V'^^

'ibrani of tlje Uluseum

OK

COMPARATIVE ZOOLOGY,

AT HARVARD COllECE, CAMBRIDGE, MASS.

iFoun'Dc'B bji prfbate suiiscrfptfoii, fii 1861.

No.^^d-j.

%cl },,yU. U. IS^-)^ .

i

/ I

J

NATURAL HISTORY

OF THE

UNITED STATES.

CONTRIBUTIONS

TO

THE NATURAL HISTORY

OP THE

UNITED STATES OF AMERICA.

BY

LOUIS AGASSIZ,

THIRD MONOGRAPH.

[by ALEXANDER AGASSIZ.]

IN TWO PARTS. — I. EMBRYOLOGY OF THE STARFISH. — II. HARD PARTS OF SOME

NORTH AMERICAN STARFISHES.

VOL. V.

BOSTON:

LITTLE, BROWN AND COMPANY.

1877.

tmmxs d tl^t p;uscum of Cnmparatik ^oola^j)

at harvard college.
Vol. v. No. 1.

NORTH AMERICAN STARFISHES.

ALEXANDER AGASSIZ.

""WITH TWENTY PLATES.

CAMBRIDGE:

WELCH, BIGELOW, AND COMPANY,

!Enibccsits lluss.

1877.

PREFACE.

The Plates which accompany this volume* have now been drawn on
stone for more than twelve years. It was the intention of the late Pra
fessor Agassiz to add to them the anatomy of several of our more com-
mon species, but the duties connected with the care of the Museum pre-
vented him from accomj^lishing this task. Although during the last twelve
years several important papers have been published on the anatomy of
Echinoderms which would necessitate a complete re-examination of the
anatomy of Starfishes, it has been thought best, since there was no proba-
bility of being able to finish Avithin a reasonable time the necessary ana-
tomical investigations to complete this volume as originally planned, to
publish the Plates as they were left by Professor Agassiz; all that has
been added to them is the lettering necessary for their proper explana-
tion. However incompletely the subject of Starfishes is thus presented,
these Plates cannot fail to be of value not only as illustrations of a
number of our American Starfishes, and as showing the systematic value
of characters thus far almost completely neglected, but also as determining
the homology of several genera not previously figured, the solid parts
of which are given in detail. As several European naturalists are at the
present moment engaged upon the study of the Starfishes, it appeared
judicious to issue these Plates before they became antiquated.

* They were intended to accompany the text of the fifth volume of the " Contributions to the Natural
History of the United States," by L. Agassiz.

iv PEEFACE.

The Memoir on the Embryology of the Starfish, Part I., has been repub-
lished substantially as it originally appeared in 1864, in advance of the
remainder of the volume. I have added notes in brackets on the points
where additions have been made by subsequent investigations for the sake
of calling attention to the present condition of| the subject, and I beg the
reader to remember that it was written thirteen years ago.

ALEXANDER AGASSIZ.

Museum of Comparative Zoology, >
Cambridge, April, 1877. |

CONTENTS.

PEEFACE.

PAET I. Embryology of the Staefish.

CHAPTER FIRST. Artificial Fecundation, and History of the Development of the
Larva, p. 3.

CHAPTER SECOND. History of the Development op the Starfish proper, p. 30 ;
Recapitulation, p. 36.

CHAPTER THIRD. Embryological Classification of Starfishes, p. 59.

CHAPTER FOURTH. Examination of the Investigations of former Observers, p. 66.

CHAPTER FIFTH. On the Plan of Development of Echinoderms, p. 75.

PAET II. On the Solid Parts of some North American Starfishes.

Homologies of Echinoderms, p. 87 ; Description op the Hard Parts op some North
American Starfishes, p. 94 ; Asterias, p. 94 ; Echinaster sentus, p. 97 ; Crossaster,
p. 98 ; Pycnopodia helianthoides, p. 100 ; Brisinqa, p. 102 ; Luidia Gdildingii, p. 105 ;
Asterina folium, p. 106 ; Asteropsis imbricata, p. 106 ; Pentaceros reticdlatus, p. 108 ;
Solaster endeca, p. 112 ; Cribrella sangdinolenta, p. 113 ; Astropecten articulatus,
p. 114 ; Luidia clathrata, p. 117 ; Fascioles of Starfishes, p. 119.

NOTE, p. 121.

EXPLANATION OF THE PLATES, p. 122.

PA KT I.

EMBRYOLOGY OF THE STARFISH.

By ALEXANDER AGASSIZ.

[Part r., together with Plates I. -VIII., was published in December, 18G4; the text included in
brackets has been added with the subsequent part.]

EMBRYOLOGY OF THE STARFISH.

CHAPTEE FIRST.

ARTIFICIAL FECUNDATION, AND HISTORY OF THE DEVELOPMENT OF THE LARVA.

Differences of the Sexes. — Since the existence of different sexual organs
in separate individuals was first pointed out among the lower animals,
the tendency of every additional advance in our knowledge of their struc-
ture has been to bring out more fully the differences of sex between them.
But recently, we did not even know that among the Medusae there were
male and female individuals; and yet, at the present day, it is a com-
paratively easy task to distinguish, among the larger Jelly-fishes, the
males from the females. The difference of coloring is very striking. The
spermaries of the males are often brilliantly tinged, while the ovaries of
the females are of duller hues. We thus find among Jelly-fishes the fii'st
indication of an almost universal law in the animal kingdom, and which is
nowhere carried out to so great a degree as among Birds. A casual ob-
server could not fail to distinguish a male from a female Aurelia, — though
the great difference in the coloring of the males and females had not been
perceived by naturalists till it was first pointed out by Professor Agassiz,
in Aurelia flavidula Per. et Les. In Melicertum, in Turris, in Staurophora,
in Circe, a glance will suffice to determine the sex of the individual ;
while a single look through a magnifying-glass will reveal to us the sex
of the smaller species, such as Eucope, Pennaria, Euphysa, and the like.
The difference of the sexes of some Echinoderms is easily perceived by
their difference of coloring at the time of spawning; among them are
our common Starfishes and our Sea-urchins.

The males and females of our common species of Starfishes, Asteracan-
thion pallid us Agass. (A. vulgaris Stimp.?), and Asteracanthion berylinus
Agass., can readily be distinguished by their difference in coloring : all

2

4 EMBRYOLOGY OF THE STARFISH.

those having a bluish tint being invariably females; a reddish or reddish-
brown color indicating a male. Among the many specimens I have had
occasion to open, I have thus far never found a single exception. When
cut open, so as to expose the genital organs, the difference between the
males and females is still more striking. The long grape-like clusters of
reproductive organs extending from the angle of the arms, on both sides
of the ambulacral system, to the extremity of the rays, present very
marked differences in the two sexes. The ovaries are bright orange,
while the spermaries are of a dull cream-color. At the time of spawning,
which is very different in the two species mentioned above, the genital
organs are distended to the utmost, filling completely the whole of the
cavity of the ray; the abactinal system itself being greatly expanded by
the extraordinary development of these organs.

Artificial Fecundation. — If we take a male and female Starfish in this
state, and cut a portion of the genital organs into small pieces, we shall
find that the eggs and spermaries escape in such quantities as to render
turbid the water in which they are placed. Throwing these small pieces
of the genital organs into shallow dishes containing fresh sea-water, and
stirring the mixture thorovighly to insure the contact between the sper-
maries and the eggs, will be sufficient to fecundate the latter. In order
to make the operation perfectly successful, some precautions are neces-
sary : all the pieces of the genital organs, which are left after repeated
stirring, must be carefully removed ; there must not be too many eggs
in one dish, so that the water can have free access to them in every
direction. The removal of the remnants of the ovaries and spermaries
is very necessary, as the pieces which remain clotted together decom-
pose very rapidly, and endanger the safety of the eggs, even when the
water can be changed with the greatest facility. As soon as the fecun-
dation is fulfilled, the water in the dishes must be i-epeatedly changed
until it becomes perfectly clear, for the presence of too many spermaries,
rendering the water milky, prevents a favorable result. It is best only
to use one male and one female for the mixture in each vessel, as eggs
taken from many individuals lessen the chances of success. The eggs
sink to the bottom, so that the water can be poured off" and changed
without much danger of throwing them away. Immediately after the
mixture is made, the water should be changed three or four times in
succession; after that, every half-hour, until the fourth hour, when an

CHANGES IX THE EGG. 5

interval of two to four hours may elapse before renewing the water. As
it is extremely difficult to change the water after the embryos have
hatched and are swimming freely about in the jar, without losing many
of them, it is advisable, before they hatch, which is about ten hours after
the fecimdation, to reduce the water to a minimum volume, and then
simply to add a little fresh sea-water and remove the contents of the
vessel to larger and larger jars. In this way the water can be main-
tained sufficiently pure, until the young embryos have taken the habit
of swimming near the surface, when it may all be drawn off by means
of a siphon. A great deal of time and trouble will be saved by this
mode of procedure, and fewer specimens lost. The jars containing the
eggs should be kept in a cool place ; the most convenient method of
securing a low and even temperature is to place the small jars in large
tubs filled with cold water.

Changes in the Egg. — At the time of spawning, the eggs in the ovaries
are so closely packed that they are pressed into all sorts of shapes, tri-
angular, polygonal, elliptical ; but when placed in water, and allowed to
remain a short time, they soon become perfectly spherical (PI. I. Fig. 1).
The following numbers are the ratios of the diameters of the yolk, the
germinative vesicle, and the germinative dot, the outer envelope being 1 :
the 3'olk is 0.75, the germinative vesicle 0.22, and the germinative dot
0.08. The formation of the egg in the ovary, and its changes up to the
time of spawning, I have had neither time nor opportunity, thus far, to
examine.

The spermatic particles, which swim about with great rapidity on
escaping from the spermaries, soon find their way to the outer envelope
of the egg to which they attach themselves, beating about very violently
the whole time. The jDarticles remain imbedded in the thickness of the
outer envelope, and are sometimes so crowded as to form a halo round
the egg (PI. I. Figs. 1-4). I have not, in a single case, seen any of
the particles penetrate through the outer envelope and reach the yolk
itself.

Probably a great deal of the diffijrence of opinion prevailing among
Physiologists, as to whether the spermatic particles penetrate through
the successive envelopes of the egg to the yolk itself, is due to the want
of precision still existing in our knowledge concerning the envelopes
of the yolk in the different branches of the animal kingdom. We do not

6 EMBRYOLOGY OF THE STAEFISH.

know whether what we call the outer envelope of the egg of an Echino-
clenn is homologous to the outer envelope of the egg of an Acaleph, of a
Polyp, or of Worms, Insects, or Crustacea, or how for these envelopes are
found in the ovarian eggs of Mammals, Birds, Eeptiles, and Fishes. And
before we can come to a satisfactory result as to the place in the egg
which the spermatic particles reach before changes can be observed to
take place in the yolk, the eggs of the different classes of Animals must
be carefully compared Avith reference to this point. The first phenome-
non which precedes any change in the egg is a rotary motion given to
the whole egg by the constant beating of the spermatic particles; the
germinative vesicle disappears (PI. I. Fig. 2) soon after this, and next the
germinative dot (PI. I. Fig. 3). The yolk has then all the appearance of
an egg which has undergone segmentation, and the yolk of which should
consist of innumerable small spheres. The yolk has the same granular
structure previous to segmentation which has usually been considered to
belong to it only after the segmentation is complete. [The phenomena
preceding segmentation, the structure of the yolk, the mode of forma-
tion of the Richtung's-Blaschen, the manner in which the germinative
vesicle disappears, are subjects which since the preceding investigations
were made have all received considerable attention. The explanations
given of these points are therefore all subject to revision and to correc-
tion. See more particularly the papers by Ludwig, C. Semper, Lan-
KESTER, Hartwig, Fol, Auerbach, Balfour ; sundry Embryological Me-
moii-s by E. Van Beneden, Composition de I'oeuf, 1870; Kowalewsky
A. Mem. Akad. St. Peters, XVI., 1871 ; Blutschli, Die Eizelle ; Haeckel
E., Die Gastrsea Theorie ; Strassburger, Die Zelle.] The resemblance
between these two stages is still more marked in the eggs of Cteno-
phorae, where the ratio between the diameter of the yolk and that of
the outer envelope is large, and in which the segmentation is carried
on until the whole yolk consists of such minute spheres that it is impos-
sible at first sight to distinguish an egg of a Ctenophorous Medusa, which
has undergone complete segmentation, from one in which the segmenta-
tion has not even begun, after the germinative vesicle and dot have dis-
appeared. The disappearance of the germinative dot is accompanied by
a separation of the yolk from the inner w\all of the outer envelope of
the egg (PL I. Fig. 3); this is the first step towards segmentation, and
the presence of such a marked interval would greatly facilitate the detec-

CHAXGES IX THE EGG 7

tion of spermatic particles upon the surface of the yolk, if any of them
had penetrated through the outer membrane. The first trace of segmen-
tation consists in a depression of the yolk, visible on one side of the
sphere (PI. I. Fig. 4), and is soon followed by a similar change on the
opposite pole.

The segmentation takes place very rapidly, passing in about eight
hours from the stage represented by PI. I. Fuj. 3 to that of PI. I. Fig. 21,
immediately before the escape of the embryo from the egg. The spheres
in the earlier stages of segmentation are well separated (PI. I. Figs. 7, 9,
11, 13). They have a centrifugal tendency, and, as they increase in
number, arrange themselves in a shell-like envelope, which eventually
becomes the wall of the embryo. This tendency is already apparent
when there are not more than eight spheres (PI. I. Figs. 13, 14) ; and
as early as the stage represented on PI. I. Fig. 16, where there are only
thirty-two spheres, the envelope is quite prominent. The rotation of the
spheres of segmentation commences before this (PI. I. Fig. 6), and is
entirely independent of the motion given to the whole egg by the sper-
matic particles ; this stops soon after the rotation of the spheres of segmen-
tation has commenced.

As the egg of the Starfish presents nothing peculiar in its process of
segmentation beyond what has been just remarked, I refer the reader to
the explanation of the plates for the details concerning every successive
step of this process, as observed in Asteracanthion berylinus.

The Richtung's-Blaschen of Schultze, which he first noticed in the seg-
mentation of Mollusks, and which were afterwards seen by Lacaze-Duthiers
and by Robin, who traced their mode of development, were also observed
in the segmentation of the yolk of our Starfish. They are noticed, before
the yolk has been divided into halves (PI. 1. Fig. 5), as three or four
small granules, situated at the extremity of the axis which is to divide
the yolk into two portions (PI. I. Fig. 6). They are developed from the
yolk itself as a slight swelling, which afterwards becomes entirely distinct
from the mass of the yolk (PI. I. Fig. 7), retaining always throughout
the whole process of segmentation the same relative position to the axis
of segmentation (PI. I. Figs. 9-17). "What part they play in the subse-
quent history of the embryo I have not been able to ascertain. Without
doubt they always hold the same relation to the first axis of segmenta-
tion, and are, as far as I have observed them in the segmentation of

8 EMBRYOLOGY OF THE STARFISH.

Asteracanthion and of Toxopneustes, invariably at one pole of the first
axis of segmentation.

The Emhryo after IlaicUng. — At about the end of the tenth hour after
fecundation, the segmentation has been carried so far that the walls of the
future embryo have become quite conspicuous, and it is now ready to hatch
(PI. I. Fig. 21). When the outer envelope is torn, the young rotate slowly
about, around a shifting axis, by means of very minute cilia placed over
the whole surface; the walls are everywhere of the same thickness, and
the embryo is perfectly spherical. A difference soon becomes evident ; the
walls thicken at one pole of the sphere (PI. I. Fig. 22, a), and the thick-
ening is accompanied by a flattening of the same side (PI. I. Fig. 23, «);
the embryo has lost its regular spherical shape and its homogeneous walls
(PI. I. Fig. 23, a). The next change consists in a slight depression at this
flattened pole (PI. I. Fig. 24, a); the wall bends inward, forming a very
shallow depression, growing deeper and deeper, until it forms a pouch
extending half the length of the embryo (PI. I. Figs. 25, 26, d, 27, d).
[This stage has become well known as the gastraja stage of Haeckel ; for
a fuller discussion of the gastrtea theory see my Memoir on the Embry-
ology of the Ctenophorje, Mem. Amer. Acad., 1874, p. 379.] While a
cavity (rf) is thus formed by the simple folding in of the outer wall, the
embryo is constantly lengthening and becomes more cylindrical ; the walls
of the extremity opposite the pouch becoming attenuated, while, imme-
diately round the opening of the cavity, the walls have not lost their
original thickness (PI. I. Figs. 26, 27, a). Water flows freely into and
out of this cavity; currents are established, running in different directions
along opposite walls of the pouch, showing this opening to be for the
present a mouth ; the pouch, or digestive cavity, sustains the same relation
to the Avhole body as in the most regular and circular radiated animals,
such as young Actinise, or young Porites. The motion of the embryo,
which immediately after escaping from the Qg'g is an extremely slow
rotation, increases in rapidity as it lengthens, and by the time the cavity
equals half the length of the embryo (PI. I. Fig. 27, d\ the motion is
much accelerated. Instead of a simple slow rotation, with scarcely any
motion of translation, the latter is now quite rapid, and is accompanied
by a slow rotation round a vertical axis, through the centre of the longer
diameter of the animal ; the opening leading into the coecum is foremost
during; their motion.

THE EMBEYO AFTER HATCHING. 9

At the end of about twenty hours after fecundation the embryo has
reached the condition just described ; it is noAV somewhat pear-shaped,
with rounded extremities (Ph I. Fig. 27), having at one end an opening
(«), leading into a pouch {d), which extends half the length of the cylin-
der.* We have now the embryo in a condition which can best be com-
pared to the embryos of other Eadiates ; for there is as yet nothing of
the complication hereafter introduced in the subject by the development
of bilateral parts, obscuring the plan upon which the embryo is built.
It is an embryo closely I'esembling those of the other Radiates, in which,
however, the class-characters, distinguishing it from the embryos of the
other classes of the type, are already developed beyond question. In
the young Polyps the earliest appearance of the class-characters is de-
noted by the presence of a few radiating partitions, dividing the cavity
of the embryo into distinct chambers. In the Acalephs, in the most
rudimentary stages, we already find the chymiferous tubes pushing their
way through the spherosome ; while in our larvte the echinodermoid
class-character, that of having distinct walls, forming the different organs,
is already plainly visible from the mode of formation of this digestive
cavity. What unites all these embryos in one great type is, that we
have in them all an axis around which are arranged the different elements

* So far, the changes which have been observed do not differ materially from what we know of
the earlier stages of Echinoderra larvae, from the observations of Derbes, Miiller, and Krohn. As I
have shown, in the Memoirs of the American Academy for 1864, the earlier stages of the Echinus larvse,
as they have been figured by Derbes, agree in the main points with what has been observed of the
earlier stages of our American Echinus larvie (Toxopneustes drobachiensis). With the exception,
however, that Derbfes, not having followed all the intermediate stages between his figures 15 and 16
in the Annales des Sciences Naturelles for 1847, did not see the transformations the digestive cavity
undergoes, and committed, therefore, the very natui-al mistake of supposing that the first-formed opening,
which we have described as a mouth, retained the same function afterwards. He, however, correctly
noticed the separation of the three cavities, the oesophagus, the stomach, and the alimentary canal,
into which this primary cavity is gradually differentiated, and has given a correct description of their
relation to each other. Miiller has taken up this same subject rather where Krohn and Derbes have left
it, and although he has traced the development from the egg of several Echinoderm larva, yet he has
not given us as detailed descriptions and figures of the earlier stages, as of those which were more
advanced, and says simply, that in the main points his observations coincided with those of Krohn
and Derbes. Krohn, who has artificially fecundated Echinus lividus, gives us in his figures some of
the missing links in the chain of the observations of Derbes, and shows distinctly for E. lividus,
that the first-formed opening becomes the anus eventually, and in what way this is brought about by
the bending of the bottom of the digestive cavity towards one side of the larva, as is the case in our
Starfish, and the formation at that point of a second opening, which becomes the true mouth, while
the first-formed opening henceforth assumes the function of an anus.

10 EMBRYOLOGY OF THE STARFISH.

of which they are composed. Our young Echinodei'm in this condition
(PI. I. Figs. 23-28) can be strictly homologized with the earlier stages
of a Polyp at the time when the digestive cavity is first formed, before
the appearance of the partitions ; and with an acalephian embryo, where
the digestive cavity alone is developed, previous to the pushing of the
chymiferous tubes through the gelatinous mass. The stages subsequent
to the condition of the embryo here described, represented in PI. I. Fig.
24, not having been traced very carefully by previous observers, we have
not had before us the means of forming a true conception of the mode
of development of the Echinoderms ; for to obtain a clear and precise
idea of the functions of those problematic bodies which have puzzled
Mviller during the whole of his investigations, it is necessary to follow,
step by step, the changes taking place in the pouch of the embryo, which
is in this early stage its digestive cavity {d) ; for it is as much a diges-
tive cavity as that of a young Actinia or a Scyphistoma, where the
same opening serves as mouth and anus. The mode of formation of the
digestive cavity is entirely diflerent in the two classes; in the Polyp
it is hollowed out of the interior of the embryo, while in the Echino-
derm the bending in of the wall forms the stomach. Hence the two
cavities are not homologous, and the openings which lead into them,
though performing similar functions — those of mouth and anus — are
likewise in no way homologous, though they are in all built upon the
plan of radiation. This opening always retains its double function in
the Polyps and some of the Acalephs, while in the Echinoderms it becomes
the anus after the true mouth has been formed, and the currents have
ceased to cii'culate in the extremity of the pouch and to pass out through
the same opening which admitted them.

If there is any doubt that Echinoderms, Acalephs, and Polj'ps belong
to the same great type of the animal kingdom, a comparison of the
young Echinoderm, Acaleph, or Polyp in their earlier stages of growth,
at a time when the spherosome has not yet been divided into its com-
ponent spheromeres, will show how great is their identity of development,
and how little there is in nature to justify the separation of this most
natural great division of the animal kingdom, the Radiates, into Echino-
derms and Coelenterata. I shall I'eturn to this point when speaking of the
homologies of the larvfe of Echinoderms.

Formation of the Mouth. — The perfect symmetry of the larva (PI. I.

FORMATION OF THE MOUTH. H

Fig. 27) is soon modified, and in the next stages of development (PI.
II. Figs. 2, 4), the digestive cavity (d) no longer runs in the centre of
the larva, but is bent slightly to one side. If we examine one of the
embryos about forty hours old (PI. II. Figs. 5, G), we find that great
changes have taken place in the thickness of its walls. The outer wall
has everywhere become much thinner, except near the opening thus
far called mouth, where the decrease is not so marked. The walls of
the digestive cavity, which were of an equal thickness for the whole
length, have become exceedingly attenuated at the bottom of the sac,
and have dilated to a considerable extent, forming a sort of reservoir
with very thin walls at the extremity of the pouch (PL II. Figs. 4, 6, d,
magnified and isolated. Fig. 1, d). These changes in the thickness of
the walls, and in the form of the internal cavity, are also accompanied
by corresponding changes of form in the embryo as a Avhole. The ex-
tremity opposite the so-called mouth has increased in bulk, and greatly
exceeds in size the perforated extremity (PI. II. Figs. 4, 6) of the body.

When seen in profile (PI. II. Figs. 2, 4, 5), still greater changes are
visible j there is a decided difference between the two sides of the em-
bryo, forming what is to become above and below ; calling that part
below, where the mouth is situated in the adult larvte, and which is car-
ried downward in its natural attitude while moving. The dorsal portion
of the larva projects beyond the so-called mouth, so that the perforated
extremity has become bevelled ; the narrowing of the central portion
of the larva has increased, and the digestive cavity which, in younger
embryos, occupies the centre of the cylinder (PI. I. Figs. 27, 28), is bent
towards the lower side (PI. II. Figs. 2, 4, 5, d). The outer wall has be-
come thickened at a point opposite the bent extremity of the digestive
cavity, and the thickening of the wall, together with the bending of the
digestive cavity, goes on till the closed end touches the lower side at m.

The changes which have taken place during the time elapsed since the
twentieth hour have been very gradual. The embryo now enters into a
state where the changes are exceedingly rapid and important; so much
so that at the end of the third day the embryo has, in a rudimentary
state, all the parts characteristic of older, fully developed larvse.

At the end of the second day the reservoir at the extremity of the
digestive cavity has changed its outline from a circular to a lobed one
(PL II. Fig. 8, o) ; the lobes widen towards the sides, almost forming

3

12 EMBEYOLOGT OF THE STARFISH.

diverticula (tv, to'), from the digestive cavity. During this time the main
digestive cavity has entirely lost its cylindrical form ; it has become nar-
rowed at the extremities and bulging in the centre (PI. II. Fiff. 8, and
isolated, Fit/. 9). When seen in profile, and comparing it with earlier
stages (PI. 11. Fiffs. 2, 4, 5, 7, isolated, F/'ff. 10, «), it is at once noticed
that the opening at one end, the present mouth of the larva, has little
by little changed from a position at one extremity of the embryo (PI. I.
Fiffs. 27, 28, a) to a slightly eccentric one (PL II. Fiffs. 4, 5, 7). While
the present mouth is changing its position from a terminal to an eccentric
one, and while the digestive cavity has been expanding at the bottom
into a large reservoir, its closed end is bending more and more towards
one side (PI. II. Fiffs. 2, 4), until it finally touches the outer wall of the
embryo at m (PI. II. F(()i. 5). At this point of junction an opening is
formed, leading into the bottom of the digestive cavity (PI. II. F/'ff. 7) ;
this second opening (w?) is now the true mouth, and performs hereafter
all the functions of a mouth, while the first-formed opening of the young
embryo {a, PL II. Fiffs. 2, 4, 5, 7) is restricted in its functions, and per-
forms hereafter only those of an anus ; although in the eai-ly stages
(PL I. Fiffs. 25, 26, 27, 28; PL IL Fiffs. 2, 4, 5, 6) it had performed the
functions of a mouth. We have thus an apparent anomaly in the fact
that the first opening becomes the anus, while the true mouth is only
formed afterwards ; but this difficulty is readily explained if we compare
the functions of this first>formed opening, the so-called mouth, with what
we find among Polyps, where one and the same opening performs the
double functions of mouth and anus throughout life.

The diverticula {w, w', PL II. Fiffs. 7, 10) do not extend, as would seem
when seen from above (PI. II. F/'ff. 8), at right angles from the main
cavity, but trend obliquely upwards, as seen in profile (PL II. F/'ff. 7),
towards the other extremity of the embryo, as in F/ffs. 7, 10, PL II. The
outer wall, which had formed a connection with the closed extremity of
the digestive cavity, on the lower side, has been drawn out in the shape
of a slender cone (o, PL II. F/'ffs. 7, 10, 11, 14, 17), and becomes the
oesophagus, which leads to an opening {m, the mouth), connecting the
ventral side with the digestive cavity.

Nomenclature. — It will materially assist in the explanation of the sub-
sequent changes of form, and obviate a great deal of circumlocution, if we
at once call the different organs by their true names. The original open-

rOEMATIOX OF THE WATEE-TUBES. 13

ing (a), which performed at first the functions of the mouth, is hereafter
the anus («) ; the second opening, the true mouth (?«), is not formed until
the embryo has arrived near the end of the second day ; it is placed in
the middle of the lower surface, and from this time forward the former
mouth assumes the function of an anus. That portion of the digestive
cavity which leads from the mouth to its bulging portion is the oesopha-
gus (o), the bulging portion is the true digestive cavity, or stomach proper
(d), the short tube leading from the stomach to the anus is the intestine
(c), while the diverticula [to, w) are the two branches of the future water-
system. The reasons for calling these parts mouth, anus, oesophagus,
stomach, intestine, and water-system will become apparent as we trace
the development of the embryo in its more advanced stages, in the fol-
lowing pages.*

The currents, which before had entered through the mouth {a), passed
to the extremity of the cavity (a), and been expelled again through the
same opening (a), now change their course completely ; there is a current
which enters the mouth {m), flows through the oesophagus (o) into the
diverticula {tv, to'), then into the true stomach (d), and is finally rejected
through the anus («). From this time forward it is quite an easy thing
to observe the course of the food ; it is taken into the mouth by means
of the currents produced around its opening, passes rapidly through the
oesophagus, rotates for some time in the spherical stomach {d), and then
passes out slowly through the opening (a) of the alimentary canal (c).
As these currents are more and more distinct as the larvfB grow older,
there can be no doubt that the function of the first-formed opening is
eventually confined to that of an anus, after having performed the func-
tion of mouth during the first stage of growth of the larva.

Formation of ihc Wcder-Tiibes. — By water-tubes I mean the bodies which
have received from Miiller the name of problematic bodies, in their earlier
stages of growth, and which he has called Schlauchsystem, when they
appear, in the older lai'viB, as broad tubes running on each side of the
oesophagus and stomach. These parts he considered as independent sys-
tems, but as they are only different stages of the same thing, as will
appear below, they have received here the name which denotes most

* Other terms are also frequently used, to denote the different parts of radiated animals, which are
not usually adopted ; they will be found fully explained in the third volume of the Contributions to
the Natural History of the United States, by Prof. Agassiz, p. 73, and seq.

14 EMBRYOLOGY OF THE STARFISH.

appropriately the function they assume of circuhating water through the
body of the harva.

The water-tubes («-, w'\ at first (PI. II. Fig^. 7, 8, 9, 12, 13, 14) only
diverticula from the main digestive cavity (t?), become less and less con-
nected with it; and by the end of the second day the constriction at the
point of attachment has almost entirely separated them from the diges-
tive cavity (PI. II. Figs. 15, 16, iv, w). A marked difference is noticed
in the rapidity of growth of these two bodies ; the right-hand one {io'\
when the anus is placed in advance, and the mouth downwards, increases
more rapidly, extending towards the dorsal side, which it eventually reaches,
opening into the surroiuiding medium by a small aperture (PI. II. Fig. 17, h),
the water-pore, or, as Miiller has called it, the dorsal pore. A comparison
oi Figs. 8 and 18 of PL II. Avill perhaps render more evident the trans-
foi-raation of the diverticula {tv, tv') from the digestive cavity into two
separate bodies. All we have to do is to swell out the lobed pouches
{w, w') of I\g. 8, PI. n., then cut them off, removing them a short distance
from the digestive cavity, and we shall have the two independent bodies
{to, 7v') of PL 11. Fig. 18, which have little by little been changing their
relation to the digestive cavity, as described above. This transformation
I have actually observed in every stage of its progress, as it is repre-
sented here isolated (PL II. Figs. 9-16).

The walls of the oesophagus (o), of the digestive cavity {cl), and of the
intestine (c), which up to this time are of nearly the same thickness, quite
rigid, capable of very limited expansion and conti'action (PL II. Figs. 2, 4,
5, 7, isolated. Figs. 10, 11), lose their uniform character with the gradual
circumscription of these three regions. The walls now become quite dif-
ferent in their appearance, and the more marked the separation between
these three organs, the greater the difference in the character of the walls
which circumscribe them (PL II. Figs. 17, 19, 21, 23). In proportion as
the stomach {d) grows more spherical, the angle between it and the intes-
tine (c) is more acute, and the intestine (c) becomes a longer and narrower
tube, with walls much less thick than those of the stomach {il). The walls
of the oesophagus (o) are even more flexible ; the conical tube, leading
from the mouth to the stomach, widening and taking a jjistol-shaped form,
the walls have become so movable, that the opening leading into the
stomach can be closed and opened by the greater power of expansion and
contraction of this part of the walls (PL II. Figs. 23, 25). The mouth (m),

rOEMATION OF THE AVATER-TUBES. 15

as it increases in size, grows triangular, with rounded corners ; the de-
pression in which it is placed divides the larva into two very distinct
regions (PI. II. Figs. 19, 23, 25). Since the formation of the mouth, and
the change of position of the first-formed opening to an eccentric one, we
find the mouth and anus placed on one side of the larva. These open-
ings present, at this stage (PI. IT. Fig. 17), the same relations as the mouth
and anus of Clypeaster and Scutella-like Echinoids, while at a much earlier
period they are more like Pygorhynchus.

If we now return to the water-system, we find that the two diverticula
(«', w), mentioned above (PI. II. Figs. 15, 16), have entirely separated
from the digestive cavity (PI. II. Fig. 18), and are now distinct cavities,
having no connection whatever either with the cavity from which they
originated or with one another ; one of these cavities is entirely closed
{iv), the other («/) connects with the surrounding mediiun by means of
a very small opening, the dorsal pore {b, PI. II. Fig. 23, and isolated.
Fig. 17). Such is the appearance of an embryo at the close of the second
day after fecundation.

Miiller never knew the origin of the water-tubes ; in his last paper
only he becomes aware that they are independent at first, but subse-
quently unite. It must he remembered, in reading his earlier papers, that
he sets at rest, in his last memoir, the doubts he expressed concerning
the independence of the two branches of the water-tubes ; in fact, to
obtain a clear conception of Miiller's views, it is advisable to read his last
memoirs first, to be able to adopt at once the corrections he himself
makes during the laborious course of his investigations. The problem-
atic bodies, however, still remained a puzzle to him, even at the time of
his last memoirs, as he was never aware that they were simple diver-
ticula of the digestive cavity, and were finally transformed into the two
independent branches of the water-tubes, uniting, in subsequent stages
of growth, to form the Y-shaped water-system. Van Beneden saw, in the
young Bipinnaria (Brachina Van Ben.), that the water-tubes are at first
separate, but he did not trace their mode of formation, and no other
observer has since returned to this subject.

[Metschnikofi" states that in some cases there is but a single water-tube,
and that I have mistaken an accumulation of cells for a second water-
tube. I can only state that I have frequently repeated my observations
on the Pluteus of Starfishes, Ophiurans, and Echini, and have invariably found

16 EMBRYOLOGY OF THE STAEFISH.

two water-tubes present, but I bave also seen in Starfisbes and Opbiurans,
as be bas well sbown in Opbiurans alone, that tbe wbole rosette of tbe future
ambulacral system is developed only upon tbe surface of one of tbese, tbe
one communicating with the exterior through the dorsal pore, tbe future
madreporic body.]

Appearance of the Chords of vihratile Cilia. — Tbe cilia, spreading over the
wbole surface, which moved tbe embryo so rapidly at first, have almost
entirely disappeared, and are no longer capable of propelling such a large
mass; consequently, at this last-mentioned stage (PI. II. Fig. 20), the larva
is very sluggish, advancing but little, and rotating slowly about a longi-
tudinal axis at the same time. During tbe third day, the movements
become still more sluggish ; it is then that we find tbe first aj^pearance
of the organs which are to propel the larva in future. The general out-
line does not change during the third day; the principal transformations
are tbe greater bending and extending of tbe oesophagus and alimentary
canal, the increase in size of tbe mouth, of tbe water-tubes, and the
appearance of slight projections, small clusters of vibratile cilia, near tbe
anterior and posterior sides of the mouth, which are the beginning of
rows, extending in older larvae in continuous lines all round the body, and
their only means of locomotion (PI. II. v, v, Figs. 20-28). Tbese rows
are at first two very short arcs {v, v, PI. II. Fig. 22), with their convexi-
ties placed opposite one another on each side of tbe depression in which
tbe mouth is placed [v, v', PI. II. Fig. 21).

Tbe general outline of the larva bas, up to this stage (PI. II. Fig. 20),
midergone but slight modifications, tbe changes taking place principally in
the digestive organs. Tbe phases through which tbe larva passes in the
next three days are of a very different character ; the alimentary canal,
the stomach, and the oesophagus become more circumscribed by the
increasing difference noticeable in tbe walls of these regions. Tbe stom-
ach {cl) is always marked by the greater thickness of its walls ; while,
with increasing age, the walls of the oesophagus (o) become more attenu-
ated, and capable of greater expansion and contraction (PI. II. Figs. 25,
0, 27). We observe, also, a rapid increase in tbe growth of tbe water-tubes
{w, tv), which by tbe end of tbe sixth day (PI. II. Figs. 27, 28) extend
as far as the corners of the mouth and along tbe edge of tbe walls of
tbe stomach, towards the anal extremity (PL II. Figs. 24, 26, tv, w').
When viewed in profile (PI. II. Figs. 25, 27), it will be seen that the

APPEAEANCE OF CILIAEY CHOEDS. 17

plane in which these water-tubes run is not parallel to the longitudinal
axis, but inclined to it in such a manner, that the oesophagus passes
between these two tubes. It is in these stages, represented in PI. II.
Figs. 20-28, that the passage from the initial, truly radiate form to a
bilateral one is the most obvious, and it may be well to dwell for a mo-
ment on the changes which are going on here, and compare them to what
we find in other Radiates. Miiller has always maintained that, the Echi-
noderm larvie being bilateral, we had a passage from a bilateral symmetry
to a radiate type, while in reality this seeming bilaterality is subordinate
to a truly radiate plan of structure. The first question to settle with
regard to this is, whether we have a strictly bilateral form among the
larvje or not, and whether we do not find here a repetition of what is so
constantly met with in the animal kingdom, — the undue preponderance
of some parts, hiding effectually the plan upon which the whole animal
is built; in fxct, the engrafting of a subordinate type upon the type which
remains predominant. With the gradual development of the plastrons
alluded to, as formed from the chord of vibratile cilia, the embryo assumes
more and more a shape which renders it quite difficult to perceive the
original plan of radiation, concealed, as it gradually becomes, by the sym-
metrical arrangement of the edges of these plastrons, which leads one
involuntarily to mistake their mode of execution for the plan upon which
the animal is built. This apparent passage from a strictly radiating form
to a seeming bilateral one is nothing more than what we find constantly
among the adults of this same class, and j-et no one has attempted, for
that reason, to make bilateral animals of the Echinoderms. The Spatan-
goids might as well be called bilateral, and not radiating animals, on
account of the perfectly regular symmetrical arrangement of the fascioles,
extending over all the spheromeres composing the body of such Sjiatan-
goids, and in which even the ambulacral system presents marked fea-
tures of bilateral symmetry. The case is exactly a parallel one ; this
chord of vibratile cilia, and the chord of fascioles, arranged so regularly,
simply conceals in both cases the plan upon which the animal is built,
but does not, in either case, change the plan of radiation into that of
bilaterality. As little should we be justified in removing some of the
Holothurians, such as Cuviera and the like, from the Radiates, simply
because the greater preponderance of some of the ambulacra has brought
out, in these animals conspicuously, a dorsal and a ventral side, and an

]8 EMBEYOLOGY OF THE STAEFISH.

anterior and posterior one. In the embryo of our Starfish, which told
so plainly, in its early stages, of the plan upon which it is built, that
plan is now lost sight of in the extraordinary bilateral development of
some of the parts. But, until Spatangoids and flai>soled Holothurians are
proved to be truly bilateral animals, and not genuine Radiates, with sub-
ordinate bilateral features, these seeming bilateral Echinoderm larva? must
be considered as truly radiate, with bilateral features engrafted upon
them.

Development of the Plastrom. — The cylindrical shape, characterizing the
earlier stages of the larva, disappears soon after the appearance of the
first trace of the appendages which give to these larv^ such a peculiar
appearance, and they now assume the features of the adult. The depres-
sion (PI. II. Figs. 25, 27, m), in which the mouth is placed, becomes more
marked ; we have a greater separation of the oral [v') and anal (v) swell-
ings of the vibratile chord, little by little changed into two independent
breastplates, the edges bound with chords of powerful vibratile cilia, be-
coming the locomotive organs of the larvce (PI. II. Figs. 20, 22, 24, 26,
28). These plastrons, at first mere crescent>shaped shields (PI. II. Figs.
20, 22, 24), extend gradually towards either extremity, become elliptical,
and then somewhat triangular. The outline of the anal shield becomes
sinuous, slight indentations point out the position of the future arms (PI.
II. Fig. 26, e e, Fig. 28, e e, e" e"') ; the rows of cilia creep gradually
round the edge of this anal shield, turn towards the mouth again, and
extend, on the dorsal side, along the whole length of the larva (PI. II.
Fig. 25); this chord of cilia makes a complete circuit, while the cilia,
extending along the edge of the oral plastron, do not meet.

The formation of these plastrons is attended with great changes in the
general outline of the larva ; the anal extremity becomes pointed, trian-
gular, with rounded edges ; the body, on each side of the oral opening,
bulges out beyond the general outline, and the oral plastron is more and
more pointed, as it separates from the rest of the larva. This change of
shape can perhaps be better appreciated when seen in profile, and by
comparing the drawings of larvae three days and six days old ; compare
PI. II. Fig. 19 with PI. II. Figs. 25 and 27 seen from opposite sides. The
great elongation of the oral extremity and the marked separation made
by the opening of the mouth between the anal and oral plastrons cannot
fail to be noticed.

COMPAEISON OF LAEY.E OF ASTEEACANTHION. 19

Comparison of Larvae of Asteracanthion paUidus and A. hcri/limis. — Up to
this time all the larva3 described were raised by artificial fecundation
from eggs taken out of the ovaries of Asteracanthion berylinus Ag.
When I first discovered the larvse of our Starfishes, I immediately
examined the ovaries of our two most common species, the A. bery-
linus Ag. and A. pallid us Ag. I found that the eggs of the former
were not sufficiently advanced to be fecundated, while those of the second
species (A. pallidus) had all escaped. I am, therefore, positively certain
that all the larvte I am about to describe belong to the second species,
as they were all found swimming about previous to the time of spawn-
ing of the A. berylinus. As the interval between the time of spawning
of these two species is not less than three weeks, I had been able, during
this period, to make a general sketch of the whole development, from the
3'oungest larva found (PL III. Fig. 1), to the time when the Starfish is
formed, before beginning the artificial fecundation of the species just
described, the A. berylinus Ag.

I thus obtained a general knowledge of the changes these larvae un-
dergo, and was enabled, when making the artificial fecundation, to pay
special attention to the develojjment of those parts, the origin of which
was not easily traced in older larvas. I was able in this way to carry
on, at the same time, the comparative study of the development of two
closely allied species, belonging, undoubtedly, to one and the same genus,
and to see how far differences could already be noticed in their early
stages of growth ; a glance at the figures of the young of one species (A.
pallidus Ag.) on Plate III., compared with the figures of Plate II. of the
second species (A. berylinus Ag.), will show how far the development- of
allied species diverges. What is particularly characteristic is the fact that
specific differences make their appearance so early. Soon after it became
evident that the embryos we were studying belonged to Echinoderms, it
was apparent that they were different species. The order of appearance
of the characters of the classes, the orders, the families, and genera, is one
of the greatest importance in a zoological point of view; and we owe to
Professor Agassiz to have pointed out, that the characters which make
their appearance first are by no means those which have been usually
supposed to take precedence ; in the present case we do not find it
possible to discern the class, the ordinal, the family, the generic and the
specific characters, in the order in Avhicli they are here mentioned. On

20 EMBRYOLOGY OF THE STARFISH.

the contrary, the specific characters are early stamped upon the embryo,
and did we but know how to recognize individual differences among the
lower animals as well as we can already in some of the Fishes, we might
find that with Echinoderms, as has been shown for Fishes by Professor
Agassiz, the stamp of individuality is very early impressed upon the
embryo. Long before we can tell that a young Perch belongs to the
genus Ctenolabrus, we can already say with certainty whether it will be
colored red or gray or brown or green.

The time of spawning of Starfishes is very short, as, three or four days
after the A. berylinus began to spawn, it was quite difficult to find females
with eggs ; and a week after the beginning of the spawning, I never suc-
ceeded in finding a single one. Owing to this great difference in the
time of spawning and its short duration, there can be no doubt, from the
date at which I first caught the Starfish larvae floating about, to which of
our two species they belong. A careful comparison of the youngest speci-
mens also shows very striking differences, and Avill always enable an
observer to distinguish readily the larva3 of the two species, even in their
earliest stages. Compare PI. III. Fiffs. 1, 2, 3, 4, 5, with Figs. 22-28 of
Plate 11. These differences become more marked as they grow older, as
will be seen when we describe adult larvte. In fact, the larva of A. bery-
linus is pear-shaped, with the thick end at the oral extremity, while in
the larva of A. pallidus the thick end of the equally pear-shaped, but
relatively shorter body is at the anal extremity.*

The principal points of difference in the young larvae of this second
species (the A. pallidus), from those previously described, are differences
of -proportions. The larvae of the A. berylinus are elongated cylindrical ;
the oral extremity is somewhat broader and more prominent than the
anal. The larva of A. pallidus can at once be recognized by its short-
ness ; the small size of the oral extremity, when compared to the anal,
the latter being by far the most prominent,

• Watcr-System. — Before going on with the description of more advanced

* Thougli we now consider the further progress of development of our lars'se in a different species
from the first, we proceed without interruption, as the phenomena of growth are identical in both ;
and we link them liere together only because our most complete observations for the younger stages
relate to A. berylinus, and to A. pallidus for the older stages. Had we presented these changes for
a single species only, the one would have been defective in the beginning, the other in the end. As
it is, our history is tolerably complete, the course and nature of the changes being identical in both
species.

CHANGES OF FORM OF THE LARVA. 21

stages, I will take up the development of the water-tubes at the point to
which we had traced them (PI. II. Fig. 28) in the larvae of A. berylinus.
After the ends of the water-tubes have extended beyond the oral opening
(PL III. Fig. 4), the tubes increase rapidly in diameter (PI. III. Figs. 6, 8,
w, w'), bending at the same time towards the longitudinal axis (PI. III.
Figs. 4, 5, 6, 8, 10, iv, iv), the other extremity of the tubes creeping round
the stomach until they touch, but without uniting (PI. III. Figs. 6, 8, 10,
IV, tv). The tubes at the oral extremity bend towards each other (PI. III.
Fig. 4), come in contact (PI. III. Fig. 6), and, soon after, a communication
is made, the water-system assuming the shape of an elliptical ring (PI. III.
Fig. 6, ivw') ; and the water which enters into the right tube through the
dorsal pore (PI. III. Figs. 2, 5, 7, b) passes into the other branch on the
opposite side of the stomach, through the fork at the oral extremity, and
not round the stomach, where the water-tubes simply touch, but do not
communicate. The small tube leading from the dorsal pore to the main
branch of the water-system widens and becomes funnel-shaped as it ap-
proaches the main tube. The dorsal pox-e is cut obliquely across the end
of this small tube, giving it an elliptical shape. By the time the two
branches of the water-system have joined (PI. III. Fig. 6) at the oral ex-
tremity of the larva, it has assumed an entirely different outline from
au}^ we have met with in the former species. The anal extremity is
very much flattened, the corners of the anal plastron project slightly be-
yond the general outline, the indentations have become very distinct,
the oral plastron has grown rectangular with rounded angles and concave
sides, the oral triangular opening leads into a deep pouch. The sides of
the body are marked by three strong indentations (PI. III. Fig. 8). The
oral extremity of the water-system changes rapidly from a rounded to a
pointed outline (PI. III. Fig. 8, ivtv) ; it advances more and more towards
the oral extremity. In proportion as the dorsal region projects beyond
the oral plastron, the water-system extends into this projection, sending
off, at the same time, two branches leading into small appendages (PI.
III. Figs. 10, 11,/,/), (only developed in more advanced larvce), which
have, in the adult larvae, a peculiar structure (PI. IV^. Figs. 4, 5, 6).

Changes of Form of the Larva. — The prominent changes now going on
are only changes of degree. The larva has completely lost its cylindrical
shape, and even the pear-shaped form it assumed afterwards ; it has be-
come rectangular, with deep indentations, gradually assuming the char-

22 EMBEYOLOGY OF THE STAEFISH.

acter of short arms. The transformation from the pear-shaped (PI. III.
Fig. 1) to the rectangular flattened larva, with undulating outline (PI. III.
Fiy. 6), can readily be traced by comparing the successive stages here
represented. After the digestive cavity of the younger embryo (PI. 11.
Fig. 7) is bent at the extremities, bringing the mouth and the anus on
the same side of the larva, the anal and oral extremities increase rapidly
in bulk, and the larva, when seen from above (PI. II. Fig. 18) or in profile
(PI. II. Fig. 19), becomes somewhat dumb-bell shaped. The depression
thus formed grows deeper, especially on the lower side, at the time when
the chords of vibratile cilia make their appearance (PI. IT. Fig. 21), and
the mouth (PI. II. Fig. 21, m) is placed in the convexity of a deep curve.
As the oral and anal vibratile chords extend towards the oral extremity,
slight grooves arise (PI. II. Fig. 23), starting from the depression in which
the mouth is placed, and extending towards the oral extremity. These
grooves are gouged out from the oral extremity ; they extend but little
way towards the stomach, forming a very well-marked channel separating
the anal from the oral vibratile chord (PI. II. Figs. 25, 27). The oral is
less broad than the anal plastron ; the former retains its shield-like shape,
while the sides of the latter become somewhat undulating from the bend-
ing of the ciliary chord (PI. II. Figs. 26, 28). These slight imdulations,
as the larva grows older and more elongated, increase in size, giving it
more and more a rectangular outline (PI. II. Figs. 27, 28 ; PI. III. Figs. 3,
4). With its quadrangular shape, the larva assumes also a more flattened
chai'acter, and loses its cylindrical form, as will be readily seen on com-
paring Figs. 21 and 27, of PI. II. These slight undulations of the ciliary
choi-d are formed at points where accumulations of pigment cells have
taken place. The ciliary chord, at first simply a wavy line (PI. III. Fig.
4), soon becomes quite deeply indented by the formation of loops at these
indentations (PI. III. Fig. 6). The loops, at first, scarcely project beyond
the general outline of the larva (PI. III. Figs. 6, 7). Little by little they
increase in length (PI. III. Figs. 8, 9), extending slightly beyond the edge
of the outline, like short arms ; until, passing through somewhat older
stages (PI. III. Fig. 10), these loops are gradually transformed into larger
and larger arms (PL III. Figs. 11, 12), and finally attain the shape of the
long, slender arms of the adult Brachiolaria (PI. IV. Figs. 1, 2, 4 ; PI.
VII. Fig. 8). During the process of the formation of the arms, the cut
in which the mouth is placed becomes deeper (PI. II. Figs. 25, 27 ; PI.

DEVELOPMENT OF THE AEMS. 23

III. Figs. 2, 5, 7, 9, 12 ; PI. IV. Fig. 4), as well as the groove extending
along the sides of the larva, which runs from the median anal arms (c')
to the oral extremity, and separates the anal from the oral plastron. In
all these larvfe the ventral part of the anal and the oral plastron are
much narrower than the dorsal portion of the anal plastron. This differ-
ence is at first slight (PI. II. Figs. 26, 28 ; PI. III. Figs. 3, 4) ; it becomes
more marked with advancing age, passing through the different stages
represented in PL III. Figs. 6, 8, 10, 11) ; PI. IV. Figs. 1, 2 ; PI. VII. Fig.
8 ; and in proportion as all the ridges and edges are more prominent,
the surfoces circumscribed by them become flattened and more spreading.

Nomenclature of the Anns. — For the sake of brevity, I shall call the rudi-
mentary appendages by the names proposed for them in the adult larva?,
and shall adopt the names given by Miiller, with slight modifications, viz.
ventral side, that on which the mouth is situated; dorsal, the side on
which the watei"-pore is placed ; anal plastron, Avhat Miiller has called
"anales Bauchfeld," or " hinteres Bauchfeld"; oral plastron, what he
calls "antorales Feld," or "vorderes Bauchfeld"; the oral region [m) is
situated between these two plastrons- The arms are designated accord-
ing to their position by the following names : the median anal pair
[c e) ; the dorsal anal pair (e" c") ; the ventral anal pair [e" e") ; the
dorsal oral pair {e"" e"") ; the ventral oral pair (e* e") ; the odd anterior
arm (e*), from which projects, at the base, a single arm of a different char-
acter from the others ; the odd brachiolar arm (/") ; and another pair
of smaller brachiolar arms (//), connected with the oral ventral pair
(e* e^) of arms (PL III. Fig. 11). The brachiolar arms are provided at
their extremity with wart-like appendages (PL IV. Figs. 4, 5, 6 ; PL VII.
Fig. 8) ; the other arms have nothing of the sort, but are surrounded by
chords of vibratile cilia, making a complete circuit from the anal extrem-
ity round the dorsal side, while on the oral side it is not closed.

Development of the Anns. — In adult larviB the arms have, at their ex-
tremity, clusters of orange pigment cells. These colored cells make their
appearance early in the younger stages, and it is easy to trace the first
appearance of the arms by the presence of these pigment cells. Before
the appearance of the arms, the course of the chord of vibratile cilia is
very sharply defined; it is like a narrow binding extending round the
outline of the larva, seen either from above or from below (PL III. Figs.
3, 4, 6, and PL II. Figs. 26, 28). When seen in profile (PL III. v, v,

24 EMBEYOLOGY OF THE STAEFISH.

Fir/s. 2, 5, 7, and PI. II. v, v, Fkjs. 25, 27), it follows the two edges of
the deep groove which separates the dorsal from the ventral side. The
median anal arms (e' c') are the first to make their appearance (PI. III.
Figs. 2, 3, 4, 6, 7) ; these arms take the greatest development in the
adult larvjB; the other arms appear also at the same time, but as simple
bulgings of the ciliary chord. The anal ventral pair (e'" e'") and the
odd dorsal arm (e*) are both developed about the same time (PI. III.
F'kjs. 8, 9, e^) ; the odd anterior arm increasing in size, and changing its
shape more rapidly at first than the median anal pair. The next set of
arms formed is the dorsal pair (c" c") ; then follows the oral dorsal pair
(e'" e'"), and next the ventral oral pair {e" &). These develop very rap-
idly, and soon attain as large a size as the dorsal oral pair, which had
preceded them (PI. III. Ficj. 10). In this same figure we see the first
trace of a small thick arm (/"), cut ofi" square at the extremity, placed
at the base of the odd anterior arm (e"), and also a similar arm (//) at
the base of each of the ventral oral pair (e^ e*) ; the water-system branches
into this small pair of arms which are not surrounded with vibratile cilia
(PI. III. Figs. 9, 10, 11). Of the brachiolar arms, the one which is odd
precedes the two that form a pair.

The chord of vibratile cilia keeps pace with the growth of the arms,
and extends to their very extremity ; the most important change Avhich
takes place, from the time when the median arms first appear, is the
extraordinary increase of one of the diameters of the water-tubes. The
portions («', w) extending along the stomach become much flattened ;
when viewed from above (PI. III. Figs. 8, 10, 11), their great increase in
size is not seen, and it is only when examined in profile that the changes
the water-system has undergone in the vertical diameter, compared to the
transverse, can best be appreciated (PI. III. Figs. 9, 12, ?c).

It is in this condition that Miiller has seen the greatest number of his
larvse ; struck by their symmetry, he has, throughout his memoirs, in-
sisted upon the bilateral symmetry of the Echinoderm larva, as contrast-
ing directly wich the radiate structure of the adult animals. It appears
to me that this interpretation of the form of the larvae of Echinoderms is
incorrect ; they are radiate animals, and are no more bilateral than a large
number of Radiates exhibiting, as will be shown hereafter, bilateral char-
acters, such as Arachnactis, the Ctenophorai, the Spatangoids, and the
Holothurians.

developme:xt of the arms. 25

The larvfe figured on this plate (PI. III.) correspond to the larvte
observed by Van Beneden, and called by him Brachina ; the latter resem-
ble more our larvjB than any figured by Miiller. I am strongly inclined
to believe that Van Beneden's Brachina will eventually prove to be the
larvc^ of the Asteracanthion rubens 31. T., or of a closely allied species.
The more advanced specimens of his Brachina began to show signs of
the brachiolar appendages, though Van Beneden did not notice them.
See Fiff. 8 of the Plate accompanying his notice in the Bulletin de
1' Academic des Sciences de Belgique for 1850. These larvae are easily
distinguished from ours by the shortness and thickness of the arms, as
well as the less elongated shape of the larva. The time of breeding is
also different ; the European species spawning during the end of March
and beginning of April. The A. berylinus spawns in the last part of
July ; by the 26th no eggs could be found in any of the females, and
the other species (the A. pallidus) spawns during the third week in
August. These facts are additional proofs of the specific difference be-
tween our species of Asteracanthion and the Asteracanthion rubens of
Europe.

[I have retained in this memoir the specific names adopted in 1863.
At that time no description had been published of Stimpson's A. vulgaris ;
his name has subsequently been adopted by writers on American Starfishes,
although the figure given on PI. VIII., had it been baptized and described
as a new genus and species, and subsequently proved to be the young
of A. vulgaris, would have obtained jirecedence ; but failing to give it
the mythical diagnosis, this memoir was not entitled to recognition by
the strict rules of systematic zoology !

It is only comparatively recently that A. berylinus and A. arenicola
of Stimpson have both proved to be probably identical with A. Forbesii
of Desor, so that the name of pallidus would at any rate have to give
way to that of Desor.]

When seen in profile (PI. III. Fiffs. 9, 12, to, wiv' ; PI. IV. Fiff. 4, w, wiv),
the water-system runs in an arch, from the alimentary canal to the opening
of the mouth ; here the diameter increases, forming a reservoir {ivw),
from which are sent off small pouches (f'f), leading into the brachiolar
arms (//); the whole of the oral opening is placed below the water-
system. When seen from above or below (PI. III. Fi'ffs. 6, 8, 10, 11 ;
PI. IV. Fiffs. 1, 2; PI. VII. Fi(/. 8) the water-system is an elliptical ring

26 EMBRYOLOGY OF THE STARFISH.

tapering to a point in the odd brachiolar arm, enclosing the stomach and
oesophagus, which form, as it were, a soUd axis to this elliptical envelope.
On one side of the stomach appears a large hole (PI. V. Fi(). 7, h, anal
part only; PL VII. Fig. 8), the opening of a cul de sac of one branch of
the water-system passing between the stomach and the intestine. The
portions of the water-system extending along the stomach appear to be
made up of distinct chambers (PI. V. Figs. 6, 7, 8, w, vl) : these cham-
bers are merely the result of an optical delusion, arising from the greater
or less flattening of certain parts of the tube ; this gives it the appear-
ance of having been divided off into segments.

The Adult Larva. — The anal part of the larva, in its adult condition
(PI. IV. FigH. 1, 2), has become pointed ; the general shape is still some-
what rectangular; the ventral and dorsal side are separated by a deep
groove (PI. IV. Fig. 4), extending from the stomach, from the base of
the median anal pair of arms, to the base of the ventral oral arms, thus
separating the larva into very distinct dorsal and ventral regions (PI. IV.
Fig. 4), from the earliest stages of its growth. The body of the larva
itself is capable of great motion ; nothing is more common than to see
the larvae almost broken in two, by the strange habit they have of bend-
ing the oral extremity upon the opening of the mouth as a pivot, to such
an extent as to make quite an angle with the anal part (PI. III. Fig. 5).
The larvae generally assume this position when disturbed, and usually
remain stationary in the same attitude, simply striking violently up and
down with their extremities (compare Fig. 5 and Fig. 2, where the larva
is at rest). The whole substance of the body is tinged with yellow, and
is made up of large transparent cells with irregular nuclei, giving the
mass about the consistency of a Salpa; very minute granular epithelial
cells cover the whole surfoce. The powerful contraction of portions of
the body is simply that of the cells themselves, and what has frequently
been mistaken by Miiller, when describing these larvre, for muscular stria?,
are strings of such contracted cells. The extremities of the arms are
tipped with orange, the stomach and the alimentary canal are of a slight
yellowish-brown color, the chords of vibratile cilia are somewhat darker.
The oesophagus is perfectly transparent, capable of violent movements;
it expands and contracts by sudden jerks, forcing open violently the
passage leading into the stomach, when the contents of the oesophagus
rush in, and are set slowly rotating in the stomach. The interior surface

MOTION AND HABITS OF THE LAEVJ5. 27

of the oesophagus is covered with vibratile cilia, so closely crowded that
the walls appear striated from the regularity of these rows (PI. IV. Fig. 1 ;
PI. VII. Fig. 8) ; they are particularly powerful round the opening of the
mouth.

The rejection of the digested food takes place quietly, and there are
none of those violent jerks attending its admission into the digestive
cavity. The anal opening simply expands, and the fecal matter is forced
out slowly, in a constant stream, until the whole of the contents of the
alimentary canal, which had become very much distended before the
operation, has been cleaned out.

Motion and Habits of the Larvce. — The adult larvfe move about rapidly
by means of the cilia; their natural position is more constant than when
young. The oral extremity is kept in advance while in motion, and
the larva still rotates about a longitudinal axis, though not frequently ;
it generally moves with either the dorsal or ventral side uppermost, and
quite frequently in such a Avay as to show the lateral groove.* When
at rest, the larvae invariably assume one and the same position ; the
anal extremity is the lowest, and the oral extremity inclined to the ver-
tical ; in this attitude they often remain a long time, drifting about with
the currents ; their only movements being the expansion and contraction
of the oesophagus and the play of the arms. The movements of the
arms are exceedingly graceful ; comparatively longer and more slender
than the tentacles of the Tubularians, they have none of the stiffness of
their movements, the constant curving and thrusting in every direction

* The position in which the larvse figured in this memoir have been placed requu-es a short ex-
planation. To be able to compare readily the different stages, it is necessary to have them all in
the same position, and this should, if possible, be the natural attitude. But in the younger stages
of the larva the body of the embryo is not loaded down at one extremity by the young Starfish,
thus compelling the larva to assume always one and the same general attitude when in motion. It is
more common, in the younger stages, to see the embryo moving with the anal extremity uppermost ;
it would be as unnatural to turn these younger stages upside down, as it would be to represent an
adult larva in anything but its natural attitude (PI. VII. Fig. 8) with the anal extremity downward.
I have therefore compromised, by representing all the stages in the same position in which they are
generally represented by Miiller, to facilitate the comparison with his figures, and have given one
figure of an adult Brachiolaria, in its natural attitude (PI. VII. Fig. 8), with which the others can be
readily compared in their theoretical position. The figures here given are drawn from the larvte as
they appear swimming through the water; and I have endeavored, as much as possible, in repre-
senting them, to give an accurate idea of the mobility of the arms ; avoiding, in this way, the un-
natural stiffness which characterizes drawings made under compression, like the majority of those of
Miiller.

5

28 EMBEYOLOGY OF THE STARFISH.

reminding us rather of the motions of the tentacles of Phyllodoce and
similar Annelids. They are never at rest, being always kept in motion
to produce currents round the mouth of the larvie ; and, in addition to
the action of the powerful vibratile cilia placed round the mouth, are
continually bringing fresh water into the oesophagus.

The large triangular mouth (PI. IV. Figs. 1, 2, 4, m ; PI. VII. Fig. 8)
opens into a rectangular pouch (PI. IV. Fig. 4, m, m"), extending back
from its posterior edge ; from this pouch the oesophagus tapers rapidly,
and attains, near the apex of the mouth, the size (o) which it retains till
it joins the stomach. The surface of the oesophagus (o) presents a more
or less corrugated appearance near its junction with the digestive cavity,
owing to the somewhat greater thickness of the walls (PI. IV. Fig. 1).

Brachiolar Arms. — The brachiolar arms (//,/") are appendages belong-
ing only to adult larvte. Our larva has three of them (PI. IV. Figs. 1, 4,
5, 6), one pair (//), and a somewhat larger odd arm (/"), placed at the
base of the odd anterior arm (e") ; the branches of the water-system ter-
minating in these arms proceed from a large pouch {wiv) in the oral ex-
tremity (PI. IV. Fig. 4). The brachiolar arms are, like the others, tipped
with orange, but have, in addition, wart-like terminal appendages, each
having six to eight nipples, according to the age of the larva (PI. IV.
Figs. 4, 5, 6, 8 ; PI. VII. Fig. 8). These knobs give to the short arms
the appearance of the hind feet of Sphinx larvae. In the hollow between
the base of the brachiolar arms there is a small elliptical disk (/'", PI.
IV. Figs. 4, 5, 6 ; PI. VII. Fig. 8), reminding us of the madreporic body
of a Starfish, and a row of similar disk.s, two or three on each side of the
odd brachiolar arm, the pair of small brachiolar arms having no such ap-
pendages. It has been found convenient to retain for these jjeculiar arms
the name of brachiolar, used by Miiller to distinguish one of his genera
(Brachiolaria) of Echinoderm larvas. I have not succeeded in ascertain-
ing the functions of the disks ; the terminal buttons imdoubtedly are used
in the last stages of growth of the larva as supports, by means of which
they can attach themselves, while the young Starfish is resorbing the
larva ; for during that process the larvfe never float about, but invari-
ably sink to the bottom of the jar in which they are kept, and remain
attached, apparently by means of the brachiolar arms, during the resorp-
tion of the larval appendages.

These larvje are found floating in large numbers at night near the

BEACHIOLAR ARMS. 29

surface, among cast-ofF skins of barnacles, -which furnish them -with food
during the time Avhen they swim freely about, in company with multi-
tudes of small Crustacea, Annelids, and H3'droids. They seem to be noc-
turnal, as I have only found here and there single specimens when fishing
for them under exactly the same circumstances of tide and wind during
the daytime.

CHAPTER SECOND.

HISTORY OF THE DEVELOPMENT OF THE STARFISH PROPER.

We have thus far described the changes the embryo undergoes from
the time it leaves the egg, and have traced its gradual transformation
into the complicated being called Brachiolai'ia. All the phases through
which the embryo passes thus far have not the least resemblance to a
Starfish, nor have we yet alluded to any of the changes which must take
place to produce the Echinoderm proper. However wonderful the process
by which an animal seems to pass from a radiate form to an apparently
bilateral one be, the changes we shall now see taking place, by which
this seeming bilateral animal is again reduced to a strictly radiate struc-
ture, are perhaps still more remarkable.

For the development of the Starfish itself, we must turn back and ex-
amine the larva in some of its younger stages, in order to trace the first
changes in its anal extremity. There alone transformations take place
affecting the development of the Echinoderm proper, until the whole of
the complicated framework upon which the Starfish is fastened has dis-
appeared, being resorbed by the very Echinoderm it has helped to
raise. The Brachiolaria is completely drawn into the body of the young
Starfish, before it leads an independent existence. This is contrary
to the observations of Miiller and of Koren and Danielssen respecting
Bipinnaria asterigera; where it is said that the Starfish and the Bipin-
naria separate, both becoming free. [Metschnikoff's and my own observa-
tions on this point seem to throw doubt on the separation of the Pluteus
and of the Echinoderm, so that renewed observations are necessary regard-
ing the Bipinnaria of Koren and Danielssen to establish the fact which thus
far is contrary to all the observations of Miiller, MetschnikofF, and myself
on the Starfish Pluteus, and on the other orders of Echinoderms.] The
process by which the young Starfish eventually resorbs the Brachiolaria
(PI. IV. Figs. 7, 8, 9) is similar to that observed by Sars in the develop-

riRST APPEAEAyCE OF THE STAEFISH. 3]

ment of Echinaster. -where the whole Lirva and all its appendages are
gradually drawn into the bod}-, and appropriated during the growth of
the young Stai-fish.

It has alread}' been shown that the anal portions of the water-system,
as they increase in size, spread little by little over the surface of the
stomach ; the edges creeping towards each other and surrounding the
stomach on both sides, like a cap, yet without miiting. The funnel leading
from the dorsal pore shortens as the water-system extends towards the
dorsal region, and the anal extremities of the water-tubes come so near
together (PI. V. Figs. 1, 2, 3, 5, w, ?//) that we might almost be tempted
to believe the}- join, like the oral portions, and thus form a complete
circuit (PI. m. Fig. 10) ; this, however, is not the case, as an examina-
tion in profile of the above figui-es readily shows.

First Appearance of the Staijish. — In the drawings here given to illus-
trate the development of the Starfish, only a small portion of the Brachio-
laria is figured, that which has direct reference to the Starfish itself; as
this part is limited to the anal extremity of the larva immediately sur-
rounding the stomach, the anal extremity alone of the Brachiolaria is
drawn, with the arms cut ofi", somewhat beyond the opening of the anus.
To make the references to the figures of Plate V. more satisfiictory, a
reference has also been made to a drawing of a whole Brachiolaria, in
a stage of growth nearly identical, in order to show more readily the
relation of the Starfish to the whole framework of the Brachiolaria.
Tliese stages are so similar that, with this explanation, it will always be
possible to refer the anal extremities, upon which we are tracing the
development of the Starfish iu its diflerent phases of growth, to some
figure of Brachiolaria, very nearly representing its actual condition. The
stages of development figured in Plate V. have been selected simply
for the sake of the young Starfish, without reference to the Brachiolaria,
and would, if drawn on the same scale as the other figures of the
Brachiolaria here given, show no differences, which would make the mode
of growth of the vouno- Echinoderm more intelliorible. For instance, the
earlier stages of the development, such as Figs. 1-7, correspond to the
stage of PL III. Fig. 10 ; while the more advanced Fig. 8 corresponds to
that of PI. in. Fig. 11, and the others to the adult stages of the Brachio-
laria on Plate IV., when the Starfish undergoes extensive changes, while
none take place in the general appearance of the Brachiolaria.

32 EMBRYOLOGY OF THE STAEFISH.

Up to the stage of the larva represented on PI. III. Figs. 6, 7, the out-
line of the kji water-title (left when seen from above in its natural attitude),
in a profile view, is that of a flattened cylinder (PI. V. Fig. 1, w'), Avith
the end slightly bent towards the anal opening. Near the point where
the upper line of the water-tube bends downwards, a marked indenta-
tion is formed, having in the centre a slight projection. There appear,
soon after starting from the anal edge of this depression, five very
faintly defined folds, the first trace of the future ambulacral sj^stem, ex-
tending obliquely across the water-tube (?y') (PI. V. Fig. 2, t ; PI. III.
Fig. 8). If we examine the other side of the anal extremity, we find
deposited opposite the angles of these folds (PI. V. Fig. 2, /), five rods
of limestone ; the anal jiart of the larva having at the same time lost
its former transparency, and assumed a dull yellow color. These two parts
ai'e the first traces of the future Starfish. The limestone rods, and the
whole of the granular surface covering the rigU water4ube, with the dorsal
pore, forms eventually the abactinal area of the adult Starfish ; while the
folds, running obliquely across the left irater-tiibe, are the first rudiments
of the future rows of suckers extending along the lower side of the fu-
ture rays ; the rods are placed exactly opposite what will hereafter be
the extremity of the rays.

It is apparent, from the above description, that the abactinal area (the
rods), and the suckers (the folds across the water-tube), are not situated
in one plane, or even in parallel planes. The arc containing the rods
and the arc passing through the folds make an acute, nearly a right
angle, as is better understood by referring to older stages. It will also
be seen, by a glance at the drawings (PI. V. FiffS. 1, 2, 3, 5; PI. III.
Figs. 7 - 10, t), that the folds denoting the place where the suckers will
make their appearance, and the rods (/, r") marking the position of the
future rays, are neither of them closed curves, but are always open, forming a
sort of twisted crescent-shaped arc. When describing the young Starfish
immediately after it has resorbed the larva, and is ready to crawl about
by means of its suckers, I shall show how these curves become closed ;
and also point out the changes these parts undergo to form diverging
rays, as well as the manner in which the warped surfaces developing the
actinal and abactinal regions are brought into parallel planes.

Relative Position of the Actinal and Abactinal Areas. — The folds of the water-
tube {w'), which forms the actinal area, are not contained in one plane,

EELATIVE POSITION OF THE AEEAS. 33

but are placed upon a spiral; the same is the case with the five lime-
stone rods situated on the surface of the other water-tube (w), which
forms the abactinal region. When we look at the Brachiolaria from the
side, that is, when facing the groove which separates the ventral from
the dorsal side, as in PI. IV. Fig. 4, or in the corresponding profiles, from
the side of the right and left water-tubes of PI. V. Figs. 1, 2, 3, 5, 10,
11, 12, we see either the actinal or abactinal side of the Starfish. We
look in one case at the water-tube (to) upon which is developed the
abactinal system; while in the other profile, drawn from the opposite
side, we see the water-tube {w') which develops the actinal system ; the
two water-tubes are placed on different sides of the stomach, and have
no connection whatever at this extremity, but are separated by the Avhole
diameter of the stomach, over parts of which these tubes have spread
like a cap. It will at once be noticed that, in any of these figures, each
side of the future Starfish makes an independent open curve ; these curves
form what appears to us, when seen from the profile view, part of a cir-
cular arc. On looking, however, at the same sides from the ventral or
dorsal view of the larvse, as in PI. IV. Figs. 1, 2, or the corresponding views
of PI. V. Figs. 4, 6, 7, 8, 9, 13, 14, we do not see the arc formed by these
sides projected as a simple straight line, as it would be were it all con-
tained in one plane. The extremities of the arc, both of the actinal and
abactinal area, — that is, the two ends of it which are nearest, one to the
water-pore, and the other to the anus, — are seen, as in PI. V. Figs. 1,
2, 3, 5, 10, 11, 12, one on one side of an axis passing through the centre
of symmetry of the Brachiolaria, and the other on the other side. The
only curve which fulfils the conditions of such a projection is that of a
Avarped spiral, so that, in reality, when passing (in PI. V. Fig. 10) from
r"i, along the edge of the disk, to r'^, r"^, r'i', r"^, we do not move in a
plane, but are constantly winding, somewhat as when ascending a spiral
staircase; this is seen in PI. V. Fig. 9, when passing from r'^, the arm
placed nearest the anus, along the edge of the abactinal area, to r'i, the
arm next to the water-pore (b). It is the same for the actinal arc,
Avhich forms a spiral identical to that of the abactinal area, only bent in
the opposite direction.

The actinal and abactinal regions are, in reality, two Avarpod spiral
surfaces, making an angle with one another, separated by the whole
width of the stomach. This is best seen in a view from the dorsal or

34 EMBEYOLOGY OF THE STAEFISH.

oral side (PI. YII. Fig. 8), when the folds are distinctly visible one above
the other, but so arranged as to be all seen at the same time (PI. V.
Figs. 4, 6, 7, 8; PI. HI. Figs. 8, 10, 11). Three of the folds are near the
edge, while the other two are placed close to the digestive cavity on the
ventral side. This spiral, seen from the dorsal or from the ventral side,
has all the appearance of the foot of a bivalve {t, PI. V. Figs. 4, 6, 8).
The spii'al position of the five rods indicating the position of the future
rays of the Starfish {r'i' -r"^) is also apparent from the same point of
view. Two of the rods are placed on the dorsal side of the larvae, run-
ning somewhat obliquely (r'i', r"^), the three others {r"z, r"l, r"^) turning
away still more from the median line ; the last (r's') placed very near
the edge, on the ventral side, close to the base of the median arms
(PL V. Figs. 3, 5, 6, 8, 9, r'-l-r'l'); the nearest distance between these
two spiral surfaces being fully as great as the width of the water-tube :
in fact, it seems as if the rudimentary tentacles and tlie dorsal system
had as yet no connection whatever with one another (PL V. Figs. 6, 8).

It is very important that this oblique position of the actinal and abac-
tinal areas, as Avell as their great distance apart, should be distinctly
kept in mind ; as it will explain many of the errors committed by pre-
vious writers on this subject, and greatly assist us in correctly understand-
ing many points in the anatomy of Echinoderms hitherto unexplained.

From what has been shown thus far, it is self-evident that the water-
tubes, the problematic bodies, as MUller has called them in their early
condition, are the surfaces from which the future Starfishes are developed,
and not the surface of the stomach. The spiral of tentacles is developed
by folds placed on one side of the stomach (PL III. Figs. 6, 8, 10, 11), on
one of the water-tubes {to'), that with the water-pore [b) ; while round
the other water-tube {w), placed on the other side of the stomach, is
formed the spiral surface of the abactinal system. The stomach has re-
mained as it was before, and has in no way contributed to the foi'mation
of the young Starfish. A glance at any figure of the larvfB, either in pro-
file or from above or from below, will show that no change has taken
place in the shape of the stomach, or any part of the alimentary canal,
as Miiller believed (PL V. Figs. 1, 8; PL III. Figs. 1-11), but that a
kind of cap has been formed lound it by the water-tubes. Owing, how-
ever, to the accumulation of very fine granules of limestone, the anal
extremity has by this time lost its transparency ; this would be easily

FORMATION OF THE AIMBULACEAL SYSTEM. 35

mistaken for an encroachment on the stomach itself. In proportion as
the abactinal region becomes sohdified (PI. III. Fi<j. 11; PI. IV. Figs. 1,
2; PI. VII. Fig. 8), the stomach loses its globular shape, and becomes
from this time forward flattened and pear-shaped. Previously to the for-
mation of the Starfish on the surfoce of the two water-tubes, placed on
opposite sides of the stomach, we could trace no change of form in the
stomach itself From the time, however, when the Starfish encroaches
little by little upon the anal extremity of the larva^, it pushes the stom-
ach and the intestine slightly to one side, owing to the great increase in
bulk of its actinal and abactinal areas. The anal portion of the water-
tubes now swells and contracts in such a way as apparently to divide that
portion of the water-tubes into chambers ; but, on watching the circula-
tion of the fluid in the water-tubes for any length of time, the currents
can be followed flowing from one of these elliptic chambers to the other,
plainly showing the different planes in which the ventral and dorsal part
of the tubes are placed to be the only cause of this delusion.

Miiller has distinctly stated, over and over again, during the course of
his investigations, that the young Echinoderm was formed by encroaching
upon the stomach itself; I am satisfied, from repeated observations of this
point, in Starfish, Sea-urchin, and Ophiuran larvfe, that this is not the
case. The mistake arises from the fact that the water-tubes, by their
extension and increase, cover and conceal part of the stomach, forming a
sort of hood over it ; while the two sides of the young Echinoderm,
separated by the whole width of the stomach and the thickness of the
two water-tubes, form upon the outer surface of the latter, and do not in
an}' way encroach upon the stomach, which is simply enclosed by the
actinal ;ind abactinal areas of the Echinoderm. Had I not traced this
with the greatest care, I should scarcely venture to doubt the statements
of Miiller, but I am satisfied that he was mistaken in this explanation of
the mode of the formation of the Echinoderm.*

Formation of the Ambulacral System. — We have already seen that the
very first changes which take place in the water-system {w, iv) consist
of the five folds (t, PL V. Fig. 2) extending obliquely across the exterior
surface of one of the water-tubes («''). From the fiict that these folds

* It may not be out of place to sav, that Professor Agassiz, during this investigation, satisfied him-
self of the accuracy of every point which seemed in the least contradictory to the statements of
Miiller.

36 EMBEYOLOGY OF THE STARFISH.

develop across the surface of an elliptical tube, the five folds naturally
form a twisted spiral, with a pentagonal outline, each side of this spiral
forming the first nucleus of the five ambulacral tubes. I speak con-
stantly of pentagonal spirals, pentagonal ambulacral system, and pentago-
nal abactinal system. In using these terms, I do not mean a pentagon
with five equal sides, the adjacent sides making equal angles with one
another and surrounding a closed surface, but simply that we have five
sides limiting an open space, the two extremities of this five-sided figure
being separated by the whole vertical diameter of the water-tubes. One
extremity of the ambulacral five-sided open figure is placed at the water-
pore {b, PL v. Fig. 2), the other at the opposite side of the water-tube
on the surface of which the ambulacral system is developed. The two ex-
tremities of the abactinal open five-sided figure are placed, one above the
water-pore {b, PL V. Figs. 8, 9, 13, ?•'['), the other on the opposite side of the
water-tube, which develops the abactinal surface on one side of the anus
(a, PL V. Fig. 14, ?•'•'). A glance at the figures of the Brachiolaria from the
dorsal or ventral side (PL IV. Fig.s. 1, 2 ; PL VII. Fig. 8 ; PL V. Fiffs. 8,
9, 13, 14) shows that the two surfaces, upon which the actinal and abac-
tinal areas are developed, do not correspond to one another, or fit into
each other as in the full-grown Starfish. That is, if the ambulacral system
were projected upon the abactinal system, in order to bring these two
surfaces into the same relation which they hold in the adult Asteracan-
thion, we should find the ambulacral system projecting beyond the out-
line of the abactinal system, and placed nearer the mouth of the Brach-
iolaria, while a portion of the abactinal system — that which is placed at
the anal extremity of the larva — would, in the same way, project be-
yond the outline of the ambulacral system.

The sides of this twisted pentagonal spiral are somewhat concave, and
the apex of the angles of adjacent sides are rounded. It is in conse-
quence of the changes taking place at the apex of the sides of this irregu-
lar ambulacral pentagon that we have the simple apex transformed, by
its gradual extension beyond the general outline of the open pentagon,
into the five-folded loops (PL V. Figs. 10, 12), each of which corresponds
to an ambulacral tube and its accompanying suckers in an adult Star-
fish.

The ambulacral pentagon with concave sides and rounded angles, seen
in profile (PL V. Figs. 2, 5 ; PL III. Figs. 7, 9, t), changes its shape

FOEMATION OF THE ABACTINAL SYSTEM. 37

rapidly ; the convex cavity becomes greater, the apex of each angle of
the pentagon more prominent and less pointed, a double line is formed
by the ruffling of their folds (PI. III. Fig. 11), and each apex of the pen-
tagon has the appearance of a small loop projecting beyond the curved
sides ; the loops grow larger and larger, until they have reached a size
somewhat less than one third of the diameter of the water-tube, when
they stand out freely from the pentagon, and seem to form no part of
the water-tube (PI. V. Figs. 10, 11, 12, t ; PI. IV. Fig. 4). When seen
either from above or from below, the folds appear as small flaps on the
broad side of the fooUike appendage projecting from the surface of the
stomach, formed by the folding of the water-tube (PI. V, Figs. 4, 6, 8, 9,
1.3, 14, t ; PI. III. Firjs. 10, 11 ; PI. IV. Fujs. 1, 2 ; PI. VII. Fig. 8). These
small folds are, in reality, nothing but open bags communicating with
the main water-tube {/'/) ; small pouches leading from it. The outer and
inner fold of each loop do not remain concentric, and we can soon trace,
in the inner fold, changes similar to the first folding of the water-tube.
The rounded end of the inner fold becomes triangular ; this is the first
indication of the formation of the separate suckers (PI. V. Fir/s. 10, 11,
12, t t t t t). The ambulacral pentagon remains in this state until the
Starfish has resorbed the many appendages of the larva.

Formation of the Abactiiial Systan. — Let us now follow the corresponding
changes of the abactinal system, accompanying the modifications, just de-
scribed, of the ambulacral pentagon. On examining the anal extremity,
at the time when the larva has reached the state represented on PI.
m. Fig. 10, we are at once struck with the fact that the outline of the
abactinal system has undergone analoo-ous changes to those of the actinal
pentagon. Instead of remaining a miiform spiral, the two ends of Avhich
are separated by the whole height of the water-tube, while the two areas
are divided by the combined width of the stomach and the two water-
tubes, it has a slightly lobed pentagonal outline, the convexities corre-
sponding to the apex of the pentagon of suckers (PI. V. Fig. b, r"^-r"' ; PI.
III. Fig. 10). The rods, simple at first (;•', PI. V. Fig. 2), have increased
in size ; small Y-shaped appendages have developed at their extremities.
We also see that in the intermediate spaces, corresponding to the con-
ca:vities of the lobes of the actinal system, a second set of small rods (r",
PI. \. Fig. 5), of a similar character to the large ones, have developed.
The whole of the abactinal system has become coated with a very fine

38 EMBRYOLOGY OF THE STARFISH.

granular deposit of limestone ; and the edge of the surface, connecting
the two extremities of the abactinal pentagon, can readily be seen in
profile (PI. V. Fici. 5). The five large rods placed in the middle of the
sides of the spiral abactinal pentagon, and the five small ones placed in
the angles of this same pentagon, are the first trace of the plates com-
posing the abactinal surface of the young Starfish.

The water-pore {h, PI. III. Fig. 10 ; b, PI. V. Fir/s. 7, 8) remains open,
the only change being an accumulation of limestone matter round the
opening, forming a sort of solid tube to protect it. This water-pore, as
we shall see hereafter, eventually becomes the madreporic body ; and
the canal formed by the deposition of limestone is the stone canal of the
full-grown Starfishes.

Abactinal System. — The double line on the edge of the abactinal pen-
tagon (PI. V. Fig. 2) is formed by the thickness of the surface of the
abactinal system. This double line, at first only slightly vnidulating, be-
comes gradually more indented (PI. V. Figs. 3, 5) ; at the same time,
additional rods arise round the primary ones with such rapidity that we
soon find a complicated network of limestone rods, forming ten clusters
(PI. V. Figs. 8, 9, 13, r, ?•"), five large (?•') and five smaller ones [r") round
the original rods. This network is produced by the addition of a Y-
shaped rod, at each extremity of a simple primary rod ; presently, eight
Y-rods arise upon the shanks of the first set of Y-rods, followed by a
third set upon the shanks of the second set, and so on ; in this manner
are formed the closed polygons composing the clusters of the patches of
limestone deposit (PI. V. Fig. 9, r, r"). The small granular cells, filling
the larger meshes of the network, increase in number, rendering the
whole abactinal system somewhat opaque ; Avhen the larva is seen in
profile from the abactinal side, the outline of the stomach (PI. V. Fig.
5) can be traced exactly as it was before the Starfish had begun to form ;
and outside of it, the edge of the future back is distinctly visible (PI. V.
Fig. 5).

As the two water-tubes are placed on opposite sides of the larva, it fol-
lows that when seen in profile (PL V. Figs. 11, 12), from the left or from
the right, it presents, in the one case, a full view of the tentacular pen-
tagon (t), and only the lower oral edge of the abactinal system, the net-
work of limestone meshes being quite indistinct, as seen through the
thickness of the abactinal surface (PI. 111. Fig. 7 ; PI. IV. Fig. 4 ; PI. V.

FORMATION OF THE SPINES. 39

Figs. 10, 12) ; while, in the other case, a full view of the abactinal pen-
tagon (PI. ni. Fig. 10; PI. IV. Fig. 4; PI. V. Figs. 5, 11) is obtained, and
the arrangement of the different rods forming the plates of the limestone
network is distinctly seen. A view of the larva from the dorsal side
(PI. III. Fig. 11; PI. VII. Fig. 8; PI. V. Figs. 8, 9, 13) shows the abac-
tinal system extending in such a way as to surround the stomach en-
tirely on one side, while the tentacular pentagon covers it on the oppo-
site side. This attitude gives us the position of the lobes {r"-r"l), the
future rays of the Starfish, next to the water-pore (r'i', r'a), while a view
from the oral side (PI. V. Figs. 4, 14) indicates the trend of the lobes
on the opposite extremity of the spiral of the abactinal system {r"i, r',').

[Metschnikoff was able to trace most satisfactorily the development of
an Ophiuran in which the formation of the abactinal system was shown
to be identical in every respect with that here given of the Starfish. In
the Ophiuran of which I had previously* traced the growth the forma-
tion of the plates of the disk could not be seen. These I only traced
at that time in one of the viviparous species.]

Formation of the Rays of the future Starfish. — The plates of the abactinal
system early reach a condition when the changes they undergo are merely
quantitative, and the only modifications affecting the appearance of the
Starfish take place on the edge of the disk. A depression is formed in
the middle of the convexity of the lobes of the abactinal area; this is
soon followed by two other depressions in the middle of the small arcs
thus formed, dividing each lobe of the pentagon into four smaller lobes ;
at the same time the indentations between the original sides of the pen-
tagon have grown much deeper, separating these five lobes in a very
marked manner. We can now no longer mistake the true character of
the lobes ; they are the five rays of the Starfish, but as the actinal and
abactinal regions are not yet fitted together, as in the adult (PL V. Figs.
10, 11, r"l-r"-^; PI. IV. Fig. 4), they represent only the dorsal sides of the
rays. A glance at Fig. 9 of the same Plate (PI. V.), seen from the dorsal
side, will show how fiir the suckers {t) are removed from the abactinal
portion of the arm which is to protect them. The position of the water-
pore {!)) is immediately on the edge of the disk, at the extremity of the
dorsal end of the pentagon (PL V. Figs. 10, 13, I)).

* Mem. Am. Acad. 1864.

40 EMBRYOLOGY OF THE STARFISH.

Formation of the Spines. — Such is the state of the abactinal system
when the pentagon of tentacles is composed of simple loops ; let us now
examine this system in more advanced larvte, at the time when the inner
fold of the loops has become triangular at the extremity. When seen
from the ventral side (PI. VII. Fhj. 8), we find that the small lobes have
become wart-like projections, surrounding the whole edge of the abactinal
system (PI. V. Fig. 9). These projections are composed of accumulations
of Y-shaped rods, connected with the system of network in the larger
plates. The surface of the abactinal system has also become covered
with these wart-like projections, rendering the outline irregular. In an
abactinal profile, smaller tubercles are seen on each arm, identical, in every-
thing except size, with those of the edge ; the tubercles are young spines,
arranged in regular lines (PI. VI. Figs. 2, 4, 6) ; one row of four alternat-
ing on the edge of the abactinal system with one row of three, this again
with one of two, followed by single tubercles, forming a pentagon, placed
in the apex of adjoining rows, in the angle between two arms; the older
tubercles are those nearest the edge.

When the young Starfish has reached this state, it has the rudiments
of nearly all the external parts of the adult. I shall, therefore, apply
to these rudimentary organs the names usually given to them. The spines
are warts, not rising much above the general level of the abactinal region,
and they are arranged in regular rows. The position of the network
of limestone meshes has become well circumscribed, the plates formed by
them occupying the position of the original rods. The five smaller plates
in the angles of the arms are arranged round a central plate, the larger
plates alternate with them and occupy nearly the whole of the surface
of the arm ; this arrangement is identical with that of the plates of the
abactinal surface, as shown in PI. VI. Fig. 10, /, l^, l^. The indentations
of the rays are now so well marked (PI. V. Figs. 12, 13) that there is
quite a large open space between the outer spines on the edge of any
two adjoining arms. On examining the plates formed by the network
of limestone meshes, we see that the cells are polygonal ; they are usually
hexagonal, and are more or less quadrangular near the exterior of the
plate. The original rod can be recognized by the larger cell it has de-
veloped (PL V. Figs. 9, 13, ;•') ; <i"<^ it is from this central cell that the
others diverge, growing smaller and smaller as they approach the edge.

Tn the present stage of the young Starfish the anal extremity of the

EESOEPTIOX OF THE BRACHIOLAEIA. 41

Bi-achiolaria (PI. VII. Fig. 8) lias almost entirely disappeared, and the
embryo Starfit^h has taken its place (PL V. Figs. 9-14). This embryo
is so heavy that, when floating about, it loads down the anal part, which
is always the lowest, and the larva is compelled to move always more
or less obliquely, having to drag this great weight after it. The water-
pore remains in the position in which it w^as at first, in the angle of the
arm {i\), which opens the pentagon, and is encased in a stronger deposit
of limestone.

Resorption of the BraduoUria. — While the Starfish is growing upon the
outer surfaces of the two opposite water-tubes, and is gradually becoming
a part of the Brachiolaria, no changes take place in the external appear-
ance of the larvjB (PI. IV. Figs. 1, 2 ; PI. VII. Fig. 8). But when the
Starfish has become so far advanced as to occupy a very prominent posi-
tion at the anal extremity of the larva (PI. IV. Fig. 4; PI. VII. Fig. 8),
the complicated appendages designated as arms, which have served for the
development and for the locomotion of the Starfish, are resorbed by the
little Echinoderm.

"We now come to a most interesting period in the history of our Star-
fish. The larvne, very active up to this time, grow sluggish ; the body,
which, with the exception of the anal portion, is, in the early stages,
perfectly transparent and clear, becomes cloudy and opaque. Changes
are first visible in the side arms (PI. IV. Figs. 7, 8, 9); thej' contract,
and apparently divide into many large cells. Next in turn the anal ven-
tral arms, and, lastly, the dorsal arms, contract in the same manner. This
contraction of the arms is accompanied by a corresponding shrinking of
the anal part of the larva, beyond the mouth (PI. IV. Fig. 9j, so rapid
that in a few hours the anal arms have shrunk to quite a small compass
(PL IV. Fig. 9); the oral dorsal arms and the oral ventral arms contract
in their turn, until there remains nothing but the brachiolar arms, brought
close to the Starfish by the shrinking of the mass of the body (PL IV.
Fig. 8). They soon follow the rest, and we can actually see the gradual
disappearance of this complicated fabric. It has served its purpose of
developing and feeding the young Starfish, which has now reached a state
when, in a few hours, it will move about independently, having resorbed,
for what purpose is not known, the whole of the framework. Not a single
part is dropped off, the wliole of the larva passes into the Starfish, and, before
twelve hours have elapsed from the commencement of the first sign of

42 EMBEYOLOGY OF THE STAEFISH.

contraction of the anal tentacles, nothing is to be seen of the larval ap-
pendages, except a few indistinct swellings on the actinal side of the little
Starfish (PI. VI. Fig. 1).

The Starfish after the Eesorption of the Bipimntria. — The process of resorp-
tion, which I have frequently had the opportunity to examine and trace
in all its stages, leaves no doubt, at least in this case, that the young
Starfish does not separate from the Bi-achiolaria. We cannot, therefore,
consider the Starfish and the framework (the Brachiolaria) as two indi-
viduals, leading a separate existence at different stages of growth, but
must regard them both as one and the same thing. This is in direct
contradiction to the statements of Mliller, and of Koren and Danielssen,
with regard to the Echinoderm, the development of which they have had
occasion to watch. I must add that my own observations concerning the
development of Echinoids and of Ophiurans have led nie to an entirely
different opinion from the one they have expressed ; see my remarks on
the Embryology of Echinoderms, in the Memoirs of the American Academy
for 1864.

Closing of the Actinal and Abactinal Areas. — Although the young Starfish
has noAv resorbed all the appendages of the Brachiolaria (PI. VI. Figs.
1, 2, 3, 4, 6, 7), it is very different from the adult; the rays do not yet
make a comjjlete circuit, nor are they similar to each other; the penta-
gon of tentacles is still open, and the first step, preceding any other great
change, is the closing of the actinal and abactinal areas, by which the
two regions are brought into their proper relations. While the arms of
the larva are shrinking aAvay, the tentacular and abactinal pentagons are
drawn closer together by the contraction of the water-tube. The ex-
tremities of the two open pentagons approach each other simultaneously
by the flattening, in opposite directions, of the two pentagonal spirals,
until the surfaces are brought into parallel planes, and the space, still
separating the two ends of the pentagon (PI. VI. Fig. 4) gradually dimin-
ishes, when they finally join ; the Starfish is then in its normal con-
dition, and the circuit is completed, though the embryo is by no means
synniietrical.

[Metschnikoff' has since also shown the same thing in his development
of an Ophiuran. See 1. c. PI. IV.]

Development of the Avihilacral Tentacles of the Starfish. — While the closing
of the spiral goes on, the pentagon of the tentacular side is undergoing

AMBULACEAL TENTACLES. 43

great changes. We will follow these until the tentacles have acquired
their normal shape, and then return to the changes of the abactinal sur-
face. The points of the inner folds of the tentacular pentagon, as seen
in PI. V. Fi(is. 11, 12, t t t, become rounded, forming a rosette, dividing
each loop into five lobes. The terminal lobe in its turn goes through
the same process ; two smaller lobes are developed on each side of it
(PI. VI. Figs. 3, 5), thus dividing the original simple loop into seven lobes,
a terminal one {t'), and three pairs {t t t) arranged symmetrically on the
sides. The first^formed lobes retain their greater size until the tentacles
are well developed, which at first is always in proportion to their prox-
imity to the base of the loop. The odd lobe, from which the last pair
of tentacles was formed, does not participate in the rapid growth of the
others, and is soon outstripped by all the lobes formed along the side of
the original loop (PI. VI. Figs. 3, 5). The point at which additional
tentacles are formed is plainly seen in this early stage of growth ; a pair
is always added at the outer extremity of the arm, immediately at the
base and on the side of the odd tentacle (the eye-bearing tentacle), which
remains at the termination of the ray during the whole life of the Star-
fish. It is quite the reverse with the additional spines of the abactinal
surface of the disk; they are always formed upon the dislv, and are
pushed out upon the arms by younger spines growing up nearer the
centre of the disk. This will be plainly seen when describing more ad-
vanced conditions of the young Starfish. As the loops increase, they ex-
pand, lose their character of simple folds, and soon become quite extensive
sacs {t t t, PI. VI. Fig. 8), opening into the main tube {t"), from which
they were formed, until, finally, they attain the shape represented upon
PI. VI. Fig. 9. They soon grow long enough to be quite movable; they
contract at the base, the walls thicken towards their extremity, and they
become club-shaped. The result of this contraction is a change of the
tentacular cavity into a rudimentary radiating tube {t"), with the ten-
tacles attached to it; it also draws together the first pair of tentacles,
which are usually seen in such a way as to appear like knobs (PI. VI.
Fig. 5). This basal pair does not lengthen so rapidly as the second pair,
which in a couple of days becomes the longest (PL VI. Fig. 9). Before
the base of the radiating tube {t") has contracted, the adjacent basal ten-
tacles of adjoining loops are placed nearer together than those of the
same basal pair, the basal tentacles thus forming five pairs of tentacles

7

44 EMBRYOLOGY OF THE STAEFISH.

(PI. VI. Fi(j. 8, i t), separated by the radiating tube (/"). In proportion
as the tentacles elongate, the separation between them and the radiating
tube is more distinct, and very soon the tentacles appear like club-shaped
branches projecting from it (PI. VI. Fig. 9) ; the first pair of tentacles
are somewhat shorter and stouter than the second, which is the longest,
while the three terminal tentacles have nearly the same size, the odd
tentacle {t') not showing as yet the slightest tendency to become club-
shaped, though developed so much earlier than the larger basal pairs at
its base.

Formation of the Sucker of the Tentacles of the Starfish. — When the ten-
tacles have reached the state of PL VI. Fig. 9 they develop rapidly ; the
walls at the extremity of each tentacle thicken so much, that the cavity
becomes a pointed tube set into a somewhat conical head, which grows
more club-shaped, and projects beyond the walls of the tentacles as they
increase in length, so that, when the basal pair of tentacles equals again
in length the second pair (PI. VI. Fig. 12), the clubs at the extremities
are supported upon comparatively narrow bases. This club-shaped termi-
nation is the future disk of the tentacle, the sucker, by means of which
the Starfish adheres so firmly to rocks. From an early period, even
when there is only one large pair of tentacles at the base of the ray,
and when the others exist only in the most rudimentary condition (PI. VI.
Fig. 5), these tentacles are used by the embryo in adhering to the sur-
faces upon which it is placed ; and, though they are not provided with a
regular sucking disk, they fasten themselves so firmly, by means of these
loops, that it requires considerable force to make them loose their hold.

Formation of the Eye. — We have seen that, unlike the others, the odd
terminal tentacle does not become club-shaped, but increases slowly in
length alone, the walls retaining a uniform thickness. It is not till all
the pairs of tentacles are well developed that we begin to perceive slight
changes (PI, VIII. Fig. 5). The opening leading into the radiating canal
contracts, the basal portion of the tentacle swells, and it assumes a some-
what pear-shaped form, the swelling at the base increases, principally on
the oral side, and we soon trace in it an accumulation of pigment cells
(PI. VII. Fig. 6, e), which, by the time the other tentacles have developed
knobs, and equal in length the diameter of the arms, has become a brill-
iant carmine spot (PI. VI. Fig. 12, e; PI. VII. Fig. 6, e, and PI. VIII.
Fig. 5, e). This odd tentacle, placed at the extremity of the radiating

MOUTH OF THE STAEFISH, 45

tube, is the ocular tentacle. Ehrenberg discovered the presence of eyes
in Starfishes, but their true relations to this odd terminal tentacle was
first pointed out by Professor Agassiz, in his Homologies of Radiata.

[The nature of this terminal tentacle in the young Starfish and all
young Echinoderms seems to have been entirely overlooked by all writers
who have described the eye of the Starfishes, which they have usually
represented as an organ totally unlike any other Echinodermal ap-
pendage.

The Embryology of Echinoderms certainly shows most distinctly that
the eye of the Starfish is only a modified tentacle, an organ of sense, such
as we find at the base of the marginal tentacles of Acalephs.]

Formation of the Mouth of the Starfish. — From the manner in which the
tentacles are formed by folds of the water-tube, it is plain that, in the
younger stages of the Echinoderm, the two ends of the circular tube must
remain disconnected ; the rapid accumulation of limestone particles on
the lower surfiice prevents us, however, from ascertaining this point.
Soon after the larva has disappeared, the whole actinal surface between
the pentagon of tentacles is covered by a membrane ; this membrane, in
the centre of which is placed the mouth, is the remnant of that part of
the larva situated in the groove between the anal and oral plastrons
{m, PI. VI. Fig. 12; PI. VII. Fig. 1). The mouth of the Starfish, how-
ever, is not in reality the mouth of the larva. During the shrinking of
the larva the long oesophagus has become shortened and contracted,
bringing the opening of the mouth of the larva to the level of the open-
ing of the oesophagus, which becomes eventually the true mouth of the
Starfish.

Before the limestone particles have accumulated sufficiently to cover
the base of the radiating tubes, the mouth is movable, shifting its posi-
tion from one side to another indifferently (Pi. VI. Figs. 3, 7, 8, 12, m ;
PI. VII. Fig. 1), though by the time the deposit of limestone has formed
a small pentagon inside of the base of the radiating tubes, it has lost its
mobility. The water-pore (PL VI. Fig. 12, h\ or the madreporic body,
connects with the circular tube through a long, narrow tube, and is
placed on the actinal side in the angle between two rays; it is, as yet,
only a simple opening, protected by a thick funnel-shaped limestone pro-
jection (PI. VI. Fig. 12, I)). The young Starfish has no other anus than
that of the larva, which is placed on the very edge of the disk; but,

46 EMBEYOLOGY OF THE STARFISH.

with the rapidly increasing deposit of limestone cells, it is soon hidden
from view, and I have not been fortunate enough to find it again in
more advanced young. I am therefore unable to say where the anus opens
outside, though it undoubtedly discharges, at this time, through one of the
many limestone cells. Owing to the difficulty of tracing its ojiening in
the dfedalus of round cells, I am not able to state this positively, never
having seen, from any point, discharges of fecal matters. Like the mad-
reporic body, it is not yet upon the abactinal area, but on the actinal
side, near the edge of the disk. The madreporic body itself would
have been lost in a similar manner, had it not been possible to track it
by means of its connection with the circular tube (PI. VI. Flrj. 12) ; and,
even then, it was only by the closest attention, and at moments when
the position of the young Starfish was especially favorable for the inspec-
tion, that the opening of the madreporic body could be distinguished
from that of the surrounding limestone cells.

[With regard to the functions of the mouth of the Pluteus and its sub-
sequent fate in the young Starfish and Ophiuran, my observations as well as
those of Metschnikoff would show that it becomes the mouth in both.
This does not seem to be the case in Auricularia, and the fate of the
openings (both the anal and oral) of the Pluteus of Echinoderms is not
yet definitely known for all the orders. Additional observations are needed
on this point. Embryological studies on Mollusca would seem to favor
the formation of a new mouth distinct from that of the early stages of
the embryo, but the direct observations on Echinoderms all tend to prove
that there is no new opening formed, and that the mouth of the Pluteus
passes directly into that of the young Echinoderm.

Selenka shows for Holothuria also that the original opening of the Plu-
teus becomes the permanent anus.]

Formation of the Actinal Limestone Surface. — The actinal side of the disk
is at first a narrow flat band (PI. VI. Fig. 3), following the general out-
line of the rays. This band increases in breadth, loses its convex outline,
and soon reaches the terminal tentacle, when the actinal band has assumed
a pentagonal shape. Inside of this small pentagon is situated the ambu-
lacral system, entirely independent, as yet, from the limestone deposit on
the actinal surface, the whole rosette of tentacles expanding and contract-
ing, with perfect liberty, in every direction. This freedom soon ceases;
the points of the limestone pentagon develop ra23idly towards the centre

SPIKES OF THE YOUNG STAEFISH. 47

of the disk, and soon reach the base of the radiating canal (PI. VI. Fig. 7).
There they unite by bridging the" intei'vening spaces, and form five tri-
angular openings, enclosing the tentacles, which are still at liberty, with
the exception of this band across the base of the radiating tubes (PI, VI.
Fifj. 9). The additions made to this deposit of limestone take place more
rapidly near the bridge, where additional limestone cells are sent out,
enclosing at first the basal pair of tentacles, but leaving the remaining
five still unconfined. The next pair is then imprisoned by a similar pro-
cess, without, however, interfering with the terminal tentacles. Finally,
the last pair of tentacles is surrounded in a like manner, and all the ten-
tacles are now confined somewhat as we find them in the adult (PI. VI.
Fig. 12; PL VII. Fig. 1). A roAV of limestone cells, extending along
the median line, separates the base of the suckers, while transverse bands
join the larger cells of adjoining spaces. It is plain that the transverse
bands correspond to the ambulacral plates of the adult, and that, in the
earlier stages, the embryo Starfish has no trace whatever of any inter-
ambulacral system. This mode of formation of the ambulacral system may
explain the absence of interambulacral plates in the Crinoids and Ophi-
urans. The deposit of limestone is not sufficiently transparent to allow a
good view of the radiating canal, or of the formation of the vesicles of
the tentacles.

Formation of the Spines of the young Starfish. — We have seen that, at the
time of the closing of the young Starfish, the abactinal region is already
covered with regular rows of spines (PI. VI. Fig. 4). These spines are,
however, simple wai'ts, slight protuberances, in which limestone cells are
formed, connecting with the general network. The cells of these spines
are arranged in regular tiers one above the other; the younger cells,
formed at the base, being always more numerous, and pushing up the
older ones. All the cells send off Y-shaped appendages, which unite,
forming stories (PI. VII. Figs. 3, 4, 5) of circular cells; the cells of the
spine near the edge do not close, but project beyond the margin, giving
the spines the appearance of small Gothic spires.

The spines of the first row — viz. those immediately on the edge
of the rays — increase rapidly, curving sideways, expanding at the tip,
and assuming as fantastic shapes as those of Ehabdocidaris Orbygniana
(PI. VI. Figs. 10, 11, 12, p p). The other rows of spines, diminishing in
size as they approach the centre, are exactly similar to the former ij)^,

48 EMBRYOLOGY OF THE STARFISH.

^2)5 but not so broad at the extremity, and somewhat more slender.
New spines are always added between those originally at the extremity
of the rays and the centre of the disk ; the latter always remain the
most advanced and most prominent of the spines, even when the young
Starfish has assumed many more of the features of the adult than it has
at present, and has reached a stage when it would not be mistaken for
anything but a Starfish, closely allied to our common species.

Network of Limestone Cells. — As we have seen in the earliest stages of
the Starfish, there are, on the abactinal area, rods from which, by the
addition of Y-shaped processes, clusters of polygonal cells are gradually
formed (PI. VII. Fig. 7); one cluster in the middle of each ray (PI. VI.
Fiy. 10, 4)j one around the smaller rod placed in the angle of the rays
(4), and a still smaller one round the rod placed in the very centre of
the abactinal area {l\ The large clusters extend and unite along the
edge of the rays, forming a continuous network ; it is from the cells of
the edge that the limestone deposit is formed, which extends over the
abactinal surface. The clusters of cells placed in the angle of the rays do
not unite laterally, though they become indirectly connected in the more
advanced stages of our Starfish, joining with the plates of the rays by a
few cells (PI. VI. Fig. 10). The central plate remains unconnected with
the others in the most advanced of the young which I have raised from
the Brachiolaria. The whole of the network is quite movable, and the
plates, before they become united, are capable of independent motion by
the contraction of different portions of the abactinal area.

[Loven has given excellent figures of the young of Asterias glacialis,
corresponding to some of the stages here figured. They differ, however,
in having the plates more distinctly separated even in the young stages
(PI. VI. Fig. 10). The reticulation is compact, so that it is oidy in cer-
tain stages of expansion that the original composition of the abactinal
surface can still be traced.

It is from the careful comparison of these young stages (Pis. VI. VII.
VIII.) with the corresponding stages of the 3'oung Ophiurans, given by
Metschnikoff, PI. IV. of his memoirs, and in my memoir on the Embrj'-
ology of Echinoderms, Figs. 29, 32 (Mem. Am. Acad.), and of young
Echini with the young of Comatida figured by Allman, Carpenter, and
Thomson, that we can make out a satisfactory homology of the test of
Echinoderms as has been so successfully done b}' Loven in his superb

OUTLINE OF THE YOUNG STARFISH. 49

Memoir on Sea-Urchins,* where he has most thoroughly proved the homol-
ogy of the basal and radial plates of Crinoids with their corresponding
plates, still readily to be traced in the young Starfish, and with their
homologies in the apical system of Echini.

An admirable paper by Selenka in the Zeits. f. Wiss. Zool. (June, 1876)
gives us a complete history of the development of Holothuria, showing
an entire agreement in its general features with the Embryology of other
Echinoderms in the mode of formation of the water-system as diverticulum
for the alimentary canal, forming eventually (as in the Starfish) the cir-
cular canal with the ambulacral system.]

Change of Outline of the young Starfish. — With advancing age, the outline
of the young Starfish is greatly modified ; at first, when the actinal and
abactinal areas are not yet closed, while the larval appendages are still
visible on the lower side of the young Starfish (PI. VI. Figs. 1, 2), im-
mediately after the larval appendages have disappeared, and the surfaces
of the actinal and abactinal areas are brought nearer together (PI. VI.
Figs. 3, 4), it is hardly more than an irregular pentagon, with slightly
convex sides, and small rounded notches cut in at the angles (PI. VI.
Figs. 3, 4). These notches become deeper, the arms of the Starfish as-
sume more the appearance of a Greek cross (PI. VI. Figs. 6, 7) ; the sides
of the rays are strongly concave, and the concavity is increased with the
development of the spines to such a degree that the extremity of the
ray is almost twice as broad as its base (PI. VI. Figs. 10, 11, 12). The
outline of the inner wall of the disk can be easily seen through the lime-
stone network. The pentagonal form, so different from that of the
adult, is still less like it when seen in profile (PI. VII. Fig. 2). The abac-
tinal area rises like a high, rounded cone, supported upon the spines {p)
of the edge of the disk ; the tentacles project far beyond the edge on
every side (PI. VII. Fig. 2). In fact, the regular rows of spines, their great
size, the convexity of the disk, are features so unlike our usual concep-
tion of a Starfish that, without closer examination, one would readily mis-
take this Echinoderm, at first sight, for a young Sea-urchin, like the flat,
conical Ecliinocidaris.

The tentacles are longer than the rays, extending far beyond the edge
in front and on the sides. The pairs of tentacles move in every direc-

* Kongl. Svenska Vets. Akad. Hamll. XI. No. 7. Etudes sur les Ecliinoidees par S. Loven.
Stockholm, 1874.

50 EMBRYOLOGY OF THE STARFISH.

tion ; but the odd tentacle is always curved upward, and carried between
the two middle spines of the extremity of the rays. When we see the
Starfish in profile (PL VII. Fig. 2), the red eye-siieck appears prominent
near the edge of the diskj surmounted by the upturned tentacle {f t'),
of a slight rosy hue. This manner of carrying the terminal tentacle re-
minds us strongly of the way in which ^Eginopsis, as well as the young
of so many of our Hydroid Medusse, carry their marginal tentacles : Ne-
mopsis, Staurophora, Turritopsis, Willia.

This is the most advanced stage of the young Starfishes (PI. VI. Fig.
11) which I have succeeded in raising in confinement. When we com-
pare this with an adult, having long, slender-pointed rays, four rows of
suckers, its surface covered with pedicellariae and water-tubes, surround-
ing individual spines, like so many wreaths, we cannot foil to be struck
with the astonishing changes of form which must still take place to bring
this pentagonal star to any shape resembling a slender five-rayed Star-
fish. In foct, when we remember how rarely embryologists continue the
study of the egg beyond the moment of hatching of the embryo, it is
not to be wondered at that this same young Starfish should be intro-
duced to us again and again, in its different stages of growth, under half
a dozen new names, both generic and specific. It is only by a thorough
knowledge of all the changes of form through which these young em-
bryos pass, from the first moment of their existence till they are full-
grown, that we can hope to remedy this evil.

The next state of our young Starfish is, when magnified (PI. VIII.
Fig. 1), even more different from the adult than the pentagonal state
of PI. VI. Fig. 11. The young Starfishes figured on this Plate (PI.
VIII.) were all found attached to roots of Larainaria, thrown up on
the beaches, in the neighborhood, after a storm ; and from their differ-
ent stages of growth, as compared with the oldest Starfish raised from
a Brachiolaria (PI. VI. Fig. 11), specimens of which were also found
upon these roots, it is probable that the sizes here figured are one {Fig.
1), two {Fig. 8), and three {Fig. 10) years old. A considerable number
of specimens were picked up in this way, and they could all be arranged
into very distinct groups, representing the Starfishes of the present and
of two previous seasons. There seemed to be no gradation from one
group to another, such as we have among the young Sea-urchins, which,
in consequence of their manner of breeding during the whole year, form

OUTLilSrE OF THE YOUNG STARFISH. 61

series, the relations of -which it is impossible to determine. In this con-
nection I wonld say, that by arranging the Starfishes found upon our
rocks into series according to their size, we are able to obtain a rough
estimate of the number of years required by them to attain their full de-
velopment ; this I presume to be somewhere about fourteen years.*
They begin to spawn before that time, as sjieciraens have been success-
fully fecundated which evidently were not more than six or seven years
old. It is during the fourth year that the rate of growth seems to be
most rapid. A young Starfish, measuring one and a half inches across
the arms, was kept, during five months, alive in Mr. Glen's tank at the
Museum, and during that space of time it grew to three inches.

In the youngest specimens (PI. VIII. Fig. 1) it is easy to see how the
young Starfish has changed its outline from a pentagonal cross (PI. VI.
Fig. 11) to the one here represented. The original plates are sufficiently
distinct to enable us to trace the process. The arm-plates at the ex-
tremity have been pushed away from the body by the addition of new
spines formed at the base of the ray, and on each side of the interradial
plates (4) (the ovarian plates ?). The terminal plate (4) is perfectly well
defined at the extremity of each ray, and, by cutting out the remainder
of the arm, and bringing the extremity of the ray close upon the disk,
we should have our former pentagonal Starfish almost identically the
same ; the only change being the greater stiffness of the suckers, the
more rounded character of the spines, as well as their greater number
upon the original radial plates. The spines have almost entirely lost
their fan-shaped embryonic type, and are gradually assuming the aspect
of the full-grown rounded spines of the adult Starfish. Here and there,
however, a spine still occurs which has retained its fan-shaped out-
line.

Owing to the elongation of the ray, the single median line of spines
stands out very prominently, and this, together with the rows of large
spines extending from the interradial plates on each side of the rays,
gives to the young Starfish the appearance of a small Oreaster. The
median line of spines is supported by a long, narrow limestone plate, ex-
tending distinctly from the basal plate almost to the terminal radial,
plates totally independent, also, of the prolongation of the ovarian plates

* For an account of the method adopted by Professor Agassiz for ascertaining the age of many
of our marine animals, see Proceed. Esse.x Inst., 1863, p. 252.

8

52 EMBEYOLOGY OF THE STARFISH.

iPc), which make a broad binding on each side of the raj, uniting with
the terminal plate so as to form a continuous limestone cord round the
edge of the Starfish. The interradial plate projects from the angle of the
rays towards the basal plate, spreading somewhat, to fill up the space be-
tween the median arm-plates. We find, in this stage (PI. VIII. Fir/. 1),
the first dorsal water-tubes {d') ; there are five pairs, one tube on each
side of the ovarian plate (pj. But, as jet, no pedicellarioe have appeared.

From the lower side, no trace of the plates of the interambulacral sys-
tem can be seen, beyond the spines which have formed at the extremity
of the ambulacra. The ambulacral pores are arranged in a single row
on each side of the median line, and the slender last-formed tentacles
are placed at. the extremity of the ray, nearest to the odd ocular ten-
tacle ; while the tentacles nearest the mouth are quite short and stout,
having a large sucking disk, and resembling, in all respects, those of the
adults. The separation of the different ambulacral plates is very faint, and
does not become well marked till a later stage. The odd ocular tentacle
has retained its function ; the eye-speck has increased greatly in size, as
well as the bulb to which it is attached, while the walls of the tentacle
are nearly as thin as in the younger stages (PI. VIII. Fiff. 5), exhibiting
no trace of the formation of any sucking disk. Nearest to this are found
the last-formed tentacles, easily recognized by their length, and the some-
what less developed sucker. These and subsequent stages of the young
Starfish show undoubtedly that new tentacles are formed at the extremity
of the rays, while new portions of the upper part of the arm are formed
at the base ; that is, the actinal system is developed at its periphery,
while the abactinal system is developed at the centre.

In young Starfishes of two years (PI. VIII. Fig. 8) the median plate
is longer, more closely crowded with spines ; the terminal plate being
less prominent, though still distinct, while the processes from the median
and lateral plates are quite large. No additional dorsal water-tubes have
been formed since the last stage (PI. VIII. Fig. 1). When examined from
the oral side, the median line is becoming more strongly marked, and
the lateral and ambulacral spines more prominent. These features give
to the young Starfish a more pointed appearance, and the resemblance to
the adult now becomes more apparent.

In somewhat older specimens (three years old) (PI. VIII. Fig. 10), we
finally trace the first appearance of pedicellarite (PI. VIII. Figs. 2, 3, 4,

OUTLINE OF THE YOTJXG STARFISH. 53

v, p"), the dorsal tubes (PI. VIII. Fig. 10, d" d") are found arranged
in greater number along certain portions of the rays ; while the median
and lateral plates have increased so much in size that the terminal plate
has lost entirely the preponderance which it had in younger stages, and
the exti-emity of the arm actually assumes a rounded outline. The dorsal
tubes (d'") are found numerous on both sides of the median arm-plate,
and along the edge of the oral lateral plates [d' ), diminishing somewhat
in size as they approach the extremity of the ray ; they are not open
at the tip. The central basal abactinal plate is still distinct from the
others.

The development of the pedicellariaj around the base of the spines
gives us no clew as to the function which they perform in Starfishes
(PI. VIII. Fijjs. 2, 3, 4). At first a simple projection, they early assume
the character of the head of pedicellaria? without stems, the rounded
swelling becoming conical, after which the fork of the head begins to be
distinguished. In Plate VIII. Figs. 2, 3, 4, we have the different stages
of the spines (^.i), and the pedicellarijB [p, jj"), found at their base. It
was impossible in these young Starfishes to discover the place of the
madreporic body.

[Professor E. Perrier has published a very elaborate and beautifully
illustrated memoir on the Pedicellarite of Echinoderms in the Annales des
Sciences Naturelles. For the discussion of the nature of Pedicellarite see
an account in the Revision of the Echini, Part IV., by A. Agassiz, and an
article in the American Naturalist, Vol. VII.]

From the oral side these Starfishes (PI. VIII. Fig. 9) exhibit scarcely
any difference from th'ose of the stage last described, with the exception
of the somewhat more crowded ambulacra. There is a row of median
ambulacral spines («'), quite small, defining the plates distinctly, as well
as the presence of a very distinct row of spines {u), the ambulacral
spines, along the edge of the ambulacral plates. In the most advanced
of these Starfishes we must specially call attention to the absence of a
well-defined interambulacral system. The young Starfish is still emi-
nently ophiuroid in its most important embryonic features.

Professor Sars, in his Norge's Echinodermer, has described a new
genus, which he has named Pedicellaster. I think there can be but
little doubt, on comparing the figure he has given of his Starfish and the
diflferent stages of our Asteracanthion, that his Pedicellaster will turn out

54 EMBRYOLOGY OF THE STARFISH.

to be the young of one of the species of Asteracanthion of the northern
coast of Europe. The single row of ambulacral pores, the ocular ten-
tacle, the arrangement of the pedicellaria^, the size, all confirm the idea
of its being only the young.

Successive Phases of Devehpmeitt of ihc Larvce of Starfishes. — Before apply-
ing the information thus far obtained to the solution of more general
problems, it may be well to consider what are the normal stages of growth,
at different periods, in the history of our Starfish larvfB. During the
earlier stages of its existence, the young developed from the egg (PI. I.
Figs. 22 - 28) laid by one of our Asteracanthion has no resemblance what-
ever to the future Starfish. This first condition we might call the pyri-
form, or Scyphistoma stage ; when it is simply a symmetrical radiate
animal, reminding us of earlier stages of Polyps and Acalephs. It then
assumes the shape of a dumb-bell, becomes slightly one-sided (PI. II.
Figs. 2-19), and has, in its most advanced state, no other appendages
but the simple crescent^shaped, slightly undulating, vibratile chord (PI. II.
Figs. 20-24). The simple, straight digestive cavity is now differentiated
into three distinct regions. This second stage we might call the Tornaria
stage, from its resemblance to the Echinoderm larvte, called Tornaria by
Miiller, in which all the parts of the adult larva are simply hinted at in
the most rudimentary form, and during which it is eminently cylindrical.
[This Tornaria has been proved by Metschnikoff and myself to be a
young Balanoglossus.] Another well-marked epoch is that during which
the larva passes from the cylindrical, or, as we have called it, the Tornaria
stage, into a quadrangular, somewhat compressed form ; and the compli-
cated system of locomotive appendages, so greatly developed in the Brachi-
olaria, is gradually laid out, thus preparing the larva for the last stages
of its existence, characterized by the development of the young Echino-
derm. This third stage, corresponding to that observed by Van Beneden,
may appropriately be called the Brachina stage. During this period the
former independent water-tubes («•, iv) of the Tornaria stage (the prob-
lematic bodies of Miiller) become united, and are gradually transformed
into the Y-shaped, elliptical water-system (the Schlauch-System of Miil-
ler); this present stage (the Brachina stage) is therefore marked by the
great modifications of the water-system (PL II. Figs. 25-28; PI. III.
Figs. 2-10). In the last stage, which we shall call, with Miiller, the
Brachiolaria stage (Pi. III. Fig. 11; PL VI. Fig^. 1, 2, 4 ; PL VII.

CHAEACTER OF THE DE^^XOP]SIEXT. 55

FUj. 8), the rudimentary locomotive organs, laid out during the Brachina
stage, attain their greatest development, as long, slender arms. The great
changes which take place on the anal extremity of the water-tubes on
both sides of the stomach, characterize the present stage (the Brachiola-
ria stage). These changes upon the surface of the two branches of the
water-tube lead to the formation of the future Starfish. But the incipi-
ent Starfish is, as it were, a part of the Brachiolaria, or rather the Bra-
chiolaria is undergoing local transformations w'hich lead to the formation
of a Starfish. They present thus, for a time, the appearance of a double
existence, as if a new being were forming in one which had completed
its growth. This third period, during which the twofold nature is pre-
sex'ved, is the one which constitutes the Brachiolaria stage. In the Bra-
chiolaria stage there are several marked periods : the parts which appear
at first on the surfaces of the water-tubes have no connection, and stand
in such indefinite relation to each other, that they do not seem to tend
towards a common result. But in proportion as the young Echinoderm
progresses in its development, the relations of the two areas, formed on
the surfaces of the two water-tubes, are more apparent ; and we finally
reach the last of the strictly larval stages, when the Brachiolaria, with
its complicated system of locomotive appendages, becomes secondary to
the young Echinoderm and is completely resorbed by it, when the em-
brj'O enters into its truly echinoderraoida: condition (PI. VI. Fig. 1), the
different stages of which we have already described.

Examination of the Character of tlie Development. — The mode of develop-
ment of Starfishes, thus divided into phases as observed in our Astera-
canthion, cannot be called a case of alternate generation, nor is it a
metamorphosis in the ordinary sense of the word. It is a mode of
development peculiar to Echinoderms, entirely different from that of
any other class of Eadiates. It is not an alternate generation, for the
Brachiolaria can in no way be called a nurse, as each Brachiolaria pro-
duces but one Stai-fish, and the whole of the larva is resorbed by the
Starfish, not an appendage being left out. Nor is it strictly a metamor-
phosis, as the changes which take place are so gradual that at no time
can the line of demarcation be drawn between two stages with any
degree of precision, as in Crustacea or Insects, where the casting of an
envelope marks distinctly different epochs. There is, however, something
m these successive phases of development which reminds ns of the meta-

5G EMBRYOLOGY OF THE STARFISH.

morphoses of Insects. There is a sort of general similarity between this
process of resorption and the growth and changes in the chrysalis of
Lepidoptera, ending in a butterfly. In the latter case, the chrysalis,
though retaining its character throughout the whole growth and develop-
ment of the Insect, has an earlier stage when it seems to be purely chrysa-
lis, and a later one immediately before the hatching of the perfect Insect,
when the butterfly seems to be gaining the ascendency, and the whole
outline of its form may be seen through the chrysalis, which now seems
to be only its envelope. And yet the character of the development of the
Starfish during its Brachiolaria stage recalls also vividly the phenomena
of alternate generations. It is, nevertheless, strictly echinodernioid, and
whether we observe it in the Ophiurans, the Sea-urchins, or the Holo-
thurians and Crinoids, there seems no doubt, from the observations of
Miiller, Busch, Thomson, Krohn, and Agassiz, that it is carried on according
to one and the same plan in all the orders of the class, where we have
corresponding differences in their various modes of development. With
reference to the separate existence of the larva and of the Echinoderm,
urged by other observers, I can only say that nothing of the kind has oc-
curred in those Echinodenris the changes of which I have ti'aced, whether
it be an Ophiuran, an Echinus, a true Starfish, or a Holothurian.

RECAPITULATION.

I shall, in a few words, recapitulate the development of these Star-
fishes, in order to be able more fully to compare my observations with
those of previous writers, and to explain the differences, when they exist.

Changes of the Yolk. — The yolk, after fecundation, separates slightly
from the outer envelope. The segmentation takes place rapidly; as soon
as the yolk has divided into eight portions, they arrange themselves in
such a manner as to enclose the remaining space, which is more and
more separated as the spheres increase in number, until, finally, there is
a complete envelope formed of spheres of segmentation.

Sct/phistoma, or pyriform Stage. — At the time the young escapes from
the egg, it is spherical, and the walls of the envelope are of the same
thickness. 0«e side becomes thicker, the embryo flattening at this ex-
tremity, which is bent in so as to form a slight cavity, in which fluids
circulate. This cavity extends half-way the length of the larva, then

EECAPITULATIOX. 57

swells at the extremity, the walls become thinner, the pouch formed at
the end of this cavity develops laterally, forming two smaller pouches,
which afterwards become hollow bodies, entirely separated from the main
cavity, whence they originated (the problematic bodies of Miiller).

Tornaiia Stage. — The main cavity bends slightly towards one side, and
eventually unites with a depression formed there. This depression be-
comes the mouth ; the other opening, which was the first to be devel-
oped, and served the purpose of a mouth, is changed to an anus. This
agrees with the observations of Krohn, who shows that in an Echinus
larva the mouth is formed after the anus. The bent tube, or cavity,
divides into three distinct regions, forming the cesophagus, the stomach,
and the alimentary canal.

Brachina Stage. — The small disconnected hollow bodies (the water-
tubes, the problematic bodies of Miiller) are not alike ; the left one {left,
when seen from above) connects with the surrounding medium by means
of an opening, the water-pore. This opening in the Starfish is the mad-
reporic body. The water-tubes elongate so as to reach beyond the
mouth, when i\\Qy approach each other and unite, forming a Y-shaped
tube.

Brachiolaria Stage. — Arms are developed from the sides of the larva,
edged with rows of vibratile cilia. Some of these ariiis are of a different
character, having peculiar appendages, the so-called brachiolar arms. It
is on the outer surface of the water-tubes that the Starfish is developed
(not from the stomach, as stated by Miiller) ; one of the tubes, the left,
when seen from above, developing the actinal or ambulacral side, the
other developing the abactinal area. These two areas are open, pentago-
nal, warped, spiral surfoces, making almost a right angle with each other.
The open pentagons do not close till after the Starfish has resorbed the
whole of the larva.

Echinodermoidal Stage. — The complicated system of arms and the whole
of the Brachiolaria are resorbed by the Starfish, which does not separate
from the larval stock, as seems to be the case of Bipinnaria, from the
statements of Miiller and of Koren and Danielssen. The arms of the
Starfish are broad and short in the young, and not symmetrical; the
suckers are pointed, have no terminal disk, and are arranged in two rows,
the sucking disk being developed later. The embryo, if compared to Aca-
lephs, might then appropriately be said to be in its Ephyra stage. The

58 EMBRYOLOGY OF THE STAEFISH.

odd terminal tentacle has an eye at its base, and no disk is ever formed
at the extremity of this tentacle. The abactinal surface is very arched,
the spines are arranged in regular rows, and the arrangement of the plates
reminds us of the plates of Crinoids ; the plates first formed retaining
their embryonic or crinoidal character. The anus opens near the edge of
the disk, on the lower side ; the madreporic body is situated on the
edge, but moves to the abactinal area, in more advanced stages. About
a fortnight is required for the egg to pass through its different stages,
or the embryo to be hatched, and the larva to have reached the condi-
tion when the young Starfish is ready to resorb the Brachiolaria ; and
another week must elapse before it reaches the stage represented in PI.
VI. Fiff. 11. Those which I raised from eggs artificially fecundated re-
tained this shape four months.

CHAPTER THIRD.

EMBRYOLOGICAL CLASSIFICATION OF STARFISHES.

The study of the young forms, or morphological embryology, if I may
so call it, is destined to play an important part in Systematic Zoology;
thouo-h investigations of this kind can only be carried on under peculiar
advantages not easily obtained. The fact that many marine animals live,
in their early stages, \mder stones, or firmly attached to roots of Lami-
narians, in deep water, and are only occasionally thrown upon the beaches
after storms, when their small size prevents us from obtaining them in
any great number, increases the difficulty of this kind of observations.
We must, therefore, limit ourselves to those animals which pass the greater
part of their lives near the surfece of the water, or within the limits of
tide-marks. A commencement has already been made in this direction, in
the study of Fishes, the young of which live among the eel-grass, and in
that of the young of the several species of Ctenophor?e, so abundant
during the summer months along our coasts. For an account of these
investigations, I would refer the reader to the Illustrated Catalogue of the
Museum of Comparative Zoology.*

Comparison of Young and Adult Starfishes. — The difference in appearance
between the young and the adult of our Starfishes is so great, that they
would not be placed in the same family by one unacquainted with their
transformations. The young has characters which, if taken singly, recall
a variety of families; in fact, the combination of characters belonging to
different families is almost always a sign that these features will disap-
pear, or become modified with age.

Here I must again insist on the importance of the constant comparison
of the younger stages of growth with the adult. We are but little accus-
tomed to consider these younger stages in our description of animals,

* [See also my papers on Young Stages of Annelids, Ann. Lye. Nat. Hist. ; Young Echini, No.
VII. 111. Cat. Mus. Comp. Zodl. ; Embryology of Ctenophora;, Mem. Am. Acad. 1874.]
9

60 EMBRYOLOGY OF THE STARFISH.

and we necessarily lose many elements of the greatest importance, when-
ever we attempt to associate the .adults of any class in natural groups,
without taking into account the characters of their young. Natural-
ists, who have not yet entered upon this method of study, cannot con-
ceive what extraordinary facilities this kind of investigation aflbrds for
tracing the more comjjlete affinities among animals. One of the principal
reasons why embryologists have overlooked these investigations may be
found in the fact that they rarely examine more than one species of each
type at a time. Who would place the young Echinus, with its Cidaris-
like spines and straight simple ambulacra, among the true Echinidae, or
take a young Spatangoid for anything but an Echinus? What has the
pear-shaped outline and long tentacles of a young Bolina — which is, in-
deed, a diminutive picture of a Pleurobrachia — in common with the
adult, with its long, twisting rows of ambulacra, and wing-like jirojections
of the spheromeres beyond the actinostome ? Yet these embryonic char-
acters remind us of familiar forms, and cannot fail to give us an insight
into the relative standing of the forms through which they pass.

Let us commence with our embryo Starfish at the time when it is just
forming, and when the first outlines of the abactinal region can be traced.
Suppose its development were to stop there (PI. V. Fig. 5), and that
the slight lobes should close soon after the formation of the coating of
limestone granules over the abactinal area, we should then have a condition
strongly reminding us of a Culcita, with its arched back, its almost circu-
lar outline, and the total absence of any very prominent spines. In the
next stage (PI. V. Fig. 12), the cuts between the rays have become some-
what more marked, the plates of limestone cells are well developed, and
there are tubercles in place of future spines. The resemblance of this
stage to such forms as Anthenea, Pentagonaster, and the pentagonal
Starfishes, in which we find a great development in the abactinal plates,
is at once apparent. In a somewhat more advanced stage, the rays are
slightly more marked, the spines quite well developed ; this type is rep-
resented among living Starfishes by such forms as Pteraster, Paulia, Pen-
taceros, Ai-tocreas, Oreaster : unless it were known beforehand that PI.
VIII. Fig. 1 represents a highly magnified young Starfish, the figure
would readily pass for a new species of Oreaster. The corresponding
changes of the actinal surface are not the less important. In the early
stages the tentacles are pointed, they have no disk (PI. VI. Figs. 3, 9);

EMBEYOLOGICAL CLASSIFICATION". 61

it is only afterwards that they are developed (PI. Vll. Fig. 1; and PI.
VI. Figs. 10, 11, 12). In fact, the tentacles of our young Starfish, in its
earlier stages, resemble those of Astropecten, Luidia, and Ctenodiscus.
We are, therefore, at once provided with a set of characters taken from
the young, enabling us to decide the comparative value of the various
features, and the order in which they are to be taken. From the tenta-
cles alone we are fully justified, upon embryological data, in placing
Starfishes with pointed tentacles lower than those which have disks, like
Asteracanthion. Another embryological feature is the fact that the em-
bryo has only two rows of tentacles, while in the adult Asteracanthion
we find the tentacles arranged in four rows. [The arrangement of the
ambulacral tentacles into furrows seems due simply to the crowding to-
gether of adjoining plates in consequence of increasing age, and has not
the systematic value formerly assigned to it.] Combining these charac-
ters, as we find them in the adults, we have at once good and conclu-
sive reasons for placing all those Starfishes which have, like Ctenodis-
cus, a pentagonal outline, and at the same time pointed tentacles, low-
est in the scale ; next in order would come the Starfishes with pointed
rays and pointed tentacles, without suckers, like Luidia and Astropecten ;
above them pentagonal Starfishes, with plates like Anthenea and Hippas-
teria, and two rows of tentacles, provided with suckers; then those with
more prominent rays, and tentacles also ending in suckers, like Penta-
ceros and Artocreas; higher still, the Starfishes, with long slender arms,
and only two rows of tentacles with suckers, such as Cribrella, Ophidi-
aster, and the like ; while highest in the order we should place the gen-
uine Asteracanthion, with four rows of tentacles, with suckers, and highly
developed spines on the abactinal area.

The same principles applied to the different fomilies would place Star-
fishes having plates without spines lower than those in which the net-
work of limestone is covered with spines on the abactinal surface. This
classification is not very different, as far as regards the order from that of
the three families j^roposed by Midler and Troschel. It differs materially,
however, from the standing given to pentagonal Starfishes in a short paper
by Professor Agassiz, in the Proceedings of the Natural History Society
of Boston. From this it is plain, that the mere study of the adult is not
a sound foundation for a natural classification. The echinoid characters
of the young Starfishes were not known at that time, which would natu-

62 EMBEYOLOGY OF THE STARFISH.

rally give the pentagonal Starfishes an entirely different position. Nor
is it always sufficient to have traced the development of any one species ;
unless it happen to stand highest in its group, its different phases would
not tell us anything of the relative standing of the other members of the
group Avith Avhich the adult is associated. Embryologists should, there-
fore, whenever it is possible, select those species for investigation which,
upon anatomical evidence, stand highest in their group.

There are other embryonic features, recalling not simply families of the
same suborder, but characters of other lower orders. The situation of
the anus on the actinal side, the presence of the madreporic body on
the same area, are features of the Crinoids and Ophiurans. These pecul-
iarities are soon lost, and the madreporic body gradually finds its way
to the abactinal area. The opening of the anus next to the mouth is
eminently crinoidal, and it is accompanied by other structural details
reminding us still more of that order. Were there a stem on the
central plate of its abactinal area, the young Stai-fish, when seen from
the abactinal side, would have all the appearance of a Crinoid. The
central plate corresponds to the basal plate (PI. VI. Fi<j. 10), the set of
five plates in the angles of the arms to the interradial plates, and the
arm-plates themselves to the radial plates of a Crinoid ; and, to make
the resemblance still stronger, the anus opens near the mouth, on the
same side with it, as in Comatula. This analogy had already been
pointed out by Professor Agassiz, in his Lectures on Embryology ; and
it shows conclusively that Starfishes are built upon the same plan with
other Echinoderms, contrary to the views long entertained by Johannes
Miiller. This comparison to the plates of a Comatula can be carried out
to its fullest extent, and is exceedingly instructive if made with the
young Comatula, of which an admirable figure has been given by Pro-
fessor Allman, in his valuable memoir on the prebrachial stage of Coma-
tula, in the Memoirs of the Eoyal Society of Edinburgh for 1863. The
arrangement strikes one, at once, as identical, though the plates are by
no means homologous. The central plate occurs in both, but the most
prominent plates, occupying indeed the greater part of the abactinal re-
gion of the young Starfish, are the same plates which eventually develop
with others at the base of the arms, those at the angle of the arms
being but little developed. It is quite the reverse with Comatula, in
which the arm-plates are but small at this stage ; though, according to

POSITIOX OF THE MADREPOEIC BODY. 63

Professor Allman, who quotes Carpenter, these small radial plates event-
ually encroach upon the others, at the time of the appearance of the
arms, the rest of the calyx being formed by the five large interradial
plates. I cannot agree with Professor Allman in considering the central
plate otherwise than as a solidified homologue of the basalia of the other
Crinoids figured by him; the only difference being that in some cases
the plates composing this piece are soldered together, as in Comatula,
while in others they are kept distinct, as in Coccocrinus, and the like.
From the peculiar way in which young tentacles are formed in Starfishes
may not the strange toothed plates noticed by Professor Allman, at the
base of the tentacula (or cirri, as he calls them), be young tentacles?
Their position seems to me to make this very probable.

PosUiou of the Madreporic Body. — There has lately been a great deal
of discussion among writers on Echinoderms, as to whether the madre-
poric body was, or was not, a proper point to start from in determining
the axes of the body ; Agassiz, on one side, maintaining that the madre-
poric body was constantly in the same relation to the different parts of
the Echinoderms, while Miiller, Cotteau, and Desor have warmly opposed
this view. The mode of formation of the madreporic body seems to me
to decide this question in favor of the former view. The madreporic
body is invariably formed on the left water-tube of the Brachiolaria, and
is placed, during the development of the Starfish, at the angle of the
upper arm. The future position of the madreporic body opposite the
third arm of the open pentagon is therefore, after it has closed, the nat-
ural consequence of its position. The opening of the anus, on the con-
trary, has no such clear and precise relation to the middle arm. At any
rate, however this may be, one thing is perfectly apparent, viz. that
the madreporic body is always placed in the suture of the terminal arms
of the pentagon, which brings it opposite the third arm. Thus the mad-
reporic body gives us the means of dividing the Stai'fish into symmetrical
halves, and of determining the position of the odd arm. The case of the
Echinometrada3 and Salenidae is constantly quoted to show that the mad-
reporic body is not connected with any definite axis. But might it not
be that a stage which is embryonic in the young Starfish — viz. that
preceding the closing of the actinal and abactinal areas — is probably re-
tained in those Echinoid families in which the process of closing is not
completed ? And may not the unsymmetrical position of the madreporic

04 EMBEYOLOGY OF THE STARFISH.

body in such cases be owing to the continuance of this embryonic char-
acter?— the natural result of which would be, to throw the madreporic
body slightly on one side of the middle line, so that, though still retain-
ing its position opposite the third arm, an axis passing through them
both would not divide the spherosome into symmetrical portions. If
there were in nature such forms as asymmetrical Starfishes, analogous to
the Echinometradae, they would be represented by the embryonic Star-
fishes of PI. VI. Figs. 1-6, in which a line drawn through the madre-
poric body and the middle of the odd arm would not divide the Starfish
into symmetrical halves. Suppose the flattening of the young to be
completed without the loss of this want of symmetry, and we have a
form representing Echinometra-Starfishes, if any such exist in nature.
The fact that in some of these Echinometradge the axis, passing through
the madreporic body and this long arm, crosses the median line from
opposite sides, could be easily explained on the supposition that the
former is placed on the ventral instead of the dorsal side of the larva, —
an assumption which is not unfounded, as this occurs in Ophiurans and
in young Starfishes. In this way the change of position in the direction
of the axis which is found in Acrocladia and Podophora on one side,
and in Echinometra on the other, could be easily explained. [For fuller
discussion of the bearing of the positions of the madreporic body deter-
mining the anterior axis of the Echinoderras, see my Revision of the
Echini, Part IV., and the description of Salenia. Consult also the Memoir of
Loven (Etudes sur les Echinoidees).] In Echinoids the actinal and abac-
tinal areas are formed upon the exterior surfaces of the water-tubes, as
in Starfish larva?. This I have shown in the paper referred to above,
published in the Memoirs of the American Academy for 1864. The
earlier appearance of the tentacular pentagon in Echinoids and in Ophiu-
rans is that of a spiral on the surface of the water-tubes, similar in plan
to that observed in our Starfish larva; it is evident that the additional
plates formed in a young Sea-urchin arise spirally, and from what is
known of the mode of formation of the young Echinus and young Ophi-
uran, it follows, necessarily, that the ambulacral sj'stem in both must
have been open pentagons, becoming connected only by the closing of
the surfaces upon which the young Sea-urchin or Ophiuran were de-
veloped.

An examination of the figures of our young Starfish, just after the re-

POSITION OF THE MADEEPOEIC BODY. 65

sorption of the larva (PL VI. Figs. 2, 3, 4), in -which a line, drawn from
the maclreporic body through the middle of the odd arm, would by no
means divide the Starfish symmetrically, confirms the above explanation
of the eccentricity in Echinometi^a. Supposing the spiral to have been
formed from the other side, the obliquity would be in the opposite direc
tion. Of course this is simply a supposition on my part, which future
examination alone can verify ; but it seems to me such a natural expla-
nation of the whole difficulty, that I give it here for what it is worth.
The multiplication of madreporic bodies in many Starfishes need not in-
validate the view I have taken of its value, as we need only ascertain
which is the original one, the others being supplementary. I have found
larvfB with two water-pores (madreporic bodies?), but have never succeeded
in raising them.

CHAPTER FOURTH.

EXAMINATION OF THE INVESTIGATIONS OF FORMER OBSERVERS.

Eeviciv of 3Iiiikr's Observations. — It is with the greatest diffidence that
I enter upon this part of my subject. It seems the height of presump-
tion, for one who has scarcely any claim to recognition, to begin by
criticising so many statements of one of the great masters of our science.
Yet I hope to show, from Miiller's own figures, that the observations I
have made upon the development of our Starfish, though they do not
agree with his earlier memoirs, yet coincide entirely with a few figures
which he has given on the last plate of his great memoirs on the em-
bryology of Starfishes ; and that it is only because Miiller neglected the
earlier stages of develojsment, that he failed to arrive at the conclusions to
which I have been led by the above investigations. I trust that I have
succeeded in describing the successive stages in this development with clear-
ness enough to enable me now to draw a comparison, which the reader may
easily follow, with the last drawings made by Miiller, and to show that,
had he had the good fortune to see so complete a series as that which
I have traced, he would undoubtedly have entirely remodelled his former
views, with the same frankness which has characterized all his memoirs.
No preconceived theories, no observations, however careful, have ever
been allowed by him to interfere in the least with his subsequent obser-
vations. Hence the great difficulty of following Miiller in his intricate
discoveries ; each memoir modif\dng, correcting, and sometimes entirely
contradicting, the previous ones, so that we must, as it were, begin his
book at the end, in order rightly to understand his meaning. Any one
who has tried to follow the development of a single animal, so that noth-
ing should be wanting in the evidence of the successive stages, will easily
imderstand how later observations continually modify and explain what
had previously been considered as well understood.

Although Sars was the first who followed the development of an Echi-

BIPINNAEIA AXD BRACHIOLAEIA. 67

noderm, which, at first sight, did not seem to differ very materially from
what was known of the development of other Radiates, yet Miiller was
the first to trace the wonderful changes of the yomig Echinoderms ; his
memoirs have been the basis of all subsequent investigations, which are
insignificant when compared to the immense amount of labor involved in
his i-esearches. Not alone the history of a single animal, but the history
of a whole class, is gradually unfolded in his successive memoirs. The
very fact that so little has been done in the embryology of Echinoderms
since the days of Miiller — for, in fact, with the exception of Krohn
and Thomson, no one has followed these transformations — is a sufficient
proof of the great difficulty attending investigations of this kind. It
must also be remembered that these animals are so small that it re-
quires the most practised eye to detect their presence; their habits also
are such that we may spend days in watching for them, without obtain-
ing a single specimen, and again be overwhelmed with such an amount
of material as to be at a loss where to begin. This can but heighten
our admiration of the untiring zeal and perseverance of Miiller in fol-
lowing out the development of so large a number of species, in a field
where everything was unknown, and where his powers as an observer
must have been taxed to the utmost.

Bipinnana and BracJiiolaria. — A glance at the figures of Bipinnaria and
of Brachiolaria of PI. IX. of MUller's seventh Memoir will show how dif-
ferent they are, with few exceptions, from the figures of the same larvae
in his former memoirs ; compare PI. VII. of his third Memoir and PI.
II. of his second Memoir. From the figures and explanations given by
the author, it is evident that he had observed, in the last larvae of Star-
fishes found by him, the very charactei'S which have enabled me to
correct his observations. He has seen the two Y-shaped water-tubes ex-
tending the whole length of the Bipinnaria. He has seen, also, that
the pentagon of the future back of the Starfish was open in its younger
stages, though he did not succeed in tracing the position of the tentacu-
lar pentagon, nor does he perceive the connection of these two pen-
tagons wdth the water-tubes. And, finally, if he had kept his Bipinnaria
alive but a short time longer, he would have seen brachiolar appendages
develop, and have satisfied himself that Brachiolaria is only an adult
state of what he calls Bipinnaria. It must be remembered, however,
that the original Bipinnaria of Sars, the Bipinnaria asterigera, has en-

10

68 EMBRYOLOGY OF THE STAEFISH.

tirely different characters from the Bipiiinaria of Miiller. Judging from
the development of our Starfish, it seems to me that Miiller's Bipinnaria
von Helsingor, second Memoir (PI. I. Fi(js. 1-7), is probably nothing
but a younger stage of his Brachiolaria von Helsingor (PI. II. Figs. 4,
5; and PL III.). Van Beneden's Brachina, in its turn, is a still younger
stage of the same thing, or of an allied species. A comparison of the
above figures of Miiller, and of the figures of PI. III. of this Memoir,
will leave no doubt on this subject. For the same reasons the Brachio-
laria of Marseilles is probably only the adult of a Bipinnaria, closely
resembling that of Marseilles (second Memoir, PI. I. Fifis. 8, 9), if it is
not the same species. In the Brachiolaria figured on Plates II. and III.
of the second Memoir of Miiller, the young Starfishes are evidently on
the point of resorbing the arms. The larvae present all the appear-
ance of contraction and distortion usually accompanying this process, and
Miiller's figures agree entirely -with the various attitudes which they as-
sume during this resorption.

If we now turn to his fourth Memoir, which contains the fullest de-
scriptions, we shall see that although in many of the figures of Miiller
the Starfish, or at least one side of it, has been drawn correctly, yet his
statements and some of the figures wliich he gives cannot be reconciled
with one another. On Plate II. Fi(js. 5, 6, of his fourth Memoir, we
have the evidence, from his own drawings, that his Bipinnaria had two
water-tubes ; yet, in the subsequent stages, Miiller says positively that
it has only one water-tube, the one with the water-pore, — a statement
entirely contrary to the earlier stages of his Bipinnaria. From what I
have shown of the mode of development of these water-tubes, of
their increase in size in proportion to the age of the larva, it is quite
improbable, notwithstanding the statement of Miiller, that one of them
should disappear ; he also says that they are not to be confounded with
what he calls " wimpernder Schlauch," while our observations of Astera-
canthion go to show that these two systems are but one.

The discovery of the water-pore in Miiller's Bipinnaria was a great
step towards solving the question of the origin of the madreporic body,
which he rightly conjectures to be nothing but the water-pore. He also
notices the rosette of tentacles, or, more properly speaking, the five radi-
ating tubes from which the tentacles eventually branch. He fails, how-
ever, to notice that this rosette, like the cap of the Starfish, as he calls

BIPINNAEIA ASTEEIGEEA. 69

the back, is open ; and although he has occasionally represented it as
such, he has not perceived the true relation between the positions of
these two areas. He says distinctly that the cloak-like envelope, or the
abactinal area, originates upon the surface of the stomach, whereas it lies,
in reality, upon the surface of the second water-tube, which he says does
not exist in his Bipinnaria ; while the water-system, or the ambulacral
system, originates on the water-tube in such a way that the two open
Avarped pentagonal surfoces, the actinal and the abactinal areas, make a
very large angle with one another; Miiller, however, did not notice that
they were open and warped surfaces.

Van Beneden's observations, in which he says that the two branches
of the Y-shaped water-tubes are separate in the young, and become united
in the adult, are fully confirmed by my observations. Miiller has called
these small bodies, while they are still separate, problematic bodies ; he
says they disappear in older larvae, and have nothing to do with the
" Schlauch-System." It is evident, from my observations, that the Schlauch-
System is only the advanced condition of tlie problematic bodies, which
are isolated on each side of the body in the young larvte (see Pis. II.,
III. of this Memoir, and Van Beneden's Brachina), and become united
in a Y-shaped water-system (Schlauch-System), when they reach the con-
dition of Bipinnaria of Miiller. It wo\ild seem, from his figures, as if
the abactinal pentagon closed, while the Bipinnaria is still visible. I am
rather inclined to think that more advanced larvaj will be found to be
Brachiolaria-like, as is the case with our Starfish and the Brachiolaria
from Messina ; and that this apparent closing up is due to the fact that
the larva is not in its normal state, or that the drawings are made some-
what foreshortened. In the second Memoir of Miiller, on Plate I., we
see that the Y-shaped water-system (Schlauch-System) has been noticed
in two of the larva? {Figs. 4, 7), while in the intermediate stages, and
in younger larva\ it has escaped his notice. It is midoubtedly to Miiller's
want of acquaintance with the earlier and later stages of his Bipinnaria
that we must ascribe the discrepancies in his observations. Many of the
more important points in the structure of the young larvaj naturally es-
caped Derbes and Krohn, who were not familiar with the adult larvte ;
neither of these observers tells us anything of the presence of the water-
tubes, or of the first appearance of the young Echinoderm.

Bipinnana asterigera. — Miiller's views concei-ning the different organs of

70 EMBRYOLOGY OF THE STAEFISH.

Bipinnaria asterigera of Koren and Danielssen are undoubtedly correct.
What they took for a respiratory opening, leading into the cavity, is the
mouth ; they had correctly seen the anus, as well as its connection with
the intestine of the Starfish. Judging from the figures of Miiller, anu
of Koren and Danielssen, there are evidently striking differences in the
termination of the intestinal canal, from that of our Starfish. In Bipin-
naria asterigera the anal opening is on the abactinal side of the Starfish,
while in our young Starfish it is still on the actinal side. The position
of the young Starfish, with reference to the stomach of the larva, seems
still to require further investigation, as it is not j'ossible to say, from
the figures of Miiller, or from those of Koren and Danielssen, what is
its true relation, and whether it has the same oblique position which it
occupies in our young Starfish. The investigations of younger specimens
than those examined by Miiller, or Koren and Danielssen, will at once
settle this point, as well as determine the mode of formation of the mouth
of the young Starfish, and the question of its separation from the Bipin-
naria. From the figure given by Miiller, in his third Memoir (PI. YII.
Ficis. 5, 6, 7), I am led to think that the position is also an oblique one ;
and that, though the Starfish may separate from the Bipinnaria, yet it is
undoubtedly the opening of the oesophagus into the stomach, which be-
comes the future mouth of the Starfish, as in our Asteracanthion. In
his third Memoir Miiller shows conclusively that the madreporic body is
not the scar left by the junction of the young Starfish with the Bipiimaria,
but corresponds to an opening leading into a short tube between two
of the arms ; and also points out the probability of its correspondence with
the opening leading into one of the water-tubes which he had noticed in
Auricularia. This supposition is fully confirmed by the observations we
have made of the coincidence of the water-pore and of the madreporic body.
The slit in the Starfish, noticed by Miiller and by Koren and Danielssen,
was probably owing to the fact that in their young specimens the sj)iral
was not yet closed and flattened, as is the case in older Starfishes.

From the drawings of Sars, and of Koren and Danielssen, it would seem
as if a large tube extended into the long appendage opposite the arms.
If this is truly so, it leaves no doubt that the long, tail-like appendage
of the Bijjinnaria is homologous to the brachiolar appendages of our larvae,
only developed to a much greater extent, and with all the arms placed
nearer together, immediately round the mouth. A comparison, after care-

DIFFEEEXT TYPES OF LAEVJi. 71

fill examination of the position of the Starfish in the Bipinnaria astcrigera
■with the mode of development as noticed in Echinaster (Cribrella) A. flac-
cida, and A. Miilleri, will give the means of settling the true affinities of the
singular ventral appendage of these larvae, and of deciding whether they
are, as I have suggested, the homologues of the brachiolar appendages, —
a result which seems probable from the observations made by Professor
Agassiz, of a circulation in this peduncle, in a species of Asterias (A. flac-
cida, Ag.) closely allied to Asteracanthion Miilleri, the mode of develop-
ment of which is identical with that observed by Sars in Echinaster.

Professor Thomson, who has had occasion to study the sedentary mode
of development of several Echinoderms, has given us the most accurate
description of the structure of this peduncle, in a species which he calls
Asterias violaceus. A glance at his figures and descriptions will suffice
to show us the complete identity between the brachiolar appendages and
this peduncle, in which there is a circulation arising from a branch of the
water-tube, and at the base of which, at the point of junction of the three
arms, we find a peculiar disk, having the same structure as the elliptical
disk, noticed at the base of the brachiolar arms in our Starfish larv*. But
we cannot agree with Professor Thomson, that this peduncle is the first
sign of an ambulaci-al tentacle, the ambulacral tentacles being developed
at a totally different part of the watei'-tube*

Different Ti/pes of Lurvce. — Miiller did not suspect that his Bijiinnaria
and Brachiolaria Avere the larvae of different species of Asteracanthion.
The observations of Sars, who had traced the embryology of Asteracan-
thion Miilleri, in which the eggs attain their full development without
leaving the mouth of the parent, seemed to preclude the possibility of
these nomadic larvjB belonging to the same genus. He even went so far
as to say that his Bipinnaria? belonged to the same genus as the Starfish
of the Bipinnaria asterigera. This is undoubtedly an error, for the Star-
fish of the Bipinnaria asterigera, as figured by Miiller, and by Koren and
Danielssen, has already the characters of a Pteraster ; and it is evident
that the Bipinnaria of Miiller, being a young Brachiolaria, which I have
shown to be the larva of an Asteracanthion, cannot belong to that
genus.

The larvEe which I raised by artificial fecundation from Asteracanthion

* [See also a most interesting paper by Thomson in the Journal of the Linnsean Society, Vol. XIIT.
p. 57, 1876.]

72 EMBRYOLOGY OF THE STARFISH.

berylinvis and Asteracanthion pallidus — species which have their representa-
tives in Europe, and which have, up to the present time, been included in
the same genus with Asteracanthion Miilleri — are free-swimming larvae,
resembling the Bipinnaria of Miiller. These facts can, therefore, leave but
little doubt that MuUer and Van Beneden have observed the larvae of
Asteracanthion rubens M. T., and of allied species, the larv^ of which
have been called by them Bipinnaria, Bi-achiolaria, and Brachina, and are
only different stages of one and the same generic type. The difference of
the two modes of development of A. Mulleri and A. pallidus is so great,
that these two groups of species have been separated into two genera by
Professor Agassiz. [Verrill has subsequently placed A. Mulleri in a sejoarate
genus (Leptasterias), to which have been added Asterias tenera and A,
compta. The former is probably what I have seen called A. flaccida. See
also Memoirs Am. Acad. Fifi. 34, 1864, for an account of its mode of de-
velopment. Compare also tlie development of Pteraster militaris, M. Sars,
Norges Echinodermer, 1861. (PI. VI. Figs. 3-13).] The Brachiolaria from
Trieste and Messina present very striking differences from the northern
Brachiolaria. These larvae are probably the young of Asterias tenui-
spinus, so common in the Mediterranean. In his revision of the Starfishes,
Professor Agassiz has also separated this species from the true Asteracan-
thion, under another generic name. We have next the Bipinnaria asteri-
gera, still another type of larva, belonging in all probability to another
family, differing from both the other larval forms. As Bipinnaria asteri-
gera can only be the larva of a Pteraster, a Ctenodiscus, an Astropecten,
or of an Hippasteria, either of which belong to families distinct from the
Brachiolaria type of larvae, we find differences in form, modified by struc-
tui'al features, characterizing the larval conditions, as well as the adult
stages of families of the same order ; while structural peculiarities in the
larvae characterize the different generic divisions more plainly than in the
more advanced conditions. It is evident, from the observations of Pro-
fessor Agassiz and of Sars, that the Asterias violaceus of Thomson, the
embryology of which he has traced in the Microscopical Journal, must
be placed in the same genus with A. Mulleri, and may, perhaps, be iden-
tical with it, unless the true A. violaceus L. has also a similar mode of
development. [This would most certainly prove that A. violaceus, at least
what the English call A. violaceus, cannot be the male of the European
of A. rubens, as has been suggested by several European writers on Star-

DIFFEEENT TYPES OF LARV.E.. 73

fishes.] There is still another type of Echinoderm larvne,. which in all
probability are the larvte of Starfishes, viz. the Tornaria type. [For
the history of Tornaria, which has been proved to be the embryo of Bala-
noglossus, see my paper in the Memoir of the American Academy, Jan. 1873,
where the relations of Balanoglossus and Tornaria to Echinoderms and their
mode of development are fully discussed.] In this type there is not the
excessive development of the ciliary chord into long, slender arms, charac-
teristic of the Brachiolaria ; there are only slight, wavy indentations, cor-
responding to the position of the arms of the Brachiolaria, as we find
them in the younger stages of the larva? (PI. III. Fig. 4; PL 11. Fig. 26).
In fact, this type of larva, in its adult condition, seems to be a permanent
embryonic type of the younger stages of the Brachiolaria. I would infer
from this that the Tornaria will probably prove to be the larva of Cteno-
discus, Astropecten, or Luidia, or of some Starfish with pointed ambulacral
suckers. Having had the opportunity to examine several of the Tornaria
type of larvae at Naushon, in different stages of development, I hope to
return to this subject at a future time.

[The only important embryology relating to Echinoderms in general pub-
lished since the distribution of copies of this Memoir in 1864, is that of
Metschnikoff.* He has confirmed the explanation I had given of the mode
of development of the Echinoderm upon the surfiice of the water-tubes,
the spiral nature of the young embryo, the mode of development of the
water-tubes as diverticula of the original imaginated cavity, and the re-
sorption of the pluteus by the young Echinoderm in Starfishes ; he has
also been able to follow very carefully the mode of development of the
water-system of an Ophiuran, and showed its entire agreement with the
changes I have traced in the development of the Starfish as far as relates
to the formation of the abactinai system and the ambulacral system.
Although Metschnikoft' has added some new points to the development of
Echinoids, still the mode of formation of the original plates composing
the test of the young sea-urchin is not yet clearly shown, beyond the
very earliest stages. As far as relates to the development of Auricularia,
we can form a better idea than formerly of the nature of the change from
the Auricularia to the young Synapta ; and certainly in a general way this
development is different from the normal growth of some of the other

* Studien iiber die Eatwickelung der F.rhinodermen und Ntmei-tinen, Mem. Aead. St. PetersD.
XIV. No. 8. 1869.

74 EMBRYOLOCxY OF THE STARFISH.

Echinoderms (Slarfisli, Echini, or Ophiurans). The changes in the rela-
tive position of the organs remind us strongly of the mode in which the
Toruaria gradually passes into a Balanoglossus by a mere diflference in
the topography of the organs of the Piute us and of the enclosed Synapta
The mode of development of Holothurians seems to be intermediate be-
tween the pelasgic pluteus, with its gigantic arms (Starfish, Ophiuran,
Echinus) and the sedentary or viviparous development of certain Ophi-
in\ans, Starfishes, and Echini. See also the excellent paper by Selenka. on
Holothuria in Zeitscrif. Wiss. Zool., Vol. XXVIl.]

FIFTH CHAPTER.

ON THE PLAN OF DEVELOPMENT OF ECHINODERMS.

We have constantly insisted, during the whole of this Memoir, upon
the radiate plan of our Starfish larvse in their different stages of growth.
We have, however, seen that this radiate plan of structure, at certain
periods of their existence, is so far hidden by the apparent bilateral
arrangement of the locomotive appendages as readily to escape notice.
We have also had occasion, in discussing the development of these appar-
ently bilateral appendages, to show that Miiller's views of the bilateral
nature of these larvae were founded upon mistaken analogies. It now
remains for us to examine, somewhat in detail, the theory put forth by
Huxley, in his review of Miiller's observations, concerning the articulate
nature of the Echinoderm larvfe. The fiicts already stated respecting
the development of these larvae show that they have only a very remote
analogy to some of the larval forms, quoted by Huxley in order to strength-
en his interpretation of the investigations of Miiller. Misled, perhaps, by
the names which Miiller has given to some of these larviB (" Wurmfdrmige
Larven"), he has allowed this analogy to influence him so far that he revives
the old opinion of Oken, and refers the Echinoderms to the type of Articu-
lates. [See my Memoir on Balanoglossus, for a later review of the views
of Huxley, Haeckel, and others, who have urged the affinities of Echino-
derms with worms.] Huxley has given us no observations of his own,
bearing upon the subject, but endeavors to justify his assertion by redu-
cing all these forms to one hypothetical type, having an elongated form,
a straight intestine, with the mouth at one extremity, the anus at the
other, and girded by a circular ciliated fringe, just like the larvae of some
Annelids. The region in front of the ciliated fringe he calls prcefrocluil,
and the region behind the fringe postrochal ; and then, by an ingenious
process, he shows how all these different forms might be produced by

the greater or less development of one or other of these region. He
11

76 EMBEYOLOGY OF THE STAEFISH.

then attempts to prove, furtlier, that there is an intimate connection be-
tween the point where the young Echinoderm is developed, and the po-
sition of the rows of vibratile fringes ; Starfishes being, according to him,
developed in the postrochal and the Echini in the pra^trochal region. Any
one who has observed these larva) alive cannot fail to see that whatever may
be the position of these vibratile fringes, the young Echinoderm, whether it
be an Echinus, a Starfish, or an Ophiuran [also a Holothurian Selenka], is
developed in exactly the same spot on the sides of the stomach, upon the
outer sui'face of opposite water-tubes, one of them forming the actinal, the
other the abactinal surfoce of the future Echinoderm. The hypothetical
form of Huxley is indeed one which has never been observed, as in all
larvfe of Echinoderms the mouth and anus are always on the same side,
viz. on the lower surface of the larva. It is only during the first few days,
after hatching from the egg, that the so-called mouth is placed at one
end ; this, however, is not observed beyond the time when this open-
ing performs the double function of mouth and anus, and leads into a
very short digestive cavity. By the time the true mouth begins to be
formed, the future anus, which has served the purpose of mouth thus far,
has already changed its position to the lower side. The mouth is, in fact,
never formed at one extremity, but always in the centre of the lower
surface, and only some time after the anus, which performs the functions
of a temporary mouth. This has been demonstrated by Krohn and my-
self, with reference to the Echinus larvae, and I trust that the preceding
pages have shown it to be also the case with our common Starfish. [See
also Selenka for Holothuria.] The division into rings, of what Miiller calls
the Wurmfdrmige Asteridenlarve, is only an optical delusion, due to the
lines formed upon the abactinal surface during the closing of the pentagon.

The radical difference in the mode of formation of the oesophagus,
stomach, and intestine, in the Echinoderm larvte, as compared with the
larva? of Annelids, a number of which, including those most resembling
Echinoderm larvte, I have examined myself, will, perhaps, be the strong-
est proof that they do not belong to one and the same type. The diges-
tive cavity of Annelid larvte is formed by the liquefaction of the interior
of the larva, while in the Echinoderm larvae the digestive cavity is formed
by the bending in of the outer wall of the larva itself. The superficial
resemblance of Annelid larvaB to those of Echinoderms is due to the append-
ages surrounding the mouth, while the principal appeirdages of the Echini

PL AX OF DEVELOPMENT OF ECHINODEEMS. 77

and Starfish larva3 are developed from the vibratile chord developed round
the anus. Nothing is more characteristic of the Echinoderms among Ea-
diates than the isolation of the digestive cavity by means of distinct walls.
This feature is so strongly marked that a larva can be recognized as an
Echinoderm larva before its radiate characters are developed. It is only
later that the circular tube, the water-system, is formed, Avhile the ciliary
appendages, which have nothing to do with the formation of the Echino-
derm, make their appearance later still long after the first rudiments of
the Echinoderm (the water-tubes) are present.

It seems to me that the different modes of development in Holothu-
rians, Echini, true Starfishes, Ophiurans, and Crinoids, different as they are
apparently, may easily be reduced to a single type. [See the Memoirs
of Metschnikoft' on the affinities of Echinoderms, in Siebold's Zeitschrift for
1874. Since this paper was written Haeckel has put forth his views of
the relationship of the Sponges and Coelenterates, and of the Echino-
derms and Worms. As the whole subject is intimately connected with the
history of Tornaria and Balanoglossus. I would refer to my Memoir* for
an analysis of the modifications which these views are likely to bring
about rea;ardin": the classification of Echinoderms and of Coelenterates ;
also to my Memoir on the Embryology of CtenophorfP, Mem. Am. Acad.
1874.] We have in Ophiurans two different modes of development, — one
by means of the Pluteus, the other by means of the viviparous mode of de-
velopment observed by Krohn and Schultze. We have two similar modes of
development in the Starfishes, — the one as observed by Sars and Agassiz
in Echinaster, the other in which the embryo assumes the shape of a Bi-
pinnaria or Brachiolaria ; and, finally, in the Holothurians we have these
two modes represented by the Auricularia type and the tj'pe of the
"Wurmformige Holothurienlarve." [See also for Echini Thomson's paper
in Journal Lin. Soc. and A. Agassiz, Viviparous Echini from Kerguelen,
Proc. Am. Acad. 1876.] The difierence between these two modes seems
to be one of time ; in one case, the eggs are retained by the parent
until they have passed through many of their changes, and are freed
in a stage corresponding to that of our young Echinoderm after it
has resorbed its Pluteus, its Brachiolaria, or its Auricularia. In the
other case the egg goes through all these changes after it has left the
parent, developing this complicated system of arms, which seems to be

* A. Agassiz Balanoglossi s and Tornaria in Memoirs American Academy, 1873.

78 EMBRYOLOGY OF THE STAEFISH.

simply a means of locomotion for the young Starfish till it shall have
acquired a sufficient size to be able to take care of itself, and use its
suckers as organs of locomotion.

Have we not here, in Echinoderms, something analogous to what we
have in Discophorous Medusae ? In Cyanea and Pelagia, for instance,
where, in one case, the young Acaleph passes through a Scyphistoma
stage before it reaches the Epliyra condition, while in Pelagia, on the
contrary, the Ephyra is at once produced from the egg, without passing
through the Scyphistoma stage.

I think it can be easily shown that there is, in reality, no difference be-
tween these two modes of development ; it is merely a question of quantity.
In Cribrella, in Pluteus, in Brachiolaria, or in Auricularia, the young Echi-
noderm is developed on the outer surface of the water-system. The
water-tubes obtain a great prominence in Auricularia, in Brachiolaria,
and in the Pluteus-like form of the Ophiurans and Echini, Avhile in types
of development like those of Echinaster they remain more rudimentary ;
the only appendages developed in this last type being those which cor-
respond to later periods of growth in the Starfish larvte, viz. the
brachiolar appendages. The peduncle and its ajipendages, by means of
which the young Echinaster fastens to the rocks, are strictly homologous
to the brachiolar appendages of our Starfish larvaj. In fact, when the
young Starfish has resorbed all the arms, and there is nothing left of
them, except a few swellings on the actinal side, to mark their former
position, the brachiolar appendages are in exactly the same position
as that occupied by the peduncle of the Echinaster larva. Had we known
nothing of the previous modes of development, and found those young
Starfishes in the open sea in this stage, nothing would have been more
natural than to have assumed that they had reached this condition by the
same mode of development. The cavity noticed in the peduncle of the
Echinaster larvae is part of the water-system, corresponding to the branch
of the water-system leading into the brachiolar arms of our Asteracan-
thion larva.

The same is the case with the two modes of development of Ophiurans
and of Holothurians ; they are shorter ways of arriving at the same point,
whether they pass through what we shall call hereafter the Pluteus type
of development or the Echinaster type ; in either of the orders it is one
and the same thing differently carried out. The larviB of our Cribi-ella,

PLAN OF DEVELOPMENT OF ECHINODEEMS. 79

■which I have had frequent occasion to examine, have satisfied me that
the process of development is the same, with the exception that it is
shorter. The larvte of Ophiurans examined by Professor Agassiz at
Charleston would lead to the same conclusion with reference to the
Ophiurans; while, from the drawings of MuUer, it is easy to satisfy one's
self, with the above data, that the two types of development of Holo-
thurians examined by him are only modifications of each other. As the
only larvfB of Holothurians which I have seen belong to the " Wurm-
fbrmiger" type, I am unable to state this from actual observation. It is
evident that we have also in Comatula these two types of development.
Professor Agassiz frequently observed that in a species of Comatula found
in Charleston, S. C, the young embryos remain attached to the par-
ents ; while Thomson and Busch have found the larvae swimming freely
about.

[An important paper on the development of Comatula by Goette, in
the Archiv fiir Microscopische Anatomic for April, 1876, gives us the
early stages of its embryo. Goette shows conclusively that the type of
the crinoidal development is echinodermoid. We have, as in Echini,
Starfishes, Ophiurans, and Holothurians, an original digestive cavity, from
which arises the water-system, as diverticula. This observation is in
direct contradiction to that of MetschnikofF, who distinguishes the Cri-
noids from the other Echinoderms by the absence of these processes.
Goette's observations of the early stages are, however, in complete agree-
ment with the usual mode of development among Echinoderms. Unfortu-
nately the subsequent stages are all figured from embryos preserved in
osmic or chromic acid, and while I have the greatest respect for Goette's
technical skill, the very fiict that in so many general points he differs,
both from MetschnikofF and myself, throws considerable uncertainty on
the whole of his memoir. He begins by stating that the larval mouth
(the subsequent anus of the other Echinoderms) is entirely obliterated. As
he has not followed this from living embryos, we must be pardoned if,
knowing, as we do, the difficulty of tracing the gradual changes in
living embryos of the most transparent kind, we doubt many of his
conclusions obtained from the study of opaque embryos acted upon by re-
agents. Although Goette has derived his knowledge of the present paper
from the excellent abstract in Leuckart's Jahresbericht, he has not only
credited me with a very indifferent treatment of the embryology of

80 EMBRYOLOGY OF THE STARFISH.

Ecliinoderms, but also with several errors contained neither in the ab-
stract nor in the original. He does not appear to know my paper on
the Embryology of Ctenophora?, nor on Balanoglossus, published in the
Memoirs of the American Academy in 1873 and 1874, and distributed
at tlie time ; consequently, in what he now writes, the older views re-
garding the affinities of Echinoderms with Coelenterata and Annelids, which
had been discussed from a different standpoint, do not receive the least
recognition.

The mode of development of these two types having been shown to be
on one and the same pattern, modified in such a way that a like re-
sult is reached either by fewer stages or by a greater or less i-apidity
in the process, it remains for me to show that the larvte we have had
before us, in the complicated form of a Brachiolaria or a Pluteus, is really
built upon tlie radiate plan. We find a good starting-point in the water-
tubes, which, as I have shown, become the circular tulje of the 3'oung
Starfish, from which the ambulacral system is afterwards developed. This
Avater-tube, it is true, is not circular ; it is not continuous, and yet it is
the homologue of the circular tube of Acalephs, the radiating tubes being
developed oidy afterwards, when the pentagon of tentacles is formed.
The mouth is placed wnthin this circular tube ; and the fact that the
mouth of the larva is brought, by the contraction of the oesophagus, close
upon the stomach, does not change its position with reference to this
circular tube. The water-system contracts with it, changes its position,
and surrounds eventually the new opening, by the flattening and closing
of tlie Starfish.

The Brachiolarian and Plutean stages are the Acalephian stages of the
Echinoderms, corresponding to the Hydrarium forms of the Acalephs, in
their Polyp stage ; Avhile the arms of the Pluteus stage, Avith their cords
of locomotive cilia, recall strongly the strange filiform appendages of por-
tions of the spheroinere, covered with locomotive flappers as in Euramphnea,
and other Ctenophora?. The resemblance of the larva^ of Echinoderms to
Ctenophorce had already been pointed out by Baer, and more recently by
Professor Agassiz, who was not then acquainted with the observations of
Baer. This comparison seems to have found but little favor with more
recent investigators. Leuckart, in his Bericht for 1862, simply says that
no further proof has been adduced by Professor Agassiz to show that the
homology holds good. A writer in the Natural History Review for 1861

PLAN OF DEVELOPMENT OF ECHINODERMS. 81

seems to consider the whole comparison so puerile as not to be worth
even a moment's consideration ; and the off-hand way in which he dis-
misses the whole subject shows his total want of appreciation of the argu-
ments by which this view is supported. If the writer of the said article
had ever seen the young of Brachiolaria, of Pluteus, or, still better, the
young of Tornaria, swinnning about amongst crowds of young Ctenophoroe,
such as Idyia, Pleurobrachia, Mertensia, or Bolina, he would not have
passed such a sweeping judgment on this comparison. The motions of
a Tornaria are so similar to those of young Ctenophorae, that I venture
to say that many a skilful naturalist would be deceived as to their true
nature, on first seeing them moving about together in the water. The
Tornaria has no appendages developed into long arms as in the adult
Brachiolaria or Pluteus. The appendages remain always abortive, the
larvae in their adult condition resembling young Ctenophora?. From an
examination of drawings given by Miiller, Professor Agassiz Avas induced
to make the same comparison already hinted at by Baer, and we have
seen that it is sustained in every particular. Gegenbaur has also noticed
the resemblance between 3'oung Trachynemae and Echinoderm larvae.

From what has been said, it is evident that the plan of radiation un-
derlies this apparent bilaterality of the Brachiolaria and of the Pluteus.
The throwing of the whole of the stomach and the alimentary canal on
one side, the complicated system of arms arranged with perfect symmetry
on each side of the axis, passing through the mouth and the anus, does
not change, though it partially conceals, the radiate plan. We have Holo-
thurians which always creep upon thi-ee of their ambulacra, where a dorsal
and a ventral side, an anterior and a posterior region, are subordinate to
the plan of radiation ; and the same takes place to a less extent in Spa-
tangoids. Among Polj-ps even, which are, as it were, the simplest type
of radiate animals, an anterior and a posterior region are strikingly shown
in the case of Arachnactis. The additional spheromeres are all added at
one extremity of the mouth-slit, and yet the Actinia is made up of radi-
ating spheromeres. The earliest stages of the larvae of Echinoderms, be-
fore the appearance of the water-tubes, reminds us forcibly of the young
Actinia soon after it has escaped from the egg, or of the first stages of
growth of a Scyphistoma, after it has attached itself to the ground, pre-
vious to the formation of tentacles. Let us now consider what constitutes
the difference in the structure of these animals in their primary stages

82 EMBEYOLOGY OF THE STAEFISH.

of growth, as far as the different classes of the type are concerned. They
are all built according to one and the same plan, yet this plan is so car-
ried out as to be eminently echinodennoid in one instance, acalephian in
another, and polypoidal in a third. In young Echinoderms, as in young
CtenophorjB, we find nothing of the remarkable preponderance of certain
parts Avhicli gives these young their bilateral appearance in more advanced
conditions. Their radiate character is extremely prominent at first, but
becomes gradually obscured and hidden under the guise of this bilateral-
ity, which is, after all, due only to the excessive development of certain
spheromeres as compared with the others.

The case of these larvce is only an additional example of what we find
so often in nature, that a plan of structure which seems to prevail is
in reality only an external analogy produced by great predominance in
certain parts, but subservient to the primary plan, even though the latter
be perceived only on closer examination. This view solves a question
which has hitherto perplexed all investigators of this subject, viz. how
it was possible that a larva, which has always been considered as bilateral,
should produce a radiate animal by a process of internal gemmation. It
is, indeed, a bilateral larva, but built upon a radiate plan ; a larva recall-
ing a lower class of this branch of the animal kingdom, an acalephian
larva giving rise to an Echinoderm, which, from its very beginning, is a
radiate animal, having all its spheromeres developed at the same time,
and equally.*

These ti-ansformations are, however, peculiar to the class of Echino-
derms ; they constitute neither a metamorphosis nor a case of alternate
generation. The egg becomes the embryo larva, nothing essential is lost
during the process, no intermediate individual comes into the cycle. It
is the yolk which becomes the larva, the latter being, in its turn, trans-
formed into the young Echinoderm. This larva is, in short, an Acalephian
larva, reminding us somewhat of the twin individuals of free Hydroids,
the Diphyes, though adapted to the mode of development of the Echino-
derms. But in the latter we have no intermediate condition corresponding
to the Polyp-like Hydroid in Acalephs from which the Medusa? or repro-
ductive individuals arise, and in their turn, bring forth the Hydroid again,
which completes the cycle by developing another set of Medusae.

* For a closer comparison of young Ctenophorse and Echinoderm Larvie, see tlie Chapter on Cte-
nophoroe, Illustrated Catalogue of the Museum of Comparative Zoology, No. II.

PLAX OF DEVELOPJIEXT OF ECHIXODEEMS. 83

If the views here taken of the plan of development of Echinoderuis be
correct, they inti'oduce a new set of facts respecting their affinities with
the Polyps and Acalephs, which cannot fail to have an important bearing
on the question of the separation of the Echinoderms as a distinct type
from the two latter groups. The Echinoderm plutean form, with its mouth,
stomach, intestine, and with its water-system originally forming a part of
the digestive cavity, bears, as it seems to me, the same relation to the
Ctenophorae which the Hydroid Polyps hold to the true Pohps. The Cte-
nophorje may be considered, as it were, the prototype of the Echinoderms,
as the Polyps are the prototype of Acalephs. We have in the Ctenophorne
a digestive cavity, from which branches the water-system, and that peculiar
funnel ojjening outwards, through which the fecal matters of the Cteno-
phoroB are discharged, reminding us at once of the almost identical arrange-
ment of an Echinoderm Pluteus, in the relations of the intestine to the
stomach. The plutean forms certainly show that the plan upon which the
Echinoderms are built does not differ from that upon w'hich the Acalephs
are built, and that we have between the Echinoderms and Acalephs the
same connection, based upon identity of plan, as exists between the Acalephs
and Polyps. We cannot, therefore, admit that the views so frequently
urged and so universally admitted, in support of the separation of the
Acalephs and Polyps as a distinct type (Ccelenterata), from the Echino-
derms, have any real foundation in nature ; and still less can we concur
in them when we remember that the main argument in their favor rests
upon the assumed total want of connection between the ambulacral system
and the digestive system. This connection has been shown by Pro-
fessor Agassiz to exist in the adult of many Echinoderms, while the facts
above stated prove that it also exists in the early stages of the embryonic
development, where, in fact, the water-system is formed from the digestive
system. With this evidence falls the strongest argument for the validity
of a classification by which the type of Radiates would be broken up, and
the Polyps and Acalephs separated from the Echinoderms, as a dis-
tinct type, imder the name of Ccelenterata. We are, therefore, justified
in affirming that the tj-pe of Radiates constitutes an independent type
of the animal kingdom, coiitaining three equivalent classes, — Echinoderms,
Acalephs, and Polyps.

12

PART II.

ON THE SOLID PARTS OF S031E .NORTH AMERICAxN

STARFISHES.

HOMOLOGIES OF ECHIISODERMS.

Comparatively little use has been made thus far, in the study of
Starfishes, of the structure of their hard parts. With the exception of
the short paper by Gaudry, the occasional references to them in sundry
memoirs of Miiller, Duvernoy, Agassiz, Perrier and others, only show
that we know but little of the solid frame concealed under the mass
of appendages covering a starfish. The study of these solid parts is
instructive, as it throws new light on their homologies with other Echi-
noderms, and enables ns to form a better idea of the I'elationship be-
tween Ophiurans, Echini, and Starfishes. The homology between these
orders, as usually understood, can be stated in a general way as follows ;
the great development in Starfishes of what has been called the tergal
system, covering the centre and arms, forming a system where no special
regular arrangement could be traced, joined to the presence of distinct
ambulacral and interambulacral plates, limited to the furrows occupying
the lower face of the arm ; also the absence of specialized genital plates,
or anal plates, and the presence (in some genera) of a special ocular
plate. The essential character in a general way of the Ophiurans as
distinguished from Starfishes is the presence of genital plates, and the
limitation of the tergal plates to a comparatively simple casing, con-
sisting of few plates enclosing an ambulacral system, no interambulacral
system having been traced ; while in Echini the tergal system is re-
duced to a minimum, ocular and genital plates being present, and the
ambulacral and especially the interambulacral plates greatly developed.

From the more careful examination of Echini a hypothetical Echino-
derm was formerly established, which has served as the type to which
the other Echinoderms were to be reduced, and with which they should
homologize. In the absence of embryonic data, however, the fundamental
facts derived from the study of young Echinoderms already point to

88 HOMOLOGIES OF ECHIXODEEMS.

a different interpretation of the Echinodermoidal homologies. Taken
in connection with our knowledge of the hard parts of Starfishes as
compared with those of Ophiurans and Echini, they throw much light
upon many imperfectly known structural features of Crinoids. The liv-
ing Crinoids, on account of their small number, have till recently seemed
to promise but little help in explaining the fossil forms. The collec-
tions of the Challenger include, however, a variety of stalked Crinoids,
and until the information to be derived from them is on hand it does
not seem advisable to extend the comparison of the hard parts beyond
the more common orders.

All young Echinoderms, while still in the Pluteus stage, or soon after
its resorption, are strikingly alike. They all have an actinal and an
abactmal area. The actinal area is occupied almost entirely by the
ambulacral canals radiating from the central ring enclosing the actino-
stome, with their latei\al ambulacral tubes existing as mere loops, the
different tubes not being as yet encased in any limestone plates. The
abactinal surface consists of the outer integument, in which rudiment-
ary plates begin to appear, made up in the early stages merely of Y-
shaped rods, more or less closely connected together, so as to form
patches of reticulated network to become in the future the solid plates
of the Echinoderm.

Thus far all Echinoderms are alike, and show no structural difierence
between the different orders. We shall greatly facilitate our examina-
tion of them by beginning our comparison at this early and imiform
stage, so that we may see how far we can, by merely tracing the devel-
opment, explain the mode of differentiation by which the orders gradu-
ally assume the structural features of the adults.

Before proceeding any further in this comparison, I must state that
I have given already in detail* my reasons for considering the Echi-
noderms as more closely related to the Polyps and Acalephs in oppo-
sition to the view lately revived of their affinity to Worms. I have
stated the objections mainly on embryological grounds, by comparing
the development of the most Echinodermoid larva among Annulata, that
of Balanoglossus, with other vermiform Echinoderm Plutei. Haeckel has
recently strongly urged, on theoretical grounds chiefly, their annulate

* See A. Agassiz, Embryology of Ctenopliorse ; Revision of the Echini ; The History of Bananoglossus
and Tornaria.

HOMOLOGIES OF ECHINODERMS. 89

affinity, and has assumed the composite nature of Echinoderms, which he
considers as a colony of five persons xmited at the buccal extremity in a
somewhat similar way to that of a colony of compound Ascidians hav-
ing a common cloacal opening.

He considers each arm of a Starfish, for instance, — and it will apply
equally well to any Sea-urchin, — as made up of a series of distinct ar-
ticulations, just as well marked as the articulation of any Annelid. To
a certain extent this analogy is correct ; we find a repetition of very
similar parts, a remarkable vegetative capacity in all Echinoderms, which
at first glance might seem to be of great importance as confirming their
articulate affinities. Yet the earliest stages of the young Echinoderms
in the Pluteus show beyond doubt that they have nothing in common
with a community.

As well might we compare the simple chymiferous tube of an Aca-
leph with a single individual, and make a many-rayed Zygodactyla a
community of individuals with a single central digestive cavity. The
very fact that we can trace the passage between an Acaleph with a
polymeral chymiferous system like Zygodactyla and a Siphonophore zoid
in which we can trace but a single chymiferous tube, shows, at any
rate, that the number of ambulacra! tubes should not be taken as any
proof whatever of a composite structure. When we come to the ar-
ticulation of the arms, can we consider that as anything beyond the
adaptation of the ambulacral system to the deposition of limestone plates,
allowing certain limited movements? In the whole order of Echinoderms
traces of this adaptation can be seen, as in some genera of Echini,
which, like Astropyga and the Armored Echini, retain a more or less
movable test, while in Holothurians the limestone deposition is reduced
to a minimum, the latter showing the range of a dermal and closed
ambulacral system, while the Ophiurans show the limits within which
thfc articulation can be developed in the Starfishes proper (Crinoids not
being in question at present). In consequence of this articulation and
their presumed Annulose affinities, Haeckle does not hesitate to derive
Echinoderms from worms, but as far as the orders of Echinoderms now
known are concerned, it seems impossible to imagine, even with the
light palseontology has thrown upon their appearance, how they have
succeeded one another, much less whence they have been derived. We
can readily see from the presence of several of the orders of Echino-

90 HOMOLOGIES OF ECHINODEEMS

derms in the older geological deposits, that, if any development from
one order to another has taken place, it must have been during much
earlier geological periods. As far as we now know, palaeontology throws
no light whatever upon such a transition, however possible it may seem
from embryological data. Starfishes, Ophiurans, Echini, and Crinoids ex-
isted in the oldest-known Echinodermoid fauna, having all the typical
features of Echinoderms of our day, or only so for modified as to be
readily homologized with them. If there has been such a thing as a
single ancestral Echinoderm, his primordial descendants early assumed
different lines of development diverging to a great degree, and retain-
ing their characteristics from the earliest-known geological period. This
at least appears to be the case with Starfishes * and Ophiurans ; while
the different groups of Crinoids which have appeared and vanished are
numerous as compared to those of the other orders.

The Echini again continued to develop until the secondary period with
very little modification, and only after the Jurassic period did the marked
changes begin through which they subsequently pass ; changes fully equal-
ling those of the Crinoids in their earlier geological history. t And yet, great
as these changes have undoubtedly been, were we to measure them sim-
ply by pateontological evidence, we must remember that they are not
greater in degree than the changes known to take place among the
Echini of the present day during their embryological development. But
while the successive appearance of the great types of Echini in geologi-
cal time — in other words, their paliBontological development — is in the
strictest harmony with what we know of their embryological develop-
ment,t we as certainly know nothing whatever of the causes which have
brought about their sequence in time, in such striking agreement with

* The attempt made by G. O. Sars to prove Brisinga to be the living representative of the palaeo-
zoic Starfishes seems to be very far-fetched, and I must acknowledge I have been unable to see
any such radical difference between Brisinga and ordinary Starfishes (Solaster, Crossaster, and Pycno-
podia, for instance) as Sars insists upon in liis Memoir on Brisinga. The type of Starfishes, as I have
already shown, has been remarkably persistent from the earliest geological periods to the present
day, and there is no indication that the Starfishes now living have undergone such changes as to
make the agreement of Brisinga or other genera with the older forms a matter worthy of especial
notice as survivals or representatives of the earlier types.

See also the view taken by Liitken in regard to the affinities of Protaster in his Ophiur. Add. III. ;
he does not agree with the view taken by Sars.

t See A. Agassiz's Revision of the Echini, Part IV.

J See Part IV., Revision of the Echini, by A. Agassiz.

HOMOLOGIES OF ECHIXODERMS. 91

the sequence in their phases of growth. All we can say at present is
that the course of the embryological development of the Spatangoids is
such that we can, as it were, read off upon it the sequence of echinoidal
development since the Jurassic time in the developmental history of
some genera of that group. In the one case, however, this development
is accomplished in the course of a few years, in the other it stretches
over a comparatively infinite period. We have no data for any such
comparisons in the other orders of Echinoderms.

The case of successive modifications of the ancestral horse, which has
so often been brought forward as conclusive regarding the genealogy of
the group, although more familiar, is far less complete and much more
limited in time than the succession to be traced from the palteontological
evidence of Echini. But while natural selection gives a plausible explana-
tion of like problems among Vertebrates, it foils utterly when applied
to the majority of the Invertebrates, and we have completely failed
thus far to find any causes for their palaeontological development dif-
fering from those acting upon their successive embryological stages at
the present day, of which we know absolutely nothing.

Let us return now to the comparison of the changes undergone by
the embryo Echinoderm from its earliest post-Pluteus stages, vmtil the
structural features characteristic of the several orders are clearly differ-
entiated. The actinal and abactinal surfaces of the embryo Echinoderms
in the different orders are, as has been stated, identical ; and it would
be impossible to characterize them from early stages immediately follow-
ing the resorption of the Pluteus, in the same way as from the adult.
The abactinal surface consists, in all cases, of a central plate, round which
are arranged radial and interradial plates, while the actinal surface is en-
tirely occupied by the pentagonal rosette of the water-system, held by the
abactinal system as it were in a cup, — a combination which is strictly
crinoidal. It is only later that ordinal distinctions appear, but in such
succession as to show that the homologies of the several orders as usually
understood are not correct.

In the case of the young Starfish the radial plates of the abactinal system,
which form the dorsal part of the arms, gradually extend towards the edge
of and down on to the actinal side, enclosing the water-system little by
little, and finally, as has been described, covering the ambulacra! tube, leav-
ing only openings for the passage of the tentacles. This is a stage which

13

92 HOMOLOGIES OF ECHIXODEEMS.

is passed through by Ophiurans and Echini as well as Starfishes, the only
difference in the subsequent development being that the Ophiurans always
remain in an embryonic condition, closely resembling the one just described.
In the Starfishes the actinal plates formed by the bridges separating suc-
cessive pairs of tentacles become resorbed along the central line, the edges
forming inwardly by spurs the true ambulacral plates, and the plates which
little by little develop so as to form the edge of the arms are likewise
formed from the plates originally a part of the abactinal system. Those
Avhich are on the outside of the tentacles become the interambulacral
plates, but differ in no way from the plates forming the sides of the arms.
In the case of the Starfishes these side arm-plates are often very numerous ;
in the case of the Ophiurans they are reduced to a minimum, the upper
arm-plate being, as in young Starfishes, very prominent and distinct, while
the lower arm-plate is formed by the junction of opposing spurs of the
interambulacral plates, as can readily be imagined from a comparison with
Brisinga, where we find a spur from the interambulacral plates extending
nearly one third across the arms. Vie must only remember that in Ophiu-
rans the lower arm-plates represent the original plates derived from the
abactinal side extending across the tentacles, Avhile in Brisinga and Star-
fishes the median part of the plate has become resorbed, so that the
tentacles passing between the ambulacral plates are inside of the inter-
ambulacral plates, while in Ophiurans they pierce the connected inter-
ambulacral plates (or the lower arm-plate).

Something analogous to what takes place in Ophiurans occurs in Echini.
The plates which cover the water-system never become resorbed (as in
Starfishes) ; there is no internal ambulacral system of plates developed,
from the fact that new plates in the Echini are always developed near
the basal plate (the apical system), while new plates in Ophiurans and
Starfishes are invariably formed at the extremity of the arms. In Echini,
therefore, the extremity of the water-system (the ocular tentacle) remaining
connected with the original apical system, the water-system thus forms a
loop, one end of which is attached to the so-called ocular plate, while the
other connects with the circular canal at the mouth, and hence, both ends
being fixed, the new jilates must necessaiily cover the water-system, while
in Ophiurans and Starfishes, one end alone being fixed, it is possible, as in
the case of Starfishes, for the water-system, owing to the resorption of the
central part, to appear in a peculiar position. But in spite of this similarity

HOMOLOGIES OF ECHIXODERMS. 93

in the position of the water-tube in Echini and Ophiurans, the hxtter are
really more closely allied structurally to Starfishes than to Echini. This
will readily account for the position of the water-system inside of the test
in Ophiurans and Echini, contrasted to the Starfishes. AVe can also
homologize Holothurians with Echini by supposing that in that group
the limestone plates never form ambulacral and interambulacral plates,
but that the abactinal system of the embrj-o, as it elongates, covers
irregularly the water-system, the suckers of which pierce the plates as
they do in the embryonic stages of other Echinoderms. In fact, the
external limestone plates forming the test of a Sea-urchin, the reticulated
network of the actinal and abactinal surface of a Starfish together with
the ambulacral and interambulacral plates and the plates forming the disk
of an Ophiuran, the upper, lower, and side arm-plates, as well as internal
skeleton, are all directly derived from the simple system of limestone
plates of the abactinal surface of the Echinoderm embryo. This system
consists, in all cases, of a basal plate, five radial and five interradial
plates. In Ophiurans the genital plates are formed from the angles
of the five interradial plates ; similar plates can still be traced in the
young Starfishes, while in the full-grown Starfishes their presence is
shown by the interbrachial partition, on each side of which the ovaries
discharge. Thus there exists a complete homology between the genital
plates of Ophiurans and the interbrachial partitions of Starfishes, a ho-
mology fully carried out in its details when we examine the relations
held by the genital jjlates to the ovaries in Ophiurans and by the inter-
brachial partitions to the ovarian openings in Starfishes.

From the primitive number of plates existing in the disks of all em-
bryo Echinoderms, it is evident that palaeontologists have laid altogether
too much stress upon the arrangement of the plates of the arms in
Crinoids. The study of the -solid parts of Starfishes, while valuable as
accessories, would certainly furnish no very satisfoctory data for a classi-
fication, at least if this were based entirely upon an examination of the
hard parts of the abactinal system alone, as is so frequently the case in
Crinoids.

HARD PARTS OF SOME NORTH AMERICAN STARFISHES.

ASTERIAS.

In the genus Asteracanthion (Asterias) the true character of the plates
of the abactinal and actinal surfaces is far more difficult to trace than in
other genera where the plates retain more or less homogeneous features.

In Asteracanthion, although in the younger stages (as shown in Plate
VIII.) the reticulation consists entirely of plates readily distinguished
one from the other, yet in the adult the plates have become changed
to a mere irregular network anastomosing in all possible directions (PI.
IX. Fig. 3), and thus rendering it quite difficult, if not frequently impossi-
ble, to trace the connection of the actinal and abactinal reticulation with
the interambulacral plates.

In the majority of species of this genus the plates adjoining the inter-
ambulacral plates are cross-shaped (PI. IX. Fig. 6), connecting with ad-
joining plates at three ends, in front, behind, and towards the abactinal
surface ; the other end connects with the interambulacral plates. These
plates lose their regularity as they ascend towards the abactinal side on
the edge of the arms, the prongs becoming gradually short processes,
and finally simply rods or irregularly shaped plates all more or less
imbricating (PI. IX. Fig. 4). The spines are generally attached to the
rods by a very shallow socket fitting into a rudimentary tubercle and ring.
The spines of the interambulacral plates are movable, those of the actinal
and abactinal less so, and frequently soldered to the reticulation.

Asteracanthion berylinus.

Asteracanlhion berylinus* Ag., A. Ag. 1863. Proc. Amer. Acad. Boston.
Asterias Forbesii Des. 1848. Proc. Bost. Soo. Nat. Hist., UI. p. 67.

PL IX.

The base of each of the interambulacral plates at its junction with
the ambulacral plates is marked by a pore for the passage of a water-

* For the typography used to explain the synonymy, see A. Agassiz, Kevision of the Echini, p. 26_

NOETH AMEEICAX STAEFISHES. 95

tube (PI. IX. Figs. 5, 6) between this plate and the following reticula-
tions, forming a part of the sides of the arms ; similar pores are found
arranged like the first row of pores in a line parallel to the longitudinal
axis of the arms. The other water-pores are ii-regularly placed over the
surface of the arms.

Near the mouth the interambulacral plates come together in the angle
of the amis and form the mouth-papillae so called. The}' are readily
seen in a Starfish examined from the lower side (PI. IX. Fig. 5) when
denuded of spines. Seen from above (PI. IX. Fig. 6, an interior view),
the connecting plate between adjoining ambulacral systems is formed
by the rising of the outer edge of the plate (the outer pore not being
present) towards the limestone network formed by the junction of the
interambulacral imbricating pieces which constitute the framework of the
abactinal system.

This structure is best seen in large specimens of A. vulgaris, in which
the alternate arrangement of the ambulacral plates commences at once at the
base of the arms, and where the interbrachial fold at the angle of the arms
is high and well set off from the pores left for the passage of water-tubes.

The spines placed on the junction of the interambulacral plates (a", PI.
IX. Fig. 5) form the papillte (PI. IX. Fig. 6), near the actinal opening ;
they differ in no respect from the other spines.

The arrangement of the pores in double rows (PI. IX. Figs. 5, 6) for
the passage of the ambulacral suckers is, as is Avell known, only due to
age, owing to the crowding of adjoining plates ; in large specimens there
is no trace of the original linear arrangement of the ambulacral poi-es
beyond the plates nearest the actinostome or at the extremity of
the arms. But while the ambulacral pores thus alternate, the plates
themselves extend entirely across from the median line to the interam-
bulacral plates ; they are wedge-shaped, the broad and pointed ends
of adjoining plates alternately extending to the median line of the arm
or to the edge of the interambulacral plates (PI. IX. Figs. 5, 6). See
a note on the fecundation of A. berylinus and A. vulgaris in Archives de
Zool. Exper., which suggests a plausible cause for the great number of
varieties of the genus Asterias. A. Forbesii (berylinus) extends from Hali-
fax, N. S., to Florida, while A. vulgaris (pallidus) has a more limited
southern range, and extends farther noi-th, from Labrador to Long Isl-
and Sound. lu Massachusetts Bay the two species are about equally
common.

96 DESCEIPTIOX OF THE HARD PAETS OF SO]ME

Asterias ochracea.

Asterias ochracea Br. 1835. Prodrom.

JPl. XI.

The striking diiFerences which apparently exist on a cursory examina-
tion of the species allied to A. ochi-acea are not found to be of suffi-
cient importance, when analyzed, to warrant us in considering the genus
Pisaster, as recognized by Professor Agassiz, anything more than a con-
venient systematic generic subdivision ; the special points of difference are
the great width of the ambulacral system, its elongated plates, the breadth
of the furrow forming the median ambulacral ridge (seen from the
interior) (PI. IX. Fig. 5), the proximity of the openings for the passage
of the ambulacral tubes on each side of the median ridge, with the cor-
responding slender interambulacral plates carrying only one row of lono-
spines at the outer extremity. When denuded of spines the reticulation
of the actinal surface of the arms adjoining the interambulacral plates
forms a close pavement with small interstices (PI. XI. Fig. 4); the tubercles
carrying the spines ai-e arranged in three or four rows at right angles to
the longitudinal axis of the arm ; they have a deep slit at the top of
the boss ; these tubercles are connected laterally by a comparatively low
ridge. In the reticulation of the abactinal surfixce of the arms the
primary spaces are quite large, but these are greatly subdivided by a
secondary system (PI. XI. Figs. 1, 2) (more or less prominent), consisting
of smaller plates, most irregular in shape, which encroach upon the
primary areas and subdivide them again, or materially reduce the area
through which the water-tubes can be protruded. The reticulation of
the actinal surface carries large club-shaped spines of moderate length,
Avhile the spines of the upper surface are shorter but similarly shaped,
presenting the appearance of having been ground down so as to form
nearly continuous walls on the separating ridge of the reticulations (PL
XI. Fig. 1). The interambulacral papillse are generally cylindrical, some-
times pointed or somewhat club-shaped at the tip, contrasting with the
generally flattened and slightly spatulate interambulacral spines of
Asteracanthion proper. The interbrachial partition (PI. XI. Fig. 5) is
naturally very well developed, owing to the great number of narrow
interambulacral plates from which the brachial reticulations arise. The
whole reticulation of the arms is far more solid than m any other
group of species of Asteracanthion (Asterias) ; compare PI. IX. Fig. 6, the

NOETH AMEEICAN STAEFISHES. 97

corresponding interbracliial partition of A. berylinus, and also PI. XI. Fir/. 4,
the prolongation of the plates separating adjoining ambulacral systems
in A. ochracea and in A. berylinus (PI. IX. Fir/. 5). To show the differ-
ence in the thickness of the limestone reticulation of the abactinal and
actinal systems compare PI. XL Fig. 3, and PI. IX. Fig. 4, which are simi-
lar views of the interior of the abactinal systems of A. berylinus and A.
ochracea. or compare the horizontal sections shown in PI. XI. Fig. 5, and
PI. IX. Fig. 6. The range of Asterias ochracea is from Sitka to San
Diego, California ; it is the most common species of Starfish on the coast
of California.

Echinaster sentus.

Asleiias sentus Say, 1825. Journ. Acad. Nat. Scien. Phila., V. 143.
Echinaster sentus Verr. 1867. Notes on Radiata.

Fl X.

The meshes of the abactinal limestone network are larger than in
Asteracanthion, especially near the centre of the disk, where the irregu-
lar polygonal spaces covered by the abactinal membrane are quite large
(PI. X. Fig. 3). The same loose structure extends a short distance along
the abactinal surface (PI. X. Fig. 4) and the sides of the arms ; but
towards the extremity the meshes become smaller, and on the actinal
side, immediately adjoining the interambulacral plates, the limestone work
is quite compact (PI. X. Figs. 5, 6), and leaves only a few small openings
for the passage of the water-tubes.

In addition to the water-tubes in the actinal surface of the arms,
there is a row of very large tubes (PI. X. Fig. 1') passing between the
interambulacral plates. The madreporic body differs considerably from
that of Asteracanthion, and is not as well separated or as distinct from
the general abactinal surface as is the case in that genus.

The interambulacral plates, forming the so-called teeth, are larger than
the others ; they form the extremity of the single lateral rows (PI. X. Fig.
5), and do not make a partition or division-wall between adjoining am-
bulacra, as in Asteracanthion proper, the actinal part of the limestone net-
work extending nearer the actinostome. The solid character of the
actinal part of the limestone network covering the arms is well shown
in an interior view (PI. X. Fig. 6). This figure also shows how far the
ambulacral and interambulacral plates become soldei'ed together with the

08 DESCEIPTIOX OF THE HAED PAETS OF SOME

actinal limestone network. The small size of the first set of ambulacral
plates is characteristic of the genus as well as of other Starfishes with
two rows of suckers ; the plates of the actinal are scarcely more promi-
nent than the other ambulacral and interambulacral plates, forming a strik-
ing contrast to the immense development they take in Asteracanthion
and allied genera. The interambulacral plates, as is well shown on PI. X.
Fi(/. 5, are remarkably uniform in size ; the secondary ambulacral plates
forming the brachial limestone network adjoining them are compactly
soldered together. In this genus the spines of the limestone network
are completely sheathed by the outer membrane covering the whole abac-
tinal and actinal system (PI. X. Figs. 1, 2) ; they are large, sharply pointed,
generally placed only at the angles of the limestone polygons, and form
irregular longitudinal rows, from the central part of the abactinal part
of the disk, gradually diminishing in size towards the extremity of the
arms. This species is particularly abundant in the West Indies and Florida,
and extends northward to New Jersey.

CROSSASTER.

Crossaster M. T. 1840, Monatsb. d. Akad. Berlin, (emend.) A. Ac.

The genus Crossaster, as originally established by Miiller and Troschel
in the Monatsbericht d. Akad. d. Wiss. of Berlin, was identical with So-
laster of Forbes, which had the priority of a year. In the System d. Asteri-
den, Miiller and Troschel adopted Forbes's genus. From an examination
of the hard parts, it is evident that Solaster papposus and Solaster endeca
should not be included in the same genus, having reiilly nothing in com-
mon beyond the great number of arms.^ The accompanj-ing descriptions
will fully show my reasons for placing these two species in different gen-
era. In order not to multiply names, I have retained the genus Cros-
saster, which is quite closely related to Pycnopodia, only limiting it to S.
papposus, and have kept Solaster for S. endeca and its allies, which are
more nearly related to Cribrella.

NOETH AMEEICAX STAEFISHES. 99

Crossaster papposus.

Crossaster jxijijiosus 51. T. 1840. Monatsb. d. Akad. Berlin.

FL XII.

In Crossaster the membrane covering the abactinal system, like that
of Pycnopodia, forms a mere film, bnt it is strengthened by a regular
reticulation, with open meshes, carrying, at the points of junction of the
horizontal limestone plates, prominent club-shaped processes, upon the tip
of which are attached minute slender spines, forming more or less promi-
nent tufts (PI. XII. Fi(/s. 3, 4). The interbrachial partition consists of a
membrane, without limestone jilates, extending towards the base of the
arms, connecting the few limestone plates reaching from the actinal plate
to the abactinal surfiice with the triangle formed by the rising of the
actinal floor at the point of junction of adjoining arms (PI. XII. Fir/. 3).
The actinal floor, with the exception of the plates of the interambulacral
area, is entirely composed of a compact pavement formed of small irregu-
larly shaped imbricating plates, gradually passing into the open reticula-
tions of the abactinal surface, along the sides of the arms (PI. XII. Fi^. 2).

The ambulaci'al plates of this genus are broad and well separated; the
ambulacral groove is broad and prominent (PI. XII. F/'ff. 2) ; at the junc-
tion of the ambulacral and interambulacral plates the former are well
separated ; they are pointed, bulging in the central portion, leaving a
wide opening for the passage of the sucker. The basal plates take an
unusual development, forming a prominent ring round the actinostome ;
they are well separated by the interbrachial basal plates, forming the
base of attachment to the limestone plates, which constitute the basal
part of the interbrachial arch (PI. XII. Fiffs. 2, 3). The actinal side of
the interambulacral plates forms a series of slightly curved plates, at
right angles to the ambulacral groove, carrying tubercles diminishing
in size as they recede from the edge of the arms ; these plates form a
prominent row along the edge of the arms on the actinal surface (PI.
XII. Fiff. 2) ; the tubercles of the interambulacral plates, arranged in
narrow belts, carry slender spines, similar to those of the tufts of
minute spines found on the abactinal surftice. The basal interambulacral
plates, like their corresponding ambulaci-al plates, are immensely developed,
projecting far into the large actinal ring, and carrying, like all the
interambulacral plates, long, slender spines; these form powerful papilla3,

14

100

DESCEIPTIOX OF THE HAED PAETS OF SOME

' P

Britain,
on the

^'^' ^■' surrounding the mouth ; though their use, as in

all Starfishes, is evidently very limited, the prin-
cipal Avork of digestion being done by the
stomach itself, which folds over the substance
to be introduced, and thus gradually dissolves it.
The accompanying woodcut (F/'ff. 1) shows
somewhat more plainly than F/(/. 3 on PL XII.
the plates composing the parts round the acti-
nostome and base of the arms.

This species is conmion to both sides of the
Atlantic ; it is found in Norway, Denmark, Great
on the west coast of France, in Ireland, Greenland, and extends
east coast of America as far south as Massachusetts Bay.

Pycnopodia helianthoides.

Pycnopodia helianthoides Stimps. 1861. Proc. Boston See. Nat. Hist., Vni.
Aslerias helianthoides Br. 1835. ProJrom.

PI. XIII.

In Pycnopodia the opening at the end of the large ambulacra! plate
near the actinostome is best seen in profile {Fig. 2) ; it differs in no
way from the structure of the corresponding plates in Asteracanthion,
though apparently, on first examination, the actinal plates forming the
actmal ring seem quite peculiar, owing to the disappearance near the
mouth of the interbrachial membrane, and the isolation of the interbra-
chial partition ; this connects the actinal and abactinal I'eticulated surfaces
by a mere film only. The large plate at the actinal ring, forming the
base of the interbrachial partition (PI. XIII. Fig. 6 i p), is entirely discon-
nected from the interambulacral system, as can easily be seen by an
examination of the actinal extremity of the arm from the inside of

* Fig. 1. Crossa.iler papposus. — Internal view, seen from above, the abactinal surface of the plates at
the base of one of the arms, round the actinostome, removed, c c, attachment of muscular bands con-
necting adjoining ambulacral plates ; a o, a o, openings for passage of ambulacral suckers ; a'b, a'b, am-
bulacral plates with projecting spur ; i m, termination of film forming the interbrachial partition ; c'p, spur
from interambulacral plates forming connecting floor of actinal surface ; c'J, large muscular plate at base
of the arm ; ip, interbrachial plate to which the film forming interbrachial partition is attached (when
it extends to that point) at the base of the arms ; s p, spur of interambulacral basal plates forming
the base of attachment of the interbrachial partitions ; in p, interambulacral plates forming the so-called
mouth-papilla; }>', spines of mp.

NOETH AMEEICAX STARFISHES.

101

OlS

'OLtf)

the actinal ring, showing the plate rising up on the side of the two
large ambuLicral pLates of the actinal ring. The
interambulacral plates (PI. XIII. Fig. 7) form small
scale-like plates near the base of the arm, carrying slen-
der pointed spines ; they increase somewhat in size
at a distance from the base. They are followed on
the edge of the arm by two lozenge-shaped plates,
with extended points, carrying large club-shaped
spines forming a thin low wall for the support of the line of at-
tachment of the abactinal membrane covering the abactinal surftice of
the arms. This membrane extends also over the central part of the
disk ; over the abactinal surface it is strengthened here and there by
a few small limestone plates or rods, placed at the base of the large
spines ii-regularly scattered on the surface of the abactinal region ; these
plates sometimes form in the disk a very irregular disconnected retic-
ulation, the lines of which are composed of small irregularly shaped
rounded plates. "Within the space where the arms are united the am-
bulacral plates rise nearly vertically, but towards the extremity they
gradually slope more and more, inclining towards the actinostome, so that
the ambulacral plates form a hard flat area, occupying neai'ly the whole
of the actinal surface of the arms.

The genera Pycnopodia and Crossaster are specially interesting on
account of the close relationship they have to Brisinga. In fact,
•compared with Brisinga, they prove conclusively that the latter genus,

* Fifj. 2. — Profile view of actinal extremity of arm of Pycnopodia (the dermal covering of arm removed).
c'r, small plates forming the ridge, covering the ambulacral groove,
repeattd along the whole length of the arm. a'c, the correspond-
ing plate of the second interambulacral plate. The upper projecting
part of this plate is the support of the basal plate of the interbr.ichial
partition, and below it is seen the plate which strictly corresponds to
it. This plate a'c also covers in part the plate a's of the basal am-
bulacral plate a, at the base of which is situated the interambulacral
plate a ip carrying the short spines forming the moxith-papilla^.

The relative position of the plates a's to the, basal plate of the
interbrachial partition is well shown in Fig. 3, which represents an
inside view, seen from above, of a part of the actinal ring, a's, a'c,
the same as in Fig. 2, i]), the interbrachial partition, reduced in this
genus to a few scale-like plates, supported on spurs of the interam-
bulacral plates, c, longitudinal groove, line of junction of ambulacral plates, c'r, ridge of connecting
plates forming groove c'. a o, opening for passage of ambulacral suckers.

102 DESCRIPTION OF THE HAED PAETS OF SOME

far from being so exceptional in its structure as has been generally
supposed, is structurally very intimately connected with Pycnopodia and
Crossaster. We might readily transform a Pycnopodia or a Crossaster
into a Brisinga by reducing the actinal and abactinal interbrachial spaces
into a minimum, which would give us a Starfish with a small disk, in
which the ambulacral plates adjoining the actinostome assume a great
development, and thus the numerous arms would appear quite discon-
nected (as in Brisinga). The connection of the arms in Starfishes does
not depend so much on the greater or less development of the ambu-
lacral and interambulacral systems, as upon the greater or less increase of
the limestone network forming the interbrachial spaces, which, although a
feature greatly affecting the physiognomy of the Starfish, yet influences
but slightly its internal structure.

The range of this species is fi'om Sitka to Mendocino City, California. In
the Gulf of Georgia and at Mendocino it is a very common species in shallow
water and at low-water mark.

BEISINGA.

The genus Brisinga, with its long slender arms, the whole actinal side
of which, with the exception of the large interambulacral plates forming
the edge of the arm, is occupied by the ambulacral plates, shows vis very
distinctly how we can pass from the Starfish to the Ophiuran by the
joining of the large interambulacral plates on the lower surface, and their
becoming soldered together into one plate to form a lower arm-plate, so
that the absence of interambulacral plates, which has always been cited
as the great difference by which Starfishes and Ophiurans could always
be distinguished, is readily explained ; the lower arm-plates of Ophiurans
being only modified interambulacral plates. We further find, on examin-
ing, in Brisinga, the secondary imbricating plates forming the arches
which support the abactinal membrane covering the arms, each of which
carries only a single spine, and is arranged in more or less regular
curves, that we have some ajjproach already to the side arm-plates
* so characteristic of Ophiurans, the separation of the so-called disk from
the arms, — which, although so striking a feature of the genus, is less im-
portant than it seems at first sight, — being merely brought about by the
reduction to a minimum of the lateral spreading of the actinal part of
the secondary and interambulacral plates. In the case of Brisinga this

XOETH AMEEICAX STAEFISHES. 103

foi'ms at once an arch over the arms without expanding, as in other
Starfishes, into a flat actinal floor of greater or less extent on the side of
the interambulacral plate, from the extent and shape of which the vari-
ous families are determined. This obtains its maximum of development
in the extreme forms like Culcita and Palmipes ; in one case the actinal
and abactinal floors are well separated, in the other closely united hy
vertical shafts and walls. The analysis of the structure of Brisinga gives
a most satisfiictory explanation of the general homologies existing be-
tween Starfishes and Ophiurans, and reduces the gap hitherto unfilled
between Starfishes and Ophiurans to a comparatively unimportant method
of development.

As the madreporic body of Starfishes is placed in one of the interbrachial
arches, and this arch reduced as it often is to a minimum (Solaster). or lim-
ited sometimes to the mesenteric support of the stone canal, we have a ready
explanation of its position in the Ophiurans on the homologous plates in
the interbrachial spaces, namely, the single plate in the continuation of
the line of the interambulacral plates; at the same time the homology
between the genital plates and the single interbrachial plate found in
some Starfishes at the angle of the arms is fully carried out, as it is
well known that it is on each side of the interbrachial arch that the
ovarian openings are found. An examination of the base of the arms of
Brisinga near its junction with the disk shows already quite a constric-
tion, and of course a corresponding reduction in the length of the inter-
brachial arch. The great extension of the interambulacral plates across
the space covering the ambulacral canal reduces it to its minimiuu at
that point. The mode of articulation of the ambulacral plates of joints
in the arms of Brisinga has been compared rather with Ophiurans than
with Starfishes, but the articulation of the internal skeleton in Ophiurans
is not specially different, although somewhat more perfect, from the artic-
ulation of the joints of arms in the Starfishes proper, and the homology
between the internal skeleton of Ophiurans and the ambulacral system
of Starfishes can be clearly established. If we imagine for each primary
interambulacral plate but a single row of secondary interambulacral plates
composed of a small number of plates, we shall of course have a side arm-
plate and an upper arm-plate ; the lower arm-plate being formed of
the opposed interbrachial plates, which have become soldered together,
and through which the tentacles have pierced their way.

104 DESCEIPTIOX OF THE HAED PARTS OF SOME

The abactinal membrane of the disk of Brisinga is eminently Asterian ;
it is only sHghtly strengthened by a few minute limestone plates, as is
the case in Crossaster and Pycnopodia, and in spite of the general re-
semblance, at first glance, of this well-defined disk to an Ophiuran disk,
we have nothing whatever corresponding to the arrangement of the cen-
tral i^lates so characteristic of the disk of Ophiurans. But we have in
a great many genera of Starfishes the central part of the disk, show-
ing in the young stages only, as regular an arrangement of the plates
of the abactinal system as in any Ophiuran, though it is lost in the
adult. Such a young stage is figured in PI. VIII., a corresponding stage
has also been recently figured by Loven in his Memoir on the Echini
(1875), and a similar structure of the disk will undoubtedly be found to
exist in the very youngest stages of each genus, as it seems to be a
general structure of the young of all Starfishes, as far as observed.

While Brisinga is a most important form, as showing the relationship
between Starfishes and Ophiurans, there certainly is nothing in its struc-
ture or in its affinity to Pro taster to warrant the paloeontological importance
ascribed to it by the younger Sars ; and it cannot be considered, any more
than several other genera of Starfishes now living,* as the representative
at the preseiit day of the oldest-known Echinoderm. I think we can
show from the study of the hard parts of Starfishes that they have been
a remarkably persistent tj^pe, and that the apparent changes of form
due to the excessive increase or diminution of the interbrachial lime-
stone deposit is a very secondary feature, which, though greatly modi-
fying the external appearance of the Starfishes, yet does not affect the
main structure, which, as has been stated, is remarkably uniform through-
out the order. "While fully admitting the many important points (so
well brought out by Sars in his Memoir on Brisinga) wherein the genus
differs from the other Starfishes, yet I must call his attention to the
fact that many of the structural details which he strongly insists upon
as specially characteristic of Brisinga t are common to the other Starfishes,
and do not constitute features by which this family can be contrasted
with the remaining Starfishes.

* Pycnopodia, Crossaster.

f For an opportunity of examining both dry and alcoliolic specimens of Brisinga, I must thank Sir
C. Wyville Thomson, and Dr. G. O. Sars. Brisinga endecacnemos is found in deep water off the Lofoten
Islands, Norway. It has been collected by the " Challenger " in eighty fathoms, on the La Have Bank off
Kova Scotia.

XOETH AMEEICAX STARFISHES. 105

Linckia Guildingii.

Lhtckia Guildingii Gray, 1840. Ann. Mag., Vl.

PI. XIV. Fifjs. 1-6.

A longitudinal section of one of the arms (PL XI\\ Fig. G) shows the
great thickness of the irregularly shaped polygonal plates composing the
limestone network of the abactinal surface. The plates (as seen in Fig. 4,
PI. XIV., when they are denuded of the finer granulation covering them,
PI. XIY. Fig. 1, and the intervening spaces) are very closely packed ; the
processes connecting plates laterally often do not exist in the median space
of the arm, and appear only as short rods along the sides of the arms and
on the outer edge of the lower surface (PI. XIV, Fig. 3), where they are
closely packed, forming in older specimens a compact pavement, and
losing on this surfiice the imbricating aiTangement to be traced only
along the sides of the arms or to be seen in a transverse section. The
fine granulation mentioned above extends over the whole actinal surface
of the arms, concealing almost completely the three to four longitudinal
rows of small plates • immediately succeeding the interambulacral plates
(compare Figs. 2 and 3, PI. XIV.; see also Fig. 2').

The top of the large papilliB (PI. XIV. Fig. 6) attached to the interambu-
lacral plates forms a close pavement when seen from the actinal side ;
these large pajiillte pass very rapidly into the minute granules covering the
lower side of the arms (PI. XIV. Fig. 2). Toward the actinostome the pa-
pillae flare inwardly, forming several rows placed one behind the other, and
appear, when seen in section, as if there were a series of secondary inter-
ambulacral plates forming the mouth-papillte, but on examination we find
that the structure of the actinal interambulacral plates is that of other
Starfishes. Seen from the interior on the actinal floor, the interbrachial
plate is sunk far below the level of the ambulacral groove ; the inter-
brachial arches are reduced to the thickening produced by the junction
of the arms, which extend in wedge shape a short distance toward the
actinal ring ; the space in which the limestone canal is situated alone
connecting by a low ridge with the actinal ring.

This is specially a West-Indian and Florida species.

106 DESCEIPTIOX OF THE HAED PARTS OF SOME

Asterina folium.

Aslerbia folium Lutk. 1859. Vidensk. Meddel.

PL XIV. Figs. 7-9.

In Asterina, as in the bulk of the pentagonal Starfishes, the great
lateral development of the secondary interambulacral plates introduces
some modifications in the structure and position of the hard parts in this
genus. The plates forming the actinal and abactinal floors are irregularly
lozenge-shaped, imbricating at the extremities of the adjoining points,
leaving thus a greater or less free space for the passage of the water-
tubes ; the plates of the two floors at the ridges of the disk become sol-
dered together, thus forming, as it were, a new system of plates, which in
some genera are regularly arranged and often furnish characteristic spe-
cific distinctions. At the actinal ring the interbrachial arches exist only
as columns rising directly from the actinal to the abactinal floors in the
interambulacral spaces. As in all the species of the genus, the lozenge-
shaped plates of the actinal and abactinal surface carry short slender
spines, with a more or less regular fan-shaped arrangement on the abac-
tinal side ; the spines are less numerous on the actinal side, some-
what longer on the interambulacral plates, especially near the actino-
stome, forming mouth-papillis of considerable prominence (PI. XIV. Figs. 7
and 7').

The simple structure of the short, pointed tei'minal tentacle at the ex-
tremity of the arm is well seen in this genus (PI. XIV. Figs. 8', 8); adjoin-
ing tentacles are, as in Asterias, long, slender, without a prominent suck-
ing-disk. The water-tubes are specially large in this genus (PI. XIV.
Fig. 9).

This species has the same range as Linckia Guildingii.

Asteropsis imbricata.

Aslerupsis imbricata Grube, 1857. Wiegm. Archiv., XXIII.

PI. XV.

The abactinal limestone reticulation of this species consists of flat, irreg-
ularly star-shaped plates, from which diverge longer flat pieces, connects
ing the adjoining centimes of radiating plates (PI. XV. Fig. 2); the plates
and their connecting links are all imbricating. Towards the central part
of the disk the larger spaces between the rods are partially closed by
shorter spurs, and are further separated by disconnected plates into ellip-

NORTH a:meeicax staefishes. 107

tical areas, through which the water-tubes are protruded (PI. XV. Figs. 1, 2);
the network becomes closer towards the tip of the arms, and there are a
great number of small areas for the passage of the water-tubes (PI. XV.
Fiff. 3). In the living state the limestone skeleton is deeply imbedded in
a thick epidermis, completely covering the upper and lower surface of
the disk. Compare PL XV. Figs. 1, 1', with the figures PI. XV. Figs. 2, 3,
showing preparations of the plates of abactinal surface from the exterior
and interior. The interbrachial arches are I'educed in this genus to a
mere vertical column, consisting in portions of not more than a single
plate placed close to the actinal ring, and leaving a large open space be-
tween it and the edge of the arm (PI. XV. Fig. 6).

The plates of the actinal floor form a regular pavement, diverging from
the interbrachial angle parallel to the axis of the arms ; the actinal and
abactinal systems of plates form, at their junction on the edge of the arms,
a double row of large plates forming a binding at the periphery (PI. XV.
Fig. 4), one row placed on the actinal side, the other on the abactinal
side (see PI. XV. Figs. 4 and 2). There is a well-marked abactinal ori-
fice near the centre of the disk (PI. XV. Fig. 1).

In Gymnasteria, otherwise closely related to Asteropsis, there is no
special difference between the plates of the actinal and abactinal systems ;
they are more distinct, and not arranged quite so regularly as to form a
pavement.

In the greater number of the pentagonal Starfishes we find the same
general distinction between the pavementrlike plates of the actinal
side, extending to the junction of the actinal with the abactinal sys-
tem, although we do not always find so regular a peripheric series of
plates. This is the case in Culcita (see Figs. 4, 5). Where the actinal
plates acquire a great thickness, forming a lower floor through which
the passages between the plates and beams make an intricate system of
openings placed at different levels {Fig. 4), while the abactinal system is
reduced to a comparatively simple series of rods having the general ar-
rangement of triangular network with the longer or shorter rods separatr
ing them set on edge and imbricated {Fig. 5). The whole limestone sys-
tem is, as in Asteropsis, entirely imbedded in the thick epidermal layer
in which the plates have been deposited, so that but a trace of the lime-
stone network appears when seen either in a natural condition or merely
in dried specimens. The interbrachial arch of Culcita is reduced to a few

15

108

DESCEIPTION OF THE HARD PAETS OF SOIVIE

vertical plates rising close to the actinal ring from the interbrachial in-
terambulacral plate.

Fig. 4* Fig. 5.t

This species of Asteropsis occurs on the West Coast, from Vancouver's
Island to San Francisco.

The general arrangement of the limestone plates of the European As-
teropsis pulvillus does not differ materially from the one here figui'ed.
The European species carries spines on the edge of the marginal as
well as of the interambulacral plates, and the interbrachial arch is more
fully developed, separating adjoining arms more completely.

Pentaceros reticulatus.

Pentaceros reticulatus Linxk, 1733. De Stellis Marinis.

PL XVI.

Figs. 1, 2, 4, 5, on PI. XVI., show the extent to which the lime-
stone network of the actinal and abactinal surfaces is hidden by the
spines and granules covei'ing the plates and rods of the two surfaces.
Seen from the abactinal side, the network consists of a central plate
more or less hexagonal, with projecting angles connected together by
short stout rods, overlapped by the plates, so as to form an open tri-
angular network, covered by minute granules, and in their interstices
giving passage to the thickly clustered water-tubes (PI. XVI. Fig. 1).
The same granulation covers the plates and rods as well as the greater

* Fig. 4 is an interior view of tbe actinal plates of a Culcita, as seen in an alcoholic specimen, of
■which the abactinal floor has been removed to show the intricate system of passages lying between the
limestone rods, a b, abactinal system ; s, section across a limestone rod ; f, interior muscular sheath con-
necting limestone rods; o, passages between adjoining plates.

f Fly. 5, edge of a Culcita, to show the gradual passage of the open reticulation of the abactinal
surface into the closely packed vertical wedges mp, forming the outer edge of the disk of a Culcita.

NOETH AMERICAX STARFISHES. 109

part of the spines scattered over the abactinal area, leaving but a short
piece of the end of the spines bare. The meshwork of course becomes
closer towards the extremity of the arms ; the plates and rods are of
considerable thickness, as is seen in their section along the edge of the
arm (see Fig. 5, PI. XVI.).

The junction of the actinal and abactinal systems forms a double row
of large contiguous plates carrying a heavy spine (see Figs. 4, 5, PI. XVI.).
The pavement-like plates of the actinal surface are arranged in rows
parallel in a general way to the longitudinal axis of the arms (PI. XVI.
Fig. 4), and also in indistinct rows at right angles to this (PI. XVI.
Fig. 2). They are well covered by a coarser granulation than that of
the abactinal surface, the central part of the plate carrying a cluster of
three to five larger granules, becoming in some cases nearly fixed spines ;
these granules, on the actinal surface of the interambulacral plates, become
a large flat pointed movable spine, with smaller flat lateral spines, rounded
at their extremity. Round the edge of the interambulacral pieces forming
the jaws they increase materially in size, becoming very prominent mouth-
papillEB (PI. XVI. Figs. 2, 3).

In all the pentagonal Starfishes the fact that the jaw pieces are simply
the modified interambulacral plates of the last joint is very apjDarent, as
well as that the interbrachial plates forming the base of the interbrachial
arch are also only a modified part of the interambulacral plates formed
by the soldering together of the inner lateral spaces of the opposite
interambulacral plates of the joint of the jaw.

The interbrachial arches are composed of comparatively few lai'ge solid
plates ; their breadth varies materially in difierent specimens, either nearly
filling the whole space between the actinal ring and the angle of the
arms, or limited to a shorter wall next to the mouth. The ambulacral
system is composed of tall plates rising well above the actinal floor,
forming a broad median groove, seen from the abactinal side (PI. XVI.
Fig. 5) ; when seen in profile (PI. XVI. Fig. 7), large elliptical spaces are
left for the passage of the powerful ambulacral suckers (PI. XVI. Fig. 2).
The interambulacral plates are large, distinct, and of great thickness, with
their actinal face well developed (see Figs. 4, 7, PI. XVI.) ; the last joints
of the plates of the actinal ring are prominent, raised high above the
interbrachial plates. The jaws are large, projecting far into the central
actinal space (PI. XVI. Figs. 4, 5, 7); the papillae when extended meet.

110 DESCEIPTIOX OF THE HAED PARTS OF SOME

nearly closing the actinostome, only leaving (PI. XVI. Figs. 2, 3) a small
pentagonal opening.

The interambulacral spines, when the suckers are drawn in close, com-
pletely cover the ambulacral furrow (see Mg. 2, PL XVI., where they
are closed over a portion of the ambulacra at the base of one of the
anns).

Pentaceros reticulatus is found on both sides of the Atlantic, at Cape
Verde Islands, and in the West Indies, extending north to South Carolina.
Several other West India species of Echinoderms are also found at Cape
Verde Islands and on the main coast opposite.

In the pentagonal Starfishes the plates forming the so-called jaws are
huge interambulacral plates extending far towards the centre of the
mouth, where they nearly meet, to form, with the papillae, the so-called
jaws and teeth of Starfishes. So far we have not been able in any way
to homologize the teeth of Echini with any of the solid parts of Star-
fishes or Ophiurans; the auricles of the regular Echini and the peculiar
spur of the interior of the test near the mouth of some Spatangi being
the only processes which appear to have analogous position. For a com-
parison of the Starfish mouth parts with those of Ophiurans compare the
figures here given with those of Lyman in the Bull. Mus. Comp. Zool.,
Vol. III., from which it is evident that in Starfishes and Ophiurans the
mouth parts are strictly homologous, and are formed by the terminal
ox-al interambulacral plates. Compai-ing profile figures of the oral ex-
tremity of one side of an arm in Culcita {Fig. 7), Acanthaster {Fiff, 6),
and Solaster {Fig. 8), we cannot fail to be struck with the great size
of the terminal oral interambulacral plate aip, carrying the mouth-
papillte mp.

In the views of the arms, seen from the interior (the abactinal sys-
tem being removed), the great development of the oral terminal plate
(a'c) is well shown. In Fig. 6', Acanthaster, Fig. 8', Solaster, and Fiff. 9,
Anthenea, the lettering corresponds to the profile figures. The only addi-
tional notation introduced is //> for the interbrachial partition, and i p b
for the spur forming the basal plate of the interbrachial partition. The
mobility of the arms of Starfishes depends entirely upon the comparative
width of the ambulacral and interambulacral plates compared in their
length, upon the solidity and extent of the interbrachial partition, and
the extent to which the abactinal system corresponds in its articulation

K^OETH AMEEICAX STAEFISHES.

Ill

Fig. 6.*

OL't

to the number of plates of the ambulacra! system. Genera where theie
are large marginal plates, as in Astropecten, and the pentagonal Star

Fig. 8'.

M"" C' fteV ■*■«■

(A I.

Pig. 9.

«lT

ai/p

at at v.-^ I

* In Figs. 6, 7, 8, c' r are the small plates or spurs forming the ridge covering the ambulacra! groove ;
a' c, the corresponding plates of the ambulacral plates a' ; ao, the opening for passage of ambulacral ten-
tacle ; a i, the interambulacral plates carrying papiUae p. The large terminal plate a'c at the actinostome
encroaches upon the corresponding ambulacral plate, so as to cover it by a spur a' s, so much so that,
when seen from above, this plate appears to be directly connected with the terminal interambulacral
plate a ip, but the profile view clearly shows in all these genera that the terminal plates, although of
different proportions, do not differ in their arrangement from those forming the body of the arm..
The spur a' s of the terminal oral plate a'c sometimes forms a secondary cup-shaped plate along the
edge of the anus ; in Figs. 7, s a', and 8, a's, this plate is quite prominent, while in Fig. 6 it is less
developed.

112 DESCEIPTIOX OF THE HARD PARTS OF SOME

fishes, or Starfishes in which the abactinal system is stifFenecl by heavy
interbrachial partitions extending from the oral ring to the angle of the
arms, or Starfishes where the abactinal reticulation is extremely solid, as
in Ophidiaster, are all capable of but slight movements. On the contrary,
Starfishes in which, as in most of the Asteracanthidse, the
reticulation is loose, the interbrachial partitions reduced often
■g^ to a film, or to a mere arch of limestone plates, and in wliich
the ambulacral plates are high, are much more flexible.
The extremes are found in such forms as Anthenea and
Brisinga. In some genera the arms are rendered more stiff
by long flat spurs extending on the inner side of the actinal surface from
the sides of the ambulacral plates towards the edge of
the arms (see a b, Fig. 1, and a h, Fig. 8'). These spurs
are also highly developed in Cribrella {Fig. 10), where,
as in the preceding figures, they conceal the interam-
bulacral plates, which are small compared with the
ambulacral plates of Culcita, Anthenea, Acanthaster, and
the like. The interambulacral plates retain their pre-
ponderance even towards the extremity of the arms
quite near the tip, where the ambulacral and interam-
bulacral plates become separated from the abactinal system proper (see
Fig. 11, the tip of an arm of Culcita).

Solaster endeca.

Solaster endeca Forbes, 1839. Mem. Wern. Soc.
Asterias endeca Lin.

PI. xrii.

In Solaster endeca the arrangement and general structure of the am-
bulacral and interambulacral plates are identical with those of Crossaster ;
the plates are, however, more closely articulated, and the basal ambu-
lacral plates attain a still greater prominence even than in Crossaster.
The mouth-papillse are also more powerful. The fundamental difference
between these genera, Crossaster and Solaster, lies in the structure of the
abactinal floor (compare PI. XVII. Fig. 1, and PI. XII. Fig. 4). The ac-
.tinal floor between the arms is composed of small, somewhat elongated
plates, arranged in more or less regularly diverging rows, quite similar
to those of Crossaster. The interbrachial partitions can hardly be in-

XOETH AMEPJCAX STAEFISHES. 113

tended for the support of the abactinal floor, either in this genus or in
Crossaster. In Solaster it forms a broad band when it connects with the
abactinal surface, and is gradually changed into a mere chord at the point
of attachment to the interbrachial basal plates. These partitions are
all exactly similar to the one supporting the stone canal. At the base
of the arms the sides of adjoining arms come together, forming rounded
angles, and do not, in the specimen examined, form an interbrachial par-
tition for the support of the abactinal floor (see PI. XVII. Fig. 3). The
reticulation of the sides of the arms and of the abactinal region is com-
pact, composed of small meshes forming diagonal lines across the arms,
and more or less irregularly radiating lines from the centre of the disk.
All the plates of the actinal floor carry tufts of small spines (PI. XVII.
Fig. 2), arranged usually in parallel rows, corresponding to the long axis
of the plates ; so that on the actinal side the spines of the interambu-
lacral plates are at right angles to the arms; on the plates forming the
triangular interbrachial space, the spines diverge from the actinostorae,
while those of the plates at the angles of the arms, of the arms themselves,
and of the abactinal surface, form more or less circular tufts arranged
on the lines of the plates of these surfaces.

Solaster endeca and Cribrella sanguinolenta are both found on the two
sides of the Atlantic, occurring in Norway, Denmark, Great Britain, the north-
west coast of France, Iceland, Greenland, Labrador, and as far south as Massa-
chusetts Bay ; C. sanguinolenta extending as far south as Long Island Sound.

Cribrella sanguinolenta.

Cribrella sanguinolenta LiJTK. 1857. Vidensk. Meddel.
Asterias sanguinolenta O. F. Muli- 1776. Zool. Dan. Prod.

Fl. XVIII.

The genus Cribrella is most closely allied to Solaster. It has, like So-
laster proper, a compact system of limestone network, forming, when de-
nuded of spines, small meshes on the abactinal surface (PI. XVIII. lYg. 1),
while the actinal surface and a part of the edge of the arms are covered
with larger plates, forming longitudinal rows parallel to the longer axis
of the arms, with more or less irregular shorter rows at right angles to
the axis (PI. XVIII. Fig. 4). The arrangement of the spines on this net-
work is very similar in the two genera, consisting of short sharp spines
placed on the abactinal surface, either in clusters or in semicircular fan-

114 DESCEIPTION OF THE HAED PAETS OF SOME

shaped rows as they approach the edge and lower surface of the arms ;
the sharp spines often become quite blunt in larger specimens. On the
actinal surface the spines are longer and sharper, usually arranged in lines
parallel with the longitudinal axis of the plates upon which they are
carried (PI. XVIII. Fig. 2). They gradually increase in size towards the
ambulacral furrow ; the spines of the interambulacral plates are still longer,
and those which form the actinal papillas attain the greatest develop-
ment (PI. XVni. Fig. 3). The above features this genus has in common
with Sblaster, differing from it, however, in not having in the interbrachial
angles the sharp line of demarcation between the arrangement of the
plates and rods forming the actinal and abactinal surfaces. The genera
differ also greatly in the structure of the interbrachial arch. In Cri-
brella the arch is well developed (PI. XVIII. Fig. 7), starting from
the angle of the arms and extending the whole way, between the
two floors, towards the actinal ring, while in Solaster the arch is lim-
ited to a free loop, swinging between the abactinal surface and its
basal interbrachial plates at the actinal ring in the interambulacral space.
The last actinal joint of the ambulacral system is large, the ambulacral
plates distant, and the interambulacral plates prominent, with a wide
actinal face, upon which are placed numerous spines of different sizes,
arranged in rows at right angles to the ambulacral furrow (PI. XVIII.
Fig. 2).

On the actinal surface two to four water-tubes pass through the free
space enclosed by the limestone rods ; the water-tubes on the actinal sur-
face are less numerous, but longer.

Astropecten articulatus.

Astropecten articulatus M. T. 1842. Syst. d. Ast.

Aslerias articulatus Say, 1825. Journ. Acad. Nat. Scien. Phila.

PI XIX.

On account of the great prominence of the marginal plates of the
actinal and abactinal surfaces in this genus, the limestone network is
reduced to a small surface. This is particularly the case on the actinal
surface, where the reticulation corresponding to the actinal surface of
the arms is reduced to a few minute plates between the interambulacral
and marginal plates placed at the angle of the arms near the base of the
jaws (see Figs. 4, 7, PI. XIX.). The remainder of the lower side of the

XOETH AilEEICAX STAEFISHES 115

arm is occupied by the marginal plates ; these project beyond the mar-
ginal plate of the abactinal surface ; forming, when seen from above, what
appears like a second row of marginal plates (see Fig. 3, PI. XIX.). The
abactinal limestone network extends over the disk and over the narrow
elongate space left on the upper side of the arms between the marginal
plates (PI. XIX. Figs. 1, 3, 6). The marginal plates are firmly soldered
together, leaving no space between the floors where they are placed,
■with the exception of a single large opening for the passage of water-
tubes along the line of junction of two plates, across the arms; the whole
space in the arms between the plates being thus reduced to a narrow
flattened space, of which the larger part is occupied by the ambulacral
plates (PI. XIX. Fig. 5).

The interbrachial arches are reduced to a thin partition at the angle
of the arms, where the abactinal marginal plates attain their greatest
height. The abactinal limestone network is, when seen from above, found
to be closely covered by short club-like spines, often with a broad base
and constriction in the middle below the head, attaining their greatest
diameter a short distance from the base of the arms, passing gradually
into mere granules towards the extremity of the arms and the centre of
the disk ; these spines are attached to the abactinal limestone network
(PI. XIX. Fig. 3) by a very shallow sucker, shaped like a saucer, with
edges slightly turned up. On the tip of these spines are arranged concen-
trically a number of minute spines more or less cylindrical, with rounded
ends, often completely filling the interval between adjoining spines, so
that they appear to form at first glance a smooth surface (PI. XIX. Fig. 1)
over the whole space Ij'ing between the marginal plates. The gi'ooves
between the adjoining marginal plates are lined by similar, but even
more delicate spines, which appear to perform the same functions as the
delicate spines on the fascioles of Echini, namely, to sift the foreign matter
contained in the water admitted to the water-tubes.

Seen from the actinal side, the abactinal floor consists of small circular
plates (PI. XIX. Fig. 6) corresponding to the flat saucer-like plates of
the centre of the disk ; seen from the opposite side, these gradually pass
into the flattened plates closely soldered together, which extend into the
arms, leaving only, however, on each side of the solid central band, a
number of passages for the water-tubes (see PI. XIX. Fig. 6). The gen-
eral surface of the marginal plates of the abactinal side is covered by

16

116 DESCRIPTIOX OF THE HAED PARTS OF SOME

short rounded spines passing mainly into granules similar to those cover-
ing the abactinal surface of the arms ; they are in addition provided with
one or two long flat triangular movable spines similar to those covering
the outer edge of the actinal side of the marginal plates, which are
arranged in irregular diagonal lines across the plates, varying in size,
generally flattened and triangular ; but we find with them, along the
edge of the furrows separating the plates, slender spines similar to those
of the grooves of the abactinal side.

The interambulacral plates carry flat spatula-shaped spines placed at
right angles to the longitudinal axis of the arms ; these plates, when
denuded, are seen to be in contact with the marginal plates, except near
the actinal ring (PI. XIX. Fig. 5). Spines similar to those carried by the
interambulacral plates, only shorter, cover the actinal side of the jaws
(PL XIX. Fig. 2).

The ambulacral plates seen from the interior of the arm occupy, with
the base of the interambulacral plates, the whole space between the mar-
ginal plates ; the median furrow formed by the junction of adjoining
plates is deep ; the ambulacral plates themselves are narrow, elongate,
spreading somewhat at their junction with the interambulacral plates,
leaving a wide space for the passage of the ambulacral feet.

It is not uncommon in this genus to find the ambulacral and interam-
bulacral plates soldered together, either wholly or in part, so that it be-
comes difficult to trace the line of contact.

In Astropecten and Luidia the interambulacral plates of the last basal
joint of the adjoining arms are connected together, forming a prominent
point at the angle of the arms ; but those plates which carry the mouth-
papilljB are not, as in other families of Starfishes, at a lower level than the
adjoining interambulacral plates. The jaws are on the same level, in direct
continuation of the other interambulacral plates, only somewhat more
prominent (see PI. XIX. Figs. 7, 8).

Astropecten articulatus and Luidia clathrata extend from New Jersey
to the West Indies. Luidia clathrata is one of the most common Star-
fishes of the sandy coasts of North and South Carolina.

XORTH AMERICAX STARFISHES. 117

Luidia clathrata

Luidia clalhrata LtJTK. 1859. Vidensk. Meddel.

Asterias clathrata Say, 1825. Journ. Acad. Nat. Scien. Pliila.

PI XX.

The genera Astrojiecten and Luidia are most closely allied, not only
by their possessing but two rows of pointed ambulacra! suckers (PL XX.
Fig. 2), but also by the structure of the limestone network of the two
surfaces of the spines and other appendages covering them. As in Astro-
pecten, the actinal limestone network is limited to a small triangular
area close to the mouth in the angle between two arms ; this area reminds
ns of the interbrachial space on the actinal side covered by small plates in
such genera as Solaster and Crossaster ; with the latter they are closely
connected. The rest of the actinal surface of the arms is covered by
the narrow elongated marginal plates which correspond in number to the
ambulacral and interambulacral plates (PI. XX. Fig. 4).

The actinal marginal plates are, as in Asti'opecten, separated at the
surface by deep grooves edged by minute spines less numerous along
the main lines of the grooves than in the grooves of Astropecten, but
much more crowded at the openings near the interambulacral plates, form-
ing a regular sieve from plate to plate. The spines when removed leave
upon the face of the plates markings exactly similar to those found as
bands upon Echini, and known as fascioles. The spines carried upon
these minute granules are similar in every respect to the spines of the
fascioles of Echini. Their function is evidently identical, namely, that of
filtering and clearing the water before it reaches the water-tubes. Their
use is much more apparent than in the Spatangoids, where the bands
of fascioles are really of use only when lining the edges of the sunken
ambulacra of such genera as Hemiaster, while their extension from the
tip of one ambulacral rosette to the other seems to be a remnant of a
structure having at the present day in Echini but little if any use, while
in Spatangoids it still performs its function of accumulating minute
muddy particles floating in the water, which would to a certain extent
impede the access of clean water to their lobed ambulacral tentacles.
I have not observed these fascioles in other genera besides Astropecten
and Luidia. The presence, however, in some genera of minute spines
arranged in tufts on a solid basis projecting above the general surface
shows us a regular transition from the closed area formed by them on

118 DESCRIPTIOX OF THE HAED PARTS OF STARFISHES.

the abactinal surfiice between the marginal plates in such genera as
Astropecten and Luidia, to a somewhat looser arrangement in Solaster
endeca and Cribrella. This arrangement is still further modified in
Crossaster papposa and Pycnopodia, and leads to such spines as are
found in Asterina and Palmipes, where the tufts consist of a smaller
number of minute spines more uniformly scattered over the surface, thus
forming an approach to the usual distribution of spines in Asteracan-
thion. Finally we pass to genera where the spines are long and few in
number, and do not, as in the genera of the Asteriadae proper, perform
the part of sieves.

In place of the single row of large marginal plates along the abactinal
edge of the arms, we find in Luidia a series of much smaller plates, coi*-
responding in number, as on the actinal side, to the number of ambula-
cral plates. There are sometimes four or five rows of such plates, form-
ing regular longitudinal and transverse rows (PL XX. Fig. 3), followed
towards the median band of the arm by more irregulai'ly arranged plates.
These plates form the base of prominent pillars, somewhat constricted in
the centre, flaring at the extremity, surmounted at the tip by short spines
or merely granules articulating in a shallow socket. These spines are
so closely packed as to leave but very narrow passages between ad-
joining rows (see PI. XX. Fig. 1), generally mere slits edged by
minute spines, so that longitudinal and transverse passages run the
whole length of the arms for the passage of water, which must be all
carefully sifted before it enters either through the passages protected by
the fiiscioles or through those screened by the minute spines of the abac-
tinal surface.

The plates of the abactinal limestone networks are completely sol-
dered (PI. XX. Fig. 6), leaving but few irregular rows of holes for the
passage of the water-tubes to the abactinal side, where they are com-
pletely sheltered under the floor formed by the minute spines of the
abactinal surface of the arms.

In no other genera of Starfishes do we find so great a simplicity in
the structure of the plates of the actinal ring as in Astropecten and
Luidia. Usually the ambulacral and interambulacral plates of the arras
differ in no essential way except at the actinal ring formed in most Star-
fishes by such a modification of the last joint as to make it somewhat
difficult to trace the homology of all the parts. This last joint is extremely

FASCICLES OF STAEFISHES. 119

simple in Astropecten, being but slightly modified and difTering from the
others mainly in length. Thus the homology I have attempted to trace
between the jaws can there be seen in its simplest form (PL XX. Fic/s.
4, 5, 8). The plates of the extremity of the arms are soldered together
when seen from above (PI. XX. Fi()s. 9-11), forming a prominent knob
with a deep groove on the actinal side for the passage of the ambulacral
tentacles.

The spines of the actinal side increase slightly in length towards the
outer edge of the arms, where there is found a prominent row of lai'ger
flattened spines fringing the edge of the arms.

The actinal face of the jaw-plates is prominent and thickly studded
with irregularly arranged minute spines, forming a marked feature at
the actinal angle of the arms, between adjoining ambulacral rows (PL XX.
Fig. 7).

The madreporic body is often irregular in outline (PL XX. Fig. 12),
and is frequently completely hidden by the surrounding spines of the
abactiual surface.

FASCIOLES OF STARFISHES.

The description of the accompanying figures of Luidia and of Astro-
pecten will explain the disposition of the minute spines of those genera
which I have homologized with the fascioles of Echini. Fig. 12 repre-
sents a transverse section of an arm of Astropecten.

^ Fig. 12.

a' being the ambulacral, a i the interambulacral, plate, ^„ a--

with its spines p. Ip is the plate on the edge of ?^|^f---^jiS^®^'2iSiSj;2!^(n ,
the lower side of the arm, and Ip the correspond- ^fe"\_^xS^^"^^^M~^

ing plate of the upper edge of the arm, ip being
the small columnar plates surmounted by tufts of '' «

minute spines forming the close covering of the central part of the
abactinal side of the arm. The surfaces s of the upper and lower plates
on the edge of the arm are the articulating surfaces which rise somewhat
above the surrounding edge of the plate, leaving a flat space /' on the
lower arm-plate, g" on the upper arm-plate, and from g to / between these
two dilates, through which water from outside can circulate as in a groove
all round the articulation, and thus find its way between the columnar
plates of the abactinal surface of the arm. The small papillte which

120

FASCIOLES OF STARFISHES.

Fig. 13.

cover these marginal plates, but more especially the minute spines
crowded upon the surfaces of the grooves g, g, g" , and g" form an effective
sieve, and in thus freeing the water from its impurities before it circu-
lates through the channels between the abactinal plates, act exactly-
like the fascioles of Echini. Only in Starfishes we can much more
readily see their great use in the economy of the animal, while their
action in the Echini is much less eflicient. In a profile view of a part of
the edge of the arm of Astropecten {Fig. 13), the openings, left for the
passage of the water, which are lined by these so-called fascioles are very
plainly defined. Ip and // are the lower and upper marginal plates, with
the deep grooves /' between the lower plates and the furrows /" be-
tween the upper plates, these furrows being completely arched over by
the minute spines acting as meshes of a sieve. At the angle of the

junction between the shallower horizontal grooves
and the deeper vertical grooves a prominent open-
ing g is formed for the passage of the bulk of
the water, which is thus admitted to be sifted.
The lettering of Fig. 13 is the same as in Fig. 12.
In Luidia the only difference in the mechanism
of the fascioles is the greater number of openings through which the
water is admitted to circulate between the columnar plates covering the
abactinal surfiice and a part of the arms. In Fig. 14 we have a section
of the arm of a Luidia, corresponding to Fig.
12 of an Astropecten. The lettering is the same,
onl}^, there being a larger number of upper
marginal plates, the passages between them [g,
g\ g", g", g"") are more numerous. The lower
marginal plate alone is as prominent as in

Astropecten. The articulation forms a continuous ring
round the arm, broken by the columnar plates sur-
mounted with their tufts of minute spines. These tufts
are so thick as to form a uniform shield almost solid
and unbroken on the abactinal surface of the arms.
Seen from below {Fig. 15), the deep groove / of
the lower marginal plate edged with minute spines, the fascioles, is well
shown. A view of the edge of the arm of Luidia {Fig. 16, correspond-
ing to Fig. 13 of Astropecten) shows the small rectangular areas into

Ot, (^ «<l Q

FASCICLES OF STAEFISHES. 121

which the edge of the arm is divided by the deep furrows, allowing
the passage of the water; at their crossing, the furrows form larger.
more prominent openings ; the edges of all these rectan- j,jg ^q

gular spines are crowded with fascioles. The genus Cri- 1!**^ jc^j^swtj ,t|b'
brella is interestino- as showing the gradual transition of VjjJl'l^tp'

the interambulacral and marginal papillae into tufts of **^f%f^^''y

such minute spines that the difference between them '

and true fascioles is hardly appreciable. In fact, in So-

laster we have already certain parts of the surface covered by such

minute spines that we must consider them as rudi-

^_^ mentary fascioles and as probably acting as such.

4\ jAK'*' ^ Fig. 17, a cross section of Cribrella, shows a close

fe,Cr^L-^'-CJ. <;^ approximation to the cross section of Luidia as far as

'^0^fW'^'i the tufts of spines are concerned ; these need to be

'"A but slightly more crowded to form a

. . . . ^ie- 18.

most effective sieve. Seen in profile in a section {Fig. 18), ^

the tufts of spines, the interambulacral papillae, are seen ^"^.^'^-^^

to be somewhat more crowded into tufts than is the case
in such genera as Asteracanthion. A similar arrangement
is also found in Solaster (see Fig. 8), where the spines of the interam-
bulacral plates, with the exception of those of the so-called jaws, are
arranged in closely crowded tufts.

NOTE.

The arrangement of the Starfishes into families from the study of their
hard parts does not differ materially from the families adopted by Perrier
in his Revision of the group.* He himself has in a general way made
use of the characters furnished by the skeleton to limit the families he
has recognized. The modifications we should suggest go so far as to
transfer Pycnopodia from the Asteriadi^ proper, and Crossaster from the
Echinasteridge, placing them in close proximity to Brisinga, while Solaster
(limited) and Cribrella would be placed with the Asterinidae.

The disposition of the digestive cavity and its appendages does not
appear to furnish systematic characters of great value. The anatomy of
the ovaries of the coecal appendages of the digestive cavity proper with

* Asteriadse, Echinasteridae, Liackiadse, Goniasterid^, Asterinidae, Astropectinidae, Pterasteridae,
Brisinaidae.

192 FASCIOLES OF STARFISHES.

its abactinal pouch is remarkably uniform in groups apparently differing so
widely as the extreme pentagonal Starfishes, and the long slender-armed
genera like Ophidiaster, Asteracanthion, or even the apparently abnormal
group to which Brisinga, Pycnopodia, and Crossaster belong.

I do not give a list of our North American Starfishes, much less a
Sj'nonymic Catalogue, as it would be most incomplete and premature.
Quite a number of species collected by Mr. Pourtales in deep water be-
tween Florida and Cuba are at present in the hands of Professor Perrier
for determination ; of these several are undoubtedly new to our fauna.
Numerous additions have recently been made by Professor Verrill, while
eno-ag-ed on the dredffino-s made in connection with the United States
Fish Commission. In addition the " Challenger " expedition, while cruising
in the Atlantic from Halifax to Bermudas, hence to New York, and then to
St. Thomas, added quite a number of remai-kable forms to our American
species. As these collections ai'e either in process of identification or
about to be worked up, any general list now given would soon become
antiquated. The Starfish fauna of North America, as far as now known,
can be made out with sufficient accuracy from the articles by Professor Per-
rier on the " Stellerides du Musseum in the Archives de Zoologie Experi-
mentale " for 1875 and 1876, although the Synonymy he has adopted
for several of our species will probably be modified when larger material
than is now available has been collected. The principal localities of speci-
mens in the Museum collections are added.

EXPLANATION OF THE PLATES.

To avoid useless repetitions in the description of the Figures, the same letters are used, throurjhout these Plates
[I. -VIII.], to denote identical parts. It will greatly facilitate the reading of this memoir to heconie familiar with
the notation here adopted.

Explanation of the Lettepjng on Plates I. -VIII.

a, anus.

6, dorsal or water pore, madrejioric opening.

c, alimentary canal.

d, digestive cavity, stomach.

d', abactinal water-tubes in angle of rajs of young
Starfish.

d", water-tubes of lateral hne of rays of young Star-
fish.

d'", water-tubes of median line of rays of young Star-
fish.

e, eye of Starfish at base of odd tentacle {I'),
e', median anal arms of Bracliiolaria.

«", dorsal anal arms of Bracliiolaria.
e'", ventral anal arms of Brachiolaria.
e"", dorsal oral arms of Brachiolaria.
e^, ventral oral arms of Brachiolaria.
e', odd terminal oral arm of Brachiolaria.
y, brachiolar arms.

f, branch of water-tube (w id') leading intoy^
/", odd brachiolar arms.

/'", surface- warts at base of odd brachiolar arra (/").

h, hole of cul de sac of water-tube w.

I, central abactinal plate of yoimg Starfish.

'ij ^11 'i) ■ ■ interbrachial abactinal plates of young

Starfish.
2, I2, l^, • . brachial plates of young Starfish.
m, mouth.

m', pistol-shaped oral pouch of oesophagus.
m", anal pouch of oesophagus,
n, opening for passage «f ambulacral sucker.
0, oesophagus.

p, spines on edge of ray of Starfish.
pi, spines of exterior rows along abactinal surface of

rays.
P2, spines of middle row on abactinal surface of rays.
Pa, central spiue of abactinal surface of Starfish, on

central plate (l^).

Pc, plate at jimction of adjacent rays (ovarian plate).

p', ])", different forms of pedicellarijE.

r, abactinal surface.

r', first set of five limestone rods which appear on

abactinal surface, and which eventually become

the brachial plates (/j).
r", second set of five interbrachial limestone rods,

which eventually become the interbrachial plates

r'" -)■"', rays of young Starfish; r'l' being ray next
to madreporic body, when Brachiolaria is seen
from the dorsal side.

s t and s, actinal region.

t, t, t, . . tentacles of the young Starfish.

t', odd tentacle.

t", ambulacral tube.

u, lateral ambulacral plates, siu-mounted by spine.

u', mc<lian ambulacral plates with very small spines.

V, vibratile chord, anal part.

v', vibratile chord, oral part.

!(', water-tube, developing abactinal area.

ic', water-tube of Brachiolaria leading to madreporic
opening (i), developing actinal area.

w tv', portion of water-tube of Brachiolaria formed by
junction of w and w'.

In all the figures of the Brachiolaria (Plates L-IV.),
the attitude which has been given to them is not
a natural attitude. This has been done pur-
posely, for the sake of making the comparison
with the memoirs of Miiller easier. The only
figure of a Brachiolaria which is in its natural
attitude is that of PI. VIII. Fig. 8. The young
BracliioLiria does not, however, move with the
anal part below, till the latter is loaded down by
the development of the Starfish, and we see
them swimming about, before that time, almost
in every possible attitude.

124 • EXPLANATION OF THE PLATES.

Plates I., II. Embryology of Asteracantiiion berylinus Ag.

PI. I. Figs, 22-28, PI. II. Figs. 2-19, Scyphistoraa stage; PI. II. Figs. 20-24, Tomaria stage ; PI. II. Figs. 25-28, Brachina stage.

PLATE I.

Fig. 1. A mature egg, surrounded by spermatic particles, soon after the artificial fecundation. The egg
has assumed a spherical shape, and contains the germinative vesicle and dot. Tliere is no trace of
any interval between the yolk and the outer envelope.

Fig. 2. The germinative vesicle has disappeared, but the germinative dot remains.

Fig. 3. The germinative dot is no longer visible ; the yolk has contracted, and is separated by a slight
space from the outer envelope. The egg has all the appearance of having already gone through the
segmentation ; the whole yolk being made up of small spherical cells, resembling very minute spheres
of segmentation, although the segmentation has not yet commenced. Two hours after fecundation.

Fig. 4 shows the first trace of segmentation, consisting in a depression on one side of the yolk.

Fig. 5. The yolk has become flattened on opposite poles ; the Richtungsblaschen are visible on one side
of the yolk.

Fig. 6 shows the yolk divided into two united ellipsoids, the whole yolk rotating slowly, alway.s in one
direction, from right to left. The Richtungsblaschen are at one pole of the axis of segmentation.

Fig. 7. The two segments of the yolk have entirely separated. The Richtungsblaschen are likewise iso-
lated at one pole of the axis of segmentation.

Fig. 8. First trace of a further segmentation ; one half of the yolk is partially divided.

Fig. 9. The two yolk segments are about to separate into four.

Fig. 10. The four yolk segments are all distinct, and almost transformed into regular spheres.

Fig. n. Different view of Fig. 10, showing the position of the segments.

Fig. 1 2. The yolk about to separate into eight spheres.

Fig. 13 shows eight spheres of segmentation, all of which are more or less spherical; the spheres are
arranged in two clusters of four, on opposite sides of the envelope.

Fig. 14. This view of the egg shows the tendency of the spheres of segmentation to airange themselves
on the circumference.

Fig. 15. The yolk is divided into sixteen spheres.

Fig. 16. '.. he shell of segmentation is composed of thirty-two spheres ; owing to the position from which
the egg is viewed, only half the shell of segmentation is visible.

Fig. 1 7. The thirty-two spheres are again subdivided.

Fig. 18. The spheres of segmentation are still smaller than in the preceding figure.

Fig. 19. These spheres have become so small, that the walls of the spherical shell formed by them can be
readily distinj;uished

Fig 20. The walls have become still more distinct in consequence of the close packing of the small
spheres, which are now somewhat polygonal, owing to their pressure upon each other.

Fig. 21 represents an egg ten hours after segmentation; the spheres are still more polygonal; the rotation
of the yolk is quite rapid, and the embryo is ready to break ttrough the outer membrane ; the shell
envelope is very distinct from the inner contents, and has a uniform thickness.

Fig. 22. An embryo after its escape from the egg ; the wall is no longer of the same thickness through-
out, but has become very much thickened at one pole (a), while the spheres of segmentation are
somewhat indistinct.

Fig. 23. The embryo has been slightly flattened at the pole (a), where the wall is thickest ; the planula,
if we may so call it in its present condition, reached this stage at the end of about eleven hours.

Fig. 24. The wall of the flattened pole has been pressed in so as to curve slightly inward (a).

Fig. 25. The depression (a) has become much deeper, and the spheres of segmentation have entirely dis-
appeared, twelve hours after iecundation. The depression at a assumes here somewhat the aspect of
a digestive cavity.

Fig. 26. Seventeen hours after fecundation ; the embryo has lost its spherical shape and has become
somewhat pear-shaped ; a transverse section is still circular. Tlie depression made by the thickened
walls has increased in depth; the opening («) performs the functions of a mouth and anus; Vindi-
cates the bottom of the digestive cavity.

EXPLANATION OF THE PLATES. 125

•

Fii". 27. Twenty hours after fecundation; the depression lias the appearance of a small pouch (</) hang-
ing in a pear-shaped body with circular section, showing no deviation from the absolute radiate type;
the opening (a) still performing the double functions of mouth and anus. Currents of water circulate
in this cavity, as they would in the digestive cavity of any Polyp or Acaleph in .about the same stage
of development.

Fig. 28. Twenty-two hours after fecundation ; the embryo has become somewhat more cylindrical, losing
its pear-shaped form, but is still circular when seen in a transverse section. The cavity (rf) has
slightly expanded at the closed e.ttremity, and is comparatively deeper and wider; the walls of the
body are much reduced in thickness, except at the perforated region. The body is somewhat translu-
cent, and slightly tinged with ochre color. The opening (n) still serves as a mouth, altliough, in
more advanced stages, a second opening is formed, which is the true mouth, at which time the present
mouth then becomes the anus.

PLATE II.

In Figs 1, 3, 9-17, the digestive cavity alone is represented.

Fig. 1. The digestive cavity of Fig. 2, seen by itself from above, has expanded into a large reservoir at
the extremity, the walls of which are quite thin.

Fig. 2. The embryo of Fig. 1 seen in profile ; the cavity is no longer in the axis, but is bent to one side.
The larva has also lost its symmetrical outline, and the dorsal part of the perforated extremity projects
somewhat beyond the opening of the present mouth (the future anus).

Fig. 3. The digestive sac of a larva somewhat more advanced than Fig. 2, in which the present mouth (n)
(the future anus) has been brought to the lower side.

Fig. 4. The larva of Fig. 3 seen in profile ; the pouch at the closed extremity of the bent digestive cavity
is now nearer the lower side than in Fig. 2, having approached the slight depression (m) placed in the
middle of the larva.

Fig. 5. A larva somewhat more ailvanced, seen in profile, in which the pouch has actually come in con-
tact with the wall of the lower side at in. The dorsal region of the perforated extremity projects
still more beyond the opening of the present mouth (a) (the future anus) than in the preceding
stage (Fig. 4). The digestive cavity is not yet divided into distinct regions.

Fig. 6. The same larva as Fig. 5, seen from above, forty-two hours after fecundation ; large epithelia-
cells have appeared on the surface.

Fig. 7. A somewhat more advanced larva, seen in profile ; the digestive cavity is no longer a simple bent
tube, as in Fig. 5 ; it is strongly contracted near the extremities, one of them projecting upwards (w).
At the point of contact of the digestive cavity with the outer wall at m, a second opening has been
formed, connecting by a short tube with the pouch of the digestive cavity. This second-formed open-
ing (m) is the true mouth, while the first-formed opening (a) now becomes tlie anus, after having, up
to this stage, performed the functions of mouth and anus ; end of the second day.

Fig. 8. The same larva as Fig. 7, seen from above, to show the position of the lobes (w, iv') formed on
each side of the pouch of the digestive cavity (r/), which in Fig. 7 appear like projecting angles (tc).

Fig. 9. Isolated digestive cavity of a more advanced larva, showing still more plainly the transverse con-
tractions of the digestive cavity by which the oesophagus (»), the stomach ((/), and intestine (c) are
gradually formed, and also the greater projection of the earlets of the pouch, which have become quite
elongated laterally; the opening (o) in the centre is the tube leading to the mouth.

Fig 10. The same as Fig. 9, seen in profile; the tube (o) now connects very freely with the mouth (m),
formed in the depression, mentioned in Figs. 4, 5, and with the digestive cavity ; the currents now
change their course, and circulate in the opposite direction. While the larva was in the state repre-
sented by Fig. 6, the cuiTents of water enter at the mouth, the future anus (a), circulate in the pouch
(d), as well as in the earlets formed from the thickening of the wall, and then issued again from the
same opening (a). Now the water enters through the mouth (m) (the last -formed opening), passes
through the narrow conical tube (o) into the digestive cavity (</), cummunicating with the earlets
(w, iv'), and out through the anal opening («), which was the first formed, and furmerly performed
the functions of mouth.

Fig. 11. Isolated digestive cavity, seen in profile, showing the tube leading from the mouth (m) to the
digestive cavity (</), and earlets (w, ?«'), more developed than in Fig. 10.

Fig. 1 2. The same seen from above.

Fig. 13. Oral end of an isolated digestive cavity, in which the earlets, formed by the pouch, are more

126 EXPLANATION" OF THE PLATES.

distinct from the digestive cavity tlian in any of the former staj;es. Tliere is a sliglit constriction at
tlieir base of attacliment, tlie first imlication of tlieir final separation from tlie alimentary canal.

Fig. 14. Isolated digestive cavity seen endwise, to show the tube leading from the mouth to the digestive
cavity, at right angles to the pouch of the earlets.

Fig. 15. Isolated digestive cavity seen from above, in which the earlets (70, w') (the future water-tubes)
are so far differentiated as to be quite distinct from the digestive cavity. The walls of the earlets are
exceedingly attenuated, and are scarcely connected with the main digestive cavity.

Fig. 16. The same as Fig. 15, seen from below, to show the position of the mouth and anus on the same
side of the larva.

Fig. 17. Part of the same larva seen in profile: on account of the obliquity of the earlets, one of them
(w'), as it increases in size more rapidly than the other, soon reaches the outer surface of the larva
and opens into the surrounding medium by means of a small aperture (6). The walls of the tube
(oesophagus) leading from the mouth to the first swelling of the digestive cavity (rf) (the stomach),
and of that part of the tube leading from the stomach to the anus, have a very different thickness.
They are sufficiently distinct in their character to enable us to distinguish readily three regions ; forty-
eight hours after fecundation.

Fig. 18. The two small bodies (mi, w'), the former eai-lets of younger stages formed from the pouch at the
closed end of the digestive cavity (the problematic bodies of Miiller), have entirely separated from
the digestive cavity from which they were formed ; seen from above, the three divisions of stomach,
intestine, and oesophagus are plainly marked out.

Fig. 19. The same larva in profile.

Fig. 20. The same figure from below, shows the presence of short crescents of vibratile cilia (f, v') placed
in opposite directions near the mouth and anus; sixty-five hours after fecundation.

Fig. 21. A somewhat more advanced larva, seen in profile; the anal crescent («) of vibratile cilia is seen
as a small wart between the mouth of the anus, the oral crescent (t^') projects beyond the general out-
line. The division into cesophagus (0), stomach (</), and intestine (f) is quite prominent. The
stomach has a tendency to approach the anal dorsal extremity.

Fig. 22. The same as Fig. 21, seen from below, to show the triangular shape of the mouth (m). Tlie
greater size of the problematic bodies (w, !»') (the water-tubes), which increase independently at an
unequal rate, and also the position of the oral and anal vibratile crescents.

Fig. 23. The same larva seen in a profile, to show the position of the mouth in a strongly marked depres-
sion ; the great increase in size of the oral part of the cesophagus ; the swelling out of the stomach,
and the bending of the intestine back towards the mouth, so as to make a small angle with the
trend of the stomach ; at the end of the third day after fecundation.

Fig. 24. Larva seen from above. The only difference in this stage from the preceding is in the greater
increase of the vibratile crescents, forming two small plastrons, and of the water-tube. The intes-
tine also bends so as to make almost a right angle with the stomach, which is pushed out further
towards the anal extremity.

Fig. 25. More advanced larva, seen from the left profile, in which the oral pouch has assumed its ch.arac-
teristic pistol-shape. The stomach and intestine make a sharp angle with each other, the latter being
much longer than the stomach proper. In its present aspect it closely resembles a retort, the stomach
being the receiver, the intestine the tube. The anal and oral vibratile crescents have greatly ex-
teniied, the oue on the oral and the other on the dorsal side, to the extremity of the body.

Fig. 26. The same as the preceding, seen from below ; the oral plastron is quite large, projects beyond
the sides of the body; slight indentations can already be traced in the anal plastron, indicating the
position of the future arms («'). The water-tubes have increased in length, and extend half-way
from the base of the stomach to the oral plastron.

Fig. 27. A larva six days after fecundation, seen from the right profile, the water-tubes extend beyond the
opening of the mouth. The tube leading from the water-pore (6) (dorsal pore) to the water-tube
(it)'), is quite distinctly seen.

Fig. 28. The same larva as Fig. 27, seen from below ; the intestine, as in Fig. 26, .is thrown to one side
of the axis of the larva. The water-tubes extend also along the sides of the stomach towards the anal
extremity ; the sinuosity of the anal ciliary chord indicates the position of the future anus.

EXPLANATION OF THE PLATES. 127

Plates III. -VIII. Ejibryologt or Asteeacanthioi^ pallidus Ag.

PLATE III.

Owing to the transparency of these larvse, it is not easy to ascertain whether they are seen with the mouth
downwards or upwards, unless we ascertain the position of the madreporie body. In all these fig-
ures, whenever the water-tube v:' is on the left of the figure, the mouth is turned upwards.

Fio-. 1. The youngest larva of this species, seen from the mouth side, corresponding to PI. II. Fig. 20;
a comparison of these two figures will show the great difference between the larvEe of these two
species of Starfishes. In the former, the chords of vibratile cilia appear much earlier, and the oral
plastron is well defined ; while, in the other species, it is not before it has reached the condition of
PI. II. Fig. 2(1, that the oral plastron is as well developed.

Figs 2-10. Brachina stage.

Fig. 2. A larva seen from the left profile, corresponds to the stage of PI. IT. Fig. 27, of A. berylinus ;
with the exception of the size of the water-tubes, the larva of this species is much stouter, shorter,
and the anal portion is the most prominent, while the larva of the A. berylinus is quite slender and
elongated.

Fig. 3. The same larva as Fig. 2, seen from the dors.al side.

Fig. 4. A more advanced larva, seen from the dorsal side ; the undulations of the ciliary chord indicate
the futiu-e arms, the water-tubes extend beyond the mouth, and have already begun to bend towards
each other.

Fig. 5. The same larva seen from the left profile, to show the bent attitude frequently assumed by the
larva when disturbed.

Fig. 6. This larva, seen from the mouth side, is more developed than any raised by artificial fecundation
from the eggs of A. berylinus. The water-tubes have greatly increased in diameter; they have united
beyond the mouth, and extend on each side of the stomach so as almost to meet, but without uniting.
The mere indentations of the previously figured larva; correspond to accumulations of pigment cells,
and to the thickening of the vibratile chord, accompanied by the formation of rudimentary lobes,
which indicate plainly the position of the median arms (e'), the dorsal anal {is"), the ventral anal (e'"),
and dorsal oral arms (e""). The greatest accumulation of pigment cells, and the thickening of the
vibratile chord, is found at the rudimentary median arms (e'). The anal ventral pair of arms (e'") is
especially well marked.

Fig. 7. The preceding figure seen in profile, the mouth to the right, shows the great development which
the oral position of the water-tube has taken ; also tlie mode of formation of the oral ventral pair of
arms {e"), as well as the first sign of the odd brachiolar appendage (/").

Fig. 8. Larva seen from the dorsal side. The arms have increased greatly in size since the stage repre-
sented in Fig. 6. The oral portion of the water-tube has become very pointed ; it extends into the
odd oral arm (e°), which has also elongated, and stands out prominently beyond the oral plastron.

Fig. 9. The same figure seen in profile, with the mouth downward. The vibratile chord is a deeply
undulating line, following the edge of the arms, which extend beyond the general outline. The
water-tube, it will be seen, forks also at the oral extremity ; one branch extending into the odd
arm (e"), the other toward the angle made by the base of this arm and the pair of oral ventral
arms {«•). The great increase in size of this odd arm will be seen when compared to Fig. 7 of
this plate.

Fig. 10. Larva seen from the mouth side. Thus far the arms had altered but little the character of the
outline of the larva. In this figure, however, some of them are sufficiently developed to be capable
of extensive motion. Tlie median arms (e') especially are far in advance of the others. All the anal
arms develop so as to become more slender at first, and assume their true character sooner than the
oral arms, which, diu-ing the early stages, are always more heavy, and take theu- final shape later than
the anal arms. At the angle, where the oral ventral arms and the odd arm come together, at the
base of the oral arms, slight swellings are formed (/), which are the first trace of the pair of brachi-
olar arms (//), the odd brachiolar arm being only seen in the profile view (Fig. 12,/'), though it can
be traced as a double outline of the odd arm (/"). We can already see a constriction in the
water-tube as it passes into the odd arm, and from this (nearer the mouth) are sent off two small
pouches (/'/'), which enter into the brachiolar pair of arms (/). The first trace of the aetinal area
of the future Starfish is also plainly visible (f) on the water-tube (ic), on the left of this figure.

;i28 EXPLANATION OF THE PLATES.

Fi"-. 11. A more advanced larva than Fig. 10, seen from the mouth side, in whit-li the oral arms have
assumed all the eharacters of the anal ajipendages. The brachiolar arms are quite well developed ;
the intestine and the stomach are slightly crowded to one side by the greater increase of the actinal
area (;) of the Starfish ; the ambulacral pentagon of the future Starfish is still more marked (t) than
in previous stages. Brachiolaria stage.

Fio-. 12. The same as Fig. 10 seen in profile, with the mouth downwards.

PLATE IV.

Fio-. 1. Seen from the mouth side. A larva with its arms fully developed and in full activity ; no fur-
ther changes take place in the general aspect of the larva, with the exception of those of the anal
part where the Starfish is developing, and those of the brachiolar arms. All the arras are nearly
equally advanced, with the exception of the median arms (e'), which still retain their greater size.
The odd terminal arm (e') has also greatly increased in length, as well as the brachiolar arms (ff),
which are capable of motion, and into which the branches of the water-tubes can easily be traced.
Brachiolaria stage.

Fio-. 2. The same larva, seen from above, on a somewhat smaller scale, shows in what way the stomach
and the intestine have been pushed to one side, by the great development of the actinal part of the
Starfish, on the right of the figure (s /)• The shape of the mouth (m) is particularly well seen, in a
dorsal view, at this stage of growth.

Fio-. 3. The same larva on a different scale, seen endways, from the oral end, to show the connection
between the pair of brachiolar arms (//) and the oral ventral pair (c'), as well as the position of the
odd brachiolar arm (/") at the base of the odd terminal arm (e").

Fi"^. 4. An adult larva seen from the right, actinal profile ; the arms are in the position which they take
when moving rapidly, arched towards the median arms, the brachiolar arms alone being curved in the
opposite direction from the others. Here the crescent-shaped ambulacral pentagon, as well as the
lobed pentafional outline of the abactinal area are plainly seen.

Yi<r. 5. A magnified profile view of the brachiolar arms.

Fig. 6. The brachiolar arms seen from the ventral side of the larva, to show the position of the single
disk and of the double ro-iv of disks at the base and on each side of the odd brachiolar arm, somewhat
less magnified than Fig. 5.

Fig. 7. The anal part of the larva soon after the shrinking of the arms has begun. The whole of the
terminal anal part of the larva has gradually been absorbed, so that the disk of the Starfish occupies
the whole of the space between the median arms, seen from the ventral side ; the oral extremity of
the Brachiolaria is unchanged and not represented.

Fig. 8. The shrinking has gone so far that the whole of the anal part has been affected, and the oral ex-
tremity alone, with the brachiolar and the terminal arms, retain their original shape and proportions.

Fig. 9. A different view of the anal part of a larva from that of Fig. 7 ; in a slightly more advanced
condition than that of the preceding figure, showing the great height of the abactinal region of the
young Starfish ; the oral extremity of the Brachiolaria is omitted, as it remains almost unchanged.

PLATE V.
Development of the Starfish froper.

The Figures on this Plate show the gradual development of the actinal and abactinal regions of the Star-
fish, and the figures represent simply the anal part of the Brachiolaria, which is alone affected during
this development.

Figs. 1-7 correspond to a Brachiolaria, having reached a state about as advanced as that of PI. III.
Fig. 10.

Fig. 8 is a Starfish developed on the Brachiolaria of PI. IV. Fig. 11; while Figs. 9-14 are stages of
development -which are only found on Brachiolaria; having their full complement of arms, and in
■which, except these changes of the Starfish, but slight modifications take place.

Figs. 1, 2, 10, 12, reijresent that profile of the anal pai-t of the Brachiolaria, in successively more advanced
stages, -which shows the water-tube upon which is developed the actinal area.

Figs. 3, 5, 11, represent the opposite profiles of the anal extremity of the Brachiolaria, showing the water-
tube, upon which is developed the abactinal area.

EXPLANATION OF THE PLATES. 129

Fics. 4, 14, represent the ventral side of the anal extremity of the Braehiolaria, showing the extremities of
the actinal and abactinal areas of the Starfish.

Fin-s. 6-9, 13, represent the dorsal side of the anal extremity of the Bratliiolaria, in the successive
stao-es of wrowth of the young Starfish, showing the opposite extremities of the actinal and abactinal
areas of the Starfish.

Owinc to the partial transparency of the Braehiolaria, either the actinal or the abactinal area is always
projected upon the other, when the larva is seen in profile. In the dorsal or ventral views, the angle
made by the actinal and abactinal areas becomes visible.

Fi"-. 1. Actinal profile of the anal part of the water-tube («>') of the Bracliiolaria, previous to the appear-
ance of the pentagon of lobes. In stage of PI. III. Fig. 7.

Fie. 2. Somewhat more advanced actinal profile, showing the ambulacral pentagon, as well as the posi-
tion of the five rods of limestone, opposite the angles of the actinal pentagon, seen through the thick-
ness of the larva on the surface of the other water-tube (w). In Stage of PI. III. Fig. 8.

Fif. 3. Tlie same larva seen from the opposite profile, to show the abactinal area ; small Y-shaped rods
have appeared at the extremities of the simple rods.

Fin-. 4. The same larva seen from the ventral side of the Braehiolaria, to show the relative position of
the pentagons of the two areas; only two of the rods of the abactinal side are seen, wliile the edges
of three of the actinal folds Q) can be perceived, one above the other, on the foot-like projection
formed by the folding of the water-tube w'.

Fl"-. 5. A more advanced Starfish, in stage of PI. III. Fig. 10, from the abactinal profile ; the Y-shaped
appendages of the original rods have increased in number ; smaller independent Y-shaped rods have
made their appearance in tlie intervals between the larger ones, in the spaces corresponding to the
middle of the pentagon of the actinal side. The angles of the actinal pentagon are formed of a double
fold, the sides of which are concave ; the stomach is almost concealed by the great accumulation of
limestone granules on the abactinal area.

Fif. 6. Tlie anal part of a larva from the dorsal- side, to sliow the apparent dividing into elliptical com-
partments of the water-tubes (ic, to'), made by folding and the bending of the extremities of these
tubes (PI. III. Fig. 10).

Fi<T. 7. The same larva from the dorsal side, to show the manner in which the first fold (I) is made on
the exterior surface of the water-tube (w'), and the greater size of the right water-tube extending
over the digestive cavity to the madreporic opening (i).

Fig. 8. A Starfish from the dorsal side of the Braehiolaria (PI. III. Fig. 11) ; shows the lobes formed by
the two arms which are in view, with the large cluster of rods in the centre of the lobe, and the
small cluster in the space opposite the angle of two lobes.

Ficf. 9. The same view of a more advanced embryo, somewhat older than PI. IV. Figs. 1, 2; the lobes of
the arms have become indented, the arms themselves are separated by a deep cut, the Y-rods extend
so as to form almost a continuous network over the whole abactinal area. The actinal pentagon has
assumed the shape of prominent loops projecting beyond the foot-like, oblique fold of the water-tube.

Fig. 10. The same embryo, seen from the actinal profile; the tentacular loops stand out independently
from the surface of the water-tube ; the stomach and nearly the whole length of the intestine are
enclosed by the abactinal area.

Fig. 11. Seen from the abactinal profile in stage of PI. VII. Fig. 8; tubercles have formed upon the sur-
face, the Y-shaped rods extend into them, the lobes of the edge of the disk are deeper, the second
set of clusters of limestone cells have greatly increased.

Fig. 12. The same embryo from the opposite profile; the inner tentacular folds have become tipped with
a triangular point. The thickness of the abactinal surface prevents the network of cells on the edge of
the arms from being seen.

Fig. 13. A view of the embryo from the dorsal side of the Braehiolaria; the madreporic body (6), the
opening of the water-pore, is placed at the edge of the upper arm (?","'), the tubercles on the edge of
the arms are well shown by the great accumulation of small Y-sh:iped rods.

Fig. 14. The same from the ventral side of the Braehiolaria (PI. VII. Fig. 8). This figure shows, per-
haps better than any other, the relative position of the extremity of the two pentagonal warped sur-
faces. The rough outline of the Starfish is due to the manner in which the tubercles of the abactinal
surface project above. The Starfish in this condition is at the point of resorbing the larva, and of
closing the actinal and abactinal areas.

130 EXPLANATION- OF THE PLATES.

PLATE VI.

The young Starfish after the Brachiolaria has been resorbed.

Fig. 1. A young Starfish, seen from the actinal side; the anal and oral chisters of arms of the Brachio-
laria appear like small knobs, placed on opposite sides of the new mouth. The future rays are mere
lobes, and are not symmetrical.

Fig. 2. The same embryo, seen from the abactinal side, to show the arrangement of the netwoi-k of lime-
stone meshes.

Fig. 3. A more advanced embryo, in which all traces of the appendages of the larva have entirely
disajjpeared. Each side of the pentagon of suckers is a rosette made up of seven loops; the limestone
particles are deposited so as to project at the angle of the arms between these loops. The mouth is
movable, the pentagon is not closed, and the Starfish is not yet symmetrical ; the shape of the different
rays is not identical.

Fig. 4. The same embryo, seen from the abactinal side, showing the arrangement of the successively
formed rows of rounded spines and of the plates. The two ends of the open pentagon have approached
nearer than in Figs. 1,2; the outline is not yet regular.

Fig. 5. Magnified view of one of the ambulacra! tubes, with its rudimentary tentacles.

Fig. 6. The young Starfish, in which the two pentagons have almost closed, and been brought into
parallel planes. There has been a great increase in the size of the cut between adjoining rays ; the
spines also have grown longer and more pointed ; the limestone points of the angle of the rays
have advanced nearer the centre. The Starfish is not quite symmetrical, nor are the arms exactly
alike.

Fig. 7. The same embryo, from the actinal side, shows the great increase of the arabulacral system, the
tentacles being distinct pouches on each side of the main tube. The basal tentacles of one system are
much further apart than all the others, and this is the last indication that the ambulacral pentagon is
not closed.

Fig. 8. A more magnified view of the actinal side, when the ambulacral pentagon is entirely closed, and
the Starfish has become symmetrical, and all the basal suckers are equally distant.

Fig. 9. The ambulacral system of one arm, when confined by the circle of limestone which has been
formed round each ambulacral system ; the first two pairs of tentacles begin to develop disks ; they
become club-shaped; the three terminal tentacles are still closely connected, and show no sign of
any disk.

Fig. 10. An abactinal view of one ray and the centre of a young Starfish, in which the spines
project far beyond the edge of the disk. The arm-plates and the interradial plates have be-
come connected by a narrow bridge. The limestone rods are so much thickened by additional
deposits, that they form elliptical cells which have entirely lost the polygonal character of the
younger stages.

Fig. 11. One arm and a portion of the centre, from the abactinal side of the most a<lvanced of the young
Starfishes which have been raised by artificial fecundation. The spines are very prominent, long,
somewhat spreading, and becoming even fan-shaped. The limestone cells are gradually assuming the
character of the limestone cells of the adult, small cells within larger ones ; the cut between the rays
is very deep.

Fig. 12. The same young Starfish as Fig. 11, seen from the actinal side; the three pairs of tentacles
have suckers ; the deposit of limestone of the actinal area having the same cellular structure as that of
the abactinal area, though formed by the increase of small cells instead of rods. This Starfish also
shows the position of the madreporic body, immediately on the edge of the disk of the lower side ; the
eye is very prominent at the base of the odd terminal tentacle. The young Starfish (Figs. 11, 12)
is about four months old.

PLATE VII.

Fig. 1. Two rays and the centre of the Starfish (PL VI. Fig. 10), seen from the actinal side. All
the tentacles are encased separately by the limestone deposit of the actinal region. Tlie ten-
tacles have grown so long that they extend beyond the edge of the arm. The pair of terminal
tentacles has, as yet, increased but little in comparison to the other pairs. The odd terminal
tentacle has, at its base, a bright carmine spot, the eye, which appears about this time. The

EXPLANATION OF THE PLATES. 131

moutli, limited by the limestone deposit, takes the shape of a pentagonal opening; the ambulacral

tube is concealed.
Fig. 2. The same Starfish as PI. VI. Fig. 11, seen in profile, to show the great development of the abac-

tinal area, and the Echinus-like arrangement of the spines in the young Starfish. The odd tentacle

is seen turned up, between two of the spines, with the eye at its base.
Figs. 3-5. Spines of the young Starfish in different stages of growth.
Fig. 6. An enlarged view of the terminal tentacle, to show the position of the eye at the base of the odd

tentacle.
Fig. 7. An enlarged view of the meshwork of limestone cells, to show the mode of formation of additional

cells, by means of Y-shaped rods.
Fio'. 8. A "reatly masfnified fisjure of a full-<rrown Brachiolaria in its natural attitude, at rest, with the

Starfish almost ready to resorb the larva; the obliquity of tlie planes, in which the actinal and abac-

tinal pentagons are situated, is especially well seen in the pointed anal extremity of this Brachiolaria.

No letters have been added to this figure, as the different parts can readily be distinguished by com-
paring it with PI. IV. Figs. 1, 2, 4.

PLATE VI.II.

Fig. 1. Young Asteracanthion about one year old, seen from the abactinal side.

Figs. 2-4. Magnified views of spines (;>), and of rudimentary pedicellaria; (p', ;>")•

Fig. 5. Odd terminal tentacle of a Starfish in the stage of Pi. VIII. Fig. 10, at the extremity of the arm

with the eye-speck (e).
Fig. 6. One of the abactinal water-tubes ((/') at the angle of the rays.
Fig. 7. One of tlie abactinal water-tubes {<t") along the edge of the rays.
Fig. 8. Abactinal view of the arm of a young Starfish, probalily two years old.
Fig. 9. Actinal view of an arm of a young Starfish in its third year.
Fig. 10. Abactinal view of a young Starfish, in which the rudimentary pedicellarise have made their

appearance, also having median and latei-al lines of abactinal water-tubes along the arm. Probably

three years old.

PLATE IX.

Asteracanthion berylinus.

Fig. 1. Living specimen, seen from the actinal side.

Fig. 2. Living specimen, seen from the abactinal side.

Fig. 3. Preparation showing the calcareous network, abactinal side.

Fig. 4. Abactinal calcareous network seen from the interior.

Fig. 5. Preparation showing the connection of the solid parts, seen from the actinal side.

(a i) Interambulacral plates, with two rows of pores at base.

(a') Ambulacral plates, sliowing the alternating arrangement of the ambulacral pores penetrat-
ing between the ambulacral plates.

(a") Base of interbracliial partition.

(c) Ambulacral groove.

(i) Lateral imbricating pieces forming the calcareous network of the abactinal surface.
Fig. 6. The same, seen from the interior. Lettering as in Fig. 5.

(c) Dorsal groove of ambulacral system.

{ip) Interradial partition formed by soldering of the imbricating pieces attached to the interam-
bulacral plates.
Fig. 7. Longitudinal section of preparation of arm, to show the formation of the interradial partition by
the soldering of the imbricating lateral pieces of the interambulacral plates.

All Figures natural size.

The color of this species, as of all the species of the genus Asteri.is, varies greatly; it ranges from dark chocolate (on
the abactin.il side) to light violet. The actinal side is of a much paler shade of the same color. The general tint of the
abactinal side depends also greatly upon the state of expansion of the water-tubes and the development of the light-colored
pedicellariae clustered around the spines.

132 EXPLANATION OF THE PLATES.

PLATE X.

ECHINASTER SENTUS.

Fig. 1. Living specimen seen from actinal side.

Fig. 1'. Portion of arm of Fig. 1, somewhat more magnified.

Fig. 2. Same from the abactinal side.

Fig. 2'. AVater-tubes of part of abactinal surface.

Fig. 2". Madreporic body.

Fig. 3. Calcareous network of abactinal side.

Fig. 4. Internal view of abactinal surface.

Fig. 5. Calcareous network of actinal side (same as Fig. 3).

Fig. 6. Inner view of actinal calcareous network.

All Figures natural size, except Figs. 1', 2', 2", which are somewhat enlarged.

Tlie color of tliis species varies from a dark reildish-brown to a pale violet, sometimes more or less yellowish-brown or
purple. The water-tubes are light pink or violet.

PLATE XI.

ASTERIAS OCHRACEA.

Fig. 1. Single arm and disk, seen from the abactinal side.

Fig. 2. Single arm, seen from the abactinal side, with the spines of the limestone network removed.

Fig. 3. Interior view of limestone network.

Fig. 4. Actinal view of the disk and arm, to show the narrow ambulacral plates, the marginal interam-

bulacral plates, and the adjoining actinal limestone network.
Fig. 5. Inner view of the same, showing the liuge spaces left between the pillars forming the marginal

support of the limestone work adjoining the interambulacral plates.
Fig. 6. Portion of half of the arm, to show the arrangement of the ambulacral and interambulacral plates,

seen from the actinal side near the base of the arm-plates forming the median groove on top.
Fig. 7. Profile view of a similar portion of the arm (as Fig. G) toward the central part of the arm, seen

from the interior of the arm.

All Figures natural size, except Figs. 6 and 7, which are somewhat magnified.

This is often a very brilliantly colored species. Brandt has separated as species the extreme variations in color. The
most common coloring is a dark orange, passing in some specimens to an almost pure yellow, or in the other direction to
a rich chocolate color. We find also frequently violet as the prevailing tint. The ridges, on the abactinal network, are
invariablv of a ligliter tint than the ground-color.

PLATE XII.

Cbossaster papposus.

Fig. 1. Seen from the actinal side, with the spines of the interbrachial surface, and of the lower surface
of the arms. (The abactinal surface has been removed.)

Fig. 2. Seen from the actinal side, with the spines removed to show the structure of the plates carrying
the spines along tlie edge of the arms, round the actinostome, and of the limestone plates strengthen-
ing the interbrachial membrane ; the limestone network of the inner surface of the abactinal surface
is seen through the opening of the actinostome.

Fig. 3. Fig. 1, seen from the interior (from the abactinal side), showing the portion of the membrane ex-
tending as a division wall between arms, and forming the support as well as the connection between
the actinal and abactinal surfaces ; this membrane is often a mere film strengthened with limestone
plates only at its outer and inner e.^tremities, where it connects by more numerous and stronger
plates the two surfaces of the interbrachial space. The plates near the actinostome are frequently
drawn out into a long comma-shaped support on the abactinal part of the connecting membrane.

EXPLANATION OF THE PLATES. 133

Fio-. 4. Same as Fig. 2, seen from tlie abactinal side, to show the network of the abactinal surface and

the projecting knobs forming the support of the clusters of spines of that surface.
Fif. 5. Lonwitudinal section through the median line, seen in profile.

All figures natural size.

The coloring of tliis species is of all shades, between a brilliant red and a light orange or a dark violet.

PLATE XIII.

Pycnopodia helianthoides.

Fig. 1. Portion of disk, seen from .ibactinal side, -with papillte fully expanded (from life).

Fig. 2. Same as Fig. 1, seen from the actinal side.

Fig. 3. Actinal view of central part of the disk, showing the connection of the arms around the central

opening.
Fig. 4. Limestone network of part of the abactinal membrane, with the pillar separating adjoining arms

seen from the interior.
Fig. 5. Profile view of the extremity of one arm.
Fig. 6. Interior view of the arm ; the abactinal membrane is removed, showing the mode of connection of

adjoining arms at actinostome, ambulacral vesicles all removed.
Fig. 7. Same as Fig. 6 ; seen from below, the soft parts all being removed.
Fig. 8. Profile view of Fig. 6.
Fig. 9. Section across Fig. 6.

Fig. 10, 10". Profile views of two of the large spines of the abactinal surface.
Fig. 10', 10'". The same spines, 10, 10", seen from above.

Figs. 1-5 are natural size ; all others slightly enlarged.

The color of tlie abactinal surface varies greatly in this species, from a brilliant carmine to yellow, violet, or bright
vermilion, with the intermediate shades of orange.

PLATE XIV.

LiNCKIA GUILDINGII.

Fig. I. Seen from above.

Fig. 1'. Enlarged tip of one of the arms.

Fig. 2. Fig. 1, seen from the actinal side.

Fig. 2'. Magnified portion of arm of Fig. 2.

Fig. 3. Preparation of actinal side, showing the limestone plates after the granulation is removed.

Fig. 3'. Magnified -view of opening of actinostome of Fig. 2.

Fig. 4. Preparation of abactinal surface, showing the limestone plates of that surface.

Fig. 4'. Enlarged view of madreporic body.

Fig. 5. Interior view of abactinal surface of one of the arms, showing the small openings left between the

nearly united plates.
Fig. 6. Section across one of the arms, to show the depth of the ambulacral furrow, with its single line of

ambulacral pores.

Linckia is generally of an ashy-violet color, with darker spots scattered over the abactinal surface of the arms.

ASTERINA FOLIUM LiUk.

Fig. 7. Actinal view prepared to show the plates of that surface.

Fig. 7'. Enlarged view of plates, forming edge of actinal opening, in Fig. 7.

Fig. 8. Somewhat enlarged view, natural attitude, with suckers expanded.

Fig. 8'. Enlarged view of arm, showing ocular tentacle, at ba=e of pointed terminal ambulacral tube.

Fig. 9. Water-tubes of abactinal surface somewhat enlarged.

Figs. 1-7, natural size ; others somewhat enlarged.

The abactinal surface of Asterina is of a pea-green color. The actinal surface is more yellowish. Specimens frequently
occur of a yellow color on both sides.

134 EXPLANATION OT" THE PLATES.

PLATE XV.

ASTEROPSIS IMBRICATA.

Fit. 1. Seen from above; in two of the arms the water-tubes of the abactinal surface are represented as
fully expanded, while they are drawn in on the others.

Fif. 1'. Actinostome with the tentacles drawn in, taken from life ; the plates, except the marginal ones,
are all imbedded and liidden in the membrane of the actinal surface.

Fig. 2. Preparation showing the irregular limestone plates and needles of the abactinal surface.

Fi<'. 3. Portion of abactinal surface, seen from the interior, showing the original reticulation, which is lost
in the exterior view from the abactinal side.

Fif. 4. Fig. 2, seen from the actinal side, to show the arrangement of the limestone plates.

Fio-. 5. Interior view of the actinal floor, showing the broad ambulacral groove, the connection of the
ambulacral plates round the actinostome, and the position of the pillars connecting the actinal and
abactinal surfaces.

Fig. 6. Same as Fig. 5, seen in profile, to show the interbrachial arches and the great height of the am-
bulacral plates.

Fig. 7. Section across the arm near the tip ; the ambulacral plates almost touch the abactinal surfaces.

All figures are natural sizes.

On the actinal side Asteropsis is of a brownish color, with yellow edge along the ambulacral furrows. The abactinal
surface is most brilliantly colored witli large patches, irregularly aiTanged, of vermilion, bright green, blue, yellow, with
prominent carmine spots enclosing the areas for the passage of the water-lubes.

PLATE XVL

PenTACEROS RETICULATU.S.

Fig. 1. Arm and portion of the disk with water-tubes fully expanded and ambulacral tubes extending
beyond the edge of the arms near the tip, seen from the abactinal side.

Fig. 2. Same as Fig. 1, seen from the actinal side, the two rows of tentacles drawn in and ambulacral
furrow almost closed near the face.

Fig. 3. Actinostome natural size, with the ambulacral tentacles at base of furrow fully expanded.

Fig. 4. Actinal view of the lower surface, showing the limestone plates of the margin of furrow sup-
porting the papillse, and the plates covered by the granulation of Fig. 2.

Fig. 5. Interior view showing the ambulacral system, the connection of the ambulacral plates round the
actinostome, the thick abactinal surface with the nearly solid interbrachial limestone arches.

Fig. 6. Central portion of the abactinal surface of the disk, natural size, showing the massive reticulation
of the surface.

Fig. 7. Section through the centre of the ambulacral system, seen in profile, with the interambrachial
arches.

Figs. 3 and 6 somewhat enlarged ; others, natural size.

The general coloi-ing of this species is yellowish or pinkish brown, sometimes bright carmine. The ridges separating
the spaces for the passage of the water-tubes are a darker shade of the general color. On the actinal edge the plates are
of a darker brown color, while the .actinal surface itself is faintly colored gray.

PLATE XVII.

SOLASTER ENDECA.

Fig. 1. Limestone network of the abactinal surface.

Fig. 2. Dried specimen, with spines bordering the ambulacral furrows and covering the actinal surface.

Fig. 3. Interior view of the actinal floor, showing the narrow furrow of the ambulacral system, its con-
nection round the actinostome, the absence of a prominent interbrachial arch separating the central
arm-spaces.

EXPLANATION OF THE PLATES. 135

Fio-. 4. Same as Fig. 2, only denuded of spines to show tlie plates of the actinal surface supporting

them.
Fig. 5. Longitudinal section through the median line of the ambulacral furrow.

All Figures natural size.

This species is generally, on the abactinul side, a dirty orange yellow, passing into red; more yellowish and lighter in
shade on the actin.il surftice.

PLATE XVITL

Cribrella sanguinolenta.

Fi<T. 1. Seen from the abactinal side, to show the minute spines covering the whole disk and arms.

Fif. 2. Fig. 1, seen from the actinal side.

Fin-. 3. Enlarged view of spines round the actinostome and base of the arms.

Fio-. 4. Actinal view ; specimen denuded of its spines to show the close system of limestone plates form-
ing the actinal surface.
Fi"-. 5- Interior view of the abactinal surface, showing its close reticulation.

Fig. 6. Longitudinal section across the central line of the ambulacral furrow, somewhat enlarged.

Fig. 7. Interior view of the ambulacral system of one of the arms and its central connection, somewhat

enlarged.

Figures 3, 6, 7 are somewhat enlarged ; all others, natural size.

This species varies in color from brilliant orange-yellow to dark purple or dnll violet.

PLATE XIX.

ASTROPECTEN ARTICULATUS.

Fig. 1. Dried specimen, from the abactinal side.

Fig. 2. Fig. 1, seen from the actin.il side.

Fig. 3. Abactinal view of denuded specimen.

Fig. 4. Actinal view of denuded specimen.

Fig. 5. Interior view of actinal floor, showing ambulacral system, the interbr.ichial arches, the connection

of ambulacra round the actinostome.

Fig. 6. Interior view of the abactinal surface.

Fig. 7. Actinal region of specimen somewhat larger than Fig. 4.

Fig. 8. Longitudinal section through arm, showing the great thickness of the marginal plates.

All Figures are natural size.

The marginal plates are of a light brown color with darker edges; the abactinal surface of nearly the same shade, some-
what lighter than the plates. The color is sometimes arranged in darker rectangular patches along the edge of the arms
on the abactinal side. The marginal plates are also bright yellow, enclosing a violet abactinal surface.

PLATE XX.

LUIDIA CLATHRATA.

Fig. 1. Abactinal view from life.

Fig. 2. Same, seen from the actinal side, showing the two rows of pointed tentacles.

Fig. 3. Same view as Fig. 1, denuded specimen.

Fig. 4. Preparation showing the plates of the actinal surface.

Fig. 5. Interior view of actinal floor, showing ambulacral system and mode of connection round actinal

opening.
Fig. G. Interior view of abactinal surface.
Fig. 7. Actinal view of central part of disk, with tentacles contracted, showing the spines along the edge

of ambulacral furrows and at their junction.
Fig. 8. Longitudin.al section of aim.

136 EXPLANATION OF THE PLATES.

Fig. 9. Terminal knob of arm seen from above.

Fig. 10. A terminal knob, seen from the actinal side.

Fig. 11. The same, seen from the end.

Fig. 12. Madreporic body.

All Figures natural size, with the exception of Figs. 9-12, which are somewhat enlarged.

The coloring of this species is quite dull; it has a grayish hue, with large square patches of brown arranged along the
margin of the arms. The color of the abactinal surface is frequently iu lines, one in central part of the arms, the others
along tlieir margin. The ambulacral tentacles are yellow.

ERRATA.

On page S, 13th lino from bottom for BlUtscMi read Biitschli

25, 9tli " " for pallidus read herylinus

33, 7th " " for r," read r,'"

33, 8th " " for r/' read r^'"

58, 5th " " for or iJte read for the

CI, 1 2th " from top {or furroics rcxid four rows

74, 9th " " for Zeitscrif. read Zcilsclirifl

110,18th " " ior analofjoMS read an analogous

113, 2d " " lor when read lohere

Ey>-

l^ARFISHES

PL I

10.

20.

IS

Q^i"^^.

fi

1£.

ZZ.

13.

9

18.

23.

19.

E4.

A-i^asaiL.aei

ASTERAGANTfllON BEraLfflUS Ag

'•.RFIgHUS

pi TT

ASTERACANTfflON Bi^IHyLfflUS Ag

" ASTERAnANTBIOJr PAJ.LIDUS A^.

"ACMTnf!?

Ai.

"dl.on stoi-.

A Agassis,

'nil I ed by A Meise!. bos ton

.•iLLIDUS A^.

B'rAilFISHl';.:

;'- BamlioTi st.onv

r'rmted by AMeisei.iSoslon

AKTJiRACANTHIOJir PAI,LID1JS Ai.

iedtyAMeiselBoslon.

.tLMF"

PIM,

. A^SBSIE, ad

PrmtedbyA.Meisei, i. . .. .

ASTERACANTHIOJT lALLLJlJS A^

XnnTH American Starfishes

PI IX

nd Biirr:!'. from :

PtiTitby AMtisel

ASTERACANTHION BEP.YLINUS Ag"-

XnRTH American Starfishe?

PI X,

nd Burkf.ard-^ fra

Prin L "by A Meisel .

EcHiNASTER SENTUs Verr

\oRTH A>iERic.« Starfishes

PI XI

PrirvL.ty A Mti

ASTERIAS OCHRACEA Bf.

'X'ORTH American Starfishes

PI Xll

*r^f-

^;^^ :^sT-^t^ . V fX':'--->^^^^

ruT7iL. Irorr. r.at-j-

Prinlby A Meistl

Crossastzr" papposus MT

XoRTH American Starpishes

FM xni

A Agassiz froitiivaLt.

Print "by A Musel

Pycnododia helianthoides . SLimps

N^ORTH Americ.w Starfishes

PL XIV.

AAs'asEi:: dii.d Butril! ftorr. r.a.:u.:

PrmLly A Mci

-6 LiNCKiA GuiLDiNGii Gray, .7-9 Astertna folium Liitk

NopjH American Starfishes

P 1 , X\',

/fjh'- "f

' * ^ 0 ft i'»

Cv^'^'^

--.e'ass'.z Iron-, rs ■

AsTEROPSis iMBRicATA Grube .

A^ORTH AmEFUCAN STARFISHES

PlXAl

AAga-£si2 ari.d E-u.r.K>.ardi frorr

Print, by A Meisd

Pentaceros reticuiatus Link

XoRTH Amfric-W Starfishes

PI XVI!

BuiTiil fro:n nat c

Prmity A M«stl

SOLASTEPv- ENDECA ForteS

XoRTH American Startishes

XVI

A-Ag'assis and Eurr.:] frarr, r.a^ur-^

Cribrella sang-uinolenta Lutk.

XoRTH American Startishes

PI XIX

'««^.

Prini.'by A Meisel

ASTROPECTEN AKTICUIATUS M.T.

\oRTH America Starfishes

PI. XX

nd Bumll from nature

PrinL.V/ AMcisel

LUIDIA CLATHRATA LuLk .

Harvard MCZ LIbrar

3" 21344 066 300 724