more ele es
figuras she tesre nee ee
oe ne .

a Ae et ne whe Spm
ee ere

o>
een
ms “ - “seer . < ee
Facto —< ~ - ‘
paedinn aed nds

: . ewe
- r- ae ae . ~~ oe rene.

% oe ;

Oe ee ney og : ingens ~ e
Ee wm ee ~— :

nee ware ee) oma

Se ee ve, mat
Oban, —se=-

+

my oar

'-e~e~o
ere mee

.

‘
hh

4)e54

Pere et Or et

aetna ci die Dhoni ee act,
a +e ®

hind de
‘ ?
Tore me ong

or ot a ar ee ee 7
oe ee ee ee

428 doe be dt ata ete ete”

oe orale vw hime

tJ)

fF ;

Digitized by the Internet Archive
in 2009 with funding from
University of Toronto

http://www.archive.org/details/quarterlyjournal28comp

Noa i

ee hy

7 4 if
y Py)
” bie any Af o oe
Hi iy ‘ TN i
jul AO
Pe Ls

VAN HE Bld |
od ee hia AAT (Lar

bab hae |
Peis coc
Se ee

4 : hi 7]

a Oy VA

tc.

ee

A
iy
Ot
i
y ny
i]
ny
)
a

QUARTERLY JOURNAL

OF

MICROSCOPICAL SCIENCE:

EDITED BY

E. RAY LANKESTER, M.A., LL.D., F.R.S.,

Fellow of Exeter College, Oxford, and Jodrell Professor of Zoology in University
College, London ;

WITH THE CO-OPERATION OF
W,. T. THISELTON DYER, M.A., C.M.G., F.RB.S.,
Assistant Director of the Royal Gardens, Kew ;

EE KLEEN, M.D. PRS.

Joint- Lecturer on General Anatomy and Physiology in the Medical School of
St. Bartholomew's Hospital, London ;

H. N. MOSELEY, M.A., LL.D., F.RB.S.,
Linacre Professor of Human and Comparative Anatomy in the University of Oxford,
AND

ADAM SEDGWICK, M.A., F.R.S.,
Fellow and Assistant-Lecturer of Trinity College, Cambridge.

VOLUME XXVIII.—New Sezzizs,

With Aithographic Plates and Engrabings on Wood,
Can

a [87 5%

LONDON:
J. & A. CHURCHILL, 11, NEW BURLINGTON STREET,
1888,

CONTENTS.

CONTENTS OF No. CIX, N.S., AUGUST, 1887.

MEMOIRS :

The Anatomy of the Madreporaria: III. By G. Herpert Fow.er,
B.A.Oxon., Ph.D., Berkeley Fellow of the Owens Eee Man-
chester. (With Plates I and IT)

On the Anatomy of Mussa and Euphyllia, and on the ee
of the Madreporarian Skeleton. By G.C. Bourns, B.A., F.L.S.,
Assistant to the Linacre Professor in the University of Oxford.
(With Plates III and 1V)

On the Intra-ovarian Egg of some Osseous Fishes. By Roperr
ScuarFr, Ph.D., B.Sc. (With Plate V)

Observations on the Structure and Distribution of Striped and Un-
striped Muscle in the Animal Kingdom, and a Theory of Muscular
Contraction. By C. F. Marsuauz, M.Sc., Platt Physiological
Scholar in the Owens College, Manchester. (With Plate VI)

On the Fate of the Muscle-Plate, and the Development of the Spinal
Nerves and Limb Plexuses in Birds and Mammals. By A. M.
Paterson, M.D., Senior Demonstrator of Anatomy, and Lecturer
in Dental Anatomy and Physiology, in the Owens College, Man-
chester. (With Plates VI1 and VIII) . : :

Note on the Ciliated Pit of Ascidians and its Relation to the Nerve-
ganglion and so-called Hypophysial Gland; and an Account of the
Anatomy of Cynthia rustica(?). By Liman SHELDON,
Bathurst Student, Newnham gi y aeheee: pie Plates
IX and X) : :

The Tongue and Gustatory Organs of Mephitis aapiieiae By
Freprerick Tuckerman, M.D.Harv., Amherst, Mass., U.S.A.
(With Plate XI)

On the Quadrate in the Mammalia. By Dr. G. Baur, of New
Haven, Conn. : : : : : :

PAGE

109

131

149

169

iv CONTENTS.

On the Hemoglobin Crystals of Rodents’ Blood. By W. D. Hattr-
BuRTON, M.D., B.Sc., Assistant Professor of peared Uni-
versity College, London . ;

An Easy Method of obtaining Méthesiinglabi Crystals ae Micro-
scopic Examination. By W. D. Haturmurton, M.D., B.Sc.,
Assistant Professor of Physiology, University College, London .

CONTENTS OF No. CX, N.S., NOVEMBER, 1887.

MEMOIRS :

On the Development of Peripatus Nove Zealandie. By Lit1an
SHELDON, Bathurst Student, Newnham College. (With Plates
XII, XIII, XIV, XV, and XVI)

On Some Points in the Anatomy of Polycheta. By J. T. Cun-
NINGHAM, B.A., F.R.S.E., Fellow of University College, Oxford.
(With Plates XVII, XVIII and XIX) .

On Temnocephala, an Aberrant Monogenetic esate by
Wittiam A. Haswet, M.A., B.Sc., Lecturer on Zoology and
Comparative Anatomy, Sydney University. (With Plates XX,
XXI, and XXII)

Notes on Echinoderm Morphology, No. XI. On the Deydopaeat
of the Apical Plates in Amphiura squamata. By P.
HERBERT CARPENTER, D.Sc., F.K.S., F.L.S., Assistant Master
at Eton College .

PAGE

181

201

239

279

303

CONTENTS OF No. CXI, N.S., FEBRUARY, 1888.

MEMOIRS :

The Photospheria of Nyctiphanes Norvegica, G. O. Sars. By
RuereRt VALLENTIN and J. T. Cunnincuam, B.A., Fellow of
University College, Oxford. (With Plate XXII1)

On the Early Stages of the Development of a South American

Species of Peripatus. By W. L. Scrater, B.A., F.Z.S. (With
Plate XXIV) . : 2 ‘ :

319

343

CONTENTS.

On the Anatomy of Allurus tetraedrus (Eisen). By Frank E.
Bepparp, M.A., Prosector to the Zoological Society, and Lec-
turer on Biology at Guy’s Hospital. (With Plate XXV)

The Development of the Cape Species of Peripatus. Part IV. The
Changes from Stage G to Birth. By Apam Szpewick, M.A.,
F.R.S., Fellow of Trinity College, Cambridge. (With Plates
XXVI, XXVII, XXVIII, and XXIX)

On the Occurrence of Numerous Nephridia in the same Segment in
Certain Earthworms, and on the Relationship between the Kx-
cretory System in the Annelida, and the Platyhelminths. By
Frank E. Bepparp, M.A., Prosector to the Zoological Society
of London, and Lecturer on Biology at Guy’s ere (With
Plates XXX and XXXI) : ,

The Anatomy of the Madreporaria, IV. By G. eaten Towees.
B.A., Pa.D., Assistant to the Jodrell Professor of Zoology in
University College, London. (With Plates XXXII and XXXIIJ)

CONTENTS OF No. CXII, N.S., APRIL, 1888.

MEMOIRS:

A Monograph on the Species and Distribution of the Genus
Peripatus (Guilding). By Apam Srpewicx, M.A., F.R.S.,
Fellow of Trinity College, Cambridge. (With Plates XXXIV,
XXXV, XXXVI, XXAVII, XXXVIII, XXXIX and XL)

Notes on the Anatomy of Peripatus capensis and Peri-

, patus Nove-zealandiz. By Lizian Sueupon, Bathurst
Student, Newnham College, Cambridge . 5

On the Construction and Purpose of the so-called Tabyrinthine
Apparatus of the Labyrinthic Fishes. By Doctor NicHoxas
ZocRaFr, of Moscow. (With Plate XLI)

Studies on the Comparative Anatomy of Sponges. I. On the
Genera Ridleia, n. gen., and Quasillina, Norman. By
Artuur Denny, M.Sc., F.L.S., Demonstrator and Assistant
Lecturer in Biology in the University of Melbourne. (With
Plate XLII) ; ‘ : , :

Kleinenberg on the Development of eed ORE By G. C.
Bourne, B.A., F.L.S., Fellow of New College, Oxford, and As-
sistant to the Linacre Professor in the University

V

PAGE

365

373

413

431

4.95

501

513

531

The Anatomy of the Madreporaria: III.
By

G. Herbert Fowler, B.A.Oxon., Ph.D.,
Berkeley Fellow of the Owens College, Manchester.

With Plates I, II.

Tue present memoir deals with the anatomy of Turbinaria
(p. 1), a colonial Perforate coral; of Lophohelia (p. 6),
an Imperforate form, colonial but with separate calyces; and of
the two aberrant Imperforate genera, Seriatopora (p. 10)
and Pocillopora (p. 13), in which the calyces are merged
in coenenchyme. ‘To these descriptions is appended a note on
the skeleton of Flabellum.

The most important facts now described for the first time
are—l. 'The absence of directive mesenteries in Lophohelia,
which thus differs from all Hexactinie hitherto described. 2.
The retraction of the tentacles of Seriatopora by introversion,
of which no other instance is known among the Madreporaria.
3. The presence of centres of calcification in the theca.

As in previous memoirs (2. 3), I have endeavoured to let the
‘figures speak for themselves rather than to give detailed
descriptions of structure.

TURBINARIA, sp. (figs. 1—3).

For the opportunity of investigating this form, as in pre-
vious instances, I am indebted to the liberality of my teacher,
Professor H. N. Moseley, who procured the material during
the voyage of H.M.S. “ Challenger.”

i, Corallum.—The colony is crateriform or goblet shaped,

VOL, XXVIII, PART ]1.—NEW SER. A

2 G. HERBERT FOWLER.

the calyces of the polyps being placed on the inner face and
on the brim of the goblet. The coenenchyme is porous, in the
manner characteristic of the Perforata, but the echinulations
_ are not arranged in costze with the regularity observable in
some genera, except on the actual thece of the polyps. The
latter project outwards from the ccenenchyme only abaxially,
i. e. inwards towards the centre of the goblet, the axial half being
almost level with the general surface. Specimens of the coralla
of this genus are not uncommon in museums; a detailed des-
cription and figures are therefore unnecessary, and may be found
in the works of the authors appended below (p. 6).

Owing to the small amount of the material at command,
none could be spared for the determination of the species. It
appeared, however, to belong to the type of T. mesenterina.

The septa, in fully-grown polyps of this particular species,
vary much in number, but are generally from seventeen to
twenty-two; they are entocclic only. It is worthy of re-
mark that the number of septa appears to bear no relation to
any multiple of six, nor can any division into orders be
effected, since all are approximately of the same length. A
loose and incomplete columella, occurring deep down in the
calices, appears to be referable to fusion of the septa.

Part of a transverse section through the corallum (made
according to the balsam and ether method introduced by von
Koch) is represented in fig. 1, showing sections through at
least five polyp cavities. Of these one, a, is cut obliquely,
owing to the sharp angle at which the polyp cavities are in-
clined to the general axis of the colony; of the others, which
are cut through at varying distances from their orifices, that
lettered 0 is a nearly transverse section, of a typical character,
exhibiting eighteen septa; while the three others, c, show the
reduction of the septa in the deeper parts of the cavities. The
upper part of the figure represents the abaxial, the lower the
axial, surface of the crateriform colony. The echinulations and
canal system are also well shown in this section.

ii. Anatomy.—The whole colony, both inside and outside of
the goblet, is clothed with an external body wall of ectoderm,

THE ANATOMY OF THE MADREPORARIA. a

mesoglcea,! and endoderm, exactly as has been described in
Stylophora (7), Madrepora (3), &c.; its continuity is only
broken by the mouth orifices of the polyps.

What the exact relation of this body wall to the tissues
actually apposed to the theca may be, 1.e. whether it agrees
with the undoubted relations described in Astroides embryo
(8), in Dendrophyllia (5), and Rhodopsammia (2), or with the
apparently equally accurate relations recorded for Stylophora
(7) and Madrepora (3), is exceedingly difficult to determine.
In my specimens, both of Madrepora and Turbinaria, the con-
traction produced by preservation in alcohol has forced the
body wall so tightly upon the echinulations that they project
in many cases through it. Again, in Heteropsammia mul-
tilobata, a form as closely allied to Rhodopsammia as one
with ccenenchyme can be to one devoid of it, the relations
appear to be identical with those in Madrepora and Stylophora,
and such as I have here (fig. 2) drawn for Turbinaria. If we
are justified in crediting the appearance of the tissues in Sty-
lophora, Madrepora, and Turbinaria, which implies that the
body wall is supported upon the echinulations (fig. 2), it
is not perhaps too much to infer that these relations are of
secondary significance, and have arisen contempo-
raneously with the development of cenenchyme,
for the support of the external body wall, owing
to the inadequacy of the peripheral sections of the mesenteries
to effect this support elsewhere than immediately round the
theca (where this is exsert from the ccenenchyme). In other
words, the mesenteries are necessarily confined to the polyp
cavity, and their peripheral sections to the small part of it
which is cut off (?) from the rest by the upward growth of the
theca; while therefore they are amply sufficient for the support
of the body wall in a form with separate free calicles (e.g.
Rhodopsammia), they could not extend over the cenenchyme
of a form with fused or sunken calicles (e.g. Heteropsammia

1 The substitution of this word for the misleading “‘ mesoderm” we owe to
Bourne (1).

4 G. HERBERT FOWLER.

multilobata), which is, by its very nature, outside of, and in a
manner independent of, the polyp cavities.

Of the more primitive (?) condition, Astroides embryo (8),

_Rhodopsammia (2), Dendrophyllia (5), and Fungia (1) stand
as admitted examples among the Perforata; Cladocora (4),
and Caryophyllia (6) among the Imperforata, all being forms
with free calyces ; while of the secondary condition, Madrepora
(4), Heteropsammia multilobata (which I hope to describe
in a future memoir), Turbinaria among the Perforata, and
Stylophora (7), Seriatopora and Pocillopora (described below)
among the Imperforata, are the recorded instances, all possess-
ing well-developed ccenenchyme.

In a minor point only do my observations differ from those
of von Koch (7), viz. that he figures no mesoglea between the
echinulations and the ectoderm ; in other words, according to
his figure the persistent ectoderm of the body wall is at those
points continuous with the calicoblast layer (fig. 5). The
reason which leads me to believe in the existence of a meso-
gloea lamina between calicoblasts and external ectoderm, is
that at the points where through shrinkage the echinulations
have pierced the external body wall, they have carried with
them this mesoglea, which in sections of decalcified specimens
preserves accurately their outline, projecting far beyond the
shrunken ectoderm.

Between this external body wall and the corallum lies the
system of approximately longitudinal canals with transverse
commissures ; in other words, the space between the body wall
and the theca is broken up into canals by the points of contact.
These canals communicate, as is usual in Perforata, with the
canals which permeate the corallum and run also into the
polyp cavities.

The polyps are built on the normal Actinian type. As
the calices are placed only on the inner side and on
the lip of the crateriform colony, an easy identification of
bilaterality is thus afforded, the dividing plane being a radius
directed to the centre of the goblet. Approximately at the
ends of this dividing plane are placed the axial and abaxial

THE ANATOMY OF THE MADREPORARIA. 5

pairs of “ directive ”’ mesenteries, distinguished by the arrange-
ment of retractor muscles on their ectoccelic faces. The
polyps are not, however, rigidly bisymmetrical, inasmuch as
the pairs of mesenteries lying right and left of the dividing
plane are not equal in number.

The total number of pairs of mesenteries is not constant,
but does not appear to depend upon the size (age) of the parti-
cular polyp. It varies generally from 17 to 22. The asym-
metry of the polyps can best be seen in a tabular form :

A B c
Number of pairs of mesenteries . realy, 20 22
Number on right side of “ directives ” ELEY f 10 9
Number on left side of “ directives ” as 8 aul

The three polyps here quoted were within a few millimetres
of each other, and all were nearly of the same size.

The number of pairs of mesenteries is of course the same as
that of the septa, the latter being entoccelic only, though a
misleading appearance of ectoccelic septa is produced by the
fact that some pairs of mesenteries die out after a very short
course, while their septa are still recognisable at a much greater
depth in the polyp cavity. The mesenteries with a longer
course are in all respects perfectly normal, and in my specimens
bore huge ova, the structure and relations of which call for no
special comment (fig. 3.) The length or shortness of the
mesenteries appears dependent on no particular system, such
as has been observed in some other forms (8).

The tentacles are probably entoccelic only, but are so retracted
as to render the point somewhat obscure. In this condition
they are covered by a ring-fold formed of the indrawn margins
of the disc, a method of protection common among the Acti-
niaria.

The histology, though much spoilt by prolonged decalcifica-
tion, agrees with that of the typical forms already described.
The muscular pleats of the mesoglea of the mesenteries are
only very slightly and irregularly developed, but entirely
normal. Nematocysts are closely packed together in the

6 G. HERBERT FOWLER.

tentacles; they are not however, arranged in knobs or
‘‘ batteries.”

Zooxanthelle are present abundantly in the canals exterior
.to the theca, in the tentacle cavities, and immediately under
the mouth disc; elsewhere they are comparatively rare.

iii, Summary.—The following are the most important points
elucidated :

1. The polyps are of the normal Actinian type, and are
bilateral, but not rigidly bisymmetrical.

2. The septa and tentacles (?) are entoceelic only.

3. The number of septa present is inconstant, and bears
no relation to any multiple of six.

4 The general body wall of the colony is supported upon
the echinulations of the cenenchyme; a condition
which may be of secondary significance, acquired for the
purpose of such support, contemporaneously with andin
consequence of the development of cenenchyme.

iv. Memoirs referring to the Genus:

Mitne-Epwarps and Hatme, ‘ Hist. Nat. des Coralliaires,’
ini, 164, pl. £1, figs. la, 10.

KiunzinceR, ‘ Korallthiere des Rothen Meeres,’ ii, 50.

LorHOHELIA PROLIFERA (figs. 4—8).

The material for a study of this form was entrusted to me
by Professor E. Ray Lankester, who had dredged it off Lervik,
Stordoe, Norway, and whom I am glad to be able thus to thank
for his generosity. Owing to the great density of its corallum,
and the consequent damage to the tissues produced by pro-
longed decalcification in a strongly acid medium, the work has
been long delayed. Part of the material had been placed
directly in absolute alcohol; part was passed from corrosive
sublimate through successive strengths of spirit to 90 per cent.
alcohol. Both sets were in excellent preservation, but the
latter method appeared to be preferable, as resulting in less
shrinkage of the tissues.

i. Corallum.—Of all corals this is probably the most generally

THE ANATOMY OF THE MADREPORARIA. 7

familiar, and requires here no systematic description. The
theca, which terminates the branches of the corallum, is solid,
as in all such Imperforata. The septa, which are exsert
above the lip of the theca, are both ectocelic and entoccelic,
but are only irregularly arranged in orders. In a polyp with
forty-eight septa, for instance, of which twenty-four are ecto-
ceelic, the remaining twenty-four entoccelic septa are probably
divisible into six primaries, six secondaries, and twelve ter-
tiaries ; but, as they all are approximately of the same length,
this division is founded more on analogy than on distinctive
differences. The total number of septa, which probably varies
with the age of the individual polyp, is not necessarily a multiple
of six or twelve.

Transverse sections of the corallum show, as has been
recorded for other forms, e. g. Cladocora (4), Caryophyllia (6),
that a dark line, indicating its earliest formed part, runs down
the centre of each septum, and may be termed a “ centre of
calcification.” In addition to these lines, however, sections,
so made that they shall just cut the extreme lip of the actual
theca, exhibit other “centres of calcification”? between the
enlarged ends of the septa, i.e. they lie in the theca itself
(fig. 4). In sections at a lower plane the centres of calcifica-
tion in the theca and in the exoccelic septa are found to have
run into a continuous dark line (fig. 5), which at a yet lower
level is joined by those of the entoccelic septa. There are
thus three separate centres of active coral secretion at three
different levels.

From fig. 5 it is also obvious that by far the greatest thick-
ness of the coral is laid down peripherally, i. e. by the calico-
blasts of the extrathecal part of the polyp. About six-sevenths
of the thickness of the theca is due to these calicoblasts, while
the remaining seventh is formed by those internal to the
theca.

ii. Anatomy.—In spite of the great length of the branch on
which it is borne the polyp is often comparatively short, mea-
suring from 5 mm. to 20 mm.

As will probably prove to be the case in all the Imperforata

8 G. HERBERT FOWLER.

with free calyces (cf. Cladocora (4), Caryophyllia (6), &c.), the
polyp is so continued over the lip and outer side of the calyx
as to form a covering for its exterior surface to a varying dis-
tance (the “ Rand-platte” of v. Heider). In Lophohelia this
continuation may extend for about 15 mm., or even more;
often it measures much less, and in it the relations of the
various body layers are such as have been already described
in the forms referred to above; the part of the colenteron
enclosed in this “ Rand-platte” is divided up into exoceelic and
entoccelic spaces nearly corresponding to those inside the calyx
by the peripheral lamellae which were at a former time con-
tinuous with the more central mesenteries, but which have
been mainly cut off from them by the gradual growth of the
theca upwards, though the continuity is maintained above the
lip (fig. 6). This explanation, due originally to Dr. von Koch,
must undoubtedly apply to this and many other adult
forms.

The general anatomical relations of the polyp, and its agree-
ment with forms already described, are shown in the diagram-
matic segment of a transverse section (fig. 6). The “ Rand-
platte,”’ the mouth disc, tentacles, and stomatodeeum are all in
accordance with the normal type. The celenteron of the living
polyp is, as usual, lined by endoderm and mesogloea, apposed
directly (except for scattered calicoblasts) to the corallum.
At the point, however, where the living polyp ceases, its coelen-
teron is separated off from the cavity in the coral which it pre-
viously occupied by a plug of decaying (?) tissue, in which no
cell-elements or organic structure are recognisable, except occa-
sionally the remains of the mesoglcea lamina of a mesentery.
Into this the living tissues pass gradually.

The tentacles, which are both ectoccelic and entoccelic, i. e.
one over every septum, are knobbed, each knob being such a
battery of nematocysts as has been described in Flabellum (2),
Stephanotrochus (11), &c.

The mesenteries which, like the septa, vary in number in
different polyps, all bear retractor muscles on their entoccelic
faces, i.e, there are no pairs of “directive” mesenteries

THE ANATOMY OF THE MADREPORARIA. 2

at the opposite ends of the long axis of the oval stomatodeum,
thus differing from those of all other Hexactinize or Madre-
poraria yet described. The significance of this fact cannot, of
course, yet be understood, as nothing correspondingly abnormal
occurs in any other part of the polyp with which it might be
correlated. As, however, the mechanical or other function of
the directive mesenteries is itself not yet explained, the mean-
ing of the variation from the common type is naturally not
appreciable. The number of the pairs of mesenteries, like that
of the septa, is not necessarily a multiple of six.

The only point in the histology that appears worthy of note
is the great length of the calicoblasts, as compared with that
of the other cell-elements. A group of them from the edge of
a growing septum is represented in fig. 7. Still more marked
is the great length of these cells in fig. 8, which represents a
transverse section through the tissues at that pqypt where the
upward growth of the theca divides the mesenteries into a cen-
tral portion within the calyx, and a peripheral portion outside
of it. Here they measure as much as ‘054 mm. The large
plate of mesogloea in the centre of the figure is merely that
which immediately overlies the lip of the calyx, and is cut in a
direction parallel to its flattened surfaces, while the section
passes nearly at a right angle to the other tissues. The point
here figured is such a “ centre of calcification” in the theca as
has been already referred to ( vide p. 7).

iii, Summary.—The most important facts thus obtained
are—

1. The polyps agree with the normal A ctinian type, except
for the absence of “directive mesenteries.” They
possess a well-developed “ Rand-platte.”’!

2. The septa and tentacles are both ectocelic and ento-
cclic, the number of septa not being necessarily a multiple of
SIX.

3. Three series of centres of calcification are recognisable in
the skeleton, of which one lies in the theca itself, and the

1 It is, perhaps, unnecessary to coin an equivalent for this till its morpho-
logical value is better understood,

10 G. HERBERT FOWLER.

other two at the summits of the ectoceelic and entoccelic septa
respectively.

iv. Memoirs referring to the Genus :

Mitne-Epwarps and Hare, ‘ Hist. Nat. des Corall.,’ iu,
116.

Stuper, Steinkorallen auf der Reise S. M. “ Gazelle”
gesammelt, Monatsb. Akad. Berlin, 1877, p. 631, pl. 1, fig. 8.

Mose zy, “ Challenger” Rep. Zool., 11, 178, pls. vir, rx.

SERIATOPORA SUBULATA (figs. 9—13).

For the material for the study of this coral and of Pocillo-
pora I again owe my thanks to Professor H. N. Moseley, who
has already investigated the general anatomy of both forms (10).
As, however, no structural details have yet been figured, and
these somewhat aberrant forms are of great interest, no apology
is necessary.for a second account of them. The specimens of
Seriatopora were obtained by Mr. Gulliver from Zanzibar.

i, Corallum.—The characteristic feature of the skeleton
which caused both Seriatopora and Pocillopora to be ranked in
the now abandoned group of Tabulata is the presence of
tabulz, i. e. successive floors of coral, by which the living polyp
shuts off its coelenteron from the cavity it previously occupied,
a condition the opposite to that described above in Lopho-
helia prolifera. The calyces are therefore nearly confined to
the outermost part of the colony, and are not continued
deeply into it, as was the case in Turbinaria. These shallow
calyces project but slightly above the ccenenchyme, and at a
very short distance below the orifice are divided into two
halves by the fusion of the two larger septa. These two septa,
the axial and the abaxial, are the only two that are developed
to any extent, though traces of the other ten may be recog-
nised in many cases (cf. the condition of Madrepora Dur-
villei (3). When all are present there are six entoccelic and
six ectocelic. It is perhaps more accurate to speak of the
calyx as divided into two halves by the fusion of these septa
than to regard the two chambers thus formed as special pits
for the reception of the two longer mesenteries (10), since

THE ANATOMY OF THE MADREPORARIA. i

they are simply downward continuations of the conical celen-
teron, and mesenteries other than the two longer ones are
sometimes attached to their sides. Other details of the
skeletal structure do not especially bear on the anatomy of the
polyps.

ii. Anatomy.—As was shown by Professor Moseley, Seria-
topora is undoubtedly a Madreporarian, and is even more in
accordance with the normal types than could be inferred
without the aid of sections.

The whole of the colony is clothed in the customary body
wall of ectoderm, mesogloea, and endoderm (fig. 9), which is
supported on the echinulations of the ccoenenchyme (vide
supra, p. 3). The space between body wall and theca is
broken up by these spines into a superficial series of canals
(figs. 9, 10, 13), which ramify over the ceenenchyme and place
the polyp cavities in communication with each other, but do
not, of course, extend into the corallum in the manner cha-
racteristic of the Perforata. The body wall is continuous with
the mouth disc, and from the centre of the latter rises a slight
hypostome, through which opens the stomatodeum. This
latter is crucial in transverse section, the longer arms of the
cross being in the dividing plane of bilaterality indicated by
the axial and abaxial septa (fig. 10).

The tentacles, which are twelve in number, being both ecto-
ceelic and entoccelic, are simple evaginations of the ccelenteron,
tipped with a terminal swelling, which is a single “ battery”
of nematocysts fig. 11). There is, I believe, no instance yet
recorded of the occurrence among Madreporaria of the method
of tentacular retraction which distinguishes Seriatopora,
namely, that of introversion (figs. 12, 13), the tentacles
being invaginated in such wise that the battery is still pointed
upwards. In fig. 13 the ectocelic tentacles are expanded,
while the entoceelic are introverted, a condition not uncom-
mon in wy specimens. Probably owing to the minuteness of
the polyp, no special muscular apparatus for effecting this
retraction could be detected.

The mesenteries, which are twelve in number, are arranged

12 G. HERBERT FOWLER.

in pairs! on the normal type. In the diagram (fig. 9) they
are numbered in the same manner as those of Madrepora (8) ;
the two mesenteries marked 5 and 10 respectively are compa-
ratively long, extending to the bottom of the polyp cavity, and
possess the thickened edge known as a mesenterial filament ;
of the rest, those numbered 1, 5, 8, 12, though generally de-
void of a “ filamentar”’ thickening, are recognisable in trans-
verse sections for some distance below the stomatodzeum ;
while the others, 2, 4, 6, 7,9, 11, are rudimentary, and are
visible only in the highest sections. It is worthy of remark
that the six rudimentary mesenteries last mentioned are those
which in the one type of polyp of Madrepora Durvillei,
are pierced by a special ectodermal canal, and which in the
other type of polyp of the same species, and in all the polyps
of M. aspera, are distinguished from the remaining six by a
greater length and the possession of a filamentar thickening ;
in other words, of the total twelve mesenteries the
six which in the one form are the best developed
are in the other quite rudimentary.

The histology agrees with that of the normal types. I have
found no trace of generative organs in my specimens.

iii, Summary.—The interesting points in Seriatopora are :

1. The polyps are Actinian in structure.

2. The septa when all are present, and the tentacles, are
both ectocelic and entocelic.

38. The tentacles are retracted by introversion.

4, The body wall is supported upon the echinulations of
the cenenchyme.

5. Of the twelve mesenteries, six (and more especially two
of these) are of some length, and six are rudimentary; but

1 Professor Moseley (10) states that in Seriatopora and Pocillopora the
mesenteries “ are not disposed in pairs with regard to the septa;” and the
remark reappears in a misleading form in Professor Martin Duncan’s “ Re-
vision of the Madreporaria” (‘Journ. Linn. Soc. Zool.,’ vol. xviii), to the
effect that “the genera differ from other Madreporaria in not having their
mesenteries arranged in pairs.” The original statement was correct because
the possibility of ectoccelic septa in a coral had not been demonstrated.

THE ANATOMY OF THE MADREPORARIA. 13

those which here are well developed are, in the Madrepore
mentioned above, rudimentary, and vice versa,

iv. Memoirs referring to the Genus :

Miine-Epwarps and Haring, ‘ Hist. Nat. Corall.,’ iii, p.
311, pl. F4, fig. 3.

Kuunzinesr, ‘ Korallthiere des Rothen Meeres, ii, 69, pls.
vii, Vili.

Acassiz, ‘Nat. Hist. United States, iv, p. 296, pl. 15,
fig. 15.

PocILLOPORA BREVICORNIS (figs. 14, 15).

The anatomy of this species agrees so closely with that of
Seriatopora subulata that only points of difference between
the two need be quoted.

The corallum is, of course, different in its mode of growth,
as upon this the distinction between the two genera is hased,
but this difference does not affect the anatomical relations of
the polyps. The method of support of the external body wall
is identical with that in Seriatopora; the tentacles agree in
the two forms, though, as they are fairly well expanded in my
specimens, it does not appear whether they are capable of in-
troversion or not; the stomatodzum is less distinctly conical
than in the cognate genus.

As regards the mesenteries, the only points of difference
noticeable are, that in Pocillopora those denoted in the dia-
gram (fig. 9) by the numbers 3, 10, are not proportionately so
much longer than those marked 1, 5, 8, 12, and that a mesen-
terial filament may sometimes be detected on the four last
mentioned ; in other words, the tendency observed in both
Seriatopora and Madr. Durvillei towards the exclu-
sive assumption of function on the part of six mesen-
teries and towards a correlated retrogression (?) on
the part of the other six, has not attained to such a
pitch in Pocillop. (and Madr. aspera) as in the other
two forms.

The statement of Professor Moseley (10), that “the mesen-
terial filaments are not enclosed in prolongations of the

14 G. HERBERT FOWLER.

chamber walls” is not justified by the examination of sections ;
the two longer and more developed mesenteries with their
filaments, 3. 10, lie, as in Seriatopora, for their whole length
in the coelenteron.

Apparently, any of the mesenteries may bear generative
organs, and it is worthy of remark that the polyps are
monecious. The ovaries and testes, though surrounded by
a thin capsule of mesoglea and endoderm, as in typical forms,
do not lie, as is generally the case, in the plane of the mesente-
ries (cf. fig. 8), but project from their sides in a manner more
characteristic of certain AJcyonaria, so that in transverse
sections of the colony they frequently appear to lie free in the
celenteron. Two stages in the development of the sperma-
tozoa are figured as well as the preservation of the material
would allow (figs. 14, 15).

In both this and Seriatopora there was left, after decalcifica-
tion, a residue in the position occupied by the corallum, which,
though staining faintly both with hematoxylin and with borax
carmine, showed no distinctly organic structure.

In transverse sections of the polyps it is just possible to
detect, at the point of the insertion of the mesenteries in the
corallum, structures similar to those described by Sclater (11)
as calicoblasts. Their excessive minuteness in Pocillopora
rendered an accurate investigation impossible, but they cer-
tainly appeared to me to be rather connected with the attach-
ment of the mesentery to the corallum than with the secretion
of coral.

Mitne-Epwarps and Haine, ‘ Hist. Nat. des Corall,’ iii, p.
301, pl. F4, figs. 1, 2.

Agassiz, ‘ Nat. Hist. United States,’ iv, 295, pl. xv, 14.

Kxunzincer, ‘ Korallthiere Roth. Meer.,’ ii, 66, pls. vil, vill.

Note ON THE SKELETON OF FLABELLUM. ;
In his latest addition to the literature of the subject (9), the
main part of which I do not propose at present to discuss,
Dr. von Koch treats, amongst others, of the skeleton of
Flabellum.

THE ANATOMY OF THE MADREPORARIA. 15

1. He states that the dark line of growth, visible in trans-
verse sections of the calyx, which indicates the earliest formed
part of the coral at that level, is in Flabellum placed peri-
pherally (fig. 16), and consequently that the skeleton is laid
down from without inwards.

2. Elsewhere in the same paper he infers, from his researches
on the development of Astroides calycularis (8) that the
epitheca of all corals, originally deposited outside the lateral
body wall of the embryo, also increases in thickness on the
inner side only.

3. Finally, we find that “diese Koralle bildet einen ganz
eigenen Typus, wegen des giinzlichen Fehlens der Innen-platte.
Die Aussen-platte! ist gut entwickelt . . . Die Homologisirung
der Aussen-platte mit der Innen-platte (Theca) der vorhin
beschriebenen Korallen wird aus der Struktur derselben als
irrig erkannt.”

The implied argument may thus be be expressed in the
syllogism :

1. The skeleton of Flabellum grows in thickness from
without inwards.

2. An epitheca grows in thickness from without inwards.

3. Therefore the skeleton of Flabellum is an epitheca—an
example of what is characterised by logicians as the fallacy of
the undistributed middle term (“‘ Medium non Distributum”).

Dr. von Koch is no doubt correct in asserting that the calyx
of Flabellum is laid down from without inwards ; but till clearer
evidence be adduced to the contrary it is far simpler to regard
it as a theca entirely homologous with the theca of typical
Madreporaria (or at least with a part thereof), than to conceive
that the epitheca, which we know elsewhere only as an incon-
stant and inconsiderable structure, should have replaced the
solid theca, merely to achieve the same physiological
end.

Nor is there anything in the structure of the corallum really
inconsistent with the idea that it isa theca. The embryonic
Flabellum patagonicum attaches itself to an Arenaceous

1 Tie, Epitheca.

16 G. HERBERT FOWLER.

Foraminifer, or some similar body (v. Moseley, ‘ Rep. Chall.
Zool.,’ ii, Madrep., pl. xv, figs. 1, 2) ; but the adult is entirely
free, and therefore more or less at the mercy of natural acci-
dents, such as currents. Correspondingly with this condition,
but unlike that of the attached forms (Lophohelia, Caryo-
phyllia, &c.) it developes no “ Rand-platte,” (vide p. 8), but
the polyp can be wholly retracted within the calyx (cf. (2)
fig. 2). The absence of the “ Rand-platte”’ implies almost neces-
sarily the absence of extracalicular calicoblasts; the calyx
must therefore be deposited by those internal to the corallum.
As a consequence of these facts, the calyx of Flabellum, if it
be not an epitheca, would be homologous with that part of
the theca of Lophohelia, &c., which lies internal to the dark
line of growth mentioned above (p. 7) ; and a comparison of
fig. 16 with figs. 4, 5, will show that there is no discordance
between the two structures.

In both forms, as is generally the case, the regions due to
separate centres of coral secretion are bounded by sutures ;
and of these regions those marked T. (thecal), and S. (septal),
in all three figures certainly appear to be respectively homo-
logous. Even the way in which the dark line in Lophohelia
curves inwards to the ectocclic septum (owing to the fact that
at the lip the latter does not project so far peripherally as an
entoceelic septum) agrees with the involution of the septa in
Flabellum. The fact that in fig. 5 the centre of calcification
of the entoccelic septum projects outwards through the line of
growth, is of course attributable to the pseudo-coste occurring
at the lip of the calicle of this species of Lophohelia which are
produced by extra-calicular calicoblasts and are not therefore
represented in Flabellum.

In fig. 17 is drawn a transverse section through a part of the
pedicle, that is to say a section through the corallum of an
embryo Flabellum measuring about 2°25 mm. in diameter and
possessing six primary and six secondary septa. The relations
indicated by the sutures are the same as in the former section.
The successive laminz showing the conversion of the embryonic
calyx into a nearly solid pedicle are well marked.

ANATOMY OF THE MADREPORARIA. 17

In conclusion, I have to express my thanks to Professor

Milnes Marshall for his assistance; and to the anonymous
donor of the Berkeley Fellowships in the Owens College,
whose generosity has enabled me to carry on my studies.

co™

10.

ll.

List or MEMOIRS QUOTED.

. Bourne, G. C.—“ The Anatomy of Fungia,” ‘ Quart. Journ. Mier. Sci.,’

XXXVIi.

. Fowier, G. H.—* The Anatomy of the Madreporaria,” I. ‘ Quart. Journ.

Micr. Sci.,’ xxv.

. Fowter, G. H.— The Anatomy of the Madreporaria,” II. ‘ Quart. Journ.

Micr. Sci.,’ xxvii.

. von Herper, A.—“* Die Gattung Cladocora,” ‘Sitz. k. Akad. Wiss.

Wien,’ lxxxiv.

. von Heiper, A.—* Korallenstudien,” ‘ Arb. Zool. Inst. Graz.,’ i.
. von Kocu, G.—‘ Bemerkungen iiber das Skelet der Korallen,” ‘ Morph.

Jahrb.,’ v.

. von Kocu, G.—“ Mittheilungen iiber Colenteraten,” ‘ Jen. Zeitschr.,’ xi.
. von Kocn, G.—‘‘ Entwicklung des Kalkskelettes von Astroides calycu-

laris,” ‘ Mittheil. Zool. Sta. Neap.,’ ii.

. von Kocu, G.—‘ Ueber das Verhaltnis von Skelet und Weichtheilen bei

den Madreporen,” ‘ Morph. Jahrb.,’ xii.

Mosrtzey, H. N.—“ Notes on Seriatopora, Pocillopora, &c.,” ‘ Quart.
Journ. Micr. Sci.,’ xxii.

Sctater, W. L.—‘ On a new Madreporarian Coral (Stephanotrochus
Moseleyanus),” ‘ P, Zool. Soc.,’ 1886.

VOL, XXVIII, PART 1.—NEW SER, B

18 G. HERBERT FOWLER.

EXPLANATION OF PLATES I and II.

Illustrating Mr. G. Herbert Fowler’s Paper on “The Anatomy
of the Madreporaria,” III.

Fic. 1.—Transverse section through a small part of the crateriform colony
of Turbinaria sp. (vide p. 2). 4d. Abaxial, inner, or ventral surface of
the colony. Az. Axial, outer, or dorsal surface of the colony. a@. Oblique
section through a polyp cavity. 4. Transverse section of a polyp cavity near
its orifice. c—c. Similar sections further from the orifices. (Camera lucida.)

Fic. 2.—Diagrammatic transverse section of a polyp of Turbinaria sp.
(p. 3). In this and similar diagrams the ectoderm is represented by
“blocked”? black and white, the mesoglea by a dark line, and the endoderm
by a light line, the calcareous skeleton being dotted. The thicker and con-
torted ectoderm at the upper part of the stomatodzum represents the taller
cells, interspersed with plentiful nematocysts, occurring in the neighbourhood
of the tentacles, i. e. mouth disc rather than stomatodeum. Twenty-two pairs
of mesenteries occur in this polyp, of which eleven are on the right and nine
on the left of the “directives.” A bit of the external body wall is drawn to
shew its relations to the echinulations. 4.D. Abaxial directives. Az.D.
Axial directives. S¢. Stomatodeum. cet. Ectoderm. Me. Mesoglea. En.
Endoderm. (Cam. luc.)

Fig. 3.—Transverse section of a mesentery of Turbinariasp., bearing an
ovum with nucleus, nucleolus, and nucleoleoli. The lengthening of the endo-
derm cells round the ovum is noticeable.

Fie. 4.—Transverse section of the calyx of Lophohelia prolifera near
the lip (vide p. 7). The darker parts represent “centres of calcification,”
or the earliest deposited portions, which become enlarged into the regions
marked respectively 7. (thecal), or S. (septal), according to their origin. The
regions are bounded by ‘‘sutures.” et. s. Ectoccelic septa. vt. s. Ento-
ceelic septum.

Fic. 5.—Similar section of Lophohelia prolifera at some distance
from the lip of the corallum. The “ centres of calcification” of the ectoccelic
septa and of the theca have run into one line, owing to the growth of the
latter upwards to the former. Lettering as in Fig. 4.

Fig. 6.—Diagrammatic transverse section through a segment of Lopho-
helia prolifera. ‘The septa are seen to stand in both ectoccelic and ento-
ceelic spaces. The peripheral sections of these spaces and the peripheral
lamellee of the mesenteries, cut off by the upgrowth of the theca, are also
apparent. Lettering as in Fig. 2.

ANATOMY OF THE MADREPORARIA, 19

Fic. 7.—Tissue from the growing edge of a septum of Lophohelia, ob-
tained by a longitudinal section of the polyp. cd. Calicoblasts. me. Meso-
gloea. ez. Endoderm.

Fie. 8.—Tissue surrounding a thecal centre of calcification, obtained by a
transverse section of the polyp (vide p. 9), showing the separation of a
mesentery into central and peripheral parts in process. Cel. Intrathecal
celenteron. Cel’. Extrathecal celenteron. M. The central, and WM’. the peri-
pheral part of the mesentery. Other letters as before. (Cam. luc.)

Fig. 9. Transverse diagrammatic section of a polyp of Seriatopora sub-
ulata. The mesenteries are numbered 1—12, in the same manner as the
Madrepore before described (8). Letters as in Fig. 2. (This diagram is
also good for Pocillopora.)

Fic. 10.—View of a polyp of Seriatopora from above. The clearer spaces
in the body wall of the colony represent the positions of the echinulations on
which the body wall is supported, they having been dissolved away by acid.
Through the body wall are seen the pair of louger mesenteries, 3 and 10.
(Cam. luc.)

Fie. 11. Longitudinal section of a partly expanded tentacle of Seriatopora,
showing the single battery of nematocysts at the tip, interspersed with a few
deeply staining gland-cells.

Fie. 12.—Diagram of a longitudinal section of an introverted tentacle of
Seriatopora. B. The battery of nematocysts, pointed upwards.

Fic. 13.—Diagram of an ideal longitudinal section through a polyp of Seriato-
pora, along the line a—a in Fig. 9. Of the six tentacles, three are expanded
and three are introverted, one of the latter being cut longitudinally. Of the
mesenteries, that on the right of the fig. (10) is one of the two longest; that
on the left (5) is much shorter; while 6 and 7 are rudimentary, and do not
reach as far as the end of the stomatodeum. ‘The cavity is divided into two
halves by fusion of the axial and abaxial into one median septum.

Fic. 14.—Early stage in the development of spermatozoa in Pocillopora.

Fic. 15.—Later stage of the same. The testis is surrounded on all sides
by endoderm, owing to the projection of the capsule outwards from the plane
of the mesentery (vide p. 14).

Fic. 16.—Transverse section through the calyx of Flabellum (vide p. 16).
The numbers ii, ili, iv, indicate the orders to which the septa respectively
belong. Other letters as Fig. 4.

Fic. 17.—Transverse section through part of the pedicle of Flabellum,
showing the conversion of the embryonic theca into a nearly solid pedicle.
i, ii, Primary and secondary septa.

a

J
y
J
5
iy
J
a
a
J
5
J
'
;

Ect. §.

wrt. Vol: AAV N.S. FOL

F Huth, Lith? Edin®

Pt

: ‘ m

~ a
©
J
ts 2
= 5
S S
Ss
SS SS
i, (fook li Le iii: i
m~N
a
ss RY
® - 3
3
[=
=
[Le

ANATOMY OF MUSSA AND EUPHYLLIA. 21

On the Anatomy of Mussa and Euphyllia,
and the Morphology of the Madreporarian
Skeleton.

By

Ge Cc. Bourne, B.A., F.L.S8.,
Assistant to the Linacre Professor in the University of Oxford.

With Plates ILI and IV.

Tue following paper contains a description of the anatomy
of two genera of Madreporaria aporosa, Mussa and Eu-
phyllia, concluding with a general account of the morphology
of the Madreporarian skeleton in the light of the most recent
researches on the group. My thanks are due to Professor
Moseley, who kindly gave me the specimens of Euphyllia with
which I worked, and assisted me with his advice; to Mr. W.
Hatchett Jackson, whom I frequently consulted on the more
general morphological questions contained in the latter part of
this paper; and to Dr. G. H. Fowler, who kindly lent me for
reference proofs of his latest paper on the Madreporaria before
it had appeared in public print.

Mussa (figs. 1—5 and fig. 17).—For the investigation of
this form I had a number of specimens of Mussa corymbosa,
collected by me during my visit to Diego Garcia (S. lat. 7° 13’,
E. long. 72° 23’). These corals grow in large quantities in
the more sheltered parts of the lagoon in shallow water. The
specimens I collected were covered by three to five feet of water
at low spring tides. The colonies are czspitose, the aggregate
of the polypes forming what is apparently a very solid mass,

22 GILBERT ©. BOURNE.

which proves, however, to be exceedingly fragile, since the
polypes are borne on long calcareous stems, which readily
break off at their bases. Mr. A. Dendy, of the British Museum
of Natural History, has kindly identified the species for me,
which is described by Milne-Edwards and Haime (‘ Nat. Hist.
des Coralliaires,’ tom. 1i, p. 333) as follows:

*Corallites sometimes entirely free, sometimes united in
small series of three or four. Spines on the theca widely
separate from one another. Coste well defined in the region
of the calyces only. Columella rudimentary; four cycles of
septa, the principal septa subequal, often decurved towards the
inner margins, which are scarcely at all dentate; above they
bear three stout, diverging spines. The smaller septa have
tolerably regular, short, and pointed teeth. The polypes are,
according to Ehrenberg, of a pale brown, with a golden-yellow
disc. The margin is covered with bursiform papille which
surround a small number of short digitate tentacles.”

To this I have to add that the number of septa is very
inconstant, and bears no relation to a multiple of six. In one
calyx I counted thirty-two septa, in another forty-six, in a
third (fig. 3) forty. In the last case twenty-four septa are of
conspicuously larger size than the remainder, but do not show
any characters which enable them to be classed as primaries,
secondaries, &c. Their shape is accurately shown in section in
fig. 8; the broader peripheral part can be distinguished from
the more slender central portion, and the processes connecting
the peripheral ends are easily seen. They are extremely
exsert, standing as much as 9 mm. above the lip of the calyx.
In the majority of cases one or sometimes two smaller septa
are found between each pair of larger septa; these are thinner
and smaller, and of approximately the same breadth through-
out their course. Finding it impossible to separate the septa
into regular systems, I shall refer to the first as the principal,
the second as the secondary septa. The theca is formed by
fusion of the peripheral ends of the septa (see fig. 2), and is
only consolidated at some distance below the edge of the calyx.
The upper part of the corallite is smooth and polished, corre-

ANATOMY OF MUSSA AND EUPHYLLIA. 23

sponding to the region that is covered externally by soft
tissues. Below this the corallum is invested by a distinct coat
of thin granular epitheca. The septa are united within the
calyx by dissepiments, thin obliquely placed lamelle of cal-
careous tissue placed at different heights in the interseptal
loculi, and meeting towards the centre of the calyx (see
fig. 4 d).

The colour of my specimens and the tentacles differ from
Ehrenberg’s account. The Mussa of Diego Garcia is of a dull
brown colour, with olive-green disc and tentacles, the latter
being numerous and moderately long. The polyps contract
energetically on being handled, and I was unable to preserve
any specimens with the tentacles expanded; indeed, all my
specimens are so much contracted that the tentacles are no
longer recognisable even in section. A view of a system of
three polyps borne on a common stem is given in fig. 1; the
drawing is taken from a specimen contracted in spirit, aud is
half the natural size.

Fig. 2 is a transverse section of a corallum of Mussa taken
just below the lip of the calyx. In each septum may be dis-
tinguished dark centres of calcification, around which are seen
a number of concentric, dark, and light lines of growth, mark-
ing successive additions of calcareous tissue. Evidence may
also be seen of centrifugal growth of each septum, due to the
deposition of calcareous matter on its peripheral extremity.
The septa are seen to be joined together by wing-like out-
growths, which fuse to form a theca. At a lower level the
theca would be seen to be consolidated by a continuation of
the peripheral lamelle of the septa over the intervening
bridges of tissue, which thus externally cement the conjoined
septa into a continuous whole.

Anatomy of the Polyp.—A glance at fig. 1 will show that
the soft tissues of the polyp extend downwards for a consider-
able distance on the outside of the corallum. There is in fact
a well developed ‘ Randplatte,” as it is called by German
authors, and it contains extrathecal continuations of the
exoceeles and entoceles separated from one another by peri-

24 GILBERT ©. BOURNE.

pheral continuations of the mesenteries, as has been described
for other forms by von Koch, Fowler, and others. Fig. 4
shows that the “‘ Randplatte ” extends downwards considerably
farther than do the soft tissues within the calyx. There can
be no question that the “ Randplatte” is a distinct structure
in these forms.

As decalcification proceeds it becomes obvious that the
corallum lies wholly external to the polyp, being, as it were,
dovetailed into it from below; or, to use another illustration,
the polyp appears to be drawn over the corallum much as a
glove is drawn over the hand. After decalcification, the intra-
calicular soft tissues appear to be divided up into wedges by
the spaces previously occupied by the septa.

Examination of the internal structure shows that it is of the
normal Actinian type except in one important point, there are
no directive mesenteries, the longitudinal muscles of each pair
of mesenteries being placed vis 4 vis. This character, which
has already been discovered by Fowler in Lophohelia, is also
characteristic of Euphyllia, as will be seen later on. The
arrangement of the mesenteries and their relations to the septa
are shown in fig. 5. Each pair of mesenteries embraces a
septum, all the septa are entoccelic, no septa occurring between
pairs of mesenteries. The mesenteries embracing the prin-
cipal septa are all inserted on the stomodzum, which is of
moderate length; the mesenteries embracing the secondary
septa are free throughout their extent. All the mesenteries
bear well-developed filaments on their free edges, and below
the stomodzum their edges are drawn out in long, sinuous,
ribbon-shaped prolongations, around the whole edge of which
the filament is continued, the whole structure being coiled up
to one side of the mesentery in an exoccelic or entoccelic space.
To such coiled filaments I erroneously gave the name of
Acontia in my paper on Fungia, but I have since satisfied
myself that they differ essentially from Acontia as defined by
Gosse and the Hertwigs. Since the septa are all entocclic
the number of pairs of mesenteries is the same as the number
of septa, and therefore not necessarily a multiple of six,

ANATOMY OF MUSSA AND EUPHYLLIA. 25

The tentacles were so completely retracted in my specimens
that I could not determine anything about them. In some
longitudinal sections there are infoldings on the surface of the
peristome, such as are represented in fig. 4, but I could not
say with any certainty whether these are tentacles or not.
The ectoderm lining the involutions does not differ from that
of the rest of the body surface.

Histology.—This does not differ from the normal Actinian
type. The ectoderm is of the ordinary character, and contains
numerous nematocysts ‘002 mm. in length when inverted.

The calicoblasts in most situations agree with the form
described in Fungia, and the majority of cases, namely, rounded
or polygonal, soft-looking granular cells, which do not stain
easily, possessing nuclei which stain but slightly in borax
carmine. They occur everywhere, in a scattered condition or
forming a distinct layer, between the mesoglcea and the corallum.
The calicoblasts which clothe the uppermost and peripheral parts
of the septa are of a different character, being drawn out into
very long, narrow, columnar cells, just like those described by
Fowler for Lophohelia; and it will be observed that they cor-
respond very closely in position to the similar cells in that form.
In both cases they are found at the seat of the greatest activity
of coral secretion. In several of my preparations I found
pyramidal or oval cells exhibiting a longitudinal or radial
striation, which exactly resemble those drawn by Sclater for
Stephanotrochus and von Heider for Astroides, and described
by them as calicoblasts. I have no hesitation in saying that in
Mussa they are not calicoblasts, since they differ entirely from
the cells above described, which from their position are undoubt-
edly calicoblasts, and they are never found in the regions of
coral secretion. On the contrary, in my specimens they are
always and only associated with the mesoglea of the mesen-
teries. Fowler at the close of his account of Pocillopora
describes similar structures, and considers that they are rather
connected with the attachment of the mesentery to the corallum
than with the secretion of coral. My observations fully con-
firm his views. It is important to bear in mind that these

26 GILBERT C. BOURNE.

structures are not calicoblasts, since von Heider, from his
observations on them in Astroides, has recently inferred that
the calcareous tissue is deposited in the form of crystals within
the calicoblasts, just as the spicules are formed within the sub-
stance of the skeletogenous cells in Alcyonaria and calcareous
sponges, and that the radial striation seen in these pretended
calicoblasts is due to the presence of minute crystal of carbo-
nate of lime. Now we know that animal tissues on the verge
of calcification are extremely resistant to the action of acids,
but in this case the crystalline form being assumed, that stage
would have been passed, and the crystals, if they were such,
would have been readily soluble in acids. Yet the acids which
dissolved the whole coralla of the specimens which von Heider
examined did not suffice to dissolve the minute crystals also.
It might have been inferred, one would have thought, that
whatever the striation was due to, it was not due to the pre-
sence of crystals of calcium carbonate. There is then nothing
to disprove von Koch’s statement that the calcareous tissue is
elaborated by and secreted by the ectoderm (i.e. calicoblasts),
and that the Madreporarian skeleton differs wholly in this
respect from that of Alcyonaria.

The mesogloea in Mussa is perfectly structureless; I failed
to detect even a fibrillar arrangement in it, except were it is
drawn out into pleats for the attachment of the longitudinal
mesenterial muscles. In the upper part of the polyp, that part
which occupies the spaces between the enormously exsert septa,
the mesoglea is enormously developed, giving a considerable
amount of stiffness to the tissues in this region (fig. 4).

The endoderm is of the normal character, and is packed, as
is usually the case in Actiniaria, with zooxanthelle. The cells
covering the mesenterial filaments are long and columnar, and
contain many nematocysts of different kinds; the one kind
small, similar to those found in the ectoderm, the other kind
large, measuring ‘005 mm. in length when the thread is not
ejected.

Ova were present in my specimens in the normal position,
towards the lower extremity of the mesenteries, embedded in

ANATOMY OF MUSSA AND EUPHYLLIA. 27

the mesoglea. They are of large size, measuring as much as
‘O11 mm. across. They are only to be found on a few, and
those invariably the shorter mesenteries; but they were not
constant on these, and I am unable to say whether the shorter
mesenteries alone are reproductive. As I found no traces of
spermatozoa in the specimens which I examined it may be
concluded that Mussa is moneecious.

Euphyllia.—The specimens of this genus were kindly
supplied to me by Prof. Moseley, and form a part of the reef
corals collected by H.M.S. “Challenger.” The corallum is
described by Quelch in the ‘Challenger Reports,’ vol. xvi, p.
74, as Euphyllia glabrescens. The description of the
species by Milne-Edwards and Haime is as follows :—“ Corallites
sometimes unite in small series of three or four, but ordinarily
separating early. Theca covered with closely set and extremely
fine grains. Cost thin, rising slightly near the edges of the
calicle, and subcristiform. Calices with irregular boundaries,
narrow, and very deep fossa. Septa very irregularly arranged
in orders, scarcely exsert, very thin, moderately close together,
their faces very finely granular, and presenting parallel striz.
Greatest width of the calices 2 centimetres; the diameter of
the corallites is somewhat smaller beneath the calices.”

Quelch arranges the septa in five orders, but I found their
number very inconstant, and could not make out more than
three definite cycles, which are, however, arranged with con-
siderable regularity. The septa of the first order are large,
and reach in the deeper parts of their courses nearly to the
centre of the calyx. Those of the second order are rather
shorter, and alternate with those of the first, whilst the septa of
the third order are very short and are only to be found in the
upper part of the calyx; they occur in every loculus between a
septum of the first and a septum of the second order. A
transverse section across part of the calyx, including septa of
all three orders, is given in fig. 7. The calcareous tissues are
seen to be much thinner and more fragile than in Mussa, and
the concentric lines of growth seen in the latter form are not
present. The centres of calcification are present as con-

28 GILBERT C. BOURNE.

spicuous dark lines running down the centre of each septum.
The primary and secondary septa are but slightly thickened
towards the peripheral ends, the theca being mainly composed
of the heads of the tertiary septa. Fig. 7 shows that in the
upper part of the calyx the tertiary septa project from a stouter
thecal piece, the two together forming a T, of which the thecal
portion is the cross-piece. There are no sutures separating _
the septal from the thecal portion. Lower down in the calyx
the tertiary septa die out altogether, but the cross-pieces,
representing their thecal portions, remain, and then the
section has precisely the appearance figured by Fowler in
Lophohelia. From the relations which obtain in Euphyllia I
am not disposed to think that the intercalated pieces figured
by him are essentially thecal structures sharply distinguishable
from septa,

The structure of the dissepiments is quite similar to that of
Mussa.

The specimens with which I worked were killed with the
tentacles fully expanded, but they were unfortunately too
much damaged to admit of a careful study. The tentacles are
numerous, short, and apparently correspond in number to
the septa, i.e. they are all entoccelic, as is the case with the
latter.

A well-developed “ Randplatte” is present as in Mussa,
extending down the outside of the corallum for a distance of
15 centimetres. The structure and relations of the “ Rand-
platte ”’ are the same as those in Mussa, and do not require a
detailed description, the only difference of importance is that
whereas the longitudinal muscles are generally absent on the
extrathecal portions of the mesenteries, they are present,
although in a rudimentary form, in Euphyllia.

The appearance of the polyp on decalcification is quite similar
to that of Mussa, and affords ocular demonstration of the fact
that the corallum is wholly external to the polyp. The
internal anatomy of the polyp is of the highest interest. I
have not had the time to make so thorough an examination of
ts peculiarities as I should have wished, but hope to give a

ANATOMY OF MUSSA AND EUPHYLLIA. 29

fuller account of them in a future communication. At present
I will confine myself to a description of what I have seen.
There are no directive mesenteries, Euphyllia agreeing in
this respect with Lophohelia and Mussa. Their absence is a
striking and important fact, adding as it does another type of
mesenterial arrangement to those we know already. We have
the normal Actinian arrangement, giving a bilateral symmetry
along the long axis of the stomodeum, the Edwardsian and
Alcyonarian arrangements in which the bilateral symmetry is
marked equally well, but in a different manner. The arrange-
ment in Zoanthus is clearly nothing more than a modification
of the Actinian type, and the arrangement in Cerianthus is
probably connected with an atrophy of the longitudinal muscles
in the same type. All these forms show a bilateral symmetry ;
Mussa, Lophohelia, and Euphyllia alone are perfectly radial.
This may either be a primitive condition or may be connected
with fissiparity, for it is impossible to conceive how two
polypes can be derived by fissiparity from one with directives,
and yet the arrangement of directives be carried over into the
daughter polyps, especially when an examination of an Astreid
colony reveals the fact that the calices are constricted in their
centres at right angles to their long axes, and the long axes of
resulting calices may either be continuations of the long axis
of the original calyx, from which they were derived, or may be
set at right angles or any other angle to it. To state shortly
the extraordinary features in the remainder of the anatomy,
the stomodzeum is very long, reaches nearly to the bottom of
the polyp, and is converted into a ramifying and inosculating
system of canals; it functions as the chief digestive cavity of
the polyp. The endoderm is greatly vacuolated, and converted
into a reticulated tissue filling up the cclenteron, in the
meshes of which are numerous nematocysts and symbiotic
alge. Mesenterial filaments are feebly developed on the
primary and secondary mesenteries, which are attached
throughout the greater part of their course to the stomo-
dzeum, and do not reach a great development on the tertiary
mesenteries, which are free for the greater part of their course,

30 GILBERT C. BOURNE.

These peculiar modifications are shown in fig. 8. The histo-
logical details have been greatly simplified, but the outlines
are correctly drawn with a camera lucida. The reticular endo-
derm en’, is seen filling up all the spaces corresponding to
the celenteron. The stomodzal canals are shown at st. c.
These canals are lined throughout by an epithelinm exactly
resembling that of the ectoderm on the free surfaces of the
body, but staining rather more deeply in borax carmine.
Great numbers of nematocysts are embedded in this epi-
thelium, and may be seen in all stages of development. Fig.
9 is an unripe nematocyst, and corresponds with the stages
drawn by Mobius in his account of the development of these
structures. (Mobius, ‘ Ueber den Bau, den Mechanismus und
die Entwickelung der Nesselkapseln,’ Hamburg, 1866, Taf. 11,
figs. 22 and 23). The ripe nematocyst, with its axial body
ejected, is shown in fig. 10, and in fig. 11 it is drawn with the
thread ejected, but not the axial body. The thread bears a
peculiar armature at its extremity. The stomodzeum below
the mouth is a simple tube, but at a short distance lower it is
produced into a number of horns running out towards the
attachments of the mesenteries; at a lower level these horns
are seen to have anastomosed with one another in such a
manner that the stomodzum is converted into a highly com-
plicated system of canals occupying the centre of the calyx.
This is shown in section in fig. 8. At a lower level the stomo-
deum again becomes more simple, and finally ends in a simple,
much compressed tube with a narrow lumen. I was unable to
determine to my satisfaction whether the stomodzum opens
below into the axial cavity, or whether it is completely closed.
I am inclined to think that it opens by a very narrow passage.
As the stomodzum reaches nearly to the bottom of the polyp
cavity, the axial cavity below it is very small, and such as it is,
it is entirely filled up with mesenterial filaments. The stomo-
dzeal canals contained numerous fragments of vegetable matter,
apparently pieces of leaves. The presence of vegetable food in
these canals is interesting, firstly, because I believe it is the
only recorded instance of a coral feeding on a vegetable diet ;

ANATOMY OF MUSSA AND EUPHYLLIA. 31

secondly, because it shows that this enormously and peculiarly
developed stomodzum is digestive in function, which might,
indeed, have been inferred from its extent, and from the almost
complete absence of a ceelenteron as a cavity,

The mesogloea is a thin but perfectly distinct, structureless
lamina, everywhere separating the ectoderm from the endo-
derm. The endoderm cells lining the extrathecal parts of the
celenteron are vacuolated, and do not fill up the cavities, but
in the intrathecal parts of the polyp the endoderm cells are
entirely converted into a parenchymatous tissue, filling up all
such parts of the celenteron as are not occupied by mesenterial
filaments. My specimens were not sufficiently well preserved
to admit of my giving an account of the histology of this
tissue. It is filled with nematocysts, one of which is repre-
sented in fig. 12, and may be seen to be of an entirely different
character to those found in the ectoderm, being of smaller size
and different shape. The axial tube is distinguishable, and the
thread coiled in an oblique spiral within the capsule. I could
not find any of these nematocysts with the thread ejected.
Numerous zooxanthelle also exist within the meshes of this
tissue.

The muscles on the mesenteries are well developed, and
exhibit the relations of the exocceles and entoceles. The septa
are all entoccelic.

Ova are borne on the tertiary mesenteries and on them only.
The ovaries are of large size and bulge out the mesenteries to
such an extent that they nearly fill up the exocceles and ento-
ceeles in that region. They are always developed towards the
peripheral ends of the mesenteries. Each ovum is surrounded
by a number of large granular masses the granules of which
stain very deeply in borax carmine. At first I mistook these
for testes, but from their relation to the ova and their
resemblance to nutritive cells in Pennaria Cavolinii and
Tubularia mesembryanthemum, I have no doubt that
they are nutritive cells destined to be absorbed by the ovum
(vide fig. 8, ov.). That one out of several primitive ova should
develop into a mature ovum at the expense of the others has

32 GILBERT 0. BOURNE.

long been known in Hydroids, but has not previously been
described for the Anthozoa. The cells of the mesenterial
filaments are vacuolated, but their boundaries are clearly
recognisable. ;

The structure of Euphyllia as above described is without
parallel in the Madreporaria, or indeed in the Anthozoa, the
ramified digestive tract and parenchymatous tissue suggest a
comparison with Dendrocele planarians; but the com-
parison will not hold good for a moment when we consider that
the alimentary tract of the latter is formed from hypoblast, whilst
that of Euphyllia, formed by the stomodeum, is ectodermic.

Any attempt to explain such isolated phenomena as these
must be peculiarly liable to error, but possibly the following
explanation may throw some light on the peculiarities. The
endoderm of Euphyllia as of most Madreporaria is filled with
symbiotic alge. These contribute an important share towards
the nutrition of the polyp, and it is conceivable that the
symbiotic nutrition, if one may express it so, became of such
primary importance to the economy of the animal that the
endoderm lost its original digestive function and became vacuolar
in order to accommodate greater numbers of zooxanthelle.
A comparison between the vacuolated endoderm of Euphyllia
and the extracapsular protoplasm of the Radiolaria, will explain
my meaning. In both cases the vacuolation is probably con-
nected with the presence of zooxanthelle. The ectoderm is
known to retain the power of amceboid digestion in many
Celenterata, including Actinia mesembryanthemum and
Bunodes sabelloides (Metschnikoff, ‘ Researches on the
Intracellular Digestion of Invertebrates,’ this Journal, new
ser., xxiv, p. 93). There is nothing surprising then that at the
same time that the endoderm became modified in connection
with symbiosis, that the ectoderm of the stomodum should
have taken a more active part in alimentation, and that its
surface should be greatly increased to meet the remaining
necessities of the organism.

It has long been felt that a classification of Madreporarian

ANATOMY OF MUSSA AND EUPHYLLIA. 33

corals based on a study of the corallum alone is unsatisfactory,
and that any attempt to remodel the old classifications should
depend on a systematic study of the relations between the
corallum and the polyp. Owing to the difficulty of obtaining
material, and of dealing with it when obtained, the number
of forms examined is as yet small, and the results of recent
researches have not advanced us very far towards an improved
classification. The most recent attempt to remodel the
systematic treatment of the group is Professor Martin Duncan’s
“ Revision of the Families and Genera of the Madreporaria,”
‘Linn. Soc. Journ.,’ xviii, in which the older classifications
are amended in some important particulars, several old families
have been struck out or merged with other families, and the
Fungide are raised te the rank of a group equal in value to
the Perforata and Aporosa. But whilst all the definitions, by
far the greater part of the classification, depend on the old
distinctions in the characters of the corallum, scarcely any
weight is given to the development and anatomy of the polyp.
Although but little has been done even in working out the
anatomy of adult forms, and although our knowledge of the
development of the Madreporaria is miserably insufficient, we
have sufficient information about the group to enable us to
make certain generalizations about it.

The pricipal workers on Madreporaria have been de Lacaze
Duthiers, Moseley, G. von Koch, von Heider, and Fowler,
whose separate memoirs are referred to in the course of this
paper. In all, the anatomy of some twenty forms has been
worked out more or less completely, and the development of
one species, Astroides calycularis, has been followed by
two observers, H. de Lacaze Duthiers (‘ Arch. de Zool. exper.
et. gen.,’ ii, 1873, p. 269), and G. von Koch (‘ Mitth. der Zool.
Stat.,’ Neapel, 1882).

Anatomy of the Polyp.—In the majority of the forms
examined the structure of the polyp, both in grosser anatomy
and in histology, is essentially that of an Actinia. The mesen-
teries are arranged in pairs, and frequently if not usually in
cycles of six pairs each. But recent observations have shown

VOL, XXVIII, PART 1,—NEW SER. c

34. GILBERT 0. BOURNE.

that the number of mesenteries, and with them of the septa, is
inconstant in several genera, and that a hexameral arrange-
ment of the parts is by no means an invariable rule, so that
the name Hexactinia as applied to the group is extremely mis-
leading. In most cases the mesenteries are marked out into
pairs by the situation of their longitudinal muscles, which are
so arranged that in all but two pairs these muscles are
attached to that side only of a mesentery which is turned
towards its fellow. At the two ends of the long axis of the
stomodzum, however, this arrangement is reversed, and these
two pairs of “directive’’ mesenteries have the muscles on their
opposite faces. This arrangement, which is typical of the
Actinie, is found in nearly all the Madreporaria hitherto
examined ; but the latest researches of Fowler and myself show
that directive mesenteries are absent in Lophohelia among the
Oculinide, and in Mussa and Euphyllia among the Astreide.
Excepting for this peculiarity Lophohelia and Mussa do not
differ from the Actinian type; but Euphyllia is modified in an
extraordinary manner, as has been described in the first part of
this paper, and need not be recapitulated here. The other
forms in which there is any important deviation from the
normal type are Pocillopora, Seriatopora, and Madrepora
Durvillei.

The anatomy of Seriatopora and Pocillopora was first inves-
tigated by Professor Moseley (this Journal, new series, xxii,
p. 391), and his results have since been confirmed by Fowler.
In these forms the main deviation from the normal Actinian
type consists in the fact that of the twelve mesenteries only
two bear mesenterial filaments, and these two do not belong to
the same pair, but are the dorsal mesenteries of the right and
left infero-lateral pairs respectively. In correlation with the
development of the filaments on these two mesenteries the
intermesenterial chambers or entocceles of the pairs to which
they belong project far deeper into the calyx than the remain-
ing chambers in Seriatopora; but this feature is not so well
marked in Pocillopora. If the mesenteries are numbered
from left to right, beginning with the left ventral or abaxial

ANATOMY OF MUSSA AND EUPHYLLIA. 35

mesentery, Nos. 3 and 10 are those with filaments and are
longer than any others; Nos. 1, 5, 8, and 12 are recognisable
for some distance below the stomodzum, Nos. 2, 4, 6, 7, 9, 11
are very short and rudimentary. In Pocillopora the difference
in size between 3 and 10 and the other longer mesenteries, 1,
5, 8, 12, is not so marked as in Seriatopora, and the last
named sometimes have rudimentary filaments. In both forms
the polyps show a well-marked bilateral symmetry with regard
to the dorsoventral axis, and in both there is a system of
superficial radiating canals by which the cavities of adjacent
polypes are put into communication with one another.

In Madrepora Durvillei, according to Fowler (this
Journal, new ser., xxvii, pl. i), there is a well-marked dimorphism
in the polyps composing a colony. In the one form of polyp
there are twelve simple mesenteries, of which six have a long
course and a better developed filament than the remainder, and
of these two are longer than the others. In the second form
of polyp are also twelve mesenteries, which in the higher parts
of the polyp are perfectly normal, but at the lower end of the
stomodzum six of them become modified, in all cases the same
six, namely, the 2nd, 4th, 6th, 7th, 9th, and 11th, using the same
method of counting as in Seriatopora. The modification consists
in the greatly increased size and vacuolation of the endoderm
cells of the mesentery, and in the presence of a canal lined by
endoderm, which runs right through its centre and is bent
sharply upon itself. For a description of the course of this
canal and its relations to the polyp cavity, Fowler’s memoir
should be consulted. Only the modified mesenteries bear
filaments, and of them two, viz. four and nine, are longer than
the others. It is important to observe that the modified
mesenteries of the second type of polyp correspond with the
longer mesenteries of the first type, and that the longest
mesenteries of all occupy the same positions in both cases.

In both Seriatopora and Madrepora Durvillei the two
longest mesenteries alone bear gonads.

The differentiation of certain mesenteries in Madrepora and
Seriatopora, and the specialisation in both cases of two mesen-

86 GILBERT C. BOURNE.

teries on opposite sides of the polyp for reproductive purposes,
suggests a close affinity between the two forms; but there is
this difference between them, that whereas the reproductive
mesenteries in Madrepora Durvillei are 4 and 9, those in
Seriatopora are 3 and 10, whilst the short mesenteries of the
former correspond to the long mesenteries of the latter, and
vice versd. There are, however, other similarities in structure
which ally the Pocilloporid with the Madreporine, although
they are usually classed far apart, as Aporosa and Perforata
respectively. In Madrepora aspera and M. variabilis
the differentiation of the mesenteries does not appear to have
advanced so far as in M. Durvillei. In the first named,
according to Fowler, filaments are present on the abaxial
directive mesenteries as well as on the same six as in M. Dur-
villei, the remaining mesenteries being devoid of filaments.
Von Koch does not give any account of these structures in
M. variabilis.

The Corallum.—In dealing with the hard tissues of
Madreporarian corals I shall try to point out that there are
only three (or possibly four) distinct structures in the corallum,
viz. the septa, the theca, and the epitheca (and perhaps the
basal plate). The columnella, the pali, the costae, the dissepi-
ments, synapticula, exotheca, and peritheca may all be traced
to modifications of the first named, and useful as they may be
for the determination of genera and species, they have no
morphological value. Associated with these names is the
view that some of these structures differ fundamentally from
the others; some were considered by Milne-Edwards and
Haime to be of dermic origin, others to be of epithelial origin,
whereas, as we shall see in the sequel, all the hard parts are
of essentially similar origin.

The first point of interest connected with the corallum is to
determine from which layer of the polyp it is derived, from the
ectoderm, endoderm, or as a calcification of the mesoglea.

Milne-Edwards and Haime, in their classical work on corals,
stated that the calcareous tissue was deposited in a layer
which they called the dermis, which they defined as one of the

————

ANATOMY OF MUSSA AND EUPHYLLIA. 37

deeper layers of the ectoderm. This view, which for a long
time held ground, was disputed by de Lacaze Duthiers, who
states in his paper on the development of Astroides caly-

- cularis, that the calcareous tissue first makes its appearance

in the endoderm! (‘ Arch. de Zool. Expér. et Gen.,’ 11, 18738,
p- 269). But since he did not recognise the existence of a
third structureless layer—the mesogloea—between the ecto-
derm and endoderm, and since his observations were not made
by means of sections, de Lacaze Duthier’s statements on this
head are somewhat perplexing. It seems clear, however, that
he considered the calcareous tissue to be deposited in the
situation which further researches showed to be occupied by
the mesoglcea, and it was from this layer that the corallum
was thought to be formed by succeeding writers on the subject
for some years.

The more exact relations of the corallum to the polyp have
been chiefly worked out by G. von Koch, to whose works
complete references will be found in my own and Fowler’s
papers on corals in this Journal.

From his observations on the development of Astroides
calycularis, von Koch established the fact that the corallum
is a product of the ectoderm, secreted by it, and makes its first
appearance between the basal ectoderm and the surface to
which the young Astroides is attached. His method of obser-
vation is at once ingenious and convincing. Pieces of cork
were floated in the aquarium, to which free swimming larve
soon attached themselves. As soon as it was apparent that
a deposit of calcareous matter was being formed the embryos
were killed in situ, and sections were cut through them and

1 Fowler is in error in attributing to de Lacaze Duthiers the statement
that the corallum is formed externally to the polyp. It is true that the figure
to which he refers (fig. 27 of the memoir above quoted) fully bears the inter-
pretation which he has given to it, but de Lacaze Duthiers’ own words show
that he took a very different view of the matter: “ C’est au milieu, et dans
Pépaisseur de cette couche interne, toujours plus épaisse dans le milieu et au
bas des loges que se montrent les premiers nodules calcaires,” and elsewhere

he expressly defines what he means by the “couche interne;” it is the en-
doderm.

38 GILBERT ©. BOURNE.

the cork to which they were attached. A diagrammatic
drawing of a section made in this way, tangential to the
surface of the polyp, is given in fig. 13; it is slightly modified
from von Koch’s original drawing for the sake of greater
clearness.

Previous to von Koch’s researches on Astroides, von Heider
(‘Sitz. der Kais. Akad. in Wien.,’ lxxxiv, 1881) had shown that
in Cladocora there exists everywhere between the structureless
mesogloea and the corallum a layer of rounded cells, which
apparently have the function of secreting the coral substance ;
to these cells von Heider gave the appropriate name of calico-
blasts, and considered them to be a product of the ‘ meso-
derm” (mesogloea), the layer from which he supposed the
corallum to be derived. It is clear, however, from von Koch’s
researches, that the layer of calicoblasts is nothing more than
the persistent ectoderm which secretes the corallum, and lies
everywhere between it and the mesoglea. Unfortunately we
are not acquainted with the development of any Madreporarian
other than Astroides, but since calicoblasts have been found in
all corals recently examined, we may assert with certainty that
the corallum is a product of the ectoderm and is
always external to the polyp.

In describing the formation of the theca, the septa, and the
coste, the statements of de Lacaze Duthiers and von Koch
differ widely from one another. The former states distinctly
that the theca appears independently of the septa, the latter
arising as twelve radiating rods, bifurcated at their peripheral
extremities. As far as the number and shape of the rudiments
of the septa are concerned, von Koch’s statements are in ac-
cordance with those of de Lacaze Duthiers, but he gives a very
different account of the formation of the theca. The skeleton,
he says, first appears as a ring-shaped basal plate, incomplete
in its central portion, this plate always making its appearance
between the basal ectoderm and the surface to which the larva
is attached. It is composed of an aggregation of small spheri-
cal nodules, each of which is made up of rhomboid calcareous
erystals grouped in concentric layers. As growth proceeds the

ANATOMY OF MUSSA AND EUPHYLLIA. 39

basal plate becomes consolidated, and its central vacuity filled
up. No doubt it was the earliest formed ring-shaped part of
the basal plate which de Lacaze Duthiers took for the theca.
The first signs of septa are twelve radially disposed ridges of
the basal endoderm, which soon project upwards into the
cavity of the polyp as a series of folds, the mesoglcea and
ectoderm being included in the folds. Between the limbs of
the folds of the ectoderm are deposited calcareous nodules,
similar to, and continuous with, those of the basal plate (see
fig. 18). The septa increase in height and thickness as growth
proceeds, but always remain covered over by a triple layer of
ectoderm, mesogloea, and endoderm, the ectoderm forming the
ealicoblast layer of the adult polyp. The peripheral extremi-
ties of the septa become forked, and, according to von Koch,
the forked extremities of adjacent septa unite with
one another to form the porous theca.

These statements of von Koch accord very well with certain
peculiarities of structure previously observed by him in Caryo-
phyllia, Galaxea lampeyrana, and Mussa, corals belong-
ing tothe Madreporaria aporosa. He found that in these
the theca lies apparently within the body of the polyp, free
from the lateral external body wall, and separated from the soft
tissues outside by a space which is a part of the ccelenteron.
The theca is formed by the thickening and fusion of the peri-
pheral ends of the septa and is not a separate structure; where
coste are present they are nothing more than continuations of
the septa external to the theca. The septa, theca, and costz
are everywhere covered by the triple layer of ectoderm (calico-
blasts), mesoglcea, and endoderm described above. From these
relations it will be understood that a portion of the ccelenteron
lies external to the theca, and this portion may conveniently
be described as extrathecal cclenteron. It is divided into
chambers by mesenteries corresponding to those which divide
the intrathecal ccelenteron into radical chambers, the originally
continuous mesenteries having been, according to von Koch,
cut in two by the fusion of the peripheral ends of the septa,
and thus divided into extrathecal and intrathecal portions. The

40 GILBERT 0. BOURNE.

septa usually project higher into the body cavity than does their
product the theca, and the extrathecal mesenteric chambers,
(exocceles and entecceles of Fowler) are continuous with the
corresponding intrathecal chambers over the lip of the calyx.
I have attempted to show these complicated relations in the
diagram fig. 14, Fig. 15 exhibits diagrammatically von Koch’s
view of the formation of the corallum (without epitheca).

It is worthy of remark that, according to von Koch’s obser-
vations, the basal plate in corals has a different developmental
history from the theca, and is morphologically distinct from it.
As far as I am aware, no one has yet called attention to this
fact. The basal plate, however, is, according to the same author,
continuous, if not identical with the structure known as
epitheca.

The epitheca is formed in Astroides as a secretion of the
ectoderm of the body wall at a spot where the lateral walls of
the polyp pass into the basal portion ; it is connected with the
basal plates and form a thin and tolerably smooth lamella in-
vesting the lower parts of the polyp (vide fig. 13, ep.).

Unfortunately, the further development of the epitheca has
not been studied, and we are even deficient of an exact know-
ledge of its structure in the adult. In his most recent contri-
bution to the subject (‘ Morph. Jahrb.,’ xi, 1886, p. 154) von
Koch has given a diagram which professes to show clearly all
the relations between soft and hard tissues in the adult coral
polyp. In it the epitheca is figured as a complete and inde-
pendent wall of calcareous tissue lying parallel with the theca,
and separated from it by a considerable space of extrathecal
ceelenteron, this space being bridged over at intervals by the
cost, which in the drawing abut upon and are fused with the
epitheca so as to connect it with the remainder of the corallum.
Such relations as are shown in this figure occur, as far as I am
aware, in no coral either living or extinct; they can only be
considered as theoretical, and form a part of the system which
von Koch is attempting to construct on the subject of the
coral skeleton. I have never seen or heard of a coral in which
the soft tissues outside the upper part of the theca are them-

ANATOMY OF MUSSA AND EUPHYLLIA. Al

selves invested by a calcareous lamina. In all the forms with
a persistent epitheca which I have examined it is invariably in
the form of a more or less well-defined layer of calcareous
tissue, investing the basal parts of the corallum, and nowhere
extending above the level of the soft parts covering the exte-
rior of the latter. Its relations and general appearance suggest
its having been formed from the free edge of the soft tissues on
the exterior of the corallum, as they retreat farther and farther
from the original surface of attachment.

The names exotheca, peritheca, coenenchyme, epitheca, are
all applied to laminar, ring-shaped, or encrusting calcareous
investments of the theca, and the distinctions drawn between
them are so subtle or so vaguely expressed that I am quite
unable to distinguish the difference between them in ordinary
cases. In point of fact no essential difference exists. If these
structures are deposited by the same parts of the polyp in each
case they are morphologically similar to one another ; variations
of form count for nothing. To determine the mode of forma-
tion of corallum, when embryological data are not to hand, a
study of the microscopical characters of the corallum by means
of sections, and a consideration of the relation of the soft parts
to the corallum is necessary. In the first place, it must be
kept in mind that throughout the region of the living polyp
the corallum is invested by an active secreting layer of calico-
blasts. Fowler’s figures of Lophohelia, and mine of Mussa,
Euphyllia, and Astrea, show that there are present in the
septa dark lines or centres of calcification marking the central
point from which calcification has taken place. In Mussa it
can be easily seen that concentric layers have been formed
around the centre in each septum (fig. 2), and that as two
contiguous septa became joined together the area of active
secretion was confined to the external or peripheral part of the
septum, which therefore increased in length centrifugally. A
section taken somewhat lower down than in fig. 2 shows that
calcareous tissue is also added over the bridges connecting the
septa until they are connected together by a very solid theca.
Compare with this section the diagrammatic figure of a coral

42 GILBERT ©. BOURNE.

(fig. 14) and it will be easily seen that this peripheral thickening
of the septa and their fusion to form a theca is due to the
activity of the calicoblasts of the inner wall of the extrathecal
soft tissues. A reference to Fowler’s paper in this number
(Pl. I, fig. 5) shows that in Lophohelia the sutures become
indistinct in the lower sections, owing to the addition of this
peripheral ring of calcareous tissue from the inner wall of the
“ Randplatte,” as von Heider calls the extrathecal tissues.
The same thing is noticeable, but to a much more marked
degree, in Oculina, and in it forms the tissue in which the
calyces are embedded, i.e. the ceenenchyme. An examination
of fig. 6 of this paper shows that in Euphyllia (and it is equally
true for Mussa) that the “ Randplatten”’ of adjacent polyps,
where the latter have not separated widely from one another,
are continuous, and form a covering for the valley separating
the two calices. When in a compound coral all the polyps are
thus connected the connection is known as the ccenosare, and
the coenenchyme is obviously the product of the calicoblasts of
the lower layer of the ccenosarc. Fig. 17 is the drawing of a
specimen of Mussa distans in the Oxford Museum, a species
in which the cespitose type shows a tendency to form a
Meeandrine type of colony. Some of the calices are seen to be
perfectly separate from one another, and are surrounded by
little rings of calcareous matter (ep.); whether one calls it
peritheca, exotheca, or epitheca does not matter. Where two
calices are closely apposed this tissue may be largely developed,
and may fill up the valleys completely to the lips of the
calices, as in the left-hand corner of the figure (c@.). . This is
then a ceenenchyme, and a comparison of this figure with fig. 6
must irresistibly lead to the conclusion that the two are formed
in one and the same manner. In the one case it is looser and
has a more adventitious appearance (Mussa), in the other it is
more solid and resembles in texture the rest of the corallum
(Caryophyllia oculina), that is all. The views here
expressed accord very well with Professor Martin Duncan’s
description of epitheca in some serial coralla (‘ Linn. Soc.
Journ.,’ xviii, p. 361), and with what is known of its develop-

ANATOMY OF MUSSA AND EUPHYLLIA. 43

ment. Beginning at the stage of development drawn in fig. 13
and mentally following the epitheca through its succeeding
stages, we see that it is really a basal structure to begin
with, and that as growth proceeds it follows the edge of
the “‘ Randplatte” as the latter retreats farther and farther
from the base. Where a compound colony is formed by
lateral budding, and the ccenosare represents the united
*Randplatten” of all the polyps, the epitheca will form a
lamellar structure at the bases of the polyps. Where, as in
the serial coralla Porites and Leptoria, the septa of adjacent
polyps fuse together, and there is no theca proper separating
the interseptal loculi of the two, it can easily be understood
how the epithecal nodules described by Professor Duncan
would be formed in early stages of growth.

It is obvious from the foregoing that the differences between
the tissue which is early laid down to consolidate the theca,
and ccenenchyme, and epitheca, depend on quantity and texture,
and not on the region of the polyps from which they are
formed.

In my description of Mussa and Euphyllia I have referred
to the existence of dissepiments. An examination of figs. 3
and 4 shows that these oblique partitions running across the
interseptal loculi are formed from the calicoblasts of what
were originally the interseptal parts of the base of the polyp.
The soft tissues do not occupy the whole of the cavity of the
calyx except in its uppermost part, but slope off to a point
below. An examination of fig. 4 shows the relations of these
parts to the dissepiments, and fig. 3 shows that the spaces
between the dissepiments and the theca are not occupied by
any soft tissues whatever. There are probably periods of
active coral secretion alternating with periods of reproduction
in these polyps. During the latter period the thin dissepi-
ments are formed by the basal tissues, whilst in the former
period the septa increase greatly in height, the polyp is, as it
were, moved higher up upon its stem, and deserts the old
dissepiments upon which it was resting. Then follows a new
period of reproduction, during which new dissepiments are

44. GILBERT ©. BOURNE.

formed. The process is comparable with the formation of
tabulee in the Pocilloporide and the growth of Tubipora.

Fowler shows in Lophohelia, as I do in Astrea (fig. 16), that
there are pieces of calcareous tissue intercalated between the
peripheral ends of the larger septa, each possessing its own
dark centre of calcification. These interstitial pieces, to
which Fowler gives the value of true thecal pieces, are clearly
formed from the calicoblasts in the angle where the soft tissues
dip down between the exsert septa. That this is a region of
specially active coral secretion is shown by the large calicoblasts
found there by Fowler in Lophohelia and myself in Mussa.
In the highest parts of the calyx, as growth proceeds, these
interstitial pieces may develop keels on their inner surfaces
which presently project into the calyx as a new cycle of septa
(vide fig. 7, Euphyllia).

The Costz.— These are clearly shown, from von Koch’s
account of the development, and from the study of such forms
as Astrea and Euphyllia, to be nothing more than the peri-
pheral ends of the septa projecting beyond the theca. But
there are structures called coste in Madrepora (Fowler) and
Leptopenus (Moseley) which do not correspond with the septa,
and clearly cannot be continuations of them. Such cost
alternate regularly with the septa in Leptopenus and in some
extinct forms (several species of Zaphrentis). In the latter
they are said to be epithecal in structure; unfortunately we
have no specimens in the Oxford museum which illustratg the
point. But, from the figures of Leptopenus, I am inclined to
think that the so-called costz in this form are epithecal in
origin, formed, that is, by the representative of the “ Rand-
platte” in this form (which presumably has the same relations
as in Fungia). Until we know more about the development
of the skeleton in the Perforata, it would be rash to dogmatize
about the “ cost ”’ in Madrepora.

I have treated the questions relating to the corallum at
length, because every fresh form that is examined convinces
me that the expectations formed of founding a new classifica-
tion of the Madreporaria on the anatomy of the polyp are to

ANATOMY OF MUSSA AND EUPHYLLIA. A5

meet with disappointment. There is singularly little variation
in the forms hitherto examined. Hence I believe that a re-
modelled classification must depend on a much more intimate
study of the structure of the corallum than has hitherto been
attempted.

Von Heider (‘ Zeit. fiir wiss. Zool.,’ xliv, p. 507) recognises
this, and attempts to found two new divisions, Euthecalia
and Pseudothecalia, on the characters of the corallum.

His Pseudothecalia would include all those forms whose
theca is formed by fusion of the peripheral ends of the septa;
his Euthecalia all forms in which the theca is a separate and
distinct structure (according to him, Astroides, and perhaps
Flabellum). His classification is based upon his account of
the structure of Astroides, which I cannot accept without
further evidence. It stands in direct contradiction to von
Koch’s account of the developmeut of the same genus. Von
Heider states that the corallum is clothed externally by a
single layer of ectoderm, mesoglea, and endoderm, the latter
layer abutting on the corallum. How then is the corallum
external to the polyp, as it undoubtedly is, if von Koch’s
account of the development be true, and there is no reason to
doubt that itis? Thus there is no proper “ Randplatte,’’ no
extrathecal coelenteron, in Astroides, according to von Heider;
yet von Koch expressly states that the theca cuts the mesen-
teries in two, and divides the polyp into an outer moiety and
an inner moiety. Finally, in the ‘‘ Euthecalia’’ founded on
Astroides, we learn from development that the theca is formed
from the fused ends of the septa; yet, by definition their theca
is a distinct structure. Unfortunately the specimens of
Astroides which I have examined were not well preserved
enough to admit of accurate observation. The external surface
of all was closely invested by a calcareous sponge, which had
in some cases apparently destroyed all traces of polyp external
to the corallum. Might not von Heider have been deceived
by the existence of a similar sponge on his specimens ?

In attempting to collect material for a new classification
from the observations already made, we must pay attention

46 GILBERT ©. BOURNE.

(1) to any marked tendency to departure from a radial sym-
metry or from the normal Actinian symmetry as shown in the
arrangement of the mesenteries; (2) To the presence or
absence of a ‘‘ Randplatte.” Iam not disposed to attach any
importance to exoceelic or entocelic septa for classificatory
purposes, since it appears that in one and the same colony
entoccelic septa only, or both exoccelic and entoccelic may be
found (Madrepora).

With regard to the symmetry. Owing to the presence of
“ directives,’ Actinia and many Madreporaria show a well-
marked bilateral symmetry, more marked in the case of many
Madreporaria by an irregularity in the arrangement of the
septa laterally, whereas the regularity is maintained at the
ends of the chief axis. But in Mussa, Euphyllia, and Lopho-
helia, the radial symmetry is perfect; and this is probably a
more primitive condition of things. In Madrepora aspera
there are signs of increased differentiation along the long axis,
since there are no filaments on mesenteries 3, 5, 8, 10 (Fowler’s
numeration), and in M. Durvillei there is a strongly marked
bilateral symmetry. This symmetry is even more marked in
Seriatopora and Pocillopora, and it can scarcely be doubted
that the latter forms are allied to the Madreporine, although
the one family is aporose the other perforate.

Recent researches have made it doubtful whether any sharp
distinction can be drawn between the two groups. Von Koch
has shown that the septa are formed from the basal ectoderm,
and that the theca is formed by fusion of the peripheral ends
of the septa in Astroides, a perforate Madreporarian. A study
of the adult anatomy of Astrza, Mussa, Lophohelia, Eu-
phyllia, Fungia, and others demand a similar explanation.
Fowler has shown that a well-developed “ Randplatte” with
extrathecal mesenteric elements exists in Rhodopsammia, a
perforate. Where a ccenenchyme with its correlative cceno-
sarc is present the extrathecal parts of mesenteries are not
present. The foregoing part of this discussion has shown that
a common ceenosare is due to nothing more than a persistent
connection between the ‘“ Randplatten” of adjacent polyps

ANATOMY OF MUSSA AND EUPHYLLIA. 4.7

(vide fig. 6), and that the two structures are homologous.
Where a coenosarc is present it develops a secondary connec-
tion with the echinulations of the coenenchyme and is sup-
ported on them, the extrathecal part of the mesenteries being
at the same time aborted. This is equally the case in the
aporose forms Seriatopora and Pocillopora, and the perforate
Turbinaria and others. No distinction then can be drawn
between aporosa and perforate on account of the presence of
a well-developed ‘“ Randplatte ” containing extrathecal ccelen-
teron divided up by mesenteries into exocceles and entocceles.
The difference between the two groups depends upon the
nature of the connection established between the peripheral
ends of adjacent septa. It is solid and continuous, an aporose
theca is the result ; it is loose and trabecular, a perforate theca
is formed. The manner in which the mesenteries may be
pierced, as it were, by outgrowths from the walls of adjacent
septa, is well shown by the formation of synapticula in Fungia.
In this form processes arise from the walls of the septa which
grow towards similar processes from the contiguous septa, and
meeting them fuse with them to form synapticula. This may
easily be understood by a study of a carefully macerated
corallum. I have shown that the mesenteries are actually
perforated by these synapticula, and that a simple canal
system exists connecting the extrathecal with the intrathecal
celenteron. This process is carried out further and in a more
complicated manner in the perforate. My researches on
Fungia have shown that the elevation of the Fungide into a
group Fungida is altogether groundless. Their anatomy does
not differ from that of other Madreporaria either in the
arrangement of tentacles, as Dana erroneously described, or in
the internal structure as Professor Duncan inferred from Pro-
fessor Moseley’s perfectly correct description of the soft parts
of a Bathyactis after decalcification. I have not yet had the
opportunity of investigating Bathyactis, but Professor Mose-
ley’s description of the appearance of its soft parts after
decalcification exactly tallies with the appearance of Fungia
under the same circumstances,

48 GILBERT ©. BOURNE.

The one form which must be placed in a separate group is
Flabellum. It possesses no “ Randplatte,” its calyx therefore
cannot grow peripherally, and in section it presents an entirely
different appearance to any other coral. Fowler gives accurate
drawings of sections of Flabellum in his latest paper
(« Anatomy of the Madreporaria,” iii), and it is not worth
while to reproduce them. His explanation, however, is not
quite adequate to explain the phenomena. If we assume that
Flabellum develops on the type of Astroides—and from the
anatomy of the adult we have no right to assume that it
develops otherwise, it is impossible to see how the theca could
have been formed by fusion of the peripheral ends of the septa,
and yet no soft parts left external tothe theca. To understand
this the reader must mentally follow the processes of growth
with the aid of figs. 18, 14, and 15, up to the adult condition.
My view of the so-called theca of Flabellum is that it is really
a basal structure which has grown upwards to form a calyx, in
a manner analogous to the opposite process by which the theca.
of Fungia has flattened out to form an apparent basal struc-
ture. The base would always lie actually, as well as appar-
ently, external to the polyp, and would entirely enclose it. It
could only be added to on its inner side, and thus Fowler’s
interpretation of the appearance of the centres of calcification
would continne to hold good. Moreover, the peculiar appear-
ance of infolding towards the centres of the septa, figured but
not discussed by him, would receive a sufficient explanation
when we consider that the septa are formed continuously with
the basal plate, and within folds of the three layers, each of
which necessarily includes two layers of calicoblasts. The
form of the corallum exhibits in fact the relation of the
primary folds of tissue within which the septa are developed.
The sutures—which, it must be observed, do not reach to
the exterior of the “theca”—would necessarily result
from the further growth of the septa. If I am right in consi-
dering epitheca to be continuous with and indistinguishable
from the basal plate, my views on Flabellum coincide
curiously with those of von Koch, who regards the apparent

ANATOMY OF MUSSA AND EUPHYLLIA. 49

theca of Flabellum as epitheca, much as I disagree with many
of his expressed views on epithecal structures.

The only arrangement, then, which our present knowledge of
the Madreporaria permits us to make, is as follows:

1. Madreporaria with no directive mesenteries and a per-
fectly radial symmetry, Lophohelia, Mussa, Euphyllia.

2. Madreporaria with directive mesenteries and a combined
radial and bilateral symmetry, Turbinaria, Rhodopsammia,
Fungia, and many others.

3. Madreporaria with reduced radial symmetry and marked
bilateral arrangement of parts, Madrepora, Pocillopora,
Seriatopora.

4, Madreporaria with a basal pseudotheca and no “ Rand-
platte,’ Flabellum.

I do not pretend tnat this arrangement has the value of even
an incomplete natural classification, but it is the best arrange-
ment that the facts warrant us in making. Finally, it must be
observed that the skeleton of the Madreporaria differs widely
from that of Alcyonaria. In the former the calcareous tissue
is nearly certainly elaborated and secreted by the ectodermic
cells ; it is always external to the polyp. In the latter, ecto-
dermic cells early separate from their primitive position, be-
come embedded in the mesoglea, and develop spicules within
their substance. The skeleton then is within the polyp. In
some cases skeletal structures may in the latter group be
developed within endoderm cells (E. B. Wilson, “ The develop-
ment of Renilla,” “ Phil. Trans.,”’ clxxiv, p. 723).

Unsatisfactory as the conclusions as to classification given
above may be, I hope that I have succeeded in presenting the
facts known about the morphology of the Madreporaria in a
manner comprehensive enough to be of use to future investi-
gators in a difficult field.

VOL, XXVIII, PART 1.—NEW SER. D

50 GILBERT C. BOURNE.

EXPLANATION OF PLATES III & IV.

Illustrating Mr. G. C. Bourne’s paper “ On the Anatomy of
Mussa and Euphyllia, and the Morphology of the Madre-
porarian Skeleton.”

Fig. 1.—Portion of a colony of Mussa corymbosa, of which the polyps
are retracted and shrunk by the action of alcohol. Half natural size. 7. Limit
of soft tissues external to the corallum. (Randplatte of von Heider.)

Fic. 2.—Transverse section of the calyx of Mussa corymbosa, showing
two principal and one secondary septum. The dark lines or areas are centres
of calcification, aud around them are concentric lines of growth. s. s. Sutures.
d, Dissepiments.

Fic. 3.—Transverse section across a single polyp of Mussa corymbosa,
to show the arrangement of the septa and dissepiments. The corallum is
represented in white and the soft tissues in black. The spaces between the
dissepiments and the theca not occupied by any soft tissues are shaded. 74.
Theca. sp. Principal septa. sp*. Secondary septa. d. Dissepiments. +.
Randplatte.

Fic. 4.—Diagram of a longitudinal section through Mussa corymbosa.
On the right side the section passes close to one of the principal septa. On
the left side a mesentery is exposed. The mesoglea is represented in black,
the ectoderm shaded with vertical lines. s¢. Stomodeum. ec. Ectoderm.
mg. Mesoglea. ex. Endoderm. cy. Calicoblast layer. m/f Mesenterial fila-
ments. m. Mesentery, with longitudinal muscles. 7. Randplatte. d. Dis-
sepiments.

Fic. 5.—Part of a transverse section through the hard and soft parts of
Mussa corymbosa, just below the level of the stomodeum. The section
comprises one of the ends of the long axis of the polyp. Mesoglcea, ectoderm,
and endoderm shaded as in Fig. 4. m’. Extrathecal portions of the mesen-
teries. mc. Pleats of mesoglea, to which the mesenterial muscles are
attached. Remainder of the lettering as in Fig. 4.

Fic. 6.—Euphyllia glabrescens, natural size, showing the expanded
polyp and the extension of the Randplatte over the lip of the calyx. At x
the Randplatten of adjacent polyps are seen to be continuous, forming a
ceenosare. 7. Limit of the Randplatte.

Fic.7.—Transverse section through the calyx of Euphyllia glabrescens,

showing primary (sp'.), secondary (sp*.), and tertiary (sp*.) septa. The centres
of calcification are shown by dark lines. s. s. Sutures. d. Dissepiments.

ANATOMY OF MUSSA AND EUPHYLLIA. 51

Fic, 8.—Part of a transverse section through a decalcified polyp of
Euphyllia glabrescens. This section shows the Randplatte 7, including
the extrathecal cclenteron and the extrathecal portions of the mesenteries>
m', The endoderm, ex’., of the intrathecal part of the polyp is greatly vacuo-
lated, forming a reticular tissue, which fills up the whole of the ccelenteron and
contains in its meshes zooxanthelle and nematocysts. The stomodeal canals,
st. c., occupy the axial part of the polyp, and serve as the digestive cavity.
ov. Ova. xz. c. Nutritive cells.

Fic. 9.—Developing nematocyst from the stomodeum of Euphyllia
glabrescens. Magnified 750.

Fie. 10.—Ripe nematocyst from the stomodeum of Euphyllia glabres-
cens, with the axis tube everted, but the thread not ejected. Magnified 750.

Fic. 11.—A nematocyst similar to that in Fig. 10, but with the thread only
ejected, showing the armature at the tip of the latter.

Fic. 12.—A nematocyst from the endoderm of Euphyllia glabrescens.

Fic. 13.—Longitudinal section through an Astroides embryo, copied from
von Koch, showing the basal plate, dp., being formed from the ectoderm
(calicoblast layer) of the base of the polyp, and a septum, sp., in the process
of formation. The epitheca, ep., is seen at the base of the polyp. z. The
piece of cork on which the embryo rests.

Fic. 14.—Diagram to exhibit the relations of the polyp to the corallum.
T. T. Tentacles. ec. Ectoderm. mg. Mesogloea. ex. Endoderm. s¢. Stomo-
deum. mf. Mesenterial filaments. r. Randplatte. m. Mesentery. m!. Ex-
trathecal portion of mesentery. ¢h.'Theca. Bp. Basal plate. ep. Epitheca.

Fie. 15.—Diagram to exhibit the formation of the theca from the fused
peripheral ends of the septa. cm. Columella. Remainder of the lettering as
in Fig. 14.

Fic. 16.—Section across the calyx of Astrea cavernosa, showing
alternately larger (sp'.) and smaller (sp*.) septa. In each septum can be seen
the dark centres of growth and the sutures, s., marking off contiguous septa
from one another. 7. ¢2. Intercalated pieces lying between the ends of the
septa (true thecal pieces ?).

Fie. 17.—Mussa distans. In this specimen the corallites may in some
instances be seen standing apart, in which case their lower portions are covered
by a loose epitheca, ep. In other places the corallites are set closer together,
and the valleys between them are partially filled up by a loose epitheca. In,
yet other places the calices are completely soldered together by epitheca, ca.
which fills up the valleys to the lips of the calices, and has the same functions
and relations as ccenenchyme.

MPT UT WRAY

M

ss

W 61g

LY 700) mung wy

Uf
ELL

“yr

MBS:

p suinog')'y

12

Ra esa

a~

a .
a

u?

ATTA

Mi

;

gaanuenl :

eae rnssau Ni Tpy Mj

SOUT

<>

—>

ue

UMMog? LON,
“ F?

AL
A

2A)
ia

y

Nf hAh
€ dsl hasis

TAY

20% 7;
ed
Ue

r

Pliyy

BR

"eigen la Alp ALD NA bannalaaapdan thea

ti po

a 7
% eT

w ht

5 a : 1a

4

ae WORT By pe

vay

y

INTRA-OVARIAN EGG OF SOME OSSEOUS FISHES. 53

On the Intra-Ovarian Egg of Some Osseous
Fishes.

By
Robert Scharff, Ph.D., B.Sc.

With Plate V.

Tue following researches were carried out during the past
summer while acting as assistant to Professor McIntosh, at
the St. Andrew’s Marine Laboratory. The fresh material for
the investigations was taken from St. Andrew’s Bay, which is
extremely rich in all kinds of animal life. Professor McIntosh
kindly placed at my disposal many preserved ovaries and ova
from his large collection, which served me for sections. But
my contributions to the history of development leave a good
many gaps to be filled up by future observers, as I was not
able to obtain a consecutive series of the ovaries of any species.
If an ovary be opened it is generally found to contain eggs of a
certain size only, and in order to get an entire succession of
the various stages, ovaries of the same species should be pro-
cured at different seasons of the year. It is not possible to
see all the phases of development in one and the same ovary.
There is one form, however, which abounds near St. Andrews,
and which, due to its spawning period being spread over
several months, contains ova of almost all sizes. This is the
gurnard (Trigla gurnardus). It is in many other points a
very suitable object for investigation.

The following is a list of the species of fishes whose ova or
ovaries I examined :—Trigla gurnardus, Gadus virens,
G. eglefinus, G. luscus, G. merlangus, Lophius

54 ROBERT SCHARFF.

piscatorius, Salmo salar, Anarrhichas’ Jupus,
Conger vulgaris, Blennius pholis, Hippeglossoides
limandoides.

In several cases the investigation was carried out on fresh
ovaries ; others were only inspected in a preserved condition.
I cut sections of all of them. They were hardened in weak
chromic or in picrosulphuric acid.

With regard to the result of my researches, I may mention
that there were two features which seemed to me of special
interest. Firstly, the development of the yolk, and secondly,
the origin of the egg-membranes and the follicle. I think I have
been successful in tracing the first, and also answered part of
the latter question. But the smallness of the objects presents
great difficulties, and the ova, after they pass a certain size,
become so opaque that their structure has to be studied entirely
from sections of hardened specimens. I propose to divide this
paper into five chapters, beginning with the nucleus and its
changes, and finishing up with a general account of the
development of the intra-ovarian egg. An abstract of this
paper was read before the Royal Society at their meeting in
the beginning of December last. It will be found in this
year’s Proceedings of the Society.

The various stages of the growing ovum have all, or nearly
all, been seen in the gurnard’s egg; but, in order to show that
they aiso occur in the other forms, where it was possible,
illustrations have been copied from sections of the different
species examined.

I. Tue Nucievus AND Its CHANGES IN THE SMALLER OVA.

I found the smallest ova, measuring 0°01] mm. in diameter,
in the ovary of the Haddock. In these eggs the nucleus
occupies almost the whole of the interior (fig. 1). A very
narrow zone of protoplasm, which, as far as I could ascertain,
was not bounded by a membrane, surrounded the nucleus.
The wall-less ovum lies in the endothelial or connective tissue

INTRA-OVARIAN EGG OF SOME OSSEOUS FISHES. 55

stroma of the ovary. The nucleoli (fig. 1, »’.)—there are a
great number of them—rest as a rule on the inner surface of
the nuclear wall, barring a few, which take up a central
position. With an ordinary high power (Zeiss F*) nothing
more of an internal structure can be detected.

The next size (fig. 2) shows a division of the protoplasm into
two distinct layers or zones, an outer lighter and an inner
darker or more denser one. The egg in this stage is still
tolerably transparent. Fig. 2 represents a small gurnard’s
egg (Trigla gurnardus) 0°030 mm. in diameter. The
nucleoli show an inclination to gather still more towards the
periphery of the nucleus, and the central portion rarely reveals
a germinal spot, while there appears instead an intra-nuclear
network which will be more minutely described in larger eggs.
In ova of this bulk, and also in somewhat larger ones, one or
more of the nucleoli become larger than the others, and in
their interior highly refractive specks are visible which have
sometimes been described as endonucleoli. Another pecu-
liarity about the large germinal spots is that they are always
surrounded by a light portion which does not stain with
carmine.

In somewhat larger ova, of a diameter of 0:080 mm. (fig. 3),
the dark zone round the nucleus is seen to have increased con-
siderably while the light one remained stationary. The large
nucleus which is represented in fig. 4 shows that in some
cases the big nucleoli disappear almost completely, leaving an
unstained part around them. Frequently one spot was seen in
the centre of the nucleus among what is generally known as
the “chromatic” substance. This consists of very minute
granules distinguished from the rest of the nuclear substance
by its greater consistence, a higher refractive power, as well as
by its capability of assuming a strong tint in certain solutions
of colourmg matter which are used among microscopists for
staining nuclei. The small granules are suspended in a net-
work of threads, which has so often been described in both
animal and vegetable cells, and which plays a conspicuous part
in the karyokinetic figures of the dividing nucleus. It is

56 ROBERT SCHARFF,

specially well seen in fig. 9, n. f., representing a nucleus and its
surrounding protoplasm of a middle-sized gurnard’s egg.

I agree with my friend Dr. Will in not attaching any morpho-
logical significance to the nucleoli. They must be regarded as
large masses of chromatic substance. In some instances they
are entirely absent; in others one or more may be present.

To return to the large nucleoli which, as has been men-
tioned, are occasionally present in the ova of the gurnard, they
are never wanting in the eggs of the conger eel (Conger
vulgaris). They stain slightly darker than the small ones.
In several cases (fig. 5, »’) I noticed a small nucleolus being
constricted off from a large one. In an egg of Gadus virens
(fig. 6), measuring 0°105 mm. in diameter, the dark protoplasm
(pr.1) surrounding the nucleus in smaller ova had been sepa-
rated in form of a ring, and internally to it another narrower
ring (p7r.*) of protoplasm was frequently present. The zone
of light protoplasm externally had increased considerably
meanwhile. In some instances, however, the dark zone had
invested the whole of the ovum, and the light portion had
entirely disappeared. Another feature which came under my
notice now was that the dark zone contained the faint outlines
of spots corresponding in size to nucleoli (fig. 7, sp.), none of
which, however, were seen outside in the light protoplasm.
Only in Hippoglossoides did I occasionally observe similar
spots close to the surface of the egg.

There can be little doubt that these spots are nucleoli which
have travelled through the nuclear membrane into the sur-
rounding protoplasm, and are gradually dissolved there. I
notice here incidentally that possibly some find their way to
the surface of the egg to form the nuclei of the follicular epi-
thelium. This is only a supposition; but it will be referred to
again more fully in a subsequent chapter.

In a larger egg of Gadus virens, 0°132 mm. in diameter
(fig. 7), we still find a granular protoplasmic ring round the
nucleus ; however, a great change has come over the nucleoli,
which are now no more closely attached to the nuclear wall.
They seem to have become broken up or dissolved in some

INTRA-OVARIAN EGG OF SOME OSSEOUS FISHKS. 57

way, assuming all sorts of shapes. Some are crescent-shaped,
some are like rods, and others again are reduced to small
specks or angular fragments. Now what is the significance of
all this ?

To commence with the relation of the dark protoplasm
towards the light one, I may previously state that this division
of the contents of the ovum has been seen by several authors
not only in fishes and amphibians, but also in invertebrate
forms. Bambeke! observed a division of the protoplasm into
two zones in the young ova of Leuciscus rutilus, Hippo-
campus antiquorum, and Lota vulgaris. Eimer? believes
in the idendity of His’ “ Rindenschicht” with His’ “ Zonoid
layer ;”’ his figures, however, show clearly that in young ova, at
any rate, it corresponds tothe light protoplasm which I described,
in distinction to the more granular part internally to it. The
fact that the yolk is divided into zones has also been observed in
Sepia by Lankester.®? Finally, my friend Dr. Will‘ has noticed
the same phenomenon in Orthoptera. No doubt the forma-
tion of protoplasmic rings round the nucleus of the growing
egg has been seen in a still greater number of animal groups
than I have just enumerated.

I think there can be little hesitation in regarding the dark
central protoplasm as owing its origin to the nucleus, although
there appear to be cases, such as recorded by Lankester,3 in
which the protoplasm is nourished entirely from without.

But a difficulty presents itself here. Has the dark part
originated by a simple transformation of the light portion, or
has another substance been added to the protoplasm from the
nucleus causing this change? The latter seems the most
probable of the two cases, This view is considerably strength-

1 Bambeke, “ Recherches sur l’embryologie des poissons osseux,” *‘ Mém,
cour. Acad. Belg.,’ vol. xi, 1875.

* Eimer, “ Untersuchungen iiber d. Hid. Reptilien und Fische,’ ‘ Archiy
f. mikr, Anat.,’ vol. viii, 1872.

* Lankester, E. Ray, “ Contributions to the Developmental History of the
Mollusca,” ‘ Philosophical Transactions,’ 1875.

* Will, ‘‘ Bildungsgeschichte und morphologischer Werth. d. Hies yon N epa
und Notonecta,” ‘ Zeitschrift f. wiss. Zool.,’ vol. xli, 1885.

58 ROBERT SCHARFF.

ened by an observation which was published by Ransom! in
1867, in the ‘ Philosophical Transactions.? He says: ‘ The
action of the water is not the same on the free uninjured ger-
minal vesicle in young ova as it is on those still within the
egg. I found that the results were in a great measure due to
the influence exerted by the constituents of the yolk, which
were carried into the vesicle by osmose. When the ovum was
acted upon by water, the germinal spots were gradually seen
to become pale, and finally disappear. These facts strongly
suggest the notion that the germinal spots are soluble in some
of the constituents of the yolk, and we may thus explain their
disappearance in ripe ova.

Ransom likewise observed that in the earlier ovum of
Gasterosteus leiverus and pungitis—the two species
which he examined—the germinal spots, which were embedded
in the colloid matrix of the vesicle, were to be seen at the
periphery of the vesicle only, so as to be in contact with the
inner surface of the nuclear wall. The germinal spots were
often ‘‘ tailed and vacuolate.”

In an older stage, such as that represented by fig. 8, the
dark ring encompassing the nucleus has evidently been ab-
sorbed, and the latter has dimished in size while the egg itself
has continued to grow larger. Indications of the boundary of
the old germinal vesicle are still seen (fig. 9, n’.),and the space
between it and the reduced one is filled with dark granular
protoplasm, which seems still to be produced by the action of
the nucleus. A new process, however, begins at this stage,
namely, the formation of the yolk spherules, which will be
described in the next paragraph.

Il. Toe Larcer Ova anp THE FoRMATION OF THE YOLK
SPHERULES.

If we look at figure 9, which represents the nucleus of a
large egg surrounded by a zone of protoplasm indicating the

1 Ransom, ‘ Observations on the Ovum of Osseous Fishes,” ‘ Phil. Trans-
actions,’ vol. clvii, 1867.

INTRA*OVARIAN EGG OF SOME OSSEOUS FISHES. 59

outlines of its predecessor, we are at once struck by the peculiar
protuberances which make their appearance all over its outer
surface. They were best seen in sections through the hardened
ova of the gurnard. These measured somewhere about 0:130
mm. in diameter. These diverticula, buds or ‘“ stolons,’’ as
they have been called by Balbiani, are pushed out by the
nucleus. Part of the nucleoli resting upon the diverticula is
drawn into them and carried away towards the exterior of the
egg. They have the appearance now of minute vesicles or
cells containing a nucleus. For such, indeed, they have been
mistaken even by the most recent writer on the subject, Ovsi-
annikov. The vesicle is either a portion of the nuclear wail
which has become constricted off, or it may be a later forma-
tion. Sometimes the vesicles thus formed do not contain any
nucleolar matter and remain unaffected by staining reagents.
Like the others they travel towards, but they do not quite reach
the surface of the egg, leaving a cortical layer of protoplasm
which is the “ Rindenschicht ” of His. The vesicles with their
nucleolar contents are the yolk-spherules. The solid mass
in their interior soon breaks up into fine granules, and it is in
this condition that the yolk-spherules are found in the largest
intra-ovarian eggs. The granules, however, are at first of a
dark colour, which they only lose on the ovum becoming ripe.
In the mature egg the yolk is perfectly transparent.

The nuclear network which has been mentioned above is
specially well seen at this stage (Fig. 9, ”.f.). The threads
connecting the minute granules in the interior of the nucleus
seem to be made up of fine dots rather than solid fibres.

The vesicles with clear contents might possibly be what is
known as “ oil globules,” on the significance of which my friend
Mr. Prince? has recently published an interesting paper. Their
oily contents would thus originate from the clear nuclear sub-
stance. I merely throw this out asa suggestion, but if it should
ultimately be proved correct, it would form another addition to
these most interesting and instructive phenomena which the

? Prince, Kd. E., ‘On the Presence of Oleaginous Spheres in the Yolk of
Teleostean Ova,” ‘ Ann. Nat. Hist.,’? 1886.

60 ROBERT SCHARFF.

ovum of osseous fishes is undergoing during its growth. In
the ripe gurnard’s egg a large oil-globule is to be found at the
periphery of the yoke, and similar ones occur in many other
marine Teleosteans. The largest yoke spherules in the gurnard
(fig. 10) have a diameter of 0015 mm. In the frog-fish
(Lophius piscatorius), however, their size varies from
0-018 up to 0:090 mm. (fig. 11). The contents in the latter
case consist, besides the small granules, of peculiar crescent-
shaped bodies and vacuoles. In many of them dark crystal-
line bodies were seen enclosed in a vesicle. Ovsiannikov!
likewise makes mention of crystals and peculiar oval bodies as
occurring in the yolk spherules of Osmerus eperlanus.
With regard to the process of budding from the nucleus
which I just described as leading to the origin of the yolk-
spherules, similar changes of the nucleus have been noticed by
various observers; but only Will? derives the yolk-spherules
from these buds. He investigated the intra-ovarian egg of
Amphibia. No other writer seems to have seen anything
corresponding to this process in Vertebrates. Roule,® Fol,* and
Balbiani® also describe the formation of diverticula from the
surface of the nucleus. The first two studied the ova of Asci-
dians, the latter those of Myriapods. However, the ultimate
fate of these diverticula containing nucleolar portions is to
become cells of the follicular epithelium, which in Teleostean
ova has been formed before those changes began. That the
yolk-spherules originate within the egg has been rigorously
maintained by some of the greatest authorities, such as Balfour
and Gegenbaur, but neither of the two mentions the budding

1 Oysiannikov, “Studien iiber d. Ei hauptsachlich d. Knochenfische,”
‘Mém. Acad. Imp. St. Petersb.,’ vol. xxxiii, 1886.

2 Will, “ Ueber d. Entstehung d. Dotters und d. Epithelzellen bei Amphi-
bien und Insekten,” ‘ Zoologischer Anzeiger,’ 1884.

3 Roule, “La structure de l’ovaire et la formation des ceufs chez les Phal-
lusiadées,” ‘ Comptes Rendus,’ 1883.

4 Fol, “Sur Porigine des cellules du follicule chez les Ascidiens,” ‘ Comptes
Rendus,’ 1883.

6 Balbiani, “Sur lorigine des cellules du follicule chez les Géophiles,”
‘Zoologischer Anzeiger,’ 1883.

INTRA-OVARIAN EGG OF SOME OSSEOUS FISHES. 61

process of the nucleus. Balfour! writes on this subject:
“ Yolk-spherules arise as extremely minute, highly-refracting
particles in a stratum of protoplasm some little way below the
surface, and are always most numerous at the pole opposite the
germinal vesicle.” ‘It deserves to be specially noted that
when the yolk-spherules are first formed, the peripheral layer of
the ovum is entirely free from them.” ‘Two points about
the spherules appear clearly to point to their being developed
in the protoplasm of the ovum, and not in the follicular epi-
thelium. 1. That they do not make their appearance in the
superficial stratum of the ovum. 2. That no yolk-spherules
are present in the cells of the follicular epithelium, in which
they could not fail to be detected.”” Balfour’s opinion is that
Gegenbaur’s* account of the formation of the yolk-spherules in
birds is probably correct. ‘ Protoplasmic molecules,” says
Gegenbaur, grow into larger granules. These become larger
again, and are transformed into vesicles, which increase con-
tinuously in volume. Newsolid products now take their origin
in the interior, and dissolve into fine granules.” “ The yolk-
spherules originate in the interior of the egg, and not from the
follicular epithelium.” These observations, which were made
in birds’, agree with those in reptiles’ eggs.

On the other hand, yolk-spherules appear also sometimes to
be formed from without. At any rate,> Beddard regards it as
highly probable. His view, which was deduced from the study
of the ovarian ovum of Lepidosiren (Protopterus), is supported,
among other facts, by their being no external membrane in this
stage. The follicle lies immediately on the yolk, and masses
of migrating cells were seen in course of being budded off
from it.

To return to the nucleus again; it rapidly degenerates as

1 Balfour, F., “Structure and Development of the Vertebrate Ovum,”
* Quart. Journ. Micr. Sci.,’ vol. xviii, 1878.

2 Gegenbaur, “ Ueber d. Bau und d. Entwicklung d. Wirbelthier Eier mit
partieller Dottertheilung,” ‘ Miiller’s Archiv,’ 1861.

2 Beddard, F., “Observations on the Ovarian Ovum of Lepidosiren,”
‘ Proceedings of the Zoolog. Soc.,’ 1886.

62 ROBERT SCHARFF.

the egg grows older. Whether the variously shaped particles
of the remaining nucleoli, which we have seen in figure 7,
subsequently congregate in the centre, in order to become
round again, I have not been able to observe. Such a trans-
formation, however, according to Iwakawa,! seems to obtain in
the young ovaof Triton. At any rate the nucleus becomes
smaller and smaller, to the benefit of the yolk. In an almost
ripe ovum very little of it is left (fig. 12). It then lies excen-
trically, and consists of clear contents, with the exception of a
few round nucleoli. More of the latter are seen just outside
the nucleus in the surrounding yolk. Large vacuoles (va.)
have also made their appearance in the immediate neigh-
bourhood of the nucleus, but their presence may be due to
the fixing reagent. It is important to note that no trace ofa
membrane can be made out at this stage of the germinal
vesicle.

The gradual decomposition of the nucleus in the growing
egg has been seen by many authors, and we have descriptions
not only from vertebrate, but also from invertebrate ova.
Thus Hertwig,? in his researches on the formation of the egg
in Toxopneustes lividus, says, “ When the ovum becomes
ripe, the nucleus undergoes regressive metamorphosis, and is
driven by means of protoplasmic contractions to the surface of
the yolk. Its membrane dissolves, and its contents break up
and are reabsorbed by the yolk. The nucleolus, however, seems
to remain unchanged, and appears to travel into the yolk mass,
becoming the permanent nucleus of the ripe ovum.”

With regard to the question as to the presence or absence of
a membrane in the nucleus, I could see a double contour dis-
tinctly in the section represented by figure 7. i believe it is
now generally recognised that the nucleus is surrounded by a
membrane, and most of the references I can find support this
fact. Thus Hertwig, in the paper quoted above, speaks of the

1 Twakawa, “The Genesis of the Egg in Triton,” ‘Quart. Journ. Mier.
Sce.,’ vol. xxii, 1882.

2 Hertwig, “ Bildung d. thierischen EHies,” ‘ Morphol. Jahrbuch,’ vol. i,
1875.

INTRA-OVARIAN EGG OF SOME OSSEOUS FISHES. 63

nuclear membrane as being of a distinct double contour and
well defined from the surrounding part, as well as from the
contents of the nucleus. I could cite many other authorities,
but it would lead me too far.

III. Tae Eocco-MemsBRANEs.

At the recent meeting of the British Association at Bir-
mingham, I submitted a paper on the egg-membranes of
osseous fishes. I pointed out that, in a section of the gur-
nard’s egg, I had seen a thin membrane internally to the egg
capsufé or “zona radiata,” and that it had been already noticed
by Ravsom and other observers.

I think the egg-membranes constitute that part of the
Teleostean ovum on which most has been written. I shall
attempt to throw some light on the discrepancies which we
find among zoologists in regard to these membranes. All
zoologists agree as to the existence of one more or less thick
layer, which most of them believe to be pierced by minute
pores. Now I find that this layer has no less than seven
different names, as will be seen by the subjoined list:

1. Zona radiata.—Balfour, Beneden, Brock, McIntosh, Ovsiannikov,
Reichert, Solger.

2. Vitell. membr.—Aubert, Beddard (?), Cunningham, Haeckel, Kolliker,
Waldeyer.

. Egg-capsule.—His, Miler, Prince.

Yolk-sac.—Ransom.

. Zona pellucida.—LHimer.

. Chorion.—Leuckart, Rathke.

. Egg-shell.—Oellacher, Vogt.

NO oP w

Leaving out the follicular layer, or “ granulosa”’ as it has
been called especially by continental investigators, we have
besides the above-mentioned membrane, descriptions of one or
two or even three others. As I am here only dealing with the
intra-ovarian egg of a few marine forms, some of which were
not even ripe enough to possess egg- membranes, my observa-
tions will necessarily be somewhat incomplete in this respect.
I will commence with a description of the “zona radiata,”

64 ROBERT SCHARFF.

following Balfour and cthers in adopting this name for the
principal membrane. As it is probably in all cases pierced by
radiating pores, the term “zona radiata” indicates its most
prominent morphological character. The name “ vitelline
membrane ”’ is misleading, additional membranes having been
described, some of which no doubt owe their origin also to the
vitellus. I believe one of the principal causes of the very
great divergence of opinion as to the number of membranes, is
that in some cases only intra-ovarian and in others ripe ova
have been examined. Balfour! has shown in his careful
researches on the intra-ovarian egg of Elasmobranchs, that in
many cases an absorption of part or the whole of the mem-
branes takes place. The disappearance of the ‘‘zonoid
layer” in the ripe ovum of the gurnard, shows that such an
absorption may also occur in Teleosteans. The ova of Lepi-
dosiren, described by Beddard,* afford another example of an
imbibition of the membrane taking place. I may mention
here incidentally that I am inclined to look upon his zona
radiata as the above-mentioned zonoid layer.

In Trigla gurnardus the zona radiata in section does not
show a distinct striation, but in the fresh egg it is well visible.
Its thickness (figs. 8, 13, 14, 15, z.) averages about 0:008 mm.
It is often granular and stains darkly as a rule in carmine and
hematoxylin. Internally to it is a much broader layer (figs.
8, 18, 14, 30), which in section appears to be the inner por-
tion of the zona, the stripes being apparently continued through
both. The width of the latter is about 0:025 mm. Both layers
are striped, i.e. provided with minute radial pores. I was in-
clined at first to consider these two layers as belonging to one
membrane, namely, the “‘ zona radiata.”

However, its semi-fluid condition distinguishes it from the
much firmer and elastic zona radiata. Hitherto it has always
been looked upon as the outer portion of the yolk, and has been
described by Gegenbaur in the ova of birds, reptiles, and
Elasmobranchs as the “ helle Randschicht,” and by His as the

1 Balfour, loc. cit.
2 Beddard, loc. cit,

INTRA-OVARIAN EGG OF SOME OSSEOUS FISHES. 65

“zonoid layer.” It stains only slightly, and I have found it
as a rule devoid of granules or vesicles. In the ripe ova this
zonoid layer disappears entirely. I cannot agree with His,
according to whom it is replaced by the egg-capsule or zona,
as the latter is always formed before the zonoid layer.

Brock } thinks that this thick layer is a peculiarity, probably
in some relation to the nourishment and growth of the egg.

According to Kupffer,” the membrane surrounding the yolk
in the herring’s egg consists of two layers, viz. an inner finely
striated and an outer one into which the striz are not continued.
The striez, he says, might be radial pores, however, he must
point out that, as the pores are not found in the external layer,
they certainly do not open exteriorly. Eimer’ first noticed the
stripes in the zonoid layer, which he believes are due to pro-
cesses from the follicle. He saw the stripes in Alburnis
lucidus, Salmo fario, and Perca fluviatilis, but they
only occupied the outer half and are not so delicate as those
in the zona. That this layer belongs to the yolk, he says,
is seen by the fact that in separating it follows always the
latter. My own observations are in direct opposition to this
last statement. Eimer also speaks of a membrane externally
to the zona. He believes that it originates from the follicular
layer and therefore might be looked upon as a chorion.

The same author regards the radial striation in the zona
radiata as being due to little rods between which are found
pores. He likewise examined the ovum of Reptiles, in which
he noticed besides the zona and the external “ chorion,’ an
internal membrane, of which I shall have to speak presently.

K6lliker* says “there is in all fishes’ eggs an outer, more
resistant, thinner layer externally to the zona which may even
preserve the striation in some cases.

’ Brock, “ Beitrage z. Anatomie und Histologie d. Geschlechtsorgane d.
Knochenfishe,”’ ‘ Morphol. Jahrbuch,’ vol. iv, 1878.

? Kupffer, ‘ Die Entwicklung d. Herings im Ei,’ Berlin, 1878.

3 Himer, ‘‘ Untersuchungen iiber d. Kid. Reptilien und Fische,”’ * Archiv f,
Mikr. Anatomie,’ vol. viii, 1872.

4 Kolliker, ‘‘ Untersuchungen z. vergleichenden Gewebelehre (also on

Ovum),” ‘ Verhandl. d. Phys. Medic. Gesellschaft,’ Wiirzburg,’ vol. viii, 1858.
VOL. XXVIII, PART ]1,—NEW SER, E

66 ROBERT SCHARFF,

Leuckart! makes mention of an outer membrane-like boun-
dary of the zona, through which canals are continued, and
Aubert? as well as Remak? speak to the same effect.

A few other observers such as His, Lereboullet,* Solger,®
and Ransom,* do not make any reference to a layer externally
to the zona, but admitting at the same time the porous struc-
ture of the latter; while Haeckel,’ speaks of a structureless
vitelline membrane only, having delicate dark spots. Ovsian-
nikov® again, the most recent writer on the ova of osseous
fishes, mentions three layers as occurring in Perca fluvia-
tilis. Prolongations from the follicular cells into the canals
of the zona were seen distinctly. No doubt they serve for
nourishing the egg. There is a distinct zona radiata externa
in Osmerus eperlanus, while in Gasterosteus there is only
one layer.

Lindgren,? who has recently examined the structure of the
mammalian ovum, found that the zona pellucida, which corre-
sponds to our zona radiata, is pretty often completely homo-
geneous in ripe eggs. He believes that this is due to different
physiological conditions of the egg. Author used the highest
powers available.

Johann Miiller!? was to my knowledge the first to identify

1 Leuckart, ‘“ Ueber die Mikropyle bei Insekteneiern (also Fishes),”’
‘Miiller’s Archiv, 1855.

2 Aubert, “ Beitrage z. Entwicklungsgeschichte d. Fische,”’ ‘ Zeitsch. f.
wiss Zool.,’ vol. v, 1854.

3 Remak, “ Ueber Hihillen und Spermatozoen,” ‘ Miiller’s Archiv,’ 1855.

4 Lereboullet, ‘“‘ Résumé d’un travail d’embryologie comparée sur le dével.
du brochet, &c.,” ‘ Ann. Sci. Nat.,’ vol. i, 4th ser., 1854.

® Solger, “ Dottertropfen in d. intracapsularen Flussigkeit v. Fischeiern,”
‘ Arch. f, Mikr. Anat.,’ vol. xxvi.

6 Ransom, loc. cit.

7 Haeckel, “ Ueber d. Hier d. Scomberesoces,” ‘ Miiller’s Archiv,’ 1855.

8 Ovsiannikov, loc. cit. :

9 Lindgren, “Ueber d. Vorhandensein v. wirklichen Porenkanalchen in d.
Zona pellucida v. Saugethieren,” ‘Arch f. Anat. und Phys. Anat. Abth.,’

1877. >
10 Miller, ‘“‘ Ueber Zahlreiche Porenkanale in d. Hi-Kapsel d. Fische.’
‘Miiller’s Archiv,’ 1854.

INTRA-OVARIAN EGG OF SOME OSSEOUS FISHES. 67

the punctures on the surface of the zona radiata with pores
piercing it. Since then, although we occasionally find state-
ments to the contrary, the great bulk of zoologists maintain
that pores exist in all cases. May not the fact that some of
the recent observers have not noticed them be due to the
peculiar physiological condition of the egg which Lindgren
referred to? Careful investigation of a large number of both
intra- and extra-ovarian ova of marine fishes would no doubt
help to clear up the matter.

In the zona radiata of the ripe egg of Gadus morrhua
(spirit preparations) no trace of a striation could be made out
with a magnifying power of about 800 diam. Cunninghan,!
however, has seen the stripes in ova of the same species which
were prepared with Perenyi’s fluid, and I firmly believe they
exist in all cases. As the same observer has recently pointed
out, the pores may occasionally (Myxine glutinosa) be
branched.

To return to His’ zonoid layer, I stated my reasons for con-
sidering it as one of the egg-membranes, although all previous
observers looked upon it merely as the modified external part
of the yolk.

The ova of Blennius pholis have one layer only. The
zonoid layer is always absent.

In speaking of the zonoid layer, His* says “It is probable
that it belongs to the zona radiata, but the detailed history of
both formations has still to be done.”

A question which has occupied the attention of many modern
observers on this subject is whether the pores of the zona
radiata receive processes from the surrounding follicular cells.
It seems to have first been suggested by Waldeyer® that the
egg-membranes are secretions from the follicular epithelium,
and that the latter sends prolongations through the pores into

1 Cunningham, “On the Structure and Development of the Reproduct.
Elements in Myxine glutinosa,” ‘Quart. Journ. Micr. Sci.,’ 1886.

2 His, W., ‘ Untersuchungen iiber d. Hi und d. Hientwickl. bei Knochen-
fischen,’ Leipzig, 1873.

3 Waldeyer, ‘ Hierstock und Hi,’ Leipzig, 1870.

68 ROBERT SCHARFF.

the interior of the egg. Eimer! is quite positive in his statements
of having seen prolongations from the follicular cells project
into the pores of the zona. This, however, was only observed
in the eggs of the ringed snake. Brock? again makes mention
of similar processes in the ovum of the barbel and that of
Servanus hepatus. Finally, the theory of the egg being
nourished by means of prolongations from the follicular cells
finds a strong supporter in Lindgren.? He shows, moreover,
that in the egg of the rabbit even follicular cells occasionally
travel through the pores of the zona. Some were observed
inside it, and others half way out. I myself have not been
able to trace any processes from the follicular cells into the
ovum, but it seems to me quite probable that such prolonga-
tions do exist. At any rate I believe that the ovum receives
its raw material as it were through the radial pores from the
follicle. It is then assimilated and transformed by the nucleus
and the egg nourished in this manner.

Before I proceed to a description of the follicular layer, I
must mention another membrane which has been described by
some authors, while its existence has been denied by others,
I am referring to an extremely delicate membrane covering
the yolk internally to the zona radiata. I have seen such a
membrane in some cases in the young gurnard’s egg (fig. 13,
m. i.), that is to say, only in sections of hardened specimens,
and not in the adult egg. It has been first described by Ransom,
who calls it the inner yolk-sac,” in distinction to the outer
yolk-sac (zona radiata). He also isolated it in Gasterosteus.
Oellacher did the same in the trout, after treatment with
chloride of gold. Allen Thomson, as well as Kolliker, Eimer,
Lereboullet, and Aubert, describe an inner delicate membrane.

Although Ovsiannikov saw this layer in Perca fluviatilis,
he found no trace of it in Lota vulgaris, and comes to the
conclusion that it is an artificial product.

1 Himer, loc. cit.
2 Brock, loc. cit.
3 Lindgren, loc. cit.

INTRA-OVARIAN EGG OF SOME OSSEOUS FISHES. 69

His, Waldeyer, Brock, and others, deny the existence of an
inner membrane.

In theory, the presence of such a membrane would explain
much which seems at present very difficult. If it really existed,
we would naturally regard the zona as being secreted from it.
It would also place the pores of the former into the same
category as those occurring in cuticular formations, with which
they indeed have much resemblance.

IV. Tse Fouuicutar Layer.

The follicular layer, or granulosa, surrounds the egg-mem-
branes which I have just mentioned. In the next paragraph
I shall speak about its development; hence I need only
describe its appearance in the ripe intra-ovarian egg.

In the gurnard’s egg it consists of a layer of closely-set
cells, which has an average thickness of 0:006 mm. Seen from
above, the cells present hexagonal outlines with a central
nucleus (fig. 16). The egg lies in a stroma of connective
tissue.

In the shanny’s egg (Blennius pholis) the follicle is
peculiarly modified (fig. 15, f). The depth of the cells, which
in one half of the egg is only about 0°007 mm., gradually
increases until it reaches 0:032 mm. at the opposite side. The
cells at that side become drawn out and taper towards the
surface of the egg. The space between the cells is filled up
with interstitial substance (7. s.). Another feature about the
follicle in this case is that it touches the zona in all parts,
except in a circular portion (c.p.), where it is not in immediate
contact with it. This space is filled with an apparently viscid
substance, which no doubt is secreted by the follicular cells.

Similar peculiar modifications of the granulosa have been
observed by Eimer and Brock. McLeod! likewise mentions
a granulosa composed of elongated cells as occurring in
Gobius niger.

* McLeod, “ Recherches sur la structure et le développement de l’app.
réproduct. femelle des Téléostéens,” ‘ Arch. de Biologie,’ vol. ii, 1881.

70 ROBERT SCHARFF.

Although I have never seen prolongations from the follicular
cells pass through the zona, as has been described by Eimer
and Lindgren, I strongly believe that the follicle forms the
most important source for the nourishment of the egg.

Lindgren! observed in the mammalian ovum that granulosa
cells occasionally travelled into the egg by means of the pores
of the zona. Some were seen inside the egg, and others half way
out, andprolongations from the follicular cells were frequently
traced to the interior. I may mention that nothing of the
kind was noticed by me in the eggs of osseous fishes; indeed,
the pores of the zona are much too narrow to allow of their
being traversed by the follicular cells.

V. DEVELOPMENT.

On account of my not being able to obtain all the different
ovarian stages, I have made no observations on the origin of
the egg. Whether the ovum originates from a simple trans-
formation of an epithelial cell, or whether several unite, as in
Elasmobranchs, has not been observed. I believe, however,
that probably only one cell is concerned in the formation of
the egg. This view is held by Brock? and Kolessnikov.2 The
former makes the following statement :—“ We often find im-
mediately under the epithelium, and frequently in direct com-
munication with it, numerous masses of cells. These masses
are seldom utricular—more often they are round or wedge-
shaped. The cells show every gradation from epithelial cells
to the smallest ova. The whole process lasts only a very short
time.”

Kolessnikov also mentions that the primary eggs are formed
by the growth of some epithelial cells, one cell giving rise to
one ovum. A division of the ovum, as described by Waldeyer,
could not be found by the same author in either Amphibians or
Teleosteans.

1 Lindgren, loc. cit.

2 Brock, loc. cit.

3 Kolessnikoy, ‘“ Ueber d. Eientwicklung bei Batrachiern und Knochen-
fischen,” ‘Arch, f. mikr. Anatomie,’ vol. xv, 1878,

INTRA-OVARIAN EGG OF SOME OSSEOUS FISHES. 71

The smallest ova I saw were those of Gadus eglefinus,
which measured only 00°11 mm. in diameter (fig. 1). By far
the greater portion of the egg was taken up by the nucleus,
and many of the numerous nucleoli were already ranged
round the inner surface of the nuclear membrane. A small
zone of protoplasm covered the nucleus outside. Apparently
there was no cell-membrane. The follicular layer was like-
wise absent.

In the next larger size, such as we find in fig. 2, which was
taken from the gurnard’s ovary, we find that the protoplasm
surrounding this nucleus is divided into a dark and light
zone. The explanation of this peculiar feature was given in
Chapter II. Among the nucleoli some are frequently of a very
large size. In figures 6 and 7 we now notice a follicular layer
surrounding the yolk and bulging somewhat into it, but the
cells composing it are large and not numerous. In fact at first
a few cells cover the whole of the ovum.

I have not arrived at any definite conclusion as to the origin
of the follicular layer. ‘There are three possible ways in which
it might have arisen:

1. From outside the egg, by an aggregation of epithelial
cells round the ovum, as in Elasmobranchs (Balfour).

2. From outside the egg by connective tissue or endothelial
cells collecting at the periphery of the ovum. Teleosteans
(His, Ovsiannikov). Cephalopoda (Lankester).

3. From inside the egg:

a. From the vitellus as in Ascidians (Sabatier).
6. From the nucleus as in Ascidians (Roule, Fol).
Myriapods (Balbiani). Orthoptera (Will).

Although some of my observations seem to show that the
follicular epithelium takes its origin from the interior of the
egg, others point the opposite way, and, on the whole, I
think it would probably be more correct to assign its origin
to the connective tissue. At any rate it is ready formed
before an egg-membrane can be discovered. Brock can give
no opinion as to the origin of the follicle. According to
His the very young ova are often surrounded by a double

72 ROBERT SOHARFF.

membrane of endothelium, while there is no trace as yet of a
granulosa.

In a young egg (fig. 6), as I mentioned before, a few large
flat cells cover its whole surface. As the ovum increases in
size they become more numerous and much smaller, until in
the ripe eggs they have the appearance as seen in figures 6,
145015 5f-

I have already spoken of the peculiar modification of the
follicular layer in Blennius pholis, and need not refer to
it again.

With regard to the origin of the egg-membranes, it was
stated that they appear after the follicle. The first membrane
is the zona radiata (figs. 8, 13, 14, 15, z). When it is fully
developed the inner ‘ zonoid ” layer is formed in those eggs in
which it is found. In Blennius pholis no zonoid layer.
ever appears. In the gurnard an inner layer was seen in
sections of middle-sized eggs, but in later stages no such layer
could be discovered. I have had occasion to refer to this layer
in a previous paragraph. I myself have no doubt that the
egg-membranes originate from the yolk. Ovsiannikov’,
however, holds that the zona is derived from the follicle; and
Cunningham,” although he has not had an opportunity of in-
vestigating the subject more closely, comes to a similar con-
clusion.

According to Eimer an outer layer chorion, of which I
saw nothing in the ova which I examined, originates from the
follicular cells. He, as well as most other authors who have
dealt with this subject, agree in the vitelline origin of the zona
radiata and the zonoid layer. In the ripe egg the zonoid layer
has disappeared completely.

Before concluding this short note on the intra-ovarian egg
of osseous fishes, I must refer to a few points which have not
been dealt with. An opening appears in the zona radiata
before the extrusion of the ovum from the ovary. This “ micro-
pyle,” as it has been called by its discoverer “ Doyére,” may

1 Ovsiannikoy, loc. cit.
2 Cunningham, loc. cit.

INTRA-OVARIAN EGG OF SOME OSSEOUS FISHES. 73

represent an enlarged canal of the zona, but I have made no
direct observation to prove this assertion. According to Cun-
ningham the micropyle in Myxine glutinosa is formed by
the growth of a cellular process from the follicular epithelium
towards the vitelline while the vitelline membrane is being
formed.

Another point which has still to be made out, is the ultimate
fate of the nucleus and the appearance of a “ discus proli-
gerus”’ in the unfertilised egg. I hope, however, to make
additional researches at some future period, in order to clear
up these matters.

EXPLANATION OF PLATE V,

Illustrating Dr. R. Scharff’s paper “ On the Intra-ovarian Egg
of some Osseous Fishes.”

Fie. 1.—Intra-ovarian ova of Gadus eglefinus. (Zeiss D3.) wz.
Nucleoli.

Fie, 2.—Intra-ovarian ovum of Trigla gurnardus. (Zeiss D‘.) WN.
Nucleus. z’. Nucleoli.

Fie, 3.—Intra-ovarian ovum of Trigla gurnardus. (Zeiss D+.)
NW. Nucleus. ' Nucleoli. pr. Dark protoplasmic ring. 7. p. Light proto-
plasm.

Fic. 4.—Nucleus of intra-ovarian ovum of Hippoglossides limandoides
(Zeiss D+.) zx’. Nucleoli.

Fic. 5.—Intra-ovarian ovum of Conger vulgaris. (Zeiss D.) wz’,
Nucleoli.

Fic. 6.—Intra-ovarian ovum of Gadus virens. (Zeiss D*.) x’. Nu-
cleoli. pr’. External protoplasmic ring. pr®. Internal protoplasmic ring.
f. Follicle. 72. ». Light protoplasm.

Fic. 7.—Intra-ovarian ovum of Gadus virens. (Zeiss F%.) z/. Nu-
cleoli. 2. f Nuclear figure. NV. m. Nuclear membrane. pr. Protoplasmic
ring. sp. Spots (nucleoli?). fA Follicle.

74, ROBERT SCHARFF,

Fic. 8.—Intratovarian ovum of Trigla gurnardus. (Zeiss D*.)
a. f. Nuclear figure. z. Zona radiata. zo. Zonoid layer. f. Follicle.

Fie. 9.—Nucleus of intra-ovarian ovum of Trigla gurnardus. (Zeiss
F*,) 2’. Nucleoli. W. Old nucleus. xz. Nuclear figure.

Fie. 10.—Yolk-spherules (Trigla gurnardus). (Zeiss D*.)

Fic. 11.—Yolk-spherules (Lophius piscatorius), (Zeiss D*.)

Fie. 12.—Nucleus of nearly ripe ovum of Trigla gurnardus. (Zeiss
D3.) W. Nucleus. 2’. Nucleoli. va. Vacuole.

Fic. 13.—Middle-sized egg of Trigla gurnardus. (Zeiss D*®) z.
Zona. zo. Zonoid layer. 7. m. Internal membrane.

Fic. 14.—Intra-ovarian ovum of Trigla gurnardus nearly ripe. (Zeiss
D’.) z. Zone radiata. zo. Zonoid layer. f. Follicle. y. sp. Yolk-spherules.

Fie. 15.—Intra-ovarian ovum of Blennius pholis. (Zeiss D’,) f.
Follicle. 7. s. Interstitial substance. . p. Circular portion filled with viscid
matter. z. Zona radiata.

Fic. 16.—Follicular cells of Trigla gurnardus (seen from above).
(Zeiss D3.)

Fic. 17.—Follicular cells of Blennius pholis (seen from above).
(Zeiss D*.)

R. Scharff del.

er fe

Moor bourrt Vou XIVUNS FY V

F. Huth, Lith Edin™ ©

i 4 a a
-"
a
.
>
.
a
" .
.
‘ -

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 75

Observations on the Structure and Distribution
of Striped and Unstriped Muscle in the
Animal Kingdom, and a Theory of Muscular
Contraction.

By

Cc. F. Marshall, M.Sc.,
Platt Physiological Scholar in the Owens College, Manchester.

With Plate VI.

Srripep muscle has long been known to occur widely dis-
tributed in the animal kingdom, but the details of the
structure of the striped muscle-cell have been the subject
of much controversy. Various descriptions have been given,
widely differing from one another, and none of them afford-
ing a satisfactory basis of comparison with other cells. The
demonstration of an intracellular network in the muscle-
fibre by several recent observers appears to afford the most
rational clue to its structure, for it not only explains all the
appearances seen in the muscle-fibre, including those seen
with polarised light in the living fibre by Briicke, but it also
renders possible a comparison with other cells, and shows that a
muscle-fibre is to be regarded as of essentially the same struc-
ture as an ordinary cell, and must not be considered as an
enigmatical structure, the details of which do not correspond
to those of any other cell in the animal economy.

It is necessary first to examine the descriptions of the several
observers who have described a network in the striped muscle-
fibre and to consider the interpretation that they have put
upon it. In the following account I have only referred to those

76 C. F. MARSHALL.

observers who have described some form of network in the
muscle-fibre.

Dr. G. Thin! appears to be the first observer who described
an intracellular network in the fibre of striped muscle. He
examined frog’s muscle treated with gold chloride in the fol-
lowing manner. After staining with gold chloride the muscle
was exposed to light in acidulated water and then kept in
strong acetic acid, at a temp. of 38° C., from 6—24 hours. By
this method he demonstrated a network of fine fibres, concern-
ing which he says “ this network was composed of exceedingly
fine fibres, and its meshes accurately corresponded to Cohn-
heim’s areas” (p. 252). He states that he demonstrated in
the muscle-fibre, by the process of isolation, (1) the existence
of flat cells (the muscle-corpuscles), (2) a network connected
with central cellular protoplasm, and (8) parallel rows of
spindle elements. Further on he states that he was.“ compelled
to associate the transverse markings with the existence of this
network, without attempting to explain the connection between
them more definitely ” (p. 258).

Gerlach? has written two papers on this subject. In the first
paper he states that the contractile contents of the sarcolemma
is traversed by a retiform substance continuous with and
identical with the axis cylinder of the nerve. He thus regards
the network as of a nervous nature.

In the second paper he states that (i) an intravaginal nerve
network is present within the sarcolemma. (11) In good
specimens striz are seen which behave in a similar manner to
the intravaginal network, and can be traced into continuity
with it. He divides muscle into an anisotropous contractile
matrix, and an isotropous nerve network. His results were
obtained by the following method : the gold preparations were
left several days in a mixture of 1—2 parts hydrochloric acid,

1 On the Structure of Muscular Fibre,” ‘ Quart. Journ. Mier. Sci.,’ vol.
xvi (N. S.), 1876, pp. 251—259.

2 «Das Verhiiltniss der Nerven zu den Muskeln der Wirbelthiere,’ Leipzig
(Vogel), 1874. ‘Ueber das Verhaltniss der nervosen und contractilen
Substanz der quergestreiftes Muskels,” ‘Arch. f. Mik. Anat.,’ Bd. xiii, 1877,

p. 399.

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 77

20 parts glycerine, and 20 parts water. This method brought
out the longitudinal striz. On treating gold preparations as
above and subsequently treating with 1 per cent. potassium
cyanide, he states that the sarcolemma gives way and the con-
tents escape, partly as fine particles and partly in larger pieces ;
in such pieces the transparent substance bounding Cohnheim’s
areas was stained red, and was thickened at the nodal points.
The muscle-corpuscles always lay in the stained substance and
not in Cohnheim’s areas; they consisted of a central oval nucleus
and a stained peripheral substance continuous with the stained
network of longitudinal striz. He states that the longitudinal
striz are very variable in thickness and always zigzag; never
straight. He regards them as thickenings of fine sheaths of
nervous matter enclosing the fibrille of the muscle; these
sheaths corresponding to the boundaries of Cohnheim areas.

He therefore concludes that the intravaginal nerve plexus
and the longitudinal striz are continuous, and together make
up the isotropous part of the muscle-fibre, and are to be con-
sidered as nervous. To them belong also the muscle-corpuscles
and the nuclei of the intravaginal nerves. He regards the
anisotropous matrix as the contractile part. Gerlach thus
appears to view the isotropous part of the muscle, stained by
the gold, as a honeycomb and not a true network of fibrils.
He has apparently failed to observe the transverse networks,
and does not attempt to explain the relation between the
network and the transverse striation.

The existence of an intravaginal nerve plexus in the muscle-
fibre, and also the continuity of the nerve end plate with the
isotropous part of the muscle, have been denied by Ewald! and
Fischer.”

Engelmann?’ regarded the isotropous part of the fibre as a
structure ‘‘ das in Physiologisches Hinsicht von einem Nerven
nicht wesentlich abweichen Wunde,” and suspected a connec-
tion with the axis cylinder.

1 ¢ Arch. fiir Mikr. Anat.,’? Bd. xiii, pp. 365—390.

* « Pfliiger’s Archiv,’ Bd. xii, pp. 529—548.
3 * Pfliiger’s Archiv,’ Bd. xi, p. 462.

78 O. F. MARSHALL.

In a later paper Engelmann! states that the contractility of
the muscle-fibre is always connected with fibrillar elements in
the fibre ; and he compares these with fibrillar elements in the
protoplasm of some of the Rhizopods and Infusoria, &c.

Retzius? describes very carefully a network in the muscle-
fibre of Dytiscus and other forms. This network consisted of
(i) transverse networks placed at regular intervals, and corre-
sponding in position to Krause’s membranes. (ii) Longitu-
dinal bars parallel to each other, and apparently running the
whole length of the muscle-fibre, and connected with the
transverse networks. His results were obtained partly by
transverse and longitudinal sections, and partly by teased
preparations. He also employed the following method of
gold staining: the specimens were placed for twenty-five
minutes in 1—4 per cent. gold chloride, either with or without
previous immersion in 1 per cent. formic acid; then in 1 per
cent. formic acid for ten to twenty hours, and exposed to the
light. He gives the following description of the muscle-fibre
of Dytiscus: In the axis of the fibre there are one or more
rows of muscle-corpuscles, the protoplasm of which is produced
into several (2—5) processes from which finer processes arise
forming the transverse networks. Each muscle-corpuscle
is in connection with five or six successive transverse net-
works. The longitudinal bars of the network he describes
rather doubtfully as consisting of rows of dots (p. 8), but he
describes and figures them projecting freely in some prepara-
tions. The matrix is structureless, and is only slightly
stained by the gold. The sarcolemma is apparently closely
attached to, but probably independent of, the network. The
nerve endings appear to be in close connection with the
transverse networks. The function of the network he was
unable to determine, but states that it is probably not merely
a supporting framework, but actively concerned in contraction.
He does not, however, regard it as the true contractile part,

1 ¢ Pfliiger’s Archiv,’ Bd. xxv, 1881, pp. 538—565.

2 «Zur Kentniss der quergestreiften Muskelfaser,” ‘ Biologische Unter-
suchungen,’ 1881, pp. 1—26, Pls. i, ii.

eee

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 79

as, according to him, it does not undergo important changes
in form during contraction and extension. He thinks that
the network is probably concerned in conveying the stimulus
from the nerve to the muscle-fibre. In support of this
he mentions that the fibre, as a rule, contracts simul-
taneously throughout its whole thickness; also that the
nerve-fibres are apparently connected with the transverse
networks.

Retzius also examined muscle from Musca, Oestrus, Noto-
necta, Locusta, Astacus, Rana, and Triton. In Locusta the
transverse networks had more polygonal meshes than in
Dytiscus. In Astacus, Rana, and Triton the longitudinal bars
of the network were thicker than the transverse.

In some cases he states that the longitudinal bars of the
network were not straight but zigzag. From the descriptions
and figures it is very probable that this appearance was due to
pressure, and is not normal. .

Bremer! also describes a well-defined network in the striped
muscle-fibre, evidently identical with that described by
Retzius. He states that the longitudinal lines are true fibrils,
and not part of cylindrical sheaths as Gerlach maintained.
He further traces the axis cylinder of the nerve into direct
continuity with the muscle-corpuscles.

He considers the longitudinal strie of Gerlach to be
identical with the longitudinal bars of the network, and
explains the irregular dotted appearance, or “ Sprenkelung,”
of Gerlach’s longitudinal striz as being due to imperfect
staining.

Bremer’s results, though published subsequently to those of
Retzius, were obtained quite independently.

B. Melland® has recently investigated the structure of the
striped muscle-fibre, and has arrived independently at results
agreeing very closely with those of Retzius and Bremer. This
close correspondence between the accounts of these three

1 © Arch. fir Mikr. Anat.,’? Bd. xxii, 1883, pp. 318—356.

7 “A Simplified View of the Structure of the Striped Muscle-fibre,”
‘ Quart. Journ. Micr. Sci.,’ July, 1885.

80 O. F. MARSHALL.

observers affords satisfactory evidence of the correctness of
their observations.

The points of difference between the networks described by
Melland and Retzius are slight, the chief one being that
Melland figures the transverse networks in Dytiscus with
more polygonal meshes, and furnished with nodal thickenings
at the points of junction with the longitudinal bars of the
network. Retzius figures these in Locusta, but in Dytiscus
he describes the transverse networks as generally composed
almost entirely of radial fibres with very few transverse
connections; and in place of nodal dots he describes several
thickenings or nodes placed irregularly, and much fewer in
number than the nodal dots described by Melland. Melland
does not trace any connection between the network and the
muscle-corpuscles nor with the nerve endings. He shows how
the optical appearances of striped muscle are caused by the
network. He considers the network to be intimately con-
nected with the sarcolemma, and to be homologous with the
intracellular networks which have been described in other
cells. It is evident from fig. 6 of his paper that we have
to do with a true network and not a honeycomb, a fact
which is not so apparent from the figures of Retzius and
Bremer. I have reproduced the figure from Mr. Melland’s
paper (fig. 18).

Mr. Melland’s results were obtained partly in conjunction
with myself, and the object of the present paper is a continua-
tion of this investigation. I have endeavoured to trace the
distribution of this intracellular network of the striped muscle-
fibre in the animal kingdom, and also, so far as possible, to
determine its function.

The striation of muscle must not be confounded with a
transversely striated appearance caused by a corrugated outline
of the fibre, possibly due to a state of over-contraction. Such
a false striation is met with occasionally in some fibres in the
Echinus, Leech, &c., and is the cause of the muscles of these
animals having been described as striped. I shall, therefore,
only describe muscle as being striped when the striation is

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 81

due to the presence of the intracellular network, described by
Retzius, Bremer, and Melland.

I have examined muscle taken from representatives of the
chief groups of the animal kingdom with the special object of
investigating the presence of an intracellular network in the
muscle-cells, either such as that of the striped muscle-fibre,
or, when this does not exist, an intracellular network of any
kind.

The Amceba and Hydra have been included in this investiga-
tion; for it is an important point to determine the existence
of an intracellular network in such a primitive and eminently
contractile cell as the Amcba; it is also important to in.
vestigate the structure of the muscular processes of the ecto-
derm cells of the Hydra, as they are supposed to represent the
first beginning of a muscle-cell.

In all cases the outlines and main details of the figures
were drawn with the camera; in most cases under the ;4,th
immersion objective of Beck with No. 2 eyepiece, giving a
magnifying power of 1100 diameters.

Methods of Preparation.—The chief method of pre-
paration used was the method of gold staining employed by
Melland. The gold stains and renders evident the intracellular
network of most cells, especially the network of the striped
muscle-cell ; hence it is at once a test whether the striation of
the fibre is due to the presence of the network, or whether it
is merely the false striation mentioned above.

Various modifications of the gold method were employed
according to the delicacy of the tissue under investigation.
The method employed by Melland consists in placing the
muscle in 1 per cent. acetic acid for a few seconds; then
in 1 per cent. gold chloride for thirty minutes; then in
formic acid, 25 per cent., for twenty-four or forty-eight hours
in the dark.

This answers well for vertebrate and insect muscles. But
for the more delicate organisms, such as the Hydra, Daphnia,
&c., and the heart muscle of invertebrates, I found a one
hour’s immersion in formic acid, exposed to strong sunlight,

VOL. XXVIII, PART 1.—NEW SER. F

82 OC. F. MARSHALL.

to be the best treatment; or, in some cases, a warm chamber
(40° C.) was used. A longer immersion than one or two
hours in the formic acid in these cases leads to disintegration
of the tissues. In addition to the gold preparations, osmic
acid preparations were made in most cases, and compared with
those made by the gold method. Osmic acid is well known for
its property of fixing the histological elements in their natural
state.

The examination of fresh tissues was in many cases of very
little use ; for the cells of the striped muscle of many of the
animals investigated are so small that under a high power they
barely appear striped, and no network can be seen at all. In
these cases it is only by softening the fibre and so swelling it
out, and at the same time staining the network, that the latter
can be demonstrated ; this is the special action of the method
of gold staining.

It is necessary to mention that the results obtained by the
gold method are somewhat uncertain. In some cases the net-
work will come out distinctly, but in others, especially when
the preparation has been left for a longer time than usual in
the acetic acid, the network appears to consist of rows of
granules instead of definite lines. This uncertainty was
noticed also by Retzius, Gerlach, and Bremer, and is no doubt
the cause of the different appearances described by these
authors.

In order to avoid the monotony and interruption of re-
peatedly stating the treatment used for the muscle of each
animal, the exact method used is given with the description of
the figure of each animal in the plates at the end of the paper.

I shall now take the chief group of the animal kingdom in
their zoological order, describing the muscle found in each
case.

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 83

STRUCTURE AND DIstTRIBUTION OF MUSCLE IN THE
ANIMAL KinGpom.

Protozoa.

Amcba.—Klein! states, on Heitzmann’s authority, that
under suitable conditions the protoplasm of the white blood-
corpuscles can be seen to contain an intracellular network
composed of fine fibrils. Dr. Klein has, however, recently
informed me that he does not find an intracellular network in
the Ameba, nor in the majority of white blood-corpuscles.

On examination of very large specimens of Amcba
princeps in the fresh state the constant flowing movement of
the protoplasm renders it difficult to conceive of any permanent
intracellular network. I have, moreover, made gold prepara-
tions of these Amcebe in the following manner :—The Amceba
was placed in a drop of water with a little cotton wool under-
neath the cover glass to prevent the animal being washed away
by the reagents. A few drops of 1 per cent. acetic acid were
then run in under the cover glass for a few seconds. Gold
chloride was then run in, and the animal left in this for fifteen
minutes. Formic acid was then added, and the animal left
exposed to light for about one hour ; by this time the gold was
reduced and the animal stained. The preparation was then
mounted in dilute glycerine.

Amecebe prepared in this way showed no trace of an intra-
cellular network ; the protoplasm simply presenting a mottled
granular appearance.

Although there is no definite intracellular network, com-
parable to that of an ordinary epithelial or gland cell, known
to exist in any of the Protozoa, yet a vacuolated condition of
the protoplasm is well known to occur in many of them. This
attains a high degree of development in many forms, e. g. Noc-
tiluca. These vacuoles are certainly not all food vacuoles, and
may possibly indicate the starting point of the differentiation
of an intracellular network, i.e. a differentiation of the cell into

1 «Atlas of Histology,’ p. 2, diag. 1.

84 OC. F. MARSHALL.

firmer and less dense parts, the former of which takes on the
form of a network or reticulum. For although it is not abso-
tutely certain that the structures described as intracellular and
intranuclear networks are in all cases denser than the rest of
the protoplasm of the cell, they are, I believe, generally
assumed by histologists to be so, and also to be protoplasmic in
nature.

VORTICELLA.

The stalk of the Vorticella contains a spiral protoplasmic
fibre, which is eminently contractile. This fibre, when treated
with the gold staining, shows no trace of the presence of fibrils,
having simply the appearance of undifferentiated protoplasm.

C@LENTERATA.

Hydra.—The peculiar ectoderm cells of the Hydra are
important to investigate, since they are generally held to repre-
sent the first commencement of a muscle. Here the one cell
is differentiated into two parts to perform two functions, the
one portion to act as a sensory cell, the other to act as a
muscle.

Hamann! describes, in the epithelial muscle-cells of the
hydroid polypes, a network in the body of the cell, but in
fibrillation in the muscular process.

My own observations on the cells of the Hydra agree with
those of Hamann. Gold preparations of these cells show a
network in the body of the cell, but no continuation of it into
the muscular process (fig. 1).

Medusa.—Striated muscle has been described as occurring
in the disc of Aurelia by Max Schultze, Bricke, and Virchow,
and in Pelagia by Kélliker.®

In gold preparations of muscle from the disc of Aurelia I
find distinct transverse striation, which, under the ;4, immer-
sion objective, is found to be due to the presence of a network

1 ¢Organismus der Hydroid Polypen,’ p. 15.
2 *Stricker’s Handbook of Histology,’ vol. iii, p. 551.

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 85

similar, in all respects, to the network described by Retzius
and Melland in striped muscle (fig. 2).

Actinia.—Muscle taken from the base of the Actinia and
treated with gold was found to consist of elongated fusiform
cells, non-striped, and showing no trace of any intracellular
network, or of any fibrillation.

Hence the conclusions obtained are that in the muscular
process of the Hydra cell there is no form of network or fibril-
lation, although a network is present in the body of the cell.
In the more highly organised Medusa the typical network of
striped muscle is found to be present, but in the equally highly
organised Actinia no network is present, nor is there any
fibrillation in the muscle-cells.

These results agree with those obtained by Hamann! by the
method of isolating the cells by maceration in various reagents.
He states that the muscles of Hydroid polypes are always
smooth, and quite distinct from the striated muscles of the
Medusze. Hamann thinks that where transverse striation has
been described in Hydroid polypes it is probably due to the
action of reagents.

The Hertwigs,” in their observations on the Actiniz, describe
no fibrillation in the muscle-fibre. They investigated the
tissues of Sagartia and Anthea.

EcHINODERMATA.

The muscle of the Echinoderms has been described as striped
by several observers. Firstly, by Schwalbe*® in the muscle-
cells between the ambulacral plates of Ophiothrix, and
more recently by Geddes and Beddard.* From the figure
given by the latter observers it is evident that the striation
they describe is the false striation mentioned above as being
due to annular constrictions. Schwalbe, however, describes
double oblique striation.

1 Loe. cit., p. 20.

2 * Die Actinien.”

3 * Archiv fiir Mic. Anat.,’ 1869, p. 205.
4 * Proc. Royal Soc. Edinburgh,’ 1873.

86 Cc. F. MARSHALL.

I have made gold preparations of muscle taken from the
“lantern of Aristotle ” of Echinus, and find no trace of the
network of striped muscie or of any fibrillation. The cells are
remarkable for the clearness and transparency of their proto-
plasm.

These results again agree closely with those obtained by
Hermann.t He describes the muscle of the Asterids as
smooth, and very seldom showing fine longitudinal fibrillation.
In the Holothurians he describes the muscle as non-striped,
but states that longitudinal fibrillation is to be seen in it.

VERMES.

Hirudo.—The muscle-fibres of the Leech are peculiar:
they consist of an outer clear portion and a central granular
part. In gold preparations the outer part stains the more
deeply of the two portions of the cell and appears quite
homogeneous, showing no trace of a network. In osmic
acid preparations the outer layer appears very faintly fibril-
lated, but I could not identify any distinct fibrils differentiated
from the rest of the cell even under the ;4, immersion
objective.

Transverse sections of the muscle of the Leech show a
radiating appearance of dark and light bands in the outer
portion of the cell. This is, I believe, caused by the method
of preparation, for in some sections the outer portion of the
cell is broken up into pieces arranged in a radiating manner
and corresponding to the light portions between the radiating
dark lines in the better preserved specimens. I find nothing
corresponding to this appearance in muscle prepared by the
gold or osmic acid methods, which are the methods generally
recognised as maintaining the true histological characters of
cells intact.

Wagener,’ from transverse sections of dried specimens of
Leech, states that the muscle-cells consist of a central medul-

1 ¢ Asteriden,’ p. 94, plate iii; ‘ Holothuranien,’ p. 38, plate il.
2 «Archiv f. Mic. Anat.,’ 1869.

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 87

lary substance round the nucleus, and a cortical substance
splitting into fibrils. This is also described by Schwalbe! in
the fibres of Aulostoma.

Lumbricus terrestris.—Gold preparations of the muscle
of the Earthworm show large elongated cells which on close
examination show longitudinal lines; these under the +>
immersion objective present a dotted appearance (fig. 3).
At first sight it might appear that we have here the network of
striped muscle; but this is not the case. In the first place,
there is no appearance of transverse striation at all; iu the
second place, the dots are not arranged transversely but are
quite irregular ; lastly, so far as I could observe, the dotted
lines are superficial and do not extend into the body of the
cell.

One of the so-called “ hearts” of the worm was treated in
the same way ; the muscle-cells were found to resemble almost
exactly the muscle-cells described above.

In the Polyzoa and Rotifera striped muscle is well known to
occur. I have not, however, been able to determine with cer-
tainty whether the striation is due to the presence of a net-
work or not. Nitsche! says that the striation in the retractor
muscle of the Polyzoa is not due to any wrinkling of the sar-
colemma. The retractor muscle appears to be the only muscle
that is striped from Nitsche’s observations.

Striped muscle has been recently described by Haswell? in
the gizzard of Syllis one of the Polychxte worms.

Mo.uuvsca.

According to Schwalbe,’ double oblique striated muscle is
present in Solen, Ostrea, and Helix.

Anodon.—Preparations of the adductor muscle of the Ano-
don treated with gold show that the muscle consists of small
elongated cells of the unstriped type, showing no fibrillation

1 «Kenntniss der Bryozoen,’ Heft 2, p. 55.
2 © Quart. Journ. Mier. Sci.,’ 1886.
3 ¢ Arch. f. Mic. Anat.,’ 1869.

88 C. F. MARSHALL.

or transverse striation. The muscle treated with osmic acid
shows faint fibrillation, but no distinct fibrils.

Patella——The Limpet was chosen for an investigation of
the structure of the heart muscle. In gold preparations of
the ventricle I find the network of striped muscle present
(fig. 5).

In the heart of the Anodon I could not determine with cer-
tainty whether the network was present or not, although faint
indications of it were obtained.

Ostrea.—The adductor muscle of the Oyster consists of two
portions: a white opaque portion, and a more gelatinous por-
tion. Gold preparations were made of each of these. The cells
of the ‘white muscle” are large, with clear outlines, and re-
markable for the clearness and transparency of their proto-
plasm. The cells of the “gelatinous muscle” are smaller
and less transparent. Neither of these showed any network or
fibrillation.

Helix pomatia.—Gold preparations of the muscle of the
foot show that it consists of very small cells of the unstriped
type densely massed together.

The muscle of the odontophore, however, shows transverse
striation, which under the high power is seen to be caused by
the presence of the typical network of striped muscle.

Pecten.—The Pecten differs from most of its class by per-
forming rapid movements of its adductor muscle whereby it
propels itself through the water. Gold preparations of the
adductor muscle made by my friend Mr. J. T. Cunningham
show the network of striped muscle very plainly (fig. 4). I
have not observed the double oblique striation described by
Schwalbe in the muscle of Molluscs and Echinoderms. As
this is not seen in gold and osmic acid preparations, I think it
must be an optical effect. Schwalbe, indeed, admits that in
Ophiothrix the transverse striation is due to folds in the
sarcolemma (loc. cit., p. 211).

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 89

ARTHROPODA.

Representatives of the Crustacea and Insecta, viz. the Lob-
ster, Dysticus, and the Bee, were investigated by Mr. Melland,’
the network (of striped muscle) being found in each case.

Astacus, Heart-muscle.—Gold preparations of the heart
of the Crayfish show the network to be present in this muscle
as in the body muscles ; the network is, however, much finer
and more difficult to demonstrate (fig. 6).

The muscle-fibres of the heart are intimately blended with
what appear to be large masses of granular protoplasm enclos-
ing nerve-cells ; these may possibly be of the nature of nerve-
endings.

Daphnia.—As a representative of the minuter forms of
Crustacea, I examined the Daphnia. The muscle-fibres of this
animal, when examined in the fresh state, only show transverse
striation faintly. After many attempts I succeeded in obtaining
a satisfactory gold preparation, where the muscle-fibres were
much softened and pressed out to many times their normal
diameter. These fibres show the network very plainly (fig. 7).

In this case the animal was placed whole in 1 per cent. acetic
acid for ten minutes, and left in the formic acid in a warm
chamber at 40° C. for two hours.

Insect Larva.—To determine if striped muscle is present
in the larval insect as well as in the imago, I made prepara-
tions from the larva of the Ermine Moth (Spilosoma lubri-
cepeda). Muscle was taken both from the jaws and from
the legs. In both cases the muscle was found to be striped,
the network in the muscle of the jaw being especially well
developed.

ARACHNIDA.

Muscle taken from the leg of the Spider and treated with
gold showed the network of striped muscle.

1 Loe. cit.

90 Cc. F. MARSHALL.

VERTEBRATA.

The vertebrate animals examined by Mr. Melland were the
Frog and the Rat. These will serve as examples of the Am-
phibia and Mammalia. I have examined the muscle of animals
taken from the other chief groups, viz. the Cyclostomata,
Elasmobranchia, Teleostea, Reptilia, and Aves, taking
as representatives of these groups respectively: Myxine, Scyl-
lium, Gastrosteus, Testudo, and Turdus.

Muscle taken from each of these animals was treated with
the usual gold method, and in each case a network was found
identical with that described by Melland. On comparing
these networks with one another and with those described
above in the striped muscle of the several invertebrate animals,
they are found to agree in all respects.

With regardto Amphioxus, I have not had the opportunity
of examining fresh specimens. The muscle has, however, been
described as striped,! and (from analogy) I see no reason to
think that the striation is due to any other cause than a
network. In the Ascidian, I have examined the muscular
bands of Salpa and find striped muscle present.

Carpiac MuscLE.

The heart muscle has long been described as faintly striped
transversely, but whether this striation is due to the same cause
as that to which ordinary striped muscle owes its striation, has
not been determined with certainty.

In order to investigate this point, I made gold preparations
of muscle taken from the Rat’s heart. The cells are seen to
contain a network similar to that of ordinary striped muscle.
The network is more delicate and with much smaller meshes
than the network in the body muscle of the same animal, and
is therefore more difficult to demonstrate by the gold method
(fig. 9). I have also prepared muscle from the heart of the

1 Grenacher, ‘ Zeit. fiir wiss. Zool.,’ p. 577.

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 91

Frog (fig. 10) and Bird (fig. 11); the network in the latter
animal is much plainer than in the others.

The striation of cardiac muscle therefore appears
due to an intracellular network similar to that of
ordinary striped muscle.

Unstrirpep Muscie oF VERTEBRATES.

Klein! describes in the unstriped muscle of vertebrates a
bundle of longitudinal fibrils which are in connection with the
intranuclear network. This description of the structure of
the unstriped muscle-cell is as follows: ‘Thus we may regard
the unstriped muscle-fibre as composed of a sheath with annular
thickenings and a bundle of delicate fibrils which at one more
or less central point forms a delicate network. This, sur-
rounded by a special membrane—except where the network is
in connection with the bundleof fibrils—represents the nucleus.”
Klein regards the bundle of fibrils as the contractile part of
the cell, and thinks that by their shortening the muscle-fibre is
caused to contract, elongation being produced by the elastic
rebound of the sheath. Flemming? has observed these fibrils
in the living muscle.

Dr. Klein informs me that he regards the bundle of
longitudinal fibrils as representing an intracellular network
homologous with other intracellular networks. Dr. Klein
has been kind enough to show me his preparations of unstriped
muscle prepared from the mesentery of the newt by twenty-four
hours immersion in 5 per cent. ammonium bichromate, and
afterwards stained with logwood. These preparations show
clearly the longitudinal fibrils and their connection with the
intranuclear network.

I have made preparations, by the gold method, of muscle
from the mesentery of the newt and from the bladder of the

1 Klein, ‘ Atlas of Histology,’ p. 74, pl. xv; also ‘Quart. Journ. Mier.
Sci.,’ 1878.

> « Beobachtungen tiber die Beschaffenheit des Zell Kerns,” ‘Arch. fir
Mikr. Anat.,’ 1876, Bd. xiii, pp. 714, 715.

92 C. F. MARSHALL.

Salamander, in both of these the fibrils in the muscle-cells are
very evident, but the intranuclear networks do not show at
all distinctly, which is a most unusual result in this mode of
preparation. However, in preparations from the mesentery
of the newt, made by Klein’s method, the intranuclear net-
works come out very distinctly in many fibres ; and in one case
I could trace the connection of the intranuclear network with
the fibrils of the cell. The longitudinal fibrils do not show
so well in these preparations as in those made by the gold
method (figs. 12, 13). It thus appears that the vertebrate un-
striped muscle differs from all the invertebrate unstriped
muscle that I have investigated, in that the cells contain an
intracellular network in the form of longitudinal fibrils.
This may perhaps represent a form of network intermediate
between the typical irregular network of other cells and the
highly modified network of the striped muscle-cell.

From these investigations it appears that the peculiar intra-
cellular network of striped muscle is developed in all muscles
which have to perform rapid or regular movements.

A brief review of the chief animals mentioned in the
preceding pages will make this clear. Commencing with the
Actinia and the Medusa, these are both highly organised
Ceelenterates, but the Actinia is a sluggish animal which
exhibits slow and irregular movements, while the Medusa
propels itself through the water by rapid and regular contrac-
tions of its disc. Now, in the Actinia we find no striped
muscle, but in the Medusa the network is present. In the
worms such as the Leech and Earthworm striped muscle is
absent; these animals only performing comparatively sluggish
movements. In the Polyzoa the retractor muscles of the
stomach, and in the Rotifers the retractors of the trochal disc,
perform rapid movements, and have been described as striated
transversely ; this is probably due to the network, although,
as stated above, I have so far been unable to determine this
myself. In the Mollusca the movements are as a rule
sluggish, and unstriped muscle is the prevailing type in this
group. But in the odontophore muscles of the Snails the

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 93

movements are more rapid, and in these we find the network
developed. Also in the hearts of these animals which perform
rapid and regular contractions I find the network present, at
any rate in the case of Patelia.

In the Pecten we have a Molluse which differs from the
majority of its class by performing rapid movements by the
contraction of its adductor muscle, and here we find the
network present. This is a most important fact in favour of
the view that the peculiar network of striped muscle is deve-
loped when rapid movements are to be performed; for here
we have the Mussel and the Pecten, both belonging to the
same division of the Mollusca, and both having adductor
muscles moving the valves of the shell. In the Mussel the
adductor muscles only act at irregular intervals and compara-
tively slowly; but in the Pecten they perform rapid and
frequent contractions when the animal swims. In the Mussel
we find unstriped muscle, but in the Pecten the network of
striped muscle is present.

In the majority of Arthropods and Vertebrates the move-
ments are chiefly rapid and of frequent occurrence, and in
these groups there is a wide distribution of striped muscle.

It is quite possible that in some animals of sluggish habits,
such as some adult insects, the presence of striped muscle may
be due to inheritance.

We should expect on this view to find striped muscle present
in all well-developed hearts, since they execute rapid and
regular contractions. However, in the so-called “ hearts” of
the Earthworm the muscle is unstriped. This can, I think, be
explained as follows. These so-called “hearts” represent the
earliest and most primitive form of heart in the animal
kingdom, being simply local hypertrophies of the blood-vessels
which perform rhythmic contraction. Now, the muscle of the
blood-vessels is unstriped, therefore we should scarcely expect
to find striped muscle in what are simply local hypertrophies
of those vessels. Moreover, the contraction of these “ hearts ”
is slow and peristaltic in nature. It is only when we come to
the more highly developed hearts, such as those of the Patella,

94 C. F. MARSHALL.

Snail, &c., which have to perform much more rapid and regular
contractions than the “hearts” of the worm that we find
striped muscle developed.

I may here state that I have not yet been able to determine
the nature of the connection between the network of striped
muscle and the nerve end-plate, which must exist if the com-
bined results of Retzius and Bremer are correct. This I
hope to do in a subsequent paper. I have, however, recently
observed the connection between the network and the muscle-
corpuscles described by Retzius.

Turory or Muscunar ContTRACTION.

The general conclusions arrived at in the preceding part of
the paper are as follows :

(1) An intracellular network of a definite character is
present in the fibre of striped muscle throughout the animal
kingdom.

(2) This network is developed where rapid and frequent
movements have to be performed.

(3) The striped muscle-fibre consists of sarcolemma, network,
and sarcous substance ; and, so far as at present determined,
there is no other structure present in the fibre (excepting the
muscle-corpuscles and nerve-endings).

The question now before us is to determine if possible the
nature and function of the network, and what relation it bears
to the contractility of the muscle-fibre.

Changes in the Network during contraction.—In
order to investigate this point I teased out some perfectly fresh
muscle from the leg of a Dytiscus and placed it on an inverted
cover-glass over a gas-chamber. Alcohol vapour was then
blown over the preparation when most of the fibres contracted
owing to the chemical stimulus. The vapour was passed over
the muscle for about a quarter or half a minute. The fibres
were then fixed in their contracted state by plunging them into
5 per cent. acetic acid for half a minute, and then treated with

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 95

gold and formic acid in the usual way. Many fibres were thus
obtained completely contracted and also many fixed waves of
contraction.

I also made preparations of relaxed muscle from a Dytiscus
killed with chloroform. However, as the fibres vary so much
in appearance according as they are more or less pressed out
in the gold preparation, comparisons of the muscle stimulated
with alcohol vapour, with that reduced by chloroform, though
they may give the general effect of the difference, are not abso-
lutely trustworthy. The only way of really proving this point
is to examinea fibre, one portion of which is in the relaxed
condition and the other contracted, or, in other words, a fixed
wave of contraction.

On careful examination of the network in one of these fixed
waves of contraction with the 4, immersion objective the longi-
tudinal fibrils of the network were always straight inall
parts of the fibre and appeared slightly thicker inthe
contracted part of the fibre although it was difficult to judge
accurately of the difference in thickness. The nodal dots,
however,werethesamesize in boththecontracted and
relaxed portions of the fibre. The dots appeared in many
cases even smaller in the contracted than in the relaxed muscle.
This is, I believe, due to their being more separated from each
other laterally, whereby the refractive effects which somewhat
obscure the real size of the dots in the relaxed muscle are
diminished (fig. 14).

It therefore appears from gold preparations that during con-
traction the nodal dots do not alter in size but that the longi-
tudinal bars of the network increase in thickness. The apparent
enlargement of the nodal dots when the fibre is seen in the
fresh state is due to optical effect. Moreover, if the nodal dots
do not alter in size it follows necessarily that the longitudinal
bars must increase in thickness ; for since they keep straight
during contraction if they do not increase in thickness there
must be a diminution in the volume of the fibre, which is known
not to occur.

These results differ from the account given by Schafer of the

96 C. F. MARSHALL.

changes during contraction. He states! from observations on
the living fibre that during contraction his “ muscle-rods ”
(which correspond to the longitudinal bars of the network)
become compressed in the centre and their substance tends to
accumulate towards the ends, 7. e. that the knotted ends of the
muscle-rods, which correspond to the nodal points of the net-
work, increase in size at the expense of the shafts connecting
them. On examination of the living fibre this certainly appears
to be the case, but the optical effects of reflection and refrac-
tion are so great as to obscure the real change that takes
place.

NaTuRE AND FuncTION oF THE NETWORK.

In discussing the theory of contraction, I shall assume that
the intracellular network of striped muscle, and the longitu-
dinal fibrils of the vertebrate unstriped muscie, are of
the same nature as other intracellular networks; and, in
accordance with the views of modern histologists, that they
are protoplasmic in nature, and denser than the rest of the
cell.

We have first to consider the nature of intracellular net-
works in general, and whether the function is an active or a
passive one. In the case of intranuclear networks the changes
which the network undergoes in karyokinetic division of the
nucleus point to their being of an active nature. The extra-
nuclear network (intracellular) is apparently of the same
nature as the intranuclear, since the two have been shown to
be continuous in many cells; and also they have the same
behaviour towards stains and reagents. Moreover, if intra-
cellular networks are developed by a process of vacuolation of
the protoplasm of the cell, or a division into denser and less
dense parts, as described previously when treating of the Pro-
tozoa, it is obvious that in these cases the network must be
the active and the contractile part of the cell.

1 “Qn the Leg-muscles of the Water-beetle,” ‘ Phil. Trans.,’ 1873.

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSOLE. 97

The continuity and identity of nuclear and extranuclear
networks is strongly supported by Sedgwick’s remarkable ob-
servations on the early stages of Peripatus.1 He not only
demonstrates the continuity of the extranuclear and intra-
nuclear networks, but he also shows that during segmentation
of the ovum the cells do not become completely separated, but
remain connected by their protoplasmic networks, 1. e. that
the intracellular networks of all the cells are continuous.

He also states that the so-called nuclear membrane is reti-
cular in nature and not a true membrane, being, in fact, part
of the general reticulum of the cell. In the cells described by
Sedgwick there is no doubt that the reticulum is the active
portion of the cell, for the rest of the cell consists simply of
vacuoles.

Flemming states,? as Sedgwick also noticed, that the first
change observable in a cell whose nucleus is about to divide
is in the extranuclear protoplasm. Strasburger® further states
that the fibrils which form the nuclear spindle originate in the
surrounding “ cytoplasm” at the time ‘of division. This ap-
pears to be direct evidence of an active function in the intra-
cellular network.

These considerations show that the function of intracellular
networks is very probably of an active nature.

We have now to consider the networks of striped and un-
striped muscle. Both these forms of network are non-essential
to contraction, for we have seen that many muscle-fibres of
invertebrates are devoid of a network of any kind; but that
they modify the nature of the contraction is very probable.

We have seen that the network of striped muscle is developed
when rapid movements are to be performed; this shows that
the function of contraction is intimately associated with the
presence of the network.

The chief points of difference in the contraction of striped

1 © Quart. Journ. Mier. Sci.,’ vol. xxvi, 1886, pp. 175—212.

* *Zellsubstanz, Kern u. Zelltheilung,’ Leipzig, 1882.

3 © Arch, f. Mikr. Anat.,’ Bd. xxiii, “‘ Die Controversen der indirecten Kern-
theilung.”

VOL. XXVIII, PART 1.—NEW SER. G

98 O. F. MARSHALL.

and unstriped muscle respectively are the great length of the
latent period and the long duration of the contraction in the
unstriped muscle. The velocity of the contraction-wave in
striped muscle is in the Frog, 3—4 metres per second, while
in the unstriped muscle «(ureter) the velocity is only 20—30
mm. per second.! This seems to indicate that the peculiarly
arranged network of striped muscle may be associated with
the rapidity of its contraction.

In nearly all the specimens I have examined both the
transverse and longitudinal bars of the network
remain perfectly straight in all conditions of con-
traction and relaxation of the muscle. Hence the
network, or part of the network, must either con-
tract to the full extent that the muscle-fibre does,
or else be elastic and so follow the movements of
contraction of the fibre.

Retzius? figures a specimen in which the longitudinal bars
are zigzag. However, from his description, and from com-
parison with my own preparations, I believe this to be due to
disturbance during the preparation and not to be a normal
condition.

We have now to consider whether the network is actively
contractile or merely a passively elastic structure; or whether
one part of it is contractile and the other passive. That both
network and sarcous substance are contractile is improbable;
for if the function of the network and the sarcous substance is
identical, there is no apparent reason for the presence of the
network. Differentiation in structure always implies differen-
tiation in function.

AcTION OF THE LONGITUDINAL Bars oF THE NETWORK.

We have seen that the longitudinal bars of the network
diminish in length and apparently increase in thickness during
contraction, and that they always remain straight in alll

1 ¢Text-book of Physiology,’ Dr. Michael Foster, 4th ed., p. 101.
2 Loc. cit., plate i, fig. 19.

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 99

conditions of contraction and relaxation of the fibre.
The question now before us is to determine if they are
actively contractile or passively elastic. The following
considerations are opposed to the latter view.

(a) If the longitudinal bars of the network are passively
elastic they must be on the stretch in the relaxed condition
of the fibre, and resemble stretched elastic threads running
the whole length of the muscle-fibre. Now, when a muscle is
cut out of the body, and thereby removed from its attachments,
it does not contract to any considerable extent; therefore,
supposing the longitudinal bars to be elastic, something must

keep them on the stretch.
_ (i) This cannot be the sarcous substance, for as it is semi-
fluid in nature it can hardly keep elastic threads on the
stretch.

(ii) It cannot be a nervous impulse, continually acting on
the longitudinal bars, for if it were so a muscle would contract
on section of its nerve.

(iii) The only force which can keep the bars on the stretch
must be that of the transverse networks. On this supposition
the uncontracted muscle is not in a state of rest, for there is a
continual force exerted against the transverse networks by the
tendency of the longitudinal bars to shorten. It is very diffi-
cult to conceive that the muscle, in its uncontracted condition,
should be in a state of extreme tension, and not of comparative
rest.

(4) In the unstriped muscle-fibre there are no transverse
networks present, and hence no force to keep the longitudinal
fibrils on the stretch, except the sarcolemma, which would be
scarcely adequate to do so.

It therefore appears improbable that the longitudinal bars
of the network are passively elastic, and if this is the case the
only conclusion remaining is that they are actively contractile,
and hence, presumably, the cause of contraction of the fibre.
This view is also supported by the following considerations ;

In the muscle-cell the part which performs the contraction
is evidently the most fundamental part of the cell, and this we

100 Cc. F. MARSHALL.

should expect to be differentiated first. In the embryonic
development of striped muscle it is found that the longitudinal
striation appears first, i.e. that the longitudinal bars of the
network are differentiated before the transverse. ‘This is also
the case in regenerating muscle. Again, in tracing the phy-
logeny of muscle, we found that the first indication of an
intracellular network was in the vertebrate unstriped muscle
in the form of longitudinal bars only. Hence both the phy-
logeny and the ontogeny of the network favours the view
that the longitudinal bars are the contractile part of the
cell. |

ACTION OF THE TRANSVERSE NETWORKS.

Similarly to the longitudinal bars the transverse networks
always remain straight in all conditions of contrac-
tion relaxation of the fibre. Hence they become necessa-
rily extended when the muscle-fibre contracts, and return to
their original form on relaxation of the fibre. The question
now remains as to whether the return of the transverse net-
works to their original position is due to active contrac-
tility or toelastic rebound, The following arguments, for
the first of which I am indebted to Mr. Melland, are in
favour of the latter view.

(a) An elastic thread, if stretched and then allowed to
rebound, will always return to its original length, i.e. will
always shorten to the same extent. The transverse networks
behave in this way ; they always shorten to the same extent,
viz. to the normal diameter of the fibre. This speaks in favour
of their being passively elastic, for if they were actively con-
tractile there is no reason why the fibre should not be com-
pressed to less than its normal diameter, elongation at the
same time taking place; whereas the fibre always relaxes to
the same extent.

(b) If the statements of Gerlach, Retzius, and Bremer are
correct, both parts of the network are connected with the end-
plate and with the axis cylinder of the nerve, the longitudinal
bars being connected indirectly through the transverse net-

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 101

works, the latter being in direct connection with the nerve.
It is therefore difficult to conceive that the transverse net-
works can contract actively after the longitudinal bars have
begun to relax, for the nervous impulse will apparently reach
the former first, and hence they must contract at the same time
as or before the longitudinal bars; and yet if the relaxation of
the fibre is held to be due to active contraction of the trans-
verse networks this is what must occur.

CoNCLUSION.

The conclusion to which I am therefore led is that the con-
traction of the striped muscle-fibre is due to the active con-
traction of the longitudinal bars of the network, and that the
transverse networks are probably passively elastic, and by their
rebound cause relaxation of the muscle-fibre. That the
transverse networks and the muscle-corpuscles, with which
they are said to be continuous, possibly furnish paths by which
the nervous impulse is conveyed from the nerve ending to the
longitudinal bars. That the contraction of the unstriped
muscle-fibre is due to the active contraction of its longitudinal
fibrils when these are present (as in vertebrate muscle). In
the case of unstriped muscle which possesses no fibrils the
contraction is due to the whole protoplasm of the cell, there
being no special part differentiated to perform this function.

Should these conclusions prove to be correct, we may
imagine the changes that occur in the striped muscle-fibre
during contraction to be as follows:

The nervous impulse reaching the end-plate of the nerve is
conducted by the transverse networks to the longitudinal bars,
and causes them to shorten; it does not cause the transverse
networks to contract, because they are passively elastic and
non-contractile. The longitudinal bars shorten according to
the strength of the nervous impulse, and remain so as long as
it lasts. By fluid pressure the transverse networks are ex-
tended and remain so as long as the longitudinal bars remain
contracted ; when these cease to contract the elasticity of the

102 C. F. MARSHALL.

transverse networks comes into play, and they shorten to their
original dimensions, and by fluid pressure extend the longi-
tudinal fibrils to their original length, the elastic sarcolemma
aiding in the process.

The alternate action of the longitudinal and transverse net-
works no doubt causes the special features of the contraction
of striped muscle, viz. the quick response to stimulus and
the rapid contraction; and we have seen that the network
is developed wherever rapid movements have to be per-
formed.

In connection with the foregoing considerations, the results
of Gerlach, Retzius, and Bremer, should they prove to be
correct, are of importance. I think there is little doubt that
the longitudinal striz described by Gerlach are identical with
the longitudinal bars of the network figured by Retzius,
Bremer, Melland, and myself. Gerlach traced these striz into
connection with the nerve endings. Retzius showed the con-
nection between the muscle-corpuscle and the transverse strie,
and Bremer traced the axis cylinder of the nerve into direct
continuity with the muscle-corpuscles. It therefore appears
that the network is connected with the nerve, and that the
longitudinal bars are connected with it indirectly through the
transverse networks. The direct continuity of the network with
the nerve does not necessarily imply that the network is itself
nervous; in fact, it really supports the view that it is the part
actively concerned in contraction; for we should expect a
priori, that if a differentiation occurred in muscle it would be
with the contractile part that the nerve would be in continuity.

On the other hand, with regard to the transverse networks,
it is possible that they may be in part nervous in nature, and
have for their function the more rapid conveyance of the
stimulus through the muscle ; and that the more rapid response
to stimulus, the special characteristic of striped muscle, may
be partly explained in this way.

There are two obvious objections to the theory of contrac-
tion we have arrived at, which I shall proceed to discuss:

1. It necessitates a difference between the longitudinal and

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSCLE. 103

transverse bars of the same network. This is an objection,
the real nature of which it is impossible to determine in the
present state of our knowledge of the nature and import of
intracellular networks in general. In unstriped muscle the
longitudinal fibrils are alone present, and in the development of
striped muscle the longitudinal elements of the network appear
first. The transverse networks are described and figured by
Retzius as direct processes of the muscle-corpuscles; the mode
of their development is as yet unknown, but should they prove
on further investigation to develop as processes of the
corpuscles, it would follow that the two elements of the
network are, in spite of their close connection in the adult, of
entirely independent and different origin. And then a differ-
ence of function would become not only possible but highly
probable. Further, the action of different reagents in split-
ting the fibre in different directions (alcohol, &c., causing
longitudinal, and acids transverse splitting) lends some
support to the same view. Haswell! in his observations on the
striped muscle of the gizzard of Syllis, states that after treat-
ment with hematoxylin, and then glacial acetic acid, the
transverse networks are stained, but not the longitudinal; he
says this may point to some differeuce in the substance of
which they are composed.

2. This theory attributes the function of contraction to the
network which forms much less of the buik of the fibre than
does the sarcous substance, the latter being far greater in
amount than the network. In reference to this it should be
borne in mind that contraction is not the only function per-
formed by muscle. The muscles, as stated by Dr. Michael
Foster,” are continually undergoing metabolism, giving rise to
a certain amount of heat; the metabolism during rest being
slow, but suddenly increasing during contraction. The energy
involved in the work done in a muscular contraction is only
about one tenth the total energy expended, the rest going out
as heat. Hence the muscles must be regarded as the chief

1 ¢ Quart. Journ. Micr. Sci.,’ 1886.
2 «Text-book of Physiology,’ 4th ed., p. 461.

104 C. F. MARSHALL.

sources of heat of the body, and are, “par excellence, the
thermogenic tissues.”

It thus appears that the thermogenic function of muscle
absorbs a far greater amount of its energy than does the con-
tractile function, and if we attribute the thermogenic func-
tion to the sarcous substance and the contractility to the net-
work, the above objection appears to receive a satisfactory
answer.

The following quotation from Prof. Michael Foster’ is
curiously in accordance with the view of the structure and
function of muscle maintained above, and may fitly conclude
this paper.

“Tt is quite open for us to imagine that in muscle, for
instance, there is a framework of more stable material, giving to
the muscular fibre its histological features, and undergoing a
comparatively slight and slow metabolism, while the energy
given out by muscle is supplied at the expense of more fluc-
tuating molecules, which fill up, so to speak, the interstices of
the more durable framework, and the metabolism of which
alone is large and rapid.”

SUMMARY.

1. In all muscles which have to perform rapid and frequent
movements, a certain portion of the muscle is differentiated to
perform the function of contraction, and this portion takes
on the form of avery regular and highly modified intracellular
neiwerk.

2. This network, by its regular arrangement, gives rise to
certain optical effects which cause the peculiar appearances of
striped muscle.

8. The contraction of the striped muscle-fibre is probably
caused by the active contraction of the longitudinal fibrils of
the intracellular network ; the transverse networks appear to
be passively elastic, aud by their elastic rebound cause the
muscle to rapidly resume its relaxed condition when the longi-

1 Dr, Michael Foster, loc. cit,, p. 475.

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSOLE. 105

tudinal fibrils have ceased to contract; they are possibly also
paths for the nervous impulse.

4. In some cases where muscle has been hitherto described
as striped, but gives no appearance of the network or treatment
with the gold and other methods, the apparent striation is due
to optical effects caused by a corrugated outline in the fibre.

5. In muscles which do not perform rapid movements, but
whose contraction is comparatively slow and peristaltic in
nature, this peculiar network is not developed. In most if not
all of the invertebrate unstriped muscle there does not appear
to be an intracellular network present in any form, but in the
vertebrate unstriped muscle a network is present in the form
of longitudinal fibrils only ; this possibly represents a form of
network intermediate between the typical irregular intra-
cellular network of other cells and the highly modified network
of striped muscle.

6. The cardiac muscle-cells contain a network similar to that
of ordinary striped muscle.

The investigations connected with this paper were partly
carried on in the laboratories of the Owens College and
partly at the Scottish Marine Station at Granton. I must
here express my thanks to my brother, Professor Milnes
Marshall, for his kindness in revising the paper, for much
advice in its production, and for obtaining the literature of
the subject; all the controversial points were discussed with
him and the preparations submitted to his examination. My
thanks are also due to Dr. Klein for kindly showing me his
preparations and for examining several of my own. I must
also thank Mr. J. T. Cunningham for the use of the Scottish
Marine Station, and for obtaining several of the animals.

106 O. F. MARSHALL.

DESCRIPTION OF PLATE VI,

Illustrating Mr. C. F. Marshall’s paper on “ Observations on
the Structure and Distribution of Striped and Unstriped
Muscle in the Animal Kingdom, and a Theory of Muscular
Contraction.”

[In all cases the gold chloride used was 1 per cent., the formic acid
25 per cent. The “usual gold method,’ when stated, means: 1 per cent.
acetic acid for a few seconds, gold chloride thirty minutes, formic acid twenty-
four hours in the dark.]

Fic. 1.—Ectoderm cell of Hydra: gold preparation. (1 per cent. acetic
acid for a few seconds, gold chloride thirty minutes, formic acid one hour,
exposed to sun.) a. Intracellular network. 4. Muscular process. ec. Intra-
nuclear network. -4, immersion obj.

Fic. 2.—Portion of muscular fibre of Medusa. Usual gold method.
qo Imm. obj.

Fic. 3.—Portion of muscle-cell of Earthworm, showing longitudinal rows
of dots. Usual gold method. 4; imm. obj.

Fic. 4.—Muscle-fibre from adductor of Pecten, showing network. Usual
gold method. 5 imm. obj.

Fic. 5.—Muscle-fibre from heart of Patella. (Acetic acid, 5 per cent.,
two minutes, gold thirty minutes, formic acid two hours, at 40°C.) Zeiss,
J obj.

Fic. 6.—Crayfish heart, Gold preparation of. (Acetic acid, 5 per cent., a
few seconds, gold twenty minutes, formic acid one hour, at 40°C.) +4 obj.

Fic. 7.—Muscle-fibre of Daphnia. (Acetic acid, 5 per cent., ten minutes,
gold thirty minutes, formic acid two hours, at 40° C.) 5 imm. obj.

Fie. 8.—Muscle of Bird (left in formic acid three days). 5 imm. obj. -

Fie. 9.—Muscle-cell from Rat’s heart, showing network. Usual gold
method. Zeiss, D obj.

Fic. 10.—Muscle-fibre from Frog’s heart. (Osmic acid, 1 per cent., half an
hour.) 5 imm. obj.

Fic. 11.—Network from heart muscle of Bird. (5 per cent. acetic acid
fifteen minutes, gold thirty minutes, formic acid twenty-four hours.) 3,
imm. obj.

Fic. 12.—Portion of unstriped muscle-cell from mesentery of newt, show-

OBSERVATIONS ON STRIPED AND UNSTRIPED MUSOLE. 107

ing intranuclear network and its connection with the fibrils of the cell.
(5 per cent. amm. chromate twenty-four hours; logwood.) 3; imm. obj.

Fie. 13.—Part of fibre from bladder of Salamander, showing fibrils. Usual
gold method. 4, imm. obj.

Fie. 14a@.—Muscle-fibre of Dytiscus, stimulated with alcohol vapour.
Portion a relaxed. 6. Contracted. 3, obj.

Fic. 14 4.—Network of relaxed portion. 4, imm. obj.
Fie, 14 ¢.—Network of contracted portion. 5 imm. obj.

Fie. 15.—Diagram of a muscle-fibre, showing change in network during
contraction.

Fie. 16.—Diagram of the intracellular network of striped muscle. a.

The transverse networks. 4. The longitudinal bars of the network. (Copied
from Melland, loc. cit., Diag. 1.)

Fic. 17.—Portion of network on a larger scale. (Copied from Melland,
loc. cit., Diag. 2.)

Fie. 18.—Portion of muscle-fibre of Dytiscus, showing the network very
plainly. One of the transverse networks is split off, and some of the longi-
tudinal bars are shown broken off. (Copied from Melland, loc. cit., fig. 6.)

Fie. 19.—Hypothetical diagram of the termination of nerve in muscle-
fibre and the connection with the network ; based on views discussed in the
paper. &S. Sheath of Schwann, continuous with sarcolemma.
cylinder branching and connected with muscle-corpuscles.
corpuscles connected with transverse networks.

nm. Axis-
m. Muscle-

C.F. Marshall del.

Fig 19.

ae See

Mor Bourn Vet, VU NS. ZY, Ve

47 Vinee eae ins
za

- , r
TF ety -
“Th tee }
+—b..¢.4} a °

mzaltae ;
fet ie es
4—folete etree cs
~ Tit ~4— © qt
rtp d 4 Tt
Port seoeeay
Seed ae
mee pales

ma fo peat
ee bo +t ee

i
ni pase
H44en yet]

Heecttt +
th beteers |

°o
°
=
LS
x

>

PIpT
Ss
Kt | |

~<

<=~,
—

ce
=!
S

X 1100.
X 1100.

-——---9-

et See (ey

F Huth, Lith? Edin®

b')

ON THE FATE OF THE MUSCLE*PLATE. 109

On the Fate of the Muscle-Plate, and the De-
velopment of the Spinal Nerves and Limb
Plexuses in Birds and Mammals.

By

A. M. Paterson, M.D.,
Senior Demonstrator of Anatomy, and Lecturer in Dental Anatomy and
Physiology, in the Owens College, Manchester.

With Plates VII and VIII.

Tue late Professor Balfour! showed that the spinal nerves in
Elasmobranchs spring entirely from epiblastic origins, and the
same has been proved conclusively regarding the roots at least
of the spinal nerves in birds and mammals, by the researches
of Milnes Marshall,” His,? and others. The most complete
account of the early stages in the development of the nerves
in higher Vertebrates is that of Marshall.? He has traced the
roots of the nerves from their origin from the spinal cord to
the point when they unite together to form the mixed nerve.
From that point onwards there is uncertainty. Though it is
considered highly probable that the further growth of the
nerves consists of an extension towards the periphery of the
original epiblastic elements, still it has not been proved that
this is so. It has not hitherto been shown that the nerve-
trunks, after the junction of the two roots, are not formed from
the cells of the mesoblast.

1 * Monograph on the Development of Elasmobranch Fishes,’ London, 1878,

2 «Journal of Anatomy and Physiology,’ vol. xi, p. 491.

3 * Ueber d. Anfange d. Peripherischen Nerven Systems,” ‘ Archiv f. Anat.
u. Phys.,’ 1879.

110 A. M. PATERSON.

The present investigation has been undertaken with the
object of tracing the nerves in their development from the
condition in which previous embryologists have left them, to a
point at which they may be fairly compared with the adult
state.

Chick embryos, artificially incubated, have been used for
the most part. By means of series of continuous sections, cut
in different directions, the nerves have been traced from end
to end in the different stages of development, from their first
appearance up to the formation of trunks having an arrange-
ment closely similar to that found in the adult. These sections
have been compared with sections of mammalian embryos of
different ages, with the result that the condition of develop-
ment of the nerves has been found to be identical in both at
periods in which the development of other parts and organs of
the body is the same.

The methods adopted were in almost all cases the same.
For hardening the Mammalian embryos, Kleinenberg’s solution
of picric acid was used; for the Chick embryos, a cold saturated
solution of corrosive sublimate. A solution of borax carmine
was the staining agent employed, prepared according to
Balfour’s directions. The sections were cut with a Jung’s
microtome, fitted with an ordinary razor.

The earliest stages in the development of the spinal nerves
in the Chick have been described by Marshall. He has shown
that they spring from the spinal cord as buds, of which the
dorsal are the first to appear, arising from the summit of the
cord. The more anterior of the dorsal roots arise from a
“ neural ridge,’’? an elevation continued back from the hind
brain. By the interstitial growth of the dorsal portion of the
spinal cord these roots become separated, and their attach-
ments to the cord more laterally placed. The ventral roots
appear at a later date. Projecting outwards directly, they
unite at an acute angle with the dorsal roots to form the
mixed nerve. The spinal ganglion on the dorsal root is
evident before the fusion of the two roots occurs ; it is formed
by the proliferation of the cells of the bud which form the

ON THE FATE OF THE MUSOLI-PLATE. 111

root. The spinal nerves are thus formed in pairs, which occupy
the intervals between the muscle-plates.

Before tracing the further growth of the nerves from this
point, it is necessary to describe the destination of the muscle-
plates and the mode of formation of the limbs, as in their
onward development the nerves present differences according
as they occur in relation to the limbs, or in the intervals
between them.

I. Fate or tHe Muscie-Piate.

The dorsal and ventral roots of the nerves unite towards the
end of the third day. But even before this time the limb buds
have begun to appear (at sixty hours). The growth and
differentiation of these buds, and the relations of the muscle-
plates in the different regions of the body, complicate the pro-
cess of nerve development.

In a Chick embryo, at the age of three days, when transverse
sections are made through the trunk between the limbs, the
muscle-plate (m. p., fig. 1) is seen as an elongated column of
cells lying directly beneath the epiblast, and separated from
the spinal cord by the spinal ganglion (sp. g.) and roots of the
nerve (N.). The lower end lies outside the angle (a.), between
the somatopleure and splanchnopleure. The nerve-roots alter-
nate with, and lie at a deeper level than, the muscle-plates.
The muscle-plate itself consists of a double layer of cells, con-
tinuous at the ends, and separated from each other by a very
evident line, the remains of the original cavity between the
two strata. The outer layer, whose thickness is made up of
several cells, consists of ovoid, spindle-shaped, or rounded cells,
fitting closely together, and with their long axes directed from
without inwards. They are sometimes multinucleated, and
stain deeply with carmine. Round the ends of the muscle-
plate they merge with the cells of the inner layer. The inner
layer of cells has different characters. As seen in longitudinal
sections, it is composed of spindle-shaped cells, which lie close
together with their long axes directed from before backwards,

112 A. M. PATERSON.

Several of these cells occur in one somite in a line from front
to back. In other words, the fibres are shorter than the
thickness of the somite. In transverse sections the fibres
appear rounded with large nuclei, and are more separated from
one another. The cells at the upper and lower ends of the
muscle-plate stain most deeply.

The bud which gives rise to the fore limb has at this
date attained considerable size. It projects almost directly
outwards from the side of the trunk (fig. 2), growing partly
from the mesoblast above the angle between somatopleure and
splanchnopleure, and partly from the somatopleure itself. It
consists of a mass of mesoblastic cells, densely packed together,
especially at the surface and distal end. Towards the origin
of the limb the cells become more scattered. This mass of
mesoblast is covered by a layer of epiblast, which presents two
thickenings where the cells are in most active growth, one
forming a cap covering the pointed outer end, the other forming
a ring round the root of the limb, and appearing in transverse
sections as thickenings above and below the root of the limb.
The mesoblastic cells are entirely undifferentiated as yet; they
are rounded, and stain deeply. Embryonic blood spaces are
found here and there. The position of the muscle-plate is
different from what has been described in the region between
the limbs. By the growth of the limb bud it has become
separated from the somato-splanchnopleuric angle (a.). Its
lower end reaches to about the centre of the upper half of the
base of the limb, that part which is continuous with the inter-
mediate cell mass. Moreover, the plate does not he external
to the angle of the body cavity: otherwise it has the same
relative position as in the dorsal region to the parts which
constitute the trunk at this period. Histologically, also, it
has very similar characters. The differences are seen in the
centre of the plate. The inner stratum of longitudinally
arranged spindle-cells is thicker. It stains badly, and the
nuclei of the fibres are large and well defined. The outer
layer is thinner in the centre, becoming thicker when traced
towards the end of the plate. The cells of this layer stain

ON THE FATE OF THE. MUSCLE-PLATE. 113

deeply. They are scattered in the centre, becoming more
closely packed at each end. .

In Chick embryos at three days six hours (fig. 3), in the
dorsal region the muscle-plate (m. p.) has extended a short
distance down the body wall in the somatopleure. The central
portion of the plate is thicker, owing to an accumulation of
the longitudinal fibres of the inner layer. The cells of the
outer stratum are still more separated in the centre, and do
not stain so deeply. The layer as a whole is thinner. The
ends of the plate still present the primitive condition of the
cells, which are angular, packed close together, and stained
deeply. In the region of the fore limb at this date (fig. 4) the
muscle-plate (m. p.) has the same characters ; but its position
is different. Its lower end reaches no farther than midway
between the dorsal attachment of the limb to the trunk, and
the somato-splanchnopleuric angle.

Twelve hours later (at three days eighteen hours) these
changes are seen to be more pronounced. In the dorsal
region (fig. 5), owing partly to the increase in vertical extent
of the embryo, the muscle-plate (m. p.) has become elongated,
and at the same time thinned out. There is no distinct trace
of an outer layer to be found, except at the ends of the plate.
Here the cells retain their primitive character, and stain
deeply. The rest of the plate consists entirely of longitudinally
arranged fusiform cells. The muscle-plate has a peculiar
bend, passing almost vertically downwards towards the body
cavity, and then suddenly sweeping outwards to enter the body
wall. Its lower end has passed still farther down the somato-
pleure, lying close to the inner side. In the fore-limb at this
date (fig. 6) the muscle-plate retains its original position. It
does not extend outwards farther than the somato-splanchno-
pleuric angle. Its histological characters are the same as in
the dorsal region. The limb bud itself has increased in size,
and is now directed downwards. The cells which compose it
are still undifferentiated.

In embryos at four days, in the regions of the trunk between
the limbs (fig. 7), the muscle-plate (m.p.) has extended down

VOL, XXVIII, PART 1,—NEW SER, H

e

114 A. Me PATERSON.

through a third of the length of the body wall, lying close to
the outer side of the body cavity. Its relations and structure
are the same as before. Each end is surmounted by a cap of
cells, which retain their primitive characters ; rounded, fusi-
form, or angular, they stain deeply, and are plainly separated
off from the main part of the muscle-plate. These cells
can be traced on to the outer surface of the muscle-plate,
where they gradually become lost. The main part of the plate
consists of elongated fusiform cells. In the region of the fore
limbs (fig. 8) the mesoblastic tissue of the limb still presents
the same characters. The cells stain deeply, are round, ovoid,
and often multinucleated; but still undifferentiated. The
foetal vessels are better marked. The muscle-plate (m. p.)
occupies its primitive position, ending below at the root of the
limb. It has the same structural characters as in the dorsal
region ; but the undifferentiated cells, which stain deeply, are
best marked at the upper end of the plate.

In embryos at four days twelve hours, in the trunk between
the limbs, the muscle-plate (fig. 11, m.p.) has passed half way
down the body wall, lying close outside the cavity. It now
consists almost entirely of elongated spindle-cells. In the
regions of the limbs further changes are taking place, but in
which the muscle-plate takes no part. It remains within the
body cavity, and ends below at the same limit as before (fig.
12, m.p.). Structurally it is the same. In the limb buds
themselves the first changes occur at this date, in the formative
blastema, which result in the production of the muscular and
osseous systems. The nerve plexuses, as we shall see below,
have been produced ; and the resulting trunks have passed into
the limb, in two groups, one dorsal (d.) the other ventral (v.).
Between, above, and below the nerves, the mesoblastic cells are
taking on a characteristic arrangement. The cells immediately
above and below each trunk (1, 2) are more closely packed
together, forming thick layers, each several cells deep. They
are histologically the same as before. In the centre of the
limb bud (38), between the nerve trunks, the cells are now
arranged in a concentric and symmetrical fashion, and are

ON THE FATE OF THE MUSCLE-PLATE. 115

separated from one another by a small amount of intercellular
substance. Towards the dorsal and ventral surfaces of the
limbs the cells are more scattered, so that the central portion
of the limb in transverse sections appears darker than the
superficial parts. At the free end of the limb (4) there is still
no differentiation of tissue elements. The cells here form a
simple mass, without any distinction into layers or groups.

In a five days’ embryo the condition of the muscle-plate in
the region between the limbs (fig. 13, m. p.) is much the same
as at the last-mentioned period. It shows still further exten-
sion in a ventral direction down the body wall. In the same
embryo, in the regions of the limbs the muscle-plate (fig. 14,
m.p.) has clearly no connection with the muscular system of
the limb itself. It consists now of elongated fibres, forming,
in transverse sections, a column lying just outside the spinal
cord and nerves, and separated from the surface of the body
and from the limbs by a considerable thickness of ordinary
blastema. The several tissues of the limb are formed from
mesoblastic elements, developed in situ. A central core of
cellular cartilage (3) is very evident at this date. The cells
are arranged regularly in a concentric manner in transverse
sections; in transverse rows in longitudinal sections. This
cartilaginous cylinder is found in longitudinal sections to be
broken up into segments, corresponding to the skeletal ele-
ments of the limb. The intercellular substance between the
cells has largely increased in amount. At the periphery of
this central mass the cells gradually become changed in cha-
racter, being more deeply stained in mass, and individually
becoming oatshaped or fusiform. The long axes of the cells
are directed from within outwards. Two layers of cells thus
appear, one above, the other below the cartilaginous bar. The
cartilage, however, is not yet distinctly marked off, but is con-
nected to these groups of cells by a definite and still more
deeply-stained (perichondrial) layer. The nerves to the limbs
derived from the plexuses have a very definite relation to these
central groups of cells, which are enclosed within the nerve-
trunks. The dorsal nerves pass over, and the ventral trunks

116 A. M. PATERSON,

beneath, this central area. Above and below the nerves two
other more distinct groups of ovoid cells (1 and 2) are now
apparent, collected each into more or less separate subsidiary
bundles, and easily distinguished from the surrounding undif-
ferentiated mesoblast by their shape, and by the fact that,
being closely packed together, they stain more deeply en
masse. The four simple, unsegmented strata of ovoid cells—
of which two are dorsal, and lie above and below the dorsal
branches of the nerves; two ventral, and having a similar
relation to the ventral trunks—are evidently the precursors of
the muscular elements of the limbs. They are quite distinct
from the muscle-plates, and are separated from them by a
large quantity of undifferentiated mesoblast. The blood-vessels
of the limb have now attained a large size, and are regular in
their arrangement. One large artery (a7t.) passes down the
centre of the limb on its ventral aspect, lymg among the
nerves, and accompanied by a vein (V.).

It is unnecessary to follow the muscle-plate further. It has
been shown that while it grows into the body wall between the
limbs, and forms the basis of the longitudinal muscles of the
trunk ; in the region of the limbs, it remains in its primitive
position and has no share in the formation of the limb-muscles.
It is merely concerned here in forming the longitudinal
muscles of the back. In the limbs themselves the muscles
are produced by the further differentiation of the four dorsal
and ventral strata, which have been described as appearing
from the mesoblast-cells, at first undifferentiated, and forming
the original limb bud. In Chick embryos at six days, these
strata of ovoid cells have become more fusiform, and are col-
lected into more definite and separate systems. Two days
later the muscles throughout the body are seen distinctly.

Il. DEVELOPMENT OF THE SPINAL NERVES AND Limp
PLEXUSES.

The later development of the spinal nerves naturally divides
itself into two parts: firstly, in relation to the limbs; and
secondly, in the trunk between the limbs. In essential points

a

aaah

ON THE FATE OF THE MUSCLE-PLATE. 117

the processes are the same in all regions of the body. The
formation of the limbs, however, and the peculiarities in the
position of the muscle-plates give rise to certain differentiations
in the arrangement of the nerves in those regions.

In Chick embryos at three days (figs. 1 and 2), both in the
trunk between the limbs and in the regions where the limbs
are being formed, the nerves have reached the same stage of
development. The nerve-roots, which lie within and alternate
with the muscle-plate, have joined together. From their fusion
a slender, finger-like process of cells results (N.), which repre-
sents the commencing nerve-trunk. The dorsal root is oval
in transverse section, the ganglion, which is very large, form-
ing nearly the whole of it. The nerves and their roots consist
of large ovoid cells, containing often two or three nuclei, the
long axes of the cells being directed outwards from the cord.
They stain more deeply than the mesoblast cells in which they
lie, and are surrounded by a slight amount of feebly-stained
intercellular substance.

Six hours later (at three days six hours), the slender stalk
(N., figs. 3 and 4), retaining the same position within the
muscle plate, has grown downwards and outwards as far as the
somato-splanchnopleuric angle (a). It has the same relative
position in the trunk and in the regions of the limbs, passing
between the muscle-plate and cardinal vein (¢.v.), But, owing to
the difference of growth of the muscle-plates in the two regions,
it has reached its lower end in the regions of the limbs (fig. 4) ;
while in the trunk, the muscle-plate, having by this time
entered the body wall (fig. 3), extends farther than the nerve.

This description corresponds with the condition of develop
ment of the spinal nerves and muscle-plates in the Rabbit
embryo of seven or eight days. Both histologically and mor-
phologically the nerves have reached the same state of
development.

In Chick embryos of three days eighteen hours there is not
much difference in the relative amount of the growth of the
nerves. The whole embryo has, however, increased in size. In
the trunk (fig. 5) the nerves cannot yet be traced into the

118 A. M. PATERSON.

body wall. In the region of the limbs (N., fig. 6) they have
passed out beyond the lower end of the muscle-plate and beyond
the angle of the body cavity. The marked change, however,
at this date is in the histological structure of the nerves. The
spinal ganglia are well-formed ovoid masses, composed of large
ovoid cells with two nuclei, the cells having a general arrange-
ment in vertical rows. The cells forming the nerve-trunks
have become elongated, fusiform, with fibrillar processes at
each end. The body of the cell does not stain well; the
nucleus, large, oval, and with several nucleoli, lies in the centre
of the cell, and stains deeply ; the distal ends of the nerves in
the regions of the limbs present a ragged appearance, due to
the protrusion and separation of these spindle-shaped cells into
the mesoblastic tissue.

At four days the histological change is more marked. The
cells forming the nerve-trunks have become more fibrous, the
nuclei are less numerous, and the trunks stain yellow en
masse. Both between the limbs and in relation to the limb
buds the growth of the nerves has continued. In relation to
the limbs (N., fig. 8), the nerves sweep round between the lower
ends of the muscle-plates and the body cavity, and, reaching
the base of the limb, spread out, and then divide into a sheaf
of branches, which diverge in the formative blastema of the
limb. In the trunk between the limbs the nerves (X., fig. 7)
have extended a great way down the body wall, lying between
the muscle-plate and the body cavity, but not reaching as far
as the lower end of the plate. They divide, as in the limb, into
branching processes, some of which pass directly outwards into
the muscle-plate and divide again; some pass on, lying within
the muscle-plate.

It is at this period that I have first been able to make out
satisfactorily the existence of the trunk passing to join the
sympathetic. A slender cord arises from the spinal nerve mid-
way between the junction of the roots and the distal end. It
courses inwards at right angles to the main trunk, and is soon
lost. Now also the formation of the superior primary division
of the nerve is first seen distinctly. It is constructed in the

ON THE FATE OF THE MUSCLE-PLATE. 119

same way in mammals, and is seen still more clearly in Rat
embryos at fifteen to seventeen days (fig. 16). Each root of
the nerves divides into two unequal branches—the dorsal root
beyond the ganglion, the ventral root directly. Of these
branches the smaller is superior, the larger inferior in both
cases. The larger branches unite to form the main trunk of
the nerve, or the inferior primary division; the smaller
branches combine to form the superior primary division. This
is directed upwards and outwards, and subdivides as it passes
towards the surface.

In Chicks at four days six hours, the condition of the nerves
in the trunk between the limbs is slightly more advanced, but
presents no change of any note. In the regions of the limbs,
however, the plexuses are now formed. In transverse sections
through the embryo, the nerves are found, on entering the
limbs, to divide into two fairly well-defined strands, separated
by a central mass of mesoblast, which is in active growth, and
preparing to form the cartilaginous basis of the limb. The
nerves, in fact, spread out around this central core, and arrange
themselves into two sets, one dorsal the other ventral. These
dorsal and ventral branches of the nerves only pass a short
distance into the limbs, and are not so well defined as in em
bryos a few hours older. But even now the process of plexus
formation may be seen.

When longitudinal vertical (sagittal) sections are made con-
tinuously through the body, it is seen that the nerves to the
limbs, besides forming the dorsal and ventral branches above
mentioned, unite with adjacent nerves at the root of the limb
to form a well-defined plexus. In the Fowl three main trunks
form the brachial plexus,'—the first thoracic and the last two
cervical nerves, with in addition a small branch from the more
anterior cervical nerve. When sagittal sections are made, the
limbs being divided transversely at their roots (fig. 9, a), the
axillary artery (art.) and vein (v.), with the three main nerves
(N., 1, 2, 8), are divided just outside the body cavity (B. C.),
and as they lie in the body wall (B. W.), before their entrance

1 Macartney, Art. “ Birds,” ‘ Rees’s Cyclopedia,’

120 A, M. PATERSON.

into the limb bud, and below the terminations of the muscle-
plates. In successive sections, from within outwards, these
nerves can be traced to their terminations in the limbs. They
first spread out, and approach one another, as described ; in
doing so they unite with adjacent nerves, so that the next step
in the proceedings (fig. 9, 6) is the formation of a plexiform
mass of nerve-tissue (plex.), which encircles the artery. At
the same time that this plexus formation occurs the division
into dorsal and ventral branches is beginning. In the figure
the axillary artery in the centre, with a group of mesoblast
cells running up and down from it, shows the commencing
separation of the mass into these two portions. In sections
made a little farther outwards, the plexus gradually separates
completely (fig. 9, c) into a dorsal and a ventral mass (d. and
v.), each consisting of a broad, flattened band, separated by
the artery and some mesoblastic cells. Still farther out, just
before the limb is completely separated from the trunk
(fig. 9, d), the nerves appear as two distinct cords (d. and v.).
These gradually divide, become attenuated, and disappear as
they are traced towards the distal portion ofthe limb. Exactly
the same process occurs in relation to the nerves of the hind
limb. When oblique longitudinal sections are made at this
date, so as to cut through the length of the nerves (fig. 10),
they are seen to spread out and divide laterally, so as to unite
with similar branches (dorsal or ventral) of adjacent nerves, to
form the plexus at the root of the limb. As already stated,
there is no trace whatever at this date of either cartilage or
muscle in the limb, which consists entirely of undifferentiated
blastema.

In embryos, four days twelve hours old, the nerves have
reached a more advanced stage of development. In the trunk
between the limbs (fig. 11, NV.) the nerve, which twelve hours
earlier divided into a sheaf of branches in the body wall, now
splits into two well-defined and unequal branches at a point
just beyond the somato-splanchnopleuric angle. The larger
branch continues the direction of the main nerve, and can be
traced for some little distance between the muscle-plate and

‘om

7 ee 2

ON THE FATE OF THE MUSOLE=-PLATE. 121

body cavity. The smaller nerve represents the lateral branch
of the adult, and is directed downwards and outwards through
the muscle-plate, on the outer side of which it divides and is
finally lost. The cord to the sympathetic nerve can be followed
farther than before ; but I have been unable to trace its con-
nections with the roots and trunk of the spinal nerves.

In the limbs the early changes occurring in the blastema,
which lead to the production of the osseous and muscular
systems, have already been described. The nerves are of large
size, aud can be traced more than halfway through the limb
towards the distal end. At the root of the limb the main
trunk (fig. 12, N.) can be seen in transverse sections to divide
into two well-defined trunks, which are clearly homologous
with the terminal branches of the nerve in the region of the
trunk at this date (fig. 11). These two large nerves are re-
spectively dorsal (d) and ventral (v); and enclose between
them the densest portion of the blastema (38). This enclosed
portion has already been described as consisting of three parts
—a central part, which is going to form cartilage, and a dorsal
and ventral part, the elements of muscular tissue. On the
dorsal surface of the dorsal nerve, and on the ventral surface
of the ventral nerve, are other layers of mesoblast cells (1 and
2) undergoing division preparatory to the production of muscles.
In longitudinal sections the three main nerves supplying the
fore limb can be traced as before in successive sections, each
dividing into dorsal and ventral branches; and these again
uniting with adjacent nerves to form two flattened bands,
which pass to the dorsal and ventral surfaces of the limb.

The nerve-trunks are almost entirely fibrous now, with rows
of deeply-stained nuclei arranged alone and among the wavy
fibres. These are evidently the connective-tissue elements of
the nerve-trunks. Towards their terminations the fibres are
fewer and the fusiform nerve-cells more abundant.

In Rat embryos, between twelve and fourteen days old,
exactly the same condition of development of the nerves is
found as in the Chick at four days twelve hours. The state of
development of the body generally is the same, the limbs exist

122 A. M. PATERSON.

as buds projecting downwards and outwards from the trunk,
and composed, for the most part, of undifferentiated blastema.
In the centre of this the cells are more closely massed together
than at the periphery, and are being arranged concentrically to
form the cartilaginous basis of the limb. Lach of the roots of
the nerve divides into upper and lower branches, which respec-
tively unite; the upper branches to form the superior primary
division, the lower to form the inferior primary division of the
nerve. The latter is the main trunk. Passing downwards and
outwards below the muscle-plates, it reaches the base of the
limb where it divides into two branches, one dorsal the other
ventral, with regard to the cartilaginous core. These branches
can be followed through the limb, actually as far as the epi-
blastic surfaces and almost to the distal end. In minute
structure the trunks very much resemble the nerves in the
Chick, the chief difference being that the fusiform cell elements
are more evident throughout.

In Chick embryos, at five days, the changes in the nerves
between the limbs are not marked (fig. 13, N.). The lateral
and inferior branches are well defined, and the whole nerve
has passed farther down the body wall. From this time on-
wards these trunk nerves present no marked differences in
morphological arrangement from what is found in the adult.

In the regions of the limbs at this date, as already described,
the cartilaginous basis and muscular elements have begun to
make their appearance. The nerves themselves occupy a
position with regard to these elements which is highly charac-
teristic. The dorsal and ventral trunks (fig. 14, d. and v.) are
each covered above and below by masses of specialised, oat-
shaped cells, which represent the layers of dorsal and ventral
muscles. These double dorsal and ventral muscular layers
are also separated by the cartilaginous framework of the limb.
The nerves themselves stain yellow; consist of extremely wavy
fibres, and present no distinct nuclei. Deeply-stained connec-
tive-tissue corpuscles lie among the fibres.

At five days twelve hours the process of muscle and cartilage
formation in the limbs is more advanced. The appearance of

ON THE FATE OF THE MUSOCLE-PLATE. 123

the nerves in transverse section is much the same as before.
In successive longitudival (sagittal) sections (fig. 15) the
nerves can be seen at the root of the fore limb, undergoing
division and union in the brachial plexus in the same way as,
but more definitely than in younger embryos (four days six
hours, fig. 9). The three main trunks are seen first (fig. 15, a,
N., 1, 2,3) in company with the axillary artery and vein. They
then divide, in successive sections (figs. 15, 6B—15, e), into
dorsal (d.) and ventral (v.) branches. The dorsal branches unite
with dorsal branches, the ventral branches with ventral branches,
to form the nerves of distribution to the limbs. The plexus-
formation is now complete, and from this time onwards there
is no change in the essentials of its formation, which tally with
the condition of the adult brachial plexus. It is to be borne
in mind that, though the plexus is completely formed, yet the
muscular elements are in a simple condition. Muscles are not
formed in anything approaching the complexity of the adult
limbs sooner than the ninth day.

III. Concivusions.

1. On the fate of the muscle-plate and development of the
muscles of the limbs.

In Elasmobranchs the muscles of the limbs are formed by
the muscle-plates which grow down, and as they pass the roots
of the limbs give off small buds, which become separated off
and form the starting point of the muscle formation in the
limbs.! In higher Vertebrates, while it has been held probable?
that the muscle-plates do not enter into the formation of these
muscles, it has not been shown satisfactorily how they do arise
and what becomes of the muscle-plates. It is evident from the
foregoing details that the muscle-plates, in the regions of the
limbs, stop short in their downward growth, do not pass farther
than the base of the limb, and are not concerned in any way

' Balfour, ‘A Monograph on the Development of Elasmobranch Fishes,’

London, 1878.
2 Kolliker, ‘ Entwicklungsgeschichte d. Menschen u. der hdheren Thiere,’

Leipzig, 1879.

124 A. M. PATERSON.

with the production of the limb muscles. These are formed
by the differentiation of the mesoblastic cells forming the
primitive limb buds. These cells form, in the first place, a
central cartilaginous bar, above and below which, in the second
place, are developed a double dorsal and a double ventral layer
of simple muscle, which later on becomes more complex in
its arrangement and forms the muscles of the adult.

In the region of the trunk, between the limbs, a different
disposition of the muscle-plates occurs. They grow down in
the body wall and eventually become converted into the longi-
tudinal muscles of the trunk. They do not, however, appear
to assist in the formation of the sub-vertebral (hyposkeletal)
muscles.

In both regions the growth and differentiation of the parts
of the muscle-plates are alike. The outer layer disappears
eradually ; the inner layer of the plate being converted into
the longitudinal fibres. The disappearance of the outer layer
is possibly due to the conversion of the cells into longitudinal
fibres, which merge with those of the inner layer; but this is
not certain. In Elasmobranchs each fusiform cell extends
from end to end of the muscle-plate In birds and mammals
this is not so. Each fibre is considerably shorter than the
breadth of the somite.

The chief point of interest here is in connection with the
development of the limb muscles. They first appear as double
layers of dorsal and ventral cells, which layers are simple,
without segmentation, and derived from the mesoblast cells of
the primitive limb bud. In Elasmobranchs* this stage in the
development of the limb muscles is a secondary one, and is
preceded by events which are omitted in higher Vertebrates.
The process of evolution of the limbs in birds and mammals is
therefore shortened. In Elasmobranchs a downward growth

and a cutting off of part of certain muscle-plates occurs; the
portions cut off undergo further growth, passing into the limb
bud, fusing together, and becoming differentiated into dorsal

1 Balfour, ‘Comparative Embryology,’ p. 552.
2 Balfour, ‘ Monograph on Elasmobranchs.’

— a

eso ee a eee Yee ee

j

ON THE FATE OF THE MUSCLE-PLATE. 125

and ventral strata. In birds and mammals the same end is
reached without these preliminary steps, and witlout the
intervention of the muscle-plates. The definite relations
which these simple muscular layers bear to the nerves of the
limbs, throw light on the evolution of the limb plexuses. Each
nerve, passing into a particular region of the limb bud, divides
into dorsal and ventral branches, to supply the dorsal and
ventral surfaces respectively of that particular portion. As
the mesoblast forming the limb bud becomes more differen-
tiated so as to give rise to the muscular layers, the portions
opposite to, and originally derived from, the same somites as the
nerves become fused, forming simple muscular layers in the
first place. The nerves therefore fuse together; the dorsal
branches forming a dorsal band, and the ventral branches a
ventral band, which pass out, and are finally lost in these
simple muscular layers.

2. On the growth and development of the spinal nerves.

It is difficult to demonstrate clearly, but it is next to impos-
sible to deny, that the spinal nerves are developed from
epiblast throughout their whole length. From the numerous
sections which I have examined at different periods of growth,
I have traced the spinal nerves, not only the nerve-roots, but
also the trunks and the plexuses, as a centrifugal growth from
the spinal cord. The growth of the nerves is both interstitial
and terminal. Consisting at first merely of rounded cells, in
an active state of proliferation ; in older embryos these become
first ovoid and then fusiform, at the same time being less
deeply stained with borax carmine. These fusiform cells, by
the alteration of their protoplasm, become converted into
nerve-fibres. Moreover, while this interstitial growth goes on,
the trunk of the nerve is elongated by means of proliferation
of the cells at the periphery, which retain a primitive cha-
racter longer than those in the more proximal portion of the
trunk. For example, when the cells are fusiform in the nerve
near the cord, they are oatshaped at the distal end; when
they are fusiform at the distal end of the nerve, they are fibrous
in the proximal part of the trunk.

126 A. M. PATERSON.

8. On the homologies of the spinal nerves.

Their development shows that the nerves which form the
limb plexuses are homologous with the whole nerves in
the regions between the limbs, where their arrangement is
simplest, and not merely with the lateral branch, as Goodsir
supposed.! The nerves in both regions first spread out into
a ragged bundle. These bundles at a later period arrange
themselves into two well-defined cords, the division from the
main trunk having the same relative position in both. In the
regions between the limbs these trunks represent the lateral
and inferior branches; in the regions of the limbs they are
dorsal and ventral in position.

4, On the development of the limb plexuses.

I have elsewhere shown? that in mammals, and as far as 1
have been able to make out, the same holds good for birds also,
the limb plexus are formed on a definite plan, which is essen-
tially the same in all the animals examined, and in relation to
both fore and hind limbs. The nerves which form the plexus
divide, in the first place, into dorsal and ventral branches.
These divisions subdivide, and the secondary cords, whether
dorsal or ventral, combine with the cords formed by the
division of adjacent (dorsal or ventral) trunks to form the
nerves of distribution. Any given nerve to the limb may be
derived from any number of the spinal nerves constituting the
plexus, but it is always formed by a combination of either
dorsal or ventral nerves.

This mode of arrangement of the nerves in the plexuses is to
be explained by a reference to their embryology, and the mode
of development of the different parts of the limbs. The plexus
formation is complete, and the nerves of distribution are
formed in the embryonic limb long before the appearance of
muscles. In the development of the nerves in the limbs the

1 «Hdinburgh New Philosophical Journal,’ New Series, vol. v, Jan., 1857;
‘ Anatomical Memoirs,’ vol. 2, p. 201, 1868.

2 Graduation Thesis, Univ. Edin., 1886, ‘On the Spinal Nervous System in
Mammalia ;’ ‘‘The Limb Plexuses of Mammals,” ‘Journal of Anatomy and
Physiology,’ vol. xxi, 1887, p. 611.

es

ON THE FATE OF THE MUSCLE-PLATE. 127

following steps occur. The primitive nerve, in the first place,
grows out beyond the lower end of the muscle-plate, and
reaches the root of the limb. It there, secondly, spreads out
into an irregular series of processes, which pass into the undif-
ferentiated tissue of the limb. Thirdly, these branches at a later
date arrange themselves in two trunks, one dorsal, the other
ventral, which extend still farther into the limb and enclose
between them a mass of blastema, from which the cartilaginous
basis of the limb is formed. Fourthly, the dorsal and ventral
trunks fuse with adjacent dorsal and ventral trunks to form
two broad flat bands, from which, still later, the individual
nerves as found in the adult are produced.

The development of the muscular system of the limb,
occurring after the formation of the nerves, corresponds with it
exactly. The muscles appear first as simple double dorsal and
ventral layers, among which the nerves pass as dorsal and
ventral bands, formed by the fusion of adjacent dorsal or
ventral divisions of the nerves of origin. As these muscular
strata lose their simplicity and take on the complex arrange-
ment of the adult, the nerves at the same time become more
and more subdivided, until in the adult the primitive characters
of both are considerably masked.

Still, in the adult mammal, it is evident that the more pre-
axial nerves in the series supply the more preaxial portions,
the postaxial nerves the postaxial portions of the limb,! and
the combinations of dorsal divisions and ventral divisions of
the nerves are distributed to those muscular and cutaneous
areas which are derived respectively from the primitive dorsal
and ventral surfaces of the embryonic limb bud.”

In conclusion, I wish to express my deep indebtedness to
Professor Milnes Marshall, of Manchester, for much advice
and encouragement during the prosecution of the above re-
searches, and for his kind assistance in the preparation of the
present memoir.

1 Herringham, “On the Human Brachial Plexus,” ‘ Proc. Roy. Soe.,’ Jan.,
1887.

* «Journal of Anatomy and Physiology,’ vol. xxi, July, 1887.

*

128 J A. M,. PATERSON.

EXPLANATION OF PLATES VII & VIII,

Illustrating Dr. Paterson’s paper “On the Fate of the
Muscle-Plate, and the Development of the Spinal Nerves
and Limb Plexuses in Birds and Mammals.”

Fie. 1.—Semi-diagrammatic view of transverse section through the trunk
of a Chick embryo at the end of the third day. Both spinal nerve (J.), with
its roots and ganglion (Sp. g.) and muscle-plate (m.p.) are shown. The spinal
cord, notochord (Wo.), aorta (4o.), and cardinal vein (C. V.), are also indicated.
The muscle-plate is just entering the body wall. The section is taken between
the limbs.

Fie. 2.—View showing same structures in an embryo of three days in the
region of the fore limb.

Fie. 3.—From an embryo of three days six hours old, showing the growth
of the muscle-plate (m.p.) and spinal nerve (1V.) in the trunk between the
limbs.

Fic. 4.—From the same embryo in the region of the fore limb.

Fic. 5.—From an embryo aged three days fifteen hours, showing the further
growth of the muscle-plate (m. p.) and nerve (JV.) in the trunk between the
limbs.

Fic. 6.—Shows the same structures in the region of the fore limb.

Fic. 7.—From an embryo aged four days, with the muscle-plate presenting
growing points at the two ends, and the nerve dividing in the body wall int
a ragged bundle. The formation of the superior primary division (s.) of the
nerve aud the cord to join the sympathetic (sy.) are also seen,

Fic. 8.—From the same embryo, through the fore limb, showing the relative
position of the muscle-plate and nerve. The nerve is seen dividing in the limb
into a sheaf of branches.

Fie 9, a.—Longitudinal section through the body of a Chick embryo aged
four days six hours in the region of the fore limb, cutting through the body
wall (B. W.) just below the level of the muscle-plate. The body cavity (B. C.)
is seen, and in the body wall are the three nerves (1. 1, 2, 3), and the artery
(art.) going to the limb. %.—This section is made farther out at the root of
the limb bud, and shows the thickening of the body wall, with the formation
of the plexus, around the artery (avt.). The vein (v.) is also seen, The
nerves in the plexus are here on the point of separating into two bundles,
c.—From a section made stili farther out in the limb. The nerves, after
forming the plexus, have more completely separated into a dorsal and a ventral
bundle (d, and v.), with the artery in the middle. d.—Here the limb bud is

ON THE FATE OF THE MUSCLE-PLATE. 129

more apparent ; it is becoming separated from the body wall; and the dorsal
and ventral bands of nerve are well defined in the blastema.

Fie. 10.—From a longitudinal section through the body of the same embryo,
in the region of the hind limb (Z. Z.), to show the lateral division and union
of the dorsal and ventral divisions of the nerves to form the plexus (plez.).
The spinal cord (Sp. C.) and spinal ganglia (Sp. g.) are also shown.

Fie. 11.—Transverse section through trunk of Chick embryo aged four days
twelve hours, between the limbs, to show the muscle-plate and the spinal
nerve. ‘The division of the latter into its two terminal branches ((. 7.) is seen.

Fic. 12.—Transverse section through the same embryo in the region of
the fore limb. The spinal nerve is seen dividing at the root of the limb into
dorsal (d.) and ventral (v.) branches. These enclose and are surrounded by
masses of formative blastema (1, 2, 3), the precursors of the muscles and
skeleton of the limb. The muscle-plate (m. p.) occupies its original position.

Fic. 18.—Transverse section through the trunk (between the limbs) of an
embryo aged five days, to show the relative position and growth of the
muscle-plate (m. p.) and nerve (J.).

Fic. 14.—Transverse section through the trunk. and root of the fore
limb of an embryo of the same age. m. yp. Muscle-plate. W. Spinal nerve.
d. and v. Dorsal and ventral divisions. 3. Commencing formation of carti-
lage in centre of limb, with a layer of dense blastema surrounding it. 2 and
1. Commencing formation of muscles above and below the nerve-trunks.
art, Axillary artery. v. Axillary vein.

Fig. 15.—Successive longitudinal sections through the root of the fore
limb of a Chick embryo aged five days twelve hours. The defined borders
represent the body cavity. From a to e the nerves (WV. 1,2, 3) are seen to
divide into dorsal and ventral branches (d. v.), which unite laterally to pro-
duce secondary dorsal and ventral trunks. 4r¢. and v. Axillary artery and
vein.

Fic. 16.—Diagram to illustrate the formation of the superior (§.) and
inferior (I.) primary divisions of a spinal nerve from a Rat embryo. Sp. C.
Spinal cord. Sp. G. Spinal ganglion. No. Notochord. 4o. Aorta.

VOL, XXVIII, PART ]1,—NEW SER. I

A.M. Paterson del.

7A © »
“ Be) 9 6

2)
eo oO
9 © 900
0 @e 99

©

Moor bourn! VL IVUNS PY.

e
®eo
\® @G0e,

Dy
2 es008

jo
[2 ©

F Huth, Lith? Edin?

oo

e

P~ 0% °°
© ego
o 32

ga 9

25 8

eq °

at ©

: aes
BSAdw" 9 ¢

PD FALARODY lhe

BS © NY

Or%5

%

A.M Paterson del.

F Huth, Lith” Edin?

:
ea
x
5
‘

)
“] Pores

® is 5

NOTE ON THE CILIATED PIT OF ASCIDIANS. 131

Note on the Ciliated Pit of Ascidians and its
Relation to the Nerve-ganglion and so-called
Hypophysial Gland; and an Account of the
Anatomy of Cynthia rustica (?).

By

Lilian Sheldon,
Bathurst Student, Newnham College, Cambridge.

With Plates IX and X.

In the adult Ascidians which I have examined, I find tour
main variations of the ciliated pit.

(1) In Clavellina the ciliated pit (fig. 1, C.P.) is simple
in shape, its opening into the mouth being round in section.
It lies ventral to the nerve-ganglion (N.G.), into the solid sub-
stance of which it leads by a wide opening (B.), situated near
the anterior end of the ganglion. Behind the opening it
narrows, and passes on into a canal (C.), lying immediately
ventral to the ganglion. The cells lining this canal are flatter
than those at its orifice, and are not ciliated. A large number
of glandular tubes (G/.) lie ventral to it and open into it.
Seeliger’ states that the hypophysial gland never attains to
the complicated condition, which Julin®? describes in some
simple Ascidians. I find, however, that the gland is large and
made up of branching tubes which open into the backward

1 Seeliger, O, “Die Entwicklungsgeschichte de socialen Ascidien,” ‘Je-
naische Zeitschrift fir Naturgewissenschaft,’ 1885.

2 Julin, Charles, ‘ Recherches sur lorganisation des Ascidies simples, sur
Phypophyse et quelques organs qui s’y rattachent,” ‘Archives de Biologie,’
Tome ii, 1881.

132 LILIAN SHELDON.

prolongation (C.) of the ciliated pit in a way precisely similar
to that described by Julin.

(2) In Amarecium proliferum the ciliated pit (fig. 2,
C.P.) is shorter and simpler than in Clavellina. It consists of
a funnel, communicating by a circular opening with the
mouth, and is lined throughout by ciliated columnar cells. It
lies immediately ventral to the ganglion, with which it has
no communication. At its apex it opens (P.) into a mass of
spongy tissue (Zi.) lying ventral to it, which has a definite
boundary, but is not glandular in structure, appearing rather
to consist of degenerated tissue, somewhat resembling the
notochordal tissue of Vertebrate embryos.

(3) Ascidia and Ciona.—I have examined several species
belonging to these genera, and have found that the condition
is similar to that described by_Julin’ in several species of
Corella, Phallusia, and Ascidia. The pit consists of a ciliated
funnel passing into a canal. A mass of glandular tissue lies
ventral to the canal and opens into it by a number of ducts.
The opening into the mouth is sometimes circular, but more
often horseshoe-shaped.

(4) The condition in Phallusia mammillata has also
been described by Julin.? A large reservoir lies ventral to the
ganglion, communicating with the mouth by a comparatively
small orifice. A large number of small canals open into the
reservoirs and also communicate with the atrial cavity by a
number of secondary funnel-shaped openings. Round the
funnels are situated masses of cells of a deep yellow colour.

CoNDITION IN THE EMBRYO.

Amarecium Embryo.—The embryo remains in the atrial
cavity of the parent until it has attained to the fully-deve-
loped tadpole stage.

The nervous system then consists of four parts :

1 Julin, C., loc. cit.
2 Julin, C., loc. cit., ‘‘ Deuxigsme Communication.”

NOTE ON THE CILIATED PIT OF ASCIDIANS. 133

(1) An anterior dorsal part (fig. 3 (1) ), exactly resembling
in structure the ganglion of the adult.

(2) A mass (fig. 3 (2) ) lying ventral and posterior to (1),
composed of very large ganglion cells with very distinct nuclei
and nerve-fibres.

(3) A nerve-cord (fig. 3 (3) ) passing off from (2) into the
tail.

(4) A hollow sense-vesicle lying to one side of (2), and con-
sisting of a vesicle with thin anterior and thick posterior walls.
The unpaired eye is embedded in the wall at its antero-dorsal
angle, and the otolith is situated on its floor and projects
upwards into its cavity. This is the only part of the nervous
system which is hollow at this time.

The ciliated pit opens into the buccal cavity and thence
passes back, lying ventral to the anterior part of the nervous
system, penetrates the junction between (1) and (2), and con-
tinues its course on the dorsal side of (2), ending blindly (£.)
not far from the atrial pore (A.P.). At two points it opens
into the solid nervous substance: first (B.), at about the
middle point of the ventral surface of (1); secondly (R.), on
the dorsal surface of (2).

Maurice and Schulgin! failed to find the ciliated pit in the
embryo of Amarezcium, and state that there is no connection
between the buccal cavity and the nervous system. This con-
nection was quite clear in all my sections in which the
stomodzum and nervous system were definitely established.

Kowalevsky? states that in Phallusia mammillata the
mouth communicates with the hollow anterior end (viz. sense-
vesicle) of the nervous system by a pore, which eventually
gives rise to the ciliated pit. It is possible that he may have
overlooked the existence of the ciliated pit in the embryo, as
in optical sections, with which he worked almost entirely, it
might appear as a simple pore.

1 Th. Maurice and Schulgin, ‘“ Embryogénie de ’Amarcecium Proliferum,”
* Annales des Sciences Naturelles,’’? 1884.

2 Kowalevsky, A., “ Weitere Studien tiber die Entwicklung der EHinfachen
Ascidien,” ‘ Arch. fiir Mikr. Anat.,’? 1871.

134 LILIAN SHELDON.

Seeliger! observed the pit in the embryo Clavellina, though
he found no connection between it and the nervous system.

Salensky? describes the ciliated pit in the embryo of Salpa
democratica as a short tube forming a communication
between the pharynx and the anterior end of the nervous
system, which is at first hollow, but afterwards becomes solid.

Kupffer? found no communication between the mouth and
nervous system in the embryo of Ascidia mentula.

Van Beneden and Julin‘ describe the ciliated pit in the
embryo of Clavellina rosaceus, and also state that on its
ventral side it communicates with a mass of tissue lying
ventral to it which gives rise to the gland. From their figures
I believe this mass of tissue to be homologous with what I
have described above as the posterior ventral part of the
nervous system in Amarecium. In the latter I assume that
this structure is a part of the nervous system on account of
its histological features, and also from the fact that it connects
the dorsal part of the nervous system with the nerve-cord of
the tail.

CoMPARISON OF CoNDITION IN EmBrRyo AMARACIUM WITH
THE VaRiIous Tyres IN ADULT ASCIDIANS.

I have not worked out the development from the oldest
unhatched embryo into the adult form owing to lack of
material, so that the views here put forward must be merely
provisional.

There can be little doubt that the part of the embryonic
nervous system, which is found persisting in the adult, is the

1 Seeliger, O., loc. cit.

2 Salensky, W., ‘“‘ Ueber die Embryonale Entwicklungsgeschichte der Sal-
pen,” ‘ Zeit. fur wissen. Zool.,’ 1876.

3 Kupffer, C., “ Die Entwicklung der Hinfachen Ascidien,” ‘ Arch. fiir
Mik. Anat.,’ 1872.

4 Rd. van Beneden et Ch. Julin, ‘ Le systéme nerveux central des Ascidies
et. ses rapports avec celui des larves urodéles,” ‘ Arch. de Biologie,’ tome v,
1884.

NOTE ON THE CILIATED PIT OF ASCIDIANS. 135

anterior dorsal part, this being rendered probable both by the
histological resemblance of the two structures, and also by the
fact that their position with regard to the ciliated pit is the
same in both cases.

The adult condition most nearly resembling that of the
embryo Amarecium is found in Clavellina. Here the ciliated
pit retains its connection with the nervous system, com-
municating with the ganglion in precisely the same way
as it communicates with the anterior part of the nervous
system in the embryo Amarecium. In both cases it then
becomes narrower, passes on and ends blindly—the difference
being that in Clavellina it communicates on its ventral side
with the gland, in the Amarecium embryo with the posterior
ventral portion of the nervous system. The relation of the
ciliated pit to the gland in Clavellina is identical with that of
the ciliated pit to the posterior ventral part of the nervous
system in the Amarecium embryo.

In the adult Amarzcium the position occupied in the
embryo by the posterior part of the brain and in Clavellina by
the gland is filled with a mass of degenerated tissue, and the
communication between the ciliated pit and the nervous system
has been lost. In Ascidia and Ciona the degenerated tissue
is replaced by a mass of somewhat complicated glandular tissue
lying under the ventral wall of the ciliated pit and communi-
cating with it.

In Phallusia mammillata the condition is still more
complicated, the gland communicating with the peribranchial
cavity by a number of secondary funnels, its opening into the
mouth being comparatively small.

SUGGESTIONS AS TO THE FUNCTIONS OF THE CrLIATED Pir.

Judging from the fact that in the embryo Amarecium the
ciliated pit is connected exclusively with the brain, it seems
probable that its original function was the aeration of the
brain; this mode of aeration being similar to that found in
Nemertines. It is doubtful whether it originally opened to

136 LILIAN SHELDON.

the exterior, and was subsequently involved in the stomodeum,
or whether its opening into the mouth is primitive. In
Amarecium, at any rate, it is almost certainly epiblastic in
origin, as it is derived from the epithelium of the stomodeum
and not from the pharynx, as has been stated by Seeliger? for
Clavellina.

Since, as mentioned above, it has in the late embryo no con-
nection with any other structure than the brain, any other
connection which exists in the adult is probably secondary.

In the adult no trace of the posterior part of the brain is found,
but occupying its place in Amarecium is a mass of degenerated
tissue, which is connected with the exterior by means of the
ciliated pit.

In Ascidia and Ciona, and apparently most other simple
Ascidians (cf. Julin, loc. cit.), the function of the ciliated pit
is to act as a duct for the so-called hypophysial gland (Julin)
which lies in the position occupied in the Amarecium embryo
by the posterior part of the brain.

In Clavellina the ciliated pit has a twofold function.

(1) It communicates with the brain, and _ probably
aerates it.

(2) Its posterior part acts as a reservoir to carry off the
secretion of the gland.

There is thus a gradual transition from one function to
another in the different types; the primitive condition, as an
organ for the aeration of the brain, is found in the Amareecium
embryo, and is retained in Clavellina, while in the latter the
secondary function, viz. that of an excretory duct, is also
acquired.

In most adult forms (e.g. Amarzcium, Ascidia, and Ciona)
the primitive function is lost, the secondary one only being
retained.

The gland is possibly an altogether secondary structure,
being developed to supply the need of an excretory organ in
the anterior part of the body. It reaches its highest degree of
complication in Phallusia mammillata.

1 Seeliger, O., loc. cit.

NOTE ON THE CILIATED PIT OF ASCIDIANS. 137

If the excretory function of the ciliated pit be merely
secondary no homology can exist between it and the pro-
boscis pore of Balanoglossus,! or the external opening of the
left anterior pouch from the fore-gut, described by Hatschek?
in Amphioxus.

It is more probably homologous with the hypophysis of
Vertebrates, the original functions of which may have been the
aeration of the brain. When a complete blood-supply to the
head was effected aeration by this means would no longer be
required, and since a definite and complete excretory system
had been at the same time developed, there would no longer
be any necessity for an excretory organ in this position. Thus
the hypophysis at the present time may represent merely a
rudimentary condition of the gland and ciliated pit in Ascidians,
having almost atrophied, and quite lost its function as a con-
sequence of the development of the ordinary Vertebrate
excretory system.

It is possible that the pineal gland of Vertebrates may
represent the dorsal continuation of the ciliated pit in the
embryo Amarecium (fig. 3, E).

THE ANATOMY OF CYNTHIA.

Whilst investigating the condition of the ciliated pit in
various genera of Ascidians, I observed several features in
Cynthia, which, as far as I know, have not hitherto been
described. I have therfore thought it worth while to publish
a short account of the general anatomy and histology of the
species I have studied, which I believe to be Cynthia
rustica.

N.B.—Some of the generic characters, in which the Cyn-
thiide differ from the Molgulide, have been pointed out by

1 Bateson, W., “ The Later Stages in the Development of Balanoglossus
Kowalevskii, &c.,” this Journal, 1885.

2 Hatschek, B., “Studien tiber Entwicklung des Amphioxus,” ‘ Arbeiten
aus dem Zool. Inst. Wien,’ 1882.

133° - LILIAN SHELDON.

Lacaze-Duthiers ;! and Herdman? gives a short account of some
forms found by the Challenger.

EXTERNAL CHARACTERS.

The individuals are small, varying in length from half to
little more than an eighth of an inch, and each is almost as
broad as it is long. It is wide and flattened at its base, by
which it is attached to the rock, there being no stalk or
peduncle. A great number of individuals live upon the same
piece of rock, and are very closely applied to one another by
their sides, so as to form a compact mass; but any one can be
easily separated from the rest. The lateral pressure to which
they are thus subjected causes the individuals to vary much in
shape.

In colour they are a bright brownish red. They are quite
opaque, so that it is not possible to make out any of the inter-
nal structure without removing the test.

The oral and atrial apertures are four-lobed.

Tue Test.

The test, which gives the red colour to the animal, is fairly
thin, and in life is so closely applied to the body wall that it
is difficult to remove it without tearing the epidermis. After
preservation it can be removed with comparative ease, although
it still adheres to the body wall. Except at its base, where
fragments of the rock on which it lived generally remain
attached to it, the test is free from sand, calcareous spicules,
or other foreign matter.

As seen in sections it is composed of a homogeneous matrix
with a few scattered cells. It also possesses a complicated
system of vessels, which are lined by short columnar cells. In
picrocarmine-glycerine preparations of the test these vessels
are clearly seen, as well as numerous pigment cells, which are

1 Lacaze-Duthiers, Henri de, ‘‘ Histoire des Ascidies Simples des Cotes de
France,” ‘ Archives de Zool. Exp. et Gen.,’ tome vi, 1877.

2 Herdman, W.A., ‘Challenger Report on the Tunicata,’ Part I, “Ascidie
simplices.”

NOTE ON THE OJLIATED PIT OF ASCIDIANS. 139

of two kinds, one appearing dark brown or black, and the other
a bright orange.

Bopy Watt ann Bopy Caviry.

Beneath the test and closely applied to it is a layer of
columnar cells (figs. 12 and 15, Ep.) with very definite nuclei.
Their internal ends are rounded and they do not rest upon any
basement membrane.

Beneath the epidermis is a thin layer of circular muscle-
fibres (fig. 15, c. m.), then a thin layer of longitudinal fibres
(fig. 15, 2. m.). Within the latter is a very thin layer of cir-
cular fibres (fig. 15, c. m.), and then a thick layer of longitu-
dinal and oblique fibres (fig. 15, 7. m.), which are arranged in
large irregular masses, not definitely marked off from one
another, or from the surrounding tissue (fig. 15). All the
muscles are unstriated.

A mass of connective tissue, constituting a meshwork of
fibres, fills up the spaces between the muscles and extends
inwards as far as the lining of the atrial cavity (figs. 12 and 15
Ps.). Nuclei are sometimes visible (fig. 15, c. ¢. ¢.) at the
points of junction of the fibres; these apparently are the
nuclei of the connective-tissue cells, the cell protoplasm being
so drawn out and branching as not to be apparent round them.

Among the meshes of the connective tissue and lying freely
in it are three kinds of cells:

1. Cells tilled with black pigment (fig. 15, p.c.). A nucleus
is present in each.

2. Large, coarsely granular cells (fig. 15, g. ¢.), in which I
have not been able to discover nuclei.

3. Cells which stain faintly (fig. 15, 6. c.), and have large,
deeply-staining nuclei. The cell protoplasm is very finely
granular.

Thin lamine of this connective tissue (fig. 12, Ps.,) pass
across the atrial cavity into masses of similar tissue (fig. 12,
Ps.,), which surround the alimentary canal, thus acting as
mesenteries.

Large, irregularly-shaped processes of this connective tissue

140 LILIAN SHELDON.

(figs. 12 and 15, Ps. p.) project into the atrial cavity along its
outer wall, being only separated from it by the epithelial
lining of the atrial cavity (figs. 12 and 15, A. Ep.) which
covers them. Figure 12 is a diagrammatic transverse section
through Cynthia, and represents the relation of these processes
to the atrial cavity. They appear very conspicuously when
the atrial cavity is opened, especially in the fresh animal.
They are then seen as very large, opaque, white bodies, which
fill up a considerable portion of the cavity. I am not aware
that any analogous structure has hitherto been described in
any Ascidian.

The connective tissue extends from the body wall inwards
to the atrial epithelium (fig. 12, Ps.), and also surrounds the
alimentary canal. There is therefore no body cavity, its place
being occupied by what appears to be a system of sinuses, the
third kind of cell described above being the blood-corpuscles.
For convenience of description I shall hereafter speak of the
sinuses as the pseudoceele ; but I do not wish to imply that it
is necessarily homologous with the so-called pseudoceele of
some Invertebrates. There are no definite blood-vessels, the
heart opening out at both ends into the sinuses. Strands also
pass across the atrial cavity to the pharynx, the spaces in the
branchial bars being continuations of the sinuses, plentifully
supplied with blood-corpuscles (figs. 5 and 7, dr. 6.).

It seems probable that the function of the processes pro-
jecting into the atrial cavity is to expose a greater surface of
the sinus to the influence of the oxygen contained in it.

THe Arriat Cavity.

The atrial cavity is very capacious, extending beyond the
posterior end of the alimentary canal. For the greater part
of its extent it is divided completely into two halves by a par-
tition, which passes along from the outer wall of the atrial
cavity to the walls of the pharynx. ‘This partition starts just
behind the buccal cavity, follows the line of the endostyle, and
curves round the posterior end of the pharynx on to its dorsal
surface, where it passes along the line of the dorsal lamina,

NOTE ON THE CILIATED PIT OF ASCIDIANS. 141

and stops just behind the atrial pore. The whole of the ali-
mentary canal behind the pharynx is situated on the left side
of the partition (fig. 12, St. and Jnt.). Thus the alimentary
canal, embedded in its own portion of the “ pseudocele,” is
completely surrounded by the atrial cavity except at those
points where strands pass off connecting its “ pseudoccele ”
with the “ pseudocele” in the body wall outside the atrial
cavity, and forming the mesenteries described above.

The atrial cavity is lined throughout by a layer of flattish
cells, lying upon a basement membrane (figs. 12 and 15,
A. Ep.). The cells are rounded at their free ends, and have
large nuclei. This layer, of course, extends over the portions
of the pseudoceele surrounding the alimentary canal, and over
the mesenteries.

The atrial pore is placed not far from the mouth on a
short papilla.

ALIMENTARY CANAL.

The mouth leads into the buccal cavity, which passes into
the large pharynx (fig. 4, Ph.). The endostyle extends round
the posterior end of the pharynx, ceasing at the point where
the cesophagus is given off. The cesophagus (fig. 4, Oes.) is
short, and soon opens into the large sack-like stomach (fig. 4,
Sz.), which, on its external surface, is marked by deep longi-
tudinal ridges. The stomach lies on the left side of the
pharynx, its long axis being at right angles to that of the
latter. It passes into the intestine (fig. 4, Jnt.), which almost
at once bends upon itself and runs back across the pharynx,
almost parallel with the stomach. On reaching the level of
the dorsal side of the pharynx it turns forwards almost at a
right angle, and runs straight to the anus (fig. 4, A.), which is
situated at a short distance from the atrial pore, at the point
where the partition dividing the atrial cavity ceases.

Shortly after leaving the stomach the intestine appears to
be dilated for a short distance (fig. 4, Int. L.) ; this appearance
is due to its being surrounded here by a layer of liver-cells.

The buccal cavity is lined just within the mouth by a

142 LILIAN SHELDON.

portion of the test which grows into it. This cellulose lining
acquires a covering of thin epithelium, and forms a circle of
short, blunt, unbranched tentacles. Further down, the cavity
is lined by high ciliated columnar cells, which are thrown up
into papille. At the junction of the buccal cavity with the
pharynx there is a circle of cirri, which are unbranched, and
in transverse action are four-lobed (fig. 10), one lobe being
much larger than the other three. Two of the small lobes
bear tufts of cilia.

Outside the epithelial lining of the buccal cavity there is a
layer of connective tissue, and around this a thick layer of
longitudinal and circular fibres.

The ciliated pit is simple, having the conditions found in
Ascidia and Ciona. Its opening into the mouth is crescent-
shaped, and it passes back thence, as a simple tube lined by
high columnar ciliated cells, below the ganglion, at about the
middle of which it opens out into the hypophysial gland,
which lies immediately ventral to the ganglion. It has no
communication with the latter.

The Pharynx.—The endostyle starts from the point where
the buccal cavity joins the pharynx, and passes round its
posterior end, ceasing where the cesophagus passes off (fig. 4,
End.). It is very similar to the endostyle of other Ascidians,
bearing a tuft of very long cilia at its base and shorter ones at
the sides of the groove (fig. 9, c.c.). Between the masses of
cells, which bear the cilia, are situated groups of much higher
cells (fig. 9), while on each side, between the most external
masses of ciliated cells and the ciliated epithelium of the
pharynx, there is a row of cells which appear to secrete mucous
(fig. 9, m. c.), as they appear when seen in transverse section
to be throwing out irregular processes towards the cavity. At
the anterior and posterior end of the endostyle its open edges
are fused, so as to form a tube (fig. 8). This is an interesting
point, as tending to confirm the theory of its homology with
the thyroid body of the higher Vertebrata.

A thin lamina (the “dorsal lamina”), covered by low
columnar ciliated cells (fig. 5, D. L.), projects from the median

NOTE ON THE OILIATED PIT OF ASCIDIANS. 143

dorsal line of the pharynx a long way into it; a rod of skeletal
tissue (fig. 5, sk. 6.) is present near its base, and extends
throughout its whole length.

The gill-slits are very numerous. The bars between them
are composed of connective tissue, which is a part of the
pseudoceele, and contains a great number of blood-corpuscles
(figs. 5 and 7, dr. b.).

The skeletal system is somewhat complicated. The slits
(fig. 7, dr. s.) are elongated longitudinally, and arranged in
transverse rows. ‘The rows are separated from one another by
thick skeletal rods, and divided into sets of five by similar rods
(fig. 7, sk. 6.), which meet the former at right angles. In
addition to this main skeletal system, there is a system of finer
rods (fig. 7, s. sk. 6.) which accompany the larger ones, and are
connected together by longitudinal rods passing between every
two gill-slits, and a transverse rod lying across the centre of
each row of slits. This secondary system of rods lies on the
internal face of the pharynx.

The bars themselves in transverse section (fig. 5, dr. b.) are
seen to be covered on the surfaces turned towards the slits by
columnar cells provided with cilia, and on their inner and
outer walls by flat non-ciliated cells. Their internal cavity,
as already stated, contains a blood sinus.

The Gsophagus (fig. 4, Gs.) is very short and simple,
being lined with short, columnar, ciliated cells.

The Stomach.—The walls of the stomach are thrown up
into deep glands, which run in a longitudinal direction along
its whole length.

The mouths of the glands are lined by high columnar
ciliated cells (fig. 6, c. c. s.), very closely packed together, with
a smallish nucleus placed at about the centre of each cell.
The portion of the cells towards the lumen stains very deeply.

These cells are separated from those forming the deeper part
of the gland by a slight constriction (fig. 6, c.).

The latter (fig. 6, s. ¢. s.) are somewhat higher than the
former, and are not ciliated. The nuclei are placed quite near
the bases of the cells. They appear to be secretory cells, as

144 LILIAN SHELDON.

their proteplasm is rich in granules, and at their free ends
many of them are prolonged into blunt processes (fig. 6), which
project into the lumen, and only stain very faintly. In most
of the cells these processes are not visible, but since they other-
wise resemble those that have them, the difference would seem
to be due to their being in a different stage of secretory
activity. In such cells the end abutting on the lumen stains
more deeply than the rest of the cell, and the two regions are
separated in preserved specimens by a row of very deeply-
staining dots. Among the cells a few leucocytes are scattered
(fig. 6, 2.). The glands are separated from one another by a
portion of the pseudoccele.

The intestine is lined throughout by rather short, ciliated
columnar cells (figs. 11, Ep. Jnt.).

When the intestine is viewed as a whole it is seen to be
enlarged for a short distance (fig. 4, Int. L.) soon after leaving
the stomach. This appearance is due to its being surrounded
by the so-called liver-cells. These are large and oval, are
closely applied to the wall of the intestine, and form a compact
mass round it. In the fresh state they are bright orange in
colour. Each cell is pear-shaped with its thin end in contact
with the epithelium of the intestine (fig. 11, L.), and is sur-
rounded by a somewhat thick fibrous coat (f./.), which is
thickest at the broad end of the cell. In preserved specimens
the cells are filled with a coarsely granular substance (fig. 11,
g.1.), which is mostly aggregated at the narrow ends, and con-
tains highly refractive concretions.

I could not find any connection of these cells, either with
one another or with the intestine, but it seems possible that as
they lie so closely applied to the wall of the intestine, their
contents may pass into its lumen between the epithelial cells,
the passages being so small as to have escaped my notice. I
am unable to offer any suggestion as to their probable function.

Tue Heart.

The heart is a long, thin-walled vessel, lying in a pericardium,
which is considerably larger than itself.

NOTE ON THE CILIATED PIT OF ASCIDIANS. 145

It curves round the posterior end of the body, following the
line of the partition between the two halves of the atrial
cavity. For the greater part of its length it is situated in the
pseudoceele external to the atrial cavity, but the anterior end
of it lies in the partition, i. e. immediately ventral to the endo-
style. It extends backwards nearly as far as the cesophagus.

At both ends it opens out into the pseudoceele, anteriorly by
a long narrow slit in its dorsal wall, and posteriorly by a
terminal pore, so that there is apparently no closed system of
blood-vessels, but the circulation is carried on only through the
blood-sinuses which surround all the organs and fill all the
spaces usually occupied by the body cavity.

The pericardium is entirely closed.

The same condition of a heart situated in a closed peri-
cardium, and itself opening at both ends, is said by Seeliger!
to exist in Clavellina. The gradual shutting off of the heart
from the body cavity in the development of the embryo, as he
describes it, explains how the adult condition may have been
brought about in the case of Cynthia.

THe GENERATIVE ORGANS.

The generative organs are unpaired, and lie near the posterior
end of the body in the pseudoceele, outside the right half of
the atrial cavity (fig. 12, O. and és.).

They are arranged in two long masses on the outer and inner
sides of a cavity (fig. 14, G. D.), which is lined by flat cells,
and opens posteriorly into the atrial cavity. The mass on the
outer side consists of testes (figs. 12 and 14, ¢s.), that on the
inner of ova (figs. 12 and 14. O.).

The testes (fig. 13) are large oval sacs filled with spermatozoa
(fig. 13, sp.); each sac is lined by a thin membrane, outside
which are bundles of muscles (MM. ¢s.), and, especially at its
outer end, large collections of pigment-cells. Atits inner end,
i. e. towards the cavity, it is drawn out into a short duct (figs.
13a, 14, D. ts) lined by low columnar cells, which terminates

1 Seeliger, Oswald, “ Die Entwicklungsgeschichte der Socialen Ascidien,”

‘ Jenaische Zeitschrift fir Naturgewissenschaft,’ 1885.
VOL, XXVIII, PART 1.—NEW SER, K

146 LILIAN SHELDON.

close to the wall of the cavity, but does not appear to open into
it. It is possible that when the spermatozoa become ripe they
pass down the duct into the cavity by an opening, which is
temporarily established into it, the communication being closed
at other times.

The ovary (fig. 14, O.) lies on the opposite side of the cavity.
It has no limiting membrane, the ova lying freely in the
pseudocele. There is no oviduct, but the ova, when ripe,
probably break through the wall into the cavity, which thus
serves as the common generative duct, conveying both male
and female products into the atrial cavity.

Each ovum is surrounded by a single layer of cells, which
either are follicle cells or give rise to the test.

Tue Nervous System.

The nervous system has the form generally found in adult
Ascidians. It consists of a large ganglion lying between the
mouth and atrial pore, and is composed of fibres, which are sur-
rounded by a peripheral layer of ganglion cells. Anteriorly it
sends off two nerve-trunks, which pass, one on each side of the
mouth ; posteriorly it also sends off two trunks, which very
soon divide again into several small branches, whose course I
could not follow for any distance.

Tue HyropnysiaAL GLAND.

The hypophysial gland is a compact mass of tissue, lying
immediately ventral to the ganglion. It consists of a great
number of fine glandular tubes, which towards the dorsal side
of the gland unite, forming larger ducts, which open into the
canal of the ciliated pit.

NOTE ON THE CILIATED PIT OF ASCIDIANS. 147

EXPLANATION OF PLATES IX & X,

Illustrating Lilian Sheldon’s “Note on the Ciliated Pit of
Ascidians, and its Relation to the Nerve Ganglion and so-
called Hypophysial Gland ; and an Account of the Anatomy
of Cynthia rustica (?).”

List of Reference Letters.

A. Anus. A. Zp. Epithelium lining atrial cavity. 4¢. Atrial cavity. A/. P.
Atrial pore. 8. Communication between ciliated pit and ganglion. J. ¢.
Blood-corpuscles. 6r. 6. Branchial bar. 47. s. Branchial slit. ©. Canal of
ciliated pit. C. P. Ciliated pit. c¢. Constriction between the two kinds of
cells lining glands of stomach. c. c. Ciliated cells of endostyle. .c. s. Ciliated
cells lining mouths of glands of stomach. c.m. Circular muscles. . ¢. c. Con-
nective-tissue corpuscles. D.Z. Dorsal lamina. JD. ¢s. Duct of testis. 2.
Blind end of ciliated pit. Hxd. Endostyle. 2p. Epidermis. Zp. Int. Epi-
thelium lining intestine. Zp. ph. Epithelium lining pharynx. (f Z. Fibrous
coat of liver-cells. G.D. Common generative duct. G/. Hypophysial gland.
g-c. Granular cells of pseudocele. g./. Granular contents of liver-cells.
H. Heart. Int. Intestine. Int. L. Part of intestine surrounded by liver-cells.
LZ. Liver-cells. 7. Leucocytes. 7. m. Longitudinal muscles. m. ce. Mucous
cells of endostyle. M. ¢e. Muscles surrounding testis. WV. G. Nerve ganglion.
O. Ova. is. (Hsophagus. P. Communication of ciliated pit with mass of
tissue in Amarecium. Pe. Pericardium. Ph. Pharynx. Ps. Pseudoceele.
Ps.p. Processes of pseudocele. p.c. Pigment-cells. 2. Communication of
ciliated pit with (2). sp. Spermatozoa. s¢. Stomach. S¢m. Stomodeum.
s.c.s. Secreting cells of stomach. sk. b. Skeletal rod. s.sk.. Secondary
skeletal rod. 7. Test. Z%. Mass of tissue lying ventral to ganglion in
Amarecium. ¢s. Testis. (1) Anterior dorsal part of nervous system of
Amarecium. (2) Posterior ventral part. (3) Nerve cord to tail.

All the figures, except Figs. 12 and 4, were drawn with Zeiss’s camera
lucida. The objective and ocular with which they are drawn are mentioned in
the description of each figure.

Fic. 1.—Longitudinal section through the ciliated pit, ganglion, and gland
of Clavellina, showing the communication of the ciliated pit with the ganglion
and gland. (Oc. 2, obj. A.)

Fie. 2.—Longitudinal section through the ciliated pit, ganglion, and gland
of an adult Amarecium, showing the ciliated pit quite shut off from the
ganglion but communicating with the mass of tissue lying ventral to it. (Oc.
2, obj. D.)

148 LILIAN SHELDON.

Fie. 8.—Longitudinal section through the ciliated pit and nervous system
of a larval Amarecium immediately before hatching. The ciliated pit com-
municates with the two portions of the nervous system. (Oc. 2, obj. D.)

Fic, 4.—Alimentary canal of Cynthia rustica, seen from the left side.
Drawn from a dissection.

Fic. 5.—Transverse section through the dorsal lamina of C. rustica and
three branchial bars and slits. (Oc. 2, obj. C C.)

Fig. 6.—Transverse section through one of the stomach glands of C.
rustica, showing the two kinds of cells by which it is lined.

Fic. 7.—A portion of the wall of the pharynx of C. rustica, showing
the arrangement of the branchial slits and skeletal system. Drawn from a
picrocarmine-glycerine preparation. (Oc. 2, obj. B.)

Fie. 8.—Transverse section through the anterior end of the endostyle of
C. rustica, where its edges have fused to forma tube. (Oc. 4, obj. B.)

Fie. 9.—Transverse section through the endostyle of C. rustica in its
middle region. (Oc. 4, obj. B.)

Fie. 10.—Transverse section through one of the cirri in the buccal cavity
of C. rustica, showing its four lobes, two of which are provided with
cilia. (Oc. 2, obj. D.)

Fig. 11.—Transverse section through a small portion of the wall of the
intestine of C. rustica, showing the adjacent liver-cells with their fibrous
walls and granular contents with concretions. (Oc. 2, obj. E.)

Fie. 12.—Diagrammatic transverse section through a C. rustica, near
its base, showing the various orgaus and their relations to one another.

Fie. 13.—Section through a testis of C. rustica, showing it filled with
spermatozoa and surrounded by a delicate membrane, outside which are
muscles and pigment-cells. At one end it is seen to be continued into a duct.
(Oc. 2, obj. C C.)

Fic. 14.—Transverse section through the generative organs of C. rustica,
showing the testes, with their ducts, and the ova lying respectively on the
outer and inner side of the generative duct. (Oc. 2, obj. A, reduced half.)

Fic. 15.—Transverse section through a small portion of the body wall and
pseudoccele, showing the muscle layers and connective-tissue meshwork, with
the different kinds of cells in it. A process from the pseudoceele projecting
into the atrial cavity is cut through. (Oc. 4, obj.C C.)

"

i

7 «2
(sc
. “Is

=

G.G.Scsmas

Mtr bourd We WU NS PEEK

: fy
ALL KSoohys (slolnlclololololoy=p. s.
Fig. 2. Se rai oP
se Nes

mal Ad P00 AR
Se wy wu
— rei
7 re
=a
Ley
bs
ae |
ase |
(Cet
|
4 (3) Ep.pk,

br.b.

F Huth, Lith* Edin”

GX.

4

UMS:

Mor Lourn! U.N

F Huth, Lith® Edin®

TONGUE AND GUSTATORY ORGANS OF MEPHITIS MEPHITIOA,. 149

The Tongue and Gustatory Organs of
Mephitis mephitica.

By

Frederick Tuckerman, M.D.Harv.,
Amherst, Mass., U.S.A.

With Plate XI.

Ir is now twenty years since Christian Lovén (26)! and
Gustav Schwalbe (37) discovered and described, independently
of each other, the peripheral end-organs of the nerves of taste
in the tongue of Mammalia. Subsequent investigators have
studied the distribution and minute anatomy of these organs
in many animals, and in all essential points they confirm the
results reached by Lovén and Schwalbe.

Before passing to the consideration of my own observations
I will briefly review what is known regarding the position and
structure of the taste organs of Vertebrates.

Bellini, nearly two hundred years ago, considered the papillz
of the tongue to be organs of taste.

In 1846, Waller (50) investigated the epithelium of the
fungiform papille of the frog, and also studied the cilia and
ciliary movement. In 1847, he (51) concluded, from the
experiments of Longet, that the glosso-pharyngeus was the
nerve of taste for the base of the tongue and the lingual for
the tip and anterior third. He succeeded in tracing nerves
into the base of the fungiform papillz of the frog, and into the
papillary elevations on the tongue of the toad. He believed
the fungiform or “neurovascular” papille to be the chief

1 These figures refer to the bibliography at the end of the paper,

150 FREDERIOK TUCKERMAN.

organs of taste in the frog. The soft palate has also, he says,
the power of taste. The conical papille he considered tactile
organs. In 1849 he (54) redescribed these organs, and also
speaks of the gustatory nerves terminating in the fungiform
papillz on the dorsum of the tongue, and of a gustatory area
situated at the summit of each papilla.

In 1851 Leydig (23) described in the external skin of fresh-
water fishes certain beaker- or flask-shaped bodies, which he
was disposed to regard as organs of a tactile nature. In 1857,

~he showed (24) that the epithelium covering the end surfaces
of the fungiform papille differs from the rest of the epithelium.
Later investigators, with the exception of Fixen, have con-
firmed this.

In 1863, J. E. Schulze (34) redescribed the beaker-shaped
bodies of fishes, and considered them organs of taste. He
found them in greatest number where the fibres of the glosso-
pharyngeal nerve are most thickly distributed, i.e. in the
mucous membrane of the palate, upon the gums and tongue
rudiment, on the inner side of the gill arches, and upon the
lips. In structure he found them to agree, in most respects,
with the organs of taste of the frog. The beakers he described
as composed of two kinds of cells, viz. Sinneszellen and Stiitz-
zellen, or sensory and supporting cells; the former having a
peripheral and central process. In 1867, he stated (35) that
the peripheral extremity of the taste-cell bears a fine hair-like
process as in mammals. In 1870, Schulze (86) described, in
the papille of the mouth of a larval amphibian (Pelobates
fuscus), bodies resembling in structure the beaker-shaped
organs of fishes, which he considered taste organs.

In 1872, Todaro (44) described, in the papille covering the
rudimentary tongue of Trygon pastinaca, a number of club-
shaped bodies connected with the ultimate ramifications of the
glosso-pharyngeus nerve, which he regarded as organs of taste
and analogous to those of mammals. At the base of the gus-
tatory organ the nerve loses its sheath, and the fibrillz of the
axis cylinder separate and join the central processes of the
taste-cells.

TONGUE AND GUSTATORY ORGANS OF MEPHITIS MEPHITIOA. 151

Jourdan quite recently (18) has pointed out on the gills and
in the buccal cavity of the malarmat, cup-shaped bodies com-
posed of central and peripheral cells, which, in structure and
situation, differ completely from the organs of touch, and
which he regards as taste-bulbs.

In 1858, Billroth (4) described the peculiar epithelium of the
taste papille of the frog, and believed that certain of its cells
were continuous with nerve-fibres. The smaller papille he
thought were unprovided with nerves.

Hoyer (16) differed from Billroth in supposing that the
nerves terminate bluntly beneath the epithelium.

In 1861, Key (19) described in the frog two kinds of cells
at the summit of the fungiform papilla,—epithelial cells and
taste-cells. He speaks of the penetration of the axis cylinder
alone into the papilla, and its division into fibres which enter
the taste-cells at its summit. The “nerve-cushion” of Engel-
mann he considered an enormous enlargement of the neuri-
lemma and called it the “ nerve-shell.”

Hartmann (10) thought that the nerves ended in plexuses
beneath the epithelial cells of the fungiform papille.

Beale (2), Engelmann (7), and Maddox (29) supported Key,
believing in a structural continuity between the cells at the
top of the papilla and the nerve-fibres in its axis. The former,
however, did not consider the cells to be of epithelial origin.

In 1868, Engelmann (8) described, in the fungiform papilla
of the frog, numerous dichotomous subdivisions of the nerve-
fibres, which form a close network and spread out in the lower
half of the “ nerve-cushion ” in nearly a horizontal direction.
The upper part of the papilla consists of a solid disc composed
of non-nucleated connective tissue, which he calls the “ nerve-
cushion,” and upon which rests the taste disc. The latter is
composed of three distinct kinds of cells, viz. cup-shaped,
cylinder, and forked. The two former he considered were
epithelial cells only. The forked cells he regarded as the end-
organs of the gustatory nerve, probably being directly con-
tinuous with pale nerve-fibres, which in their chemical reaction
they resemble. Engelmann says that the nerve-fibres just

152 FREDERIOK TUOKERMAN.

before entering the “ nerve-cushion” lose their medullary
substance and neurilemma.

In 1869, Beale (3) redescribed the epithelium of the papille
of the frog, and reiterated his disbelief in the existence of
structural continuity between nerve-fibres and epithelial cells.
He figured fine nerve-fibres ramifying in the connective tissue
of the simple papille, and connected with them oval-shaped
masses of germinal matter or nuclei, formerly supposed to be
connective tissue. He believed in a connection between the
cells upon the summit of the fungiform papilla and the nerve-
fibres in its axis, but did not consider the former epithelial in
structure. He figured a nervous plexus containing nuclei at
the top of the papilla, the fibres cf which are derived from the
nerves in its axis, and from which fine fibres may be traced
into the special organ composed of “ epithelial-like” cells.
He found that the bundle of nerve-fibres distributed to a
papilla always divides into two, which pursue opposite direc-
tions; this division taking place either at the base of the
papilla or at some distance from it.

In 1869, Maddox (29) regarded the fungiform papille as the
chief organs of taste in the frog, and described the nerves of
taste as possessing terminal organs consisting of nerve matter.

In 1867, Szabadféldy (43) described the nerves of taste as
terminating in mammals in pear-shaped bodies lying in the
mucous membrane of the tongue. ‘Two years later Letzerich
(22) called attention to a peculiar way in which these nerves
end in the papille of the cat, ox, and weazel. In neither case
have the results reached by these observers been verified.

In 1867, Lovén (27) described the taste-bulbs (Geschmacks-
zwiebeln) or taste-buds (Geschmacksknospen) of mammals. He
studied them in the circumvallate papille of the cat, rabbit,
pig, sheep, calf, dog, horse, and man, and found them to con-
sist of central and peripheral cells. The outer or cover-cells,
for support and protection; the inner or taste-cells, bearing a
central and peripheral process, the former being continuous
with a nerve-fibril, the latter terminating in a delicate hair-like
extremity which projects a short distance beyond the opening

TONGUE AND GUSTATORY ORGANS OF MEPHITIS MEPHITIOA, 153

of the bulb. He says that in man and the calf the gustatory
nerve-fibres lose their medullary sheath in the outer layer of
the mucous membrane, the axis cylinder being prolonged into
the bulb, where it divides into terminal branches which are
distributed to the taste-cells. In the calf the peripheral process
of the taste-cell carries no hair at its extremity. Lovén found
taste-bulbs and cells on the upper surface of the fungiform
papille of the calf, rat, and rabbit, the small ones containing
a single specimen only. He also detected in the rat and rabbit
a few taste-bulbs in the outer wall of the trench encircling the
circumvallate papilla.

In 1867, Schwalbe (37) published the preliminary report of
his investigation of the ‘‘ Schmeckbechers” in the papille of
the sheep, ox, horse, dog, cat, and rabbit. His detailed account
(38) of the taste-goblets of these animals, including also those
of the deer, pig, guinea-pig, hare, and man, appeared the fol-
lowing year. His description of their location and structure
agrees essentially with that given by Lovén. He found taste-
bulbs in man and the dog on the outer wall of the trench sur-
rounding the circumvallate papilla, and in the fungiform
papillz and lateral organ of taste of the pig. He notesin man
and the sheep two kinds of taste-cells, namely, staff-cells
(Stabzellen) and needle-cells (Stiffchenzellen). He found in
the sheep at the apex of the bulb, after treatment with per-
osmic acid, a circle of fine short hairs or cilia, which appeared
to spring from the end of the cover-cells. In the sheep also,
at the base of the circumvallate papilla, is a richly-developed
nervous plexus. He speaks of the small branches of the glosso-
pharyngeus being provided with ganglia, especially where the
nerve divides at the base of the papilla.

Engelmann (9) found bulbs in the fungiform papille of the
mouse and cat, and described them in the lateral organs of
taste (papille foliatez) of the rabbit and hare. He says that
usually the central process of the taste-cell divides, at a short
distance from the nucleus, into two branches. He speaks of
groups of ganglion-cells in the branches of the glosso-pharyn-
geus ramifying beneath the taste-ridges of rodents, and points

154 FREDERICK TUCKERMAN.

out the resemblance in physical characteristics and chemical
reaction of nerve-fibrils with the central processes of the taste-
cells. :

In 1869, v. Wyss (56) described the taste-bulbs in the
papilla foliata of the rabbit and hare, and called attention to
the analogous organs of man. In 1870, (57) he studied them
in the circumvallate and fungiform papille of many mammals,
including the hedgehog and squirrel; but failed to find them
in the fungiform papille of man.

Krause (20) observed taste-bulbs in the fungiform and
foliate papille of man, and v. Ajtai (1) described them in the
papilla foliata of man and other mammals.

Hénigschmied (12 and 13) has shown the distribution of the
taste-bulbs in the circumvallate and foliate papillze of various
mammals, and, in some, has found them occurring on the
summit of the papilla, though these were usually smaller than
those on the sides. By means of chloride of gold, Honig-
schmied traced the nerve-fibrils directly into the taste-cells in
the fungiform papilla of the cat, the cover-cells not being
stained, while the taste-cells were. Vintschgan and Honig-
schmied (47 and 49) found in the rabbit that, after section of
the glosso-pharyngeus nerve, the taste-bulbs degenerate, while
the cover-cells become changed into epithelial cells in a few
months. Ranvier (32) repeated this experiment, and has met
with a similar result.

Sertoli (41) states that he has traced nerve-fibrils directly
into the taste-bulbs in the papilla foliata of the horse. In the
fungiform papille of the same animal he says that the nerves
terminate in an intra-epithelial plexus of fine non-medullated
nerve-fibrils.

Hoffmann (14) found taste-bulbs on the summit of the fungi-
form and circumvallate papille of man, and also in some papillz
of the soft palate and upper part of the uvula. He failed to find
in the epiglottis what he considered genuine taste-bulbs.

In 1868, Verson (45 and 46) described in the second fourth
of the posterior surface of the epiglottis of man, certain “ bud-
like structures,’ which resembled mammalian taste-bulbs.

TONGUE AND GUSTATORY ORGANS OF MEPHITIS MEPHITIOA. 155

Krause (21) observed them on the dorsal surface of the epi-
glottis of the sheep and rabbit. Schofield (33) has described
them in the lower half of the posterior surface of the epi-
glottis of the dog and cat, arranged in horizontal and ver-
tical rows. He says that with each “ goblet”’ is associated
the duct of a mucous gland. In 1873, v. Ebner (6) pointed
out that the serous glands always occur in the parts of the
tongue that contain taste organs, and their ducts open into the
furrows lined by the taste-bulbs. Davis (5) studied the bulb-
shaped organs in the epiglottis of the cat, dog, calf, pig, rabbit,
and man, and found them on the upper and lower parts of the
posterior surface of that organ. In the dog he found them on
the inner side of the arytenoid cartilages, on the aryepiglottic
folds, and in the epithelium of the true vocal cords. He re-
garded them as terminal organs of the glosso-pharyngeal nerve.
Simanowsky (42) has also seen them on the true vocal cords
of man and the dog. In the vocal cord of the dog and rabbit
he figures the nerve-fibres as terminating in pencil-shaped ex-
tremities. Hoénigschmied has observed bulbs on the epiglottis
of the deer and calf.

Poulton (30) has described, in a highly interesting and sug-
gestive manner, the taste-bulbs of Perameles nasuta. The
circumvallate papillz, he says, are highly developed in shape
and structure, and in the abundance of nervous and glandular
tissue which they possess ; but the terminal organs he considers
of a low type. Within the papillary body is a large ganglion,
which divides into branches running towards the sides of the
papilla containing taste-bulbs. The nerve-fibres are chiefly
non-medullated, but possess a distinct sheath of Schwann.
He found near the top of a single fungiform papilla two bulbs
of alow order. Poulton (31) found in the posterior region of
the tongue of Ornithorhynchus paradoxus two pairs of
gustatory areas. The anterior pair lie below the surface in a
furrow, the floor of which is invaginated upwards into a ridge,
which carries the taste-bulbs. The ridges of the posterior pair
reach the surface. In both regions the bulbs are situated on
the sides and upper surface of the ridges, and the ducts of

156 FREDERICK TUOKERMAN.

serous glands open into the spaces around them. The centre
of the ridge is nearly filled by non-medullated nerve-fibres,
which radiate outwards to end in the bulbs. The bulbs are
developed at the ends of long papillary processes. Nerve-
fibres can be followed into the bulbs where they pass between
the cells in various places. Poulton suggests “ that the primi-
tive type of bulb was papillary in position and subepithelial in
structure, and has gradually given way to a bulb that was
interpapillary and epithelial.”

Among recent observers who have studied the taste organs
of mammals, especially with regard to development, are Lustig
(28) and Hermann (11). The former has described the de-
velopment of the taste-bulbs of man and the rabbit, and the
latter has investigated the papille, circumvallate, and foliate
of the foetal and newborn rabbit. In the papilla circumvallata
of the latter Hermann found taste-bulbs on the summit.
Boulart and Pilliet (58) have, within a short time, examined
the tongues of a number of mammals, with special reference
to the presence or absence of the papille foliatz.

Holl (15) has lately studied the taste organs of Sala-
mandra maculata. Gcblet-shaped sense organs, or end-
bulbs have been described by Leydig (25) in the skin and
mouth of various snakes, and Wiedersheim (55) says that they
are present in the lizard and blindworm on the inner sides of
the upper and lower jaws. IThlder (17) has described the end-
ing of nerve-fibres in the tongue of birds. He traced them
into oval, concentric, club-shaped bodies, like those seen by
Krause and Koélliker in the lingual papille of mammals.

Tue Toncue or MEPHITIS MEPHITICA.

The tongue about to be described was taken from quite a
young animal, and the following method was adopted in pre-
paring it for histological examination. As soon as removed
from the body it was placed in a mixture of five parts Miiller’s
fluid and one part alcohol. After remaining in this mixture
for ten days it was washed for thirty-six hours in running
water, and then transferred to strong alcohol, where the har-

TONGUE AND GUSTATORY ORGANS OF MEPHITIS MEPHITIOA. 157

dening was completed. Subsequently portions of the organ
were embedded in celloidin in the usual manner. My efforts
to obtain sections stained with chloride of gold were not
successful.

General Description of the Tongue.—The organ is
44 mm. long, 19 mm. wide, and 9mm.in thickness. It is per-
fectly free for 15 mm. from the frenum, or a trifle more than
one third of its length. Its length is therefore a little more
than twice its width, which is quite uniform from the base to
near the apex. Here it becomes thinner and somewhat ellip-
tical in form. The upper surface is slightly convex posteriorly,
and more or less flattened near the anterior extremity. There
is a very slight raphé running forwards from the middle third
of the tongue to its junction with the anterior third; here it
takes the form of a shallow groove, which disappears just before
reaching the tip. There is also a well-defined mesial groove
at the posterior part of the dorsum, commencing midway
between the two circumvallate papille, and running backwards
for 10 mm. Im the anterior third, beginning a little back
from the apex, and reaching to the lateral margin, are four or
five transverse ridges with corresponding depressions curving
backwards, which give to this part of the tongue a corrugated
appearance.

The upper surface, including the lateral margins in front of
and lying between the circumvallate papille, is covered with
tactile and mechanical papille, the points of which are directed
backwards. These become gradually smaller as they approach
the anterior extremity. The region of the tongue lying behind
the gustatory area has projecting from its surface large cone-
shaped papille, the apices of which are directed backwards and
inwards. These increase in size, but decrease in number, as
they recede towards the posterior limits of the organ.

Papille of the fungiform type are scattered quite uniformly ~
over the dorsum and upon the sides of the middle third of the
tongue, but terminate posteriorly in front of the area of the
circumvallate papillae. The average distance between them in
this region is about 0°5 mm. They are very sparingly dis-

158 FREDERICK TUCKERMAN.

tributed to the anterior third, and are much smaller than those
of the middle portion.

On each lateral half of the tongue is situated, near the ante-
rior margin of the posterior third of the papillary surface, a
large circumvallate papilla. These two papille le in the same
plane, and are 6 mm. apart. The left one is a little the larger
of the pair. They are elliptical in form, and are placed at
an oblique angle to the long axis of the tongue, with their
anterior extremity directed outwards. ach papilla is en-
circled by a deep and rather wide trench, in which it is quite
movable. Around the trench, and also forming part of its
wall, are large conical-shaped papille. The larger of the two
circumvallate papille measures, at its summit, 2 mm. in its
long diameter, and 14 mm. in its short. The upper surfaces
of both present an uneven or ridged appearance. No papilla
foliata was found.

The under surface of the tongue is perfectly smooth except
at the borders and tip, at which points are distributed nume-
rous small tactile or mechanical papille. Anteriorly it is
marked by a deep median groove, commencing at the frenum
and running towards the apex. This becomes, however, rapidly
superficial, and disappears altogether about 9°5 mm. from
the tip.

The Circumvallate Papille.—These papille are dis-
tinctly lobate posteriorly; their upper and lateral contours
are necessarily, therefore, somewhat uneven and irregular.
Still, in some sections, I found the sides comparatively symme-
trical. The diameter of the papilla at its base is always less
than at the summit. In many sections the sides are vertical,
or slightly oblique for about half their length, and then bend
inwards and downwards.

The relation of the trench to the shape of the papilla is
peculiar and, so far as I am aware, quite unusual (see figs. 2
and 3). Posteriorly, it is wide and deep, and passes directly
beneath the papilla, thus giving it (the papilla), for about half
its diameter, a free under surface.

Anteriorly and externally the trench becomes narrower

TONGUE AND GUSTATORY ORGANS OF MEPHITIS MEPHITICA. 159

(though this is not constant) and more shallow. This pedun-
culated arrangement of the base of the papilla accounts for its
free mobility, first noticed in the superficial examination. In
some sections the trench is so extremely narrow that the
surfaces of the opposing walls are almost contiguous. The
width of the trench is usually greater in its upper portion,
becoming, in most cases, gradually narrower as it curves down-
wards and inwards. ‘The outer wall reaches nearly to the level
of the upper surface of the papilla. Serous glands are present
in the body of the papilla, but both mucous and serous glands
(the latter being the more numerous) are very abundant in this
region of the tongue. The ducts of the serous glands open
into the trench, either at its sides or where it passes beneath
the papilla. They are very plentiful. In one small horizontal
section I counted twelve separate (?) ducts. Towards its
upper part the papilla carries many secondary papille, the
depressions between them being filled by epithelium. The
nerves are mainly non-medullated. They form a network in
the upper part of the papillary axis, from which branches
radiate outward towards the lateral margins, terminating, appa-
rently, at or near the bases of the taste-bulbs.

The taste-bulbs are very numerous in the circumvallate
papillae. They are distributed along the sides in a zone of ten
or twelve tiers or rows, but are most thickly placed at the
under surface of the papilla facing the bottom of the trench.
The bulbs in this situation, although protected to a remarkable
degree, are often smaller and less fully developed than else-
where. I counted here, in one horizontal section, a group of
forty-five on a surface 0°3 mm. square. The estimate of the
number of these structures cannot be very exact. In one
quarter of a horizontal section, made at the lower third of the
papilla, I counted fifty-five bulbs. If we allow 200 in each
tier, and allow ten tiers, we shall have 2000 bulbs for each
papilla, or 4000 for the two. Bulbs are also present in the
epithelium bordering the mid-trench (fig. 3) between the two
large divisions of the papilla. In one section I met with them
(fig. 3) on the free upper surface. I likewise found them in

160 FREDERIOK TUOKERMAN.

the lower half of the outer wall of the trench. In some
vertical sections they are arranged along the sides and lower
surface in several rows.

The bulbs are quite irregular in size, and exhibit some
variation in shape (fig. 5 shows the structure of the bulbs
magnified 240 diameters. Their average length is about 0:06
mm.). The neck is very short and narrow, and only in
a single bulb did I observe hair-like processes protruding
through the pore. The nuclei of the peripheral cells stain
quite deeply in hematoxylin. The outer layer of epithelium,
at the point of its perforation by the bulbs, stains a uniform _
yellow in picro-carmine. I did not succeed in identifying
gustatory as distinct from covering cells, though by teasing I
was enabled to dislodge several bulbs from their position in the
epithelium, and, in one case, to isolate a bulb with a nerve-
fibril attached to or entering its base.

The Fungiform Papille.—These papille offer nothing
very unusual in their general appearance. In shape they
resemble the human type. A variation from the normal,
however, is seen in their distribution, they being more nume-
rous and larger over the middle of the dorsal surface than
elsewhere.

In a few cases I found isolated taste-bulbs in the epithelium
at the upper part of the papilla. The best specimen is repre-
sented by fig. 6. This papilla contains two bulbs, but they
are neither of them of a high order. They are arranged
obliquely near the summit, with their apices directed outwards
and upwards, and measure about 0°05 mm. in length, and
0:032 mm. in breadth. In none of my sections do they appear
to reach the surface of the epithelium.

Non-medullated nerve-fibres are quite abundant in the upper
part of the body of the papilla, and nerve-fibrils can be seen
running directly beneath the epithelium containing the taste-
bulbs. Beyond this point I was unable to trace them. A few
collections of ganglion-like cells are scattered along the course
of the nerves. In some of the papille the nerve-fibres
terminate in end-bulbs, as already pointed out by Krause and

TONGUE AND GUSTATORY ORGANS OF MEPHITIS MEPHITIOA. 161

Kolliker. Neither serous nor mucous glands were observed
near the fungiform papille.

The tactile and mechanical papille are very numerous,
covering the upper surface of the tongue from the base to the
apex. They are largest at the posterior part of the dorsum
(behind the gustatory structures), and diminish in size, but
increase in number and density, as they approach the anterior
extremity. One that I measured, from the posterior part of
the dorsum, was 1‘8 mm. in height. About the middle of the
tongue I counted twenty-five on a square millimetre of surface.
Here they are 0°25 mm. in height. They present considerable
variation in shape. Anteriorly they are flattened, or slightly
convex on the top, with their sides vertical, forming either a
right angle with the upper surface, or having their upper edges
slightly rounded. Occasionally the sides are prolonged
upwards for a short distance, terminating in spiniform pro-
cesses. Interspersed among these papille, but chiefly confined
to the gustatory area and posterior surface, are a few cone-
shaped ones, the points of which are directed backwards and
inwards. Hach papilla is usually seated upon two papillary
upgrowths of the mucosa. The free surface is covered with a
thick layer of cornified epithelium, which, in the cone-shaped,
papille, presents an imbricated arrangement. In their internal
structure these papille do not differ materially from ordinary
conical papille.

It is probable that the cone-shaped papille of the anterior
and middle dorsal surface of the tongue are mechanical rather
than tactile in function.

On the posterior surface of the epiglottis, near the line of
union of the first and second fourth, I noticed in the stratified
pavement epithelium a few isolated bulb-like structures, I
did not, however, meet with them below this point. In the
same region, on the right side, I found the bulb-like structure
shown in figure 7. It will be seen at once that it is entirely
subepithelial in position and structure. It occupies a cavity
of the mucosa, with its apex resting against the base of the
deep layer of columnar cells of the epithelium. Its length is

VOL, XXVII1, PART 1,—NEW SER, L

162 FREDERICK TUCKERMAN.

about 0:045 mm., and its greatest transverse diameter about
0:025 mm.

I examined a very large number of sections of the soft
palate and uvula, but did not succeed in finding any bulb-like
structures.

To sum up briefly in conclusion, there are situated at the
posterior part of the tongue of Mephitis two large circumvallate
papillee. Upon superficial examination the most striking features
are their size, the ridged appearance of their upper surface,
and their rather unusual shape. Under the microscope each
papilla is seen to be divided posteriorly into two unequal lobes
or divisions, and to have in the same region a free under sur-
face, the trench passing directly beneath its base, thus afford-
ing great protection to the bulbs occurring here. Anteriorly
the papilla is connected with the tongue by a pedicel-like
attachment.

The taste-bulbs of the circumvallate papille are very
numerous, especially in the epithelium of the under surface,
and offer considerable irregularity in shape, size, and distribu-
tion. A few bulbs occur in the epithelium of the upper sur-
face of the papillz, and in that of the outer wall of the trench.
In a single instance I observed a bulb with a nerve-fibril at-
tached to its base. Glands of the serous and mucous types
are very abundant, but the former are chiefly limited to the
gustatory area of the tongue. Serous glands are also met with
in the papillary body itself.

A few fungiform papillz possess isolated bulbs lying in the
epithelium at their summit, and they also occur in the upper
part of the posterior surface of the epiglottis; but in both
these regions they are primitive in character and position.

TONGUE AND GUSTATORY ORGANS OF MEPHITIS MEPHITICA, 163

—

il.

12.

13.

14,

LITERATURE REFERRED TO.

. v. Avtar, A. K.—‘‘ Ein Beitrag zur Kenntniss der Geschmacksorgane,”

‘Arch, f. mikr. Anat.,’ Bd. vii, 1872, 8. 455.

. Beate, L. 8.—‘ New Observations upon the Minute Anatomy of the

Papille of the Frog’s Tongue,” ‘ Philos. Trans. Roy. Soc.,’ 1865, p.
443, plates xxi and xxii.

. Bratz, L. S.—‘* New Observations upon the Minute Anatomy of the

Papille of the Frog’s Tongue,” ‘Quart. Journ. Micr. Sci.,’ vol. ix,
1869, p. 1, plates i—iv.

. Britrotu, T.—“ Ueber die Epithelialzellen der Froschzunge, sowie tiber

den Bau der Cylinder und Flimmerepithelien und ihr Verhaltniss zum
Bindegewebe,” ‘Arch. f. Anat. u. Physiol.,’ 1858, 8. 159, Taf. vii.

. Davis, C.—** Die becherformigen Organe des Kehlkopfs,” ‘ Arch f. mikr,

Anat.,’ Bd. xiv, 1877, S. 158, Taf. x.

. v. Epner, V.—‘ Die acinésen Driisen der Zunge und ihre Beziehungen

zu dem Geschmacksorgane,’ Graz, 1873.

. Encetmann, To. W.—“ Ueber die Endigungsweise der Geschmacksnerven

des Frosches,”’ ‘ Céentralb. f. d. med. Wiss.,’ Nr. 50, 1867, 8. 785.

. ENGELMANN, TH. W.—“ Ueber die Endigungen der Geschmacksnerven

in der Zunge des Frosches,” ‘ Zeitschr. f. wiss. Zool.,’ Bd. xviii, 1868,
S. 142, Taf. ix.

. Eneztmann, To. W.—Article, “The Organs of Taste,” ‘ Stricker’s

Manual of Histology,’ New York, 1872, p. 777.

. Hartmann, R.—* Ueber die Endigungsweise der Geschmacksnerven in

der Zunge des Frosches,” ‘Arch. f. Anat. u. Physiol.,’ 1863, 8. 634,
Taf, xvii u. Xvill.

HERMANN, F,.—“ Beitrag zur Entwicklungsgeschichte des Geschmacks-
organs beim Kaninchen,” ‘ Arch. f. mikr, Anat.,’ Bd. xxiv, 1884,
S. 216, Taf. xiii.

HoéyicscumieD, J.— Hin Beitrag iiber die Verbreitung der becher-
formigen Organe auf der Zunge der Saugethiere,” ‘ Centralb. f. d. med.
Wiss.,’ Nr. 26, 1872, 8. 401.

HoniescuMieD, J.— Beitrage zur mikroskopischen Anatomie tber die
Geschmacksorgane der Siugethiere,” ‘Zeitschr. f. wiss. Zool., Bd.
xxiii, 1873, S. 414, Taf. xxiv.

Horrmann, A.—‘‘ Ueber die Verbreitung der Geschmacksorgane beim
Menschen,” ‘ Virchow’s Arch.,’ Bd. Ixii, 1875, 8. 516, Taf. xi,

164, FREDERICK TUOKERMAN.

15.

16.
17.
pike

19:

20.
21.
22.
23.
24.
25.
26.

27.

28.

29.

30.

Hout, M.—‘ Ueber das Epithel in der Mundhéhle von Salamandra
maculata,” ‘Sitzb. d. k. Akad. d. Wiss. Wien,’ Bd. xcii, Abth. iii,
1885, 8. 161, i Tafel.

Hoyer, H.—“ Mikroskopische Untersuchungen iiber die Zunge des
Frosches,” ‘ Arch. f. Anat. u. Physiol.,’ 1859, S. 481.

IntpEr.—< Die Nerven-Endigung in der Vogelzunge,” ‘ Arch. f. Anat.
u. Physiol.,’ 1870, p. 238.

JourpaNn, E.—* Sur les organes du géut des Poissons osseux,” ‘ Comptes
Rendus,’ T. xcii, 1881, p. 743.

Key, H. Axgx.—‘ Ueber die Endigungsweise der Geschmacksnerven in
der Zunge des Frosches,” ‘Arch. f. Anat. u. Physiol.,’? 1861, S. 329,
Taf. viii.

Kravsz, W.—“ Die Nervenendigungen in der Zunge des Menschen,”
‘ Gottinger Nachrichten,’ 1870, 8. 423.

Kravusz, W.—‘ Allgemeine und mikroskopische Anatomie,’ Hannover,
1876.

Letzertcu, L.—‘ Ueber die Endapparate der Geschmacksnerven,”
‘Virchow’s Arch.,’ Bd. xlv, 1869, 8. 9, Taf. i.

Lerypie, F.—‘“ Ueber die aussere Haut einiger Siisswasserfische,”
‘ Zeitschr. f. wiss. Zool.,’ Bd. iii, 1851, 8. 1, Taf. i.

Lrypie, F—‘ Lehrbuch der Histologie des Menschen und der Thiere,’
Frankfurt, 1857.

Leyopie, F.—“ Zur Kenntniss der Sinnesorgane der Schlangen,” ‘ Arch.
f. mikr. Anat.,’ Bd. viii, 1872, 8. 317, Taf. xv u. xvi.

Lovin, Cur.— Bidrag till kannedomen om Tungans Smakpapillen,”
* Medicinskt Archiv,’ Stockholm, iii, 1867, 8S. 9.

Lovin, Cur.— Beitrige zur Kenntniss vom Bau der Geschmacks-
warzchen der Zunge,” ‘Arch. f. mikr. Anat.,’? Bd. iv, 1868, S. 96,
Taf. vii.

Lustie, A.— Beitrage zur Kenntniss der Entwicklung der Geschmacks-
knospen,” ‘ Sitzb. d. k. Akad. d. Wiss. Wien,” Bd. Ixxxix, Abth. iii,
1883, 8. 308.

Mavpox, R. L.— A Contribution to the Minute Anatomy of the Fungi-
form Papillee and Terminal Arrangement of Nerve to Striped Muscular
Tissue in the Tongue of the Common Frog,” ‘Monthly Microsc.
Journ.,’ vol. i, 1869, p. 1, pl. i.

Pouttoy, HE. B.—“ The Tongue of Perameles nasuta, with some
Suggestions as to the Origin of Taste-Bulbs,” ‘ Quart. Journ. Micr.
Sci.,’ vol. xxiii, 1883, p, 69, pl. i.

4

TONGUE AND GUSTATORY ORGANS OF MEPHITIS MEPHITIOA. 165

31

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42,

43,

44.

45.

46.

. Poutroy, E. B.—“The Tongue of Ornithorhynchus paradoxus:

the Origin of Taste-Bulbs and the Parts upon which they occur,”
‘Quart. Journ. Mier. Sci.,’ vol. xxiii, 1883, p. 453, pl. xxxii.

Ranvier, L.— Traité technique d’histologie,’ Paris, 1882, p. 949.

Scuorienp, R. H. A.—“ Observations on Taste-Goblets in the Epiglottis
of the Dog and Cat,” ‘Journ. of Anat. and Phys.,’ vol. x, 1876,
p. 475, pl. xxvii.

Scuuuzze, F. E.—‘‘ Ueber die becherfoérmigen Organe der Fische,”
‘Zeitschr. f. wiss. Zool.,’? Bd. xii, 1863, S. 218, Taf. xxiii.

Scuuuze, F. E.— Epithel- und Driisenzellen,” ‘Arch. f. mikr. Anat.,’
Bd. iii, 1867, S. 137, Taf. vi—xii.

Scuuuze, F. E.—‘‘ Die Geschmacksorgane der Froschlarven,” ‘ Arch. f.
mikr. Anat.,’ Bd. vi, 1870, S. 407, Taf. xxii.

ScuwatBe, G.—* Das Epithel der Papille Vallate,” ‘Arch f. mikr.
Anat.,’ Bd. iii, 1867, S. 504.

Scuwa.sBe, G.—‘‘ Ueber die Geschmacksorgane der Saugethiere und des
Menschen,” ‘ Arch. f. mikr. Anat.,’ Bd. iv, 1868, S. 154, Taf. xii
u. xiii.

Scuwa.BeE, G.—“ Zur Kenntniss der Papilla Fungiformes der Saugethiere,”
Centralb. f. d. med. Wiss.,’ Nr. 28, 1868, S. 433.

Scuwa.se, G.—‘ Lehrbuch der Anatomie der Sinnesorgane,’ Erlangen,
1883, S. 38.

Sexton, E.—“ Osserrazioni sulle terminazioni dei nervi del gusto,”
* Gazzetta Medico-Veterinaria,’ anno iv, Abth. in ‘ Centralb. f. d. med.
Wiss.,’? Nr. 55, 1874, S. 871.

Srmanowsky, N.—“‘Beitrage zur Anatomie des Kehlkopfs,” ‘ Arch. f.
mikr. Anat.,’ Bd. xxii, 1883, S. 690, Taf. xxvi.

SzaBaDFOLDy, M.— Beitrage zur Histologie der Zungenschleimhaut,’
‘ Virchow’s Arch.,’ Bd. xxxviii, 1867, 8. 177, Taf. iv.

Toparo, F.—“ Die Geschmacksorgane der Rochen,” ‘ Centralb. f, d. med.
Wiss.,’ Nr. 15, 1872, 8S. 227.

Verson, E.—“ Beitrage zur Kenntniss des Kehlkopfes und der Trachea,”
‘ Sitzb. d. k. Akad. d. Wiss. Wien.,’ Bd. lvii, Abth. 1, 1868, S. 1093,
i Tafel.

Vexrson, E.—Article, “Larynx and Trachea,’’ ‘Stricker’s Manual of
Histology,’ New York, 1872, p. 428.

47, v. VintscuGan, M., and Honiescumizp, J.—“ Nervus Glossopharyngeus

48

und Schmeckbecher,” ‘ Pfliiger’s Arch.,’ Bd. xiv, 1876, S. 443.
. v. Vintscuean, M,—Article, “ Physiologie des Geschmackssinus und des

166 FREDERICK TUCKERMAN.

49

50.

51.

52.

53.

54.

55.

56.

57.

58.

Geruchssinus,” ‘ Hermann’s Handbuch der Physiologie,’ Bd. iii, Th. 2,
1880, S. 145.

v. Vintscuean, M.—“Beobachtungen iiber die Veranderungen der
Schmeckbecher nach Durchschneidung des N. Glossopharyngeus,”
‘ Pfliiger’s Arch.,’ Bd. xxiii, 1880, S. 1, Taf. i.

Water, A.—“ Microscopic Examination of some of the Principal
Tissues of the Animal Frame, as cbserved in the Tongue of the Living
Frog, Toad, &c.,” ‘Philosophical Magazine,’ vol. xxix, 1846, p. 271,
plate i.

Watter, A.—‘ Microscopic Examination of the Papille and Nerves of
the Tongue of the Frog, with Observations on the Mechanism of
Taste,” ‘ Philosophical Magazine,’ vol. xxx, 1847, p. 277, pl. iii.

Water, A.— Minute Structure of the Organ of Taste in Vertebrate
Animals,” ‘ Proc. Roy. Soe.,’ vol. v, 1848, p. 751.

Water, A.— Minute Examination of the Organ of Taste in Man,”
* Proc. Roy. Soc.,’ vol. v, 1849, p. 803.

Water, A.—“ Minute Structure of the Papille and Nerves of the
Tongue of the Frog and Toad,” ‘Philos. Trans. Roy. Soc.,’ 1849,
p. 189, pl. xii.

WIEDERSHEIM, R.—‘ Elements of the Comparative Anatomy of Verte-
brates,’ London, 1886, p. 166.

v. Wyss, H.—“ Ueber ein neues Geschmacksorgane auf der Zunge des
Kaninchens,” ‘Centralb. f. d. med. Wiss.,’ Nr. 35, 1869, 8. 548.

v. Wyss, H.—“ Die becherférmigen Organe der Zunge,” ‘ Arch. f. mikr.
Anat.,’ Bd. vi, 1870, S. 237, Taf. xv.

Bovnart, R., and Prruret, A.— Note sur l’organe folié de la langue
des mammiféres,” ‘J. de Anat. et Physiol.,’ vol. xxi, 1885, p. 337.

TONGUE AND GUSTATORY ORGANS OF MEPHITIS MEPHITIOA. 167

EXPLANATION OF PLATE XI,

Illustrating Mr. Frederick Tuckerman’s paper on “ The Tongue
and Gustatory Organs of Mephitis mephitica,”’

Fic. 1.—Vertical section through the right circumvallate papilla. S. P.
Secondary papille. 7¢ The trench. 7. Section through the ridge which
surrounds the trench. 7.4. The taste-bulbs arranged in tiers. gl. d. The
ducts of the serous glands, opening into the bottom of the trench. m. m,
Mucous membrane. (xX 20 times.)

Fic. 2.—Vertical section through the same circumvallate papilla. §. P.
Secondary papille. 7. The trench. 7. Ridge. C. P. Conical papilla, some-
what depressed. 7’. The trench passing beneath the papilla, showing the
arrangement of the taste-bulbs in the epithelium of the under surface. g/.d.
The ducts of the serous glands. (x 24 times.)

Fic. 3.—Vertical section through the posterior part of the same papilla.
t. The trench. 7. 4’. Taste-bulbs in the epithelium of the upper surface.
m.t. The mid-trench, partly separating the papilla into two divisions. g/l. d
Ducts of the serous glands. (x 20 times.)

Fic. 4.—Vertical section through the anterior part of the same papilla.
t. 6. Taste-bulbs, divided at right angles to their long axis. gi. d. Ducts of
the serous glands. (xX 30 times.)

Fic. 5.—Vertical section through the base of the same papilla, showing
the bottom of the trench and the six lowest tiers of taste-bulbs. ¢. The
trench. 7.4. Taste-bulb. g.p. Gustatory pore. ss. e, Stratified epithelium.
o. 2. Outer layer of stratified epithelium. g/.d. The ducts of the serous
glands. m.m. Mucous membrane. (xX 240 times.)

Fie. 6.—Vertical section through a fungiform papilla, from the group in
front of the circumvallate papille. 7. 4. Taste-bulbs. y. p. Papillary pro-
cesses. s.¢. Stratified epithelium. (x 220 times.)

Fic. 7.—Transverse vertical section through the upper part of the posterior
surface of the epiglottis. ..s. Free surface of the stratified pavement of
epithelium. 4.7. Deep layer of columnar cells of the epithelium. 4. J, s.
Bulb-like structure lymg in the mucosa, with its apex touching the base of
the cells of the deep layer of epithelium. mm. m. Mucous membrane. (x 240
times.)

Be ‘

K.I.Hall del.

Mor Lourn'Ve, KOUNS PM

F Huth, Lith? Edint

ON THE QUADRATE IN THE MAMMALIA. 169

On the Quadrate in the Mammalia.!
By

Dr. G. Baur,
of New Haven, Conn.

In what skeletal piece in the Mammalia is the quadrate of
the Sauropsida and Ichthyopsida to be looked for? is a ques-
tion which has received very varied answers.

During the last few years several works have appeared which
attempt to answer this question, and that ina new way. There
are several by Albrecht,” and one by Dollo.?

I will first of all set myself to discuss shortly Albrecht’s
researches.

He brings forward the various views as to the mode of

1 Translated by Wm. B. Benham, B.Sc., from the ‘ Biolog. Centralblatt,’ vi,
(1886), pp. 648—658.

2 P. Albrecht, ‘Sur la valeur morphologique de l’articulation mandibulaire,
du cartilage de Meckel et des osselets de l’orice, avec essai de pruver que
Pécaille du temporal des mammiféres est composée primitivement d’un squa-
mosal et d’un quadratum,’ Bruxelles, 1883.

Ibid. ‘Sur le crane remarquable d’une idiote de 21 ans,’ Bruxelles, 1883.

Ibid, ‘Sur le valeur morphologique de la trompe d’Eustache, et les dérivés
de l’are palatin, de arc mandibulaire, et de l’arc hyoidien des vertébrés,’
Bruxelles, 1884.

Extracts from these works are found in:

P. Albrecht, ‘‘ Ueber den morphologischen Werth der Gekérknéchelchen und
des Unterkiefergelenkes der Wirbeltiere,” ‘ Official Report of the Fifty-sixth
Meeting of German Naturalists and Doctors,’ Freiburg, 1884, p. 148.

Ibid. “‘ Ueber den morphol. Werth des Unkerkiefergelenkes, der Gehérk-
néchelchen, und des mittlern, und dussern Ohres der Saugetiere.” ‘Report of
the Third International Otological Congress at Basel, 1884,’ Basel, 1885, p. 8.

3 L. Dollo, “On the Malleus of the Lacertilia and the Malar and Quadrate
Bones of Mammalia,” ‘ Quart. Journ, Mic. Sci.,’ October, 1883.

170 DR. G. BAUR.

articulation of the lower jaw of Vertebrates, and groups them
in the following table:

Articulation of the Lower Jaw in Vertebrates,

Articulation of the Lower Jaw
except Mammals. i

in Mammals.

Quadrato-articulare articulation. | Squamoso-articulare (Huxley).
| Squamoso-dentary (Gegenbaur, Kél-
liker, Wiedersheim).

Consequently, whilst Huxley assumes the lower jaw of the
lower Vertebrates to be homologous with that in the Mammalia,
Gegenbaur, Kolliker, and Wiedersheim see in the lower jaw of
Mammals only the dentary piece of that in the remaining
Vertebrates.

In the next table Albrecht sets forth the various views as to
the development of the auditory ossicles in the Mammalia.

I.—Visceral Arch. I1.—Visceral Areh. Auditory Canal.
Reichert .| Malleus . incus Stapes
Giinther .| Malleus . incus . stapes se
Os lenticulare
Malleus=articulare (symplectic)
Gegenbaur { Incus =quadrate Stapes=hyomandi- oe
bular
Huxley . . Malleus=quadrate Incus =hyomandi-
bular, os lenticulare,
stapes
Parker’). )< 9 a » 4 sus
Parker . cus =hyomandi-
Bettany. { bular } Stapes.
Salensky
Ist Theory Malleus . incus . stapes
Salensky i
2nd Theory Malleus . incus os Stapes (outside
the periarterial
a tissue).
ae _Malleus=articulare
Kolliker. 9 Tous = quadrate } Stapes,
Wiedersheim = 3 is
Fraser . «| Malleus Incus=oslenticulare| Stapes (outside
the periarterial
tissue).

1 A. Fraser, “Onthe Development of the Ossicula auditus of the Higher
Mammalia,” ‘ Phil. Trans.,’ vol. 173, part iii, 1882.

————

ON THE QUADRATE IN THE MAMMALIA. 17}

Gegenbaur, Kolliker, Wiedersheim, therefore, regard the
malleus as the articulare, and the incus as the quadrate.

Huxley, Parker, and Bettany look upon the malleus as the
quadrate and the incus as the hyomandibular.

The incudo-mallear articulare articulation, therefore, is to
the former zoologists a quadrato-articulare articulation, and
to the latter it is a hyomandibular-quadrate articulation.

Albrecht espouses neither the one view nor the other, and
arrives, by his reasoning, at the result that in all Vertebrates
the lower-jaw articulation is the same, viz. a quadrato-articu-
lare articulation. He then reviews the relations of the auditory
ossicles. In the Sauropsida, Cecilia, and Urodela there is a
columella. It commences at the tympanic membrane and ends
at the “ membrana ovalis” (as Albrecht names the membrane
which closes the fenestra ovalis). In the Anura four more or less
ossified pieces of cartilage are found in the same position ; these
four pieces of cartilage are homologous with the columella.
In the Mammalia the malleus touches the tympanic membrane ;
the stapes reaches the fenestra ovalis. Therefore Albrecht
concludes, “the columellais homologous with the row
of auditory ossicles in the Mammalia. The columella
unquestionably forms the suspensorium of the lower jaw. The
malleus of the Mammalia belongs to the extra-mandibular
portion of Meckel’s cartilage, but this portion is homologous
with the symplectico-articular ligament of the Amphibia and
Sauropsida, therefore the suspensorium of the lower jaw is the
same as in all the Vertebrates.

If, now, the articulation of the lower jaw in Mammals is
homologous with that in the rest of the Vertebrates, in which .
a quadrato-articular articulation is present, then the quadrate
must be looked for in that part of the Mammalian skull with
which the lower jaw articulates. This, in the Mammalia, is
the squamosal: consequently the quadrate of the Sau-
ropsida and Ichthyopsida must be comprised in the
squamosal,

Albrecht, indeed, finds in a newborn child affected with
double harelip and double “ Wolfsrachen,” that this squa-

172 DR. G. BAUER.

mosal is divided into two portions: the zygomatic process and
the ‘‘scale.’” In the zygomatic process Albrecht sees
the quadrate of the rest of the Vertebrates. He finds
the same conditions in a newborn horse, and in the squa-
mosal of the right side of an idiot twenty-one years old.

Dollo uses the same arguments: first recapitulates Albrecht’s
results, and then describes his own researches. He asks the
question: Is it possible that the quadrate can form a part of
the interfenestral chain of auditory ossicles? If we succeed
in finding a Vertebrate in which the lower jaw consists of the
six normal elements, and in which, moreover, a true quad-
rate and a malleus are present, then it is impossible that the
quadrate can be homologous with any of the auditory ossicles.
Therefore—

(1) It cannot be compared with the malleus, since this is
present, and (2) it would be impossible to identify it with
one of the remaining auditory ossicles, because it would be
outside the malleus, and touch none of the other auditory
ossicles.

It depends, therefore, on finding a malleus which shall fulfil
the above conditions. Dollo maintains that he has found in
many Lacertilia (Leiolepis, Uromastix, and their allies)
a skeletal piece which has the morphological value of a malleus.

Dollo’s argumeuts in support of this are:

1. The piece has the form of a malleus, and all the charac-
teristic parts of one can be distinguished.

2. The piece has the same connections: it is fixed to the tym-
panic membrane in such a way that the manubrium is parallel
to the membrane. At the side, in the region of the cervix, it
is connected by cartilage with the rest of the auditory ossicles ;
and with the quadrate it stands in the same relation as the
malleus of Mammals with Albrecht’s “ quadrate.”

3. It is connected with the articulare of the lower jaw by a
malleo-articulare ligament—Albrecht’s extra-mandibular por-
tion of Meckel’s cartilage.

4, It can scarcely be doubted that this malleus is identical
with that described by Peters for the crocodiles.

ON THE QUADRATE IN THE MAMMALIA. 173

5. The malleus in the Mammalia serves for the insertion of
the tensor tympani: the same, according to Parker, should be
the case with the “ malleus” of the Lacertilia.

These are Dollo’s arguments, and he concludes in the follow-
ing terms: “TI believe, therefore, that I have found in the
Lacertilia a true malleus, which is homologous with that of
the Mammalia, and that we have a support for Albrecht’s
theory. The columella of the Sauropsida, therefore, will not be
homologous with the malleus and incus, and articulare and
stapes, but, as Albrecht thinks, only with the three last pieces.
Albrecht later denoted the homology of these three pieces by
the name “ Columellina.”

Let us now subject Albrecht’s work to a short examination,
First of all, it must be remarked, that his view—that the
quadrate of the Sauropsida is homologous with the zygomatic
process of mammals—is not absolutely new. Indeed, in
1810, Tiedemann says, in his well-known work, ‘ Anatomie
und Naturgeschichte der Vogel,’ vol. i, p. 191, “The two
quadrate bones of birds are analogous with the articular region
of the squamosal of man and mammals, namely, with the
glenoid fossa, the glenoid process, and the zygomatic process
of the squamosal, which have become detached from the
Squamosal as a separate bone.” Further, Platner! upheld
this view in the most decided manner.

Kostlin? is also of this opinion.

Again, the possibility of the separating of the zygomatic
process from the scale of the squamosal was recognised before
Albrecht. Duvernoy adduced an instance of this sort in the
second edition of Cuvier’s ‘ Legons d’Anatomie Comparée,’
vol. iv, p. 98: “We are led to compare the quadrate bone
to that portion of the temporal which furnishes the glenoid
fossa, and we base this comparison on the fact that this
portion of the temporal is separated from the petrous and

1 F, Platner, ‘ Bermerkungen iiber das Quadratbein und die Paukenhéhle
der Vogel,’ Dresden and Leipzig, 1839.

2 QO. Kostlin, ‘Der Bau des Knockernen Kopfes in den Vier Klassen der
Wirbelthiere,’ Stuttgart, 1884, pp. 212, 213.

174 DR. G. BAUR.

periotic, as well as from the scale of the temporal, in a skull
of Hydrocherus, which we have under observation.”

Thus Duvernoy, more than forty years ago, partly
on the same grounds as Albrecht, arrived at the
same conclusion as he did.

I now turn to Dollo’s researches. He says: ‘‘ It is sufficient
for me to find in Uromastix and its allies a malleus, which
is homologous with that of the Mammalia.”

Unfortunately, I cannot grant to Dollo the right of having
just discovered this fact. This belongs to Peters,' who had,
indeed, nearly ten years previously, found this, and very dis-
tinctly, in Uromastix.

Afterwards, Peters pointed out that in Sphenodon a true
malleus is present, which Huxley? had interpreted as the outer
stapes-cartilage.

Peters says, on pages 43, 44 of his paper: “ For the pur-
pose of this research I availed myself, for comparison, of an
exampleof Uromastix spinipes from Egypt; in this reptile,
the relation of the cartilage, described by me asa malleus, with
the lower jaw or Meckel’s cartilage, remains so very distinct
without any preparation that anyone will easily be able to
form an opinion on the point in question by simply examining
this very common form, which ought scarcely to be absent
in any collection. When the head has been removed the stapes
will readily he seen lying exposed, in the same way as in
Sphenodon, by the side of the exoccipital bone. But in
Uromastix it does not lie so near this bone as in Sphenodon,
and especially at its outer end, where it passes round under the
inner edge of the quadrate bone to unite itself in an articular
pit, with the head of the cartilaginous malleus. The body of
the malleus consists of a cylindrical piece, which fixes itself at
the tympanum, and here ends in a small plate, the longer part
of which is directed forwards, whilst the shorter end approaches

1 'W. Peters, “ Uber die Gehorknockelchen und ihre Verhaltniss zu den
ersten Zungenbeinbogen bei Sphenoden punctatus,” ‘ Monatsber. d. k.

preuss. Akad. d. Wiss.,’ Berlin, 1874, p. 40.
2 “On the Malleus and Incus, &c.,” ‘ Proc. Zool. Soc.,’ Lond., 1869.

ON THE QUADRATE IN THE MAMMALIA, 175

the edge of the ‘mastoid’ bone. But at the point where the
malleus is connected with the stapes, there springs at right
angles to it, downwards and forwards, a longer process (pro-
cessus longus mallei), which descends along the inner side of
the quadrate, passing between it and the hinder end of the
pterygoid and becomes tendinous in front of the inner
edge of the articular cavity of the lower jaw, into which it
sinks.”

Peters gives drawings of these relations. That this work by
Peters should escape Dollo is so much the more strange, since
Balfour, in his ‘ Comparative Embryology,’ vol. ii, p. 483 (foot-
note), cites him, and Hoffmann, in his ‘ Reptilia’ (Bronn’s
‘Klassen und Ordnungen des Tierreichs’), p. 605, not only
gives Peters’s contribution verbally, but also copies his figures.

Moreover, this is not the only work by Peters in which the
view that the Sauropsida possess a malleus homologous with
that of the Mammalia is upheld. Indeed, in the work quoted
by Dollo,' Peters speaks very decidedly on this point.

He found in a young alligator, with a head 13 ecm. in length,
a cartilaginous thread lying in a membranous sheath, which
proceeds from Meckel’s cartilage of the lower jaw through the
aperture which exists at the hinder inner part of the upper
surface of the quadrate. This thread he not only traced to the
hinder border of the tympanic membrane, but also convinced
himself that it is in connection with the cartilaginous plate,
which is bent inwards by its narrow middle part towards the
columella, the outer end of which is articulated with it. This
cartilaginous plate, Peters insists, is nothing but the malleus,
as, indeed, Breschet had interpreted it in birds.

Peters was able to see these relations still more clearly in an
embryo crocodile of 70 mm. in length. The same condition
of things was also found in an ostrich embryo. On p. 595
Peters expresses himself very decidedly and clearly: “ In view
of the observed facts the view that the articular portion of the

* W. Peters, “ Ueber die Gehérknockelchen und den Meckel’schen Knorpel

bei den Krokodilen,” ‘ Monatsber. d. k. preuss. Akad. d. Wiss.,’ Nov., 1868,
p. 592.

176 DR. G. BAUR.

lower jaw and the quadrate of Amphibia are homologous
with the malleus and incus of the Mammalia loses every
foundation.”

In a later communication (“ Ueber der Gehérknockelchen
der Schildkréten, Eidechsen, und Schlangen,” ‘ Monatsber. d.
Ber]. Akad.,’ January, 1869) Peters states that in an embryo of
Hemidactylus the cartilaginous thread passing from the
malleus bends round close to the quadrate, and sinks into the
lower jaw. Therefore Peters recognised the malleus in the
Sauropsida long before Dollo.

I pass now to my own researches on the subject. As is well
known, the theory is to-day nearly universally taught, especially
in England, that the columella “and its appendices” are modi-
fications of the second, and not of the first, visceral arch. Thus
Parker! has lately written, “ After long years of labour and
much vacillation of mind on the matter, I am now quite satis-
fied that the stapes, a little stirrup-bone of the ear-drum, is the
uppermost element of the second or hyoid arch.”

To these views Huxley’s* researches on the stapes of
Sphenodon especially contributed.

According to Huxley the hyoid cartilage rises up behind the
quadrate till it has nearly reached the skull, and then appears
suddenly to be bent in the form of a small scroll with a pos-
terior concavity. This scroll is due to the widening out of the
hyoid bone, which forms a cartilaginous plate. On the inner
side this plate is produced into the stem (base) of the stapes,
and soon ossifies. According to Huxley, therefore, the upper
stapes-cartilage is nothing else than the inner end of the hyoid
arch. The stapes and its appendices belong absolutely to
this arch, and have nothing whatever to do with the mandibular
arch.

On the other hand, Peters says: The connection of the hyoid
arch with the stapes-cartilage (malleus) is not a primary but

1 W. K. Parker, ‘On Mammalian Descent,’ London, 1885, p. 43.

2 T. H. Huxley, “ On the Representatives of the Malleus and Incus of the
Mammalia in the other Vertebrates,” ‘Proc. Zool. Soc.,’? London, 1869,
p. 391.

ON THE QUADRATE IN THE MAMMALIA, 177

a secondary condition. The hyoid arch applies itself to the
malleus, is bound to it by connective tissue, and perhaps even
fuses with it. The fibres of the hyoid arch are soft, and have
a different direction from those of the stapes-cartilage (malleus),
the harder fibres of which cross those of the hyoid arch. The
swelling of the hyoid arch at the point where it joins the outer-
most part of the malleus is only apparent, arising not from the
cartilage but from the connective tissue. According to Peters,
the hyoid bone is, indeed, not connected with the inner process
of the malleus, but passes below it, without adhering to it, so
that the space between the outer and inner parts of the malleus
is not, as Huxley thought, transformed into a foramen by a
junction. According to him there is no doubt that at earlier
stages Meckel’s cartilage is connected with this inner hatchet-
shaped process of the malleus by means of a fibre passing along
the inner side of the quadrate.

According to Peters, therefore, the stapes-cartilage, i.e.
the malleusof Sphenodon, is derived from the first
visceral arch.

It is seen that the opinions on this very important point are
very various. It is strange that Parker, too, in his many
works on the development of the skull in Vertebrates, makes
no mention of this work by Peters. Besides Hoffmann,
Balfour mentions this point. He says (l.c., p. 483), “ The
strongest evidence in favour of Huxley’s and Parker’s view of
the nature of the columella is the fusion in the adult Sphenodon
of the upper end of the hyoid with the columella (Huxley).
From an examination of a specimen in the Cambridge museum
I do not feel satisfied that the fusion is not secondary, but
I have not been able to examine the junction of the hyoid and
columella in section.”

Balfour was inclined to adopt Peters’s view: I can do the
same. In fact, Peters is right. The malleus (stapes-
cartilage) is not derived from the hyoid arch: the
connection with it is secondary: the malleus of
Sphenodon and all Sauropsida is a derivative of the
first visceral arch.

VOL. XXVIII, PART 1.—NEW SER. M

178 DR. G. BAUR.

My material consisted of three specimens of Sphenodon,
preserved in alcohol: for two of these (a and 4), from the
Museum of Yale College, I have to thank Mr. O. C. Marsh;
the third (ce) came from Prof. B. G. Wilder, of Ithaca.
The specimen @ measured about 360 mm., the tail being
regenerate ; 6 measured 290 mm., and c 210 mm. In the
specimen a I have dissected out on both sides the part con-
cerned, of 6 and ¢ only that of the right side. Examined with
a lens the hyoid arch is seen to be close to the cartilaginous
part of the stapes ; in fact, was in part fused with it. In order
to be quite certain of this, series of sections from the prepara-
tions of a@ and 6 were prepared. These show that the hyoid
arch is free from the real malleus, although it applied itself
closely to the front edge of the stapes-cartilage. The relations
in section are, certainly, not so distinct as I had expected;
and the examination of Sphenodon alone had made the
settlement of the point impossible. But that the hyoid has,
in fact, nothing to do with the stapes is very clearly seen in
Tarentola annularis (Platydactylus egyptiacus).
Here the hyoid arch is just as perfect as in Sphenodon, but
it does not enter into so intimate a connection with the stapes-
cartilage. From the long process of the malleus (‘ infrasta-
pedial” of Parker), however, there passes downwards a thin
fibre, to sink into the lower jaw; this is the epimandibular
portion of Meckel’s cartilage (“ceratohyal” of Parker).
Here, therefore, we have a similar condition to that described
by Parker! for the crocodile.

From the observed facts, there is no doubt that the malleus
of Sphenodon, and therefore of ali the Sauropsida, is not derived
from the hyoid arch, but from the mandibular arch. Albrecht?
has already maintained on logico-theoretical grounds, that the
wrongly so-called hyomandibular (ceratohyal) is nothing else
but the dorsal portion of the first visceral arch, that is of
Meckel’s cartilage. My own researches on Sphenodon, and

1 W. K. Parker, ‘‘On the Structure and Development of the Skull in the
Crocodilia,” ‘ Trans. Zool. Soc.,’ vol. xi, 1883, pl. 68, fig. 19.

2 P. Albrecht, ‘Sur la valeur morphologique de la trompe d’Eustache,’
Bruxelles, 1884.

ON THE QUADRATE IN THE MAMMALIA. 179

more especially on Gecko, strengthens this view. In both the
hyoid arch is complete, but has absolutely nothing to do with
the malleus. But the proof that the hyomandibular is equivalent
to the epimandibular adds strength to Albrecht’s other hypo-
thesis, that the quadrate originally belonged to the palatine
arch and not to the mandibular arch.

I will now speak of the quadrate itself. That it cannot be
looked for in one of the auditory ossicles follows clearly from
the foregoing.

According to Tiedemann, Platner, Késtlin, Duvernoy, and
Albrecht, the quadrate of mammals is equivalent to the zygo-
matic process of the squamosal. I agree fully with this view.
To the examples adduced by Albrecht and Duvernoy of an
actual separation I can add a further one. In a stillborn tiger
I found in the right squamosal very much the same condition
of things as Albrecht has described in the skull of a newborn
child. The zygomatic process is separated by a “suture,”
which nearly passes through the whole scale.

In the upper part we have the true squamosal, in the lower
we see the quadrate. All these separations in the “ squamosal ”
must be considered as atavistic. That they are so without
doubt results from Cope’s researches on the Pelycosauria of the
Permian formation. Cope looks upon these reptiles as the
ancestors of the mammals. I have in another place (‘ Morphol.
Jahrbuch’) endeavoured to show that these are somewhat too
specialised to answer this hypothesis, but that they are very
closely allied to the ancestors of the Mammalia. I give Cope’s!
remarks on the quadrate of these interesting forms in his own
words: “ Although the malar bone is out of place in the speci-
men described, examination of the skull of Clepsydrops
natalis, where it is preserved in position, shows that this
horizontal ramus of the quadrate is nothing more than the
zygomatic process of the squamosal bone of the Mammalia
forming with the malar bone the zygomatic arch.”

1 K. D. Cope, “ The Relations between the Theromorphous Reptiles and
the Monotreme Mammalia,” ‘ Proc. Amer. Assoc. Adv. Sci. (1884, Philadel-
phia), vol. 33, p. 473.

180 DR. G. BAUR.

I myself think that there is no doubt that the quadrate of
the lower Vertebrates is contained in the zygomatic process of
the mammals.

According to Albrecht and Dollo, the quadrato-jugal is con-
tained in the malar (jugal). From what I find-in a very young
skull of Dasypus, the truth of this assertion appears to me
doubtful. In this skull I find, on both sides, a perpen-
dicular fissure which tends to separate the articular surface
of the process together with the jugal. I think that this
half-separated piece represents the quadrato-jugal of the
Sauropsida; to myself, therefore, it appears as probable that
this quadrato-jugal is included in the quadrate, an assumption
which is supported by the conditions found in Sphenodon.
Here, in older specimens, the quadrato-jugal is fused with the
quadrate, while it is free in the young form.

The results of these observations I will summarise as fol-
lows:

1. The assertion, put forward by Breschet and Peters, and
again by Dollo, that the cartilaginous distal portion of the
columella (stapes) of the Sauropsida is homologous with the
malleus of the Mammalia, is true.

2. The malleus in the Sauropsida and the Mammalia belongs
to the first and not to the second visceral arch, that is, to the
epimandibular portion of Meckel’s cartilage.

3. The so-called hyomandibular or ceratohyal of Sauropsida
is nothing else than the epimandibular portion of Meckel’s
cartilage (Peters, Albrecht, Baur).

4. The “ quadrate-cartilage” really belongs, not to the mandi-
bular, but to the palatine arch. (Albrecht asserted this.)

5. The homology, asserted by Tiedemann, Platner, Kostlin,
Duvernoy, Albrecht and Cope, between the quadrate of the
Sauropsida and the zygomatic process of the squamosal, is
true.

6. Probably the anterior end of this process represents the
quadrato-jugal.

HEMOGLOBIN ORYSTALS OF RODENTS’ BLOOD 181

On the Hemoglobin Crystals of Rodents’ Blood.
By

W. D. Halliburton, M.D., B.Sc.,
Assistant Professor of Physiology, University College, London.

(From the Physiological Laboratory, University College, London.)

Tue crystals of hemoglobin since their first discovery have
been described by various observers as occurring in no less
than five out of the six crystallographic systems. Subsequent
investigators have reduced this number to two, namely, the
rhombic system, in which the hemoglobin from the blood of
most animals crystallises; and the hexagonal system, in which
that from the blood of certain rodents is said to crystallise.

This research was undertaken at Professor Lankester’s sug-
gestion, in order, first, to ascertain whether these six-sided
crystals really belonged to the hexagonal system ; and, secondly,
to find, if possible, an explanation of the difference of crystalline
form that hemoglobin presents in different animals, while in
its other chief properties hemoglobin is universally the same.

It will be convenient to take the subject under the following
heads :

1. Historical.

2. Hexagonal blood-crystals.

3. Influence of the other constituents of the blood on the
crystalline form of hemoglobin crystals.

4. The crystalline forms of hemoglobin obtained by mixing
the blood from different animals.

5. Can squirrel’s hemoglobin be obtained in any form other
than hexagonal crystals ?

6. Conclusions and remarks.

182 W. D. HALLIBURTON.

1. Historical.

Oxyhzemoglobin crystals were first described by Reichert? as
occurring in the uterus of a pregnant guinea-pig; by Leydig?
as occurring in the alimentary canal of the leech; and by
Kolliker,’ obtained from the blood of the dog, python, and
other animals. Kdlliker considered the crystals to be com-
posed of a more or less modified hematin. Funke* was, how-
ever, the first to make complete observations upon them, and
to recognise their true nature. Kunde,’ working at the same
time, made extensive observations from a comparative point of
view, and was the discoverer of the exceptional form of the
crystals in the guinea-pig and squirrel. Since then many
investigators have worked at the subject, notably Lehmann,®
Rollett,’ von Lang,’ and Preyer,® in whose exhaustive treatise
a complete bibliography of the subject up to 187] is given.

Our present knowledge of the crystalline form that hemo-
globin assumes may now be summarised as follows:

a. In the great majority of animals’? in which hemoglobin

1 Reichert, ‘ Miiller’s Archiv,’ 1849, p. 197.

2 Leydig, ‘ Zeitsch. f. wiss. Zool.,’ Bd. i, 1849, p. 116.

3 Kolliker, ‘ Zeitsch. f. wiss. Zool.,’ Bd. i, 1849, p. 266.

4 Funcke, ‘ Zeitsch. f. nat. Med.,’ N. F., Bd. i, 1851, p. 184; Bd. ii,
1852, p. 204 and p. 288. ‘De sanguine vene lievates,’ ‘ Diss. Lipsie,”
1851.

5 Kunde, ‘ Zeitsch. f. nat. Med.,’ N. F., Bd. ii, 1852, p. 276.

6 Lehmann, ‘ Ber. d. k. Sach’s Ges. d. Wissen.,’ 1852, p. 22.

7 Rollett, ‘ Sitzungsber. d. Wien. Akad.,’ Bd. xlvi, 1862, p. 65.

8 Lang, ibid.

® Preyer, ‘ Die Blutkrystalle,’ Jena, 1871.

10 To the animals falling under this rule I can add several, the crystalline
form of the hemoglobin of which have not been hitherto recorded. Iam
much indebted for specimens of the blood of these animals to my friend
Mr. F. E. Beddard, of the Zoological Gardens.

Opossum (Didelphys cancrivora).—Very large and dark red crystals,
can be readily obtained. They belong to the rhombic system.

Kangaroo (Macropus giganteus).—Crystals are more soluble, and

so less readily obtained. They are rhombic prisms, slenderer than in the
opossum,

HEMOGLOBIN ORYSTALS OF RODENTS’ BLOOD. 183

occurs, vertebrate and invertebrate, crystals of it can be
obtained in the form of prisms and plates belonging to the
rhombic system.

6. The exceptions to this rule hitherto noted are the
following :

i, Guinea-pig. Hemoglobin crystals from the blood of this
animal are tetrahedra, once supposed to belong to the regular
system, but now shown by von Lang to be in reality rhombic.

ii. Lehmann mentions that similar tetrahedra may be ob-
tained from the blood of the mouse and rat. This has not
since been confirmed. v

iii. In several birds the crystals obtained are also tetrahedra.

iv. In three animals—the squirrel, the hamster, and the
mouse—six-sided plates have been described.

v. In one of these, the hamster, rhombohedra are described
as occurring also.

2. Hexagonal Blood-crystals.

We will take the three animals in which the hemoglobin is
said to crystallise in the hexagonal form one by one.

a. Squirrel.—The discovery of the fact that hemoglobin
crystals from this animal are six-sided plates was made by
Kunde (1852). Writing in the same year, Lehmann asserts
that though these crystals are six-sided they do not belong to
the hexagonal system. He gives, however, no reasons for this
assertion. Langand Preyer arrived at the opposite conclusion
i. e. that they do belong to the hexagonal system, from the study
of their optical properties.

Belideus breviceps (a marsupial).—Crystals similar to those of the
opossum.

Seal (Phoca vitulina).—Rhombic prisms, many of them very short
and simulating hexagons. Lasily obtained.

Bear (Ursus syriacus).—Bunches of rhombic needles, easily obtained.
They are slenderer than those obtained from dog’s blood as a rule, some being
almost silken in appearance.

Hydromys leucogaster (white-bellied beaver rat).— Rhombic prisms.

Sus leucomystax (white-whiskered swine).—RKhombic prisms.

Water-vole (Arvicola aquatica).—Crystals are obtained easily by
adding water to the blood. They are of the usual rhombic shape.

184. W. D. HALLIBURTON.

My own observations are as follows:—The crystals can be
obtained with the greatest ease by simply adding a drop of
water to a drop of defibrinated blood on a slide, and covering
it; in less than a minute crystals appear. I have also prepared
them by other methods ;! but in all cases the crystalline form
is the same. When first formed the crystals are six-sided
plates, many equilateral, but many not. After recrystallisation,
however, the crystals are then all but perfectly regular. The
quetions then arises, Do they belong to the hexagonal system or
not? To this question one of the three following answers must
be the correct one.

1. They do belong to the hexagonal system.

2. They do not belong to the hexagonal system, but are
rhombic crystals, having a so-called “hexagonal habit.” In
mineralogy instances are known of such occurrences. This is
the case with copper-glance, some of whose crystals so closely
resemble hexagonal ones that several mineralogists believed
that there were two kinds, one being hexagonal. Again, mica

is an instance of a monoclinic crystal with “ hexagonal
habit.”

py.
/
ES

Fie. 1.

Fig, 2. Fig. 3.
Cc

Suppose a B c D (fig. 1) to be the basal plane of a rhombic
plate, and the angle a B c to be approximately 120°, the lines

1 The method that I have found best for the preparation of blood-crystals
in most animals is to add to defibrinated blood a sixteenth of its volume of
ether, and then to shake for two or three minutes until the liquid becomes of
a clear lake colour; in the course of time, varying from five minutes to three
days, crystals form in abundance (‘ Gamgee’s Physiological Chemistry,’ p. 87)

HEMOGLOBIN ORYSTALS OF RODENTS’ BLOOD. 185

joining ac, BD being the axes. Then if the angles D aB,
p cB be replaced, as shown by the dotted lines, a hexagon
will be produced differing but little from a regular hexagon.

3. The third alternative is that they may belong to the
rhombic system by being twins, consisting of three parallelo-
grams or six triangles, as is shown in figs. 2 and 3. -Twins
are, however, rare in the rhombic system.

In order to settle this question it is necessary to examine
the optical properties of the crystals.

Crystals may be divided, according to their optical proper-
ties, into three classes :

1. Isotropic.—Those in which there is no distinction of
different directions as regards optical properties. This includes
crystals belonging to the regular system. They have but
one refractive index, i.e. refract light like amorphous bodies
do, singly.

2. Uniaxal.—Those in which the optical properties are
the same for all directions equally inclined to one particular
direction, called the optic axis, but vary according to this in-
clination. This class includes crystals belonging to the
dimetric system (crystals with three rectangular axes, two of
them being equal) and the hexagonal system. The optic
axis corresponds with the principal crystallographic axis.
In the direction of this axis a ray of light is refracted singly,
and in other directions doubly.

3. Biaxal.—This includes the remaining three systems
of crystals, the trimetric or rhombic (three rectangular axes
all unequal), the monoclinic, and the trichinic. In these
there are always two directions along which a ray is singly
refracted.

The best test, as to whether a substance is doubly refractive
or not, is this: If between crossed nicols, which consequently
appear dark, a substance be interposed that makes the dark-
ness give place to illumination, however feeble, that substance
is doubly refractive. This action is termed the depolarisation
of the ray.

The crystals of squirrel’s hemoglobin I submitted to this

186 W. D. HALLIBURTON.

test, with the result that no depolarisation of the light can be
detected, when they are examined with the apparent basal
plane perpendicular to the axis of the instrument and rotated ;
nor when a quartz plate is inserted do they produce any modi-
fication of the tint, as the stage is turned. The instrument
used was a Zeiss polarising microscope.

Hence the presumption is that they belong to the hexagonal
system, as rhombic crystals with hexagonal habit or rhombic
twins would produce some double refraction examined in this
way.

I submitted the question as to whether this was conclusive
to Professor Lewis, of Cambridge, and he kindly wrote to me
in answer as follows:

“The observation under the microscope between crossed
nicols, so far as it goes, is rather in favour of the crystals
being hexagonal, that is, presupposing that the field remains
dark when the crystal is rotated in the field of view. However,
this is not quite conclusive, and in such cases greater certainty
would be obtained if the crystals were placed under a Ber-
trand’s polarising microscope, to see the shape of the inter-
ference rings and cross.”

It should be here stated that uniaxal crystals in the direction
of their optic axis exhibit a symmetrical cross and circular
rings ; in biaxal crystals the rings are oval, or at any rate not
circular, and the cross is not symmetrical. ‘This is the case,
because the resistance to displacement in the three cardinal
directions called the axes of elasticity are all unequal in biaxal
crystals. This is true, not only for the crystalline substance
itself, but also for the luminiferous ether that pervades it.!

Acting on Professor Lewis’s advice, I submitted the crystals
to Professor Judd, who with Mr. Fletcher’s co-operation
examined them, and gave me the following report, for which I
am much indebted to him :—‘‘I have every reason to believe
the crystals belong to the hexagonal system from their form,
and their extinction between crossed nicols. I regret, however,

1 The cardinal directions are, however, believed not to be the same for the
ether as for the material of the crystal.

HEMOGLOBIN ORYSTALS OF RODENTS’ BLOOD. 187

to find that their minute size, and especially their extreme
tenuity, prevents our applying the crucial test of the inter-
ference figures seen in convergent polarised light.

‘‘Bertrand devised a form of microscope which enables these
interference figures to be studied in the minute crystals seen
in their rock sections, and von Lasaulx has improved this
apparatus. We have what I believe to be the best form of the
Bertrand-Lasaulx apparatus constructed by Nachet; but even
employing an immersion objective magnifying 650 diameters,
the crystals are still so small as to give neither rings, nor cross,
nor brushes.

“T greatly regret that we have not been able to apply this
test. I fear that no instrument exists which will accomplish
what you desire; and Mr. Fletcher, on theoretical grounds,
doubts whether it would be possible under any conditions to
apply the test to such minute crystals.”

The largest crystals of squirrel’s hemoglobin that I have
obtained were those formed by the addition of water to the
defibrinated blood; they varied in size from ‘001 to ‘005 m. in
breadth.

Since receiving Professor Judd’s report, I have tried to
obtain larger crystals by Gscheidlen’s! method. He seals
defibrinated blood in narrow glass tubes, which are then kept
at a temperature of 37° C. for several days. On opening these
tubes and emptying their contents into a watch glass, crystals
of great size are formed from dog’s blood after evaporation has
occurred.

With squirrels’ blood, however, I have not obtained larger
crystals by this method than by the first. ‘The reason for this
seems to be the extreme readiness with which squirrels’ hzmo-
globin crystallises. It is a well-known fact that bodies that
erystalise rapidly crystallise in small and numerous crystals.
If some method could be devised for retarding, but not pre-
venting, the crystallisation of squirrel’s hemoglobin, we might
then be able to obtain crystals of it large enough to which to
apply this crucial test.

1 * Physiologische Methodik,’ p, 361,

188 W. D. HALLIBURTON.

The matter must therefore be left incomplete up to this
point for the present. The probability, however, is greatly in
favour of the crystals being true hexagons.

We have seen that in order to have a rhombic plate with
hexagonal habit, it is necessary that one of its angles be
approximately 120°. I measured the angles in the rhombic
plates found in the rat, and found that they averaged 129°.

I shall also presently show that it is possible by the inter-
mixture of the blood of different animals to obtain crystals
closely resembling hexagons, but which are not so, as is shown
by their optical properties. .

6. Mouse.—Kunde was the first to describe the hemoglobin
crystals of this animal. He made eighteen observations, and
the crystals he found were fine needles and prisms.

Bojanowski! was the next to make observations on these
crystals. He describes and figures them as six-sided plates
resembling in form those from squirrel’s blood, of a flesh
colour, and very soluble in water. He prepared them by the
addition of a mixture of equal parts of alcohol and ether to
the blood. No description of their optical properties is given.
He remarks, “I have not been able to observe the fine needles
described by Kunde.”

Preyer repeated these experiments, and confirmed the obser-
vations of Kunde, not those of Bojanowski. He obtained
small prismatic crystals.

I have myself experimented with the blood of eighteen
mice, and the result has been again to confirm Kunde’s obser-
vations. The crystals are exceedingly difficult to obtain, and
in some cases I have had to repeat the process of freezing and
thawing many times after the addition of alcohol, before suc-
ceeding in obtaining them. They are very soluble in water.
The crystals are exceedingly small rhombic prisms. They are
nearly colourless, and it is only when they are heaped together
that any red tinge at all can be perceived in them. In one
case in which by the addition of ether to the blood I obtained
crystals of fair size after allowing the mixture to stand for five

1 Bojanowski, ‘ Zeitsch. f. wiss. Zool.,’ Bd. xii, 1863, p. 333.

HAHMOGLOBIN ORYSTALS OF RODENTS’ BLOOD. 189

days, the crystals still showed this same peculiarity, namely, in
being nearly colourless. I have successfully employed Boja-
nowski’s method for the preparation of the crystals, namely,
the addition of a mixture of alcohol and ether tothe blood; but
in no case did hexagonal crystals form. Mouse’s hemoglobin
also differs from squirrel’s in being very soluble in water ; this
is admitted by Bojanowski; one would therefore expect a
priori that its crystalline form would be different.

c. Hamster (Cricetus vulgaris).—My remarks under
this heading will be only historical. I have not myself been
successful in obtaining one of these animals. The crystalline
form of the hemoglobin was first described by Lehmann, who
found rhombohedra and six-sided plates. His experiments
were repeated by Preyer,! whose observations on the subject are
very complete. He fornd both crystalline forms, viz. six-
sided plates, and rhombohedra. ‘This is interesting since the
rhombohedron belongs to the hexagonal system. By examina-
_tion between crossed nicols he found that the six-sided plates
had no action in “ depolarising ” the ray, and he therefore con-
cludes that they, like squirrel’s hemoglobin crystals, are true
hexagons.

d. Conclusions.—The presumption in favour of the
hemoglobin crystals of the squirrel and hamster being true
hexagons is exceedingly great. In the case of the mouse, it
seems to be almost equally certain that the crystals are not as
arule hexagonal. I should not like, however, to deny that
hemoglobin may sometimes in the case of the mouse crystallise
in this way, because of some observations I have made on the
hemoglobin crystals of the rat.

Crystals are obtained from the blood of this animal with
great ease; mere addition of water to the blood causes almost
immediately an abundant crop of crystals. On this account
the blood of this animal is used by the students in the practical
classes at University College for the preparation of hemoglobin
crystals. Professor Schafer told me that on looking over the
students’ preparations he had occasionally seen hexagons to-

1 * Die Blutkrystalle,’ p. 262.

iige

190 W. D. HALLIBURTON.

gether with the ordinary rhombic prisms and plates. In
order to verify this, I have niade numerous specimens of the

crystals from the blood of about fifteen rats. As a rule, no

hexagons were present; but on three occasions I have detected
hexagonal plates—very few in number, perhaps not more than

one or two on the slide—among the rhombic crystals. There

appeared to be nothing special either about the animal used
or the method employed in these cases. The diameter of
these crystals averaged about the same as in squirrel’s blood
(002—:003 m.). Between crossed nicols they also behaved
the same as squirrels’ hemoglobin crystals, viz. remained dark.
in all positions.

In addition to this, if crystallisation be watched under the
microscope, a single corpuscle wil] often be observed to set
into a minute hexagon. This is what Preyer calls intraglobular
crystallisation. He describes it as occurring in the blood of
the hamster. It can also be observed in the blood of the
rat. The crystals apparently so formed last but a few seconds,
the corpuscles then becoming shrunken, or irregular, and very
often under the subsequent action of water, globular. It is
therefore possibly a stage in the crenation of the corpuscle.
But, apart from this, it is undoubtedly the fact that hexagonal

_ crystals are occasionally found in the blood of the rat. It

1 Since writing the above, I have received the following ina letter from Mr,
Sheridan Lea, of Cambridge. He says :—‘ When I was showing a class how
to put up permanent specimens of hemoglobin crystals from rat’s blood, we
obtained uniformly hexagons, instead of prisms. This I have neither ever -
noticed or heard of before, and I thought it might be of interest to you.
The method employed was that of Stein (‘ Centralb. f. d. med. Wiss.,’ 1884,
No. 23, and ‘ Virchow’s Archiv,’ 97, 483).” I had myself occasionally used
Stein’s method of preparing crystals from rat’s blood, but had always obtained
the usual rhombic prisms. On receiving Mr. Lea’s letter I made a large
number of preparations of hemoglobin crystals by this method. The method
consists in simply mounting a drop of defibrinated blood in a drop of Canada
balsam. In the case of some animals, among which were man and the mouse,
I was not able to get any crystals at all. In the commoner mammals, dog
and cat, the crystals obtained were very fine specimens of rhombic prisms.
In the guinea-pig and squirrel they presented the usual tetrahedral and
hexagonal shapes respectively. With rat’s blood, however, the results were

HMMOGLOBIN CRYSTALS OF RODENTS’ BLOOD. 191

would therefore be possible that such crystals occasionally may
occur in the blood of other animals, such as the mouse, the
usual form of whose blood-crystals is, however, rhombic.

The rats employed in the above experiments were the

common house rat, and also tame rats.

3. Influence of the other Constituents of the Blood on the
Crystalline form of Hemoglobin Crystals.

These experiments, as well as those in the next section of
this paper, were undertaken at the suggestion of Professor
Schafer. ©

The blood-crystals of an animal have the same form whether
they be obtained from the fresh blood, or from the blood from
which the fibrin has been removed. Fibrin, or its precursor
in the blood-plasma fibrinogen, has then no influence on the
form of the blood-crystals.

The following experiments were undertaken to ascertain
whether the other constituents of the blood-plasma, which are
all contained in the serum, have any effect in influencing the
form of the crystals.

The method of experimentation was as follows:— Defibrinated
blood is taken in a tube and centrifugalised for about half an

very strange. In the majority of cases the usual rhombic needles were formed ;
but_in a few cases I confirmed Mr. Lea’s observations, and obtained perfectly
‘regular hexagons; in some. cases the hexagons would occupy one part of the
slide only, while the remainder was filled with the ordinary prisms. Hexagons
seemed to form where the proportion of blood to balsam was small, and they
were formed especially at the edges of a preparation where the drop of blood
had probably had time to dry somewhat before being covered with Canada
balsam. These hexagons remained dark in the dark field of the polarising
microscope. After a day or two they cracked in a peculiar way, and seemed
then to be made up of minute needles radiating from a centre. This may or
may not indicate the way in which they are formed. The fact that they
occurred most in parts of the field where there was least water seems, how-
ever, to confirm the theory advanced later in the paper, viz. that the difference
of crystalline forms in hemoglobin is due to different amounts of water of
crystallisation.

wae

192 W. D. HALLIBURTON.

hour; the corpuscles settle at the bottom of the tube, and
the supernatant serum is pipetted off. To the corpuscles the
blood-serum of some other animal is added, the mixture shaken,
and the mixture again centrifugalised; the serum is again
pipetted off, and more added. After repeating this process
several times, the corpuscles of one animal are obtained in the
serum of another animal without any of the serum of the first
animal being in the mixture. Hemoglobin crystals are then
prepared from this mixture. In some cases the foreign serum
dissolves the hemoglobin and disintegrates the corpuscles.
This was first pointed out by Landois.?

Mere addition of the blood-serum of one animal does not
as a rule cause the formation of blood-crystals. It does so,
however, sometimes.? This is explicable on the assumption
that the blood-serum used is very watery, and the hemoglobin
of the other animal crystallises very readily. I have myself
come across no case in which it occurred.

My results may be best given in the form of the following
table. I have given not onlythe effect of the foreign serum on the
crystalline form of hemoglobin, but also the effect on the cor-
puscles themselves, as to whether they are disintegrated or not.

Mouse Cat. Little dissolved

Corpuscles of | In Serum of Effect on the Corpuscles. | otis Heeeeee Form
Rat Squirrel § Much dissolved
Squirrel | Rat Very little dissolved
~ | Squirrel | Dog Very little dissolved
| Rat | Guinea- -pig Little ifany dissolved |.
| Guinea-pig | Cat Nearly entirely dissolved
Guinea-pig | Dog Much dissolved

The result of these experiments is to show that the serum of
one animal has no influence in causing a change of the hemo-
globin crystals of another animal.

I next examined in a qualitative manner the serum of certain

1 ¢ Die Transfusion des Blutes,’ Leipzig, 1874.

2 An instance of such action is recorded by Professor Schafer (‘ Blood
Transfusion,” ‘ Trans. Obst. Soc. London,’ 1879, p. 317).

HEMOGLOBIN CRYSTALS OF RODENTS’ BLOOD. 193

rodents with regard to the proteids or albuminous. substances
contained in it. I obtained similar results in all animals,
results which show, too, that the serum proteids of rodents
agree with those in other mammalian animals which I had
previously investigated.1_ The proteids, the most important
bodies in the blood-plasma, being similar, the serum would
not on a priori grounds be suspected of influencing the
crystalline form of hemoglobin. The results I have obtained
with regard to the heat-coagulation temperatures of these
bodies is shown in the following table.

Temperatures of Coagulation of the Proteids in the
Blood of certain Rodents.

| a | =
Name of Proteid. ) cael == | Mouse. | Guinea-pig. Squirrel. 2
fea] Wy losis | |
s | = 4
|
Globulins— ICCTA Gri cee Oia el 8 Meera C. C.|
| Fibrinogen . 56°} 56°) 56° (56° 56° 5(°
Serum globulins .| 78° 75° 75° = |75° 75° 75>
| Albumins— | |
Wie ce, (| 78° [70% FOMIe79° 720-3° 73°|
| eae mittee eben tae Poe ArT 77° (small inamount) 77°
Y. + + « « «| 84°) 84°| 84° 87° (trace)|84° (very abundant) |84°

The stromata of the red blood-corpuscles might, however,
possibly be supposed to have some influence on the crystalline
form of the hemoglobin. We have seen that crystallising the
hzmoglobin of one animal from the serum of another yielded
negative results; squirrel’s hemoglobin remained hexagonal,
rat’s and guinea-pig’s rhombic prisms and tetrahedra respec-
tively, whatever the serum in which they had been dissolved.
A similar result followed crystallisation from a fluid consisting
of serum plus the dissolved stromata of the corpuscles of some
other animal. This was obtained by adding to the blood one
sixteenth of its volume of ether, and letting it stand; the
crystals of hemoglobin which formed were filtered off, and the
ether evaporated from the filtrate which consisted of the serum
with the stromata of the corpuscles dissolved in it.

1 Halliburton, “ Periods of Serum,” ‘ Journal of Physiology,’ vol. v, p. 152.

VOL, XXVIII, PART 1.—NEW SER, N

194 W. D. HALLIBURTON.

So far then these experiments seem to show that the differ-
ence of crystalline form is due to some inherent quality of the
hemoglobin itself, and not due to any agency in the blood
external to the hemoglobin.

4, The Crystalline forms of Hemoglobin obtained by
mixing the Blood from different Animals.

By mixing the defibrinated blood from two animals, whose
hemoglobin crystallises differently, and then preparing crystals,
I thought I might obtain some new forms resulting from the
mixture. Here my experiments have yielded mostly negative
results, but the one positive result 1 have obtained from such
experiments warrants me in recording the whole. The blood
of two animals were mixed in about equal proportions, shaken
thoroughly, and then hemoglobin crystals prepared by the
ether method.

It will be convenient here again to give my results a tabular
arrangement.

Blood of | Mixed with

that of Form of Hemoglobin Crystals prepared from the Mixture.
Rat | Squirrel Both rhombic prisms and hexagons present.
Rat _Guinea-pig No rhombic prisms of the shape usually seen in rats’

blood present. No tetrahedra. Crystals are all
rhombic prisms with hexagonal habit.

Squirrel | Guinea-pig | Hexagonal plates and tetrahedra both present. Many
_ tetrahedra imperfect. The tetrahedra were all re-
| duced to about half the size of those prepared from
|

|

the unmixed blood of the same guinea-pigs.

| Dog Squirrel — Fine rhombic needles and hexagonal plates both pre-
seut in abundance.

Dog _Guinea-pig The greater number of the crystals formed are very
|

| prepared from the blood of the same guinea-pigs.
| The optical properties are, however, the same.

Rhombic prisms very slender, like those of dog’s

|
|
| small tetrahedra, about a quarter the size of those
|
|
blood, also seen.

The second case, that of mixing blood from the rat and
guinea-pig, is interesting, and demands further description.
It shows that it is possible to obtain a new form of hemoglobin

HEMOGLOBIN CRYSTALS OF RODENTS’ BLOOD. 195

by mixing that from two animals in which the crystalline form
is different. It also shows that rhombic hemoglobin crystals
may assume a hexagonal type (fig. 4). These crystals are not,
however, perfect or equilateral hexagons, two of the sides
being longer than the other four.

Fre. 4.

The side a B = E D = ‘0019 m. (average).
The sides BC = CD = EF = FA = ‘00125 m. (average).

This irregularity is possibly to be accounted for by the fact
that, in rats’ hemoglobin crystals, the angles corresponding to
BCD, AFE, are 51°. In order to obtain perfect hexagons
of a rhombic type it is necessary, as before stated, that this
angle be 60°.

Under crossed nicols these crystals appear perfectly bright,
so contrasting with the true hexagons obtained from the blood
of the squirrel and hamster.

This result was not, however, always obtained ; in one or two
cases I obtained as a result of mixing the blood of these two
animals a mixture of crystals; that is prisms and tetrahedra.

5. Can Squirrel’s Hemoglobin be obtained in any form other
than Hexagonal Crystals ?

Another set of experiments was performed with the object
of breaking down the hexagonal constitution of the hemo-
globin of squirrels’ blood. The first method tried was that of
driving off the water of crystallisation, and of then adding
water to the dehydrated hemaglobin.

The hemaglobin was obtained in a state of purity and dried
over sulphuric acid until it lost no more weight. Then it was
examined, and found to have its normal spectroscopic proper-

196 W. D. HALLIBURTON.

ties. It was heated to 100° C. in a water oven, and again
examined. It had lost but a slight amount of weight. It was
rather more insoluble in warm water than previously, but the
spectroscopic properties, and the form of the crystals obtained
from the solution, remained as before. This confirms the obser-
vation previously made by Hoppe-Seyler that dry hemoglobin
is not decomposed by a temperature of 100° C. It was again
heated in the water oven at 100° C. until there was no further
loss of weight. It was then heated to 120° C. in an air-bath,
and again examined. It was found to have lost considerably
in weight, to have lost its crystalline lustre, to be brown in
colour (hematin) and to be insoluble in water. That is, it
parts with its water of crystallisation at a temperature which
decomposes it, with the formation of hzmatin, the proteid
matter becoming at the same time coagulated and insoluble.

Experiments were then tried with the object of ascertaining
whether a lower temperature will remove the water of crystal-
lisation ina Torricellian vacuum. This I did by means of a
Pfliiger’s mercurial air-pump. The action of the vacuum alone
converted the dried hemoglobin, at any rate partially, into the
conditions of methemoglobin. The water of crystallisation
seemed to be completely lost at a temperature of 50°—60° C.,
as subsequent heating to 120° C. produced no further loss of
weight. But this temperature was also sufficiently high to
decompose the hemoglobin in such a way as to render it
insoluble, or almost so, in water, and therefore no crystals could
be subsequently obtained from it.

The next method adopted was to convert the hemoglobin by
various reagents into methemoglobin ; then by reducing ageuts
to form once more hemoglobin, and then obtain crystals of
this. But the reducing agents used were found to hinder the
formation of crystals.

The third and simplest method was to repeatedly recrystallise
the hemoglobin, when it was found after three or four re-
crystallisations that no six-sided crystals were obtained, but a
mixture of rhombic needles and tetrahedra, and in some cases
the latter were absent. This is interesting in connection with

HEMOGLOBIN CRYSTALS OF RODENTS’ BLOOD. 197

the reverse experiment already related, in which crystals simu-
lating hexagons were obtained by mixing together the blood of
the rat and guinea-pig, and in which the same result was
obtained from a mixture of the solutions of the pure hemo-
globin of the same animals.

6. Conclusions and Remarks.

What the difference between the various forms of hzmo-
globin may be, it cannot be a very deep or essential one. The
difference in crystalline form is associated with a difference of
solubility in water and other reagents; but the spectroscopic
characters, the decomposition products, the compounds it forms,
of which hzmin is a readily obtained example, are universally
the.same. Not only so, but Hoppe-Seyler has shown! that in
various animals dried hemoglobin has the same or nearly the
same elementary composition.

Have we then to deal with a case of polymorphism? The
terms dimorphism and polymorphism cannot be applied to any
substance which crystallises in two or more forms, unless the
composition of that substance be exactly the same in all cases.
Instances of dimorphism in the mineral world are carbon and
sulphur among the elements, and sal ammoniac, potassium
iodide, cuprous oxide, &c., among compounds. ‘The conditions
on which dimorphism depend are two: first, temperature,
secondly, the solvent from which the substance crystallises.
If, as in the case of many mineral salts, the compounds are
united with different proportions of water of crystallization, we
have to deal with different hydrates, and the case is not one of
true dimorphism ; an instance of this is sulphate of soda.

The case seems to me to narrow itself down to this in the
case of hemoglobin; either we have here a case of poly-
morphism, or the crystalline forms are due to the combina-
tion with varying proportions of water of crystallisation. In
the absence of a rational formula for hemoglobin it would
be unsafe to affirm the former of these two alternatives. More-
over, the conditions that are known to produce dimorphism in

1 « Physiologische Chemie,’ p. 377.

198 W. D. HALLIBURTON.

minerals, namely, differences of temperature and of solvent, have
in the case of hemoglobin no influence.

If we then fall back on the latter alternative, the question
which arises is whether there are any facts to support it. The
explanation that the varying form of oxyhemoglobin is due to
varying quantities of water of crystallisation may be other-
wise expressed by saying that we have to deal with different
hydrates of oxyhemoglobin. This would account for the
varying solubilities of these substances in water and other
reagents, and at the same time is not such an essential differ-
ence as to prevent the chief properties of hemoglobin from
being universally the same.

Turning to Hoppe-Seyler’s researches on this subject of water
of crystallisation, it is seen that its amount varies consider-
ably. The following is his table:!

Percentage of Water of Crystallisation.

Dog’s hemoglobin . - : : ; 3 to 4
Guinea-pig’s ,, : ; ’ ; : 7
Squirrel’s s : : : : ; 9-4
Goose’s 3s . ; : : ; 9-4

In an earlier paper,” the same author gives rather different
percentages, viz. for guinea-pig’s hemoglobin 6, for goose’s
hemoglobin 7, and for squirrel’s hemoglobin 9. Dr. Christian
Bohr® has more recently made observations on the water of
crystallisation of dog’s hemoglobin, and as the result of
thirteen experiments he finds that its amount varies from 6°3
to 1:2 per cent. It is thus seen that great variations occur in
the numbers obtained by these experiments. The reason for
this variation seems to me to be the great difficulty of obtain-
ing hemoglobin in a pure state, and also possibly because the
method adopted, which is the same as that carried out in
similar investigations on inorganic salts, is not applicable to
such a complex and much less stable organic compound as

1 *Physiologische Chemie,’ p. 377.

2 «Med. Chem. Untersuchungen,’ Heft iii, 1868, p. 370.

3 *«Experimentale Untersuchungen iber die Sauerstoffaufnahme des Blut-
farbstoffes,’ Kopenhagen (Olsen and Co.), 1885.

=i
i

HEMOGLOBIN CRYSTALS OF RODENTS’ BLOOD. 199

hemoglobin; in other words, the temperature necessary to
drive off the water of crystallisation is also sufficient to cause
certain decomposition changes in the pigment.

My experiments have shown that squirrel’s hemoglobin will
under certain circumstances crystallise in forms other than the
usual hexagonal form. A crucial experiment in order to see
whether this is due to union with different amounts of water of
crystallisation would have been first to ascertain the amount
of this water in the hexagonal crystals, and then in the
rhombic crystals obtained by recrystallisation. I have per-
formed three such experiments, but the results obtained are so
conflicting, and exhibit variations as great as in Bohr’s experi-
ments, that it is impossible to draw any conclusions from
them, except the negative one that we cannot by our present
methods of research make any definite statement with regard
to the water of crystallisation of hemoglobin.

Even if it be found ultimately that the difference in crystal-
line form is dependent on varying amounts of water of crystal-
lisation, the difficulty is only explained up to a certain point.
What is left unexplained is the nature of the agency that
causes the hemoglobin of some animals to unite with a certain
amount of water of crystallisation, and that of other animals
with a different amount. That some such substance or agency
does exist would seem to be the inevitable result of the recrys-
tallisation experiments which have been reiated.

METHOD OF OBTAINING METHHMOGLOBIN ORYSTALS. 201

An Easy Method of obtaining Methemoglobin
Crystals for Microscopic Examination.

By

Ww. D. Halliburton, M.D... B.S8e.,
Assistant Professor of Physiology, University College, London.

(From the Physiological Laboratory, University College, London.)

METH#MOGLOBIN is a derivative of the red colouring matter
of the blood, concerning which a number of theories have been
held. According to Sorby,! it is more highly oxygenated than
oxyhemoglobin; that is, it is a per-oxyhemoglobin. Hoppe-
Seyler,? on the other hand, regards it asa sub-oxyhzemoglobin,
as it can be obtained under conditions which remove at least
part of the oxygen of oxyhemoglobin. According to both these
views, however, the oxygen is regarded as being more firmly
combined with the hemoglobin than in the case of oxy-
hemoglobin.

More recently, however, Hiifner and Kiilz? have advanced
a third theory concerning the constitution of methzemo-
globin, and that is that it contains the same amount of oxygen
as oxyhemoglobin, only in a closer state of combination.
They are able to make this assertion from actual analyses;
and these analyses were possible, inasmuch as they succeeded
in obtaining methemoglobin in a crystalline form. The
method of obtaining these crystals is as follows :}—Three or
four cubic centimetres of a concentrated solution of ferri-

1 © Quart. Journ. Micr. Sci.,’ 1870, p. 400.

? «Zeit. Physiol. Chemie,’ vol. ii, p. 150.

3 * Zeit. Physiol. Chemie,’ vol. vii.

* G. Hiifner, “ Ueber Krystallinisches Methamoglobin vom Hunde,” ‘ Zeit.
Physiol. Chem.,’ Bd. viii, p. 366,

VOL. XXVIII, PART 1.—NEwW SER. 18)

202 W. D. HALLIBURTON,

cyanide of potassium are added to a litre of concentrated solu-
tion of hemoglobin. A quarter of a litre of alcohol is added,
and the mixture frozen. After one or two days’ exposure to
this low temperature abundant crystals of a brown colour,
which give the absorption spectrum of methemoglobin, are
deposited. They were obtained in this way from the hemo-
globin of the dog, pig, and horse, and their form is the same
as that of the oxyhemoglobin crystals of the same animals,
i.e. rhombic prisms. Dr. Gamgee! had prepared these crystals
from dog’s blood many years previously, but their true nature
was not at that time recognised. His method was much the
same as Hiifner’s, the chief difference being that the nitrite
of potassium or amyl was employed instead of ferricyanide of
potassium.

Jiaiderholm? has also obtained these crystals from dog’s
blood by the ferricyanide method, and confirms Hifner’s
statement that they are rhombic prisms. He also figures
some crystals of methemoglobin obtained by Profossor Ham-
marsten from the horse by the same method, which were regular
six-sided plates, and showed no double refraction if lying flat ;
they therefore presumably belonged to the hexagonal system,
and were more insoluble in water than the crystals of dogs’
methemoglobin. I can find no previous reference to the
methemoglobin crystals of rodent animals.

Hiifner’s ferricyanide method is applicable when one wishes
to obtain large quantities of the crystals for analysis. I now
wish to describe a much simpler method of obtaining these
crystals for purposes of microscopic observation. I have tried
this method with the blood of the ox, dog, cat, rabbit, rat,
guinea-pig, and squirrel, but only successfully in the three last-
named animals. In other words, methemoglobin crystals are
obtained with ease from the same animals as yield oxyhemo-
globin crystals with readiness.

The method consists in taking a few cubic centimetres of

1 A. Gamgee, “The Action of Nitrites on Blood,” ‘ Philos. Trans.,’ 1868,
p. 589, et seq.

> «Zeitschrift fir Biologie,’ Bd. xx, p. 419,

METHOD OF OBTAINING METHEMOGLOBIN CRYSTALS. 203

the defibrinated blood of the animal, adding an equal number
of drops of nitrite of amyl in a test-tube, and shaking the
mixture vigorously for a minute or two. The colour changes
to the dark chocolate tint of methemoglobin, and spectro-
scopic observation shows the typical absorption bands of that
compound. A drop of this liquid is then placed! on a slide
and covered; in a few minutes crystals form, which observa-
tion with the spectroscope shows to be composed of methemo-
globin. The edges of the cover-glass may then be sealed, and
the crystals kept unchanged for several months.

The crystals obtained from guinea-pig’s blood by this
process are tetrahedra, which differ only in colour and spec-
troscopic appearances from those of oxyhzmoglobin from the
same animal.

The crystals obtained from squirrel’s blood are perfectly
regular hexagonal plates, which remain dark between crossed
nicols.

The crystals obtained from rat’s blood are also perfectly
regular hexagonal plates, which remain dark between crossed
nicols, and which consequently are precisely similar to those of
squirrel’s methemoglobin. This remarkable fact helps to show
that the difference between the oxyhemoglobin of these two
animals cannot be a very deep or essential one.

In the case of rat’s metheemoglobin there were, in addition
to the hexagons, a few other plates of various shapes scattered

6 4 4.
5 3 ;
5 Ts
1
1
B Cc D

A. Regularhexagon. 3. Equilateral triangle. c. Intermediate stage between
Aand B. D. Parallelogram.

1 This must be done immediately after the formation of the chocolate-
coloured liquid; as in about a quarter of an hour the whole liquid sets into
a gelatinous mass of the same colour, from which no crystals are obtainable.

204 W. D. HALLIBURTON.

' in different parts of the field. These are depicted in the

accompanying cuts.

Mr. Fletcher kindly examined these for me, and expressed
it as his opinion that the triangular and rhombic forms were

merely variations of the hexagons. In the case of B faces 1,

8, and 5, and in the case of p faces 3 and 6 have virtually

disappeared.

On the Development of Peripatus Nove-
Zealandiz.

By

Lilian Sheldon,
Bathurst Student, Newnham College, Cambridge.

With Plates, XII, XIII, XIV, XV and XVI.

Tue account to be given in this paper is unfortunately by
no means a complete one, owing to the difficulties attendant on
working at the subject in this country. The great drawback
is the difficulty of obtaining the creatures, and as, so far as I
know, there is no way of keeping them alive in England, it is
necessary to kill them and remove the embryos as soon as they
arrive, so that one is by no means certain of obtaining the stages
which are required, when one does have the good fortune to
obtain a supply of the creatures.

All the embryos which I have worked upon were given to
me by Mr. Sedgwick, and most of them were taken out of the
uterus and preserved by him before he handed the material
over tome. So far we are not able to state at what times of
the year the different events in the development take place,
but it is possible nevertheless that there may be a definite
sequence, as we have at present only received the material in
December, July, and April, and so have not many data to go
upon.

The material which arrived in July contained seven females ;
three of these were without embryos in the oviduct or uterus,
and the other four contained embryos varying in age from that

VOL. XXVIII, PART 2,—NEW SER. P

205 LILIAN SHELDON.

figured in figs. 25 and 26 to those which were just ready to be
hatched.

The second supply arrived in November, but most of the
creatures were not opened till December. One female, which
was opened on November 27th, contained several fully deve-
loped embryos, while in one opened on November 30th the
uterus was empty. Seven females which were opened in the
middle of December contained only unsegmented and seg-
menting ova in their uteri.

In the last supply, which arrived last April, there were nine
females, which were opened on the 18th day of the month.
Of these five had no embryos in the uteri, one had several old
embryos, one segmenting ovum, and two embryos, one of
which is shown in sections in fig. 15; another contained
several old embryos, one segmenting ovum (which is shown
in section in fig. 11), and two of the stage represented in fig.
17; another contained the embryos, sections of which are
shown in figs. 13, 18, 19, and 20, and also one unsegmented
and several segmenting.

The New Zealand species, like all the others which are so
far known, is viviparous, the embryos undergoing the whole
course of their development in the uterus of the mother.

The ripe ovum is very large as compared with those of
Peripatus capensis and P. Edwardsii, measuring about
1:5 mm. initslong axis. This large size is due to the enormous
amount of food-yolk with which the egg is charged.

The egg is enclosed in a thick tough shell, which in the
fresh state adheres closely to it; after treatment with certain
reagents it becomes somewhat distended, and can be pricked
or removed. This is especially the case with eggs which are
preserved in hot corrosive sublimate, which causes the shell to
swell up, to become less tough and to lie at a greater distance
from the egg, so that it can be quite easily removed without
damaging the surface of the egg. This is not so easily accom-
plished in cases where the corrosive sublimate was not heated,
and the surface of several of my eggs was more or less injured
in the process of removing the shell. It is necessary to prick

DEVELOPMENT OF PERIPATUS NOV#-ZEALANDIA. 207

the egg-shell before placing the eggs in spirit, as otherwise it
collapses and crushes the embryo. Before cutting sections of
the eggs I almost always removed the shell, since it was too
hard to cut well, and was also apt to prevent the paraffin from
penetrating the ovum.

Within the shell the ovum is enclosed in a vitelline mem-
brane, which adheres closely to it, and is thin and membranous.
The two are easily distinguishable in eggs stained with picro-
carmine, as the shell stains yellow and the vitelline membrane
red.

The only accounts which have been hitherto published of
the development of P. nove-zealandiz are by Hutton! and
Kennel,” both of which are very brief. These observers state
that the segmentation is holoblastic, which is not the case, it
being rather on the centrolecithal than on any other type of
segmentation.

Mertuops.

The most satisfactory preparations obtained were from eggs
which were preserved in hot or cold corrosive sublimate and
glacial acetic acid mixed in the proportion of two toone. Other
ova were preserved in Kleinenberg’s picric acid, but these were
not satisfactory. The eggs were all stained in picrocarmine,
and afterwards passed through the various strengths of alcohol
in which a small amount of picric acid was dissolved; by this
method the yolk is stained yellow, the protoplasm light, and
the nuclei deep red, so that they are easily distinguishable
from one another. I am indebted to Mr. Harmer for the
knowledge of this method.

The embryos were all removed from the uterus in the living
state, and were preserved at once.

I do not purpose in this paper to enter into the subject of
the ovarian ovum and the changes undergone by it in its

1 Hutton, Capt. F. W., “On Peripatus nove-zealandiz,” ‘The
Annals and Magazine of Natural History,’ 4th Series, Nov., 1876.

2 Kennel, Dr. T., ‘‘ Entwicklungsgeschiclte von P. Edwardsii, Blanch,
und P. torquatus, n, sp.,” ‘Semper’s Arbeiten,’ Band vii, 1885,

208 LILIAN SHELDON.

passage into the ovum with the segmentation nucleus, as I
hope to be able later to make some further investigations on
that subject. It will be enough to state here that in several
cases there were ova in the uterus which possessed no nucleus
whatever ; there was a small amount of protoplasm present as a
very loose and not always easily-recognisable reticulum lying
among the yolk-spheres. This protoplasmic reticulum was
sometimes scattered throughout the egg, but was more often
only present at the periphery; while in some cases it was
aggregated at one point only. In one ovum there was a very
large compact mass of protoplasm at one point near the
periphery, but no trace of a nucleus could be discerned.

SEGMENTATION.

Immediately before the segmentation begins the ovum con-
sists of a great mass of yolk-spheres, and contains a single
nucleus. The position of the nucleus in the ovum varies
somewhat in different cases; in fig. 1 it is seen to be situated
at some distance from the periphery of the ovum; in this case
it is round in form, and contains a deeply staining wall and
also a single mass of chromatin. The protoplasm in which it
is embedded is compact and dense, and contains at its periphery
several chromatin particles. In the ovum from which fig. 2 is
taken the nucleus and its surrounding protoplasm had a some-
what different position and form. ‘The nucleus is situated
near the periphery of the ovum, being separated from the
vitelline membrane by only a thin layer of yolk. The nucleus
has a peculiar lobed form, and consists of three masses of
deeply staining material, between which is a portion of nuclear
substance which stains less deeply. It is surrounded by a very
small amount of protoplasm, which forms a loose reticulum,
the strands of which pass in and are lost among the yolk-
spheres.

The next stage is that in which two nuclei are present in
the egg; two sections from such an egg are figured in figs.
3aand36. One nucleus is situated at the periphery of the
ovum and the other somewhat deeper, but both lie in the centre

DEVELOPMENT OF PERIPATUS NOVHE-ZEALANDIAZ. 209

of one of the long surfaces on the same side of the ovum, and
each is surrounded by a protoplasmic area. The peripherally-
situated nucleus had a peculiar lobed form, while the other
seemed to be in the act of dividing, the chromatic particles
representing the spindle-fibres cut through transversely. In
another ovum, in which two nuclei were present, both were
situated quite near the periphery; but in this case the sections
were too thick for the structure of the nuclei to be made out.
In the next stage, in which three nuclei are present, other
changes have also taken place in the ovum. These concern
the segmentation of the yolk. At the pole where the nuclei
are situated the yolk is broken up into segments, which vary
considerably in shape and size. The yolk-spheres at this pole
are smaller than those over the rest of the egg. The yolk
segmentation does not bear any definite relation to the proto-
plasmic and nuclear segmentation, but takes place quite inde-
pendently of it. A nucleus is present in each of three of the
yolk-segments, in others there is a considerable protoplasmic
reticulum, but no nucleus, while in others again there is no
trace of either protoplasm or nucleus. The nuclei are all
situated near together. Fig. 4a is a section through the whole
egg. The section passes through one nucleus which is round
in form and contains a chromatin network, and is surrounded
by an area of protoplasm. It is situated in a yolk-segment,
the spheres composing which are very small. The section
passes through several other yolk-segments, four of which
contain no recognisable protoplasm or nucleus; another, which
is not completely segmented off from the mass of the yolk,
contains a small compact mass of protoplasm. The greater
part of the yolk is unsegmented, and is composed of very large
yolk-spheres. In future that part of the surface of the ovum at
which the nuclei are situated will be spoken of as the proto-
plasmic area. Figs. 46 and 4c are taken from other sections
through the same egg, and represent sections of the proto-
plasmic area seen under higher power. In fig. 46 there is a
considerable amount of protoplasm in two of the yolk-segments
besides that in which the nucleus is present. The nucleus is

210 LILIAN SHELDON.

embedded in a large mass of protoplasm, and is very much
lobed in form. In fig. 4 ¢, which passes through the third
nucleus, two of the yolk-segments possess a small amount of
protoplasm, and there is a very large amount in the segment
which contains the nucleus. The nucleus itself is very peculiar
in shape, and is made up of a large number of lobes.

At the next stage the yolk is segmented throughout the
whole egg, but the nuclei are still confined to the protoplasmic
area. In the fresh state the yolk-segments are very clearly
seen over the whole surface, so that in a view of the whole egg
it appears to be made up of a number of round segments, all
resembling one another in size and shape; a surface view of
such an egg is figured in fig. 23. It is this appearance which
probably led Hutton and Kennel to state that the segmenta-
tion was holoblastic, but the fact that it is only due to the
yolk segmentation is quite clear when sections of the egg are
examined.

The yolk-segments are much smaller at the protoplasmic
area, and the yolk-spheres composiag them are also smaller
than they are over the rest of the egg.

The protoplasm is still mainly confined to the protoplasmic
area, but small quantities are present in other regions; it con-
sists of a reticulum very indefinitely segmented, the whole
being intimately connected by strands passing over from one
aggregation of protoplasm to another. Nuclei are scattered
about very irregularly through the protoplasm, in some places
two or three lying close together, while in others there is
a considerable tract of protoplasm devoid of any nucleus.
Sections through the protoplasmic area of such an egg are
shown in figs. 7 a and 7 0.

A surface view of an egg slightly older than the preceding
is shown in fig. 24; the yolk segmentation on the surface of
the egg has been obliterated in the course of preservation, but
at the protoplasmic area the segments are clearly seen, the
presence of the protoplasm having rendered the surface at this
point less easily disintegrated. This egg bears a very close
resemblance to those figured by Mr. Sedgwick in the first part

DEVELOPMENT OF PERIPATUS NOV#-ZEALANDIA. 211

of his work on the ‘ Development of P. capensis’ (fig. 7). A
section through the protoplasmic area of this egg is shown
in fig. 5. The protoplasmic masses, which in surface view
appeared to be separate from one another, are seen to be very
closely connected by strands; in two places two nuclei are seen
lying close to one another, and in another a single nucleus is
cut through. The yolk, situated below the protoplasm, is
segmented. Fig. 6 represents a section of a small portion of
the protoplasmic area drawn under higher power, in which a
large number of nuclei are crowded close together in a small
area of reticulated protoplasm.

In the next stage the protoplasmic segments with their
nuclei extend over a somewhat larger area of the surface of
the egg, the nuclei being still very irregularly scattered
through the protoplasm. This extension of the protoplasm is
shown in fig. 8, which is drawn from the protoplasmic area of
a section of this age. The segmeuts are rather more distinct
from one another than they are in the eggs so far described,
but they are still connected by protoplasmic strands. The
yolk is very definitely segmented. Fig. 9 is from a section
through an egg of about the same age; in it the protoplasmic
segments are much more distinct than is usually the case at
this stage.

In all the above-described stages many of the nuclei show
indications of karyokinetic figures, so that it is probable that
all the nuclei are derived by division of the first segmentation
nucleus, and it is not necessary to suppose that any process of
free-nuclear formation has taken place.

In the latest segmentation stage which my material has
provided the protoplasmic segments extend over rather more
than half the surface of the ovum. They are arranged in a
regular layer near the periphery, and appear to be more
definitely separated from one another than in the previous
stages, although it is probable that they are connected by
strands which are hidden by the yolk which separates them.
Nuclei are present in many of the segments, although some
are devoid of them. In the centre of the protoplasmic area

212 LILIAN SHELDON.

there is a mass of protaplasm lying quite on the periphery, and
extending over it fora small area. In it there is a large round
nucleus with very evident traces of a karyokinetic figure.
Small irregular branched masses of protoplasm are present
near the periphery at the lower side of the egg, but they
contain no nuclei. In the centre of the egg there are two or
three very definite protoplasmic masses, none of these contain
nuclei, but in one there are three very definite chromatin
granules. Fig. 10 represents a section through this ovum.
Between the last-described ovum and the next one which I
have, there is a considerable gap. In this ovum, a section
through which is represented in fig. 11, the yolk is still seg-
mented, and ruclei are present scattered irregularly through-
out the ovum, being more plentiful near the periphery than
towards the centre. In one region there is a special aggrega-
tion of nuclei lying in a loose protoplasmic reticulum; this
mass of nuclei is situated on one surface of the egg, its long
axis being parallel to the long axis of the latter, and extending
through about the middle third of its length. In transverse
section it is irregularly triangular in shape, the apex being
directed towards the centre, and the base forming the peri-
phery of the egg in this region. The protoplasmic reticulum
passes without any sharp line of demarcation at its edges into
the yolk. Fig. 12 represents the protoplasmic portion of the
ovum ; it is drawn from the same section as fig. 11, but under
Zeiss’ obj. D instead of obj. A. The nuclei vary much in
size, and some stain very much more deeply than others; they
are extremely irregularly arranged, there being in some places
a group of several crowded close together in a small area of
protoplasm; this is very noticeable in one place in fig. 12,
where eleven small nuclei lie together in a small oval mass of
protoplasm ; in other places they are much farther apart, and
sometimes there is a fairly large area of protoplasm devoid of
nuclei. A few yolk-spheres are present among the meshes of
the reticulum. There is no trace of any cell boundaries, the
protoplasm forming a very loose reticulum, the whole mass
being everywhere connected together by strands. The impos-

" , .
Ee Le

eo, rN

DEVELOPMENT OF PERIPATUS NOVZH-ZEALANDIA. 213

sibility of fixing cell limits is rendered still more obvious by
the irregular arrangement of the nuclei. Traces of a chromatin
network, more or less distinct, are visible in most of the
nuclei; the smaller ones as a rule stain more deeply than the
larger.

A section of an ovum of the next stage is figured in fig. 13,
in it the reticulum of the protoplasmic area has become much
more compact, and is flattened against the surface of the egg.
At the same time its width laterally has increased, so that it
spreads over a larger surface at the periphery of the egg. In
fact the protoplasmic area might be described as forming a
flattened plate on one surface of the ovum, throughout rather
more than one third of its length; the lateral edges of the
plate show a slight tendency to turn inwards away from the
periphery of the egg. A trace of the triangular shape
presented by the protoplasmic area in the last stage still
persists in a low pointed ridge which runs along the middle of
the plate, and projects inwards towards the centre of the ovum.
Fig. 14 represents the protoplasmic area, it is drawn from the
same section as fig. 13, but under a higher degree of magni-
fication. In it the protoplasm is seen to consist of a fairly
close reticulum, in which the nuclei are packed very near
together. The nuclei themselves possess the same characters
as those shown in fig. 12, and like them are of various sizes.
There is still no trace whatever of any cell divisions, the
protoplasm forming a continuous mass in which the nuclei lie
quite irregularly. The tendency of the lateral edges of the
plate to turn inwards is shown in this figure. Nuclei are still
present scattered through the yolky part of the ovum, and, as
in the last stage, are more numerous towards the periphery.
Traces are visible of the segmented character of the yolk, but
this is not very clearly shown owing to the yolky part of the
egg having broken and fallen out to some extent in the course
of cutting the sections.

Between this stage and the one now to be described there
is again a large gap. A transverse section through the middle
region of the egg is shown in fig. 15 c; in it the appearance is

914 LILIAN SHELDON.

as follows:—The egg is bounded externally by the vitelline
membrane (v. m.); beneath this and closely applied to it is a
peripheral layer of yolk (p. y.), in which are present a number
of small round, highly refractive bodies, which stain very deep
red with picrocarmine. Within this peripheral yolk layer and
forming a ring round the egg is a thin layer of protoplasm
(Ec.), with clearly defined inner and outer boundaries, and a
single layer of nuclei arranged regularly in it. At one point
there is a great proliferation of nuclei forming a conspicuous
mass (p. n.) on the outer side of the protoplasmic ring; its
boundaries are not very sharply defined, so that it passes at its
edge into the yolk without any clear line of demarcation
separating the two. The space inside the ring of protoplasm
is filled with yolk.

In passing through the series of sections from the middle
one just described towards one end of the egg, the proliferating
mass of nuclei is found to gradually thin out, and finally dis-
appear, so that, as is shown in fig. 15 d, the protoplasmic
band (c.) comes to be of the same thickness along its whole
circumference. At the same time the diameter of the ring
gradually diminishes, and it finally ends not far from the
extremity of the egg as a cul-de-sac, the blind end of which is
enveloped in the peripheral yolk.

Passing from the central section towards the other end of
the egg, the proliferating mass of nuclei increases in size, and
then becomes divided into two masses (vide fig. 156, p. n.).
These masses are not completely separated from one another,
but are connected above and below by a layer of protoplasm,
in which nuclei are present, so that a second cavity is produced
(fig. 15 6, p.) lying below that enclosed by the protoplasmic
ring, and bounded laterally by the proliferating mass of nuclei,
and above (Sep.) and below by the bands of protoplasm which
connect these together. Near the extremity of the egg both
these cavities end blindly, the secondary one, i.e. that which
lies between the proliferating masses, ending a little before the
original cavity. The blind end of the latter is enclosed in the
peripheral yolk, which at this end of the eggs shows signs of

(Se Ra yg lmmmniane,

DEVELOPMENT OF PERIPATUS NOVA-ZEALANDIA. 215

the original yolk segmentation, and contains a good many
nuclei (fig. 15 a).

Cell outlines are not distinct at this stage, but indications
of them are present between the nuclei of the protoplasm
bounding the sac. There are no traces of any cell boundaries in
the proliferating mass of nuclei.

The structure of the egg at this stage may be briefly described
as follows :—The egg is surrounded by a vitelline membrane ;
beneath this is a peripheral layer of yolk containing small
round, highly refractive bodies; within this is a sac ending
blindly at both ends, bounded by a layer of protoplasm, which
is one cell thick, except along one line, where, throughout most
of its length, there is a longitudinal ridge composed of a mass
of proliferating nuclei, among which many of the small round,
highly refracting bodies are scattered. At one end this ridge
thins out and gradually disappears; towards the other it in-
creases in thickness, and by the parting of its lateral walls
comes to enclose a secondary cavity, which lies below the
primary sac, and ends blindly shortly before the latter. The
sac is filled with food yolk, in which a few scattered nuclei are
present.

By a comparison of this embryo with those of later stages
it is found that the internal sac, together with the proliferating
ridge, represents the embryonic part of the ovum, the single
layer of protoplasm being the ectoderm of the embryo. The
mesoderm and ectoderm are not yet definitely differentiated.
The peripheral yolk does not form any part of the future
embryo, but seems to be absorbed as food material.

The most tenable hypothesis, whereby this stage can be con-
nected with the previous one, is that the cells at the edges of
the protoplasmic plate of the latter grow round the ovum ina
normal epibolic manner, except that, instead of spreading over
its surface, they grow round slightly internal to it, so as to leave
a peripheral layer of yolk outside them. A small quantity of
this peripheral yolk inserts itself between the protoplasmic
plate and the vitelline membrane, so that the whole embryo is
surrounded by yolk. The protoplasmic plate itself probably

216 LILIAN SHELDON.

forms the proliferating ridge. The small round, deeply staining
bodies found in the peripheral yolk have no obvious rudiment
in the previous stage; they present no definite structure, and,
except for their property of staining deep red with picrocarmine,
they resemble the yolk-spheres. It is possible that they may
be derived by the breaking down and alteration of the nuclei
which were present in the yolk in the previous stage. This
view as to their origin is supported by the fact that they are
very much more numerous in the peripheral than in the central
yolk, which was also the case with the nuclei. Whatever their
origin may be, they probably function as food material, as to
a certain extent at this stage, and very largely in later ones,
they are found lying among the cells of the embryo.

By the next stage again the ovum has undergone considerable
changes ; the peripheral yolk is mostly absorbed, but the small
round bodies are still very numerous, lying both outside the
embryo, i. e. between it and the vitelline membrane, and also
among the cells of the embryo, thereby rendering the exact
boundaries of the latter difficult to distinguish. Ina transverse
section through the egg, near its anterior end, the embryo is
seen as a sac surrounded by a layer of ectoderm, which is ren-
dered somewhat indefinite by the intrusion of the small round
bodies. At the ventro-lateral corners there is a pair of pro-
liferating masses of nuclei, which are the rudiments of the pre-
oral lobes. As in the last stage, the whole embryo is filled
with yolk. A section slightly posterior to this is shown in
fig. 17 a; in it the rudimentary preoral lobes (p.o./.) are
present, though they are rather smaller than they were in the
section last described ; lying between them, on the ventral face
of the embryo, is the transverse section of the tip of a small
second sac (post. Em.), which is bounded by a fairly definite
layer of nuclei, and contains in its interior yolk-spheres, small
round bodies, and a few large nuclei. In a section through
the middle region of the ovum, such as is figured in fig. 17 8,
the second sac is found to have increased in diameter, and to
lie on the ventral face of the primary sac, from which it is
separated by a protoplasmic septum. Proliferating masses of

DEVELOPMENT OF PERIPATUS NOVM-ZEALANDI®Z. 217

nuclei (Mes.) are present at the ventro-lateral corners of the
primary sac, and there are also indications of a proliferation of
the nuclei at those corners of the secondary sac which are
apposed to the thickenings on the primary one.

In a section through the posterior region of the egg (such
as is represented in fig. 17 ¢), the septum dividing the two sacs
from one another has disappeared, so that they are in free
communication with each other. ‘Thus the central cavity of
the embryo is continuous from the anterior extremity of the
embryo round the posterior end of the egg to the tip, which
was found lying on the ventral face of the head of the embryo
between the preoral lobes (vide fig. 17 a, post. Em.) The
longitudinal thickenings along the ventro-lateral borders of
the sacs coalesce shortly behind the point where the septum
disappears ; that is, the thickening on the right of the primary
sac coalesces with that on the right of the secondary one, and
similarly those on the left; so that on each side of the embryo
there is a thickened ridge which starts just behind the przoral
lobes, and is continued round behind the septum along the
sides of the ventral sac. These thickenings are the mesoderm.
The endoderm is only represented by a few scattered cells in
the yolk which fills the embryo. The embryo, therefore, con-
sists of a sac which, except at the posterior end of the egg, is
divided into a dorsally- and a ventrally-lying one by a longitu-
dinal horizontal septum. The whole embryo is filled with yolk,
and is surrounded by a thick layer of the small round bodies,
outside which is the vitelline membrane.

Several intermediate stages are obviously wanting between
this embryo and the previous one, and it is therefore not possible
to state positively how the one developes into the other. It
seems possible, however, that the cavity (figs. 15 6, p) which
was present in the proliferating mass of nuclei in the previous
stage, corresponds to that which constitutes the ventral sac in
the anterior portion of the egg last described (fig. 17 a, post.
Em.). If the proliferating mass were anteriorly to divide
completely, only remaining attached by a string of cells on the
ventral surface of the embryo, constituting the ventral ectoderm,

218 LILIAN SHELDON.

a condition would be attained similar to that found at the
extreme anterior end of the embryo, the now paired prolife-
rating masses heing the preoral lobes. Farther back the con-
dition would be similar to that found in fig. 15 4, in which the
proliferating mass has divided centrally, but the divided masses
remain connected above and below by a string of cells, so as to
enclose a secondary cavity lying on the ventral face of the
primary one, the two being separated only by a thin layer of
protoplasm. This may be seen by comparing fig. 15 6 with
fig. 17 6; in the latter section the proliferating mass on
each side has divided, part being applied to the ventro-
lateral corner of the ventral and part to the apposed corner of
the dorsal sac (fig. 17 6, Mes.). The condition found in the
posterior region of the last-described embryo (fig. 17 ¢) would
be attained if the proliferating mass divided, only remaining
connected by a layer of cells above, so that no septum would
be present dividing the cavity of the embryo into two.

The main difference between this stage and the next may be
best seen on examination of a section through the middle
region of the ovum. In such a section (fig. 184) the two
cavities, which were before only separated from one another
by a single layer of protoplasm, are entirely distinct ; each
being bounded on all sides by a definite protoplasmic layer,
and the walls, which are apposed to one another, being com-
pletely separated by a narrow space in which are found some
of the small round elements, which are present in the space
between the embryo and the vitelline membrane. Posteriorly
the two sacs communicate as before (fig. 18c) ; anteriorly the
przoral lobes are very prominent (fig. 18 a, p. o. /.), there is a
slight invagination of ectoderm (S¢.) in the middle ventral
line, where the mouth will be found later, and there is a pair
of definite hollow somites, which are the somites of the
preoral lobes, lying one on each side of the yolk (fig. 18 a, S.1.).
The proliferations of nuclei which constitute the mesoblast
are much larger and more clearly defined than in the last
stage (Mes.) The peripheral yolk is entirely absorbed, but
the whole embryo is still surrounded by the small round

DEVELOPMENT OF PERIPATUS NOV@-ZEALANDIA. 219

elements, which are also very plentiful lymg among the cells
of the embryo, especially in the przoral lobes, and in the
somites. These bodies, which are represented drawn under a
high power in fig. 16, have a somewhat different form in this
egg to that which they have in the previously described ones ;
they are still round in shape, but they contain in their interior
a larger or smaller amount of vacuoles. The region in which
the embryo is doubled on itself is shorter in this egg than in
the preceding stage.

As was mentioned before the main difference between this and
the previously described embryo is in the complete separation of
the two sacs in that region where they are superposed upon
one another. This change seems to have been effected by the
ingrowth of the surrounding tissue, which by pushing in the
septum causes it to become double. This process had already
begun in the anterior region of the ovum of the previous stage,
where in fig. 17 a the ventral sac is seen to be surrounded by
a complete layer of ectoderm, while more posteriorly in the
egg it is only separated from the dorsal sac by a single septum,
as is shown in fig. 17 6.

In the next stage, sections of which are figured in figs. 19
a—d, the separation between the cavities has progressed still
farther, the anterior tip of the ventral cavity, i.e. the posterior
tip of the embryo, lying at some distance from the ventral wall
of the dorsal one, as is shown in fig. 195. The region in
which the embryo is doubled on itself is also shorter than
before, so that it seems to be gradually straightening itself out.
The embryo has also advanced considerably in other respects
—the mouth is present as an ectodermic invagination (fig. 19
6, M.), the inner end of which forms the pharynx ; the przoral
lobes are united in front of the mouth by the cerebral commis-
sure (fig. 19a, Cer. Com.) ; the somites are present as a series
of paired, hollow, thin-walled vesicles lying on the lateral faces
of the body below the ectoderm, which in this region is slightly
thickened. Inthe posterior portion of the body the somites are
not yet present (fig. 19 c, Mes.), the mesoblast being still in
the form of a pair of proliferating ridges of cells. The endoderm

220 LILIAN SHELDON.

is now for the first time clearly differentiated, and consists of
a layer of nuclei surrounded by a very loosely reticulate layer
of protoplasm, around the periphery of the inner yolk mass, and
just within the ectoderm, except in the region of the somites,
where it is subjacent to the splanchnic wall of the latter (figs.
19 a—d, End.). The small round bodies are still present, but
in much smaller quantities than hitherto, both outside and
among the cells of the embryo, from which fact it may be
inferred that most of them have been absorbed; this reduction
is shown in all the four figures (19 a—d).

In the next stage the straightening of the embryo within
the shell had considerably progressed, the posterior end only
being bent at an angle to the main part of the body. There is
as yet no anus, so that the stomodzum is formed considerably
earlier than the proctodeum. The somites are very distinct,
with a thin splanchnic (fig. 20, sp.) and a thick somatic wall
(fig. 20, so.) ; they do not contain in their cavities any of the
small round elements found in them in previous stages. The
ectoderm of the lateral body wall, i.e. that covering the
somites, is thickened, and in the ventral regions of the thicken-
ing rounded elements are present, which will give rise to the
future nerve-cords (fig. 20, n. s.). In the central yolk of this
embryo (fig. 20) there are traces of the yolk segmentation, some
of the segments containing nuclei; whether these traces have
really been retained in this particular embryo longer than usual,
or whether in the last few eggs which have been described they
have been destroyed by the action of reagents is doubtful; the
latter interpretation, however, seems possible, since in young
segmenting embryos whose yolk could in the fresh state
be very clearly seen to be segmented, the segmentation was
only partially discernible ia sections of them when pre-
served.

In all the eggs of stages subsequent to the segmentation
which have been described hitherto, it was not possible to make
out anything in a surface view of the embryo either in the
fresh state or after preservation, owing to the peripheral yolk
which surrounded it and completely obscured its external

DEVELOPMENT OF PERIPATUS NOVA-ZEALANDIA, 221

characters. But in the next stage the peripheral layer has all
become absorbed, and after the removal of the shell the
features of the embryo can be clearly made out. Figs. 25 and
26, which represent an embryo of this stage seen from the side
and front respectively, were drawn from an embryo after it had
been preserved. The przoral lobes are very prominent, and
are separated from one another ventrally by a rather shallow,
wide median groove; the antenne are beginning to bud out as
small protuberances on their anterior dorsal corners. Along
each side of the body is a longitudinal ridge, which is very
clearly discernible by its prominence and also by its opaque
white colour. This ridge is the origin of the appendages, and
is anteriorly divided into distinct lobes, which are the rudi-
ments of the appendages of the anterior segments of the body.
The posterior end of the embryo is bent up almost at a right
angle to the rest of the body, and at the tip where the lateral
ridges meet there is a small papilla, which is perforated by the
anus. The mouth is visible, situated on the ventral surface
immediately behind the przoral lobes. Except on the preoral
lobes, the lateral ridges, and the anal papilla, the body of the
embryo is a dull yellowish colour, which is due to the yolk
shining through the thin ectoderm.

In a series of transverse sections (figs. 21 a—c) through this
embryo the following points are noticeable :

The ectoderm, except over the przoral lobes and the
appendicular ridges, is a thin layer of flat cells. The cerebral
lobes of the nervous system (fig. 21a, 47.) are connected in
front of the mouth by a transverse commissure (fig. 21 a, Cer.
Com.). The ventral cords are continuous with the brain, and
form a pair of longitudinal ectodermic thickenings on the
inner ventral angles of the leg ridges (figs. 21 6 and 21 ¢, n. s.);
they are not definitely separated from the ectoderm, but are
distinguishable from it by the round elements of which they
are composed. The mouth opens into a thick walled pharynx
(fig. 21 a, Ph.), which is in communication with the gut. The
anus is present (fig. 21 c¢, an.) on a terminal papilla, and is
formed by a simple invagination of ectoderm. Immediately

VOL, XXVIII, PART 2,—NEW SER, Q

222 LILIAN SHELDON.

behind the anus in the middle line there is a mass of undif-
ferentiated cells (fig. 22, pr. st.), which is the primitive streak,
a groove running down its centre being the primitive groove
(fig. 22, pr. gr.). This is the earliest stage at which these
structures are present, and the cells do not appear to be in a
state of great activity.

The Mesoderm.—The somites are present as a series of
paired hollow cavities on each side of the embryo. Their inner
or splanchnic walls are very thin (figs. 21 6 and 21, Sp.), but
their outer are much thickened, and are composed of several
layers of nuclei (figs. 214 and 21c, So.). In the anterior somites
(fig. 21 4) the somatic wall has pushed out the thickened over-
lying ectoderm so as to form protuberances, which are the rudi-
ments of the legs. A slight ventral outgrowth of the somatic
wall towards the ectoderm is the rudiment of the duct of the
segmental organ (fig. 21 6,7. d.). The posterior pair of somites.
communicate with one another across the middle line, and
their posterior walls are fused with the undifferentiated tissues
of the primitive streak.

The endoderm is present in the form of a layer of fairly
regularly arranged nuclei lying in a protoplasmic reticulum
round the periphery of the yolk (figs. 21 a—c, End.). There
are a few nuclei in the central yolk, which latter is much less
voluminous than in previous stages.

GENERAL CoNSIDERATIONS.

The Segmentation.—So far as my material has allowed
me to investigate the subject, the segmentation of Peripatus
nove-zealandiz resembles closely that which has been
described in some other Arthropoda. Quite lately Henking,?
among others, has described the segmentation in the eggs of
certain Phalangide, and his account in many respects agrees
with that which I have given, noticeably in the formation of
the blastoderm and in the irregular arrangement in the young
stages of the nuclei composing it. The segmentation of the

1 Henking, H., “‘ Untersuchungen iiber die Entwicklung der Phalangiden,”
Theil i, ‘ Zeit. fiir wissen. Zool.,’ xlv, 1886.

DEVELOPMENT OF PERIPATUS NOVM-ZEALANDIZ. 223

yolk, however, differs in that he considers each yolk-segment
as a single cell, whereas in P. nove-zealandiz I find no
relation existing between the yolk and the nuclei, it being
quite a matter of chance whether or not a yolk-segment
possesses a nucleus. Moreover, I do not think it necessary to
suppose any free nuclear or cell formation to exist in the
formation of the blastoderm, as there seems to be no reason
why all the nuclei forming it should not be derived by division
of the original segmentation nucleus. As to the nuclei which
appear in the central part of the yolk, it is more difficult to
account for them as originating from any pre-existing nucleus,
as they are very far removed from any one, and the three
chromatin particles in a mass of protoplasm, which are figured
in fig. 10, rather point to the formation of nuclei by a pro-
cess of aggregation of such particles. The segmentation
also bears a considerable resemblance to that of Peripatus
capensis; in fact the differences between the two may in all
probability be accounted for by the presence of the yolk in the
New Zealand species. They also resemble one another in the
absence of any cell outlines, the protoplasm in P. nove-
zealandiz, asin P. capensis, forming a perfectly coutinuous
reticulum in which the nuclei are embedded; this is very
especially noticeable in certain stages in the former species, as
is shown in fig. 12. The curious forms assumed by the nuclei
also resemble those found in P. capensis, although owing to
the difficulty presented by the yolk in cutting sections of the
young stages I was not able to get sections sufficiently thin to
enable me to examine them in detail. It is nevertheless per-
fectly obvious that the nuclei often present the same curious
phenomenon of being divided by septa into numerous compart-
ments. These two points, however, viz. the continuity of the
protoplasm and the forms of the nucleus, have been sufficiently
discussed by Mr. Sedgwick in his papers on P. capensis, and
need not be further considered here.

The mode of development from the segmentation up to the
two last stages described in this paper presents many very
curious facts, and indeed, so far as I know, is without any

224 LILIAN SHELDON.

exact parallel in the animal kingdom. It is particularly un-
fortunate that so many stages are wanting in my material, so
that the exact sequence of events cannot be stated with any
certainty. It seems strange that the early stages of all the
three species of Peripatus should differ so remarkably from one
another, while the later course of development seems to be
nearly similar in all three.

It might be said of the mode of development of P. nove-
zealandiz that the embryo is formed by a process of crystal-
lising out in situ from a mass of yolk among which is a proto-
plasmic reticulum containing nuclei.

The two most remarkable features in the development are
perhaps the mode of nutrition of the embryo and the mode of
formation of the posterior part of the embryo, and it will be
more convenient to discuss these two separately.

Mode of Nutrition of the Embryo.—As has been
shown the embryo derives nutriment from two sources, (a)
the yolk contained within its body; (0) a peripheral layer
of yolk in which are embedded numerous small round,
highly refractive bodies. The former of these sources need
not be considered, as it is similar to that which occurs in many
other eggs, which are loaded with food-yolk, being present in
the hypoblast, but entirely absent in the mesoblast. It is with
the peripheral yolk that we are concerned. The complete
envelopment of the ovum in a thick peripheral layer of yolk is
a very remarkable and unusual mode of embryonic nutrition,
but its object evidently is to supply the ectoderm with a
constant source of nourishment, the yolk first and the small
round bodies eventually being completely absorbed by the
ectoderm cells. It seems possible to regard it as ectodermal
yolk, and it is very probably homologous with the peculiar
arrangement in the ectoderm cells of Peripatus capensis
which Mr. Sedgwick has described in the region of the hump.
He says (Part III, p. 472, ‘Quart. Journ. Mic. Sci.,’ vol.
xxvii) : “ This increase in thickness” (i.e. of the ectoderm) “is
mainly due to the appearance, outside the nuclei, of a layer of
vacuolated protoplasm. The vacuolation . . . . . is 4

DEVELOPMENT OF PERIPATUS NOVM-ZEALANDIA. 225

very striking feature. The surface of the dorsal ectoderm,
particularly of the hump, is very rough in these stages, and in
the best preserved embryos without a definite external boun-
dary. It presents very much the appearance which a bath-
sponge would present in section, fraying out, as it were, into
the surrounding fluid; and one may fairly conclude that
during life it possesses the power of sending out processes into
the fluid surrounding the embryo, and that the superficial
vacuoles open to the exterior. In short, I am inclined to think
that this surface ectoderm during stages © to F has a nutritive
function, absorbing the fluid in which the embryo lies, and it
seems to me conceivable that the placenta described by Kennel
in the Trinidad species may be a more specialised organ of the
same nature.” It seems also conceivable that the peripheral
layer of yolk in P. nove-zealandie may be a more
specialised organ of the same nature, and that originally when
the ovum of P. capensis was provided with yolk, the space
between the egg membrane and the ectoderm was filled with
yolk, as is the case in P. nove-zealandiz, instead of as now
with fluid. I have not been able to find processes on the
external surface of the ectoderm cells, but the boundary is not
very sharp, and the protoplasm passes without any definite
limits into the peripheral layer. This is specially the case in
the stages in which the internal yolk is divided by a longitu-
dinal horizontal septum (vide fig. 17 6 and 17c). The modes of
nutrition of Arthropod embryos are, as is well known, very
variable, and an arrangement somewhat comparable to this is
described by Ganin! as existing in Platygaster, where a layer
of protoplasm containing nuclei surrounds the embryo, both the
protoplasmic layer and the embryo being derived from precisely
similar elements. He describes the first nucleus as arising as
a new formation in the egg; from this another nucleus arises
by division, and from this second one a third. The original
nucleus gives rise by a process of complete segmentation to the
embryo, two later ones undergo division, and becoming sur-

1 Ganin, M., “ Beitrage zur Erkenntniss der Entwicklungsgeschichte bei
den Insecten,” ‘ Zeit. fiir wissen. Zool.,’ xix, 1869.

226 LILIAN SHELDON.

rounded with protoplasm, arrange themselves as a layer round
the embryo, in the formation of which they play no part. He
does not say that this layer is used as food material, but con-
siders it as a protective layer comparable in physiological
significance to the amnion of an ordinary insect development.
This process is very similar to that which occurs in P. nove-
zealandiz, with the exception that the yolk is entirely wanting
in the eggs of Platygaster. The peripheral yolk layer probably
serves both as a nutritive and a protective layer, acting as a
shield for the embryo in its young stages, since it does not
become finally absorbed until the embryonic tissues have
acquired considerable consistency, and so would no longer
require such protection. Thus P. nove-zealandiz seems
to have acquired by an extremely simple method an external
layer which serves at once the double purpose of nourishing
and protecting the embryo in its young stages.

A somewhat similar result is also brought about, although
the means by which it is effected are quite different, in those
insects which undergo an internal development, and in which
the embryo is completely embedded in the yolk. The method
of effecting this is considerably simpler in P. nove-zealandiz
than in these insects, nothing corresponding to the amnion
being present. Itis possible that the amnion is a late develop-
ment, acquired for the protection of the embryo, and that on
its establishment it became involved with the external nutritive
mass. However this may be, it is clear that there are various
modes existing in Arthropods for the protection of the embryo
and the nutrition of the ectoderm, and that these differ very
largely in their mode of origin and in their structure, although
they resemble one another in their physiological functions.

With regard to the small round bodies, which are so con-
spicuous a part of the peripheral layer, I have, as was said
before, no definite knowledge as to their origin. From their
form and structure one would be inclined to believe them to be
derived from the yolk, but this is militated against by the fact
that they stain a very deep red, whereas the yolk-spheres stain
bright yellow, and it is difficult to imagine that some of the

DEVELOPMENT OF PERIPATUS NOV#-ZEALANDI®. 227

yolk-spheres should suddenly change their properties towards
staining reagents. The only other possible mode of origin for
them is from the nuclei, which are present throughout the egg
at the stage before these bodies are present. If this is the
correct solution the nuclei must have been broken down and
considerably altered before they were converted into their
present form, since in the latter they are much smaller, instead
of being granular they are homogeneous and highly refractive,
and they stain much more deeply. Whatever their origin may
be, they are undoubtedly an important factor in the nutrition
of the egg, as they are found very plentifully scattered in the
ectoderm in the comparatively early stages, and are afterwards
absorbed without leaving a trace.

Mode of Formation of the Posterior End of the
Embryo.—As was said in the descriptive part of this paper,
I am unable to state with certainty the exact method by which
the posterior end of the embyro is formed out of the egg; but
however this may be effected, the condition attained in which
the yolk contained in the anterior and posterior portions are
separated only by a single thin layer of protoplasm is very
remarkable, since at that stage the embryo possesses no definite
ventral ectoderm, the ventral surfaces of the anterior or poste-
rior halves being so closely applied to one another that the
single protoplasmic septum belongs equally to each, and cannot
be referred to one more than to the other.

In no other known type of development, so far as I know,
does any process similar to this occur. It seems to have been
acquired as a part of the peculiar crystallising-out process
mentioned before as constituting one of the characteristic fea-
tures in the development of this creature. It is no doubt a
simple method for the formation in situ of the embryo, since
it involves no doubling or growth in length within the egg ;
also, owing to the large space occupied by the peripheral
nutritive layer, the amount of room within the egg is limited
at this stage, and it would not be possible for the embryo to
grow to any extent in length, so that the production of a
doubled-up embryo in situ from a single embryonic mass

228 LILIAN SHELDON.

would doubtless be an easy and rapid mode of formation. It
is not until after the peripheral nutriment has been mainly
absorbed, so that the amount of space within the egg-shell is
increased, that the posterior part of the body loses its close
adherence to the anterior, aud the embryo begins to straighten
out. Also, apart from the space occupied by it, the peripheral
yolk would probably act as a resistant force against a normal
lengthways growth of the embryo. Ganin, in his account of
the development of Platygaster, referred to above, describes a
process whereby a result somewhat similar to that effected in
P. nove-zealandiez is brought about:—the embryo after
the segmentation is completed consists of a solid mass of
cells, the peripheral layer being distinguished from the
central mass by their more columnar form. An invagina-
tion of the ventral surface then occurs, forming a deep trans-
verse fissure extending about half way across the embryo, and
dividing it into an anterior cephalothoracic and a posterior
caudal portion. So that at this stage the embryo has much
the same characters as that of P. nove-zealandiz after the
anterior and posterior regions of the body have acquired their
own ventral walls and have become definitely distinct from one
another. The stage in which the two are separated only by a
single septum does not appear to possess any parallel in the
development of Platygaster. But, apart from this, the forma-
tion in situ of an embryo doubled upon itself from a primi-
tively single and solid mass is very remarkably similar in the
two cases. It would appear to have been acquired as a simple
process, the conditions being, in the fact of the enclosure of the
embryo in a peripheral layer, somewhat similar in both ova.
Origin of the Endoderm.—As I said in my remarks on
the segmentation, the first endodermal nuclei seem to arise by
a process of free nuclear formation. The same may perhaps
be the case with the later endodermal nuclei, since in no case
have I found any trace in them of karyokinetic figures, which
latter are extremely common in all the other nuclei. At the
stage before the embryo is definitely formed, when the flattened
protoplasmic plate is present along one side of the egg, there

+ £4 > ane

DEVELOPMENT OF PERIPATUS NOVM#®-ZEALANDIA, 229

are a few nuclei present in the central yolk, and these probably
are the early endodermal nuclei. At the stage when the
embryo is first definitely formed, and lies as a sac within the
peripheral layer, there are nuclei present within the body
of the embryo; these are irregular angular bodies with a
granular structure, scattered irregularly in the central yolk, often
containing one or two chromatin particles; their boundary
is often indistinct, so that the nuclear passes imperceptibly
into the protoplasmic substance. Later, the endoderm nuclei
are much more numerous, and are arranged round the peri-
phery of the yolk in a regular manner. Since in no case
whatever at any stage have I found the least trace of any
karyokinetic figures in these endodermic nuclei, and as also I
often find in the central yolk all stages, from small masses of
chromatin up to definite, large, fully-formed nuclei, I am
inclined to believe that they arise by a process of free nuclear
formation, and that no nuclear division takes place, at all
events, till after the stage at which the endoderm is present as
a definite layer at the periphery of the central yolk mass.

SUMMARY.

1. The ripe ovum of P. nove-zealandiz is very heavily
charged with food-yolk, which causes it to be of comparatively
large size.

2. The segmentation is on the centrolecithal type; the
protoplasm is mainly at one pole of the egg, and in this proto-
plasm nuclei arise, probably by the division of the original
segmentation nucleus. The protoplasm forms a loose reticulum
containing nuclei on the surface of the egg, which first extends
over only a small area, but later spreads over the surface, until
in the latest stage which I have, it covers about half the
periphery of the egg.

3. In the latest segmenting ova there are small masses of
protoplasm in the centre of the egg; occasionally one of these
may contain a nucleus, and in one case three chromatin masses
are present in one of these protoplasmic areas.

4, Shortly after the segmentation begins the yolk becomes

230 LILIAN SHELDON.

divided up into a number of rounded segments, which, however,
bear no relation to the true segmentation.

5. The protoplasm is in the form of a reticulum, and there
are no traces of cell outlines.

6. In the next stage, which I have examined after the segmen-
tation, there is aspecially marked area of reticulate protoplasm,
containing a large number of nuclei extending through about
one third of the length of the ovum, and having in transverse
section an irregular triangular shape, the base of the triangle
resting on the surface of the egg. Nuclei are present through-
out the whole of the yolk, being more numerous at the
periphery than at the centre.

7. The triangular-shaped protoplasmic area becomes more
compact and flattens itself out, forming a plate-like mass of
protoplasm densely packed with nuclei on the surface of the
egg. Its lateral edges turn slightly inwards away from the
periphery. The nuclei over the rest of the egg have undergone
no change.

8. The embryo is present as a closed sac, the walls of which
are separated from the vitelline membrane by a thick layer of
yolk, in which small round, highly refractive bodies are present.
The embryo is enclosed in a thin layer of protoplasm, with
nuclei, which represents the ectoderm. Along one line, in a
longitudinal direction, there is a prominent ridge on the outer
side of the ectoderm, composed of proliferating nuclei. Ante-
riorly this ridge divides into two, which remain attached to one
another above and below, in such a way as to enclose a cavity
between them.

9. At the next stage the rudiments of the preeoral lobes are
present in the form of a thickened mass of cells at the ventro-
lateral corners of the embryo. Atavery short distance from the
anterior end of the embryo the yolk is divided by a protoplasmic
septum, which runs in a longitudinal horizontal direction,
and separates the body of the embryo into two sacs, one lying
above the other. The septum stops short at a short distance
from the posterior end, so that the two sacs communicate freely
round its end. At the regions where the septum joins the body

ee

~~ a

DEVELOPMENT OF PERIPATUS NOVMH-ZEALANDIA. 231

wall on each side of both sacs there is a thickening of the cells,
which is the rudiment of the mesoderm. The peripheral yolk
is mostly absorbed, but the small round bodies are still present
in large quantities before, outside the embryo, and among its
cells.

10. The septum has become divided into two layers by a
process of ingrowth of the surrounding tissue, so that each
sac is completely enclosed by a protoplasmic layer, and the
embryo now consists of a sac doubled on itself in such a way
that the ventral face of the anterior part of the body is opposed
to that of the posterior part of the body. Indications of cavi-
ties have appeared in the mesoblastic bands, which are the
rudiments of the somites.

11. The embryo begins to straighten itself out, the ventral
surface of the posterior end gradually removing itself farther
from that of the anterior. The mouth is present as an invagi-
nation of ectoderm on the ventral surface just behind the
preoral lobes. The endoderm is present as a layer of nuclei
surrounded by a reticulum of protoplasm lying at the periphery
of the yolk. The somites are present in the anterior region
in the form of a series of definite cavities at the sides of the
body, The small round bodies outside the body of the embryo
are almost entirely absorbed.

12. The embryo is still further straightened out, so that the
only indication of the doubling is in the fact that the posterior
end of the body is bent up at an angle to the anterior. The
anus is not yet formed. The somites are present throughout
the whole length of the embryo, and in the anterior ones the
somatic wall is thicker than the splanchnic.

13. The peripheral food material is completely absorbed, so
that the embryo lies just within the vitelline membrane and
egg-shell. The appendages begin to appear as blunt rounded
protuberances on a lateral ridge which runs along each side of
the body. The antenne arise as buds on the preoral lobes.
The anus is present, situated on a papilla at the posterior end
of the body, and consisting of a simple ectodermic invagination.
A primitive streak and groove are present, anterior to and on

232 LILIAN SHELDON.

the dorsal side of the anus. The central yolk mass is much
reduced in bulk. The somatic wall of the somites is much
thickened, and in the anterior segments pushes out the ecto-
derm covering it, so as to form the leg portion of the somite ;
a small ventral outgrowth represents the rudiment of the
nephridial duct. The ectoderm over the leg ridges is thickened,
and at the internal ventral angles of this thickening there are
special rounded elements which are the origin of the nerve-
cords. The cerebral lobes of the nervous system are joined
together in front of the mouth by a cerebral commissure.

In conclusion, I wish to express my thanks to Mr. Sedgwick
for his kindness in providing me with my material, and for the
assistance which he has given me throughout my work.

EXPLANATION OF PLATES XII, XIII, XIV,
XV and XVI,

Illustrating Lilian Sheldon’s paper “On the Development of
Peripatus nove-zealandiez.”

List of Reference Letters.

an, Anus. Ap. Appendage. ¢. Antenna. 47. Brain. Cer. Com. Cerebral
eommissure. Ch. Chorion. He. Eetoderm. xd. Endoderm. J. 8. Leg
portion of somite. 2. Mouth. Mes. Mesoblast. x. Nucleus. x. d. Ne-
phridial duct. a. s. Nerve cord. Ph. Pharynx. Pm. Protoplasm surround-
ing nucleus. Pm. 4. Protoplasmic area. Pm. §. Protoplasmic segments.
p. Cavity in proliferating mass of nuclei. p. 2. Proliferating mass of nuclei.
p.o.l, Preoral lobe. post. Em. Posterior end of embryo. pr. g. Primitive
groove. pr. st. Primitive streak. yp. y. Peripheral yolk. §. Somite. Sep.
Septum between the two cavities. So. Somatic wall of somite. Sp. Splanchnie
wall of somite. S¢, Stomodeal invagination. V.m. Vitelline membrane.
Y. Yolk in the embryo. Y. S. Yolk segments.

All the figures, except Nos. 23, 24, 25 and 26, were drawn with Zeiss’s
camera lucida; the power under which it was drawn is stated after the
description of each figure.

DEVELOPMENT OF PERIPATUS NOV#Z-ZEALANDIZ. 23838

Fic. 1.—Transverse section through an unsegmented ovum, in which the
nucleus is at some little distance from the periphery. The yolk is unsegmented.
m. Nucleus. Y. Yolk. Pm. Protoplasm surrounding the nucleus. Oc. 2,
obj. B.

Fie. 2.—Section through a small portion of an unsegmented ovum, showing
the nucleus situated close to the periphery. 2. Nucleus. Pm, Protoplasm
surrounding the nucleus. Y. Yolk. V. m. Vitelline membrane. Ch. Chorion.
Oc. 2, obj. D.

Fics. 3a@ and 34.—Transverse sections through an ovum, in which two
nuclei are present, and in which the yolk has not begun to segment.

Fig. 3a passes through one nucleus, which is situated at the periphery of
the egg, and has a lobed form.

Fig. 34 passes through the other nucleus, which is at the same pole as
that in Fig. 3a, but lies deeper in the egg. It is in a state of division,
the section passing transversely trough the spindle. x. Nucleus.
Pm. Protoplasm surrounding the nucleus. Y. Yolk. Oc. 2, obj. B.

Fies. 4a—c.—Three transverse sections through an ovum containing three
nuclei, and in which the yolk has begun to segment.

Fic. 4a shows the whole egg. At one pole a round nucleus is present,
and the yolk has begun to segment, the yolk-spheres at that pole
being smaller than over the rest of the egg. Oc. 2, obj. CC.

Fig. 44 shows only the pole of the egg containing the second nucleus
and the yolk segmentation. The nucleus is much lobed. Oc. 2, obj. D.

Fig. 4c passes through the third nucleus, and shows only a small portion
of the egg. The nucleus is lobed and very irregular in shape. xz.
Nucleus. Pm. Protoplasm surrounding the nucleus. YF. Yolk. FS.
Yolk segments. Oc. 2, obj. D.

Fic. 5.—Transverse section through the protoplasmic pole of the ovum,
which is shown in surface view in Fig, 24. The protoplasm is seen to consist
of a reticulum, in which the nuclei lie irregularly. 2. Nuclei. Pm. Proto-
plasm. Y. 8. Yolk segments. Oc. 2, obj. CC.

Fic. 6. Section througi a small portion of the protoplasmic pole of the
same egg as Fig. 5, highly magnified, to show the irregular arrangement of
the nuclei, twelve of them being closely packed together in a small reticulate
area of protoplasm. Oc. 4, obj. E.

Fic. 7a¢.—Transverse section through the protoplasmic pole of an ovum,
slightly older than that from which Figs. 5 and 6 were drawn, showing how
the protoplasm has spread over a larger portion of the surface of the egg; it
still forms a perfectly continuous reticulum. Oc. 2, obj. D.

Fic. 7 4.—Section through the same ovum near the limit of the protoplasmic

area, to show how the protoplasmic areas are connected together by strands,
Oc. 2, obj. D.

234 LILIAN SHELDON.

Fic. 8.—Transverse section through the protoplasmic pole of an ovum, in
which the protoplasm extends over a larger area of the surface of the egg
than it did in the preceding figures. The protoplasmic segments are rather
more distinct from one another than they were in the preceding figures, but
are still connected by strands. Pm. 8. Protoplasm segments. Y. 8. Yolk
segments. V.m. Vitelline membrane. x. Nuclei. Oc. 2, obj. D.

_ Fie. 9.—Section through halt of an egg of about the same stage as that shown
in Fig. 8, in which the protoplasmic segments are more distinct from one
another than is usually the case. z. Nuclei. Pm. S. Protoplasmic segments.
Y. §. Yolk segments. . V. m. Vitelline membrane. Oc. 2, obj. CC.

Fic. 10.—Transverse section through an egg, in which the protoplasmic
segments have extended fully half-way round the periphery. The protoplasmic
areas are separated from one another by considerable tracts of yolk ; one area
lies quite at the surface, and contains a large round nucleus which appears
to be about to divide. Three protoplasmic masses are present in the central
yolk, one of which contains three chromatin particles. The yolk does not
appear to be segmented, but this may be due to the action of reagents. This
figure was compounded from two sections. Pm. §. Protoplasmic segments.
Y. Yolk. xz. Nuclei. Oc. 2, obj. CC.

Fic. 11.—Transverse section through the middle of an ovum, in which
there is a special area of protoplasm at one pole forming a reticulum, in
which many nuclei lie. Nuclei are also present scattered through the yolk.
The yolk is segmented. Pm. A. Protoplasmic area. Y. S. Yolk segments.
Oc. 2, obj. A.

Fie. 12.—The protoplasmic area of the ovum shown in Fig. 11, more highly
magnified, to show the reticulate arrangement of the protoplasm, the absence
of cell outlines, and the irregular arrangement of the nuclei. Oc. 2, obj. D.

Fic. 13.—Transverse section of an ovum, rather older than that from which
Fig. 11 was drawn. The protoplasmic area (Pm. 4.) has become more com-
pact and flattened. The nuclei in the rest of the egg are more numerous
round the periphery than in the centre. The ovum is broken at two points.
Pm. A. Protoplasmic area. V.m. Vitelline membrane. Oc. 2, obj. A.

Fic. 14.—The protoplasmic area shown in Fig. 13, more highly magnified,
to show the absence of cell outlines. Oc. 2, obj. D.

Fries. 15a—d.—Four transverse sections through the youngest ovum, in
which the embryo is definitely formed. Hc. Ectoderm. p. Cavity in the
proliferating mass of nuclei. yp. z. Proliferating mass of nuclei. p. y. Peri-
pheral layer of yolk. V.m. Vitelline membrane. Y. Yolk within the embryo.
Oc. 4, obj. A.

Fig. 15a. Section through the anterior end of the ovum, in front of the
embryonic region, showing the segmented condition of the peripheral
yolk in this region.

DEVELOPMENT OF PERIPATUS NOVA-ZEALANDIA. 2ad

Fig. 15. Section through the anterior part of the embryonic region,
showing the embryo surrounded by the peripheral yolk layer.

Fig. 15¢. Section through the middle of the embryonic region, showing
the embryo surrounded by the peripheral yolk and enclosed in the
ectoderm, on one point of which is the proliferating mass of nuclei.
The small round bodies are shaded very dark.

Fig. 15d. Section through the posterior end of the embryo, shortly an-
terior to its termination and behind the region of the proliferating
ridge.

Fic. 16.— Shows a group of the small round bodies of the peripheral yolk
layer from the embryo shown in Figs. 18a—c, highly magnified. They are
vacuolated. Reichert’s ;; oil immersion.

Fies. 17a—c.—Three sections through an embryo, somewhat older than that
from which Figs. 15a—d were drawn. Hc. Ectoderm. Mes. Mesoblast.
p.o.l. Preoral lobes. post. Hm. Posterior tip of the embryo. Sep. Septum.
V. m. Vitelline membrane. Oc. 4, obj. A.

Fig. 17a is a somewhat oblique section through the anterior end of the
ovum. It passes through the posterior tip of the embryo (post. Hm.),
which is distinct from the ventral wall of the anterior end, being sur-
rounded by a complete layer of ectoderm. Owing to the obliquity of
the section the right preoral lobe is considerably larger than the left.

Fig. 174 is through the middle of the ovum, where the anterior and
posterior ventral surfaces of the embryo are only separated from one
another by a single protoplasmic septum (Sep.)

Fig. 17¢ is through the hind part of the egg, behind the region of the
septum, where the anterior and posterior portions of the embryo are
in free communication with one another.

Figs. 18a@—c.—Three transverse sections through an ovum, rather older
than that figured in Figs. 17a—c. He. Ectoderm. Mes. Mesoblast. p. o. /.
Preoral lobes. p.y. Remains of peripheral yolk. §. Somites. S¢. Stomodeal
invagination. V.m. Vitelline membrane. Y. Yolk. Oc. 3, obj. A.

Fig. 18a@ passes through the anterior end of the embryo, in the region of
the preoral lobes. The ectoderm has begun to invaginate in the middle
ventral line to form the stomodeum, and the somites of the preoral
lobe segment are present (S.1.). The space between the anbryo and
the vitelline membrane is occupied by a large number of the small
round bodies, which are also present among the tissues of the embryo.

Fig. 184 passes through the middle of the ovum. The ventral surfaces
of the anterior and posterior regions of the body are completely
separate from one another. The somites of the trunk are beginning
to appear (S.)

Fig. 18¢ passes through the hind of the embryo, where the anterior and
posterior portions of the embryo are continuous.

236 LILIAN SHELDON.

Kies. 19a—d.—Four rather oblique transverse sections through an ovum,
in which the embryo is still doubled on itself, though rather less so than in
the ovum from which Figs. 18a—e were taken, and the space between the
apposed ventral surfaces is greater. In this ovum the endoderm is first
definitely differentiated as a distinct layer. A few small round bodies are
still present between the vitelline membrane and the ectoderm, but most of
them have by this time been absorbed. Cer. Com. Cerebral commissure.
Ee. Ectoderm. Had. Endoderm. MM. Mouth. Mes. Mesoderm. p.o.l.
Preoral lobes. post. Hm. Posterior tip of the embryo. §. Somite. V.m.
Vitelline membrane. Oc. 3, obj. A.

Fig. 19a passes through the anterior end of the embryo in front of the
mouth, in the region of the cerebral commissure.

Fig. 194 passes through the mouth of the embryo. The posterior tip
is shown lying on the ventral side of the ovum, separated by a con-
siderable space from the mouth.

Fig. 19¢ is from a section posterior to the above. It passes through
the anterior part of the embryo behind the mouth, on one side passing
through the posterior end of the preoral lobe; and through the pos-
terior part of the embryo in the region behind that where the somites
are present, the mesoblast (J/es.) being solid.

Fig. 19d passes through the posterior end of the ovum, in the region
where the anterior and posterior portions of the embryo are con-
tinuous with one another.

Fic. 20.—Transverse section through an embryo which has become almost
completely straightened out. The somatic wall (So.) of the somites is
thickened, as also is the ectoderm lying over it. Hd. Endoderm. z.s.
Rudiments of ventral nerve-cords. §. Somite. Sp. Splanchnic wall of
somite. So. Somatic wall of somite. Oc. 4, obj. A.

Ties. 21a—c.—Three transverse sections through the embryo, which is
shown in surface view in Figs. 25 and 26. az. Anus. Cer. Com. Cerebral
commissure. zd. Endoderm. dr. Brain. JZ. S. Leg portion of the somite.
z.d. Rudiment of nepbridial duct. z.s. Ventral nerve-cord. PA. Pharynx
S. Somite. Sp. Splanchnic wall of somite. So. Somatic wall of somite.
Oc. 4, obj. A.

Fig. 21a is taken just in front of the mouth, in the region of the cerebral
commissure. The pharynx is seen in communication with the gut.

Fig. 214 passes through the appendage of the left side, and shows the
somite dividing into a leg and a body portion, and also the ventral
outgrowth which will form the nephridial duct.

Fig. 21¢ passes through the anus, and the lateral ridge behind the region,
where it is divided into appendages.

Fic, 22,—Shows the primitive streak, and is taken from a section a little

DEVELOPMENT OF PERIPATUS NOV#E-ZEALANDIZ. 237

posterior to that shown in Fig. 2le, but it is more highly magnified. zd.
Endoderm. pr.g. Primitive groove. pr. st. Primitive streak. Oc. 2, obj. D.

Fie. 23.—Surface view of a segmenting ovum, to show the yolk segmen-
tation. Ch. Chorion. YF. §. Yolk segments.

Fie. 24.—Surface view of a segmenting ovum. The protoplasmic segments
are seen lying on the surface of the egg. The yolk segmentation is not seen,
owing to the surface of the egg having been slightly disintegrated by the
preserving reagents. The chorion and vitelline membrane have been removed.

Fig. 25.—An embryo, in which all the peripheral nutritive layer has
been absorbed, viewed from the side; showing the large preoral lobe with
the antenna budding out from it, and the lateral ridge with five distinctly
formed appendages (4p.).

Fic. 26.—The same embryo seen from the ventral side, and showing the
mouth and anus and primitive groove, in addition to the structures seen in
the last figure. These two drawings were from an embryo preserved in
pieric acid. az. Anus. Ap. Appendages. d?¢. Antenne. M. Mouth. p.o. J.
Preoral lobe. pr. gr. Primitive groove.

VOL, XXVIII, PART 2 —NEW SER. R

Pié
iY >"

\\|

Moor Sour Veo NSF? XL

F Huth, Lith? Edin?

‘

Geen on

06,005 (ony
£599 on 0020

0,989 G90 Soo
fo) £0 00,99 096

[0 A
0% $203
Qe o. ol

F. Huth, Lith’ Edin®

2

thy

OO Mor LOUNT. VOU AA VL, HD. APV. ALL.

os S%, 0 8o °

Jo p

i)

. ~ lane. . gO ee 0 ox
fg. L7 6. FS) @ 2 6. 2°00 95
3 / 3 - Oo

62.900 0° O000?
605°, oF, elas Ue
9090,0° 0," Ow
%96° O ‘ae oo
Mes fe 80 =, = Of
e ed 5) 80. =
0007
oe

ee)
ie)

OG
OO?

AG 29
OO¢ :

OOS 90 O oa
200 Oo O
Ob a
450 ae) on

QO
0G oO One
ore (e)
O OG 7G) o
a) = fc)
~~ a. 0 AG
Oe
\ i) yO Tg
A oAO *.
e i. O a
Weg e”

Mice Lournt Ves KUM NSPE AN

Vem.

Fig. 16.

Oo,
@)
SX
else)
Be
(@)
OQ

Os
O Oo O85 0002
3 e HO POCO 880 02 BOG O
fig 7a BORGO OVO OS

F. Huth, Lith Edint®

- toe ex,

= 7 ve
he oe ————

ee . . =

“oO ete pi! ey

Wittnees: sees PO

7 Q oe 4 ce

/ 9 “poses 8 agp ct 6 2580 Lc rN
J O92 296 og ods *, 00009 OsRos% Sas
O50. Ons 2 nn GO POOPIE SEPA,
x

a tS

eS

.

ewe) * Rone iO)
Oo Oo» : 5 ARO RO!

pe

Mor Bourn XUN PI

O) ®) ais" ee
rate wee
joe oes beye\
L } °
i et BD feo te e
A: 2 os?

e ‘3 .
4) wee eh
r) O88 “|

5)
26
ele
one
=A ~
Se
sees
ce?

©

0
heya ;
poooe
P8965, O
oP on,>
sat ee © * .
oes
ae oe
Ss @' ?
Py SS

= °
Oe pe @
ON & *e/ Mes.
eYoy Yee.
Ce Ss
gee

. Se) ee - 0 OMF
Fig 19d. fo0200.9 9 58 a

F Huth, Lith? Edin®

w

WA,

3

Mor Sourn el HU ASP xy

é
DomgTaEPa».

ee i
oa? A Oe ®
D9 BU eS
we Sa Pane

2OO
OOo

MEEDO
- Fass ss

ae

Pe.

ae

F Huth, Lithé Ediné

SOME POINTS IN THE ANATOMY UF PULYUHATA, 239

On Some Points in the Anatomy of Polycheta.
By

J.T. Cunningham, B.A., F.R.S.E.,
Fellow of University College, Oxford.

With Plates XVII, XVIII and XIX.

1. NePHRIDIA AND GONADs.

Brrore the year 1880 it had been conclusively shown by the
results of various researches that a “‘segmental organ” was a
glandular tube, either simple or convoluted, ciliated through
the greater part of its length, having internally a ciliated
funnel-shaped opening by which it communicated with the
body cavity, externally an opening to the exterior. ‘The tunc-
tion of the segmental organ had been shown to be excretory,
depurative, but it had also been proved that in some cases at
least the organ served as efferent duct for the reproductive
elements. The segmental organs in their typical form had
been found principally in the Chetopoda, but it had been
shown that the tubular glands known as the “Organs of
Bojanus” in Mollusca conformed in all essential features to
the plan of a segmental organ, though they did not asa rule
act as genital ducts.! Notwithstanding that these facts had

1 'The generalisation that excretory tubes similar to “segmental organs,”
were homologous structures in whatever division of the animal kingdom they
occurred, was formulated by Lankester in 1877, in his ‘‘ Notes on Embryo-
logy and Classification,” published in this Journal. To the morphological
element so defined he gave the name Nephridium, as applicable in every case.

This view of the original morphological identity of excretory organs was con-
firmed by the discovery of typical nephridia in Peripatus, which was made by

240 J. T. CUNNINGHAM.

been firmly established as important landmarks in zoology, in
the year I have mentioned appeared a memoir bearing the
name of L. C. Cosmovici,! in which the whole question of
segmental organs was thrown into confusion. Overlooking or
rather wilfully disregarding the more recent investigations
which had brought to light the true relations of the segmental
organs, and which had resulted in a generally applicable
definition of the organs, this memoir takes the original work
of Williams as a foundation, and carefully compares every case
examined with Williams’ original description. Because
Williams did not perceive the glandular nature of the organs,
but only their connection with the reproductive functions,
Cosmovici defined a segmental organ as a generative duct,
while a glandular tube or pouch must be considered and called
an organ of Bojanus. As the real segmental organs, or as they
are now called nephridia, are usually both glands and genital
ducts, they were stated in this extraordinary paper to be usually
compound, consisting of a glandular tube, the organ of
Bojanus, and a ciliated non-glandular funnel, the segmental
organ; while certain cases were described in which the primi-
tive distinctness of the two structures was, as the author
believed, retained. Perhaps the most surprising point in this
phenomenal paper is that the absence of an internal or coelomic
opening is taken as characteristic of the “organ of Bojanus”
in the Polycheeta, and this in 1880 when to every commencing
student of zoology was demonstrated the internal or coelomic
opening in the molluscan organ to which the name organ of
Bojanus was originally given. It would be best if it were
possible to exclude the paper of Cosmovici entirely from the
Balfour in 1879 (this Journal). The Hertwigs have argued (‘Die Coelom-
theorie,’ 1881), that the excretory organs of Mollusca are not perfectly homo-
logous with those of Chetopoda, Vertebrata, &c., but it has been shown that
the pericardial space in Molluscs belongs to a system of mesodermic cavities,
distinct from the vascular cavities, and it is pretty generally agreed now, that

thesé mesodermic cavities in Molluses as well as those in Phatyhelmia re-
present a ccelom.

1 “Glandes Génitales et Organes segmentaires des Annélides Polychétes,”
‘Arch. Zool. Exp.,’ Tom. viii.

SOME POINTS IN THE ANATOMY OF POLYCHMTA. 241

literature of the subject with which it professes to deal, and
regard it as of merely psychological interest; but the paper
contains detailed descriptions and drawings of the anatomy of
several forms of Polychzeta, and these have been by some later
writers accepted as trustworthy. In a critical paper by R. 8.
Bergh,! published in 1885 which reviews the relations of the
nephridia in the various classes of Vermes, the facts concerning
the Polycheta are mostly taken from Cosmovici’s memoir.
But a great many of Cosmovici’s statements are quite erro-
neous, and although he has described other facts correctly the
false is so mingled with the true, and the actual descriptions of
structure are so mingled with false theoretical views, that it is
not safe to accept anything in his paper without re-examination
of every case.

Arenicola marina, Linneus.

In 1868, when Claparéde’s ‘ Chétopodes du Golfe de Naples’
was published the gonads of Arenicola were unknown. In that
work it is stated that most authors had taken the nephridia to
be reproductive organs, some describing them as ovaries, others
as testes. Quatrefages (‘ Hist. Nat. des Annelés,’ 1865) called
the nephridia simply genital organs. Grube (‘ Zur. Anat. u.
Phys. der Kiemenwurmer’) had assured himself that the ova-
ries were not to be sought in these organs, for he had seen (as
he thought) the ova formed on the exterior of vascular ceca
of the lining of the celom. He was inclined to regard the
nephridia as testicles. This was an impossible view, because,
as Claparéde says, the sexes of Arenicola are separate, but it
seemed very probable to Claparéde that Grube was right as to
the origin of ova, and in support of the opinion of Grube he
gives a figure of one of the cecal pseudhemal vessels surrounded
by a layer of cells. Claparéde then quits the question of the
genital organs, and proceeds to give a not very accurate de-
scription of the nephridia in Arenicola Grubii, Clap. He
says the organs previously described as generative organs are
really segmental organs of a very peculiar structure, and are

‘ «Die Exkretions-organe der Wiirmer Kormes,’ Bd. ii, 1885.

242 J. T. CUNNINGHAM.

only connected with the phenomena of reproducticn as efferent
ducts. There are five pairs placed in the fourth to the eighth
setigerous segments. Three parts are distinguishable in each
organ, the funnel, the gland, and the vascular reservoir. He
describes correctly the fringed and ciliated dorsal edge of the
funnel; but his account of the shape of the glandular part,
which he compares to a comma, does not agree perfectly with
what is seen in A. marina. His figures are not good, though
creditable for the date at which they were made. Claparéde
never saw ova in the interior of the nephridia, but once saw
spermatozoa in the funnel. Of Arenicola marina, Lin., he
says merely that the segmental organs have a great analogy
with those of A. Grubii.

Casmovici’s description of the nephridia and gonads in
Areniola marina is correct in almost every particular; his
figures of minute structure are not satisfactory, but his ana-
tomical figures are clear and accurate; and if it were not for
the absurd manner in which he has separated the nephrosto-
mata as distinct organs from the nephridia themselves, his de-
scription would be worthy of a permanent place in zoological
literature. To him, in any case, is due the credit of having
been the first to discover and describe the true gonads in
Arenicola. He states there are six pairs of nephridia, each
with its external opening, and situated in the third to eighth
somites. The external opening of each organ is situated close
behind the upper end of the corresponding uncinigerous torus.
The funnel or nephrostome is correctly described by Cosmovici ;
it is provided with a free membranous dorsal border, fringed
and ciliated. Along this border runs a pseudhemal vessel, a
branch from the branchial artery, which is given off by the
ventral vessel. (Cosmovici believes the ventral vessel to be con-
nected with the heart, and to be arterial; it is more probably
venous; that is, it probably receives the blood from the
branchiz.) The vessel which runs along the dorsal border of
the nephrostome is continued diagonally backwards and out-
wards across the nephridium, and in this part of its course,
posterior to the nephrostome, it runs through the gonad ; the

SOME POINTS IN THE ANATOMY OF POLYCHATA. 243

tissue of the gonad is, in fact, continuous with the posterior
angle of the nephrostome.

There is not, then, a great deal to be added to Cosmovici’s
account, but there are one or two corrections to be made, and
a re-examination of the subject was necessary because the false-
ness of his interpretation causes his description to be received
with doubt. His account of the position of the nephridia is
inaccurate ; they are situated in somites v—x (the fourth to
ninth chetigerous) inclusive. The first or buccal somite of
Arenicola is destitute of bristles: the following six somites bear
each a dorsal fascicle of capillary chetz, and a ventral torus
uncinigerus, but no branchie; the next thirteen somites bear
each both fascicle and torus, and, in addition, a pair of plumose
branchiz ; the rest of the body, which is of different lengths in
different individuals, is thinner, and devoid of fascicle, torus,
and branchia; it is cylindrical and covered uniformly with
papille. Behind the first four somites are incomplete septa.
The nephrostome of the first nephridium is on the anterior
face of the fourth septum. Between any two successive para-
podia, behind the third, are five constrictions, of which the fifth
is the deepest. The septum, when present, is attached to the
body wall opposite the second constriction. Between the fifth
constriction and the parapodium is a prominent ridge ending
in a sharp edge. The first nephridia of the first pair are some-
what smaller than the rest. The relation of the gonad te the
nephridium is shown in Pl. XVII, fig. 1, which gives a view of
the internal face of the nephridium. In the natural position the
nephridia are covered dorsally by the bands of oblique muscles,
which pass from the sides of the nerve-cord to the line of dorsal
bristles; and in this condition the dorsal lip of the nephro-
stome is horizontal, and its edge is directed downwards and
inwards,

Cosmovici says he searched in vain for a long time for the
genital organs, and discovered the ovary by accident when
examining a piece of the nephridium. I discovered the ovary by
tracing the origin of the ova in the body cavity. In February last
I found two or three specimens which had ova in the body cavity ;

244, J. T. CUNNINGHAM.

in these and in others in which mature ova were not present,
loose cellular masses were often seen in the neighbourhood of
the nephridia. After careful search, several times repeated,
these masses were traced to the cord of cellular tissue already
described as attached to the nephridium; the cord of cells is
merely, as usual, a local development of ccelomic epithelium.
In most specimens the cells of this cord were so undifferen-
tiated that it was not easy to be certain it was a gonad, but in
specimens which contained a few ova in the body cavity young
ova could be recognised in the cord. The reproductive cells
leave the gonad at a very early stage of development, and reach
maturity while floating freely in the body cavity.

Cirratulus cirratus, Malmgren (O. F. Miller).

Keferstein' in 1862 gave a slight description of the seg-
mental organs in Cirratulus filiformis, Kef., bioculatus,
Kef., and borealis, Lam. He made out the relations of the
organs most completely in the first-mentioned species, in which
he describes them as a single pair of ciliated tubes, each bent
on itself, extending through segments 1 to 5, and having an
internal and an external opening. He gives a figure of the
organ as seen in the living animal, and the figure shows both
the external and internal openings in the first setigerous or post-
buccal somite. Less complete descriptions and figures are
given of the organs in the other two species. No other
nephridia are mentioned by Keferstein except this pair at the
anterior end.

Claparéde also saw but a single anterior pair of nephridia in
species of Cirratulus. He says (‘Ann. Chet. Naples’): “C.
chrysoderma, like all the Cirratuliens, has only a single pair
of segmental organs, opening at the second segment (first seti-
gerous) by an oval aperture situated on the inner side of the
ventral bristles. The organ is rolled in an angular spiral. Its
external part is narrow, but soon enlarges suddenly into a wide
ciliated tube. The internal opening has escaped me.”

1 ¢Zeits. f. wiss. Zool.,’? Bd. xii. 1862.

SOME POINTS IN THE ANATOMY OF POLYCHATA. 2465

Cosmovici (loc. cit.) was unable to see the large anterior
nephridia described by Keferstein and Claparéde, but states
that in C. filiformis, Kef., segmental organs are present in
pairs in nearly all the segments, especially of the middle and
posterior region: that they are attached to the anterior face of
each diaphragm: the figures of this species which he gives do
not show the organs mentioned. The Errantia, among which
he places Cirratulus, according to Cosmovici’s peculiar views,
have only segmental organs and no organs of Bojanus, and he
denies altogether that the internal opening of the segmental
organ communicates with the cavity of the somite in front of
the one which contains the organ itself.

In Cirratulus cirratus both the large anterior pair of
nephridia described by Keferstein and Claparéde, and the
series of pairs in the middle and posterior region mentioned by
Cosmovici, are present. The nephrostome of the first nephri-
dium opens into the cavity of the buccal somite, being situated
on the anterior face of the first septum, somewhat ventrally,
in the angle between the septum and the lateral body wall.
The proximal part of the bent tube passes backwards from the
nephrostome tillit reaches the second septum (P1l. XVII, fig. 3),
then passes upwards to the dorsal body wall, where it opens
into the wider distal part of the tube which opens to the
exterior beneath the neuropodium of the second somite. The
posterior nephridia are smaller and simpler; they appear first
in the twelfth somite, and are repeated hence to the end of the
body. Each of them has a nephrostome of the typical form,
an elongated funnel with its aperture directed forwards. The
nephrostome has a similar position to that of the large anterior
nephridium, that is to say, it is placed in the lower external
corner of the anterior face of its septum. The lips of the
funnel are composed of a columnar ciliated epithelium resting
on a thin fibrous membrane; this membrane is continued on
the one hand into the transverse septum, on the other into the
body wall; the lips of the funnel project inwards and forwards
into the cavity of the somite. The nephridial tube when
traced from the nephrostome (in a series of horizontal sections)

246 J. T. CUNNINGHAM.

is seen to pass obliquely backwards and downwards, curving
over the dorsal edge of the ventral longitudinal muscle, and
opening beneath the neuropodial bristles. The internal open-
ings of the simple nephridia are shown as seen in a horizontal
section in fig. 4, while fig. 5 shows the external opening in a
transverse section of a female specimen.

The simple nephridia act as efferent ducts for the reproduc-
tive elements in both sexes. I found a number of specimens
distended with the genital products at the end of March in the
current year, lying under stones on the banks of Granton
Quarry, where they are pretty abundant at all seasons. These,
when placed in a basin of sea-water, commenced to shed eggs
and spermatozoa. The ova were fastened together, after their
escape, by transparent gelatinous mucus, which formed a soft
mass without any definite shape adhering to the stones and
mud among which the worms were lying. The escape of the
ova could be observed without much difficulty.

Fig. 6 shows the appearance presented by two somites of the
worm viewed under a low power by reflected light, as the ova
were escaping; the apertures seen are the openings of the
simple nephridia previously described. In sections of a male
specimen these nephridia are seen full of spermatozoa. The
larger nephridia of the second somite do not, as far as I have
been able to ascertain, transmit the sexual products; indeed,
no ova (or spermatozoa) are produced in the first eleven
somites where simple nephridia are absent.

I have not discovered in my sections any satisfactory indica-
tion of the places where the germinal cells are developed; the
position of the gonads is doubtful. The reproductive cells
undergo the greater part of their development in the body
cavity. The presence of a complete longitudinal vertical
septum above and below the intestine in Cirratulus is remark-
able,

Srionip#a.—Nerine cirratulus, Clap.

This form has not hitherto been recorded as occurring in
the North Sea, either on our own coast or other parts of

SOME POINTS IN THE ANATOMY OF POLYCHATA. 247

north-west Europe. It is, however, common enough between
tide marks in the sand at Granton. Claparéde! does not give
any description of the segmental organs, he merely mentions
that the ovaries are attached to them, and render their study
difficult. There is, of course, no great difficulty in making
out the relations of the nephridia in longitudinal and transverse
sections. These relations are in some small points exceptional,
and longitudinal sections are the most instructive. The nephro-
stome is a wide ciliated funnel, the lips of which are simple
and entire, not produced into lobes. The aperture of the
nephrostome is large and gaping as in Cirratulus. The lower
lip of the nephrostome projects into the cavity of the somite,
while the upper is continuous with the transverse septum, on
whose front face the nephrostome lies. The nephrostome is
situated at the lower and outer corner of the septum, and leads
into a narrow tube, which pierces the septum and then dilates
into a small spherical vesicle, which on the median side is
produced into a point, and to this point the gonad is attached
(fig. 7). A little more to the exterior side the vesicle gives off
a long duct, which passes upwards to the body wall and opens
to the exterior (fig. 8). As a general rule the efferent duct of
the nephridium in Polycheta passes downwards ventrally, and
opens below the level of the neuropodium. In Terebellidz the
external aperture is near the upper end of the uncinigerous
torus, but in this case the ventral band of muscle extends
some distance upwards, and thus, although the nephridial
aperture has an exceptiona] position with regard to the neuro-
podium, it has its normal relation to the ventral muscles. In
Nerine there is no such reason for the extremely dorsad position
of the nephridial aperture; the ventral band of longitudinal
muscles on each side is folded in at its borders, and the two
bands occupy only the ventral surface of the transverse section ;
the dorsal bands in like manner occupy only the dorsal wall ;
the whole of the lateral region of the transverse section, which
approaches a parallelogram in shape, is destitute of thick
muscular bands, and the fascicles of bristles are widely sepa-
1 © Ann. Chét. du Golfe de Naples,’ Geneva, 1868.

248 J. T. CUNNINGHAM.

rated. The efferent duct of the nephridium is therefore not
confined in a narrow space between the edges of the dorsal and
ventral muscle-bands, as is usually the case, and there is no
apparent reason why the duct should not pass between the
ventral muscle and the neuropodial bristles as it does in most
Polycheta.

It is certain that in this species the nephridia act as sexual
ducts in both sexes. I have frequently seen these organs in
the male distended with spermatozoa, and the ova doubtless
pass out in the same way.

Nerine coniocephala, Johnston.

In this species the nephridia have the same character and
somewhat similar relations to those described in the preceding,
but the efferent duct is by no means so long, and the external
aperture is therefore more ventral in position; it is on the
same level as the upper neuropodial bristles, and lies in front
of the neuropodium in the constriction between adjacent
somites (fig. 9). Sections of ripe males show the lumen of the
nephridium full of spermatozoa.

Lanice conchilega, Malmgren.

Several accounts of the nephridia of Terebella conchi-
lega have been given. H. Milne-Edwards (‘ Ann. d. Sci. Nat.
(2) Zoologie, x,’ 1838, p. 220), in a paper published in 1838, on
the circulation in Annelids, describes the vascular system in a
species to which he gives this name, and gives a figure of the
animal opened along the dorsal median line. In this figure
four looped nephridia are distinctly shown, situated behind the
branchial region. The representation of the position and cha-
racter of these organs is perfectly correct, so far as it goes ;
they are the upper parts of the four nephridia belonging to
somites vi—1x. But the paper I refer to does not describe the
nephridia, as it deals with another subject: they are shown in
the figure, and that is all; and in the description of the figure
the organs are referred to as organs of generation.

SOME POINTS IN THE ANATOMY OF POLYCH#TA. 249

Keferstein (‘ Zeitschrift fiir wiss. Zoologie,’? Bd. xi, 1862)
mentions that the structure and number of the nephridia in T.
conchilega are the same as in T. gelatinosa, Kef.; in both
cases he says there are six pairs, each organ consisting of a tube
bent on itself, of which one half is darker, the other lighter ;
the organs belong to segments 3—9.

Cosmovici! gives an erroneous description of the organs; he
says there are two pairs without internal openings, which he
calls “‘ organs of Bojanus,” one of these situated in front of the
cephalic diaphragm, the other immediately behind it, each organ
having an external opening ; and two other pairs, each of which
has an internal as well as an external opening, and is shaped
like an urn; the internal opening is large, and surrounded
with a circular lip. The gonad is attached to the posterior
part of each of these latter organs, which Cosmovici calls seg-
mental organs, and which he says serve as efferent genital
ducts.

The species referred to by these three authors is the Nereis
conchilega of Pallas, Terebella conchilega of Gmelin;
and this is called Lanice conchilega by Malmgren. My
specimens were identified from Malmgren’s description, and
there is no doubt of their identity with the species of that author ;
but there is room for some uncertainty regarding the specific
identity of the specimens referred to by the authors I have men-
tioned. For instance, Cosmovici identified his species by means
of Quatrefages’ ‘ Histoire des Annéles,’ 1865, and there it is
stated that the tube of Terebella conchilega possesses no
hollow fringes at its mouth; these fringes are always present in
the tube of Lanice conchilega, Malmgren. This species is
distinguished by some marked characters; two of them are the
presence of a large vertical lobe on the third somite (second
branchiferous), and the coalescence of the ventral scutes usually
present into a continuous ventral plate.

The fact that in Lanice conchilega the nephridia of a side
communicate together so as to form a longitudinal tube, has

' “© Glandes génitales et Organes segmentaires des Annélides polychétes,”
* Arch. de Zool. Exp.,’ t. viii, 1879-80.

250 J. T. CUNNINGHAM.

been observed by Edouard Meyer of the Zoological Station at
Naples ; and his discovery is mentioned by his permission in a
single sentence in Lang’s ‘Monograph on the Polycladen,’
published in 1884.! But the accessible information concerning
these organs in this species is so inadequate, that R. S. Bergh,?
in a general review of the excretory system in worms, cites
both Lang’s mention of Meyer’s observation and Cosmovici’s
account of the organs as if they were both equally correct.

The true relations of the excretory system are as follows :—
Enumerating the somites from before backwards, and counting
the buccal as the first, we find that the branchiz belong to
somites 11, 111, and Iv: the first notopodial fascicle of capillary
chaetz is on the fourth somite, the third branchiferous; the first
neuropodial uncinigerous torus is on the fifth ; ‘the neuropodial
tori are repeated on every succeeding somite to the end of the
body ; the notopodial fascicles occur only on seventeen consecu-
tive somites. There are traces of transverse septa behind the first,
second, third, and fourth somites, but none in the rest of the
thoracic region, which bears the notopodial fascicles. On dis-
section, four long double nephridial tubes are seen projecting
dorsalwards into the body cavity; the lower parts of these tubes
are covered by strands of the oblique muscles which pass from
the nerve-cord to the neighbourhood of the notopodial bristles ;
careful examination shows that these tubes belong to somites v1,
VII, vil, and rx. Their internal openings can be found imme-
diately behind the fascicle of bristles belonging to somites v, VI,
vu, and viii respectively, but their efferent tubes are seen to pass
down beneath the fascicle of somites vi, vil, vi1I, and 1x. The
lower parts of these efferent tubes are very wide, and it is impos-
sible to separate them from one another. Beneath the fascicles of
the following four somites (x to x11 inclusive) are seen membra-
nous nephridial sacs, which externally at least are inseparable
from one another. These sacs are simple, that is, they are not
composed of a tube bent on itself like the anterior nephridia ;
they scarcely extend above the level of the oblique muscles, and

1 «Fauna u. Flora des Golfes von Neapel,’ xi Monographie.
* «Die Exkretions-organe der Wirmer’ Kosmos, Bd. ii, 1885,

SOME POINTS IN THE ANATOMY OF POLYCHATA. 2051

no internal opening or nephrostome can be found in them. In
front of the most anterior nephridium, that belonging to
somite vir, are seen traces of a rudimentary nephridium (see
fig. 10). In order to trace out the relations of these nephridia
more accurately, the anterior part of a specimen was cut into a
series of horizontal longitudinal sections commencing with the
ventral surface, and the reason why the successive nephridia
could not be isolated from one another was seen on examina-
tion of these sections; the lower parts of the efferent limbs of
the four anterior normal nephridia, in somites vi to 1x, and the
whole of the nephridial sacs in somites x to XIII are in open
communication, forming a wide continuous longitudinal tube
extending from somite vi to somite x111 (see figs. 11 and 12).
Openings to the exterior from this tube were found in somites v1
to 1x inclusive, corresponding to the four large looped nephridia ;
each of these openings was close behind the upper end of an
uncinigerous torus. The internal openings of the same four
nephridia could be traced with ease and certainty; they are
attached to the body wall close behind the notopodial fascicles
of somites v to vill. These openings are wide, and are overhung
dorsally by a longitudinal lip furnished with a series of small
ciliated digitate processes ; lower down, the anterior and posterior
lips of the opening are simple, thick-walled, and ciliated. The
aperture leads into a thin tube, which passes inwards and back-
wards, curving round the inner end of the fascicle of bristles
behind the aperture, and then, crossing the continuous tube,
passes up on the inner or mediad side of the loop, at the apex of
which it is continued into the efferent wider limb of the loop,
which passes down on the outer side to open into the longi-
tudinal tube. Neither internal nor external openings could
be found in that part of the longitudinal tube which is
behind the loops; it seems evident that this part of the
tube represents four somewhat reduced nephridia which have
coalesced, but whose openings have disappeared. Anteriorly
to the four looped nephridia are traces of three others; the
longitudinal tube extends forwards into somite v as if it in-
cluded a nephridium belonging to that somite, but I could find

252 J. T. CUNNINGHAM.

no external opening in this somite; at the angle between the
septum behind somite 1v and the body wall is a very obvious
nephrostome, which ought to lead into the longitudinal tube,
into that part of it corresponding to somite v, but the connec-
tion could not be traced. Nephrostomes were also present
attached to the anterior face of the septa behind somites m
and 111 (the first and second branchiferous), and leading into
tubes seen in somites 111 and rv, but I could find no external
openings in these somites. I could find no nephrostome in
somite 1 (the buccal), nor any trace of a tube in somite 1.
Gonads are present in the form of clumps of deeply-staining,
small, indifferent cells, attached to the exterior of all the
nephrostoma mentioned, seven in all (see fig. 13). The
germinal cells, when still quite undifferentiated, separate from
the gonads, and undergo further development in the celom.
But I found no reproductive elements in the cavity of the
nephridial system, though the body cavity contained them in
quantity, and it is probable that at the mght season they are
expelled through the nephridial cavities. The body cavity
contains, besides the reproductive elements, a large number of
spherical, vacuolated, nucleated cells. This is the only case in
which a communication between successive nephridia has ever
been discovered in any adult invertebrate. It is true that in
the development of Polygordius, according to Hatschek, each
nephridium gives off backwards a prolongation of itself, from
which the next nephridium is formed, and the two remain in
communication for a time; but the connection is soon severed,
and in the adult the successive nephridia are isolated and
independent. In Lanice conchiiega the nephridia have
coalesced together after coming in contact from before back-
wards, the separating membranes having disappeared. The
case is extremely interesting in the fact that we have in it an
approximation to the condition of the excretory system in
Vertebrata; the presence of a metameric series of nephro-
stomata in vertebrate embryos has long ago been seen to
constitute a resemblance between them and Chetopoda, but
no other Chetopod is known which resembles the verte-

SOME POINTS IN THE ANATOMY OF POLYCHATA. 253

brate in having a number of nephridia coalesced to form a
continuous longitudinal tube.

It is surprising to find that, as far as I have been able to
discover, no resemblance to the condition seen in Lanice con-
chilega occurs in any of its near allies. The only species of
the genus Terebella as defined by Malmgren that occurs in the
Firth of Forth is Terebella Danielsseni, but of this I have
only one specimen, and have not examined its nephridia. Of
Amphitrite there are two species in the Forth; Amphitrite
cirrata I have not examined anatomically; in Amphitrite
Johnstoni there are a large number (fifteen to seventeen) of
nephridia forming long loops projecting dorsalwards into the
body cavity in the anterior region; each has its own internal
and external openings, and is isolated and independent. In
Terebellides Stremi there is one pair of large dark-
coloured nephridia in the anterior end, and three pairs of small
rudimentary ones posterior to this.

In Pectinaria belgica there are three pairs, all indepen-
dent; they are described in the following section. In
Melinna cristata there are several pairs all separate.

Pectinaria belgica, Lamarck (Pallas).

Mr. Harvey Gibson! has by carefully neglecting the distin-
guishing differences between this species and Amphitrite
auricoma, Miller, attempted to prove that the two forms are
identical. He points out that in the original figures of Pallas,
of Nereis cylindraria, var. belgica, “the stiff golden comb
shows one continuous and uniform series of teeth, not two series
as in P. auricoma,” and that figures by subsequent authors,
e.g. M‘Intosh and Malmgren, show the two combs in P.
belgica with perfect distinctness. Moreover, certain refer-
ences in Pallas’s text imply that his species had two distinct
combs. Mr. Harvey Gibson concludes that ‘either Pallas’s
draughtsman made an error in most of the figures of P.
belgica, and failed to represent the comb with sufficient

' «Notes on some of the Polycheta,” ‘First Report on the Fauna of
Liverpool Bay,’ Lond., 1886.

VOL, XXVIII, PART 2,—NEW SER. s

254 J. T. CUNNINGHAM.

accuracy, hence leading Miiller into error when comparing his
form with that of Pallas; or Pallas’s figures are correct
(although his references in the text are wrong), and his species
is distinct from that of Miiller (for the condition of the comb
appears to be the only important difference between the two).
Looking at the inaccuracy of the drawings as compared with
var. capensis in Pallas’s work, and taking into account the
indistinctly double series of teeth shown in figs. 5, 8, and 9 of
var. belgica, I think that probably the former view is most
likely to be the correct one. In that case P. auricoma of
Miiller disappears, and becomes P. belgica of Pallas.” How
a zoologist, after actually referring to the description and
figures given by Miiller of Amphitrite auricoma, and to
the description and figures of both species given by Malmgren
in his ‘ Nordiska Hafs-Annulater,’ could suppose the condition
of the comb to be the only important difference between the
two, it is difficult to understand. The two distinguishing
features given by Miiller are (1) the curvature of the tube; (2)
the serration of the margin of the flattened area behind the
palmule. Malmgren mentions both these characters and
figures them, and he examined specimens of both species ; only
Malmgren made the two characters generic instead of specific,
and called Miiller’s species Amphictene auricoma. We
can state with certainty that in our specimens the tube is
perfectly straight, and the margin of the post-palmular area
perfectly entire. The presence of two distinct palmule, as
Mr. Harvey Gibson would have seen if he had studied the
Latin descriptions of Malmgren, is common to the whole
family Amphicteuea.

There are three pairs of nephridia in Pectinaria belgica,
of which the first are the largest; all the organs are of the
usual type, each consisting of a tube bent upon itself and pro-
vided with a nephrostome and an opening to the exterior.
There is a transverse septum separating the buccal somite from
the following; the nephrostome of the first nephridium is on
the anterior side of this septum. The nephridia are brown or
black in colour. The tube of the first reaches dorsalwards

SOME POINTS IN THE ANATOMY OF POLYOCHATA. 255

above the intestine, and its external opening is a little ventral
to the origin of the first branchia. Between the nephridial
opening and the root of the branchia is the aperture of a
peculiar glandular organ whose function we have been unable
to ascertain. On dissection of a fresh specimen this gland is
seen as a milk-white, opaque, cylindrical body about one eighth
of an inch long, free everywhere except where it is continuous
with the body wall round its opening to the exterior. The
efferent duct of this gland is lined by a high columnar epithe-
lium, of which the component cells are solid and columnar;
throughout the rest of the gland there is a layer of long solid
nucleated cells next to the basement membrane, but these are
covered by other layers of large vacuolated cells whose walls
form a network nearly obliterating the lumen of the organ.
The wall of the gland is well supplied with pseudhzmal vessels.
The third somite (i.e. second branchiferous) and the fourth
are unprovided with nephridia, but the latter contains a
nephrostome belonging to the nephridium of the fifth somite ;
the sixth somite is likewise provided with a nephridium whose
nephrostome is in the fifth somite. The nephrostomata are
simple elongated funnels with their apertures directed for-
wards; they are not provided with such a series of digitate
processes as is seen in Lanice and Arenicola. The gonads are
of the usual type, masses of undifferentiated cells attached to
the exterior of the nephrostomata on the mediad side. The
reproductive cells become detached at a very early stage and
pass through the rest of their development in a free condition
in the celom. It is certain that the spermatozoa reach the
exterior by passing through the nephridia; in a series of
sections of a ripe male I saw the nephrostomata and nephridial
tubes distended with them. Between the two posterior
nephrostomata and the body wall pass membranes, which are
rudiments of transverse septa. There is also a rudiment of a
septum between the second and third somites (first and second
branchiferous). The external apertures of the two posterior
nephridia are ventral and posterior to the notopodial set of
their somites.

256 J. T. CUNNINGHAM. .

Nereis virens, Sars.

Claparéde did not study the nephridia in the Nereide.
Ehlers! has given a description of the organs as he saw them
in Nereis cultrifera, Grube, and other species. He says
the segmental organ lies close behind the entrance into the
cavity of the parapodium, on the inner surface of the ventral
body wall, near the lateral border of the ventral muscular band :
that it consists of an easily noticed, almost spherical body,
and an efferent duct, which runs on the ventral body wall
towards the hinder border of the segment, where it opens to
the exterior: that the body of the segmental organ in a large
epitokous female of N. virens was ‘189 mm. by ‘108 mm. in
size. Inthe inside of the body of the organ he states there were a
number of clear cavities, which, when the organ was compressed,
were discovered to be portions of a continuous convoluted canal
which was embedded in the mass of the organ; the inner
surface of the walls of this canal was ciliated. On the upper
surface of the body of the organ was a slightly curved, cleft-like
aperture, with thickened edges, which carried cilia; this cleft
was the internal opening of the organ; on the opposite side of
the spherical body the thin efferent duct passed off. Ehlers
concludes his description thus: ‘‘In support of the view
expressed by myself, that the segmental organs serve as
efferent ducts for the sexual products, I may point to the fact
observed by me that these organs in N. virens contained
perfectly mature ova, which were found in the efferent duct of
the organ as well as in its body, and that not seldom a single
ovum lay in the external aperture.” Ehlers gives a figure of
an isolated segmental organ which corresponds fairly well with
his description.

But that description is erroneous in one important point.
The segmental organ or nephridium in Nereis virens does
consist of a somewhat spherical mass composed of a number of
glandular ciliated tubes, which probably all form a single
convoluted tube, and a straight, thin, efferent duct, passing off

1 «Die Berstenwirmer,’ 1864—1868.

SOME POINTS IN THE ANATOMY OF POLYOHHTA. 257

from the spherical mass to open to the exterior on the ventral
side of the base of the parapodium. But the internal aperture,
the nephrostome, is not a cleft such as Ehlers figures and
describes in the wall of the spherical mass. The nephrostome
is situated at the end of a simple thin ciliated tube, which
projects out from the spherical body at the side opposite to the
efferent duct. The nephrostome is funnel shaped, as usual,
but the edges of the funnel are produced into numerous finger-
like lobes, and from the lobes project a forest of delicate,
pellucid, branched processes, the surface of which is beset with
extremely long cilia, which move somewhat slowly. The mouth
and neck of the funnel seem to be divided longitudinally by a
partition. But it is probable that the partition is incomplete.
(See figs. 15 and 16.) The somites of Nereis virens are
separated by transverse mesenteries, and at the line along
which these mesenteries are united to the intestine the latter
is considerably dilated, while in the centre of the somite it
is contracted. In order to conform to the typical nephridium
in its relations, the nephrostome in Nereis virens ought to
be on the front side of the septum, behind which the body of
the nephridium is situated. Whether this is so or not I have not
been able definitely to ascertain, but I believe it to beso. The
septum is close to the front edge of the base of the parapodium,
and the nephridium lies slanting across the entrance to the
cavity of the parapodium, and above the edge of the ventral
band of longitudinal muscles. Thus the distance between the
septum and the body of the nephridium is not too great to be
bridged by the tube leading from the nephrostome. It is to
be noted that the nephridium lies as usual below the oblique
muscles.

Ehlers’ observation above quoted, if it were perfectly accu-
rate, would prove beyond a doubt that the nephridia in
Nereis virens serve to convey the sexual products, or at
least the ova, to the exterior. But that observation is in con-
tradiction to all the evidence that has come under my notice.
The breeding habits of the Nereide still remain in some
particulars mysterious, but it seems clear that the worms of

258 J. T. CUNNINGHAM.

this and other errant families, becoming provided with more
efficient swimming organs at the breeding season by the meta-
morphosis of some of their parapodia, leave their burrows and
swim about freely in the water while they are discharging their
reproductive elements. I have only met with two forms of
the family Nereide in sexually mature condition—Nereis
pelagica and Nereis virens. Specimens of the latter are
not unfrequently thrown up on the shore in the Firth of Forth
about the month of April. Solitary female specimens thus
cast up have occasionally been brought to me, but I have not
examined these very minutely. On May Ist this year we found
about 150 specimens among the débris at high-water mark
within about a quarter of a mile, close to the Granton Labora-
tory. Every one of these specimens proved on examination to
be male; they were all alive, though some had been half
desiccated by the warmth of the sun after the tide had left
them, but they were not vigorous, and from their position and
condition must have been dead before the tide returned.
These specimens were in most cases distended with milt, and
when handled they burst or broke into pieces on the least
provocation and discharged the milt copiously. I cut sections
of some of the nephridia of these in situ, and saw not a trace
of spermatozoa within the organs. If the dehiscence is normal
it seems most probable that the animals of both sexes die after
discharging their sexual products, and the dehiscence is so
constant that it cannot be other than a normal process. If
the sexual products normally escape by dehiscence, why should
they pass through the nephridia, or if they passed through the
nephridia why should dehiscence occur at all ?

I have not seen specimens of Nereis pelagia cast up by
the waves upon the shore in an enfeebled condition, but we
found the epitokous form of the species abundantly at the
beginning of February; some we found under stones between
tide marks, but the greater number among the roots of
Laminaria and under stones in the Laminarian zone. We
kept these often for some time in captivity, and whenever they
were handled, or even without being touched, they discharged

SOME POINTS IN THE ANATOMY OF POLYCHATA. 259

ova and spermatozoa by dehiscence, and they invariably died
after being kept some time. Sections revealed no ova or
spermatozoa in the nephridia, in fact it seems impossible that
the ova should be found in the lumen of the nephridium, for
the ovum is many times larger than the tube of the latter in
diameter, being nearly equal in size to the whole body of the
nephridium.

2. Tur ‘‘ Carpiac Bopy.”

Dr. R. Horst, of Leyden, in ‘Zool. Anz.,’ viii, published a
discussion of this curious and problematic structure. He gave
an account of his own examination of the heart in some speci-
mens of the genus Brada, belonging to the family Chlorhe-
mide, and from the structure of the cardiac body in that
genus, and a comparison of it with certain structures in the
Oligocheta, draws the conclusion that the cardiac body in
Polycheta is originally derived in the embryo from the intes-
tinal epithelium, is, in fact, an evagination from the intestine.
Dr. Horst gives a rapid sketch of the history of our knowledge
of the heart in the Chlorhemide. The organ was first men-
tioned by Otto in 1821,' and considered by him to be a second
cesophagus. Claparéde gives an erroneous description of the
organ (‘Ann.Chét. de Naples’); he says it is a tubular structure,
which appears to open anteriorly in the dorsal wall of the buccal
cavity. Delle Chiaje also considered the heart to be con-
nected with the digestive system. Gab. Costa, Dujardin,
Max Miiller, and Quatrefages have all recognised the organ
as a true heart, whose function is to propel the blood into the
branchiz. Claparéde, however, seems to have been the first to
discover the curious dark-coloured cells containing granules,
which occur in the heart, while those zoologists who recognised
the true function of the heart were unaware of anything pecu-
liar in its structure. Claparéde met with the cells of the car-
diac body, and thought the organ was entirely glandular, while

1 «De Sternaspide et Siphostomate,” ‘Nova Acta Acad. Caes. Leopold
Nat. Cur.,’ x, pars 2.

260 J. T. CUNNINGHAM.

others saw that the organ was a heart, and were unaware that
it contained a glandular body.

Dr. Horst found that in Brada, as in Serpulide, Ammo-
charide, &c., there is a blood sinus round the intestine, and
the heart is continuous with this sinus, and therefore a true
dorsal vessel, through which the blood is conducted from the
walls of the intestine to the gills. It is to be remarked that
this is a confirmation of the account given by Quatrefages in
‘Hist. des Annelés, 1865, i, p. 54. Max Miiller, on the
other hand, supposed the posterior end of the heart to be
blind. Dr. Horst then points out the agreement between the
arrangement described by him in Brada and that described by
Vejdovsky in the Enchytreide, which also possess a dorsal
vessel only in the anterior somites, its place being supplied
posteriorly by a blood-sinus in the wall of the intestine; and
further says the researches of Salensky on the development of
Terebella (‘ Arch. de Biologie,’ t. 4) have shown that such an
arrangement is in other Annelids only embryonic.

I have examined the vascular system in Trophonia
plumosa, and although I find Horst’s statements for the
most part correct, there is one feature which he does not men-
tion which might give rise to another explanation. There is,
namely, a thin vessel running in the dorsal median line on the
inner surface of the body wall, unaffected by the convolutions
of the intestine, receiving metamerically arranged transverse
vessels from the walls of the latter, and opening into the
dorsal side of the heart at a point a third of its length from
the hinder end (see fig. 17). It is thus a question which
represents the typical dorsal vessel—this separate vessel that I
have described, or the blood-sinus in the walls of the intestine. I
am inclined to think the former, and that the posterior end of
the heart may be taken as representing a vessel passing from
the intestinal blood-sinus to the dorsal vessel, while the ante-
rior end of the heart is the direct continuation of the dorsal
vessel. These relations are shown in fig. 17.

However this may be, the thin dorsal vessel mentioned above
is not represented in Terebellide, Ampharetide, and Amphic-

SOME POINTS IN THE ANATOMY OF POLYCHATA. 261

tenide. In those three families an anterior heart similar to
that of the Chlorhemide is present, and its posterior end is
connected with a blood-sinus in the walls of the intestine,
which is the only representative of the typical dorsal vessel.
The intestinal blood-sinus is connected, on the ventral side of
the intestine (e.g. in Amphitrite Johnstoni), with a large
definite vessel, which at the level of the posterior end of the
heart divides into two branches ; these pass up one on each side of
the cesophagus, and unite to form the heart. Thus the paradox
is here true that the typical dorsal vessel is in these families
chiefly represented by a ventral vessel. The usual subintes-
tinal or ventral vessel is of course present in addition. Thus
Horst’s remark, that the presence of a free dorsal vessel in the
anterior somites only is merely embryonic in Terebellide, is
far from correct. Salensky, it is true, describes the arrange-
ment as existing in the larva of Terebella; but he does not
assert that any change occurs in later development, and, as a
matter of fact, the arrangement is, as I have said, especially
characteristic of the Terebellidz, Amphictenide, and Ampha-
retide, in the adult condition.

But the most interesting point about this heart is the cel-
lular body it contains. Claparéde, in his ‘Ann. Chét. de
Naples,’ mentions this body in Terebella multisetosa,
Grube, and in Audouiniafiligera. In describing the latter
species he says there are three brown cords in the walls or in
the lumen of the dorsal vessel, and similar structures are
common to all the Cirratulide. Of the body in Terebella
multisetosa he says that the dorsal vessel contains a sub-
stance of a deep black colour distributed in irregular cords.
It is curious that Claparéde should have recognised the nature
of the heart in Terebellidz, and not seen the similarity between
that organ and the heart in Chlorhemidz. In his ‘ Structure
des Ann. Sédentaires,’ 1873, Claparéde figures transverse sec-
tions of the heart of Audouinia filigera and of Terebella
flexuosa, and gives a fuller account in the text of the glan-
dular cords seen in these sections. He says that in his pre-
vious work he had supposed the cardiac organ in Audouinia

262 J. T. CUNNINGHAM.

filigera to be formed of several longitudinal bands, but on
studying sections saw that it was really a cylinder (boyau)
with longitudinal folds. In Terebella flexuosa the brown
substance forms two lobed masses, one applied to the superior
part of the vessel the other to the inferior. These two masses
are not independent, but united at intervals by thick con-
necting cords. In the brown organ of Audouinia the highest
powers only enabled him to distinguish very fine coloured
granules scattered in a fundamental mass. He thinks it pos-
sible that these brown cords are similar to the chloragogenous
substance which surrounds the exterior of the ventral vessel in
the Sabellide, remarking that in species where the cardiac
body is present chloragogenous tissue is wanting, and there
would thus be internal as well as external deposits of
chloragogen.

Horst gives some account of the minute structure of the
cardiac body in Brada. He describes it as made up of irregular
cords, each of which has usually an oval section, and is made
up of cells filled with brown granules. The limits of the cells
were not always clear, and in adult specimens only a network
of fibres could be seen in a section, nuclei being visible at the
nodes of the network, and brown granules scattered through
the ground substance of the meshes. I have examined the
minute structure of the cardiac body in several species, with
results slightly different to those of Horst.

(a) Fam. CHLORHAMID2.

In Trophonia plumosa, when the heart is examined by
means of either transverse or longitudinal sections, the some-
what cylindrical cords of which the cardiac body is composed
are seen not to be composed entirely of cells as described by
Horst, but in most cases to be tubular, each possessing a lumen
(figs. 18, 19). The cells around the lumen form a glandular-
looking epithelium composed of several layers of cells; those
nearest the basement membrane are solid and nucleated, and
contain a large number of small, round, brown grains. The
more internal cells are clear and vacuolated, and the nucleus

—

SOME POINTS IN THE ANATOMY OF POLYCHATA. 263

cannot usually be seen in them; the most internal, those
nearest to the lumen, are almost spherical, and they project
separately and at various levels into the lumen, just as do the
similar cells in a section of a nephridium. Indeed, the whole
structure recalls that of a nephridium very forcibly. Usually
the lumen of the tube contains débris which is stained by car-
mine, and among this can be recognised spherical cells similar
to those which project from the epithelium in process of dis-
integration. It is difficult to resist the conclusion that these
tubes are glands, but I have been unable to discover any trace
of an opening from the cardiac body either to the exterior of
the body or into any other organ. In many of the cross
sections of the cords no lumen can be seen; in some cases this
is obviously due to the fact that the plane of the section is too
near the surface of the cord, and has therefore passed only
through the epithelium ; in other cases the plane of the section
is median longitudinal, or transverse, through the widest part
of a cord, and yet no lumen is seen, the epithelium on one side
being so thick as to come into contact with that of the other.
It is probable that the obliteration of the lumen is partly due
to the contraction produced by reagents. The sections of the
cords in any section of the heart of Trophonia are small and
numerous, and the whole cardiac body almost fills up the entire
cavity of the heart, from its thick posterior portion to its thin
anterior region ; the channels left for the passage of the blood
are in consequence very small. The blood contains small oval
corpuscles, each showing a relatively large, well stained nucleus.
These corpuscles are not numerous.

The cardiac body in Flabelligera affinis (Siphonostoma)
presents a great contrast to that of Trophonia in its size re-
latively to that of the heart; in the former species the organ
constitutes an irregular flat, folded band, running longitudinally
through the cavity of the heart and occupying only a small
portion of that cavity. Between the cardiac body and the wall
of the heart is a wide space occupied by blood. The lower
edge of the band is in the central line of the ventral side of the
heart, whence it rises like a longitudinal partition, its upper

264 J. T. CUNNINGHAM.

branched part coming into contact with the dorsal and lateral
sides of the heart. The organ also differs from that of Tro-
phonia in minute structure. In transverse section the band is
narrow and branched; that is to say, the main band gives off
other longitudinal bands of less extent than itself, so that in
transverse section it appears as an irregularly branched narrow
tract. No distinct lumen is visible in the centre of this tract,
but there is a distinct central line which separates the epithelia
of opposite sides where they come into contact. The clear
vacuolated cells seen in Trophonia are here absent, the epi-
thelium consisting of elongated columnar nucleated cells only ;
the granules are smaller and less numerous (fig. 20).

Fam. TEREBELLIDZ.

In Amphitrite Johnstoni the cardiac body occupies
nearly the whole cavity of the heart, the channels left for the
passage of the blood being very small. It is composed of
cylindrical cords which generally have a longitudinal direction.
In prepared sections no lumen is visible in the cords, each
being completely filled with a mass of cells whose outlines are
somewhat indistinct, but the nuclei are large, spherical, and
deeply stained. The cells are small, so that the nuclei are
closely crowded together. In Amphitrite cirrata and
Terebella Danielsseni the cardiac body exists, but I have
not specially examined it.

In Lanice conchilega the cardiac body is smaller in rela-
tion to the heart than in Amphitrite Johnstoni. The
cords are thinner, and confined to the immediate neighbour-
hood of the walls of the vessel, so that a large central space is
left for the passage of the blood. In the cords a lumen
is frequently but not always visible. The cells have a
similar character to those of Amphitrite Johnstoni
(fig. 14).

In Terebellides Stremi there is but a single cord in the
cardiac body, which runs longitudinally and fills up very nearly
the whole cavity of the heart. In prepared sections a lumen is

i eat tt bh) te eee ee

SOME POINTS IN THE ANATOMY OF POLYCHATA. 265

visible in the centre of this cord, and the cells are so arranged
as to form radii passing from the wall of the cord towards the
centre.

Fam. CIrRRaATULIDA.

In Cirratulus cirratus there are three longitudinal
cylindrical cords occupying nearly the whole cavity of the
dorsal vessel. Two of the cords occasionally anastomose and
then separate again. The cells filling the cords are elongated,
pale, and nucleated; the nuclei stain but the rest of the cells
remains uncoloured in stained preparations. There is no
lumen, and the cells are not arranged in regular radii, but are
placed so that the longer axis passes from the dorsal side of the
cord to the ventral. The cells contain large numbers of the
usual granules, which are brown and spherical, and usually
more numerous near the exterior of the cord than in the
central part. In one specimen I found only two cords present
of which the dorsal was the larger. The cords are quite free
in the interior of the vessel, and have no connection with the
walls of the latter (fig. 21).

The cardiac body is present in the families Chlorhemide, Cir-
ratulide, Amphictenide, Ampharetide, and Terebellide. The
three last are closely allied, but neither of them has any other
points of close agreement with either of the two first mentioned ;
nor are the Chlorhzemide and Cirratulide at all connected with
one another. The cardiac body is present in every species
belonging to the families mentioned, though it varies in details
of structure in different species.

In the heart of Polyophthalmus pictus Edouard Meyer! has
described an organ which Horst claims as the homologue of the
cardiac body we have been considering. It is, of course,
probable enough that Horst is right, but it must be remembered
that Meyer’s description is not complete enough to decide the
question whether the organ in the heart of Polyophthalmus is
similar in structure to the true cardiac body. Meyer says
that a peculiar organ having the form of a thick, short tube,

1 ¢ Arch. f. Mik. Anat.,’ Bd. xxi.

266 J. T. CUNNINGHAM,

which is provided with strong cellular walls, and a canal run-
ning through its axis, occurs in the cavity of the heart. The
organ projects with its posterior half, at the end of which is
the broad entrance opening into the axial canal, into the intes-
tinal sinus, and is here, by means of special small muscular
bundles which arise from processes round the opening, fastened
on to the intestinal epithelium. The front end of the organ
contains the anterior opening of the axial canal and is fastened
to the anterior wall of the heart.

From this account it follows that the blood passes in at one
end and out at the other of the heart organ of Polyophthalmus,
while as yet no opening at all has been demonstrated in the
cardiac body in the families above mentioned.

I have been unable to find any facts which go to support
Horst’s view that the cardiac body is homologous with an organ
which exists in some of the Enchytrzide, and which arises as
an evagination from the intestine. The account given by
Salensky! of the development of the cardiac body in the larva
of Terebelia is not very full, but he states that the cadiac body
is at first a tube passing from the posterior extremity of the
heart and terminating blindly in its interior; and this tube,
from his description and figure, seems to have an opening from
the exterior posterior surface of the heart. It would seem
probable, therefore, that the cardiac body in Terebella is derived
from an invagination of the wall of the heart, and not from
the intestine. It is quite certain that in the adult Trophonia
there is no connection between the cardiac body and the intes-
tinal epithelium, closely as they come into relation. I cut a
continuous and complete series of longitudinal vertical sections
through the posterior part of the heart of Trophonia, and the
intestine it rested upon, and found that there was nowhere any
connection between the intestinal epithelium and that of the
cardiac body.

If Horst’s view were correct one would be tempted to
homologise the cardiac body of Chztopoda with the cellular
structure which grows out from the intestine and projects into

' « Etudes sur le Devel. des Annélides,” ‘ Arch de Biol.,’ iv.

SOME POINTS IN THE ANATOMY OF POLYCH@TA. 267

the heart in Balanoglossus, the structure which Bateson
believes to represent the vertebrate notochord. Professor T.
J. Parker! has already pointed out, in opposition to Bateson’s
arguments, that the vessel which hes on the same side of the
intestine as the notochord in vertebrates conveys the blood
from before backwards, while the dorsal vessel in Balanoglossus
conveys the blood from behind forwards. In this respect the
dorsal vessel of Balanoglossus agrees with the dorsal vessel of
all other Invertebrata, and I am strongly of opinion that
Balanoglossus is constructed on the same plan as a Chetopod.
I would consider the proboscis as the preoral lobe; the nerve-
cord in the collar and in the proboscis as the enlarged repre-
sentative of the supracesophageal ganglia. The circum-
cesophageal commissures are present at the posterior region of
the collar, and they unite into a well-developed ventral nerve-
cord. The great obstacle to this view is the presence of the
dorsal nerve-cord in Balanoglossus. But it may be pointed
out that this dorsal nerve-cord is much thinner and more
Insignificant than the ventral, and that the ventral is in shape
and character the real continuation of the nerve-cord in the
collar. The absence of nephridia and the meaning of the
proboscis pore and proboscis gland are points which cannot at
present be explained on the view I have advocated.

3. Neurat CaAnaALs.

The texts for the few words I have to say on this subject
are two papers, one by Professor McIntosh,’ published in 1877,
the other by Dr. Emil Rohde,’ published in 1886. The first
gives a brief account of the anatomical relations of the ventral
nerve-cords in the families of marine annelids, while the other
discusses the giant fibres of the nerve-cord in Aphroditide.
These structures are asserted by Rohde to be nerve-fibres,

1 On the Blood-vessels of Mustelus antarcticus,” ‘ Phil. Trans.,’
vol. elxxvii (Pt. IT, 1886), p. 719.

2 “ Arrangement, &., of Great Nerve-Cords in Marine Annelids,” ‘Proce.
Roy. Soc.,’ Ed., 1877.

3 «§. B. d. Konigl. Preuss. Akad. d, Wiss.,’ July 29th, 1886.

268 J. T. CUNNINGHAM.

which commence as processes of certain colossal ganglion cells
occurring in definite positions in the brain or ventral cord.
He gives the following account of the giant-fibres in the genus
Sthenelais. There are three kinds of colossal nerve-fibres:
(1) some traversing the whole nervous system from the
anterior to the posterior extremity ; (2) some running from the
posterior to the anterior extremity; (8) some starting from
the nervous system on each side, and running to the periphery.
He further says that if the nervous system be traced through a
series of transverse sections he finds in the posterior part of
the brain a colossal ganglion cell on each side which sends off
a large process. This process passes first forward for some
distance into the brain, then through the cesophageal commis-
sure into the ventral cord. These two nerve-fibres unite into
one which runs ventrally on one side of the ventral cord to the
posterior extremity of the body. This colossal nerve-fibre is
enveloped by a fibrous sheath, which is at first closely applied
to it, but in its further course separates from it, and then
encloses a cavity which becomes larger posteriorly, and in the
middle of the body attains an enormous diameter.
Unfortunately, no figures illustrating these descriptions have
yet appeared, and I have therefore had to confine myself to a
comparison of my own sections with the above description. I
have been totally unable to see the connections which Rohde
declares to exist. I have prepared series of sections from
different parts of the body of Sigalion boa, Johnston, which,
according to McIntosh (Invert. Fauna of St. Andrews), belongs
to Kinberg’s genus Sthenelais. In the middle region a pair
of colossal fibres, or as I shall usually call them, neural canals,
appearing under a low power like tubes, are conspicuous.
One of these is situated on the inner side of each cord,
towards the dorsal region, and at the periphery of the cord
(fig. 22). The neural canal is internal to the layer of ganglion
cells, and is partially occupied by a shrunken homogeneous
substance. Processes can be often seen passing off from the
ganglion cells transversely, and entering the substance of the
cord where they are seen to branch into fine fibrils, exactly in

SOME POINTS IN THE ANATOMY OF POLYCHATA. 269

the same way as that illustrated so well by F. Nansen! in his
memoir on the Myzostomide. But I have been unable to
trace any connection between the neural canal, alias giant-
fibre, and a ganglion cell. Indeed, in Sigalion the canal or
tube becomes very small long before the brain is reached, and
I cannot even distinguish it in the csophageal commissures,
or in the cord immediately behind them.

In the central part of each cord in the middle region of the
body are one or two tubes, which are similar in structure to
the large neural canals, but much smaller.

A few words as to the character of the nerve-cords in Siga-
lion boa. The cords-are nowhere separated from the epi-
dermis. Ganglion cells are abundant beneath the ventral
cords, both in the ganglia and between the successive ganglia.
Above the cords is a striking development of a very peculiar
tissue whose function is problematic. In the middle of the
body this tissue consists of waved fibres or lamine, which
are often arranged in parallel curves. These form a network
whose meshes occasionally contain cells with nuclei, but
usually are filled with a stained granular substance. Close
behind the head this mass of tissue is of very great size,
and is much more cellular. In the cesophageal cords it is
reduced to a very small quantity, but it forms a thick envelope
round the brain. The tissue stains with difficulty. It is in all
probability a kind of connective tissue not directly concerned
in nervous functions.

With regard to Aphrodite, I entirely agree with Rohde that
colossal fibres or neural canals are altogether absent; and in
this genus the nerve-cords are quite separated from the epi-
dermis.

Polynode I have not examined, but in Harmothoe im-
bricata I find a pair of neural canals corresponding in posi-
tion to those of Sigalion boa, but I have not seen any in a
ventral position, such as those mentioned by Rohde in Polynée.
In Harmothée the ventral nervous mass is distinctly defined,
but not separated, from the epidermis.

1 « Bidrag til Myszostomernes Anatomi und ey Bergen, 1885,

VOL. XXVIII, PART 2.—NEW SER. T

270 J. T. CUNNINGHAM.

Of the Nereidee I have examined Nereis virens, Sars, and
here cannot confirm the account of the neural canals given by
McIntosh. In many sections three or four neural canals are
seen, which are not quite symmetrical; these are sections
through inter-ganglionic transverse commissures. In the
cords between successive ganglia there are seen to be a single
pair of canals, one of which is often divided into two. The
pair occupy an exactly similar position to that of the neural
canals in Sigalion. McIntosh states that in Nereis virens
there are several neural canals, viz. two large infero-lateral, a
single superior median and a smaller, a little below the latter
on each side. This apparently means five in all, two pairs and
one median. Probably he examined sections of the ganglia in
which several canals are often seen. But these have not a
constant relative position, and are, I believe, due to the sub-
division of the two canals which are seen in the separated
cords.

In Nereis pelagical find canals placed in the positions
ascribed by McIntosh to those of N. virens. There is one
dorsal median in the fibrous partition between the cords, a
pair corresponding to the typical pair of Sigalion, and another
pair consisting of one on the external border of each cord.

In Nephthys, instead of the typical canal on the inner side
of each cord, there are two large canals, one above the other,
with a smaller one between them. There is also a smaller
canal in the external side of each cord, and still smaller ones
in the substance of the cords. The nerve area here is not
separated from the epidermis.

In Phyllodoce no well-marked neural canals can be distin-
guished. The cords are widely separated from the epidermis.

We pass now to the examination of the families of
Sedentaria.

In the Sabellide there is a pair of canals or tubes of much
greater size than any seen in the Aphroditide. Only a few
somites behind the head these tubes reach a thickness equal to
or even slightly greater than that of the nerve-cords themselves,
and they retain an almost uniform thickness in their course to

SOME POINTS IN THE ANATOMY OF POLYOHATA. 271

the end of the body. These tubular fibres have been well
described by Claparéde! as they occur in SpirographisSpal-
lanzanii. He found several transverse connecting branches be-
tween the two tubes in the thoracic region immediately behind
the union of the esophageal commissures, and he traced the tubes
into these commissures, where each divided into two branches.
These branches passed forwards into the cerebral ganglion,
where they ramified into thinner and thinner branches, but the
ultimate terminations Claparéde could not discover.

I have endeavoured to trace the anterior extremities of the
tubes in Sabella penicillus. I found a transverse connect-
ing tube in the first transverse commissure, like those described
by Claparéde, but could not find more than one, and it is
noticeable that while Claparéede mentions a number of these
anastomoses in the text he figures only one. I found that the
tube, much diminished in diameter, passed up the cesophageal
commissure, but could not discover that it branched. In my
sections it seems to become smaller and smaller, and simply
end blindly.

In Sabella the mass of ganglion cells representing the cere-
bral ganglion is situated on each side of the cesophagus, and
below it is the fibrous cesophageal commissure, which is con-
tinued by an arch above the cesophagus into its fellow of the
opposite side. The tubular fibre ends below this mass of
cerebral ganglion cells, and, as far as I can see, sends no
branches towards the mass, nor could I see any trace of a
connection between the end of the tube and any ganglion
cells.

Although these structures are spoken of as tubes, they are
not actually empty. Their interior in the sections is par-
tially filled by a transparent gelatinous-looking mass, which
for the most part does not stain; but there are lines in the
mass which are stained, and which somewhat resemble a net-
work of fibres. In my opinion these stained lines are due to
the unequal coagulation of the gelatinous mass, which during

1 “ Structure des Annélides Sédentaires,” ‘ Mem. Soc. de Phys. de Genéve,’
tom. XXil.

272 J. T. CUNNINGHAM.

life is homogeneous and semi-liquid. On the dorsal side of the
tube is a space between the gelatinous medulla and the wall of
the tube, and the edge of the medulla below this space is deeply
stained. All this is, in my opinion, the consequence of con-
traction and coagulation.

Myxicola in the greater part of its body has a single tubular
fibre similar in structure to one of the pair which exist in
Sabella. According to Claparéde (loc. cit.) the tubular fibre
of one side, at a point a little behind the cesophageal commis-
sures, opens into that of the other side, and the latter proceeds
as the unique fibre, while the nerve-cord corresponding to the
first tube joins the other cord without fusing with it and
finally terminates. My series of sections of this animal is not
quite perfect, but, as far as I can judge, Claparéde has been
somewhat deceived. It is true that only one tubular fibre
persists, but it seems to me that the other terminates and does
not open into its fellow. Some distance behind the cesopha-
geal ring the right hand cord is seen thicker than the other
and destitute of a tubular fibre; the tubular fibre is seen on
the left-hand side of the ventral median mesentery, which is
pushed considerably to the right. Farther back the tubular
fibre becomes central, and the two nerve-cords, one on either
side of it, are equal in thickness. I have seen no indication of
the disappearance of one nerve-cord. In my opinion, all that
has happened in Myxicola is that one of the large tubes has
disappeared except in the extreme anterior region, and the
other has increased in size, and in the greater part of the body
become central. One very interesting point to be seen in
Myxicola is that the two tubes are continuous with one
another in the lower part of the cerebral commissure;
the tubes in each cesophageal commissure, which are of con-
siderable diameter, can be seen to become continuous with one
another in the section of the cerebral commissure. In my
sections of Myxicola the tube in the posterior part of the body
is entirely empty, except in the ventral part, where a thick
stained band occupies the cavity; this is due to a greater
shrinking of the contents of the tube in the process of prepa-

SOME POINTS IN THE ANATOMY OF POLYCHATA. 278

ration. It may be pointed out here that the nerve-cords in
Sabella are not separated from the epidermis, while in Myxi-
cola they are completely so.

We pass now to consider another family in which colossal
tubes or neural canals are greatly developed, the Spionide.
In Nerine coniocephala, Johuston (fig. 23), the nerve-
cords are differentiations in a thickened epidermis, less dis-
tinctly defined from the surrounding cells than is the case in
Sabella. In the median line between the two cords is an
enormous neural canal, larger in sectional area than the two
nerve-cords together, but having a structure similar to the
tubes in Sabella. The appearance of this canal in section is
seen in fig. 23. The canal contains a shrunken, gelatinous-
looking mass as, in other cases. On tracing this canal for-
wards it is found to become smaller near the anterior end, and
to cease altogether much sooner than is usually the case. I
have found no indication of its anterior division into two
canals. The last trace of it seen in approaching the head is
shown in fig. 24.

In Scolecolepis vulgaris, Malmgren, there are two
neural canals, one on the inner side of each nerve-cord.

In Magelona, which forms a family by itself, there is a very
large median neural canal resembling that of Nerine, but lying
below the nerve-cords instead of above them (fig. 25).

In the Ariciidz I need only confirm the account given by
McIntosh, that in the middle of the body the nerve-cords are
thrust inwards by the great ventral longitudinal muscles,
which contain between them a narrow lamina preserving the
connection between the nerve-cords and the epidermis. A
single median neural canal runs above the nerve-cords as in
Nerine.

In Arenicola (Telethuside) the nerve-cords are separated
from the epidermis by the layer of circular muscles; there is a
small neural canal, entirely filled with a homogeneous mass, at
the dorsal and inner side of each cord.

In Trophonia (Chlorhemide) McIntosh does not mention
the existence of neural canals, but one of these exists in each

274 J. T. CUNNINGHAM.

cord on its inner side dorsally. The cords are entirely free
from the epidermis. The neural canals are similar in size and
appearance to those in Sigalion boa.

Among the Terebellide I find a median neural canal in
Lanice conchilega, Malmgren: it is of considerable size,
but has not such well-defined fibrous walls as are usually
present. In Amphitrite Johnstoni I have been unable to
detect any canal, nor in Terebellides Stremi.

In the Ampharetide, however, which are, so to speak, on the
way towards the Terebellide, the neural canals are large and
conspicuous, and have the typical structure. In Melinna
cristata there is one on the inner side of each nerve-cord in
the thoracic region.

In the Capitellide (Capitella and Notomastus) the nerve-cords
lie in the epidermis posteriorly, while in a few of the anterior
somites they are entirely separated from it by both the circular
and longitudinal muscles. In this anterior region there is a
large median neural canal on the dorsal side of the double cord
in Notomastus. In Capitella the canal is absent.

Among the Maldanide I have examined Nicomache lum-
bricalis and Axiothea catenata. In the former the nerve-
cords are not separated from the epidermis, and there is a
considerable single median neural canal above the cords in both
species.

Among the Hermellide, in Sabellaria, as pointed out by
McIntosh, the two cords are at a considerable distance from one
another; they are completely separated from the epidermis and
lie on the upper and inner side of the ventral longitudinal
muscles. Each has a large neural canal, similar to that of
Sabella, on its inner side (fig. 26).

In Serpula the neural canals are similar in structure and
position to those of Sabella.

McIntosh mentions a small and indistinct neural canal in
Ammotrypane aulogaster, H. R. In some of my sections
it can be made out, but always with difficulty, as it is exceed-
ingly ill-defined. In Cirratulus cirratus, also, neural
canals are absent.

SOME POINTS IN THE ANATOMY OF POLYCHATA. 275

It would be very startling, if not even absurd, to maintain
that such neural canals as are seen in Sabella, and in Nerine
are colossal nerve-fibres, and that their contents form a nervous
medulla which commences as a process from a ganglion cell. I
have entirely failed to trace any connection between these
canals and any ganglion cells. At the same time it is difficult
to refuse to admit that the neural canals of the Sedentaria are
completely homologous with those of the Errantia.

In my opinion, in both cases they are supporting structures
which serve to prevent the nerve-cords being bent at a sharp
angle, causing them always to remain in curves, and so to
escape injury during the wriggling and burrowing of the worm.
It is noticeable as a support of this view that the canals always
reach their greatest development in worms which are extremely
long in proportion to their thickness. Sabella and Nerine are
both extremely long, as compared, for instance, to Ophelia or
Cirratulus. Another fact, which seems to have some signifi-
cance, is that where the neural canals show their maximum
development the nerve-cord is not separated from the epi-
dermis, and is therefore more exposed to the danger of being
injured than where they have reached a more internal position.
The origin of the vertebrate notochord from the hypoblast
seems so well established that a comparison of it with the
neural canals of the Chetopoda will scarcely be regarded
seriously by morphologists. At the same time, seeing that the
notochord at a later stage is separated from the intestine by
the aorta, it is very difficult to understand how the former
structure could have been derived phylogenetically from the
intestine. The neural canals are remarkably constant through-
out the Cheetopoda, those in the Polychzta being obviously
homologous with the three giant-fibres in the earthworm and
other Oligocheta. They have a position in relation to the
nerve-cords and ventral blood-vessel which is similar to that of
the notochord in relation to the neurochord and aorta. Their
origin in the embryo has not so far as I know been investi-
gated, so that it is doubtful whether they are intercellular or
intracellular in origin. I hope to devote further attention to

276 J. I. CUNNINGHAM.

the subject, at present I can only say that the evidence
adduced in favour of their specifically nervous nature is quite
inadequate, and that the possibility of a phylogenetic connec-
tion between these neural canals in the Chetopoda and the
notochord of the Chordata, cannot at present be altogether
dismissed.

In concluding this paper I have to explain that my attention
was attracted to the points discussed in the course of a
systematic examination of the Polychzeta, which I carried on
at the Granton Marine Station, in collaboration with my friend

Mr. G, A. Ramage, Vans Dunlop Scholar in Edinburgh |

University.

For the facts and views I have set forth I am alone respon-
sible, but the collection and identification of specimens were
chiefly carried on by Mr. Ramage, and he rendered much
valuable assistance in preparation and dissection. The draw-
ings for the paper were all executed by myself.

EXPLANATION OF PLATES XVII, XVIII, and XIX.

Illustrating J. T. Cunningham’s paper “On Some Points in
the Anatomy of Polycheta.”

Fic. 1—An entire nephridium of Arenicola marina, seen under a low
power, in the fresh state. The membranous funnel has been turned back, so
that the ventral side is seen. 4. Anterior, P. Posterior end. 4/. Blood-
vessel, which joins the branchial vein. d. 4. Dorsal fringed border of nephro-
stome. ze. Nephrostome. go. Gonad.

Fic. 2.—Optical section of a portion of the wall of the nephridium of
Arenicola, after treatment with osmic acid. 47. Blood-vessel. EH, Zeiss, oc. 3.

Fic. 3.—Horizontal section of 2nd and 3rd and part of lst somites of
Cirratulus cirratus. me. Nephrostome of anterior nephridium, opening
from buccal somite. a@.2. Ascending part of the nephridium. d.z. Descend-
ing part. se, Transverse and longitudinal septa. A, Zeiss, oc. 2.

_,

SOME POINTS IN THE ANATOMY OF POLYCHATA. 277

Fic. 4.—Horizontal section through three somites, from middle part of
body of the same species. ze., se., as before. An ovum is seen in one of the
nephrostomata. A, Zeiss, oc. 2.

Fic. 5.—Transverse section of same species, middle part of body, passing
through external openings of a pair of nephridia. A, Zeiss, oc. 2.

Fie. 6.—Two somites of Cirratulus cirratus, seen living by reflected
light while the eggs were escaping. In one somite an ovum is shown just
emerging from the nephridial aperture.

Fic. 7.—Portions of two consecutive somites, from a vertical longitudinal
section of Nerine cirratulus. ze. Nephrostome. ov. Ovary. xo. Notopo-
dial chetze. mu. Neuropodial chete.

Fic. 8.— Portion of a section of the same series as the preceding, nearer
the surface of the body. o., zw., as before. mp. External opening of the
nephridium.

Fic. 9.—Longitudinal vertical section of Nerine coniocephala, showing
relations of the nephridium. Reference letters as before.

Fic. 10.—Fresh preparation made by cutting open the anterior part of a
specimen of Lanice conchilega along the dorsal median line, taking away
the intestine and cutting through the oblique muscles on each side. x. Ne-
phridia, with looped tubes. 2’. Reduced nepbridia. x. 7. Rudimentary
nephridium of 5th somite. zo. Notopodial chete, their inner ends.

Fie. 11.—Horizontal section of anterior nine somites of Lanice conchi-
lega, at a level near upper end of the uncinigerous tori. 2. Continuous
tube formed by nephridia of Somites vi to 1x. 2x. a. Ascending part of
nephridial tube. 2. p. Nephridial aperture. v. g/. Tissue of ventral glandu-
lar epidermis.

Fig. 12.—Similar horizontal section of Somites x1 to xiv. z'. Coalesced
reduced nephridia. zz. Neuropodium, z. e. uncinigerous torus.

Fic. 13.—Horizontal section of somites 11 to v of Lanice conchilega,
from the same series as that shown in Fig. 11]. — ze., ze., ne. Nephrostomata
of Somites 11, 111, and Iv, with gonads attached to the posterior two. x.
Section of nephridial tube in Somite 111.

Fic. 14.—Transverse section of Lanice conchilega in Somite x. 2’
Cavity of reduced nephridium. d. 6. Dorsal blood-vessel. v. gi. Ventral
glandular epidermis.

Fic. 15.—Nephridium of Nereis virens, as seen in fresh state, when
dissected out from the animal. ze. Nephrostome with its fringe of branched
ciliated processes. . Globular mass, formed by the convoluted nephridial
tube. x. p. Cylindrical portion leading to external aperture. A, Zeiss, oc. 2.

Fre. 16.—Nephrostome of Nereis virens, more highly magnified; fresh
condition. Zeiss, CC, oc. 3.

278 J. T. CUNNINGHAM.

Fic. 17.—Pseudhemal system of Trophonia plumosa, as seen on dis-
section. d.v. Dorsal vessel. v. v. Ventral vessel. At. Heart. int. Intes-
tine. o.p. Subcesophageal pouch. ov. Ovary.

Fie. 18.—Vertical longitudinal section through the heart and adjacent part
of the intestine of Trophonia plumosa. 4/. Blood or pseudhzmal fluid.
c. b. Tubes of the cardiac body; some with an open lumen, others full of
transparent cells. ep. Epithelium of the intestine. A, Zeiss, oc. 2.

Fic. 19.—A single tube of the cardiac body shown in previous figure,
More highly magnified. Shows the nucleated cells lining the wall and a large
central lumen. KE, Zeiss, oc. 2.

Fic. 20.—Transverse section of heart of Flabelligera affinis, showing
cardiac body in the interior. A, Zeiss, oc. 2.

Fic. 21.—Transverse section of dorsal vessel and contained cardiac body in
Cirratulus cirratus. c.6., b/., as before. H, Zeiss, oc. 2.

Fre. 22.—Transverse section of ventral nerve-cords of Siglion boa, from
middle of body, interganglionic region. x. co. Nerve cord. x.c. Neural-canal.
s. 2. Supraneural connective tissue. ed. Epidermis. ct. Cuticle.

Fig. 23.—Transverse section of nerve-cords of Nerine coniocephala.
”. CO., ”. C., as before.

Fic. 24.—Transverse section from same series as preceding figure, near the
anterior end of the animal. Letters as before. CC, Zeiss, oc. 3.

Fic. 25.—Transverse section of Magelona papillicornis. 2.¢., as
before.

Fic. 26.—Transverse section of Sabellaria spinulosa, middle of the
body. . co. Nerve cord, with its neural canal.

EE Sap eN Tihs y ag
yee sti Si

Se AN
Annan

sess |

~ nt i ; :
~ Thats \ ro :
Sy ie ip pi%
NS ; 7 : :
iS ta ge

te
oy

28,

Th

ote

Mucor

7 eo ee
Sournl Ue BUN SPE, YUL |

F Huth, Leh? Edin®

5
wo
cs)
&
4a
re

2
os
fu

a

Méor Journ Vel UMS, HE

i. F Huth, Lith? Edin®

s
:
iv -
.
: +
‘
‘
‘ Py
.
.
\
=f
*
+
'
A
’ ~
i
- =
-
e
.
x
-
7
’
s *
‘
oe
§
7 é 7
- 7

bo
~J
Fa)

ON TEMNOCEPHALA.

On Temnocephala, an Aberrant Monogenetic
Trematode.

By

William A, Haswell, M.A., B.Se.,
Lecturer on Zoology and Comparative Anatomy, Sydney University.

With Plates XX, XXI, and XII.

Historica.

Axsout the year 1849 Gay discovered, in the environs of
Santiago, on the surface of certain crayfishes, a leech-like
animal, which, in a letter to Blainville, he described briefly
under the name of Branchiobdella chilensis.! The genus
Branchiobdella was first instituted by Odier, and, though
apparently the name was applied by him to a species of the
genus Branchellion of Savigny, it has been very generally
adopted since for an external parasite of the fresh-water crayfish
of Kurope—the Branchiobdella astaci of Rudolphi.

Branchiobdella astaci, as is well known, isa well-marked
Leech ; it has an elongated body composed of about eighteen
distinct rings, with an anterior and a posterior sucker, an anal
aperture above the posterior sucker, and a median ventral
nerve-cord. In Gay’s ‘Zoology of Chilé,’?? Blanchard described
Gay’s species under the name of Temnocephala chilensis,
recognising that the differences between it and Branchi-
obdella astaci are too great to admit of both species being
placed in one genus. ‘‘ Las Temnocefalas se distinguen aun

‘ Tam not aware that the letter has been published, but it is quoted by

Moquin-Tandon in the ‘ Monographie des Hirudinés,’ p. 300.
ahaap. OL.

280 WILLIAM A. HASWELL.

del género Branchiobdella por la presencia de los ojos y de las
divisiones cefalicas de que no existe traza alguna en él.”

In 1870, Philippi! published some observations on Temno-
cephala, based on an examination of living specimens which
he found on the surface of Chilian fresh-water crayfishes of
the genus Aiglea. He described the external form, the
colouration, and the movements, and notices certain of the
internal organs which he is able to see through the wall of
the body, though without being able to give any precise
account of their nature. He concludes that Temnocephala
ought to be placed among the worms in the neighbourhood
of Malacobdella.

During his scientific explorations in the Phillipine Islands,
Carl Semper found, on the surface of fresh-water crabs, speci-
mens of an external parasite, which proved to be the Temno-
cephala of Blanchard; but a more detailed examination
showed him that the affinities of the animal were much more
with the ectoparasitic Trematodes than with Malacobdella
or the Hirudinea.?

Wood-Mason found, in 1875,a number of Temnocephalez
in a bottle containing some New Zealand crayfishes (Parane-
phropssetosus), to which they had evidently been attached,
and some also in bottles containing specimens from the north-
east frontier of India.

At the beginning of last year, when on a visit to Tasmania,
my attention was directed by Mr. Alexander Morton, the
curator of the Tasmanian Museum at Hobart, to certain re-
markable animals observable on the surface of a specimen of
the large fresh-water crayfish of the northern waters of Tas-
mania, which he had alive in a tank. These proved to be

1 ¢ Archiv fiir Naturgeschichte,’ 1870.

2 «Zeitschrift f. wiss. Zoologie,’ xxii Band (1872). I have not been able
to see this paper, which is a very short one (three pages), and am indebted for
my knowledge of it to Leuckart’s “Bericht” in the ‘Archiv f. Natur-
geschichte,’ x1 Band (1874).

8 “Qn the Geographical Distribution of the Temnocephala chilensis
of Blanchard,” ‘Ann. Mag. Nat. Hist.’ (4), vol. xv, p. 3386 (1875).

ON TEMNOCEPHALA. 281

Temnocephalz; and I have since found that other species
of this remarkable genus infest the fresh-water crayfishes of
the rivers of New South Wales.

I must here acknowledge my indebtedness to various friends,
through whose kindness I have been able to procure ample
supplies of specimens, more especially to Mr. Alexander Morton,
of Hobart; Mr. E. C. Merewether,and Mr. Harry Merewether,
of Bondi and Mount Wilson; Mr. Alexander Hamilton, of
Mudgee; Mr. J. D. Cox, of Mount Wilson; Mr. Charles
Chilton, of Dunedin, and Mr. J. J. Fletcher.

GENERAL DescriPTION oF TEMNOCEPHALA.

Temnocephala (Pl. XX, figs. 1—5) is a leech-like animal, the
largest about half an inch in length. In outline the body is
ovate or pyriform, but much compressed from above down-
wards; the anterior end narrower than the posterior, and the
lateral border fringed with a narrow, delicate fold. At the
narrower anterior end there is, in the middle, in the case of
the Tasmanian species, a rounded, dorso-ventrally compressed
lobe, about a fifth of the length of the rest of the body, and on
either side of this two very long and slender tentacles, which
are filiform, and a half to two thirds of the total length of the
body when fully extended, but are capable of being greatly
retracted. In the case of the New South Wales species, and
that from New Zealand, there are five equal slender tentacles.
Close to the broader posterior end on the ventral aspect is a
very large sucker of circular outline supported on a short stalk.
Near the middle line of the dorsal surface, placed near together
a little behind the bases of the tentacles, is a pair of small
black eyes. On the ventral surface, not far from the anterior
end, is the mouth—a well-marked aperture. Some distance
behind it, a little way in front of the sucker, is the common
genital aperture—a short transverse slit leading into the genital
cloaca.

By incident light against a dark ground, the body of the
larger New South Wales species (fig. 2) and of the Tasmanian

282 WILLIAM A. HASWELL.

species (fig. 3) has a dark grey ground colour with rich brown
mottling. In the New South Wales species there are several,
usually three, broad transverse dark bands, separated from one
another by lighter intervals. In the middle of the most anterior
of these, towards its front edge, about a fifth of the total length
behind the base of the tentacles, is the dark spot on which the
eyes are situated, and in front of this is a lighter interval,
succeeded in front again by a dark space about the bases of the
tentacles; the latter are of a nearly uniform brown, rather
lighter towards the tips. Slightly behind the eyes, and rather
nearer the lateral margin than the middle line, there will be
seen on either side a minute white spot, which marks the posi-
tion of the opening of the excretory system.

Little is to be seen of the internal organisation on a surface
view, the pigment of the integument being very dense. <A few
light reticulating lines may be observed, which indicate the
position of the elements of the dorsal nerve-plexus, the ali-
mentary canal may be discerned as a lighter patch of squarish
outline towards the middle of the body, and a few, irregularly-
placed, fine transverse lines may be observed, which marks the
position of rudimentary intersegmental septa.

On the ventral side the colour is light grey; the sucker is
colourless. The squarish intestine, of a darker colour, is seen
through the body wall, and transverse divisions of its substance
indicate the intestinal pouches or ceca.

Here and there among the full-grown Temnocephale, in
the case of the larger New South Wales species, will be found
smaller immature specimens, distinguished by their whiteness ;
and when these are examined under the compressorium a good
deal is to be learnt regarding the internal organisation. They
are almost entirely devoid of pigment, except a little in the
neighbourhood of the eyes and along the back, and, strange to
say, they have invariably six tentacles, whereas the adult has
only five.

When undisturbed the Temnocephalz adhere to the sur-
face of the crayfish by means of the sucker, with the body
extended and inclined at an angle of about 45° to the surface

ON TEMNOOEPHALA. 283

of attachment, the long slender tentacles stretched to their
utmost and waving about the water in search of food. Their
food consists of small crustacea (small Amphipoda in the case
of the New Zealand species, small Aselli and Entomostracans
in the case of the New South Wales species) and insect larve.
These they capture by means of the tentacles, and it is
readily to be understood that living on the crayfish must be of
advantage to the animal, the movements of the crayfish in
searching for food amongst the stones and dead leaves and
sticks on the bottom of the stream doubtless starting many of
these Arthropods, and enabling the Trematodes to secure them.

The Temnocephale move from place to place with a sort
of “ looping ” action very like the movements of a leech. The
body is applied to the surface of the crayfish and then stretched
out ; the tentacles become shortened and flattened out on the
surface of the crayfish and play the part of anterior suckers ;
the large sucker is relaxed and the body drawn forwards by
means of the tentacles, until the sucker is fixed again close up
to the latter, the body being now bent double; the tentacles
then let go their hold, the body is stretched out again and so
on. The body can be rotated from side to side through a very
large arc by the action of the sucker, which is capable of a
considerable amount of rotation around its short. stalk.
The rapidity of all these movements and the extreme sensitive-
ness of the animals are surprising. A slight touch from an
instrument wiil cause an instantaneous turning aside of the
body, drawing in of the tentacles, and frequently rapid flight
in a different direction from that in which the attack was
made. In turning aside from a touch the little animals show
a very definite sense of direction, and if a succession of taps
be made on the bottom of a vessel in which they are living,
alternately on opposite sides of them, they will turn from one
side to the other according to the direction of the tap. If
they are detached from the surface of the crayfish—which it is
very difficult to do owing to the slippery character of their
integument and the firmness with which the sucker adheres—
they contract themselves for a moment into a ball, then stretch-

284 WILLIAM A. HASWELL.

ing themselves out to their full length and extending their
tentacles, they are often able by strong flexions of the body
while falling through the water to seize on some other part of
the surface and regain their position on the crayfish. The
movements thus made in the attempt to regain a hold are not
unlike the movements made by the leech when swimming, but
the Temnocephalz are incapable of directed movement or
even of sustaining themselves in the water by this means.
When the body is pinched or cut, the tentacles are very often
turned backwards and clasp the instrument used as if to ascer-
tain its nature or to repel the attack, and the tentacles are to
be regarded as special organs of touch as well as instruments
of prehension and aids to locomotion.

Four very distinct species of Temnocephala frequent the
Australian and New Zealand crayfishes, and all of them, so
far as I am able to judge from Philippi’s figures, differ from
T. chilensis. That found on the surface of the larger
Tasmanian crayfish is distinguished by the possession of a
short compressed lobe instead of the median tentacle. All
the others have, like T. chilensis, five equal or nearly equal
tentacles; but are to be distinguished from one another by
certain differences in the structure of the reproductive organs
to be noticed later on. I propose the following names for the
four species :

1. TeMNocEPHALA FascIaTa, on Astacopsis serratus,
streams of New South Wales.

2. T. quapRicornis, on Astacopsis Franklinii, northern
rivers of Tasmania.

3. T. minor, on Astacopsis bicarinatus, streams of
New South Wales.

4, T. Nov#-ZELANDIZ, on Paranephrops setosus, rivers
of New Zealand.

InTEGUMENT, MuscLes, AND PARENCHYMA.

The body wall is composed of the following layers—cuticle,
epidermis, basement membrane, circularly arranged layers

ON TEMNOCEPHALA. 285

of muscular fibres, longitudinal layer of muscle, nervous
layer.

The cuticle (Plate XXI, fig. 1, c.) attains a thickness of
‘006 mm. in T. fasciata. It is beset with numerous vertical
pore-canals, the presence of which gives it the appearance in
sections of being made up of close-set papilla. Ona surface
view it appears slightly corrugated, the ruge being exceedingly
minute, irregular, and mostly transverse.

The epidermis (e) is of nearly equal thickness with the
cuticle. It is composed of a thin layer of protoplasm, with
regularly distributed nuclei, but without a trace of cell-
boundaries. The nuclei are spherical, of a diameter nearly
equal to the whole thickness of the epidermal layer, and with
a finely granular interior. In vertical section the substance
of the epidermal layer appears to be divided into a series of
vertical columns by a number of parallel lines. This appear-
ance is produced by the presence of very numerous pore-
canals, which run perpendicularly through both epidermis and
cuticle to open on the surface of the latter; and a surface view
of the epidermis shows the numerous and closely-placed
rounded openings. In T. minor the cuticle is smooth, and
the nuclei in the epidermis much less numerous than in T.
fasciata.

The basement membrane (8.) on which the epidermis
rests is a perfectly homogeneous membrane of, comparatively,
considerable thickness, being as thick as the epidermis itself.
It stains deeply with carmine dyes—much more deeply than
the epidermis, but is not readily stained with hematoxylin.
It is devoid of nuclei or other evidence of structure; like the
superficial layers it is perforated by the pore-canals, but it is
difficult to see them unless in parts where the secretion of the
subcutaneous glands, which is readily acted on by staining
agents, is passing out. Like the cuticle and epidermis the
basement membrane varies little in thickness in different parts
of the body. In T. minor cuticle, epidermis, and basement
membrane are very delicate, and together are only ‘008 mm.
in thickness.

VOL. XXVIII, PART 2.—NEW SER. U

286 WILLIAM A. HASWELL.

The external muscular layer (c. m.) is a thin stratum of
circularly arranged fibres, not more than one or two fibres in
thickness. It is situated in a stratum of finely fibrous matter
containing a layer of pigment cells, which in T. fasciatus
on the dorsal side form with their anastomosing processes a
dense and regular network ; this pigment-bearing stratum is
much more strongly developed on the doral side than on the
ventral.

The internal layer of muscle (J. m.), the fibres of which run
longitudinally, is much more strongly developed than the
external layer, and is composed of thicker fibres; it is most
powerfully developed on the ventral side, but extends over the
whole surface and is specially well developed in the tentacles
on the ventral side. The fibres on the dorsal side are arranged
in bundles, separated from one another by very narrow inter-
spaces occupied by interstitial fibrous matter containing pig-
ment; on the ventral side they are arranged in thin fasciculi.

The muscular fibres of these two layers are usually angular,
sometimes oval in cross section, "004 mm. in diameter in the
internal layer, rather less in the external. They are finely
striated longitudinally, non-nucleated, and are separated from
one another by finely fibrillar interstitial matter. In cross
section they are seen to consist of two substances, one forming
a narrow darker central core, the other clearer and constituting
the principal bulk of the fibre.

The nervous layer is a thin layer of the parenchyma, dis-
tinguished by the presence in it of a considerable amount of
pigment, and by being the seat of the superficial nerve-plexuses.
It passes insensibly into the general parenchyma of the centre
of the body.

In the structure of the body wall Temnocephala resembles
the marine ectoparasitic Trematodes such as Tristomum
and Onchocotyle, and differs essentially from the Dis-
tomide in the presence of the pore-canals and the absence of
distinct cell-boundaries in the epidermis ; the inner cell-layer’
described by Sommer! in the liver-fluke is not represented,

1 “Die Anatomie des Leberegels,” ‘ Zeitschr. f. wiss. Zool.,’? xxxiv Band.

ON TEMNOCEPHALA. 287

but its existence in the latter species has been denied by
Mace.! The layer of oblique muscular fibres found in the
body wall of some Polystomidz, though absent in others, is
not here represented.

The interstices between the various organs and the wall of
the body are occupied by the parenchyma, which consists of a
variety of areolar fibrous tissue with very delicate anastomosing
fibres with plates and nuclei. In the interspaces of this
network are occasionally cells which having no distinctive
character may be called parenchyma cells; but these only
occur irregularly and are entirely absent in many places. To
be regarded perhaps as modified cells of the parenchyma is a
layer of very large cells (Pl. XXI, figs. 3 and 4) lying
between the longitudinal layer of muscle of the body wall and
the testes, and extending from the region of the pharynx in
front to that of the genital aperture behind. These are the
subcutaneous gland-cells whose function is the secretion
of slimy and viscid matter to be discharged on certain areas of
the outer surface. The fibrous tissue of the parenchyma
forms wide meshes in which these cells are contained. The
cells themselves are of colossal size, averaging ‘066 mm. in
diameter, with a large vesicular nucleus like that of an ovum,
and a spherical solid-looking nucleolus. The substances of
the protoplasm of these cells presents three principal modifica-
tions. In the first of these forms there is only a very delicate
protoplasmic network in which the threads have a tendency to
radiate outwards from the nucleus to the periphery. In a
second form there occur in the interstices of the network a
greater or smaller number of rounded granules, ‘(004 mm. in
diameter. In the third variety the place of the network is
more or less completely taken up by numbers of bacilliform
bodies, about ‘02 mm. in length, slender, and with a slight
enlargement at one end. Intermediate stages between the
two last forms are also to be observed, the contents of the cell
consisting of oval bodies, each of which is enclosed in a clear

* “Recherches anatomiques sur la grande douve du foie (Distoma
hepaticum).” [Abstract in ‘Zool. Jahresb.,’ 1882, i, pp. 230, 231.]

288 WILLIAM A. HASWELL.

spherical space. The secretions of these cells have two distinct
destinations. In the case of the more posterior cells of the
gland (Pl. XXI1, fig. 2, d.) the secretion passes out of the
cell by a narrow elongated neck or duck without well-defined
walls, and reaches the exterior by perforating the muscular
layers and passing through basement membrane, epidermis,
and cuticle by means of the delicate pore-canals. The
numerous “ducts” of these cells, which branch and anas-
tomose in their course, pour their secretion over a considerable
area of the ventral surface in front of and behind the genital
aperture, and all over the ventral surface of the sucker. The
function of these cells which discharge around the genital
opening is, doubtless, the secretion of the viscid matter by
means of which the eggs adhere together, while that of the
cells discharging on the sucker is the formation of a similar
sticky secretion adding to the adhesive power of the organ.
The most anteriorly placed of the cells discharge their contents
into a system of narrow anastomosing channels which run
between them. These channels unite on each side anteriorly
to form a larger duct which runs forwards into the region of
the head, past the excretory sac, and breaks up in front into
branches which open on the exterior (through the pore-canals)
on the ventral aspect of the tentacles. The secretion here
voided is a viscid matter similar to the secretion of the pos-
terior part of the gland, and its function is obviously to add to
the prehensile power of the tentacles.

Scarcely to be distinguished from those subcutaneous uni-
cellular glands are certain gland-cells at the base of the penis,
secreting a substance containing spherical granules which
reaches the interior of the ejaculatory duct to become mingled
with the testicular secretion ; but, though apparently merely a
part of the subcutaneous system of unicellular glands, this
group of cells on account of the special destination of its
secretion is best considered along with the reproductive organs.

The system of muscular fibres of the parenchyma
is highly developed. In general they occur either singly or in
narrow bands, running (Pl. XXII, fig. 17, d. v. m.) either trans-

ON TEMNOCEPHALA. 289

versely or obliquely through the parenchyma from the dorsal
to the ventral aspect, and inserted at their extremities into the
basement membrane. In the region of the intestine, however,
the muscular fibres of the parenchyma form a series of about
twelve incomplete transverse dissepiments (Pl. XXI, fig. 8,
sept.), constricting the intestine at regular intervals, and
dividing the peri-intestinal region into a series of incompletely
separated segments. Continuous with these are the layers of
muscular fibres investing the alimentary canal and the repro-
ductive organs.

The muscular fibres of the sucker are derived both from
the muscle of the body wall and from that of the parenchyma.
There are six sets of fibres to be distinguished, viz. (1) tibres
which pass from the dorsal wall of the body to near the centre
of the concavity of the sucker; (2) oblique fibres which run
through the substance of the outer part of the sucker from the
dorsal to the ventral surface; (3) fibres which run longi-
tudinally from the ventral body wall obliquely through the
lateral parts of the stalk ; (4) radial fibres; (5) circular fibres
running round the margin; (6) accessory fibres,

ALIMENTARY CANAL.

The mouth is situated on the ventral surface, near the
anterior end of the body, rather behind the plane of the eyes
and the excretory openings. It is a transverse opening of
considerable width, leading directly into the cavity of the
muscular pharynx. The latter is a subspherical organ with
thick walls and a relatively small cavity. The wall of the
organ (Plate XXI, fig. 6) is constituted as follows. Most inter-
nally is a thick membrane (ep.), almost homogeneous in cha-
racter, but very finely granular, and very finely striated in a
vertical direction. The substance of which this membrane is
composed is not cuticular, but, though entirely devoid of
nuclei, is to be taken as the equivalent of an internal epithe-
lium. Running right through the substance of the wall of the
pharynx is a series of radially arranged muscular fibres (rad.) ;

290 WILLIAM A. HASWELL.

each of those divides both externally and internally into two
or three parts ; internally, these pass into the internal layer, in
which they are traceable for a little distance as vertical lines;
externally, they are affixed to the most external layer. Between
the radiating fibres, which are not closely placed, is a finely
fibrous material, which is little acted on by staining agents, and
embedded in this, here and there, is a large ganglion-cell.
Besides these radiating fibres the wall of the organ comprises
four layers of circularly arranged fibres; the most internal
layer (2. ¢. /.) is longitudinal in direction, the second and third
(e.c.¢, and 7. ¢.¢.) transverse, and the fourth (most external
e.c./.) longitudinal.

The intestine (Plate XXI, fig. 5) is a wide, dorso-ven-
trally compressed sac of rectangular outline, nearly as broad
as long, which occupies about the middle third of the length of
the body and more than half its breath. It is surrounded on
all sides, and to some extent also in front and behind, by the
rounded lobes of the vitelline glands ; in the middle in front it
comes into relation with the pharynx, while behind the recepta-
culum seminis and ovary lie in the concavity of a sort of recess
between two very slight postero-lateral prolongations.

The walls of the intestine are deeply sacculated, a number
of incomplete partitions running inwards from the body wall,
as already described, and producing a series of circular con-
strictions, so that the internal cavity consists of a central
space and a series of annular ceca lying between successive
constrictions.

The intestinal wall (Plate XXI, fig. 7) is of considerable
thickness. It is composed of greatly elongated epithelial cells
(e.) with rounded inner ends which are devoid of cilia. In the
granular substance of the cells themselves, as well as in little
special reservoirs (¢.), which appear to be intracellular, are
very numerous large granules which show a distinct concentric
structure ; these, which colour darkly with hematoxylin, but
are little affected by carmine, are most numerous towards the
bases of the cells. This epithelium is supported externally by
a thin musclar coat (m.).

ON TEMNOCEPHALA. 29]

Excretory System.

The excretory system of Temnocephala differs from that
of most Trematodes! in opening on the exterior by two aper-
tures. These are situated on the dorsal surface slightly behind
the eyes, as already described. The surface in this position is
usually found to be elevated on each side into a rounded
eminence, and in the centre of this is easily to be distinguished
the minute circular opening. This leads into the cavity of a
non-ciliated sac (Plate XXI, figs. 9 and 10) with thick walls,
of a pyriform shape, but slightly bent on itself. From the
narrower posterior end proceed two large vessels which pass
backwards along the lateral edge of the alimentary canal, and
two (or a single one which speedily bifurcates) passing forwards
towards the bases of the tentacles. As ordinarily seen these
canals are twisted into spirals owing to the contraction of the
whole body, but when the body is fully extended they are straight-
ened out. The arrangement and final destination of these
vessels and their branches I did not succeed in tracing. It
would seem probable that the branches open directly into the
spaces and channels in the parenchyma, and the “‘ Schwingende
Lippchen ” found in other forms do not seem to be present.?

The above-described arrangement of the excretory system
would seem to be very characteristic. Paired external open-
ings, where they occur in other Trematodes, are to be found on
the ventral and not on the dorsal surface, with the single
doubtful exception of Polystomum.

The wall of the terminal sac (fig. 10) is composed of a thick
layer of very finely fibrous substance (e.) without a trace of
nuclei, but with occasional rounded vacuoles, and a thin layer
of a similar substance lines the longitudinal vessels (fig. 11).
External to this protoplasmic layer is a thin sheet of the
parenchyma muscle (fig. 10, m.) which completely invests the

1 Vide Fraipont’s “ Recherches sur l’appareil excréteur des Trematodes et
des Cestodes,”’ ‘ Archives de Biologie,” ii (1881).

2 Taschenberg states that they are absent also in Tristomum (see ‘ Zoolo-
gischer Jahresbericht,’ 1879, i, p. 319),

292 WILLIAM A. HASWELL.

organ. The external aperture is capable of being dilated or
contracted by specially arranged fibres of the muscular layers
of the body wall. At the narrow end of the sac, where the
branches are given off, are two large ganglion-cells (figs. 9
and 10, g.).

Nervous System.

The cerebral ganglion (Pl. XXI, fig. 12, and Pl. XXII,
fig. 1) is a six-sided body situated immediately beneath the
longitudinal layer of muscle on the dorsal aspect, just in front
of the pharynx. It consists of a clump of non-nucleated granular
material, having ganglion-cells symmetrically arranged around
it, with lateral, anterior, and posterior commissures of nerve-
fibres. Laterally, it gives off in front a pair of nerve-trunks,
each of which divides into three branches entering the tentacles
—the middle tentacle being supplied by a branch from each
side. Lateral branches pass outwards to supply the sides of
the anterior region of the body. Posteriorly, the ganglion
gives origin to three pairs of nerve-cords, which pass back-
wards towards the posterior end of the body. The ganglion
bends downwards laterally towards the ventral aspect of the
body, but the origin of the branches are still all distinctly
dorsal in position. The first of these—the dorsal (Pl. XXI,
fig. 12)—-are the smallest, and are rather more superficial in
position than the others. They leave the cerebral ganglion at
its posterior and lateral angles, and run along on the dorsal
aspect of the body immediately beneath the longitudinal layer
of muscle in the pigmented “nervous” layer, the distance
between the two cords being less than their distance from the
lateral border. Branches are given off from these cords as
they pass backwards, both internally and externally at frequent
and fairly regular intervals. The internal branches of each
cord pass inwards at right angles to the long axis of the body,
and unite with the corresponding branch of the opposite side,
sometimes after dividing into two. The external branches
bifurcate again and again, the system of fine twigs thus pro-
duced anastomosing freely, and giving rise to a fine meshwork

ON TEMNOCEPHALA. 293

of fibres. Finally, the longitudinal cords, much diminished in
size, unite together near the posterior border.

The whole course of these dorsal cords and their branches
can be made out much more readily than in the case of the
others. If an alcoholic specimen of T. fasciata be macerated
for a day or two in weak bichromate of potash and weak
alcohol, and the cuticle, epidermis, and muscular layers, with
the outer layer of pigment carefully stripped off, the nerves
are readily traced by the light lines which they form among
the pigment of the nervous layer. In young specimens of the
same species in which the outer layer of pigment is imperfectly
developed, more or less of the course of this series of nerves
can usually be seen, and in the New Zealand T. nove-
zelandiz in which the outer layer of pigment is little deve-
loped or absent, the nerves are readily traceable even in the
adult. In the case of the other sets of nerves resort must be
had in tracing them to the study of series of sections, and I
am not able to give more than a general account of their dis-
tribution.

The second pair or dorso-lateral cords (Pl. XXI, fig. 4, and
Pl. XXI, fig. 12, d. /. n.) run backwards on the dorsal aspect
of the body immediately outside the testes, not far from the
lateral border. It is of larger size than the dorsal, and sends
off branches between the organs. The ventral cords (wv. x.)
are the largest. From the ganglion they curve round the
pharynx and run along the ventral aspect in the angle between
. the testes externally, and rather farther from the median line
than from the lateral border. The ventral and lateral longi-
tudinal cords are connected by transverse branches, and the
ventral] cords of opposite sides are united similarly by numerous
- transverse commissures.

The nerve-fibres (P]. XXII, figs. 2 and 3) of which these
- longitudinal cords and their branches consist, are large fibres
averaging ‘01 mm. in diameter, of very delicate material, which
is not readily acted on by staining agents, enclosed in a more
resistent sheath. The central material shrinks greatly in pre-
served specimens, and this with its being little affected by dyes

294 WILLIAM A. HASWELL.

very often gives the fibres the appearance of hollow tubes;
when perfectly preserved it presents a reticulate appearance
such as is represented in fig. 3; but this is very rarely to be
observed. Within the sheath, besides the delicate substance
constituting the nerve-fibre, there are also in the commissural
nerves that form an important part of the cerebral ganglion,
though not, so far as I have observed, in the course of the
peripheral nerves, numbers of bipolar ganglion-cells.

This arrangement of the nervous system is, as regards the
posterior part of it—the six longitudinal cords—very similar to
that described by Lang as observed by him in the Tristo-
mide. The development of the tentacles in Temno-
cephala, and the consequent greater relative importance of
the anterior region of the animal, are accompanied by a greater
development of the nerves running forwards from the cerebral
ganglion.

The single pair of eyes (Pl. XXI, fig. 13) are of extremely
simple structure. They lie almost over the brain, so that the
nerves which pass up to them are very short. The eye consists
of a cup-shaped mass of dense pigment (p.), at the mouth of
which are one or two nerve-cells (g.) not differing from those
of the cerebral ganglion. Enclosed in the cup, the mouth of
which is directed upwards and outwards, is a highly refracting
body (r.) of a spherical form. This stains with difficulty and
not very darkly, and is obscurely fibrillar in minute structure ;
it contains a nucleus near the mouth of the cup, and towards
the base shows a trace of division into separate segments.
Completely enclosed in the substance of the pigment of the cup
on its inner side is a spherical cell (¢.}, of nearly the same dimen-
sions as the body contained in the cavity of the cup; this
exhibits a fine protoplasmic network and contains a solid-
looking nucleus.

1 « Untersuchungen zur vergleichenden Auatomie u. Histologie des Nerven-
systems der Platyelminthen,” ‘ Mittheilungen aus der Zool. Station zu Neapel,’
ii Band. A very similar arrangement is described in Distomum isostomum
by E. Gaffron (“Zum Nervensystem der Trematoden,” ‘Zool. Beitriage,’

herausg. v. A. Schneider, 1884, known to me through an abstract in the
Biologisches Centralblatt,’ iv).

ON TEMNOCEPHALA. 295

The tentacles are also to be regarded as sense organs as
well as aiding in locomotion and prehension. There is nothing,
however, in the structure of these organs, except the presence
of large nerves, specially connected with their sensory func-
tions; the epidermis and the muscular layers resemble those
of other parts of the body.

REPRODUCTIVE ORGANS.

The common genital aperture (Pl. XX, fig. 5, g.) is a tolerably
large slit-like opening situated a little distance in front of
the sucker. It is surrounded by a special set of muscular
fibres which serve the purpose of a sort of sphincter. It
leads into a common genital cloaca, into which on one side
projects the penis, while on the other is situated the female
opening. This cavity is lined by a continuation of the cuticle
and epidermis of the outer surface, the cells of the latter being,
however, considerably elongated, forming an almost columnar
epithelium, external to which is a thick layer of muscle.

The testes(¢e.in Pl. XX, fig.6; Pl. XXI, fig. 4; and Pl, XXII,
fig. 17) are two pairs of large glands of cylindrical form, with
the long axis longitudinal, lying at the sides of the alimentary
canal, and extending throughout the length of the body from
the pharyngeal region to some distance behind the sexual
aperture. The two testes of the same side are connected by a
slender duct. They are invested with an extremely delicate
layer of muscle, which is continued into the wall of the duct
and of the vas deferens. Though there are only two pairs
of testes, these partake to some extent of the segmented
character of the animal—being partially subdivided at the
sides by a deep transverse incision opposite each of the mus-
cular partitions through which, however, the main substance
of the gland is continued uninterrupted. The spermatozoa
have pear-shaped heads, about ‘0046 mm. in diameter, and
slender flagella, ‘083 mm. in length.

The two vasa deferentia are slender tubes, which, passing
inwards from the posterior testes towards the middle line of
the body, meet to form a large seminal vesicle or spermatic

296 WILLIAM A. HASWELL.

reservoir (Pl. XXII, fig. 17, e.7.), which is always found to be
distended with spermatozoa. This is an elongated sac, much
dilated proximally, which runs almost transversely from near
the middle line of the body where the vasa deferentia open
into it towards the right, and opening into the base of the canal
of the penis. The latter organ (Pl. XX, fig. 6, p.) is con-
tained, when retracted, in an elongated muscular sac lying
transversely with the mouth directed towards the left and
towards the dorsal side. It is a cylindrical, slightly curved,
chitinous body, having a wider proximal and narrower distal
end. In T. fasciata (Pl. XX, figs. 5—7) and T. quadri-
cornis (Pl. XX, fig. 8) the distal end is provided with a knob
or glans (g/.); in T. minor and T. nove-zelandiz, this is
merely represented by a slight rim (Pl. XXII, figs. 9 and 10).
The whole is enclosed in a sheath composed of circular and
longitudinal muscular fibres—the latter the stronger, and
enclosed by the former. At the opening of the penis the
sheath is continuous with the proper chitinous wall of the
organ, it is continuous also proximally with the sheath of the
spermatic reservoir and distally with the muscular investment
of the genital cloaca. In T. fasciata, where it turns back to
become continuous with the proper wall of the penal cylinder,
it is provided with a number of chitinous spines, which, when
the penis is retracted, lie in the interior of the terminal knob
or glans in a radiating manner, their outer ends, which are the
broader, embedded in the sheath and the acute inner ends
pointing into the narrow lumen. When the penis is pro-
tracted this inverted part of the sheath will become everted,
and the spine project on the exterior of the end of the penis,
thus enabling the organ to retain a firm hold during the act
of copulation. In T. minor and T. nove-zelandiz there
is only a slight rim to effect this purpose; but in the latter
species (Pl. XXII, fig. 19) the female opening is provided with
inwardly directed spines, which doubtless effect the same
object. The interior of the penis and spermatic reservoir is
lined with a protoplasmic layer containing nuclei, but without
cell boundaries. Into the lumen of the ejaculatory duct there

ON TEMNOCEPHALA. 297

is discharged the secretion of the unicellular glands already
mentioned as apparently forming a part of the system of sub-
cutaneous glands. The secretion consists of or contains little
spherical, highly refracting bodies, which stain darkly with
hematoxylin; probably the foreign particles found in the re-
ceptaculum seminis with the spermatozoa are the products of
this secretion.

The female organs consist of receptaculum seminis, or
single ovary, oviduct, uterus, vitelline, and uterine glands,
together with certain of the subcutaneous glands opening
around the sexual orifice, which probably secrete the viscid
matter by means of which the eggs adhere together.

The receptaculum seminis (Pl. XX, fig. 6, 7e., and Pl.
XXI, fig. 8, rec. sem.) is a large rounded sac which lies in
the middle line in a deep bay of the posterior wall of the in-
testine. Its walls are formed of a granular protoplasmic
matter, without differentiation, into cells, but with large nuclei
here and there. External to this is a thin layer of muscular
fibres. In all the specimens examined the cavity of the sac
was found to be full of spermatozoa, with frequently small
particles of amorphous matter, probably derived from the
accessory glands of the male organs.

Opening out of the left-hand corner of the receptaculum
seminis is the oviduct (o.d. in Pl. XX, fig. 6; Pl. XXI, fig.
8; and Pl. XXII, fig. 12), a rather narrow, curved tube which
opens below into the uterus. The wall of the oviduct (PI.
XXII, fig. 12) consists of an external circular and an internal
longitudinal layer of muscular fibres, and is devoid of epithe-
lium, but is lined internally with a homogeneous layer of some
delicate non-nucleated material, which is not readily acted on
by staining agents. Into its lumen open the ducts of a few
of the shell-glands.

The wall of the uterus (Pl. XXI, fig. 8) resembles that of
the oviduct in structure; but the muscular layer is thicker,
and the fibres cross one another in all directions. Most of the
shell-glands open into the uterus. Lach shell-gland (s. gl.) is a
single, irregularly.shaped cell of very large size, with a large

298 WILLIAM A. HASWELL.

nucleus ; passing from it to the oviduct or uterus is the narrow
duct (d.), which is essentially a process of the cell substance.
As the duct passes through the wall of the uterus it acquires
definite boundaries, and presents a little vesicular enlarge-
ment (d.) just before it opens into the uterus. Leading from
the uterus to the genital cloaca is the vagina, a short, narrow
passage, which in T. nove-zelandiz is provided at its mouth
with a series of chitinous teeth.

The vitelline glands (Pl. XXII, figs. 15 and 16) consist of
a number of rounded lobules arranged in narrow branching
lobes, which, for the most part, take a transverse direction.
These lobes are very closely applied to the wall of the intestine,
which they almost entirely cover, both on the dorsal and on the
ventral aspect. The muscular septa pass outwards from the
wall of the alimentary canal, and through between the lobules
of the vitelline gland in such a way as to bring about an
imperfectly metameric arrangement in irregular transverse
lobes, which, however, branch and anastomose with one another.
The arrangement of the lobes and the degree to which the
glands are developed vary somewhat in different individuals.
The lobules are invested in a thin layer of the parenchyma
muscle, which is continuous with that constituting the septa,
and similar to the layer investing the intestine. The central
substance of the lobules is composed of large irregular cells
with ill-defined boundaries, whose protoplasm is clear and
colourless in the fresh state, but with a number of large
granules which become more evident in hardened specimens ;
the nuclei resemble those of the alimentary epithelium;
throughout the protoplasm there are frequently large rounded
vacuoles.

The ovary (Pl. XXII, fig. 11) is an oval solid body ‘16 mm.
in length, attached to the right wall of the receptaculum seminis,
close to the beginning of the oviduct. The ova are narrow
pyramids about ‘083 mm. in length, each of which passes trans-
versely through the entire thickness of the ovary. The
impregnated ova are received singly into the uterus, where they
become surrounded by a considerable quantity of vitelline

ON TEMNOCEPHALA. 299

matter, and become enclosed in a chitinous shell secreted by
the shell-glands. The egg is now compared with the animal,
a very large structure (as much as a sixth of the length of the
animal), and greatly distends the walls of the uterus. When
extended (Pl. XXII, fig. 18) it has a short stalk, by means of
which it becomes attached to the shell of the crayfish, and is en-
closed in viscid matter, which when it hardens serves to cement
the eggs together. The eggs are found in considerable numbers
from October to February, attached chiefly to the under surface
of the abdomen, some also at the sides of the mouth and the
lower edges of the branchiostegites. The development of the
embryo has not yet been studied; there is, as in other ecto-
parasitic Trematodes, no metamorphosis. Temnocephalea,
perfect in every respect, being found still enclosed in the egg.

AFFINITIES OF TEMNOCEPHALA.

Though most nearly related to the Tristomide, Temno-
cephala presents so many special peculiarities that it becomes
necessary to regard it as the type of a distinct family. The
principal characteristic features in its structure may be sum-
marsied as follows :

The cephalic end of the body is produced into four, five, or
six slender, filiform tentacles, which are capable of being used
for prehension and touch, and in locomotion take the place of
anterior suckers, their adhesive powers being increased by the
secretion of certain special unicellular glands. There is a
single, large, radiated posterior sucker without hooks. The
body presents traces of a rudimentary form of segmentation
in the shape of incomplete transverse dissepiments formed by
specialised portions of the parenchyma muscle. The intestine
is constricted at regular intervals by these septa ; its epithelium
is not ciliated. There are three pairs of longitudinal nerve-
trunks, a dorsal, a dorso-lateral, and a ventral, connected by
numerous commissures. The excretory system opens by two
apertures, placed far forwards on the dorsal surface. There is
a single genital aperture leading into a genital cloaca, into

300 WILLIAM A. HASWELL.

which the ejaculatory duct and the vagina open ; there are two
pairs of lobed testes, vitelline glands, which partake of the
imperfect segmentation of the body, a single ovary, recepta-
culum ‘seminis, oviduct and uterus.

The broader questions suggested by the imperfect segmen-
tation of Temnocephala, and by other features in its organization
which seem to point to a possible genetic relationship with the
segmented worms, cannot well be dealt with until the develop-
ment has been investigated.

EXPLANATION OF PLATES XX, XXI, & XXII,

Illustrating Mr. William A. Haswell’s paper “On Temnoce-
phala, an Aberrant Monogenetic Trematode.”

PLATE XX.

Fic. 1—Temnocephala fasciata in various positions. Natural size.
Fic. 2.—Temnocephala fasciata, from living specimens. Magnified.
Fic. 3.—Temnocephala quadricornis. Magnified.

Fic. 4.—Temnocephala minor, dorsal view, from preserved specimen.
Magnified.

Fic. 5.—Temnocephala minor, ventral view. g. Genital aperture,
m. Mouth. s. Sucker. ¢e. Tentacles.

Fie. 6.—Diagram of the general organization of Temnocephala. ph.
Pharynx. 7. Intestine. ex. Excretory sac. ea’. Anterior canal of excretory
system. ew'’. Posterior canal of excretory system. 7. Brain. c/. Genital
cloaca. ¢e. Testes. v.d. Vas deferens. jp. Penis. re. Receptaculum
seminis. ov. Ovary. od. Oviduct and uterus. vé¢. Vitelline glands. s.
Sucker.

PLATE XXI.

Fic. 1.—Transverse section through the body wall of Temnocephala
fasciata. ec. Cuticle. e. Epidermis. 4. Basement membrane. cc. m, Cir-
cularly arranged layer of muscular fibres. py. Pigment layer. /. m. Longi-
tudinal layer of fibres. par. Parenchyma.

ON TEMNOCEPHALA. 301

Fic. 2.—Longitudinal section of the ventral body wall, in the neighbour-
hood of the genital opening, to show the ducts of the subcutaneous glands,
d. Other letters as above.

Fie. 3. Cells of subcutaneous glands.

Fic. 4.—Transverse section of a young specimen of Temnocephala
fasciata, in the region of the genital cloaca and penis. c. Cuticle. e. Epi-
dermis. 4. Basement membrane. p. Pigment layer. 7. m. Longitudinal
layer of muscle. par. Parenchyma, with subcutaneous gland-cells and dorso-
ventral muscular fibres. d. /. x. Dorso-lateral nerve. v.. Ventral nerve.
cl. Genital cloaca. pe. Penis. vit. Lobule of vitelline gland. ¢e. Extremity
of testis (here imperfectly developed).

Fic. 5.—Outline of the alimentary canal of a young specimen of Temno-
cephala fasciata. ph. Pharynx. i. Intestine.

Fic. 6.—Vertical section of the wall of the pharynx, from a longitudinal
section of the body. ep. Internal layer of granular matter. 7. c. 7. Internal
circular longitudinal layer of muscle. 7. c. ¢. Internal circular transverse
layer. e.c.¢. External transverse. .c. 7. External longitudinal. — rad.
Radiating fibres. g. Ganglion-cells.

Fic. 7. Section of the wall of the intestine. m. Layer of muscle. e. Epi-
thelium. c. Reservoirs of granules.

Fic. 8.—Longitudinal and vertical section of Temnocephala minor.
s. Sucker. v. Ventral body wall. d. Dorsal body wall. iz¢. Cavity of the
intestine. sept. Septa. vit. Lobules of vitelline gland. g/. Subcutaneous
gland-cells. od. Oviduct. uw. Uterus. vec. Receptaculum seminis.

Fic. 9.—Excretory sac of Temnocephala fasciata in the fresh condi-
dition, viewed from the dorsal aspect. o. External opening. J/. ce. Longitudinal
canals. g. Nuclei of ganglion-cells.

Fic. 10.—Section of excretory sac, on a level with the external opening.
e. Proper wall of sac, with vacuoles. m. Investing layer of muscle. g.
Ganglion-cell.

Fic. 11.—Longitudinal section of posterior excretory canal.

Fie. 12.—Diagram of dorsal portion of nervous system, showing the arrange-
ment of the tentacular nerves, and the dorsal longitudinal trunks with their
branches and commissures.

Fic. 13.—Transverse section through the eyes of Temnocephala fasci-
ata. p. Pigmentcup. 7. Contained substance. g. Ganglion-cells opposite the
mouth of the cup. ¢. Peculiar cell enclosed in the pigment of the optic cup.
zu. Large cells lying between the eyes.

VOL, XXVIII, PART 2.—NEW SER. x

302 WILLIAM A. HASWELL.

PLATE XXII.

Fie. 1.—Transverse section through the brain ganglion of Temnocephala
fasciata.

Fic. 2.—Transverse section of a portion of the ventral nerve.

Fic. 3.—Longitudinal section of nerve-fibre from dorsal nerve-plexus.

Fic. 4.—Elements of testis.

Fic. 5.—Penis and penis-sheath of Temnocephala fasciata. cl. Genital
cloaca. sh. Penis-sheath. g/. Terminal enlargement of “ glans.”

Fic. 6.—Transverse section of the distal end of the penis of Temnoce-
phala fasciata, through the “glans.” sh. Sheath. sp. Inverted chitinous
spines.

Fic. 7.—Oblique section through spermatic reservoir, penis, and penis-
sheath of Temnocephala fasciata, from a horizontal section of the body.
cl. Genital cloaca. sh. Sheath. ej. Spermatic reservoir (ejaculatory duct).

Fic. 8.—End of penis of Temnocephala quadricornis.

Fic. 9.—End of penis of Temnocephala minor.

Fie. 10.—End of penis of Temnocephala nove-zelandie.

Fic. 11.—Ovary of Temnocephala fasciata.

Fic. 12.—Transverse section of the oviduct of Temnocephala fasciata.
ce. m. Circular layer of muscle. JZ. m. Longitudinal layer of muscle. p. In-
ternal homogeneous substance. / Investing reticulum of fibrous tissue.

Fie. 13.—Vertical section of the wall of the uterus in the same species,
with the shell-glands and their ducts. s. g/. Shell-glands. d. Ducts. d’.
Terminal enlargements of the ducts. m. Muscle. jp. Thin internal homo-
geneous layer.

Fic. 14.—Oblique section of uterine wall nearly parallel with the surface,
showing the terminal dilatations of the ducts of the shell-glands with the
investing plexus of muscular fibres.

Fic. 15.—Outline of the ventral part of the vitelline glands of Temnoce-
phala fasciata. 7. Outline of the wall of the intestine.

Fic. 16.— Section through a lobule of the vitelline glands. m. Investing
layer of muscular fibres. v. Vacuoles in the protoplasm of the gland-cells.

Fic. 17.—Portion of a transverse section of a young specimen of Temno-
cephala fasciata, just behind the pharynx. ce. Cuticle. ¢. Epidermis.
&. Basement membrane. J. m. Longitudinal layer of muscle. v. /. m. Ventral
longitudinal layer of muscle. par Parenchyma, with subcutaneous glands.
d. v. m. Dorso-ventral bundles of muscular fibres. . /. x. Dorso-lateral nerve.
v. m. Ventral nerve. izé. Intestine. ¢e. Anterior end of testis.

Fic. 18.—Eggs of Temnocephala fasciata. Magnified.

Fic. 19.—Opening of the vagina of Temnocephala nove-zelandie.

Mor Sourn! Vb UU NSF

Fig. I.

. \
iF
Te,
\
~

B
a
“Speac
Aw
,

Easwell del.
F Huth Lith Edin®

as
3
bg
‘A
by

e
By Moor fourn' VA. AAU NS. DY, EL

ry
j
ues

)) 4

S ‘aii
AAA

——-~

F. Huth, Lith? Edin®

z

- Ag Vvyrrzsy, Tr Zz, VL;
Mlscr, Zourn Vol XM NS PYIL

ge

F Huth, Lith? Edin®

NOTES ON ECHINODERM MORPHOLOGY. 303

Notes on Echinoderm Morphology, No. XI. On
the Development of the Apical Plates in
Amphiura squamata.

By

P. Herbert Carpenter, D.Sc., F.R.S., F.L.S.,
Assistant Master at Eton College.

Some very important observations “ On the Development of
the Calcareous Plates of Amphiura” have been recently pub-
lished by Dr. J. W. Fewkes,! who has also entered into a dis-
cussion respecting the morphological relations of these plates.
Amphiura squamata is viviparous, and the young undergo
a direct or abbreviated development without passing through a
definite Pluteus form, so that it is easy to follow the various
larval stages and to study the mode of appearance of the
skeletal plates.

Ludwig took up the subject in 1881, and made the very
striking discovery that two rings of radially situated plates are
developed around the dorsocentral. Those of the distal ring
are carried outwards from the disc at the ends of the growing
arms, and are known as the terminals (fig. 1, T) ; while the
five plates of the proximal ring (4) remain on the disc close to
the dorsocentral (1), and with it develop into the rosette of six
primary plates which is a prominent feature on the abactina.
surface of many adult Ophiurids. In Amphiura squamata
two more rings of plates, the intermediate plates of Ludwig,
eventually appear between these radials and the dorsocentral ;

1 «Bull. Mus. Comp. Zodl.,’ 1887, vol. xiii, No. 4, pp. 107—150, pl. i—iii.

9

2 « Zur Entwicklungsgeschichte des Ophiurenskelettes,”’ ‘ Zeitschr. f. wiss.
Zool.,’ 1881, Bd. xxxvi, pp. 181—200, Taf. x, xi.

304 P, HERBERT CARPENTER.

and I expressed my belief in 1882! that they respectively
represent the under-basals and basals of the dicyclic Crinoids
(fig. 11, 2,8). Two years later I pointed out the presence of
this dicyclic base in various adult Ophiurids?; while the
important discovery was announced by Sladen that it likewise
occurs in many Asterids, both larval and adult, and also that
in this group, just as in the Ophiurids, there are primary
radial plates which remain on the disc while the terminals are
carried outwards on the growing arms.®

These primary radial plates of Asterids and Ophiurids,
including Amphiura, were homologised by both Sladen and
myself with the first radials of the Crinoidea, the correspond-
ence of which with the ocular plates of the Urchins had been
long ago pointed out by Agassiz, Lovén, and others. No one
had questioned this latter comparison, and it appeared equally
indisputable that the primary radials of Amphiura were also
homologous with the ocular plates of an Urchin, despite the
relation of the latter to the eye-spot, a relation which is not
characteristic of the primary radials in either of the three
. groups of brachiate Echinoderms.

The subject has been lately reopened by Fewkes,* who
says :—

“The question which plates in the young Amphiura cor-
respond to the oculars of sea-urchins, assumes a new phase in
the light of what we know of the permanent retention of the
radials in the abactinal hemisome of the body of Amphiura,
and the relation of the terminals to the primitive tentacle.
Now that we know that the primary radials of Amphiura are

1 «Notes on Echinoderm Morphology,” No. V: “On the Homologies of

the Apical System, with some Remarks upon the Blood-vessels,” ‘ Quart.
Journ. Mier. Sci.,’ 1882, vol. xxii, New Ser., p. 380.

2 “ Notes on Echinoderm Morphology,” No. VIL: “On the Apical System
of the Ophiurids,” ‘ Quart. Journ. Mier. Sci.,’ 1884, vol. xxiv, New Ser., pp.
4,11.

3 “On the Homologies of the Primary Larval Plates in the Test of Brachiate
Echinoderms,’’ ‘Quart. Journ. Micr. Sci.,? 1884, vol. xxiv, New Ser., pp.
29, 34.

4 Loe. cit., p. 124.

Rites mae

NOTES ON ECHINODERM MORPHOLOGY. 305

not pushed out to the extremity of the rays, but always remain
in the disk, and that another set of plates (terminals) do suffer
this change, we have this difficulty in a comparison of the
young Echinoid with the young Amphiura. The terminals of
Amphiura are independent centres of calcification from the
radials. If terminals and radials in Amphiura lie in the same
radius, how can the one or the other, especially the former, be
the same as the oculars of the sea-urchin ?”

I must confess that I do not quite understand what “ new ”
phase this question has assumed since I wrote on the subject
in 1882. All the facts to which Fewkes refers in the above
passage were then known, having been discovered by Ludwig
in the previous year ; and it was these very facts which led me
to homologise the primary radials of Amphiura and of the
Ophiurids generally with the oculars of an Urchin. Sladen
has accepted this view, and I am not aware of its ever having
been disputed.' I fail to see therefore what the new phase of
the question may be to which Fewkes refers; for the fact
which he mentions, that the radials of Ophiurids are not per-
forated for the eye-spots, like the oculars of an Urchin, is
familiar to all Echinoderm students.

The radials of a Crinoid are equally devoid of any relation
to the ambulacral structures, and yet they have been always
regarded as homologous with the oculars of an Urchin by
every writer who has dealt with the subject; and Fewkes2
arrives at the same conclusion as Sladen and myself had pre-
viously done, that ‘it seems more natural to compare radials
in Amphiura with oculars in sea-urchins, notwithstanding the

‘ Fewkes refers, somewhat unnecessarily, “ to those who compare the ter-
minals of starfishes and brittle stars without pluteus with the ocular plates
of the sea-urchin.” I know of no writer who has done so since the publication
of Ludwig’s and Sladen’s discoveries respecting the development of both radials
and terminals in Ophiurids and Asterids. Neither has anyone, that 1 am
aware of, suggested that either terminals or radials of Amphiura are com-
parable to the genitals of the Urchins, though Fewkes takes some trouble to
point out the very obvious impossibility of such an homology (p. 125).

? Loc. cit., p. 126.

306 P. HERBERT CARPENTER.

position of the eye-spot.”” One or two further arguments may
be adduced in favour of this view.

1. The embryonic radials of Ophiurids always remain upon
the abactinal surface of the adult in the neighbourhood of the
dorsocentral, and in many species develope into large plates,
the primaries, which form a closed ring around the dorsocentral,
just as in the early stages of the young Amphiura (fig. 1, 4),
though they are sometimes separated from it by intermediate

Fie. I.—(‘ Quart. Journ. Micr. Sci.,’? 1882, vol. xxii, p. 379.) Apical system
of the young Amphiura squamata; StagelI, after Ludwig. 1. Dorso-
central. 4, Radials. 5. Orals. T. Terminals. ad,. Second adambula-
cral plates. w. p. Water-pore.

plates (fig. 11, 2,3). The oculars of an Urchin are also confined
to the abactinal surface, and are separated from the dorso-
central by one ring of plates; while in most Crinoids the
radials form a closed ring as in the earliest Amphiura,
sometimes with one and sometimes with two rings of plates
within them, just as in the two subsequent stages of the deve-
lopment of Amphiura.

eee ce

NOTES ON ECHINODERM MORPHOLOGY. 307

On the other hand, the terminals of Ophiurids never form a
closed ring of plates, but are well separated laterally from their
very first appearance, and become more and more separated as
the growth of the arms carries them away from the disc,

2. Sladen has shown that there are both terminal and radial
plates in the young Asterid, just as in the young Ophiurid,
and, like all his predecessors, he regards the terminals of
Asterids as homologous with those of Ophiurids, although the
latter have no eye-spot. If this homology be admitted—and I do
not know of its ever having been disputed—we have an exactly
similar case to that of the Ophiurid radials and the oculars of an
Urchin, the one set of plates being associated with an eye-spot
and the other not, although they are mutually homologous.

We may then, I think, assert, without fear of contradiction,
that the radial plates are mutually homologous in Ophiurids
and Urchins, Asterids and Crinoids, and also that the relative
time of their appearance is of no general morphological im-
portance. They appear before the basals in the Ophiurids, but
after them in the other three groups, being also anticipated in
Asterids' and Crinoids,*? and probably in the Urchins as well,
by the dorsocentral.

Thus then, so far as concerns the primary radial plates of

1 On pp. 124, 127 Fewkes refers to the works of Agassiz and Balfour as appa-
rently proving that the radials of Starfishes appear before the dorsocentral. He
forgets, however, that the plates which were formerly described as the radials
of Starfishes are now known to be the terminals, and that the true radials do
not appear till after the basals and the dorsocentral, as shown by Sladen.

2 Fewkes remarks that “it is difficult to compare the centrodorsal of
Crinoids with the dorsocentral of the Ophiuran” (p. 127). But why make
the attempt? There is a dorsocentral in Crinoids perfectly homologous
with that of the Ophiuran, as pointed out by Sladen and myself. But it
is at the bottom of the larval stem, and altogether different from the cirrus-
bearing top stem-joint or centro-dorsal. The distinction between the two
was explained in this Journal as long ago as 1878; and the homology of
the terminal plate at the base of the Crinoid stem with the dorsocentral
of the Echinozoa has been frequently noticed since then by Liitken, Duncan
and Sladen, Wachsmuth and Springer, Etheridge and myself. I am at a loss
to understand therefore why Fewkes refers to the Crinoid centro-dorsal, but
omits all notice of the dorsocentral.

308 P. HERBERT CARPENTER.

the apical system, Fewkes’s views are those of his predecessors,
as indeed he himself admits. But he adds a suggestion which
he considers as a direct sequence of the homology of the
primary radials in Ophiurids and Crinoids, respectively! It is
this: ‘ The homology of the radial shields of the Amphiura
with the first brachials of the Crinoid would seem not unrea-
sonable. The only paired plates of the arms with which they
could be compared are the adambulacral. With these, however,
they have little resemblance save in their double origin.” If,
as I suppose, Fewkes uses the term “ first brachials ” to denote
the lowest joints of the two arms which are borne upon the
radial axillaries of Pentacrinus, Antedon, and most recent
Crinoids, there are at least three objections to an homology
between them and the radial shields of the Ophiurids.

1. Many Crinoids, e.g. Hyocrinus, Rhizocrinus,
Eudiocrinus, Cupressocrinus, &c., have no paired first
brachials at all, for there are only five arms, one on each
primary radial.

2. In those Crinoids which have ten or more arms there may
be from one to seven radial plates between the primary or calyx-
radials and the paired first brachials. The only genera in
which the latter plates ever rest directly on the primary radials
are the aberrant Allagecrinus and Tribrachiocrinus.
But this is not the case all round the cup, so that there are
never ten first brachials to correspond to the ten radial shields
which are so constant in Ophiurids. Then, again, each of the
two larger radials in Catillocrinus may support as many
as thirty arms. Where are the homologues of the two radial
shields among all these first brachials ?

3. The radial shields of Ophiurids are invariably present, but
they have no permanent relations to the primary radials. It is
not unlikely that they are always developed immediately outside
the primaries, where they remain in many species. But, on
the other hand, they are often separated from the primaries by
a series of intermediate plates which exhibit no general
constancy of arrangement, and cannot therefore be directly

» hoc: cit, p: 130,

NOTES ON ECHINODERM MORPHOLOGY. 309

compared to the single row of outer radials which presents
itself so frequently in the Crinoids.!

These three considerations seem to me to indicate pretty
clearly that it is useless to seek for the homologues of the
radial shields in Ophiurids among the post-radial plates of
Crinoids, more especially as we have not yet been able to
identify them in any Asterid; and I prefer to regard them,
like the terminals of both Ophiurids and Asterids, as plates
which have no representatives in the Crinoidea.

Fewkes also discusses the homology of those intraradial or
adaxial plates in the young Amphiura which I have regarded
as homologous with the basals of Crinoids (figs. 1—111, 3) ; and
he suggests some doubts as to whether these plates “ are basals
in preference to other interradial plates.’

The plates in question have an interradial position within
the ring of radials, i.e. between them and the dorsocentral;
and at one stage of development they are the only adaxial
interradial plates (fig. 111, 3). They thus correspond exactly
to the basals of monocyclic Crinoids, and to the so-called
genitals of Urchins and Asterids. In a large number of
Ophiurids they develop into large plates which form a closed
ring round the dorsocentral, with the radials immediately out-
side them, precisely as in the apex of an Urchin, or the calyx
of a stemless Crinoid lke Uintacrinus. Fewkes must
surely be aware of this fact, for there are numerous figures il-
lustrating it in Lyman’s report on the “ Challenger” Ophiurids ;
and it has formed the subject of much discussion by Sladen
and myself. He supports his doubts by no arguments what-
ever, and entirely overlooks the important fact that the inter-
radials of a Crinoid never form a closed ring within the
circle of radials; for the only inter-radially situated plates
which occupy such a position are the basals. It was this fact

1 T am quite ready to admit, however, that the outer radials and the brachials
of a Crinoid are in a general way represented in the Ophiurid by the upper
arm-plates, and the row of median plates beyond the radial primaries which

occur in Ophiomusium, Ophioglypha, and other genera.
2 Loc. cit., pp. 129, 146.

310 P. HERBERT CARPENTER.

which led me to suggest that these first formed plates between
the radials and the dorsocentral of the young Amphiura are
the homologues of the Crinoidal basals and of the so-called
genitals in Urchins and Asterids ; and the six years’ work by
myself and others since this suggestion was made has only
served to make me more confident of its truth.

Ludwig gave a very clear description! of the appearance of
these inter-radially placed basals in the young Amphiura,
and he pointed out how at a later stage a set of radially
situated plates appears between them and the dorsocentral.
He illustrated both conditions by figs. 24 and 25 of his series,
the latter showing five and the former ten plates between the
dorsocentral and the radial primaries. Fig. 24, which I here
repeat (fig. 11), was copied by myself for the purpose of

eG OY: Qs &
PSN aaa 4

Te a Ts

Fic. I1.—(‘ Quart. Journ. Mier. Sci.,’ 1882, vol. xxii, p. 380.) Apical system
of the young Amphiura squamata; Stage III, after Ludwig.
1. Dorsocentral. 2. Under-basals. 3. Basals. 4. Radials. r,s. Ra-
dial shields.

demonstrating that plates corresponding in position to the
under-basals of a Crinoid (2) were developed in the young
1 Loe. cit., p. 195.

NOTES ON ECHINODERM MORPHOLOGY. Bi

Amphiura, as well as those corresponding to the basals (8).
Fewkes! thinks, however, that it would have been better “ in
considering the relationship of the basals in the young
Amphiura and their homologues in the projection of the calyx
of an Antedon larva” if I had copied Ludwig’s fig. 25
instead of his fig. 24, i. e. that showing the monocyclic and not
the dicyclic stage of Amphiura. It is curious that he fails
to see the reason why I copied the figure of the older, in pre-
ference to that of the younger stage. Homologues of the
Crinoidal basals had long been recognised in the so-called
genital plates of Urchins and Starfishes. But up to the time
of the publication of Ludwig’s memoir, representatives of the
under-basals of a Crinoid? were not known to exist in any of the
Echinozoa ; and it was for the purpose of demonstrating their
presence in Amphiura that I copied the figure of the older
and dicyclic condition rather than that of the younger and
monocyclic one, with the remark,’ “The plates in the outer
ring (fig. 11, 3) are inter-radial, while those of the inner ring
next the dorso-central are radial in position (fig. 11,2). In
these plates we have, I believe, the representatives of the
diycyclic base of Marsupites and other Crinoids, viz. a
proximal ring of under-basals hitherto unknown in any of the
Echinozoa, and a distal ring of interradial plates correspond-
ing to the basals of the Crinoid and the genitals of the Asterid
or Urchin, which have not been previously discovered in an
Ophiurid.”

The doubts which seem to have occurred to Fewkes respect-
ing the homologies of the adaxial ring of interradial plates in

1 Loc. cit., p. 129.

2 While this paper was in the press an important discovery was announced
by Mr. H. Bury at the Manchester meeting of the British Association. He
has found under-basals in the ciliated larva of Antedon rosacea; but they
soon fuse with the top stem-joint (centro-dorsal), and all trace of them is
lost when the cirri appear. This is a very striking confirmation of the views
of Messrs. Wachsmuth and Springer, whose paleontological studies had led
them to express the belief that under-basals might be present in the early

larva of Comatule.
3 ‘Quart. Journ. Micr. Sci.,’ 1882, vol. xxii, New Ser., p. 380.

312 P. HERBERT CARPENTER.

the young Amphiura which I have described as basals (figs.
Il, 111, 3) appear to me to be due to the fact that the sequence
of development of the apical plates is not the same in the
American as in the European variety of Amphiura squa-
mata. According to the description and figures of Ludwig
the order in the latter form is as follows:—i. Radials (4);
ii. Dorsocentral (1); iii. Basals (8); iv. Primary interradials
(I.), and Under-basals (2) ; and v. Radial shields (r. s.). The
under-basals and primary interradials seem to develop almost
ccntemporaneously, fig. 111 showing two under-basals (2) in
two rays between the dorsocentral (1) and the radials (4), and

Fie. I1].—Apical system and one arm of the young Amphiura squamata;
Stage II, after Ludwig. 1. Dorsocentral. 2. Under-basals. 3. Basals.
4. Radials. 5. Orals. I. Primary interradials. T. Terminals. ad;., ad,.,
ad,, Side arm-plates (adambulacrals).

two very unequal abaxial or primary interradials (I.) each of
them between a basal (3) and what I take to be an oral (5).
This figure also shows that the fifth pair of side arm-plates

1 Tt is just possible that this interpretation may be erroneous, and that the
plates which are marked as orals (5) in the above diagram are really the
primary interradials, and should be marked (I.)._ I do not think this probable,
however, as this individual is so little older than that figured by Ludwig in his
fig. 25, in which the orals are still quite distinct on the margin of the dorsal
surface. But if the plates in question be interradials appearing thus early,
the contrast between the times of development of the interradials and radial
shields in the American and European forms respectively is even greater
than I have described above. Fewkes agrees with me in thinking that they
are probably orals.

NOTES ON ECHINODERM MORPHOLOGY. 313

(adambulacrals, ad;.) appear about the same time! It is
evident therefore from fig. 111 that the basals (3) reach a com-
paratively large size before the appearance of either inter-
radials (I.), under-basals (2), or radial shields, and that the
latter do not appear till after the formation of two at least of
the under-basals and interradials.

From Fewkes’s descriptions and figures, however, it would
seem that the radial shields appear much earlier in the
American variety, and that the developmental sequence is
altogether different from that described by Ludwig for the
European form. Fewkes’s account is a little difficult to follow,
owing to a want of precision in his terminology. Thus, for
example, he three times uses the names “ basals” and “ inter-
radials” as if they were synonymous; though every previous
writer upon the calyx of the Crinoids and its relation to that of
other Echinoderms has regarded the interradials and basals
as fundamentally distinct in their morphological relations.

The earliest plates to appear in the inter-radial areas are the
orals (figs. 1, 111, 5) which, as Fewkes himself describes on p.
128, are “‘ forced to the actinal surface of the disc before the
interradials arise.’ He speaks of the latter, the first inter-
radials, as forming “ on the periphery of the abactinal hemi-
some on interradii between contiguous radialia. They are
triangular in shape, and occupy a triangular interspace
between adjoining primary radials.” He calls them, however,
abaxial basals or abaxial interradials, and they are designated
in his figure by the letters 7 p*. (fig. rv, I). But as they are
situated on the periphery of the abactinal hemisome outside
the closed ring of radial primaries, it is altogether incorrect to
speak of them as “ basals,”’ This term is only applicable to
the ring of interradial plates which are situated adaxially to
the primary radials, i.e. between them and the dorsocentral
(figs. 11, m1, 3). The basals of a monocyclic Crinoid, as
implied in their name, which was given to them by Johannes

1 The reader will do well to remember that the first two pairs of adambu-

lacrals are on the ventral surface of the disc at this stage, and are there-
fore not visible in its dorsal aspect.

314 P. HERBERT CARPENTER,

Miiller, are the nearest to the vertical axis of the calyx of all
the plates in the interradial areas ; and the term abaxial basals

Fic. [1V.—Dorsal view of one ray of the young Amphiura squamata;
Stage II, after Fewkes. 1. Dorsocentral. 4. Radials. I. Primary inter-
radials. 7. Other interradials. 7.s. Radial shields. am. Ambulacral
plates. d,., d,. Upper arm-plates (dorsals). ad,., ad,., ad;. Side arm-
plates (adambulacrals). s. Their spines.

is thus self-contradictory. The proximal ring of interradially
situated plates or basals (figs. 11, 111, 8) cannot be called
abaxial; and the interradial plates belonging to a distal or
abaxial ring (fig. tv, I) are most certainly not basals. The
term adaxial interradials is an expressive (but lengthy) name
for the basals (3), as it implies that they le within the ring of
radials (4). But ‘abaxial basals” as a synonym for the
interradial plates (1) beyond this ring is an impossibility ; and
the use of the term by a trained morphologist like Fewkes has
surprised me considerably.

According to his description! the true basals or adaxial inter-
radials seem to be the second set of plates to be developed inter-
radially in the apical system of the American Amphiura,
arising “in the corners left between the dorsocentral and con-
tiguous radialia.” But before this happens the radial shields
and two upper arm-plates have made their appearance (fig.
Iv, r. s.; @. d*.). At any rate this is how I should interpret
Fewkes’s fig. 19, which is copied in fig.1v. But it has puzzled

1 Loc. cit., p. 128.

NOTES ON ECHINODERM MORPHOLOGY. 315

me considerably, as he has omitted to letter the two plates
next to the radial primary of the ray which he figures. I
have marked them (4), regarding them as the primaries of the
adjoining rays. But from his remarks on p. 128, which I
have just quoted, it seems possible that they may be his adaxial
interradials, i.e. the true basals, which appear before the
radial shields, and not after them, as I have—perhaps erro-
neously—supposed. On the other hand, he states on the
same page that a third ring of interradials “ arises between
the last-formed interradials and the periphery of the dorso-
central.”” Which does he consider as representing the Cri-
noidal basals? The second, or the third ring which is the
most adaxial? There is not a single reference to his figures
in the whole course of his description on p. 128 of the develop-
ment of the plates in the interradial areas, and I have there-
fore found much difficulty in understanding it. I say this
because I wish to apologise in advance if I have misinter-
preted his statements.

The under-basals, which are formed comparatively early in
the European Amphiura squamata, do not seem to appear
in the American variety till the radial shields have reached
some size. If we take the number of adambulacral plates as
a criterion, we find that the European form with three of these
plates (fig. 111), has five basals, two under-basals, and two inter-
radials,! but no radial shields ; while the American form at the
same stage has small shields and large interradials, but neither
basals (?) nor under-basals (fig. tv). The differences in the
sequence of formation of the apical plates in the two types
may be summarised as follows:

EUROPEAN : AMERICAN :
Radials (4). Radials (4).
Dorsocentral (1). Dorsocentral (1).
Basals (adaxial interradials, 3). Abaxial interradials (1.).
Abaxial interradials (1.). * Radial shields (7. s.).
Under-basals (2). 2 Basals (adaxial interradials, 3).
Radial shields (7. s.). Under-basals (2).

1 See note 1 on p. 312.
> It is possible that these two should be interchanged.

316 P. HERBERT CARPENTER.

Curiously enough, this considerable difference in the order
of formation of the principal apical plates in the American
and European varieties of one and the same species does not
seem to have attracted Fewkes’s attention. Had it done so, I
cannot but think that several passages in his memoir would
have been differently expressed. Thus, for example, in discuss-
ing the nature of the dorsocentral on p. 123, he says: “ If,
however, in Echinoids this plate forms before the ocular and
genitals, and in Amphiura after the same, one is tempted to
ask whether they are homologous. One might, of course,

avoid the difficulty by the truism that the relative time of.

development is of little consequence, and that the appearance
of the plate in Amphiura is simply retarded. Such an escape
from the difficulty does not give much satisfaction, even if we re-
member the abbreviated development of Amphiura.” But since
the under-basals (and basals?) appear after the radial shields
in the West Atlantic, and before them in the Mediterranean,
the late appearance of the dorsocentral in Amphiura as
compared with the Urchins, is no argument whatever against
the view that this plate is homologous in the two groups; and
in fact if the relative time of appearance of the Apical Plates
is to be taken as a criterion of homologies, it is scarcely worth
while for us to attempt to arrive at a general understanding of
the Apical System of Echinoderms.

Another point of the same nature is the discrepancy between
the descriptions given by Fewkes and Ludwig respectively of
the relative times of formation of the radials and terminals.
Ludwig! thinks it probable that the terminals appear before the
radials; while Fewkes” believes the reverse to be the case,
from the comparative sizes of the plates, though he admits
that he has never seen a young Amphiura “with radials
and without terminals.” But if we may judge from the
analogy of the under-basals and radial shields it would appear
that both observers may be in the right.

This is still more probable with regard to the development

1 Loe. cit., p. 187.
2 Loe. cit., p. 139.

Y

NOTES ON ECHINODERM MORPHOLOGY. 817

of the upper arm-plates. Fewkes says on p. 145: “Iam led
to suppose that the dorsals have been inadvertently omitted in
certain of the figures of a young Amphiura by Ludwig (pl.
xi, figs. 21, 25), for he has not represented these plates in a
young specimen in which three pairs of side arm-plates are
represented (pl. xi, fig. 21, ad*., ad*., ad°.).1 In a young
Amphiura of about the same age (pl. iii, fig. 19)? at least one
dorsal plate is formed, and in another as old as that repre-
sented in his fig. 25 (same plate) the dorsals have increased in
number.’ In none of Ludwig’s figures are dorsals represented,
though in figs. 21, 25, they must have been already formed.”
Fewkes makes substantially the same statement in his con-
cluding summary on p. 147, ‘ Dorsals are omitted in all
Ludwig’s figures of the arm from the abactinal side. My
figure is younger than his (pl. xi, fig. 21), in which a dorsal
ought to be represented.”

It was, however, expressly noted by Ludwig* that “In
Stadien welche nicht alter sind als das in fig. 21 gezeichnete,
sind noch gar keine Dorsalplatten vorhanden, obgleich schon
drei freie Armglieder angelegt sind.”

This passage must have altogether escaped Fewkes’s notice,
or he would otherwise have scarcely have hinted at an inad-
vertent omission on the part of Ludwig, or have written so
positively as to what plates ought or ought not to be repre-
sented in a particular developmental stage ; while he makes no
reference on his own part to the differences in the time of
appearance of the radial shields, which are revealed by a com-
parison of his own observations with those of Ludwig, a fact
which may eventually turn out to be of very considerable
interest.

‘ This is copied as fig. 111 of the present communication (p. 312).

2 Fig. Iv, on p. 314.

There is something wrong about this comparison, For Ludwig’s fig. 25
represents a younger and not (as Fewkes implies) a later stage than fig. 21.
There are not likely to be more dorsal plates developed in a form with two
adambulacrals (fig. 25) than in one with three (fig. 21).

4 Loc. cit., p. 190.

VOL. XXVIII, PART 2.—NEW SER, Y

The Photospheria of Nyctiphanes Norvegica,
G. O. Sars.

By

Rupert Vallentin
and

J.T. Cunningham, B.A.,
Fellow of University College, Oxford.

With Plate XXIII.

Ir has long been a familiar fact to zoologists that in nearly
all the Schizopodous genera which form the family Euphaus-
lide are present ten small globular, reddish organs, having a
resemblance in many respects to such eyes as those of Verte-
brata and some Mollusca and Chetopoda. The organs in
question have not the least resemblance to the typical com-
pound eye which is so characteristic of Arthropoda, but
because whe one of them is examined in the fresh state, a
doubly con x structureless lens, and a cell layer suggesting
the idea of « retina, are easily seen, therefore the organs were,
until a recent date, generally denominated accessory eyes, or in
German “ Nebenaugen.”

One of the earlier descriptions of the organs is in a paper
by Claus! published in 1863. He says there that one of the
conspicuous ch racters of Euphausia is the presence of acces-
sory eyes, some of which are median unpaired, some lateral
and paired. In proceeding to give an account of the structure

1 “Ueber einige Schizopoden und niedere Malacostraken Messina’s,”
‘Zeit. f. wiss. Zool.,’? Bd. xiii, 1863.
VOL, XXVIII, PART 3,—NEW SER. Z

320 RUPERT VALLENTIN AND J. T. CUNNINGHAM.

of these “ eyes,’ Claus points out that they had previously
been seen by Dana! and Semper,” the latter of whom judged
them to be eyes, and by Kroyer,? who found them in a form
which he named Thysanopoda inermis, and thought they
were probably auditory organs. The species in which Claus
examined the organs was one occurring abundantly at Messina ;
he named it Euphausia Miilleri. He states that the organs
were present on the basal joint of the second and seventh pair
of thoracic appendages, and between the two members of each
of the four anterior pairs of abdominal swimming feet, so that
they were eight in number, two pairs and four median. They
were reddish, pigmented, cylindrical bodies, whose immediate
relation to the ganglia of the nerve-cord testified to their
importance as sense organs. Although Claus did not succeed,
notwithstanding earnest and repeated endeavours, in tracing
out the nerve-endings, he concluded from the whole structure,
and from the presence of lenses and muscles which rolled the
organs to and fro, that in all probability they were movable
organs of vision. The structure of all the organs was similar.
Each buib lay in a hemispherical protuberance of the body
covering, fastened by slender threads, and movable by several
oblique muscle-strands. The external wall of the bulb was
formed by a cuticular envelope to which the threads and
muscles were fastened, while the internal parts were more
complicated. The front part of the contents’ usisted of a
transparent kind of vitreous body, bounded behind by a
glistening ring containing a lens. Bebind the lens, followed in
the centre of the “eye,” a likewise highly refractive striated
body composed of a number of closely packed rods. This
body was enveloped in a clear spherical layer, whose posterior
half fitted into a course pigmented fibrous membrane. This
latter was in immediate contact with the wall of the bulb, and
had the form of a hemispherical pigmented cup, open in front ;

t «Expl. Exped. of the United States,’ “ Crust. I,” p. 639.

2 « Reisebericht,” ‘Zeit. f. wiss. Zool.,’? Bd. xii, 1862.

3 * Forsog til en monog. Fremst. af slaegt. Sergestes,” ‘ Kon. Dansk. Vid.
Selsk. Skrifter,’ p. 294, 1859.

ee

PHOTOSPHERIA OF NYCTIPHANES NORVEGICA. 821

it held the position of a choroidea with regard to the trans-
parent nucleated sphere. Claus could say nothing as to the
significance of the transparent layer with its bundle of rods,
but believed it to be the percipient element, although he could
not trace any relation between it and nerves.

The pair of similar organs which are present behind the
facetted cornea in the eye-stalks in the adult, escaped Claus’s
notice ; but in the young larve he mentions a bundle of rods
behind each eye, which were obviously the anterior pair at an
early stage of development, though he did not recognise the
identity. Indeed, he was not likely to do so, for he could not
have expected to find an accessory eye in immediate contact
with the principal eye.

His description is illustrated by small figures which are
drawn from the appearance of the organ in the fresh state in
situ, and accordingly do not adequately exhibit the minute
structure.

The species Nyctophanes norvegica, on which our ob-
servations were exclusively made, was first defined by Michael
Sars in 1863! as Thysanopoda norvegica, and his diagnosis
of the red spherical organs agrees exactly with the account
given of them by Claus in ‘Kuphausia Miilleri:’ ‘ Organa
in ventre existunt octo sensitiva (haud dubie oculi simplices)
spherica, cornea transparente semiglobosa, ceterum laete
purpure pig tata, intus lente discreta crystallina len-
ticulari.”

It is surprising that in the numerous works published between
1840 and 1880, in which species belonging to the Euphausiidz
are described and their accessory eyes meutioned, no allusion
is made to any emanation of light from the animals, or to the
possibility that the supposed additional eyes are really phos-
phorescent organs, notwithstanding the fact that J. Vaughan
Thomson, long before 1840, described his discovery of a
Schizopod which was brilliantly luminous. The third memoir
of that author’s zoological researches is entitled, “On the

‘Om Slaegten Thysanopoda, og dens Norske Arter, Christiania Vidensk
Forh., 1863.

O22 RUPERT VALLENTIN AND J. T. CUNNINGHAM.

Luminosity of the Ocean, with descriptions of some remarkable
species of Phosphorescent Animals, and particularly of the
four new genera—Noctiluca, Cynthia, Lucifer, and Podopsis of
the Schizopode.” In this memoir he relates how he captured
the species he calls Noctiluca in the Atlantic, on his voyage
home from the Mauritius, and says the animal gave out bril-
liant scintillations in the dark when disturbed. It was in most
parts perfectly transparent, but turgid here and there, with
orange-red pigment, particularly on its anterior feet.

Thomson identified his specimens with a form described by
Sir Joseph Banks, and observed by him to be luminous, and to
be the chief cause of the phosphorescence of the sea on a cer-
tain occasion. Banks gave the animal the name Cancer
fulgens, which Thomson altered. The latter was quite
unaware that the luminosity was due to and limited to special
organs, and makes no mention of any ‘‘accessory eyes,”
although there is little doubt that hig Noctiluca (which has not
yet been identified with one of the species now recognised)
belonged to the family Euphausiide.

The naturalists on board the ‘“ Challenger” were familiar
with the phosphorescence of the Euphausiide, and observed the
connection between the emission of light and the organs
known as accessory eyes.

Mr. John Murray paid particular attention to animal phos-
phorescence and to the captures of the tow-net, and therefore
was repeatedly impressed with the brilliancy of the luminosity
of members of this family, and its exclusive origin in the
special organs.

The account given in the ‘ Narrative of the Cruise of the
“ Challenger” ’ (vol. i, p. 748) is as follows:

“The phosphorescent light emitted by the species of
Enphausiide was frequently under observation during the
cruise. If one of these be taken up by a pair of forceps when
newly caught, a pair of bright phosphorescent spots will be ob-
served directly behind the eyes, two other pairs on the trunk,
and four other spots situated along the median line of the tail.
These can all be quite well seen by the naked eye. The pair

SS

PHOTOSPHERIA OF NYCTIPHANES NORVEGICA. 323

close to the eyes are first and most brilliantly illuminated, and
then the light, which is bluish white, spreads to the other
organs on the trunk and tail. After a brilliant flash has been
emitted from the organs they glow for some time with a dull
light. The light is given out at will by the animal, and usually,
but not always, when irritated. Subsequent flashes become
less and less bright till the animal appears to lose the power
of emitting light. If the organs be removed with the forceps
the points will glow brightly for some time, and when the
animal is dying the whole body is frequently illuminated by a
diffused light. These phosphorescent organs appear under the
microscope as pale red spots, with a central, clear, lenticular
body. The phosphorescent light comes from the red pigment
surrounding the lenticular space.

“In August, 1880, Mr. Murray observed at night on the
surface of the sea in the Farée Channel, large patches and
long streaks of apparently milky-white water. The tow-nets
caught in these immense numbers of Nyctiphanes (Thysa-
nopoda) norvegica, M. Sars, and the peculiar appearance
of the water seemed to be due to the diffused light emitted
from the phosphorescent organs of this species.”

Professor G. C. Sars mentions the light producing function
of the organs in his “ Preliminary Notices of the Schizopoda
of the ‘ Challenger,’” published in 1883,' and in his “Report on
the Schizopoda ” (1885), he discusses at some length their struc-
ture and function in his account of the genus Euphausia. He
believes, after careful examination of the structures both in
spirit specimens and in the living animal, that they are highly
differentiated luminous organs. It is unnecessary to repeat
his description of the position of the organs, it confirms that
already quoted from Claus, with the addition of the pair of
organs situated in the eye peduncles and mentioned in the
“ Challenger” narrative. He says the organs are globular
bodies with a very complicated structure, bearing in some par-
ticulars great resemblance to that of the eyes in Vertebrates.
A rather thick and elastic cuticle forms the outer envelope of

1 Christiania Videns. Selsk. Forh., 1883.

O24 RUPERT VALLENTIN AND J. T. CUNNINGHAM.

the organ, which, moreover, in fresh specimens is coated with a
beautiful red pigment in its posterior half, whereas the front
portion remains quite pellucid. On closer examination the
two portions are found to fit as it were into each other without
being actually connate. At the junction a glistening ring may
be seen internally, encompassing in the middle a highly refrac-
tive lenticular corpuscle. The posterior hemisphere is filled
up with cellular matter, in the midst of which lies embedded a
flabelliform bunch of exceedingly delicate fibres, exhibiting
in fresh specimens a most beautiful iridescent
lustre. “Tothe equatorial zone of the organ, moreover, two
or three thin muscles are attached, admitting to a certain
extent of its being rolled to and fro.”

Sars then refers to the observation of J. Vaughan Thompson!
that these Crustacea are highly luminous at night, and states
that he himself has observed their luminosity in the Norwegian
species, doubtless principally in Nyctiphanes norvegica
and he found that the animal was able by varying the
movements of the organs to increase or diminish the light
at will. He believed that the chief light-producing matter was
the fibrous fascicle lying in the centre of the organ. ‘‘ Even if
the organ be crushed and this fascicle be extracted it still con-
tinues to give forth a comparatively strong phosphorescent
light when seen in the dark.” The lens he believes to act asa
condensor for the light, and the pigment coating behind to
prevent the light being radiated in all directions. He gives
four reasons which show that the organs are not eyes; (1) that
the nerve to the organs is very thin and does not give rise to
any special retinal expansion; (2) that the structure of the
posterior portion of the organ is not that of an eye; (38) that
the position of the orgaus is ill adapted for vision, and (4) the
presence of one of the organs in the eye peduncle.

He points out that the ocular organ is immobile, and entirely
lacks the front hemisphere with its lens, and that the light from
it is more intense and more steady than that from the others.

Professor Sars’s description of the position of the organs

? Quoted by mistake as W. Thomson,

ole —>

PHOTOSPHERIA OF NYOTIPHANES NORVEGICA. 825

on the body leaves nothing to be added; it is necessary here
simply to indicate their position briefly. The organ in the eye
peduncle is dorsal to and somewhat on the outer side of the
eye, and in close contact with the latter. The thoracic organs
are contained in the coxal segments of their respective limbs ;
the anterior pair are on the internal side of the limb, and cannot
be seen in the ordinary condition and position of the animal ;
they are directed downward and inward when the animal is on
its ventral surface. The posterior pair are on the external side
of the limb, and are directed downward, outward, and backward.
This pair are quite conspicuous in the living animal, and in the
dead specimen without any operation being performed on it.
The abdominal organs are simply directed downwards.

The position of the organs is constant throughout all the
genera of Euphausiide with two exceptions. In Stylocheiron
there are only five, one ocular pair, the posterior thoracic pair,
which have an additional lens, and a single caudal organ. In
Bentheuphausia, which is the only one brought up by the
“ Challenger ” from deep water (LO00—1800 fathoms), accord-
ing to Sars there are no luminous organs and the eyes are
imperfectly developed, while Willemoes Soehm believed that
organs he found on each of the thoracic limbs were the
luminous organs.

We have examined, at Mr. Murray’s suggestion, the struc-
ture of the luminous organs, or as they may more conveniently
be called, photospheria, of Nyctiphanes norvegica, G. O.
Sars. Before proceeding to describe the result of our studies
we will say a few words as to the distribution of the species.
It is a distinctly northern form, being absent from the Mediter-
ranean and the warmer parts of the Atlantic. It is abundant
on the west coast of Norway, having, as we have already men-
tioned, been first defined from Norwegian specimens by M. Sars
under the name Thysanopoda norvegica. But in consi-
dering the localities where specimens have been taken, it is
necessary to mention whether the capture was made from the
surface waters or from the bottom. The adult, so far as our
information allows of a decision, lives on the bottom, and never

326 ROBERT VALLENTIN AND J. T. CUNNINGHAM.

swims far from the ground, while the young, up to half or
three quarters the size of the adult, occur abundantly at the
very surface, and at all intermediate depths. As mentioned
above, Mr. Murray found swarms of individuals at the surface
in the Farde Channel, but none of these were full grown, and
very few more than half the adult size. Within the last two
or three years the species has been recognised as common
enough in the Clyde sea-area.

The place where we obtained it in abundance for the
purpose of the examination described in this paper, was a deep
area ninety to ninety-five fathoms from the surface, situated off
Brodick Bay, and running north and south. We captured it
in a shrimp trawl worked from the Steam Yacht ‘“ Medusa,”
of the Scottish Marine Station, and we took large numbers of
living specimens to the small house-boat laboratory called the
“ Ark,” stationed at Millport. Young specimens up to a
quarter of an inch long were taken by Mr. now Professor J. R.
Henderson in the Firth of Forth in 1884, and in June of the
present year we captured a number of specimens of the same
size at the surface by means of the tow-net, in the neighbour-
hood of St. Abb’s Head. The eggs and larve have never
been described, and, as far as we know, never captured. We
are unable to state whether the adult exists anywhere in the
Firth of Forth or its neighbourhood, or in the North Sea. As
the young occur at the surface in the region of the Firth of
Forh, it is natural to conclude that the adults occur at no
great distance; but this inference requires verification. It
is clear that the adult is pretty widely distributed in depths
over, say, sixty fathoms off the north-western shores of
Europe.

It is obvious from the preceding historical survey that no
complete histological analysis of the structure of the photo-
spheria in Euphausiidz has yet been made. All the organs in
our species, except the pair in the eye peduncles, have the same
structure, and fig. 1 showing a vertical transverse section of
the photospherion of the first abdominal segment in situ illus-
trates the following description: the section passes through

PHOTOSPHERIA OF NYCTIPHANES NORVEGICA. 327

the principal axis of the organ. The posterior part of the
organ is bounded by a layer composed of wavy fibres or
laminz, which are generally parallel in direction, but are to
some extent interlaced and anastomosed ; this layer is of con-
siderable thickness, and forms a hemispherical cup open in
front, and in front only; it is perfectly continuous everywhere
behind its free edge, being pierced by no apertures whatever.
This is the layer called by Claus a coarse pigmented fibrous
membrane, and by G. O. Sars a thick elastic cuticle. The
layer is not pigmented except on its internal surface, which will
be referred to later. It is non-cellular, that is, contains no
nuclei nor distinguishable cell-areas. This layer resembles to
some extent a tapetum, and as one of its functions is un-
doubtedly the reflection of light from its internal surface, we
may adopt for it the name used for a similar layer in the phos-
phorescent organs of fishes by Von Lendenfeld, that of reflector.
Covering the exterior surface of the reflector is a flat, mosaic-
like epithelium of polygonal red pigment-cells. The appear-
ance of these, as seen after slight compression in a fresh organ,
is shown in fig. 2.

Red pigment is present in patches and dots in other parts of
the animal’s body, and is everywhere contained in stellate
mesoblastic chromatophores, situated close beneath the epi-
dermis. It is obvious that the epithelial covering of the
reflector is a special development of these chromatophores.
Where the chromatophores are not present the body of the
animal during life is beautifully pellucid. Internal to the
reflector is a layer of large cells, somewhat higher than broad,
and each containing a large nucleus. These cells are, in some
places, in two layers, though the layers are not regularly super-
imposed. The largest cells are near the surface of the reflector,
and smaller ones exist above them in some parts. The
internal surface of the cellular layer is perfectly smooth and
hemispherical in shape, and in the hollow contained by it is a
curious fibrillar mass. The fibrils of this mass are for the
most part straight, and in its external part are perpendicular
to the surface of the cellular layer, while the core of the mass

328 RUPERT VALLENTIN AND J. T. CUNNINGHAM.

consists of straight fibrils in two bundles crossing at right
angles, and other bundles in other directions. In front of the
fibrillar mass are seen one or more flat cells which belong to
the cellular layer. In front of these is the biconvex lens (/.,
fig. 1), perfectly homogeneous in structure, and highly refrin-
gent; its diameter exceeds that of the fibrillar mass, so that it
rests on the edges of the cellular layer. In front of the lens
is a layer of cellular tissue, which contains a ring of circular
fibres, running round the edge of the lens. The cells of this
layer, which may be called a cornea, are much smaller and
more regular than those of the posterior cellular layer.
Between the fibrous ring and the posterior part of the organ,
outside the lens, is a kind of cleft occupied by a few small
cells, which separate the ring from the anterior edge of the
reflector.

The red pigment of the cells coating the reflector disappears
in spirit, and the peculiar colour of the internal surface of that
layer cannot be seen in sections. The reflector and the fibrous
ring absorb carmine very slightly, and the only deeply-stained
parts of the sections are the cell nuclei. The organ is situated
immediately below the epidermis (ep., fig. 1), and connected
with it by cellular strands, which, passing in at the cleft be-
tween the edge of the reflector and the fibrous ring, place the
posterior cellular layer in structural continuity with the epi-
dermis. It is possibly by this cellular communication that
nervous impulses are conveyed to the organ. The organ is sur-
rounded by a blood space (dac., fig. 1), and the cellular strands
cross this space. In the thoracic organs there are also thin
bands of muscle passing across the space to be inserted in the
surface of the organ, but we have not found such muscles con-
nected with the abdominal organs. The subneural artery
passes between each of the abdominal organs and the pair of
nerve ganglia which lies close to it dorsally.

The structure of the photospherion of the ocular peduncle
differs considerably from that of the rest. The difference con-
sists in the absence of the lens, and of the cornea as a separate
and distinct layer. The organ (fig. 3) is continuous with the

3
|
.
.
|

PHOTOSPHERIA OF NYCTIPHANES NORVEGICA. 329

epidermis, from which it projects inwards in the form of a
somewhat elongated knob. The principal axis of the organ is
oblique to the external surface. The organ is placed imme-
diately outside the tissues of the compound eye, a layer of
dense pigment separating the reflector from the ocular elements,
as is seen in fig. 3, representing a longitudinal section through
the photospherion and part of theeye. The reflector is present,
and has the same structure as in the other organs, but it extends
nearer to the external surface, its free edge being immediately
beneath the epidermis. As in the other organs, it is coated ex-
ternally by a mosaic of polygonal red pigment-cells. Internal
to the reflector is the posterior cellular layer, also having the
same structure as in the other organs, but becoming at its
edge continuous with the epidermis. Internal to the cellular
layer is a layer of straight fibrils. This layer is of a uniform
thickness, equal to that of the cellular layer, and it is limited
internally by a hollow surface. The hollow is filled up by a
mass, which is almost homogeneous at the base near the surface
of the fibrillar layer, but passes by gradual transition into a
layer of somewhat elongated cells, which themselves are con-
tinuous with the epidermis.

There is sometimes visible in the section a blood space
between the exterior of the organ and the surrounding tissues,
but we have never seen any muscles or cellular strands like
those occurring in the other organs.

It is to be noted that, with the exception of the layer of
straight fibrils and the reflector, every layer in the photosphe-
rion of the ocular peduncle is continuous with the epidermis.
It seems a necessary inference from this that the organ is pro-
duced by differentiation of parts from a simple thickening of the
epidermic layer of cells. The reflector is probably a specialisa-
tion of subepidermic mesoblastic tissue ; the posterior cellular
layer is a specialisation of the deepest portion of the epidermic
thickening, and remains continuous with the epidermis at its
free edge. Similarly, the central portion of the thickening is
modified into the almost homogeneous mass filling the cavity
enclosed by the layer of straight fibrils, This layer itself is

330 RUPERT VALLENTIN AND J. T. CUNNINGHAM.

also, in all probability, produced by differentiation of other cells
of the thickening.

It seems also clear that a few steps farther in the same pro-
cess of molification would produce an organ exactly like the
photospheria in the thorax and abdomen, and that we are jus-
tified in regarding the organ of the eye peduncle as permanently
retaining a condition which in other organs is merely a stage
of development. The more complicated structure of the other
organs is probably derived from the condition seen in the
ocular organ in the following manner:—The edges of the
reflector are turned somewhat inwards; the layer of straight
fibrils is compressed into a solid mass with slight modifications
in the arrangement of the fibrils in the centre of this mass.
The homogeneous mass in the hollow of the fibrillar layer in
the ocular organ may be taken as representing the lens, which,
when the fibrillar layer was compressed, would be removed
farther outwards, and lie in front of the fibrillar mass. This
supposition involves the view that the lens in the posterior
organs is derived from the bodily conversion of cells, and not
by extracellular secretion. But it may be a “ cuticular”
structure deposited by the cells in front of it. The cells which
in the ocular organ lie between the fibrillar layer and the epi-
dermis obviously only need to separate from the epidermis and
form a distinct cap in order to become the cornea of the more
complicated organ, while the fibrous ring at the base of the
cornea seems to be produced by the transformation of some of
the cells of the cornea, for in sections nuclei are often to
be seen in the substance of the ring. The separation of the
cornea would, of course, push the posterior part of the organ
into a deeper position in the body, and separate the edges of
the cellular layer to a great extent from the epidermis. We
say to a great extent, for it is to be borne in mind that in the
posterior complicated organs the cellular layer remains con-
nected with the epidermis by cellular strands. It is probable
that the nervous stimulation of the organ reaches it by these
cellular strands.

We have not been able yet to follow out from the earliest

PHOTOSPHERIA OF NYCTIPHANES NORVEGIOA. 331

stage the actual development of the organs, and so test the
above hypothetical account, but we have examined the con-
dition of the posterior organs in very young specimens, which
were only a quarter of an inch in length, but which were
already similar in external characters to the adult. In these
young specimens the organs already possess all the different
parts present in the adult organs, but in asomewhat embryonic
condition. A section of one of these organs is shown in fig. 5.
It is from one of the anterior thoracic pair. It will be seen
that the continuity of the cornea and of the posterior cell
layer with the epidermis is well shown, and the whole organ is
evidently arising by differentiation in an epidermic thickening.
The fibrillar mass is different in appearance from that in the
adult, the striations being all parallel to the principal axis of
the organ. The cells of the posterior cell layer still retain
their primitive character. Fig. 6 shows asimilar section of the
first abdominal organ at the same stage.

Function of the Organs.

Nearly everybody who has written about the luminosity of
the Euphausiide has mentioned that the emission of light is
intermittent, and is directly affected by stimulation. It is
usually also stated that the activity of the photospheria is under
the control of the will of the animal, and is soon exhausted by
repeated exercise. Our observations go to confirm these state-
ments, but we regret to say that they are not altogether con-
clusive, and it is desirable that experiments more rigidly exact
should be made on the living animals. The observations are
as follow:

The behaviour of the animals when alive, and, as far as can
be judged after capture, in a healthy normal condition, is pecu-
liar. They are in a state of almost incessant activity, swim-
ming restlessly forwards and struggling vigorously against any
obstacle they come into collision with. Their motion is due
almost entirely to the limbs, the sudden backward movement
produced by flexion of the abdomen is scarcely ever observed.
And it is curious to note that they are as often on their dorsal

302 RUPERT VALLENTIN AND J. T. CUNNINGHAM.

as on their ventral surface when swimming, or perhaps oftener
in the latter position. This is of some importance in relation
to the ventral position of the luminous organs. The contrast
between the apparently excited, hurried, heedless motions of
Nyctiphanes, and the graceful, wavy movement of Mysis,
which is always either suspended in perfect balance with a
regular motion of its limbs, or escaping by a swift, well-
directed, backward dart, is very striking.

In total darkness the animals swimming about in a glass jar
of sea-water gave out short flashes of light from time to time.
Each flash was of short duration, but sometimes lasted longer
than at others; when several animals gave out light simul-
taneously or in rapid succession, the effect was very brilliant
and beautiful, but nothing like continuous luminosity was ever
observed.

Handling.—When the hand was plunged in the water
among the animals, any one of them when touched immediately
gave forth a flash. When an animal was caught and removed
from the water between the finger and thumb, all the organs
emitted a brilliant light for five to ten seconds, while the
creature was flapping its abdomen vigorously and trying to
escape. Then followed an interrupted series of flashes lasting
ten seconds more, and then the animal would become quiet and
no light could be seen. But when slight pressure was adminis-
tered, all the organs flashed again, the duration of the flash
being longer when the pinch was stronger. When the animal
was crushed between the fingers and the tissues rubbed
between the hands, certain particles were luminous and re-
mained continuously so until they were dry. When an
organ was dissected out from the abdomen the light ceased,
and by the time it was mounted and placed under the micro-
scope all luminosity had vanished. But when the organ under
the microscope was crushed the field was lit up and continued
so for some time. When an eye-stalk was cut off by scissors
the ocular organ became luminous for an instant when the
division took place. After the animals had been in captivity
twenty-four hours, they were by no means so easily excited to

PHOTOSPHERIA OF NYOTIPHANES NORVEGIOA. 339

give out light. Only about one in four became luminous on
being removed from the water.

Chemical Stimulation.—When an animal was dropped
into a saturated solution of bichloride of mercury all the
organs shone most brilliantly from five to seven seconds, when
the muscles were being violently exerted ; in one case the light
lasted for thirty seconds. A similar result followed immersion
in nitric acid, 4, per cent., but death ensued more quickly.
In both cases the posterior organs ceased to shine first, and the
ocular organs were the last to be extinguished.

One of us spent nearly a whole day in the laboratory
examining the fresh organs with the microscope in order to
ascertain which part of the photospherion produced the light,
and the results of this examination were afterwards verified by
both of us. It was found, by repeated trials, that it was occa-
sionally possible by crushing the organ under a cover-glass to
separate all the component layers from one another. The red
pigment was usually dispersed in the operation. All other
parts of the organ were seen to be perfectly transparent,
including the greater part of the thickness of the reflector, but
excepting the internal surface of that layer. That surface in
transmitted light glowed with a beautiful luminous-looking,
rosy-purple colour, reminding one of a sunset tint. When the
light from below was cut off and the preparation viewed by
reflected light, the colour was changed to its complementary
tint, namely, yellowish-green. The appearance lasted as long
as the preparation remained moist, over half an hour, but as
time went on the purple colour became more and more tinged
with blue, and the complementary colour more distinctly
yellow. The appearance of the most successful preparation is
faintly indicated by fig. 7. The most striking fact about the
matter was that when the light from the mirror was shut off,
and the preparation viewed through a low power, illuminated
only by diffused daylight, every part of the preparation was
invisible except the coloured surface of the reflector, which
appeared to give off a green light in a dark field. But when
the daylight was entirely cut off by a cloth coat placed over the

334 RUPERT VALLENTIN AND J. T. CUNNINGHAM.

microscope and the head of the observer the green colour dis-
appeared, and nothing could be seen. It is evident then that
the inner surface of the reflector possesses in a marked degree
the property of fluorescence, and the appearance described is
due to this property, not to a pigment. When white light
falls upon the surface in question its more refrangible rays are
made less refrangible and given off as a greenish light, while
the purple colour seen by transmitted light is due to the
absorption of the blue and green rays, and the transmission of
the remainder.

As stated above it appeared from the examination under the
microscope that the surface of the reflector gave off no light
under the conditions described, in perfect darkness. But we
afterwards fouid that when a slide containing a crushed organ
was viewed by the naked eye in the dark there was always a
luminous spot in the preparation, which shone with an intrinsic
light, and on examination the luminous portion always proved
to be the reflector, whose inner surface gave off the light. It
is to be noted that the colour of the surface of the reflector
was seen equally well, whether the surface was viewed directly
or seen through the thickness of the reflector, for the latter
was transparent.

It was stated above that when an animal was completely
crushed by rubbing between the hands certain particles in the
scattered tissues were continuously luminous in the dark. It
was easy to pick up one of these particles with the forceps,
place it on a slide and examine it with the microscope; this
we did repeatedly, and always found that the particle was the
whole or a portion of a reflector, which was always purple by
transmitted light.

It is evident that our results differ from those of Sars with
regard to this question of the light-producing part of the
organ. We found no evidence whatever that the central mass
of straight fibrils gave out light, and it never showed any
colour phenomena: it was always transparent and colourless.

When the organs are examined in the living animal in day-
light, the interior of each is seen to have a yellow-green

PHOTOSPHERIA OF NYOCTIPHANES NORVEGICA. 335

glistening sheen. This is doubtless due to the surface of the
reflector, and is the same thing as the yellowish-green colour
possessed by that surface in a crushed organ in reflected light
under the microscope.

Unfortunately, the observation of the field of the microscope
being lit up when a fresh organ was crushed beneath the
objective in the dark, was not repeated after the above dis-
coveries were made concerning the reflector; and, therefore,
we cannot at present say what relation the above-described
properties of the reflector have to the emission of flashes of
light by the living animal. It is certain that these flashes are
more intense than the continuous light given out by the
reflector after the organ has been crushed. But the question
which demands an answer is this: Is the sudden flash due to
a sudden intensification of a phosphorescence always existing
in the surface of the reflector? or is the sudden flash produced
elsewhere (perhaps in the posterior cellular layer or in the
central mass of fibrils), and the fluorescence of the reflector
merely a property accessory to its principal function of
reflecting the light so produced ?

In attempting to understand the mechanism of the photo-
spheria of the Euphausiide it is natural to endeavour to get
some enlightenment from a comparision between these and
luminous organs in other animals. The best known luminous
organs are those of the glowworm, and their structure and
mechanism have been investigated by Max Schultze! and
others. Schultze says that researches previous to his own
have shown that oxygen is necessary to the emission of light,
and that the intensity of light is under the control of the
nervous system, while all attempts to isolate phosphorus from
the organs have failed. He then proceeds to describe the
structure of the organs, which consist of plates composed of
two layers of cells. The deeper layer is opaque and non-
luminous ; its opacity is due to granules containing uric acid,
and probably consisting of urate of ammonia, and the layer

1 « Teuchtorgane von Lampyris splendidula,” ‘Arch. f. mik. Anat.,’
Bd. i, 1865.

VOL, XXVIII, PART 3.—NEW SER. er

336 RUPERT VALLENTIN AND J. T. CUNNINGHAM.

acts merely as a reflector. The anterior layer is composed of
pellucid polygonal cells, among which are numerous trachez
whose ultimate ramifications end in stellate cells. Schultze
found that the tracheal end-cells in the fresh condition reduced
osmic acid very powerfully, ‘and he concludes that they have a
great affinity for oxygen, and that they are thus the principal
agents in the production of the light, though the other cells
of the layer are also luminous. At the same time he points
out that tracheal end-cells occur in other organs besides the
light organs, and wherever they occur powerfully reduce osmic
acid. There is obviously at first sight very little agreement
between the light organs of the glowworm and the photo-
spheria of Nyctiphanes. If we compare the reflectors, we find
that of the former, cellular; that of the latter, fibrous. Perhaps
the posterior cellular layer of Nyctiphanes is similar to the
superficial cellular layer in Lampyris, but then we have not
yet proved that the former is luminous.

We next have to compare the photospheria of Nyctiphanes
with the luminous organs of fishes, which have been carefully
and ably investigated by R. von Lendenfeld and Professor
Moseley.! There is nothing in these organs of fishes which
resembles the structure seen in Nyctiphbanes at all closely. In
the fish the phosphorescent organ usually contains structures
of two kinds, a system of glandular tubes and a layer of super-
ficial specialised epithelial cells. The glandular portion is
always the deeper, the epithelial more superficial. Most of the
organs have, in all probability, been developed in connection
with the slime canal system. The nervous supply as a rule
does not present any very striking peculiarity ; there are certain
large suborbital organs which are innervated by an enlarged
branch of the trigeminus having a special lobe at its origin ;
the other organs are supplied by ordinary superficial nerves.
There is usually au enveloping capsule with an internal reflect-

1 ¢ Report on the Deep-Sea Fishes of the “ Challenger,” App. A; ‘ Report
on the Structure of the Peculiar Organs in the Head of Ipnops,’ by Professor
H. N. Moseley, F.R.S., App. B; ‘Report on the Structure of the Phospho-
rescent Organs of Fishes,’ by R. von Lendenfeld, Ph. D., F.Z.S,

PHOTOSPHERIA OF NYCTIPHANES NORVIEGIOCA. 337

ing surface, and this is morphologically a modified dermic
scale. There is a special kind of cell often present which
Lendenfeld regards as a specialised gland-cell. Each of these
contains a highly refractive vesicle which represents the secre-
tion of the gland-cell.

To take one or two examples. One of the most interesting
of the organs described is that called “ ocellar, regular, com-
posite.” Some of these are provided with a reflecting layer
internal to the pigment, composed of calcareous spicules, and
morphologically comparable to one or more scales. These
organs are all formed on one general plan. There isa sphericai
deep portion surrounded externally by pigment, and continued
towards the surface of the animal into a superficial portion,
which is either a paraboloid or a parabola in some of its sec-
tions. The interior of the spherical portion is occupied by
gland-tubes usually arranged radially, and lined by a single
layer of gland-cells. Where the gland-tubes converge there is
a space into which they pour their secretion. In the organs
with reflectors cords composed of blood-vessels and nerves
pierce the calcareous reflecting layer and extend vertically to
the surface. These vertical columns are surrounded by ra-
diating, slender cells, closely packed, so that each column of
nerves and muscles forms the core of a cellular prism. Some
of these slender cells are thin and inconspicuous, but among
them are large club-shaped cells, which contain an oval, highly-
refractive body, the latter apparently consisting of a cavity
with a very fine wall, and containing fluid. The cellular
prisms are separated from each other by a kind of packing
composed of small cells, and the surface is also covered by
layers of these cells. These organs are almost the only ones
from which light has actually been observed to be emitted.
Guppy! saw the light emitted from them in Scopelus. Len-
denfeld thinks that the deeper gland-tubes pour out a slimy
secretion into the distal portion, and that a mutual chemical
action takes place between this slime and the typical phos-
phorescent clavate cells above mentioned, at the will of the fish,

1 Ann. and Mag. Nat. Hist.,’ ser. 5, vol. ix.

338 RUPERT VALLENTIN AND J. T. CUNNINGHAM.

so that light is produced. This is not very clear. In another
place he speaks of the secretion produced by glands in the
lower part of an organ serving as a fuel, at the expense of
which the slender cells above it may produce light. Again, in
the case of the simple ocellar organs, he speaks of a special
phosphorescent apparatus above, which produces light at the
volition of the fish by using up or burning the secretion supplied
by the gland, and stored in the space below.

These views, indefinite as they are, are yet inconsistent with
anything that we know concerning animal physiology. It
seems to me that they are of the same kind as the ancient
notion that the heat of the body was due to the combustion
of carbohydrates in the lungs by the oxygen taken in respira-
tion. But that the emission of light is really dependent on
oxidation in the luminous organs seems to have been con-
clusively proved in the case of the glowworm, in which the
light was observed to disappear when the animal was placed in
a medium destitute of oxygen. Lendenfeld regards the cellular
layer at the back of the photospheria in Euphausiide as com-
posed of gland-cells. I cannot see any particular reason for
this, as I see no evidence of secretion being produced by them.
At the same time it is probable enough that these cells are
really the active agents in emitting light, the fluorescent
surface of the stratified layer being only an accessory adjunct.
All that can at present be said in the way of comparison is
that the cells of this cellular layer in the Euphausiide are similar
in general appearance to the cells of the luminous layer in
Lampyris splendidula, and that some or other of the cel-
lular elements in the luminous organs of fishes are the active
light-producing agents. With regard to the refractive globules
in the club-shaped cells in fishes, it is possible that these really
function as lenses, a number of lenses here perhaps being
more advantageous than a single large lens, such as that of
Nyctiphanes.

It is much to be desired that a careful investigation of the
production of light in definite organs should be undertaken by
a combination of physicists and physiologists, for hitherto the

PHOTOSPHERIA OF NYCTIPHANES NORVEGIOA. 339

most profound physical and physiological knowledge has not
been applied to the problem. We content ourselves, in the
above pages, with having worked out the internal structure of
the organs of Nyctiphanes in detail, and having called atten-
tion to the interesting physical properties of the surface of the
stratified layer.

P.S.—In the above brief review of researches on the phos-
phorescence of otker animals than Euphausiide we have
omitted to refer to the important papers of Panceri. The
principal of these are memoirs published by the Academy of
Naples, namely, on the phosphorescence of the Pennatulide,
1871, on that of Pyrosoma and Pholas, and on that of Phylir-
rhe, 1872. Besides these there are valuable abstracts in the
‘Comptes Rendus of the Acad. Roy. de Naples,’ of the years
1871 and 1872. These abstracts are to be found in a French
translation in the ‘ Ann d. Sci. Nat.,’ tome xvi, 1872, and it is
to this source that we owe our knowledge of Panceri’s views.

The production of light in all the cases investigated by him
is, according to Panceri, exclusively due to a special granular
substance having apparently the nature and properties of a fat.
He found a species of fish, Trachypterus iris, to be luminous
in all parts ; light emanated from almost the whole external sur-
face, from the muscles when they wereex posed, and from theviscera
shortly after the abdominal cavity wasopened. The liquid which
drained from the flesh was lumiuous and gave that quality to
every surface which it covered. This liquid was merely the oil
of the animal, and the light was due to oxygen, being extin-
guished by carbonic acid and intensified by oxygen, not imme-
diately but after an exposure of some hours.

In Medusz Panceri found that the property of luminosity was
confined to the epithelium either of the external or of internal
surfaces, and was due to a substance contained in the cells of
the epithelium, which substance resembled fat.

In Pennatula there are on each zooid eight cords or ridges
on the external surface of the stomach. These ridges are com-
posed principally of a substance of a fatty nature contained in

34.0 RUPERT VALLENTIN AND J. T. CUNNINGHAM.

the cells of which the ridges consist. Light emanates exclu-
sively from these ridges, and in the natural state of the Pen-
natula appears only in consequence of a stimulation ; a single
touch at any point of the colony may cause a luminous current,
so to speak, to run through the whole. But the luminous
matter can be caused to give out light after separation from
the Pennatula by rubbing or by the application of fresh
water.

In Pyrosoma there are two small luminous organs in each
zooid near the external end of the branchial chamber. Each
is composed of spherical cells and is attached to the inner
surface of the outer layer of the integument. The eells contain
a substance soluble in ether.

Pholas possesses several patches of excreting epithelium on
the internal surface of the mantle; the cells of these patches
give off a mucus which is in great part composed of fat and is
luminous.

In Phyllirrhde the cells which emit light are not superficial
but are the ganglionic cells of the peripheral nervous system ;
but the light is due, not to the nervous matter properly speak-
ing, but rather to a substance associated with the nervous
matter, a substance which is soluble in alcohol and ether, and
which gives out light when agitated even after the death of the
animal.

It is evident then that there is no very clear connection
between our observations and the views of Panceri. If, as the
latter seems to think, the property of luminosity is always
confined to a substance of a fatty nature, we connot say in
which part of the photospheria of Nyctiphanes the fatty
substance occurs; we can only repeat that we saw light emitted
only from the surface of the reflector.

PHOTOSPHERIA OF NYCTIPHANES NORVEGICA, 341

EXPLANATION OF PLATE XXIII,

Illustrating Messrs. Vallentin’s and Cunningham’s Memoir on
“The Photospheria of Nyctiphanes Norvegica, G. O.
Sars.”

List of Reference Letters.

et. Cuticle. co. Cornea. ep. Epidermis. fv. Fibrous ring. f.m. Fibril-
lar mass. ga. Pair of ventral ganglia. 7. Lens. Jac. Blood lacuna. p.e.
Posterior cellular layer. re. Reflector.

Fic. 1.—Section of the photospherion of the first abdominal somite iz situ.
The plane of the section is perpendicular to the principal axis of the animal’s
body, and passes through the principal axis of the organ.

Fic. 2.—Excernal surface of the posterior thoracic photospherion, seen after
slight compression in the fresh condition. (Zeiss A, oc. 2.)

Fie. 3.—Section of the photospherion of the ocular peduncle. The plane of
the section passes through the principal axis of the organ.

Fie. 4.—Section of the photospherion of the ocular peduncle perpendicular
to the principal axis of the organ.

Fie. 5.—Section of anterior thoracic photospherion from a young individual
a quarter of an inch long.

Fic. 6.—Similar section of the organ of the first abdominal somite at the
same stage.

Fie. 7.—Fresh preparation of a photospherion carefully crushed beneath the
cover-glass. Transmitted light. (Zeiss C C, oc. 2.)

\i

Fe ee)

ee PT
a ta - > 4

Meior Fournt Ves KUUNSPL IY

22 ‘ €\_@

F. Huth, Lith? Edin®

4 i
g
:
: ‘ a anni
i ‘ Z wed : 7
iL ¥
” .
‘
.
‘
.
‘
ms
Com
4
5 -
-
i
2a
~~

Soa
=

A SOUTH AMERICAN SPECIES OF PERIPATUS. 343

On the Early Stages of the Development of a
South American Species of Peripatus.

By

Ww. L. Sclater, B.A., F.Z.8.

With Plate XXIV.

ConTENTS.
I. Introductory Remarks. IlI. Details of the Development.
II. Structure of the Uterus. IV. Conclusions.

I. InrRopuctory Remarks.

Tue development of the South American form of Peripatus
has been worked at by one author only, Kennel, who brought
his material from Trinidad. In his two published papers (8)
his results are so much at variance with the results arrived
at by Mr. Sedgwick (6), who has worked solely at the South
African form of the same genus, that it seems worth while
to go through the early stages of the South American Peri-
patus again. This I have been enabled to do by means of
specimens of Peripatus brought home by me alive from
Demerara last winter, which have supplied me with a fairly
complete series of embryos from the earliest stages onwards.

The species of Peripatus occurring in Demerara seems to
me to differ materially from the other South American species,
as I have pointed out in a note in the ‘ Proceedings of the
Zoological Society’ (5). But the absence of well-preserved
material makes it very difficult to settle this question with any
degree of certainty. Also it is to be hoped that Mr. Sedgwick,
in a memoir which he is about to publish on the various species

344 W. L. SCLATER.

of the genus Peripatus, will be able to clear up the diffi-
culties of the subject.

It will be convenient, however, for me to distinguish the
species now under investigation from those treated of by
Kennel, and I propose, therefore, to allude to it as Peripatus
imthurni, since it was Mr. E. F.im Thurn who first discovered
Peripatus in British Guiana, and it was through his kindness
and hospitality that I was enabled to procure my specimens.
To Mr. Sedgwick also I owe very many thanks for all his
kindness and help to me in my work on this subject. When
I arrived in England it was he who preserved the specimens
and their embryos, and afterwards helped me with many
suggestions, besides allowing me to use all the many resources
of his laboratory at Cambridge for the prosecution of my
researches. -

II. Srructure or THE UTERUS.

When the uterus of Peripatus is examined it is found to -

consist of a long and very much coiled duct, which commences
at the ovary as a slender tube, and, gradually widening, joins
its fellow to form a short vagina, and opens to the exterior at
the penultimate somite.

The uterus in the lower part is divided up by constrictions,
the space between the constrictions being occupied by an
embryo. Advancing towards the ovary the constrictions are
longer and the swellings are smaller, till, for some distance
from the ovary, the swellings entirely disappear. These swell-
ings mark the position of the various embryos, and are usually
eight to ten in number.

The position of the embryos is also marked by deposits of
pigment, which appear as a pinky-red colour when seen in the
solid uterus ; these patches are found not at the actual position
of the embryo itself but just in front and behind. In one or
two instances, in the case of a very young embryo (?.e. the one
nearest the ovary), the pigment was found to envelop it com-
pletely, so that it seems that the pigment is first formed all

A SOUTH AMERICAN SPECIES OF PERIPATUS. 345

round the embryo, and that it is then gradually divided into
two patches, one in front, the other behind the embryo.

The most noticeable point about the uterus of Peripatus
is that the contained embryos, which are from eight to ten
in number, are all at different stages, the youngest near the
ovary being perhaps in the segmentation stage, while the oldest
will be completely formed and ready to be born.

All my specimens were collected at one season of the year
(i, e. November and December), so I am unable to say whether
they are pregnant all the year round, but it seems probable
that this is the case. In this relation Peripatus imthurni,
as also P. torquatus and Edwardsii, differ from the South
African and New Zealand Peripatus, in which cases the de-
velopment of the embryos, though going on all the year round,
commences at one particular season, so that all the embryos
found in the uterus of the female are approximately of one age.
The structure of the uterus of Peripatus will be best seen by
examining figs. 1 and 5.

Fig. 1 represents a longitudinal section through a piece of
uterus not far removed from the ovary. At e is seen a young
embryo lying in a cavity of the uterus, this cavity, which is the
widened lumen of the uterus, is difficult to trace in front and
behind, though in some places remnants of it can be detected ;
but on the whole it is generally obliterated.

The outer wall of the uterus is formed by a very slightly
differentiated single layer of cells (¢.), which though distinctly
nucleated are without cell walls, so that they cannot be
separated from one another very easily. Within this is the
uterine epithelium, which simply consists of a mass of proto-
plasm in which is embedded a large number of nuclei. Slight
traces of cell walls can be seen in fig. 5, which is a transverse
section of a young embryo with its surrounding uterus. The
inner line of the uterine epithelium is folded, so that it has a
crinkled appearance (ck.) ; this is, doubtless, for the purpose of
increasing the absorption surface, and it is by means of this
crinkled appearance that the boundary between the uterus and
embryo can be always detected. This folding of the inner layer

346 W. L. SCLATER.

of the uterine epithelium is not seen very well in the longitu-
dinal section (fig. 1), but it is better seen in fig. 5.

At either end of the embryo is seen a mass of pigment (pg.),
in the form of black, star-shaped masses, which seem to block
up the lumen of the uterus; around and near them are certain
darkly-staining irregular masses of protoplasm (z), in which no
structure whatever can be distinguished. Between the em-
bryos is seen the very curious vacuolated tissue described by
Kennel, of which itis very difficult to understand the meaning.

This tissue (v. c.) consists of at first a more irregular, after-
wards of a more regular mass of vacuoles, separated from one
another by thin lines of protoplasm which stain but slightly ;
along the outer border of these vacuoles there is a line of
nuclei (z.) which are considerably larger than those of the
ordinary uterine epithelium.

The appearance of this vacuolated tissue in transverse
section is very well shown by Kennel (Part 1, Pl. VII,
fig. 42).

The lines of protoplasm separating the vacuoles run straight
from the central point of the section, where occasionally some
remains of the lumen of the uterus can be seen, to the circum-
ference where the vacuolated tissue comes in contact with the
uterine epithelium. Just at this point, but in the vacuolated
tissue, are found the numerous large nuclei generally arranged
in a single row (z. in fig. 1).

This vacuolated tissue is never found near the embryo;
where the embryo is present the vacuolated tissue is entirely
replaced by the uterine epithelium, as is seen in fig. 1.

This vacuolated tissue is probably simply modified uterine
epithelium, though how the change has been effected I am
unable to suggest.

The explanation of this curious histological structure in the
uterus of P. imthurni seems to me connected with the
entire absence of yolk, and the small size of the ovum of this
form. The difference in size between the youngest embryo (a
sphere measuring ‘04 mm. in diameter) and the fully formed
one, which is often an inch long, is enormous, and there must

A SOUTH AMERICAN SPECIES OF PERIPATUS. 347

be a very large quantity of food material absorbed in order to
account for this increase in size. This food material must
necessarily be derived from the uterine walls, and it appears to
me that it is principally derived from these vacuolated regions,
and that the pigment and the structureless protoplasm figured
x are concerned in this phenomenon.

III. Deratts or THE DEVELOPMENT.

The youngest embryos I have found are all of approximately
the same age and are in the segmenting stage, but owing to
their small size I have never been able to get them, except in
a series of sections of unsplit uterus; and in this case the
embryo is never so satisfactory, owing probably to the contrac-
tion of the uterine tissues due to the action of the reagents.

Fig. 2 represents what I take to be the youngest of all in
the segmenting stages ; it measures ‘04 mm. across. It is not
easy to assert definitely, but it probably contains eight nuclei
embedded in a mass of unsegmented protoplasm, the whole
lying free in the cavity of the uterus, which is always present
where the embryo is, and for a short distance on either side
of it.

In fig. 4, which represents an embryo (‘07 mm. in diameter)
at a rather more advanced stage, the cells—or rather the nuclei,
since there is as yet no sign of any cell partitions—have begun
to arrange themselves in a ring round a central cavity. This
embryo probably consists of twenty-four cells or rather nuclei.
This embryo was also peculiar in that the uterus did not have
the masses of pigment usually found in it on either side of the
embryo. This embryo measures ‘(08 mm. long by :07 mm.
across.

Figs. 3 and 5 are approximately of the same age as fig. 4.

In fig. 5 the uterus has been drawn on one side to show the
usual arrangement of the nuclei of the uterine epithelium ;
towards the embryo traces of cell outlines can be detected, but
in the outer part of the uterine wall the uterus consists simply
of a clear protoplasm with dark staining nuclei embedded in

348 W. L. SCLATER.

it. The nuclei of the uterus stain much more darkly than the
nuclei of the embryo itself.

In fig. 3 the only striking peculiarity is the presence of the
two bodies (p. 0.); it seems possible that these may be polar
bodies, though beyond their appearance and position I have no
further evidence to offer.

Kennel has also figured these early segmenting embryos,
and my results do not differ materially from his; he has also
figured what he believes to be polar bodies, but in no case do
they seem to have separated from the embryo itself, but
remain still buried in its substance.

The next stage, which is represented in fig. 6, presents a con-
siderable difficulty ; it seems to resemble in some respects
fig. 51, Pl. viii, of Kennel.

This embryo (°105 mm. in diameter), which at first I took to
be a vesicle with the embryo inside, must, I think, be
regarded as a stage previous to the one next to be described,
i.e. the pseudogastrula stage.

The embryo consists of a single layer of cells marked out
from one another only by their nuclei; it is approximately
spherical. The nuclei on one side of the embryo, on being
traced out, are found to form a small patch on that side of the
sphere, and are considerably larger and more numerous, and
it is at this spot, I take it, that invagination will take place.

That this embryo must be placed at this point in the series
is evident from the size; as the embryos hitherto described
varied from ‘04 mm. to ‘(08 mm. in diameter, this one (i.e.
fig. 4) is ‘105 mm. in diameter and approximately spherical,
while the pseudogastrula, to be described below, is*112 mm. in
diameter and ‘190 mm. in length; this of course forms a
fairly regular gradation of increase in size.

The ovum, therefore, of Peripatus imthurniis holoblastic,
and the segmentation is fairly regular, the result being a
blastosphere (fig. 6), that is, a hollow vesicle one cell thick.
There is no sign of any attachment of the embryo to the wall
of the uterus, and the inner wall of the uterus is marked by a
thickened and crinkled line, which is very characteristic, and

A SOUTH AMERICAN SPECIES OF PERIPATUS. 349

which is very useful in later stages for determining the
boundary between the embryo and the uterus. The size of the
segmenting ova of Peripatus imthurni varies from ‘04
mm. to ‘07 mm. in diameter, and that of the blastosphere is
‘105 mm. in diameter.

The next stage, measuring ‘112 mm. x ‘190 mm., represented
in figures 7 a and 7 B, is a most important stage; during it the
blastosphere, which was described above, is invaginated so as to
form what appears to be a gastrula. I have several series of
sections of embryos in this stage, but those figured, which are
both from the same embryo, are by far the best; fig. 7 Bisa
transverse section through the embryo in the middle of its
length, showing the invagination; fig. 7 a is a section through
one end of the embryo beyond the point of invagination.

The cells are large and generally fairly well defined, more
especially the outer layer ; in the inner layer it is more difficult
to distinguish cell-outlines; the nuclei of the outer layer show
a distinct reticulum, and those of the inner layer are rather
more chromophilous. The opening of the gastrula is situated
on one side of the vesicle, that is, it opens at right to the long
axis of the uterus.

This stage, which may be called the pseudogastrula stage,
seems to me of great interest, and to be really the key of the
whole matter.

The outer layer of the pseudogastrula forms in later stages
the wall of the embryonic vesicle ; the embryo proper is formed
solely from the inner layer of the pseudogastrula.

Of the significance of this stage, and of its relations to the
mammalian pseudogastrula, I will say more later on in the final
part of the paper.

This stage doubtless corresponds to Kennel’s stage, figured
in Part I, Pl. viii, fig. 56, where he makes the outer wall
of the gastrula w.e., which he interpretes to be uterine epi-
thelium; he also letters the point of invagination as 0., but
neglects to give, as far as 1 can see, any explanation of o.

The result of this is that Kennel throughout his paper con-
siders what I have termed the wall of the embryonic vesicle to

350 W. L. SCLATER.

be part of the uterine epithelium and a purely uterine struc-
ture, whereas to me it seems evident that the wall of the
embryonic vesicle has nothing whatever to do with the uterine
epithelium, but is derived solely from the outer layer of the
pseudogastrula exactly in the same way as the surrounding
layer of the mammal’s blastoderm, which afterwards forms the
chorion, is derived from the outer cells of the Mammalian
pseudogastrula.

From the pseudogastrula stage onward the embryo is always
found lying in a hollow space, the embryonic vesicle, and
attached on one side to the wall of the embryonic vesicle, which
is formed from the outer layer of the pseudogastrula stage.

The embryo during this stage is at first sessile, after-
wards a stalk is formed ; and it is during the formation of the
stalk that the two structures termed by Kennel ammion and
placenta are found.

Figs. 8 and 11 represent the youngest embryos in vesicles
which I have met with.

Fig 11 is drawn from an embryo lying in its vesicle still in
the uterus, the uterus has been slightly split and the object
drawn after being rendered transparent in benzole. The most
noticeable point about fig. 11 is the enormous increase in size
of the vesicle which is represented in the previous stage only
by the slight split between the inner and outer layers of the
pseudogastrula; the embryo, which is seen to form a small dark
patch on one side of the vesicle, is oblong in shape and sessile.
The size of the embryo is ‘16 mm. long by ‘06 mm. across ;
this is approximately of the same size as the embryo itself in
the previous stage; but the vesicle in figure 11 measures *24
mm. across as against ‘11 mm. in the pseudogastrula stage
(fig. 7 B).

Fig. 8 shows a section through an embryo measuring ‘084mm.
in diameter with part of its vesicle of the same age as fig. 11.
The wall of the vesicle consists of a band of clear protoplasm
with definite nuclei at intervals. The embryo itselfis composed
of two parts, the basal part, in which the nuclei resemble those
of the vesicle wall, and the embryo proper, consisting of large

A SOUTH AMERICAN SPECIES OF PERIPATUS. 351

cells in which cell outlines can only with great difficulty be
distinguished.

These embryonic cells have a very peculiar appearance, due,
as it seems, to the diffusion of the nuclear substance or chroma-
tin throughout the cell substance; it therefore follows that
no definite nucleus can be detected. Careful focussing, however>
seems to indicate clear lines of protoplasm where no chromatin
is present, and these lines of clear protoplasm seem them-
selves to demarcate the various cells of which the embryo is
made up.

In the next stage the embryo is still sessile, that is, it is
attached to the vesicle wall along its whole length.

Fig. 12 represents an embryo of about this stage, measuring
about ‘1 mm. across, lying in its vesicle, the vesicle again lying
in the cavity of the uterus ; the two latter have been split open;
so as to expose the embryo.

Figs. 9 a and 9 B represent two sections, an embryo and vesicle,
of approximately the same size and age as fig. 12; the embryo
*14 mm., the vesicle ‘28 mm. in diameter.

The vesicle wall (v. w.) is rather thicker in this stage, and
the individual cells composing it are very much better defined
than in the previous stages. The embryo proper consists, as
before, of large cells with diffused chromatin and obscure cell
outlines, and of supporting cells (sp. c.), whose nuclei stain
very deeply, and whose cell outlines are invisible.

Between the two forms of cells of the embryo there has now
appeared a cavity (0). This I believe to be arti-fact, and due
to the action of reagents, especially as it is very inconstant in
its appearance in embryos of this stage.

The only other noticeable feature of this stage is the so-
called amnion (am.) of Kennel ; this consists in the region of
the embryo of a few scattered nuclei embedded in strings of
protoplasm, in some cases surrounding and fusing with the
embryo, in others fusing with the vesicle wall.

Fig. 9 B represents a section of the same vesicle behind the
embryo proper. Here a complete ring of nucleated protoplasm
is seen surrounding the space where farther forward would be

VOL, XXVIII, PART 3 —NEW SER, BB

852 W. L. SCLATER.

found the embryo. This is the highest development of the so-
called amnion.

The amnion springs from the basal supporting cells of the
embryo, as is asserted by Kennel.

This growth, which I have called an amnion, following
Kennel’s nomenclature, does not seem to me to fulfil the con-
ditions of an amnion at all. An amnion may be described as
a double fold of the non-embryonic area of a blastoderm (=
vesicle wall), which is caused by the sinking of the heavy
embryo into the cavity (= yolk-sac or vesicle) filled with fluid,
the double folds finally fusing at the top.

This definition is true both for the amnion of the vertebrate
and of the insect. In the case of Peripatus the outgrowth is
not a double fold, but a single and thin string of protoplasm; it
cannot possibly be explained by the mechanical descent of the
embryo into the vesicle, since in that case the amnion would
be formed on the other side of the embryo from folds in the
vesicle wall.

It seems, therefore, that this so-called amnion of Peripatus
has no sort of homology or analogy to the true amnion of
insects and vertebrates. As to the use of this structure in
Peripatus, it is at present impossible to dogmatize, but it
seems to me that, like other embryonic organs, it has some
part in the conveyance of nourishment to the embryo from the
vesicle and uterus.

Another embryo—measuring the embryo ‘12 mm. the vesicle
*25 mm. respectively—of about the same size and age as the
one above described, is represented in fig. 10; it is remarkable
for the thinness of the vesicle wall, which consists of a quite
slender string of protoplasm with very few nuclei. This and
several other examples which I have met with, of the same sort
seem to show that the vesicle wall varies much in thickness at
different times.

During the next stage the primary layers begin to form, and
soon after that the legs begin to grow out, and the embryo
begins to assume the form of the adult.

Figs. 13, 14a, and 14 8 represent whole embryos at this

A SOUTH AMERICAN SPECIES OF PERIPATUS. 353

stage. Figs. 14.4 and 148 are drawn from the same embryo,
the latter by reflected light, and made transparent by benzole,
the former by direct light in alcohol; fig. 13 is also drawn with
reflected light.

Figs. 14.4 and 148 represent an acorn-shaped embryo cor-
responding to Kennel’s fig. 12, lying in the unbroken embryonic
vesicle; the band across the embryo and the shaded part of
fig. 14a, from which springs the stalk of the embryo from the
thickened area of the vesicle wall, called by Kennel the pla-
centa. The stalk, which is represented only in fig. 14 B, is drawn
too slenderly ; it should be considerably thicker.

The other embryo (fig. 13) resembles fig. 14 in every way,
except that it has not been removed altogether from the
uterus, of which the split half is stiJl seen attached to the
vesicle.

These embryos both measure *24 mm. across, and the vesicles
are respectively ‘8 mm. and °6 mm. long.

When the embryo has got to this stage the layers begin to
be differentiated. The mesoderm and endoderm are formed
by a proliferation of cells which takes place at what will after-
wards be the hind end of the embryo.

This process is illustrated in figs. 15 a, 15 B, and 15 c, which all
represent sections from different parts of one embryo; fig. 15 a
being at the hinder end and fig. 15 c towards the front end of
the embryo. These figures show the embryo proper alone,
neither stalk nor vesicle wall have been represented.

In fig. 15 a the embryo is seen to consist of a double layer
of long-oval nuclei (ec.), from which are afterwards formed the
ectoderm cells, and from which at one point (p7.) there isa pro-
liferation of cells (en.) filling up the greater part of the cavity
of the embryo; from these cells the endoderm and mesoderm
are subsequently formed; the endoderm cells have a more
granular and more rounded appearance than the ectoderm
cells.

Fig. 15 8 shows the appearance of the same embryo some-
what further forward; here all connection between the endo-
derm and ectoderm is lost, and the endoderm is growing

B54 W. L. SCLATER.

forward in the middle of the embryo at the expense of the pro-
liferating cells behind.

In fig. 15 c, still farther in front, a cavity (mes.) begins to
appear in the hitherto solid endoderm ; this is the first com-
mencement of the enteron.

Farther still in front the endoderm thins out somewhat, so
as to form a narrow thin band encircling the enteron; still
farther it gradually disappears, so that nothing is left at the
extreme head end of the embryo but the ectoderm. From
these proliferated cells the mesoderm is also formed later on.
Another feature of this stage is the thickening of the ectoderm
on one side, or rather, ventrally to form the nerve-cord. This is
not marked at the hind end of the embryo, where the thicken-
ing extends all round, but is more marked at the front end
(fig. 15 c), where on one side the ectoderm is seen to consist
of a double layer (n.c.), on the other of only one; it is from
part of this double layer that the nervous system will be sub-
sequently formed.

The proliferation of cells takes place on the ventral side
of the embryo distally to the stalk, which is attached to the
dorsal side of the embryo at the head end. This proliferation,
therefore, exattly corresponds to the primitive streak of P.
capensis as described by Sedgwick; and all that has to be
conceded is that in consequence of the extraordinary changes
in the early stages due to the small size of the ovum and the
absence of yolk, the blastopore of P. capensis described by
Sedgwick has disappeared from P. imthurni, although its
position is still marked by the primitive streak, which replaces
the primitive streak + the blastopore of P. capensis.

After this stage I have not worked at the development of
this form; for one reason I have not had time, for another
because the later stages seem to me to resemble those of
P. Edwardsii and P. capensis as arrived at by Kennel
and Sedgwick respectively.

After a few words on the so-called placenta and amnion I
will proceed to my conclusions.

The organ described by Kennel as a placenta does not

A SOUTH AMERICAN SPECIES OF PERIPATUS. 355

appear till somewhat later. I shall not call it a placenta,
because it does not seem to me to bear any analogy to the
mammalian placenta.

It seems to me that the best words to express the organs are
the embryonic and vesicular thickening corresponding to the
two parts of the placenta of Kennel (i. e. embryonic and uterine
placentas).

An early stage in its development is shown in fig. 16, which
represents part of the vesicle wall of a stage of the same age as
fig. 15; here it is seen to consist of a large mass of cells (pl.)
formed by the proliferation of the wall of the vesicle, which
under ordinary circumstances consists of a single row of cells
only. This is the vesicular thickening (= uterine part of the
placenta) as distinguished from the embryonic thickening (= the
embryonic placenta), which is merely the swollen part of the
vesicle wall from which the stalk of attachment arises (z.).

The vesicular thickening is found in its fullest development
rather towards the hinder end of the embyro, whereas the stalk
of attachment and its swollen base are at the head end of the
embryo.

The histological structure of the vesicular thickening at this
stage is not remarkable; it consists of a mass of nucleated
cells. The outlines of the cells are very apparent, and their
protoplasm is in parts much vacuolated.

This stage corresponds in age to the embryo last described
(fig. 15), and is the earliest stage at which the vesicle swelling
is of any great size or importance.

Fig. 17 shows a much further development of the vesicular
thickening ; owing to its large size only a small portion of the
vesicle wall (v. w.) is represented.

The histological structure of the vesicular thickening has here
completely changed, all traces of cell walls have disappeared,
the nuclei are darker and more distinct, and the protoplasm
presents a very peculiar granular appearance which I have not
seen elsewhere.

The vesicular thickening seems to be fused to the uterus wall
itself, but of this I am not certain, since in all my sections the

356 W. L. SCLATER.

uterus wall has been in each case separated from the vesicle
wall, and on opening the fresh uterus the vesicle always comes
away by itself.

The meaning of the cells marked f in fig. 17 has puzzled
me; I think it highly probable that these nuclei and their
adjacent protoplasm are food material for the embryo lying in
the vesicle (not shown in the fig.), since the nuclei resemble
nuclei found in the embryo, and the mass of cells, if followed
through adjacent sections, are found to form a patch or cap of
cells lying on the vesicular thickening; the vesicular thickening
itself is seen when followed out to be directly continuous with
the vesicle wall, so that the theory that the patch of cells (f) is
merely a continuation of the vesicle wall, and the vesicular
thickening (p/.) a product of the uterine epithelium, seems to
me untenable.

This vesicle swelling persists till quite the end of the uterine
life of the embryo as a thickening at the hinder end of the
embryonic vesicle.

The placenta in the case of mammals is a vascular plexus
formed by the uterine epithelium, which is in connection with
a vascular plexus formed by part of the embryonic membranes.

In the case of Peripatus imthurni there is certainly, as
far as I have been able to observe, no plexus of blood-vessels at
all; and Kennel, I think, makes no mention of this matter.

But apart from that, since it may be argued that the word
placenta can be applied to any uterine nourishing organ, whether
vascular or non-vascular, the swelling of Peripatus is alto-
gether an embryonic organ, the uterus takes no part in its
formation whatever.

The vesicular thickening of Peripatus is formed entirely by
the proliferation of the cells of the wall of the embryonic
vesicle, which vesicle is originally derided from the outer wall
of what I have called the pseudogastrula, so that it has nothing
whatever to do with the wall of the uterus.

Se

:

A SOUTH AMERICAN SPECIES OF PERIPATUS. 307

IV. Concuvusions.

On comparing the early stages of Peripatus imthurni
as described above with the early stages of P. capensis, of
which we now have a very complete account by Mr. Sedgwick
(6), there will be seen to bean extraordinary discrepancy between
the two forms.

Apart from the South American forms P. Edwardsii and
P. torquatus, which resemble very closely in their anatomy
and development P. imthurni, there is yet one more form,
P. nove-zealandiz, about which a little is known through the
researches of Moseley (4), Hutton (2), and Sedgwick (6), and
at the development of which Miss Sheldon, of Cambridge, is
now working.

A continuous series is made by the size of the ova of these
three forms.

In P. nove-zealandiz the ovum measures 15 mm. long
by 1:0 mm. across; the ovum consists almost entirely of a mass
of yolk, and the segmentation is meroblastic.

In P. capensis the ovum is smaller, measuring only *17
mm., and though there is no yolk and the segmentation is total,!
yet the spongy nature of the ovum, which consists of a very
loose meshwork of protoplasm, clearly shows that the ovum has
in this case only recently lost its yolk, and that with the loss
of yolk a gradual reduction of its size is taking place.

In P. imthurni (as also in P. torquatus and P.
Edwardsii, as shown by Kennel) the ovum is still smaller ;
Kennel gives the size of the ovum of P. Edwardsii as (04 mm.,
and this is also the diameter of my youngest embryo, which I
believe to consist of about eight segmentation spheres; I have
not met with any fully ripe ova among my sections, but I
imagine that the embryo does not increase much in size during
the early stages of segmentation.

In these forms the segmentation, as shown above, is complete,
and there is no appearance of sponginess such as described by

1 Sedgwick, on page 517 of the third part of his paper, would prefer to
term the segmentation of P. capensis meroblastic.

358 W. L. SCLATER.

Sedgwick in P. capensis; nor would one suspect, from the
nature and size of the ovum, that it had been derived from a
meroblastic ovum, and had only comparatively recently lost
its yolk.

The only other instance of a holoblastic and alecithal ovum
which has been derived from a meroblastic and telolecithal
ovum, of which we have any knowledge, is the ovum of the
placental mammals which has doubtless passed through a mero-
blastic and telolecithal stage such asisnow presented by the ovum
of Ornithorhynchus, our knowledge of which is due to Caldwell.

The results on the mammalian ovum of the loss of yolk are
(1) a large diminution of the size of the ovum; (2) total seg-
mentation; (3) the formation of a blastodermie vesicle which
corresponds to the yolk-sac of meroblastic forms, the embryo
proper being formed from only a small portion of the original
mass of segmentation spheres which is attached to one side of
the blastodermic vesicle.

Now, it seems to me that the loss of yolk has had precisely
the same effect on the ovum of Peripatus that it has on the
ovum of placental mammals, i.e. (1) diminution of the size of
the ovum ; (2) total segmentation, and (3) the formation of what
I have termed the embryonic vesicle, which appears to me to
be exactly analogous to the blastodermic vesicle of mammals.

The ovum of Peripatus has a stage directly comparable to
the gastrula and blastopore stage of Van Beneden, by means of
which the embryonic mass proper is separated from those cells
which form the wall of the embryonic vesicle, and it is because
this stage in Peripatus, as in the mammal, has nothing to do
with the true gastrula and blastopore, such as is found in
Amphioxus, that I have termed it the pseudogastrula stage.

It seems to me that this point is one which has never been
sufficiently dwelt upon in morphology, namely, that like results
may be due to the same mechanical cause,! and that it is not

1 Professor Lankester some years ago (‘ Aun. and Mag. of Nat. Hist.’ 1870)
discussed this point and applied the terms ‘‘ homoplasy ” and ‘ homoplastic ”
to such resemblances as distinguished from those due to heredity for which
he proposed the terms “‘ homogeny ” and “ homogenetic.”

+ NA aia Nt le

A SOUTH AMERICAN SPECIES OF PERIPATUS. 309

therefore necessary to suppose that because a certain process
is brought about in the same way in two dissimilar forms, that
there must be some genetic connection between these two
forms. For instance, the same mechanical cause (i.e. loss
of yolk) has brought about remarkably similar results in
two entirely different groups of animals (i. e. mammals
and Peripatus).

A point in the development of Peripatus imthurni, to
which I have not yet alluded, and which I am unable to
explain satisfactorily, is what may be termed the inversion of
the layers.

An inspection of the figures will at once show that the
epidermis of the young Peripatus is apparently formed from
the inner layer of cells, and conversely, that the hypoblast is
formed from the cells that are, morphologically speaking, part
of the outside layer of the embryo. I have not been able to
find an explanation of this inversion, though I have made great
efforts to do so. I think, perhaps, that the so-called amnion
may be concerned in its explanation ; until, however, I am able
to procure more material, I cannot offer any definite explana-
tion of this curious phenomenon.

It may be interesting to note the differences between the
adult Peripati, since they differ so immensely in their
development.

The species of Peripatus about whose anatomy anything is
known seem to fall into three groups:

(1) The New Zealand species (P. Nova#-zeaLanpim), which
stands by itself.

(2) The Cape species, three in number (P. capensis, P.
BREVIS, and P. Batrouri).

(3) The South American species (P. Epwarpsi1, P. ror-
quatus, and P. imrHuRNI).

The only really important anatomical difference between the
groups 2 and 3 is that in the South American species there is
present between the ovary and the receptaculum seminis
another closed thin-walled vesicle (“ ovarian funnel” of Gaf-
fron (1), receptaculum ovorum, Kennel (3)), which Sedgwick

360 W. L. SCLATER.

regards as homologous with the nephridial funnel of the genital
segment and its vesicle. This structure is entirely absent in
the Cape species of Peripatus.

Also in the South American species the generative opening
lies between a pair of well-developed legs, while in the Cape
species the legs of the generative segment are rudimentary or
represented only by the anal papillz.

In the New Zealand species the legs of the generative
segment are well developed, but the question of the presence
or absence of the receptaculum ovorum does not seem to be
quite settled.

These are the only really important anatomical differences
between the three groups of Peripatus as far as is at present
known.

The only other instance of such great variation in develop-
ment between such nearly allied forms which I can eall to
memory is that of Bateson’s Balanoglossus, which in its deve-
lopment differs largely from the ordinary Tornaria balano-
glossus.

This, however, is easily explained by the difference of the
habits of the two forms; Bateson’s larva is a mud-living animal,
while Tornaria is pelagic.

The curious thing about Peripatus is that, as far asis known,
its habits and mode of life are much the same all over the world,
so that the striking differences in the development of the three
forms cannot be explained by change of habits modified by
external conditions.

In conclusion, I wish to apologise to my readers for the in-
completeness of my work, and for my uncertainty on many im-
portant points; my excuse is the want of time and the absence of
material, since a portion of that which I brought from Demerara
was not satisfactory. And as it seemed unlikely that for some
time at least anyone would be able to procure more material,
either alive or with the uterus adequately preserved, I have
thought it better to publish the small contribution that I have
been able to make towards our better knowledge of that most
ancient aud interesting of all living Arthropods—Peripatus.

:

A SOUTH AMERICAN SPECIES OF PERIPATUS. 361

List oF LITERATURE REFERRED TO.

(A complete bibliography of Peripatus will be found in ‘Proc. Zool. Soc.,’
1887, p. 133.)

(1) Garrroy, E.— Beitrage zur Anatomie und Histologie von Peripatus,”
‘ Zool. Beitr. (Schneider),’ i, Taf. vii—xii, pp. 3360, 1883; also tom.
cit., Taf. xi, xii, xili, pp. 145—162, 1885.
(2) Hurron, F. W.—“* On Peripatus nove-zealandiz,” ‘Ann. Mag.
Nat. Hist.,’ (4) xviii, pl. xvii, pp. 361—369, 1876; also op. cit. (4)
xx, pp. 81—83, 1877; and op. cit. (5) i, pp. 204206, 1878.
(3) Kennet, J.— Entwicklungsgeschichte von Peripatus Edwardsii,
Blanch., und Peripatus torquatus, n. sp.,” Theil i, Taf. v—xi;
* Arbeit. zoot.-zool. Inst. Wirzburg,’ vii, pp. 95—228, 1885, Theil ii,
Taf. i—vi; ‘Arbeit. zoot.-zool. Inst. Wiirzburg,’ viii, pp. 1—93,
1886.

(4) Mosetey, H. N.—* Remarks on Observations by Capt. Hutton, Director
of the Otago Museum, on Peripatus nove-zealandiz, with notes

on the Structure of the Species,” ‘Ann. Mag. Nat. Hist.,’ (4) xix, pp.
85—91, 1877.

(5) Scrater, W. L.— Notes on the Peripatus of British Guiana,” ‘ Proe.
Zool. Soc.,’ 1887, pp. 130—137.

(6) Sepewicx, A.—“ The Development of the Cape Species of Peripatus.”
Part I, with pls. xxxi, xxxii, ‘Quart. Journ. Micr. Sci.,’ xxv, pp.
446—449, 1885; Part II, with pls. xii—xiv, ‘ Quart. Journ. Micr.

Sci.,’ xxvi, pp. 175—212, 1886; Part III, with pls. xxxiv—xxxvii,
‘Quart. Journ. Micr. Sci.,’ xxvii, pp. 467—550, 1887.

362 W. L. SCLATER.

EXPLANATION OF PLATE XXIV

Illustrating Mr. W. L. Sclater’s Memoir, “On the Early
Stages of the Development of the South American Species
of Peripatus.”

Complete List of Reference Letters.

am. So-called amnion. c. Cuticular layer. c&. Folded inner edge of the
uterine epithelium. e. Embryo proper. ec. Ectoderm. ez. Endoderm, //.
Cells of doubtful function, probably nourishing. mes. Mesenteron. x. Nuclei
of vacuolated epithelium. 2.c. Nerve-cord. o. Artificial cavity due to re-
agents. .6. Polar body. py. Pigment. pl. Vesicle swelling (placenta).
pr. Primitive streak. sp.c. Supporting cells. w.e. Uterine epithelium. wt.
Uterus. v. Vesicle. v.e. Vacuolated epithelium. .w. Vesicle wall. <z.
Embryonic swelling (embryonic placenta) whence the stalk arises.

All the figures, except Figs. 1 and 11 to 13, are drawn with Zeiss’s camera
(obj. D, oc. 2), but Figs. 16 and 17 have been reduced to one half the original
size.

Fie. 1.—Longitudinal section through the upper part of the uterus, showing
a young embryo and the curious histological structure of the uterus. Drawn
with Zeiss’s camera. (Obj. A, oc. 2.)

Fig. 2.—Youngest embryo met with, measuring ‘(04 mm. in diameter. It
seems to consist of eight nuclei with surrounding protoplasm.

Fic. 3.—Embryo segmenting rather more advanced with polar bodies.

Fic. 4.—Embryo with commencing central cavity.

Fic. 5.—Segmenting embryo drawn with the surrounding uterus to show
the structure of the uterus and its relation to the embryo.

Fie. 6.—Young blastosphere stage before gastrulation and after the com-
pletion of the segmentation.

Fries. 7A and 7B.—Two sections of an embryo in the pseudogastrula stage.
Fig. 74 shows the condition of the embryo at one end beyond the influence of
the invagination; Fig. 7B is a transverse section through the gastrula and its
blastopore.

Fig. 8.—Youngest embryo found in a vesicle. The wall of the vesicle is
much broken, only a very small part being represented in the figure.

Fies. 94 and 98.—Two sections through a vesicle and embryo of moderate
size. The embryo is still sessile. Fig. 9a represents a section through the
middle of the embryo, and shows the so-called amnion in part and the sup-

A SOUTH AMERIUAN SPECIES OF PERIPATUS. 363

porting cells (sp.c.) ; Fig. 98 is cut through the vesicle behind the region of
the embryo, and shows the amnion at its highest development.

Fie. 10.—Section through an embryo of almost the same age as Fig. 9,
probably a little older. The vesicle wall in this case is reduced to a very thin
string of protoplasm with a few nuclei embedded in it.

Fies. 11, 12, 13, 144 and 148 are drawings of solid embryos; in the case of
Figs. 11, 12, and 13, lying in the uterus, which has been split open to expose
them ; in the case of Figs. 144 and ]48 the embryo has been removed from
the uterus. All, with the exception of Fig. 14, have been drawn with a Zeiss
A after being made transparent by soaking in benzole. Fig. 144 was drawn
when lying in absolute alcohol by refracted light.

Fig. 11. Stage about the same age as Fig. 8. The embryo is entirely
sessile; there is no sign of a stalk. The embryo measures ‘16 mm. x
‘06 mm.

Fig. 12. Stage of about the same age as Figs. 94 and 9B. The embryo
is shown lying in the vesicle, which is in turn shown lying in the
uterus ; both uterus and vesicle having been split open.

Fig. 18. A stalked embryo in a vesicle, the vesicle lying in the split
uterus. Embryo measures ‘24 mm. long.

Figs. 144 and 148. The same embryo, drawn, the former by refracted,
the latter by reflected, light ; Fig. 144 a ventral view, Fig. 148 a side
view; show a vesicle with an acorn-shaped embryo lying within it.
The embryo is stalked, the stalk springing from the darker area (so-
called embryonic placenta). The embryo measures about ‘24 mm.
across.

Figs. 15a, 153, and 15c.—Three sections through different parts of the
same embryo. Fig. 15a is the posterior of the three, Fig. 158 the middle one,
and Fig. 15c the most anterior. Fig. 15a shows the primitive streak (jpr.),
Fig. 158 the clump of cells growing forward from that point, Fig. 15c the
commencing mesenteron.

Fie. 16.—Part of the vesicle wall of an embryo of the same age as Fig. 15,
showing the commencement of the formation of the so-called placenta.

Fie. 17.—Later stage in the development of the placenta.

Chr are

ee
¥

327 bo nob,

aOR aa
a
7

a
a

‘ Jae .

Bs
-

W.L.Sciater del.

Mucr Leura Vea XUN NSPE YY

7h

atin if
wr Mtn

F. Huth, Lith? Edin™

ON THE ANATOMY OF ALLURUS TETRAEDRUS. 365

On the Anatomy of Allurus tetraedrus
(Eisen).

By

Frank E. Beddard, M.A.,

Prosector to the Zoological Society, and Lecturer on Biology at Guy’s
Hospital.

With Plate XXV.

Allurus was first recognised as a distinct genus by Eisen,!
who described the worm from specimens found in Sweden. It
has since been recorded by Dr. Rosa,” from Northern Italy ;
as it is probably identical with Hoffmeister’s* Lum bricus
agilis, it occurs also in North Germany and in Great Britain
and France.

The specimen described in the present paper comes from the
island of Tenerife. I received it alive with a number of other
specimens of Lumbricus and Allolobophora from the
Rev. C. V. Goddard, who was good enough, at the request of
Mr. Stogdon, of Harrow, to furnish me with a sample of the
earth-worm fauna of Tenerife.

The papers already referred to contain no account of the
internal anatomy of Allurus, and I am not aware that we at
present possess any knowledge of its structure.

I have thought it worth while, therefore, to publish the
following notes.

With regard to external characters, I have little to add to

' «Ofv. af Kongl. Vetensk. Akad. Forh.,’ 1870, p. 966.
2 «1, Lumbricidi del Piemonte,’ p. 51.
3 «Die bis jetzt bekannten Arten aus der Familie der Regenwiirmer,’ p. 36.

366 FRANK E. BEDDARD.

the descriptions of Rosa and others. In every particular my
specimen agreed with the descriptions given by those authors.
The first dorsal pore lies between segments 4 and 5, opening
of course into the latter segment.1_ The most important point
mentioned hitherto among the external characters of Allurus
is the forward position of the male reproductive apertures and
the clitellum ; the former are upon the 13th segment, the latter
extends from Segments 22—27.

Anatomy.—tThe single specimen at my disposal was care-
fully prepared for histological purposes; the anterior region was
investigated by means of longitudinal sections; the clitellar
region was studied by a continuous series of transverse sections.

Reproductive Organs.—The testes are two pairs
(Pl. XXV, fig. 2 ¢); they are situated on the anterior me-
sentery of Segments 10 and 11; they correspond, therefore,
in number and position as well as in structure to the testes of
Lumbricus and many other genera.

The seminal reservoirs (fig. 2 v. s.) occupy Segments
9, 10, 11, and 12. The two anterior pairs are smaller than the
two posterior, and are attached to the hinder septum of their
segment; the posterior pairs of seminal reservoirs are attached
to the anterior wall of their segment.

The disposition of the vesiculz is, in fact, precisely that of
Allolobophora? and of Criodrilus;’ as in those genera
the funnels of the vasa deferentia float freely in their seg-
ments (10 and 11), and are not in any way enclosed by the
seminal reservoirs.

The vasa deferentia unite in the 11th segment to form
a single tube with the usual structure, which opens on to the
exterior in Segment 13. The termination of the vas deferens
is stated by Rosa to be furnished with an atrium, confined to
the 13th segment. The body which Rosa terms an “ atrium ”
does not appear to me to be really comparable to an atrium;

1 *Ude (‘ Zeitschr. f. wiss. Zool.,’? 1885, p. 139) states that the first dorsal
pore is between 3 and 4.

2 Bergh, ‘Zeitsch. f. wiss. Zool.,’ 1886, p. 314.

3 Benham, this Journal, 1887, p. 567.

eer fF ail

ON THE ANATOMY OF ALLURUS TETRAEDRUS. 367

it is formed by a mass of glandular cells, exactly like those of
the clitellum, which form a rounded mass (see fig. 5) through
which the vas deferens passes without losing its distinctness ;
there is no cavity into which the vas deferens opens. It appears
to resemble in every way a similar structure in Criodrilus,
’ which Benham! has more correctly spoken of as “ prostate.”

Rosa’s account of the organ in question differs from that of
Benham. My own observations indirectly confirm the descrip-
tion given by Benham.

In Allurus the clitellum does not commence until the
22nd segment, whereas in Criodrilus it begins in the 15th
segment; the “prostate” gland has, however, the same
minute structure as the clitellum, and has every appearance
of being a solid ingrowth of the epidermis. In the succeeding
segments, including those of the clitellum, there is (fig. 8) a
solid ridge of glandular cells on each side of the body, just
above the ventral pair of sete; this ridge is continuous in
front with the “ prostate”’ glands, and consists of a mass of
cells exactly like those of the prostate ; it is plainly connected,
as shown in the figure cited, with the epidermis.

The ovaries? (figs. 2,5, ov.) are in Segment 13, and occupy
the usual position.

The oviducts (figs. 2, 5, od.) open by a much plicated
funnel into the same segment; their external aperture is in
Segment 14 (fig. 1, ?).

A receptaculum ovorum (figs. 2, 5, ro.) is present in
Segment 14. Its cavity is continuous with the oviducal funnel,
through the disappearance of part of the septum.

The oviduct, therefore, opens on to the exterior of the body
behind the male reproductive pores, as also occurs in Rhyn-
chelmis and other Lumbriculide; this arrangement is not
met with in any other earthworm, except M oniligaster.®

1 Loe. cit., p. 568.

* My specimen possessed an additional pair of ovaries in Segment 14. This
may or may not be an abnormal condition.

3 Horst, ‘ Notes from the Leyden Museum,’ 1887.

4 Beddard, ‘ Ann. and Mag. Nat. Hist.,’ Feb., 1886.

VOL. XXVII1, PART 3,—NEW SER, C.¢

368 FRANK E. BEDDARD.

Spermathece.—I examined my sections for a long time
without finding any evidence of the presence of spermathecz.
Finally, I succeeded in discovering in the 8th segment a struc-
ture which is illustrated in fig. 4; this consists of a minute
pouch (c.p.), lined with epidermic cells, and plainly formed as an
invagination of the epidermis, with which it is continuous.
This pouch would not, however, have been seen in a dissection
of the animal, because it only extends for a little way into the
thickness of the muscular layers of the body wall. I am dis-
posed to believe that it does represent a spermatheca, though
presumably in a very immature condition. This presumption
is strengthened by the fact that one at least of the ventral pair
of sete, near to which the pouch opens, is different in form
from the other sete of the body, being thinner and longer (fig.
3a). I found no traces of spermathece in any but the 8th
segment. Without assuming that there is only a single air
of spermathece present in the adult Allurus, I may point out
that the position of these organs—situated in front of the testes,
vasa deferentia funnels, and seminal reservoirs—is only found
in such Lumbricide as Allolobophora complanata where
there are a large number of pairs of these organs; further-
more, it is important to note that they do not in Allurus open
on to the intersegmentail furrow, but in the middle of the
segment to one side of the sete.

The alimentary tract presents no features of special
interest. The gizzard is placed at the junction of the cso-
phagus and the intestine, as in the Lumbricide, and is
preceded by a crop. It differs, however, from that of
Lumbricus in only occupying a single segment, the 17th.
Its structure, as shown in longitudinal section, is illus-
trated in fig. 9. At present there are no materials for a
comparison of the minute structure of the gizzard in different
forms. I give the figure because it is unlike any figure con-
tained in Claparéde’s ‘ Memoir! on the Histology of the Earth-
worm,’ and also because it does not agree with preparations of

1 ¢ Zeitschr. f. wiss. Zool.,’ 1869.

ON THE ANATOMY OF ALLURUS TETRAEDRUS. 369

my own of the gizzard of Urocheta. The intestine has a
typhlosole.

The buccal cavity (see fig. 10) occupies the first three
segments. The pharynx lies in 4, 5, and 6.

The cesophagus, from the middle of Segment 10 to the end
of Segment 14, is furnished with a structure which corresponds
with the calciferous glands of Lumbricus and other earth-
worms; the esophagus is here rather wide, and the lateral
region is much folded; these longitudinal folds (fig. 2 ca.)
are quite continuous from segment to segment; there is no
question of any separation into distinct glands. This arrange-
ment is possibly to be regarded as a primitive one.

The epithelium of the cesophagus, which differs in character
from that which covers the lateral folds, is very distinctly
ciliated in the segment where the folds are developed (fig. 2 a.).

T& segment J0, at the junction of the anterior section of
the cesophagus with that part which has just been described,
is a short diverticulum on either side ; this diverticulum marks
off the non-ciliated anterior region of the cesophagus, which
also differs from the ciliated region in the shape of its cells.
They are not quite so tall and narrow as the ciliated cells. In
all these particulars the diverticulum agrees in structure with
the anterior section of the cesophagus.

The crop occupies the 15th and 16th segments.

With regard to the nervous system, I may mention that
the cerebral ganglia are in the 4th segment (fig. 10), and that
the ventral nerve-cord has cells along its whole length.

Nephridia.—These organs consist for the most part, as in
the Oligocheta generally, of a series of “drain-pipe”’ cells ;
as in Lumbricus and other genera, the calibre of the inter-
cellular duct, and of the cell which contains it, is at first less
than it becomes afterwards. Fig. 11 is a transverse section
through a nephridium which illustrates this fact; @ are the
smaller drain-pipe cells, 2 the larger. The nephridium is sur-
rounded by a number of peritoneal cells (p.) ; these are oval in
form, the protoplasm appearing almost homogeneous and have
a darkly staining nucleus. The larger drain-pipe cells »’ appear

370 FRANK E. BEDDARD.

to be embedded in the substance of other cells (n, fig. 11);
these latter stain very slightly with borax-carmine, and the
protoplasm is arranged in a reticulate fashion. The nucleus of
the cell is small but conspicuous in that it is deeply stained ;
it is perfectly plain, as also shown in a longitudinal section
(fig. 12) that the drain-pipe cells actually perforate these cells.

I cannot, however, be perfectly certain as to the correctness
of the above description ; it may be that the larger cells are
simply specialised into a denser area surrounding the lumen,
and a less darkly staining region situated peripherally. There
is, however, a very sharply-marked boundary line between the
_two regions.

If the supposition that these two regions are really distinct
cells, be correct, there is an obvious resemblance to Clepsine.!

Summary of structural differences from Lumbricus
and Allolobophora.

(1) Position of male reproductive folds upon Segment 13
and therefore in front of female generative orifice (fig. 2).

(2) Presence of a single (?) pair of spermathece opening
on to the middle of their segment a little to one side of
seta (fig. 4).

(3) Calciferous glands of consecutive segments not distinct
from each other, occupying Segments 10—14.

(4) Gizzard confined to a single segment (No. 17).

(5) Structure of nephridia (figs. 11 and 12).

(6) The presence of a continuous glandular fold on either
side of the body of the same structure as the clitellum, ex-
tending from thef4th to the 24th segment, and interrupting
the muscular layers (figs. 2 and 8).

(7) The special development of this glandular mass in
Segment 13, round the orifice of the vas deferens (fig. 5).

1 A. G. Bourne, this Journal, 1885, p. 482.

ON THE ANATOMY OF ALLURUS TETRAEDRUS. 371

EXPLANATION OF PLATE XXV,

Illustrating Mr. Frank E. Beddard’s paper “On the Anatomy
of Allurus tetraedrus, Eisen.”

Fic. 1—Allurus sp., lateral view of anterior segments. 9 oviducal pore.
3. Pore of vasdeferens. sp. Spermathecal pore. d. Clitellum. ~. Nephridipores.
Fic. 2.—Dissection of genital region of one side of the body. v.s. Seminal
reservoirs. ¢. Testes. v. d. Funnels of vasa deferentia. ov. Ovary. od. Oviduct.

r. o. Receptaculum ovorum. ca. Calciferous glands. p. Glandular fold of in-
tegument. :

Fic. 3.—Elongated seta (a) from Segment 8 compared with one of the
ordinary sete (4).

Fic. 4.—Section through body wall of 8th Segment, to show the rudimen-
tary spermatheca (c. p.).

Fig. 5.—Section through part of Segment 13. ». d. Vas deferens. cl.
Mass of glandular cells. ov. Ovary. od. Oviduct. 7.0. Receptaculum
ovorum.

Fies. 6 and 7.—Isolated glandular cells from region marked e/. in last
figure.

Fic, 8.—Part of a transverse section through one of clitellar segments.
s. Sete of ventral pair. 2. Nephridium, showing external aperture. p. Mass
of glandular tissue.

Fie. 9.—Longitudinal section through gizzard. m. Longitudinal muscles.
¢. Transverse muscles. e. Epithelium. c. Internal cuticular lining. 4/7. These
spaces were in many sections filled with blood.

Fic. 2 a.—Transverse section through cesophagus and calciferous glands.

Fic. 10.—Longitudinal section (diagrammatic) through anterior segments,
to show position of cerebral ganglia. 4. Buccal cavity. pf. Pharynx.

Fic. 11.—Transverse section through nephridial tubule. yp. Peritoneal
cells. x’. Drain-pipe cell, perforating, ., another cell.

Fic. 12.—Longitudinal section, similar to preceding.

—e YS ——=— Ve UC UL ae on eee — en a a ne rie aaah i |

\

aS

———
7

SPU EE INET ADRIENNE NCS NI tf erer eee er eee meaner
Pes rare -
ee

&

Xi
xu |

xv |}

F. Beddaréd del.

tv

,
4

Mor. Four VALIMUNS FA

F. Huth Lith® Edin

DEVELOPMENT OF THE CAPE SPECIES OF PERIPATUS. 3873

The Development of the Cape Species of
Peripatus.

PARE EV:
THE CHANGES FROM STAGE G TO BIRTH.
By

Adam Sedgwick, M.A., F.R.S.,
Fellow of Trinity College, Cambridge.

With Plates XXVI, XXVII, XXVIII, and XXIX.

Tue changes which take place during and subsequent to
Stage e are mainly changes of growth and histological differ-
entiation, The most important organs which have not yet
made their appearance by the close of Stage F, are the crural
glands and trachee. The origin of the latter is, I regret to
say, still hidden from me. The remaining organs have acquired,
in all essential respects, the adult relations by the close of
Stage Fr.

Tue EcropEerm.

In Stage c, the ectoderm retains the charact2rs already
described in Part III, p. 472. It forms an extremely thin,
much vacuolated layer over the greater part of the body, and
the nuclei are far apart and in a single layer (fig. 5). In
embryos of this age, the ectoderm does not contract when the
embryo is preserved, and no doubt its extreme tenuity is due
to this fact. On the ventral organs the ectoderm is thicker
and the nuclei in more than one layer and close together.
On the dorsal hump also, the ectoderm still remains thick,
with a large amount of protoplasm external to the nuclei.

374 ADAM SEDGWICK.

The dorsal hump, however, has already begun to atrophy ; it
eventually completely vanishes.

The general ectoderm possesses a large number of vacuoles,
and in certain places the nuclei are aggregated together in
masses, and are smaller than elsewhere, forming the rudimeuts
of the future spiniferous sense-organs. The latter give rise to
the white spots seen on the skin of embryos of this age.

In P. Balfouri, and, to a slight extent, in P. capensis,
the dorsal ectoderm contains a number of highly refractile
globules (Pl. XXVI, figs. 1—4). These are probably yolk
globules. They seem to be most numerous in the dorsal hunip.
The whole dorsal ectoderm of P. Balfouri is thicker than the
ventral, and partakes, to a certain extent, of the character of
the ectoderm of the dorsal hump.

In the later stages, in embryos just before birth, the dorsal
ectoderm is highly protoplasmic and much striated, and con-
tains very few, if any, vacuoles (Pl. XX VIII, fig. 12), while the
ventral ectoderm is much vacuolated, and retains the cha-
racters it possessed at Stage c. That is to say, the nuclei lie
in the outer part of the layer, the inner parts being reduced
to fine unstained strands passing between the vacuoles.

The further changes in the general ectoderm need no special
description. The nuclei come closer together, the vacuoles
disappear, a cuticle is formed on the outer surface, and the
adult condition gradually acquired. The claws of the jaws
and legs, and the spines of the sense-organs, are special de-
velopments of the cuticle.

The following ectodermal organs require a special description,
and will be considered separately and apart from the general
ectoderm :

1. The ventral organs.

2. The nervous system.

3. The slime-glands and crural glands.

On the origin of the tracuzex I have no observations.
They seem to arise very late, and have hitherto escaped my
observation.

DEVELOPMENT OF THE CAPE SPECIES OF PERIPATUS. 375

Toe VENTRAL ORGANS.

For the origin and general history of these organ I must
refer back to Part III, p. 476, and to Kennel (No. 1). They
consist of segmented thickenings of the ectoderm, placed be-
tween the appendages and composed of two halves, which are
in contact in the middle ventral line (Pl. XXVI, fig. 5).
During Stage « they possess two kinds of nuclei: the surface
layer of oval, and an inner mass of more rounded elements. The
latter are much inclined to drop out in the sections, leaving a
surface layer of nuclei and thin protoplasmic strands passing
inwards.

From what has been said as to their history it is obvious that
the ventral organs represent a portion of the ectoderm, from
which the central nervous system was constricted off. They
correspond in number with the segments, and are therefore
twenty in number in Capensis.

1. The ventral organ of the first somite is probably,
as Kennel has suggested, represented by the cerebral grooves.
These become completely cut off from the surface ectoderm and
form the hollow appendages attached to the ventral side of the
brain of the adult (Pl. XX VI, figs. 2,3, and No. 2, fig. 19, c, p).
The walls of these vesicles appear to consist of nervous tissue.

2. The ventral organs of the jaws (Pl. XXVI, fig. 4,
v. o. 1) come to lie in the buccal cavity on each side of the
mouth at the base of the jaws. They differ from all the
posterior ventral organs in not coming into contact with one
another in the middle ventral line. They remain in the ecto-
derm, and appear to retain a connection with the posterior
lobe of the brain, or anterior part of the circumpharyngeal
commissure.

3. The ventral organs of the oral papille join one
another ventrally and become divided into two parts by the
lips—an anterior contained in the posterior region of the
buccal cavity, into which the salivary glands open ; and a pos-
terior part on the ventral side of the body just behind the
mouth. The intrabuccal part remains in connection with the

376 ADAM SEDGWIOCK.

lateral nerve-cords (No. 2, fig. 14), and these two connections,
together with the interposed ventral organ, contribute what
Balfour has called the second commissure between the ventral
cords.

The posterior part of this ventral organ behaves exactly as
do those about to be described.

4. The two halves, of which each ventral organ of
the seventeen ambulatory legs at first consist, join one
another ventrally, remain as part of the ectoderm, and appear
to retain a cellular connection with the lateral nerve-cords
(Pl. XXVI, fig. 5). I could not be certain of this connection
in the case of every ventral organ; but Kennel asserts that it
exists, and I am inclined to agree with him.

Nervous System.

The early development of the central nervous system is
described in Part III, pp. 473—475 and 481. In Stage F the
cerebral grooves are still open (Part III, fig. 33), and the
ventral cords are in close contact and still continuous with the
thickened ventral ectoderm. The white matter has also made
its appearance along the whole length of the dorsal side of
both brain and spinal cord.

In Stage G two important changes have taken place.
(1) The cerebral grooves have become converted into closed
vesicles (P].XXVI, figs. 2 and 3) and entirely cut off from the
superficial ectoderm. (2) The ventral cords have withdrawn
themselves from the ventral ectoderm, though they still appear
to be attached to the latter by marked cellular processes (fig. 5,
Pl. XXVI).

The Brain.—The structure of the brain in Stage ¢ is shown
by the series of sections figured (figs. 1—4, Pl. XXVI). Ex-
cepting the increase in the amount of white matter and the
closure of the cerebral grooves it is essentially the same as in
Stage Fr. In front the two lobes of the brain, though in close
contact, are separate from one another (Pl. XXVI, fig. 1).
They are continued forwards into the antenne as the anten-
nary nerves (vide Part III, p. 480). The white matter is

DEVELOPMENT OF THE CAPE SPECIES OF PERIPATUS. 3877

dorsal and extends somewhat into the centre of the lobes. It
may be described (vide Balfour, No. 2) as consisting of three
horns: viz. a dorso-lateral (a), a ventro-lateral (4), and a dor-
sal (ec). It is continued forwards along the dorsal sides of the
tentacular nerves.

At a little distance behind the eyes the central lobe of white
matter and the cells dorsal to it become continuous with the
same structures on the opposite side (Pl. XXVI, fig. 2). A
little farther back the connection between the two brain lobes
is effected only by the white matter, the dorso-median patch
of cells entirely disappearing (Pl. XXVI, fig. 3). A few sec-
tions farther back the latter again appear, and their appear-
ance is soon followed, in the section series, by the separation
of the cerebral lobes (Pl. XXVI, fig. 4).

The ventral appendages of the brain have already been
dealt with (p. 375).

The Eyes.—At the close of Stage & the eyes are closed
vesicles, connected by their ventral corners with the brain
immediately external to the white matter (Part III, fig. 22 a).
The connection, which is at first a broad one, has in Stage ¢
become constricted to a narrow pedicle—the optic nerve—
connecting the inner wall of the optic vesicle with the brain
(Pl. XX VI, fig. 1); at the same time the nuclei withdraw
from the optic nerve and from the portions of the wall of the
optic vesicle and of the brain, which connects the optic nerve
(fig. 1). In this manner the white matter of the so-called
optic ganglion of the adult is established, and the optic nerve
comes to consist entirely of white matter. The layer which
will form the rods in the adult eye appears at a very early
stage as a result of the withdrawal of the nuclei of the thick
inner wall of the vesicle from the internal surface (Part III,
fig. 22 a, and Pl. XXVI, fig. 1, rods). The pigment and lens
have not yet been formed. The eye, therefore, in Stage e
(Pl. XX VI, fig. 1) consists of a vesicle with a thin outer wall
closely subjacent to the epidermis, which is very thin over the
eye, and of a thick inner wall lying close to and connected with
the brain by a cord of nerve-fibres. The inner wall further

378 ADAM SEDGWIOK.

presents a patch of white matter at the point of entrance of
the optic nerve, and a layer of white matter (the rods) next the
lumen of the vesicle. The pigment appears in January at the
junction of the layer of rudimentary rods and the nuclei. The
lens is formed at about the same time, as a secretion of the
wall of the vesicle. It lies within and fills up the cavity of
the vesicle. This condition of the eye is practically that of
the adult.

Toe VENTRAL Corps.

The early history of the ventral cords is given on p. 475 of
Part III. In Stage F they are still in close contact with the
ectoderm, but an indistinct line of separation can generally be
seen between them (Part III, fig. 39). In Stage e the separa-
tion is complete and distinct, though they still remain con-
nected at intervals by cords of cells with the ventral organs
(Pl. XXVI, fig. 5). It isin Stage ¢ also that the commissures
between the ventral cords and the main nerves first become
apparent. The commissures between the two nerve-cords are
very numerous. They extend in this stage from the ventro-
median corner of the cords towards the ectoderm, where they
lie in close connection with some rather loose fibrous tissue,
which is found at this stage everywhere beneath the ectoderm.
They consist of fibrous matter, and can be easily traced into
the white matter of the cords.

The efferent nerves arise from the outer border of the cords,
directly from the white matter. They are very numerous
(Balfour, No. 2), but there are, opposite each leg, two—the
pedal nerves—which are much larger than the others and more
easily observed. These two arise, the one immediately in front
of the nephridium, and the other behind it. In Stage e, when
they are first apparent, they consist of close bundles of fibres
passing out from the white matter (Pl. XXVI, fig. 5, nerve)
and continuous with a loose plexus of fibres placed imme-
diately within the ectoderm of the ventral side of the legs
(neuro-musc.). It therefore appears that the commissure be-
tween the nerve-cords, the efferent nerves, and the fibrous

eT

i ee

DEVELOPMENT OF THE CAPE SPECIES OF PERIPATUS. 379

matter beneath the ectoderm, all become distinctly apparent
at about the same time in Stage a; but how and when they
are developed I am unable to say. As may be seen from an
inspection of the sections (Part III, figs. 837—-39) there is in
Stage F a certain amount of this fibrous tissue, especially at
the ventro-lateral corners of the body, close to the outer border
of the nerve-cord, and in the nerve-cords themselves as the
white matter, and I have no doubt that it is present at a still
earlier stage, though masked by the large amount of nuclei
present. In fact, it may be said of this tissue generally, that
it does not become a marked feature of the sections until the
organs separate from one another and leave room for the pre-
viously closely-packed nuclei to spread out, and, as in the case
of the white matter of the nerve-cord, partly to withdraw them-
selves from it (cf. Pl. XX VI, fig. 5, and Part III, fig. 39). In
whatever manner this tissue may be developed, I think there can
be little doubt that it is from its first appearance a continuous
tissue, that is to say, the circular fibres at circ. musc. in Pl.
XXVI, fig. 5, are continuous with the network at newro-musce.,
which, in its turn, is continuous with the bundleof fibres forming
the nerve, and so with the fibrous matter of the nerve-cords. It
thus appears, so far as I have been able to observe the develp-
ment, that the nerves are not formed as outgrowths from the
central nervous system, but are parts of a network which
originally existed when the nerve-cords were part of the
surface ectoderm. In Stage ec, the network is clearly con-
tinuous with the surface ectoderm (Pl. XXVI, fig. 5). With
regard to the commissures connecting the ventral nerve-cords,
it seems to me that they also are differentiated in situ from the
median ventral ectoderm at a time when the nerve-cords were
still parts of the surface ectoderm. I have already said that I
do not know the manner in which this network develops ; part
of it is undoubtedly formed around the ectodermal nuclei, e. g.
the white matter of the cords, the commissures between the
cords; some of it, on the other hand, has, from the first, a
relation to the mesodermal nuclei, e. g. the circular fibres at
circ. musc., and the network on the ventral side of the feet at

380 ADAM SEDGWICK.

Pl. XXVI, fig. 5, meuro-musc. The nerves, therefore, are
to be regarded as special differentiations of a pre-existing
network, the origin of which is not known, but which at first
pervades and is continuous throughout the whole ectodermal
and mesodermal tissues of the body.

StrmE-GLanps AND CruraL Guanps.

The slime-glands are entirely ectodermal products. Their
early development has been described in Part III, p. 482, and
the later changes being simply processes of growth, I have
nothing to add to the account there given.

The crural-glands appear very late. In embryos of April
almost ready to be hatched they have the form of shallow
invaginations of ectoderm immediately external to the opening
of the nephridium (P]. XXVII, fig. 11). They seem to be
entirely derived from the ectoderm, but I have no details as to
their development. I could find no trace of the enlarged
crural gland of the last leg of the male in the oldest embryos
which I have examined.

The Stomodeum and Proctodeum—tThe early history
of the sestructures has already been given in Part III, pp.
482, 485, 486.

The mesodermal investment of the anterior part of the
stomodzum becomes very thick, while that of the posterior
part remains comparatively thin. The lining cells secrete a
cuticular layer. The anterior part becomes the pharynx, and
the posterior the cesophagus of the adult.

The proctodzeum also acquires a cuticular lining and a well-
marked mesodermal investment. It becomes the rectum of
the adult.

THe ENDODERM.

In Stage ce the endoderm is reduced to a layer of extreme
tenuity (Pl. XXVI, fig. 5). It soon, however, begins to in-
crease in thickness, and in embryos almost ready for birth has
the form represented in Pl. XXVII, fig 11. The nuclei are
placed in the deeper parts of the layer, and the protoplasm

—-

DEVELOPMENT OF THE CAPE SPECIES OF PERIPATUS. 381

stains deeply and contains a large number of granules. The
endodermal part of the alimentary canal is without glandular
appendages of any kind. In old embryos the enteron generally
contains a deeply-staining material with a number of highly
refractile particles in suspension. This substance is probably
a secretion of the endoderm cells. The contents of the ali-
mentary canals of the free-living adults is permeated by a
number of similar highly refracting bodies.

Tue MeEsopERM.

The later history of the mesoderm, i.e. the tissues derived
from the walls of the somites, will best be considered under
the following four heads, viz :

1. The muscles.

2. The vascular system.

3. The nephridia.

4. The generative organs.

Whether any part of the cutaneous mesodermal structures
are ectodermal in origin is, as I have already hinted in dealing
with the nervous system, impossible to decide, because of
the intimate connection which is established between the ecto-
derm and the somatic walls of the somites, at almost the first
appearance of the latter (vide pp. 487, 488, Part III),
and which remains during the whole development (see
above (p. 379).

Tue Musctes.

The cutaneous muscles arise from the subectodermal fibrous
network which has been already mentioned, and which in Stage F
was crowded with the nuclei of the ventro-lateral corners of
the somites. The fibres of the outer part of this network
arrange themselves in a circular manner, and form the circular
muscles of the body wall. At first the fibres are extremely
scanty (Pl. XXVI, fig. 5, circ. musc.), but they soon become
more numerous. Nuclei are found at intervals amongst them
(Pl. XXVIII, fig. 13). The longitudinal muscles appear some-
what later within the circular muscle in seven patches, viz. two

382 ADAM SEDGWICK.

dorsal (Pl. XXVIII, fig. 13), two lateral, two ventro-lateral,
and one medio-ventral between the nerve-cords, These patches
gradually enlarge into the corresponding muscular bands of the
adult. They contain nuclei which have often a peculiar, irre-
gular shape (PJ. XXVIII, fig. 13). The muscles, both longi-
tudinal and circular, are deposited outside the commissures
connecting the ventral nerve-cords.

The muscles of the feet seem to be derived from the fibrous
plexus shown in Pl. XXVI, fig. 1, and are shown at a later stage
in Pl. XXVII, fig. 11). The origin of the transverse septa
dividing the body cavity into a central and two lateral com-
partments has already been described (Part III, pp. 493 and
495, figs. 9, 2la, 24, 39, &c., v.s.). They arise as outgrowths
of the ventral corners of the somites.

The contractile tissue of the gut wall and internal organs
generally, is derived from the wandering cells, which them-
selves appear to be derived from the walls of the mesodermal
somites.

Tue Bopy Cavity anp VaAscuLAR SysTEM.

As was first pointed out by Lankester, the Arthropoda are
distinguished from all other animals by the possession of paired
ostia, perforating the wall of the heart and putting its cavity
in communication with the pericardium. ‘The pericardial
cavity of Arthropoda, therefore, contains blood, and in this
respect differs fundamentally from the similarly named cavity
in other animals. Not only does the pericardial division of
the body cavity contain blood, but the general body cavity,
and in Peripatus all the compartments of the latter are also
vascular tracts; and it is important to distinguish by a special
name this vascular type of body cavity from the non-vascular
or c&Lomic type which is found well developed in Annelida
and Vertebrata. The term “ Ha#Moc@Le,” which has been
suggested by Lankester for this purpose, seems a convenient
one, and I propose to adopt it.

The development of the hemoccele of Peripatus has been
already fully described in Part III, p. 499, et seq., and p. 510;

<<

~
9

DEVELOPMENT OF THE OAPE SPECIES OF PERIPATUS. 383

dl

and I have but little to add to the account there given. It
is derived in part from a system of species developed within
the mesoderm, and in part from spaces arising between the
ectoderm and endoderm. The diagrams (Pl. XXIX, figs.
14—17) will enable the reader to understand at a glance
the origin of the various parts of the vascular tracts and their
relation to the ccelom.

In Stage c the heart becomes a tube with thin walls and
flattened nuclei, lying freely in the pericardial cavity, with
cords of cell projecting from its walls into the latter. These
cords, the origin of which I have not been able to make out—
they first appear in Stage r (Part III, figs. 43, 45, 46, c.c.)
when the dorsal divisions of the anterior somites are disap-
pearing—seem to become transformed in the later stages into
a very remarkable tissue. The structure of this tissue, which
may be called the pericardial network, or reticular tissue of
the pericardium, will be best seen by reference to Pl. XXVIII,
fig. 13, which represents a transverse section through the
dorsal partof an embryo shortly before birth.!. The wall of
the heart is prolonged into delicate processes, which are con-
tinuous with a network occupying a considerable part of the
lateral region of the pericardium. This network contains
round nuclei in its nodes, and is continuous with the floor and
roof of the pericardium. Sometimes the nuclei occur singly, but
often they occur in masses (right hand side of fig. 13), which
often have the appearance of multinucleated cells lying freely
in the pericardium. A careful examination, however, shows
that they are nodes of the network already described, and that
they only differ from the other nodes in possessing more than
one of the round nuclei. Occasionally an apparently free cell
with one nucleus may be seen lying in the spaces of the reti-
culum. ‘This reticular tissue is, I think, derived mainly from
the strings (c. c.) of the earlier stages, and it persists as the pecu-

1 The apparent fusion of the dorsal and ventral walls of the heart to the
dorsal and ventral walls of the pericardium in this figure is due to the con-
traction of the specimen. The heart at this stage, excepting for the network
about to be mentioned, lies quite freely in the pericardium.

VOL. XXVIII, PART 3.—NEW SER, DD

384 ADAM SEDGWICK.

liar cellular tissue which has been described by Gaffron (No. 3)
in the pericardial cavity of the adult. Gaffron compares this
tissue to the fat bodies of other Tracheates, a comparison with
which I am inclined to agree, although I am not aware that
fat bodies are as a rule present in the pericardium.

Mr. Heathcote, however, informs me that in the Myriapoda
a portion of the fat body does lie in the pericardium, and
resembles in its relation to the heart the pericardial tissue of
Peripatus. A tissue exactly like the pericardial tissue is found
in the lateral compartment of the body cavity (Pl. XXVII,
fig. 11). It has been noticed by Balfour and Gaffron. It
seems to me probable that this tissue, which lies in the vascular
system, is of the same nature as the lymphatic tissue of the
Vertebrata, with which it undoubtedly presents many points of
resemblance. The botryoidal tissue of Leeches (Lankester) and
the brown cells of Chetopoda may possibly fall into the same
category. The former presents very much the same relations
to the vascular system, but the latter differs by lying in the
coelom.

In Stage c the horizontal septum which divided the cerebral
compartment of the body cavity into a dorsal (0. d. ¢.’) and
ventral (0. 4.c., Part III, fig. 48) chamber breaks down.

Both the pericardial and lateral compartments of the body
cavity (hemoceele) seem to communicate with spaces amongst
the muscles of the body wall. One such set of spaces is especially
conspicuous between the ectoderm of the ventral body wall and
the circular muscles (Pl. XX VII, figs.11,17). This system of
spaces, which is probably segmentally arranged, communicates
with the spaces in the legs. It is, I think, the blood in these
ventral vascular channels which exudes through the ventral
organs when the animals are contracted by the action of
chloroform.

The ostia of the heart appear to arise in Stage c. I have
no satisfactory observations of them. They are, I think, con-
fined to the posterior end of the heart in the Cape species.

The main vascular tracts, therefore, are five in number, or,
to put it in another way, the HZMoc@LE is divided into five

DEVELOPMENT OF THE CAPE SPECIES OF PERIPATUS. 385

main chambers: (1) the central compartment of the body
cavity; (2) the heart; (3) the pericardial cavity; (4) the two
lateral compartments or lateral sinuses (in which the nerve-
cord and salivary glands lie). In addition to these there are
the leg cavities, which contain the nephridia and communicate
with (4). Of these the central compartment, lateral sinuses,
and heart are free for the most part from traversing tissue,
while the pericardial chamber and the leg cavities are broken
up by reticular tissue, and the leg cavities by muscles as well.

Tue NeEPHRIDIA.

An account of the nephridia up to Stage F will be found in
Part III. By Stage ce they have practically attained the adult
condition, and to complete my account of their history it will
only be necessary to describe their final condition, for which
purpose I have chosen a stage shortly before birth. I propose
at the same time to give a short recapitulatory account of their
whole history, under the head of the somites from which they are
respectively derived. The general changes which the somites
undergo will be rendered clear by a glance at the diagrams
on Pl. XXIX, figs. 14—17. It must be remembered that I
am only dealing here with the cavity contained in the somites,
i,e. the c@Lom, and its immediate lining. The walls of the
somites, particularly the somatic walls, become greatly thick-
ened and hollowed out. The tissues and cavities so formed
give rise to muscles, connective tissue, and parts of the vascular
system, as has already been fully described.

Somites oF THE AnTENN# (Part III, p. 504).—The first
somites send down a diverticulum outside the brain towards
the skin (Part III, figs. 19 6, 50, s. so. 1) and then divide into
two parts (Part III, fig. 51, s.1). They seem to have com-
pletely disappeared by Stage c. The ventral diverticulum of
Stage E is obviously the rudiment of a nephridium.

Somites ofr THE Jaws (Part III, p. 506).—The second
somites do not give rise to even a rudiment of a nephridium,
They seem to disappear.

386 ADAM SEDGWICK.

SoMITES OF THE ORAL Papity# (Part III, p. 507).—The
third somites send out a ventral diverticulum (Part III, fig.
21 c.), which acquires an opening to the exterior (Part III,
fig. 23 e.). They become divided into a dorsal and ventral
part (diagram, Pl. X XIX, fig. 14), of which the dorsal vanishes,
while the ventral persists in connection with the opening above
mentioned, and is at first placed in the appendage (Part III,
fig. 23e). This ventral or appendicular part is the nephridium
of the somite, and becomes the salivary gland of the adult.
The general form of this nephridium at the close of Stage F is
well shown by the diagram (Pl. XXVII, fig. 6). It consists of a
tubular part (/. s.¢. 3, sal. gl.), opening in front on the ventral
surface of the body (o0.s. 3) within the lip and ending blindly
behind, and of a vesicle (/. s. v. 3) opening into the tubular
part a little in front of its termination. The structure and
relations of these parts to each other are illustrated by the
three transverse sections figured in Part III, figs. 37, 38, and
39, aud taken along the lines marked 37, 38, 39 in the diagram
(Pl. XXVII, fig. 6). The subsequent changes which this organ
undergoes are unimportant. They are illustrated by the diagram
Pl. XXVII, fig. 7, and by figs. 8 and 9, taken from transverse
sections of an embryo shortly before birth, along the lines
8 and 9 in fig. 7.

The tubular part has become much elongated (Pl. XXVII,
fig. 7, sal. gl.), so that it now extends a considerable distance
behind the point of communication with the internal vesicle.
It constitutes the salivary gland of the adult, and lies, as is
well known, in the lateral compartment of the body cavity
(lateral sinus). The walls of the vesicle (/. s. v. 3) have be-
come much thicker. They consist (Pl. XXVII, fig. 8) of a
layer of nucleated, richly vacuolated protoplasm. Finally, the
portion connecting the tubular part (/. s. ¢. 3) and the vesicle
(Z. s. v. 8) has become elongated into a tube running forwards
from the tube to the vesicle, as shown in the diagram (Pl.
XXVII, fig. 7). This communicating portion, as shown in
Pl. XXVII, fig. 9, is closely applied to the dorsal side of the
tubular part.

ee le

DEVELOPMENT OF THE OAPE SPECIES OF PERIPATUS. 387

The internal vesicle, which, together with the communica-
ting tube, has hitherto been overlooked, persists in the adult,
and probably constitutes an important functional part of the
salivary gland.

SoMITES OF THE FIRST, SECOND, AND THIRD Lees (Part III, p.
507 et seq.).—Thethird, fourth, and fifth somites divide early into
a dorsal and ventral portion, of which the dorsal vanishes, while
the ventral acquires an opening to the exterior and persists as
the nephridium. The condition of these nephridia in Stage F
is shown by Part III, fig. 40. The subsequent changes are
very slight, and may be gathered from an inspection of
Pl. XXVII, fig. 10, which is from a transverse section through
the third leg of an embryo almost ready for birth. The
nephridium (Pl. XXVII, fig. 10) consists of a thin-walled
internal vesicle contained in the leg compartment of the body
cavity and communicating by a straight tube with the exter-
nal opening on the ventral surface. The wall of the vesicle
consists of a ragged protoplasmic layer, with here and there a
round nucleus.

Somites oF Lees 4 to 12.—The early history of the seventh
to the fifteenth somites inclusive is similar to that of the somites
of the first three legs ; but in the later stages the tubular part
of the nephridium becomes elongated, coiled, and divided into
at least three regions (Pl. XXIX, fig. 17, diagram). (1) The
part next the external opening is dilated into a vesicle—the
external vesicle—which is connected with the external opening
by a narrow tube (Pl. XXVII, fig. 11). (2) The vesicle opens
into a long coiled tube, which forms the greater part of the
nephridium. It is cut across twice in the transverse section
from which fig. 11 was taken. It is continuous with (3) a
short terminal portion in which the nuclei are very closely
packed together. This terminal portion opens with everted
lips into the thin-walled, internal vesicle, and constitutes the
so-called funnel of the nephridium. The external opening
of the nephridia of the fourth and fifth legs are at first imme-
diately outside the nerve-cord, as in the case of the others.
Their adult position is due to a secondary shifting.

388 ADAM SEDGWICK.

Somites or Lees 18 to 17 (Part ITI, p. 510).—The
dorsal divisions of these somites persist as the generative
organs, and will be described below (Pl. XXIX, diagrams
figs. 15—17). The ventral divisions develop as in the legs
immediately preceding.

Somires oF THE ANAL Papriu# (or in P. Balfouri of the
eighteenth legs).—I have nothing to add to the description given
on p. 513 of Part III. They persist entirely as parts of the
generative ducts. For descriptions and figures of the isolated
nephridia ot the seventeen legs of the adult I must refer the
reader to Balfour’s memoir (No. 2), pp. 832—385, and Pl. XIX,
figs. 27, 28. I have nothing to add to his description, except-
ing the fact that the terminal portions of the nephridia do not
open into the body cavity, which is a vascular space and not
celomic, but, as shown in Pl. XXVII, fig. 11, and in diagram
fig. 17, into a thin-walled vesicle, which is directly derived
from the original somite.

I think there can be no doubt that the vesicle of the
nephridia of the first three legs is homologous with the
internal vesicles of the posterior nephridia and not with the
collecting, or external vesicle. A comparison of figs. 10 and
11 on Pl. XX VII, shows that the tubular part of the first three
nephridia is very different from the narrow tube leading outward
from the external vesicle in the posterior nephridia; though it
is without the closely-packed nuclei in the terminal so-called
funnel. Further, the structure of the wall of the vesicle itself
resembles that of the internal vesicle of the posterior nephridia
and not that of the collecting vesicles.

The external cuticle is only prolonged for a very short
distance into the neck of the collecting vesicle.

THe GENERATIVE ORGANS. .

The early history of these organs has already been fully
described in Part III, pp. 511—515, and I have but little to add
to that description. They first appear in the endoderm as large
round nuclei (Part III, figs. 26, 27), which migrate into the
splanchnic mesoderm (Part ITI, fig. 41) of the dorsal divisions of

ee ee a

DEVELOPMENT OF THE CAPE SPECIES OF PERIPATUS. 389

the sixteenth to the twentieth somites, where they acquire a pro-
toplasmic investment (Part III, figs. 43,47). The parts of the
somites containing them persist as the generative tubes, and
become continuous behind with the twenty-first somite (Part
III, figs. 42, 44), which does not divide into a dorsal and ventral
part but acquires a ventral opening in the same position as the
preceding somites. The opening soon, however, shifts to the
middle line, where it joins its fellow, so as to form the single
generative opening of the adult. In all probability the greater
part, if not all of the ducts of the adult, are derived from the
twenty-first somite ; the dorsal divisions of the five preceding
somites forming the generative glands only.!

In Stage c the generative organs form two tubes lying in
the central compartment of the body cavity and closely applied
to one another in the middle line (Part III, figs. 47, 48).

I cannot say when the generative cells begin to show sexual
differences. The appearance of the sections referred to (Part
III, figs. 47, 48) would lead one to suppose that the specimen
was a female, and I have but little doubt that it was. At the
same time, I must mention that I have never seen anything at
this stage which I could call a male.

In January sexual differences are undoubtedly manifested by
the generative tubes. Those of the females presented very
much the appearance of the earlier stage. In the male the
nuclei were smaller and more numerous, and the lumen was
narrower. In fact, the organs differed very much in the same
way that they do in ripe embryos. Pl. XXVII, fig. 12, is from
a transverse section of a female embryo almost ready for birth,
and fig. 13 from a male embryo of the same age. At this

1 This view would be confirmed in the case of the female if it could be
proved that my suggestion (Part III, p. 514) that the receptaculum ovorum
of the neotropical species is part of the somite which gives rise to the genera-
tive opening and outer part of the generative ducts, and homologous with
the internal vesicle of the nephridia. The case of the male is more difficult,
but probably the testes (prostates of Moseley and Balfour) only, ¢. e. the
parts in front of the swollen vesicule seminales (testes of Moseley and
Balfour), are alone derived from the dorsal divisions of the generative somites,

390 ADAM SEDGWICK.

stage the ovarian tubes communicate with one another at their
extreme front ends, and behind where they pass into the ovi-
ducts. In the male the tubes are considerably twisted ; I could
not make out any distinct trace of the vesicula seminalis.

I regret to say that I have not paid much attention to the
histological development of the sexual glands. The first trace
of the sexual organs is the round nuclei of the endoderm.
When and how these acquire a cell body I cannot say. They
certainly have the latter by Stage e (Part III, figs. 47, 48).
The follicular nuclei are the nuclei of the splanchnic meso-
derm, which closely apply themselves to the germinal nuclei as
soon as the latter emerge from the endoderm. The follicular
nuclei appear, therefore, before the protoplasm of the sexual
cells (Part III, fig. 26). In Stage eG areas of protoplasm, in-
distinctly marked off from one another, could be distinguished
round the larger nuclei (Part III, figs. 47, 48). In the
females of the stage just before birth the boundaries of these
areas were slightly more marked, but still indistinct (Pl.
XXVII, fig. 12). In the male of this stage there are no lines
separating the protoplasm round the granular nuclei of the
testes into areas (Pl. XXVIII, fig. 13).

The general bearing of the facts of development of
the ccelom and body cavity of Peripatus is fully dwelt
upon in Part III of this series. I have but little
doubt that the same method of development will be
found in other Arthropoda. If I am right in this view
it must be admitted that the Arthropoda are cclomate
animals, that their generative cells are products of the
coelomic epithelium, and that the generative ducts are modi-
fied nephridia.

The ccelom of Peripatus does not extend into a perivisceral
or body cavity, but remains small, discharging only the func-
tions of excretion and reproduction. The functions of a peri-
visceral cavity are discharged by the vascular system, in which
indeed the cceelom is contained (Pl. XXVII, figs. 10, 11, 13,
and diagram, fig. 17) in exactly the same way as the intestine
of a mammal is contained in the ccelomic body cavity. The

:

OO

DEVELOPMENT OF THE CAPE SPECIES OF PERIPATUS. 391

condition of the ceelom and vascular tracts in the adult and
the relation of the ccelom to the vascular body cavity is clearly
illustrated by the diagram (Pl. XXIX, fig. 17). Itis commonly
said that in the Arthropoda the generative ducts are continu-
ous with the glands, and in this they are contrasted with the
Annelida and Vertebrata. As a matter of fact, however, the
generative ducts, in Peripatus at least, present exactly the
same relation to the generative glands as do the oviducts of a
dogfish or earthworm to the ovaries of these animals; that
is to say, like the latter, the generative ducts open into the
ceelom, and the ova are products of the ccelomic epithelium.

It is important to notice that in Peripatus the nephridia
are parts of the celom (Pl. XXIX, diagram, fig. 17), just as
they are in Elasmobranchs. They are commonly spoken of in
a manner which implies that they have but little to do with
the celom beyond opening into it. This way of speaking
of them is calculated to mislead. The nephridia are
direct differentiations of part of the celom (dia-
grams, figs. 13—17, and figures in Part III illustrating their
development).

A negative feature, which has often been put forward as
characteristic of the Arthropoda, is the apparent absence of
nephridia. The nephridia of Peripatus have generally been
considered as a primitive and peculiar feature. Lankester,!
however, some time ago (No. 54, p. 516), suggested that the
coxal glands of Limulus and the antennary glands of Crus-
tacea were nephridia, and that the peculiar ‘‘ end-sacs”
described by Gulland in the coxal glands of the young
Limulus, and the internal vesicle of the Crustacean anten-

1 Lankester’s words were: ‘‘ The observations here recorded on the struc-
ture and connections of the immature coxal gland of Limulus tend to render
it probable that the green glands of Crustacea are also to be regarded as a
pair of modified nephridia;” and he goes on to say that ‘it seems not
improbable that the so-called end-sac of these glands is not part of the
nephridium, but is developed from the connective-tissue space (coelomic space)
into which the true tubular nephridium opens.” ‘The distinction implied by
these last-quoted words between the nephridium and celom is not justified
by embryology, as I have just pointed out.

392 ADAM SEDGWICK.

nary gland described by Grobben (No. 55), were part of a true
coelomic space. The discovery of the end-sacs in Peripatus,
and of their method of development, entirely confirms Lan-
kester’s view. And it is interesting to notice that the end-
sac of the Crustacean green gland, as figured by Grobben,
resembles somewhat in the structure of its wall the end-sacs of
the Peripatus nephridia.

In the Leeches the nephridia are stated by Bourne (No. 4)
and other observers to communicate with the vascular system.
I think, however, it is worth while to bear in mind the pos-
sibility of there having been a mistake on this point. It is
possible that the nephridia in the Leeches may, as in Peri-
patus, end in a closed vesicle which lies in, but does not
open, into the vascular system. Such vesicles might quite well
have been overlooked by the able observers who have dealt
with the subject, as they undoubtedly were overlooked in
Peripatus.

It is interesting to notice the resemblance which would
exist between the transverse section of Peripatus (Pl. XXIX,
fig. 17) and the transverse section of a Leech, if the blood
tracts of the former were more broken up, and the nephridia
of the latter did not open internally into the vascular
system.

List oF PAPERS REFERRED TO.

1. Kennet, J.—‘‘ Entwickelungsgeschichte v. Peripatus Edwardsii u.
torquatus,” Theil ii, ‘Semper’s Arbeiten,’ Bd. vii.

2. Batrour, F. M.—‘ Anatomy and Development of Peripatus Capen-
sis,” ‘Quart. Journ. Micr. Sci.,’ vol. xxiii.

3. Garrron, E.— Beitrage zur Anatomie u. Histologie v. Peripatus,”’
‘ Schneider’s Zoologische Beitrage,’ Bd. i.

4. Bourng, A. G.—‘ Contribution to the Anatomy of the Hirudinea,”
‘Quart. Journ. Micr. Sci.,’ vol. xxiv.

ad

wae “YD
ee ee

a

—

DEVELOPMENT OF THE OAPE SPECIES OF PERIPATUS. 393

EXPLANATION OF PLATES XXVI, XXVII, XXVIII,
and XXIX,

Illustrating Mr. Adam Sedgwick’s paper on “ The Develop-
ment of the Cape Species of Peripatus. The Changes from
Stage eG to Birth.”

List of Reference Letters.

a. Dorso-lateral horn of white matter of brain. J. Ventro-lateral horn of
white matter of brain. . 4. app. Body cavity of appendage, a blood-space.
b. lat. Lateral compartment of body cavity (space formed in parietal meso-
derm); a blood-space. cer. ves. Cerebral vesicles or ventral appendages of
brain. circ. muse. Circular muscles of body wall. crur. gi. Rudiment of
crural gland. 2. Central lobe of white matter of brain. J/. s. ¢. 3. Tubular
part of nephridium of third somite. 7. s. v. 3. Internal vesicle of nephridium
of third somite. xeuro-musc. Network of fibres, so-called because it gives rise
to nerves and muscles; it is continuous with the lateral pedal nerve, and is
apparently composed of a substance like the latter and the white matter of the
cord. It is derived from a compact mass of nuclei present in the previous
stages at the same point. or. pap. Oral papilla. o. s. 3. External opening of
nephridium of the third somite. sad. gl. Salivary gland, — s/. g/. Slime-gland.
v.o. Ventral organ. v.o.1. Ventral organ of jaws. v.s. Septum separating
the central from the lateral compartments of the body cavity, called in earlier
stages the ventral sheet of somatic mesoderm.

Fics. 1—4,—A series of transverse sections through the head of an embryo
of Peripatus Balfourii of Stage G (removed from the uterus on 12th
December). One side only of each section is completely drawn. The dorsal
ectoderm possesses a large number of highly refractile spheres, probably yolk-
spheres, and has contracted on to the brain so as to render indistinct the
mesoderm structures between. a. Dorso-lateral. 4%. Ventro-lateral. e.
Central lobe of white matter.

Fig. 1. Through the region of the eye and anterior lobes of the brain.
One side only is figured, and the nuclei of the ventral part of the brain
are omitted. The section goes through the optic nerve and centre of
the optic vesicle. Zeiss’s camera, obj. D, oc. 2.

Fig. 2. A little farther back, through the anterior part of the cerebral
vesicle (cer. ves.) and the region of junction of the two halves of the
brain. The dorsal mass of nuclei and the central lobe of white matter
have united with the corresponding structures of the opposite side.
Veutrally the two halves of the brain are separate.

394 ADAM SEDGWICK.

Fig. 3. A little farther back, through the centre of the cerebral vesicles,
and still through the joined part. The dorsal mass of nuclei are absent,
and the white matter is broadly exposed dorsally.

Fig. 4. Through the posterior lobes of the brain and the region of the
buccal cavity. The ventral organ of the jaw (v. 0. 1) and the jaw are
shown. The oral papilla (or. pap.) and slime-gland (s/. g/.) are also
visible.

Figs. 2—4 drawn with Zeiss’s camera, obj. C, oc. 2.

Fic. 5.—Portion of transverse section through the middle region of the
body of an embryo of Peripatus capensis of Stage @ (removed from
uterus 16th December). The section passes through the anterior part of a
pair of legs. The details are filled in on the left-hand side of the drawing.
The section passes through a ventral organ (v. o.), with which the nerve-cords
are connected by a cellular process, a persistent trace of the original complete
continuity between these two structures. The anterior of the two large pedal
nerves is shown (zerve), leaving the cord as it immediately passes forwards out
of the plane of the section; the continuity between it and the neuro-muscular
network (xeuro-musc.), which will eventually develop into the nerves and
muscles of the foot, and between it and the circularly disposed network from
which the circular muscles (circ. musc.) of the body wall will develop, could
not be shown in this figure. The walls of the alimentary canal are very thin.
6. lat. marks the lateral compartment of the vascular body cavity (space
formed in parietal mesoderm of early stages). 4. app. The vascular body cavity
of the appendage. Drawn on the table, Zeiss’s new camera C, oc. 2.

Fic. 6.—A diagram of the ventral portion of the third somite at Stage F,
The vertical lines indicate the planes of the sections Figs. 37, 38, 39 of
Parhlt i.

Fie. 7.—Diagram of the ventral portion (nephridium) of third somite at
birth. The hinder part of the tube of the preceding stage has elongated
backwards to form the long tubular salivary gland (sad. g/.).

Fics, 8—18 are all from embryos just before birth (removed from uterus
19th April). They were all drawn on the table with a Zeiss’s new camera,
obj. C, oc. 2.

Figs. 8 and 9. Transverse sections through the same structure, along the
lines marked 8 and 9 in the preceding diagram (Fig. 7). The vesicle
(¢.s.v.3) has thick, much vacuolated walls, and is placed dorsal to
the tubular part (/.s. ¢. 3). In Fig. 9, the small portion connecting
the vesicle and tube is shown closely applied to the dorsal wall of the
salivary gland.

Fig. 10. Portion of a transverse section through the region of the third
leg of P. capensis just before birth. The section passes through
the opening of the nephridium, and what I take to be the rudiment of

DEVELOPMENT OF THE CAPE SPECIES OF PERIPATUS. 395

the crural gland (erwr. g/.). The whole nephridium is shown in the
section. In the actual section the internal vesicle was somewhat more
collapsed than in the figure. The funnel and internal vesicle I take
to be homologous with the similarly named structures of the posterior
nephridia.

Fig. 11. Portion of a transverse section through the region of one of the
posterior legs of P. capensis just before birth. The section passes
through the opening of the nephridium and the rudimentary crural
gland. The tubular part of the nephridium is cut in two places, and
its terminal portion with the closely-packed nuclei is shown opening
into the internal vesicle. The cavity of the leg is traversed by a con-
siderable amount of muscular tissue external to the nephridium, and
contains some of the reticular tissue in its dorsal part. The septum
separating the lateral sinus from the leg cavity is absent in this region.

Fig. 12. Section through the ovaries of an embryo of P. capensis just
before birth. On the right side there is an attachment to the peri-
cardium. ‘The dark elongated nuclei are follicular nuclei.

Fig. 18. Dorsal part of a transverse section of amale of P. capensis just
before birth. In most embryos of this age the wall of the heart is
separate (always connected by filaments) from the dorsal body wall and
the pericardial floor. The apparent fusion in this section was pro-
bably due to contraction of the body. The two patches of dorsal
longitudinal muscles are shown.

Fics. 14—17 are a series of diagrams to show the relations of the ccelom
and body cavity at successive stages. The lining of the ccelom is shaded dark,
the light shading indicates the general mesoderm.

Fig. 14. Earliest stage : ccelom as a series of separate spaces. The ecto-
derm and endoderm still in contact: no trace of the vascular space or
hemoceele.

Fig. 14a. The endoderm has already separated from the dorsal and ventral
ectoderm. On the right-hand side the somite has not yet divided into
a dorsal and ventral portion. The first rudiment of the lateral sinus
is present in the thickened mesoderm. On the left side the ccelom has
divided into a dorsal part, which in the anterior part of the body
vanishes, but in the posterior part becomes the generative organ;
and into a ventral part which becomesa nephridium. The lateral sinus
has increased in size, and a space has appeared in the mesoderm of
the leg.

Fig. 15. The dorsal division of the ccelom has passed dorsalwards, and
considerably encroached upon the dorsal of the two blood-spaces formed
by the separation of the endoderm and ectoderm. This median dorsal
blood-space becomes the heart. Two spaces have appeared in the
ventral wall of the dorsal division of the somites; the dorsal of these

396 ADAM SEDGWIOK.

becomes the pericardium, the ventral the dorsal part of the central
compartment of the body cavity. The ventral blood-spaces have in-
creased in size, but are otherwise unchanged.

Fig. 16. The dorsal divisions of the somites have relatively diminislied in
size, and been overlapped dorsally by the greatly increased pericardium.
The dorsal part of the central compartment of the body cavity has
increased in size, but is still separated ventrally by a septum. The
ventral blood-spaces have increased in size. The ventral division of
the ccelom is assuming the form of a nephridium.

Fig. 17. Diagram of arrangement at birth. The two halves of the peri-
cardial cavity have coalesced dorsally and ventrally to the median-
dorsal blood-space which forms the heart. The dorsal divisions of the
ccelom have become constricted off from the floor of the pericardium
and the dorsal wall of the enteron, and now lie in the central compart-
ment of the body-cavity and constitute the generative tubes (in the
anterior part of the body they atrophy in the floor of the pericardium).
The horizontal septum separating the dorsal division of the central
compartment of the body cavity from the ventral has vanished. The
ventral or nephridial division of the ccelom has assumed the form of a
nephridium. On the left hand the whole course of the nephridium is
diagrammatically shown.

ws

cer. ves

“J
\
\
\
—~
Bi
“
J
<—— 5
~ 9 6

Te ee

or. pap.

d Mor Fouurry Veh. XXVI NSFLKLVL.

Enteron | :

*@e2,
2a9 @ 22 98 9

“oy,

F Huth, Lith? Edin*

=
=

é

ti

at

0.S.3 12
i

A : eae 2,
F : : g
crur. gl. i Epidermis 605 Peg

opening of nephridium.

-- CITC. Mus.

Mor Sourn, Vol XU NS FL XYVI

--\-_.

external opening of
nephridium.

crur. él.

a muscles of leé

F Huth, Lith? Edin?

athe

Mor Fourn VA XIV NS LLL

has

‘trem Apog jo

wiapopue
~ Sa[DSNUL IepNdIL9

Sl bug ne
‘Kyneo Kpoq jo es ce,
juswyreduioo [esjuad / oe

~ : : s1jsaq

\
umpacatiad

“OSTUL “O419

wiapojoa

‘wnipeotied Jo ‘anssty

JeaponNUurjjnwU % Ie[Noye1 ‘g]osnu peuipnytfuoy

it ad

7 oywios jo Y
f7 Woistatp [esyuan

H- STULS [2.1972]

ooeds pooyg

s$eq jo Auweo Kpog -WaX--- aoeds poorg

STLULS ]e1a}e] ---¥

A fyrweo Kpog jo uaujseduios [esjueo
ainyng jo yred yestop jo JUSWIPNL

* fyaeo yetprestiad wawtpns 4st

--- @}IULOS JO UOISIAIP [esiop

rec fyravo Apo jo woujsedus0s
Daf

ae 7 ...419y ailnyoy

ajIULOS JO UOISIATP [PSAOP -
S}IULOS jo UOISLALP [esp UsA

Kytweo jeipreotsed ---
aoeds poojq
qreey aanynz - x eae by

Fm yeeed seepuadde yo Aytneo Apog

CARES a”
P
a
/

U019}U8
-1.9W

‘yred uois}Uusa
[e1jUsA ¥ ]esLop @
OyUL SULPIAIp a}LULOS

‘oyLULOS Jo
UOTSIArp TesLOp

or

nw soeds poorq

umipliydau
= W0[Mo Jo
UOISIATP [B4}UIA

oo. \ ( =e | - 4 1009
“ge, “waspapae Y |
/ aoeds poojq : ‘ > . | '

‘Sal | soeds poo
jo Ayaeo Xpog --f- na}... E a

Moor Fourn Cet: XX NS TY, HE

SNULS ]219}R[ -
uo1ajue

‘puRTs awUiys ~\

Ayraeo Kpog jo
jueujseduios peayuas---

EO

S9QN} Uolye19u9s = WO]Hd

JO WOTSTAID [@SLOp-~ WARRRepgS 78 -raoen nee reteeeacenceseneen caceesencnze Ad 7 bog

a et tt re rn ie na ee ee Ye ee A PO en aren as

wnipieoised weeeeee - ERR beak
X aoeds poolq

aoeds poolq

NEPHRIDIA OF CERTAIN EARTHWORMS. 397

On the Occurrence of numerous Nephridia in
the same Segment in Certain Earthworms,
and on the Relationship between the Excre-
tory System in the Annelida and in the
Platyhelminths.

By

Frank E. Beddard, M.A.,

Prosector to the Zoological Society of London, and Lecturer on Biology at
Guy’s Hospital.

With Plates XXX and XXXI.

Acanthodrilus multiporus, F. EF. B.

Tue nephridia of this worm differ in their appearance from
those of any other species of Acanthodrilus at present
known. In a dissection of the worm the nephridia can be
recognised as a series of glandular tufts closely adherent to the
body wall and to the intersegmental septa; each of the eight
setze often appeared to have a nephridial tuft specially related
to it: thus by dissection there seemed to be eight separate
nephridia in each segment. This, however, is not invariably
the case, and in the anterior segments no relation between
the setz and the nephridia could be detected by dissection.

I have also studied these organs by transverse and longi-
tudinal sections.

By these means I have been able to ascertain that there
are more than one pair of nephridiopores in each seg-
ment of the body.

These observations depend upon the study of very well-

1 © Proce. Zool. Soc.,’ 1885.

398 FRANK E. BEDDARD.

preserved material, for which I am indebted to Professor
Parker, of Otago. One of the numerous specimens which he
sent me was so preserved that the nuclei, not only of the
nephridia but of the tissues generally, were stained dark yellow,
and could not be subsequently stained by borax carmine ;
accordingly the nuclei were extremely conspicuous in such pre-
parations ; and as the nuclei of the nephridial tubules differed
in size and shape from other nuclei (e.g. those of peritoneal
cells and blood-capillaries), the course of the tubules through
the body walls could be easily followed. It must not, how-
ever, be inferred that.such drawings (e. g. fig. 1) in the plate
as illustrate the structure of the nephridial tubules are imagi-
nary, except as concerns the nuclei. They are in every case
accurate drawiugs, so far as I could make them ; but the con-
spicuous nuclei drew my attention to the delicate inconspicuous
walls of the tubule, which would otherwise almost inevitably
have been overlooked.

The nephridia are arranged, as already stated, in tufts of
tubules, closely adherent to the body wall (see figs. 2, 8);
their structure is much like that of other earthworms; they
consist of tubules with an intracellular duct, some (fig. 2’)
larger than others (fig. x). The nephridial tufts are supplied
with abundant blood-capillaries, upon some of which are those
curious dilatations to which Lankester first drew attention in
Lumbricus. In spite of the most careful search I have
hitherto failed to find any trace of the internal openings of
the nephridia.

The external orifices, however, are obvious enough. The
disposition of these is illustrated in figs. ], 2, 3. Figs. 1, 2,3
represent actual sections, or rather are compiled from a number
of sections. Fig. 15 is schematic, and represent what I
believe to be the general arrangement of the nephridia in
certain segments of the body.

Figs. 2, 3 are copied from sections through certain of the
post-clitellar segments, fig. 1 the 8th or 9th.

It will be noticed that the external pores are decidedly more
numerous in the anterior segments of the body, while in the

NEPHRIDIA OF CERTAIN EARTHWORMS. 599

posterior segments they tend to be arranged in accordance
with the number of the setz, i.e. eight per somite. This dis-
position of the nephridiopores is associated with a correspond-
ing regularity of the nephridial tufts, which tend to break up
in the posterior segments into eight separate nephridia. There
is, however, a connection between them ~.s shown in figs. 2, 3.
A series of tubules (figs. 1 2 and 5») run from seta to seta, and
unite the nephridial tufts of these setz ; this tubule for the most
part runs within the peritoneum. ;

The appearances presented, in fact, can nardly be explained,
except on the assumption of a network of nephridial
tubules. In no Annelid, except in Pontobdella (Bourne
5) has the nephridial system been recorded to be a network.
I can, however, find no evidence that the nephridial network
of one side of the body is continuous with that of the other ;
nor does there appear to be any connection between the
nephridia of successive segments.

The tubules leading to the exterior are, in the posterior
segments, nearly invariably related to the sete; occasionally
(fig. 2a) a tubule was observed to perforate the body wall
between the sete, but in this case it must be noted that the
tubule passes up a septum of connective tissue, which at this
point breaks the continuity of the longitudinal muscle layer.
It may be that such septa, which occur here and there, represent
the last trace of setee which have now disappeared. Sometimes
(fig. 3 aa’) I noticed two nephridial tubules belonging to the
same seta. I did not succeed in observing with certainty whether
or not they opened on to the exterior by distinct orifices. The
ductule is usually embedded in the lax connective tissue which
surrounds the insertion of the seta (fig. 4); it passes up as far
as the circular muscular layer without undergoing any changes
in character, that is to say, the duct is intracellular. At
this point the tubule often runs for some distance along
the junction between the circular and longitudinal muscles
before perforating the former, and opening on to the exterior ;
in other cases (fig. 4 5) the tubule at once perforates the circular
muscular layer and the epidermis, and opens on to the exterior.

VOL, XXVIII, PART 3,—NEW SER, EE

4.00 FRANK FE. BEDDARD.

The distal section of the nephridium where it traverses the
circular muscles is somewhat wider than the rest, and is lined
by a single layer of small, delicate cells ; the duct in this region
is inter-cellular. Occasionally the distal section of the
nephridium is branched (fig. 2 4) and opens by several (three)
distinct orifices.

In the anterior region of the body the nephridiopores of
each segment are more numerous; they could be easily ob-
served by a simple examination of the cuticle, and lhe between
the setz in a more or less regular row. There were over
one hundred apertures in each of several segments
that I examined in this way.

In transverse sections the nephridial tubules were seen to
pass through the body wall, not only in the immediate neigh-
pourhood of the seta (fig. 1), butelsewhere. Very constantly
the nephridial tubule did not open by a single pore, but
branched, at the junction of the circular and longitudinal
muscles, and opened on the exterior by a number of distinct
orifices (fig. 10). This branching of the nephridial duct
within the substance of the body wall is paralleled in the case
of Capitella, where Fischer (8) has recorded a similar
branching.

Fig. 15 represents a diagram of what I believe to be the
arrangement of the nephridia in one of the posterior segments
of A. multiporus; it may be compared with fig. 14, which is
a corresponding diagram of Pericheta. I am not able to
give a satisfactory diagram of the anterior part of the body in
Acanthodrilus.

Pericheta, Schmarda.

The peculiar “ tufted” condition of the nephridia in this
genus was first made known by Perrier, and has since been
commented upon by others. This appearance of the nephridia
led me to expect that I should find numerous nephridiopores
in each segment. I was, however, unable to put this suppo-
sition to the test, until I received quite recently a number of

NEPHRIDIA OF CERTAIN EARTHWORMS. 401

excellently preserved examples from Bermuda, through the
kindness of Mr. Shipley, Fellow of Christ’s College, Cam-
bridge. These specimens appear to be identical with Perrier’s
P. aspergillum. The nephridiopores could be easily ob-
served, as in the case of Acanthodrilus multiporus, upon
the cuticle when stripped off and examined in water. I may
take this opportunity of saying that there are other structures
which might possibly be confounded with the nephridiopores ;
these are the impressions of sense cells upon the cuticle!
which in Pericheta often lie between the sete. A careful
examination, however, soon serves to discriminate between the
two ; the circular bulgings of the cuticle, due to the prominent
sense organs, although presenting the appearance of pores
under a low power, are readily seen under a high power not to
be pores at all, while the nephridiopores are clearly holes in
the cuticle.

The latter form a more or less irregular row surrounding
each segment and lying between the sete (fig. 12). Often
there are four or five between two sete. In transverse sections
each pore is seen to be continuous with a nephridial tubule ;
the actual orifice is surrounded (fig. 8) by a very few large
cells—smaller, however, than the cells of the epidermis ;
below these are smaller cells; that section of the tubule
which lies within the circular muscle layer, has a wide lumen
whose walls are made up of excessively delicate flattened cells.
At the junction of the circular and longitudinal fibres the
duct becomes intracellular (fig. 10). The nephridial tubules
are abundantly supplied with blood-vessels (figs. 10, 11) in their
passage through the body walls. The nephridial tubules agree
with those of A. multiporus in the fact that the duct becomes
inter-cellular at the junction of the longitudinal and circular
muscles (cf. figs. 6 and 10). Fig. 9 represents the actual
number of nephridial tubules and their external orifices in a
small portion of the body wall; the figure would be equally
correct, so far as my observations go, for any part of the body
wall. I should state, however, that I have at present only

1 Vejdovsky (13), pl. xv, fig. 13a.

402 FRANK E. BEDDARD.

studied the anterior region of the body comprising the first
sixteen or seventeen segments.

I have never observed any branching of the nephridial
tubules in the body wall like that which occurs in Acantho-
drilus (see fig. 1); but I am not prepared to deny that it
may occasionally happen.

So far the nephridial system of Pericheta is closely
similar to that of Acanthodrilus, with only some small
differences in detail.

An important point of difference is illustrated in fig. 7; in
this figure it will be seen that the nephridial network of
one segment is continuous through the septum with
that of the next. ‘The nephridial system in fact differs from
that of any other earthworm, or indeed from that of any other
Annelid except Pontobdella (Bourne) (5), in that it forms a
continuous network uninterrupted by the inter-
segmental septa. Furthermore, there appears to be a
perfect continuity between the nephridial system of the right
and left halves of the body. The nephridial system of one
segment communicates with that of the neighbouring seg-
ments at numerous points; there is no question here of a pair
of longitudinal ducts such as has been described recently in
Lanice conchilega (Lang (11), Cunningham 6). The
resemblance to Pontobdella is very striking, the only dif-
ference being that the external apertures in each segment are
numerous instead of being only a single pair. With regard
to internal apertures (funnels) I can say nothing. I have
been entirely unable to find any trace of them.

The diagrams (fig. 13) illustrates what I believe to be the
arrangement of the nephridial network and the external pores
(0) in a few segments. I should say that I cannot pretend to
accuracy in the course of the tubules within the celom. To
reconstruct from a series of sections the course of the com-
paratively simple nephridium of Lumbricus is a_ hard
enough task ; to map out accurately the extremely complicated
coils of the nephridial network of Pericheta is for me an
absolute impossibility. I can, however, assert most positively

NEPHRIDIA OF CERTAIN EARTHWORMS. 403

that the external apertures do not each correspond to a
nephridium isolated from its fellows ; the nephridial tufts form
a continuous layer coating the body wall; there must therefore
be a network, although it may not be of course so complete a
network as I have sketched in the figure just referred to.

Typheus, F. E. B., and Dichogaster, n. g.

I have observed numerous nephridiopores in a single seg-
ment in both these genera; at present I can only record the
fact without giving any details, which I hope to be able to do
later.

The Excretory Organs of Annelids and
Platyhelminths.

Two views have been held respecting the relationship of the
nephridia in Annelids to those of the Platyhelminths. Ge-
genbaur has held (Comp. Anat.,’? Engl. trans., p. 177) that
there can be no direct homology between the nephridia in the
two groups, because, in the Hirudinea at least, these organs
are preceded by several pairs of embryonic excretory tubes.
On the other hand, Lang and others believe that “ the excretory
system of the Platyhelminths is the starting point of that of
all higher animals,” that in fact the nephridia of Annelids are
strictly comparable to those of the unsegmented worms.

The former view has recently been extended and defended
with great ability by Dr. R. S. Bergh, of Copenhagen (4).
The latter view is the one which perhaps finds the most
general acceptance among comparative anatomists. Bergh
uses arguments which tend to prove that the provisional larval
excretory organs of Polygordius, of Hupomatus, and other
Polycheta, of the Hirudinea, of Echiurus, of certain Oligo-
cheta (lately discovered by Vejdovsky), and of Mollusca, are
homologous with each other and with the excretory system of
the Platyhelminths ; the permanent nephridia of these various
groups are therefore to be regarded as new formations (in the
Annelida), or possibly to be derived from the genital ducts of

4.04. FRANK E. BEDDARD.

the flat-worms. Great stress is laid upon the structural
identity of the Platyhelminth excretory system with the
“head kidney” of larval Polygordius, Chetopods, and
Gephyreans. In both there is a system of fine branched tubes
formed of perforated cells, which unite to form a pair of wider
tubes opening on to the exterior ; the internal apertures open
through a single cell, the so-called ‘flame cell;” in many of
the larval forms the peculiarly modified flame cell is absent,
but the aperture is plugged by a single cell. In the Hiru-
dinea and Oligocheta there is no longer a special anatomical
resemblance to the Platyhelminth; but this is due to the
rudimentary condition of the larval excretory organ, rudi-
mentary because there is no free swimming larva.

Another argument is drawn from the apparent independence
of the larval and permanent excretory organs in Annelids ; the
fact that in many cases there is no connection between the
larval nephridia and those of the adult, and the want of any
confirmation of Hatchek’s statements to the contrary, indi-
rectly favour the hypothesis.

Again, it is pointed out that while the larval nephridia of
Polygordius lie in the head cavity, the persistent nephridia
lie in the celom ; and these two cavities are not similar. On
the other hand, the Platyhelminths have no true celom, only
a series of minute clefts and spaces in the mesodermic tissue
with which the flame cells are connected, and which form a
“primary body cavity,” comparable to the primary body cavity
or head cavity of the Annelid larva.

Dr. Bergh’s conclusions largely depend upon his own
researches into the nephridia of MHirudinea; these have
been recently criticised by Whitman (15), who meets Bergh’s
assertion that the larval nephridia have no connection with the
permanent nephridia by an equally positive assertion that
they have. With regard to the larval excretory organs of the
Oligocheta, which have been discovered by Vejdovsky (13) in
a variety of types, it is to my mind a significant fact that they
occur at the anterior end of the body where no permanent
nephridia are developed ; furthermore, these organs lie in the

NEPHRIDIA OF CERTAIN EARTHWORMS. 405

celom perforating the mesentery which separates the first
from the second segment;! hence Bergh’s objection to the
homology between the larval and permanent nephridia, on the
score that the former do not lie in the true ceelom, is removed.

It is not yet a universal belief that a cclom is absent in
the Platyhelminths ; in any case, as Lang says, “the entire
mesoderm of the Platyhelminths may very well be the equiva-
lent of the entire mesoderm in the Enterocela,’ apart
altogether from the question of the enclosed cavity. In
Capitellide the nephridia seem, from Eisig’s description, to lie
in the somatopleure, i.e. not in the coelomic cavity at all.
The greater portion of the nephridium in Polygordius is
similarly situated ; and Fraipont in his work upon the Archi-
annelida (9) remarks upon this as an archaic character.
Again, Weldon (14) has brought forward reasons for believing
that in Dinophilus, “ which is literally no more than a larval
Annelid with reproductive organs” (Lang), the body cavity
may be really strictly comparable with that of Annelids, seeing
thatin Saccocirrus—an undoubted Annelid—the body cavity
may be secondarily (?) invaded by a ramifying network of
mesodermic cells.

In Dinophilus there are excretory organs of the Platy-
helminth type, branched and ending in flame cells, which are
in D. gyrociliatus (according to Ed. Meyer, quoted by
Lang) metamerically arranged, with a pair of external aper-
tures to each segment. These facts taken together appear to
me to do away with many of the difficulties urged by Bergh
against the comparison of the Platyhelminth and Annelid
excretory system.

A difficulty in comparing the Platyhelminth with the
Annelid excretory system is undoubtedly in certain structural
differences; there is no trace in any Platyhelminth of a
“funnel;’ the excretory tubules end in the well-known
“flame cells.” This difficulty has led Lang (11) to regard the
funnel in the Annelid as a new structure unrepresented in the
Platyhelminths; this opinion is supported by the observations

1 Vejdovsky (13), pl. XVI, fig. 6.

4.06 FRANK E. BEDDARD.

of E. Meyer and of Vejdovsky (13), that in Polychzta and
Oligochzta the funnel is developed perfectly independently of
the rest of the nephridium. It was important, however, to
note that not only is the nephridial funnel formed by the
proliferation of a single cell, but that in Stylaria this cell
becomes ciliated and acquires a lumen before it
undergoes division ; repeating in fact the presumed ancestral
condition. With regard to the presence of a single flagellum in
the ‘‘ flame cells,” Hartog has recently pointed out (10 4) that
the optical effect produced by the motion of minute cilia might
readily give the impression of a single flagellum. In the
adult Clepsine (Bourne 5) the funnel consists of only two
cells, so that it is but a stage removed from the Platyhelminth
funnel. The above facts lend at least some support to the
views that it is unnecessary to regard the funnels of
the Annelida as new structures.

The facts recorded in the present paper fit in very well with
the supposition that the Annelid excretory system is directly
traceable to that of the Platyhelminth; I shall suggest, how-
ever, a course of development rather different from that put
forward by Lang.

Lang has pointed out that in certain Planarians a commenc-
ing metamerism (parallel with the commencing metamerism of
the generative organs, &c.) is visible in the excretory system.
Here and there “secondary ” external orifices are formed
through branches given off from the longitudinal trunks of
2ither side; at first irregular, these secondary pores finally
become regularly disposed, and a paired arrangement even is
seen; the disappearance of at least the greater part of the
network (and the development of internal funnels) brings
about the conditions which occur in Lanice conchilega; in
this Annelid, Meyer (11) and Cunningham (6) have shown that
the longitudinal trunks connecting the nephridia of successive
segments persist ; Vejdovsky also (13) has figured traces of this
same longitudinal duct in Anacheta; and Wilson has
recently (16) been able to prove that the embryo Lumbricus,
like the embryo Criodrilus, possesses a similar longitudinal

NEPHRIDIA OF CERTAIN EARTHWORMS. 407

duct, one on each side of the body. Finally, the disappearance
of the longitudinal duct leaves the paired nephridia of the
majority of Annelids. The remarkable disposition of the
nephridia of Pontobdella, described by Bourne (5), is dif-
ferent in that there is not a single longitudinal duct on each
side, but a continuous network. It resembles, in fact, as Lang
has pointed out, the branched and anastomosing canals of the
Trematode nephridium ; it must also be remembered that the
longitudinal canals of Gunda segmentata break up here
and there into a network of anastomosing canals. It is this
condition which has been inherited by Pontobdella, and
also by Pericheta; Pericheta appears to me to represent
a more archaic structure than Pontobdella since the external
pores are more numerous; in Pontobdella they have become
reduced to a single pair in each somite. Acanthodrilus
multiporus offers the next stage; here the nephridial net-
work is not continuous from segment to segment, but there is
a separate network for each segment opening by numerous
external pores. The Capitellide (Hisig 7) agree in many particu-
lars with Acanthodrilus; the nephridia of each segment are
numerous, and occasionally are connected; the ducts opening
on to the exterior are often (Fischer 8) branched and open by
several distinct pores. In the posterior region of the body
of Acanthodrilus the external apertures are fewer, and are
more particularly connected with the setz, while the nephridial
network tends to become arranged in a series of tufts, corre-
sponding in number to the setz; finally, the nephridial net-
work of each half of the segment is independent. From
Acanthodrilus to Lumbricus there is a considerable gap,
only very partially bridged over by such forms as Plutellus
(Perrier 12), where the irregularity in the position of the
nephridial pores is, perhaps, to be regarded as a last trace of
the numerous nephridial pores of Acanthodrilus and
Pericheta. The above considerations necessitate a further
inquiry into the nature of the longitudinal canal in Lanice,
&e.; if it be assumed that this latter is the homologue of the
longitudinal canal in the Platyhelminths, it must follow that the

408 FRANK E. BEDDARD.

Annelida have been derived from at least two Platyhelminth
ancestors, i.e. one with a pair of longitudinal ducts, and another
with a network of ducts. This would not perhaps be an un-
reasonable supposition were it not for the fact, that in this
case Lumbricus must have had a different origin from
Pericheta. The nature of this longitudinal duct has been
put in a fresh light by the observations of Wilson (16) ; accord-
ing to this author it arises from the epiblast, and is therefore
comparable to the archinephric duct of Vertebrates, to which
the observations of several writers have concurred in assigning
an epiblastic origin. Haddon has recently (10) ingeniously
suggested how a continuous groove, into which the nephridia
opened, may have been converted into a canal opening into
the cloaca. But while in the Vertebrata it seems to be generally
agreed that the nephridial tubules are of mesoblastic origin,!
Wilson states, that in Lumbricus they are budded out from
the longitudinal duct, and are therefore epiblastic, with the
exception of the funnel, which is mesoblastic. In any case his
investigations only relate to early stages; and there is therefore
nothing to negative the supposition that the portion which
arises from the longitudinal duct is the vesicular region of the
adult nephridium, which is known to be epiblastic, while the
rest is mesoblastic.

Lang (11) dwells upon the identity of the longitudinal duct

of Annelids with the longitudinal canals of Platyhelminths, .

but apparently in the belief that the former is of mesoblastic
origin like the latter. There are, however, now considerable
reasons for believing that the longitudinal duct of Lumbricus,
and possibly of Lanice, is epiblastic; this homology cannot
therefore be at present definitely accepted.

In the meantime, therefore, I would submit (1) that the
origin of the Amnnelid excretory system from that of the
Platyhelminths is as has heen suggested in the present paper,
(2) that the longitudinal duct of Lumbricus, Lanice, &c.,
has not any relation to that of the Platyhelminths.

1 The pronephros has been described as formed by outgrowths of the archine-
phric duct, and therefore epiblastic.

—

NEPHRIDIA OF CERTAIN EARTHWORMS. 409

List oF MEMOIRS REFERRED TO.

(1) Bepparp, F. E.—* Preliminary Note on the Nephridia of a New Species
of Earthworm,” ‘ Proc. Roy. Soc.,’ 1885, and of ‘ Ann. Sci. Nat.,’ 1885.

(2) Bepparp, F. E.—“ Preliminary Note on the Nephridia of Pericheta,”
‘Proc. Roy. Soc.,’ 1888.

(3) Bennam, W. B.— Studies in Earthworms,” ‘ Quart. Journ. Mier. Sci.,’
1886 and 1887.

(4) Beren, R.$.—‘ Die Exkretions Organe der Wiirmer,” ‘ Kosmos,’ 1885.

(5) Bournz, A. G.—“ Contributions to the Anatomy of the Hirudinea,”
‘Quart. Journ. Micr. Sci.,’ 1884.

(6) Cunnincuam, J. T.—‘“‘On the Nephridia of Lanice conchilega,”
‘Nature,’ June 16th, 1887.

(7) Etste, H—* Die Segmental Organe der Capitelliden,” ‘ Mitth. a. d. Zool.
Stat. Neapel,’ 1879.

(8) Fiscuer, W.—“ Ueber Capitella capitata,” ‘ Zool. Anz.,’ 1883.
(9) Fratpvont, J.—‘* Polygordius,” ‘ Naples Monographs,’ 1887.
(10) Happon, A. C.— Suggestion respecting the Epiblastic Origin of the
Segmental Duct,” ‘ Proc. Roy. Dublin Soc.,’ 1887.
(10 a) Hartoc, M.—‘ The True Nature of the Madriporia System of Echino-
dermata, with Remarks on Nephridia,’’ ‘Ann. and Mag. Nat. Hist.,?
Nov., 1887.
(11) Lane, A.—* Die Polycladen,”’ ‘Naples Monograph,’ 1884.
(12) Perrier, E.—* Etudes sur un genre nouveau des Lombriciens,” ‘ Arch.
Zool. Exp.,’ 1873.
(13) Vespovsky, F.—‘ System und Morphologie der Oligochaeten,’ Prag.,
1884.
(14) Wexpoy, W. F. R.—“‘On Dinophilus gigas,” ‘ Quart Journ. Micr.
Sci.,’ 1886.
(15) Wuirman, C. O.—* A Contribution to the History of the Germ Layers
in Clepsine,” ‘Journ. Morph.,’ 1887.
(16) Witson, E. B.—‘ The Germ Bands of Lumbricus,” ‘ Journ. Morph.,’
1887.

410 FRANK E. BEDDARD.

EXPLANATION OF PLATES XXX and XXXI,

Illustrating Mr. Frank E. Beddard’s paper ‘‘On the Occur-
rence of Numerous Nephridia in the same Segment in
Certain Earthworms, and on the Relationship between the
Excretory System in the Annelida and in the Platy-
helminths.”

PLATE XXX,
Fies. 1—6.—Acauthodrilus multiporus.

Fig. 1. Transverse section through body wall of 8th segment. o. Hx-
ternal orifice of nephridium. 4. Nephridium branching in the body
wall and opening by several (three) orifices.

Fig. 2. Transverse section of one half of circumference of body wall of
one of post-clitellar segments. s. Sete. 2. Nerve-cord. p. Peri-
toneum. «a, 4. Nephridial tubes.

Fig. 3. Similar section of a neigbouring segment; lettering as in last
figure.

Fig. 4. Highly magnified section to show the course of a single nephridial
tubule through the body wall. ep. Hpidermis. m. Circular. m!. Longi-
tudinal muscles. p. Peritoneum. c. Connective tissue. a, 4. Nephri-
dial tube.

Fig. 5.—Highly magnified section to illustrate the course of a nephridial
tube within the substance of the peritoneum. Lettering as in last
figure.

Fig. 6. Section through a single nephridial tube passing through the
body wall. a, 4. Nephridium. Other letters as in Fig. 4.

Fies. 7—1]1.—Pericheta aspergillum.

Fig. 7. Longitudinal section through part of two adjacent segments.
cp. Spermatheca. 2. Nephridia. s. Intersegmental septum. Other
letters as in Fig. 4.

Fig. 8. Transverse section of nephridiopore, highly magnified. xz. Ne-
phridium. Other letters as in Fig. 4.

Fig. 9. Transverse section through a portion of the body wall to show
the nephridiopores. zp. Nephridiopores. e. Epidermis. s. Sete.
Fig. 10. Section through a single nephridial tube, passing through the

body wall. 4/. Blood-vessels. Other letters as in Fig. 6.

NEPHRIDIA OF CERTAIN EARTHWORMS. 411

Fig. 11. Section through a portion of nephridial tube, highly magnified.
bl. Blood-vessel. 2. Marks the boundary between the region of the
nephridium with an intracellular duct and that with an intercellular
duct.

PLATE XXXI.

Fics. 12—14.—Pericheta aspergillum.
Fig. 12. Diagram of the relations of the nephridiopores (0) to the sete
(s), as seen on a superficial view.
Fig. 13. Diagram of the nephridial network in several segments. o. Ne-
phridiopores. yp. Nephridial network. s. Intersegmental septa.
Fig. 14. Diagram to illustrate the disposition of the nephridia in a single
segment as seen in transverse sections. s. Sete. 0. Nephridiopores.
Fie. 15.-Acanthodrilus multiporus.—A diagram corresponding to
the last. s. Sete. 2. Nephridiopores.

arwiedtd “wie dn) igen

a
j Le mys yor) Ay ~—rs esi AE
J { a L rte ptt :
' ind » eh oe
j t
7 Sp »
aa |
~
. *
- {Rs
a é : ;
: thos ‘ye
i A 3
- F : a
- i ye | Se

ee

_

‘
1
q

eset ah 2

5 = ~ =
ae Seas s a ier
eS = ge

F. Huth, Lith? Edim™

SS
R
x
:
:

a it
Pi
7 -
§
a Y 7

Fig. 12.

® Seddard del.

WEE

ey)
< auekis ,
= LATA

Sy

t

Mor Sourn! VA XIUNS PY, LU

Py eT
By = = Se Ny aa 8

AKON ANTES

A if 1) Se ee

SITET

vay Ll YY Zee

F. Huth, Lith? Edin™

THE ANATOMY OF THE MADREPORARIA. 4138

The Anatomy of the Madreporaria: IV.
By

G. Herbert Fowler, B.A., Ph.D.,
Assistant to the Jodrell Professor of Zoology in University College, London.

With Plates XXXII and XXXIII.

As was pointed out in a former paper (“ Anat. Madrep.,’’
ili), the relations between the external body wall and the
corallum of Madreporaria appear to yield a distinctive mor-
phological character, and to depend upon the presence or
absence of ccenenchyme. In all the genera yet examined in
which the individual polyps are more or less free and indepen-
dent of each other, the body wall is supported upon laminz of
mesogleea and endoderm (generally termed the “ peripheral
lamelle”’), which are continuous over the lip of the calyx with
the mesenteries, and bear a constant relation to them. But in
all the genera which form ccenenchyme, whether belonging to
the Perforata or to the Imperforata, the body wall rests upon
the little spikes or echinulations which stud the surface of the
corallum. From this it would appear that a distinction, between
_forms with and without coenenchyme, is justified by anatomical
differences at least as great as those which differentiate Per-
forata from Imperforata; but our knowledge of the group is so
slight that any generalisation such as the above is to be
accepted with great caution. Of the 878 genera recognised by
Professor P. Martin Duncan, in his recent “ Revision of the
Madreporaria” (‘ Journ. Linn. Soc. Zool., xvii), we have now
a more or less complete account of the anatomy of only sixteen
genera and a few more species, and the very foundations of a
true morphology of the group have yet to be laid. Indeed,

414, G. HERBERT FOWLER.

the interest of some of the forms which I am about to describe
lies in the fact that they exhibit a certain divergence from
both the structural types mentioned above. An account of the
following species will be found below.

Madracis asperula, p. 414. Stephanaria planipora, p. 424.

Amphihelia ramea, p. 416. Pocillopora nobilis, p. 425.

Stephanophyllia formosissima, |Seriatopora tenuicornis, sp. n.,
p. 418. p- 426.

Sphenotrochus rubescens, p.421.

MapRaciIs ASPERULA (fig. 1).

For the material for a study of this coral I am indebted to
the courtesy of Mr. John Murray, who has placed in my hands
the remainder of the spirit collection of Madreporaria obtained
during the voyage of H.M.S. “ Challenger.”

In systematic position this coral is ranged by Martin Duncan
close to the genus Stylophora, of the anatomy of which Dr. von
Koch, of Darmstadt, has already published an account (‘ Jen.
Zeitschr.,’ xi). From this, however, Madracis differs in an
important point.

The corallum, which requires here no systematic descrip-
tion, branches in digitate lobes, above the general surface of
which the thece of the polyps project slightly. As in Seria-
topora, Stylophora, and other such, it is Imperforate; the
living tissues are confined to the external surface, and do not
penetrate into the depth of the colony, the cavity formerly
occupied by each polyp being shut off by a succession of
“ tabule” as growth proceeds.

The septa are eight in number, Madracis thus adding
another to the several exceptions to Milne-Edwards’ law,
recently described. With this divergence, however, are not
correlated in Madreporaria such marked structural differences
as characterise the Monaulee, Edwardsiz, &c., which simi-
larly diverge from Hexactinian symmetry. The septa are
entoccelic only.

The polyps are in the main of Actinian structure; the
body wall clothing the whole colony is composed of the ordi-

THE ANATOMY OF THE MADREPORARIA. 415

nary three layers. The tentacles are apparently both ecto-
ceelic and entoccelic, and are simple evaginations tipped each
by a single battery of nematocysts, as in Seriatopora (“ Anat.
Madr.,” iii, fig. 11). The septa lie each between a pair of
normal mesenteries. No differentiation of particular mesen-
teries is recognisable, and all extend to about the same depth
into the polyp cavity. The two pairs of directive mesenteries
are well marked, but it is worthy of note that the plane in
which they lie does not always coincide with the axial-abaxial
(dorso-ventral) plane which generally governs the orientation
of similar colonies, but is here, in the case of some polyps, at
right angles to it.

Madracis is of especial value as affording an intermediate
condition between the two Madreporarian types mentioned
above, in that the body wall is supported on the echinulations
of the ccenenchyme only in certain more or less limited arez
between the polyps, whereas immediately round the calyces are
recognisable such peripheral lamelle as are usually charac-
teristic of forms devoid of ccenenchyme (fig. 1, B). This is
probably due to the fact that the polyp calyces are slightly
exsert above the general surface of the colony.

Here, therefore, we apparently have a condition morpho-
logically intermediate between that of a solitary imperforate
coral, as, for example, Caryophyllia, and sucha coenenchyma-
tous form as Seriatopora. The condition in Caryophyllia
(von Koch, ‘Morph. Jahrb.,’ v, fig. 4) and other simple forms
is probably the more primitive. The next stage in the series
seems to be indicated by the existing condition of Lophohelia
; prolifera (‘‘ Anat. Madrep.,” iii, fig. 6), where the polyps and
calyces are free and independent of each other, although a
colony is formed by gemmation. Next it would appear that,
as coenenchyme was developed (presumably in order to increase
the solidity and mass of the colony), and the spaces between the
various polyps of the colony filled gradually with coral from
below upwards, the peripheral lamellz, as being necessarily
confined to the immediate neighbourhood of the calyces, would
be inadequate to bridge over the intercalated arex of coenen-

VOL, XXVIII, PART 3,——NEW SER. FF

416 G. HERBERT FOWLER.

chyme, and hence a new means of support for the body wall,
namely, by means of the echinulations of the cenenchyme
itself, became necessary. The acquirement of this secondary
condition is apparently represented by Madracis; but as the
calyces are still slightly exsert above the ccenenchyme, the
peripheral lamelle are retained in the aree immediately round
the polyps. Lastly, in Seriatopora (‘“‘Anat. Madrep.,” ili,
fig. 9), which may represent the final term of the series, the
calyces are merged in ccenenchyme, and no trace of the peri-
pheral lamellee is to be found.

As in other ccenenchymatous forms, by fusion of the body
wall with the echinulations, the section of the ccelenteron lying
below it is broken up into canals which form the only com-
munication between the various polyp cavities.

A figure and description of the corallum may be found in
Milne-Edwards and Haime, ‘Ann. Sci. Nat.’ (8) xiii, p. 101,
pl. iv, fig. 2.

AMPHIHELIA RAMEA (figs. 2, 3).

Some specimens of the live coral were dredged off Lervik
in Norway, by Professor E. Ray Lankester, who has been kind
enough to entrust them to me for investigation.

Like Lophohelia, this genus forms by gemmation a branching
colony without any development of ccenenchyme, the separate
polyps being practically independent of each other. The
theca is thin and imperforate, and marked externally by
irregular longitudinal ridges and furrows which are of some
morphological interest. The septa are generally twenty-four
to twenty-eight innumber, of which twelve to fourteen aresmaller
and ectoceelic, while a similar number are large and entoceelic.
The latter unite below into a loose and irregular columella.

The body wall, which is composed of the usual three
layers, forms a continuous sheet over the whole colony, and
this, with a canal system to be shortly described, constitutes
the sole connection between the adult polyps; that is to say,
there is no continuity between the mesenteries of any given

THE ANATOMY OF THE MADREPORARIA. 417

polyp and those of its parent polyp below, though such a
connection might have been expected. The tentacles are
apparently both ectoccelic and entoceelic; but in these, as in
other spirit specimens, it is difficult to determine the point
with certainty. They are covered with batteries of nematocysts,
as in Flabellum (“‘ Anat. Madr.,” i, fig.9) and others. The mouth
disc and stomodeum are normal, the lumen of the latter being
nearly filled with coils of craspedon or the contorted edge of the
mesentery. The mesenteries are generally in twelve to
fourteen pairs of normal appearance, and in my specimens bore
abundant ova. Unlike the condition recently described in
Lophohelia (“ Anat. Madr.,”’ iii), there are present two pairs
of well-developed “ directive ” mesenteries; the more remark-
able as these two genera are most closely allied, and gemma-
tion is of a similar character in both cases. All the mesenteries
extend downwards nearly to the parent polyp, but do not
touch it.

The most interesting point in the anatomy of Amphihelia
lies in the relation existing between the external body wall and
the corallum. For a very short distance round the lip of the
calyx are recognisable, as in other free forms previously de-
scribed, the characteristic peripheral lamelle of the mesen-
teries (fig. 2,4). Just below the lip, however, the coral grows
outwards at the points to which these lamelle are attached,
and the mesogloea lamina of the body wall comes to lie at
these points directly on the broad ridges of coral thus formed
(fig. 2,B). A series of canals lined by endoderm is left be-
tween these ridges, which are continuous over the lip of the
calyx with the coelenteron. These canals (fig. 3) are, in the
main, parallel to the longer axis of the polyp, but are also
connected by transverse branches ; they come to lie in about
the same radius as the mesenteries, and are generally equal to
them in number, but are very irregular, both in position and
formation. Like the body wall which bounds them peripherally,
they appear to be continuous with the similar canals of the
polyps above and below.

An examination of the corallum shows that the ridges be-

418 G. HERBERT FOWLER.

tween which the canals lie are not homologous with true costz,
since the latter (which are generally regarded as the ends of
the septa, projecting radially outwards beyond the theca) are
also represented round the lip of Amphihelia. ‘They may per-
haps bear some relation to the echinulations of ccenenchy-
matous forms; but here, as in most questions of Madreporarian
morphology, a close investigation of a large number of allied
forms is necessary for a true explanation of structure.

A figure and description of the corallum may be found in
Milne-Edwards and Haime in‘ Ann. Sci. Nat.’ (3) xiii, p. 86,
pl. iv, fig. 3.

STEPHANOPHYLLIA FORMOSISSIMA (figs. 4, 5, 6).

Professor H. N. Moseley has kindly handed to me for further
investigation part of a decalcified specimen of this polyp, of
which he has already published a partial account in his mono-
graph of the Deep-Sea Madreporaria obtained by the “ Chal-
lenger” (“Chall.””? Rep. Zool. ii, p. 203, pl. xvi). Owing to the
fragmentary nature of the material, | am unable to give a
detailed account of the general anatomy, but it is still possible
to make out some points of interest.

As is the case in Fungia, and in the embryo Astroides de-
scribed by von Koch and Lacaze-Duthiers, the corallum is
plano-convex, resting free and unattached upon its plane
surface. The latter constitutes the theca, and from it the
septa rise perpendicularly upwards. The theca consists of
a large number of concentric and radial trabecule, and from
the radial trabecule rise alternately a mesentery and a septum
(fig. 4). The septa are thus both ectocelic and entocelic,
and, like the concentric trabecule from which they spring, lie
free in the ceelenteron, not touching the body wall. On the
other hand, those radial trabecule to which the mesenteries
are attached, lie directly on the basal body-wall, and are united
with it just as are the ridges and spikes in Amphihelia and
coenenchymatous forms.

Stephanophyllia therefore appears to stand in much the same
relation to the family Eupsammidz, with which it is generally

THE ANATOMY OF THE MADREPORARIA. 419

classed, as does Amphihelia to its family the Oculinide ; that
is to say, it is possible that in the former, as certainly in the
latter, the peripheral lamelle, which are found in such allied
genera as Rhodopsammia and Lophohelia respectively, have
been obliterated and functionally replaced by a radial out-
growth of coral from their points of attachment. The tra-
beculz of coral to which the mesenteries are attached are of
course not in this genus any more than in Amphihelia to be
regarded as true costz, since they bear no relation to the septa,
and are a simple outgrowth from the theca.

In several respects Stephanophyllia exhibits a resemblance
to the Fungia recently described by Bourne (‘ Quart. Journ.
Mier. Sci.’ xxvii). The plano-convex shape of the corallum,
and the correlated basal position of the theca are of course
characteristic of both genera; besides this, a view of a com-
plete mesentery shows that a few synapticule brace together
the septa of this coral, and that between the synapticule the
muscular pleatings of the mesoglea are gathered into strong
bundles in a very characteristic manner hitherte described only
in Fungia (fig. 5). The histological appearance of the cras-
pedon is also similar and characteristic in both forms. On
the whole, Stephanophyllia appears to bring Fungia into
closer relations with the Eupsammide than have been gener-
ally allowed. At the same time it is to be remembered that
in Fungia the body wall is, at any rate partially, supported on
peripheral lamelle.

The tentacles are ectoccelic as well as entoccelic, and are
covered with batteries of nematocysts. As has been pointed
out in other forms, the nematocysts are of two kinds, of which
the smaller alone are to be found in the tentacles. In the
craspedon, however, the larger kind is of exceptional dimen-
sions, measuring as much as ‘144 mm. by ‘012 mm.; while
the smaller form is only (048 in length. A pair of directive
mesenteries occurred in the small fragment used for transverse
sections.

420 G. HERBERT FOWLER.

ON sOME STRUCTURES PREVIOUSLY DESCRIBED AS CALICOBLASTIC.

The only other point of interest noticed in Stephanophyllia
is a structure for the firmer attachment of the mesentery to
the corallum, a structure which occurs in many other genera,
though a different significance has been hitherto assigned to it.
In Flabellum alabastrum it is especially well developed, and
consists of a series of processes given off radially into the
corallum by the mesoglea lamina of the mesentery (fig. 7).
These processes are generally laminated, presumably for a
further increase of the surface of attachment, and are directed
obliquely downwards. In transverse sections of decalcified
specimens of Flabellum and similar forms they often appear
as a layer of polygonal bodies, striated as a rule radially, and
lying between the mesoglcea and the space occupied before
decalcification by the corallum (fig. 8). Such bodies were re-
cently described by W. L. Sclater (‘ Proc. Zoo]. Soc.,’ 1886) in an
account of Stephanotrochus Moseleyanus as calicoblasts
or coral-forming cells ; but vertical sections, cut parallel to the
broader plane of the mesentery, show that such an explanation
is untenable. A calicoblast is ontogenetically derived from
the basal ectoderm of the embryo, and presumably has the
characters which we ordinarily associate with a cell. A layer
of such calicoblasts, obviously true cells, may be recognised at
the growing points of any coral. The structures in question,
however, are merely offsets of the homogeneous mesoglea of
the mesentery, and possess neither nucleus nor cell wall ; nor
is the mesogloea of the Anthozoa, itself a secretion, known to
exhibit secretory activity. Again, these processes do not occur
at the growing points of the corallum. On the other hand, as
they are to be met with only in the neighbourhood of the lines
of attachment of the mesenteries to the corallum, their posi-
tion, as well as their shape and lamination, indicate that their
function is to provide an increased surface for fixation of the
mesentery, and a firmer fulcrum for the action of the powerful

THE ANATOMY OF THE MADREPORARIA. 421

retractor muscles. The reason that their occurrence has been
hitherto overlooked is, that the mesoglea is stained only when
the colouring matter, whether carmine or hematoxylin, is
strong and diffuse; but I have now detected them in Flabellum
(2 sp.), Amphihelia, Seriatopora (2 sp.), Pocillopora (2 sp.),
Stephanophyllia, and Sphenotrochus : and Bourne has recently
described their occurrence in Mussa (‘ Quart. Journ. Mier. Sci.,’
XXvili).

Dr. von Heider, of Graz, who was the first to call attention
to the existence of calicoblasts, describes in a recent account
of Dendrophyllia (‘ Zeitschr. wiss. Zool.,’ xliv) a structure which
he regards as calicoblastic, in terms which apply exactly to a
transverse section of these processes. His figures (pl. xxxi,
figs. 9, 11) do not, however, clearly show whether the two
structures are identical or not, though it seems probable that
they are so, and that what he describes as needles of calcium
carbonate inside the “cells,” are the lines of lamination (cf.
Bourne, ‘ Quart. Journ. Micr. Sci.,’ xxviii, p. 25). It should
be noticed, on the other hand, that he explicitly states
that, in some at least of these so-called cells, a nucleus was
present.

In Stephanophyllia these processes are to be found at the
base of each mesentery, and form a band running down the
sides of the radial trabecula to which the mesentery is attached ;
but their structure is not different from those occurring in
Flabellum and other forms, and their position a natural con-
sequence of the different position of the theca (fig. 6).

Figures and description of the corallum may be found in
Moseley, “ Chall.” Rep. Zool., ii, p. 198, pls. iv, xiii, xvi.

SPHENOTROCHUS RUBESCENS (figs. 9—12).
A study of some fragments of this coral, entrusted to me by
Professor Moseley, has produced one or two points of con-
siderable interest, though, as with Stephanophyllia, the account

is necessarily incomplete.
The corallum is not unlike that of Flabellum, near which

422 G@. HERBERT FOWLER.

it is placed by Martin Duncan (‘Journ. Linn. Soc.,’ xvii). From
the latter it differs, in that soft tissues are present on the
outside of the theca. The septa are both ectoccelic and
entoceelic, and to both sets correspond true cost, which
extend downwards for some distance over the external surface
of the corallum.

As was suggested by Moseley (‘‘Chall.”? Rep. Zool., 11, p. 159),
in complete retraction the tentacles are covered over by the
mouth-dise, which is drawn inwards by a stroug sphincter
muscle (fig. 9). This constitutes the first recorded occurrence
of ‘ Rotteken’s muscle”? among Madreporaria; in longitudinal
section it has the same ‘diffuse ” appearance as that figured by
the Brothers Hertwig,in Sagartia parasitica (‘Jen. Zeitschr.,’
xiii, pl. xix, fig. 18.) The mouth-disc itself is much corru-
gated, the rugz consisting of lengthened columnar ectoderm
cells supported by solid outgrowths of the mesoglea, which
is at this point very thick (fig. 10, a). The ectoderm of the
tentacles consists almost entirely of nematocysts, showing
but slight arrangement into batteries. The tentacles are ap-
parently entoceelic only. The mesenteries appear to be
normal, but in the fragment used for microscopic sections no
directives were present. Both retractor (longitudinal) and
protractor (oblique) muscles are exceptionally well developed,
the former applied to such arborescent pleatings of mesogloea as
have been described among both Madreporaria and Hexactinie.
It is worthy of note that there is so little cohesion between the
muscle-fibres and the mesogloea which affords them support,
that they are frequently seen in transverse sections to tear
away with the endoderm, leaving the mesoglea bare. Over
the whole of the polyp the mesoglea is unusually strongly
developed, and it exhibits at some points, the position of which
appears to imply a more recent secretion than elsewhere, a
great affinity for carmine, and a markedly vacuolated structure.
The vacuoles are often empty, but contain, in many cases,
brilliantly refractive crystals, the nature of which I was unable
to ascertain. Cells are rarely to be seen in the mesoglea.
The processes for attachment of the mesentery to the corallum

THE ANATOMY OF THE MADREPORARIA. 423

(cf. p. 8, supra) are well developed, not only on the mesentery,
but also on the thickened mesoglcea at the sides of the pseudo-
coste (v. infra, and figs. 10, 11, a).

Encapsuled in the mesoglea lamina of the mesentery and of
its craspedon are numerous Ova in various stages of matura-
tion. So far as I am aware, all Anthozoan ova as yet
described are formed in the endoderm, and then migrate
into the mesoglea, the latter forming a capsule appressed
closely round them. In Sphenotrochus, however, the ovum is
surrounded by a series of deeply staining bodies which lie
between its membrane and the mesogleeal capsule (fig. 12).
At first sight these corpuscles appear to consist of extruded
yolk, but in most, if not in all, it is possible to detect a faint
nucleus. Nearly triangular in section, their bases are pressed
closely against the capsule, from which they sometimes shrink
away, allowing their outline to be clearly seen. It is hardly
possible to regard them as other than follicular cells. They
are not present in the capsule of the smallest ovum observed
(1x°07 mm., diam. nucleus ‘008 mm.), and their number
increases up to a certain point with the size of the ovum. It
is probable therefore that they migrate from time to time into
the mesogloea, as does the ovum itself; and their appearance
indicates that their function may be to supply yolk-material
for the ripening egg-cell. More deeply staining and larger
bodies of irregular outline are found scattered in the endoderm
of the craspedon, apparently identical with those described in
Euphyllia (Bourne, ‘ Quart. Journ. Micr. Sci.,’ xxviii, p. 31) ;
these may bear some relation to the follicle cells.

The soft tissues external to the theca are also of some
interest. No peripheral lamelle of the mesenteries are recog-
nisable, but at the points where they might be expected to
occur are formed outgrowths of corallum comparable to those
in Amphihelia (v. p. 5, supra.), on which the body wall
rests (fig. 10, B. a.). The costz corresponding to the ento-
ceelic septa come also into direct contact with the body-wall
(ibid., a'), while those of the entoccelic septa are free. Ata
lower level, however, both sets of costz regularly serve for the

424, G. HERBERT FOWLER.

support of the body wall (fig. 11, a1), but here and there they
fail to touch it (ibid., a”), thus allowing of a cross commu-
nication between the longitudinal canals. The number of
longitudinal canals is thus double that of the ectoccelic and
entoceelic spaces, since the body wall is in contact both with
true costz and with pseudocoste (i. e. those which replace the
peripheral lamelle). In the single perfect specimen of the
corallum in the British Museum, the pseudocoste are dis-
tinctly visible on the upper two thirds of the exterior of the
calyx, but below this are lost in the true coste. Unfortu-
nately my fragments did not show the relations of the soft
tissues in the lowest third.

Figures and a description of the corallum may be found in
Moseley, ‘“ Chall.” Rep. Zool., ii, p. 157, pl. vi.

STEPHANARIA PLANIPORA (fig. 13).

A specimen of a Psammocorid coral, for which I am again
indebted to the generosity of Professor Moseley, appears to be
identical with specimens of the above-named species in the
British Museum. It is not recorded by Quelch (“ Chall.”
Rep. Zool.), and therefore adds another to the ‘‘ Challenger ”
species of Reef-corals. For the identification of this and
many other corals, and for much other assistance, I desire to
acknowledge my great obligation to Mr. Dendy, late of the
British Museum.

The boundaries of the various polyps which compose the
colony are so ill defined that it is impossible to decide how
much of the soft parts seen in a transverse section should be
referred to a particular individual. A glance at a figure of a
section cut transversely to the axes of the polyps will show
this (fig. 18). Each polyp possesses a stomatodeum, to which
are attached a certain number of mesenteries, generally seven
to ten; between these mesenteries lie entoccelic and ectoccelic
septa, and over each entoceelic (?) septum is placed a tentacle
of the simple character already described for Seriatopora,
Madracis, &c. (“ Anat. Madr.,” iii, fig. 11). The muscles,

THE ANATOMY OF THE MADREPORARIA. 425

though obviously present, are not sufficiently developed to
admit of a distinction of the mesenteries into pairs. The con-
tinuity of the mesenteries radiating outwards from the polyp
is constantly broken by synapticule. The whole of the surface
of the colony is practically composed of these radiating slips of
mesentery lying between radiating septa, and interrupted by
synapticule ; and, except in the immediate neighbourhood of a
polyp, it is often impossible to decide to which individual a
mesentery is referable, since neither their direction nor mus-
culature afford a clue. Contrary to what one might expect
(cf. the more or less comparable condition in Fungia; Bourne,
‘Quart. Journ. Micro. Sci.,’ xxvii, pl. xxv, fig. 13), even such
slips of- mesentery as are at no point in contact with the
stomatodeum often exhibit a filamentar (craspedal) thickening.
Even at the uppermost ends of the branches, at a distance of
perhaps a third of an inch from the nearest polyp, the same
arrangement of mesenteries between septa (septo-coste) is
found. On these mesenteries the body wall is supported,
although in places it appears to rest also on echinulations of
the septa; as will be seen from the figure, there is practically
no cenenchyme. ‘Through the inner parts of the colony
ramifies a system of canals by which the various polyps are in
communication with each other.

This genus appears to exhibit a distinct degeneration,
implied by the low development of the mesenterial filament (a
slight lengthening of the ordinary endoderm), and an indi-
viduality hardly more marked than in a Poriferan colony.

A figure and description of the corallum may be found in
Verrill, ‘ Trans. Connect. Acad.,’ 1, p. 545, pl. ix, fig. 4.

PocILLOPORA NOBILIS.

For material of this species I am again indebted to Pro-
fesssor Moseley, who obtained it during the voyage of the
“* Challenger.”

The anatomy agrees almost entirely with that of P.
brevicornis, recently described (‘ Anat. Madr.,” iii). The

426 G. HERBERT FOWLER.

differentiation of the mesenteries is similar, though perhaps
hardly so well marked as in the cther species, since the mesen-
teries numbered 3 and 10 in the diagram of P. brevicornis
are often not appreciably longer than 5 and 8. The reticular
tissue mentioned as filling in Seriatopora subulata and
P.brevicornis the spaces from which corallum has been
removed by decalcification, is in this species very much more
clearly recognisable. It appears to consist of thin strands of
(?) protoplasm (? mesogloea), staining deeply with carmine; it
retains the shape of the crystalline ellipsoids of which the
corallum is ultimately composed, and forms an accurate cast
of the dead internal parts of the colony.

SERIATOPORA TENULCORNIS, sp. n. (figs. 14, 15).

For a small fragment of a colony of this species I owe my
thanks to Dr. S. J. Hickson, who collected it with other corals
in the Celebes group.

In anatomy this form agrees exactly with the Seriatopora
subulata, already described ( Anat. Madr.,” iii). It appears
to constitute a new species, of which the following are the
diagnostic characters:—The branches are thin and finely
tapering, exhibiting no tendency to unite in the manner cha-
racteristic of most Seriatopore ; they are very solid in section,
and bear four or five rows of calyces. The coenenchyme is very
regularly ribbed, the echinulations of which the ribs are
composed being markedly short and blunt. The calyces pro-
ject slightly above the surface of the ceenenchyme; the septa
are generally about ten in number, short, and very irregular.
The columella does not reach to the surface of the calyx; it
is spinous, and not so plate-like as in some other species of the
genus.

From S. caliendrum, to which it is approximate, it differs
in that the calyces are farther apart, and project more from the
ceenenchyme than in that genus; the echinulations are further
apart, blunter, and shorter; the branches more slender. The
specimen has been deposited in the British Museum,

THE ANATOMY OF THE MADREPORARIA. 427

CONCLUSIONS.

The most important morphological points elucidated by a
study of the species above described are as follows:

]. The apparent law that the body wall, when present, is
supported in accenenchymatous forms upon “ peripheral
lamelle”’ of the mesenteries, and in ccepenchymatous species
upon the echinulations of the ceenenchyme, is seen to require
modification. The two methods of support may coexist in a
coenenchymatous form (Madracis), and to a certain extent
in an accenenchymatous species (Amphihelia); further, the
body wall in accenenchymatous species may rest upon pseudo-
costz (? homologous with echinulations), either mainly (Am-
phihelia) or entirely (Stephanophyllia) ; or upon both pseudo-
cost and true coste (Sphenotrochus). At the same time it is
doubtful whether such exceptions to the law, formulated above,
will not eventually prove to be modificatious due to exceptional
conditions, of which we are at present ignorant.

2. The modifications of the mesenteries exhibited by pre-
viously described species of Seriatopora and Pocillopora are
seen to extend to other species of the same genus.

3. The ultimate attachment of polyp to corallum consists, in
many genera, in a series of laminated offsets of mesoglea in
the neighbourhood of the mesentery, such structures having
been previously described as calicoblastic in function.

4. A sphincter muscle, comparable to the ‘ Rotteken’s
muscle” of Hexactiniz, may occur among Madreporaria (Sphe-
notrochus).

5. Follicle cells, which are perhaps immigrants from the
endoderm, may surround the ripening ovum as it lies in its
mesoglceal capsule (Sphenotrochus).

6. A case of degradation, tending to obscure the individuality
of the polyps, is presented by Stephanaria.

428 G. HERBERT FOWLER.

List OF THE PAPERS CHIEFLY QUOTED.

Bourne, G. C.—‘ Quart. Journ. Mier. Sci.,’ xxvii (Fungia).
Bourne, G. C.—‘ Quart. Journ. Mier. Sci.,’ xxviii (Euphyllia, Mussa).

>)

Duncan, P. M.--‘ Journ. Linn. Soc.,” xvii (revision of genera).

Fowter, G. H.—‘ Anat. Madr.,” i, ‘Quart. Journ. Mier. Sci.,’ xxv; ii, ibid.,
XXVii; iii, ibid., xxviii.

von Heiper, A.—‘ Zeitschr. wiss. Zool.,’ xliv (Dendrophyllia).

>

von Kocu, G.—‘ Jen. Zeitschr. Naturw.,’ xi (Stylophora).
von Kocu, G.—‘ Morph. Jahrb.,’ v (Caryophyllia).
Mosetey, H. N.—‘ Chall.” Rep. Zool., ii (Stephanophyllia, Sphenotrochus).

Sctater, W. L.—‘ Proe. Zool. Soc.,’? 1886 (Stephanotrochus).

EXPLANATION OF PLATES XXXII and XXXIII,

Illustrating Mr. G. Herbert Fowler’s paper on “The Anatomy
of the Madreporaria,’ IV.

Fic. 1. MapDRACIS ASPERULA.—Diagrammatic transverse section through a
polyp at three different levels, showing the eight entoccelic septa and eight
pairs of mesenteries. 4, above the level of the theca; the exsert edges of the
septa only are cut in this section. B, through the theca; showing the peri-
pheral lamelle. C, below the surface of the ccenenchyme. The support of
the body wall on the echinulations agrees essentially with that of other forms
previously figured (cf. “ Anat. Madrep.,” ii, fig. 4; or ili, figs. 2, 9). d@.,
* directive” septa. co/. Columella. cez. Ccenenchyme. Ectoderm, “blocked”
black and white; mesogloea, black; endoderm, grey; corallum, dotted (v.
pp. 2—4).

Fic. 2. AMPHIHELIA RAMEA.—Diagrammatic transverse section through
two quarters of a polyp, showing half of the twenty-four septa and of the
twelve pairs of mesenteries. 4, through the theca in the region of the ten-
tacles, showing the peripheral lamelle. 2, below the stomodeum, showing
the external canals between body wall and corallum. Ectoderm, quasi-
cellular; endoderm, mesoglcea, and corallum as in Fig. 1 (v. pp. 4—6).

\
i]
ws.

ler del.
Fowler de
G owl
G.H.F

Mor Pourn Ve MIM NSP, KH

*\*jelelol«

Pode 7 vA : ir
=! , aa ek [as _-
Vas 8 in | Bene ae :
7 > 2h To |

Edin™

F. Huth, Lith™

Mier Sourn: Vol XIVUNS Pe, KEM.

THE ANATOMY OF THE MADREPORARIA. 429

Fic. 83. AMPHIHELIA RAMEA.—The body wall seen by transmitted light.
The darker parts represent the canals, the lighter the ridges of coral between
them.

Fig. 4, STEPHANOPHYLLIA FORMOSISsIMA.—Ideal diagrammatic representa-
tion of the relations of septa, mesenteries, costs, and body wall to each other,
in asmall cube cut out from the base of the polyp. Hctoderm, ‘‘blocked”’ black
and white; mesogloea, thick black line; endoderm, thin black line; corallum,
dotted. Two radial trabecule (to which transparent mesenteries are attached)
are cut across in the front part of the diagram, showing their connection with
the body wall. From two other radial trabecule rise two septa, not touching the
body wall at any point; of these septa the left-handmost is cut between two
concentric trabecule, and the other at the point where concentric trabecula,
radial trabecula, and septum meet. A synapticulum braces together the two
septa, perforating the right-hand mesentery (v. p. 6).

Fic. 5. STEPHANOPHYLLIA FORMOSISSIMA.—View of a complete mesen-
tery, with the basal (thecal) tissues attached. In the latter is seen the row of
holes left by decalcification of the concentric trabecule ; and in the mesentery
four perforations produced by synapticule, at the sides of which the muscle
pleatings are gathered into characteristic bundles (v. p. 7).

Fic. 6. STEPHANOPHYLLIA FORMOSISSIMA.—Two transverse sections taken
at the base of a mesentery. In 4, the higher of the two, the concentric
trabecule have cut the mesentery across, so that a series of plates resembling
the one here drawn (JZ) is produced along the whole radius (cf. the condition
in Fungia, Bourne, ‘Quart. Journ. Micr. Sci.,’ xxvii, Pl. XXV, fig. 13). In
the plane of B, the section passes through a radial trabecula also, and is thus
taken below the edge of the mesentery. In both sections are seen the pro-
cesses for the firmer attachment of the mesentery (v. p. 8). The arrows
indicate the direction of #., the radial, and C., the concentric trabecule.

Fic. 7. FLABELLUM ALABASTRUM.—Lateral view of the distal edge of the
mesoglcea lamina of a mesentery, i.e. of the edge which is attached directly to
the corallum (cf. “Anat. Madrep.,” i, fig. 2), showing the laminated pro-
cesses of attachment of the mesentery (v. p. 8).

Fic. 8. FLABELLUM ALABASTRUM.—Transverse section of the base of a
mesentery (J/.), showing the processes of attachment lying in the space from
which corallum has been removed. Those which are radially laminated are
probably seen in transverse section, while parallel lamination indicates oblique
section (cf. Sclater, ‘P. Z. 8.,’? 1886, pl. xiv, figs. 11—13).

Fic. 9. SpHENOTROCHUS RUBESCENS.—Transverse section through the
sphincter muscle (“‘ Rotteken’s muscle”), which draws the mouth-disc over the
tentacles.

Fic. 10. SPHENOTROCHUS RUBESCENS.—Transverse section (incomplete)
through a pair of mesenteries and the surrounding tissues, Ectoderm and

ty mn

430 G. HERBERT FOWLER.

endoderm, light grey; mesoglcea, dark gray; space formerly occupied by
corallum, dotted. On each side of a central entoccelic septum lies a mesen-
tery, external to which is an ectoceelic septum. Corresponding to the mesen-
teries are the pseudo-coste (a) in direct contact with the body wall (4. w.),
with which the ectoccelic coste are also in contact (a'). The ruge of the
mouth-dise (m. d.) are seen in transverse section, supported on solid offsets of
mesoglcea (v. pp. 9—12).

Fic. 11. SPHENOTROCHUS RUBESCENS.—Transverse section at a lower plane
than is represented in Fig. 10, 4. The body is in contact with pseudo-coste
(a), and with both ectoccelic and entoccelic coste (a!). The latter, however,
at certain points do not reach the body wall (a*), thus allowing of transverse
communication between the longitudinal canals. (The theca has been
diminished in breadth by one half.)

Fic. 12. SPHENOTROCHUS RUBESCENS.—Section through a portion of the
ovum and surrounding tissues. The vacuolated ovum, bounded by a thin
vitelline membrane, has shrunk in the preservative fluid away from the follicu-
lar cells. The latter lie closely appressed to the thin mesogloeal capsule,
which is covered by endoderm cells externally (v. p. 11).

Fic. 13. StEPHANARIA PLANIPORA.—Section through a part of the colony
at right angles to the polyps, of which three are included (4, B,C). The
rest of the section is composed of the mesenteries and septa (septo-costex)
which radiate from them (v. p. 12).

Fies. 14 and 15.—SERIATOPrORA TENUICORNIS, sp. n. (v. p. 14). The
drawings for these two figures are due to the skill of Miss Stone.

ERRATA.

On page 444, fifth line from bottom, for ‘‘p. 29” read ‘“p. 458.”

On page 450, last line of note, for “p. 25” read “ p. 455.”

On page 455, fifteenth line from bottom, for “twenty-one legs” read
“twenty-one pairs of legs.”

On page 457, three lines from top, for “c. 7.” read ‘ cf.”

On page 467, sixth line from bottom, for “are” read ‘‘ is.”

On page 484, tenth line from top, for “ Strunstrip”’ read ‘‘ Steenstrup.”

On page 485, sixteenth line from top, for ‘“‘constructed” read “constricted.”

On page 487, last line (above footnote) dele the ‘*!” after ‘ (Sclater).”

QQ, ss Sts”

—, *”

|
| -
wk ( , { MEPs, F4ib ES ian? oo

er) es ai , na ; is? | &-anie ah

ah nites VE y Pane payer ‘
Cpe at ta EF ne re ain't = F
ta ‘ f a Ae

his) id pease, GT akateew ayy ee ba r
i a sp T

SR oe wi "ntesh *
ele? [og | gay Sel YG Hed dure), lake
eres 5. me" nen itl yal ae es a
“a ie 7 a. Le "th me lait” + pall tila iy
Ay ele "saat Vieni reed eae

A Monograph on the Species and Distribution
of the Genus Peripatus (Guilding).

By

Adam Sedgwick, M.A., F.R.S.,
Fellow of Trinity College, Cambridge.

With Plates XXXIV to XL.

INTRODUCTION.

THE editors of the posthumous memoir of Professor F. M.
Balfour on the anatomy and development of Peripatus
capensis stated, in a note at the end of the memoir, their inten-
tion of preparing a complete monograph of the known species
of Peripatus. This intention has at length been carried out,
and the present monograph is the result of a laborious exami-
nation of all the specimens of Peripatus to which it has been
possible to get access.

To my great regret Professor Moseley has been obliged, by
his most unfortunate illness, to withdraw from active participa-
tion in the work; and the whole responsibility for the state-
ments and descriptions falls upon me alone. But he has given
me much valuable assistance, and has made some substantial
contributions to the monograph. The most important of the
latter relate to Peripatus nove-zealandie. He examined this
species with great care, and the drawings from which figs. 16,
17—20, and 30 were copied, were executed under his
supervision,

The material at my disposal has been as follows:

VOL. XXVIII, PART 4,—NEW SER, GG

432 ADAM SEDGWICK.

1. The material left by Professor Balfour. This comprised
a large number of the Cape species collected by Mr. Lloyd
Morgan and by the late Mr. Oakley of the South African
Museum, and some specimens of the New Zealand species col-
lected by Professor Moseley and Professor Jeffrey Parker.
Balfour also had fourteen specimens from Caracas, sent him,
I believe, by Professor Ernst, and one specimen from South
Africa, of the exact locality of which I am ignorant, with
twenty-two pairs of legs. This specimen was found by
Mr. Mansel Weale, and given to Balfour by Mr. Wood
Mason.

2. Alarge number of specimens from the Cape, collected partly
by myself in 1883, and brought to England alive, and partly by
Mr. C. Stewart, of the Royal Hotel, Wynberg, who in the
winters of 1884-85 sent me a large additional supply of live
animals collected on Table Mountain. I am under great
obligations to Mr. Stewart, not only for these specimens,
but for the great help which he gave me when I was at the
Cape.

3. A large number of living specimens from Wellington, New
Zealand. These are the only specimens of this species which
have ever been brought to England alive, and I owe them to
the kindness of two gentlemen, who were personally unknown
to me until they began to help me in my Peripatus work. Mr.
Noel Barraud, of Wellington, at the request of my friend Mr.
G. E, Anson, M.A., of Trinity College, began to hunt for
Peripatus, and was successful in finding them. His specimens
were, after two unsuccessful attempts, brought to England
alive by Mr. Edgar J. Evans, chief officer of the Shaw, Savill,
Albion Company’s magnificent steamer “Tainui.’” The first
two lots all died soon after leaving Rio Janeiro, but in the
third attempt Mr. Evans was successful in finding a place near
the cold chamber, where the temperature in the tropics was
not too high for the animals. Since the third attempt Mr.
Evans has been successful every voyage. My most sincere
thanks are owing to both of these gentlemen, who, though not
specially interested in natural history, have put themselves to

MONOGRAPH OF THE GENUS PERIPATUS. 433

very considerable trouble and inconvenience to satisfy what
would have seemed to most people an absurd whim of a perfect
stranger. The living animals brought by Mr. Evans have
enabled me to give a much more complete account of the New
Zealand species than if I had had to rely only on the somewhat
shrunk and contracted specimens in Balfour’s material; and
the embryos found in them are now being used by Miss Sheldon
(No. 45), of Newnham College, who is engaged in preparing a
memoir of their development.

4. Several specimens of the neotropical group of species
from the Museum of Copenhagen, most kindly sent to me by
Professor Steenstrup, in response to the appeal for specimens
at the end of Balfour’s posthumous memoir.

5. Dr. J. Kennel, of Wiirzburg, was kind enough to send
me, in exchange for two living specimens of Capensis,
two specimens of the smaller species, which he found in
Trinidad.

6. Four specimens from near Williamstown, South Africa,
belonging to the Indian Museum.

7. Ali the specimens in the British Museum. These have
not, unfortunately, been of so much use to me as I had hoped,
on account of their small number and contracted condition,
and of the difficulty of getting a sufficiently strong light in
the rooms set apart for the examination of spirit specimens
in the British Museum, for the minute examination which is
required. I am greatly indebted to Professor Jeffrey Bell, of
the British Museum, for his courtesy in giving me every
facility in his power to examine the specimens.

8. Two specimens from Queensland, Australia, belonging to
Professor Jeffrey Bell. He has been most kind in putting
them entirely at my disposal for the purposes of this mono-
graph.

9. About twenty specimens from Demerara, brought alive
to England by Mr. W. L. Sclater. Mr. Sclater has himself
given a short description of these specimens, and has kindly
placed the greater number of them at my disposal for the
purposes of this monograph. ‘These specimens were all killed

434 ADAM SEDGWICK.

and opened by me shortly after their arrival in order that
their embryos might be cut out and preserved. The result
is that they are somewhat contracted and not so favorable
for observation as they would have been had they been
drowned.

Of the figures illustrating this monograph, the beautiful draw-
ings of Capensis on Pl. XXXV, of Edwardsii, Moseleyi,
and Nove-zealandiz on Pl. XX XVII, were made by Miss A.
B. Balfour in Professor Balfour’s lifetime and under his
direction. The figures on Pl. XXXVI were made by Miss
Balfour after her brother’s death. To enable me to complete
the illustrations required and to assist in the publication of
the monograph, the Government Grant Committee of the
Royal Society granted the sum of £50. Figs. 16, 17—20, and
30 were made at Oxford by Mr. W. H. Hill, under Professor
Moseley’s direction and supervision. The remainder of the
drawings, including that of the living animal of Pl. XXXIV,
were made by Mr. E. Wilson, of the Cambridge Scientific
Instrument Company. My best thanks are due to these
gentlemen for the care and skill with which they have exe-
cuted their work.

Finally, I have again to acknowledge my indebtedness to
Professor Jeffrey Bell for his assistance in preparing the
Bibliography. Several papers which had escaped my notice
were first pointed out to me by him.

For an account of the general anatomy and characters of the
genus Peripatus, I must refer to the memoirs of Moseley
(No. 18), Balfour (No, 28), and Gaffron (Nos. 34, 35).. In
this monograph only those features of a specific differential
value are dwelt upon. I have, however, made a partial ex-
ception to this rule in the case of Capensis, the external
characters of which have been described at considerable
length. The reader will be able to gather from this descrip-
tion a sufficient knowledge of the general external features of
the genus to enable him to understand the short descriptions
of the other species.

MONOGRAPH OF THE GENUS PERIPATUS. 435

The chief result of my observations has been to establish a
definite series of characters which distinguish quite sharply all
the species found in one region of distribution from those found
in the others. Excluding the doubtful case of the Sumatran
species, Peripatus has been found in the Ethiopian region
(South Africa), the Australasian, and the Neotropical regions,
and in each of these regions the genus is represented by more
than one species. I have been able to establish a certain
number of new species, but on the whole I must confess to
failure in this respect. My failure chiefly relates to the species
from the Neotropical region, and is due to the insufficient
number of specimens at my disposal. It is remarkable of the
species from this region that the number of the walking legs
varies considerably within the same species, and it is only
possible to determine the limits of the variation by examining
a large number of individuals. Inasmuch as the specific
characters other than those afforded by the legs are extremely
inconspicuous, the importance of having a large and well
preserved material is obvious—large in numbers to enable one
to establish the limits of leg variation, and well preserved that
the more minute specific differences may be made out.

How inconspicuous the specific characters are is well shown
by contrasting the South African species Capensis and
Balfouri. That these are distinct species is proved by the
fact that the number of legs is constant in all the large number
of specimens examined, and by the fact that it is preserved in
the reproduction of the species. Embryos removed from
P. capensis have invariably seventeen pairs of legs, while
embryos removed from P. Balfouri have invariably eighteen
pairs. The other differences relate simply to the texture and
tint. of the skin, and are so minute as to escape any but the
experienced eye.

Before concluding this introduction, I am desirous of pointing
out how extremely loose and inaccurate have been the observa--
tions of some professed zoologists on the members of our
genus. In several cases has it happened that the observer
(sit venia verbo) has not been at the trouble of counting the

436 ADAM SEDGWICK.

legs of his specimens, though he has not refrained from making
statements on this point, and in more than one case the number
of legs in the specimen figured does not correspond with the
author’s statement in the text. If one may draw conclusions
as to these zoologists’ ideas of accuracy in observation from
such instances in which only the most obvious external features
are concerned, one would be inclined to infer that but little
value can be attached to their statements with regard to the
more inconspicuous details, which require some nicety of obser-
vation.

Peripatus, Guilding.

Soft-bodied vermiform animals, with one pair of ringed
antenne, one pair of jaws, one pair of oral papille, and a vary-
ing number of claw-bearing ambulatory legs. Dorsal surface
arched and more darkly pigmented than the flat ventral surface.
Skin transversely ridged and beset by wart-like spiniferous
papille. Mouth anterior, ventral; anus posterior, terminal.
Generative opening single, median, ventral, and posterior. One
pair of simple eyes. Brain large, with two ventral hollow
appendages ; ventral cords widely divaricated, without distinct
ganglia. Alimentary canal simple, uncoiled. Segmentally
arranged, paired nephridia are present. Body cavity is con-
tinuous with the vascular system, and does not communicate
with the paired nephridia. Heart tubular, with paired ostia.
Respiration by means of trachez. Dicecious; males smaller
and generally less numerous than females. Generative glands
tubular, continuous with the ducts. Viviparous. Young born
fully developed. They shun the light, and live in damp places
beneath stones, leaves, and bark of rotten stumps. They eject
when irritated a viscid fluid through openings at the apex
of the oral papille.

Distribution: South Africa, New Zealand, and
Australia, South America and the West Indies [and in
Sumatra ?].

The genus Peripatus was established in 1826 by Guilding
(No. 1), who first obtained specimens of it from St, Vincent

MONOGRAPH OF THE GENUS PERIPATUS. 437

in the Antilles. He regarded it as a Mollusc, being no
doubt deceived by the slug-like appearance given by the
antenne. Specimens were subsequently obtained from other
parts of the Neotropical region and from South Africa, and
the animal was variously assigned by the zoologists of the day
to the Annelida and Myriapoda (vide Moseley, No 22, and
Sclater, No. 41). Its true place in the system, as a primitive
member of the group Arthropoda, was first established in 1874
by Moseley (No. 18), who discovered the trachee. It was
reported from Australia in 1869 by Saenger (No. 15), and
from New Zealand by Hutton (No. 19) in 1876. The
nephridia were first discovered by Saenger (No. 15), but they
were re-discovered and more fully described by Balfour (Nos.
21 and 28). Gaffron was the first to observe the cardiac ostia
and the cilia in the generative tracts. The development has
been worked at by Moseley (No. 18), Hutton (No. 19),
Balfour (No. 28), and more in detail by Kennel (Nos. 32 and
33), Sheldon (No. 45), Sclater (No. 46), and myself (No. 89).
There can be no doubt that Peripatus is an Arthropod, for
it possesses the following features, all characteristic of that
group, and all of first-class morphological importance. (1) The
presence of appendages modified as jaws. (2) The presence of
paired lateral ostia perforating the wall of the heart and putting
its cavity in communication with the pericardium. The import-
ance of this feature as an Arthropod character was first pointed
out by Lankester. (3) The presence ofa vascular body cavity and
pericardium (hzemoccelic body cavity). (4) The inconspicuous
character of the celom in the adult. Finally, the trachez,
though not characteristic of all the classes of the Arthropoda,
are found nowhere outside that group, and constitute a very
important additional reason for uniting Peripatus with it.
Peripatus, though indubitably an Arthropod, differs in such
important respects from all the old-established Arthropod
classes, that a special class, equivalent in rank to the others,
and called Prototracheata, has had to be created for its sole
occupancy. ‘This unlikeness to other Arthropoda is mainly
due to the Annelidan affinities which it presents, but in part

438 ADAM SEDGWICK.

to the presence of the following peculiar features: (1) the num-
ber and diffusion of the tracheal apertures, (2) the restriction
of the jaws to a single pair, (3) the disposition of the generative
organs, (4) the texture of the skin, and (5) the simplicity
and similarity of all the segments of the body behind the
head.

The Annelidan affinities are superficially indicated in so
marked a manner by the thinness of the cuticle, the dermo-
muscular body wall, the hollow appendages, that, as already
stated, many of the earlier zoologists who examined Peripatus
placed it amongst the segmented worms; and the discovery
that there is some solid morphological basis for this determin-
ation constitutes one of the most interesting points of the
recent work on the genus. The Annelidan features are: (1)
the paired nephridia in every segment of the body behind the
first two (Saenger, Balfour), (2) the presence of cilia in the
generative tracts (Gaffron). It is true that neither of these
features are absolutely distinctive of the Annelida, but when
taken in conjunction with the Annelid disposition of the chief
systems of organs, viz. the central nervous system, and the
main vascular trunk or heart, may be considered as indicating
affinities in that direction. Peripatus, therefore, is zoologically
of extreme interest from the fact that, though in the main
Arthropodan, it possesses features which are possessed by no
other Arthropod, and which connect it to the group to which
the Arthropoda are in the general plan of their organisation
most closely related. It must, therefore, according to our
present lights, be regarded as a very primitive form; and this
view of it is borne out by its extreme isolation at the present day.
Peripatus stands absolutely alone as a kind of half-way
animal between the Arthropoda and Annelida. There is no
gradation of structure within the genus; the species are very
limited in number, and in all of them the peculiar features
above mentioned are equally sharply marked.

We may, therefore, with some justice, regard Peripatus as
an animal which has persisted for a long time, with but little
structural modifications; as the representative of an ancient

MONOGRAPH OF THE GENUS PERIPATUS. 439

group, once widely diffused,' and probably rich in species and
genera, and closely related to the ancestors of living Arthropoda.
It probably has owed its preservation, as so many of the survivals
of ancient types seem to have done, to the peculiar habits of life
which are shared by all the living members of the class, viz.
the habitual avoidance of the light of day, and the habit of
seeking the obscurity and protection afforded by the spaces
beneath stones and beneath the bark of trees.

Peripatus, though a lowly organised animal, and of re-
markable sluggishness, with but slight development of the
higher organs of sense, with eyes the only function of which
is to enable it to avoid the light—though related to those
animals most repulsive to the esthetic sense of man, animals
which crawl upon their bellies and spit at, or poison,
their prey—is yet, strange to say, an animal of striking
beauty. The exquisite sensitiveness and constantly changing
form of the antennz, the well-rounded plump body, the eyes
set like small diamonds on the side of the head, the delicate
feet, and, above all, the rich colouring and velvety texture
of the skin, all combine to give these animals an aspect of
quite exceptional beauty. Of all the species which I have
seen alive, the most beautiful are the dark green indi-
viduals of Capensis, and the species which I have called
Balfouri. These animals, so far as skin is concerned, are
not surpassed in the animal kingdom. The drawing on
Pl. I. is from one of the dark green specimens of Capensis.
Clever as the drawing is, the artist has failed to catch the
exquisite velvet of the skin ; but this could hardly be expected
in a lithograph. I never shall forget my astonishment and
delight when on bearing away the bark of a rotten tree-stump
in the forest on Table Mountain, I first came upon one of
these animals in its natural haunts, or when Mr. Trimen
showed me in confinement at the South African Museum a fine
fat, full-grown female, accompanied by her large family of thirty
or more just-born but pretty young, some of which were

1 That the class had once a world-wide diffusion is indicated by the wide
and discontinuous distribution of the living species,

440 ADAM SEDGWICK.

luxuriously creeping about on the beautiful skin of their
mother’s back.

THE SOUTH AFRICAN SPECIES OF PERIPATUS.

The following is a list of the distinct species of Peripatus
which have so far been found in South Africa: Capensis
(Table Mountain), Balfouri (Table Mountain), brevis
(Table Mountain), Moseleyi (near Williamstown). In ad-
dition to these there are a certain number of other possible
species, concerning the distinctness of which, however, I cannot
be certain.

General Characters of the South African Peripatus.

Peripatus with three spinous pads on the legs, with two
primary papillz on the anterior side of the foot (fig. 2), and a
small tooth at the base of the main tooth on the outer blade of
the jaw (fig. 28). The last fully developed leg of the males is
provided with a white papilla on its ventral surface (fig. 4),
and an enlarged crural gland. The generative opening is
always subterminal and behind the last pair of fully developed
legs. The ovaries are attached to the floor of the pericardium by
a ligament which passes off from their front end. The terminal
unpaired portion of the vas deferens of the male is short. The
ova are large, but with little food-yolk. The portion of the
proximal pad of the fourth and fifth legs, which carries the
opening of the nephridium, is separated by faintly marked
grooves from the rest of the pad (fig. 9). The legs appear,
with rare doubtful exceptions, to be constant in number in
all specimens of the same species. The median line of the
dorsal surface is destitute of pigment.

Peripatus capensis being the best known and the most easily
procured will be taken as the type of this group of species.

MONOGRAPH OF THE GENUS PERIPATUS. 441

Peripatus capensis.

South African Peripatus with seventeen pairs of claw-bearing
ambulatory legs.

The females are, on the whole, larger than the males, but
the difference between them is not very marked. A large female
would measure about 65 mm. (23 inches), and a large male
about 48 mm. There are, however, other external differences
between the sexes. The last leg of the male is smaller than
the preceding, and rarely touches the ground when the animal
is walking, while in the female it is as large as the others, and
used in walking. Further, the last leg of the male possesses,
on its ventral surface, a small white papilla (fig. 4), at the apex
of which opens its crural gland, which is much enlarged.

In the living animal (Pl. I) the skin has a_ beautiful
velvety texture. This is especially noticeable in the darker
specimens.

Colour.—The colour varies in different individuals. But in
general it may be said that the ventral surface has a light colour,
and that the dorsal is darkly pigmented. The principal colours
are two in number, which present every variation in tint in
different individuals and in different parts of the body of the
same individual. They are (1) a dark green, graduating to a
bluish grey ; (2) a brown, varying to a red orange.

The ventral surface is almost entirely free from the green
pigment, but possesses a certain amount of light brown. This
brown pigment is more conspicuous and of a darker shade on
the spinous pads of the legs. The only part of the ventral
surface where the green pigment is always present is the ven-
tral side of the foot, where it has a blue tint, and round the
lips (fig. 5). In the latter situation there are a number of green
papille, with which are intermingled a few of an orange
colour. Very rarely there is a suspicion of green pigment
along the middle line of the ventral surface, and in one spe-
cimen I found the distal pad of the leg to contain green
pigment,

442 ADAM SEDGWIOK.

While the colour of the ventral surface is practically the
same in all individuals, that of the dorsal differs in almost all.
The differences are due to the varying proportions in which
the green and brown pigments are present.

To facilitate matters I will describe two extreme cases: (1)
a very dark green specimen (fig. 1), in which the brown is
very inconspicuous, and (2) a red specimen, in which the brown
predominates.

(1) The skin between the close-set papillz, so far as it can be
observed, is a bluish grey ; but on the papille the pigment has
a very dark green colour, except on a few, in which it is
brown (even these may be absent). The ground colour, i.e.
the colour of the skin between the papille, varies in shade
in different places. On the dorsal sides of the legs and along
a dorso-lateral band at the bases of the legs, extending the
whole length of the body, it is lighter than elsewhere, while
on each side of the median dorsal white line it is much darker
than elsewhere. These differences may be partly due to the
closer aggregation of the papille in one place than in the other.

(2) The pigment of the skin and of most of the papille
is a reddish brown, except on each side of the dorsal white line
and on the dorsal sides of the legs where it is green. Scattered
amongst the brown papillze are a considerable number of green.
The brown colour is of a lighter shade along a dorso-lateral
band, extending the whole length of the body at the base of the
legs, and from this band green papille are almost entirely absent.
The brown pigment is, however, almost entirely absent from
the dorsal side of the legs on each side of the dorsal white line.

The conditions intermediate between these two extreme cases
are due to the variations in the number of the brown papille.
As a peculiarity of these intermediate cases may be men-
tioned the fact that the brown pigment extends into the skin
round the base of the brown papille, giving the brown
papille a brown setting, so that when a number of them occur
together the skin between the papillze has an entirely brown
colour (as in the brown specimens). Brown papille are most
numerous in the light band at the bases of the legs, and are

MONOGRAPH OF THE GENUS PERIPATUS. 443

sometimes so numerous that the ground colour in that region
is brown, though green elsewhere on the back.

The pigment is always present, whether on the papille or
between in minute square, pentagonal, and hexagonal patches.
The darkness of the skin is probably mainly due to the number
of these patches present in any given area.

The antenne are always green, the brown being almost
entirely absent from them, and they are the first to acquire the
green colour in the embryo. In fact the young at birth are
almost quite white excepting the antenne.

The colour seems hardly at all affected by the action of spirit.
The flesh-coloured brown of the ventral surface is sometimes
slightly reddened when the animal is first put into spirit, but
the red tinge soon vanishes, being apparently dissolved out
by the spirit, which in such cases becomes slightly coloured.

Ridges and Papille of the Skin.—The skin is thrown
into a number of transverse ridges, along which the primary
wart-like papille are placed.

The papille, which are found everywhere, are especially
developed on the dorsal surface, less so on the ventral. The
papille round the lips differ from the remaining papille of the
ventral surface in containing a green pigment. Each papilla
bears at its extremity a well-marked spine.

The ridges of the skin are not continued across the dorsal
middle line, being interrupted by the whitish line already
mentioned. Those which lie in the same transverse line as
the legs are not continued on to the latter, but stop at the
junction of the latter with the body. All the others pass round
to the ventral surface and are continued across the middle line;
they do not, however, become continuous with the ridges of the
other side, but passing between them gradually thin off and
vanish. The ridges on the legs are directed transversely to
the long axes of the legs, i.e. are at right angles to the ridges
of the rest of the body.

The papille of the dorsal surface are not arranged in a
single row in the ridges, but in more than one row, in fact a
ridge varies in thickness in different parts of its course.

4AA ADAM SEDGWICK.

Further, the dorsal ridges are interrupted by thin and sharp,
less coloured lines, which are somewhat diagonally arranged, and
divide the ridges into lozenge-shaped areas (vide figs. 1 and 10).

The antenne& are ringed and taper slightly till near their
termination, where they present a slight enlargement.

The rings consist essentially of a number of coalesced primary
papille, and are, therefore, beset by a number of spines like
those of the primary papille. They are more deeply pigmented
than the rest of the antenna.

The free end of the antenna is covered by a cap of tissue like
that of the rings. It is followed by four or more rings placed
close together on the terminal enlargement. There appears to
be about thirty rings on the antenne of all adults of this
species. But they are difficult to count, and a number of small
rings occur, between them, which are not included in the thirty.

The antenne are prolongations of the dorso-lateral parts of
the anterior end of the body.

The eyes are paired and are situated at the roots of the
antennze on the dorso-lateral parts of the head. ach is placed
on the side of a protuberance which is continued as the an-
tenna, and each presents the appearance of a small crystalline
ball inserted on the skin in this region.

The rings of papilla on that part of the head from which
the antennz arise lose their transverse arrangement. They
are arranged nearly concentrically to the antennal rings, and
have a straight course forwards between the antenne.

The oral papille are placed at the sides of the head. They
are attached ventro-laterally on each side of the lips. The
duct of the slime gland opens through their free end. They
possess two main rings of projecting tissue, which are especially
pigmented on the dorsal side ; and their extremities are covered
by papille irregularly arranged (vide description of oral papilla
of New Zealand species, p. 29).

The Buccal Cavity.—The buccal cavity has the form of
a fairly deep pit, of a longitudinal oval form, placed on the
ventral surface of the head, and surrounded by a tumid lip.

The lip is covered by a soft skin, in which are numerous

MONOGRAPH OF THE GENUS PERIPATUS. AAD5

organs of touch, similar to those in other parts of the skin,
having their projecting portions enclosed in delicate spines
formed by the cuticle. The skin of the lips is raised into a
series of papilliform ridges, whose general form is shown in
fig.5; of these there is one unpaired and median behind, and
a pair, differing somewhat in character from the remainder,
in front, and there are, in addition, seven on each side. The
cutaneous papille round the front of the lips are raised up
and appear like a second outer lip concentric with the anterior
part of the real lip, with the posterior part of which it is
continuous.

The structures within the buccal cavity are shown as they
appear in the surface views in fig. 5. In the median line of
the buccal cavity in front is placed a thick muscular protuber-
ance, which may perhaps conveniently be called the tongue,
though attached to the dorsal instead of the ventral wall of the
mouth. It has the form of an elongated ridge, which ends
rather abruptly behind, becoming continuous with the dorsal
wall of the pharynx. Its projecting edge is armed by a series
of small teeth, which are thickenings of the chitinous covering
prolonged from the surface of the body over the buccal cavity.
Where the ridge becomes flatter behind, the row of teeth
divides into two, with a shallow groove between them.

The Jaws.—On each side of the tongue are placed the
jaws, which are a pair of appendages, modified in the character-
istic arthropodan manner, to subserve mastication. They are
essentially short papille, moved by an elaborate and powerful
system of muscles, and armed at their free extremities by a
pair of cutting blades or claws. The latter structures are, in
all essential points, similar to the claws borne by the feet, and,
like these, are formed as thickenings of the cuticle. They have,
therefore, essentially the characters of the claws and jaws of
the Arthropoda, and are wholly dissimilar to the sete of
Chetopoda. They are sickle-shaped, and, as shown in fig. 5,
have their convex edge directed nearly straight forwards, and
their concave or cutting edge pointed backwards. The inner
cutting plate has five to eight teeth (fig. 27). The outer plate

44.6 ADAM SEDGWIOK.

has one main tooth (fig. 28), at the base of which is a small
tooth. This accessory tooth is found on the outer blade in
all South African species. Posteriorly, the behaviour of the
two blades is very different. The epithelial ridge bearing the
outer blade is continued back for a short distance behind the
blade, but the cuticle covering it becomes very thin, and it
forms a simple epithelial ridge placed parallel to the inner
blade. The cuticle covering the epithelial ridge of the inner
blade is, on the contrary, prolonged behind the blade itself as
a thick rod, which, penetrating backwards along a deep pocket
of the buccal epithelium, behind the main part of the buccal
cavity for the whole length of the pharynx, forms a very
powerful lever, on which a great part of the muscles connected
with the jaws find their insertion.

The Ambulatory Appendages.—The claw-bearing legs
are seventeen in number, and with the exception of the fourth
and fifth pairs in both sexes, and the last in the male, they all
resemble each other fairly closely. A typical appendage will
be first described and the small variations found in the appen-
dages just mentioned will then be pointed out. Each consists
of two main divisions, a large proximal portion the leg, and
a narrow dorsal, claw-bearing portion, the foot.

The leg has the form of a truncated cone, the broad end of
which is attached to the ventro-lateral body wall, of which it
appears to be, and is, a prolongation. It is marked by anumber
of rings of primary papille, placed transversely to the long
axis of the leg, the dorsal of which contain a green and the
ventral a brown pigment. These rings of papille at the
attachment of the leg, gradually change their direction and
merge into the body rings. At the narrow end of the cone
there are three ventrally placed pads, in which the brown
pigment is dark, and which are covered by a number of spines
precisely resembling the spines of the primary papille. These
spinous pads are continued dorsally, each into a ring of
papille.

The papille of the ventral row next the proximal of these
spinous pads are intermediate in character between the primary

MONOGRAPH OF THE GENUS PERIPATUS. 447

papillze and the spinous pads. Lach of these papille is larger
than a normal papilla, and bears several spines (fig. 2). This
character of the papille of this row is even more marked in
some of the anterior legs than in the one figured; it seems
probable that the pads have been formed by the coalescence of
several rows of papillze on the ventral surface of the legs.
On the outer and inner sides of these pads the spines are absent,
and secondary papille only are present.

In the centre of the basal part of the ventral surface of the
foot there is present a group of larger papille, which are of
a slightly paler colour than the others. They are arranged so
as to form a groove, directed transversely to the long axis of
the body, and separated at its internal extremity by a median
papilla from a deep pit which is placed at the point of junction
of the body and leg. The whole structure has the appearance,
when viewed with the naked eye, of a transverse slit placed at
the base of the leg. The segmental organs open by the deep
pit placed at the internal end of this structure. The exact
arrangement of the papille round the outer part of the slit
does not appear to be constant.

The foot is attached to the distal end of the leg. It is
slightly narrower at its attached extremity than at its free end,
which bears the two claws. The integument of the foot is
covered with secondary papille, but spines and primary papillee
are absent, except at the points now to be described.

On each side of the middle ventral line of the proximal
end of the foot is placed an elliptical elevation of the integu-
ment covered with spines. Attached to the proximal and
outer end of this is a primary papilla. At the distal end
of the ventral side of the foot on each side of the middle line
is a group of inconspicuous pale elevations, bearing spines.

On the front side of the distal end of the foot, close to the
socket in which the claws are placed, are two primary papille,
one dorsal and the other ventral.

On the posterior side of the foot the dorsal of these
only is present. The claws are sickle-shaped, and placed on
papille on the terminal portion of the foot. The part of the

VOL. XXVIII, PART 4,—NEW SER. HH

448 ADAM SEDGWICK.

foot on which they are placed is especially retractile, and is
generally found more or less telescoped into the proximal part
(as in the figure).

The fourth and fifth pairs of legs exactly resemble
the others, except in the fact that the proximal pad is broken
up into three, a small central and two larger lateral. The
enlarged segmental organs of these legs open on the small
central division.

The last (17th) leg of the male (fig. 4) is characterised
by possessing a well-marked white papilla on the ventral
surface. This papilla, which presents a slit-like opening at
its apex, is placed on the second row of papille, counting from
the innermost pad, and slightly posterior to the axial line of
the leg.

The anal papille, or as they should be called, genera-
tive papilla, are placed one on each side of the generative
aperture,

The generative aperture is subterminal and on the
ventral surface. It is inconspicuous in most specimens.

Internal Anatomy.—The points of internal anatomy
which require to be noted in an account of the species relate
entirely to the generative organs. In the male the ductus
ejaculatorius (posterior unpaired part of vas deferens, penis of
Moseley) is short, and the crural glands of the seventeenth
pair of legs are much elongated, reaching forward for a
considerable distance in the lateral compartment of the body
cavity.

In the female the ovaries are closely approximated and
short. They are united to the floor of the pericardium by a
single ligament passing off from their front end. Receptacula
seminis are absent. The ova contain but little food-yolk.
They are oval in shape, and the greatest length of an unseg-
mented ovum which has passed into the oviduct is ‘56 to 6 mm.

Habits.—They live beneath the bark and in the crevices
of rotten stumps of trees, and beneath stones. So far they
have only been found, so far as I can ascertain, in the woods
on the slope of Table Mountain. They require a moist atmo-

MONOGRAPH OF THE GENUS PERIPATUS. 449

sphere, and are exceedingly susceptible to drought. They
avoid light, and are therefore rarely seen, and it is owing to
this fact that, though fairly numerous, they were for so
long unknown to the inhabitants of the Cape Peninsula.
They move with great deliberation, picking their course by
means of their antenne and eyes. It is by the former
that they acquire a knowledge of the ground over which
they are travelling, and by the latter that they avoid the light.
The antenne are extraordinarily sensitive, and so delicate
indeed, that they seem to be able to perceive the nature of
objects without actual contact. When irritated they eject with
considerable force the contents of their slime reservoirs. The
force is supplied by the sudden contraction of the muscular
body wall. They can squirt the slime to the distance of almost
a foot. The slime, which appears to be perfectly harmless, is
extremely sticky, but it easily comes away from the skin of the
animal itself.

I have never seen them use their apparatus for the capture
of prey. So far as I can judge it is used as a defensive weapon ;
but this of course will not exclude its offensive use. They will
turn their heads to any part of the body which is being irritated
and violently discharge their slime at the offending object.
Locomotion is effected entirely by means of the legs, with the
body fully extended.

Of their food in the natural state we know nothing; but it
is probably mainly, if not entirely, animal. Those which I kept
in cavity eagerly devoured the entrails of their fellows, and
the developing young from the uterus, They also like raw
sheep’s liver. They move their mouths in a suctorial manner,
tearing the food with their jaws. They have the power of
extruding their jaws from the mouth, and of working them
alternately backwards or forwards. ‘This is readily observed in
individuals immersed in water.

The young are born in April and May. They are almost
colourless at birth, excepting the antenne, which are green, and
their length is 10to 15 mm. A large female will produce
thirty to forty young in one year. The period of gestation is

450 ADAM SEDGWICK.

thirteen months, that is to say, the ova pass into the oviducts
about one month before the young of the preceding year are
born. They are born one by one, and it takes some time for a
female to get rid of her whole stock of embryos; in fact, the
embryos in any given female differ slightly in age, those
next the oviduct being a little older (a few hours) than
those next the vagina.

The mother does not appear to pay any special attention to
her young, which wander away and get their own food.

There does not appear to be any true copulation. The male
deposits small, white, oval spermatophores, which consist of
small bundles of spermatozoa cemented together by some glu-
tinous substance, indiscriminately on any part of the body of
the female. Such spermatophores are found on the bodies of
both males and females from July to January, but they appear
to be most numerous in our autumn.

The testes are active from June to the following March.
From March to June the vesiculz of the male are empty.

Peripatus Balfouri (n. sp.).

South African Peripatus, with eighteen’ pairs of claw-bearing
ambulatory legs, of which the last pair ts rudimentary, with
white papille on the dorsal surface.

Peripatus Balfouri resembles very closely P. capensis.
The points of difference are as follows:

The dorsal skin has an olive-green tinge. The largest
papille are white, except at their free extremities, which are
green (fig. 10). The white spreads out a little round the base
of the papille. Brown tints are entirely absent in all the
specimens which I have examined except one.

The ventral surface is whiter than in Capensis, but the
papille are faintly green. The same remark applies to the
ventral surface of the legs.

The Ambulatory Appendages.—The claw-bearing legs
are eighteen pairs. The legs of the eighteenth pair are smaller

1 Peters (No. 25) states that there are in the Berlin Museum specimens
from the Cape with eighteen pairs of legs (see p. 25).

MONOGRAPH OF THE GENUS PERIPATUS. A451

than the rest (fig. 24). The remaining ambulatory appendages
resemble, except in the following points, those of Capensis.
The three spinous pads are green, and the middle one is broader
than the other two; the ventral surface of the proximal
part of the leg is white, and the papille are a bluish green;
the groove at the base is less marked than in Capensis.

The foot is rather more delicate than in Capensis, and the
spines on the ventral side of the base of the claws are placed
on two pairs of small papille (fig. 9).

The male is distinguished from the female, as in Capensis,
by the possession of a white papilla on the ventral side of the
legs of the seventeenth pair, in the same position as in
Capensis, and the legs of the eighteenth pair are smaller than
in the female. So small are they, indeed, that they are hardly
distinguishable from the large papille found near the hind end
of the body ; but they bear two claws, and a rudiment of the
foot may be made out.

In the female the legs of the eighteenth pair (fig. 24) present
the following features :—The foot seems to be normal and un-
reduced, but the leg is much reduced, presenting on the ventral
side only three rows of papillae and one spinous pad, which
indeed shows, in some specimens more than others, its consti-
tution of separate papilla. The pad and papille are all tinged
with green.

The embryos are much smaller than in Capensis. In pre-
served specimens the length of the fertilised ovum is *4 to °48
mm.; and a full sized adult specimen may reach the length of
43 mm. The generative orifice is between the rudimentary
legs of the eighteenth pair. As a peculiarity in habits may be
mentioned the fact that the individuals of this species nearly
always coil themselves into a spiral when touched, while Ca-
pensis simply contracts and shortens itself.

Locality—Table Mountain,

452 ADAM SEDGWIOK.

Other South African species.

In addition to these two South African species from Table
Mountain, the following varieties, some of which at least are
probably distinct species, are known.

1. One with fourteen pairs of legs, already named P. brevis
(Blainville). This species was found by M. Goudot beneath
a stone in the woods on Table Mountain. It has been
shortly described by Blainville in a note on p. 38 of Gervais’
«Etudes pour servir A Vhistoire naturelle des Myriapodes”
(Ann. d. Sci. Nat.,’ series ii, vol. vii) as follows :—‘ Corps
subfusiform pourvu de quatorze paires de pattes, noir
velouté en dessus, blanchatre en dessous; longeur totale en
comprenant les antennes, 43 mill.; plus grande largeur, 4
mill.”

2. Another with nineteen pairs of legs,’ reported by Mr.
Roland Trimen from Plettenberg Bay, Cape Colony, but
hitherto undescribed.

3. A third, in my possession, from Table Mountain with
twenty pairs of claw-bearing legs, I have found one specimen
only. Peters (No. 25) records the existence of specimens from
the Cape with twenty pairs of legs (see below p. 455).

4. A fourth, with twenty-one! pairs of legs from near
Williamstown, South Africa, I have only seen three speci-
mens. They are in the possession of the Indian Museum.

5. A fifth, with twenty-two” pairs of legs, of which two
specimens are known to me. One of these isin the possession
of the Indian Museum ; the locality is marked “ near Williams-
town, S. Africa.” The other was found by Mr. J. P. Mansel
Weale, and given by him to Mr. Wood Mason, who in his
turn gave it to Professor Balfour. This specimen, of which I
have not been able to ascertain the exact locality, is in my
possession, and is figured on Pl. XXXVII fig. 8.

1 There are specimens in the Berlin Museum with nineteen pairs of legs
(Peters, No, 25).

? Peters (No. 25) records the existence of specimens from the Cape with
twenty-one and twenty-two pairs of legs (see below, p. 455).

MONOGRAPH OF THE GENUS PERIPATUS. 453

Of these five varieties I have not seen the first two. I have,
however, had full opportunity of examining preserved speci-
mens of the last three and will now shortly describe my obser-
vations on them.

Peripatus with twenty pairs of claw-bearing legs.—
One specimen only—a female—is known to me. Locality,
Table Mountain. Length of spirit specimen 23 mm.

The specimen very closely resembles P. Balfouri, and would
be mistaken for the latter were not its legs counted. The skin
presents an identical appearance. The last pair of legs are
very small and rudimentary, and the generative opening, which
is subterminal just in front of the anus, is between them.

The first nineteen pairs of legs are all normal and resemble
exactly, so far as I can judge, those of P. Balfourt. In the
legs of the twentieth pair, while the foot is normal the leg is
much reduced in size. It is entirely without the spinous pads,
and possesses only three rows of papillz of which the row next
the foot is slightly tinged with green, the other two being
white.

The only other difference which I was able to detect between
this specimen and P. Balfouri consisted in the very small amount
of green on the ventral surface, which is almost white.

On the whole I am not inclined to establish at present a
distinct species for the reception of this specimen, but would
prefer to regard it provisionally as a variety of P. Balfouri.

Peripatus Moseleyi.

South African Peripatus, with twenty-one and twenty-two
pairs of legs.

All the specimens under this head presented the same
general appearance (fig. 24). Were it not for the number of
legs they would be taken for specimens of Capensis. The
ventral surface is light brown and the dorsal an olive green,
with scattered brown patches. Green is entirely absent from
the ventral surface, excepting on the foot and distal pad, and
sometimes a very little on the middle pad. On the dorsal sur-
face there is a band on each side at the base of the legs, in

454, ADAM SEDGWICK.

which the brown papille are so numerous as to cause the
appearance of a brown band. Elsewhere on the dorsal surface
the brown papillee are very sparsely scattered.

They are all reported from a part of South Africa far removed
from Table Mountain, the home of Balfouri and Capensis,
viz. near Williamstown (with the possible exception of the
one figured, the locality of which I do not know). Unfortu-
nately I have only been able to see preserved specimens, which,
on account of the great contraction they had undergone in
dying, were not very favorable for observation.!

Specimens of Peripatus Moseleyi with twenty-one pairs of
legs.—Skin as in Capensis. Foot and /egs as in Capensis.
At the bases of some of the legs (no constancy in the different
specimens), immediately internal to the opening of the segmental
organ, a small white patch of a tumid appearance is present.
It occupies the same position as the large tumid papillz on the
ventral side of the leg of Capensis (see Pl. XXXVI, fig. 2),
and has been noticed by Wood Mason (No. 23). The generative
opening is subterminal, and on each side of it there is an in-
conspicuous anal papilla. The dorsal side of the foot is marked
with streaks of green pigment, arranged parallel to its long
axis. The streaks are much less distinct on the anterior than
on the posterior feet.

Two of the specimens were smaller than the third, from
which they differed by possessing a distinct white papilla on
the last (twenty-first) leg, exactly resembling in appearance
and position the papilla on the last leg of Capensis. I
opened one of these smaller specimens, and found it to bea
male; while the larger specimen turned out to be a female.

The female was about 26 mm. in length, the male about
20 mm.

The specimens of P. Moseleyi with twenty-two pairs of legs
were both females. They resemble the specimens with twenty-
one legs, so far as I could see from a study of the contracted

1 Drowning (twenty-four hours or more) and then spirit is the best method

of killing Peripatus for museum purposes and observation of external
characters.

MONOGRAPH OF THE GENUS PERIPATUS. 455

specimens at my disposal, except in two points: (1) in the
absence of the anal papilla, and (2) in the fact that the streaks
on the dorsal side of the feet were entirely absent in the first
nine pairs of legs. The legs of the last pair resemble, in all
respects, the preceding, and the genital opening is behind
them.

One of these specimens measured 26 mm. and the other
30 mm. in length.

Inasmuch as I have not been able to find any marked charac-
ters associated with the character afforded by the number of
legs, and further, as I have had no opportunity of ascertaining
whether the latter character is transmitted in reproduction, I
am inclined not to establish two distinct species but to regard
the specimens with twenty-two legs to be a variety of the species
Peripatus Moseleyi, which is distinguished by the possession of
twenty-one legs, and a subterminal genital opening behind the
last legs.

Peters (No. 25) in a short paper on the variation of the
number of legs in P. capensis, states that the following speci-
mens were brought to him from Cape Town by a friend of his :
three specimens with 22 legs, eight with 21, eight with 20, one
with 19, one with 18, and two with 17 legs. He adds that
they were all found in the same locality, which, however, is not
mentioned. He gives no description of the specimens, beyond
mentioning the number of legs, and it is not therefore possible
to say whether heis right or not in his view that they all
belong to the species Capensis. I may add that, though I
have examined more than a thousand specimens from the
Cape Peninsula, I have only seen one specimen with more than
eighteen pairs of legs.

456 ADAM SEDGWIOK.

THE AUSTRALASIAN SPECIES.

GENERAL CHARACTERS,

Peripatus with fifteen pairs of claw-bearing ambulatory legs,
with three spinous pads on the legs, and a primary papilla pro-
jecting from the median dorsal portion of the foot (figs. 21, 21a).
The ventral organs are conspicuous, and the males are consi-
derably smaller than the females. The generative opening is
between the legs of the last pair, and there are no anal papille.
The number of legs are constant in all specimens. The ovaries
are attached by their whole length to the floor of the peri-
cardium, and each oviduct is provided with a receptaculum
seminis. The unpaired portion of the vas deferens is long and
complicated in structure. The ova are large and heavily
charged with food yolk. The portion of the proximal pad of
the fourth and fifth legs which carries the opening of the
nephridium is continuous distally with the rest of the pad
(fig. 21). A median dorsal white line is present.

Two species are known from the Australasian region;
P. nove-zealandie from New Zealand, and P. Leuckarti from
Queensland in Australia.

The former was first described by Captain Hutton (No. 19),
the latter by Saenger (No. 15).

Peripatus Nove-Zealandie.
(Figs. 7 and 17.)

Australasian Peripatus, without a small tooth at the base of
the main tooth of the outer blade of the jaw, and without a white
papilla on the ventral side of the last leg of the male.

The males are considerably smaller and less numerous than
the females. The length of a large female is 50 mm. (2 inches),
that of a large male 25 to 30 mm. in the extended condition
after drowning. There is no external difference which enables
us to distinguish the sexes. The ventral organs, owing to the

MONOGRAPH OF THE GENUS PERIPATUS. 457

character of their pigment, are much more conspicuous than
in the South African species.

Colour.—The colour varies in different individuals (c.f. figs.
7 and17). The ground colour varies exceedingly in tint ; it
consists of a bluish grey, or slate colour, or violet; it is
darker on the antenne than elsewhere, and is especially con-
centrated in small, dark, square, pentagonal, and hexagonal
patches lying close together over the whole surface of the
body. Sometimes the outline of these patches is darker than
the centre.

The pigment of the papille is also much darkened, but this
requires a separate description as the variations in the colour
of different individuals is mainly due to the papille. In all
specimens a certain number of the papille have brown or
orange pigment, which spreads out for a short distance around
the base of the papilla, as in the case of the white papille of
the South African Peripatus Balfouri, so that if many of these
papille occur close together the ground colour is brown or
orange and the slate entirely displaced ; if such are numerous,
they impart a distinctly brown aspect to the specimen. They
are scattered irregularly over the whole surface of the body,
but are most numerous, as in Capensis, in two bands on the
sides of the dorsal surface at the base of the legs, where, indeed,
in some specimens they almost completely replace the blue.

In most specimens, however, the greater number of papille
presents a pigment which resembles more or less closely that of
the ground colour. In many specimens—perhaps the majority —
the papille have a dark slate colour; but in some specimens
they may have a distinctly blue pigment, and occasionally even
a dark purple. The lips, as in Capensis, are always destitute
of pigment, and, as in that species, there isa sharp line extend-
ing along the middle of the whole length of the dorsal surface,
in which the pigment is either absent or of an extremely light
shade. On each side of this line the pigment in the papille is
much darkened. The pads of the legs vary much in colour.
In most specimens the distal one is blue, the middle one brown
or orange, and the proximal one brown or orange in the centre,

458 ADAM SEDGWICK.

and blue along the outer and inner border. In some specimens,
however, they are all blue, and in others all brown or orange.
In short, it may be said that the colour of the pad varies from
blue with hardly any admixture of brown, to brown or orange
without any blue. The distal pad is always the most blue.
The row of composite large papille next to the proximal pad
presents the predominant colour of the proximal pad. The
blue colour is always absent from the ventral organs, which are
either white, brown, or orange.

In all specimens there is a band of especially dark papill
extending from the ventral extremity of the leg towards the
ventral organ (fig. 19). The opening of the segmental organ
is placed in the outer end of this band. The ventral surface
is almost always mottled, the blue and yellow pigment being
distributed in patches; the colour in each kind of patch extend-
ing between the papille as well as on to them.

The ridges and papille of the skin are arranged as in
the South African species.

The antenne resemble those of the South African species.
They are ringed and slightly swollen near the free end (fig.
16). In none of the specimens that I examined did they
present any brown pigment. They are entirely of the blue
(violet 2?) grey colour, which forms the ground colour of the
skin. The rings are beset with spines, and are covered by
closely approximated patches of dark pigment such as have
been already described. On the anterior edge of the rings at
the front end of the antenne there is a row of hexagonal,
lighter-coloured spaces. At the bases of the spines also the
pigment is lighter than elsewhere on the rings. Between the
rings spines and patches are absent, and the pigment is of a
lighter colour. The free end of the antenna is rounded and
covered by a cap of integument resembling that on the rings
and bearing a large number of spines, as in all the species of
Peripatus that I have seen.

The eyes resemble in position and character (fig. 18) those
of the South African species.

The oral papillee resemble essentially those of Capensis.

MONOGRAPH OF THE GENUS PERIPATUS. 459

Fig. 20 shows very clearly the peculiar collapsable joints which
this appendage possesses in all the species.

The buccal cavity, tongue, and lips resemble in all
respects the same structures in the South African species.

The jaws differ from those of the latter only in being
without the small tooth on the outer blade.

The ambulatory appendages (fig. 21) are fifteen in
number in all the specimens which I have examined. They
resemble in their general features the same structures in
Capensis, so that in the following short description stress
will be laid only on the points in which they differ from the
latter.

The opening of the segmental organ at the base of the leg is
much more indistinct than in Capensis, andthe peculiar tumid
papille, which in Capensis extends from its outer horder on
to the ventral surface of the leg, are absent in this species.
There are three pads, but the large papille of the row adjoin-
ing the proximal pad are larger with regard to the ordinary
papille than in Capensis. Sometimes, indeed, they are so
large as to present the appearance, unless closely examined, of
one continuous spinous pad.

The foot differs from that of Capensis in the following
points. The two prominent papille, placed one on each
side (anterior and posterior) of the base of the foot are
absent. The dorsal side of the foot near the free extremity
possesses a papilla (fig. 21a), while the anterior face bears,
like the posterior, only one papilla. As in Capensis,
the opening of the nephridia of the fourth and fifth legs
are placed on a small portion of the proximal pad. The
part of the pad around the opening is only partly separated
from the rest (vide fig. 21). The fifteenth leg, so far as I
could ascertain, differs only in size (being slightly smaller)
from the preceding, and is without the white papilla found on
the last leg of the male of the South African species. Anal
papillz are never present.

Internal Anatomy.—<As already explained, I do not
propose to give in the monograph any detailed account of

460 ADAM SEDGWICK.

internal structure. It will be sufficient for my purpose to sum
up briefly the more striking differences between the various
species.

The internal structure of Peripatus nove-zealandie closely |
resembles that of the South African species. The differences,
so far as I have been able to observe them, chiefly concern the
generative organs, and the crural glands. It has recently been
shown by Miss Sheldon (No. 40) that the crural glands are
entirely absent from this species in both sexes.

The generative organs of the male differ from those of the
Cape species in three points, viz.: (1) In the much greater
length of the terminal unpaired portion of the vas deferens ; (2)
in the absence of any specially enlarged crural glands in the last
pair of legs; (3) in the fact (recently shown by Miss Sheldon,
No. 40), that the accessory glands, which are longer than in
the male of Capensis, do not open with the vas deferens, but
on the sides of the body outside the nerve-cord and close to the
hind end. The terminal unpaired portion of the vas deferens
is continuous with the two vasa deferentia (one of which passes
asin Capensis beneath the two nerve-cords) at the level of
the last pair of legs. Thence it is continued forwards for a
considerable distance (as far as the level of the eighth legs in
some cases) ; eventually bending round to pass backwards to
its opening between the last pair of legs. Its walls increase in
thickness from before backwards, and are of a distinctly gela-
tinous consistency in the greater part of their course.

The generative organs of the female differ from those of
Capensis in two main points, viz. : (1) The two ovarian tubes
are much longer, extending from the level of the eleventh to
that of the thirteenth, and sometimes to that of the fourteenth
leg, and are entirely separate from one another, each being
suspended throughout its entire course to the pericardial floor
by a distinct membrane. (2) There are two spherical re-
ceptacula seminis, each of which opens into the oviduct by
two ducts; and the oviduct in the neighbourhood of these
openings is slightly sacculated.

It will be remembered that in the Cape species the ovarian

MONOGRAPH OF THE GENUS PERIPATUS. 46]

tubes were closely applied together and united to the pericardial
floor only at their anterior extremities by a single band.

Spermatozoa have been found in the receptaculum and in
the oviduct near the opening of the latter. There are few, if
any, spermatozoa in the ovary. I have not been able to see,
though I have examined live specimens with great care, a trace
of cilia in any part of the female organs. It will be seen from
the above that I take exception to Captain Hutton’s! descrip-
tion of the ovary as an ovate organ.

The ova are large, oval in shape, and heavily charged with
food-yolk. They are surrounded by a membrane of the same
nature as the egg membrane of P. capensis, but much
tougher. The greatest length of an unsegmented ovum from
the uterus is about 1°5 mm., the breadth ‘8 mm. The greatest
number of embryos found in one animal was eighteen, twelve
in one uterus and six in the other. But the number varies in
the different specimens. Captain Hutton found eighteen in
one uterus and eight in the other. The same naturalist states
that ‘‘ when the embryos are numerous there is a considerable
difference in the point of development to which they have
attained.” I can confirm this statement; but the greater
number of the embryos in any given animal are of the
same age.

Habits.-—Captain Hutton (No. 19) has fully described
the habits of this species. He says:

*‘They live in decayed wood, under stones, or in crevices
of rock. They are nocturnal, but will feed in the daytime
when hungry. They feed upon animals. I have seen one
shoot out its viscid fluid from the oral papillz at a fly intro-
duced into the jar in which it was confined, and stick it down;
it then went up and sucked its juices, rejecting the whole of
the integument. This viscid fluid is for offensive and not
defensive purposes. In the winter they become half torpid,
though procreation still goes on. During this time of the year
I have never seen them feed, and they cannot emit their viscid

‘I have not been able to see any trace of the lateral vessel of Captain
Hutton.

462 ADAM SEDGWICK.

fluid, or only in very small quantity. They move with de-
liberation, entirely by means of their legs, the body being
much lengthened. When walking, the antenne are constantly
moved about as feelers. If a needle is placed upright imme-
diately in front of one, the antenna is drawn past it without
actual contact ; but the points of the hair probably touch the
needle. Although viviparous, the eggs are often extruded
before the development is complete, but these always die.”

From the study of the living specimens brought by Mr.
Evans I have been able to confirm Captain Hutton’s observa- .
tion as to the habits of the species, so far as it was possible to
do so on imported specimens.

I have not been so fortunate as to see them catching flies
with their slime, but this is not to be wondered at considering
the greatly changed conditions in which I observed them. In
fact I have failed to keep the specimens alive for any length
of time in this country.

Having received two lots, one in July and the other in
December, I am able to make some conjectures as to the period
of gestation. Captain Hutton asserts that they breed all the
yearround. The only other statement concerning the breeding
is, so far as I know, by Moseley (No. 20). He states that the
young are born in July. This is undoubtedly correct, for the
live specimens received by me at the end of July gave birth to
fully-developed young on the voyage, and directly after reach-
ing England, and those examined contained, in the great
majority of cases, either old embryos or none at all.

On the other hand, the specimens which came in December
contained, in the great majority of cases, unsegmented and
segmenting ova. But in a few (small specimens) the uterus
was empty, and again, in a still smaller number, there were
old embryos, and in some a few old embryos coexisted with the
more numerous young ova. These observations seem to me to
show that the eggs pass into the uterus in November and De-
cember, and that the young are born in July ; in other words,
that the period of gestation is eight or ninemonths. This con-
clusion is, however, not borne out by Captain Hutton’s state-

MONOGRAPH OF THE GENUS PERIPATUS. 463

ment! (No. 19), that he has ‘‘ never opened one which did not
contain embryos ;” and that he found the uterus full of embryos
in September and November. It must be admitted, therefore,
that the point cannot be settled on the evidence before us. It is
much to be regretted that none of the New Zealand naturalists
have taken the trouble todetermine a point so easy of observation.

With regard to the sexual relations, I am inclined to think
that copulation does not take place, and that the end of the
vas deferens, which I have called the ductus ejaculatorius,
is not protrusible. I have, indeed, observed in spirit specimens
small white ovoid bodies, which closely resemble the spermato-
phores of the South African species, and I think there can be
no doubt that the sexual relations are the same as in those
species. The period of the year at which fertilisation is
effected is unknown. Hutton has observed that the recepta-
cula contain spermatozoa in November, but are empty in
September. This observation distinctly confirms my deduction
that the ova pass into the oviduct in November or December.

Before leaving this subject I may mention that I can entirely
confirm Hutton’s statement that the eggs are often extruded
before the development is completed. This may possibly be a
reminiscence of the time, probably not very remote, when the
eggs were laid in the normal Arthropodan manner—a view
which receives support from the thick shell, large size, and
heavily yolked nature of the ovum of this species.

Peripatus Leuckarti.
Locality, Queensland.

Australasian Peripatus, with fifteen pairs of legs, an acces-
sory tooth on the outer blade of the jaw, and a white papilla on
the ventral side of the last leg of the male.

The following observations were made on two specimens most
kindly placed at my disposal by Professor Jeffrey Bell, to whom
they were sent by Dr. Ramsay, of Sydney. They were found

1 Tt should be noted that Hutton does not state whether his observations

were spread over the whole year.
VOL, XXVIII, PART 4.——-NEW SER. rod

464 ADAM SEDGWICK.

(vide No. 44) near Wide Bay in Queensland. The finder’s name
has not been communicated to me.

Both the specimens were much contracted and the feet
bent ventrally on the legs, so that it was difficult to get a good
view of the ventral surfaces of the feet.

The length of large specimen . : . 16—17 mm.
oe) 2” small 33 . . . 9—10 mm.

The large specimen was a female, and the small a male.

Generally it may be said of these specimens that they
resemble almost exactly the New Zealand species. After
careful search I have only been able to find three minute
points of real difference between them. These are:

1. The outer blades of the jaws have an accessory tooth at
the base of the main tooth, as in the Cape species.

2. The male has a rounded white papilla on the ventral face
of the fifteenth leg, on each side of the genital opening. It is
in the same position with regard to the leg as the corresponding
structure in the Cape males.

3. The pigment on the ventral surface is much less con-
spicuous in this than in the New Zealand species, so that the
mottled appearance presented by the ventral surface of the
latter species is not found in these specimens. The pigment
on the ventral surface of these specimens is much more
marked in the lower parts of the papille than elsewhere. In
the skin between the papille and at the apices of the papillz
the pigment is so faint as to be hardly discernable. The result
is that to the naked eye the ventral surface appears quite pale
with coloured papille projecting from it. The predominant
pigment of the ventral surface is the blue, but orange is
present. The hind end of the ventral surface in the region of
the last three legs is darker than elsewhere, in consequence of
the great number of the pigmented papille.

In addition to the above characters, it may be mentioned
that the genital papilla of the female is remarkably prominent,
and bears at its free end a longitudinally disposed slit. In the
male the genital papilla is fairly prominent, but its aperture is

MONOGRAPH OF THE GENUS PERIPATUS. 465

wider and more rounded, resembling the same structure in
both sexes of the New Zealand species. I append a short
general description of the two specimens.

There are fifteen pairs of legs. The ventral surface is pale,
dotted uniformly with pigmented papille, which are more
numerous behind. The dorsal surface is dark, and has a
median white line. The pigment is of the two colours found in
the New Zealand species, viz. bluish to green, and orange to
brown. The blue pigment is much the most conspicuous on
the dorsal surface. The antennz are blue mainly, but possess
some orange pigment arranged in rings round their basal
halves.

The genital papilla, which is remarkably prominent in the
female, is between the legs of the fifteenth pair. The feet
and legs resemble exactly, so far as could be made out, those of
the New Zealand species. The feet have the median dorsal
papilla so characteristic of that species; there are three pads
on the legs, and a patch of blue pigment round the opening of
the nephridia.

If there is any difference, it relates to a faint double row of
somewhat turgid papille proceeding outwards from the open-
ing of the nephridium along the ventral surface of the leg.
The same feature is present in a much more marked form in
P. capensis. The opening of the nephridium is_ perhaps
slightly more conspicuous than in the New Zealand species.
The last leg of the male presents a white papilla on its ventral
surface. The outer blade of the jaw has an accessory tooth.

The internal anatomy resembles, so far as I could make out,
that of the New Zealand species.

In the female the ovaries were attached along their whole
length, and possessed numerous oval white eggs of an average
length of ‘27 mm. In addition there were some larger eggs of
a yellowish colour, some of which were attached to the ovary,
and some broken away and lying in the body cavity. The
largest of these measured *75 mm. in length. They were full
of yolk and without any visible membrane.

Each oviduct possessed the receptaculum seminis in a

466 . ADAM SEDGWICK.

position similar to that of the same structure in Peripatus
nove-zealandie. The uterus was empty.

In the male the genital organs were normal, and the
unpaired portion of the vas deferens was long, and apparently
of a similar structure to that of the New Zealand species.

_ The specimens were not sufficiently well preserved for an
examination for the accessory glands.

PERIPATUS FROM THE NEOTROPICAL REGION.

Peripatus is found all over the northern part of the Neo-
tropical region. It is reported from Chili, Columbia, Cayenne,
Venezuela, Nicaragua, and from many of the West Indian
Islands, viz. Jamaica, Cuba, Trinidad, St. Thomas. I unfor-
tunately have only been able to make a complete study of the
species from Venezuela and of that from Demerara; of some
of the remainder I have only seen single specimens, or specimens
the preservation of which was not sufficiently good to allow of
the determination of specific characters. A partial exception
must be made in favour of the small species from Trinidad, of
which Dr. J. v. Kennel has been good enough to send me two
specimens ; but these were, unfortunately, somewhat contracted
and not sufficient in number to enable me to generalise as to
specific characters. I trust, however, that the careful description
of the Venezuela species, the specimens of which were collected
by Professor Ernst at Caracas, and given to Professor Balfour,
will form a groundwork on which future collectors of Peripatus
from this region will be able to work.

Although I have failed in determining the relations between
the various specimens of Peripatus which have been found in
the Neotropical region, still 1 have seen enough to be able to
establish a certain number of characters which distinguish a
great number—and probably all—of the neotropical Peripatus
from those found in other regions. Those characters are stated
in the following definition :

MONOGRAPH OF THE GENUS PERIPATUS. 467

General Characters of the Neotropical Species.

With four spinous pads on the legs, and two papille on the
anterior side of the foot. With generative aperture between
the legs of the penultimate pair. Dorsal white line absent,
and papille arranged in a single row on the ridges of the skin.
Many of the primary papille have a terminal portion slightly
constricted off from the main portion. Outer blade of jaw with
one minor tooth, inner blade with one minor tooth next the
main tooth (fig. 25), and a row of smaller minor teeth separated
from the latter by a diastema. Unpaired part of vas deferens
of greatlength. Ovary with oviducts entering its anterior end,
and attached to pericardial floor by a single band of great
length from the opposite end. Each oviduct provided with
a receptaculum semiunis with double duct, and with a thin-walled
receptaculum ovorum. Ova minute without yolk. Embryos of
very different ages in same uterus, and births probably taking
place all the year round. Males generally smaller than females,
and frequently with a smaller number of legs. The number
of legs often inconstant in the same species in the same sex;
in fact it may be said that the number of legs varies in all the
Neotropical species which are at all well known. The opening
‘of the nephridium of the fourth and fifth legs is on a papilla
which is quite separate from the third pad (fig. 11).

Eeecaa Edwardsii.

Neotropical Peripatus from Caracas with a variable number
-of legs—the smallest number being twenty-nine and the yreatest
thirty-four. Males with twenty-nine and thirty legs, and
tubercles on a varying number of the posterior legs. The basal
part of primary papille are cylindrical.

1 propose to reserve the name Edwardsii for the Neo-
tropical species, which is best known, viz. that from Caracas.
This has been described by Ernst and Gaffron. Whether the
specimens obtained by Audouin and Milne-Edwards from
‘Cayenne and named Edwardsii by Blanchard (No. 8) belong to

\

468 ADAM SEDGWICK.

this species cannot be definitely settled until more specimens
come to hand.

In P, Edwardsii the females are larger than the males
and have a greater number of legs. This fact was first noticed
by Gaffron. He found that the males possessed either twenty-
nine or thirty legs, while of his females one had thirty-four,
four had thirty-two, and four thirty-one. In my specimens,
which came, I believe, from the same place as the specimens
which Gaffron used for his! second paper (No. 35), all the
males had thirty or twenty-nine legs (four with thirty and
three with twenty-nine), while of the females three had
thirty-one, four had thirty-two, and one twenty-three
(fig. 6). Ernst states that the full-grown animal has
thirty-one pairs of legs, the new-born young but twenty-
nine; and he deduces from this that the young are born
with an incomplete complement of legs, and that new legs
make their appearance in the subsequent growth of the
animal. This, if true, would be important, as in no species of
Peripatus that I know of are the young born imperfect in
this respect. I therefore examined the number of legs of the
oldest embryos in my specimens with great care, and the result
of my observations is in entire contradiction to Ernst’s state-
ments. The embryos I found differ in the number of legs,
just as do the adults, the greatest number being thirty-two
pairs and the smallest twenty-nine. If this is so there can be
no doubt that the new-born young differ in the same manner.
To take an instance: from the lower end of the uteruses of
the four specimens with thirty-two legs I obtained in all seven
embryos, which were practically fully-developed and ready for
birth. Of these, four had twenty-nine legs, two had thirty-one,
and one had thirty-two—an embryo with twenty-nine and one
with thirty legs were found in the same mother; and I have
also found instances of a quite immature embryo (but possessed
of the full number of legs) with a greater number of legs than
the large mature embryo which occupied the part of the uterus

' The specimen which Gaffron used for his first paper was from Trinidad,
and had thirty-two legs,

MONOGRAPH OF THE GENUS PERIPATUS. 469

next the external opening. Considering the easy nature of the
observations required, Professor Ernst’s statements display a
very extraordinary method of work.

Colour.—My observations on this point were made on
spirit specimens, and cannot therefore have the value of those
of Ernst, who had the living animals before him. He says:
“ The colour is brownish black, with a diffused black line on the
middle of the back; the ventral side is dark flesh-coloured.”

In all my preserved specimens the colour was brown,
darker in some than in others ; in the specimen figured it is
as dark as in any in my possession. The ventral surface,
moreover, was of the same colour as the dorsal. As these
specimens came from Caracas, and have become distinctly
paler since I first saw them, it seems pretty clear that the
colour of this species is much affected by spirit. It will be
remembered the brown pigment of P. capensis was changed
by the action of spirit. The same fact has been observed
by Grube (see below, p. 480), who found that the pigment was
partly dissolved by the spirit, and also by myself in some
specimens brought alive from Demerara by Mr. W. L. Sclater
(No. 41).

The Ridges and Papille of the Skin.—The ridges are
more clearly marked, and the papillze of the dorsal surface are
less numerous. The dorsal white line is not present, so that
the ridges are continuous right across the dorsal middle line.
Further, there is for the most part only one row of papille on
each ridge, whereas in the South African and New Zealand
species there is considerable irregularity in this respect.
The fine diagonal lines, which break up the rows of papille
into lozenge-shaped areas, are absent in this species. The
ridges extend for the most part right across the dorsal surface,
but here and there, particularly at the level of the legs, there
are accessory ridges extending across the middle line and
stopping short a little distance on each side of it. They cause
a slight deflection of the contiguous main ridges (fig. 6).
Many of the papille—particularly those on the legs—are
divided by a constriction into two main portions (fig. 12)—

470 ADAM SEDGWIOK.

a free portion bearing the spine and a larger basal part. The
basal part is cylindrical, and the terminal portion often of
considerable size. Those immediately round the lips appear to
be without this characteristic.

The antenne present no features of specific interest. The
tongue and lips are without pigment and have the typical
form.

The jaws present differential characters. The outer blade
(fig. 26) has a well-marked minor tooth in addition to the
main one. On the inner blade the number of minor teeth
varies (generally eight), and the anterior of them is close to the
main tooth and larger than the rest, which are separated from
it by a diastema (fig. 25).

The oral papille are normal.

A typical ambulatory appendage presents the following
characters (fig. 12). The leg possesses four spinous pads, a
strongly marked, rather deep groove in the position of the
tumid papille of Capensis, i.e. a groove placed on the ven-
tral surface of the leg, and extends from the opening of the
nephridium as far as the third or fourth row of papille from
the proximal pad. This groove may be widely open as in the
leg figured, or its edges may be approximated so that it appears
as a slit. The papille at its margin are somewhat larger and
more indistinct than the ordinary papille. The foot resem-
bles that of Capensis in possessing two papille on its
anterior face, but the two basal papille are absent.

The opening of the segmental organ of the fourth and fifth
legs is on a papilla which is placed on the proximal side of, and
quite separate from, the third pad, between it and the proximal
pad (fig. 11). This feature is found in all the neotropical
species which I have examined. On certain of the posterior
legs of the males there are two and sometimes one smooth
white tubercle with an opening at their extremities (fig. 22).
They are placed close behind the groove, and are found only
on the posterior legs. Their exact arrangement varies in
different individuals. To give examples :

MONOGRAPH OF THE GENUS PERIPATUS. 471

In a male with 30 legs :

Right side-—Legs 21—24 inclusive, each had one such tubercle ; legs
25—28 inclusive, each had two such tubercles; legs 29 and 30 were
without them.

Left side—Leg 21 had one tubercle; leg 22 was without one; legs 23
and 24 each had one; legs 25—28 inclusive, each had two; legs 29
and 30 were without them.

In another male with 30 legs:

Right side—Leg 23 had one tubercle ; legs 24—28 inclusive, each had
two; legs 29 and 30 were without.

Left side—Leg 22 had one ; legs 23—28 inclusive, each had two; legs
29 and 30 were without them.

When one papilla only is present it is the distal one.

From these examples it is obvious that the arrangement of
these tubercles is different not only in the different individuals
but also on opposite sides of the same individual. The last
two legs are always without them. Gaffron found precisely
the same irregularity in the arrangement, but in his specimens
they were symmetrical. In one with thirty legs, the twenty-
second leg had one, and legs 23 to 28 each had two tubercles.
While in another male with twenty-nine legs, leg 20 had
only one, while legs 21 to 27 each had two. The pits at
the apices of these tubercles are, according to Gaffron, the
openings of glands corresponding to the crural glands of
Capensis.

The legs of the last pair are smaller than the penultimate,
and possess only two spinous pads. The legs of the penulti-
mate pair are without the nephridial opening, and the pedal
groove is inconspicuous as it is in the last pair. I could not
satisfy myself whether the legs of the last pair possessed a nephri-
dial opening ; but Gaffron states that they possess a nephridium.

Gaffron (No. 35) describes a peculiar bean-shaped papilla,
placed in a pit of the integument on the dorsal surface of the
leg near the foot. Its surface is smooth as is also the lining
of the pit in which it is placed. It is found in the Trinidad
species, and may very probably turn out to be characteristic
of the neotropical species.

472 ADAM SEDGWICK.

Internal Anatomy.—Excepting the generative organs
there is nothing in the internal anatomy of this species which
deserves notice here. The generative organs, of which we have
an excellent description by Gaffron, do, however, present some
features of interest. The generative opening in both sexes is
between the legs of the penultimate pair. The oviduct end of
the ovary is directed forwards, and the ovarian ligament, which
is attached to the opposite end of the ovary, is of great length,
being attached to the pericardial floor between the twenty-fifth
and twenty-sixth pairs of legs. A globular receptaculum seminis
(with two short ducts) opens into the anterior part of each
oviduct. Immediately in front of the receptacula each oviduct
gives off a short diverticulum, called “ cecum” by Ernst,
‘* zipfelformige Anhang,” and “ ovarial-trichter ” by Gaffron.
Gaffron, who at first thought that this process opened at its
free end into the body cavity, now accepts Kennel’s statement
that it opens into a small vesicle with extremely thin walls.
Kennel calls this vesicle the receptaculum ovorum. I have
seen the process, but, unfortunately, have no observations on
its termination; but I am strongly inclined, on theoretical
grounds, to think that Kennel is correct in his statement as to
the delicate vesicle. The generative ducts are the modified
nephridia of the segment on which the external opening is
placed: this is proved, on the one hand by their develop-
ment, and on the other by the fact that nephridia are absent
from the penultimate legs, between which the generative -
opening is placed. Now, it has been shown by me (No. 39)
that all the nephridia open internally, not into the body cavity
as has been supposed, but into a small vesicle with extremely
delicate and thin walls. It thus appears that the presence of
this delicate vesicle of the receptaculum ovorum is another
proof—if another were wanted—that the oviducts of Peripatus
are modified nephridia. No such structure has been found in
the New Zealand species, but, possibly, further investigations
may come upon it. In Capensis, for reasons which I have
set forth elsewhere (No. 39, Part 3), one would not expect to
find this structure.

MONOGRAPH OF THE GENUS PERIPATUS. 473

The male organs differ from those of the Cape species and
resemble those of the New Zealand species in the fact that the
common posterior part of the testicular ducts is of great length.
A very good description of it has been given by Gaffron. <A
pair of accessory glands is present in the male. They open on
each side of the anus (Gaffron).

Nephridia are present in the legs of the last pair but are
absent from the penultimate legs, between which the genera-
tive opening is placed (Gaffron, No. 35). With regard to the
crural glands, Gaffron states that they are absent from the
female, and only present in the males in those legs provided
with the tubercles described above.

The ova are small and without yolk. Their development has
been described in a closely-allied species from Trinidad by
Kennel, according to whom the embryos acquire a placental
connection to the uterine wall and an amnion. These struc-
tures are, however, said to disappear after a certain stage is
reached, and there is reason to doubt whether they have the
relations, significance, and method of development which
Kennel ascribes to them (Sclater, No. 46).

The uterus contains embryos in all stages of development,
and the young, which are fully developed at birth, are pre-
sumably born at different times of the year.

The length of mature embryos of Peripatus Edwardsii, lying
stretched out in the uterus with head near generative opening,
is about 20 mm.

The length ofa large adult female is 55 to 60 mm. The males,
of course, are rather smaller.

Habits.—The habits of this species are apparently much
the same as in the other species. A large number of speci-
mens were found by Ernst in a yard of the University building
of Caracas under heaps of rubbish.

Peters (No. 24) mentions specimens from the following
localities in Venezuela :—Caracas, Puerto Cabello, Laguayra.
He states that some of the specimens from Puerto Cabello
have thirty and others thirty-two pairs of legs.

474, ADAM SEDGWICK.

Peripatus from Demerara.

In January of this year (1887) Mr. W. L. Sclater brought
to England twenty female specimens of Pe ripatus collected at
Maccasseema, on the Pomeroon River. The specimens, when
they came into my hands, were torpid and apparently at the
point of death, and it was necessary to open them at once and
remove the embryos. I was unable therefore to make a de-
tailed examination of them in the fresh state.

Mr. Sclater has already (No. 41) given a short description
of the specimens. To his description I add here notes of my
own observations, made on the first arrival of the animals, and
an account of those which I have since made on their preserved
bodies.

All the specimens (twenty in number) were females. The
colour was a dark brown on the dorsal surface, with a median
diffuse dark stripe, such as Ernst describes. The antenne
were of a darker colour than the rest of the body. The ven-
tral surface was higher than the dorsal—a kind of flesh colour.
The animals turned quite red in spirit, and the red colouring
matter was gradually dissolved by the spirit leaving them a
lighter brown.

In well-grown specimens the uterus contained ten embryos
in each horn, of which the fifth from the ovary was gene-
rally in the spiral stage. The receptacula ovorum seemed to
contain ova, which were’‘038 mm. in diameter. The large
eggs in the ovary were the same size.

In one specimen, which I carefully examined for the pur-
pose, there were cilia in the receptacula seminis in the
position described by Gaffron. There can be no doubt of
their presence. I saw them in active movement. I am very
glad to have had the chance of confirming Gaffron on this
point. The older embryos had the same colour as the adult.
I could not be certain of the presence of spermatozoa in
either the receptaculum seminis or in the oviduct. If present
at all, they must have been few in number.

MONOGRAPH OF THE GENUS PERIPATUS. A475

To these observations I have now the following to add:
The colour, under the prolonged action of spirit, has become
lighter. The antennz, oral papille, jaws and legs, resemble in
all respects the same structures in the Caracas specimens.
The grooves on the legs were for the most part closed and
therefore slit-like. None of them possessed tubercles.

Mr. Sclater has the following statement on the slits and
tubercles. ‘In my specimens and in that from Dominica, the
openings (i. e. the slits) are in many cases roundéd, and some-
times have attached to them a bladder-shaped appendage.”
I donot quite understand this passage, but if it means that
the slits are round openings and that there are tubercles in the
specimens he brought from Demerara, I cannot confirm his
statement. It is unfortunate that no males are to hand, as it
is important from a systematic point of view to know if they
have the tubercles such as are found in the Caracas species,
and if they differ from the females in the number of legs. The
length of a large specimen was 55 to 60 mm.

My. Sclater states that all the specimens examined by him,
including those taken from the uterus, had thirty pairs of legs.
Mr. Sclater’s observations must have been confined to a very
small number of his specimens. I examined fourteen adults:
of these seven had thirty pairs of ambulatory legs, six had
thirty-one, and one had twenty-seven. Out of thirteen
embryos examined seven have thirty pairs and six have
thirty-one. Unfortunately, I did not notice that the adults
varied in the number of their legs, until after the embryos had
been removed from all except the specimens with twenty-seven
pairs of legs; so that it was not possible to determine, except-
ing in this case, whether the young resembled their parents in
this respect. Out of four embryos which had already developed
the full complement of legs and were removed from the speci-
men with twenty-seven pairs, three had twenty-seven and one
had twenty-eight pairs of ambulatory legs, so that it appears
that the number of legs varies in the species.

The only other difference between these specimens and those
from Caracas that I could detect, related to the primary

476 ADAM SEDGWICK.

papille on the skin. In the Caracas species, as already men-
tioned, these have comparatively narrow cylindrical bases,
and the diameter of the tops is often almost as great as that
of the basal portion. Inthe Demeraran specimens, on the other
hand, the lower portion of the papillae have the form of trun-
cated cones with very broad bases, while the tops are relatively,
_and I think absolutely, much slenderer than in the Caracas
specimens. The papille figured by Gaffron (No. 34, Pl. VII,
and here reproduced fig. 27) resemble those of the Demerara
specimens, and will serve for comparison with the papillze of
the Caracas specimens shown on Pl. XXXVIII, fig. 12. I
propose provisionally to regard these specimens as belonging
to a distinct species, and to call it P. demeraranus! with the
following characters. Neotropical Peripatus with twenty-seven
to thirty-one pairs of ambulatory legs and cylindrical primary
papille. Locality Maccasseema, Demerara.

Peripatus from Trinidad.

Dr. J. v. Kennel (No. 31) found two distinct species of
Peripatus in Trinidad; one of these he calls P. Edwardsii
and the other P. torguatus. His description of both is unfor-
tunately extremely meagre.

The species which he calls Edwardsii possesses twenty-
eight to thirty pairs of legs (No. 32). The generative opening
is between the legs of the penultimate pair, and the generative
organs present the characters of the Neotropical species.

Dr. Kennel was kind enough to send me two of this species
in spirit, and I am able to supplement his description.

' Sclater (No. 46) gives the name im Thurmi to the specimens with thirty
pairs of legs, which he has observed. It is of course quite possible that the
specimens with thirty pairs may be specifically distinct from those with
twenty-seven and thirty-one pairs. This, however, as stated above, I do not
regard as probable. On account of this uncertainty, and also because of the
further uncertainty as to whether the Demeraran specimens are specifically
distinct from species already determined and named, I propose the provisional
name of Demeraranus to include all specimens from Demerara, whether the
number of legs be twenty-seven, thirty, or thirty-one pairs,

MONOGRAPH. OF THE GENUS PERIPATUS. 477

One of these specimens had thirty-one pairs of legs and the
other thirty, from which it appears that Kennel, like so many
other zoologists who have examined Peripatus, has not been
very careful in counting the legs. The dorsal surface was of a
chocolate colour, the ventral surface being a light brown.
The papille and ridges of the skin presented the features
characteristic of the Neotropical species. The bases of the
primary papille are conical asin Demeraranus. The jaws
also presented no points of difference from those of the species
from Caracas, excepting that possibly the number of minor
teeth was rather larger: in one I found as many as eleven.

I think there can be no doubt that this is a distinct species,
and I propose to call it and define it as follows:

Peripatus Trinidadensis (Hdwardsii, Kennel).

Peripatus from Trinidad, with twenty-eight to thirty-one
pairs of ambulatory legs, and a large number of minor teeth on
the inner blade of the jaw. The basal portions of the primary
papille are conical.

Peripatus torquatus (Kennel),
Peripatus from Trinidad of large size, with forty-one to
forty-two pairs of ambulatory legs. The head is marked off
from the body by a bright yellow band on the dorsal surface.

The larger species is named /. torquatus, and Kennel gives
the following description of it. ‘‘ The females reach the length
of 15 cm., with a diameter of 8 mm., while the males have a
length of about 10 cm. The colour of the dorsal surface is
red brown, the middle line of the back being somewhat darker,
and paling off towards the sides. The head with the tentacles
is black and is marked off from the body on the dorsal side
by a bright yellow band, which often shows a small interrup-
tion in the middle line. The ventral surface has a dark flesh
colour. There are forty-one or forty-two pairs of legs.

This completes the list of the Neotropical Peripatus of
which we have anything like detailed knowledge. The remain-

473 ADAM SEDGWICK.

der of this monograph will be devoted to a statement of all that
is known with regard to the specimens found in other localities.

1. The original species found by Guilding (No. 1) in the
forests of the Island of St. Vincent in the Antilles, and called
by him P. juliformis, possessed thirty-three pairs of legs, and a
dark line down the centre of the back. The generative opening is
apparently immediately in front of the penultimate legs. The
animal was of a fair size, being three inches in length by three
lines in breadth. It is apparently similar in all essential
respects to other neotropical Peripatus, and I am inclined to
maintain for the present the species, and to define it as
follows :

Peripatus juliformis (Guiiding).

Neotropical Peripatus from St. Vincent, with thirty-three
pairs of ambulatory legs.—This definition is exceedingly un-
satisfactory because it is based on the number of legs, which,
as I have stated, varies in all the species which have been
closely examined.

2. The species described by Audouin and Milne-Edwards
(No. 2) possessed thirty pairs of ambulatory legs, and came
from Cayenne (on the banks of the River Approuague, three
leagues from its mouth). The specimens were found “ unter
faulem, im Schlamme versunkenem Holze, an den Ufern des
Approuage im Brackwasser.”

The description is very imperfect, as may be judged from
the fact that the generative aperture is not even mentioned.

The species was regarded by the authors as identical with
Guilding’s P. juliformis, but subsequently Blanchard (No.8)
gave it the name of P. Edwardsii. I propose to retain the
latter name and to regard it as belonging to the same species
which I have fully described above from Caracas.

It must, however, be remembered that the characters of P.
Edwardsii, as given in this monograph (p. 467) are based on
the Caracas specimens; and it may quite well happen that the

MONOGRAPH OF THE GENUS PERIPATUS. 479

Peripatus found at Cayenne, when better known, will turn out
to be a distinct species.

3. Wiegmann (No. 4) obtained a specimen of Peripatus
from near the Valentia Lake in Columbia, with thirty pairs
of legs. It is quite impossible to say whether this is a distinct
species or not. It possesses, according to Wiegmann’s descrip-
tion, four spinous pads on its legs and a generative opening
between the legs of the penultimate pair.

4. C. Mority (No. 5) obtained a large number of Peripatus
from the Island of St. Thomas. He gives no details.

There is a specimen in the British Museum from St. Thomas.
It has twenty-eight pairs of ambulatory legs, and is of a
yellowish-brown colour, but is unfortunately too ill-preserved
for determining any specific characters.

5. Peters (No. 24) mentions specimens from Utuado, Porto
Rico, and gives the following particulars.

Specimens 21 mm. in length had 27 pairs of legs.

33 33 mm. bed 33 30 bP) 3
23 3 8 mm. 33 ” 3 29 >
” 42 —4.8 mm, ” ”? 32 ” »

6. Blanchard (No. 8) has described a Peripatus found in
Chili by M. Claude Gay, with nineteen pairs of legs. His
description is as follows :—‘‘ Le corps est long de 30 a 32
mill., et large de 5 a 6, légérement atténué aux deux extrémités,
mais surtout vers la partie postérieure. Sa couleur est noire,
un peu variée irréguligrement de taches roussatres. La téte
est presque carrée avec les antennes amincies vers le bout,
présentant des annulations trés serrées. L/orifice buccal est
ovalaire. Les pattes sont au nombre de dix-neuf paires, ciliées
de poils raides comme de petites pointes, et terminées par des
crochets.”

There is obviously nothing in this description which enables
us to say whether the three specimens at the author’s disposal
possessed the characters of the Neotropical species or not.
It is extremely probable, considering the remoteness of the
locality, that this is a distinct species; but unfortunately

VOL, XXVIII, PART 4,—NEW SER. K K

480 ADAM SEDGWICK.

Blanchard has not, with the exception of a name, assigned
to it any feature which can be in the least degree regarded
as specifically distinctive. He calls it P. Blainvillei, and says
that it has nineteen pairs of legs. The name I propose to
discard, and the statement of fact I am inclined to doubt, for
this reason:—In Gay’s ‘ Historia de Chile,’ vol. iii, “ Zoologia,”
p. 58, there is a description of this proposed new species, and the
possession of nineteen pairs of legs is given as a character. I
presume Blanchard is responsible for this statement as it coin-
cides with that given in No. 8. In the description reference
is made to some figures in the Atlas. These turn out to bea
dorsal, ventral, and side view, &c., of the specimen described.
Will it be believed that not only does each of these figures
show a different number of legs, but in the case of the dorsal
and ventral views, the numbers on the right and left sides are
different ? The details are as follows:
Dorsal view, right side, 27 legs and oral papilla.
ae 5» wit. 5 26
Ventral ,, right ,, 33
»” pa RAED Bia Pose thre
Side view of left side, 20

I do not know who is responsible for these figures. The
draftsman’s name on the plate is Spinola. I need hardly say
that, if they are a fair sample of the drawings in the Atlas, the
zoological plates are not worth the paper they are printed on.

It will, perhaps, be convenient to denote the fact that there
is a Peripatus in Chili, by introducing for it the provisional
name of Peripatus chiliensis.

7. Blanchard refers to a Peripatus found in Cuba by M.
Macleay. He regards it as belonging to the species juliformis.
I have been unable to find any account of this Cuban species.

8. Grube (No. 11) describes three specimens of Peripatus
from near Colonia Towar, in Venezuela, and referred them to
P. Edwardsii. One of the specimens possessed twenty-nine
pairs of legs and the other two thirty each. But one cannot
regard his statements on this head as being trustworthy, in-

MONOGRAPH OF THE GENUS PERIPATUS. 481

asmuch as the specimen he has figured has thirty-one pairs
(in addition to the oral papille).

He found a number of embryos in the uteruses of his
specimens, all of which, excepting one with thirty, possessed
thirty-one pairs of legs.

His statement on the colour is interesting, as tending to
show that the pigment in this species is affected by the pro-
longed action of spirit. He says:

“Die Farbung war an einem sehr frisch erhaltenen Wein-
geist exemplar ein dunkles unreines Kirschroth, der Weingeist,
indem es lag, hatte sich blassroth gefarbt, bei denen die
lingere Zeit aufbewahrt waren, ging der Ton in’s Braunlichgrue
tiber, doch blieb die Ruckenseite immer sehr viel dunkler als
die Bauchseite, auch zeigte sie bestandig die schon von den
friiheren Beschreibern erwahnte mittlere Langsfurche von
noch dunklerer Farbe, rechts und links von ihr in einiger Ent-
fernung sieht man gewohnlich noch eine dunkle Seitenlinie.”

I cannot be quite certain from Grube’s figures whether the
papille have the form characteristic of the Caracas species or
of the Demerara form.

The species possesses all the Neotropical characters, viz. inner
blade of jaw with minor teeth separated by diastema from the
first small tooth, legs with four spinous pads, generative open-
ing between the penultimate pair of legs, oviducts with recep-
taculum and process, embryos in uterus of very various ages.

I propose, therefore, to retain provisionally Grube’s name
for the specimens from Colonia Towar, and to regard them as
belonging to the species found at Caracas, and described above
as P. Edwardsii.

9. Mr. Thomas Belt found a specimen of Peripatus at Santo
Domingo, in Nicaragua. The specimen (dried) is referred to
in his work, ‘The Naturalist in Nicaragua’ (p. 140), and has
been examined by Professor Moseley, who found that it
possessed thirty-one pairs of ambulatory legs.

482 ADAM SEDGWICK.

P. quitensis (Schmarda).

10. Professor Jeffrey Bell has recently (No. 43) drawn
attention to a reference by Schmarda in his ‘ Zoology’ (No. 42)
to a species with thirty-six pairs of legs from Quito, in Ecuador.
Schmarda gives a figure of the specimen, which came from an
elevation of 9000 feet.

Neotropical Peripatus in the British Museum.

1. Specimen from Dominica found by Mr. G. F. Angas.
This specimen is in excellent condition, and has twenty-nine
pairs of ambulatory legs. It has been shortly described by
Professor Jeffrey Bell (No. 29), who says that it has thirty
pairs of legs. This may be so, but I could not make out more
than twenty-nine. The dorsal surface is brown, and there is
a dark streak (chocolate-coloured with a dash of purple) run-
ning along the sides of the body just dorsal to the legs. The
legs are without tubercles. The pedal grooves are widely open.
The papillz are, I think, conical in form; but the light was
not good enough to enable me to obtain certainty on this point.

2. A specimen marked P. Blainvillei, without locality. This has
thirty-three pairs of ambulatory legs, and is of a reddish-brown
colour. It is very much contracted. There were no tubercles,
and I was not able to make out the shape of the papille.

3. A specimen marked P. Blainvillei, without locality (see
above, p. 479), with twenty-eight pairs of ambulatory legs.

4. There are three specimens in a bottle labelled “ From
Jamaica,” collected by Gosse.

They are all a yellowish brown. Two of them have thirty-
one and one thirty-seven pairs of ambulatory legs. The
latter is remarkable as being the smallest of the three, mea-
suring in the contracted condition 22 mm. Of the two with
thirty-one pairs of legs, the largest measured about 48 mm.,
and the other about 22 mm. The papille were conical and
there were no tubercles. Mr. Gosse, in ‘A Naturalist’s
Sojourn in Jamaica’ (P. H. Gosse, London, Longmans, 1851,

MONOGRAPH OF THE GENUS PERIPATUS. 483

p. 62), refers to these Peripatus in the following terms: “ Peri-
patus found at Bluefields mountain above Bluefields House,
near the town of Savanna lo Mar. The mountain height is
four or five miles from Bluefields. Here, around a piece of
burnt ground just reclaimed from the forest, but not yet
planted, were found, under stones, five or six specimens of
Peripatus, one twice as large as any of the others. The piece
of ground lay at the foot of a conical peak of considerable
elevation, but not the very loftiest, covered with original forest.
It is a curious creature, and I] think rather allied to the
Annelida than the Mollusca. It is of a velvety appearance, of
a blackish-brown hue, the tentacles tipped with white. From
these latter organs there exudes, when the animal is touched, a
thick glutinous substance, as adherent as birdlime.” He con-
cludes that it is of a different species from that found by the
Rev. L. Guilding at St. Vincent.

5. A specimen labelled “ Peripatus juliformes, West Indies,
Mr. Gibson, Nereis viridis, Adams, ‘Linn. Trans.,’ feet only
thirty-one pairs.”

This specimen was about 65 mm. in length, 53 in breadth,
5 in dorso-ventral depth; i.e. it was cylindrical in form. It
possessed thirty-two pairs of ambulatory legs, and has a very
pale brown colour (almost white). Its skin is much smoother
than is generally the case.

The legs have four spinous pads, and are without tubercles ;
the generative opening is between the legs of the penultimate
pair; the integumentary papille are constricted; the legs of
the last two pairs are very small. It clearly, therefore, belongs
to a typical Neotropical species, but more than this cannot be
said.

6. A smaller specimen with thirty pairs of ambulatory legs of
very much the same colour and form. It was labelled, “ Peri-
patus juliformis, Guild., W. Indies? Sloane collection.”

It possesses thirty pairs of legs. The generative opening
is between the legs of the penultimate pair. The grooves on
the base of the legs fairly well marked. Feet not sufficiently
well preserved for study (claws broken away). The integu-

484, ADAM SEDGWICK.

mentary papillze constricted, and arranged on the dorsal
surface in regular rows. Length about 48 mm.; body cylin-
drical in shape with a diameter of about 4mm. The legs are
without tubercles.

7. Finally there is a specimen labelled “ Peripatus Santarem,
Wickham, purchased of W. H. J. Carter.” It has thirty-one
pairs of ambulatory legs, and presents, so far as its external
features are concerned, the Neotropical characters. The papillze
are conical, and the legs are without tubercles.

Professor Strunstrip was kind enough to send me for exam-
ination the specimens in his museum. I desire to take this
opportunity of thanking him for his courtesy and kindness in
the matter. The Copenhagen specimens were in four bottles :

(1) Label “ Peripatus Edwardsii, Bl., Vestindien, Kroyer.”
This was a fine, well-preserved specimen, with thirty-one pairs
of ambulatory legs, and a brown colour. The dorsal surface was
darker than the ventral. The dorsal papille were remarkably
large (fig. 14) and constricted, as were also the ventral, but less
markedly. The generative opening was between the legs of
the penultimate pair, and the spinous pads of the legs were
four in number.

(2) Label “ Peripatus Edwardsii, Blanch., St. Croix, Kroyer.”
This specimen, which in general appearance resembled the first,
but was smaller, possessed twenty-seven pairs of ambulatory
legs. Spinous pads and generative opening as in (1).

(3) Label “ Peripatus Vestindien, Hombek (?).” With thirty-
two pairs of ambulatory legs.

(4) The specimen in the fourth bottle was not sufficiently
well preserved for observation.

It is unfortunate that the exact localities of the above were
not recorded. They are obviously all Neotropical species, but
to which of these they belong cannot be at present settled.

Four specimens of Peripatus, of which one had thirty-one
pairs of legs, are reported from Demerara (Hoorubea Creek,
twenty miles from Georgetown, on east side of Demerara river),
by Mr. J. J. Quelch (No. 86). No details are given.

MONOGRAPH OF THE GENUS PERIPATUS. 485

A single specimen was found at Breves, on the Island of
Marajo, at the mouth of the Amazon, by Mr. J. C. Branner
(No. 37). No details are given.

PERIPATUS SUMATRANUS.

A single specimen of Peripatus, stated to have come from
Sumatra, has recently been described by Dr. R. Horst (No. 38).
The evidence that the specimen was actually found in Sumatra
is not, however, conclusive. Dr. Horst states that it was found
in a bottle containing insects from East Sumatra. The name
of the finder is not given, and there is no evidence to show
Low the specimen got into the bottle. Considering the fact
that this is the only specimen of Peripatus ever reported from
the Oriental region, it will be prudent to suspend our judg-
ment as to the authenticity of the locality given by Dr. Horst.
The specimen has twenty-four pairs of ambulatory legs, and is
25 mm. in length. The papillae are constructed as in the Neo-
tropical species, and are apparently on the cylindrical type.
Dr. Horst describes them as ‘‘ appearing to consist of a trun-
cated cone, bearing on its top a small cylinder provided with a
spine.” The legs have four pads, the generative opening is
between the legs of the penultimate pair. All these are Neo-
tropical characters. The anus is not quite terminal. Colour
is dark blackish brown ; the ventral surface is paler, greyish,
“Some small white spots scattered on the dorsal surface, but
they seem only to be produced by the loosening of the cuticle
from the top of the papille.” The foot carries only two papille,
one on the anterior and one on the posterior face. This is
unique so far as my experience of Peripatus goes. The pedal
groove is absent from the two posterior legs asin P. Edwardsii
The antenne have forty-seven rings.

I think that there can be no doubt that this is a distinct
species. It is the only specimen hitherto recorded from the
oriental region, and it seems a fact of extreme interest that it
should resemble the Neotropical species more than any other.
It is a great misfortune that Dr. Horst was not able to examine
the jaws and generative organs.

486 ADAM SEDGWICK.

Peripatus sumatranus (Horst).

Peripatus from Sumatra, with twenty-four pairs of ambulatory
legs, four pads on the legs, and constricted papille. The
generative opening is between the legs of the penultimate pair.
The feet have only two papille.

SYNOPSIS OF THE SPECIES OF PERIPATUS.

South African Species.

With three spinous pads on the legs and two primary papilla on the anterior side
of the foot, and one accessory tooth on the outer blade of the jaw; witha
white papilla on the ventral surface of the last fully developed leg of the
male. Genital opening subterminal, behind the last pair of fully-developed
legs. The terminal unpaired portion of vas deferens short. (Ova of con-
siderable size, but with only a small quantity of food-yolk.

P. capensis (Grube).—South African Peripatus, with seventeen pairs of claw-
bearing ambulatory legs. Locality, Table Mountain.

P. Balfouri (Sedgwick).—South African Peripatus, with eighteen pairs of
claw-bearing ambulatory legs, of which the last pair is rudimentary. With
white papillae on the dorsal surface. Locality, Table Mountain.

P. brevis (De Blainville).—South African Peripatus, with fourteen pairs of
ambulatory legs. Locality, Table Mountain. (I have not seen this
species. Presumably it has the South African characters.)

P. Moseleyi (Wood Mason).—South African Peripatus, with twenty-one and
twenty-two pairs of claw-bearing ambulatory legs. Locality, near Williams-
town, Cape Colony.

Dovstrut Srectgs.
(1) South African Peripatus, with twenty pairs of claw-bearing ambulatory legs
(Sedgwick). Locality, Table Mountain. (Also Peters, locality not stated.)

(2) South African Peripatus, with nineteen pairs of ambulatory legs (Trimen).
Locality, Plettenberg Bay, Cape Colony. (Also Peters, locality not stated.)

Australasian Species.

With fifteen pairs of claw-bearing ambulatory legs, with three spinous pads on
the legs, and a primary papilla projecting from the median dorsal portion
of the feet. Genital opening hetween the legs of the last pair. Receptacula

MONOGRAPH OF THE GENUS PERIPATUS. 487

seminis present. Unpatred portion of vas deferens long and complicated.
Ova large and heavily charged with yolk.

P. Novee-zealandize (Hutton).— Australasian Peripatus, without an accessory
tooth on the outer blade of the jaw, and without a white papilla on the base
of the last leg of the male. New Zealand.

P. Leuckarti (Saenger).— Australasian Peripatus, with an accessory tooth on
the outer blade of the jaw, and a white papilla on the base of the last leg
of the male. Queensland.

Neotropical Species.

With four spinous pads on the legs, and the generative aperture between the legs
of the penultimate pair. Dorsal white line absent. Primary papille
divided into two portions. Inner blade of jaw with gap between the first
minor tooth and the rest. Oviducts provided with receptacula ovorum and
seminis. Unpaired part of vas deferens very long and complicated. Ova
minute, without food-yolk. (Legs not constant in number in the same species.)

P. Edwardsii.'"—Neotropical Peripatus from Caracas, with a variable
number of ambulatory legs (twenty-nine to thirty-four). Males with twenty-
nine or thirty legs, and tubercles on a varying number of the posterior legs.
The basal part of the primary papilla is cylindrical.

P. Trinidadensis (n. sp.).—eotropical Peripatus from Trinidad, with
twenty-eight to thirty-one pairs of ambulatory legs, and a large number of
teeth on the inner blade of the jaw. The basal portion of the primary
papille is conical.

P. torquatus (Kennel).—WNeotropical Peripatus from Trinidad, with forty-one
to forty-two pairs of ambulatory legs. With a transversely placed bright
yellow band on the dorsal surface behind the head.

Dovustrut SPECIES.

The above are probably distinct species. Of the remainder we do not know
enough to say whether they are distinct species or not. The following is
a list of these doubtful species, with localities and principal characters.

P. juliformis (Guilding).—WNeotropical Peripatus from St. Vincent, with
thirty-three pairs of ambulatory legs.

P. Chiliensis (Gay).—Weotropical Peripatus from Chili, with nineteen pairs
of ambulatory legs.

P. demeraranus (Sclater).1—WNeotropical Peripatus from Maccasseema,

1 This name was first applied by Blanchard (No. 8) to a species from
Cayenne (vide above, p. 478). The description, however, is very imperfect,
and it is by no means clear that the Cayenne species is identical with the
species here named Edwardsii.

488 ADAM SEDGWICK.

Demerara, with twenty-seven to thirty-one pairs of ambulatory legs and
cylindrical primary papille.

Peripatus from Cayenne (Audouin and Milne-Edwards).—With thirty
pairs of legs. Named P. Edwardsii by Blanchard. Doubtful species.

Peripatus from Valentia Lake, Columbia (Wiegmann).— With thirty
pairs of legs. Doubtful species.

Peripatus from St. Thomas (Moritz).—No description. Doubtful species.

Peripatus from Colonia Towar, Venezuela (Grube).— With twenty-
nine to thirty-one pairs of ambulatory legs. Named P. Edwardsii by
Grube. Doubtful species.

Peripatus from Santo Domingo, Nicaragua (Belt).—With thirty-one
pairs of anbulatory legs. Doubtful species.

Peripatus from Dominica (Angas).—WNeotropical Peripatus, with twenty-
nine pairs of ambulatory legs. Doubtful species.

Peripatus from Jamaica (Gosse).—With thirty-one and thirty-seven pairs
of ambulatory legs. Species doubtful.

Peripatus from Santarem.—WNeotrepical Peripatus, with thirty-one pairs
of ambulatory legs.

Peripatus from Cuba.—No details.

Peripatus from Hoorubea Creek, Demerara (Quelch).—With thirty
pairs of legs.

Peripatus from Marajo (Branner).—No details.

Peripatus from Utuado, Porto Rico (Peters).—With twenty-sevens
thirty, thirty-one, and thirty-two pairs of legs.

Peripatus from Surinam (Peters).—No details.

Peripatus from Puerto Cabello, Venezuela (Peters) —With thirty and
thirty-two pairs of legs.

Peripatus from Laguayra, Venezuela (Peters).—No details.

Peripatus Quitensis (Schmarda).—From Quito, with thirty-six pairs of
legs.

Peripatus from Sumatra.

P. Sumatranus (Horst).—Peripatus from Sumatra, with twenty-four pairs of
ambulatory legs, and four spinous pads on the legs. The primary papille
of the neotropical character with conical bases. Generative opening between
the leys of the penultimate pair. Feet with only two papillae.

SUMMARY OF DISTRIBUTION.

DistTRIBUTION OF THE SourH Arrican SPECIES—
Slopes of Table Mountain, neighbourhood of Williamstown, Pletten-
berg Bay—Cape Colony.

MONOGRAPH OF THE GENUS PERIPATUS. 489

DIstRIBUTION OF THE AUSTRALASIAN SPECIES—

Queensland—Australia.
North and South Islands—New Zealand.

OxrentaL Recion—
Sumatra.

‘DisTRIBUTION OF THE NEOTROPICAL SPECIES—

Nicaragua.

Valencia Lake, Caracas, Puerto Cabello, Laguayra, Colonia Towar—
Venezuela.

Quito—Ecuador.

Maccasseema, Hoorubea Creek—Demerara.

Surinam (Peters).

Cayenne.

Santarem, Marajo at the mouth of the Amazon—Brazil.

Chili.

and in the following West Indian Islands—Cuba, Dominica, Porto Rico
(Peters), Jamaica, St. Thomas, St. Vincent, Trinidad.

LITERATURE.

(1) Guinpine, L.— Mollusca caribbeaana: an Account of a New Genus
of Mollusea,” ‘ Zool. Journ.,’ vol. ii, 1826, p. 443, pl. 14; reprinted in
Isis vol. xxi, 1828, p. 158, pl. ii.

(2) Aupourn et Mitne-Epwarps.—“ Classification des Annélides, &c.,”
‘Ann. Sc. Nat.,’ series 1, vol. xxx, pp. 411—414, pl. xxii, 1833.

(3) Gervais, M.—‘‘ Etudes pour servir 4 l’histoire naturelle des Myria-
podes,” ‘ Ann. Sc. Nat.,’ series 2, vol. vii, pp. 35—60, 1837.

(4) Wieemany, A. F. A.—‘‘ Einige Bemerkungen uber Guilding’s Peripa-
tus,” ‘Arch, f. Naturgesch.? (Wiegmann), iii, pp. 195—200, 1837.

(5) Moritz, C.—‘‘ Noch einige Worte tiber Peripatus Guilding,” ‘ Arch.
f. Naturgesch.” (Wiegmann), v, pp. 175, 176, 1839.

(6) De Buarnvitte.—‘ Dictionnaire des Sciences Naturelles,’ Supplément,
tom. i, p. 237, Paris, 1840.

(7) Mitne-Epwarps, H.—“ Note sur le Péripate juliforme,’ ‘Ann. Sc.
Nat.,’ series 2, vol. xviii, pp. 126—128, 1842.

(8) Buancuarp, E.—* Recherches sur l’organisation des Vers,” ‘Ann. Sc.
Nat.,’ series 3, vol. viii, pp. 119—149, 1847.

(9) Quarreracss, A. pE.—‘ Mémoire sur la famille des Hermelliens,”
‘Ann. Se. Nat.,’ series 3, vol. x, pp. 5—58, 1848.

(10) Mitnz-Epwarps, QuaTreracEs, ET Brancuarp.— Recherches anato-

460 ADAM SEDGWICK.

miques et zoologiques fait pendant un voyage sur la cote de Sicile,’
part ili, p. 61, pl. i, fig. 2, Paris, 1849.

(11) Gruss, E.—“ Untersuchungen iiber den Bau von Peripatus Edwardsii,”
‘Miller’s Arch.,’ 1853, pp. 822—360, Taf. ix, x, 1853.

(12) Gay, C.—“ Historia fisica y politica de Chile,” ‘Zoologia,’ vol. iii, p.
58; ‘ Atlas, Annelides,’ Lam. iti, fig. 2, 1854.

(18) Quarreraczs, A. DE.—‘ Histoire Naturelle des Annelés,’ tom. ii, p+
675, Paris, 1865, 8vo.

(14) Grouse, E.—‘ Reise der osterreichischen Fregatte ‘‘ Novara,”’ Band 2,
Abth. iii; ‘ Anneliden,’ Wien, 1868, p. 4, pl. iv, figs. 3, 4a.

(15) Sarncer.—“ Description of a Peripatus from Australia,” ‘Transac-
tions of the Russian Assembly of Naturalists,’ held at Moscow in
1867, Moscow, 1869 (Russian).

(16) Leucxart, R.— Bericht tiber Leist in d. Naturgeschichte der niederen
Thiere wahrend der J. 1868, 1869,” ‘Arch. f. Nat.’ (Troschel),
xxxv, pt. 2, pp. 277, 278, 1869.

(17) Bett, T.—‘ The Naturalist in Nicaragua,’ 8vo, London, 1874, p.
140.

(18) Mosrtry, H. N.—‘‘ On the Structure and Development of Peripatus
capensis,” © Phil. Trans.,’ clxiv, pls. Ixxii—lxxv, pp. 757—782; and
‘Proc. R. S.,’ xxii, pp. 344 —350, 1874.

(19) Hurron, F. W.—“On Peripatus nove-zealandiea,’ ‘Ann. Mag. Nat.
Hist.,’ 4th series, xvili, pl. xvii, pp. 361—369, 1876; also, op. cit.,
4th series, xx, pp. 81—83, 1877; and op. cit., 5th series, i, pp. 204—
206, 1878.

(20) Mosrtey, H. N.— Remarks on Observations by Captain Hutton,
Director of the Otago Museum, on Peripatus nove-zealandia, with
Notes on the Structure of the Species,’ ‘Ann. Mag. Nat. Hist.,’ 4th
series, vol. xix, pp. 85—91, 1877.

(21) Batrour, F. M.—‘‘On Certain Points in the Anatomy of Peripatus
capensis,’ ‘Proc. Cambr. Phil. Soe.,’ iii, pp. 266—269; ‘Quart.
Journ. Micr. Sci.,’ vol. xix, pp. 431—433, 1879.

(22) Mosr.ry, H. N.—“ Notes on the Species of Peripatus, and especially
on those of Cayenne and the West Indies,” ‘ Ann. Mag. Nat. Hist.,’ 5th
serles, vol. ili, pp. 263—267, 1879.

(23) Woop Mason, J.— Morphological Notes bearing on the Origin of
Insects,” ‘Trans. Ent. Soc., London,’ 1879, p. 155.

(24) Perers, W.— Ueber die Arten von Peripatus,” ‘Sitzgsber. d. Ges.
nat. Freunde zu Berlin,’ 1880, pp. 28, 29, 1880.

(25) Perrrs, W.—“ Die Variation der Fusszahl bei Peripatus capensis,”
ibid., pp. 165, 166, 1880.

(26) Ernst, A.—‘ Some Remarks on Peripatus Edwardsii Bianch.,”
‘Nature,’ xxiii, pp. 446—448, 1881.

MONOGRAPH OF THE GENUS PERIPATUS. 491

(27) Packarp, A.S.—‘‘ Noteona Peripatus from the Isthmus of Panama,”
‘American Naturalist,’ August, 1883; ‘Annual Report of Peabody
Acad. Salem, Mass.’

(28) Batrour, F. M.—‘‘ The Anatomy and Development of Peripatus capen-
sis,” edited by Professor H. N. Moseley and A. Sedgwick, ‘ Quart.
Journ. Micro. Sci.,’ xxiii, pp. 213—259, pls. xili—xx, 1883.

(29) Brut, F. J—“ Note ona Peripatus from the Island of Dominica, West
Indies,” ‘Ann. Mag. Nat. Hist.,’ 5th series, xi, p. 388, 1883.

(29a) Oaxtey, H. W.—*‘ On Peripatus capensis,’ ‘Trans. 8. African Philo-
sophical Soe.,’ vol. iii, 1881—1883, p. 35.

(30) Kennet, J.—‘ Entwickelungsgeschischte von Peripatus,” ‘Zool. An-
zeiger, 1883, p, 531.

(31) Kenne, J.—“ Biologische u. faunistiche Notizen aus Trinidad,” ‘ Ar-
beiten a. d. Zool. Inst. Wiirzburg,’ vi, p. 282.

(32) Kennet, J.—‘ Entwickelungsgesch. v. P. Hdwardsii (Blanch.} und P.
torquatus (un. sp.),” Th. i, ‘ Arbeiten a. d. Zool. Inst. Wurzburg,’ Bd.
vii, 1885.

(33) Kennet, J.—‘ Entwickelungsgesch. v. P. Hdwardsit (Blanch.),” Th. ii,
‘Arb. a. d. Zool. Inst. Wirzburg,’ Bd. viii, 1886.

(34) Garrron, E.—“ Beitrage zur Anatomie und Histologie von Peripatus,”
‘Zool. Beitrage (Schneider),’ Bd. i, p. 33, 1885.

(35) Garrron, H.—‘ Beitrage z. Anat. u. Histologie v. Peripatus,” Th. ii,
ibid., p. 145.

(86) QuzLcu, J. J.—“ Peripatus in Demerara,” ‘ Nature,’ xxxiv, p. 288,
1886.

(37) Branner, J. C—‘Peripatus in the Island of Marajo, Amazons,”
‘ Nature,’ xxxiv, p. 496, 1886.

(38) Horst, R.—‘‘On a Specimen of Peripatus, Guild. from Sumatra,”
‘Notes Leyd. Mus.,’ viii, pp. 37—41, pl. ii, figs. 1—5, 1886.

(39) Srepewick, A.—“ The Development of the Cape Species of Peripatus,”
Part I—IV, ‘ Quart. Journ. Mier. Sci.,’ vols. xxv, xxvi, XXvil, XXviii.

(40) SHEtpon, L.—‘ Notes on the Anatomy of P. capensis and P. nove-
zealandia,’ ‘ Quart. Journ. Micr. Sci.,’ vol. xxviii, p. 495.

(41) Scrater, W. L.— Notes on the Peripatus of British Guiana,” ‘ Proc.
Zoological Society, London,’ 1887, p. 180.

(42) Scumarpa.—‘ Zoologie,’ vol. ii, 2nd ed., p. 77.

(43) Bett, F. Jerrrey.— Report of the British Assoc. for the Advancement
of Science,’ 1887.

(44) Bett, F. Jerrrey.— Habitatat of P. Leuckarti,’ ‘Ann. and Mag. of
Nat. Hist.,’ series v, vol. xx, 1887, p. 252.

(45) Suetpon, L.—“On the Development of Peripatus nove-zealandia,”
‘ Quart. Journ. Mier. Sci.,’ vol. xxviii, pp. 205—237.

(46) Scrater, W. L.—< On the Development of a Peripatus from Deme-
rara,” ‘ Quart. Journ. Micr. Sci.’ vol. xxvili, pp. 343—363, pl. xxiv.

492 ADAM SEDGWICK.

EXPLANATION OF PLATES XXXIV, XXXV, XXXVI,
XXXVII, XXXVIII, XXXIX, & XL,

Illustrating Mr. Adam Sedgwick’s Monograph on _ the
Species and Distribution of the Genus Peripatus
(Guilding).

‘PLATE XXXIV.

Peripatus capensis, drawn from life. Life size. From a drawing by Mr.
E. Wilson.

PLATE XXXV.

Peripatus capensis. X 4. Dorsal view of a spirit specimen. From a
drawing by Miss Balfour. This figure was originally published in vol. xxiii of
this Journal.

PLATE XXXVI.

The figures on this Plate originally appeared in vol. xxiii of this Journal.
They are from drawings by Miss Baifour.

Fic. 2.—A leg of P. capensis, ventral view. X 30.

Fie. 3.—A right leg of P. capensis, viewed from the front side.

Fic. 4.—The last leg of a male specimen of P. capensis, ventral view to
show the papilla, at the apex of which the accessory gland of the male, or
enlarged crural gland, opens to the exterior.

Fic. 5.—Ventral view of head and oral region of P. capensis.

PLATE XXXVII.

Fic. 6.—P. Edwardsit. x 4. Dorsal view of a specimen with thirty-
three legs. From a drawing by Miss Balfour.

Fic. 7.—P. nove-zealandie. X 4. Dorsal view. From a drawing by
Miss Balfour.

Fic. 8.—P. Moseleyi. x 4. Dorsal view. From a drawing by Miss
Balfour.

Fic. 9.—Fourth leg of P. Ba/fouri. Ventral view. From a drawing by
Mr. E. Wilson.

Fre. 10,—A piece of skin from the dorsal region of P. Balfourt. Froma
drawing by Mr. E. Wilson.

MONOGRAPH OF THE GENUS PERIPATUS. 493

Fic. 11.—Terminal portion of fourth leg of P. Edwardsii. From a draw-
ing by Mr. E. Wilson. (The colour of this figure is not that of the specimen.)

PLATE XXXVIII.

All the figures on this Plate, except Fig. 16, are from drawings by Mr.
Wilson. Fig 16 is from a drawing by Mr. Hill.

Fic. 12.—Leg of Peripatus Edwardsit. Ventral view.

Fic. 13.—Head of P. Hdwardsii. Ventral view.

Fic. 14.—Portion of skin of one of the Copenhagen West Indian specimens,
labelled P. Edwardsit.

Fig. 15.—Sixth left foot of P. nove-zealandie., Front view.

Fic. 16.—Terminal part of antenna of P. zove-zealandia. X 120.

PLATE XXXIX.

All the figures on this Plate, except Figs. 21 and 21a, are from drawings
by Mr. W. H. Hill, executed under the supervision of Professor Moseley.
Figs. 21 and 21a, by Mr. Wilson.

Fie. 17.—P. nove-zealandiea. A specimen of another colour. Dorsal
view. xX 6.

Fic. 18.—Head of P. nove-zealandie. X 32. Dorsal view.

Fie. 19.—Portion of ventral surface of P. nove-zealandia. X 28.

Fie. 20.—Oral papilla of P. xove-zealandie. xX 100.

Fic. 21.—Fourth leg of P. xove-zealandie. Ventral view.

Fic. 214.—Foot of the same fore-shortened, so as to show the dorso-median
papilla.

PLATE XL.

Figs. 22, 23, 24, 25, 26, from drawings by Mr. Wilson. Figs. 27 and 28,
from drawings by Miss Balfour.

Fic. 22.— Ventral view of a posterior leg of P. Edwardsit.

Fic. 23.—Ventral view of hind end of P. xove-zealandia.

Fie. 24.—Ventral view of hind end of P. Balfouri.

Fies. 25 and 26.—Jaw blades of P. Hdwardsit.

Fig. 25. Inner blade. The diastema is rather too marked.
Fig. 26. Outer blade.
Fics. 27 and 28.—Jaw blades of P. capensis.
Fig. 27. Inner blade.
Fig. 28. Outer blade.

Fic. 29.—Skin of ventral portion of leg (close to the pads) of P. Trinidad-
ensis, copied from Gaffron. ‘he figure shows the form of the simple papille,
and two of the composite papille of the row next the proximal pad.

Fic. 30.—Skin of leg of P. nove-zealandie. From a drawing by Mr. Hill-

1 uh Me Oi | ajley ae cy », ag @

wig il

‘ - 7. a y
y fe 1, yh 10 Titec ite ous 7

Va

wan. Vol SIMMS POLEY

/

ie SE: tt

WOT UM (|

J

- oe =. ee aS. a —- Hi a “| ¥ . a 2 C i — |e oe oe — Se

Fig.7

‘

2 OE

—
= =
i
i re
a4
—

e
i a
i
.
—_—
g
re >
-
7 ae
i
o
.
.

lh leg of P Balfouri .

Fig )

Fig

4” leg of P Edwardsii.
i

x c biece 2 es ied KW. fo Prep |
j a

P Novae Zealandiae x 4,

Lith & Imp Camb. Sci Inst.Co

. as

yy

*y Wiy

C)
00
29,

Soo

©)

Oyo06
oo,”

Mer Journ, Vol UNS PLEO Nh

rH,
hs =
»
hy, K
ee
5
y 3
wD, ‘
YS ee
«" ~
\
=
==

ancy 239
pretty 4

* ei Niro
oe
ay gay

ay

iv

--

Se
i
< 7
a
a) c
,
i
7 k - 4
- F ‘
; ; > t S
* a
us “
4 ¥ »
: ae
t
.
| a
=
ahs
i
>
P
¢
:
: , —
4 > é
+ :
r 7
=
J
= v
: A ~ . +,
ii . a ae
ay 4 -

LALA

}
f

PY,

We

7

Y.

Vo LIU

Ww
Cle Sour

dM

at

‘

re

Zz

ry awe

P ig
‘ a
a ° -
ae
= a
5 ~ J’
ae :
-
“4
od
j LIL ee ee
bigs
s i
} . :
&
{ -
P Edwardsii Ventral view of a posterior leg.
‘4
a
at +o.
19)
29
ee
oo

—-generative opening. a 7
Ss

: l view of hind end =
; is ee to 4 = aoe Ventral view of hind end of P. Balfourt.
, * : |
> s ? ; “—

Figab
Outer blade of jaw of P. Edwardsii.

Le
me
t
a
| 6 al
be Inner blade of jaw of P.Capensis

Fig 23 a‘

Ga gal <0 oye VD) Of of,
Sou. Col, AAMAS It Al,

aw of P.Capensis.

Skin of P Novae-Zealandiae.

‘
Lith & Imp.Camb. Sci. Inst. Co.

NOTES ON THE ANATOMY OF PERIPATUS. 495

Notes on the Anatomy of Peripatus capensis
and Peripatus Nove-Zealandie.

By

Lilian Sheldon,
Bathurst Student, Newnham College, Cambridge.

THERE are a few anatomical details in which these two
species of Peripatus differ both from one another and from P.
Edwardsii, the anatomy of which has been described by
Gaffron (2).

CRURAL GLANDS.

I have examined several legs of P. capensis, both of males
and females, and have found a crural gland in every one,
except the first pair of legs.

Except those of the fourth and fifth, and in the male of the
last pair of legs, the glands all resemble one another in their
structure, in the position of the gland and duct, and in the
point at which the latter opens, viz. on the ventral surface of
the leg external to the opening of the duct of the segmental
organ.

The glands of the fourth and fifth pairs of legs are very
much smaller than the others, and open internally to the seg-
mental organs, in the angle formed by the junction of the
leg with the body. The gland of the last leg of the male is
very long, and extends forwards through many segments of
the body, being apparently modified in connection with the
male generative organs. .

These organs have all been previously described by Professor
Balfour (1).

VOL. XXVIII, PART 4.—NEW SER, LL

496 LILIAN SHELDON.

In none of the legs of Peripatus nove zealandiz, of
which I examined sections, did I find any crural glands, either
in the male or female. I took legs from the various parts of
the body in both sexes, and in all cases the segmental organs
were present, and the legs contained a much larger supply
of muscles than those of P. capensis, but there were no
traces of crural glands in any case, even in the last leg of the
male.

In P. Edwardsii Gaffron (2) states that the crural glands
are absent in the female, but are present in some of the
segments of the male, there bemg in some of them two
pairs,

SEGMENTAL ORGANS.

In Peripatus nove zealandiz the external aperture
of the generative apparatus is placed on the ventral surface of
the body in front of the last (fifteenth) pair of legs. In this
pair of legs there are no segmental organs, so that the genera-
tive ducts are apparently the modified segmental organs of the
last segment.

In Peripatus capensis, in which the generative aperture
is situated at the posterior end of the body, immediately in
front of the anus and behind the last pair of legs, segmental
organs are present in the latter.

In P. nova zealandiz, as was described in P. capensis
by Professor Balfour (1), the segmental organs of the fourth
and fifth pair of legs are much larger than the rest.

Accressory GLANDS OF THE MALE.

In P. capensis, in addition to the enlarged crural glands
of the last pair of legs, the male generative apparatus is
provided with a pair of glandular tubes, which lie on each
side, and dorsally to the ductus ejaculatorius, into which they
open at the point where it opens to the exterior. These
accessory glands were first mentioned by Professor Moseley (4),

NOTES ON THE ANATOMY OF PERIPATUS. 497

and their position and mode of opening were described by
Professor Balfour (1).

In P. nove zealandiz these accessory glandular tubes are
also present, but their relations are somewhat different. They
lie more laterally in the body. Each gland starts as a blindly-
ending tube near the posterior end of the body, and passes
forwards for a short distance. It then bends sharply on
itself, and passes backwards to its external opening, which
is situated near the posterior end of the body at its ventro-
lateral angle, external to the nerve-cord. This position of
their openings is clearly seen, both on dissection and in sec-
tions. Thus they open quite independently of the vas deferens,
therein differing from those of P. capensis. They also differ
from the accessory glands of P. Edwardsii, which Gaffron (2)
describes as opening with the anus.

Vas DEFERENS.

To the naked eye the main difference between the vas
deferens of P. capensis and that of P. nove zealandiz ap-
pears to be in their relative lengths, that of the former being
much the shorter. This difference seems to be due to the
very great difference between the spermatophores formed by
the two species.

The vas deferens of P. nove zealandiz very closely re-
sembles that described by Gaffron in P. Edwardsii. The
whole of the posterior part of the duct is filled by an enor-
mously long spermatophore, which is surrounded by a horny
case. The internal cavity is enlarged in several places and in
these regions is filled with spermatozoa, the external case
being thin and composed of a single layer. In the remainder
of the spermatophore the lumen is very small, and the case
very thick and composed of several layers of horny substance.
The whole spermatophore has, in fact, precisely the same
structure as that described by Gaffron’ in P. Edwardsii.

The structure of the walls of the vas deferens of P. nove
zealandiz is also similar to that of P. Edwardsii, except

498 LILIAN SHELDON.

that the secretion globules are arranged in irregular masses in
the cells near their ends, which abut upon the lumen, instead
of having the regular arrangement described as occurring in
the latter. The structure of the walls of the testes in P.
nove zealandiz differs from that P. Edwardsii, in that
in the latter Gaffron (2) states that there is no epithelial lining,
whereas those of the former are lined by a layer of fairly
deep columnar cells with large nuclei, their ends, which abut
upon the lumen, being rounded. There is a layer of muscles
external to the epithelium.

In P. capensis the spermatophores are small rounded
bodies, enclosed in a thin, structureless case, and filled with
spermatozoa.

In its lower part, where the vas deferens is filled with
spermatophores, the cells lining it are somewhat flat, are much
vacuolated, and stain very slightly. Just in front of the
region where the fully formed spermatophores lie the cells are
continued into small masses of unstained matter containing
very deeply staining globules lying in the lumen of the duct.
These masses probably form the cases of the spermatophores,
and are secreted by the cells with which they are continuous.
In front of this region the cells become columnar, the nuclei
being closely packed at the bases of the cells. The cell pro-
toplasm also stains very deeply. The spermatophores of P.
capensis, after they have been shed, are easily seen by
examining with a lens the dorsal surface of a female. They
appear as small, round, whitish bodies, lying on the skin of
the animal, and when teased are found to be filled with
spermatozoa.

I have not had the good fortune to see a spermatophore of
P. nove zealandiz after it has been shed, and I only
assume that the horny case found in the vas deferens is a
spermatophore, because Professor Moseley (4) and Gaffron (2)
both describe it as such. But it is strange that it should differ
so largely from P. capensis in this respect.

The ovarian funnel (receptaculum ovorum of Kennel (3) ),
described by Gaffron (2) in P. Edwardsii as lying on the ovi-

NOTES ON THE ANATOMY OF PERIPATUS. 499

duct, between the ovary and the receptaculum seminis, is not
present in P. nove zealandiz. The receptaculum seminis,
with its two ducts, is present, and the oviduct in that region
is much coiled and vesiculated.

These observations were made at the suggestion of Mr,
Sedgwick, to whom I am indebted for providing me with the
necessary material.

PaPeRS REFERRED TO.
(1) Batrour, F. M.—“‘The Anatomy and Development of Peripatus
capensis,” ‘ Quart. Journ. Mier. Sci.,’ vol. xxiii.
(2) Garrron, E.—“ Beitrage zur Anatomie und Histologie von Peripatus,”
Theil ii, ‘Schneider’s Beitrage,’ Band i.
(3) Kennet, J. v.— Entwicklungsgeschichte von Peripatus Edwardsii,
&c.,” ‘Arb. a. d. Zool.-Zoot. Inst. Wurzburg,’ Band vii.

(4) Mosrtey, H. N.—“On the Structure and Development of Peripatus
capensis,” ‘Phil. Trans. of Royal Society,’ 1874.

Wi
423
‘
i
’
j
‘
*
tele

whey i) Ove eve ah) aie
comet

. Mix. yi Lena iL, eg iat a de
o> “¢ -

rath heh af Vane Gt aa

ey YASAGTA GLEN BRS

: ti ini i ayer Fy ct See

‘
.

idoe@s? Lot bebe) eee

'

: ani Le 7 NSB Me ba lo SEAL «"
. vol betdahub oy Lads opel

Asia gal VS

7 7 AS.

19

in
PAHAWTAL. Sa TK f =

— an

es Z s ip O® i "49054 ee

a By ). cm co
vetart A Bs Sel ag _ See rt? 4
ay

i ‘ ae
' ) ea hs oes
3 A ( Sul cer 4 ‘
> f | be: is q
*
, 5 Ao ae
+
- be
=
— oa ih
: ‘
- | 7
d
: as ‘
tes H

LABYRINTHINE APPARATUS OF LABYRINTHIO FISHES. 501

On the Construction and Purpose of the so-called
Labyrinthine Apparatus of the Labyrinthic
Fishes.

By

Doctor Nicholas Zograff,
of Moscow.

With Plate XLI.

In the year 1797 two Dutch sailors, Daldorf and John, sent
information to the Linnean Society of London as to the re-
markable capacity of Anabas scandens to crawl from one
pool of water into another, to climb up bushes and trees, to
spend several days without water, &c., before the beginning of
a dry season.! Since then almost every year one or the other
journal or paper has given various travellers’ accounts of the
interesting life and habits of these fishes. Notwithstanding
the variety of these accounts they all agree in one respect,
viz. in trying to show that the Labyrinthici are remarkable
for their highly-developed capacity of accommodation, and for
their striking viability. Cuvier and Valenciennes, for example,
in their natural history of fish, tell us how the Anabus can
spend hot seasons in the slime of a dry water-basin (very
much like the Protopteri), or how in the markets of the Hast
Indies the Ophiocephali will keep on moving a long time
after their insides have been cleaned out by the fishmongers.
Besides that, we learn from the accounts of travellers that

1 «Transactions of the Linnean Society of London,’ vol. iii, 1797.

502 NICHOLAS ZOGRAFF.

these fishes develope more mental activity than any other
species of the same class ; such instances we see, for example,
in their mode of nest building (Macropodus, Trichogaster),
their changes of basins (Anabas, Ophiocephalus), and, lastly,
in their way of getting food (Toxotes). Whilst communicating
their observations on the life of Labyrinthici, travellers have
not found it necessary to acquaint us with the anatomy of these
fishes.

I have not found in existing literature any other reports on
the inner construction (anatomy) of the Labyrinthici than those
published in the important works of Cuvier and Valenciennes,’
and in the work of Wilhelm Peters,” describing the gill-apparatus
of some of the Labyrinthici. Both these authors took interest
in that striking apparatus, which, thanks to its complicated
exterior, received the name of “ the labyrinthine apparatus.”’

It seems Cuvier was the first and probably the only author
who examined the labyrinthine apparatus, Peters having been
more interested in the relation between that apparatus and
the skeleton of the clavicles. Cuvier thought this apparatus
to be a complex of very thin bone lamelle or plates, which
served, sponge-like, to retain water for the purpose of moisten-
ing the gills of the fish when in the open air. These bone
lamelle, according to Cuvier, are nothing but projecting parts
of the pharyngeal bones; their surface is covered with a
great quantity of thin sanguiferous vessels, which appa-
rently receive the blood from the general gill-artery. How-
ever, Cuvier says he is not quite convinced of that, having
only had for his experiments samples kept in spirit, thus
making it difficult to trace such thin vessels as the artery of
the labyrinthine apparatus. Cuvier’s views as to the construc-
tion and physiological importance of the apparatus have found
credit in science, and there is not a good manual that does
not describe it in the same manner as Cuvier did. As to

1 «Histoire naturelle des poissons par Cuvier et Valenciennes,’ Paris, 1831,
vol. vii, p. 328.

? Wilh. Peters, “‘ Ueber das Kiemengeriist der Labyrinthfische,” ‘ Miiller’s
Archiv fiir Anatomie und Physiologie,’ 1853.

LABYRINTHINE APPARATUS OF LABYRINTHIO FISHES. 503

the work of Peters, this author tries to prove that the labyrin-
thine bones correspond not to the upper pharyngeal bones,
which are merely the upper fourth segments of the hinder-
most branchial arches, but are strongly developed third seg-
ments of those branchial arches. Peters does not say a word
about the construction of the soft parts of the labyrinths.
Besides these two works I have not met with any other
account of this interesting apparatus.

I have been able to make a detailed analysis of a labyrin-
thine apparatus of a Macropodus (Polyacanthus venustus,
Cuv.). For my experiments I had little fishes brought up in
my aquarium from the roe ejected by a female given to me by
our indefatigable practical zoclogist, A. S. Meschersky. Be-
sides that (thanks to the kind permission of Professor Bogden-
off, the Director), I studied aspecimen of Anabas scandens,
var. macrocephalus and Osphromenus olfax, from the
collections of the Moscow Zoological Museum.

All these fishes had the labyrinthine apparatus set inside
the gill-operculum, and surrounded on all sides by a capsule, or
walled with a thin membranous net. Its construction is the
same as the lining of the cavity of the mouth; in it can be
distinguished the epithelium of the surface, the connective
tissue cutis, and scattered here and there pigmented cells.
From the above we have the right to draw the conclusion that
the sides or partitions of this capsule are only projecting
cutaneous coverings of the inner surface of the gill-operculum.
Indeed, if we examine under a moderate power the inner surface
of a well-prepared gill-operculum of a Macropodus, we shall
see on it a little narrow slit leading to the capsule of the
labyrinthine apparatus, and connecting its cavities with the
mouth and gill-cavities of the fish.

The same relation betwixt the gill-cover and the labyrin-
thine capsule exists in the other fishes which I have examined,
viz. Osphromenus olfax and Anabas scandens.

The labyrinthine apparatus rises up into the cavity of the
capsule from its inner side, which touches the exterior sides
of the branchial arches (fig. 2). Notwithstanding Cuvier’s

504 NICHOLAS ZOGRAFF.

very accurate description of the exterior view of the labyrin-
thine apparatus in several species of the Labyrinthici, his
description seems to be deficient in thorough acquaintance
with the other sides of the apparatus, 1. e. relative to dorsal,
pectoral, front and back surfaces of the fish, the study of which
is very important, as we shall see later, for the purpose of
forming a just idea of the designation or purpose of that organ.

I shall begin with the apparatus of Anabas. Cuvier de-
scribes it as a complex of numerous very thin bone lamelle
lying over one another, and joined in the middle by a piece
by which they are fastened to the branchial arches. The space
between these lamelle, according to Cuvier, is so small as
to enable them easily to detain water.

If we take an exterior view of the labyrinthine apparatus
we can readily agree with what Cuvier has said about its con-
struction and purpose; but, taking a profile view of the same
(fig. 3), we are prone to suspect the truth of Cuvier’s explana-
tion. Indeed, the labyrinthine apparatus of Anabas does not
consist of numerous laminz, but only of three’ thin wavy
lamelle, one over the other, which become the larger the nearer
they are to the point of fastening to the branchial arches;
besides that the distance between each lamellz is much wider
than it seems to be when seen from above. Measured with a
pair of compasses these distances or intervals were found to
vary from 1:5 mm. to 2°75 mm. in breadth. We leave the
reader to judge if an apparatus consisting of three lamelle,
of which the uppermost is 5 mm. long and 4 mm. broad,
the middle one from 7 to 8 mm. broad, and the undermost
11 mm. long and 9 mm. wide, can detain much water. There
are three spaces to detain water: one between the first and
second lamelle with a surface of about 15 square mm. and
about 1:5 deep; another between the second and third lamelle
with a surface of about 60 square mm. and 2°5 mm. deep; a
third one between the hard lamella and the sides of the pouch
of the labyrinthical apparatus with a surface of about 80
square mm., and varying in depth according to wrinkles or

After ‘Giinther’s Manual of Ichthyology,’ 3—5.

LABYRINTHINE APPARATUS OF LABYRINTHIC FISHES. 505

furrows formed here by the sides of the pouch. Experiments
show that water is not detained in sufficient quantities between
small plates with such a distance between them. I have
tried to fasten together covering glasses in such a manner as
to keep their sides parallel, with a distance from 1 to 3 mm.
between each, and have found that with a surface of the
glasses equal to 324 square mm., and with a distance of 1 mm.
between each, the surface of water detained therein was equal
to about 80 square mm.; with a distance of 2 mm. the surface
of detained water was still less, and with a distance equal to
3 mm. the water was accumulated on a surface exceeding a
little 0°5 mm. just round the cork that kept the glasses
together. The quantity of water between the lamelle was
again still smaller if, instead of glasses, cartoon or British
paper was used. Consequently the apparatus of Anabas, if
applied for the purpose attributed to it by Cuvier, would in
any case function imperfectly.

The apparatus of a Macropodus is still less fit to function in
that way. It also consists of three parallel lamelle or plates
(see figure 4) which are placed in such a manner as to be
quite unable to obtain water. The (first) top plate is so bent
that its back part only remains parallel to the middle one;
the third one forms a rectangle with it; the lower part is
elongated more towards the pectoral surface of the fish,
whereas two top lamelle have grown in the opposite direction
(see drawings 4 and 1).

The labyrinthine apparatus of Osphromenus olfax is so
much like the apparatus of Anabas, that the aforesaid can be
wholly applied to it.

But what are, then, properly the functions of the labyrinthine
apparatus of the Labyrinthici ?

A study of the fine microscopic structure of this organ gives
us a right to make a supposition that will better explain its
functions than the supposition made by the great naturalist.

The projecting bone of the third segment of the last
branchial arch serves as support or basis to the labyrinthine
apparatus (see Pl. XLI, fig. 8, os.).

506 NICHOLAS ZOGRAFF.,

This bone basis is very fine and consists of a typical (partly
described by Kédlliker)! bone tissue, in which star-like bone
elements are not seen ; when the section is made in a direction
parallel with the surface of the lamella one can distinguish
in the thickness of the bone ramous canals (described by
Kostler), which make it look something like the dentine of
teeth.

The bony basis is covered by a periosteum, which consists of
very fine long and flat cells, and passes into a connecting
adipose tissue, which, on its external surface, becomes a
connective tissue (cutis) ; the cutis is covered by an epithelium ;
among the cells of this can be seen numerous goblet-like
mucous cells.

From the above one sees that the labyrinthine apparatus is
not a very complicated organ, and by its structure approaches
the general typical structure of the skin coverings of fishes ;
but, as we shall see directly, special elements of the labyrin-
thine apparatus appear to be strongly modified.

Let us begin with the layer of connective tissue which lies
next to the periosteum. This connective tissue consists of
star-like cells with very long processes, and with a small
cell-body and a little nucleus. These cellules, connected by
their outshoots, form a netted tissue (fig. 8, ctj. ad.) ; between
the meshes of this tissue are big, fat, adipose cells, that make
a fresh tissue look very odd; but generally there remain
no traces of fat in the sections, because the preliminary pre-
paration in spirit, chloroform, &c., dissolves the fat.

It is only in objects prepared in osmic acid that the fat
is retained where it is when sections are made (fig. 8); but
meanwhile an organ without fat differs very much from one
with it, as the fat in a tissue does not lie in mass, but in
small round groups (c¢. ad.). The skin covering over these
groups of fat is lifted or raised, whereas between them the
coverings are much thinner, and consequently the surface of

1 A, Kolliker, Ueber verschiedene Typen in der microsecopischen Structur
des Skelettes der Knockenfische,” ‘ Verhandlungen der _physiecalisch-
medicinischer Gesellschaft zu Wiirzburg,’ Band viii.

LABYRINTHINE APPARATUS OF LABYRINTHIC FISHES. 507

the labyrinthine apparatus under a microscope looks very
much like a counterpane or a stitched soft lining with edged
risings (fig. 10).

The meshes of the netting with fat in them get narrower
and narrower the nearer they are to the cutis, and in the cutis
the star-like connective tissue becomes fibrous. In this fibrous
tissue over the groups of adipose cells one sees numerous
capillary vessels (figs. 8 and 9, cp.), which form very pretty
and characteristic braids or interweavings. These braids
are in relation with the risings of the adipose masses or accu-
mulations ; between these accumulations capillaries are not
observable. All the above-described accumulations or gather-
ings of adipose tissue have their separate arterial and venous
branches ; these branches rise to the surface of the accu-
mulation, ramify severally, and partly pass into the capillary
network of the cutis. Thus every part of the capillary net-
work, with its separate adipose accumulation, has its special
arterial and venous branches, and is connected with the
nettings of other risings only by means of these branches.
The capillary net of every rising has the form of a very pretty
rosette (figs. 11, 12). This rose looks as if composed of
separate petals, shaped out by the bendings of the meshes
of the capillaries. The capillary windings of every petal are
of three and seldom of four rows, wherefore the petal
consists of an external, middle, and interior vessel. These
vessels communicate on one side with the arterial branches,
and on the other with the venous branch, so that in every
petal can be observed an arterial and a venous half, or, to be
more precise, one half that brings in the blood, and the
other which lets it out from the interweavings.

The connective tissue that lies next to the cutis somewhat
changes its appearance at the edges of the lamelle of the
labyrinthine apparatus. Here it is so swollen, and has such
an appearance, that it can hardly be taken for a connective
tissue (figs. 7 and 8, jt. mrg.). Here it has no adipose matter ;
its cells become much thicker, more juicy, but their offshoots
get so small that in sections they can be observed only when

508 NICHOLAS ZOGRAFF.

the tissue has been split asunder into its component parts by
means of fine needles, after it has been macerated for thirty-
six hours in weak alcohol.

On its outside the cutis is covered with an epidermis of
many layers (figs. 7, 8 and 9). This epidermis consists of
numerous very small cells, scattered among which there is a
quantity of mucous goblet-like cellules.

The latter are not seen in sections kept in alcohol and pre-
pared in paraffine or soap, but they are well observed in
sections prepared in osmic acid, and embedded in the white of
egg according to the method of Kalberla. In such sections
we observe that the number of goblet-like cells in the
epidermis is very great, and that consequently these cellules
can let out a considerable quantity of slime or mucus which
covers the labyrinthine apparatus. The bottom of such cells
is fast set in between other cells of the epidermis, and that
is the reason why when teazed they come out with irregular
edges (fig. 3); one can observe in their goblet-like cavity a
fine irregular protoplasm network and also a slime or mucus,
the traces of which are seen outside the cell.

Having described the construction of this apparatus, so com-
plicated and entangled externally, but so simple in its micro-
scopic structure, I now return to its functions.

The first view of the strongly-developed capillary networks
which are separated from the exterior by a fine epithelium of
two or three layers and have arterial and venous vessels, and
let only one or two corpuscles of blood through their cavities,
makes us suppose that these capillary nets are the principal
part of the organ. If, on the other hand, we remember that
the organism of labyrinthine fishes must have adaptations
for the characteristic, partly terrestrial or amphibious mode of
life which is described by all authors, we shall have to seek
in the organism of these fishes for such adaptations, and we
must hope to meet with organs which will admit oxygen to
the blood, not only from the air dissolved in water, but also
from the atmosphere. The labyrinthine apparatus is such an
organ. To establish a decisive view as to the physiological

LABYRINTHINE APPARATUS OF LABYRINTHIC FISHES. 509

functions of any organ would be possible either after exact or
precise physiological experiments, or after having functionally
studied the anatomy of the organ by comparing its structure
with the structure of already well-known organs, the physio-
logical functions of which are indisputable. The first method
we cannot adopt. Notwithstanding my trying to introduce
a very fine bent glass tube under the gill-operculum into the
cavity of its labyrinthine pouch, it seemed impossible to do so
without making ruptures through which little bladders of gas
do. not enter in the pipe but pass out. Perhaps if I had
examined a greater amount of living material I could have
accumulated and analysed a certain quantity of gas from the
labyrinthine pouch, but the material I had was insufficient,
and therefore I could not solve the question by mere experi-
menting. I then applied myself to the study of facts got by
anatomy. Could the capillary net that has its arteries and
veins be an organ for oxygenising blood? Did it let in venous
blood and let out arterial, or vice versi? To solve the ques-
tion I applied injections. Injections through the arteria bran-
chialis, which is very fine and slender, gave no results; injec-
tions through the spinal aorta almost the same ; I say “almost,
because once after having made some unsuccessful injections
I happened to inject some of the fluid through the aorta into
the labyrinthine apparatus, and into a part of the branchial
arches in a direction opposite to the course of the blood.. But
this not quite successful experiment, together with the results
obtained on another occasion, convinced me that the labyrin-
thine apparatus is in reality a supplementary respiratory appa-
ratus, helping the fish to breathe while in the open air or, in
the damp atmosphere of drying water-basins. The causes of
my supposition are as follows :—Cuvier had made a supposition
that the vessels of the labyrinthine apparatus are supplied with
blood from the gill-artery which passes through, a separate
branch. But for all the preciseness of Cuvier’s experiments
this great naturalist did not succeed in seeing the supposed
connection, and had to limit himself to supposition. All
my efforts seemed to fail till I happened to fall. upon a

510 NICHOLAS ZOGRAFF.

very simple and convincing method. For long I took great
pains to find a means of injecting one way or the other
some fluid into the vessels conducting blood to the labryin-
thine apparatus, but unfortunately I did not succeed. But
two or three times I thought the fine light blue mass pre-
pared according to Beal did penetrate through the gill-arteries
from the heart into the apparatus, and that is the reason why
in my remarks published in 1886 I ventured to say that Cuvier’s
hypothesis was very probable.!

Operating with chloroform upon Macropodus I observed
that both the vessels of the labyrinthine apparatus and the gill-
vessels of the fishes thus operated upon were overfilled with
blood. Then I tried to rapidly curdle the blood by pouring some
boiling water over the organ, and then making it hard with
alcohol. My experiment proved to be successful. Not only
the labyrinthine apparatus and the gill-vessels, but even the
‘‘ bulbus arteriosus ” were filled with coagulated blood, wherein
one could very well distinguish the red corpuscles. After that
I began to prepare sections of the part of Macropodus which
lies next to the gills and its labyrinthine apparatus, and found
that the blood with its red elements could be tinted with
boracic carmine so as to enable one to discern the finest blood-
vessels.

Several sections thus prepared convinced me that the artery
leading to the labyrinthine apparatus (fig. 7, a7. ld.) is a
branch of the artery of the fourth branchial arch (ar. dr), and
consequently furnishes the apparatus with venous blood, the
blood returning from the apparatus to the spinal aorta through
the junction which I had already observed before, and through
which the injected mass passed into the apparatus. So the
labyrinthine apparatus is a supplementary apparatus of the
fourth branchial arch, and the circulation of the blood of the
gill-apparatus of Labyrinthici differs from the circulation of
other Teleostei in the presence of an arterial branch (and

1 « Ueber den sogenanntess Labyrinthapparat der Labyrinthfische (Laby-
rinthici),” von Nicolaus Zograff, ‘ Biologisches Centralblatt,’ Band i, No. 22,
1886, p. 687.

LABYRINTHINE APPARATUS OF LABYRINTHIO FISHES. 511

probably of a vein too) from the fourth arch to the labyrin-
thine apparatus (fig, 13). In certain other fishes we meet
with the same branches of gill-arteries. Johann Miller, in his
famous work on Ganoid fishes, describes the artery of the un-
developed extra gill as an arterial branch of the first branchial
arch.! Therefore, that the artery leads to the supplementary
gill-apparatus from the gill-artery is not an anatomical rarity,
but a consequence of the greater development of capillaries,
and their particular disposition in the labyrinthine apparatus.
If we remember that Peters proved the labyrinthine apparatus
to be only a part of the branchial arch irregularly developed,
we must, in my opinion, have expected that the labyrinthine
apparatus would receive its artery from the artery of its arch,
and that its function would be the same as that of the other
branchial arch, but adapted to the special wants of these semi-
terrestrial fishes.

1 J. Miller, ‘ Ueber den Bau und die grenzen der Ganoiden und ueber das
natiirliche System der Fische,’ Berlin, 1876.

VOL. XXVIII, PART 4.—NEW SER, MM

512 NICHOLAS ZOGRAFF.

EXPLANATION OF PLATE XLI.

Illustrating Dr. Nicholas Zograff’s paper “ On the Construc-
tion and Purpose of the so-called Labyrinthine Apparatus
of the Labyrinthic Fishes.”

Fig. 1.—The labyrinthine apparatus of Macropodus venustus (from
the side). 4,
Fic. 2.—The labyrinthine apparatus of Anabas scandens, var. macro-

cephalus (Java) (from side). 2.

Fie. 3.—The labyrinthine apparatus of Macropodus from behind. (Natural
size.)

Fic. 4.—The labyrinthine apparatus of Anabas from behind. (Natural size.)

Fig. 5.—One of the calyciform cells of the epithelium of the labyrinthine
apparatus (Macropodus). 282 (Leitz, homogen immers. ;}).

Fic. 6.—A cell of the connective tissue of the labyrinthine apparatus
(Macropodus). 75° (Leitz, hom. imm. 33). .

Fie. 7.—Transverse section through the labyrinthine apparatus of Macropo-
dus. ar. dr,. Arteria branchialis 4. ar. Jb. Arteria apparati labyrinthiformis.-
76. Apparatus labyrinth. ar. /d., va. Jb. Arteria et vena appar. labyrinth.
ot. adp. Conjunctiva adiposa. jt. mrg. Conjunctiva marginalis. ar. dr.
Arteria arcus branchialis. ep. Epithelium. -%2, Hartnack, 2.

Fie. 8.—Part of the transverse section through the labyrinth apparatus.
ep. Epithelium. cp. Capillares labyrinthi. ¢j¢. mrg. Conjunctiva marginalis.
cejt. ad. Conjunctiva adiposa. pg. Pigmentum. cf. Cutis. v. a. Vasa
arteriosa. os. Os apparati. 282, Hartnack, 2.

Fic. 9.—Part of a section through the apparatus, fixed with osmic acid.
ep. Epithelium. c/. gl. Cellule glutinose. ct. Cutis. cp. Capillares. ad.

Conjunctiva adiposa, 235, Leitz, 2.

Fie. 10.—Part of the surface of the labyrinthine apparatus. c. ad. Groups
of adipose tissue. a. Arterie. v. Vene. -%°, Hartnack, 2.

Fic. 11.—Part of the rete mirabile of the apparatus. 2, Hartnack, 2.
Fic. 12.—Schematic figure of the rosette of the rete mirabile.

Fic. 13.—Schematic figure of the gill circulatory system of labyrinthic
fishes er. Cor. ao. Aorta. 7b. Apparatus labyrinthicus. ér. Branchie.
a. br., v. br, Arteriz et vens branchiales,

i

hs

Ss faces Seo
= TSS DS BOE
wy YL ells Van)
PEW ees ee / Oper
B05 5S 7, = on, COO
orn) BMmOols

Qawe ;
GIR ERG)

Mor Sourn. Ut SAV NS PL RLL

=e iy ()
Taolas ae!
ogeemsars

Lith? Edin?

F Huth,

STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES, 513

Studies on the Comparative Anatomy of Sponges.

I. On the Genera Ridleia, n. gen., and
QuaSsillina, Norman.

By

Arthur Dendy, M.Sc., F.L.S.,

Demonstrator and Assistant Lecturer in Biology in the University of
Melbourne.

With Plate XLII.

Some months ago, when going over the large collection of
Sponges in the British Museum, I came upon a curious little
specimen,! which a cursory examination soon showed to be of
exceptional interest. I was therefore led to make a careful
investigation of the minute anatomy, and soon discovered that
the specimen in question was the type of a new genus of
Monaxonid Sponges allied to Quasillina on the one hand,
and Suberites on the other. For purposes of comparison,
I was further led to make as complete a re-examination as
possible of the minute anatomy of Quasillina, which led to
an almost complete confirmation of what Dr. Vosmaer has
already published on the subject, and also to the discovery of
some interesting new points.

As, therefore, the present paper treats of the anatomy of
two distinct though closely allied genera, I have thought it
desirable to divide it into two main parts. These two parts
are, however, very intimately connected with one another, and
perhaps the chief interest of the paper lies in the comparison

1 Labelled, “H. M.S. ‘ Porcupine,’ Station 82,” and registered in the
collection 83, 12, 13, 69.

514 ARTHUR DENDY.

of the two types and in the conclusions to be drawn therefrom.
Hence I propose, at the end of the paper, to give a short sum-
mary in which the points of difference and of resemblance
between the two genera will be pointed out and discussed.

Some of the sections were cut in the laboratory of the
Zoological Society, at the Zoological Gardens, and others in
the laboratory of the Royal School of Mines, South Kensing-
ton, and I have great pleasure in taking the present opportunity
of expressing my thanks to the authorities of both these in-
stitutions, and especially to Professor G. B. Howes and Mr.
Beddard, for the great facilities for work afforded to me.

With regard to the figures illustrating this paper, I feel that
a few words of explanation are desirable. None of the figures,
excepting those of mere external form and of spicules, are in-
tended for exact representations of individual preparations.
What I have been unable to make out in one section I have
found in others, or vice versa, and the figures are the result
of the repeated examination of a large number of prepara-
tions. In drawing them, however, the camera has been largely
used, and, although they are to a certain extent diagrammatic,
I believe them to be far more instructive than any mere
facsimile drawings of individual sections could be. Figure 7
is a pure diagram showing what I believe to be the arrange-
ment of the canal system in Ridleia oviformis, after having
repeatedly examined a large number of sections, both transverse
and longitudinal.

In his report on the Hexactinellida of the “Challenger” Ex-
pedition,! Professor Schulze, who is undoubtedly the greatest
living authority on the anatomy of Sponges, observes: “If I
had attempted to copy the individual sections exactly as they
appeared, the essential and typical could not, as a rule, have
been distinguished from the unessential and accidental, except,
of course, by giving a larger number of illustrations than
seemed justifiable for such a slight possible advantage,” and
with this I most heartily agree.

Pos,

STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 515

I. The Genus Ridleia, n. gen.

This genus, which I wish to dedicate to my friend and late
colleague, the Rev. S. O. Ridley, who is so well known for
his work on Sponges, may be diagnosed as follows:

** Sponge corticate, with a highly-developed system of fibrous
tissue. Consisting of a single solid body terminating above in
an osculum which leads into a well-defined oscular tube with
fibrous walls. Spicules monactinal, styli or tylostyli, confined
almost entirely to the ectosome. Skeleton composed chiefly of
stout longitudinal bands of spicules situate in the inner portion
of the ectosome, and of tufts of small spicules projecting at
the surface of the Sponge. Canal system canalicular, flagel-
lated chambers diplodal.”

A generic diagnosis based upon a single species, especially
when only a single specimen is present, must be more or less
tentative, and very possibly future investigations upon fresh
material will lead to a modification of the above.

Ridleia oviformis, n. sp.

External Characters.—The single specimen in the collec-
tion (fig. 1) consists of an egg-shaped body terminating above
in a slight papilla upon which les the small osculum, and
thinning out below into a flattened peduncle. The peduncle
appears, from the existence of certain flattened surfaces (a a),
to have been attached to pebbles during life, in a manner
similar to that which occurs in Quasillina (fig. 8).

The specimen is 15 mm. in height and 7 mm. in greatest
breadth. It has a very compact, cork-like consistence, and its
colour in spirit is pale yellow. .

The osculum (figs. 1 and 2, 0) is single, very minute, and
situate on a small papilla at the summit of the Sponge.

Arrangement of the Skeleton.—The skeleton is almost
entirely confined to the ectosome, or that outer portion of the
Sponge in which there are no flagellated chambers. Its most
important constituent is a-ring of stout, longitudinally dis-

516 ARTHUR DENDY.

posed bundles of large spicules (fig. 3, /. sp.), situate in the
inner portion of the ectosome. Outside of these comes a broad
zone of irregularly interlacing, separate, large spicules, which
occupy the greater portion of the ectosome, and which give
support to the numerous bundles or tufts of small spicules
(fig. 3, sp. 6.) which project outwards at the surface of the
Sponge. The only remaining portion of the skeleton is a zone
of somewhat obliquely disposed and not very well-defined
bundles of small spicules (fig. 3,4. sp.) occurring in the outer-
most portion of the choanosome. These bundles are separated
from the well-defined longitudinal bundles occurring in the
inner portion of the ectosome by the two layers of fibrous
tissue, which I shall describe later on. The remainder of the
choanosome appears to be entirely free from spicules, a fact
which is probably to be accounted for by its dense, cork-like
consistence rendering their presence unnecessary.

The Spicules.—The spicules are all tylostyli of various
sizes, each consisting of a long, straight, or slightly curved,
cylindrical shaft, terminating at the base in a subglobular
head, and at the apex ina fine point. They are of two chief
sizes. The larger ones (figs. 4 and 5) are straight, or nearly
so, and measure up to about 0°9 mm. in length, with a diameter
in the thickest portion of the shaft of about 0°014 mm. The
small spicules (fig. 6), which are almost confined to the surface
brushes and to the innermost zone of longitudinal bundles,
differ somewhat in shape from the large ones. They are very
slender, frequently somewhat curved, and they taper very
gradually to almost imperceptible fineness at the apex. They
measure about 0°2 mm. in length by 0°002 mm. in maximum
diameter of the shaft.

The Ectosome.—In describing the ectosome and choano-
some, I shall, of course, leave out of account the spicular
elements, which have been more conveniently treated of apart.

The ectosome (fig. 3, ect.) is composed for the most part of
a transparent, gelatinous-looking tissue, containing ill-defined
fibres and other mesodermal cells. Towards the outside it
becomes more granular and stains more deeply, and no doubt

STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 517

it is limited externally by an epithelium of flattened cells,
which, however, I have not made out in my sections. The
most remarkable character about it, however, is the presence,
in its innermost portion, of two well-developed layers of fibrous
tissue. In the outer layer (fig. 3, J. f.) the fibres are arranged
longitudinally, and in the inner layer (fig 3, c. f.) horizontally
or circularly. The outermost layer is not perfectly continuous,
but is occasionally interrupted by strands of fibres given off
towards the outside of the Sponge from the inner layer. In
short, the outer fibrous layer consists of bundles of densely
packed, longitudinally disposed fibres, wedged in between the
inner fibrous layer and the stout longitudinal skeleton-bundles
described above (fig. 3, J. sp.). The inner layer is about
0:024 mm. thick, and is composed of densely packed, circu-
larly arranged fibres, giving off, as already described, occasional
outgrowths towards the circumference. At intervals one can
distinctly make out much elongated, granular nuclei. In parts
there are distinct traces of a second layer of longitudinal fibres
internal to the layer of circular fibres. The histological cha-
racters of the different layers appear to be thoroughly identical.
The individual fibres are capable of separation from one another,
and are then seen to be very slender and greatly elongated.
They are, I believe, identical with the myocytes of Sollas,!
and I attribute to them a contractile function.

While speaking of the fibrous tissue, I may also notice that
the central oscular tube of the Sponge (fig. 2) is encased in a
thick sheath of fibrous tissue, continuous in the neighbourhood
of the osculum, with the true ectosomal layer. Here, again,
we distinguish between two chief layers of fibres, circular and
longitudinal. The circular fibres occur on the inside, i.e. next
to the lumen of the tube, and the longitudinal on the outside.
Fibrous tissue also occurs to a greater or less extent in connec-
tion with the main branches of the canal system.

The Choanosome.—The term choanosome has been
applied by Sollas, in his ‘ Preliminary Report on the Tetrac-

1 Article “Sponges,” in ‘Encyclopedia Britannica,’ ed. ix, p. 419,
fig. 21,e.

518 ARTHUR DENDY.

tinellida of the ‘‘ Challenger’? Expedition, to that internal
portion of the Sponge body which contains the flagellated
chambers, as distinguished from the ectosome, in which, it will
be remembered, there are no flagellated chambers at all.

In Ridleia oviformis the choanosome (fig. 3, ch.) is
almost entirely filled up by the closely-packed flagellated
chambers which are embedded in a very sparingly developed
granular ground-substance, in which it is difficult to make out
any cellular constituents. A remarkable character in the
choanosome, and to a certain extent in the ectosome also, of
the specimen under description, is the occurrence of great
numbers of minute, highly refringent, yellowish granules, which,
from the fact that they occasionally exhibit a moniliform
arrangement, I believe to be bacteria of some description.
They are especially abundantly developed around the flagel-
lated chambers, and around the inhalant and exhalant cana-
liculi. ‘Chey occur both in stained and unstained preparations,
and I am inclined to believe that they were present in the
living Sponge.

The fibrous tissue surrounding the various parts of the canal
system has already been described in connection with the
similar tissue in the ectosome.

The Canal System.—The canal system is characterised
by its canalicular character, or, in other words, by the absence
of irregular lacune, such as exist in Halichondrine Sponges.
The pores (fig. 3, p.) are minute, and scattered apparently
without order aud rather sparsely over the general surface of
the Sponge. They may be seen in sections leading into the
narrow, elongated subdermal cavities, which penetrate the
various layers of the ectosome and terminate in slightly
expanded spaces (subcortical crypts of Sollas) beneath the
layer of circular fibres. These subcortical crypts, one of which
is shown in fig. 3, are the only portions of the canal system
which can be considered as at all lacunar. From them the in-
halant canaliculi, leading sooner or later to the flagellated
chambers, take their origin.

1 «Sci. Proc. Roy. Dub. Soc.,’ vol. v, pt. vi, p. 177.

STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 519

The flagellated chambers are more or less elongated or pear-
shaped sacs, measuring about 0°05 mm. in length by 0:03 mm. in
transverse diameter. They are provided with narrow excurrent
and incurrent canaliculi as represented in the diagram, fig. 7.
The canaliculi are extremely difficult to make out, but I have
repeatedly examined a large number of sections, both trans-
verse and longitudinal, and have satisfied myself of their
existence. Their lumina, possibly owing to the contraction of
the Sponge in spirit, are especially obscure, being only very
occasionally discernible. They commonly appear as short,
slender cords, leading to the proximal or distal end of the
chamber, as the case may be, and are chiefly visible owing to
the presence around them of the minute refringent granules
already referred to. As usual in ordinary spirit specimens little
can be made out concerning the collared cells. They appear to
be very minute, and all I have succeeded in observing are their
small nuclei and a number of irregular, gelatinous-looking pro-
cesses projecting into the lumen of the chamber, and probably
representing the remains of the flagella. The exhalant canaliculi
unite together and open ultimately into the central oscular
tube, whence the stream of water is discharged through the
single osculum at the summit of the Sponge (fig. 2, 0.).

I have endeavoured to represent in fig. 7, in a purely diagram-
matic form, what I believe to be the relations of the flagellated
chambers to the exbalant and inhalant canaliculi. It thus
appears that the flagellated chambers are of the type described
by Sollas as “ diplodal,” although the inhalant canaliculi appear
to be but short. It is worthy of note that I have seen in one
instance a flagellated chamber apparently opening direct out of
a subcortical crypt without the intermediation of a canaliculus.
It is obvious that such an occurrence would be most likely to
take place, if anywhere, where the canal system is lacunar, and,
as already pointed out, the subcortical crypts are the only
portions of the canal system in Ridleia oviformis which
are lacunar.

520 ARTHUR DENDY.

II. The Genus Quasillina, Norman.

The only species of this genus as yet described is Bower-
bank’s Quasillina (Polymastia) brevis, a common Shet-
land form, living in moderately deep water. As the somewhat
complicated history and synonymy of this Sponge have lately
been fully treated of by various authors,’ I need not enter
into the question here, but will proceed at once to the descrip-
tion of its general form and minute anatomy.

Quasillina brevis, Bowerbank sp.

External Characters.—Externally the Sponge is seen to
consist of a usually somewhat flattened, oval body, perched on
the summit of a short stalk (fig. 8). At its lower extremity
the stalk is somewhat expanded, and the expanded portion is
generally attached to a small pebble. At its upper extremity
the body terminates in a slight mammiform prominence, at
the summit of which there is a single minute osculum (figs.
8 and 9, o.).

The osculum is usually so much contracted and so difficult
to make out that its existence has, until quite recently, been a
matter of some doubt; and Vosmaer,* who has given by far
the most complete account of the genus yet published, observes
that he “never saw an opening at the top larger than those
where the sea-water enters.” Ridley and Dendy? have, how-
ever, demonstrated that an osculum is, at any rate sometimes,
present, and my recent researches on the arrangement of the
canal system in the body of the Sponge justify us in assuming
that there is always one, although it is frequently found more
or less completely closed up in preserved specimens.

According to Vosmaer (loc. cit.) the general shape of the

1 Vide Ridley and Dendy, ‘ Report on the Monaxonida of the “Challenger ”
Expedition,’ p. 225.

2 ‘Sponges of the “ Willem Barents” Expedition, 1880-81,’ p. 20.

3 Report on the Monaxonida of the “Challenger ” Expedition,’ p. 226,
woodcut, fig. 10.

STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 521

Sponge is subject to considerable variation, but such strongly
marked deviations from the ordinary type as he figures have
not come under my notice, and I believe the essential
characters remain fairly constant.

Certain rectangular marking on the surfaces of the Sponge,
due to the arrangement of the skeleton in the ectosome, are
very characteristic; their nature will be best understood by
reference to fig. 8.

One of the most obvious of the characters of the species is
the general flaccidity and emptiness of the body. ‘Thus,
Bowerbank! remarks, “ When divided longitudinally the parietes
of the Sponge did not exceed in the dried state the fourth of a
line in thickness at any part, and the internal cavity extended
the whole length of the Sponge. The greater number of them
were more or less in a compressed state, but in some there
were strong indications that this was due rather to collapse
than to natural form,” and Norman? begins his diagnosis of
the genus with the words “Sponge consisting of a single
clavate hollow body.”’ I lay especial stress upon this character
as it will be found later on to be very intimately connected
with the arrangement of the canal system, and is one of those
characters in which the two genera, Ridleia and Quasil-
lina, differ very markedly from one another. I may forestall
my account of the canal system so far as to state that the
hollow condition insisted upon by Bowerbank and Norman is
chiefly a post-mortem character, due to the shrinkage of the
loose and delicate choanosome.

A full-grown specimen is usually less than an inch in height,
and less than half an inch in greatest width. Fig. 8 repre-
sents the Sponge of the natural size.

Arrangement of the Skeleton.—This portion of my
subject has already been more or less fully treated of by
Bowerbank, Norman, and Vosmaer (loce. citt.), and, as I have
but little to add, I may dismiss it briefly.

1 *Mon. Brit. Spong.,’ vol. ii, p. 64.

2 Tiast Report on Dredging among the Shetland Isles,” ‘ Brit. Assoc.
Rep. for 1868,’ p. 329.

522 ARTHUR DENDY.

By far the larger and most important portion of the skeleton
lies within the ectosome. It may be best studied by cutting
out a portion of the body wall (ectosome) and examining it as
a transparent object with a low power of the microscope. A
number of stout longitudinal bands or bundles of large spicules
radiate upwards from the peduncle to the osculum, more or
less parallel with one another, but sometimes branching.
These bundles occur in the innermost portion of the ectosome.
They support on their outer surfaces a layer of large, trans-
versely, or obliquely disposed spicules arranged in rather con-
fused and irregular bands. The spicules of both these systems
are more or less parallel with the surface of the Sponge, but
we now come to a third system of spicules placed at right
angles to the surface of the Sponge, and consequently at right
angles also to the other spicules. This third system consists
of innumerable tufts or brushes of small spicules (fig. 9, sp. 6.),
whose apices project from the surface of the Sponge, and whose
bases rest upon the spicule bands of the second system. These
tufts are placed at a little distance from one another, and,
when viewed from the surface, are seen to be arranged in
a somewhat reticulate manner, so as to leave certain irregular
spaces free from their presence. Within the choanosome the
skeleton is very feebly developed, but is not so poorly repre-
sented as in Ridleia. It consists chiefly of separate
bundles of small stylote spicules, each bundle (fig. 11) being
composed of a number of spicules arranged parallel to one
another. These structures closely resemble, on an enlarged
scale, the well-known “ trichite bundles ” or “ trichodragmata ”’
of some Halichondrine Sponges (e.g. Esperella Murrayi),
and I attribute to them the same function as has been attri-
buted to the trichodragmata by Mr. Ridley and myself. In
our report on the Monaxonida of the “ Challenger ” Expedition
we suggest that these last serve, like straw in mortar, to bind
together the soft, gelatinous tissue in which they lie. The
trichodragmata are, of course, bundles of microsclera or flesh
spicules ; the occurrence of similar isolated bundles of mega-
sclera has not, I believe, hitherto been noted. As they are

STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 523

composed of styli the term “‘ stylodragmata ” might be applied
to them. The presence of the stylodragmata in Quasillina,
while they are absent in Ridleia, is probably to be attributed
to the much less compact character and consequently greater —
need of support of the choanosome in the former Sponge.

The Spicules.—The spicules are nearly all styli (i.e.
monactinal spicules simply rounded off at the base and pointed
at the apex), but occasionally the base is swollen into a head,
when the spicule becomes tylostylote.

They are, as already indicated, of. two chief kinds.

(1) Large, straight, or slightly curved, fusiform styli, usually
gradually sharp-pointed at the apex and narrowing consider-
ably towards the base. These measure up to about 0°9 mm.
in length, with a maximum diameter of 0:0144 mm.

(2) Small, slightly curved styli, very gradually sharp-pointed
at the apex, but not so markedly fusiform as the large ones.
These spicules are all of pretty much the same length, viz.
about 0°24 mm., but the variation in thickness is very remark-
able. Both stout and slender ones occur mixed up promis-
cuously in the surface tufts, the stout ones measuring about
0:0096 mm. in diameter, and the slender ones only about
0:002 mm. The two sizes appear to keep fairly distinct
from one another, and one cannot help being struck by the
general absence of intermediate stages. The stouter ones
appear much more distinctly fusiform than the slender ones.

The Ectosome.—The ectosome consists of a somewhat
granular but jelly-like matrix, with small, nucleated, multi-
polar cells embedded in it. It stains more deeply towards the
outside, and on the extreme outside I have detected what
appear to be traces of a single layer of flattened epithelial
cells. This epithelium appears also to line the subdermal
cavities in the ectosome. Fibrous tissue appears to be entirely
absent, at any rate in most parts.

- Here, however, we are met with a considerable difficulty in
deciding the exact limits of the ectosome. In the wall of the
oscular tube there is a very well-developed system of fibrous
tissue, which seems hitherto to have entirely escaped observa-

524. ARTHUR DENDY.

tion, and as the wall of the oscular tube, at any rate in its
upper portion, seems to me to have as much claim to be re-
garded as ectosome as it has to be regarded as choanosome,
I shall describe it in this place. The osculum (fig. 9, 0.) leads
direct into a widely-expanded canal, the oscular tube, about
0'8 mm. in diameter. The walls of this canal present a series
of well-defined circular ridges of fibrous tissue. Owing to the
fact that the oscular tube does not run parallel with the plane
of section, the section figured shows these fibrous ridges
(fig. 9, f. 7.) in transverse section in the upper portion of the
tube, and in vertical longitudinal section in the lower portion.
The ridges themselves do not necessarily run parallel with one
another but may branch and anastomose. Each consists, for
the most part, of a dense band of slender fibres (fig. 12)
running round the oscular tube, and showing here and there
elongated granular bodies, which I believe to be nuclei. The
fibres themselves exhibit a very distinct, wavy outline, strongly
calling to mind the appearance of ordinary white fibrous con-
nective tissue. This fibrous band is covered on the outside by a
layer of curious, granular, flocculent-looking tissue, the nature
of which I do not at present understand (fig. 12, f.). I have
little doubt that these rings of fibrous tissue around the oscular
tube act as sphincter muscles, whereby the diameter of the
tube is regulated. The concentration of the fibrous tissue in
distinct annular bands is probably to be regarded as indicating
a higher degree of differentiation than that which occurs in
Ridleia, where the fibrous tissue forms a continuous sheath.

The Choanosome.—Owing to the lacunar character of the
canal system the choanosome is much less dense and compact
than in Ridleia. The amount of mesodermal tissue in pro-
portion to the rest of the choanosome is also greater, and it
exhibits numerous deeply staining, irregularly shaped cells
(probably the amceboid cells) embedded in a coarsely granular
and at the same time somewhat gelatinous-looking matrix
(fig. 10, m. e.).

The Canal System.—We owe by far the greater portion
of our knowledge of the arrangement of the canal system in

STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 525

this Sponge to the researches of Dr. Vosmaer (loc. cit. supra),
and I am able to do little more than confirm the results which
he has arrived at.

As already stated, the canal system is lacunar. The scat-
tered pores (fig. 9, p.) lead into expanded subdermal cavities
(s. c.) lying between the tufts of spicules in the ectosome.
These in turn communicate with a system of more or less
Jacunar inhalant channels in the choanosome. The inhalant
lacune are separated from the exhalant lacune by strands of
mesodermal tissue in which the flagellated chambers are em-
bedded (fig. 10, f. c.). Many flagellated chambers communi-
cate directly with one and the same inhalant or exhalant
lacuna, but occasionally I have seen traces of what appear to
be short and rather wide canaliculi (fig. 10, @).. These would
seem to be developed when the situation of the chamber
prevents it from opening immediately into a wide canal or
lacuna.

The chambers themselves are usually somewhat elongated,
but their actual shape is a good deal affected by the state of
contraction of the tissues. Good-sized examples measure
about 0°045 mm. in length and 0:02 mm. in_ transverse
diameter.

The exhalant channels open into branches of the oscular
tube, whence the water is discharged through the osculum.

Thus it would appear that the canal system of Quasillina
belongs essentially, as stated by Vosmaer, to his third type,
while that of Ridleia belongs to his fourth type. I believe,
however, that the canalicular and non-canalicular types of
canal system cannot be sharply defined from one another, and
that they will be found to pass by insensible gradations into
one another.

General Conclusions.

The genera Ridleia and Quasillina are shown by their
spiculation, skeleton arrangement, and general form to be
closely allied, and it is not until we have examined properly
prepared stained sections that we are able satisfactorily to

526 ARTHUR DENDY.

comprehend the points of distinction between them. We then
see that they differ to such an extent in their minute anatomy
that we cannot include them both in the same genus.

From the foregoing description it appears that the first
exhibits the canalicular and the second the lacunar type of
canal system ; in Ridleia the inhalant and exhalant channels
are canalicular, and the flagellated chambers are provided with
special inhalant and exhalant canaliculi, while in Quasillina
the inhalant and exhalant channels are for the most part
lacunar, and the flagellated chambers open directly into them
without the intermediation of narrow canaliculi. At the same
time there is some evidence to show that these two types of canal
system cannot be sharply defined from one another, and
that flagellated chambers with, and others without special
canaliculi may coexist in the same Sponge.

It is noteworthy that in Quasillina the chambers, although
usually opening directly into the lacune of the canal system,
are elongated in form like those of Ridleia, and not, as in
the Halichondrina, where such a mode of opening is typical,
spherical, or subspherical.

Both genera are remarkable for the development of the
fibrous tissue. In Ridleia it is largely developed in the
ectosome proper, and in the wall of the oscular tube, being
arranged in well-defined layers of longitudinal and circular
fibres. In Quasillina, on the other hand, it is almost
entirely absent from the ectosome proper, but is well developed
in the wall of the oscular tube, where it forms definite annular
ridges in which the close-packed fibres (myocytes) have a dis-
tinct, wavy outline.

The mode of occurrence of the fibrous tissue in this and in
other genera in which it occurs, indicates that its function is a
contractile one, or, in other words, that the fibres are muscular
fibres. The annular bands of fibres around the oscular tube
of Quasillina are probably to be regarded as sphincter
muscles.

The genus Quasillina has been recognised for a long time
as a member of the sub-family Suberitide. It is, however,

STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 527

a very aberrant one, and the new genus Ridleia forms a
connecting link between it and the other members of its sub-
family.

Probably Quasillina is to be regarded as a more highly
modified form than Ridleia. I have already pointed out that
the arrangement of the fibrous tissue in definite bands indicates
a higher degree of modification than that which exists in
Ridleia, where it forms more or less continuous sheaths, and
the occurrence of the stylodragmata appears to me to be an
adaptive modification resulting from the lacunar character of
the canal system, and the general delicacy of the choanosome.
Stylodragmata occur, so far as I am aware, in no other members
of the group.

Professor Schulze, in his ‘ Report on the ‘‘ Challenger” Hex-
actinellida,’ and elsewhere, has expressed the opinion that the
polyactinal type of spicule is the primitive form from which
the monactinal type has been derived by abortion of the rays.
Although in a previous paper! I upheld, for reasons therein
given, the contrary hypothesis, I am now inclined to regard
Professor Schulze’s view as the more correct one.

It is, I believe, generally admitted that the swollen base, or
head, of a typical Suberitid spicule, together with the corre-
sponding enlargement of the axial thread, indicates the position
where other rays were at one time united with that one which
now alone remains. In the typical Suberitide, then, and in
Ridleia, all rays but one have disappeared, but their former
presence is still indicated by the head of the tylostylote spicules.
In Quasillina the spicules are still more modified, and even
the head has, in most cases, disappeared.

Judging, then, from the condition of the fibrous tissue
and of the spicules, we should expect the canal system of
Quasillina to be less primitive that that of Ridleia, and,
in general, the lacunar type of canal system, as it occurs in
the Monaxonida, with chambers opening directly into wide

1 Dendy and Ridley, “On Proteleia Sollasi,” ‘Ann. and Mag. Nat.
Hist.,’ ser. 5, vol. xviii, p. 153.

VOL. XXVIII, PART 4.—NEW SER. N N

528 ARTHUR DENDY.

lacunz, to be less primitive than the canalicular type, in which
the chambers are provided with special canaliculi. There is
not sufficient evidence at present to show whether or not this
is the case, but I think it not impossible that it may be so.

EXPLANATION OF PLATE XLII.

Illustrating Mr. A. Dendy’s paper “Studies on the Com-
parative Anatomy of Sponges. I. On the Genera Ridleia,
n. gen., and Quasillina, Norman.”

Figs. 1—7.—Ridleia oviformis.

Fig. 1. Entire Sponge, x 2. o. Osculum. aa. Flattened surfaces, by
which the Sponge has probably been attached to pebbles.

Fig. 2. Longitudinal section, taken at right angles to the surface shown
in Tig. 1, xX 5. o. Osculum, leading into oscular tube; the black line
indicates the distribution of the fibrous tissue. (The figure is rather dia-
grammatic ; the canals radiating towards the centre are exaggerated.)

Fig. 3. Portion of transverse section, x 44. ect. Ectosome. ch. choano-
some. p. Pore, leading into slit-like subdermal cavity. e¢. ¢. Exhalant
canals converging towards the centre of the Sponge. J/./f. Layer of
longitudinal fibres cut across. c.f Layer of circular fibres. sp. d.
Bundles of small spicules projecting at surface. J/. sp. Longitudinal
bundles of spicules cut across. 7. sp. Oblique fibres of spicules within
the choanosome cut across. (A little diagrammatic.)

Figs. 4 and 5. Two of the larger tylostyli x 190.

Fig. 6. One of the small tylostyli x 500.

Fig. 7. Diagram of diplodal canalicular canal system, as it occurs in
Ridleia oviformis, x 190. iz. c. Inhalant canal, giving off narrow
canaliculi to f ¢., the flagellated chambers, which communicate again by
narrow canaliculi with e. c., an exhalant canal.

Fics. 8—12.—Quasillina brevis.

Fig. 8. Entire Sponge attached to a pebble, nat. size. (After Ridley and
Dendy, ‘ Report on the “ Challenger” Monaxonida.’) 0. Osculum.
Fig. 9. Part of longitudinal section passing through the osculum, x 23.

o. Osculum. p. Pores. s.c. Subdermal cavities. /. Lacune of canal

STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 529

system. f. r. Fibrous ridges on wall of oscular tube. sp. 4. Bundles
of spicules projecting at surface. (Rather diagrammatic.)

Fig. 16. Part of transverse section through choanosome, x 190, showing
arrangement of flagellated chambers and inhalant and exhalant Jacune.
f.c. Flagellated chambers. J. /'. Lacune of canal system. m.c. Meso-
dermal cells, probably amceboid. At @ there is an indication of a canali-
culus at one end of a flagellated chamber. (Rather diagrammatic.)

Fig. 11. Separate bundle of small styli (stylodragma) from choanosome,

xX 284.

Fig. 12. Portion of a branched fibrous ridge from inner surface of wall
of oscular tube, seen in section taken parallel to the wall of the oscular
tube, x 284. 2. Elongated nuclei. f. Flocculent granular tissue
covering the ridge.

(The spicules, except where separately figured, are represented in blue.)

= . : ~y : ros :
' 7 . ,
«
‘
~
*
2
© |
‘
— a
‘
‘
‘ a -
.
~
a
4 -
; ;
L
ia
hal =
—
s
xy - 4
- iY

a
»
co
a
5
ao
A
ond
=
2
<=

UNS GOR

7)
d

f

LAL

ve
vd
¢

7
Vf
v

Mor. Fourw

S

F Huth, Lith? Edin®

;

KLEINENBERG ON DEVELOPMENT OF LOPADORHYNCHDS. 531

Kleinenberg on the Development of
Lopadorhynchus.

By

G. C. Bourne, B.A., F.L.S.,

Fellow of New College, Oxford, and Assistant to the Linacre Professor in
the University.

Ir is now more than a year since Professor Kleinenberg
published, in the ‘Zeitschrift fiir Wissenschaftliche Zoologie,’ a
long account of the development of a polychzte Annelid, Lopa-
dorhynchus. Whether it is because of the extreme length of
the paper, which extends over 225 pages, that the work has not
attracted the attention it deserves, or because the very nume-
rous illustrations include no diagrams, but are invariably exact
drawings of the sections studied by the author, and require to
be laboriously re-interpreted by the reader, or because the
facts and views expressed are so novel and so much opposed to
all preconceived notions of what mesoblast ought to be, the
fact remains that this paper is little known by English
readers.

The morphological importance of the questions raised by
Kleinenberg concerning the origin of the mesoblast is so great,
that a reproduction of those parts of his paper which deal espe-
cially with this subject calls for no excuse.

In an introductory chapter, the consideration of which I
shall defer until the facts of the development of Lopadorhyn-
chus have been presented to the reader, the author lays down
the proposition that there is no such thing as a mesoblast as a
specially differentiated germ layer. The existence of two
primary germ layers in the Ceelomata, ectoderm and endo-
derm, homologous with the similarly-named layers in Celen-

oon G. C. BOURNE.

terates, is admitted; the existence of a mesoderm, if by it we
understand a tissue or aggregate of tissues interposed between
the two primary germ layers, is not denied, but it is specifi-
cally denied that those intermediate tissues originate from a
separate embryonic layer, composed of undifferentiated cells.
Ectoderm and endoderm, says the author, are the foundations
of all other tissues and organs, with the sole apparent exception
of the germinal cells, not only in Coelenterata but in all
Metazoa. ‘Tissues and organs originating directly from these
layers take over with them the power of giving origin to other
tissues and organs on their own account; in no part of the
living body is the capacity for change entirely lost. By this
the genetic relationship between any organ and the primary
germ layers is not destroyed, it is only made more remote by
the intercalation of one or more intermediate developmental
phases. Say that an organ or tissue which originates directly
from one of the primary germ layers is a primary tissue ; it
gives rise to other tissues which are secondary, those give rise
to another series of tertiary tissues, and so forth. The meso-
blast, so called, is in fact nothing more than the aggregate of
these primary, secondary, and tertiary tissues, the precise origin
of which is often obscured by the tendency to precocious
development.

Let us see how this view is supported by the facts of the
development of a single form. The account of the develop-
mental phases of Lopadorhynchus is so long that it would be
impossible to give details of the whole organogeny of the
animal in the space of a short essay. I will therefore confine
myself to a short description of the general development of the
Aunelid from the larval form, and will give the details of the
formation of a few selected tissues and organs which have a
special bearing on the question in hand.

The earlier developmental history of Lopadorhynchus was
not studied; the youngest larva described is a true trocho-
sphere, with a preoral ring of cilia, called by Kleinenberg by
the convenient name of prototroch, dividing the body into pre-
oral and oral lobes of nearly equal size. The stomodeum

KLEINENBERG ON DEVELOPMENT OF LOPADORHYNCHUS. 533

marks the ventral surface, and is not yet connected with the
archenteron. The latter is a closed sac, occupying the whole
of the interior of the trochosphere, except that a pair of small
slits on either side of the stomodcal invagination represent the
remnants of the segmentation cavity. Both ectoderm and
endoderm are composed of a single layer of cells. The ecto-
derm cells are colourless, their nuclei are distinct, but the cell
outlines are difficult to see, and their size and shape vary in
different parts of the larva. Those ectoderm cells which bear
the cilia of the prototroch are, for instance, much larger than
those on the oral and preoral lobes. The endoderm cells are
coloured, and their inner ends are stuffed with yolk-granules.
Two differentiations can be seen in the ectoderm of the aboral
lobe (umbrella, KI.). On the ventral surface, further removed
from the aboral pole than from the prototroch, is a pair of
aree, where the ectoderm cells are more closely packed
together than elsewhere. These are the foundations! of the
sense-plates of the head. A pair of thickened ectoderm are
can also be seen on the ventral sides of the oral lobe (sub-
umbrella, K1.), the foundations of the ventral plates, and close
beneath the stomodzum an unpaired thickening of the ecto-
derm, forming a transient larval organ known as the ventral
shield. The further important changes in the larva concern
the fate of the ventral plates, and the appearance of the foun-
dations of a number of organs on the preoral lobe. On the
preoral lobe appear two pairs of prominences in front of the
sense-plates, which are presently put into connection with the
jatter, and the three pairs of structures form a semicircle en-
closing a field of thickened ectoderm, the head shield, which
occupies nearly the whole ventral surface of the preoral lobe.
A single symmetrical pit also makes its appearance towards
the aboral pole. At a later period the sense-plates divide to
form the foundations of the second pair of permanent antenne
anteriorly, and the foundations of the olfactory organs poste-

1 Tuse the term “foundations” as the equivalent of the German word

Anlage, which cannot be exactly translated into our language. The use of
the term “rudiment ”’ in this sense is clearly open to objection.

534 G. C. BOURNE.

riorly. The second of the two pairs of prominences above
mentioned is the foundation of the first pair of permanent
antenne.

In the oral field the ventral plates increase greatly in size,
extend anteriorly and posteriorly, and become divided by a
number of clear transverse lines into segments, but as yet there
are no external furrows marking off one segment from another.
Usually as many as eight segments make their appearance at
once, and the number increases up to sixteen, but is very in-
constant. In the next stage the number of segments is not
added to, but each segment grows greatly in size, and soon the
hinder part of the ventral plate begins to project beyond the
oral lobe of the larva. Already in each segment a median, and a
distal prominence, have made their appearance on each side of
the body. These are the foundations of the dorsal and ventral
cirrhi of the parapodia, and before long a third prominence
appears between each pair of cirrhus-foundations, which is the
foundation of the chetopodium. In further growth the vermi-
form segmented body of the Annelid, derived from the ventral
plates, becomes more and more important, and preponderates
over the oral lobe of the larva. The chetopodial elements
grow out into the parapodia of the adult, and in so doing carry
with them the dorsal and ventral cirrhi which, at first separate
from the chzetopodia, are eventually carried on the ends of the
latter. At this stage the pre-oral lobe and the prototroch retain
their primitive characteristics, and the larva has a broad, dome-
like trochosphere head, with a vermiform body bearing para-
podia. The foundations of the two pairs of antennz have,
however, grown out into processes projecting from the ventral
surface of the pre-oral field, and the olfactory pits are developed
as deep sacs, with a narrow, slit-like opening, which can be
evaginated and drawn in again.

The final changes consist in the atrophy of the preoral lobe,
the loss of the ciliated ring, the further growth of the two pairs
of antenuz, and the completion of the parapodia with their
sete, Xe.

The internal organs of the adult Annelid are developed by

KLEINENBERG ON DEVELOPMENT OF LOPADORHYNCHUS. 535

successive differentiation of derivatives of the two primary
layers, ectoderm and endoderm.

In the first place it may be stated that the adult Lopador-
hynchus has no blood-vascular system, and Kleinenberg was
unable to find its nephridia. The youngest larva observed
had a well-developed provisional nervous system, consist-
ing of a ring of nerve-fibres lying beneath the prototroch,
and a series of ganglion-cells, connected by fine processes with
the sub-prototrochal nerve-ring. It is unnecessary to enter
into the exact arrangement of these ganglion-cells: they lie in
the ectoderm of the preoral lobe of which they are differentia-
tions, and although the cephalic ganglion of the adult animal
and the sense-organs in connection with them are developed
from the ectoderm of the preoral lobe, the central nervous
system of the trochosphere has but very little share in their
establishment. The whole of the nervous structures of the
head of the adult are formed from three centres: firstly, from
nerve ganglion-cells, which make their appearance at an early
stage in the ectoderm of the preoral lobe ; secondly, from parts
of the sense-plates already mentioned; thirdly, from indif-
ferent ectoderm cells. The more anterior of the two pairs
of prominences, which have been described as appearing on
the ventral surface of the preoral lobe, and are called by Klein-
enberg the pole antenne (Scheitelantennen), prove to be
nothing more than cell-proliferations, and eventually form the
anterior lobes of the cephalic ganglion of the adult. The
remainder of the ganglion is formed from ganglion-cells and
nerve-fibres which arise in connection with the sensory organs
of the head (the two pairs of antenne, the olfactory pits, &c.),
to which are added cells derived independently from the ecto-
derm. The newly-formed cephalic ganglion is first placed in
connection with the sub-prototrochal nerve-ring by means of
commissural fibres; the angles of these commissures are after-
wards produced posteriorly into the oral lobe, and, becoming
subsequently connected with the anterior ganglion-pair of the
ventral chain, form the pericesophageal connectives. In a
later stage of growth the cephalic ganglion retires from the

536 G. C. BOURNE.

ectoderm, and lies in the cavity of the head, separated by a
wide space from the head wall, except at its anterior end, where
it retains its connection with the ectoderm. In the young
larva there is no special nervous apparatus in the oral lobe, but
the foundation of the ventral ganglionic chain can be detected
in the ectodermic thickenings already described as the ventral
plates. At an early period of development special nervous
elements are found ventrad of the inner edges of the ventral
plates, but lying outside of them; they are the first foundation
of the central nerve-chain of the body. In connection with
them rise small bundles of sense-hairs, arranged metamerically.
Just beneath these sense-hairs is a mass of large nuclei, derived
from the continued multiplication of the ectoderm cells of that
region. They lie close upon the endoderm. The inner bound-
ary of the ectoderm layer is quite distinct, but at a later
period it disappears suddenly; at this point the endoderm
becomes somewhat bent in towards the archenteron, so that a
space is formed between it and the ectoderm. Many of the
cells, proliferated from the ectoderm in the manner just de-
scribed, find their way into this space. One might say that
the ectoderm bursts through at this point, emptying its contents
as an abscess might, between the previously existing cell
layers (see Fig. 1).

As the result of this process there are established three
layers—an ectoderm, a middle layer derived from it, and an
endoderm. ‘This division of the ventral plate is only complete
in its anterior position, posteriorly it remains undivided. The
middle layer gives rise to muscular tissue only, and hence is
called muscle-plate ; the external plate, or ectoderm of the
ventral plate, may be called the neural plate. It gives rise to
the ventral ganglion-chain, sense-organs, and also to non-
nervous structures, the sacs of the sete. These seta-sacs are
formed by a proliferation of the cells of the anterior parts of
the neural plates; they grow inwards as pear-shaped projec-
tions, and enter the muscle-plates. In the lateral parts of each
neural plate is next developed the dorsal cirrhus as a prolife-
ration of cells, and a little later a similar proliferation, close to

KLEINENBERG ON DEVELOPMENT OF LOPADORHYNCHUS. 537

the seta-sac, gives rise to the ventral cirrhus. At a later
stage the neural plates of the opposite sides of the body become
connected together by a series of transverse commissures. |
These commissures are developed coincidently with the elonga-

Fic. 1.—Transverse section of a young larva of Lopadorhynchus, showing
the formation of the muscle-plate from the ectoderm. Hc. Ectoderm. end.
Endoderm. M.P. Muscle-plate. S.O. Sense organ.

tion of the body in such a way that for each pair of seta-sacs a
single commissure is developed. Each pair of seta-sacs with
its transverse commissure represents a single somite. In
addition to the transverse commissures, longitudinal fibres are
developed towards the median edges of each neural plate ; these
put the ganglionic centres of the ventral chain into communi-
cation with one another, and form the longitudinal connectives,
From their mode of formation the connectives are necessarily
paired structures ; subsequently they become intimately united
together ; but their paired structure is always distinguishable.
At first both commissures and connectives lie directly upon
the ectoderm, but subsequently become separated from it.
The foundations of the dorsal and ventral cirrhi are divided
into proximal and distal portions. The latter form bundles of

538 G. OC. BOURNE.

nerve-fibres communicating with the ventral nerve-chain. As
the seta-sac grows in size the dorsal cirrhus is pushed further
and further away from the ventral cord, so that the nerve
connecting the two becomes longer.

Before the development of the cirrhi has progressed very
far the segmentation has begun, not as a simple process of
division of mesoblastic plates into somites, but by the differ-
entiation of the already existing foundations of various organs.
The tissues composing the muscular and neural plates are at
first dense and homogeneous throughout, but now they become
divided into masses of denser tissue, separated by transverse
tracts of loose tissue, which mark the boundaries of the seg-
ments. Each segment of the neural plate contains the two
halves of a rudiment of a ventral ganglion, two pairs of cirrhi,
and a pair of seta-sacs, in addition to which are several masses
of undifferentiated cells. The more anterior segments are
further advanced than those posteriorly situated. The seg-
mentation involves only the lateral moieties of the neural
plates, it does not extend to the middle line. The ventral
ganglion-chain is not, in fact, divided up into segments at any
period of its development; and, although at a later stage the
ganglia correspond exactly with the body segments, this does
not alter the fact that the foundations of the organ, as well as
the fully-developed chain of ganglia and connectives, are at all
times continuous throughout the length of the body. A para-
podial nerve system is developed, not directly from the ventral
nerve-cord, but in connection with it; and the latter may be
said to determine its formation.

The visceral nerve system of the adult is developed directly
from the anterior ganglion pair of the ventral nerve-chain, by
the differentiation and separation of its dorsal moiety. The
cells thus separated spread over the pharynx, and in the earlier
stages of development do not extend over more than the
pharyngeal, and therefore ectodermal portion of the alimentary
canal.

The muscular system, like the nervous system, is deve-
loped as separate larval and adult elements. The two arise

KLEINENBERG ON DEVELOPMENT OF LOPADORHYNCHUS. 559

independently, and very little of the larval musculature passes
over to the adult. The larval muscles originate from two
ectoderm thickenings situated beneath the prototroch. Finally,
the larva is provided with a muscular layer, which is to be
found nearly everywhere between the ectoderm and endoderm.
The exact arrangement of the larval muscles need not detain
us: it only remains to be noticed that no part of the larval
musculature is formed from the ventral plate, whereas the
adult muscles are derived, if not entirely, for the most part
from that structure. The separation of the muscle-plate from
the neural plate has already been described: it must be borne
in mind that both owe their origin to an ectodermal cell-pro-
liferation. The muscle-plates grow rapidly in size and im-
portance, partly through the addition of wandering cells
derived from the ectoderm. The muscle-plates become seg-
mented from before backwards, so that each somite has a pair
of muscle-plates, which are separate from one another in the
mid-yentral line. A seta-sac projects into the muscle-plate
on either side of each segment, dividing it into four regions—
an anterior, a posterior, a median, and a lateral region. The
ventral muscles of the adult are derived from the median
region, the dorsal muscles from the lateral region, the parapo-
dial muscles from the anterior and posterior regions. (See
Fig. 2.) In the cavities of the parapodia in the adult are found
peculiar stellate muscle-cells, which have no connection with
the ventral plate, although they appear to be remnants of the
larval nervous system, they are far more probably derived from
separate centres determined by the muscle foundations of the
body.

The muscle-plate does not split into splanchnopleure and
somatopleure in Lopadorhynchus, nor does this happen in
many other Annelids (Asterope, Nauphanta, and others), but
the musculature of the gut is formed by the aggregation of a
number of single free wandering cells, derived originally from
the ectoderm through the intervention of the muscle-plate.
Thus we learn that the ceelom of the adult Annelid is neither
an euterocele nor a schizocele, but is an archiceele, i.e. the

540 G. C. BOURNE.

remnant of the primary body cavity existing between ectoderm
and endoderm. The muscles of the head of Lopadorhynchus
are derived from the body muscles, as is the case in Polygor-
dius and other Annelids.

Fic. 2.—Transverse section of an older larva, showing the further develop-
ment of the muscle-plates and the origin of the ventral ganglion chain, cirrhi,
and seta-sacs. £.C. Ectoderm. And. Endoderm. D.M.P. Dorsal muscle-plate.
V.M.P. Ventral muscle-plate. D.C. Dorsal cirrhus. V.C. Ventral cirrhus.
S.S. Seta-sac. V. gang. Ventral ganglion. V.S. Ventral shield.

The seta-sacs are invaginations of the ectoderm, solid at
first, but subsequently developing a lumen, the cells lining
which secrete the sete. The sac of the aciculum is formed as
a special outgrowth of the seta-sac, which becomes partially
constricted off.

The gonads of Lopadorhynchus are derived from the ecto-
derm. Earlier or later in development the ectoderm cells
lying behind and to the side of the parapodial ganglia prolife-
rate, and the mass of cells thus formed is invaginated in the
form of a pear-shaped body projecting inwards towards the
body cavity. In the female, cells are separated from the inner
ends of these ingrowths, which develop into ova, wander free

KLEINENBERG ON DEVELOPMENT OF LOPADORHYNCHUS. O41

in the body cavity, and settle anywhere on its walls. In the
male, masses of cells are shed off into the body cavity, which
become spermatoblasts.

There remains to be described the development of the pharynx
and esophageal glands of the adult. The stomodeum of the
larva does not become the pharynx, but the latter is derived
from a pair of gland-like stomodzal outgrowths, which subse-
quently fuse together and form the pharynx. The esophageal
glands are formed as three outgrowths—two lateral, one median
and dorsal, from the pharynx.

The important facts which are to be gathered from the
foregoing account are these: Lopadorhynchus has no meso-
blast in the sense of a germ layer composed of undifferentiated
cells. The larva has its own nervous system and its own
musculature, no parts of either of which pass over to the adult,
but are replaced by a new nervous system and musculature,
and are afterwards aborted. The central nervous system of
the adult is formed from a number of separate centres or foun-
dations, the situation of which is in part determined by the
pre-existing nerve centres of the larva, although the latter
have no direct share in their formation, but act only as foci,
in connection with which fresh groups of nervous elements are
differentiated.

The muscular system of the adult is derived wholly from the
ventral plates, as is also the whole ventral chain of ganglia,
with its connectives and commissures. These ventral plates
are, in their earlier stages, nothing more than thickenings of
the ordinary ectoderm, the cells of which become subsequently
divided into muscle-plates and neural plates. The muscle-
plates give rise to the whole of the muscular system, both of
the body walls, and of the gut; but they do not split into two
layers, comparable to somatopleure and splanchnopleure, the
ceelom being formed simply as an extension of the space exist-
ing between ectoderm and endoderm, the peritoneal walls of
which are formed by cells derived from the muscle- plates.
The neural plates give rise to the whole of the ventral ganglion
chain, and besides this to other and different organs, viz. the

542 G. C. BOURNE.

cirrhi, parapodia, and seta-sacs. All the tissues usually classed
as mesoblastic are derived from the ectoderm; the endoderm
appears to form nothing more than the lining of the gut.

The generalisations which Kleinenberg would found upon
these facts may be briefly stated as follows:

New organs and tissues are always developed in connection
with and through the intervention of pre-existing organs and
tissues. 'That this must be true in a phylogenetic sense is
obvious on reflection, for a new form with newly specialised
organs must always be developed from a form in which the
same vital functions were performed, less efficiently it may
be, by other or less perfect organs. Hence, if ontogeny is a
recapitulation of phylogeny it must also be true in an ontoge-
netic sense, but the fact has been overlooked because authors
have been too ready to refer developing organs to the undiffer-
entiated cells of the two primary layers, ectoderm and endo-
derm, or to a mesoblast, falsely assumed to be an undiffer-
entiated third germ layer, homodynamous with the two first
named.

In the Coelenterata the two primary germ layers are present,
and form all the tissues of the adult: each of them may give
rise to differentiated tissues or organs, to which special func-
tions are appropriate. Among the first tissues to be formed
are the nervous and muscular, and since irritability and con-
tractility are complementary to one another, these two tissues
are invariably developed in such close connection that they
may be said to be mutually determinant. But they are not
structures developed de novo; they are the result of the
specialisation of function of already existing cells, whether of
ectoderm or of endoderm, and very frequently arise from the
direct transformation of those cells. In the higher Metazoa
such tissues derived from the primary ectoderm or endoderm
because of the increased functional activity due to their spe-
cialisation, give rise to modifications in adjoining cells; these
group themselves round the tissues first formed, and, becoming
themselves specialised, lead to the formation of other tissues
and organs. In some instances the cells newly formed are

KLEINENBERG ON DEVELOPMENT OF LOPADORHYNCHUS. 543

joined to those already present to form a mass, the component
units of which, although unlike, subserve the performance of
some one physiological function (e.g. glandular and pavement
epithelia, connective tissue, muscle-fibres); the whole aggre-
gate forms a single organ, and the new differentiations only
serve to increase its efficiency. In other instances groups of
cells, subserving a special function, will induce changes in
adjoining cells, which changes are the first step towards the
establishment of new tissues and organs. In every case change
of function is the guiding principle in the formation of fresh
tissues and organs.

It is a mistake to suppose that a new organ is formed by the
gradual growth and change of a pre-existing organ. Since
Darwin showed that species through gradual modification are
transformed into new species, it has been generally believed
that organs similarly may go through slow series of uninter-
rupted modifications, the final resuit of which is the transfor-
mation of that organ into some other organ having an analogous
though not wholly similar function, or even into some wholly
different organ with entirely different functions. The phylo-
genetic and ontogenetic changes of the gill-slits in the Verte-
brata are familiar instances of this supposed progressive
modification of a single organ. As a matter of fact this process
of direct change has seldom, if ever, taken place. In animals
of simple organisation tissues and organs are formed to which
special functions are appropriate. Their functional activity is,
as it were, a disturbing element in the organism; it induces
changes in neighbouring tissues, and gives the signal for new
specialisations in them. The functional activities of the newly
specialised tissues must always bear some relation to the
function of the organ which determined their origin, and mast
either support or modify their action. The newly-formed tissues
again affect the organism, their importance increases, and they
may in their turn give rise to fresh tissues. Finally, they may
become so important that they outweigh in functional import-
ance the organ to which their origin was due; they then take
its place, and the latter dwindles till finally it may disappear

VOL. XXVIII, PART 4,—NEW SER. 0 Oo

544 G. C. BOURNE,

altogether. This process, which is of the greatest importance,
is called by Kleinenberg the development of organs by substi-
tution. The organs resulting from the changes are the organs
of substitution ; those through which the changes are effected
are the intermediary organs. In no case of substitution are
the intermediate steps represented by an indifferent germ layer,
but always by a functional and specifically differentiated organ.

One of the most striking instances of substitution is the
development of the axial skeleton of the Chordata. Both phy-
Jogenetically and ontogenetically the transformation of an
already existent specialised portion of the gut leads to the
formation of a supporting skeletal rod,—the notochord. When
once this is established it induces correlative changes in the
muscles, which act more efficiently because of the firmer sup-
port. The muscular changes react upon the supporting rod,
which no longer affords a suitable point d’appui for them,
and is replaced by, not converted into, a segmented cartilagi-
nous column. The very segmentation is caused by the muscles,
for the cartilage is primitively unsegmeuted. Finally, the
cartilage is inefficient, and is replaced by bone. It is most
important to observe that each organ is functional and com-
plete in itself for a time, and does not form any part of the
organ which subsequently replaces it, but is entirely or very
nearly aborted. The disappearance of the notochord is an
excellent example, but many others could be given, all equally
good—for instance, the successive development of pronephros,
mesonephros, and metanephros.

The author claims that the theory of the development of
organs for substitution gives an entirely new significance to the
persistance of rudimentary organs. That organs which have had
no functional importance should persist through long series of
generations is surprising and inexplicable by conventional views,
But if once it is recognised that they are intermediary organs
their use in ontogeny becomes apparent. Phylogenetically
they induced the formation of the organs which replaced them,
therefore they are retained in order that, in ontogeny, they may
give the impulse to the formation of those new organs of sub-

KLEINENBERG ON DEVELOPMENT OF LOPADORHYNCHUS. 545

stitution, having done which they may either disappear (gill-
slits of Vertebrates as such), or remain as obviously rudimentary
structures.

The above considerations must lead to an entire remodifica-
tion of the conventional assumption of a mesoblast. There is
nothing in morphology about which there is so much disagree-
ment as the homology of the mesoblast and the cavities included
in it. The most diverse origins have been assigned to this
supposed germ-layer ; it arises from the outer germ-layer alone,
from the inner germ-layer alone, from both together, from
neither, from two special blastomeres, from every part of the
living blastoderm (Sarasin), not from the blastoderm at all,
but from the yolk (His). With so many contradictions before
one, it is at least a step in advance when we can strike out on
a fresh line, and begin a new series of observations on the
assumption that there is no mesoblast. Whether Kleinenberg’s
views will stand the test of further proof is of course an open
question, but it cannot be doubted that investigations under-
taken from his point of view must be fertile of results. To
trace back organs through the intermediary organs which gave
them origin, and to establish homologies between the latter
and the permanent organs of lower forms, is a task which
offers many advantages to the researcher, and is likely to
yield much surer and more satisfactory results than the
attempt to refer an organ merely to one of the three primary
germ-layers.

It must not be forgotten that the homologies of the meso-
blast and ceelom are two of the most debateable questions in
morphology. But, as Kleinenberg says, a hole is a hole the
world over,—if we can establish the homologies of the walls
bounding a cavity, there is no need to quarrel about the empty
space itself. Already one can see the possibilities of satisfac-
tory explanations, through the theory of substitution, of the
celomic cavities of Arthropods, the greater part of which is
shown by Sedgwick to be turgid blood-vascular space, and the
various stages of ccelosis in the Hirudinea. Kleinenberg states
that the ceelom of Lopadorhyuchus is what would be conven-

546 G. C. BOURNE.

tionally called an archiceele, i. e. it is derived from the original
segmentation cavity, as is the cavity of the head in all Anne-
lids. It is possible that this statement may be extended to
all other Annelids in which the ccelom is stated to result from
the splitting of the mesoblast into somatopleure and splanch-
nopleure, but against the probability of this view must be set
the careful researches of Hatschek.

Finally, it is gratifying to find that Kleinenberg has insisted
upon the intimate connection between morphology and phy-
siology. Enough has been said in the foregoing pages to show
how strongly he insists upon the dependence of function upon
form. Inthe introductory chapter he condemns the theory of
the Hertwigs, that the origin of cells can be predicated by their
form, aud shows from numerous examples that the form of
cells is caused solely by their function and the mechanical
necessities of their position. Naturally the two formative
conditions, position and function, generally coincide ; but if
there is a contention between those conditions, function
prevails.

In an abstract which has to be confined within certain limits
of space, I am unable to give more than the most meagre
outline of Prof. Kleinenberg’s argument. It is enough, how-
ever, if I have succeeded in calling attention to his remarkable
work, and in inducing morphologists to read carefully the
formidable paper in which his results are published.

PAR Or VOd KOOV EE,

NEW

Allurus tetraedrus, anatomy of,
by Beddard, 365

Amphiura squamata, apical plates
of, by H. Carpenter, 303

Ascidians, ciliated pit of, by Lilian
Sheldon, 131

Baur on the quadrate of Mammalia,
169
Beddard on the anatomy of Allurus
tetraedrus, 365
», on the occurrence of numer-
ous nephridia in the same segment
in certain earthworms, 397
Blood of Rodents, crystals of, by
Halliburton, 181
Bourne, G. C., on the anatomy of
Mussa and Euphyllia, and on the
‘morphology of the Madreporarian
skeleton, 21
= on Kleinenberg’s re-
searches on the development of
Lopadorhynchus, 531

Carpenter, Herbert, on the develop-
ment of the apical plates of Am-
phiura squamata, 303

Cunningham and Vallentin on photo-
spheria of Nyctiphanes, 319

Cunningham on some points in the
anatomy of Polycheta, 239

Cynthia, anatomy of, by L. Sheldon,
131

SERIES.

Dendy on the sponges Ridleia and
Quasillina, 513

Earthworms, numerous nepbridia in
same segment in certain species of,
397

Echinoderm morphology, No. XI, by
Herbert Carpenter, 303

Egg of some osseous fishes, by Scharff,
53

Kuphyllia, anatomy of, by G. C.
Bourne, 21

Fishes, intra-ovarian eggs of, by
Scharff, 53

Fowler on anatomy of Madreporaria,
Part III, 1; Part 1V, 413

Hemoglobin erystals of Rodents’
bload, by Halliburton, 181
Halliburton on methemoglobin erys-
tals, 201
= on the hemoglobin erys-
tals of Rodents’ blood, 181
Haswell on Temnocephala, 279

Kleinenberg’s researches on the de-
velopment of Lopadorhynchus, 531

Labyrinthine apparatus of fishes, by
Zograff, 501

Lopadorhynchus, development of, 531

Madreporaria, anatomy of. by G. H.
Fowler, Part II], 1; PartIV, 413

548

Madreporarian skeleton, Bourne on, 21

Marshall, C. F., on striped and un-
striped muscle, 75

Mephitis mephitica, tongue of,
by Tuckerman, 149

Methzmoglobin, easy preparation of
crystals, by Halliburton, 201

Muscle-plate and development of
spinal nerves, by Paterson, 109

Muscle, striped and unstriped, by C.
F. Marshall, 75

Mussa, anatemy of, by G. C. Bourne,
21

Nephridia, numerous, in same seg-
ment of earthworms, by Beddard,
397

Nyctiphanes, photospheria of, by Val-
lentin and Cunningham, 319

Paterson on the fate of the muscle-
plate and the development of the
spinal nerves, 109

Peripatus, a monograph of the species
of, by Adam Sedgwick, 431

3 anatomical differences in
species of, by Lilian Sheldon, 495

- developmeut of the Cape
species of, Part IV, by Sedgwick,
373

= development of a South
American species, by Sclater, 343

5 nove-zealandiae, de-
velopment of, by Lilian Sheldon,
205

Photospheria of Nyctiphanes, by Val-
lentin and Cunningham, 319

INDEX.

Polycheta, anatomy of, by Cunning-
ham, 239

Quasillina, a genus of sponges, by
Dendy, 513

Quadrate of Mammalia, by Baur, 169

Ridleia, a new genus of sponges, by
Dendy, 513

Scharff on the intra-ovarian egg of
some osseous fishes, 53
Sclater, W. L., on the development
of a South American species of
Peripatus, 343
Sedgwick, a monograph of the species
of the genus Peripatus, 431
33 on the development of the
Cape species of Peripatus, Part
IV, 373
Sheldon, L., on anatomical differences
in species of Peripatus, 495
on the ciliated pit of As-
cidians, 131
2 on the development of
Peripatus nove -zealandiz,
205
Spinal nerves, development of, by Pa-
terson, 109

Temnocephala, by Haswell, 279

Tongue of Mephitis mephitica,
by Tuckerman, 149

Tuckerman on the tongue of Mephi-
tis mephitica, 149

Vallentin and Cunningham on photo-
spheria of Nyctiphanes, 319

Zograff on the labyrinthine apparatus
of the Labyrinthic fishes, 501

PRINTED BY ADLARD AND SON, BARTHOLOMEW CLOSE.

ih iy ni weg ae Oh ete ee a ek A!
ih i, y. Me an he D
ne .

en
ei M ‘Sr

t
56+
——- #
:)

Epi

m
GG, OT, Cees one we
ae

“ty ye
naderroes

a

Laelel

ey
*

‘ave

%
*

to]
~

4

ae

BINDING =r" °°. vue renm

QH Quarterly journal of micros-
201 copical science

Biological

& Medical-
S

Serials

PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET

UNIVERSITY OF TORONTO LIBRARY