenden Saume vereinigt,
sondern durch kurze Zwischenräume getrennt.

Benthodytes incerta n. sp. kann nur vorläufig und bedingungsweise
zur Gattung Benthodytes gestellt werden, da der ungünstige Erhaltungs-
zustand der zwei vorliegenden Exemplare eine sichere Entscheidung
nicht gestattet. Benthodytes sanguinolenta Théel ist in nicht weniger als
25 Exemplaren vertreten. Die bis jetzt unbekannt gewesenen Weibchen
dieser Art zeichnen sich durch eine bei Echinodermen unerhörte Grösse
ihrer deutoplasmareichen Eier aus, deren Durchmesser sich bis auf 2 mm.
beläuft; das 0.26 mm. grosse Keimbläschen kann man mit blossem
Auge wahrnehmen. Der Blinddarm dieser Art hängt (im Gegensatze zu
der Angabe Théel’s) zweifellos an der Cloake und ist der Kieme der
sogenannten Lungenholothurien homolog. Die Längs- und Quermuscu-
latur der Körperwand ist im Sinne einer Bilateralsymmetrie in den
einzelnen Radien und Interradien ungleich ausgebildet.

2) Deimatinae. Von den acht erbeuteten Arten sind nur zwei
bekannt ; die übrigen sind neu und geben zugleich zur Aufstellung von
drei neuen Gattungen Veranlassung. Die beiden bekannten Arten sind
einmal die sehr variable Oneirophanta mutabilis Théel und zweitens die

r

Pannychia moseleyi Théel; letztere liegt aber in Exemplaren vor, die
nicht ganz mit dem Arttypus übereinstimmen und deshalb eine besondere
Varietüt: henrici darstellen. Von den neuen Arten bekannter Gattungen
schliesst sich Deima pacificum n. sp. am nächsten an das: westindische
Deima blakei Théel an. Oneirophanta affinis n. sp. unterscheidet sich
von O. mutabilis namentlich durch eine andere Anordnung der Rücken-
papillen und andere Gestalt der Fühler und macht eine Erweiterung
des Théel’schen Gattungsbegriffes nothwendig. Laetmogone theeli n. sp.
unterscheidet sich durch ihre grössere Fühlerzahl (20 statt 15) von allen
bisher bekannten Zaetmogone-Arten.

MUSEUM OF COMPARATIVE ZOOLOGY. 109

Die neue Gattung Scotodeima mit der gleichfalls neuen Art setige-
rum nimmt eine Zwischenstellung zwischen den Théel’schen Gattungen
Oneirophanta und Orphnurgus ein; ihre Diagnose lautet: 20 ziemlich
grosse Fühler; in jedem seitlichen ventralen Radius eine Doppelreihe
grosser Füsschen und darüber eine Reihe langer, schlanker, nicht zurück-
ziehbarer Papillen (Flankenpapillen) ; eine Doppelreihe ähnlicher Papil-
len (Rückenpapillen) auf jedem Radius des Ritckens ; mittlerer ventraler
Radius mit einigen verkümmerten Füsschen ; Kalkkörper : vierarmige
Kreuze und Umbildungen derselben.

Die neue Gattung und Art Laetmophasma fecundum gehört in die
rüdchenführende Gruppe der Gattungen Pannychia, Laetmogone, Ilyo-
daemon und schliesst sich durch ihre Kalkkörper am engsten un Pan-
nychia an, unterscheidet sich aber durch die reiche Entwicklung von
Ambulacralpapillen auf der Bauchseite. Diagnose von Laetmophasma :
16—20 ziemlich grosse, nicht zurückziehbare Fühler; in jedem seit-
lichen ventralen Radius eine einfache Reihe grosser Füsschen ; Rücken
und Bauch mit zahlreichen Papillen übersüet; Kalkkörper ähnlich wie
bei Pannychia.

Die dritte neue Gattung endlich, Capheira, mit der ebenfalls neuen
Art C. sulcata, kann nur vorläufig und mit einigem Zweifel bei den
Deimatiden in der Nähe von I/yodaemon untergebracht werden. Sie
deutet darauf hin, dass auch einige Holothuria-Arten (H. thomsona, lactea,
murrayi) von den Aspidochiroten zu den Elasipoden hinüberleiten und
lässt sich folgendermassen characterisieren : 30 (1) Fühler ; Bauchseiten
und Rücken mit zahlreichen, gleichartigen, feinen Füsschen bedeckt, die
an den Flanken noch dichter stehen ; mittlerer ventraler Radius ohne
Füsschen ; Kalkring aus fünf Radial- und zehn Interradialstücken ge-
bildet; Kalkkörper : Stühlchen.

3) Elpidiinae. Aus der Unterfamilie der Elpidiinen liegen nur drei
Arten vor, von denen zwei neu sind und die dritte eine Varietät einer
bekannten Art darstellt. Von der typischen Peniagone vitrea Théel
unterscheidet sich die neue, als setosa bezeichnete Varietät unter Anderem
durch einen feinen Stachelpelz, der durch die Aussenfortsätze der Kalk-
körper gebildet wird. Die zur Gattung Scotoanassa hinüberleitende
Peniagone intermedia n. sp. schliesst sich in der Anordnung der Füsschen
und der Ausbildung eines hinteren Randsaumes, welcher die hintersten
Füsschen aufnimmt, sowie auch durch den Mangel freier Rückenpapillen
an die Peniagone challengeri Théel an, ist aber verschieden von ihr durch
die Körperform und durch die Form und Stellung des Nackensegels.
Von der seltenen, bis jetzt nur nach einem einzigen Exemplare bekannten

110 BULLETIN OF THE

Gattung Seotoanassa liegen mehrere Exemplare und Bruchstücke vor,
welche zur Aufstellung einer neuen Art Scotoanassa gracilis zwingen,
deren Diagnose die folgende ist: Körper etwa dreimal so lang wie breit,
nach hinten etwas verjüngt. Vordersaum (= Nackensegel) und Hinter-
saum durch einen Seitensaum verbunden ; Vordersaum mit vier Zipfeln,
von denen zwei vorwärts, zwei seitwärts gerichtet sind; Hintersaum zu
einem unpaaren, nach hinten gerichteten Zipfel verlängert. Mund
ventral und schräg nach hinten gestellt, am Ende eines kurzen, nach
hinten abgeknickten, vordersten Rumpfabschnittes. Endscheibe der
(10%) Fühler rundlich mit zahlreichen, winzigen Papillen. Füsschen
gross, nur unten an den Seitentheilen des Hintersaumes angebracht,
jederseits in der Zahl fünf. Haut sehr zart, etwas durchscheinend ; ihre
vierarmigen Kalkkórper zahlreich, aber zart und mit schlanken, nur sehr
schwach bedornten Aussenfortsützen, die fast doppelt so lang sind wie die
gleichfalls nur sehr schwach bedornten Arme selbst.

III. PELAGOTHURIIDAR.

Unter allen vom Albatross erbeuteten Formen ist eine merkwiirdige
Art am auffälligsten und interessantesten, welche sich in Bau und Lebens-
weise so sehr von allen bisher bekannten Holothurien unterscheidet, dass
sie nicht nur als Vertreter einer neuen Gattung sondern auch einer neuen
Familie betrachtet werden muss. Sie führt ein pelagisches Leben und
zeichnet sich in bemerkenswerthester Weise durch die Ausbildung eines
besonderen Schwimmapparates aus, der in einer am Rande in lange
Strahlen ausgezogenen Scheibe besteht, welche im Umkreis des Fithler-
kranzes angeordnet ist und in seiner Form etwa an die an ihrer Basis
durch eine Schwimmhaut verbundenen Arme mancher Tintenfische erin-
nert. Ich beschränke mich hier darauf eine kurze Diagnose der Familie,
rattung und Art zu geben und daran einige Worte über die Verwandt-
schaft des wunderbaren Thieres zu knüpfen, bei dessen Anblick man
zunächst an manches Andere eher als an eine Holothurie denkt.

Familie Pelagothuriidae. Füsschen (und Ambulacralpapillen)
fehlen. Mund und After terminal. Körper drehrund, rings um den
Fühlerkranz zu einer dünnen, an ihrem Rande zu langen Strahlen
ausgezogenen Scheibe ausgebreitet. Die Fühlercanäle entspringen aus
den wohlentwickelten Radialeanülen und entsenden an der Fühlerbasis
je einen (wahrscheinlich einer Fühlerampulle homologen) Canal in die
Scheibe; diese Scheibencanäle verlaufen in radiärer Richtung zur Peri-
pherie der Scheibe und treten in je einen Strahl des Scheibenrandes ein

MUSEUM OF COMPARATIVE ZOOLOGY, 111

um ihn bis zur Spitze zu durchziehen. Längsmuskeln der Körperwand
einfach ; Rückziehmuskeln fehlen ; Quermuseulatur in den Radien unter-
brochen. Weder Kiemenbáume, noch Wimperorgane, noch Cuvier'sche
Organe sind vorhanden. Genitalorgane rechts und links vom dorsalen
Mesenterium.

Pelagothuria n. g. 13—16 Fühler und ebenso viele Scheibencanilo
sind vorhanden. Die Fühler sind am Ende zweitheilig und ebendort mit
winzigen Papillen besetzt. Keine Spur eines Kalkringes. Steincanal
in der Einzahl, dicht vor dem Genitalgang zur Haut aufsteigend und
direct nach aussen mündend. Jederseits ein traubenfórmiges Genital-
organ, das linke oft viel krüftiger entwickelt als das rechte. Kalkkórper
fehlen sowohl der Haut als allen inneren Organen.

Pelagothuria natatria n. sp. Haut dünn, weich, etwas durchscheinend,
violett bis purpurn gefärbt. Rumpf etwa dreimal so lang wie dick, nach
hinten verjüngt und abgerundet. Länge des Körpers einschliesslich der
Fühler bis 47 mm.; Dicke des Rumpfes bis 13 mm. ; Durchmesser
der Schwimmscheibe bis 35 mm. ; Länge der Scheibenstrahlen bis
50 mm.

Im Mangel der Füsschen stimmt die neue Familie der Pelagothuriiden
mit den Synaptiden und den Molpadiiden überein, unterscheidet sich
aber von diesen durch das Fehlen der Kiemenbäume, von jenen durch
den Besitz von Radialeanülen, den Ursprung der Fühlercanäle aus diesen
Radialcanálen, die Unterbrechung der Quermusculatur der Kórperwand,
den Mangel der Wimperorgane, Das Fohlen der Kiemenbüume verweist
die Pelagothuriiden unter den actinopoden Holothurien, zu denen sie
zweifellos gehóren, in die Nachbarschaft der Elasipoden, von denen sie
aber wieder durch den Mangel aller Füsschen (und Ambulacralpapillen)
geschieden sind. Ich glaube in ihnen Abkömmlinge von Elasipoden
schen zu müssen, die sich aus den Verhältnissen der Tiefsee zu Bewoh-
nern der oberflächlichen Meeresschicht heraus- und emporgearbeitet und
dem pelagischen Leben durch Ausbildung eines Schwimmapparates

angepasst haben.

IV DENDROCHIROTAE.

Die Dendrochiroten sind durch neun Arten vertreten, die zu fünf
Gattungen gehören ; darunter befinden sich eine neue Gattung und
sieben neue Arten. Fine Art Phyllophorus aculeatus n. sp. stammt nicht
aus der Tiefe, sondern von der Küste von Panama. Die übrigen Arten
sind Tiefenbewohner aus den Gattungen Psolus (vier Arten), Psolidium

112 BULLETIN OF THE

(zwei Arten), Cucumaria (eine Art) und Sphaerothuria n. g. (eine Art).
Von den vier Psolus-Arten ist eine nur in einem ingenditehen, nicht
sicher bestimmbaren Exemplare vorhanden. Psolus pauper n. sp. zeichnet
sich durch vólligen Mangel von Kalkkórperchen in der Haut der Sohle
aus und steht im Übrigen den Arten squamatus und fabric nahe.
Psolus digitatus n. sp. hat ebenfalls den Habitus der beiden eben genann-
ten Arten, unterscheidet sich aber durch die Gestalt der Fühler, welche
einfach fingerfórmige, unverästelte Schläuche darstellen; die Fühler
haben demnach hier ihre jugendliche Form dauernd festgehalten — ein
Fall, den wir bis jetzt nur bei den Molpadiiden-Gattungen Lupyrgus
und Hoplodactyla kannten. Psolus diomedeae n. sp. schliesst sich durch
die deutliche Ausbildung von je fünf interradialen Oral- und Analplatten
an Ps. antarcticus, tuberculosus und ephippifer an, von denen er sich aber
in anderen Merkmalen hinreichend unterscheidet.

Die beiden neuen Psolidium-Arten, panamense und gracile, geben
Veranlassung die Diagnose der Gattung Psolidium einer Revision in dem
Sinne zu unterwerfen, dass sie nunmehr lautet: “Zehn Fühler (die
beiden ventralen kleiner als die übrigen) ; mittlerer Abschnitt des Tri-
viums zu einer deutlich umgrenzten Sohle abgeflacht und hier mit gut
entwickelten, auf die Ambulacren beschränkten Füsschen besetzt ; auf der
übrigen Körperoberfläche kleinere oder zu Papillen verkümmerte Füss.
chen, welche entweder überall zerstreut stehen oder sich an den Kórper-
enden auf die Radien beschrünken." Von der so gefassten Gattung
Psolidium unterscheidet sich Theelia lediglich durch die Zahl (15) der
Fiibler. Zu Psolidium wird auch die Théel’sche Art Psolus brasiliensis
gerechnet, so dass Psolidiwm nunmehr vier Arten: dorsipes Ludw.,
brasiliense Théel, panamense und gracile umfasst. Die Psolidium-Arten
lassen sich als werdende, in Bildung begriffene Psolus-Arten ansehen.
Die Formenreihe, welche sich von Cucumaria (und Thyone) beginnend
durch Colochirus zu Psolidium und von hier zu Theelía und schliesslich zu
Psolus verfolgen lässt, begründet die Ansicht, dass die Gattung Psolus
keine alterthümliche und ursprüngliche, sondern eine verhültnismüssig
junge, vielleicht, eine der jüngsten unter allen Dendrochiroten ist.

Die Cucumaria abyssorum Théel wurde in 64 Exemplaren aus Tiefen
von 905—2232 Faden erbeutet, welche lehren, dass die var. hyalina
Théel jugendliche, dagegen die var. grandis Théel alte erwachsene Thiere
derselben Art darstellt.

Als die interessanteste aller bis jetzt ausgefundenen Tiefsee-Dendro-
chiroten erscheint die neue Gattung Sphaerothuria mit der einen Art:
Sph. bitentaculata n. sp. Dieselbe hat eine fast kugelrunde Gestalt, und

MUSEUM OF COMPARATIVE ZOOLOGY. 113

ist mit grossen Platten bepanzert, welche je einen kräftigen, frei her-
vorstehenden Stachel tragen. Die Fühler haben wie bei Psolus digitatus
die einfach cylindrische, jugendliche Form bewahrt. Von den zehn
Fühlern anderer Dendrochiroten sind die beiden ventralen, die bei jenen
oft sehr viel kleiner sind als die übrigen, vollständig geschwunden. Von
den acht übrig gebliebenen sind sechs (vier dorsale und zwei ventrale)
verkiimmert, dagegen nur zwei (ein rechter und ein linker) wohl ent-
wickelt. Am Kalkring ist jederseits das ventrale Interradialstück mit
dem seitlichen ventralen Radialstück zu einem anscheinend einheitlichen
Stücke zusammengedrängt. Die ungemein kleinen Füsschen sind auf
die Radien beschränkt ; ihre Füsschencanäle durchbohren die Platten des
Hautpanzers. Aus einem Vergleiche mit Cucumaria (Echinocucumis)
typica. (Sars) geht hervor, dass die Sphaerothuria aus der Echinocucumis-
Gruppe der Gattung Cucumaria abzuleiten ist.

V. MOLPADIIDAE.

In der Sammlung sind sechs Arten vertreten, von denen nicht weniger
als vier neu sind. Kine Art: californica n. sp. gehört zur Gattung
Caudina, drei Arten: violaceum (Stud.), granulatum n. sp. und inter-
medium n. sp. gehören zur Gattung Zrochostoma, zwei Arten : danielssent
Théel und spinosum n. sp. zur gattung Ankyroderma, Die Caudina cali-
fornica ist durch ihre Kalkkörper scharf unterschieden von den bisher
bekannten Caudina-Arten. Trochostoma violaceum (Stud.) und Ankyro-
derma danielsseni Théel konnten bis jetzt als rein antarktische Formen
angesehen werden, während sich nunmehr herausstellt, dass sie in tiefem
kalten Wasser nordwärts den Aquator überschreiten. Dass dieses Vor-
dringen in tropische Meeresgebiete Hand in Hand geht mit dem Herab-
steigen in grössere, kalte Tiefen lässt sich besonders bei Zrochostoma
violaceum nachweisen. Die neue Art Trochostoma granulatum ist durch
ihre Kalkkörper characterisiert und entbehrt wenigstens in den vorliegen-
den Exemplaren der “weinrothen Körperchen.” Auch die andere neue
Trochostoma-Art : intermedium ist durch ihre Kalkkörper gekennzeichnet,
welche ihr eine vermittelnde Stellung zwischen den arktischen Arten :
Tr. arcticum, boreale, thomsonit und dem antarktischen Zr. antarcticum
anweisen ; an einem Exemplare konnte die innere Organisation genauer
untersucht werden. Von dem bis jetzt nur nach einem einzigen Exem-
plare aus der Ausbeute des Challenger bekannten Ankyroderma danielssent
sind im Ganzen 29 Exemplare vorhanden, welche cine Bestátigung und
Ergänzung der Théel’schen Angaben ermöglichten. In ihrer geographi-

114 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY,

schen Verbreitung bietet diese antarktisch-pacifische Art ein Gegenstück
zu der arktisch-atlantischen A. jefreysii, ohne indessen,.wie Théel ver-
muthete, nur eine Varietát der letzteren zu sein. Die neue Art Ankyro-
derma spinosum unterscheidet sich von der iibrigens nahe verwandten
A. danielssent vorzugsweise durch die Kalkkörper der Rumpfhaut und
die verhältnismässig grössere Länge des Schwanzes.

VI. SYNAPTIDAE.

Nur eine Art Synapta abyssicola Théel liegt in Bruchstücken aus 1672
und 1772 Faden Tiefe vor, welche in ihren‘ Kalkkórpern nur insofern
eine Abweichung von den typischen Exemplaren zeigen als die Anker-
arme keine Zähnchen besitzen. Wegen dieses Unterschiedes geben die
vorliegenden Bruchstücke einstweilen zur Aufstellung einer besonderen
Varietát : pacífica Veranlassung. Sie gestatten ferner nähere Mittheil-
ungen über den anatomischen Bau dieser Art, von der bis jetzt einzig
und allein die Kalkkörper bekannt waren.

Bonn, den 6. April, 1893.

No. 5.— The Development of the Scales of Lepidosteus. By
W. S. NICKERSON.!

Tue scales of Lepidosteus have been studied by several investigators.
Their most prominent peculiarity is the possession of a hard, smooth,
nearly homogeneous outer layer, which in the adult condition lies free
to the exterior not covered by other tissue. The question of most
immediate interest concerning the scales is the one in regard to the
source and nature of this outer layer. The view which has been gen-
erally held concerning it is that recently maintained by Oscar Hertwig
(79), who believes that 16 is a true enamel layer derived from the
overlying epidermal tissue.

The more recent studies of Klaatseh (90) seem to contradict this
conclusion by showing that it has a dermal origin. Closely connected
with the question concerning this layer, and in part dependent upon
the answer found for it, is that of the relationship existing between the
Ganoid scale and the scales of Selachians, Teleosts, and Dipnoi. Addi-
tional interest is also given to this inquiry by the fact that it has a
bearing upon the question of the relation of scales to teeth, ,

The writers who up to this time have investigated the scales of
Lepidosteus have, with one exception, had only adult material for study,
and therefore their conclusions regarding the outer layer were neces-
sarily drawn chiefly from the physical properties of the finished scale.
Since in the adult the tissue has disappeared from the outer surface of
the scales, it is evident that the source of the problematical outer layer
can be determined only by the study of young fishes in which the scales
are in process of formation, and still covered by the tissue which pro-
duces them.

Klaatsch (90) in his study of material from a single young gar-pike
made some very interesting and important observations, which, if con-
firmed by subsequent investigations, must overthrow the idea of Hert-
wig, that the outer layers of the ganoid scale are directly homologous
with the enamel of the teeth of higher vertebrates.

1 Contributions from the Zoölogical Laboratory of the Museum of Comparative
Zoology, under the direction of E. L. Mark, No. XXXVI,

VOL. XXIV. — NO. 5. 1

116 BULLETIN OF THE

By the study of abundant material from several (ten) different fishes
of various ages and sizes, I have endeavored to trace the course of
development with sufficient thoroughness to settle the origin of the layer
now in dispute, and if possible to throw some light upon the homologies
of the scales of the Ganoids. For the most part my observations lead
me to conclusions in agreement with those of Klaatsch, although in
regard to some points I am unable to coincide with his views.

For the greater part of the material used I am indebted to Dr.
Edward L. Mark, who very kindly placed at my service his own supply.
These fishes had been reared by him in aquaria from eggs collected at
Black Lake, N. Y., in the spring of 1882. They had been killed at inter-
vals by various fixing reagents, as indicated in the table following, and
carefully preserved in alcohol. The histological conditions were very
faithfully preserved in all except one specimen. This, the largest and
oldest fish of the lot, had died in the aquarium after attaining an age
of fifty-two months and a length of thirty centimeters. When found
it could have been dead but a few hours at most, and was at once
preserved in alcohol. On studying this fish I found the scales in a
most interesting stage of development, but unfortunately the epidermis
had so far degenerated that it was impossible to make out anything
of importance concerning its relation to the scale. I was moreover some-
what apprehensive lest the confinement in which these fishes had been
reared, together with the handling and other unnatural treatment to
which they had been subjected, might have rendered the conditions of the
scales found in this and the other larger fishes abnormal to some extent.

I was unable to obtain at once healthy material of the proper age to
show the later stages of growth, and to this fact is due a considerable
delay in the completion and publication of the results of my work. The
missing later stages were supplied by material from two young gar-pikes
which Mr. S. P. Bartlett of the Illinois State Fish Commission obtained
and very kindly sent me. They were killed according to my direction
by immersion in strong alcohol, a method that is very satisfactory for
epidermal and dermal tissues, and which has the advantage over acid
killing-fluids of leaving calcareous structures quite unaffected. For other
material which I have studied I am indebted to Prof. E. A. Birge of the
University of Wisconsin, and through Dr. Mark to Prof. Jacob Reighard
of the University of Michigan. The material from both of these sources
had been killed in alcohol.

To all of ‘these gentlemen I wish to express my gratitude, and
especially to Dr. Mark, not only for his liberal supply of material

MUSEUM OF COMPARATIVE ZOOLOGY. Ben

which it had cost him much time and labor to obtáin, but also for his
direetion and many valuable suggestions.

TABULATED List or FISHES STUDIED.

Length.| Age. Killing stis E
mm. |Months.| Reagent. Jondition. Received from

145 184- Merkel’s Local aggregations of spongy connective | Dr. Mark.

fluid,
115 18 HgCl, Bony scale beginning to be formed. au:
(hot).
150 184 Merkel’s Spines bemg formed. ER
fluid,
160 19 90 per cent | Few spines. “ u
alcohol.
175 24 Picro-sul- Edges of scale beginning to overlap. ne
phuric
acid,
190 35 2 per cent Mature spines all over scales, «e u

chromic 4-
acetic (few

drops).
287 ? Alcohol. Spines all over scale ; no ganoin layer. Prof. Birge.
300 52 Died, 90 per | Epidermis degenerated ; thin ganoin layer, | Dr Mark.
cent ale,
300 ? Alcohol, Spines on edges ; thin ganoin layer, Mr S. P Bartlett.
460 T Alcohol. Few spines on edges ; thicker ganoin layer | Prof Reighard.

The method used has been chiefly the preparation and study of
series of sections. In all except very young stages it was necessary to
decalcify the material before it could be sectioned, and even after
decalcification sections in most cases could not be cut thinner than
15 m, and often in the later stages it was necessary to make them
between 20 and 30 y thick. For deealeifying I used 90% alcohol to
which was added a small quantity of 10% hydrochloric acid (in the
ratio of about 3 to 1). The tissue was usually left in this acid alcohol
twenty-four hours or more, and then soaked in several changes of fresh
alcohol to remove all traces of the acid before staining. Sections pre-
pared by grinding down scales have also been studied, as well as scales
freed from the soft tissues by treatment with caustic potash. Only by
the use of the latter reagent was I able to get a satisfactory knowledge
of the spines which cover the scale in its immature state.

The stains which have given the best results are Boehmer's alum
hematoxylin and Kleinenberg’s hematoxylin. As a nuclear stain the

BULLETIN

OF THE

former is much the better; the latter is especially useful in the
study of the glands of the epidermis and in bringing cell membranes
into prominence,

The young Lepidosteus in which I found the earliest indication of
the formation of scales was about 145 mm. long and a little over
eighteen months old. Klaatsch has stated that the scales appear upon
the dorsal side of the body earlier than upon the ventral, and as the
material studied was taken in this specimen (as in most cases) from
near the ventral line, it is probable that the scales begin to form upon
the dorsal portion of the body at a somewhat earlier age.

Before the beginning of the formation of the scales the integument is
about 225 u thick. The epidermis is nearly twice as thick as tho
dermis and is made up of numerous layers of spheroidal or cuboidal
cells (Fig. 1). The cells of the deepest layer are somewhat elongated
in a direction perpendicular to the surface, and consequently have the
appearance of indistinctly columnar epithelium.

There are two kinds of glandular structures present in the epidermis.
Those of one kind have a spherical, or more often ovoid or oblong form,
with the greater axis perpendicular to the surface of the body, and
appear in sections as spaces of corresponding shape. The long diameter
in the larger of these spaces is not infrequently (Fig. 1) from one half
to two thirds the thickness of the epidermis (75-100 a).

The glands are quite irregularly distributed, some areas being only
very meagrely provided with them, while in others they are so crowded
as to leave little more than thin filmy partitions between them.
Sections parallel to the surface in such areas present a net-like appear-
ance, the empty gland cells corresponding to the meshes. To each
space there belongs a flattened nucleus, which is closely applied to the
wall near its deep end. The whole structure is doubtless to be con-
sidered as a single cell which has taken on a secretive function, — a
unicellular gland.

These glandular spaces sometimes appear empty, but are more often
found partly filled by an indistinctly granular colorless substance
(coagulum), probably precipitated by reagents from the fluid or semi-
fluid secretion which they held during life (Plate I. Fig. 1, Plate II.
Fig. 16). I do not find that these spaces have any means of communi-
cation with one another, and only very rarely do I find one having an
opening to the exterior. Klaatsch has described these as mucous

glands. I have no doubt that they are concerned in the production of

MUSEUM OF COMPARATIVE ZOOLOGY, 119

the slime by which the young “gars” are thickly covered. They are
not, however, the only agents concerned in its production.

The epidermal glands of the second kind are considerably smaller
than those already described; they are nearly spherical, and have an
average diameter of from 18 to 25 q. They do not appear to have been
recognized by previous observers. Some of these lie in the deeper
part of the epidermis, but much the greater number and the larger ones
occur near the surface, where many of them open.

They stain deeply with Kleinenberg’s and with Delafield’s heema-
toxylin, but in all other stains used, even in other hematoxylin dyes, the
nuclei alone are colored, and in this condition they can be distinguished
from small glands of the first kind only by very careful observation.
This probably accounts for their having been overlooked heretofore. In
Kleinenberg’s hematoxylin the whole gland takes a blue tint, while the
wall presents a reticulated appearance due to an irregular network of
lines of a much deeper blue color. This appearance is shown in Figure
2, a-e. The glands lying in the deeper part of the epidermis (d) take
less color than those situated just below the surface (c); those which
open at the surface vary, much in the intensity of their coloring, the
differences doubtless being due to the varying amounts of mucin con-
tained in them. The distribution and peculiar staining qualities of
these glands show that they originate in the deeper part of the epidermis
and migrate to the surface, where their secretion is discharged.

Hematoxylin has long been known to stain mucus deeply. Hoyer
(90) found that basic stains are those which chiefly affect mucin, and
that hamatoxylin stains which contain alum act like basic stains. Of
the four hematoxylin dyes which I have used, Kleinenberg's (basic) and
Delafield’s (alum) gave characteristic deep blue stains to these glands;
Doehmer's (alum) and Ehrlich's (acid), on the other hand, gave pure
nuclear stains. Hoyer imputed certain failures of hematoxylin stains
to act in their normal manner to lack of “ripeness,” and it is possible
that this may be the reason for Boehmer’s alum hematoxylin not col-
oring the mucin in this case.

At all events, I believe that the stains which I have obtained afford
sufficient ground for declaring that these cells act as glands, and secrete
the mucin element of the slime by which the young gar-pikes are thickly
covered. The other and larger glands first deseribed, which take only a
nuclear stain with Kleinenberg’s hematoxylin, must be considered to
have the function of secreting some component of slime other than

mucin,

120 BULLETIN OF THE

The distribution of these mucous glands, like that of the kind first
described, is somewhat irregular, and the two kinds are about equally nu-
merous; those of the larger kind are, however, much more conspicuous
by reason of their size. Both kinds of glands are found in the epidermis
as long as it persists over the surface of the scales,

Chromatophores are abundant in the epidermis. They vary greatly in
size and shape, and are found in all parts of the layer.

The boundary between epidermis and dermis is marked by the
presence of a distinct basement membrane.

The dermis (Fig. 1), before it has begun to undergo modifications pre-
paratory to the formation of the scales, is only about half as thick (75 a)
as the epidermis, and is composed of bundles of connective-tissue fibres
which may be divided into three sets according to the directions in
which they run. Two sets are parallel to the surface and diagonal to
the axis of the body, the third is perpendicular to the directions of the
other two. Of the first two'sets, the fibres of one have a direction
backward and ventrad parallel to the direction which the rows of
scales will have later. The fibres of the other diagonal set cross
these in such a way as to make the dorsal and ventral included angles
a few degrees less than right angles, the anterior and posterior included
angles being consequently a little greater than right angles. The
fibres of the third set are much less numerous than those of either
of the other two, and, as they extend perpendicularly through the der-
mis, serve to bind together the different layers. Nuclei are irregu-
larly distributed among the fibres, as in ordinary connective tissue.
Blood-vessels are numerous, and ramify chiefly in the outer part of the
layer just beneath the basement membrane. Chromatophores are
present, as in the epidermis, and are found mostly upon the two
surfaces of the layer, being especially numerous at the deeper one
(Fig. 1). Upon the outer surface they lie at nearly uniform distances
apart (Fig. 4). Beside these color cells there is associated with them
at the deeper surface of the dermis, between it and the musculature,
quite a thick but very irregular sheet of finely granular white pigment.
Concerning the origin of this pigment I have been able to learn
nothing.

The first indication of the formation of scales is a change in the outer
part of the dermis. There appear local thickenings of spongy tissue,
which are made up of many nucleated cells and a small number of fibres
lying among them without definite arrangement. These regions of
modified dermis are abundantly supplied with blood-vessels. At first

MUSEUM OF COMPARATIVE ZOOLOGY. Ti

these thickenings are quite distinct from one another, being separated
by tracts of unmodified tissue; thus each one resembles a very much
flattened papilla.

A similar condition found in the early stages of the formation of the
scales of Teleosts is described by Klaatsch (90, p. 159) and by Hofer
(90, p. 111). Both of these authors regard these dermal thickenings
in Teleosts as homologous with the dermal scale papillee of the Selachians.
This view receives a certain amount of support from the fact that in
Selachians, Ganoids, and Teleosts the first indication of the formation of
scales is the appearance of local elevations of the dermis in the form
of papille. On the other hand, it is difficult to imagine any change
due to increased local activity of the tissue taking place in the dermis
without producing modifications resembling to some extent papilla.
That the scales of all the groups of fishes owe their origin primarily to
the dermis seems therefore a sufficient explanation of the early appear-
ance of these papilla-like local modifications of this layer, and not to
require the assumption of any direct homology between the resulting
structures. The fact that in Selachians the scale is formed over the
surface of the papilla, while in Ganoids (Lepidosteus) and in Teleosts it
arises in the midst of the mass of cells forming the elevation, is a
fundamental difference not to be overlooked. Furthermore, at a later
period in the formation of the scales of Lepidosteus there arise dermal
papille just beneath the basement membrane, over the surface of which
calcareous material is laid down to form spines. These receive an ecto-
dermal enamel secretion over their tips, and later pierce the epidermis,
and thus exactly resemble in all important features the spines of the
placoid scales of Sclachians. The formation of the spines in Lepidos-
teus has been described already by Klaatsch (90, p. 130), and will be
treated of more fully further on in this paper.

The condition found in Lepidosteus is intermediate between that
presented by Selachians and that described in Teleosts. The presence
of the spines enables us to say that the papille by which they are
secreted, not the broader, earlier formed dermal thickenings, are the
homologues of the papilla formed in Selachians. As the early con-
ditions found in Teleosts so closely resemble those in Lepidosteus, there
can be no choice but to interpret these dermal thickenings as homolo-
vous with those of Lepidosteus. They cannot therefore be considered
the homologues of the scale papille of Selachians, as maintained by
Klaatsch and Tofer.

Hofer claims that in the trout the basal layer of cells of the epidermis

122 BULLETIN OF THE

becomes modified over the scale papillee to form an enamel membrane,
though this never becomes functional and the cells soon lose their
specialized character. Klaatsch studied the same form, and very
positively denies that any such modification takes place. He says
(p. 159): “Speciell die basale Epithelschicht lässt auch nicht die
geringste Veränderung wahrnehmen.” In view of this conflict of
testimony, it seems more probable that the differences in the condition
of the basal cells of the epidermis observed by Hofer were due to vari-
ations in different individuals than to changes normally taking place
in the same individual.

At a little later period these areas of dermal thickening have
extended laterally until their edges have become confluent, and the
dermis may then be described as made up of two layers, an outer
spongy layer and an inner fibrous layer, though no definite line of
demarcation separates the two. ‘This is the earliest stage described by
Klaatsch, At the time when the formation of the scale begins, the
spongy layer in its thickest parts is about equal in thickness to the
deeper fibrous layer. The places in which scale formation is to take
place are indicated by a somewhat greater thickness of the spongy
layer.

The scale first appears (Fig. 3) as a thin sheet of calcarcous secreted
matter in the midst of the outer layer of the dermis. It is surrounded
on all sides by the dermal cells, which lie thickly accumulated around it
and cover it on its outer surface from two to four cells deep, completely
separating it from the epidermis. Around its margin the cells lie closely
packed, and at the posterior edge are particularly numerous. To these
dermal secreting cells Klaatsch has given the name “ Scleroblasten,” a
term which I shall adopt in referring to them.

The scale plate is not quite parallel with the surface of the body ; its
posterior margin is very slightly inclined outward, and with the sclero-
blasts around it forms a low elevation against the base of the epidermis
(Fig. 3).

Secreted matter is not deposited where blood-vessels from below pierce
the outer or scleroblastic dermal layer, and consequently openings through
the scale result at such places. These persist to form the canals which
pierce the middle area of the adult scale (Haversian canals of authors).
The vessels which traverse them ramify upon the outer surface of the
scale to supply nourishment to the overlying scleroblasts. No difference
in character could be detected between the cells on the upper and those
on the under side of the scale.

MUSEUM OF COMPARATIVE ZOOLOGY. 128

Almost as soon as the calcareous material of the scale begins to be
laid down some of the surrounding scleroblasts become enclosed by it
(Fig. 3). This process goes on hand in hand with the increase in the
size of the scale, and as a result the scleroblasts are distributed through
all parts of it. That the distribution is a fairly regular one may be
seen from Figure 15 (Plate IL) aud Figure 21 (Plate IIL). In a scale
from which the soft parts have been removed by treatment with caustic
potash, the cavities occupied by these cells may be very distinctly seen.
Each one has leading from it a small number of canals (canaliculi)
which branch and traverse the scale to unite with similar canaliculi
from the neighboring cavities. Thus the whole scale is traversed by a
network of fine ramifying tubules connecting the osteoblastic cavities,
As may be seen in section, however, all the canaliculi from a given
cavity show a tendency to spread out in a plane parallel to that surface
of the scale within which it has been buried, so that the cavities in most
intimate connection are those which lie at the same distance below the
surface. Thus the material of the scale is divided into more or less
regular lamellie of calcareous matter alternating with successive layers
of cell cavities and their connecting canaliculi. These cavities are in
communication with the exterior by means of canals which penetrate
the scale and break up at their inner ends into fine tubules to join the
canaliculi. They penetrate from both the upper and the lower surface,
though much more abundantly from the lower (Plate II. Fig. 15, and
Plate III. Fig. 21). At the opening of each of these canals at the
surface of the scale there is a large cell (Plate I. Figs. 8 and 9, Plate III.
Fig. 22, Plate IV. Fig. 23), from which a process extends into the lumen,
These cells were called by Hertwig “Odontoblasten,” and the canals oc-
eupied by their processes * Dentinróhrchen." I must, however, agree with
Klaatsch that the names are poorly chosen, for the reasons set forth by
him. The constant character of dentine, as the term has been used,
is the absence of enclosed cells, and as the substance penetrated by
the processes of these cells contains such elements (osteoblasts), it seems
undesirable to call it dentine or the cells odontoblasts. Hertwig's nomen-
clature rests on an assumption of homology which has not been proved
true. However, as these terms have been adopted in the literature, it is
perhaps inadvisable to introduce new ones at this time. Besides the
odontoblasts with their processes extending into the dentinal tubules
which Hertwig describes and figures (79, p. 5, Taf. III. Fig. 4), he
mentions the presence of a granular substance partially filling the lumen
of the canals. Klaatsch asserts that cells are present in the dentinal

124 BULLETIN OF THE

tubules (p. 138). I have found by diligent searching a few cases in
which, as stated by the latter author, cells seemed to be present in the
lumen of the tubes. I have, however, observed other cases in which a
flattened cell lay just at the side of the tube, but separated from its
lumen by a thin layer of secreted material. I am therefore led by this
and by the infrequency of the cases to believe that the cells apparently
occupying the tube have a similar position, but that since they lie in
radii of the tube perpendicular to the plane of the section they have the
appearance of being in the lumen of the tube. Figure 8 will make this
clear. At the opening of the tube is seen a cell which is apparently
about to be enclosed between the secretion of the large odontoblast cell
and the present wall of the canal. At a little later stage this, if seen in
a plane at right augles to the present, would appear to lie in the lumen.

Each of these tubes is lined by a very thin secretion from the cell
which extends into it, — a secretion of the same character as the material
of the scale. In later stages, after the ganoin layer has begun to be
formed, this sheath persists even after decalcification, when the ganoin is
entirely destroyed (Plate IV. Fig. 23).

The scale increases both in thickness and in lateral extent by the
deposition of new layers of material secreted by the surrounding sclero-
blasts. Thus the edges of adjacent scales come together, and finally
overlap one another. Owing to the slight elevation of the posterior
margin of the scale, previously mentioned, this edge of each scale over-
laps the anterior edges of the adjacent posterior scales; for the same
reason, each scale is in turn overlapped on its anterior edge by the poste-
rior edges of the scales in front of it (Plate IV. Fig. 31).

Increase in thickness takes place upon both the outer and the inner
surfaces, though much more rapidly upon the inner one. As the scale
thickens inwardly, fibres which lay irregularly disposed beneath it in
the spongy dermis are enclosed by the secretion. As the calcification
extends deeper, the felted regularly arranged bundles of connective-tissuo
fibres of the deeper dermis are also enclosed, and help to make up the
deeper part of the scale. These fibres extend across from one scale to
another and so form a strong but flexible connection between adjacent
scales. They are the “Schuppenligamente ” of Hertwig. The ratio of
outward to inward growth is shown by the arrangement of the layers of
the scale (Plate II. Figs. 15, 16, and Plate IIT. Fig. 21) The oldest
part of the scale is that in which the dentinal tubules from the opposite
surfaces meet and break up into minute branches (Plate IIT. Fig. 21).
In their immature state, and before the outer layer has begun to be

MUSEUM OF COMPARATIVE ZOOLOGY. 125

formed, the scales bear upon their outer surface numerous slender coni-
cal spines or teeth. These were first described by Reissner (’59, p. 260).
They stand up from the surface not quite perpendicularly, but have a
slight inclination toward the tail (Plate I. Fig. 7, and Plate IV. Figs. 26,
28). They are transient structures, which frequently leave no trace
of their existence in the adult scale. They begin to be formed when the
scale is still comparatively thin (75-100 y), and like it they owe their
origin to the cells of the outer or scleroblastic layer of the dermis from
which they are principally formed.

The first step in the formation of a spine is an increase in the number
of the cells lying upon the upper surface of the scale just under the
basement membrane. A thickening is thus formed which rises up
against the base of the epidermis in the form of a papilla (Plate I. Fig. 5).
The basal layer of epidermal cells becomes arched over it, and the co-
lumnar character of the cells much more pronounced. The papilla
increases in height rapidly, pushing forward into the epidermis, with its
point directed slightly caudad, but remaining at its base little if any
larger than at first. The cells of which it is composed are the same in
character as those overlying the scale; in the axis of the papilla they
are crowded together without any definite arrangement, but at its periph-
ery, where they adjoin the basement membrane which separates them
from the epidermis, there is perceptible a certain degree of regularity in
the arrangement of the nuclei; they lie rather closely appressed, being
elongated radially to the axis of the cone.

After the papilla has attained a considerable height there appears
upon its tip a thin crust of calcareous matter secreted by its cells; this
gradually extends down over the sides, forming a conical cap (Plate IT.
Fig. 10), which is pushed forward by growth from below as the papilla
increases in height. In the early stages of its formation, the papilla
merely indents the lower surface of the epidermis, but as the point of
the spine is pushed forward it causes a swelling upon the outer surface
of that layer (Plate I. Figs. 5, 6). At the same time that this is taking
place the thickness of the spine cap is being increased by the addition
of new material from within. By the continuation of these processes
the point of the spine is made to pierce the epidermis and so lie free to
the exterior, while the basal end joins the scale already formed below.
There is no joint or hinge of any kind at the place where the two
unite, but the material of the spine is directly continuous with that of
the outer part of the scale. The cells of the papilla are thus entirely
shut off from the other seleroblastic cells which lie over the outer sur-

126 BULLETIN OF THE

face of the scale except at one point, where a small opening persists
through the side of the base of the spine at its junction with the scale
(Plate I. Fig. 7, and Plate IV. Fig. 26).

I have as yet said nothing concerning an enamel layer upon the
spines, because it could not be observed in any material of the stages
described. Both Hertwig and Klaatsch assert that the points of the spines
are covered by a cap of enamel Reissner says (p. 260): * Die ganzen
Stacheln scheinen übrigens auch einen dünnen Ueberzug von Schmelz
zu besitzen.” In scales which have reached the condition in which
spines are present the thickness and hardness are such as to make section
cutting impracticable, unless the tissue is first decalcified. In sections
of decalcified scales no trace of an enamel cap is to be seen, nor is any
space left between the point of the spine and the base of the overlying
epidermis to indicate that anything has been lost. I was for some time
inclined to believe that both Hertwig and Klaatsch were mistaken in
asserting the presence of an enamel cap. In spines, too, which were
broken off from scales which had been treated with caustic potash to
remove the fleshy matter, there was evidently no enamel present. Such
a spine is shown in Plate II. Fig. 11. Its cavity is seen to be con-
tinued up into the tip as a dendritic system of fine tubules (Den-
tinröhrchen of Hertwig), whose finest branches can be traced quite to
the surface, and so preclude the possibility of an enamel cap being
present.

I succeeded in learning the true condition only when, by the aid of a
low power of the microscope, I watched the caustic potash slowly eat
away the tissue from the surface of the spine-bearing scale. As the
epidermal tissue became clear and began to be dissolved, there could be
seen supported in it minute scattered caps unconnected with the scale
or the spines. With a little search, however, these caps were found
in some cases resting directly over the points of spines, but raised
slightly from them (Plate IV. Fig. 27). Similar conical caps were also
found in the bottom of the watch-glass after the soft tissues had been
entirely destroyed. Several of these loose caps were picked up and
transferred to glass slides. On treating them with weak acid (2% HCl)
their points quickly dissolved away, leaving the cap truncate (Plate II.
Fig. 12, a, b). The rest not only was not dissolved, but it showed no
change when stronger acid (10% HCl) was added. From this it would
seem that the only part of the spine which can be considered to be
enamel is the small apex which is dissolved away by the acid, or else
that the basal cells of the epidermis first secrete a substance which is

MUSEUM OF COMPARATIVE ZOOLOGY. 124

largely animal (not mineral) matter, a substance which closely resembles
the matrix of the scale itself. But I know of no case of true enamel in
which the secretion is of such a mixed nature.

Whether the bony material of the spine or the enamel tip begins to
be formed first I have no means of determining. I believe that the two
formations begin nearly simultaneously, although there is some reason
for thinking that the enamel is not secreted until a part of the bony
spine has been formed, for in a few cases I find peculiar abnormally
shaped tips (Plate I. Fig. 6, Plate III. Fig. 18, and Plate IV. Fig. 26)
which are more readily explained upon the former assumption. These
all oceur on individuals raised in aquaria, and I believe them to be due
entirely to artificial influences, — perhaps to handling. If the enamel
were already present over the tip of the spine, I cannot believe that its
point could be so distorted, whereas the matrix of the bony part of the
spine is probably somewhat plastic when first secreted. It is also con-
coivable that, before the commencement of secretion, some distortion of
the papilla permanently altered its form and caused the abnormal shape
of the spine. "This seems to me less probable, however, in view of the
particular forms which the points of the spines have taken.

The material of the spine is in lamelle, though these are not as dis-
tinct as in the scales. The insoluble part of the point (Plate II. Fig.
12, b) is doubtless made up of the first formed lamelle of the spine
which have separated from the rest of the cone on treating with caustic
potash. In Figure 25 (Plate IV.) is shown the splitting between lamella
caused by an oblique cut through the basal part of a spine. No selero-
blastic cells, however, are enclosed between the layers, and the material
of the spine therefore differs from that of the scale at its base in that
particular. But the absence of enclosed cells is not surprising, in view
of the thinness of the walls of the spine. In reaction toward acids there
is no difference between spine and scale.

The number of spines upon a scale is wholly indefinite, and their
arrangement for the most part not subject to any precise law. They
are more numerous near the free (i. e. posterior) margins, and here
show a tendency toward an arrangement in lines parallel with the
edges of the scale (Plate IV. Fig. 31). I have counted over thirty
spines on the posterior margins of a scale whose central area was
nearly destitute of them. Though some spines are formed near the
centre of the scale, by far the greater number arise near the posterior
edges. With the growth of the scale, however, the margin advances,
leaving them farther and farther from the edge.

128 BULLETIN OF THE

The scleroblasts overlying the scale about the bases of the spines do
not cease their secretive activity when the spines are completed, but
their secretion continues to be employed gradually to thicken the scale
by additions to its upper surface. The amount of material so added
is very little in the central part of the scale, but toward the margins,
where growth is still taking place, and where the greater number of
spines are formed, it is considerable (Plate II. Fig. 13). This results in
the basal ends of the spines being surrounded and incorporated in the
outer layers of the scale (Plate IV. Figs. 24, 25). These outer layers
also contain enclosed cells, and are composed of exactly the same mate-
rial as the layers immediately underlying them.

It is this outer part of the scale which Klaatsch believes to be the
enamel layer of Hertwig and other authors, and which he calls gano.
He says (p. 141): “Sie [die Ganoinschicht] entsteht im Anschluss an
die Zahnbildung auf der Schuppe; sie ist eine direct Fortsetzung des
Zahnbeins der kleinen Schuppenziihne.” He also says (p. 132) : “ Die
Substanz der Ganoinschicht stimmt in ihrer homogenen Beschaffenheit
mit dem Dentin der Zähnchen überein.”

In this I believe that Klaatsch is in error. It is inconceivable to me
that any one who had seen ground sections of the ganoin under the
microscope could for a moment confound with it the dentine of the
spines. Moreover the figure of ganoin given by Klaatsch (Tafel VII.
Fig. 6) represents a condition quite different in appearance from that
presented by the true ganoin. Furthermore, as shown by the table
already given (page 117), I found no trace of the ganoin layer on the
scales of a fish 289 mm. long while the fish from which Klaatsch's mate-
rial came was only 180 mm. long. Reissner (59, p. 260) says concerning
the spines “ihre Insertionsstellen unmittelbar unter dem Schmelz lie-
gen." The presence in this outer layer of enclosed osteoblasts, which are
absent from the layer described by Reissner and Hertwig, the fact that
it is not destroyed by acid, and its optical properties, all give convincing
proof that the material secreted “im Anschluss an die Zahnbildung
auf der Schuppe” is not the layer described by Hertwig and others
as enamel. If then, as I believe, Klaatsch did not sce the layer in
question, his claim in regard to its origin can have no weight, and the
question of its source remains where Hertwig loft it.!

1 Tt is only fair to state, however, that both the 17.5 em. and the 19 em. gar-
pikes which I studied were killed in acids, so that I haveno undecalcified material
of the same size as that which Klaatsch studied, for a perfect control of his state-
ment about the ganoin layer.

MUSEUM OF COMPARATIVE ZOOLOGY. 129

T have frequently found in sections of scales of different ages, killed
by various reagents, a narrow marginal zone appearing somewhat
different from the rest of the section (Plate I. Fig.. 3, and Plate III.
Figs. 17, 20). This often appears strikingly like a different layer
of material, and is sharply bounded from the deeper-lying layers. I
believe that this appearance is due to the action of reagents used in
killing or staining (in toto), or both, and not to any natural difference
between the parts of the scale. No such difference is found in material
killed in alcohol. Klaatsch refrains from stating anything concern-
ing the methods he used or the manner in which his fish had been
killed. The impression produced by the study of his paper is that he
was misled by some artificial condition such as I have just mentioned
and figured.

I have already stated that the spines frequently leave no trace of
their existence in the adult scale. To what their disappearance is due
I am unable to state positively. Before the outer layer begins to be
formed they have almost completely disappeared from all the central
part of the scale, but usually a few still remain close to the posterior
edges. Hertwig (79, p. 7, Taf, II. Figg. 1, 2, 3, 10) has described and
figured certain little knob-like elevations projecting up into the ganoin
layer, and he believes them to be semnants of spines which have been
lost. In this conclusion he agrees with that already expressed by Reiss-
ner (59, p. 260). I have carefully looked throngh series of sections
from two different adult “gars” for such structures, but without finding
them. Hence it would appear that their occurrence cannot be considered
a constant feature. In scales from one young fish 46 cm. long, however,
I found some peculiar structures similar to those figured by Hertwig.
One of these is shown in Plate TI. Figure 14 and is without doubt, as
Hertwig maintained, the base of a lost spine. It had been entirely
buried in the ganoin. There were also present a number of other
much smaller bodies lying between the top of the scale and the over-
lying tissue in the space from which the ganoin had been dissolved.
These were widely scattered, and many were little larger than an odon-
toblast cell; they were of an ovoidal or spheroidal form, and appear
to be the same in composition as the one figured (Plate II. Fig. 14),
and T am led to the conclusion that they are also remnants of the
bases of lost spines, which were probably in process of absorption when
the secretion of the ganoin began and buried them. Hence I be-
lieve that the obliteration of spines is in general due to absorption,
though it is hardly conceivable that the distal part of the spine dis

VOL. XXIV —NO. 5. 2

150 BULLETIN OF THE

appears in this way. That the disappearance of the spines is due to
resorption was suggested by Hertwig (79, p. 8).

The ganoin layer begins to be formed first. over the central area of the
scale, and covers all but a very narrow marginal zone, as a regular coat-
ing of uniform thickness. At this time the spines have mostly disap-
peared from all except the posterior margin. The ganoin is separated
from the bony part of the scale by a distinct, regular line, and shows
markedly different optical properties. It is more highly refractive than
the part below, and appears entirely homogeneous except for a very
delicate striation parallel to the surface, which probably corresponds to
irregularities in the rate of deposition. It cannot be seen in cut
sections, since it is entirely destroyed by decalcification, without
which section cutting is impossible in scales of this age. In sections
prepared by grinding it presents the appearance seen in Plate II.
Figure 15. As seon from the surface, too, the appearance is quite
unlike that of the bony part of the scale, which is not covered by it
(Plate IV. Fig. 27). A knowledge of this layer can be obtained only
by combining the results of study both of sections from decalcified
tissue and of preparations made by grinding undecalcificd scales of the
same stage of development.

I find the ganoin layer first present in a fish 52 months old (30 cm.
long) reared in confinement ; it is also present, and a very little thicker,
ina fish of the same length from Quincy, Ill. (Plate II. Fig. 15). In the
former case the relation of the scale to the epidermis could not be made
out, but in the latter all the tissues were well preserved in a perfectly
healthy condition, and the epidermis was distinctly separated from the
scale by a thin layer of dermal scleroblasts seldom over two or three
cells thick (Plate II. Fig. 16, and Plate IV. Fig. 23). The former of
the two figures cited shows about an average condition, and the latter a
place near the opening of one of the canals which pierce the central
part of the scale where the blood-vessels and surrounding tissue make
an unusually thick sub-epidermal sheet. Small blood-vessels are abun-
dant in this layer.

On the scales of a young Lepidosteus 44 cm. long, in which the
layer of ganoin had a thickness approaching that found in the adult,
the sub-epidermal layer still persisted over the scales of both the
ventral and the dorsal regions, though in the latter the epidermis
showed a perceptible decrease in thickness.

I am therefore led to the conclusion that, as maintained by Klaatsch,
the outer scale layer, called enamel by L. Agassiz, Reissner, and Hertwig,

MUSEUM OF COMPARATIVE ZOOLOGY. 131

is secreted, not by the epidermis, but by cells of dermal origin. Hence
it is not enamel in the modern sense of the term, but may better be
known by the name of ganoin, the term introduced by Williamson and
recently revived by Klaatsch.

The great difference in physical properties between this ganoin and
the underlying layers of the scale, and the striking resemblance which it
bears to the enamel of the scales of Sclachians and of the teeth of lower
vertebrates, have suggested that the secretion of the epidermal cells may
have made its way through the very thin layer of dermal tissue sepa-
rating the epidermis from the scale. If this were so, the thin layer
of sub-epidermal tissue with its rich blood supply would have to be
regarded merely as a device for providing an adequate supply of nour-
ishment to the epidermis during its work of secretion, —a device not
necessary in the case of the selachian spines by reason of the small size
of the secreting area, nor in the mammalian tooth, because of the sunken
position of the enamel organ in the gum. This theory would also account
for the absence of prismatic structure in the layer.

The condition of the cells of the basal layer of the epidermis, how-
ever, makes this hypothesis untenable. There is nothing in their form
or appearance to give any ground for comparing them with the cells
active in secreting the enamel cap upon the points of the spines in
Lepidosteus, or with the cells which in the Selachians secrete the enamel
which coats the scales. They are not more elongated than in early
stages when the scale is just beginning to form; their nuclei are not
larger, nor do they show any difference in staining quality ; in short, it is
impossible to look upon them as an enamel organ. (Compare Figure 10
with Figures 16, 22, and 23.)

We are thus compelled to admit that the dermal scleroblasts give rise
to three different products: (1) calcareous scale material with animal
matrix and included scleroblasts ; (2) ganoin ; and (3) a membrane which
Hertwig has called enamel membrane, but which may better be known
henceforth as ganoin membrane.

This membrane is clearly visible in all sections of tissue from which
the ganoin has been dissolved away (Plate II. Fig. 16). It is a struc-
ture which is entirely distinct from the basement membrane with which
it was confounded by Hertwig, and in sections appears thicker and more
prominent than the basement membrane (Figs. 16, 22, and 23).

It appears that we have to do here not with a differentiation of the
cells of one layer, but rather with a modification of the function of the
same cells at different periods in their history. The only case at all

182 BULLETIN OF THE

comparable to it with which I am acquainted is that of cartilage cells,
which in the formation of endochondral bone become transformed into
osteoblasts; i. c. the character of the secretion of the same cells is
different in the different periods of their activity, and hence to that
extent the two cases are similar.

In young scales in which the ganoin layer is still thin it forms an
even coat, and the striation visible in it is parallel to the surface. In
sections of older scales there are, near the edges, a series of notches in
the lower surface of the ganoin which conform to inequalities of the
upper surface of the underlying bony layers (Plate II. Fig. 16). Similar
conditions have already been described and figured by Williamson and
others. Figures 15 and 16 (Plate IL.) show that each notch marks the
point which was once the edge of the ganoin layer, and that the forma-
tion of ganoin on one side of this point (right, Fig. 16) and of bony
material on the other must have gone on for some time without any
lateral extension of the ganoin taking place. It is also evident that
extensions of this layer equivalent in amount to the distance between
successive notches must have taken place periodically, not by contin-
uous growth. The cause of such periodicity I have no means for
determining.

The fine striations (lamellation) in the ganoin which have already
beon mentioned have directions in this part of the layer slightly different
from those in the region farther from the margin of the scale, Instead
of being parallel with the upper surface, they are conformable to the
carlier surfaces of the layer, and so have a dip downward toward the
underlying bony material as they approach the edge of the scale. They
also gradually diminish in thickness toward the central area of the
scale, showing that the process of secretion went on less rapidly there
than it did nearer the margin of the scale.

The “tubes lepidines” of Williamson are clearly visible in ground
sections, though not present in cut sections of decalcified scales. They
are due to the presence of uncalcified connective-tissue fibres, as has
been stated by Klaatsch. These fibres in drying shrink, and so leave
minute spaces about them which the balsam does not enter (tbl. Ipd.,
Plate II. Fig. 15, and Plate IIT. Fig. 21). The course of these ‘ tubes”
is very characteristic, and is shown in Figure 15. They are absent in
the part of the scale immediately beneath the ganoin layer; they begin
in that part which was first formed, and from here they radiate, — the
directions being downward in the middle of the scale and diagonally
downward and outward near cither end of the section.

MUSEUM OF COMPARATIVE ZOOLOGY. 198

Klaatsch says concerning them (p. 129): * Untersucht man eine
Schuppe des jungen Lepidosteus in getrocknetem Zustande, nach Isola-
tion mit verdünnter Kalilauge, so findet man in ihr sehr zahlreiche zu
ihrer Oberflüche parallel verlaufende Röhrchen. Dasselbe gilt von
den senkrecht aufsteigenden Faserbündeln. Die Röhrchen, welche sie
in getrochnetem Zustande hinterlassen, sind von Williamson bei der
erwachsenen Schuppe als ‘tubes lepidines’ bezeichnet worden.”

As may be seen in Figure 15 (Plate II.), the greater number of these
“tubes” have an oblique direction ; none are exactly parallel to the
surface, though the ones near the ends of the section are nearly so, and
only a small part are vertical. Thus it is evident that their courses are
not the same as those of the fibres of dermis which has not undergone
calcification, neither do they agree with Klaatsch’s statement concerning
them. It is also noticeable that they do not have the same direction as
the dentinal tubules among which they lie, but that the two often cross
at considerable angles. Klaatsch does not account for this, neither am I
confident that I can explain the causes of the differences in direction to
which I have called attention ; but it would seem to be due to the
odontoblast cells — in common with a part of the other scleroblasts —
migrating before the advancing line of calcification in a direction per-
pendicular to the surface of the seale adjacent to them, such migra-
tion being however independent of the course of the fibres among which
they lic.

I find that the small scales from the under side of the lower jaw agree
with the description and figures of Hertwig (79, pp. 2, 3, Taf. I. Figg.
1-5). They closely resemble the larger scales in all essential respects,
differing from them chiefly in size and in having the form unmodified
by the proximity of adjacent scales. They are composed of a basal
bony plate, which is not destroyed by acid, of which the central part is
covered by a layer of ganoin, soluble in acid, about which there remains
a narrow marginal arca not covered by the ganoin, but bearing one or a
few spines. The number of the spines, upon which Hertwig laid empha-
sis, is, however, as has been stated concerning the larger scales, entirely
indefinite, and can be of no morphological importance. The scales
described by Hertwig as haying one spine owe this condition doubtless
to the others having been lost, not to the scale having been developed
as the basal plate of that single spine, as is the case in the placoid
scale. Such scales cannot therefore be considered as the complete
homologues of the placoid scales of Sclachians,

Mark (90, p. 11) has described the act of swallowing in the young

134 BULLETIN OF THE

gar-pike; this process gives a satisfactory explanation for the feebly
developed condition of the scales upon the under side of the lower jaw.
He observed that in the act of deglutition the floor of the mouth became
very much distended, so much so as to lead him to compare it to the
pouch of a feeding pelican. The frequent distention to which this part
of the skin is subjected by the greedy habits of the young fish furnishes
a sufficient reason why continuous plates of hard bony material are not
formed in this part of the dermis. The small size of the scales described
by Hertwig (79, p. 9) as lying at the bases of the fins is doubtless to be
explained in a similar manner. Instead of being atavistic conditions, as
maintained by him, they would appear to be due entirely to the action
of purely mechanical influences.

I have already spoken of the canals which pierce the central part of
the scale, and which have been called Haversian canals by Hertwig and
Klaatsch. Their usual course is from below directly through the scale
to the upper surface, where the vessels which traverse them spread out
in all directions through the sub-epidermal layer. There are sometimes
one or two vessels of considerable size which at the upper surface break
up into small branches, but quite as often the canal is filled by a large
number of minute vessels apparently distinct from one another. The
course of these canals is not, however, by any means constant. I have
found that sometimes, instead of running directly through, they extend
for some distance inside the scale parallel with the surface. Indeed, in
the scales of one fish (287 mm. long) the rule seemed to be for quite a
number of canals to run horizontally through the scales for considerable
distances. This condition seems to me to give an additional reason for
regarding these canals as Haversian, or perhaps better — since the
osteoblasts are not arranged in concentric lamelle about them — as
Volkmann’s canals. Figures 28 and 29 (Plate IV.) are from sections of
such scales. In the former figure the canal extends the greater part of
the length of the scale and opens to the surface at five points, near
those marked for, as can be seen in other sections not figured.

The surface contour shows some quite marked peculiarities in several
instances. Figure 30 (Plate IV.) represents a portion of a scale from
the same fish as the one last mentioned (287 mm. long). ‘The upper
surface bears numerous rounded elevations or hillocks, upon which for
the most part the spines are borne. The variations in shape and height
are indicated in the figure. Frequently the canals pierce the scale just
beneath these elevations and open to the upper surface at one side, as

indicated by dotted lines. These slightly abnormal conditions seem to

MUSEUM OF COMPARATIVE ZOOLOGY. 135

me to show that not only in histological structure, but also in surface
contour and in the degree to which Haversian (or Volkmann’s) canals
are developed, Lepidosteus scales stand somewhat more nearly related to
those of Polypterus than has been held to be the case.

The relation of the Ganoid scale as seen in Lepidosteus to those of
other groups of fishes remains to be considered. If my conclusions with
regard to this case be true, the idea that any Ganoids have enamel-
covered scales will have to be discarded, for Polypterus scales are so
nearly like those of Lepidosteus as to make a difference of origin
extremely improbable, and no other fishes of this group have scales so
closely resembling enamel in physical characters. What relation do the
scales of Lepidosteus bear to those of Sclachians? In the scales of these
two forms there is much that is unlike. In the former the basal plate
begins to be formed first, in the latter the spine is the part first to
appear; in the former the spines are many upon each scale, and they
are small and transient, in the latter there is only one to a scale and
they are large and persistent. In the former the scale plate contains
osteoblasts, Haversian canals, dentinal tubules, and three crossing sys-
tems of incorporated fibres ; in the latter only dentinal tubules (not in
all respects homologous with those of Lepidosteus) and in most cases :
few ends of the vertical set of fibres, though in the more highly devel-
oped cases three sets of fibres are found.

There is however one respect in which the two agree quite closely.
The spines are in all essential characteristics alike. Both have the tip
covered by enamel secreted by the basal epidermis cells; both have the
main part composed of calcareous secretion (dentine) of dermal cells
lying within and occupying a central cavity. Both have a system of
dendritic tubules extending from the cavity into the region of the tip,
and both arise by the calcification of the outside of a dermal papilla.
These must, thon, be taken as the fundamentally homologous parts, and
must serve as the basis for comparison.

In the Selachians the simple spine has remained as the typical
structure, and only in the more highly developed cases (Mustelus levis)
has the basal plate been developed to the extent of incorporating in
itself the fibrous dermis.

In the Ganoid scale two changes have taken place in the passage from
the condition in Selachians : —

1. The basal plate has increased in size and in complexity of or-
ganization until it has become the essential structure; not only has it

incorporated in itself the dermal fibres, but with them it has also taken

BULLETIN OF THE

and in connection

in the scleroblastic dermal cells to become osteoblasts,
it has developed a system of tubules for supplying
Moreover, instead of being formed simply as a
it has come

with this process
them with nourishment.
continuation of the process by which the spine is produced,
pine, for it is only in a late stage of its

to develop independently of the s
Thus the Ganoid scale plate seems

growth that the two become united.
to have arisen from the placoid basal plate by increase in size and with
important modifications.

2. The spine, on the other hand, has become reduced in size and in
Ganoids (Lepidostens) only a rudi-

complexity of structure, and is in
early, as is frequently the

mentary organ arising late and disappearing
case with degenerate. structures, the “ wisdom teeth” of man being a
familiar illustration of this.

Hertwig’s view is that the scales of Lepidosteus have arisen by the
fusion of numerous smaller basal plates of scales of the Selachian (pla-
coid) type. Each spine upon a scale of Lepidosteus therefore represents
a primitive placoid scale, and the whole Ganoid basal plate has arisen by
the fusion of as many simple scales as the total number of spines formed
Klaatsch objects to this interpretation, since the

upon its surface.
and because the

number of spines is so large and wholly indefinite,
spines lack such an orderly arrangement as that which the scales have
in selachians.

My own view in regard to t
expressed by the latter author.
regarded as a more highly developed basal plate
is not due to the fusion of many small ones, but
l in connection

his matter is essentially the same as that
Though the Ganoid scale must be
than that found in the
Selachians, its origin
rather to the calcification which in Selachians originate

with the formation of placoid spines, having become in
yon the impulse given by

Lepidosteus

an independent process no longer dependent uy
the growth of the spine. The hereditary tendency toward the growth
pille still shows itself, however, in the formation

and calcification of ]
hose aro retarded in time and but feebly

of the small spines, though t
developed.

As long as each spine had a basal plate, as in Selachians, the spatial
requirement of this plate exercised a, controlling influence upon the num-
spines). When now, as in

ber and the arrangement of the scales (=
independently of the under-

Lepidosteus, the spines have come to arise
lying plate, such restraint is removed, and we consequently find an in-
crease in the number of the spines and a lack of regularity in their

arrangement .

MUSEUM OF COMPARATIVE ZOOLOGY. 137

Upon the development of the scales of Teleosts I have made no obser-
vations, but employ for comparison with the process in Lepidosteas the
accounts given by Klaatsch and Hofer.

In the earliest stages of development a similar modification of the
dermis takes place in both cases, giving rise to local thickenings of this
layer within which the scale begins to be formed as previously described.
These I hold to be homologous structures, but not the homologues of the
dermal spine papillie of Selachians.

The resulting bony plates formed in the two cases are homologous but
that of Lepidosteus attains to a much higher degree of development
than the one formed in Teleosts. The part to be formed first in the
Teleost scale is, as in Lepidosteus, the outer more homogeneous part,
but in many cases it differs from the corresponding part of the scale of
Lepidosteus in the absence of enclosed osteoblasts. That these are
present in some species and absent in others which are very near
relatives shows that this difference cannot be of any great morpho-
logical: importance. Whether they have been secondarily acquired in
one case or secondarily lost in the other need not concern us here.
Their absence in some cases cannot prevent the layer being considered
the homologue of the corresponding layer in Lepidosteus.

The deeper fibrous part of the Teleost scale is the later formation, and
in this respect, as well as in general structure and method of develop-
ment, agrees with the deeper part of the scale of Lepidosteus.

The outer layer of the Lepidosteus scale, the ganoin, which is the part
latest formed, is absent in the scales of Teleosts. It is a layer which has
been developed within the order of Ganoids, and is not found in any
other group of vertebrates.

Throughout the series of scale structures beginning with the Selachian
type there has been a constant tendency toward reduction of superficial
parts (spines) and increase of the deeper parts which are independent of
the epidermis, In Selachians the process of scale formation begins at
the surface of the dermis just beneath the basement membrane. In
Ganoids there is the same process repeated at the base of the epidermis,
but in a much less vigorous manner, while. the principal activity is
deeper-seated, in the midst of the dermis. In the higher Teleosts the
whole scale growth is within the dermis, and the more superficial
process is entirely lost.

Thus I believe that the basal plate of the scale of Lepidosteus and the
Teleost scale have both been derived from the basal plate of the placoid

scale, and have for the most part been modified along the same lines.

138 BULLETIN OF THE

That the general course has been the same in both cases is shown by
the fact that a broad flat bony plate composed of two layers, an outer
more homogeneous and a decper fibrous one, has resulted in both cases,
that these have developed ontogenetically in a very similar manner, and
have come to overlap one another in similar diagonal rows. This has,
in both cases, involved the reduction of the spines, which in some of
the lower,Teleosts (e. g. Siluroids) and in Lepidosteus are present in a
degenerate condition, but are absent in the greater part of the Teleosts.
That in some of the lower Teleosts (e. g. Hypostoma) they do not fuse
with the basal plate, but are joined to it by connéctive-tissue fibres only,
may be considered an evidence that degeneration has here gone a step
farther than in Lepidosteus.

Klaatsch’s idea that the upper layer (Hyalodentin) of the Teleost
scale is homologous with the ganoin layer cannot be true, for the two
develop quite differently and their physical and chemical properties are
very unlike. If, as I believe, Klaatsch did not see the true ganoin
in Lepidosteus, but mistook for it the outer part of the bony scale below,
his conclusion in regard to the homology of this layer is in perfect
accord with that which I have expressed.

If my conclusions concerning the origin and nature of the outer layer
(ganoin) of the scales of Lepidosteus is correct, it will follow that no
very close and direct relationship can exist between these scales and
teeth. Their only relationship is such as arises from the fact that they
are both derived from an ancestral condition similar to that found in the
scales of Selachians. This primitive condition has been modified by
changes leading in opposite directions. In the mouth, the spines have
been developed to form the teeth ; on the surface of the body, the basal
plate has given rise to the scales. Thus cach represents at present only
a highly modified part of the early ancestral prototype.

o(

MUSEUM OF COMPARATIVE ZOOLOGY. Las

LITERATURE.

Agassiz, Louis.
'83-45, Recherches sur les poissons fossiles. Tom. I. et IL, and Atlas.
Neuchâtel.
Hertwig, Oscar.
79. Ueber das Hautskelet der Fische. Zweite Abtheilung : Das Hautskelet
der Ganoiden (Lepidosteus und Polypterus). Morph. Jahrb., Bd. V. p. 1.
Hofer, Bruno.
"90. Ueber den Bau und die Entwicklung der Cycloid- und Ctenoidschuppen.
Sitzungsber. d. Gesellsch. f. Morph. u. Physiol., München, Bd. VI. p. 103.
Klaatsch, Hermann.
"90. Zur Morphologie der Fisehschuppen und zur Geschichte der Hartsub-
stauzgewebe. Morph, Jahrb., Bd. XVI. pp. 97 and 209.
Leydig, Franz.
"54. istiologisehe Bemerkungen über den Polypterus bichir. Zeitschr. f.
wiss. Zoologie, Bd. V. p. 40.
Mark, E. L.
'90. Studies on Lepidosteus. Part I. Bull. Mus. Comp. Zool., Vol. XIX.
p 1.
Reissner, E. W.
'59. Ueber die Sehuppen von Polypterus und Lepidosteus. Arch. f. Anat.,
Physiol. u, wiss. Mediein, p. 254.
Williamson, W. C.
'49. On the Microscopie Structure of the Scales and Dermal Teeth of some
Ganoid and Placoid Fish. Philosoph. Trans. Roy. Soc. London, p. 435.

EXPLANATION OF PLATES.

All figures were drawn with the aid of an Abbé camera lucida, from specimens
of Lepidosteus osseus, L.

ba. spi.
can
chr’ph.
drm f’br.
drm. spng.
Jue,

Jac. e’drm.
Jor.

gan.

gin

gin?

mb. ba,

ABBREVIATIONS.

Base of spine.

Canal. nl.
Chromatophore. od’ bl.
Fibrous portion of dermis. pap. spi.
Spongy sr Hu pig.
Surface. 8c.
Surface of ectoderm. sc bl.
Foramen. spi.
Ganoin. tbl. de.
Gland, first kind. tbl. Ipd.
Gland, second kind. vs. sng.

Basement membrane.

mb. gan.

Ganoin membrane.
Nucleus.
Odontoblast.
Spine papilla.
Pigment.

Scale.
Scleroblast.
Spine.

Dentinal tubule.
* Lepidine tubule.
Blood-vessel.

”

NICKERSON — Scales of Lepidosteus.

a

6.

PLATE I.

Longitudinal section of skin of young gar-pike, 145 mm. long X 140.

Mucin gland cells of ectoderm ; a-d, stained with Kleinenberg's hama-
toxylin; e, stained with Boehmer’s alum hematoxylin; a-d represent
successively younger stages. X 500.

Longitudinal section through posterior end of very young scale from fish
150 mm. long. X 262.

Optical section of ectoderm, parallel with surface ; from tissue stripped
off from surface of scales of fish 300 mm. long, after decalcifying
the chromatophores lie at a deeper level than the gland spaces, and
therefore appear at different focus. X 110.

Section from fish 150 mm. long, showing two stages in the formation of
spine papilla; the section is not quite parallel to axis of larger
papilla. X 370.

A part of the section next to that represented in Figure 5, showing
summit of larger spine papilla and its peculiar shaped secreted tip
x is line in which the farther plane of the section cuts the basement
membrane. X 370.

A longitudinal section through a spine to show the canal connecting its
central cavity with the exterior. X 100.

Odontoblast with its process extending into dentinal tubule. From lower
surface of scale of fish 460 mm. long. X 500.

Another odontoblast, having two nuclei. Conditions as in Figure 8.

ete mig:

seti. se

bbc

Pay m $ 1
od bf a

y tb den

papspi.

NICKERSON. — Scales of Lepidosteus.

10:

Ll,

PLATE IL

Section through the axis of a spine papilla in which secretion has begun.
From fish 150 mm. long. X 887.

Optical section of spine broken off from scale of a young Lepidosteus
300 mm. long. x 387.

Tip of spine loosened by use of caustic potash ; a, before treating with
hydrochloric acid; b, after applying acid. X 137.

Section through posterior edge of ascale from a fish 300 mm. long (52
months old), showing portions of imbedded bases of four spines. X 200,

Base of alost (resorbed?) spine, which had been surrounded by ganoin
before the latter was destroyed by decalcification. From fish 460 mm.
long. x 198.

Bection (prepared by grinding) through scale of fish 300 mm. long, from
S P. Bartlett. x 31

Section through posterior margin of decalcified scale from fish 300 mm.

long. x 198.

1 tht de, ‘ al à
i r p à a
i x j
& `
| ie E IR
Nis €
i |
UA
| Y
|
\ 10
É 1 È
tl ke
nom x
" bi
es tbi de
m 7 [X h
/ | ,
[74 A y “
| 1
if
se
pi P
e.
v - i *

NICKERSON. — Scales of Lepidosteus.

PLATE II.

Section through axis of a spine showing the difference in appearance
between the superficial bony layers of the scale and those lying
deeper. From fish 190 mm. long. X 370.

An abnormally shaped spine tip. X 370.

Fibrillar appearance of the under surface of a scale from a fish 287 mm.
long, after treatment with caustic potash. X 275.

Section through anterior edge of a scale from a fish 190 mm. long ; show-
ing fibres of connective tissue in the scale. X 275.

Portion of ground section of scale of fish 800 mm. long. X 152.

Portion of section of scale and overlying tissue from a fish 460 mm. long.
x 198. In a ground section of an immediately adjacent scale the
ganoin layer on the corresponding part was 12-15 u thick. Both in this
and the succeeding figure (Plate IV. Fig. 23) an odontoblast is shown
lying upon the upper surface of the scale and sending its process down

through the ganoin layer in a dentinal tubule.

22

dui BIS NBI wlan.
Le A" M ig
eme teca RÀ nna

Gir

~od bl.

on
de
gen. El
m
p, an.

i i in i

\ N

li E

|

í

fc. n

PI

I del

i,

NicksRson, — Scales of Lepidosteus,

Fig. 23.

no
ES

PLATE IV.

Portion of section of scale and overlying tissue from a fish 460 mm.
long. X 862. See description of Figure 22, Plate IIT.

Base of spine beginning to be imbedded in the outer layers of the scale.
From fish 190 mm. long. X 370.

An oblique section through the base of a spine partly imbedded in the
scale. From a fish 300 mm. long. x 887.

Section through a spine having an abnormally shaped tip. X 105.

Surface view of posterior extremity of a scale freed from soft tissues by
treatment with caustic potash, showing loosened cap on point of
spine. X 75.

Longitudinal section through a scale of a fish 190 mm. long. X 44.

Longitudinal section through a scale of a fish 287 mm. long. X 25.

A representation of a portion of the surface of a scale of a fish 257 mm.
long. The dotted lines show the positions of Haversian (or Volk-
mann’s) canals beneath the surface. X 31.

Outlines of several scales of a fish 300 mm. long. The dots indicate the

positions of the spines on four scales. X 6%.

130

| mb ba.
m get.

/ 3 for
l i : 29

Vs png. gan.

per. ás
N fer.
dul y for 28
TA
\
m
& I
j * M
s , eoo,
wr
| \
| \
| \
/ | |
j
i 1
\
{ ka - Loa. PLA
j $
|
FL
À \ \
4
| A
4
N

No. 6. — Studies in Morphogenesis. — I. On the Development of the
Cerata in Molis. By C. B. DAvENPOnT.!

Waite at Mr. Agassiz’s Newport Laboratory last summer, I collected,
early in August, some spocimens of /Eolis? which oceurred abundantly
on the ITydroids. The individuals varied greatly in size, and, since Nu-
dibranch eggs had been abundant on the same Hydroids earlier in the
season, I regarded most of them as immature.

I was particularly struck by the regularity with which the cerata, or
dorsal papillee, were distributed over the surface of the body, and the
constancy in position of young cerata in relation to the older ones.

Figures 1 and 2 are drawings of two individuals showing the arrange-
ment of the cerata. They are placed in transverse rows, which are not
equidistant, however, but grouped in twos or threes. In each transverse
row the largest ceras lies dorsalmost, the smaller more ventral, the
youngest of all being often indicated by only a slight protuberance at
the base of and ventral to the next older. The regularity of the pro-
cess of budding of new cerata induced me to make sections to determine
its details,

The cerata of Æolis contain, as is well known, processes of the ali-
mentary tract, — the so called hepatic coa (Plate II. Fig. 12). They
are covered externally by the ectodermal epithelium. Between ectoderm
and hepatic cocum are mesodermal cells, which lie (1) in a thin layer at
the inner surface of the ectoderm, (2) in a thin layer over the cocum,
and (3) between these two layers, surrounding blood spaces or loosely
distributed.

The coecum of each ceras connects at the base with a lateral diverticu-
lum of the alimentary tract, — gastric diverticulum, — with which also
the coca of all of the other cerata of the same transverse row unite
(Plate I. Fig. 3, ga. dv.). There are as many of these gastric diverticula
as there are transverse rows of cerata (Fig. 4, ga. dv.).

1 Contributions from the Zoólogical Laboratory of the Museum of Comparative
Zoology, under the direction of E. L. Mark, No. XXXVII.

2 Owing to the immaturity of the individuals, I was unable to determine the
species accurately. Figures 1 and 2, however, show the external form of the
species (one or two) employed.

VOL. XXIV. — NO. 6,

142 BULLETIN OF THE

Distal to the coecum, at the apex of the ceras, lies a sac which in the
adult opens to the exterior and is connected at its proximal end with
the coecum in such a manner that the walls of the two organs are con-
tinuous, and their cavities confluent through a communicating canal.
The existence of this communicating canal, although in past times called
in doubt, has recently been correctly reaffirmed by Herdman (’90, p. 52).
The walls of the sac are composed of almost completely vacuolated cells,
which contain nettling organs or nematocysts. Since the cells of such
sacs contain nematocysts, the sacs are called nematophores or cnido-
phores. Cnidophores are characteristic of the /Eolide. Figure 13 rep-
resents a cross section through a cnidophore, showing the vacuolated
cells and the nematocysts therein (nt’cy.).

The first indication of the formation of a new ceras is a thickening of
the mesenchyme at the base of a young ceras and upon its ventral aspect
(Plate I. Fig. 5, ms'chy.”). At a slightly later stage (Fig. 6) the mesen-
chyme has become greatly thickened, and a protuberance of the ectoderm
has occurred. Karyokinetic figures indicate that the mesenchymatous
mass is growing by cell proliferation. The growing mesenchymatous
cells, as well as the adjacent cells of the gastric diverticulum, stain
more deeply than those of other regions.

At a slightly later stage (Fig. 7, III.) the wall of the gastric diver
ticulum (ga. dv.) has begun to protrude into the thickened mesenchyme,
and the ectoderm is sharply evaginated. Still later (Fig. 8, IV.) these
features become, more pronounced. The mesenchymatous cells become
arranged into three or four layers, of which one is closely applied to the
ectoderm, and another to the hepatic ccecum.

It is important to note, (1) that the ectoderm of the new ceras lies, at
at an early stage, in the angle made by the body wall with the ventral
wall of the next older ceras; (2) that the mesenchyme is directly con-
tinuous with that of the ceras, and is in fact transitional from the meso-
derm of the trunk to that of the ceras; and (3) that the budding
entoderm lies at the distal extremity of the gastric diverticulum, that
is to say, at the angle made by the gastric diverticulum, and the last
ceras arisen therefrom (Fig. 3, cer. 3). The mother cells of each of the
three layers of the incipient ceras thus lie in the outer margin of the
region whence the corresponding layer of the last formed ceras has
arisen. With the development of the ceras it gradually becomes farther
removed from the next older one (Fig. 3).

. The next stage figured (Plate IT. Fig. 9) shows a ceras whose length
is slightly greater than its diameter at the base. The mesoderm is

MUSEUM OF COMPARATIVE ZOOLOGY. 143

greatly flattened, and the entodermic wall of the coocum is composed of
cuboidal deeply stainable cells. Ata stage at which the length of the
ceras is less than twice its greatest diameter (Fig. 10), a circular fold
occurs in the distal third of the coocum, constricting the lumen. The
cells of the distal compartment stain deeply, like those of the hepatic
cecum of the preceding stage. They do not at all resemble the un-
stained cells of the eotoderm.

At a stage only very slightly older than the preceding (Fig. 11), one
finds the cells of the distal sae partly vacuolated and containing nema-
tocysts. This distal sac can be traced forward into the oldest stages,
and it becomes the enidophore. Its lumen constantly retains its con-
nection with that of the hepatic ccecum through the communicating
canal. A communication with the outside world at the apex of the
ceras is established only at a later stage by a close approximation of the
enidophore to the apex of the ceras and a disintegration of the apical
ectodermal cells.

Herdman asserts in two or three places (’90, p. 52, and Herdman
and Clubb, '89, p. 233, 792, p. 552), that the cnidophores arise by an
invagination of the ectoderm on the apex of the ceras; but although in
his last paper he says he has “shown” it in the earlier ones, I fail to
find that he has offered the slightest evidence for his statement. It
seemed indeed a priori more probable — from what wo know of the
origin of protective organs, and especially of the origin of nematocysts
in the Cnidaria— that the cells of the cnidophore had an ectodermal
origin. Butthey have not. This is the conclusion to which I am forced
by the following considerations. (1.) The development just outlined,
which has been traced in a series so complete as to leave little chance
for misinterpretation. (2.) The absence of an external opening until
quite a late stage. Since the axis of the cnidophore does not in later
stages coincide with that of the ceras, and since it is not easy to obtain
sections which pass through the entire axis of the ceras, especial care
must be exercised in determining the absence or presence of an apical
opening. Figure 12 is a strictly axial section. The apical ectodermal
cell shows signs of degeneration, and its outer surface is sunken in.
(3.) The presence of nematocysts in those cells also which lie in the
hepatic cocum proximal to the constriction, and for which no one has
maintained a derivation from an invagination of the apical ectoderm.

The cells of the hepatic ccecum, especially at a late stage, show large
numbers of nematocysts of the two kinds mentioned by Herdman ('90).
Cf, Figures 14 and 15. That these nematocysts have been developed

144 BULLETIN OF THE
in situ, and have not migrated from the cnidophore through the. com-
municating canal and become incorporated into the “hepatic” cells,
is indicated by the fact that they are found in different stages of de-
velopment (Fig. 14, né’cy.’). One finds first of all small ovoid areas,
which stain deeply ; later, one finds an elongated deeply staining central
streak, while the rest of the vesicle remains colorless. This central
streak is the proximal part of the thread. In some cases I have seen in
the “hepatic” cells the larger kind of nematocyst mentioned by Alder
and Hancock ('55, Expl. Fam. 3, Plates 7 and 8, Fig. 16); in fact,
they seem to occur quite as abundantly here as in the cuidophore,

Comparing the origin of nematocysts in the AMolide and Hydra, — for
a knowledge of which in the latter group we are indebted among others
to K. C. Schneider (90, pp. 332, 345), —the most striking difference is
that the nettle capsules in Hydra arise in the ectoderm only (so it is
maintained), and that those which do occur in the entoderm have not
been formed there, but have been devoured. In /Eolis, similar capsules
arise in diverticula of the alimentary tract, and therefore from entoderm.
Secondly, the nettling capsules of Hydra are formed in indifferent cells
lying at the base of the ectoderm ; in /Eolis they are produced in the
large “hepatic” cells. Finally, in Hydra only one nematocyst is formed
in a single cell, and this comes to occupy nearly the whole of it; in
Rolis several nematocysts (Figs. 13, 14) are produced in a single cell.

This comparison raises the question whether the so called hepatic cœ-
cum is properly named, — whether its cells have either the function of
storing up reserve stuff or of secreting digestive fluids. To this question
I can give no final answer, but it is worthy of note that the “ hepatic "
cells of /Eolis do contain numerous small granules which stain deeply in
hematoxylin. This is true even for the cells which produce nematocysts
(Figs. 14,15). A comparison with sections of the hepatic cœca of Doto
coronata treated in the same way shows, however, that the “hepatic”
cells of Afolis are smaller and contain smaller granules than do those
of Doto, in which genus, of course, no nettling organs are formed.
From their smaller size, and the smaller size of the contained granules
(whether these are stored food stuff or katabolie product), I am inclined
to regard the hepatic function of the cells forming the wall of the cocum
of /Eolis as less important than in Doto.

The sum total of the nematocysts in the hepatic cocum of any ceras
is much greater than in the enidophore. On physiological grounds the
latter term is in /Eolis applicable to the entire cecum, and not merely

to its distal part.

MUSEUM OF COMPARATIVE ZOOLOGY. 145

After having found that the cerata arose in close connection with pre-
ceding ones of the same transverse rows, and therefore ultimately in
close connection with the first or dorsalmost ceras of each row, it seemed
desirable to determine the origin of the dorsalmost longitudinal series of
cerata, and of the origin of new transverse rows. Sagittal, or slightly in-
clined longitudinal, and also frontal sections of young specimens gave the
desired information. Figure 16 represents a longitudinal section which
is not strictly sagittal, the upper edge of the section plane having been
tilted about 30° towards the right so as to pass at the same time through
the alimentary tract and the dorsal series of cerata, — the only series
as yet developed on that side. From the drawing it is clear that the
alimentary tract sends off diverticula, which pass directly into the first
cerata of each transverse row. Behind the most posterior ceras the
mesenchyme is thickened over the alimentary tract, which runs to the
posterior end of the body. In a section a little removed from this
(Fig. 17, I1.), one sees the beginning of another papilla behind number
Ill. of the series represented, — the foundation of another transverse
row. The ectoderm has already begun to fold upwards, and the alimen-
tary tract sends out a prominent pocket.

Of very great, importance is the fact that just behind the Anlage of the
ceras the testis is arising, so that the mesenchyme which seems to take
the initiative in the formation of the ceras gives rise also to the sexual
cells. Moreover, the two cell masses arise close together, and indeed in
a definite relation to each other. As is clear from an inspection of
Figure 16, the sexual glands — ovary and testis — lie between the trans-
verse planes occupied by the cerata, and in each mass the ovary lies in
front of the testis, so that successive transverse sections cut from the
head backwards pass in order through a transverse row of cerata, through
an ovarian mass, and through a testicular mass. This succession is,
however, not that in which the Anlagen of the three organs have been
established out of the mesenchyme of the tail end ; for, as Figure 17 in-
dicates, new sexual cells arise before the ceras which lies in front of them
begins to appear.

The oblique (nearly frontal) section, Figure 18, shows the same relation
of the sexual glands and the cerata. A reconstruction of the series shows
that diverticula arise from the parts of the alimentary tract indicated by
the designations cer. I, IL, HL, and IV. These diverticula correspond
in position to transverse rows of cerata. The same reconstruction shows
that between these diverticula ovary and testis follow in the way just

described. At the extreme tail end the section passes obliquely through

146 BULLETIN OF THE
the glands of the foot, then through the mass of indifferent mesenchyme
which lies under the dorsal ectoderm at x. Just in front (IV. Y) sexual
cells are being cut off from the mesenchyme as a paired mass whose two
lobes are united in the median plane.

I have above assumed, somewhat gratuitously, that the mesenchyme
takes the initiative in ceras production. The evidence for this lies in
two facts. (1.) The first indication of the formation of the new ceras
is seen in the thickening of the mesenchyme at the base of the next
older ceras (Fig. 5). It is not until after a solid mass of mesenchyma-
tous cells is produced that the ectoderm begins to evaginate, almost as
though pressed outwards (Fig. 6). The alimentary diverticulum is pro-
duced still later (Fig. 7). (2.) That the coocum does not take the ini-
tiative is indicated by the fact that I have found young cerata composed
only of ectoderm and a thickened mesenchymatous core, the entoderm
not having yet penetrated into it.

The capacity possessed by Nudibranchs of regenerating the cerata is
well known. I have not experimented with them, and have no sections
of stages in the process. The known phenomena of regeneration in other
cases makes it probable that the capacity for regeneration depends upon
the existence of embryonic tissue. We should therefore expect to find
thickened, embryonic mesenchyme lying at the base of the dorsal papilla.
As a matter of fact we do find it, as is shown in Figure 16 at the base of
cerata IT. and III. (>). The mesenchyme at the base and in front of
ceras L was torn away in sectioning; in adjacent sections the basal
mesenchyme appears thickened here also.

The foregoing study of the development of the cerata of /Eolis points
emphatically to one conclusion, namely, the embryonic or growth tissue
of /Rolis is in its origin identical with that producing sexual cells. Like
the latter, it is germ tissue; it differs from the sexual cells chiefly in this,
that it gives rise to growths constituting part of the body of the present
individual, — growths which are as mortal as any other part of the pres-
ent individual ; whereas the sexual cells play no part in the production of
the present individual, but eventually give rise to a new individual and
its germ tissue. It differs, secondly, from the sexual cells in this, that it
gives rise to one kind of organ only, —the mesenchyme to the mesenchyme
of the buds, the entodermal diverticulum to the entoderm of the buds.

MUSEUM OF COMPARATIVE ZOOLOGY, 147

The mesenchymatous growth tissue of /Eolis resembles the sexual cells,
however, in this, that while it goes to produce the mesenchyme of any
ceras, a, not all of it is used up in forming the mesenchyme of ceras a,
but some of it remains behind to form a new ceras, b, and the Anlagen
of other new cerata. Thus, as in any young individual we may distin-
guish between the differentiated tissue and the germ tissue from which
new individuals will arise, so in any ceras we may distinguish between
the differentiated tissue and the embryonic tissue from which new cerata
will arise.

While, however, the sexual cells have the capacity of reproducing new
individuals indefinitely, the mesenchyme at the base of the cerata does,
as a matter of fact, produce only a limited number of cerata. Of this
limitation there are, however, all degrees. In some cases, as in Doto,
only one ceras is produced in a transverse row; in some species of /Eolis,
on the other hand, young cerata are produced, even in adult individuals,
at the ventral end of the long transverse rows, so that here the growth is
apparently limited only by the duration of life of the individual. In all
cases the limitation'in the reproduction of cerata must be considered as
resulting, not from the limited capacity of reproduction of the embryonic
tissue, but from the needs of the species.

CAMBRIDGE, December 20, 1892.

148 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

LITERATURE CITED.

Alder, J., and Hancock, A.
'55. A Monograph of the British Nudibranchiate Mollusca. London: Royal
Society
Herdman, W. A.
"90. On the Structure and Functions of the Cerata, or Dorsal Papille, in some
Nudibranch Mollusca. Quart. Jour. Mier. Sci., XXXI. 1, p. 41.

Herdman, W. A., and Clubb, J. A.
’89. Second Report on the Nudibranchiata of the L. M. B. C. District.
Proc. Liverpool Biol. Soc., III. p. 225.
'91. On the Innervation of the Cerata of some Nudibranehiata. Quart. Jour.
Mier. Sci., XXXIII. 4, p. 541.
Schneider, K. C.
'90. Histologie von Hydra fusca mit besonderer Berücksichtigung des Ner-
vensystems der Hydropolypen. Arch, für mikr. Anat., XXXV. p. 321.

DAVENPORT, — Cerata in Molis.

PLATE L

ABBREVIATIONS.

an. Anus. ya. Stomach.
can. comn. Communicating canal. ga.” Prolongation of stomach back-
cer.1,2,etc. Cerata of the same trans- wards.

verse series. ga dv. Gastric diverticulum at base of
cn’ph. Cnidophore. transverse rows of cerata.
ca. Hepatic cecum. ms'chy/ Mesenchyme of developing
cæ! Forming hepatic coecum. ceras.
epl. Capillary vessel. CA (Esophagus.
ec'drm. Ectoderm. pd. Foot
ec'drm.^ Ectoderm of young ceras. rt. Rectum.
en’drm.’ Entoderm of young ceras.

Figs. 1,2. Two figures of young individuals of /Rolis sp., showing arrangement

of cerata in transverse rows, which are numbered with Roman
numerals. X 5.
3. Transverse section of a young JEolis showing the stomach, the gastric
diverticulum, and hepatic coca of a transverse row of cerata. X 45.
4. Outline of alimentary tract of ZBolis, copied from Alder and Hancock, show-
ing gastric diverticula and points of connection of hepatic coca.
5-8. Four vertical sections through different stages in the development of the
cerata.
First stage in formation of ceras. Mesenchyme thickened at ms'chy.
Right side. X 405.
6. Second stage in formation ot ceras. Mesenchyme thiekened still further.
Ectoderm beginning to evaginate in the axis. Left side. X 405.
7. Third stage in formation of ceras. Gastric diverticulum beginning to out-
fold to form hepatic cocum of new ceras. Left side. X 228.
Fourth stage in formation of ceras. Right side. 28,

$ e A
yr NS A 1 17
| p Vur Li X y N M 1
A
P p a att n \
~ H
7 N Pi] "m j
€ ~ ti { wo vm
E }
TUR
cert ANS
f ys et x
\ i
3
rt. é
Loy) cont
ye
gu. ya ene m
(Sr 4 oy
( ^ f \ ge etr, \\
/ i 1 wi ww UST ` {
ri ^ N
y 15) Pp + TP
| Y y PO i re
3
$i 4 ! b $ it
dt " A \
NT. d pe. Y 3 {
( : x OY N
pe 3 jf \
ge
ge I»
| Ng gado l Ù
gaol. 6. rs
K "m
moy
ma chy! 3 S \
$5 5
* LÀ
* J. " Ve
N qu
y en pl
pee, er
COC pet cn» "Hn
, e SN
\ . "
| \
| b > ;
| $
| 7 $ ee
ejos
ey *
n D. ¢

can, comn.

Davenport —Cerata in Molis.

PLATE II.

ABBREVIATIONS.

Communicating canal.

cer.L,IL,ete. Cerata of I, IL, etc.

cn'ph.

co.

ec'drm.

ga.

ga.

ga. dv.
gl. nid.
gl. pd.

gn.

mo, buc.
ms’drm.

Figs. 9-11. Three stages in the development of the cerata. Vertical sections

12.

13.

14.

15.

16,

17.

18.

transverse rows.
Cnidophore.
Cocum.
Ectoderm.
Stomach.
Posterior prolongation
of stomach,
Gastric diverticulum.
Nidamental gland.
Foot gland.
Ganglion (cerebral).
Buccal mass.
Mesoderm.

x 120.

Longitudinal section of ceras just before formation of external opening of

enidophore. X 120.

Transverse section of adult enidophore, showing nematocysts. X 405.

Bit of transverse section of hepatic cocum, showing fully formed and
developirig nematocysts in hepatic cells. X 405.

Bit of transverse section of hepatic cæcum, showing large nematophore.

x 405.

Longitudinal section of young JEolis, inclined about 30% from vertical,
passing through axis of alimentary tract and first ceras of three trans-

verse rows. X 76.

Enlarged view of the posterior of the two regions marked by asterisks in
Figure 16, but from a section adjacent to Figure 16, showing parietal
mesoderm, some of which is about to form that of the youngest trans-
verse row, and some to give rise to sexual cells, te. X 318

Longitudinal section of young Æolis. Oblique in front, nearly frontal
behind. Showing repetition of sexual organs and their relation to

x 76.

transverse rows of cerata.

ms'drm.^

ms’drm.”

nl'cy.
nt'cy.^
nt'cy."
pd.

pe. ves
rad.
rhn.

rt.

ur

L, IL, HL, 9, d. Lobes of ovary and

Mesoderm lining body wall.

Mesoderm lining hepatic
cocum.

Nematocyst, smaller kind.

Developing nematocy st.

Nematocyst, larger kind.

Foot.

Penis sac.

Radula.

Rhinophore.

Rectum.

Testis.

Ureter.

testis.

gay.

"Ls en.
^N ) BE.
Y

nt

Sgenid,

Ga.

ps hour

ico? —

No. 7. — Reports on the Dredging Operations off the West Coast of
Central America to the Galapagos, to the West Coast of Mexico,
and in the Gulf of California, in Charge of ALEXANDER AGASSIZ,
carried on by the U. S. Fish Commission Steamer “ Albatross,”
during 1891, LigUT.-COMMANDER Z. L. Tanner, U. S. N., Com-
manding.

[Published by permission of Marshall McDonald, U. S. Fish Commissioner.]

VI.

Preliminary Descriptions of New Species of Crustacea.’ By Warrer
Faxon.

Orpver PODOPHTHALMIA.
SUBORDER DECAPODA.
Family MAIID 2.

Huprognatha granulata, sp. nov.

Carapace coarsely granulate ; two erect blunt spines in median line of cara-
pace, one on the gastric, the other on the cardiac area ; a transverse row of four
or five small tubercles in front of the gastric spine ; a spine near the middle of
each branchial area, with a smaller lateral spine below and a little in advance
of it ; the hepatic area bears a short blunt spine on its most prominent part ;
there are, besides, three or four prominent tubercles on the angle which divides
the hepatic and pterygostomian regions. The antennary spine is very long,
reaching beyond the rostral horns; the three horns of the rostrum (interan-
tennulary and lateral) are about equal in length; the supra-orbital spines are
well developed, and, like the antennary and three rostral spines, are conspicu-

1 Figures of all the species described in this paper will appear in the final report,
whose publication is deferred pending the preparation of the lithographic plates.
A record of the stations occupied by the “ Albatross ” during her cruise of 1891
will be found in Bull. Mus. Comp. Zoól., Vol. XXIII. No. 1, pp, 4-8.

VOL. XXIV. — NO. 7. 1

BULLETIN OF THE
ously granulated. The post-orbital spines are even more coarsely tuberculated.
When viewed from above their margins appear laciniated. Surface of abdo-
men thickly set with bead-like tubercles ; the first segment bears a prominent
granulated spine, and there is a rudimentary spine on each of the three follow-
ing segments. Chela slender, covered with small tubercles; the remaining
segments of the cheliped and the ambulatory appendages are furnished with
small spines, tubercles, and scattered curled sete.

Length of carapace, 7 mm.; breadth, 6 mm.

Station 3369. 52 fathoms. 2 females.

Anamathia occidentalis, sp. nov.

Carapace pyriform, furnished with spines and tubercles which are arranged
as follows: four on the gastric region (two in the median line, two lateral) ;
the posterior median has.the form. of a tubercle, from which a low blunt keel
runs back to the cardiac region; one on the cardiac region, one (tubercle) on
the intestinal region, one on each hepatic region, five on each branchial region.
Of the branchial spines the one near the middle projects upward and forward,
the longest spine of the carapace, one half as long as the rostral horns. In ad-
dition to these spines and tubercles there are four or five small tubercles on
the outer border of the pterygostomian region, Rostrum produced into two
divergent awl-shaped horns, which are more than one fourth the length of the
carapace. Pre-orbital spines acute, post-orbital processes obtuse. Chelipeds
twice as long as the carapace, not much more robust than the ambulatory
limbs; chela little longer than the merus, the basal part cylindrical, distal
part gradually widening to the base of the fingers; fingers smooth, slightly
curved, less than one half as long as the basal portion; prehensile edges
regularly dentate, closing throughout their length. The merus of all the
legs has a small tubercular projection at the distal end above, most prominent
on tne anterior pair. The carpus of the chelipeds has two low ridges on the
outer face. Abdomen seven-jointed.

Length of carapace without the rostral horns, 45 mm.; length of rostral horns,
12 mm. ; breadth of carapace, 38 mm.

Station 3404, 385 fathoms. 1 male,

MAIOPSIS, gen. nov.

Carapace subtriangular, as broad as long, spinose ; rostrum produced ante-
riorly into two divergent horns with an accessory spine upon their outer
margins; interorbital space broad. Eyes small, eyestalks slender, retractile
within the orbits. Orbits large, with a forward aspect, incomplete below, the
upper margin prominent, with two deep fissures, and supra-ocular .spines.
Epistome short. Basal segment of antenne very broad, with three prominent
spines upon its anterior margin; flagellum of the antenn widely separated

MUSEUM OF COMPARATIVE ZOOLOGY. 151

from the cavity of the orbit by a broad process of the bail segment. Merus
of outer maxillipeds notched at the antero-internal angle. Legs of moderate
length ; carpus of chelipeds elongated, not carinated; chela elongated and
slender, fingers canaliculate within, but not spoon-shaped at their tips, their
prehensile edges meeting throughout most of their length, not distinctly
toothed ; ambulatory legs spinose. Abdomen (male) seven-jointed, terminal
joint short and broad.

This genus is proposed for the reception of an interesting Maioid dredged
in 182 fathoms on the west coast of the Isthmus of Panama. It combines
in one form characters of the genera Maia, Cyclomaia, Paramithrax (subg.
Leptomithrax), and Schizophrys. In its general aspect it bears a close resem-
blance to Maia, and the likeness is enhanced by the small eyes and slender
eyestalks, the elongated wrist and hand, and the short epistome. It differs
from Maia in having a much broader carapace, a less deeply cloven rostrum,
spinose legs, supra-ocular spines, and trenchantly in the exclusion of the an-
tennal flagellum from the orbit by a process of the basal segment of the antenna.
In the latter regard it agrees with the genus Cyclomaia, the basal segment of the
antenna being very broad, three-spined on its distal border, and giving off a
process which separates the flagellum from the orbit by a wide interval; it also
agrees with Cyclomaia in having a short epistome, a short and broad terminal
abdominal segment, and supra-ocular spines. It differs from Oyclomaia by
having a subtriangular carapace, longer rostral spines, and a less abruptly
declivous front. It is like Paramithrax and Schizophrys in the exclusion of
the basal antennal segment from the orbit, but this segment is much broader
than in these genera, and three-spined; moreover, Matopsis differs from Para-
mithrax and Schizophrys in having a more broadly triangular carapace, supra-
ocular spines, slenderer wrist and hand, less deeply cleft rostrum, and spinose
ambulatory legs. Like Schizophrys, it has an accessory spine on each rostra
horn. In the shape of the carapace Maiopsis is intermediate between Maia
and Cyclomaia. In the form of the fingers it stands between Maia and Para-
mithrax on the one side, and Schizophrys and Cyclomaia on the other, as the
fingers, though canaliculate within, are but slightly excavated at the tips.
The synthetic character of the species upon which this genus is based suggests
the propriety of ultimately extending the scope of the old genus Paramithrax
so that it may embrace this form, together with Cyclomaia and Schizophrys.

Maiopsis panamensis, sp. nov.

Dorsal surface of carapace thickly set with spines of various sizes and scat-
tered hooked sete. The largest spines are distributed as follows : five on the
gastric region, one on the genital, one on the cardiac, four on the intestinal,
and about seven on each branchial region. Margin of carapace armed with
twelve prominent spines, three of which are on the hepatic region. The ster-
num is ornamented with small tubercles along each side of the abdomen. The

2 BULLETIN OF THE

first abdominal segment is furnished with a bidentate tubercle. Legs covered
with numerous spiny tubercles ; the meri armed with three br more prominent
spines at the distal end, Chela long and slender, the tubercles of the hand
smaller than on the other parts of the legs; fingers nearly smooth, A deep
pit at base of movable finger,

Length from base of rostrum to posterior margin of carapace, 112 mm. ;
breadth, 113.5 mm; length of rostrum, 22 mm.; length of rostral horns,
11 mm.; breadth between eyebrows, 38 mm, ; length of cheliped, 156 mm.

Station 3355. 182 fathoms. 1 male.

Family PARTHENOPIDZE.

Lambrus hassleri, sp. nov.

This is the Pacific coast representative of Lambrus pourtalesii Stimps.
(L. verrilliv Smith) of the east coast of North America, It differs from the
latter species as follows: the carapace is broader in proportion to its length ;
the branchial regions are more expanded and inflated, the inflation extending
farther in toward the cardiac area so as to involve the oblique row of small
tubercles; that is to say, this row of tubercles, which in L. pourtalesii lies low
down in the fossa between the branchial and cardiac regions, is raised up, in
L. hassleri, on the swell of the branchial region. The spines on the edges of
the chelipeds, moreover, are not laciniated to such a degree as in L. pourtalesii.

Length of a female specimen, 27 mm.; breadth, including lateral teeth,
38 mm.

Station 3368. 66 fathoms. 1 female,

«5 2342.77" 80 is 1 male, 1 female.

This species was previously obtained during the voyage of the * Hassler " at
Magdalena Bay, Lower California, August 14, 1872. The specimens then
obtained were apparently picked up dead on the shore.

Family CANCRIDZE.

Xanthodes sulcatus, sp. nov.

Carapace granulated, granulation heaviest on the lower surface, and near the

gastric

borders of the upper surface. Deeply impressed grooves separate the
from the branchial regions, and the mesogastric from the lateral gastric lobes.
The groove which continues in the median line to the front, anteriorly to the
mesogastric lobe, is crossed a short distance behind the frontal margin by a
transverse groove, which meets on.each side another groove running parailel to
the upper margin of the orbit. In this way there are marked off a pair of

frontal and a pair of orbital areolets. The frontal margin is nearly straight,

MUSEUM OF COMPARATIVE ZOOLOGY. 108

finely denticulate, and separated from the orbital areolets by a groove. Antero-
lateral border of the carapace armed with four spines or teeth, E, N, T, S, of
Dana’s nomenclature, there being no tooth at the external orbital angle; of
these teeth the first is the smallest, the third the largest, and the second and
fourth are of about equal size; the margins of all the teeth are denticulate,
The chelipeds are short and unsymmetrical; the merus is granulated on its
outer face, spinulose on the upper edge, and grooved near the articulation with
the carpus. The carpus is granulated on its outer side, and furnished with a
median internal tooth, The meri of the ambulatory appendages are spinulose
along their upper edges.

Length of carapace, 8 mm.; breadth, 11 mm.

Station 3355. 182 fathoms. 1 male, 1 female.

& OO Les 59... 1 male, 3 females.

Panopeus latus, sp. nov.

Carapace broad, convex in antero-posterior direction, granulated, especially
on the hepatic, branchial, and cardiac regions; areolations well marked ‘and
protuberant. Front divided by a small median incision into two slightly con-
vex lobes, whose edges are simple and not produced into teeth at the lateral
angles. Antero-lateral margin cut into five teeth; post-orbital tooth small,
separated from the second tooth by a shallow granulated sinus ; the third tooth
is the broadest, and rounded off at the apex; the fourth is the most salient,
and acute; the fifth is very small, and acute; all the teeth have crenate or else
spinulose margins. The margin of the orbit is minutely crenulate; its upper
part is marked by two closed fissures; the external hiatus of the orbit has the
form of a triangular notch; the lower margin is produced to form an obtuse
tooth at the inner angle. The subhepatic region is granulous, but not pro-
vided with a tubercle. The merus of the third pair of maxillipeds is also
granulated. The carpus of the chelipeds is rough with small tubercles, grooved
along the distal margin of the outer side, and armed internally with a small
blunt tooth; the hands are robust, inflated, smooth except near their articu-
lation with the carpus where scattering granules appear ; upper and lower mar-
gins rounded; fingers long, down-curved, smooth, canaliculate, their cutting
edges irregularly armed with small and rather sharp teeth, without any promi-
nent basal tooth; when closed the fingers are separated by a slight gape, and
their tips cross one another; the color of the fingers in alcoholic specimens is
very light brown. The chelw of the right and left sides are unequal in size,
but similar in shape. The ambulatory appendages are setose, their merus
joints furnished with small teeth along their upper edges. The seventh seg-
ment of the sternum in the male is entirely concealed by the base of the abdo-
men. The coxe of the fifth pair of legs are in contact with the third abdominal
segment. The penultimate segment of the abdomen has concave sides, the
terminal segment is broad and rounded.

154 BULLETIN OF THE

Length of carapace, 6.5 mm.; breadth, 10.25 mm.

Station 3897. 85 fathoms. 1 male.

In this species the carapace is short, the posterior margin wider than usual
in the genus. In its general aspect it recalls P. santhiformis A. M. Edw.;
the meri of the ambulatory legs are denticulated on their upper margins as in
canthiformis, but in latus the carapace is much broader and more convex in
an antero-posterior sense, the front is less prominent and destitute of lobes at
the lateral angles, the carpal tooth is blunt, etc.

Panopeus tanneri, sp. nov.

Carapace moderately convex both Jengthwise and transversely, smooth and
polished in the central and posterior part, coarsely granular on the anterior
gastric and hepatic regions. The gastric and hepatic areas are well marked
off by depressions, Front nearly straight, divided by a median triangular fis-
sure, and projecting as a blunt tooth at each lateral angle ; a double edge is
formed by a groove which runs along the front, the lower edge projecting far-
ther forward than the upper edge; both upper and lower edges are finely
granulate. The antero-lateral margins of the carapace are cut into triangular
teeth ; the first and second of these teeth are small, coalesced, connected by a
shallow sinus; the third and fourth are of about an equal size; the fifth is
much smaller; all the teeth are granulous on their anterior border, smooth on
posterior border ; the depressions between the third and fourth and the fourth
and fifth teeth are continued in as furrows for some distance on the carapace,
finally uniting and meeting the cervical furrow. The upper margin of the
orbit forms a tooth at the inner angle above and outside of the external mar-
ginal tooth of the front; there is a closed fissure near the middle of the supe-
rior orbital border; the external fissure is a triangular notch; a blunt tooth at
the inner angle of the lower margin. There is no subhepatic tubercle. A
transverse row of tubercles on the front of the eyestalk (when viewed re-
tracted within the orbit) just inside the eye. Chelipeds unsymmetrical; carpus
squamoso-rugose, with a slight transverse groove, and armed with a prominent
but not very sharp internal tooth ; propodus robust, superior border rounded
and squamoso-tuberculate, outer surface smooth, fingers of larger chela gaping,
movable finger armed with a blunt tooth at base ; fingers of smaller chela slen-
derer, without prominent teeth; fingers of both hands black, hooked at extrem-
ities, tips crossing. Base of abdomen entirely covers the sternal segments ;
penultimate segment of abdomen short, posterior angles prolonged backward ;
last segment short, pentagonal. In young specimens the granules of the carpus
and propodus are more numerous and sharper pointed.

Length of carapace, 6,25 mm.; breadth, 10 mm.

Station 3405. 53 fathoms, 5 male (3 adult, 2 young).

ie 709094. OIM 1 male.

a

MUSEUM OF COMPARATIVE ZOOLOGY. LO

Family PORTUNIDJE.

Achelous affinis, sp. nov.

Carapace moderately convex both longitudinally and transversely, the frontal
region depressed, horizontal; surface rugose, granular, and pubescent. Front
not prominent, composed of four blunt teeth, not counting the internal angle
of the orbit; the two teeth near the median line are equilaterally triangular,
separated from one another by a triangular notch ; the next teeth on the outer
side are separated from those within by a shallow sinus, and fall off abruptly
on the outer side into a deep n&ch which separates the front from the inner
orbital angle. The two middle teeth of the front are a trifle longer than the
lateral teeth. The tooth formed by the projecting inner orbital angle does not
attain the level of the frontal teeth; it is double, the orbital margin just out-
side of it being slightly folded, raised, and projected forward so as to form
a secondary tooth similar to that in A. spúvimanus, but not so well marked.
The antero-lateral margin of the carapace is nine-toothed, including the tooth
at the outer orbital angle; the posterior tooth is hardly larger than those in
front of it; the posterior margin of these teeth is convex. The anterior edge
of the merus of the chelipeds is armed with five spines. The carpus has an
internal and a smaller extérnal spine. There is a spine on the propodus at the
base at the point of articulation with the carpus, and another on the upper
margin a little distance back of the articulation with the dactylus. The distal
edge of the merus of the fifth pair of legs is spinulose.

Length, 25 mm.; breadth, 39 mm.; length of merus of chelipeds, 24 mm. ;
length of propodus, including digit, 33 mm.; length of internal carpal spine,
measured from distal margin of carpus, 3.3 mm.

Station 3379. 52 fathoms. 5 males, 4 females.

tele A ar 1 male.

A large number were also taken in the tow-net at the surface at the follow-
ing stations: Off Mala Point; Hydr. 2627; Nos. 3355, 3371, 3382, 3386,
3398 ; 50 miles south of Guaymas.

The speeimens taken at the surface are small, and much darker in color than
those that came up in the trawl, but show no structural differences. I take
them to be the same species in the pelagic stage of its existence.

This species is very closely related to Achelous depressifrons Stimps., from the
Atlantic side of the continent. Compared with 4. depressifrons it is broader,
and the curve formed by the front and the antero-lateral margin of the carapace
forms an are with a longer radius; the lateral lobe of the front is broader; the
tooth of the inner orbital angle is double, as above described; the teeth of the
antero-lateral margin have broader bases, and the anterior one at the outer
orbital angle is not so prominent; the protuberances on the cardiac and bran-
chial areas are less prominent; the internal carpal spine is shorter, and the
distal margin of the merus of the fifth pair of legs is denticulate behind the

articulation of the carpus.

BULLETIN OF THE

Family CORYSTIDZE.

TRACHYCARCINUS, gen. nov.

Carapace pentagonal, moderately convex, lateral margins long, nearly straight
toothed. Front narrow, produced, three-toothed. Orbits large, with forward
aspect, imperfect, with two hiatuses above, one below, and one at the inner
angle; lower wall formed chiefly by the carapace, Anterior margin of buccal
cavity not distinctly defined, epistome short, ridges of the endostome devel-
oped. Sternum long and rather narrow. Abdomen of male narrow and five-
jointed, the third, fourth, and fifth segments consolidated. ^ Eyestalks very
small, retractile within the orbits, Antennules longitudinally folded. The
antenn lie in the inner hiatus of the orbit; their basal segment is but
slightly enlarged, not filling the hiatus at the inner angle of the orbit, nor
attaining to the front, subcylindrical, unarmed, imperfectly fused with the
carapace ; the second segment is longer and slenderer than the first, the third
segment about equal to the second in length, but slenderer ; all these segments
are furnished with long and coarse sete ; the whole antenna is less than one
half as long as the carapace. The ischium of the outer maxillipeds is produced
at its antero-internal angle ; the merus of the same appendages is rounded at
the antero-external angle, obliquely truncated but not emarginated at the
antero-internal angle, where it articulates: with the following segment. Legs
of moderate length. Right and left chelipeds very unequally developed in the
male. Dactyli of ambulatory legs styliform, straight, slender, longer than the
penultimate segments.

The pentagonal shape of the carapace recalls the genus Telmessus White.
Jut in Telmessus the front is divided by a median notch, the orbit is much
more complete, the basal segment of the antenna sending off an external pro-
cess that completely fills the hiatus at the inner angle of the orbit, In the
structure of the orbit and antenne, and in the shape of the merus of the outer
maxillipeds, Trachycarcinus is much like Hypopeltarium Miers (Peltarion
Jacq.).

Trachycarcinus corallinus, sp. nov.

Carapace irregularly pentagonal, clothed with a dark brown pubescence, and
bearing flattened tubercules of ivory whiteness arranged in groups, as follows :
two anterior lateral and one posterior median, on the gastric region; four, dis-
posed in two pairs, on the cardiac region; five or six on each branchial region;
and one, of a crescentic shape, on each hepatic region. Each group of tubercles
resembles the crown of a complex molar tooth whose cusps have been worn
down to a common level. Front tridentate, the median tooth twice as long as
the lateral. Walls of the orbit furnished with four teeth separated by deep
hiatuses ; these teeth are a pre-ocular, median superior, post-ocular, and sub-

MUSEUM OF COMPARATIVE ZOOLOGY. 157

ocular, Antero-lateral border of carapace armed with three prominent teeth;
another minute tooth just back of the largest posterior lateral tooth. Eye-
stalks slender; eyes small, unpigmented, with imperfectly faceted cornea.
Chelipeds asymmetrical in the male; in adults the larger claw is naked,
smooth, and ivory-white, like the tubercles on the carapace ; the merus has a
few small teeth along its upper margin, most of them near the two ends of the
segment; the upper margin of the carpus is armed with one strong tooth, and
is denticulated along its whole length ; the propodus is short, the immovable
finger bent down at an obtuse angle with the lower border of the palm; the
dactylus is furnished with small tubercles on its upper border; both fingers
are armed with large blunt teeth on their opposed edges. The smaller claw is
slenderer than the large claw, and its fingers are proportionally longer. In the
female both chelipeds are of approximately equal size, and resemble both in size
and shape the smaller cheliped of the male. The ambulatory limbs are un-
armed, but clad with coarse seteo.
Length of carapace, 26 mm.; breadth, 27 mm.
Station 3353. 695 fathoms. 1 female.
T c UU OMG 5 males, 4 females,
"080.000, > 1 male.

Family GECARCINIDZ.

Gecarcinus malpilensis, sp. nov.

Carapace very broad and convex anteriorly, flattened and narrowed poste-
riorly; surface microscopically granulated; antero-lateral margin rounded,
not denticulated ; the median gastric furrow is well marked; the furrow
separating the gastric from the branchial regions does not extend forward far
enough to separate the gastric from the hepatic area; the so called genital
area is separated by a pronounced groove from the branchial and cardiac regions,
but not from the gastric; there is a deep indentation at the anterior extrem-
ity of the lateral genital furrows; the furrows bounding the cardiac region on
either side are moderately developed. Front deflexed at right angles to the
axis of the body, deep, concave above the margin, margin not reflexed, granu-
lated, The merus of the outer maxillipeds is five-sided, outer side convex,
distal slightly notched, antero-internal straight and parallel to the long axis of
the body, forming an obtuse angle with the postero-internal margin, which is
slightly concave. Merus and carpus of chelipeds devoid of spines or teeth.
Dactylus of the ambulatory legs furnished with six rows of spines.

Length, 55 mm.; breadth, 76 mm.; width of hind border of carapace, 19 mm. ;
width of front, 11 mm.; depth of front, Y mm.

Malpelo Island, March 5. 1 male.

This species is very distinct from any previously described, Compared with
the previously known species from the Pacific coast, it is nearest to G. quadratus

158 BULLETIN OF THE

De Saussure, but the outline of its carapace is very different; in this regard it
is similar to G. ruricola from the) eastern coast. Moreover, the front is nar-
rower, deeper, and bent down at a sharper angle than in G. quadratus, and the
merus of the outer maxillipeds is very different in shape from that of G. quad-
ratus or any other known species.

Family PINNOTHERIDZE.

Pinnixa panamensis, sp. nov.

Male. — Carapace short and very broad, smooth and punctate for the most
part, granulated at the sides, The lateral angles form a prominent shoulder,
back of which the carapace diminishes rapidly in width. A transverse depres-
sion involves the hinder part of the gastric and the fore part of the cardiac
regions, This depression is bounded behind by a ridge which extends across
the carapace between the bases of the last pair of thoracic appendages. Back
of this ridge the carapace is deflected at an angle of about forty-five degrees.
Front depressed, divided by a median notch into two not prominent lobes ;
margin setose. The chelipeds are small, with very robust carpus and hand,
destitute of spines, but furnished with some sete along their edges; carpus
squamose on the anterior and superior parts of the outer side; propodus not
broader than the carpus, compressed laterally; a row of tubercles along the su-
perior border forms a sort of crest; a longitudinal row of setiferous squamous
tubercles runs lengthwise of the hand from the proximal end, stopping short
of the base of the immobile finger; the surface between this and the dorsal
crest is thickly beset with tubercles, but below it the surface is almost smooth
down to a row of tubercles which runs along the inferior margin; immovable
finger straight, the dactylus closing against it throughout its length; dactylus
furnished with long sete on the upper margin; no prominent teeth or tuber-
cles on the cutting edge of either finger. First and second ambulatory limbs
of moderate length, nearly naked, with slender and straight dactyli; last three
segments tuberculate on superior margin. Third pair very much enlarged;
the merus has a convex anterior border, and is pubescent on the anterior third
of the upper surface; posterior margin double, spinulose. Fourth ambulatory
legs very small, not reaching beyond the distal end of the merus of the third
pair; upper surface smooth. Daetyli of penultimate and last pairs of legs
short, acute, set at an angle with the propodi so as to form prehensile hooks,
End joint of abdomen broader than the preceding joint, semicircular. The
dactylus of the outer maxillipeds articulates with the propodus near the distal
end of the latter. The carapace is 5 mm. long by 10 mm. broad.

The female is larger than the male, the carapace is more highly polished, the
transverse depression not so pronounced, and the appendages are more thickly
clothed with sete. Length of carapace, 6.3 mm.; breadth, 13 mm.

Panama, March 12. 4 males, 6 females ovig.

MUSEUM OF COMPARATIVE ZOOLOGY, 159

Family MATUTIDZE.
Osachila lata, sp. nov.

Carapace laterally expanded; three low obtuse protuberances on the gastric
region, one on the cardiac, three or four on the branchial. Front prominent,
¡bilobed, the lobes separated by a completely closed fissure. Antero-lateral
margin of carapace sharp; behind the point where the subhepatic ridge joins
‘the margin, the margin is divided into five obscure lobes, each of which is
denticulate. Merus of chelipeds tuberculous along the upper edge and outside
of it; carpus tuberculate externally, distal and superior margins produced to a
cristiform tooth which continues back upon the carpus the crest on the superior
border of the propodus; propodus tuberculate on the outer face, superior bor-
der denticulate and eristate. Edges of ambulatory limbs slightly cristate,
dactyli pubescent on distal portion.

Length, 24.5 mm.; breadth, 32 mm.

Station 3427. 80 fathoms. 1 male.

Family DORIPPIDZE.

Æthusa ciliatifrons, sp. nov.

Carapace broader than long, branchial regions much inflated ; surface granu-
lated on the branchial and cardiac regions, pubescent on the gastric region.
Front and anterior part of the lateral border ornamented with long upturned
cilia. Front between the orbits divided by a triangular median sinus and two
slightly shallower lateral sinuses into four triangular teeth of equal length.
Branchio-cardiac lines deeply impressed, meeting together in the median line
in front of the heart. Eyes small, mounted on very short peduncles, just
reaching, when extended, to the angles of the orbital sinuses. Chelipeds equal,
small and slender; chela smooth, not more robust than the carpus; fingers
longer than the palm, laterally compressed, curved inward, longitudinally
grooved, thin prehensile edges straight and regularly denticulated. Ambu-
latory appendages very long (the second longer than the first), naked and
minutely granulated; propodus slightly shorter than the merus, compressed,
grooved longitudinally on both the upper and lower faces; dactylus one half
longer than the propodus, flattened, curved, grooved and ribbed longitudinally,
its upper edge very sharp. Last two pairs of thoracic limbs densely clothed
with sete. Sternum rather coarsely granulated. Conspicuous red transverse
bands adorn the chelipeds and first two pairs of ambulatory appendages.

Length of a male (carapace), 26.5 mm.; breadth, 29.5 mm,

Station 3387, 210 fathoms,

* 3889, 153 E
* 8891. 269 s
* 8896, 127 er

BULLETIN OF THE

Zithusa pubescens, sp. nov.

Carapace a little broader than long, densely pubescent; frontal margin
ciliated ; cardiac area open in front, the branchio-cardiac lines not meeting one
another in the median line; front four-toothed, the median teeth more widely,
but less deeply, separated from each other than from the lateral; the antero-
lateral angles reach the level of the frontal teeth. Eyes as in the last described
species (Æ. ciliatifrons), Chelipeds equal, small, with pubescent merus and
naked carpus and hand ; fingers longer than palm, compressed, curved, gaping
at base. Ambulatory*s pendages similar to those of Æ. ciliatifrons, but free
from granulation. Abdomen (of female) very broad, pubescent.

Length, 26 mm.; breadth, 29 mm.

Station 3367. 100 fathoms. 1 female.

This species resembles Æ. ciliatifrons, but the median notch of the front forms
a more open angle in Æ. pubescens, and the sinus separating the front from the
external orbital angle is not so deep. The branchio-cardiac lines do not meet
in front of the cardiac area, and the cardiac area is not so much sunken below
the level of the branchial regions ; the branchial areas, moreover, are not so
conyex, nor is the cardiac region so uneven. The whole surface is densely
pubescent, and the abdomen of the female is much broader. The fingers too
are different, inasmuch as they are separated by quite an interval at base.
The first and second pairs of ambulatory limbs are imperfect in the only
specimen at hand, but they are very similar, as far as they go, to the corre-
sponding appendages of Æ. ciliatifrons.

Ai. pubescens may prove to be the full-grown state of Æ. lata Rathbun, the
description of which has just appeared in Proc. U.S. Nat. Mus., Vol. XVI,
p. 258, 1893.

Aithusina smithiana, sp. nov.

In this species the carapace is longer than broad, and is not much nar-
rowed anteriorly. The front is four-toothed, the middle pair of teeth large,
triangular, separated from one another by a wide triangular sinus which is
broader than the antennular sinus; between these teeth the margin is bent,
down. till it meets the epistoma below ; the lateral teeth of the front are
spiniform and shorter than the middle teeth. The surface of the carapace
is clothed with a short pubescence, and is lightly granulous; the branchio-
cardiac grooves are well marked. The post-orbital teeth are spiniform, and
they project far beyond the extremity of the small eyestalks. The eyes are
smaller than the extremity of their peduncles. The chelipeds are equal,
smooth, naked, unarmed ; the merus cylindrical, the carpus short and rounded :
the fingers about equal in length to the body of the chela, compressed, prehen-
sile edges sharp and not provided with distinct teeth or tubercles, The
ambulatory legs are nearly naked, the second pair more than twice the length

MUSEUM OF COMPARATIVE ZOOLOGY. 161

of the carapace, the dactylus longer than the propodus. The last two pairs of
legs terminate in short recurved claws, which are setose on the posterior edge.
The sexes do not seem to differ in any marked degree.
Length of carapace (male), 9.3 mm.; breadth, 8 mm.
Station 3370. 134 fathoms. 3 females,
** 8880. 899 y 2 males.

This species is nearly related to Æ. abyssicola Smith, but its carapace does
not diminish so much in width anteriorly, the external frontal spines are less
developed, the external orbital spine is musu longer and brought forward so
that the orbits face more to the front. From Æ. challengeri Miers it differs in
the greater development of the frontal and external orbital spines.

Cymopolia tuberculata, sp. nov.

Carapace very broad, subpentagonal, branchial regions swollen. Front four-
toothed, the teeth blunt, separated from each other by narrow sinuses which
are rounded at the bottom; the two middle teeth are longer than the lateral
ones, and the median sinus is deeper than the two lateral. The antero-lateral
margin of the carapace is four-toothed, counting the prominent tooth at the
external angle of the orbit ; the posterior tooth of the series is the smallest.
The upper margin of the orbit has three deep fissures defining two triangular
teeth ; the lower margin of the orbit has two fissures enclosing a broad trun-
cate tooth or lobe; there is also a broad and prominent lobe just below the
inner orbital angle ; above this lobe is a single tooth at the inner angle. The
surface of the carapace is ornamented with granulated tubercles; the parts be-
tween the tubercles are more finely granulated, and when viewed under a lens
are found to be furnished with fine hairs. The chief tubercles are disposed as
follows: one pair on the frontal region behind the margin; four in a transverse
row on the anterior part of the gastric area, and five on the posterior part of the
same area arranged thus :':; of these the posterior pair is the smallest ; four
in a transverse line on the cardiac region and one median behind the transverse
series; about six on each branchial area ; six just anterior to the straight pos-
terior margin of the carapace (three on each side). There are three small
tubercles on each eye-stalk near the margin of the cornea. The chelipeds are
small, slender, equal; the carpus tuberculose, the fingers as long as the hand
proper, curved downward and inward, crossing at the tip, their prehensile
edges finely denticulate in small specimens, nearly entire in larger ones. The
second and third ambulatory limbs are very long, the second slightly longer
than the third; their merus joints are granulated and costate above, and armed
with a prominent spine at the antero-distal angle and a smaller one each side
at the point of articulation with the carpus; this holds good of all three pairs
of ambulatory appendages; the carpi are carinate on their anterior margin,
with a vestige of a tooth at each end of the carina; the anterior edge of the
propodus is also carinate. The ambulatory legs are ornamented with trans-

162 BULLETIN OF THE

verse bands of red, three of which cross the merus, The abdomen and sternum
are granulated.

Length, 13 mm, ; breadth, 18 mm. ; length of ambulatory leg of second
pair, 34 mm.; merus of do., 9.3 mm. ; carpus, 5.6 mm. ; propodus, 9 mm.;
dactylus, 7.5 mm,

Station 3355. 182 fathoms. 4 males, 1 female.

Cymopolia zonata Rathbun (Proc. U. S. Nat. Mus., XVI. 259, 1893), lately
described fiom the Gulf of California, 40 fathoms, differs from C. tuberculata as
follows. The carapace is narrower and more quadrangular, The median lobes
of the front are small and inconspicuous, while the lateral lobes are very broad
and are separated from che median lobes by a slight, shallow notch, In C. tu-
berculata the four frontal lobes take on the form of prominent triangular teeth,
clearly separated from each other by deep triangular sinuses. The antero-
lateral margin of the carapace is three-toothed in C. zonata, four-toothed in
C. tuberculata. The tubercles near the posterior margin of the carapace are
more elongated in the former species than in the latter. The hand of the
former is much broader, and is armed with prominent spiny tubercles,
Finally the meri of the ambulatory legs are much shorter in C. zonata, and are
armed at the distal end with a blunt triangular tooth, while in C. tuberculata
this tooth is transformed into a long sharp spine, and a pair of smaller spines
is present, one on each side of the proximal end of the carpus.

Family RANINIDZE.
Raninops fornicata, sp. nov.

Carapace very convex from side to side, naked, smooth or nearly so, punc-
tate. Rostrum acute, lightly carinate, the carina extending backward for a
short distance on the carapace. Superior margin of orbit armed with three
acute teeth, the second of which is curved forward ; the anterior tooth is sep-
arated from the rostrum by a deep rounded sinus, from the second tooth by an
angular notch ; the second tooth is separated from the third by a nearly
straight interval; the third tooth lies some distance in front of the posterior
end of the orbit. Back of the orbit there is a strong procurved spine on the
margin of the carapace. Eyestalks compressed, equal in length to one half
the width of the carapace, Second segment of the third maxilliped equal to
the third joint, crossed by a piliferous line; third segment notched at the an-
tero-internal angle. Cheliped: merus unarmed, microscopically spinose above,
setose below; carpus minutely rugoso-spinulose, the superior distal angle pro-
jecting as a sharp tooth; propodus lightly rugose, upper and lower borders
margined, unarmed, palmar edge irregularly and inconspicuously toothed ;
dactylus without any prominent tooth. The dactylus of the fourth pair of
legs has a very convex internal border, the dactylus of the fifth is long, narrow,
and spatulate. Abdomen setose; telson obtuse at the end,

Length of carapace, 12 mın.; breadth, 8.6 mm.

Station 3369. 52 fathoms. 1 specimen.

MUSEUM OF COMPARATIVE ZOOLOGY. 163

Family LITHODIDZE.

Rhinolithodes cristatipes, sp. nov.

The carapace is subtriangular in outline, its surface devoid of sete, but
covered with low squamiform tubercles; the whole gastric area is raised into
a conical prominence; there is also a prominent crescentie rounded ridge on
each branchial region, enclosing the cardiac area in a deep fossa open only
behind. The rostrum is straight and conical, with a vertical plate projecting
below from the proximal half down between the eyestalks; this plate is
toothed anteriorly but does not reach forward nearly to the tip of the rostrum.
The antero-lateral margin of the carapace is five-toothed; the second, third,
and fourth of these teeth give rise to long thread-like cilia, There is another
tooth at the angle between the postero-lateral and posterior margins ; posterior
margin straight. The upper surface of the eyestalk is covered with small
tubercles and a blunt spine projects over the cornea, The movable scale of the
antenna is spiniform and bears two blunt spinules on the outer side and two
smaller ones on the inner. The chelipeds are unequal (the right being the
larger); coxa granulated, setose on the lower inside margin; lower margin of
the merus armed with three or four blunt teeth, superior margin toothed, inter-
nal distal border setose, external distal border forming a bilobed crest; outer
face of carpus squamous, margins eristate, the internal crest expanded and cut
into setiferous lobes ; propodus tuberculated without, smooth within, toothed
and setose on superior margin; immobile finger, as well as the dactylus, exca-
vated within, setose ; the larger claw has blunt teeth on the fingers, while the
fingers of the smaller claw have nearly straight cutting edges. The ambula-
tory appendages have cristiform anterior margins from the merus to the propo-
dus inclusive ; the crest of the carpus is entire, but that of the merus is bilobed,
of the propodus trilobed; the posterior margins of these appendages are den-
tate and more or less setose, the dactyli are provided with curved acute black
tips, and with pencils of hair especially on anterior margins. The abdomen
is indurated, with three rows of tuberculated plates.

Length of carapace, 16.5 mm.; breadth, 16.5 mm.

Station 3354. 322 fathoms, 1 male,

Ihave seen neither specimen nor figure of the type of this genus, Rhinoli-
thodes wossnesenskit Brandt, from the coast of Alaska, and the specimen above
described possibly does not belong to the same genus. Brandt's generic
diagnosis (Bull. Phys.-Math. Acad.. Sci. St. Pétersbourg, VII. 174, 1849)
appears to include specific as well as generic characters. The specimen above
described conforms to Stimpson's diagnosis of the genus (Proc. Acad. Nat. Sci.
Phila., 1858, p. 231): * Abdomen seutis triseriatis obsessum. Antennarum
aciculum margine spinosum, Pedes mediocres.”

VOL. XXIV. — NO. f. 2

164. BULLETIN OF THE

Echinocerus diomedese, sp. nov.

Carapace subpentagonal, gastric and branchial regions inflated, the whole
surface beset with tubercles which give rise to minute sete. There is one
rather more prominent tubercle in the depression on each side of the gastric
area. Rostrum short, three-spined; one of the spines is median and inferior,
two are paired near the base above; in one of the two specimens obtained the me-
dian spine is toothed below. The antero-lateral margin of the carapace is irregu-
larly toothed. Eyestalks spinulose above, with one prominent spine projecting
forward over the cornea. The movable scale or spine of the antenna is spinu-
lose on each side (four or five spines on each margin), The merus of the cheliped
bears a spine on the inner side at the distal end; the carpus is smooth outside,
the inner border expanded into a seven-toothed crest setose within, the outer
border straight, naked, and keeled; the chele are of unequal size (the right
being the larger), setose, spiny on the upper edge; fingers spoon-shaped within,
setose and somewhat gaping. Ambulatory appendages: the meri are spinose
on their edges; the anterior border of the carpus of the first pair is furnished
with a crest whose edge is even and entire; the propodus of the first pair is
crested along the proximal half of its anterior border, while the distal half is
armed with two or three teeth; the carpus and propodus of the second and
third pairs are toothed on the anterior margin, the propodus of the third pair is
also toothed on the posterior margin; all the segments are hirsute, especially
the dactyli. When the legs are closely folded against the sides of the carapace
a wide interval is left between the carpi of the cheliped and first pair of am-
bulatory appendages, bounded by the opposite crests of these segments, and
forming a passage for the admission of water to the gills. This orifice is simi-
lar to that seen in E. foraminatus Stimps., but it is not so perfectly formed
The apex of the abdomen (in the female) is turned to the right (most strongly
in the larger specimen); the marginal plates are wanting on the left side; all
the abdominal appendages excepting the first are aborted on the right side.

Length, 64 mm.; breadth, 71 mm.

Station 3384. 458 fathoms. 1 female.

Station 3394. 511 fathoms. 1 female ovig.

In this species the acicle of the antenna is spinous on the margins only, as
in E. foraminatus Stimps.

Paralomis aspera, sp. nov.

Carapace pentagonal, as broad as long; gastric, cardiac, and branchial regions
well defined and prominent; whole surface of carapace and abdomen thickly
beset with papille or tubercles, each one of which is encircled with a crown
of stiff setae, Rostrum short, indistinctly tripartite, multispinose, A sharp
dark-tipped spine at the external orbital angle, another at the antero-lateral
angle of the carapace, and four or five, irregularly arranged, on the margin of
the branchial region.

MUSEUM OF COMPARATIVE ZOOLOGY. 165

The dorsal face of the second abdominal segment consists of a single plate,
undivided by longitudinal sutures, with a deep depression on each side of the
middle. The following segments are unsymmetrical on the two sides of the
unique type specimen (female), the abdomen being twisted to the right. The
lateral margins of these segments are laciniated.

Eyestalks spinulose above; eyes very black, with downward aspect. Distal
segment of antennule much longer than the antecedent segment, tuberculous
above. Antenne of moderate length; outer margin of first segment spinulose ;
second segment spinulose, and produced on the outer side to a long spine;
movable acicle reaching to the distal end of the peduncle, spinose, the longer
spines marginal, one spinule on the lower side and another on the upper side
near the base; penultimate and ultimate segments of peduncle bear small
setiferous tubercles. Right cheliped more robust than its fellow, thickly beset
with strong spines. Ambulatory legs long, robust, spinose like the chelipeds;
their basal segments are wellnigh covered by the overlapping margin of the
abdomen (in the female).

Length of carapace, 113 mm.; length of rostrum, 9 mm.; breadth of caras
pace, 113 mm.; length of posterior ambulatory legs, 255 mm.; merus, 68 mm.;
carpus, 39 mm.; propodus, 72 mm.; dactylus, 56.5 mm.

Station 3353. 695 fathoms. 1 female.

This species, like the one next described, is much longer legged than P. gran-
ulosa (Jacquinot), the type of the genus. In this regard it is more like the two
“Challenger” species described by Henderson. The specimen above described
was infested with a huge Peltogaster 36 mm. in breadth.

Paralomis longipes, sp. nov.

Male. — Carapace triangular; gastric, cardiac, and branchial regions well
defined, protuberant; the most prominent part of the cardiac area reaches a
higher level than the branchial areas; whole surface of carapace thickly cov-
ered with blunt papilla; viewed under a lens each tubercle is seen to be en-
circled with a ring of short, stiff set» ; one of the tubercles, situate in front of
the centre of the gastric region, assumes a spiny form. Rostrum furnished with
three prominent spines, one median and inferior, two lateral and superior; the
latter are not so long as the inferior spine; there is, too, a spinule on the
lower side of the inferior spine, and a still smaller one above, between the roots
of the superior pair. There are two pairs of long spines on the anterior margin
of the carapace, one at the external orbital angles, the other at the antero-lateral
angles of the carapace. There are also three or four prominent spines on the
side of each branchial area. Eyestalks spinulose above. Basal segment of the
antenna armed with an external spine, the second segment with severai spines,
the most prominent of which is on the outer side; acicle furnished with five
prongs or spines, the largest median, the others lateral; flagellum much longer
than the carapace. Cheliped of moderate size (the right one has been lost from

166 BULLETIN OF THE

the only male specimen obtained), coxa tuberculate on the lower face, the fol-
lowing segments armed with strong spines, fingers excavated within, slightly
gaping, penicillate, cutting edges entire. Ambulatory appendages very long,
second and third pairs of about equal length and longer than the first pair;
all of them armed, like the chelipeds, with spines; the spines tend toward a
regular arrangement in longitudinal rows, and the spaces between the spines
are smooth and naked, i. e. nearly free from spinules, tubercles, or sete.
Second segment of abdomen composed of a single calcified plate, marked by
a deep hollow on each side of the middle ; the other segments of the abdomen
are of a leathery, semi-membranaceous consistency, and are made up of five
longitudinal rows of plates, viz. one median row, flanked by a row on each
side, the latter in turn bounded externally by a marginal series ; whole sur-
face of abdomen verrucose.

Length of carapace, including rostrum, 84 mm.; breadth, 78 mm.; breadth
between the antero-lateral angles of carapace, 34 mm. ; length of last ambula-
tory appendages, 242 mm.; merus, 76 mm.; carpus, 36 mm. ; propodus, 64 mm. 5
dactylus, 47 mm.

In the female the right chela is larger than the left, and the prehensile edges
of the fingers are furnished with blunt teeth. The abdomen is asymmetrical,
the apex turned to the right, the marginal plates absent from the left side; the
right side bears but one ovigerous appendage (the first). The abdomen nearly
conceals the basal segments of the thoracic legs, which in the male are almost
entirely exposed.

Station 3371. 770 fathoms. 1 male, 1 female ovig.

The egg measures 2 mm. in diameter.

Lithodes panamensis, sp. nov.

Carapace sub,, -iform, of about equal length and breadth; gastric and
branchial regions very convex ; a deep depression on each side at the anterior
limit of the branchial areas, and another between the gastric and cardiac areas.
Rostrum eylindrical, terminating in three spines or teeth, one of which is me-
dian, the other two lateral; a long horn, slightly upeurved, is given off from
the lower side of the proximal end of the rostrum. The whole surface of the
carapace is rough with low warty protuberances ; the gastric region bears two
pairs of spines, the anterior pair separated by a greater interval than the pos-
terior pair. Two small spines on each branchial area, and two on the intes-
tinal region. The orbit is bounded externally by a prominent spine, and
there are five more spines on the lateral border of the carapace, viz, one at the
antero-lateral angle, one on the hepatic region, and three on the branchial.
Besides these there is a rudimentary lateral spine near the anterior limit of the
branchial region, The posterior margin of the carapace is tuberculated, not
spinose. The third segment of the antennular peduncle is equal to the first
segment in Jength, and considerably longer than the second segment. The

MUSEUM OF COMPARATIVE ZOOLOGY. 167

antenne are about equal in length to the carapace without the rostrum; the
second segment is armed with a long and sharp external spine; there is, more-
over, on the antenna of the right side a movable thorn-like acicle equal in
length to the last two segments of the peduncle together; the acicle is armed
with a small tooth on the external margin midway between the base and the
tip. There is no trace of an acicle on the left antenna. As the type specimen
is the only one obtained, it is impossible to tell whether the acicle is normally
present or absent. The last segment of the peduncle is nearly twice as long as
the penultimate segment. The chelipeds are furnished with scattered tuber-
cles, a few of which assume a spiny form. The left cheliped is rather slen-
derer than its fellow. The ambulatory legs are long and rather slender, their
meri sparsely furnished above with tubercles, which tend to a spiny form on
the anterior and posterior margins; at the distal end of the anterior margin
of the meri there is a prominent spine-like tooth ; the carpi and propodi are
armed with teeth, chiefly on the anterior margins; the dactyli are equipped
with four short spines (two superior and two lateral) near the proximal end.
The abdominal segments (of the female) are roughened by low tubereles, and
dentate on their margins. The lateral teeth of the marginal plates of the right
side are drawn out into long spines, The tergal plate of the second abdominal
segment is completely fused with the epimera, showing no trace of an in-
tervening suture. The marginal (episternal?) plates of “this segment are
bounded within by a distinct suture.

Length of carapace, excluding rostrum, 79 mm. ; rostrum, 14 mm.; breadth
of carapace, 79.5 mm.; length of posterior pair of ambulatory legs, 193 ram. ;
merus, 58.5 mm.; carpus, 31 mm.; propodus, 54 mm.; dactylus, 53 mm.

Station 3384. 458 fathoms. 1 female, '

This species finds its nearest relative in D. murray? Henderson (Rep. Chal-
lenger Anomura, p. 43, Plate IV.), from the distant Prince Edward Island in
the Southern Ocean. The latter species differs from L. panamensis in. having
a much longer rostrum, which is forked at the end, a more oval and spiny oarg-
pace, shorter external antennal spine, more spinose legs, ete.

Family PAGURIDZ.
Cancellus tanneri, sp. nov.

Differs from O, canaliculatus (Herbst) in having a much shorter and broader
abdomen, in the lobate character of the marginal erests of the first and second
pairs of legs, in the pubescence of the thorax and abdomen, and in the short-
ness of the antenne. From C. typus M. Edw., it is distinguished by having
the anterior border of the carapace less deeply incised on either side of the ros-
trum, and the telson squarely truncated posteriorly, not notched in the middle;
the cox; of the last pair of legs, too, present a very different shape, since they
lack the prominent anterior lobe observable in C. typus. C. parfaiti A. M. Edw.

168 BULLETIN OF THE

et Bouv., compared with C. tanneri, displays longer eyestalks, while the thorax
and abdomen are less pubescent, and the coxal segments of the posterior legs are
much more protuberant.

Length of carapace, 7 mm. ; breadth, 5 mm.; length of eyestalk, 3 mm.

Station 3368. 66 fathoms. 1 male.

Found in a cavity in a piece of dead coral rock.

The abdomen in the unique type specimen is not quite symmetrical, but this
may be an individual peculiarity resulting from the shape of the cavity in
which the animal lived.

Pylopagurus longimanus, sp. nov.

Carapace smooth, naked, polished ; rostrum ‘short, triangular, subacute, ad-
vanced farther than the rounded lateral anglos. Abdomen longer than cephalo-
th-rax. Eyestalks equal in length to the first two segments of the antennulary
peduncle. Ophthalmic segment uncovered. Ophthalmic scales separated by
a considerable interval, triangular, their tips split in a horizontal plane so that
they end in two acute teeth, one above the other, the lower tooth the longer.
Last segment of antennulary peduncle very long and slender (much longer
than the eye-stalk). Right cheliped of enormous size, much exceeding the
whole body in length ; the outer face of the merus is nearly smooth, the lower
and inner faces granulated ; carpus very large, equalling in length all the pre-
ceding segments combined, its surface granular, the granules assuming the form
of small spinulose tubercles on the dorsal face, which is limited within by a
row of larger teeth. Chela irregularly oval, the external side flattened to form
an opercular facet, which is thickly set with granules and surrounded with a
margin of denticles. Left cheliped very small, its segments more hairy than
those of the right ; basal part of propodus short and swollen, the fingers long,
gaping at the base; the outer or upper faces of the propodus and dactylus are
granulated, and definitely bounded by a line of regularly arranged granules on.
the outer margin of the dactylus and the inferior margin of the propodus; the
inner or lower face of the chela is pretty free from granules, but is furnished
with numerous sete, those on the dactylus being grouped in conspicuous tufts
or pencils, Penultimate pair of thoracic appendages almost perfectly chelate,
its rasp restricted to the distal part of the claw, broad, and composed of many
rows of scales.

The legs are yellowish, banded and mottled with red.

Length of carapace, 10 mm.; eyestalk, 4 mm.; right cheliped, 34.5 mm.
(merus, 8 mm.; carpus, 11 mm.; propodus, 14.8 mm. ; dactylus, 10 mm.).

In younger, smaller specimens the chelipeds are shorter in proportion to the
length of the body, and the major claw is shorter, broader, and of a more regu-
lar oval form. A specimen, whose carapace is 6.5 mm. long, gives the follow-
ing dimensions for the right cheliped : total length, 16.5 mm.; merus, 4.5 mm. 5

carpus, 5.5 mm.; propodus, 7.5 mm.; dactylus, 4.5 mm.

MUSEUM OF COMPARATIVE ZOOLOGY, 169

Station 3368. 66 fathoms. 5 males,

Resembles P. ungulatus (Studer), but readily distinguished from that species
by the great size of the right cheliped, the irregularly oval outline of the right
chela, the great length of the distal segment of the antennulary peduncle, etc.

Pylopagurus affinis, sp. nov.

This species is nearly related to P. ungulatus, from which it differs in the
following respects, the eyestalks are longer, and narrower at the distal end ;
the external prolongation of the second segment of the antenna is longer and
slenderer; the upper margin of the carpus of the right cheliped is armed with
two or three spines, the largest of which is elose to the anterior border ; the
outer face of the carpus is smooth save where a light tubercular ridge runs
along the middle. In P. ungulatus this face of the carpus is thickly covered
with spinulose granules which assume larger proportions and a uniserial ar-
‘angement on the superior and inferior margins, On extending the comparison
to the large chela, further differences between the two species become appar-
ent. In both species the external face is flat, covered with minute spinulose
granules, and surrounded by a border of sharp spines; but in P. affimis the
marginal spines are larger and more irregular, and the flat opercular facet is not
sharply defined at the proximal end by the regular arrangement of the marginal
spines as in P. ungulatus ; instead, one finds the marginal series of spines
broken down at this point, thus effacing any distinct limit between the oper-
cular face of the chela and the articular surface which connects the propodus
with the carpus. The inner or lower surface of the large chela is smooth in
P. affinis, granulated in P. ungulatus. The left cheliped is quite different in
ts the several segments of which it is composed give

the two species: in P. a,
rise to long sete, which give the appendage a very hairy appearance when
contrasted with P. ungulatus; the inferior border of the chela is conspicuously
toothed, while in P. ungulatus it is entire. The ambulatory legs are more
hairy in the Pacific species than in P, ungulatus, and their carpal joints are
not so distinctly dentate on the superior border, The rasps of the fourth pair
of legs are multiserial in both species, The telson is symmetrical, subcircular
in outline, its posterior border convex and entire ; in P. ungulatus, the telson
has a deep and wide posterior median notch.

Length, about 12 mm. ; length of carapace, 4.5 mm.

Station 3397. 85 fathoms. 1 male.

There are three simple unpaired abdominal appendages on the left side, in
the type specimen, The vasa deferentia are extruded from the base of the
fifth legs on each side. They appear as slender threads, the one on the right
side much longer than its fellow, and twisted into a small bunch.

BULLETIN OF THE

Pylopagurus hirtimanus, sp. nov.

This species closely resembles P. rosaceus A. M. Edw. et Bouv.,! from the
West Indian seas. Compared with the type of P. rosaceus it presents the fol-
lowing differences. The eyestalks are a little slenderer, and together with the
ophthalmic scales are separated by a wide interval in which the antennules lie,
exposed from above. In P. rosaceus, on the contrary, the ophthalmic stalks and
scales are closely approximated, concealing the antennules beneath. The outer
face of the right chela is ornamented with conical tubercles, whose bases are
expanded into circular plates; these plates are closely packed over the surface
of the chela, so that no interstices are left between them; their borders are
cut into ‘a large number of minute radiating processes; on. the basal half of
the propodus the tubercles give rise to long setze, which render that part of
the claw conspicuously hirsute, in contrast with the distal part which is naked;
furthermore, on the distal half of the propodus, especially on the concave sur-
face of the immovable finger, the tubercular processes tend to become obsolete,
leaving only the basal circular radiate plate; the bases of the conical teeth
along the outer margin of the hand and the movable finger are expanded at
the base into flattened roundish surfaces with radiate margins ; these surfaces
form a conspicuous outer border to the hand; the inner face of the hand is
tuberculated. In P. rosaceus the tubercles of the outer face of the chela are
encircled by rounded granules much less numerous than the radiating points
in P. hirtimanus; the tubercles are so loosely arranged that numerous inter-
stices are apparent between them ; the hairs on the basal half of the hand are
not so well developed as in the Pacific species ; the whole outer face of the
immovable finger is strongly tuberculated, the teeth of the lower margin are
not expanded into conspicuous plates, and the inner face of the chela is nearly
smooth. The carpus in P. hirtimanus is armed with larger spines on the
internal margin, is more hairy, and more coarsely granulated on its inner,
inferior, and outer surfaces than it is in P. rosaceus. The left chela is com-
paratively smaller than in P. rosaceus, is more strongly toothed along the
internal margin of the propodus and dactylus, and exhibits, besides, most of
the above specified. peculiarities of the right claw. The primary branch of the
anpaired abdominal appendages is shorter and broader than in P, rosaceus. The
rasp on the propodus of the fourth pair of legs is uniserial, as in P. rosaceus.

Length of carapace of largest male, 12 mm. ; length of carapace in front of
cervical groove, 7 mm. ; breadth across the branchial regions, 9.5 mm, ; length
of large claw, 11 mm. ; breadth of large claw, 7 mm.

Station 3367. 100 fathoms. 2 males, 1 female.

ISO = OO ao 5 males, 5 females.

A good deal of the color is still preserved in alcohol. The lower surface of

the eyestalk is quite a deep red, while the upper surface is a pale yellow. The

1 Mem. Mus. Comp. Zoöl., XIV. No. 3, p. 97, Plate VII. Figs. 10-17, 1893.

MUSEUM OF COMPARATIVE ZOLOGY. 171

merus and carpus of the chelipeds and all the segments of the ambulatory legs
from the ischium to the dactylus inclusive are banded transversely with bright
red on a yellowish ground.

Catapagurus diomedese, sp. nov.

The carapace is smooth and naked, and divided into an anterior and pos-
terior section by the cervical groove. The anterior margin projects slightly
between the eyestalks, forming a blunt rudimentary rostrum. The gastric
region is sharply defined, and presents an indistinct longitudinal furrow on
each side of the median line; it is lightly convex in both directions. The
branchial regions of the right and left sides are strongly inflated and sharply
separated from each other by a re-entrant angle formed by the curving for-
ward of the posterior border of the carapace on each side of the median line.

The eyestalks are rather long and slender, being about equal in length to
the anterior section of the carapace. The ophthalmic scales are very small and
minutely bifid ‘at the tip. The third segment of the antennulary peduncle is
about two thirds the length of the eyestalk, and increases in diameter from the
base to the distal end; the superior flagellum is rather longer than the distal
segment of the peduncle, and its enlarged ciliated basal portion forms rather
more than one third of its whole length, The inferior flagellum is about one
half as long as the superior, and is composed of about eleven segments. The
peduncle of the antenna surpasses the eyestalk by one half the length of its
distal segment ; the acicle is long and slender, tipped with a few sete ; it
reaches forward a little beyond the eye; the flagellum reaches beyond the tips
of the ambulatory legs. The chelipeds are nearly alike in shape and size ;
their segments are clothed with long sete, which assume a tomentose appear-
ance on the chel®; the carpus is about equal in length to the chela, its inner

face is perpendicular, the inner margin of the upper side is armed with seven

small spines; there is also a spine at the distal end of the superior margin of
the carpus ; the chela is short and thick, the fingers about the same length as
the basal portion of the propodus, slightly down-curved, meeting throughout
their length, working horizontally. The ambulatory legs are of nearly equal
length, surpassing the chelipeds, setose, the carpus armed with a sharp tooth at
the distal end of its upper side ; the propodus is twice as long as the carpus,
the dactylus is considerably longer than the propodus, and, like that segment,
is distinctly curved ; it is tipped by a small horny nail. The fourth pair of
legs is but slightly subcheliform ; the rasp is formed of a single row of scales.
There are three small rudimentary appendages on the left side of the abdo-
men; the first and second of these are two-branched, the secondary branch
being exceedingly minute.

The sexual tube, which issues from the coxal segment of the last thoracic
appendage of the right side, is very long in this species.

Length of carapace, 5.5 mm, ; abdomen, 8 mm.; cheliped, 12,5 mm.; last
ambulatory leg, 17.5 mm. ; ocular peduncle, 2.5 mm,

172 BULLETIN OF THE

Station 3355, 182 fathoms. 1 male.

This species differs much from the more typical species of Catapagurus,
C. sharrerv A. M. Edw. and C. gracilis Smith, in the shortness of its chelipeds
and ambulatory limbs, the symmetry of its chelipeds, and the length of the
protruded vas deferens. It is more nearly related to C. australis Henderson
(Challenger Anomura, p. 76, Plate VITI. Fig. 1). It would seem to have a
close general likeness to Pagurodes piliferus Henderson, but the gills in
C. diomedece are of the phyllobranchiate type.

Spiropagurus occidentalis, sp. nov.

The carapace is smooth and naked except on the sides of the branchial re-
gions, where a few hairs arise; the branchial regions are swollen, membrana-
ceous, and covered with a network of white lines; the cardiac region is long
and narrow; the portion of the carapace in front of the cervical groove is cal-
cified, produced in the median line anteriorly to form a short, broad, and obtuse
rostrum, which does not conceal the ophthalmic segment; the lateral teeth are
acute, and project as far as the rostrum does; they form a sharp demarkation.
between the front and the oblique antero-lateral border of the carapace.

The ophthalmic scales are triangular, with simple tips. The ocular peduncles
are enlarged at the distal end, hardly overreaching the distal end of the second
segment of the antennulary flagellum and the third segment of the antennary
flagellum. The last segment of the antennulary peduncle is more than twice
as long as the penultimate segment. The second segment of the antennal pe-
duncle is produced externally into a long, sharp tooth; the antero-internal angle
is likewise armed with a small spine; the acicle is long, sharp, curved, and
furnished with sete, as are also the several joints of the peduncle.

The chelipeds are subequal, the right chela being appreciably larger than the
left; the ischium is armed with about five denticles along the internal margin;
the inner margin of the merus is armed with the same number of rather larger
teeth; the inner margin of both of these segments is furnished with long and
slender sete; the surface of the carpus is rough with setiferous ruge, and is
armed with scattered spinules upon its upper face and with a row of larger
spines along the internal margin of the upper face. The external face of
the propodus is armed with spines which are regularly arranged in five longi-
tudinal rows; only at the distal end of the propodus, at the base of the fingers,
do these spines lose their regular serial arrangement; from the bases of these
spines spring long sete; the fingers are acute, tuberculo-spinose and setose, like
the hand. The fingers of the left hand are longer in proportion to the palm
than those of the right hand. The ambulatory legs are rather robust, and longer
than the chelipeds; their lateral surfaces are smooth ; the upper margins of the
carpi are armed with a row of little spines, and the corresponding margins of
the propodi are denticulated; the dactyli are longer than the propodi, but not
so long as the propodus and carpus combined ; the anterior pair of ambulatory

MUSEUM OF COMPARATIVE ZOOLOGY, 173

legs differs from the posterior pair in having the dactylus and propodus a little
shorter, and the lower edge of the merus more hairy and minutely spinulose.
The telson is deeply cut by a broad median notch in its hind margin into two
lobes, the left of which is the larger. The margins of both lobes are spinose.

There is (in the alcoholic specimens) a narrow red ring around both fingers
near their tips, a broader band of the same color around the base of the fingers,
and a red patch on the inner side of the anterior face of the hand.

Length of carapace, 6.5 mm. ; greatest breadth of carapace, 6 mm. ; length of
ocular peduncles, 2.3 mm,

Station 3368. 66 fathoms. 1 male.

Ihamadt dus Bit 1 male.

This species, like S. ris A. M. Edw., is characterized by the spiny armature
of the chelipeds. It differs from S. imis by having a smaller number of spines
upon the anterior face of the chela, and in their arrangement in a few definite
longitudinal rows, the surface between the rows being spineless and reticu-
lated. The chela does not display the lively iridescence so striking in Milne
Edwards’s species, nor are the ophthalmic scales bidentate at the end. The
merus of the third maxilliped is not armed with a distal spine, as in S. iris.

Paguristes fecundus, sp. nov.

The anterior or gastric section of the carapace is smooth in the central and
hinder parts, rugose near the front, and tuberculose in the antero-lateral region,
where a few of the tubercles assume a spiny form; from the tubercular surface
spring long slender sete. The anterior margin of the carapace is produced in
the median line so as to form an acute triangular rostrum, which projects be-
yond the subacute lateral processes ; between the rostrum and the lateral pro-
cesses the anterior border is concave and thickened so as to form a rim. The
anterior gastric lobes are clearly defined anteriorly.

The ocular peduncles are long and cylindrical, reaching considerably beyond
the antennal peduncle, but not quite so far as the antennulary peduncle. The
ophthalmic scales are of moderate size and bidentate at the tip, the external tooth
very minute. The antennal acicle reaches almost to the end of the peduncle ;
it is setose, and armed with six spines, two of which form a terminal fork,
the others being marginal. The external prolongation of the second antennal
segment is narrow, setose, and minutely spinulose; there is, moreover, a
spinule on the upper face of this segment behind the base of the acicle ; the
antennal flagellum is very short (about equal in length to the anterior section
of the carapace). The third pair of maxillipeds are closely approximated at the
base; their merus joints are armed with three or four denticles on the lower
margin, and one at the distal end of the upper margin.

The chelipeds are short and of like size and shape; the merus is smooth
within, rugose without; the two inferior margins are armed with minute
black-tipped spinules. The carpus is tomentose and spinulose, the largest

174 BULLETIN OF THE

spines occurring along the superior border; the hand is both spinulose and
pubescent, but the hair is less dense upon the fingers than upon the basal
portion of the hand, which is short and swollen below; the fingers are
short, excavated within, and terminate bluntly in a dark corneous nail. The
ambulatory limbs are pubescent, particularly on the upper and lower mar-
gins; the distal end of the merus, and also the carpus, propodus, and dac-
tylus, are armed with numerous spines; the most prominent of these spines
are arranged in a row along the upper border of the carpus and propodus ; the
area about equal in length to the propodus and carpus together. The legs
of the fourth pair are furnished with long hairs on their upper margin, and
there are a few spines on the upper margin of the carpus; the propodus is
rather longer than the dactylus, and the rasping surface on its lower margin
occupies two thirds of its length. The last pair of legs is much less hairy
than the preceding pair; the rasping surface, which is truncate posteriorly,
falls a little short of reaching the middle of the hand. The telson is divided
by a pair of lateral incisions and a median one into four lobes, those on the
left side the larger; the pair of terminal lobes are obscurely toothed on their
* margins.

Length of carapace, 9.5 mm.; breadth, 7 mm.; length of ocular peduncle,
4mm.

Station 3368. 66 fathoms. 2 females ovig.

This species is nearly related to P. lymani A. M. Edw. et Bouv.,! of the
West Indian region, and to P. subpilosus Hend.,? of New Zealand. From the
former it is distinguished by the more prominent and acute rostrum which
overhangs the ocular segment, by the smaller number of spines on the antennal
acicle, and by the armature of the telson, the margin of which is ornamented
with obscure teeth, while in P. lymani it is furnished with numerous spines
whose tips are horny and dark colored, From P. subpilosus it differs in having
shorter eyestalks and antennal acicle, fewer spines on the antennal acicle, a
longer and narrower external prolongation of the second joint of the antenna,
and a differently shaped telson.

The ovisacs of both specimens are large and filled with eggs,

Family PORCELLANIDZE.
Petrolisthes agassizii, sp. nov.

In the shape of the carapace and the front this species bears a close resem-
blance to Petrolisthes sexspinosus (Gibbes) and P. occidentalis Stimps., but the
transverse ridges are more broken anteriorly, while posteriorly they extend
without interruption across the whole width of the carapace, being here more
perfectly developed than in the two species above named, The carpus and claw,

1 Mem. Mus. Comp. Zoöl., XIV. No. 3, p 49, Plate IV. Figs. 13-22, 1893.
2 Challenger Anomura, p. 77, Plate VIII. Figs. 2, 2 a, 1888.

MUSEUM OF COMPARATIVE ZOOLOGY. Bu

moreover, are longer and narrower ; the anterior margin of the carpus is three-
toothed instead of five-toothed. The squames of the carpus and claw do not
tend to widen out into ridges or folds either on the upper or lower face, but
preserve the form of close-set imbricated scales over the whole surface, includ-
ing the space between the longitudinal depressions of the carpus and along the
depressed line of the propodus, The form of the carpus approaches nearer
to that of P. armatus (Gibbes), but the present species may be readily dis-
tinguished from .P. armatus by the prominent ruge of the carapace and the
squames of the chelipeds. From P. edwardsii (Sauss.) it is distinguished by its
longer chelipeds, by the ridges of the hinder part of the carapace extending
clear across the carapace without interruption, etc. The ridges of the frontal
lobes are much more strongly developed in P. agassiaii than in any of the
allied species,

Carapace 5 X 5 mm.; cheliped, 34 mm.

One male, taken with Petrolisthes occidentalis Stimps. on the reef at Panama,
at low tide, March 12.

Pachycheles panamensis, sp. nov.

Carapace subcircular, of equal length and breadth ; upper surface flattened,
granulate, and setose, especially on the gastric and anterior part of the branchial
regions; posterior portions of the branchial regions lightly rugose. Front broad,
produced to a rounded median lobe, lateral portions straight, inner orbital
angle rounded, not produced; outer orbital angle projecting as a triangular
tooth between the eye and the base of the antenna, Lateral border of the
carapace with a concavity behind the antenna; the margin is slightly raised as
a rim which is lost in one of the branchial ruge before reaching the hind border
of the carapace; hind border slightly concave. Chelipeds: unequal, the right
being the larger; upper surface of the merus rugose and setose, under surface
smooth, internal distal angle produced as a denticulated tooth; carpus short
and broad, upper surface tuberculous and setose, internal border armed with
a large tooth at the base, followed by one or two smaller teeth, lower sur-
face smooth and naked; upper surface of the propodus furnished with small
tubercles which bear stiff bristles, inner margin rounded, not toothed, outer
margin granulated, lower surface convex, naked and polished near the centre
(where the surface is reticulated in the larger claw), squamoso-granular and
setose near the edges; fingers gaping, crossing at tips, toothless on both the
inner and outer margins (or at most slightly denticulated on the outer margin).
Ambulatory appendages setose,

Length of carapace, 6.5 mm.; breadth, 7 mm.; breadth of frontal margin,
2 mm.

Panama, March 12. 1 female ovig.

Pachycheles rudis Stimps. is a larger species, with the posterior margin of the
carapace more concave (the concavity having almost the shape of a triangular

176 BULLETIN OF THE

notch); the carapace and chelipeds are less setose than in P. panamensis,
the anterior margin of the carpus not prominently toothed, and there is a
strong protuberance near the middle of the upper surface of the propodus. In
P. tuberculipes Lockington, the central part of the front is triangular and
deeply furrowed along the median line, the chelipeds and ambulatory legs are
knobbed so as to present “a mass of tubercles above.” P. panamensis appears
to be near P. barbatus A. Milne Edwards, from the Azores, but in the latter
species the front is broader and the carpus more denticulated.

Family GALATEIDZE.

Pleuroncodes monodon (M. Edw.)?

Compared with Milne Edwards's figure of P. monodon (Ann. Sei. Nat., 3° sér.
Zool., XVI. Plate XI. Figs. 6-9), the “ Albatross” specimens present a more
obese appearance; their greatest width is across the cardiac region, while in the
figure of P. monodon it is near the posterior end of the carapace ; the cardiac
area, too, in the examples before me, is sunk below the level of the rest of the
carapace, and the transverse piliferous lines are more broken at this point, as
well as on the gastric region, than appears to be the case in P, monodon, to
judge from the figure referred to, Unless these discrepancies are due to the
inaccuracy of Milne Edwards’s draughtsman, the “ Albatross” specimens belong
to a new species,

Station 3385. 286 fathoms. 16 males, 7 females.

* 3386. 242 i^ 9 males, 14 females.
“ 3396, 259 K 2 males, 2 females.
“3493, 94 5 18 males, 11 females.

Munida obesa, sp. nov.

In this species we see an approach to the genus Plewroncodes, as the sides or
latero-inferior walls of the carapace are somewhat swollen, so that they show
a little when the animal is viewed from above. The basal joint of the an-
tenns, too, is more exposed from above than it is in the typical species of
Munida. The lateral rostral spines, or supra-ocular spines, are curved wpward
more than the median rostral spine, and the three are nearly parallel, the
lateral spines reaching about half way to the tip of the dian ; all three are
microscopically spinulose on their upper edge. Th ve two pairs of spines
on the anterior part of the gastric region in line with vae lateral rostral spines.
Of these two pairs the anterior is the larger. There is also a longitudinal line
of spinules in the median line between the two pairs just spoken of. There
are, in addition to these, several small spines irregularly arranged on each side
of the gastric region. The cardiac arca is somewhat sunk below the level of

MUSEUM OF COMPARATIVE ZOOLOGY. 177

the surrounding parts; its anterior margin is denticulated, with a larger spine
on each side. The lateral margins of the carapace are armed with ten or
eleven spines, the one on the antero-lateral angle being the longest. The
second abdominal segment is ornamented with a transverse row of eight small
spines. The other abdominal segments are normally destitute of spines,
but in a few of the many specimens before me there are two or four small
spinules on the third segment. The pleure of the third, fourth, and sixth
abdominal segments are acute, the rest blunt. The eyes are large, and are
provided with rather long cilia on the edge of the cornea. The basal joint
of the antenna is armed with a long and sharp spine which reaches forward
beyond the eyes; the second joint also has a long spine on each side. The
chelipeds are long and hairy; the merus, carpus, and basal part of the pro-
podus are spiny, the fingers long, slender, the cutting edges straight and
finely spinulose, The ambulatory appendages are setose, the upper and lower
edges of the merus are spinulose and there is, moreover, a row of spinules on
the outer surface, this external line of spinules being best developed on the
proximal end of the segment. The carpus is armed with small spines on the
upper margin, and one spine on the distal end of the lower margin; the penul-
timate and terminal joints are unarmed.

Length, 65 mm.; length of carapace, 34.5 mm.; breadth of carapace between
epimeral sutures, 21 mm.;' length of rostrum, 11 mm., length of cheliped,
84 mm., merus, 26 mm., carpus, 10 mm., basal portion of chela, 18 mm.,
dactylus, 21 mm.

Station 3389. 910 fathoms. 2 males, 7 females.

win DOVO ou "e 5 young.

Munida refulgens, sp. nov.

In this species the sete on the ridges of the thorax and abdomen and on the
legs are resplendent with iridescent hues. The rostrum is long, triangular in
cross section, the upper surface scabrous, the lateral margins armed with two to
four spines which are generally placed unsymmetrically on the two sides. The
supra-ocular spines are short. There is a transverse line of spinules back of
the base of the rostrum, the two which lie on either side of the median line
being larger than the others. Seven marginal spines on each side of the
carapace, the ones at the antero-lateral angles the largest. There are no spines
on the abdominal segments. The abdominal pleure are acute. The basal
joint of the antenna has a plate-like expansion, but is not spinose ; the second
joint is furnished with an external spine. Chelipeds very long, squamose, and
clothed with silky sete; the merus has a row of spines on the upper margin,
another on the inner side, and a row of smaller ones on the outer side ; the
carpus is provided with three or four spinules at the distal end; the chela
is slender, the outer finger flattened, ribbed above, the outer edge rather
convex and expanded toward the base; cutting edges of fingers finely denticu-

178 BULLETIN OF THE

lated. The anterior border of the merus and carpus of the ambulatory
appendages is spinose. The general color in life is red. In the alcoholic
specimens the color is retained in the chel and particularly in the rostrum.

Dimensions of largest specimen (male). Length, 91 mm.; length of carapace,
including rostrum, 43 mm. ; breadth of carapace, 34 mm.; length of cheliped,
211 mm., merus, 90 mm., carpus, 15 mm., basal part of chela, 56 mm.,
dactylus, 42 mm.

Station 3367. 100 fathoms. 13 males, 18 females (7 ovig.).

“08918, 112 e 15 males, 19 females (14 ovig.).
« 3379. 52 “ 1 young.
Cede SU 1 young.

In M. iris A. M. Edw., the sete are iridescent, as in this species, but the
rostrum lacks the lateral spines and the supra-ocular spines are much longer,
reaching beyond the eyes ; from M. irrasa A. M. Edw., our species differs in
the shape of the hand, in the relatively shorter median rostral spine provided
with lateral spines, etc.

Munida propinqua, sp. nov.

The carapace of this species is rather flat ; the rostral spines are scabrous,
the lateral reaching to a point beyond the eyes; there is a prominent spine on
the gastric area behind each lateral rostral spine, another on each side behind
and external to these, and a pair of very small ones on the median line at the
hase of the rostrum; besides these there are about four small spines on the
anterior half of the carapace. The anterior lateral angle of the carapace is
truncated, the lateral border seven-spined. The pleure of the abdomen are
rounded, short, and broad ; the second abdominal segment is furnished with a
transverse row of about eight spines, the rest of the segments being destitute of
spines ; the terga of the second to the fourth segments are very smooth behind
the central transverse fosse, The first joint of the antenna is armed with a
long spine, the second joint with one on each side. There is a minute spine at
the antero-inferior angle of the carapace. The chelipeds are robust, setose,
and spiny ; merus spiny on upper and inner parts ; carpus spiny on all sides ;
The hand is furnished with two rows of spines on the lower side, another along
the middle of the'outer face, and three irregular series along the upper side ;
both the fingers are spinulose. Ambulatory limbs setose, spinose along the
superior and inferior edges. In small specimens the lateral spines of the
rostrum may be shorter than the eyestalks.

Length, 84 mm.; carapace (including the rostrum), 45.5 ram. ; breadth,
26 mm. ; length of cheliped, 96 mm.

Station 3384. 458 fathoms. 11 males, 6 females (1 ovig.).

Hise 8304, OLLE, e 1 male.
“ 3404, 385 “ 1 male juv.

This species resembles M. miles A. M. Edw., but the carapace of M. pro-
pinqua is flatter, the cardiac area is more distinctly circumscribed by a furrow,

MUSEUM OF COMPARATIVE ZOOLOGY. 179

the abdomen bears spines only on the second segment, and the abdominal
segments are not so much sculptured.

Munida gracilipes, sp. nov.

Carapace rather flat and quadrangular. Lateral spines of rostrum less than
one half the length of the rostrum, shorter than the ocular peduncle. Four
spinules on the gastric area arranged in the form of a square, — two behind each
lateral rostral spine ; a longitudinal row of obsolescent spinules in the median
line from base of rostrum to the cardiac area ; one spine on the cardiac region,
a pair on the intestinal region, and one on each side of the cardiac region just
back of the cervical suture. The lateral margins of the carapace are armed
with about seven spines, the first of which is the largest. The second abdomi-
nal segment is armed with a transverse row of six spines, the third with a row
of four, the fourth with a row of four and one median spine behind the trans-
verse row. This is the normal arrangement, but in one specimen out of the
four there is an additional pair of spines on the second and third segments back
of the transverse row. The pleure of the third, fourth, and fifth abdominal
segments are acute. Eyes very large, reniform. The chelipeds are very long
and slender, the merus spinose (the chief of the spines being on the inner side
of the joint); the carpus also is spinose. The hand has about eight spines on
the upper margin and one on the lower; there are several acute spines on the
outer border of the movable finger, the cutting edges of the fingers are straight,
finely denticulated or spinulose.

Length, 24 mm.; breadth, 8 mm.; length of cheliped, 34 mm.

Station 3391. 153 fathoms. 4 specimens.

This species is very near to M. stimpsont A. M. Edw., but the carapace is
flatter, less granulated, more quadrangular in outline, with more evident trans-
verse ruge ; the lateral rostral horns are shorter, the eye larger, the transverse
ridges on the abdominal somites fewer in number; the lateral spines of the
carapace and the abdominal spines are better developed, the cardiac area nar-
rower and bounded by more distinct furrows.

Munida microphthalma A. M. Enw. ?
Bull. Mus. Comp. Zoöl, VIII. 51, 1880.

Station 3370. 134 fathoms. 1 female ovig. Length, 20 mm.

Only one specimen of M. microphthalma has been returned to this Museum
from Paris. It is a very small specimen, without chelipeds. The “ Albatross”
specimen differs from this one in having the rostral median spine less upturned.
The chela, compared with that of M. microphthalma, as figured by Henderson
(Rep. Challenger Anomura, Plate IIT. Fig. 4), has no spine on the outer mar-
gin of the dactylus, and the row of spines on the outer face of the hand is
obsolete.

VOL. XXIV. — NO. T. 3

180 BULLETIN OF THE

M. microphthalma was taken by the “Blake” among the West Indies in
573-1030 fathoms, and by the “Challenger” in the same region in 390 fathoms,
north of the Kermadec Islands in 600 fathoms, and near Ascension Island in
425 fathoms.

Galacantha rostrata A. M. Epw.

Galacantha rostrata A. M. Epw., Bull. Mus. Comp. Zoöl., VIIT. 52, 1880.
Galacantha bellis HexD., Ann. Mag. Nat. Hist., 5th Ser., XVI. 418, 1885; Rep.
Challenger Anomura, p. 167, Plate XIX. Fig. 6, 1888.

Station 2362. 1175 fathoms. 1 male.

O LOL © 3 males, 2 females (1 ovig.).
war LBOU. t 1 female.

The “ Albatross” specimens differ from the typical West Indian form in the
greater divergence of the lateral spines, the anterior being more nearly parallel
with the axis of the body; the abdomen. is smoother toward the central part
of the segments; the dorsal spine of the fourth abdominal segment is smaller.
There is-considerable variation among different individuals, and the characters
pointed out by Henderson to distinguish G. bellis can hardly be deemed of spe-
cific value. The color in life is red, paler and yellowish toward the middle of
the carapace. G. rostrata has been collected in the West Indian seas in from
1098 to 1591 fathoms, and off Juan Fernandez in 1375 fathoms (G. bellis
Hend.).

Galacantha diomedese, sp. nov.

Rostrum without lateral spines; distal part turned upward at an angle of
less than 95° in most specimens, but in some cases the inclination is greater;
basal part marginate, the margin running for some distance along the anterior
edge of the carapace ; a slight keel runs back from the rostrum to the median
gastric spine. Gastric spine smaller than in G. rostrata ; anterior lateral spine
much longer than the posterior; there is an additional small spine on the side
of the carapace, just behind the cervical suture. Anterior half of the carapace
ornamented with setiferous squamous tubercles; on the posterior half of the
'arapace the tubercles assume the form of interrupted transverse ridges. The
median spines of the abdomen are small, diminishing successively in size from
the first to the third, which is obsolete in some examples. Upper surface of
abdomen rather hairy, the pleure tuberculose, angles rounded. The legs are
rough with granular setose tubercles. There are two prominent spines at the
distal end of the carpus of the chelipeds, and one at the distal end of the carpus
and merus of the ambulatory limbs. The antenne are twice and a half as long
as the body.

Dimensions of a female specimen. Length of body, 79 mm.; length of cara-
pace, 39 mm. ; breadth of carapace, not including the lateral spines, 25 mm.

This species runs into a well marked variety, in which the anterior lateral
spines, as well as the median gastric spine, are very much smaller than in the

MUSEUM OF COMPARATIVE ZOOLOGY. 181

typical form. This variety may be called Galacantha diomedew parvispina. At
one station (3429) both forms were obtained at the same haul.

This species differs from G. rostrata, areolata, spinosa, and talismanii in the
rugose nature of the sculpture of the hinder half of the carapace. In the
relative proportion of the anterior and posterior lateral spines it agrees with
G. spinosa.

Station 3357. — 782 fathoms, 1 female juv.

OB te ore. re 3 males, 3 females ovig.

E 18904 902 Y 1 female.

ee OOOO 4007 t 3 males, 1 female ovig.
BON y Ona t 5 males, 2 females (1 ovig.).
mt IST, © 1 male.

ua. 1020, O 3 males.

Me 0805. | 380 — ^ 2 males, 1 female,

* 3499. 919 “ 1 male.

Var. parvispina.
Station 3418, — 660 fathoms. 1 male.

Du wre (IB 1 female ovig.

See odad 010. S 1 male.

St COIN 1-DIUT f l male.

Tu O a OOD oV 18 males, 17 females (6 ovig.).
n Bad BD 6 males, 4 females (3 ovig.).

In both G. rostrata and G. diomedew there is a curious sexual difference. In
the male the proximal half of the telson is furnished on each side with long
amber-colored: sete, which are entirely wanting in the female. The same
difference between the sexes is found in some species of Munidopsis.

G. diomedeæ is often infested with parasites. One of the males from Station
3371 bears a Peltogaster, while seven specimens (5 males, 2 females) of var.
parvispina house a Bopyrus in the left branchial chamber.

The eggs of this species measure 3 X 2.5 mm.

Munidopsis! vicina, sp. nov.

Near M. ciliata Wood-Mason, from which it differs as follows. Tt is'a very
much smaller species, the adult ovigerous female being only twenty-nine milli-
meters long; the anterior margin of the propodus of the ambulatory appen-
dages bears two very prominent spines. The telson lacks the pair of long and
narrow plates which lie on each side of the small central plate in M. ciliata,
As in M. ciliata, the carapace is covered with squamous tubercles, the rostrum
is curved slightly upward, and the chela is short,

Length, 29 mm. ; breadth, 9.5 mm.

Station 3360. 1672 fathoms. 1 female.

ap Og. i 1 female ovig.

1 The genus Munidopsis, as here understood, includes Galatodes, Orophorrhynchus,
and Llasmonotus of A. Milne Edwards, and Anoplonotus of Smith.

BULLETIN OF THE

Munidopsis agassizii, sp. nov.

The carapace of this species is moderately convex, with a deep transverse
depression across the anterior part of the cardiac area. The rostrum is long,
slightly upturned, and armed near the middle with a pair of lateral spines.
The gastric area has four pairs of spines, the anterior pair the largest. The
cardiac area bears two or three pairs of spines, The lateral margins of the
carapace carry from six to eight spines each, and there is a longitudinal series
of small spines within the margin on the branchial area. A small spine is
situated on the anterior margin between the eye and the antenna. The pos-
terior border of the carapace is ornamented with six (in one specimen seven)
spines. There are also several spines on the sides of the carapace below the
epimeral suture, There is a very small spine over each eye. The antenne
are shorter than the body, the first joint bears a long external spine, the second
joint two lateral spines, the third joint two lateral spines and one superior. The
chelipeds are long and slender, the merus and carpus have no long spines, the
propodus carries four spines on the upper edge and several rudimentary spin-
ules, the fingers are spinulose, their cutting edges straight and denticulated.
The ambulatory appendages have spiny meri and carpi, the longest spines
being one at the distal superior border of each of these joints. The second,
third, and fourth abdominal segments bear four spines each, The abdominal
pleure are rounded.

Length, 23 mm. ; length of carapace, 12.4 mm»; breadth of carapace, 8 mm.;
length of rostrum, 4.5 mm.

Station 3389. 210 fathoms, 1 male, 1 female.

This species bears a general resemblance to M. erinacea (A. M. Edw.) and
M. spinifera (A. M. Edw.). It differs from both these in having a flatter
carapace marked by a deeper transverse depression across the cardiac area, in
having a larger number of spines on the sides of the carapace, and in the pres-
ence of spines on the pterygostomian regions and a small but distinct spine
over the eye. It also has strong spines on the superior edge of the hand
which are wanting in M. erinacea and M. spinifera. In the possession of
three pairs of gastric spines it agrees with M. spinifera, but differs from JM.

erinacea.

Munidopsis villosa, sp. nov.

The whole surface of the body and limbs is beset with seta, which arise
from low squamous tubercles and transverse ruge on the carapace, and from
the transverse ridges of the abdominal segments, The rostrum is triangular,
the distal half upturned, cylindrical, and pointed, the proximal half naked
below and slightly carinated in the median line. A pair of short, stout, blunt
spines on the gastric region. One spine at antero-lateral angle of the carapace,
one on margin of the hepatic area, and a rudimentary one on the side of the

MUSEUM OF COMPARATIVE ZOOLOGY, 183

branchial region. There is a medium spine on the second, third, and fifth ab-
dominal segments, and a rudiment of one on the fourth. The abdominal pleure
have rounded external angles. The eyes are freely movable and destitute of
spines. The second antennal segment is armed with a prominent external
spine. The chelipeds are robust, setose, and granulate ; the merus has a short
superior spine and two lateral spines at the distal end; the carpus is similarly
equipped, though on one side the superior spine is obsolescent ; the chela is
broad and strong, the fingers excavated, denticulated on their cutting edges
and at their tips, The merus of the first pair of ambulatory appendages has
an external distal spine ; the carpus of all the ambulatory limbs has two longi-
tudinal ridges, and that of the first and second pair has a spine on the upper
border at the distal end of the joint.

Length, 55 mm.; breadth, 18 mm.; length of carapace, 31 mm. ; rostrum,
8 mm,

Station, 3394. 511 fathoms. 1 male.

Munidopsis hystrix, sp. nov.

Jarapace setose and thickly covered with small spiny tubercles; three
spines of special prominence on the gastric area disposed in the form of a
triangle with apex directed backward; one on the cardiac area; two (rarely
six) on the hind margin of the carapace; one on each branchial area. There
is a spine at the external angle of the orbit, and the lateral margin of the cara-
pace is spinose. The rostrum is long, lightly curved upward from the base to
the tip, and armed with from two to five spines on each side; these spines
are unsymmetrically arranged on the two sides. The second, third, and
fourth abdominal segments are conspicuously two-ridged; the second seement
has a pair of small spines on the anterior ridge, and another pair nearer the
median line on the posterior ridge; the third segment also has a pair of spines
on the anterior ridge, and in some specimens a third spine in the median line
on the posterior ridge. The abdominal pleure are truncate, The chelipeds
are long, very spiny from the proximal end of the merus to the base of the
fingers; the chief spines of the propodus are on the upper margin of the seg-
ment; there are two spines near the base of the dactylus. The ambulatory
appendages are long, setose, and spinose except the daetylus joint. A spine
over the eye, Antenne shorter than the body; a spine on the outer side of
the first segment, one on each side of the second and third segments, and one
on the upper side of the third segment.

Length of ovigerous female, 47 mm. ; length of carapace, 25 mm. ; breadth,
15 mm. ; rostrum, 8 mm.

Station 2417. 493 fathoms. 1 male, 2 females ovig.

‘903424, 1078 =) € 4 females (2 ovig.).
ABD, RIO ot 7 males, 5 females (2 ovig.).

BULLETIN OF THE

Munidopsis sericea, sp. nov.

The whole surface of the body and limbs is covered with a silky pubes-
cence. The rostrum is long, curved gently upward, convex above, but not
carinated, armed with a prominent spine on each side near the middle, and
with three more minute spinules near the base. Gastric region swollen, armed
with two conical spines and ten or twelve small spinuloid tubercles. The
cardiac region hag a prominent transverse ridge near the centre, in front of
which is a deep depression separating it from the gastric region; the ridge is
armed with a pair of short spinules. There is a small spine on the anterior
border between the eye and the antenna, a large one at the antero-external
angle, three on the border of each hepatic region (the middle of one of these is
the largest), and one small one on the border of each branchial region just
behind the cervical suture; there are besides about ten sharp tubercles on
each. branchial area, and five or six pairs of spinules on the posterior margin of
the carapace. Pterygostomian regions granulated, There isa pair of spines on.
the second, third, and fourth abdominal segments ; besides these there are sev-
eral small spinules on the terga and pleurw of these segments; the pleura
are rather narrow, with rounded lateral angles. The chelipeds are wanting in
the unique specimen, The ambulatory appendages are spinulose, the spinules
of the dactyli restricted to the hind margin. The eye is provided with a very
minute spine. The antenne are rather longer than the body, the basal joint
has a short external spine, a longer one at the lower internal angle, and a
small one at a higher level on the inner side, The latter spine shows, when
the animal is viewed from above, between the eyestalk and the antenna, The
subsequent segments of the antenna are armed as usual in this genus,

Length, 39 mm.; length of carapace, 12 mm, ; length of rostrum, 8 mm. 5.
breadth of carapace, 12 mm.

Station 3394. 511 fathoms. 1 male.

Munidopsis margarita, sp. nov.

In this species the rostrum has a gentle upward curve near the tip; it is
carinate above, and minutely spinulous on the margins. ‘The surface of the
carapace is rough with squamous tubercles and forward-pointed spines. The
gastric and cardiac regions are prominent, and separated from one another by
a deep depression ; a pair of spines on the gastric, and one spine on the cardiac
region, attain a special prominence. A long sharp spine outside the eye forms
the outer wall of a well marked orbit. There are eight spines on each lateral
margin, six on the posterior (including those at the postero-lateral angles).
The branchial areas are iridescent. Second abdominal segment: the anterior
transverse ridge, which is broken down in the centre, bears on each side a
prominent hooked spine, which is enlarged at the base and denticulated on the
outer margin ; the posterior ridge is furnished with three hooked spines ; the

D

MUSEUM OF COMPARATIVE ZOOLOGY. 185

pleur of this segment bear each a broad, flattened, forward-pointing tooth
with denticulated edges; when the animal is viewed from above, this tooth
appears to form the lateral extremity of the pleura, which really lies below it,
and is rounded. Third abdominal segment: both ridges are spiny and dentic-
ulate, three spines being specially prominent on each ridge. Fourth abdominal
segment : armed with but one small median spinule. The sides of the carapace
below the epimeral sutures are covered with spiny tubercles, and display an iri-
descent lustre. The eye has two spines projecting over the cornea from the
inner side; the posterior of these spines is very minute. The antenne are very
slender and about as long as the carapace; the first and second joints are pro-
vided with a prominent external spine, the third joint with three spines, viz.
one external, one internal, and one superior. The chelipeds are absent in both
the specimens. The ambulatory appendages are spinulose on all the segments
except the dactyli, which are finely serrate on the hind margin. The legs, and
more especially the sternum, are iridescent, like mother of pearl. This irides-
cence is seen in a less degree in several other species of this genus.

Length, 20 mm.; length of carapace, 11 mm. ; breadth, 7 mm.; length of
rostrum, 3.5 mm.

Station 3404, 385 fathoms. 1 male, 1 female.

Munidopsis crinita, sp. nov.

The whole surface is clothed with long sete, which are longest and densest
on the chelipeds and ambulatory appendages: The rostrum is very broad at
the base and ends in three points, the middle of which is the longest; the ros-
trum is slightly carinate in the median line. ‘The carapace is roughened by
low setiferous ridges. The antero-lateral angles are obliquely truncate; a spine
over the antenna, and four on the lateral margin, the last one just behind the
cervical suture, the third one obsolescent ; hind margin unarmed. A pair of
spines on the gastric region, behind the base of the rostrum. The abdomen is
devoid of spines, and there is no spine over the eye. The antenne are slender,
shorter than the body; the basal joint is provided with a long spine on the
external side, and another on the internal side; the second joint has an external
spine, the third an internal one. Chelipeds : merus five-spined; carpus with
one prominent spine; hand unarmed, broadest at base of fingers, cutting edges
of fingers toothed. Ambulatory limbs setose, hind border of dactyli spinulose.

Length, 19.5 mm.; carapace, 11.5 mm. ; rostrum, 2.6 mm. ; breadth of cara-
pace, 7.5 mm.

Station 3384, 458 fathoms. 1 female.

This species resembles M. rosacea (A. M. Edw.), M. latifrons (A. M. Edw.),
and M. tridens (A. M. Edw.). From the first (Comptes Rendus, XCIII. 934,
figured in Recueil de Figures de Crustacés nowy. ou peu connus, 1% livr.) it differs
in having a much shorter rostrum, in the presence of a pair of spines on the
gastric region, in the different shape of the hand, the absence of prominent

186 BULLETIN OF THE
spines on the meri of the ambulatory legs, and its greater pilosity. From the
second (judging from Milne Edwards’s short description of that species) it is
distinguished by the long sete, gastric spines, and broader carapace. From
the last it differs in being very hairy, eto. M. rosacea comes from the north
coast of Spain, M. latifrons from the Barbadoes, M. tridens from St. Kitts.

Munidopsis ornata, sp. nov.

Carapace convex, the whole upper surface, including the rostrum, thickly
covered with low squamous tubercles; seen under a magnifying power the
surface of each tubercle is seen to be made up of a number of secondary scale-
like prominences; the tubercles are not lengthened out transversely to form
ridges on any part of the surface; two of the tubercles on the gastric region
take on a spiny character. The rostrum is nearly horizontal, triangular in
cross-section, the margins serrate; the anterior border of the carapace is convex
between the eyes and the antenne, but has no spine at this point; lateral bor-
der four-toothed, one of the teeth lying atthe antero-lateral angle, two on the
hepatic region, and one on the edge of the branchial region behind the cervical
suture; the posterior border is delicately festooned, but not armed with spines.
The abdomen is spineless, its surface punctate, anterior half of the pleurs of
the second segment tubereulate, all the pleura rounded. The eye has a trans-
verse granulated tubercle running over the cornea from the inner side. The
antenne are very slender, and do not exceed the carapace in length. The
chelipeds are moderately robust ; the merus tubereulate and armed with a row
of short spines along the upper edge; the carpus spino-tuberculate, with two
longitudinal furrows on the outer side; the hand almost smooth on the inner
side, outer side and superior surface roughened with low tubercles ; fingers a
little curved upward, spoon-shaped at the denticulate and setose tips. Ambu-
latory appendages: meri flattened, tuberculate, upper edge produced to a spi-
nose carina; the carpi have three denticulate ridges; propodi scabrous, with
an irregular row of spines on under side; the dactyli have black tips, and are

finely spinulose on their posterior edges.

Length, 23 mm.; length of carapace, 12 mm." breadth, 8 mm.; length of
rostrum, 3 mm.

Station 3404. 385 fathoms. 1 male,

Munidopsis scabra, sp. nov.

The rostrum is triangular, slightly curved upward, carinated above, the
lateral edges and the carina lightly denticulated. The carapace is covered
with squamous setiferous tubercles which end in spiny points. There is a
transverse row of six more prominent spiny tubercles on the gastric region.
The posterior border of the carapace is ornamented with a denticulated rim
(about eight denticles). There is a spine between the eye and the antenna

MUSEUM OF COMPARATIVE ZOOLOGY. 187

below the anterior margin of the carapace. The abdomen is devoid of spines,
the pleure have truncated lateral angles. A spine projects over the cornea of
the eye. The antennm are shorter than the body; a spine on the outer side
of the basal joint, one on each side of the second joint, and one on each side
and one on superior margin of third joint. The chelipeds are long, spinose,
except the fingers ; hand lung, the basal part longer than the fingers. All the
joints of the ambulatory appendages are spiny except the dactyli.

Length (ovigerous female), 40 mm. ; length of carapace, 13.5 mm.; breadth,
14 mm.; rostrum, 5 mm.

Station 3424. 676 fathoms, 2 males, 1 female ovig.

682159490, 08078 1 male, 1 female ovig.

Munidopsis tanneri, sp. nov.

Carapace flat, quadrangular, covered with squamous setiferous tubercles
which have a tendency to develop spiny points on the gastric region. This
is especially true of a transverse row of six on the anterior part of that region.
The rostrum is triangular and horizontal. There is a prominent spine on each
side of the anterior margin of the carapace between the eye and the antenna,
another at the antero-lateral angle, and two or three on the side of the hepatic
area; the hind border,of the carapace is denticulated. A small spine over the
eye. Antenne shorter than the body ; one spine on the outer side of the first

‘joint, two lateral and one superior on the second and third joints. Cheliped
(present in only one specimen) long, slender ; merus and carpus many-spined ;
propodus spiny along the upper and lower margins ; tips of fingers enlarged
and denticulated. Ambulatory limbs : a prominent row of spines on the upper
edge of the merus and carpus, propodus and dactylus devoid of spines. Abdo-
men without spines; pleura narrow, angles rounded.

Length, 41 mm.; length of carapace, 23.5 mm.; breadth, 15.5 mm.; rostrum,
6 mm.

Station 3396. 259 fathoms. 2 males, 1 female (1 male with Bopyrus).

ERA ps 1 male.

This species is nearly related to M. scabrosa, but differs from the latter spe-
cies in having the carapace broader and flatter, with squamous tubercles which
are not produced into points except a few on the gastric area, The spine be-
tween the eye and the antenna is longer; the propodi of the ambulatory legs
are smoother, with no well-developed spines.

Munidopsis hamata, sp. nov.

Body and limbs clothed with short, scattered sete. Rostrum long, curved.
slightly upward, basal half furrowed longitudinally, with a row of short spines
on each side of the furrow ; infero-lateral edges of rostrum also furnished with
small spines. Carapace quadrangular, anterior border forming a right angle
with lateral border, both borders spinulose ; lateral border with an indentation

188 BULLETIN OF THE

at anterior boundary of hepatic area; a deep depression back of each hepatic
area and another across the anterior part of the cardiac region; the upper surface
of the carapace is adorned with spinulose tubercles, and a median longitudinal
row of more prominent spines runs along the gastric and cardiac regions; the
anterior spine of the cardiac region overhangs the transverse depression, the
posterior spine of the row springs from the hinder rim of the carapace. There
is a median hooked spine on the tergum of the second, third, fourth, and fifth
abdominal segments and many spiny tubercles irregularly disposed on these
segments ; the pleura of the third to the sixth abdominal segments are narrow
but blunt, those of the second to the fifth are costate, The ocular peduncle is
movable and devoid of a spine. The antenne are about as long as the body ;
the basal joint has an inferior and a small external spine ; the second joint also
bears an external spine. The chelipeds are long and slender ; all the joints from
the ischium to the propodus are equipped with longitudinal rows of small
spines ; the chela is not broader than the basal part of the propodus, the fingers
are straight, their prehensile edges denticulate. The ambulatory appendages
are spinulose.

Length of male, 49 mm.; length of carapace, 25 mm.; breadth of carapace,
14 mm.; length of rostrum, 9 mm.; length of cheliped, 47.5 mm.; merus,
15 mm.; carpus, 5.5 min.; chela, 19 mm.

Station 3394. 511 fathoms. 13 males, 16 females ovig.

* 8395. 730 « 3 males.

Munidopsis aspera (Henp.).

Elasmonotus asper Hend., Ann. Mag. Nat, Hist., 5th ser., XVI. 416, 1885; Rep.
Challenger Anomura, p. 163, Plate XIX. Fig. 4, 1888.

Station 3357. 782 fathoms. 1 female ovig.

000. -UDU. alt 1 male.

2. .88104,:184 a. 6 1 female.

sty, 040%: A % 2 males, 5 females (3 ovig.).
d 3403. 384 ít 1 male.

* 3406. 551 i 2 males.

This species is subject to considerable variation. In the specimens from
Stations 3402, 3403, and 3406 the tubercles of the carapace are more numerous
and less spiny than in those secured at the other stations. The ambulatory
appendages of all the “ Albatross” examples are apparently more spiny than
in the types from the “Challenger.” The-latter came from the Straits of
Magellan, 245 fathoms.

Munidopsis quadrata, sp. nov.

Carapace quadrangular, the anterior and lateral margins forming a right an-
gle; upper surface flat, spineless, but furnished with low squamiform tubercles,

MUSEUM OF COMPARATIVE ZOOLOGY. 189

Rostrum curved upward, broad at base, narrowing anteriorly to form a long,
sharp acumen, Central part of gastric region prominent above the hepatic
region, from which it is separated by a deep pit. A prominent transverse ridge
on cardiac region, forming the posterior wall of a deep fossa. Antero-lateral an-
gles rounded. Second segment of abdomen armed. with a median spine which is
curved forward; third and fourth segments with a very prominent ridge which
bears an acute median tooth ; pleure of second segment faintly tuberculate, the
others narrow with the external angles rounded but not truncate. Eye spine-
less, almost concealed by the base of the rostrum, Antenne about as long as
the carapace ; a conspicuous spine on the upper side of the third segment.
Cheliped long, tuberculate with the exception of the fingers; chela slender,
fingers not gaping, Ambulatory legs tuberculate with the exception of the
dactyli, which are furnished with small teeth on the posterior margin.

Length of body, 29 mm. ; length of carapace, 15.5 mm.; breadth of carapace,
9 mm.; length of rostrum, 6 mm.; length of cheliped, 30 mm.

There is some variation in the length and upward curvature of the rostrum
among the different specimens. A female, from station 3424, differs markedly
from the males in having the tubercles on the carapace and appendages much
more strongly developed.

Station 3424. 676 fathoms. 2 males, 1 female ovig.

“ 3425. 680 “ 1 male.

Munidopsis depressa, sp. nov.

Closely allied to M. hamata, but differs as follows. The cephalothorax is
more swollen, so that the sides of the carapace are visible below the epimeral
sutures when the animal is viewed from above. The median row of spines on
the carapace consists of a smaller number of spines (two on the gastric region,
one on the cardiàc region, and one on the posterior margin). The spinules of the
lateral margin of the carapace are less developed. The depression on the cara-
pace involves the gastric region to a greater degree. The anterior margin of the
carapace is not so straight, and it is not spinuliferous. The antero-lateral spine
is.more prominent, the eyes smaller, and the antenne shorter (shorter than the
carapace). There is, moreover, no spine on the fifth abdominal segment,

Length, 32 mm.; carapace, 19 mm.; rostrum, 5 mm.; breadth of carapace,
12.5 mm.

Station 3425. 680 fathoms. 1 male,

Munidopsis carinipes, sp. nov.

Carapace quadrangular, flat, marked by a median tubereulated ridge on the
gastric and cardiac regions ; sides converging a little from front backward ; the
antero-lateral angles form a rounded shoulder. Rostrum broad.at base, nearly

horizontal, sides converging near tip, which is blunt; upper surface nearly

190 BULLETIN OF THE

flat, lightly granulated. The rest of the upper surface of the carapace has a
coarser granulation. There is a conspicuous hooked tooth on the third and the
fourth abdominal segment, and in some specimens there is a rudimentary one
on the second and the fifth segment; the teeth on the third and fourth seg-
ments have denticulated margins in adult specimens ; abdominal pleuræ long
and narrow. Chelipeds very long, lightly tuberculate; chela long, slender,
fingers rather short, smooth, with straight denticulated prehensile margins.
The meri of the ambulatory legs granulated, superior border produced to a
keel, the edge of which is entire; the lower margin of the meri is also en-
tire ; the carpi have three tuberculated ridges, one superior, two external ; the
propodi are lightly tuberculated ; dactyli smooth, their hind margin armed
with about five teeth. Eye spineless, nearly hidden under the rostrum. An-
tenn shorter than the carapace, first, third, and fourth joints armed with an
external spine.

Length, 30 mm.; carapace, 16 mm.; breadth, 9.5 mm.; rostrum, 5 mm.;
cheliped, 40 mm.; merus, 13 mm.; carpus, 4.5 mm.; propodus, 17 mm.;
dactylus, 7 mm.

Station 3353. 695 lathoms. 2 males, 1 female ovig.

Near M. longimana (Elasmonotus longimanus A. M. Edw.), from which it
differs in having the rostrum more nearly plane, the merus of the cheliped
much less strongly tuberculated, the meri of the ambulatory limbs more
Strongly carinated, with lower margin entire instead of denticulate ; the spine
on antennal peduncle is more prominent, while the tooth on she second seg-
ment of the abdomen is absent or at best rudimentary.

Two specimens (male) of Elasmonotus longimanus A. M. Edw., and one
(female) Æ. brevimanus A. M. Edw., have been returned to Cambridge from
Paris. I suspect that these may prove to be the male and female of one
species. The chelipeds of the female specimen of M. cristatipes are lost.

Munidopsis hendersoniana, sp. nov.

In this species, as in M. armata (Elasmonotus armatus A. M. Edw.) and
AM. marginata, (Elasmonotus marginatus Hend.), the lateral margins of the
carapace are extended as sharp crests overhanging the sides of the body. The
upper surface of the carapace is rather flat, and is clothed with a close short
pubescence ; the sides are neatly parallel. The rostrum is long, acute, nearly
horizontal, the upper surface roof-shaped. There is an acute tooth on the
anterior margin of the carapace external to the eyestalk and another at the
antero-lateral angle ; otherwise the carapace is unarmed. The eyestalks are
immovable and prolonged into a long horn one half as long as the rostrum;
seen from above, the eyéstalks appear like lateral spines of the rostrum.
The eye is rudimentary, occupying the basal part of the lower side of the
peduncle. The antenne are shorter than the body, the basal joint armed with

a well developed inferior spine. The chelipeds are short robust and tomen-

MUSEUM OF COMPARATIVE ZOOLOGY, 191

tose; there is a spine at the distal superior angle of the ischium and another
near the distal end of the lower internal edge ; five spines along the superior
margin of the merus and two inferior distal spines ; the carpus bears a superior
proximal tooth together with three teeth on the distal margin ; the chela is
short and stout, the hand without teeth or spines; the fingers are very thick
and short, meeting one another only at their spoon-shaped denticulated tips ;
there is a rounded tubercle at the base of the inner margin of the immovable
finger; the outer margin of this finger is denticulated. Ambulatory limbs :
five to seven spines on the superior and external inferior margins of the meri
(those on the superior margin the largest); upper edge of carpus three- to four-
spined ; propodi and dactyli unarmed. Abdomen without spines.

Length, 37 mm.; carapace, 20 mm.; rostrum, 6.5 mm.; breadth of carapace,
12 mm.; length of cheliped, 28 mm.

Station 3393. 1020 fathoms. 3 males, 1 female (with Peltogaster).

Nearly related to M. edwardsii (Elasmonotus edwardsii Wood-Mason, Ann.
Mag. Nat. Hist., 6th series, VII. 201, 1891) of the Bay of Bengal, but easily
distinguished from that species by the lateral margins of the carapace, which
*n Wood-Mason’s species are divided into two lobes, but in M. hendersoniana
are entire.

Munidopsis inermis, sp. nov.

In this species the whole surface of the body and appendages is naked and
free from spines and tubercles. The carapace is rather flat above, with
subparallel sides; the gastric region is protuberant and separated from the
hepatic and cardiac areas by conspicuous furrows. The surface of the carapace
is punctate and lightly granulate and rugose on the branchial regions. The
rostrum is triangular, blunt at the apex, bent strongly downward, and slightly
carinate above. The antero-lateral angle is rounded, and a rounded lobe
proiects from the anterior margin above the base of the antenna. The
abdomen is smooth, naked, devoid of spines and ridges ; the abdominal pleuræ
are rounded, Ocular peduncle free, spineless. The peduncle of the antenna
is also destitute of spines; the flagellum is wanting in the only specimen
obtained. ‘The chelipeds are also missing. The ambulatory appendages are
smooth, unarmed ; the dactyli long (equal to the propodi in length), slightly
curved, acute at the tips. The appendages of the third, fourth, and fifth
abdominal segments are simple and rudimentary. The merus of the third
maxilliped is short, its antero-internal margin three-toothed ; the palpus of
this appendage is nearly as long as the merus and ischium combined.

Length, 12 mm, ; carapace, 6 mm. ; breadth, 4 mm.

Station 3354. 322 fathoms. 1 male.

This species nearly resembles M. polita (Anoplonotus politus Smith), but
the carapace of the former is longer and narrower, the rostrum is curved more
strongly downward, and the propodi of the ambulatory limbs are much shorter
in proportion to the dactyli.

BULLETIN OF THE

Uroptychus nitidus occidentalis, subsp. nov.

Differs from the typical Uroptychus nitidus (A. M. Edw.)! as follows: the
branchial regions are more swollen, giving to the posterior half of the carapace
a more convex lateral outline; the rostrum is shorter, the chelipeds shorter
and more robust, the fingers shorter in proportion to the length of the basal
part of the propodus ; the branchial regions are more distinctly margined.
It approaches in some respects U. uncifer (A. M. Edw.), in which the rostrum
and chelipeds are still shorter. In some specimens of occidentalis there are a
few low tubercles on the inner side of the proximal end of the merus and
ischium of the chelipeds, — a condition similar to that in U. australis Hend.
which may be considered a variety of U. nitidus. U. politus Hend., another
closely related form, is distinguished by its short antennal acicle.

Length of body of a female, 29 mm. ; length of carapace, 15 mm. ; length
of rostrum, 4 mm. ; breadth of carapace between antero-lateral spines, 5 mm. ;
breadth across the branchial region, 10 mm. ; length of cheliped, 44.5 mm. ;
merus, 11 mm. ; carpus, 12.5 mm. ; chela, 17.5 mm. ; dactylus, 6 mm.

Station 3384. 458 fathoms. 2 males, 2 females ovig.

Uroptychus pubescens, sp. nov.

Carapace, without including rostrum, broader than long, pubescent; a
transverse row of spines across the gastric region from one side of the carapace
to the other; lateral border of carapace spinulose ; the anterior margin has a
deep concavity above the eye, outer angle of the concavity armed with a spine.
Rostrum one half as long as the rest of the carapace, bent downward a little,
acute, with entire setiferous margins. Eye small, not broader than the eye-
stalk, with brown pigment, Abdomen naked, smooth, pleure subacute,
Antenne equal in length to the carapace with the rostrum ; acicle shorter than
the peduncle. Chelipeds long, all the joints as far as the fingers spinulose, the
spinules with broad bases; propodus not broader than the carpus ; carpus
equal in length to the basal portion of the propodus ; fingers straight, a slight
tooth near the base of the dactylus ; the tips of the fingers cross. Meri of am-
bulatory legs minutely spinulose on the superior margin, distal end of propo-
dus spiniform on the hind margin, whole hind margin of dactylus armed with
spines; all the joints of the ambulatory limbs are furnished with long sete.

Length (female), 44 mm. ; breadth, 17.5 mm. ; length of carapace, 21 mm. ;
length of rostrum, 7.5 mm.; length of cheliped, 57 mm. ; merus, 12 mm. ;
carpus, 15 mm. ; chela, 24 mm.; dactylus, 9.3 mm.

Station 3354. 322 fathoms. 2 females ovig.

o H80DD- Tags ut 1 female ovig.

This species is. more nearly related to U. insignis Hend. than to any other

described snecies.

1 Bull. Mus. Comp. Zoöl., VIII. 62, 1880,

MUSEUM OF COMPARATIVE ZOOLOGY, 193

Uroptychus bellus, sp. nov.

Carapace broad, branchial regions inflated, upper surface naked, smooth,
and polished ; the branchio-cardiac lines meet in the median line of the
carapace ; the anterior margin has a concavity above the eye, forming an orbit
with a spinule at its external angle. There is one spine at the antero-lateral
angle, one on the margin of the hepatic area, and eight on the margin of the
branchial region; the branchial spines decrease in size posteriorly. The
rostrum is long, tapering, acute at the apex, and concave at the base aboye.
The abdomen is smooth, the pleure subacute. The eyestalks are short and
stout, the eye not wider than the peduncle, black. The antenn® are very
slender, shorter than the carapace, the acicle considerably shorter than the
peduncle. Chelipeds very long, naked except for a few sete on the fingers,
polished; the ischium bears a spine on the superior margin and several others
on the lower side ; the merus and carpus are armed with spines arranged in
longitudinal.rows ; there is a row of spines on the upper margin of the pro-
podus (the row is double at the proximal end), another series on the outer face
reaching from the proximal end about half-way to the distal end, and another
still shorter row of more rudimentary spines just outside the latter series;
the fingers are separated by a gap; their prehensile edges are denticulate,
with one or more prominent teeth near the base of the dactylus. Ambulatory
appendages : meri and carpi of the first and second pairs spinulose along the
upper edge, these joints being spineless on the third pair. All of the ambu-
latory appendages are subchelate, the distal end of the propodus being enlarged
and furnished with spines against which the spined dactylus closes.

Length (male), 17 mm. ; carapace, 10.5 mm.; rostrum, 4.5 mm. ; breadth of
carapace, 7.7 mm. ; cheliped, 31 mm.; merus, 7 mm. ; carpus, 9 mm.; chela,
14 mm. ; dactylus, 5 mm.

Station 3354. 322 fathoms. 1 female ovig.

UD A 1 male.

Family AXIIDM.

Axius crista-galli, sp. nov.

Near Axius acutifrons (Eiconaxius acutifrons Bate), but differs in the fol-
lowing regards. The margin of the rostrum is armed with prominent teeth,
The median carina of the rostrum, entire or at most but slightly serrate in the
former species, is here cut into about seven prominent teeth. The larger claw
differs from the corresponding organ in A. acutifrons in lacking the serration
on the superior margin of the propodus, in the presence of a strong tubercle
on the anterior border of the hand between the bases of the fingers, and in the
absence of prominent teeth on the prehensile edges of the fingers.

Length, 24.5 mm.; length of carapace, 10 mm.

Station 3359. 465 fathoms. 3 males, 1 female.

The female carries eighteen eggs of large size (2 X 1.5 mm.).

BULLETIN OF THE

Family CALOCARIDZE.

CALASTACUS, gen. nov.

Abdomen long, enlarged in the middle, narrowed at each extremity, pleure
broad and rounded. Cephalothorax laterally compressed, rostrum long, acute ;
eyes rudimentary, subglobose, unpigmented, unfaceted. Second antenna on
a horizontal line with the first antenna ; the second segment armed with a
long external spine (stylocerite) and a still longer articulated style-shaped
scale (scaphocerite), Third maxilliped pediform. First and second pairs of
legs chelate. First abdominal appendages of male modified to serve as sexual
organs (gonopods). Outer branch of swimmerets divided near the posterior
margin by a diagonal suture. Telson long, quadrangular. Gills composed of
a central stem which bears two rows of filaments. Branchial formula :

Somite QV sb B BR o7 AA PE A SED Sonal S

¡Epipoda Ase bib wal 1 1 1 1 1 0, mm x0
Podobranchie . . 0 1 1 1 1 0 0 = 4
Arthrobranchie . 0 2 2 2 2 2 0 ess 10
Pleurobranchie . 0 0 0 0 0 0 0 = 0

Differs from Calocaris in having a long styloid scaphocerite appended to the
peduncle of the external antennæ.

Calastacus stilirostris, sp. nov.

Carapace naked, punctate ; apex of rostrum turned a little upward; two
strong spines turned upward and forward at base of rostrum ; a light median
carina runs along the back from the base of the rostrum, fading out before
reaching the hind border of the carapace. Chelipeds long, symmetrical on the
two sides ; coxa furnished with a small spine on the anterior border of the dis-
tal end ; ischium armed with from one to four spines on the lower margin ;
merus laterally compressed, armed with a spine on the upper edge near the
distal end and a variable number (four to eight) of spines on the lower edge.
Carpus triangular, unarmed. Chela: upper and lower margins sharp, the
upper armed with five to seven spines, inner and outer faces with a few scat-
tered spinules; fingers with denticulate prehensile edges and curved crossed
extremities. Second pair of legs furnished with small chele.

Length of carapace, 22.2 ium.; rostrum, 5.5 mm.; abdomen, 30 mm. ;
cheliped, 39 mm. ; chela, 17 mm.
Station 3418. 660 fathoms. 9 males,

MUSEUM OF COMPARATIVE ZOOLOGY. 195

Family ASTACIDZE.

Nephropsis occidentalis, sp. nov.

Pubescent. Carapace cylindrical, the branchial regions convex. Rostrum
densely ciliated on the margins, armed with a pair of lateral teeth near the mid-
dle; a double row of prominent granulations on the dorsal surface, diverging
posteriorly and continued backward for some distance on the gastric region.
A small, blunt papilla in the median line of the gastric area, a pair of acute
teeth near the anterior margin at the base of the rostrum, and another pair just
above the insertion of the second pair of antenne; a small papilla in the
median line on the intestinal region. Abdominal pleure rather longer-pointed
than in N. stewarti, but not so much so as in N, agassizii and N, atlantica;
their anterior borders are finely denticulated, but are destitute of spinous pro-
cesses, Telson armed with a sharp spine in the median dorsal line, near the
proximal end.

Length, 119 mm.; carapace, 51 mm.; rostrum, 14 mm. ; second antenna,
225 mm.

Station 3418. 660 fathoms. 23 males, 32 females.

en DARAN Sd Qi. i 2 males.

Family ERYONTID 2.

Willemoesia inornata, sp. nov.

Similar to W. leptodactyla, but readily distinguished from it by the small
number of spines on the margin and dorsal ridges of the carapace. The
armature may be formulated thus : —

Marginal... 65 Se. ov D 408 —2t0 3/0 to 6
Medium age. ^76 wo UD À

The marginal spines which lie behind the cervical groove, if found at all, are
but rudimentary, while there are no spines on the submarginal carina or along
the lateral boundaries of the cardiac area, where they are present in W. lepto-
dactyla. The third maxilliped bears only a slender epipod ; the membrane
that conneets this limb with the body carries a small, but perfeetly formed
gill (arthrobranehia) According to Spence Bate, this gill is absent in
W. leptodactyla.

Station 3374. 1823 fathoms. 8 males, 6 females.

o ane + 1 male, 1 female ovig.

i SESS alba 2 males, 4 females (1 ovig.).
16.598899; O t€ 2 males, 1 female.

* — 3400. 1322 ‘i l female.

VOL, XXIV. — NO. 7, 4

196 BULLETIN OF THE

Polycheles! tanneri, sp. nov.

Orbital sinus rounded at the bottom, outer margin spinulose. Median carina
of carapace furnished with two anterior rostral spines, followed by five (or six)
spines in front of the cervical groove, the fourth (or fifth) of which is double.
The arrangement of these spines may be thus formulated: 2. 1. 1. 1. 1. 2. 1 (or,
2.1.1. 1, 2, 1). Back of the cervical groove the spines of the median carina are
2.2.2. Marginal spines of carapace arranged as follows : 5—3—13 or 14).
A longitudinal row of four small spinules on the anterior division of the
carapace midway between the median and marginal rows, and a row of twelve
or fifteen on the branchial regions inside the margin of the carapace. "There
are, besides, two or three spines on each side of the hind margin of the cara-
pace, and a few along the cervical groove.

Resembles P. nanus (Smith), but differs in the number of spines on the
median and sublateral carinee of the carapace, in the existence of a spine on
the antero-external angle of the first and second abdominal pleur®, and in the
greater number of spines on the merus, carpus, and propodus of the chelipeds.
P. nanus, moreover, is described as having the posterior pair of thoracic ap-
pendages chelate in the male, while in the males of P. tanneri that I have
examined these appendages are simple. Compared with the types of P. agassizii
(A. M. Edw.), the carapace of the present species is broader and. fewer-spined
on the margins; the first and second abdominal pleure are armed with an
anterior lateral spine; the rostral spine is double; and the orbital sinus is
broad and rounded at the bottom. P. agassizii, like P. tanneri, has non-chelate
posterior legs in the male.

Station 3354. 322 fathoms. 1 male.

a er A EE 2 males, 1 female.
ns 3403. 884 * 12 males, 14 females,
cor BAO BET 56 * 1 female,

Polycheles sculptus pacificus, subsp. nov.

Differs from the Atlantic P. sculptus Smith as follows. The carapace is
broader in proportion to the breadth of the abdomen, the lateral margins con-
verging strongly at the posterior end, where, in P. sculptus, they continue
nearly parallel to one another; there is a small spine on each branchial region
inside of and on a level with the second spine of the submarginal carina, which
spine is entirely wanting in P. sculptus; the spine on the anterior border of
the ophthalmic lobe is larger and blunter; the pleure of the second abdominal
somite have a different shape, their anterior margins being in line with the an-
terior margin of the tergum, whereas in P. sculptus they form a strong obtuse
angle with that margin.

1 The genus Polycheles, as here defined, comprehends Polycheles, Pentacheles, and
Stereomastis of Bate.

MUSEUM OF COMPARATIVE ZOOLOGY. 197

These differences, although slight, are constant, and should be recognized in
our nomenclature, if any significance is attached to geographical variation,

The last thoracic appendages are chelate in the adult female, while they are
but imperfectly so in breeding males; that is, in the male the “thumb” is
very much shorter than the index.

Station 3353. 695 fathoms. 1 male.

“8892. 1270 z 1 female.

u 8393. 1020 U 3 males, 3 females.
* .— 88904. 511 e 12 males, 20 females,
54 OH Bie „Hoss " 1 male, 1 female.
1.010, okie k 1 female.

* . 9494, 676 p 1 female ovig.

Polycheles granulatus, sp. nov.

Carapace long oval, broadest across the anterior branchial region; dorsal
surface granulated, but nearly devoid of spines; there are two small rostral
spines, and back of these, on the low granulated median carina, lies another
pair followed by one or two spinules on the gastric area, The submarginal
ridge is incurved and composed of minute spinulose granules. Orbital notch
narrow, armed with a spine at its internal angle and with another at its ex-
ternal angle. Marginal spines thus disposed : 9 (or 10) — 3 — 15. Thean-
terior abdominal pleure are rounded, gradually becoming acute as one passes
backward to the sixth. The posterior thoracic legs in the sole specimen seen
(a female) end in a small but perfect chela,

Length, 99.5 mm.; length of carapace, 45 mm. ; greatest width of carapace»
37 mm.; length of cheliped, 118 mm. ; ischium, 19 mm. ; merus, 33 mm.;
carpus, 22 mm. ; basal part of propodus, 14.5 mm.; dactylus, 22 mm.

Station 3380, 899 fathoms, 1 female.

Eryonicus cescus Barr?

Station 3375, 1201 fathoms. 1 male, 62.5 mm. long.

oF oe AHO is e 1 female, 40 mm, long.

ROO Oo l juv 37 mm. long.

“ 3388. Surface to 400 fathoms, submarine tow-net. 4 juv., 19-29 mm.
long.

Bate’s description of Æ. cecus was drawn up from a single immature speci-
men, 13 mm. long, in which the first abdominal appendages were undeveloped,
The largest of the “ Albatross” specimens is asexually mature male with well
developed gonopods. It differs from Bate’s specimen in having much shorter
spines upon the carapace and abdomen ; the spines of the lowest series on the
branchial region decrease in length posteriorly, while in the type speeimen the
posterior spines in this row are the longest. Whether these discrepancies are

198 BULLETIN OF THE

due to difference in age, or whether they denote specific diversity, cannot be
determined until more mature specimens are obtained from the Atlantic, In
the smallest of the “ Albatross” specimens (which have attained a length of
19 mm.) the spines, especially those of the abdomen, are relatively longer than
in the adult, though not so long as in the type described by Bate.

As regards the ophthalmic sinuses and lobes, the genus Eryonicus is like
Polycheles (Pentacheles).

Eryonicus spinulosus, sp. nov.

In this species the spines of the carapace, instead of being wellnigh limited
to nine longitudinal ridges, as in E. cecus, are thickly strewn over the whole
surface. The intervals between. the spines give rise to slender hair-like sete.
This is the arrangement of the spines of the median carina of the carapace :

2: LL 2 LL — 28) Le By
The sublateral carina bears fourteen small spines, the lateral,
5 — 2 —13 or 14,

The uppermost of the two carino below the lateral is denticulated anteriorly
and armed with a spine at the front end behind the second antenna. The low-
est ridge carries twelve spines, which increase slightly in length posteriorly.
The rostral spines are very small, but on each side of the rostrum the front
margin of the carapace is produced so as to form a pair of horns over the base
of the first pair of antenne. The abdomen is ornamented with seven longi-
tudinal rows of spines, one dorsal and median, the others paired and lateral,
The lowest of the lateral rows is on the upper part of the pleura,

Length, 37 mm. ; carapace, 21 X 17 mm.; abdomen, 17 mm.

Station 3403. 384 fathoms. i specimen.

Family GNATHOPHYLLID&.

Gnathophyllum panamense, sp. nov.

Closely related to G. elegans of the Mediterranean Sea, but distinguished by
a prominent conical protuberance, pigmented with black, on the upper part
of the cornea. In G. elegans this tubercle is wanting or reduced to the merest
vestige, discernible only by aid of a lens. The rostrum of G. panamense is
furnished with seven teeth above, and one below. Color entirely different
from that of G. elegans. The ground tint is dark brown, ornamented with a
multitude of light blue spots, amongst which are sixteen red spots. Rostrum,
eyes, and antenne whitish, flagellum of second antenna orange. The fifth and

sixth abdominal segments, the telson, and the swimmerets are also white. Basal
joints of second pair of chelipeds violet, the merus, carpus, and fingers whitish,
basal part of propodus orange. Third, fourth, and fifth pairs of legs violet.
Panama, March 12. 1 female ovig.

G. fasciolatum Stimps., from Australia and Amboyna, agrees closely in

MUSEUM OF COMPARATIVE ZOOLOGY. 199

form with G. elegans, but differs wholly in the pattern of its color marks
from both G. elegans and G. panamense. G. zebra Richters, from Mauritius,
is without much doubt the same as G. fasciolatum. Ortmann has lately re-
corded a Gnathophyllum from Tahiti as a new species, G. pallidum. It differs
from @. fasciolatum only in the absence of color marks, —a difference due
possibly to the action of alcohol.

Family CRANGONIDZE.

Sclerocrangon atrox, sp. nov.

Of the described species of Selerocrangon, S. ferox G. O. Sars comes nearest
to this species. These are some of the chief points of difference. In S. ferox
the upturned rostrum is simple, while in S. atrox a long acute tooth is given
off from its ventral side, a tooth which reaches as far forward as the tip of the
rostrum. In the former species the dorsal carinæ of the sixth abdominal seg-
‘ment bear two pairs of well developed spines, while in the latter species one
finds but one pair of very small spines at the posterior end of the carving, The
pleural spines of the abdomen are much longer in the former than in the lat-
ter, and on the fifth somite there are four to five spines on each pleura against
two in S. atrox. The eyes, are much smaller in S. feroz, and lack the spine
above the cornea seen in S. atroz.

Length of the largest specimen (a female), 162 mm.

Station 3418, 660 fathoms. 3 males, 2 females (1 ovig.).

* 3424, 676 “ 4 females,

Sclerocrangon procax, sp. nov.

Nearly related to S. agassizii Smith, from the Atlantic side of the conti-
nent. The rostrum of S. procax is longer than in S. agassizü, and inclined
upward at a much sharper angle ; the same is true of the antero-lateral spines
of the carapace. The most conspicuous difference is found in the anterior spine
of the median carina of the carapace, which is much longer and nearly erect in
S. procax. The two flagella of the first antenna in the male, S. procaz, are
subequal, while in the male 5. agassizii the outer flagellum is much longer
than the inner; the scale of the second antenna, moreover, is narrower in the
former species than in the latter, and the terminal segment of the inner branch
of the second abdominal appendage in the male bears on its inner margin a
short blunt stylamblys, which is wanting in S. agassizii. In neither of these
species is this seement produced into a lobe at the base of its outer margin, as
it is in the more typical species of Sclerocrangon, e. g. S. ferox and S. atrox.

Length, 49 mm.

Station 3380. 899 fathoms, 1 male juv.

ey p; MOSM tt 2 males, 3 females.
OO, 300. S 2 females ovig.

E 23400, 909 " 1 female,

BULLETIN OF THE

Pontophilus occidentalis, sp. nov.

Allied to P. abyssi Smith and P. batei. From the former it differs in
having a shorter rostrum, larger eyes, and more strongly developed caring: on
the carapace. From the latter it also differs in its shorter rostrum armed with
two pairs of lateral teeth; in the presence of a sharp spine on the sternum be-
tween the second pair of legs; in the length of the antennal scale, which in
P. occidentalis is equal to the distance from the tip of the rostrum to the car-
diac spine; in the shortness of the second pair of legs, which reach only half
way to the distal end of the merus of the first pair; and in the absence of the
spine on the outer margin of the merus of the first pair of legs.

The eyes are as large as in P. gracilis Smith(much exceeding the rostrum),
but they are nearly colorless, and unfaceted, as in P. abyssi and P. batei,

Length, 73 mm. ; carapace, 21 mm.

Station 3361. 1471 fathoms. 2 specimens.

O O m 2 A

dl 10 Tr il 2:
aboot Lila” em 4 We
ee ee 100 5
CT MBO 2/0. ie 1 Ki
tee a FOO Y 2 "m
wc HUS. "19007 E 4 d
uw OAL, 2202 5c 1 ae
PS a Lila 2 Hi

Paracrangon areolata, sp. nov.

Rostrum long, acute, upturned, inferior margin armed with two spines, A
prominent carina extends the length of the carapace in the median dorsal line;
it is armed with four spines, three on the gastric, one (obsolescent) on the
cardiac region. Orbit incomplete, bounded externally by a slender spine. Just
below the base of the second antenna the antero-lateral angle of the carapace is
drawn out into another rather stronger spine. Just behind this, and from a
little higher level, a strong, sharp horn is directed outward and forward ; this
horn is in continuity with a rounded ridge which runs inward to the external
orbital spine. A longitudinal carina on each side of the gastric region, armed
with a small spine a little way behind the middle; from this spine another
ridge runs upward and inward, meeting the median carina at the base of the
third spine. The branchial regions are traversed by a series of ridges which

1 Pontophilus gracilis Bate, Rep. Challenger Macrura, p. 487, 1888. This name
having been previously employed for another species by Smith (Bull. Mus, Comp.
Zoöl., X. 36, 1882), I have substituted the name Pontophilus batei for P. gracilis of
Bate.

MUSEUM OF COMPARATIVE ZOOLOGY. 201

anastomose in such a way as to divide these regions into cells of different sizes ;
they are armed with three small spines, the anterior of which is the spina
hepatica.

The thoracic sterna are armed with two median spines, one of which is situ-
ated on the somite which normally bears the second pair of legs (absent in this
genus), the other on the somite behind this. Abdominal pleure acute.

Chelipeds of moderate length, with a spine on. each side of the distal end of
the carpus, and another long and acute one at-the antero-internal angle of the
propodus.

Length, 85 mm.

Station 3424, 676 fathoms. 2 males, 3 females (1 ovig.).

* $8426. 680 “ 1 male.

Family GLYPHOCRANGONIDZ.

Glyphocrangon alata, sp. nov.

Rostrum armed with a pair of lateral spines on a level with the anterior end
of the eye ; posterior to this pair of spines follows a variable number of smaller
marginal spines (three to five on each side). "The anterior half of the rostrum
is unarmed, although ciliated on the margin. A light spinulose carina runs
along the median line of the rostrum from the base to the anterior third (in
some specimens this carina is obsolete). On each side of the median line there
are, on the basal part of the rostrum, four or five small spinules.

Carapace and abdomen thickly tuberculated. On the antero-lateral regions
of the carapace, and on the abdominal pleura, the tubercles assume a spiny
character. For the rest, the tubercles are mostly compressed, their tops trun-
cate and more or less eroded. It is further to be noted that these tubercles are
arranged in longitudinal rows, and that six of these rows on each side of the
carapace form, by their prominence, imperfect carinæ, corresponding in position
to those commonly found in species of this genus. Between the external or-
bital spine and the spine at the antero-lateral angle of the carapace lies a
strong spinous tooth, acute at the end and vertically compressed, its base
broadening out in such a fashion that the whole tooth forms an acute, wing-
like expansion. This is, in fact, the greatly developed anterior part of the
fourth carina (counting from the median dorsal line). Behind it the carina
continues as a low toothed ridge. A median interrupted dorsal carina runs
along the abdomen, broken up into two teeth on the base of the telson.

Length, 116 mm.

Station ? 12-+ specimens.

“3895. 730 fathoms. 1 &
tis 8418066054“ 2 3

202 BULLETIN OF THE

Glyphocrangon spinulosa, sp. nov.

Rostrum long, acute, margins armed with vertically flattened spinous teeth
from base to the level of the anterior extremity of the eyes, beyond which
point the margins are unarmed ; the anterior pair of the lateral spines are the
largest. A median longitudinal row of smaller spines extends from the ante-
rior boundary of the gastric area to the anterior fourth of the rostrum ; anteri-
orly, these spines are confluent at their bases, forming a carina which is con-
tinued forward beyond the spines to the tip of the rostrum, Just inside the
marginal spines is an irregular longitudinal row of very small spinules on each
side of the rostrum.

The carapace is thickly covered with spinules which are laterally compressed.
Along six longitudinal lines on each side of the carapace the spines are larger,
more flattened, forming interrupted carino. The third carina (reckoning from
the dorsal line) ends anteriorly at the deep branchio-hepatic sulcus. The
fourth carina is especially prominent on the hepatic region, where it is broken
up into two or three prominent teeth. The two lowermost carine are obso-
lescent, The anterior margin of the carapace is produced into a large external
orbital spine, directed obliquely upward and outward. The hind margin of
this spine is furnished with a variable number of spinules.

The abdomen, like the carapace, is spinulose on the dorsum and pleuree.
An interrupted carina extends the length of the median dorsal line. On the
sixth segment this carina is toothed along its edge. The second to the fourth
abdominal pleure are armed with two spines curved outward and backward,
the anterior one being the larger. The fifth pleura is likewise armed with
two spines, but in this case the posterior spine is the larger. The sixth
pleura ends posteriorly in a single spine directed outward ata greater angle
than those of the more anterior somites. The median dorsal abdominal carina
is continued along the basal part of the telson as a line cf four or five teeth,
which decrease in size from before backward.

Length, 105 mm.

Station 3353. 695 fathoms. 11 specimens.

“ — 8418. 660 A 91 H
ud BES Da E 1 ia
“ 3424. 676 d 19 y
* 3425. 680 be 1 pe
“ — 84895. 859 v 2 E

Glyphocrangon sicarius, sp. nov.

Rostrum as long as the rest of the carapace, armed with a pair of short,
blunt, lateral teeth, a little in advance of the front of the eyes, and with an-
other pair of obsolescent ones at the root. Between these two pairs of teeth
the margins of the rostrum are concave and slightly raised. From the anterior

MUSEUM OF COMPARATIVE ZOOLOGY. 203

teeth the rostrum tapers regularly to the point. The upper surface is plane
and smooth. A slight median carina, most obvious near the tip, runs through
the whole length of the rostrum from the anterior end of the gastric area to the
tip; the lateral margins are slightly rimmed. The lower surface of the ros-
trum is longitudinally grooved, and wholly devoid of a median keel.

Orbital spine rather short and thick, directed a little outward and upward.
The spine at the antero-external angle of the carapace is also short and thick,
slightly exceeding in length the orbital spine. From its base a low keel ex-
tends backward over the hepatic area,—a keel unarmed with spines, but
broken into two tubercles, the hinder of which is the more prominent. In the
small triangular area between this keel and the gastro-hepatic sulcus are several
tubercles disposed for the most part in a single row. The gastric region is
free from tubercles in the median line save one minute one at the anterior
boundary near the base of the rostrum; but on either side, this region is orna-
mented with tubercles ; these are not spiny, nor do they tend to form very well
marked caring. On the branchial regions there are three carine on each side;
the uppermost of these is broken up into five or six tubercles ; the top of these
tubercles, as well as those of the other carine on the branchial and hepatic areas,
exhibit a corroded surface, and a similar appearance is manifest on the whole
lower surface of the carapace where it is bent beneath the thorax. The abdo-
men is irregularly bestrewn with low tubercles. The telson is straight and
much shorter than the rostrum ; it is furnished with a small, laterally com-
pressed tooth in the median dorsal line at base; both the dorsal carin: and the
lateral margins are obsoletely dentate along their basal half.

Length, 124 mm. ; length of rostrum from tip to anterior gastric groove,
30.5 mm. ; length of carapace, including rostrum, 58 mm.

Station 3382. 1793 fathoms, 2 specimens.

Family PANDALIDA.

Heterocarpus vicarius, sp. nov.

This is the Western representative of Æ. gibbosus Bate, from the Philippine
Islands. It differs from that species as follows. The carapace is very much
longer both in proportion to its height and to the length of the rostrum, and its
dorsal margin is not so convex. The teeth on the upper edge of the rostrum
are more numerous and closely set. The lateral carinw of the carapace are
more prominent. The third abdominal somite forms a sharper angle or knee,
Finally, the antennulary flagella are longer and subequal in length.

Length, 110 mm. ; rostrum, 26 mm.; carapace, including rostrum, 55 mm.

Station 3385. 286 fathoms. 49 specimens.

d 3386. 242 “ 91 P
di 9889, 210 T f
d 3396, 259 “ 68 ii

204 BULLETIN OF THE

Heterocarpus hostilis, sp. nov.

Similar to H. alphonsi Bate, from off the Philippine Islands and J apan, but
readily distinguished from that species by the stronger upward curve of the
rostrum, by the presence of but one tooth in the median dorsal line of the
carapace back of the base of the rostrum, and by the two spines on the third
abdominal segment.

Length, 197 mm.; rostrum, 70 mm. ; carapace, including rostrum, 108 mm.

Station 3353. 695 fathoms. 22 specimens.

mi, 20008 IA 24 m
9 InODOd gue. 6 3 Hs
E BUB SAOR Y 20- “
Ar 09801 ag ene 6 e
P OBB. LUQUE TY 8 “
OOO Bl 2 A

Heterocarpus affinis, sp. nov.

Very similar to H. hostilis, from which it differs in the following particulars.
The rostrum is much shorter; measured from the tip to the posterior limit
of the orbit, it about equals the rest of the carapace in length. There are
fewer teeth on its dorsal margin (five or six) ; the number of teeth on its lower
margin varies from six to ten. The median dorsal line of the carapace is more
convex, and it is armed, back of the base of the rostrum, with two acute teeth,
where there is but one tooth in M. hostilis.

It bears a resemblance, further, to M. dorsalis Bate, collected by the “ Chal-
lenger” in the Banda Sea, but it may be at once distinguished from the
Oriental species by its shorter rostrum and the presence of two median teeth
on the third abdominal segment.

H. affinis and H. hostilis belong to different geographical areas. HH. affinis
is the more northern form, found off Acapulco and Cape Corrientes, while
H. hostilis was obtained in the Gulf of Panama.

Station 3418, 660 fathoms. 13 specimens.

" — 8494. 676 " 5 E
E 3425. 680 i 9 d

Famil NEMATOCARCINIDZE.
Nematocarcinus agassizii, sp. nov.

The rostrum is one fifth longer than the rest of the carapace (in some small
specimens only equal to the rest of the carapace), slender, nearly horizontal
for the basal two fifths of its length, the remaining portion gently upturned and
ending in a very acute point; its upper margin is continued backward in the

MUSEUM OF COMPARATIVE ZOOLOGY. 205

form of a carina, which becomes obsolete on the hinder part of the gastric
reg.on; this carina is pectinate, or armed with close-set, forward-pointing
teeth on the anterior part of the gastric region; the teeth are continued on
the upper margin of the rostrum through one third or two fifths of its length,
the distal third or three fifths of the rostrum being entirely free trom teeth
above; the lower edge of the rostrum is ciliated above the eyes and armed
with three (rarely four) teeth, separated by wide intervals, on the distal half.

The third abdominal segment is somewhat prolonged posteriorly over the
next segment, but the hind margin is rounded off and does not form a promi-
nent tooth. The telson is tipped with three pairs of spines, the intermediate
pair the longest ; there are, besides, about six pairs of small spines on the dor-
sal side of the telson.

The flagella of both the antennules and antennx are prodigiously developed,
the antennule being nearly twice, the antenna more than twice, the length of
the whole body including the rostrum; excepting the proximal part of the
organ, the annuli of the antennary flagellum are enlarged at the distal end,
giving a beaded appearance to the flagellum. The antennal scale reaches rather
more than half-way to the end of the rostrum ; it is truncate at the distal end
and armed externally with an apical spine.

The legs have the characteristic shape and proportions of the genus, The
dactyli of the third and fourth pair are slender, acute, somewhat curved, and
invested by a pencil of long hairs. The dactyli of the fifth pair are stouter
but very short, and hidden in the tuft of hairs which arises from the distal end
of tlie propodus.

The outer blade of the swimmeret slightly surpasses the telson. It is fringed.
with long hairs along the internal and distal margin, and furnished with a
minute tooth and a movably articulated spine on the external border near the
distal end. The inner blade is a trifle shorter than the telson; it is also fringed
along its whole margin with long hairs.

Dimensions of a female specimen. Length of body, rostrum included,
139 mm. ; length of rostrum, 35 mm. ; length of carapace, rostrum included,
61 mm. ; length of telson, 20 mm. ; length of antennule, 267 mm. ; length of
antenna, 315 mm.; length of antennal scale, 19 mm.

This species, like all the Nematocarcint, is very fragile. The long and slender
rostrum is often broken off during life, and the attempt to restore it sometimes
results in an abnormally small and otherwise monstrous rostrum, which might
easily be mistaken for a specific character if ample material were not at hand,

Station 3353. 695 fathoms. 2 specimens,

“ 83354, 392 =“ 4 b

uo Uae O00 T 90 ik

O 2 Oey
"ww Sob OUR Mi
Bart, > EMO
OOO, - QUE
* 8384. 458 *"

— 10a»

206

Station 3393.

* — 8895.
x: 34006.
d: 3407,
4 3418.
“ 2

1020 fathoms.

730

551

885

660
2

113

66

e

[1

BULLETIN OF THE

“

114

8 specimens.

(1 bopyrized)

Family MIERSIIDZE.

Acanthephyra cristata, sp. nov.

Differs from A. debilis A. M. Edw. (= A. gracilis Smith) in having a much
shorter and fewer-spined rostrum, two pairs of longitudinal lateral carinæ on
the carapace, a dorsal carina on the fourth abdominal segment, and by the ab-
sence of a series of denticles on the posterior margin of the dorsum of the fourth

and fifth abdominal segments.

From A. lanceolata (Systellaspis lanceolata Bate) it also differs by having the
rostrum shorter, the lateral and dorsal caring of the carapace more prominent,
and by the absence of a prominent tooth on the anterior margin of the first
abdominal segment.

This species, like A. debilis and A. lanceolata, has no carina on the dorsal
surface of the fifth and sixth abdominal segments.

Length, 78 mm.; rostrum, 13 mm.; carapace, including rostrum, 27.5 mm.

Station 3361.
e ne Boole

1471 fathoms.

1772

e

1

e

1 specimen.

Acanthephyra cucullata, sp. nov.

The integument is soft, membranaceous, and transparent in alcohol. The
carapace is carinated in the median dorsal line anteriorly ; this carina is fur-
nished with seven minute teeth, and is continued forward to a very small
acicular rostrum, which hardly reaches forward to the end of the eyes; the
infero-lateral margins of the orbit are continued downward for some distance
nearly parallel with one another in a nearly vertical direction, and then sud-

denly diverge and trend backward, forming the upper wall of the orbit; a sort
of hood is thus formed of the anterior part of the carapace, overhanging the
facial region. The infra-orbital angle is rounded, not spiniferous. The an-

tennal spine is acute, and advanced forward of the infra-orbital angle. The
branchiostegian spine is small and continuous with a longitudinal carina that

runs along the branchial region of the carapace. A low fold or ridge marks

the upper boundary of the branchial region,

The abdomen is carinated in the median dorsal line on the second to the

sixth segment inclusive ; the carina is most prominent on the third segment,

MUSEUM OF COMPARATIVE ZOOLOGY. 207

where it is produced into a strong posterior tooth which overhangs the anterior
part of the fourth segment; the three following segments are furnished with
minute posterior teeth, The posterior half of the telson in the unique speci-
men at hand is missing; there is one pair of minute marginal spines at the
hind end of the remaining proximal half. The eyes and eyestalks are well de-
veloped, the stalks broadening toward the distal end, and projecting a slender
blunt process on the inner side close to the cornea; the eye itself is as broad
as the distal end of the peduncle.

The basal segment of the antenna is armed with an acute external spine ;
the antennal scale is long, gradually narrowing distally to the apex, which is
furnished with a small spine.

The thoracic appendages have the form characteristic of the genus Acanthe-
phyra, and appear to offer no important specific characters.

Length, 87 mm.; carapace, 27 mm.; antennal scale, 16 mm.

Station 3381. 1772 fathoms. 1 male.

Notostomus fragilis, sp. nov.

Dorsal line of carapace convex and keeled from anterior to posterior margin,
anteriorly produced into a short, acute rostrum, which does not exceed the eye-
stalks in length ; the dorsal carina is armed with seven or eight minute teeth
on the anterior gastric region and the basal portion of the rostrum ; lower
margin of rostrum unarmed. A longitudinal carina on each side of the cara-
pace begins near the orbit, above the infra-orbital spine, and runs back to the
posterior margin; another carina runs obliquely downward and backward,
dividing the branchial from the hepatic regions. The inferior lateral carina is
obsolete except for a short distance behind the spine which lies near the ante-
rior margin of the carapace behind the base of the second antenna.

The abdomen is strongly compressed, the third, fourth, fifth, and sixth seg-
ments dorsally carinated ; the carina terminates in a small tooth at the hind
end of the fourth, fifth, and sixth segments ; on the fourth segment the carina
is divided into two parts by a deep notch about two thirds of the distance from
the anterior to the posterior margin of the segment. The telson is channelled
on the dorsal side, and is tipped with two long spinous sete.

The eyestalks taper from the base to the tip; their outer and upper mar-
gins are nearly straight, but their inner and lower surfaces are swollen ; on the
inner side of each stalk, a little way behind the eye, there is a blunt tubercle.
The eye itself is small and black. The integument of the eyestalk is trans-
parent, and when held to the light discloses the optic ganglion within, giving
off a nerve to the retina and another to the tubercle on the inner side of the
stalk.

The basal segment of the antennule is armed with a very small but sharp
external spine. The outer side of the second antenna, on the contrary, is
unarmed. The antennal scale is very broad, oval, and furnished with a small

208 BULLETIN OF THE

spine on the external border, near the distal end. The third maxillipeds are
robust ; they reach forward far beyond the end of the antennal scales; their
terminal segment is triangular in cross section, The first pair of legs are also
robust, about equal in length to the third maxillipeds, and their chela is strong,
with fingers about equal to the hand in length. The second pair of legs are
longer but much weaker than the first pair, the carpus and propodus much
elongated, and the fingers not more than a third as long asthe hand. The
ischium and merus of both the first and the second pair are flattened, and the
same compression is seen in the three following pairs of legs, which have
the form and proportions characteristic of the genus. The inner branches of
the swimmerets are about the length of the telson, while the outer branches
are rather longer.

Length, 70 mm.; cdrapace, 30 mm.; telson, 14 mm.; antennal scale, 11 mm.

Station 3371. 770 fathoms. 1 specimen,

Tn this specimen the integument is soft and membranaceous, and the cara-
pace is so collapsed that it is difficult to restore its true outline. In several
respects this species shows an approach to the genus Hymenodora, e. g. the
soit integument, small eye, and the reductibn of the rostrum.

Notostomus westergreni, sp. nov.

Similar to N. patentissimus Bate, with which it agrees in nearly all the de-
tails of carine, etc., but it differs much from Bate’s species in its general form
and proportions, the carapace being longer in proportion to its height, and less
convex along the dorsal line than it is in N. patentissimus. The rostrum, be-
sides, is armed with many more spines (at least twelve) on its inferior margin.
Bate says that in N. patentissimus the antennal carina terminates in the pos-
terior margin of the carapace, where it is confluent with the lowermost, sub-
marginal carina. This is not the case in N. westergreni.

Length, 127 mm.; length of carapace, including rostrum, 61 mm.; height of
carapace, 31 mm.

Station 3399. 1740 fathoms. 1 male.

Family PASIPHAEIIDZ.
Pasiphaeia cristata americana, subsp. nov.

This form is closely allied to P. cristata Bate, dredged by the “ Challenger”
in 315 fathoms, near the Fiji Islands. The ** Albatross” specimens differ in
some respects from the Fiji specimen. "They may be considered a geographi-
cal race of the same species.

On comparing the two forms it appears that the carapace of the “ Albatross”
specimens is considerably longer in proportion to the length of the whole body

MUSEUM OF COMPARATIVE ZOOLOGY. 209
N

than it is in the Fiji form, and that the dorsal crest is smaller and somewhat
differently shaped. The fifth and sixth abdominal segments are much shorter
in proportion to the length of the telson in the former. The antennulary pe-
dunele is only one third as long as the carapace, while in the typical P. cristata
it is about one half as long as the carapace. Furthermore, the flagellum of the
antenna in the former is longer than the body, whereas in the latter it is
described as being only one half as long.

Length 65 mm.; carapace, 22 mm.; telson, 7.5 mm.

Station 3385. 286 fathoms. 4 specimens,

0896: 260 yk 2 i
** 3403. 384 3 2 5
“ 3406. 551 2 3 A

Pasiphaeia magna, sp. nov.

This species is second only to P. princeps Smith in dimensions. It is most
similar to P. tarda Kroyer, from the North Atlantic. Differs from the latter
species by having the dorsal line of the carapace more convex, the posterior
part of the carapace higher in proportion to the anterior part, the dorsal keel of
the carapace rounded except on the anterior gastric region, and the anterior
rostrum-like tooth’ longer and differently shaped. The proportional length of
the segments of the third pair of legs appears to be quite different in the two
species : in P. tarda, according to Kroyer, the propodus and dactylus are of
equal length and four times as long as the carpus, while in P. magna the
dactylus is even shorter than the carpus, which is itself only one sixth as
long as the propodus. In P. tarda, following Kroyer, the lower margin of
the second segment of the second pair of legs is furnished with three spines ;
in P. magna this margin is unarmed but for the tooth at its distal end.

Length, 145 mm. ; carapace, including the anterior dorsal tooth, 55 mm.

Station 3384, 458 fathoms. 1 specimen.

Family PENEIDAS.

Sicyonia affinis, sp. nov.

This species is the Pacific coast representative of Sicyonia edwardsii! of the
Atlantic coast of America. It agrees with S. edwardsii in the form and denti-
tion of the carapace, but differs in the form and sculpture of the abdominal
segments. In S. edwardsii the first four abdominal segments are rugose, and
ornamented on the sides with two deeply impressed transverse lines, followed
by another less deeply incised near the hind margin of each segment; more-
over the pleurw of these segments are angulated below. In S. affinis these

1 Sicyonia edwardsii Miers, Ann. Mag. Nat. Hist., 5th series, VIII. 367, 1881 ;
Sicyonia carinata (Olivier) Milne Edwards, Ann. Sci. Nat., XIX. 344-346, Plate IX.
Fig. 9, 1830; nec Sicyonia carinata (Olivi).

210 BULLETIN OF THE

segments are smooth, the impressed lines are absent with the exception of the
hinder one of the deeply cut pair, and the pleure are broadly rounded below ;
the peduncle of the second antenna is not much over one half the length of
the antennal scale, and the whole appendage is but little longer than the
carapace.

Length, 62 mm. ; carapace, 20 mm.

Station 3367. 100 fathoms. 1 male, 1 female.

f^ 598300) 02 j 1 male,
ARTS LUZ ES 2 females.
“> 78810 “DY xi 1 female,

The color in life, as shown in a sketch made by Mr. Westergren, is light
greenish yellow, banded with vermilion on the branchial regions and abdomen.
Appendages red, antennary flagellum transversely banded with light and dark.
The coloration is quite different from that of S. edwardsii as given by Dana
(Crust. U. S, Explor. Exped., p. 602).

Sicyonia picta, sp. nov.

Rostrum a little shorter than the eyestalks, laterally compressed, elevated,
armed with six teeth, three superior, three terminal; inferior margin ciliated.
Carapace carinate in the median dorsal line; the carina is furnished with two
teeth, one minute, on the anterior part of the gastric region at the base of the
rostrum, the other larger, over the cardiac region. An infra-orbital and an
hepatic spine are present. General surface of carapace punctate and sparsely
clothed with sete. Abdominal segments smooth, with a median dorsal carina
which rises into a strong tooth on the first segment directed upward and
forward ; on the fifth and sixth segments the carina is drawn out into an acute
tooth which is directed horizontally backward. The pleura of the abdominal
segments are margined and armed with a tooth on the lower border; on the
second, third, and fourth segments the tooth is hamate, being directed out-
ward and backward. The general surface of the abdomen is smooth, with a
transverse groove on each side of the segments. The telson is channelled
above, acute at the tip, and armed with a pair of small lateral spines near the
distal end. The eyes are very large, horizontally flattened. Basal segment of
antennule armed with two spines on its exterior border; flagella shorter than
the peduncle. The peduncle of the second antenna reaches about two thirds
of the way to the end of the scale; the flagellum is about equal to the abdo-
men in length ; the basal segment of the peduncle is furnished with a long and
acute external spine. The sternum is armed with a long spine between the
bases of the second, third, and fourth pairs of legs, as is usual in species of this
genus. The last pair of abdominal appendages are a little shorter than the
telson.

Length, 70 mm.; carapace, 24 mm.

Station 3

06:0 g8By ELOY M 6 males, 4 females,

355. 182 fathoms. 1 male.

MUSEUM OF COMPARATIVE ZOOLOGY. 211

On the hinder part of each branchial region there is a dark (in alcohol) ring
of pigment, S. ocellata Stimpson and S. penicillata Lockington are similarly
ornamented, The flagellum of the antenna is banded alternately with light
and dark color, and there are traces of color on the margins of the rostrum, the
dorsal carine, and appendages.

Peneus balbose, sp. nov.

Integument thin and membranaceous, its surface thickly beset with minute
squamiform tubercles. The rostrum of the sole specimen procured is broken
off a little short of the anterior end of the eye; on the upper margin of the
part remaining, and on the median line of the gastric region there is a series
of eight slender acute teeth, three of which lie behind the orbit. Rostrum
continuous posteriorly with a sharp, non-sulcated carina which becomes
obsolete before reaching the posterior margin. Suborbital angle prominent,
but not armed with a spine; a small branchiostegian spine projects from the
margin on a level with the second antenna. Neither the cervical nor any
other grooves are apparent on the carapace. A faint longitudinal ridge runs
along the side of the carapace on a level with the orbit; this carina is most
conspicuous on the gastric region. Another longitudinal ridge runs from the
suborbital angle, dividing into two branches near the middle of the carapace.
A third ridge extends from the branchiostegian spine to the lower branch of
the ridge last noted. Fourth, fifth, and sixth abdominal segments dorsally
carinated, the sixth armed with a small horizontal spine. The fourth, fifth,
and sixth segments are also ornamented with a lateral ridge. Eyestalks short ;
eyes large, globular, black.

Length, 93 mm. ; carapace, exclusive of rostrum, 29.5 mm.

Station 3371. 770 fathoms. 1 female.

Solenocera agassizii, sp. nov.

Similar to S. sıphonocera (Philippi), but different from that species in
having the two antennulary flagella much shorter and subequal, and a larger
number of teeth on the upper margin of the rostrum and gastric region.
Comparison of a specimen fifty-seven millimeters long with S. siphonocera of
equal size from the Bay of Naples shows that in the former the antennulary
flagella are but four fifths the length of the carapace, and that there are eight
teeth on the rostrum and gastric region, while in the Neapolitan specimen the
antennulary flagella are as long as the distance from the tip of the rostrum to
the middle of the third abdominal segment (two fifths longer than the cara-
pace), and there are but six teeth on the rostrum and gastric region. More-
over, not only is the upper flagellum broader (1 m.) and blunter in S. agassizii
than in S. siphonocera (where it is only 4m. in breadth), but is also subequal
in breadth to the lower flagellum, while in the Mediterranean species the upper
flagellum is conspieuously narrower than the lower one.

VOL, XXIV. — NO. T. 5

212 BULLETIN OF THE

In full-grown specimens of S. agassizii, which attain a length of 150 mm.,
the antennulary flagella are only one half as long as the carapace, or even legs.
The flagellum of the second antenna is very slender and enormously long, —
more than two and two thirds times the length of the whole body, The num-
ber of teeth on the rostrum and gastric region is nine.

Length of an adult female, 149 mm.; carapace with rostrum, 54 mm.; ros-

cum, 13.5 mm.; second antenna, 410 mm.
Station 3389, 210 fathoms, 2 males, 8 females.
o A 5 males, 6 females,

A Solenocera has been recently recorded from the Bay of Bengal by Mr. J.
Wood-Mason. In this species the antennulary flagella are described as being
shorter and broader than in any previously described species. As no other
characters are mentioned, it is impossible to tell whether it is the same as the
“Albatross” species. When the remoteness of the localities is considered, it
seems hardly warrantable to assume the identity of the East Indian and Amer-
ican species of a comparatively shallow-water genus.

Peneopsis diomedese, sp. nov.

Integument hard, firm, and smooth. Rostrum long, nearly horizontal, ex-
cept near the tip, where it is bent up slightly, acute, armed with four teeth
above. A dorsal carina, armed with one tooth on the posterior part of the
gastric region, runs the length of the carapace. Cervical groove very deep, but
not cutting the dorsal carina, Another deep groove runs backward, and then
diagonally upward from the cervical groove toward the posterior border of the
carapace, stopping just short of the posterior margin. Antennal region well
defined by the cervical groove below and a gastro-antennal groove above. A
strong antennal tooth on the margin of carapace, below the orbit, another at
the infero-lateral angle, a third just behind the groove that marks the posterior
limit of the antennal region, and a fourth on the hinder edge of the cervical
groove. Above and behind the last mentioned tooth the cervical groove is
indented, and the upper angle of this indentation tends to assume the form of
a small tooth or spine.

Fourth, fifth, and sixth apdominal segments carinate on the median dorsal
line, and produced to teeth posteriorly. Telson deeply grooved on dorsal side,
and armed with a pair of spiniform lateral teeth near the tip. The antennules
with their long flagella surpass the whole body in length; the prosartema is
oval, foliaceous, reaching forward as far as the posterior border of the upper
face of the cornea. The exopods of the second maxillipeds are very small, not
longer than the short ischial segment of the limb. Upon the appendages back
of these, the exopods are reduced to the merest rudiments, discernible only by
the use of a lens. In some individuals, indeed, the exopods of the posterior
appendages are altogether wanting.

i Soleñotera hextii Wood-Mason, Ann. Mag. Nat. Hist., 6th ser., VII. 188, 1891.

MUSEUM OF COMPARATIVE ZOOLOGY. 213

Length, 215 mm.; carapace, 101 mm.; rostrum, 42 mm.; antennal scale,
33 mm.; telson, 33 mm.
Station 3353. 695 fathoms. 2 females.

9. 4.8008, 1-000, tt 1 male.

ft ob a Ode: > 408. 5% 3 females,
‘8898... 1020... 3 females,
Bl 3 males, 1 female.
000 00 1 male, 1 female,

Haliporus nereus, sp. nov.

Integument membranaceous, Carapace lightly granulated. Rostrum about
one third the length of the rest of the carapace, horizontal, armed with six
teeth above, ciliated below. Median dorsal line of carapace carinate, with two
teeth on the posterior half of the gastric region. The antennal, branchiostegian,
and hepatic spines are present, besides one behind the branchiostegian, and
another still farther back on the hind border of the cervical groove. Orbital
region definitively bounded by an hepatic and a gastro-hepatie sulcus. The
hindmost of the lateral spines of the carapace lies in the anterior angle of a
triangular field enclosed by branches of the cervical groove ; from the infero-
posterior angle of this triangle two carino run backward along the branchial
area; the upper one ends at the postero-lateral margin of the carapace, the
lower one meets the inferior submarginal carina of the branchial area before
attaining the posterior border. These two carinw, with the submarginal ridge
of the carapace, enclose a long oval area on each branchial region. Third to
sixth abdominal segments carinated ; on the sixth segment the carina ends in
a small tooth. Eyestalks about one half the length of the rostrum ; eyes large,
black, much broader than their stalks.

In the female there is a large process, covered with stiff hairs, and flat-
tened on the inner side, developed from the base of the third pair of legs.
Behind this process lies a pair of flat, setiferous sternal processes. Between
the legs of the fourth pair there hangs in the median line a nearly vertical
curtain-like partition, notched on the free lower margin, and flanked by two
lower blunt setiferous tubercles. The sternum of the posterior thoracic seg-
ment has a slightly elevated median longitudinal ridge, and a low transverse
ridge at the posterior boundary of the segment.

Length, 81 mm.; carapace, 31 mm.; rostrum, 8 mm.

Station 3353. 695 fathoms. 1 female.

Carod hOB Tare” 1 female.
“QUAN. 1700. oe 2 females.
its 199008. 1579.41 2 females.
se Baer dAYeU 1 male.

u 9400. Tage “ 2 females.
m BOr ooo.” 1 female.

« 3413. 1360 '" 1 male, 2 females.

BULLETIN OF THE

Haliporus doris, sp. nov.

A larger and more robust species than H. nereus, with a stronger upward
curve to the rostrum. The tubercular processes of the third pair of legs, to-
gether with the sternal tubercles immediately behind them, are very like the
corresponding structures in H. nereus. But here the likeness between the
sexual parts of the females of the two species ends. Betwixt the legs of
the fourth pair there is a transverse diaphragm consisting of a median tongue,
concave on its front face, its lower edge entire, supported on each side by a
strong blunt triangular process, of equal height with the median tongue-like
plate. Between the bases of the fifth pair of legs, in place of the low longi-
tudinal ridge seen in H. nereus, there is, in this species, a large setiferous tri-
gonal tooth, acute at the tip, equalling in height the transverse partition on
the sternum of the antecedent segment. The posterior median angle of this
tooth abuts against the low transverse ridge that forms the posterior limit of
the last thoracic sternum.

Length, 104 mm.; rostrum, 10 mm.; whole carapace, 42 mm.

Station 3414, 2232 fathoms, 4 females.

“ 98415, 1879 “ 1 female.

Haliporus thetis, sp. nov.

Integument membranaceous. Rostrum rather less than one third as long as
the vest of the carapace, curved strongly upward, upper margin convex, armed
with five teeth. A distinct carina, bearing three teeth, runs the length of the
median line of the gastric area, from the base of the rostrum to the cervical
groove. Back of the cervical groove this carina continues on as a low blunt
ridge to the posterior border of the carapace. Posteriorly, this ridge gives off
two pairs of lateral branches, which course diagonally backward to the poste-
rior margin of the carapace, marking off two triangular fields on the cardiac
region, one enclosed within the other. The spinous armature of the carapace
consists of an antennal, a very small branchiostegal, an hepatic, and a lateral
spine on the posterior edge of the cervical groove. The spine which lies a
little way behind the branchiostegian spine in H. nereus and H. doris is lack-
ing in this species. Antennal and orbital regions well defined. All the seg-
ments of the abdomen are carinate along their median dorsal line. The pleure
are shallow and rounded. A longitudinal furrow on each side of the abdom-
inal segments. Eyestalks one half the length of the rostrum; eyes large,
black, broader than their stalks.

Length, 94 mm.; rostrum, 9 mm.; carapace, including rostrum, 33 mm.
Station 3413. 1360 fathoms. 1 female,

MUSEUM OF COMPARATIVE ZOOLOGY, 215

Aristzus occidentalis, sp. nov.

Similar to A. antennatus (Risso) of the Mediterranean Sea, By comparison
with Duvernoy’s figures of A. antennatus it appears that the Pacific species
has a longer, more strongly upturned rostrum, and that it differs furthermore
in having the sixth abdominal segment carinated and toothed, and in having
much longer abdominal appendages. As in A. antennatus, there is no epipod
on either the last or the next to the last pair of legs.

Length, 158 mm. ; rostrum, 49 mm. carapace, including rostrum, 85 mm.

Station 3403. 384 fathoms. 1 male.

661. BELO 08 le see! 2 females.

Hemipeneus triton, sp. nov

This species, like H. spinidorsalis Bate, is remarkable for the long, curved
thorn on the dorsal side of the third abdominal segment. It differs from
H. spinidorsalis in having a much shorter rostrum, and longer, more flattened
outer antennulary flagella. The rostrum is shorter than in any previously
described species, being much shorter than the eyestalks ; it is furnished with
from two to four teeth above. The sixth abdominal somite is longer than in
H. spinidorsalis. Another difference between the two species affects the inner
branch of the second abdominal appendages of the male; in both species this
branch is triple; in H. spinidorsalis the anterior piece is a horny plate, concave
within and furnished with seta on its distal border; the middle ‘siece has the
form of a long triangular tooth, shorter than the anterior plate ; the posterior
piece is a longer multiarticulate flagellum, homologous with the inner branch
of the following pairs of appendages. In H. triton, the middle tooth-like
process is developed into a broad plate which exceeds in length the anterior
plate.

Length, 127 mm.; carapace, 44.5 mm. ; rostrum, 6 mm.

Station 3360. 1672 fathoms. 1 female.

" 3374 16 á “ 1 male, 5 females.
"s Obes 1709 ga 6 3 males, 1 female.

Benthesicymus tanneri, sp. nov.

In this species the carina on the fifth and. sixth abdominal segments termi-
nates posteriorly in a small acute tooth, whereas in B. altus Bate these segments
are devoid of teeth, the posterior margin rising to form a peculiar transverse
ridge. B. brasiliensis Bate differs from B. tanneri in having the carapace
higher, with more convex infero-lateral borders, the third, fourth, and fifth
abdominal segments toothed, the sixth toothless, and in the absence of the
tooth or tubercle on the sternum between the abdominal appendages of the

BULLETIN OF THE

216

first pair. B. bartletti Smith, from the Atlantic side of the continent,
resembles B. tanneri in many respects, but is distinguishable at a glance by
the long spine on the dorsum of the fifth abdominal segment. B. moratus
Smith, another allied species, differs in having a distinct hepatic spine, a much
broader merus joint to the second maxilliped, in the presence of small exopods
at the bases of all the thoracic appendages, etc.

Color in life, deep red, with a large patch of bright blue on the back of the
second, third, and fourth abdominal segments ; eyes black.

Length, 112 mm. ; carapace 44.5 mm. ; rostrum, 8 mm.

Station 3358. 555 fathoms. 3 males, 2 females.

ERB al e. A 1 male, 1 female.
Oo 9868) 978 i 3 males, 5 females.
«> 3964, 902 P. 2 males.

"8860." 1010 E ] female.

* 23366. 1067 dj 1 male, 1 female.
A A = 2 females.

mo DO ronan er 1 male.

& 8380. 800 ee 1 male,

“8884... 408 N 27 males, 36 females.
«9396... 1020 d 9 males, 3 females.
fey, 0400... 1022 M 1 female.

€... 0408 Bod hou 1 male, 1 female.
e — 90404. ...386 ud 1 female.

IINE PO y PIDE lo) p 5 females.

© . 3410. «100k 4 2 males, 3 females.
Seo Oblate FH 2.0 1 male, 1 female.
& SO. 000) di 7 males, 8 females.
fe A L0 bed 1 male,

* 9420, , O80 di 1 male.

* . 0480. 809 dj 3 males, 2 females.
« 3436. 905 de 5 females.

Family SERGESTIDZE.

Sergestes inous, sp. nov.

Near S. mollis Smith. Besides differences of minor importance, the follow-
ing structural, difference is apparent: in S. mollis the posterior pleurobranchia
of the antepenultimate thoracic segment is replaced by a small simple lamella,
which is concealed beneath the following gill, while in S. ¿nous the said.
pleurobranchia is well developed and unconcealed.

Length 113 mm. ; carapace, 34.5 mm.

Station 3380. 899 fathoms. 1 female.

MUSEUM OF COMPARATIVE ZOOLOGY. 217

Sergestes phorcus, sp. nov.

Carapace devoid of spines; rostrum cristiform, short, subquadrate, the
anterior margin produced to a short point in the middle. Second, third, and
fourth abdominal segments lightly sulcate in the median dorsal line ; sixth
somite armed with a minute posterior dorsal spine. Eyestalk shorter than
proximal segment of the antennulary peduncle ; eye subspherical, much wider
than its stalk. First and second segments of the antennulary peduncle of
equal length, the third considerably shorter. Thoracic appendages much like
those of S. robustus Smith.

Length, 65 mm. ; carapace, 21 mm.

Station 3382. 1793 fathoms. 1 male.

5 3386. 242 Hf 1 female.
E 3388. 1168 " 2 females.
5 3401. 395 " 1 female.
* .. 8497. -6028 es 1 female.

Sergestes halia, sp. nov.

A small species, in which the cervical groove is nearly obliterated on the
dorsal part of the carapace, and the posterior transverse furrow, which in
some species of Sergestes forms the front boundary of the cardiac area, is
obsolete. A sharp spine near the antero-lateral margin of the carapace, and
another on the hepatic area. Rostrum cristiform, tapering to a slender, acute
point, which overreaches the eye. A longitudinal ridge runs from the base of
the antenna backwards, dividing at the hepatic spine into a superior and an
inferior branch'; the former forms the upper boundary of the branchial area,
the latter courses along the middle of the branchial area and fades out before
reaching the posterior border of the carapace, Abdominal segments unarmed.

Length, 29 mm. ; carapace, 9.3 mm.

Station 3388. Surface to 400 fathoms (submarine tow-net). 3 males.

SUBORDER SCHIZOPODA.
Family LOPHOGASTRIDZ.

Gnathophausia dentata, sp. nov.

In this species a thin triangular crest, produced at the apex to a spine, is
found at the base of the rostrum, over the anterior part of the gastric region ;
a minute denticle near the anterior, and another near the posterior end of the
crest. The lower spine of the infero-posterior angle of the carapace is reduced
to a mere tooth, obsolete in some specimens. Antennal scale very broad
(breadth equal to one half the length.)

Distinguished from G, gracilis W.-Suhm by the prominent dentate gastric

218 BULLETIN OF THE

crest, the small size, or even absence, of the lower spine of the infero-posterior
angle of the carapace, the greater breadth of the antennal scale, etc.
Gnathophausia gracilis, var. brevispinis Wood-Mason, agrees with G. dentata
in the obsolescence of the lower posterior spine, but I infer from Wood-Mason’s
short description that this variety conforms to the type of G. gracilis as regards
the gastric teeth, antennal scale, etc.
Length, 60 mm.; carapace, including rostrum and dorsal spine, 33 mm.;
rostrum, 14 mm.; dorsal spine, 4 mm.
Station 3361. 1471 fathoms. 1 specimen.
an 100104. AOL Ran
“ 8400. 1822 “
« 8400. SL ©
weal, HS A

RS ee
a

Family EUCOPIIDZ.
Eucopia sculpticauda, sp. nov.

Frontal margin very prominent, forming a three-sided blunt rostram, which
projects between the eyestalks and wholly conceals the subjacent ocular seg-
ment, Anterior part of the telson deeply furrowed for a little less than one
half its length, the furrow bounded on each side by an elevated, rounded ridge;
a low median keel, beginning in the anterior furrow, runs back to the hind
end of the telson ; a constriction near the hind end divides off a terminal plate,
which is broadly rounded at the end, its lateral margin concave; the whole
dorsal face of the telson, from the posterior end of the anterior ridges to the
terminal plate, is beautifully ornamented with a network of ridges like honey-
conib,

Length, 66 mm.; carapace, measured from lower angle of orbit to posterior
end of the lateral wings, 23 mm.

Station 9619 Hydr. 1000 fathoms. 1 female juv.

“ 3407. ona 1 female.
“ — 9413. 1300 00 1 male.

Family MYSID.ZE.
Petalophthalmus pacificus, sp. nov.

Similar to P. armiger W.-Suhm,! but different in some particulars. The
rostrum is more prominent and triangular; there is a median tooth on the

1 Amongst the materia] dredged by the ** Blake" in the Atlantic in 1877-78, I find
the female of P. armiger. It agrees closely with the male, barring the usual sex-
ual differences, viz. the presence of an incubatory pouch, and the simple structure
of the caudal limbs; the mandibular palp, carapace, telson, etc., are as in the male.
The brood-pouch consists of six pairs of incubatory lamelle. The Schizopod
described by Suhm as the female of P. armiger is apparently a Boreomysis. It may
be called Boreomysis suhmı.

MUSEUM OF COMPARATIVE ZOOLOGY. 219

carapace behind the base of the rostrum : the caudal limbs are simple in the
male.

Length, 31 mm.

Station 2637 Hydr. 700 fathoms. 1 male.

SCOLOPHTHALMUS, gen. nov

Carapace rostrate, small, covering only the anterior part of the thorax, leav-
ing the last two thoracic segments exposed. Abdomen slender, cylindrical,
sixth segment the longest. Eyestalks transformed to sharp spines, visual ele-
ments wanting. Antennular peduncle rather long and robust. Antennal scale
elongate, oval; proximal fourth of outer edge smooth, naked, terminating in
an angle which bears a strong seta; the rest of the outer edge fringed with
hairs like the inner edge. Mandibular palpi reaching to about the middle of
the antennulary flagellum, terminal segment narrowly oblong, Second max-
illae with inner basal part expanded, three-lobed, terminal segment expanded
at distal end, triangular. First maxillipeds without an exopod, but furnished
with a well developed epipod. Second maxillipeds of moderate length, sub-
pediform. Legs long and slender, antepenultimate pair reaching forward
beyond the base of the antennal scale. Marsupial pouch of the female com-
posed of seven pairs of incubatory lamelle. Telson large, apex truncate, not
incised. Outer plates of the swimmerets (uropods) not divided by a transverse

suture.

Scolophthalmus lucifugus, sp. nov.

Frontal margin of carapace produced so as to form an acute rostrum ; antero-
lateral margins oblique, armed with two spines, one behind the external margin
of the antennule, the other at the anterior inferior angle. Cervical suleus
wel! marked, with a distinct linguiform dorsal area behind it. Posterior lat-
eral wings short and rounded. Abdominal segments smooth and cylindrical,
sixth segment nearly as long as the two antecedent segments combined, Telson
broad, truncate. Eyes atrophied, their peduncles assuming the form of spines.
Antennal peduncles one half as long as the carapace, second and third segments
about equal in length, a little longer than the first, the third slenderer than
the first and second ; the flagella are equal to or rather longer than the pedun-
cle, and present a uniform structure of small articulations, the basal part of the
outer flagellum not being expanded. The antennal scale and mandibular palpi
are described in the generic diagnosis ; the antennal scale is equal in length to
the antennular flagellum, Abdominal limbs (of the female) simple, increasing
in length posteriorly, the fifth pair distinctly two-jointed. ‘Inner branch of
uropods slender, lanceolate, surpassing the telson and the outer branch, the
latter of which is oval and divided by transverse suture.

Length, 42 mm.

Station 3400. 1322 fathoms. 1 female,

920 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY.

CERATOMYSIS, gen. nov.

Cephalothorax robust. Carapace short (leaving the last two segments of the
thorax exposed), spinose ; frontal margin truncate, not forming a rostrum, armed
at the external angles with a pair of long horn-like spines ; a conspicuous notch
at the anterior end of the lateral margin serving as an excurrent orifice from
the respiratory chamber. Abdomen slender, cylindrical, spinose ; sixth seg-
ment not much longer than the fifth. Eyes absent, their stalks taking the
form of slender styles, whose tips are soft and delicate, perhaps serving as
tactile organs. Antennular peduncle rather short ; flagella much longer than
the peduncle, flattened and fringed with long sete on their margins. Antennal
scale linear, ciliated on both margins. Mandibular palpi reaching beyond the
antennular peduncle ; terminal segment long oval, margins fringed with long
sete. First maxillipeds devoid of an exopod. Distal segment of second max-
illipeds oval, ciliate. Legs long and slender, propodi not segmented. Seven
pairs of incubatory lamelle in the female. Fifth and sixth abdominal limbs
elongated in the female. Telson linear, setose on the margin. Both branches
of the uropods linear, setose on each margin, subequal, shorter than the telson;
outer branch not divided by a transverse suture.

Ceratomysis spinosa, sp. nov.

In front of the cervical groove are three long erect spines in the median line,
the foremost of which,is on the frontal margin; there is, besides, a spine on
each side of the carapace in a transverse line with the middle one of the median
row; behind the cervical groove there is one spine in the median line near the
posterior margin of the carapace, two on each side of the tongue-shaped dorsal
area, and a long row of six on each side, in line with the lateral spines of the
gastric area. Abdomen armed with five longitudinal rows of spines (one me-
dian, four paired and lateral). Telson very long and narrow, nearly equalling
the length of the remaining portion of the abdomen.

Color in life, milk-white.

Length, 36 mm.; carapace, 9 mm.

Station 3357. 782 fathoms. 1 female.

Date

Harva

ONAN a

3 2044 066 303 256

Due

DO NOT C

( ld US
ntn WD M
vem

NOV 3 0 1988

IRCULATE