HARVARD: UNIVERSITY.

LIBRARY

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY.
) oar
Hot

ete 7. [qoi— pul q, 1god.

i : me ‘eg | |
se " ayt, Ali . Bink
5 = .>

ce
ae e )

QUARTERLY JOURNAL

MICROSCOPICAL SCIENCE.

EDITED BY

Kk. RAY LANKESTER, M.A., LO.D., F.R.S.,

HONORARY FELLOW OF EXETER COLLEGE, OXFORD} CORRRSPONDENT OF THE INSTITUTE OF FRANCE,
AND OF THE IMPERIAL ACADEMY OF SCIENCES OF ST. PETERSBURG, AND OF THE ACADEMY
OF SCIENCES OF PHILADELPHIA; FOREIGN MEMBER OF THE ROYAL BOHEMIAN

SOCIETY OF SCIENCES, AND OF THE ACADEMY OF THE LINCEI OF ROME;

AND OF THK AMERICAN ACADEMY OF ARTS AND SCIENCES OF BOSTON

ASSOCIATE OF THE ROYAL ACADEMY OF BELGIUM; HONORARY MEMBER

OF THE NEW YORK ACADEMY OF SCIENCES, AND OF THE
CAMBRIDGE PHILOSOPHICAL SOCIETY, AND OF THE ROYAL
PHYSICAL SOCIETY OF EDINBURGH ; HONORARY
MEMBER OF THE BIOLOGICAL SOCIFTY
OF PARIS;
DIRECTOR OF THE NATURAL HISTORY DEPARTMENTS OF THE BRITISH MUSRUM; LATE FULLERIAN

PROFESSOR OF PHYSIOLOGY IN THR ROYAL INSTITUTION OF GREAT BRITAIN-

WITH THE CO-OPERATION OF

ADAM SEDGWICK, M.A., F.RS.,

FELLOW AND TUTOR OF TRINITY COLLEGE, CAMBRIDGE 3

W. F. R. WELDON, M.A., F.BS.,

LINACKK PROFESSOR OF COMPARATIVE ANATOMY AND FELLOW OF MERTON COLLEGE, OXFORD
LATR FELLOW OF ST. JOHN’S COLLEGE, CAMBRIDGE}
AND

SYDNEY J. HICKSON, M.A., F.RS.,

BREYER PROFESSOR OF ZOOLOGY IN fHE OWENS COLLEGE, MANCHESTER,

VOLUME 45.—New Senrtes.
With Aithographic Plates and Engrabings on Wood.

LONDON:
J. & A. CHURCHILL, 7, GREAT MARLBOROUGH STREET.
1902.

CONTENTS:

CONTENTS OF No. 177, N.S., SEPTEMBER, 1901.

MEMOIRS:
PAGE
The Development of Lepidosiren paradoxa.—Part If. Witha

Note upon the Corresponding Stages in the Development of
Protopterus annectens. By Grawam Kerr. (With Plates
teeny ; ye

On the Malayan Species of Onychophora —Part II. The Develop-
ment of Eoperipatus Weldoni. By Ricwarp Evans, M.A.,
B.Se., of Jesus College, Oxford. (With Plates 5—9) . : 41

CONTENTS OF No. 178, N.S., NOVEMBER, 1901.
MEMOIRS :

The Lateral Sensory Canals, the Kye-Muscles, and the Peripheral
Distribution of certain of the Cranial Nerves of Mustelus !
levis. By Epwarp Puetrs Atus, jun. (With Plates 10—12) $7

The Anatomy of Scalibregma inflatum, Rathke. By J. H.
AsuwortH, D.Se. (With Plates 13—15) : : 2 23%

On the Pelvic Girdle and Fin of Eusthenopteron. By Epw1n
S. Goopricu, M.A., Fellow of Merton College, Oxford. (With
Plate 16). ; : : : : ; + oT

1V CONTENTS.

CONTENTS OF No. 179, N.S., FEBRUARY, 1902.

MEMOIRS: PAGE
Dendrocometes paradoxus.—Part I. Conjugation. By
Sypney J. Hickson, M.A., D.Sc., F.R.S., Beyer Professor of
Zoology in the Owens College, Manchester; assisted by Mr.

J.T. WapswortH. (With Plates 17 and 18) . : . 325
On the Oviparous Species of Onychophora. By ArTtHuR DENDY,
D.Sce., F.L.S., Professor of Biology in the Canterbury College,

University of New Zealand. (With Plates 19—22) . . 363
A New and Annectant ‘Type of ee By R. I. Pocock.
(With Plate 23) . ‘ : : ; . 417

The Trypanosoma Brucii, the canis found in Nagana, or
Tsetse Fly Disease. By J. R. Braprorp, F.R.S., and H. G.
Pummer, F.L.S. (from the Laboratory of the Brown Institu-
tion). (With Plates 24 and 25) : : ‘ . 449

Notes on Actinotrocha. By K. Ramunni Menoy, Assistant Pro-
fessor, Presidency College, Madras. (With Plate 26) . . 473

Review of Mr. Iwaji Ikeda’s Observations on the Development,
Structure, and Metamorphosis of Actinotrocha . : . 485

CONTENTS OF No. 180, N.S., MARCH, 1902.

MEMOIRS :

On the Structure of the Excretory Organs of Amphioxus.—Part I.
By Epwin 8S. Goopricn, M.A., Fellow of Merton College, Oxford,
(With Plate 27) . : ‘ ; ; , . 493

A Contribution to the Morphology of the Teleostean Head Skeleton,
based upon a Study of the Developing Skull of the 'Three-spined
Stickleback (Gasterosteus aculeatus). By H. H. Swinner-

TON, B.Sc., from the Zoological Laboratory, Royal College of
Science, London. (With Plates 28—31 and 5 Text Illustrations) 503

The Development of Admetus pumilio, Koch: a Contribution to
the Embryology of the Pedipalps. By L.H. Govan. (With
Plates 32 and 33) ‘ : ; x ‘ > 595

On the Teeth of Petromyzon and Myxine. By Ernest Warren,
D.Se., Assistant’ Professor of Zoology, University College,
London. (With Plate 34) ‘ : : . = 63m

Typhlorhynehusnanus, a New Rhabdoceele. By F. F. Larptaw,
B.A. (With Plate 35) . : ‘ : : . 637

OCT

“I

10n]

SEPTEMBER, 1901.
\SX

THE

QUARTERLY JOURNAL

OF

MICROSCOPICAL SCIENCE.

EDITED BY

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

HONORARY FELLOW OF EXETER COLLEGE, OXFORD; CORRESPONDENT OF THE INSTITUTE OF FRANCE
AND OF THE IMPERIAL ACADEMY OF SCIENCES OF ST. PETERSBURG, AND OF THE ACADEMY
OF SCIENCES OF PHILADELPHIA; FOREIGN MEMBER OF THE ROYAL BOHEMIAN
SOCIETY OF SCIENCES, AND OF THE ACADEMY OF THE ILINCEI OF ROME;
ASSOCIATE OF THE ROYAL ACADEMY OF BELGIUM, HONORARY MEMBER
OF THE NEW YORK ACADEMY OF SCIENCES, AND OF THE
CAMBRIDGE PHILOSOPHICAL SOCIETY, AND OF THE ROYAL
PHYSICAL SOCIETY OF EDINBURGH ; HONORARY
MEMBER OF THE BIOLOGICAL SOCIETY

OF PARIS;
DIRECTOR OF THE NATURAL HISTORY DEPARTMENTS OF THE BRITISH MUSEUM, LATE FULLERIAN
PROFESSOR OF PHYSIOLOGY IN THE ROYAL INSTITUTION OF GREAT BRITAIN.

WITH THE CO-OPERATION OF

ADAM SEDGWICK, M.A., F.RS.,

FELLOW AND TUTOR OF TRINITY COLLEGE, CAMBRIDGE}

W. F. R. WELDON, M.A., F.RS.,

LINACKE PROFESSOR OF COMPARATIVE ANATOMY AND FELLOW OF MERTON COLLEGE, OXFORD;
LATE FELLOW OF ST. JOHN’S COLLEGE, CAMBRIDGE 3

AND

SYDNEY J. HICKSON, M.A., F.RS.,

BEYER PROFFSSOR OF ZOOLOGY IN THE OWENS COLLEGE, MANCHESTER.

WITH LITHOGRAPHIC PLATES AND ENGRAVINGS ON WOOD.

J. & A. CHURCHILL, 7, GREAT MARLBOROUGH STREET.
1901.

New Series, No, 177 (Vol. 45, Part 1). Price 10s.

Adlard and Son,] (Bartholomew Close.

CONTENTS OF No. 177,—New Series.

MEMOIRS:

The Development of Lepidosiren paradoxa.—Part II. With a
Note upon the Corresponding Stages in the Development of
Protopterus annectens. By Granam Kerr. (With Plates
1—4) .

On the Malayan Species of Onychophora. Part II. The Development
of Koperipatus weldoni. By Ricnarp Evans, M.A., B.Sc., of
Jesus College, Oxford. (With Plates 5—9) .

PAGE

4]

Ger 721901

The Development of Lepidosiren paradoxa.
Part II.

With a Note upon the Corresponding Stages in the Development
of Protopterus annectens.

By

J. Graham Kerr.

With Plates 1—4.

Conrents.

PAGE
Introduction. : ; ‘ ; : . 1
Methods : : , é : : 3

Karly Stages in Develdome nt:
Segmentation and Origin of Segmentation Cavity. ‘ 6
Gastrulation 3 : : : ; 10
Fate of the Segmentation Cavity : ‘ : é 16
Origin of Mesoblast and Notochord . : : : 17
Origin of Colom : : : : : 20
Karly Development. of Notochord : : 5 20
Origin of Central Nervous System. : : : 22
Note upon the Early Development of Protopterus : , 23
Note upon Size of Nuclei in Lepidosiren Egg : : ; 25
General Remarks upon the Phenomena described 2 ! : 25
Summary of Facts ; ! t ; ; . 37
Explanation of Plates . : : : : : 4 38

INTRODUCTION.

THE following pages constitute a further instalment of my
description of the developmental phenomena of Lepido-
siren paradoxa.! [I have devoted much time and

1 The first instalment, containing a description of the ‘External Features
in Development,” is to be found in ‘ Phil. Trans. Roy. Soe.,’ B, vol. excii, p.

VoL. 45, paRT 1.—NEW SERIES. A

2 J. GRAHAM KERR.

labour to making a very detailed investigation of the stages
in development here treated of, and I had originally intended
in my description to go into something lke corresponding
detail. I have, however, altered my original intention in
this respect, for various reasons: amongst others because
in the interpretation of minute details of early development
one is necessarily much influenced by preconceived ideas ;
and in the second place, because I find that these details
vary to an extraordinary extent in different eggs—some of
the variations being apparently due to variation in technical
methods of investigation, but many being certainly true in-
dividual variations. So potent are these disturbing factors
that I doubt very much whether a description going into very
minute detail must not necessarily be to a great extent mis-
leading, and so do harm. I therefore propose to limit myself
in regard to the early stages of development to the endeavour
to give an adequately complete general description of the
phenomena observed, with only so much detail as may seem
necessary to make the description clear.

The investigation of a holoblastic egg 7 mm. in diameter
and packed with yolk involves great technical difficulties, as
the whole of each egg has to be converted into thin sections.
The full extent of these difficulties will only be appreciated
by embryologists who have essayed a similar task. In order
to help future workers I devote a few paragraphs to a general
account of methods. ‘Then follows an account of the
phenomena observed, in which, as in my first paper, I
reserve remarks of a general nature embodying views rather
than facts for a concluding section, so that any reader may
obtain the facts, which are naturally of greater importance,
with a minimum of trouble.

299. As in that paper I naturally did not make precise statements regarding
the interpretation to be put upon surface features without having assured myself
first by the examination of sections that they were correct, it is unnecessary
for Prof. Semon to feel the doubts about the behaviour of the blastopore in
Lepidosiren which he expresses in his latest, contribution on the de-
velopment of Ceratodus (Semon, ‘ Zoologische Forschungsreisen,’ Band i,
S. 327).

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 3

In conclusion I have to record the gratitude which I owe
to my friend Mr. J. S. Budgett for the generous way in
which he has placed his store of Protopterus embryos at
my disposal. By his kindness I am able to interweave with
my description, references to what takes place in the only
other Dipneust still surviving, and consequently to greatly
increase its value.

MeEra#ops.

The eggs and Jarve on being brought in from the swamp
were first studied alive. For permanent preservation two
fluids were used—formalin and alcohol. Of the former
solutions in water of from 5 per cent. to 10 per cent. were
used, and I found formalin an admirable preservative for
the early stages. It caused practically no shrinkage either
of capsuleorembryo. It further left the former transparent
as in the fresh condition. The material of early stages fixed
and preserved in formalin was found to be in admirable con-
dition, both as to fixation and as to consistency for section
work. This, however, only applies to the early and heavily
yolk-laden stages.

The alcohol material was fixed in a variety of ways.
Practically all the ordinary fixing agents were tried, but the
best all-round results were obtained by corrosive sublimate
and acetic acid, and Flemming’s chrom-aceto-osmic solution
(strong formula). Perenyi’s solution proved to be unreliable.

For section cutting, after many weeks of failure, the fol-
lowing three stock methods were adopted :

I. Thick sections of early eggs, where the cell elements
were very large, were cut with a “Jung” microtome after
soaking for three days in thin celloidin, three days in thick
celloidin, and thirty minutes in chloroform, followed by
treatment with cedar oil until clear. The block was kept
saturated with cedar oil, and the sections were transferred in
order to a shallow tray containing the same fluid. The
sections were then arranged upon strips of tissue-paper
3 inches by 1 inch within a space equal to the size of the
cover-slip used. ‘The paper strips with the sections lying on

4 J. GRAHAM KERR.

them were now laid in a bath of absolute alcohol, to remove
the cedar oil, and then taken up and laid sections downwards
upon slides coated with a layer of dry collodion. A finger
was now passed lightly along the paper, giving a gentle
pressure, just sufficient to cause the celloidin of the sections
to adhere to the collodion on the slide. The slide was now
removed to 90 per cent. alcohol and the ordinary process of
staining carried out. In the subsequent dehydration pre-
vious to mounting a mixture of chloroform and absolute
alcohol was used for the final stage of the process. ©

II. To obtain thin sections of yolk-laden eggs, it was
necessary to embed in both celloidin and paraffin. The pre-
liminary embedding in celloidin was done as before. The
egg was taken from celloidin solution and dropped bodily
into chloroform for 15—80 minutes. At first I was in the
habit of transferring the celloidin block to cedar oil before
embedding in paraffin, but latterly I have frequently em-
bedded at once by the chloroform-paraffin method. It is of
great importance to keep the temperature of the water-bath
as low as possible, and also to diminish the length of time
during which the object is on the water-bath to the shortest
possible.

Sections were cut with a Cambridge rocking microtome,
and flattened with warm water on a slide coated with
glycerine and egg albumen. The water was drained off
and the slides put aside to dry in an atmosphere containing
vapour of alcohol and ether. It was found that drying in
an ordinary atmosphere over the water-bath caused the
celloidin-infiltrated section to dry, curl up, and break away
from the paraffin: this was avoided by drying in the manner
described. It is important, however, not to use an atmo-
sphere completely saturated with ether and alcohol vapour,
as this, by causing the celloidin to swell, may cause wrinkling
of the sections.

III. Older embryos were embedded in paraffin in the
ordinary way and cut with the rocking microtome.

Orientation.—For the accurate orientation of embryos

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 5

during the embedding process I use a special apparatus! in
which a pool of paraffin in contact with the block holder of
the microtome is kept melted by a small loop of platinum,
nickel, or other wire of high resistance and not easily
oxidisable, heated by the current from one or two ordinary
bichromate cells.

Staining.—After many trials of different staining fluids
I adopted two stock methods.

I. Harly eggs rich in yolk were stained in Griibler’s
“Safranin 0”—a saturated solution in absolute alcohol,
diluted with an equal volume of distilled water. In regard
to formalin eggs, difficulty was found in obtaining a good
chromatin differential stain. ‘his difficulty was completely
got over by treating the eggs with corrosive sublimate
solution for a couple of hours before transference to alcohol.

II. Later embryos were stained in Heidenhain’s iron
hematoxylin followed by faint staining with eosin. By this
stain beautiful preparations were obtained showing minute
nuclear detail to perfection.

Mounting Medium.—When sections of early eggs did
not stain successfully they were mounted in colophonium,
which on account of its lower refractive index shows up
feebly stained structures better than Canada balsam.

Reconstruction.—In working out the organogeny of
Lepidosiren I have found the following method of recon-
struction from serial sections extremely useful. Sections 10”
thick are drawn with the Abbe camera lucida at a magnifi-
cation of 100 diameters upon finely ground sheets of glass
1mm. in thickness. Sheets of glass bearing drawings of
consecutive sections are then piled in position on top of one
another, a fluid of the same refractive index as the glass
being run in between adjacent sheets. The result of this
is to convert the whole into a transparent block, in which
the structures drawn are seen occupying space of three
dimensions, forming a kind of model. Different organs are
drawn in different colours, lead pencil or coloured crayons

' Made for me by the Cambridge Scientific Instrument Company.

6 J. GRAHAM KERR.

(not anilins) being used. It is best, I find, only to do one
or two systems of organs at a time, the process being so
rapid compared to ordinary modelling by Born’s method
that it can easily be repeated if necessary. When I first
devised this method I used a chemical solution having the
exact refractive index of the glass, but latterly I have used
ordinary clove oil, which is near enough for practical purposes,
With clove oil ordinary water-colour pigments may be used.!

The above method is not meant to give a permanent model
of the structures investigated, as does the Born method of
reconstruction from sections ; but, on the other hand, it in-
volves far less expenditure of time, and is to be strongly
recommended for purposes of research. The main principle
of the method—the using sheets of glass or other transparent
plates on which to draw the consecutive sections—has been
used by other workers, e.g. Strasser and Dixon, and more
recently by Vosmaer. I have not, however, come across any
mention in literature of the two details upon which to my
mind the chief beauty of the method rests, viz. the using
sheets of ground glass to draw upon, and the subsequent
rendering these transparent by an interposed fluid of high
refractive index. The first of these details provides a par-
ticularly suitable surface upon which to draw; the second
gives a perfect transparency to the mass of superimposed
plates, quite unattainable where there are numerous alternat-
ing layers of substances differing so much in refractive index
as do glass and air.

Barty DEVELOPMENT oF LEPIDOSIREN.

Segmentation and Origin of Segmentation Cavity.
—A vertical section through a mature ege of Lepidosiren
shows that the interior is filled with a mass of yolk granules,
the protoplasmic substance between being so small in quan-
tity as to be quite invisible. ‘The yolk granules are rounded
or occasionally subangular in form. Through the greater

1 Mr. Budgett, who has been recently using my method of reconstruction,
strongly recommends the use of moist water-colours,

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 7
part of the eve there are large granules, measuring, as a rule,
between ‘015 mm. and ‘02 mm.! in diameter, and of the
characteristic salmon-pink colour, while the interstices
between these are filled with smaller granules. There is no
indication of a region of specially coarse-grained yolk in the
centre of the egg, but towards the surface of the ‘‘animal”’
portion the large granules are absent, and there is present a
superficial layer in which the yolk is entirely broken up into
very minute particles, whose innumerable reflecting surfaces
give to this part of the egg a snowy white appearance when
seen by incident light. In the middle of this cap of fine-
grained yolk lies the germinal vesicle, the details of whose
structure I have not been able to make ont satisfactorily.

As segmentation proceeds, the fine-grained yolk spreads
downward towards the centre of the egg—the smaller
blastomeres being distinguished by their fine-grained yolk
from the larger lower blastomeres, where the yolk remains in
large granules.”, Kven in this latter region, however, the
division planes become marked out by a septum of fine-
grained yolk.

As mentioned in my former paper, the segmentation
cavity begins to appear very early, in the form of chinks
between the micromeres. Inan egg of Stage 8% (PI. 1, fig. 1)
the cavity within the egg still remains in the form of such

1 Although the eggs laid by one female may be said to be on the whole
more coarsely grained than those of another, yet there is much variation even
amongst the eggs laid by a single female; e.g. in four eggs taken from one
nest the large yolk granules averaged ‘018, ‘018, ‘020, and ‘022 mm. in
diameter respectively; in three eggs taken from another nest the correspond-
ing dimensions were ‘015, ‘015, and ‘02 mm.

2 This statement must be taken as true only in a general sense; every now
and then one meets with a few coarse granules within the micromeres; while
in the region of the macromeres irregular patches of comparatively fine-
grained yolk frequently appear.

3 By ‘Stage 2” I mean an egg whose external features have reached the
stage of development represented by fig. 2 of my previous paper. At Prof.
Lankester’s suggestion I have had a figure (Text-fig. 1) prepared to illustrate
the chief stages, and so to obviate the necessity of frequent reference to the
plates of my previous paper,

J. GRAHAM KERR.

SS

MLL

Trext-FiG, |.

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA.

chinks. ‘The extent of these cavities varies consider-
ably in different eggs of the same age, the blastomeres in
some being more rounded, in others less rounded and more
flattened against one another. The more rounded condition
of the blastomeres in the former case does not appear to be
associated with the nuclei being in a state of karyokinetic
activity, as has been asserted to be the case in other forms.

Tpxt-rie. 1, illustrating the course of development of the Dipneumona.—
The stages are numbered in accordance with my earlier paper. Roman
numerals indicate figures of Protopterus (after Budgett, ‘Trans. Zool.
Soc.,’ vol. xvi). The remaining figures refer to Lepidosiren. In
figs. 16—24 tle embryo is for convenience shown spread out in one plane
and viewed from the dorsal aspect. The magnification is slightly over
two diameters. 47. Rudiment of external gills. c.o. Cement organ,
Ad. Rudiment of hind limb. px. Pronephros. 8. Egg during segmenta-
tion. 10. An early stage of invagination, the invagination groove stretch-
ing round about one third of the egg’s circumference. 13. A later stage
of invagination, the large yolk-ceils being now for the most part covered
in by small cells. xtr. Corresponding stage in Protopterus. 14. Ege
at; the close of invagination, showing the crescentic blastopore. 16.
Dorsal view of an embryo in which the medullary folds have just become
visible, diverging posteriorly to embrace the blastopore. 17, Later em-
bryo where the folds have met behind the blastopore, and are approximated
in the middle region of the embryo; the rudiment of the pronephros is
now visible as a slight bulging on either side. 19. The medullary folds
are nearly completely fused; the branchial rudiment is visible as a bulging
in front of the pronephros; indications of the myotomes are seen between
the pronephros aud the neural rudiment. 22. The branchial rudiment
has greatly increased in size, the optic rudiments are conspicuous, the
pronephric ducts have grown considerably backwards. 24. Embryo in
which the branchial rudiment has become completely segmented on the
right-hand side ; the central cavity of the neural rudiment has appeared
as a dark shadow. 25. Side view of a slightly older embryo in its
natural position on the egg; the rudiments of the four external gills now
form distinct projections; the rudiment of the cement organ has appeared
ventrally. xxxv. Corresponding embryo of Protopterus. 28. Larva
three days after hatching. 31. Larva (thirteen days) in which the ex-
ternal gills have become pinnate, and the rudiments of the limbs have
appeared (anterior hidden by external gills). 35. Larva with external
gills at their maximum ; the cement organ, now in course of atrophy, is
seen beneath the throat. xxxv. Corresponding larva of Protopterus.
36. Young Lepidosiren with external gills in process of atrophy.

10 J. GRAHAM KERR.

As already mentioned, the yolk in the micromeres is reduced
to the condition of fine granules. These also become reduced
in number, and the nucleus tends to be surrounded by an
area of finely granular reticular protoplasm, almost free
from yolk granules. The transition from the finely granular
micromeres to the coarsely yolked macromeres is perfectly
gradual.

Between Stages 8 and 9 there appears an irregular chink of
larger size than the others amongst the lower micromeres
(Pl. 1, fig. 2). This, the definite segmentation cavity, increases
in size, spreading laterally, and at the same time approach-
ing close to the upper surface of the egg, being eventually
covered in by a roof of comparatively regular thickness
throughout. This roof soon becomes composed of two regular
layers of cells (figs. 3 and 4). As the segmentation cavity
further increases in size these become flattened out, until
the roof forms a thin translucent membrane through which
in the entire egg the segmentation cavity appears as a dark
shadow. The characters of the completed blastula may be
sufficiently gathered from fig. 4.

The blastomeres on the floor and sides of the segmentation
cavity are rounded, almost spherical in form, and project into
the cavity. Usually, some of these spherical blastomeres
appear to float quite free in the fluid of the segmentation
cavity. ‘I'his appearance does not of course prove that they
are not really connected up to the other blastomeres by
delicate protoplasmic strands; but such connecting threads
if present are too delicate to be seen by ordinary observation.

Gastrulation.—The process of gastrulation in Lepido-
siren may for convenience of description be divided into
three periods, which I will call 4, 8, and c.

A. In this period, which marks the beginning of gastrula-
tion, we have to do with a process of true invagination. ‘The
commencement of this process is indicated, as I have shown
in my previous paper, by the appearance of a row of little
depressions of the egg’s surface arranged in a latitudinal
direction a few degrees below the equator. These depres-

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 11

sions soon become joined up to form a continuous groove
stretching through about one third of the circumference of
the egg at this latitude (cf. Text-fig. 1, fig. 10). A section
through the whole egg at this stage is given in fig. 6 (Pl. 2),
and sections through the groove itself under a higher magni-
fication in figs. 5 and 7 (PI. 1).

In the cells lining the groove much of the yolk has passed
into a state of fine subdivision, thus pointing to cell activity.
From the open character of the groove during this stage it is
obvious that we have to do with a process of true invagina-
tion. In some series of sections one can see very well (fig. 7)
how the groove, although to the naked eye apparently coin-
cident with the boundary between the cells with small and
those with large yolk granules, lies really just within the
region of the latter. The invagination in Lepidosiren is
thus essentially a lower cell phenomenon.

The groove, as mentioned in my previous paper, gradually
becomes more limited in extent by its lateral portions be-
coming flattened out. Had it extended at any period of
ontogeny completely round the exposed area of large cells,
we should have been able to speak of a yolk-plug. As it is
probable that the disappearance of a yolk-plug bounded all
round by an invagination groove is due to increase in size
and richness of yolk in the egg, I had hoped to find it present
in Protopterus, In this I have been disappointed, the
condition in this respect being just as in Lepidosiren.

While the lateral parts of the groove flatten out and
disappear, the middle part is deepening to form the arch-
enteric cavity.

B. ‘The exact method by which this takes place in its
earlier stages forms a problem of considerable general
interest, but at the same time one the final solution of which
is attended with great difficulties.

The appearance of sections during this period is illustrated
by fig. 8 (Pl. 2). The archenteric cavity runs obliquely in-
wards from the surface of the egg, and at its inner end turns
upwards so as to run roughly parallel to the surface. The

12 J. GRAHAM KERR.

whole cavity is slit-like in form and is eminently suggestive
of having been formed by a process of splitting amongst the
large yolk-cells, after the manner described by Robinson and
Assheton in the case of the frog. Further evidence is
afforded in this direction by the fact that many sections
show the archenteric slit to end in a perfectly sharp edge
(fig. 9), which may even appear to be prolonged by division
planes along which the cells have not yet separated. Had I
had to rely upon a small amount of material, I should almost
certainly have described the archenteric formation during
this stage as being carried on by a process of splittmg. I
have, however, examined now a very large number of series
of sections, and I am disposed to think that the process is by
no means one of simple splitting. In the first place, by
looking through complete series of sections, one as a rule
finds that, in certain sections, the archenteron terminates in
a clear rounded end (fig. 10). It appears impossible te me to
imagine that this can occur if the cavity is only extending by
a splitting process. Further, it is usual to find that, round
the tip of the archenteron, the cells have assumed a triangular
shape in section, with their tips towards the archenteron,
which strongly suggests the existence of a compressing force
acting round the tip of the cavity, and of such a nature as
would be caused by growth of either roof or floor of the
cavity. On the whole, I conclude that increase of the arch-
enteric cavity does take place mainly by true invagination
during this stage also. ‘The slit-lke appearance in many
sections may conceivably be an artefact due to the roof of
the archenteron being squeezed down against its floor by the
action of the fixing agent, or possibly the process of invagina-
tion may be aided by one of splitting. ‘There seems nothing
improbable and indeed little of importance in this, notwith-
standing how much has been written on the subject. If it
does occur it is only another example of a very common
phenomenon in yolky eggs,—the formation by splitting of a
cavity elsewhere formed by invagination.

As regards the probable cause of the invagination—beyond

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 13

the use of the vague phrase “ differential growth ”—nothing
can be said. The absorption of the fluid in the segmentation
cavity which is associated by Samassa with the invaginatory
process of Amphioxus is excluded as an explanation of the
phenomenon here, as the first obvious result of such absorp-
tion would be the collapse of the very thin and delicate roof
of the segmentation cavity, and such collapse is conspicuously
absent.

In transverse section the archenteron is seen to be, in this
stage, a tube rounded in section—in other words, showing no
signs of splitting laterally, and about *2 mm. in diameter,
strikingly narrow in proportion to the diameter of the egg as
compared with most holoblastic forms.

Towards theend of period 8 the archenteron approaches
the margin of the segmentation cavity, and now we have very
distinct evidence that the growth of the archenteron is not
due to splitting, for the cells round its tip become pushed
definitely into the segmentation cavity forming a rounded
bulging into it (Pl. 2, fig. 8). As the process goes on the
large-yolk cells become laid up against the original roof of the
segmentation cavity, which, already two-layered, alters little
in character and will later become definitive epiblast. The
further stages in the obliteration of the segmentation cavity
I will deal with later.

c. In the later stages of gastrulation we have certainly to
do with a process of true invagination, the end of the archen-
teron being always quite smooth and rounded, with cuticular
lining, and there being never any trace whatever of sphtting
(cf. Pl. 3, figs. ll and 12). The precise character of this inva-
gination could only be settled definitely by experiment upon
the living egg, and such experiments, though attempted,
proved absolutely fruitless on account of the tough egg
capsule and the soft nature of the egg contents. From the
study of sections! of the eggs I am disposed to believe that

1 In my account of the external features I pointed out that against the

probability of such a backgrowth taking place, was the fact of the blastoporic
lip not assuming the form of an are of gradually diminishing radius with its

14 jJ. GRAHAM KERR.

we have to do with an invagination of the large yolk-cells of
the lower lip of the blastopore by the upper dorsal lip grow-
ing bodily down over them. The evidence upon which this
belief rests is as follows :

(a) A sagittal section through an egg of this stage fixed in
such a way as to avoid shrinkage of the capsule is shown in
outline in Text-fig. 2. It is obvious that the general outline
of the section suggests strongly that the dorsal lip of the

Text-ric. 2.—Camera outline of sagittal section through an egg in its cap-
sule at a late stage of gastrulation. The lines O A, OB, and OC are
drawn from the centre of the section so as to touch respectively the tip
of the archenteron (O A), the edge of the small-celled area (O B), and
the dorsal lip of the blastopore (O C).

concave side downwards. As a matter of fact this objection is done away
with by the fact that in Protopterus frequently the lip does become concave
downwards just as we should expect (cf. a forthcoming paper by Mr. Budgett
in ‘Trans. Zool. Soc. Lond.,’ vol. xvi). The blastoporie lip becoming convex
downwards in. Lepidosiren I attribute now to the backgrowth being more
active in the middle line than laterally.

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 15

blastopore is growing bodily downwards, wedging itself in
between the capsule and the large yolk-cells, and causing
as it does so the latter to invaginate into the floor of the
archenteron.

(3) The frequency of mitotic figures in the region over-
lying the archenteron, and more especially in the dorsal lip,
appear to indicate active growth of this region, and conse-
quent backward migration of the blastoporic lip.

(y) During the later stages of gastrulation I find that the
angle between the lines O A and O B (passing from the centre
of the section to the tip of the archenteron and to the margin
of the small-celled area respectively) remains nearly con-
stant, and the increase in the angle A O C corresponds fairly
closely with the diminution in the angle C O B.

This seems to suggest that the line O C is gradually swing-
ing through the arc between A and B. Otherwise we must
believe that the lines O B and O A are swinging with equal
velocity in a clockwise direction. It appears to me from
study of my sections that this is not the case, the forward
movement of the point B being very slow compared with the
advance of the archenteric tip.

(0) The cells of the ventral wall of the archenteron are
continuous, without any visible change in character, with the
large yolk-cells lying exposed on the outer surface of the egg
below the blastopore.

On the whole, then, I believe that the evidence, such as
it 1s, points to the view that the main factor of the increase
in length of the archenteron during this last stage is the
downgrowth of the blastoporic lip.

While these processes of formation of the archenteron
have been going on the area of yolk-cells exposed has been
gradually reduced, dorsally by the growth of the blasto-
poric lip, elsewhere by the gradual encroachment of the
small-celled area. This spreading of the smal] cell margin
over the yolk-cells is most rapid in the neighbourhood of the
blastoporic lip, least so at the point opposite to this. In this
latter region the superficial layer of small cells passes into a

16 J. GRAHAM KERR.

thickened rim, which at first I called the growing edge of
the epiblast. Further investigation showed, however, that
the chief characteristic of this rim is not its growth, which
is comparatively small, but the fact that it represents the
mass of small cells on which the roof of the segmentation
cavity rested at its margin. The thin two-layered epiblast,
in fact, from this rim for a considerable distance is nothing
else than the persistent roof of the segmentation cavity.
This is shown to be the case by the fact that within a short
distance of the rim one frequently finds the small blastomeres
beneath the epiblast retaining their rounded form with
chinks between, or we may even find the segmentation cavity
still present as a continuous slit.

What spreading of small cells over the large yolk-cells does
take place is brought about by addition to the margin of
small cells cut off from the yolk. This is well shown by
sections such as that in fig. 15, where there can be no
question of true epibole or sliding of the small cell layer
over the surface of the yolk-cells.1

The slight extent of the movement over the yolk of the
small-celled margin at the point opposite the blastopore rim
is of importance as providing a nearly fixed point in the in-
terpretation of sagittal sections. ‘The evidence of these sec-
tions is, on the whole, that the dorsal roof of the archenteron
is formed mainly by backgrowth of the dorsal lip, and as the
medullary plate at its first appearance is practically coin-
cident with the extent of the archenteron, Lepidosiren
is brought into line with the Selachians, where almost the
whole of what is commonly called the “embryo” is formed
from a similar backgrowth.

The Segmentation Cavity.—I now return to the con-
sideration of the segmentation cavity, which was left at a
period when it was beginning to be encroached upon by
the bulging wall of the archenteron. ‘The further oblitera-
tion of the segmentation cavity, although it takes place

' By an error the word ‘ epibole’”’ was used in my former paper at one
place (p. 822) when delamination was actually meant.

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 1

during the process of gastrulation, does not by any means
keep time with the latter—a further support to my asser-
tion that the former is not the direct cause of the latter.

What takes place may be said to be in general terms that
the floor of the segmentation cavity is brought up against
its roof. During this process, however, a transient phase
occurs which is not without interest. While the cavity is
still at its full development we notice a tendency for large-
yolk blastomeres to become arranged round the segmentation
cavity, and in close contact with its roof (cf. Pl. 2, fig. 8;
or better, figure of Protopterus VIII); following this,
the smaller blastomeres lying in and near the floor of the
cavity push out processes, become irregular and angular in
shape, and, attaching themselves to one another by their
corners, form a loose and irregular sponge-work traversing
the cavity completely (Pl. 3, fig. 11). As will be gathered
from the figures, the segmentation cavity during this pro-
cess, although broken up by the sponge-work, really extends
through a much larger volume than it did before. As, how-
ever, gastrulation proceeds, the fluid filling the meshes of
the sponge-work becomes absorbed, and the blastomeres
resume their spherical or, as they become pressed closer
together, polyhedral shape. We may still for a long time,
however, observe chinks persisting here and there, especially
laterally. The roof cells of the segmentation cavity remain
all through the stages we are now describing sharply marked
off from the large yolked elements which have been laid up
against them.

Origin of the Mesoblast and Notochord.—Pl. 3,
fig. 14, illustrates a section through an egg of Stage 12 and
transverse to the axis of the medullary plate region. Lying
over the archenteron and tapering off on each side is a mass
of cells distinguished from the remainder of the inner cells
by their smaller size, more finely granular yolk, and by their
rounded form. Immediately over the archenteron these
small cells are aggregated closely together, but laterally as
a rule they are separated by wide chinks—the remains of the

VoL. 45, pART 1.—NEW SERIES, B

18 J. GRAHAM KERR.

segmentation cavity. At its outer edge this mass of small
cells passes gradually into the large inner cells. The sum of
small cells in question is the rudiment of notochord and meso-
blast. It is perfectly continuous across the middle line, and
is separated from the cavity of the archenteron by a definite
archenteric roof composed of cells closely fitted together.

The cells of the notochordal-mesoblastic rudiment are the
small blastomeres which are seen in earlier stages lying
below the floor and round the edges of the segmentation
cavity, or penetrating that cavity as a sponge-work.

A little later—in Stage 14 (cf. Text-fig. 1)—a transverse
section (Pl. 3, fig. 15) exhibits very similar features, only
now the mesoblastic cells are in close contact with one
another, and the mesoblastic rudiment is found to be grow-
ing at its edges by delamination from the underlying large
yolk-cells. The rate of this growth varies much. As a
rule, in an egg of Stage 14 the mesoblast extends very little
below the level of the archenteron on each side, though
in one case I found that it had grown right round the ventral
side of the egg. The process is in any case usually com-
pleted by Stage 18 or 20. For example, in an egg of Stage
18 I find that, although the actual splitting off of the
mesoblast has taken place only to a level slightly below
that of the archenteron, the superficial layer of yolk has
become fine-grained all round the egg, and here and there a
small mesoblast cell has separated off the large yolk-cells
beneath. Such mesoblast cells are often split off far beyond
the edge of the sheet of continuous mesoblast, so that when
I speak of the mesoderm spreading over the hypoblast I must
guard against giving the impression that the sheet is neces-
sarily continuous up toa definite margin. Finally, in egos of
Stage 21 the stratum containing fine-grained yolk has been
cut off the underlying hypoblast all over, as a definite layer,
somewhat irregular in places, of rounded mesoblast cells.

Where the mesoblastic rudiment has in its early stages
largely developed intercellular spaces, the boundary be-
tween it and the large yolk-cells is sharply marked very

= = ~~~

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 19

early (except in the middle line and at its outer margin).
Where the cells composing the rudiment are in close contact
the line of demarcation may be for a time indistinct. But

=

Trxt-ric. 3.—Section through a complete egg of stage transverse to axis of
embryo. ext. Enteron. m. Mesoblastic rudiment. m.p. Ectodermal
thickening of medullary plate. . Rudiment of notochord.

in any case by Stage 14 the mesoblastic rudiment on each
side becomes separated definitely from the underlying hypo-
blast (except at its outer edge), and a little later (PI. 4, fig. 16,
and Text-fig. 3) it becomes separated in a similar way from
the axial portion which will give rise to the notochord.
This latter remains in the meantime attached to the hypoblast.

It should be mentioned incidentally that the cells added
to the edge of the mesoderm sheet tend to take ona rounded

20 J. GRAHAM KERR.

form as soon as they become separated from the hypoblast.
It consequently often happens that, when the sheet is con-
tinuous up to its edge, this edge with its rounded cells is very
sharply marked off from the hypoblast beyond. With only
such sections to go by one might well believe that the sheet
of mesoblast was quite independent of the hypoblast, and
erowing inwards over its surface from the blastoporic rim
after the manner described for various forms by Lwoff,
Brauer, and others. It is at once seen from the study of a
complete series of stages, such as the above account is
based upon, that any appearance of the kind is quite
secondary, and that originally mesoblast and hypoblast rudi-
ments are perfectly continuous. I will return to this question
later on.

With the formation of the medullary keel the mesoderm
sheet becomes thickened out to each side of it in the region
where the myotomes are to be formed.

On account of the yolk-laden character of the mesoblastic
rudiment it is difficult to make out when its segmentation
begins. Distinct protovertebree were first found in about
Stage 17, where there were six present. They were squarish
in section and were solid.

Coelom.—The first parts of the ccelom to appear are
myoceelic. In Stage 21 (Pl. 4, fig. 21) a ccelomic cavity is
seen to have appeared in the centre of the myotome. This
appears to arise by simple breaking down of the central cells,
the cavity not having at first any sharply-marked outline,
and irregular masses of yolk-laden protoplasm projecting into
it. A little later (Stage 23) the outline is quite definite
and the cavity is walled in by a single layer of regular
columnar cells. From this the ccelom spreads outwards by
definite splitting.

Karly Development of Notochord.—The Notochordal
rudiment was left (PI. 4, fig. 16) at a stage in which it remains
attached to the hypoblast on the separation of the mesoblast
from iton each side. It formsa median dorsal ridge running
along the middle line above the archenteric cavity. The

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA,. 21

yolk in the cells of this ridge is usually in a state of com-
paratively fine subdivision, though much coarser than that
of the epiblast.

A set of division planes now become so arranged as to mark
off the notochordal part of the ridge from the comparatively
thin basal layer next the cavity of the archenteron (fig. 17,
e.r.). The cells of this latter frequently, though by no
means always, retain their yolk in a coarse-grained condition.

c,
4 !
. '

hen Sel

¢

Text-ric. 4.—Transverse section through dorsal region of an embryo of Stage
23. a.d. Archinephric duct. exf. Unteron. 4.m. Mesenchyme cells being
directly split off from hypoblast. 2. Notochord. scl. Sclerotome out-
growth from mesoblast.

They are part of the definitive hypoblast, and form the roof
of the enteron. The enteric roof is thus differentiated in
situ from the cells of the archenteric roof, without any trace
of ingrowth from the sides such as has been described by
Lwoff, Brauer, and others.

The notochordal rudiment thus laid down retains for some
time its comparatively undifferentiated condition (figs. 20
and 21), showing no obvious change beyond assuming a
rounder, more definite outline as it separates off the hypo-
blast. About Stage 23 the separation is completed, and the

22 J. GRAHAM KERR.

notochord, now circular in transverse section, develops a fine
cuticular membrane which foreshadows the sheath, and in
longitudinal section its cells are seen to be becoming flat and
plate-like.

In due course the notochord becomes separated off from
neighbouring structures by mesenchymatous tissue, partly
directly cut off the subchordal region of the hypoblast
(Text-fig. 4, h.m.), but for the most part arising by pro-
liferation from the inner surface of the mesoderm at about
the level of the nephric rudiment very much as in Selachians,
except that there is no obvious trace of a segmental arrange-
ment (Text-fig. 4, scl.). I propose to postpone further con-
sideration of the mesenchyme till a later period.

Origin of the Central Nervous System,—Already in
Stage 12, as has been mentioned (cf. Pl. 3, fig. 14), the
epiblast has become somewhat thickened over the region of
the archenteron, the thickening affecting the lower layer
especially whose cells have become more regularly columnar.
By Stage 14, when there runs forward from the blastopore a
faint depression along the axis of the medullary plate, this
thickening has become more marked, and in addition the
deep layer of epiblast is becoming more than one-layered (cf.
figs. 15 and 16). The medullary plate thickening of the
epiblast, most marked along the mid-dorsal line, extends
outwards for a considerable distance, gradually thinning
away on either side. The axial portion of the medullary
plate rapidly increases in thickness, forming a deep wedge-
shaped keel, the rudiment of the neural cord. This medul-
lary keel develops from before backwards, and in some eggs
of Stage 14 it has already begun to be distinctly formed
anteriorly. By Stage 16 (cf. Pl. 4, figs. 17 and 18), where the
medullary groove is well formed but widely open, the keel
has increased much in thickness, being about five cells thick
posteriorly, and thickening out anteriorly to about three
times as much. Just about the anterior limit of the archen-
teron the keel tapers off, first suddenly, then gradually, till
the ordinary two-layered condition of the general ectoderm

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 23

is reached. The whole thickening of the keel is confined to
the deep layer of the ectoderm—the outer layer passing
unaffected over the floor of the groove. As the medullary
folds approach one another the groove shallows out and dis-
appears. Occasionally, in places, the folds come in contact
before the groove has disappeared, so that for a short time
they remain separated by a vertical chink (Pl. 4, fig. 19). As
before suggested, this may be looked on as a last trace of a
former method of formation of the spinal cord by involution,
but any trace of cavity that is so enclosed in Lepidosiren
is purely temporary and soon disappears. ‘The keel is now
(fig. 20) absolutely solid, and there is no indication of the
formation of a central canal until about Stage 20 (fig. 21),
when the cells of the interior of the neural rudiment are seen
to begin to assume a regular arrangement and columnar form
on each side of the median plane.

Along this plane the cells finally spht apart, apparently by
the secretion of fluid, the cavity in preserved specimens
showing an abundant coagulum. The split appears some-
what irregularly, but by Stage 23 it has become continuous,
forming a well-marked cavity in the region of the fourth
ventricle, and stretching back from this through about three
fourths of the extent of the neural rudiment. Anteriorly and
posteriorly the neural rudiment still is solid.

Nore vpon THE Harty DEVELOPMENT OF PROTOPTERUS.

The egg of Protopterus is much smaller than that of
Lepidosiren, measuring only about 3°5—4 mm. in diameter
(Budgett). Corresponding with this the yolk granules are
smaller, averaging about ‘(015 mm. by ‘01 mm. They have
alsoa characteristic difference in shape, being very frequently
lenticular or fusiform. The blastula of Protopterus differs
from that of Lepidosiren in the relatively greater depth
and volume of the segmentation cavity, and in the greater
relative extent of the micromeric region of the egg. ‘I'he
roof of the segmentation cavity is also thicker.

Gastrulation.—The line of invagination appears nearer

24 J. GRAHAM KERR.

the Jower pole of the egg than in Lepidosiren, about 30°
below the equator instead of about 10°. It is consequently
visible from the beginning when the egg is viewed from the
lower pole, forming part of the circumference of the small
circle bounded by the edge of the small-celled area. The
condition is exactly as in a typical Urodele’ or Anuran egg,
only here the groove never extends round the whole circle to
enclose a definite yolk-plug, but, as in Le pidosiren, shortens
up, flattening out at each end. ‘The examination of sections
shows that here as in Lepidosiren the invagination groove
is at its first appearance distinctly within the coarsely-yolked
portion of the egg.

The general features of gastrulation closely resemble those
in Lepidosiren, and it is therefore not necessary to describe
them in detail. I give, however, figures illustrating three
successive stages (PI. 2, figs. vi, vi; and Pl. 3, fig. xm). By
comparison of fig. vr with fig. vi, the vertical axis being
marked by the position of the segmentation cavity, it will
be readily seen how important a part is played by overgrowth
of the blastopore lip. The orientation of the egg during
these stages is rendered simpler than it is in Lepidosiren
by the segmentation cavity retaining its original relations
much longer.

At the close of gastrulation the appearance of the egg is
practically identical with that of Lepidosiren. I notice,
however, that very frequently an egg of Protopterus at this
stage assumes an ellipsoidal form, with the blastopore either
at one end or somewhat ventral to this. In Lepidosiren
only pathological or unfertilised eggs assume an ellipsoidal
shape.

As regards the further points of development treated of in
this paper, there do not appear to be any noteworthy differ-
ences between what occurs in Protopterus and what has
been described for Lepidosiren.

1 The Protopterus egg very frequently passes through a stage identical
in appearance with the stage in the development of Triton figured by O.
llertwig in ‘Jen. Zeits.,’ Bd. xv, Taf. xii, fig. 1.

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 25

Size of Nuclei during Harly Stages of Develop-
ment of Lepidosiren.—Owing to the small scale of the
figures it is not possible to indicate the relative sizes of the
nuclei in different parts of the egg. These bear, as one.
might expect, a rough relationship to the volume of the cell
territories over which they preside; e. g. in two eggs of
Stage 16 the nuclei of the ectoderm averaged ‘016 mm. and
‘014 mm. in diameter, those of the mesoderm ‘018 mm. and
‘016 mm., and those of the large yolk-cells ‘(022 mm. and
‘021 mm. Again, in an egg of Stage 13 the nuclei in the
region of the dorsal lip of the blastopore measured ‘015 mm.,
and those of the large yolk-cells ‘(019 mm.

The measurements are in all cases the average of ten
measurements of whole nuclei as seen in thick sections.

GENERAL Remarks.

Segmentation.—In studying the segmentation of Lepi-
dosiren Ihave beeu much struck by the readiness with which
all trace of the division planes may be destroyed in the parts
of the egg filled with large yolk-granules. The two com-
monest causes of this are, firstly, the use of a fixing agent of
inferior penetrating power, the blastomeres running together
into a continuous mass very soon after death if the fixing
agent has not reached them; and secondly, the use of two
thin sections. In cutting a section it would appear that the
yolk-granules become very slightly displaced as they strike
the edge of the knife, and if the section is very thin this is
enough to completely obliterate the division planes. During
segmentation in Lepidosiren thick sections will show an egg
to be completely divided up into blastomeres, while in thinner
sections the whole lower portion with coarsely-grained yolk
seems to form a quite continuous unsegmented mass. ‘The
mass of uncleaved yolk figured by Semon in the middle of the
Ceratodus egg, and upon which he bases the statement that
this egg in its early stages of segmentation occupies a place
intermediate between the telolecithal and centrolecithal types,
may, I think, quite possibly be an artefact of this nature, due

26 J. GRAHAM KERR.

to the fixing agent not having penetrated sufficiently rapidly ;
and it also seems by no means impossible that the lower part
of the egg of Gymnophiona may be only apparently uncleaved
for the same reason.

Segmentation Cavity.—The segmentation cavity of
Lepidosiren arises in the normal fashion from intercellular
chinks. Amia, whose segmentation otherwise so resembles
that of Lepidosiren, is said to develop its segmentation
cavity from intra-cellular spaces (Whitman and Eycleshymer’):

The mode of disappearance of the segmentation cavity, the
blastomeres permeating it as a sponge-work, and then later
rounding themselves off so as to leave the diminishing cavity
in the form of chinks between them, resembles closely what
occurs in Petromyzon as described by Nuel. It may
quite possibly occur pretty generally, as in Lepidosiren
this stage lasts such a short time that it might easily be
missed.

The two-layered character of the roof of the cavity from
an early stage is noteworthy. The roof, in fact, has taken on
its definitive epiblastic character already in the blastula
stage. In Ceratodus the roof is one-layered; and in other
cases where it is two or three layers thick, it is usual for a
one-layered condition to be passed through before it becomes
definite epiblast (Petromyzon, Axolotl, Gymnophiona).

Blastoporie Lip Downgrowth.—In Amphioxus it
has been shown that the blastopore occupies the hind end of
the embryo. So it is with Lepidosiren, so that we may
reasonably compare embryos of the two forms at the close
of gastrulation.

It is commonly said that in a heavily yolked egg the
macromeric part has become too bulky to allow of invagina-
tion. ‘This is true only in a restricted sense, there not being
room for the macromeric portion to be pushed bodily within
the other as in Amphioxus. In such a form as Lepido-
siren, however, new space is continually being provided by

1 ¢J. Morphol.,’ vol. xii, p. 336.
2 ‘Arch, Biol.,’ t. ii, p. 436.

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 27

the continued increase in area of the small-celled outer layer
of the egg due to the backgrowth of the upper lip, and
under this invagination goes on in the ordinary way. This
is, it appears to me, the real significance of the backgrowth.
It is a phenomenon directly associated with the increase in
bulk of the macromeres. If this were true, we should find it
become more and more pronounced as a developmental
feature with increase in the quantity of yolk. This is, I
think, what we do find, and we can also understand on this
view why recent observers have failed to find such a process
taking place in Amphioxus.

I do not propose to enter at length into the controversy
which has raged over the parts played by invagination, split-
ting, downgrowth of dorsal lip, etc., in the gastrulation of
vertebrates. Much of the evidence that has been brought
seems to me unreliable, resting as it does on such characters
as size of cells, size of yolk-granules, presence of pigment—
characters which appear to me to be in great part merely the
expression of greater or less metabolic activity for the time
being, and which cannot therefore safely be used as criteria
in treating of morphological questions.

Apart from these, the evidence afforded by the study of
sections is of such a character that its interpretation is liable
to be seriously affected by the observer’s preconceived ideas.
As regards observations on the living egg, many of the
methods also seem open to the influence of very serious
disturbing factors, either of a traumatic nature or of a
simple physical character, such as movement of the egg as
a whole, brought about by shifting of the centre of gravity
due to the change in the relative extent and position of
archenteric and segmentation cavities. The only really
reliable method of investigation appears to be that of
Kopsch,! where the developing egg is submitted to pro-
longed photographic exposures, and the surface-cell move-
ments worked out on the pictures so obtained.

1 “Verh. Anat. Ges.,’ 1895, p. 181; and ‘S. B. Ges. naturf. Freunde
Berlin,’ 1895, p. 21.

28 J. GRAHAM KERR.

My own conclusions with regard to the part played by back-
ward movement of the blastopore lip agree closely with
those reached by Kopsch for Amphibia, and my support of
his views is strengthened by the fact that I had not seen his
paper until I had finished my observations of the phenomenon
in Lepidosiren.

As will have been gathered from the descriptive part of this
paper, I am strongly of opinion that in Lepidosiren the
main factor in the formation of the archenteron is a process
of invagination. J am not at present, however, prepared to
deny that during what I have called Stage B of gastrula-
tion this process may not be aided to some extent by
splitting.

Communication between Archenteron, and Seg-
mentation Cavity.—The view expressed by Kupffer in
1879,! that the enteron is formed by a fusion of the two |
originally separate cavities—archenteron and segmentation
cavity—has recently been supported for the large eggs of
Salamandra maculosa and Gymnophiona. It will be
seen from figs. 8 and 11 how thin is the septum separating
these cavities, and how easily they might be thrown into one
by rupture of the intervening wall. In one or two eggs I
have found this happen. I attribute it to the fixing fluid
not having penetrated properly; but whether this be so, or
whether it really existed in the living egg, it is In any case
quite abnormal in Lepidosiren, and in all except these
few exceptional cases the two cavities remain completely
shut off.

Formation of Parts of Archenteric Roof from
“Hetodermic” or “Animal” Cells.—In the preceding
description I have made no statements regarding ingrowth
of ectodermic or animal cells along the roof of the arch-
enteron. Assertions that this occurs in other forms seem to
me to be weakened by two fallacies. In the statement by
Lwoff, Brauer, and others that the plate above the arch-
enteron which gives rise to chorda and mesoderm is ecto-

1 * Zool. Anzeiger,’ ii, p. 594.

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 29

dermic, there appears to me to lurk a confusion of ideas
between the two pairs of antithetical terms—ectoderm and
entoderm (or epiblast and hypoblast), and micromeres and
macromeres (or animal cells and vegetable cells). The latter
pair of terms are purely descriptive, and may be applied to
blastomeres at once upon the evidence of an isolated observa-
tion. The former, on the other hand, are terms associated
with definite theory; they are not to be applied on mere
observations of sizes and shapes of cells, but involve the fate
of the cells. It seems to me quite impossible to define a
layer as hypoblastic except by asking one or other of the
two questions: (1) Does it form the lining of an archenteric
cavity ? and (2) Does it become a certain part of the definitive
epithelial lining of the gut? And if during the early stages
of development a certain set of cells become invaginated
along a considerable extent of the archenteric roof, this
seems to me in itself amply sufficient reason for calling such
cells hypoblastic quite apart from what their special cha-
racters of size, shape, content, and so on may be. There is
no justification at all that I can see for calling the small fine
yolked cells towards the upper pole of the egg epiblast, and
on their extension in along the archenteric wall to found the
statement that “ectoderm becomes invaginated.” Because
these cells behave as they do they are not ectoderm, but
entoderm.

Further, a main character upon which these cells along the
archenteric roof are relegated to the category of “ ecto-
derm” or “animal cells” is the finely granular character of
their yolk. Size of contained yolk granules is a form of
evidence which must be used with the greatest caution, for
wherever metabolism is active there the large yolk granules
are broken down into fine granules to facilitate assimilation.
All the yolk is destined to be so broken down eventually, and
the fact of its having done so in some particular part of the
egg earlier than elsewhere seems to indicate merely that
metabolism is there more active. Consequently I can attach
little weight to statements on the morphological nature of

30 J. GRAHAM KERR.

particular cells based on the finely granular character of the
yolk. I should attach much greater weight to the presence
of large granules of yolk in cells, for when the yolk is in
this form in a developing embryo it seems usually to indicate
that it has remained so all through, it being at least very
unusual for yolk to be secondarily built up again into large
granules during embryonic development.

The finely granular character of the yolk frequently shown
by the roof, as compared with that of the floor of the arch-
enteron, I would look upon then as being merely a necessary
accompaniment of the active growth of this region asso-
ciated with the backgrowth of the blastopore lip.

What I have said regarding the unreliability of evidence
of the morphological nature of cells from the finely granular
character of their yolk contents apphes equally well to the
presence of black pigment in cells. I believe it to be one of
the most general reactions to light stimulus for active but
unspecialised cells to have their metabolism so affected as to
cause the formation of this particular product.! Examples
are seen in the case of comparatively undifferentiated cells
wandering into a position where they are subjected to light
stimulus, e. @. to the surface of the body, or into the vicinity
of a special light-collecting organ (e.g. Arthropod eye).
Where pigment occurs in the smaller cells of a frog’s egg it
is, I think, to be correlated simply with the more active
metabolism going on in these cells, and it is rather the
absence of pigment in special cases which demands explana-
tion; in many cases this may be due to natural selection—as
in the case of eggs which are laid in a floating mass of white
foam, where their being black would render them extremely
conspicuous.”

1 Which once produced may well be made use of as a protective agent for
neighbouring tissues against the harmful influence of light rays.

? With the criticisms in the foregoing paragraphs are to be associated those
on similar lines of Houssay (‘ Arch. Zool. exp.,’ 2nd sér., t. viii) and Samassa
(‘Verh. Deutsch. Zool. Gesell.,’ Strasbourg, p. 189; also ‘Arch. Entw.
Mech.,’ Bd, ii and Bad. vii).

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 31

Formation of Secondary Enteric Roof.—Brauer has
described in Gymnophiona the formation of the definitive
enteric roof by a backgrowth of “vegetative cells” under the
original archenteric roof. In Lepidosiren no such back-
growth takes place. It is to be noted, however, that there is
much variation in the character of the yolk granules in the
cells lining the archenteric roof immediately ventral to the
chorda. Most usually these cells have fine granules, but very
frequently, on the other hand, they are distinctly marked off
from the chorda cells by their yolk remaining in much coarser
granules (cf. Pl. 4, fig. 17). With only scanty material, in
which the later stages happened to show this difference, one
might well imagine it due to a secondary growth of large
yolked cells in beneath the chorda rudiment. In view of
this possibility of erroneous interpretation of sections I do
not feel absolutely convinced that such a backward growth
of large yolk-cells under the original cells of the archenteric
roof as has been described by Brauer and also by Lwoff
actually takes place.

In regard to Brauer’s observations I might add that in
my personal opinion the large-grained character of the cells
figured by him as growing backwards makes it unlikely that
they are multiplying with the activity which would be
necessary on his view.

In regard to Brauer’s fig. 59,' where the large yolked cells
extend right to the blastopore, it is of importance to note that
the author expressly states that it is not a median section.
In Lepidosiren it is only the median part of the archen-
teric roof that is fine-grained.

In Ceratodus Semon has described the roof of the
enteron as being formed by an ingrowth from each side
under the chorda rudiment. ‘There is no such ingrowth in
Lepidosiren. Where it does occur it may be looked on
as a cenogenetic modification bearing the same relation to
the method of chorda formation in Amphioxus, as the
method of separation of the neural rudiment from the

' Zool. Jahrb, Anat.,’ Bd. x, Taf. xxxvii,

32 J. GRAHAM KERR.

ectoderm in Amphioxus does to the method occurring more
usually by the formation of aneural groove. The method of
chorda formation found in Lepidosiren may be compared,
on the other hand, with the modification of the development
of the neural rudiment occurring in Teleosts.

Growth of Hpiblast.—In that the epiblast grows at its
edge by delamination, Lepidosiren agrees with what has
been found in various Amphibia (Houssay, Robinson and
Assheton, Grénroos), but differs from what has been found
by Brauer in Gymnophiona.

Mesoblast Formation.—In regard to the development
of mesoblast there are two features of special interest. The
first of these is the fact which cannot, I think, be doubted
by anyone in Lepidosiren—that the so-called “ gastral”
mesoderm is formed directly out of the smaller blastomeres
on each side of the archenteron. ‘These masses are con- |
nected across the middle line, and the common rudiment of
mesoderm and chorda is quite continuous. To go further
than this and say, as has been done by others, that the
notochord is derived from mesoderm, is quite unwarranted.

I do not see any possible phylogenetic explanation of this
phase in the formation of the mesoderm.

The later phase in its development which is of interest is
that in which we see the mesoderm asa sheet upon each side,
segmented or not according to its age, free at its inner
thicker edge next the chorda, and thinning away to become
continuous with the large yolk-cells or primitive hypoblast at
its outer edge, where it continues to grow by delamination.
Here we have a condition of things upon which I think a ray
of light is thrown if we regard it as a fleeting reminiscence
of the primitive method of mesoderm development in the
Chordata.

_ The phenomena, in fact, in Lepidosiren, closely paralleled
by those in Petromyzon, suggest a scheme of the steps by
which the method of mesoderm formation in the higher ver-
tebrates may have been derived from that found in Am-
phioxus, differing somewhat from that due to Hertwig.

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 33

Figs. 101 and 102 in Hertwig’s ‘ Lehrbuch’! are sufficient
to illustrate his view of the derivation of the mesodermic
rudiments in the higher vertebrates from the enteroccelic
pouches of Amphioxus. This view, as is well known,
rests mainly on Hertwig’s observations on the development
of Triton, in which he found pouches of the archenteric
cavity projecting on each side of the notochord into the
mesoblastic rudiment, which pouches he interpreted as
vestiges of the original communications between the archen-
teron and the cavity of enteroccelic mesoderm pouches like
those of Amphioxus. These observations of Hertwig
appear to have failed to find adequate confirmation, and it
seems to me that a scheme such as that represented below
fits in better with the general facts of vertebrate develop-
ment. Such a scheme, as will be seen, agrees in general
principle with the theory suggested by Lankester and de-
veloped especially by O. Hertwig, that the mesodermal
rudiments on each side of the vertebrate embryo represent
the walls of the enteroccelic pouches of Amphioxus; it
differs from the Hertwig development of the theory in the
detail that it regards the continuity often found in verte-
brate embryos between mesoderm and hypoblast on each side
of the notochord (and necessarily also the similar continuity
between mesoderm rudiment and notochord) as being a
secondary fusion rather than as representing the original
connection of mesodermic diverticulum with wall of the
archenteron.

The adjoining figures (Text-fig. 5) represent transverse
sections through the embryos of Amphioxus, Petromyzon,
Lepidosiren, and chick. In the case of the last three I
have, for convenience, represented only a small portion of
the whole section. As will be seen, the condition in P etro-
myzon agrees very closely with that in Amphioxus, and
is immediately derivable from it by reduction in the size of
the archenteric cavity.

The disappearance of the cavity of the enteric diverticulum

1 6te Auflage.
VoL, 45, pARY 1,—NbW SERIES, c

34 J. GRAHAM KERR.

seems to me of no special weight ; it is merely an additional
example of a very common phenomenon, of the fact that
hollow organs, formed primitively by involution of a cell-

Trxt-Fic. 5.—Transverse sections through embryos of various vertebrates to
illustrate the formation of mesoblast—A. Amphioxus. B. Petro-
myzon. OC. Lepidosiren. D. Sauropsida. ex¢. Enteron. m.
Mesoderm rudiment. . Notochordal rudiment. 2.7. Neural rudiment.

layer, tend, where the cells are burdened with yolk, to
arise from a solid rudiment, and to develop their cavity
secondarily.

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 30

Passing on to Lepidosiren, the difference between it
and Petromyzon is seen to be quite insignificant, consisting,
in fact, only of difference in relative dimensions.

Finally, the condition of the mesoblast in one of the higher
Vertebrata, as indicated in fig. D, seems to me to hang on
equally well to the earlier members of the series. What dif-
ference there is is merely difference in shape and relative
size. I hold, then, that in the series of Vertebrata there
exist passing phases in the development of the mesoblast
which may be readily linked on to one another, and that the
existence of these phases may be accounted for by regarding
them as reminiscences of phylogenetic stages in the modifica-
tion of the process of mesoblast development.!

Conclusion.—In general the phenomena described in
this paper fully bear out what I referred to in my earlier
communication—the extreme resemblance with corresponding
features in the Urodela. As regards external features
during the earlier periods of development this likeness is
perhaps slightly less marked in Lepidosiren than in Pro-
topterus, but as regards internal features of segmentation
and gastrulation the most remarkable resemblance is seen.

The resemblance with Petromyzon is equally striking,
and that with Ganoids only slightly less so. I do not feel
it necessary to go into detail in this matter; it will only
be necessary for the reader to turn to such figures as
Houssay’s? pl. xi, fig. 26 (transverse section of Axolotl,
showing early stage of mesoblast); Calberla’s* fig. 7
(similar section through Petromyzon); Eycleshymer’s#
pl. xx, fig. 8 (external view during early invagination of

1 It will be noticed that, on the above hypothesis, the growth of the meso-
blast at its outer side by continued delamination from the hypoblast would
correspond to a continued deepening of the groove between mesoblastic and
chordal rudiments of Amphioxus, and is therefore easily understood.

Were Hertwig’s scheme the true one this growth of the mesoblast would be
quite incomprehensible.

2 *Arch. Zool. exp.,’ 2nd série, tome viii.
3 *Morph. Jahrb.’ ili, Taf. xii.

4 «J. Morphol.,’ vol. x.

36 J. GRAHAM KERR.

egg of Rana palustris), or Dean’s! pl. iv, fig. 62 (Aci-
penser, longitudinal section of egg during gastrulation),
to see the remarkable unity which runs through these dif-
ferent types. Of the figures which I happen to have men-
tioned, the first three might have been used to illustrate the
corresponding stages of Lepidosiren or Protopterus
almost as well as the figures which I have given.

Looking at the broad facts in these three groups, and com-
paring them with what occur in other vertebrates, one cannot
but be struck with the fact that in the only two groups in
which it is almost certain that we have to do with poorly-
yolked eggs in forms descended from richly-yolked ones, viz.
the Teleostei and the higher Mammalia, we find that in neither
has the process of gastrulation reverted to its original cha-
racter. Rather by its profound modification from the normal
type it bears witness to the changes which have taken place
in its history. This being so, the comparatively simple type
of gastrulation similar in Petromyzonts, Amphibia, Dipnoi,
and Ganoids cannot but weigh strongly as evidence against
the view propounded by Rabl, that any of these groups
are descended from ancestors with large, richly-yolked
meroblastic eggs.

There is one point which I should lke, in conclusion, to
draw attention to, and that is the shunting forwards in
development of the rudiments of organ systems to an earlier
period than that to which they normally belong. Thus by
Stage 14, when gastrulation is just completed, the study of
sections teaches us that the embryo is already a compli-
cated triploblastic organism, with definite mesoblast and
chorda.

IT have obtained the small results recorded above only by
prolonged work upon a most extensive material preserved
with the greatest care and by the most approved methods.
I have been greatly impressed by the variability observed
amongst embryos of similar stages in development, much of
it probably natural, much of it certainly due to differences

1 *J, Morphol.,’ vol. xi,

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 37)

in methods of preservation, section cutting, etc.; so much
so that my final description varies in many important respects
from the rough draft made on a preliminary study of a few
embryos. My experience convinces me of the futility of
trying to give a fair description of the embryology of any
type unless one has a very large material to go upon. Much
of the discussion, involving often flat contradiction of dis-
tinguished observers’ statements, which is constantly taking
place appears to me to have a very probable cause in the small
amount of material which has been made use of.

SuMMARY OF THE MORE Important New Facts.

1. The segmentation cavity arises in Lepidosiren from
intercellular chinks.

2. The roof of the segmentation cavity early becomes two-
layered, and assumes the character of epiblast.

3. Gastrulation takes place, for the most part, by a true
invaginatory process.

4, Spreading of small cells over large takes place by de-
lamination, there being no true epibole.

5. The disappearance of the segmentation cavity is in-
augurated by its penetration by a sponge-work of small
blastomeres from its floor and sides.

6. The notochordal and mesodermal rudiments are at first
quite continuous across the middle line.

7. The notochordal rudiment remains attached to the
hypoblast for some time after the mesoderm has separated
off on each side.

8. The enteric roof is formed in situ directly from the
archenteric roof.

9. The mesoderm grows outwards on each side by delami-
nation from the large yolk-cells.

10. The myoccele arises by breaking down of cells in the
middle of the myotome.

11. Later on the myotome wall is composed of a single
layer of regular columnar cells.

38 J. GRAHAM KERR.

12. The first-formed mesenchyme arises from sclerotomic
outgrowths, assisted by proliferation to a slight extent from
the subchordal hypoblast.

18. The solid neural keel arises by thickening of the deep
layer of the epiblast.

14. The egg of Protopterus closely resembles in its
early development that of Lepidosiren.

15. The roof of the segmentation cavity is, however,
thicker.

16. And the invagination groove appears about 20° nearer
the lower pole of the blastula.

17. The early development of Lepidosiren and Proto-
pterus shows an extraordinarily close resemblance to that
of Amphibia Urodela; a close resemblance to that of Petro-
myzon; and an only slightly less close resemblance to that
of Ganoids.

EXPLANATION OF PLATES 1—4,

Illustrating Mr. Graham Kerr’s paper on “The Development
of Lepidosiren paradoxa,” Part II.

As already pointed out, by ‘‘ Stage x” I mean the stage represented by fig.
a of my previous paper on the “ External Features in Development.’ For the
convenience of readers of the present paper I have copied and selected a
number of these figures in Text-fig. 1 (p. 8). This will, I hope, obviate the
necessity on the part of the reader of having to frequently refer to a separate
publication.

Fic. 1.—Vertical section of egg of Stage 8, showing chinks between the
micromeres. IV B 681.

Fie. 2.—Vertical section through egg, showing the commencing formation
of the definite segmentation cavity, s.c. VII A 151.

Fie. 3.—Vertical section through the upper part of an egg slightly older

than the last. The segmentation cavity has here begun to spread laterally.
XXIV A 21.

THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 39

Fic. 4.—Vertical section through an egg, showing the segmentation cavity
at its full development. V 221.

Fie. 5.—Vertical section through the groove of invagination just after its
first appearance. (Stage 10) y 511.

Vie. 6.—Sagittal section through a very slightly more advanced egg. a.
Spherical blastomeres round floor of segmentation cavity. 7g. Invagination
groove. s.c. Segmentation cavity. Va 371.

Fig. vt.—Similar section through egg of Protopterus. D 192.

Fig. 7.—Part of the section drawn in Fig. 6 under a higher power, show-
ing the groove to lie within the region of coarsely-yolked elements.

Fic. 8.—Sagittal section of an egg during a later stage of gastrulation,
showing folding up of coarsely-yolked elements against the roof of the seg-
mentation cavity. VIC 222.

Fig. viut.—Corresponding section through egg of Protopterus.
A 411.

Figs. 9 and 10.—Small portions of two sagittal sections through an egg of
similar age to the last. The sections show the tip of the archenteron; the
section drawn in Fig. 9 favouring the idea of ‘‘ splitting,” that shown in
Fig. 10 negativing it. VI4 111 and V10 221.

Fig. 11.—Sagittal section through an egg of Stage 12 to show the penc-
tration of the segmentation cavity by a continuous sponge-work preparatory
to its obliteration. 3* H 232.

Fic. 12.—Sagittal section through an egg of Stage 13 in which the seg-
mentation cavity has become completely obliterated. xx D 262.

Fig. x11.—Corresponding section through egg of Protopterus.
C 263.

Fig. 13.—Portion of a similar section to that in Fig. 11, to show the
characters of the small cell margin. 3* D 211.

Fig. 14. Stage 12. R 441.

Fie. 15.—Stage 14. 7* 551.

Fic. 16.—Stage 14. XXXVIIC 542.

Fie. 17.—Stage 16. XXXVII E531.

Fig. 18.—Stage 17. XXXIII 632.

Fic. 20.—Stage 21. XXXIV B 531.

Fic. 21.—Stage 21. XXXIV C 482.

These figures form a series meant to illustrate the gradual differentia-
tion of the mesodermal, notochordal, and other rudiments. ¢.c. Indica-
tion of split to form central canal. e. Epiblast. ex¢. Enteron. e.7.
Elements of enteric roof, here with coarsely-grained yolk. 4. Hypoblast.
m. Mesoblast. m.e. Thickened ectoderm of medullary plate. — m.g.

4,0 J. GRAHAM KERR.

Medullary groove. m.#. Medullary keel. m.s. Spinal cord. myoe.
Myocele. 2. Notochord. p.z. Pronephros,
Fic. 19.—Transverse section through neural rudiment. Stage 18.
v.c. involution of outer surface of ectoderm to form a vestigial neural
canal.

all _ : LF . rs
: ; ‘
]
a)
rs ]
a :
7 As]
oi |
. Th
i) & _
e
| : ,
_ /
oe : 7 ;
\ ; .
1

F i yf sd « .
i) =
o : e. 7 :
gee, => 7 A , 4
2 ~ Samhita, *% "
. 7 ‘
_ a ; ‘
- Oo - — ~ es Ss)

Mor Sci. Vel 25S GL 1

iS)

Wilson, Cambridge.

E

& f
€ /
7.) fy
E
—E
—o7
ene
| E
l=
ori
nu
1m
| ; iS
|
|
Ww
| —
7M
cS
4 |=)
a}
| (ep)
| .
i )
™m
Le
4
fo)
lo =
xD >
ro) =)
te >
wy mn
Rie er
ue i
a
,

LG 2

ns

7

& Fon Mir Set. WEIS

i lar ie ial é ¢
Cluart. Foun. Mor Sct, Vl GEMS IE 3

3
ee
ee

mm.
I
2
=)
i
0 0

Fig xii Figs. 13-15

E Wilson. Cambridge

Figs 16,17,18,20,21. 2

Fig (oe : : e at

THE MALAYAN SPECIES OF ONYCHOPHORA,

On the Malayan Species of Onychophora.

Part Il—The Development of Eoperipatus weldoni.

By

Richard Evans, M.A., B.Sc.,
Of Jesus College, Oxford.

With Plates 5—9.

ConvEN'Ts.

I. Introduction
Il. The Ovum
III. A General Account of the Develonmentt viewed ertentally
IV. The Development of the Germ Layers, ete.
(1) The First Embryo ;
(2) The Second Embryo
(3) The Third Embryo
(4) The Fourth Embryo
V. The Development of the Mesodermal Ohenis
(1) The Development of the Mesoderm and its Garities
(2) The Development and Disappearance of the First Somite
(3) The Development and Disappearance of the Second Somite
(4) The Development of the Third Somite
(5) The Development of the Generative Organs
(6) The Development of the Last Somite (Male aeieeon
Gland) F
VI. The Development of the Nervous Spatem, and Ventral Orsi
VII. The Development of the Eye
VILL. The Endoderm ;
Conclusion
List of References
Explanation of Plates .

41

42 RICHARD EVANS.

I. Inrropvucrion.

So much good work has been already done on the develop-
ment of the Peripatide that it is necessary to justify the
publication of another account. However, it is not difficult
to do so, for not only is the development of the Malayan
species still unknown, but even their very existence has been
recently doubted. ‘Their close anatomical relation to the neo-
tropical forms renders an account of their development doubly
interesting and highly desirable. For these reasons it is
proposed to give a fairly complete account, and to pay special
attention to those points which, hitherto, have not been suffi-
ciently elucidated, and are still in dispute. It is not possible
to give an account of the segmentation stages, owing to the
material not being sufficiently well preserved. The ova being
full of yolk, the younger stages in the development must be
taken out of the uteri to preserve them properly; and even
when this precaution has been taken, owing to the presence
of a thick egg-shell, it will be difficult to ensure good preser-

vation.

Il. Tae Ovum.

When the germinal cells first appear in the splanchnic wall
of the somites, they possess a highly granular nucleus,
without a nucleolus, and their cytoplasm is in no way different
from that of the remaining cells of the somite.

When the period of growth of the ovarian ovam commences
the nucleus enlarges, its chromatic granules become, relatively
to its size, less numerous, and are connected together through
the intermediation of fine threads, which take up the chromatic
stains. he nucleolus first appears, during the early stages
in the period of growth, as a small spherical body. At first
it presents no visible structure, but it soon becomes alveolar
(P1. 8, fig. 18a). The cytoplasm, which at first resembles that
of the surrounding cells, becomes clearly alveolar in character
and remarkably uniform in appearance throughout the ovum.
When the ovum has reached an intermediate stage in size,

THE MALAYAN SPECIES OF ONYCHOPHORA. 43

the small clear spaces or alveoli of the cytoplasmic network
of the previous stages become darker than the intervening
substance, a result probably brought about by the deposition
in them of fine granules of food-yolk (PI. 8, fig. 18¢c). As
the ovum increases in size the granules seem to run together,
and consequently to form larger bodies, which in many cases
appear to fuse so as to form structures which may be de-
scribed as systems of granules which present several centres
of deposition, as well as a common surrounding coat (PI. 8,
fig. 18 d).

The nucleus, at the commencement of the period of growth,
is situated at or near the centre of the cell, and presents a
regular oval outline; but towards the end of the period in
question it moves nearer the surface. While this transference
is being effected it presents an irregular outline, and seems
to influence the general arrangement of the yolk bodies
situated in its immediate vicinity (PI. 8, fig 18 d).

The fully grown ovum possesses a fairly thick coat, pre-
sumably a vitelline membrane, and is furthermore surrounded
by a layer of cells derived from the wall of the ovary: it is
suspended, by means of a cord of cells, in the body-cavity.
Dr. Willey used the term ‘‘ exogenous” to describe this method
of origin, in contrast to that found in the genus Pevipatus
in which the ova are formed “ endogenously ” (7).

The fully grown or mature ovum is oval in shape, provided
it has sufficient space to assume its proper form; otherwise,
under pressure exerted upon it by the neighbouring organs,
it may become quite irregular in outline.

The difference existing between the modes of origin and
structure of the ovum in the closely related genera Peri-
patus and Hoperipatus is worthy of note, and is probably
the main cause of their divergence in development.

Ill. A Genera, Account oF THE DEVELOPMENT VIEWED
EXTERNALLY.

In each uterus of Hoperipatus weldoni there may be as
many as a dozen embryos, ranging in development from

44, RICHARD EVANS.

the segmenting ovum to an individual which measures from
25 to 27 mm. in length, and is coloured almost like the
mother.

The description of the external features of the development
will be limited to a number of embryonic stages, which are
illustrated in the figures found on Pl. 5. The youngest
embryo successfully taken out of the uterus is illustrated by
the first figure on the above-mentioned plate. It seems to
represent an early gastrula stage, which is oval in shape and
provided with a slit-like blastopore possessing a somewhat
irregular outline. The yolk masses situated in the interior
are distinctly seen at the sides, but they are not so evident
on the ventral surface round the blastopore, owing to the
greater development of the germ layers in that position.
In both shape and size the embryo under consideration sub-
stantially resembles the ovum.

The next stage of development to be described is repre-
sented in the second figure on Pl. 5. Besides being very
different in shape, the embryo in question is actually shorter
than the ovum. On the anterior end there are two pairs of
thickenings situated one behind the other. The blastopore
has been divided into two parts, one of which is situated im-
mediately behind the first pair of thickenings mentioned above,
but owing to the yolk which protrudes out of it and covers a
considerable portion of the ventral surface, it cannot be seen
in an external view ; the other is placed further back, and may
be similarly filled with protruding yolk. The quantity of
external yolk present seems to be highly variable, and in some
cases it appears to be wanting. When there is a great mass
of external yolk spreading over the ventral surface, nothing
can be seen save a botryoidal appearance, produced by the
yolk embedded in a sparse reticulum of ectodermal cytoplasm.
The presence of external yolk and its variability are points
in which the development of Hoperipatus resembles that
of Peripatoides (6).

The posterior portion of the blastopore presents the appear-
ance of a square, the anterior side of which is absent. From

THE MALAYAN SPECIES OF ONYCHOPHORA. 45

in front it is being gradually encroached upon by the double
layer of cells which has grown across the middle portion of
the elongated opening found in the younger embryo. From
the middle point of the posterior border of the blastopore, the
primitive groove extends backwards for a considerable dis-
tance. The groove in question is not so evident in the
immediate neighbourhood of the blastopore as it is some
distance behind it. On each side of the groove and close to
the blastopore there is a triangular-shaped thickening, in
which very active proliferation is going on. These thickenings
constitute the so-called primitive streak, and are the sources
from which the mesoblastic bands are produced; in fact, they
may be described as teloblastic spots from which the
mesoderm is derived.

The next stage of development to be considered is repre-
sented by the third figure on Pl. 5. Unfortunately there
is a considerable gap between this stage and the previous
one. The posterior end of the embryo has grown round the
head, so that it almost touches the rudimentary antenne, which
at this stage consist of three rmgs, The rudiments of perhaps
all the appendages are visible; those of the jaws and oral
papillez being specially well developed in comparison with the
others, which decrease in size from before backwards. 'The
brain lobes constitute a marked feature of the embryo at this
stage. Neither the body nor the rudimentary appendages
exhibit ring-shaped markings.

The next stage to be described is illustrated by the fourth
figure on Pl. 5. The posterior end of the embryo under
consideration has grown over the head, and the antenne,
having increased in length, consist of about a dozen rings.
The brain lobes are enormously large in proportion to the
other parts of the body, and behind them are seen rows of
papille, which represent the rudiments of the lips. The
oral papillae present at their free ends a marked depression,
which represents the opening of: the future slime-gland.
Both the body and the appendages are provided with ring-
like markings.

46 RICHARD EVANS.

‘The next stage to be considered is represented by the fifth
figure on Pl. 5. The most marked change that has been
effected, as compared with the previous stage (shown in the
fourth figure of the same plate), consists in increase in size,
the embryo in its folded condition being two and a half
times as long as the previous one. The posterior end of
the embryo is passing from the strictly dorsal and median
position to the left side of the head preparatory to unfolding
itself. When the posterior end has slipped off the back it
may become slightly coiled, so as to produce a short spiral,
and, owing to both ends of the embryo simultaneously un-
twisting themselves in opposite direction, which sometimes
happens, the whole body may for a time present the appear-
ance of being spirally twisted. ‘he ring-like markings
occurring on the body and appendages are gradually becom-
ing deeper.

The sixth figure on Pl. 5 illustrates the next stage of
development to be studied. In the embryo under considera-
tion the posterior end displays a curious twist, which it has
acquired in passing from the left side and in becoming ex-
tended. The anterior end of the embryo is so placed that
the ventral surface is turned towards the right side, but
does not slip over to the side as the posterior end does. The
flexed anterior end, besides being twisted laterally, is un-
folding itself longitudinally ; for in the embryo under con-
sideration it carries only two pairs of the walking appendages,
while in the one described in connection with the previous
stage, and illustrated by the fifth figure on Pl. 5, it was
provided with four pairs. ‘The actual length of the embryo
in its folded condition was 7 mm.

The next stage of development to be considered is repre-
sented in the seventh figure on Pl. 5. With the exception
of the head, which is bent down towards the ventral sur-
face, the body of the embryo has attained the extended
condition of the older ones which occur in the uteri, and its
colour is just beginning to turn brown. Only the first pair
of walking appendages are in any way involved in the

THE MALAYAN SPECIES OF ONYCHOPHORA. 4.7

cephalic flexure. The actual length of the embryo in its
folded condition was 17 mm.

The eighth figure on PI. 5 illustrates the last stage in the
development. The embryo is fully extended and presents
nearly all the characters of the newly-born young. Even
the colour is not very different, and the great length of the
body is a most remarkable feature. In this latter respect it
surpasses the embryos of the genus Peripatus by several
millimetres. It may be worthy of note that the embryo
under consideration, which is a male specimen taken out of
the uterus of Hoperipatus weldoni, is five millimetres
longer than one of the male specimens of H. horsti obtained
from a dead tree trunk. Consequently there must be a con-
siderable difference in length between the embryos of E.
weldoni and of EH. horsti at the time of birth.

This concludes what I have to say on the embryonic stages
of Koperipatus weldoni, viewed externally. There is
nothing new or remarkable in the various phases of outward
form through which the embryo passes in the course of de-
velopment from the egg—heavily laden with yolk—to the
young just before birth. The early stages in the develop-
ment appear to be passed through very quickly, and the
changes which occur appear to consist in the development
and differentiation of parts at the expense of yolk stored up
in the egg, the actual increase in volume being very small.
It is not until the rudiments of all the most important organs
have been developed that any appreciable increase in size
takes place. Consequently the first five embryos in the
uterus, counting from the ovary, present the appearance of
being of the same size when examined through the uterine
wall. The sixth embryo, however, is considerably larger
than the fifth, the seventh than the sixth, and the difference
between any two successive embryos goes on increasing to
the end of the series.

The uterus may contain as many as a dozen embxyos, the
second, third, fourth, fifth, seventh, ninth, tenth, and twelfth of
which are respectively represented on Pl. 5, and illustrate
corresponding stages in the development.

48 RICHARD EVANS.

IV. Tue DevELOPMENT OF THE GERM LAYERS, ETC.

The First Embryo.—tThe sections obtained from the
specimen illustrated in the first figure on Pl. 5 were not
sufficiently good to admit of the structure being made out
with any degree of accuracy and certainty. For this reason
it is necessary to commence the description of the formation
of the germ-layers from a slightly older embryo, of which
four transverse sections are represented in fig. 9 (a, J,
and d) on Pl. 6.

In the embryo under consideration there is no external
yolk, and the blastopore is as yet undivided; but both the
endoderm and mesoderm are already in process of formation.
At this stage in the development there are no nuclei in the
centre of the yolk.

The ectoderm consists of a single layer of cells except in
front and at the sides of the blastopore on the ventral sur-
face where the nuclei are already arranged two deep (PI. 6,
fig. 9a). On the dorsal surface, especially towards the ante-
rior end, the ectodermal layer seems to be incomplete. The
ectodermal nuclei of the ventral surface are oval in shape,
and are arranged close to one another; while those situated
at the sides and on the dorsal surface are circular in shape,
and placed at greater distances from one another. The
doubling of the ectodermal layer, in front and at the sides
of the blastopore, seems to represent the first rudiments of
the nervous system, which is always developed from before
backwards, a fact which explains the greater condensation of
ectodermal nuclei in the position in question than elsewhere
(Pl. 6, fig. 9 a).

The mesoderm is already in process of formation. It is
derived from an area situated immediately behind the poste-
rior end of the blastopore (PI. 6, fig. 9d). The first somite,
already present, has not yet formed a cavity, though the
nuclei are arranged in a ring (PI. 6, fig. 9 ¢). In addition
to the first somite, the rudiments of the second and third

THE MALAYAN SPECIES OF ONYCHOPHORA. 49

have already appeared, but their nuclei present no particular
arrangement (PI. 6, fig. 9d), Even at this early stage in
the development the first somite, preparatory to advancing
along the side of the embryo towards its anterior end, is far
removed from the median plane,

The endoderm is also forming, especially towards the
posterior end of the embryo (PI. 6, fig. 9¢, en.). In front of
the blastopore there are no endodermal nuclei, but at its
sides a few have already appeared. ‘l'owards the posterior
end they are more numerous, and in the region in question
an occasional nucleus may be seen halfway up the sides
(comp. figs. 9a, 9c, 9d). There seems to be no doubt that
the endodermal.elements in Hoperipatus are derived from
the lips of the blastopore, with which the endodermal layer
is continuous, and that they pass from that position through
the outer layer of the yolk. While this process of invagina-
tion is going on, the layer of yolk in question, which is
being gradually invaded by the endodermal elements, loses:
its deutoplasmic character and becomes more cytoplasmic.
The central mass of yolk presents the appearance of contain-
ing compound systems as well as separate yolk bodies,
between which there seems to exist a certain amount of
cytoplasm. Simultaneously with the increase in number of
the endodermal elements the cytoplasm grows at the expense
of the deutoplasm.

Unfortunately the quantity of material at my disposal was
not enough to enable me to form very decided conclusions
on the formation of the blastula stage in Hoperipatus,
and I do not wish, in any way, to question the accuracy of
already published accounts of its formation in other genera
of the Peripatide ; but it is necessary to point out that the
incomplete condition of the ectodermal layer on the dorsal
surface of the very early gastrula stage, already described,
tends to show that in Hoperipatus the circurcrescence of
the yolk takes place after the same plan as in so many other
Arthropoda ; that is, by overgrowth from the future ventral
surface towards the dorsal. It is well known that the

you. 45, PART 1,—NEW SERIEs, D

50 RICHARD EVANS.

various genera of the Peripatide differ from one another
to a remarkable degree as regards the early stages in
their development, and it is quite possible that the method
by which the yolk becomes surrounded by the blastula
cells in the young embryo of Hoperipatus conforms more
closely to that existing in most Arthropoda than it does
to that occurring in the other genera of the Onychophora,

On this theory, Hoperipatus, which, from the point of view
of external characters and internal anatomy, seems to be
more primitive than the other genera belonging to the family
Peripatidee, would have to be considered more primitive, as
regards the mode of circumcrescence of the yolk, unless it
be admitted that the method in question has originated twice
within the limits of the Arthropodan phylum, a view which
is in no way probable. ; |

The mesoderm seems to be formed exclusively from the
primitive streak, but the endoderm develops from the HPs
of the blastopore by invagination.

The Second Embryo.—The next embryo which will be
considered corresponds to the one represented in the second
figure on Pl. 5, and is considerably more advanced than
that described in the foregoing pages, for it possesses the
rudiments at least of eight somites. |

The ectodermal layer is complete on the dorsal surface,
though thin, and possesses nuclei which appear either circular,
or oval in transverse section. When they are oval in shape
their long axis lies parallel to the surface. The ectoderm,
which covers the ventral surface, and is situated between the
two portions of the divided blastopore, has similar characters.
(Pl. 7, fig. 10 e). Elsewhere the ectodermal layer is con-
siderably thickened, and its nuclei are arranged three or four
deep. The thickening of the ectoderm is most marked. on
the cephalic lobes, where the nuclei are arranged in three or
four layers, and on the sides, where they are arranged in two
layers. ‘The cephalic lobes and the lateral bands, both pro-
duced by the thickening of the ectoderm, are continuous
with each other, and probably represent the undifferentiated

THE MALAYAN--SPECIES OF ONYCHOPHORA. 51

rudiments of the. appendages, the nervous system, and the
ventral organs. The upper moiety represents the rudiments
of the appendages, while the lower one, later on in the
development, splits into an inner and an outer portion, the
forerunners respectively of the nervous system and ventral
organs. Inthe region in front of the anterior portion of the
blastopore, the ectodermal thickening is continuous. across
the middle line (Pl. 6, figs. 10 a, 106, and 10c) ; but in the
mid-region of the body, that is in the region where the
blastopore has been obliterated, the two thickenings are
widely separated from each other (PI. 6, figs. 10 d, 10 e, and
10f); while behind the blastopore in the region of the primi-
tive streak they converge and fuse together (PI. 6, fig. 10,7).

The disposition of these ectodermal bands presents another
feature in which the development of Hoperipatus ap-
proaches that of P. nove-zealandiz (6).

The mesodermal bands, which originate from the triangu-
lar-shaped thickenings noticed in surface view, and situated
immediately behind the blastopore, are placed exactly under
the ectodermal bands above described (PI. 6, figs. 10 a—,),
At the sides they are broken up into several somites, which
decrease in size from before backwards. In the anterior
region, in which the ccelomic cavity has already appeared,
every successive somite overlaps the one in front of it (PI. 6,
fies. 10c, 10d, and 10e). The first somite has a well-
developed ccelom, and is situated in the latero-ventral aspect
of the cephalic lobes (PI. 6, figs. 10a, 10b, and 10c).. The
second somite is considerably smaller and overlaps the first
one (PL. 6, fig. 10c). The third somite is slightly smaller
than the second and passes forwards above it (Pl. 6, fig. 10 e).
The same is true of the fourth and fifth somites, the remain-
ing ones being so small in size as to be incapable of over-
lapping (PI. 6, fig. 109).

As soon as the ccelom begins to form the two walls of
the somite present distinctive characters. In the splanchnic
wall the nuclei are placed at a distance from one another,
and are flattened in the tangential plane ; but in the somatic

52 RICHARD EVANS.

wall, they are closely packed together and are oval in shape,
their long axis being directed at right angles to the surface
(Pl. 6, fig. 10 e, som.’ and som.*). This rapid differentiation
is prophetic of the changes which take place in the somatic
wall at an early stage in the development; that is, of the
formation of the myotome and the renal outgrowth.

In the embryo under consideration the endoderm is
present as a layer which completely surrounds the central
yolk. There are no cell outlines in it, and the nuclei are not
always spherical in shape. The endodermal layer, however,
is incomplete at two points, namely, the two portions of the
divided blastopore through which the yolk protrudes and
spreads itself over the ventral surface as the so-called external
yolk, The inner limit of the endoderm is quite distinct from
the central yolk, and at the edges of the blastopore it is con-
tinuous with the ectoderm. ‘The endodermal nuclei are not
situated in the outer layer of the yolk, but in a layer which
seems to be new, and entirely different. It is true that it
contains spherical masses, presumably stored-up food mate-
rial, but there are no compound systems among them, and
they stain much less readily. ‘hey probably consist of food
masses, which the endoderm itself has elaborated at the
expense of the central yolk, and stored up within its own
substance. The endodermal layer possesses another charac-
teristic which the central yolk completely lacks, namely, a
great multitude of small refringent granules, which were not
observed in the younger stages already described. The ex-
ternal yolk differs from the central yolk in that the masses
of food material contained in it are smaller, while the amount
of cytoplasm present is larger. It seems that the ectoderm
is capable of acting on the external yolk in the same way as
the endoderm does on the central yolk, and of building and
storing up food masses for future use. This is done chiefly
on the ventral surface, where the ectoderm and external
yolk are in contact.

In the embryo under consideration there still remains to be
described a most remarkable structure, the nature and signifi-

THE MALAYAN SPECIES OF ONYCHOPHORA. ays)

cance of which must be discussed. It is the small somite
(Pl. 6, fig. 10 a, ce. som.) situated in front of and above the
somite usually described as the first. It has developed a cavity
only on one side; on the other side it consists of a mere group
of nuclei. This is all that can be said of its structure, and
ucthing is definitely known of its origin. It seems that it
cannot have been separated from the first somite, which
remains undivided until a much later stage in the develop-
ment. If this be true, it follows that the small somite in
question must be an independeut structure, produced from
the mesodermal bands at a late stage and disappearing early,
and, as such, must have a very short existence. To judge
from its position, its form and structure, and its late forma-
tion and early disappearance, it would seem that we are justi-
fied in concluding that it represents a somite which has been
reduced to the merest vestige: so vestigial is it that it may
not even develop a cavity at all, but remain as a group of
nuclei lying in the undivided cytoplasm. It may, perhaps, be
permissible to conclude, since it is not possible to regard any
other metamerically arranged organs as corresponding to this
somite—unless the dorsal lobe of the brain, that is the archi-
cerebrum, be so considered—that the structure, the nature
and significance of which is here discussed, is a true cerebral
somite which up to the present has not been discovered.

The Third Embryo.—The third embryo, the internal
structure of which will be considered, is the one shown in the
corresponding figure on P]. 5, and is represented by four
drawings of sections, marked lla, 11b, llc, and 11d on
Pl. 7, in which the first, second, third, and fourth somites
are respectively illustrated.

On the ventral and dorsal surfaces the ectoderm forms a
thin layer, in which the tangentially compressed nuclei are
situated at some distance from one another, and, in the mid-
region, as well as towards the posterior end, is still in con-
tact with the endoderm (PI. 7, figs. 1l ¢ and 11d).

‘The rudiments of all the appendages have appeared as out-
growths of the dorsal moiety of the lateral thickening of

54, RICHARD ‘EVANS.

the ectoderm, and decrease in size from in front backwards.
The common rudiment of the nervous system and ventral
organs has been separated from that of the appendages, and
in the region of the third and fourth somites the thickening
in question has been divided into two, namely, an internal
one, the primordium of the nervous system, and an external
one, the forecast of the ventral organs. ‘The ectodermal
thickenings are continuous from the cephalic lobes backwards,
—that is, the rudiment of the para-cesophageal cord is already
formed. |

Owing to the way in which they develop from in front
backwards, ‘the mesodermal somites in the embryo under
consideration illustrate in the most perfect manner the
changes through which they pass in their development up to
a certain stage, a result made possible by the most admirable
series of sections into which the embryo was cut. — But the
general remarks I have to make on these will be reserved for
another section of the paper, in which the ccelom and the
mesodermal organs will be specially considered. At present
it suffices to say that there are twenty-seven pairs of them,
and that it is impossible to state with any degree of certainty
whether one pair more would have been developed or not.
There are no traces whatever of the germinal nuclei, either
in the mesoderm or endoderm.

In the embryo under consideration the endoderm is very
different from what it was in the second embryo described.
It is no longer possible to speak of a peripheral layer of endo-
derm containing nuclei, and a central mass of food yolk devoid
of nuclei. It would be preferable to speak of a peripheral
layer and central mass of endoderm; for the endodermal
elements of the peripheral layer of the previous stage have
invaded the central mass and converted the central yolk
into a number of nucleated masses, which present the same
structure as the peripheral layer did in the previous stage:
In both the peripheral layer and the central mass spherical
masses of stored-up food materials, and almost innumerable
‘granules small and refringent in character, are present ;

THE MALAYAN SPECIES OF ONYCHOPHORA. 55

but the compound systems of yolk-masses have totally
disappeared, unless they are represented by the nucleated
masses. The nuclei of the peripheral layer are almost in-
variably circular or oval in shape, and usually much larger
than either mesodermal or ectodermal ones. Their larger
size, most probably, is correlated in some way with the
function of presiding over the transformation and elaboration
of the yolk present in the central mass of the previous stage.
This increase of size is quite comparable to that which takes
place in the case of the nucleus of the ovarian ovum, which
supervises the process of elaborating and storing up the
deutoplasm in the egg-cell. The two processes seem to be
comparable in every respect. he nuclei found in the central
mass of amoeboid wandering cells, may occasionally present
an irregular form, but they seem never to break up and dis-
integrate. ‘The wandering character of these cells more than
suffices to account for the angular outline of some of their
nuclei. In a slightly later stage in the development the
avgularity of the endodermal nuclei becomes much more
marked, even those of the peripheral layer presenting the
same characteristic (Pl. 7, figs. lla, 11b, llc, and 114d).

The Fourth Embryo.—The fourth embryo, the internal
structure of which will be described, is the one illustrated
by the corresponding figure on Pl. 5, and represented by
drawings of ten transverse sections on Pl. 7 (figs. 12 a, 12 6,
12} ¢, 12'd, 12:e,12 f, 12 9, 12h, 129, and 12 k).

On the dorsal and ventral surfaces, the ectoderm presents
the same characteristics as in the third embryo, already
described, with the difference that they are still more
emphasised, the layer being thinner, the nuclei more com-
pressed, and the space—especially on the dorsal surface—
over which it is in contact with the endoderm, being greater,
the last result being brought about, not so much by the
withdrawal of the mesoderm, as by the growth of the
embryo, and consequently more highly arched condition of
the dorsal aspect (Pl. 7, figs. 12 e and 12/f).

es.
The nervous system has undergone ouly a very slight

56 RICHARD EVANS.

change, the differentiation that has taken place beyond what
was observed in the third embryo being very small, but in
spite of this fact, the growth in size of the nervous rudiment
is considerable. Most marked of all is the increased thick-
ness of the forecast of the brain, which so far shows no sign
of demarcation into several ganglia or lobes, and is situated
in front of the renal opening of the first somite (Pl. 7, figs.
12 a, 12.6, 12.c, and 12 d).

Both the stomodeeal and proctodveal invaginations are well-
formed structures, and communicate with the irregularly
shaped enteron (PI. 7, figs. 12 e and 12k).

The mesodermal somites have attained, in a general way, a
more advanced stage of development than they had in the
previously described embryo, in which they were not divided
into appendicular and median portions. In the present
embryo the renal portion of the first somite communicates
with the exterior, and the same portion of several other
somites has reached the ectoderm, though the opening has
not been actually formed (PI. 7, figs. 12 d and 12 e, ren. op.).
“Germinal nuclei” have already appeared in the splanchnic
walls of several somites (Pl. 8, fig. 13).

In the embryo under consideration the endoderm seems
in some respects to be in a less advanced state of develop-
ment than in the third one. In connection with its structure
there are several points which should be noticed. In the
first place, the peripheral endoderm is not so well marked off
from the central mass as it was in either the second or the
third embryo, and it often contains within its substance a
number of yolk-bodies belonging to the type referred to as
compound systems, which was not the case in the third
embryo (Pl. 7, fig. 12). In the second place, the presence
of compound systems marks a decidedly less advanced state
of development, unless they are regarded as the products of
the metabolic activity of the endodermal elements them-
selves, and different from those, occurring in the second
embryo, which were directly derived from the yolk-bodies
originally found in the egg. It seems that this second alter-

THE MALAYAN SPECIES OF ONYCHOPHORA. rd

native is more probable, because in the embryo under con-
sideration they occupy a central as well as a peripheral
position in the endoderm. If this supposition be correct,
their formation from the spherical masses of deutoplasm,
which occur in the endoderm of the third embryo, would
have to be explained on the same principle as the building
of compound systems from separate masses of deutoplasm
in the case of the ovarian ovum. In the third place, the re-
fringent granules found in the third embryo do not seem to
occur in the present one. In the fourth place, nearly all the
endodermal nuclei display an angular outline; while in the
third embryo only a few of these, situated in the central
endoderm masses alone, presented the characteristic in ques-
tion. In the fifth place, the endodermal cytoplasm, apart
from the difference arising from the change in character of
the deutoplasm, is quite different in structure, in that in
many places it appears distinctly fibrous, the fibres being
arranged according to no particular plan (PI. 7, fig. 12 e).

V. Tue DeveLopMEeNtT oF THE MESODERMAL ORGANS.

In the foregoing pages an attempt has been made to
describe the formation of the Germ Layers, and to give a
brief account of what happens during the early stages of the
development; but in the following pages it will be my en-
deavour to give a more detailed account of the development
of the Oraans from the Germ Layers.

(1) The Development of the Mesoderm with its
Cavities.

The development of the mesodermal somites has been
followed both by von Kennel (4) and by Mr. Sedgwick (5).
The description given by the former of their development in
Peripatus edwardsii is so unsatisfactory as to deserve ne
further mention in this connection ; while that given by the
latter of the processes going on in Peripatopsis capensis
is so good that, even now, after an interval of fifteen years

58 RICHARD EVANS.

it is almost impossible to improve upon it. However,
the failure of Kennel’s attempt to explain the mysterious
phenomena in P. edwardsii hardly accounts for the whole
difference between the two descriptions. But before proceed-
ing any further a few words must be said regarding the
terms which Mr. Sedgwick used in his description, and those
which will be adopted in the present account. On page 493
of the third part of his account (5) Mr. Sedgwick uses the
terms “dorsal” and “ ventral,” by which he means the dorsal
and ventral portions of the almost unmodified somite, with its
fully developed ccelom. On the following page he speaks of
an “ outgrowth” from the posterior part of the somite into
the rudimentary limb, and describes it as the “lateral or
appendicular portion” of the calom, in contrast with the
‘median portion.” On the same page he labels the “ median
portion” as the “dorsal,” and on page 496 the “lateral” or
“appendicular”’ becomes, in Mr. Sedgwick’s nomenclature, the
“ventral division.” In the passages referred to, each of the
terms dorsal and ventral has been used in two different
senses, for if I understand him rightly, Mr. Sedgwick does
not apply the expressions dorsal and ventral on pages 494
and 496 as he does on page 493 (5). In the present account
the expressions dorsal and ventral will be used in the
same sense as was given them on page 493 of Mr. Sedg-
wick’s account,—that is, to refer to the upper and lower
portions respectively of the unmodified somite. The ex-
pressions median and appendicular, alone, will be applied
to the portions of the ccelom resulting from its division into
a portion situated in the body, and a portion placed in the
appendage.

In treating of a mesodermal somite and its coelomic cavity,
it will conduce to clearness if I describe the development of
a somite situated somewhere about the middle of the body
before proceeding to consider the special modifications oc-
curring at either end of the animal.

The first appearance of the somites and of the cavities
situated in them has been considered in describing the second

THE MALAYAN SPECIES OF ONYCHOPHORA. 59

embryo (p. 49). A mesodermal somite which has reached
the point of maximum development, but which is still un-
modified, is crescent-shaped (Pl. 7, figs. lla and 12k), the
splanchnic wall always being less curved than the somatic.
Though the somites situated -on cither side never meet
either below or above the gut—except in the formation of
the ovary,—at one period in the development they approach
very near to one another, especially on the ventral surface,
a condition which reminds us of that occurring in an adult
Annelid, and at one time must have existed in the ancestral
Arthropod.

By the time the somites have reached this stage, the rudi-
ment of the nerve-cord has attained considerable develop-
ment. The rudiment in question, together with the myotome
which develops on the latero-ventral aspect of the somatic
wall, exerts such a pressure on the somite that the two walls
of its ventral portion are pressed together, resulting in the
obliteration of the ccelom in that part of the somite. Only
the dorsal portion of the ccelom, together with the small
rudiment of the appendicular outgrowth, remains (Pl. 7,
fig. 116). The growth of the appendicular portion for a
time keeps pace with that of the leg, to the distal end of
which the ccelomic outgrowth reaches. The ultimate separa-
tion of the appendicular from the median ccelom is
brought about by the continued increase in size of the myo-
tome, which extends both in front of and below the canal.
which places the two portions of the coelom in communi-
cation with each other. I have been unable to observe the
septum which Mr. Sedgwick describes as growing from the
ventral wall, and finally dividing the coelomic cavity into
median and appendicular portions, and I am firmly con-
vineed that the obliteration of the communication is brought
about in Eoperipatus by pressure from without, the result
of the continued growth of the myotome and nervous system.

The median portion of the ccelom persists for a short
time after the appendicular part has been constricted from
it, but it-soon disappears, leaving absolutely uo trace of its

60 RICHARD EVANS.

former existence, that is in the first eighteen to twenty
somites. Mr. Sedgwick, in a foot-note to page 493 (5), seems
to doubt his own description as to the origin of a layer of
cells situated above and below the somite, during a particular
stage in the development of P. capensis. In Hoperipatus,
however, the ventral portion of the ccelomic cavity is ob-
literated in exactly the same way as is described by Mr.
Sedgwick in the text, with the difference that the obliteration
is more extensive, probably owing to the presence of a large
quantity of food material in the interior-—wanting in P.
capensis,—which helps to intensify the pressure brought
about by growth. The median ccelom of the later stage
disappears in the same way as the ventral portion of the
earlier stage, that is by the coming together of its two
walls.

Now that the median celom has been absolutely obli-
terated, only the appendicular ccelom remains. It was
mentioned above that the appendicular outgrowth for a
time kept pace with the rudiment of the leg in which it is
situated; but this arrangement is not long continued. Its
distal end ceases to grow, and a spot on the ventral wall of
the appendicular ceelom begins to proliferate, resulting in a
downward growth which is situated near its proximal end.
Its distal end projects into the leg rudiment. This condition
is very well marked in the appendicular ccelom of the third
appendage (oral papille, Pl. 7, fig. 12g). But this distal
projection of the appendicular ccelom, in later stages of the
development, seems to be obliterated by the formation in its
walls of the rudiments of the leg muscles, which appear to
develop to a greater extent in Hoperipatus than in Peri-
patopsis. The downward outgrowth from the ventral wall
of the appendicular ccelom soon reaches the ectoderm at
the base of the rudimentary leg, and effects a communication
with the exterior. ‘The ectodermal indent is at first ex-
tremely slight, nearly the whole tube being mesodermal, and
throughout the series of changes which give rise to the renal
organ, consisting of renal duct and cclom, I have seen

THE MALAYAN SPECIES OF ONYCHOPHORA. 61

nothing that would tend to make it necessary even to modify
not to speak of abandoning, the view that the renal duct is
almost exclusively mesodermal ; that is, that the ectodermal
portion is extremely short, though always present—a con-
clusion arrived at from the histology of the fully developed
organ,

It will be observed that the above description of the de-
velopment of the celom in Eoperipatus agrees with that
given by Mr. Sedgwick of the processes occurring in P.,
capensis, save in a few details of no material importance, a
confirmation which bears weighty testimony to the correct-
ness of that embryologist’s work. This correspondence in
the development of the ccelom in these forms goes far to
prove that the changes which take place in P. edwardsii
would be found to agree, if only they were properly exa-
mined; for P. edwardsii and Hoperipatus are much more
closely related to each other than either of them is to P.
capensis.

(2) The Development and Disappearance ot the
First Somite (som.').

The early stages in the development of this somite have
been already described, and are illustrated on Pl. 6, figs.
9c,10a,and 106, In fig. 11a on Pl. 7 it is represented at
the zenith of its development, but still undivided. At the
side of the stomodeal invagination, its inner wall, contrary
to what occurs in the somites situated behind, has consider-
ably thickened to produce the rudiment of the mesodermal
portion of the fore-gut. In figs. 12 a, 12 b, 12 c, 12d, and 12e,
on PI. 7, the first somite is shown in its divided condition, the
two portions being separated by a thin septum, which was
probably produced in the manner described by Mr. Sedgwick
as occurring in all the somites; but I have no evidence on this
point. The two portions into which the somite is divided
extend the whole length of the rudimentary antenna (PI. 7,
fig. 12 a, som.'), Posteriorly they pass backwards almost to
the level of the second pair of rudimentary appendages,

62 RICHARD EVANS.

Frora the posterior inner corner of the renal portion a fine
canal passes to the exterior, its external opening being
situated immediately in front of the rudiment of the jaw and
on the inner side of the forecast of the lip (PI. 7, figs. 12 d
and 12e). It is only fair to point out that this ecommunica-
tion with the exterior was discovered many years ago by
Miss Sheldon in P. novee-zealandiew (6), and that the
nephridium of the first somite had been seen and identified
by Mr. Sedgwick in P. capensis (5). Consequently, the
somite under consideration was clearly shown to have the
same morphological value as any one of the succeeding
somites. Owing to the great increase in size of the cephalic
ganglia and first pair of ventral organs, the communication
of the first somite with the exterior soon becomes obliterated,
and the two portions into which the somite is divided be-
come much reduced in size and finally disappear. Fig. 22-4,
som.', on PI]. 9 shows the remnants of the first pair of somites
as small spaces situated above the brain and in front of the

eye.

(3) The Development and Disappearance of the
Second Somite (som.®*).

The early stages in the history of the second somite have
been noticed in describing the first and second embryos, and
are illustrated in figs. 10 ¢ and 10d on Pl. 6, in which it
is shown that the somite in question passes above the first,
one, situated in front of it, and below the third, which suc-
ceeds it.

Fig. 11 b on Pl. 7 shows the somite under considera-
tion much compressed, and displays a tendency on the part
of its walls to come together, and consequently to obliterate
the ccelom.

Dorso-ventrally, however, the somite presents a consider-
able extension. The same figure. also shows the lateral out-
growth of the coelom, which extends the whole length of the
rudimentary second appendage. ‘The outgrowth in question
is a true appendicular ccelom, on the ventral wall of which

THE MALAYAN SPECIES OF ONYCHOPHORA. 63

the myotome, which has already reached a fairly advanced
state of development, is placed, as in all the other somites.
Fig. 12 f on the same plate shows the second somite divided
into median and appendicular portions, both of which are
much reduced in-size and ultimately disappear.

(4) The Development of the Third Somite (som.*).

The third somite passes through the early stages of its
development in a way similar to a somite from the mid-region
of the body, which has been already described. However,
some peculiar points in its history should be noticed, and are
illustrated in fig. 12 g on Pl. 7. In the first place, the
distal extension of the appendicular ccelom is much larger
than in any of the succeeding appendages, a fact which is in
keeping with the great size of what, in the adult, may be
designated the salivary celom. In the second place, the
downward outgrowth from the appendicular ccelom to
form the duct is much larger than the same outgrowth is in
any other appendage. The later stages in the development
of the appendicular ccelom of the third somite so closely
agree with the account given by Mr. Sedgwick of the course
followed in the development of the salivary gland in P.
capensis, as to make any further remarks unnecessary (8).

(5) The Development of the Generative Organs.

Before I proceed to describe the development of» the
generative system in Hoperipatus, it is necessary to
summarise von Kennel’s and Mr. Sedgwick’s accounts of the
development in Peripatus edwardsii and Peripatopsis
eapensis respectively ; for the descriptions given by these
two embryologists disagree to such an extent as to render
the development of the organs under consideration a subject
of absorbing interest.

According to von Kennel’s account, first, the germinal
nuclei arise from the mesoderm; secondly, only one pair of
somites takes part in the formation of the generative organs ;

64 RICHARD EVANS.

and thirdly, the generative ducts develop almost exclu-
sively from the ectoderm by invagination (4).

According to Mr. Sedgwick’s description, first, the germinal
nuclei arise from the endoderm, and only later acquire a
relation to the mesoderm; secondly, several pairs of somites
take part in the formation of the generative organs; and
thirdly, the generative ducts are derived from the appendi-
cular outgrowths of the somites of the anal papillee (5),

The Germinal Nucleiand their Place of Origin,—
Having made a brief statement of the present position of
our knowledge of the development of the generative organs
of the Peripatide, I shall proceed to give an account of
their origin and growth in the genus Hoperipatus, It
would seem that the germinal nuclei can be distinguished at
an earlier stage in the prospective female than in the male;
for this reason the following description applies to the
former rather than to the latter. Because the number of
metameres present in the body is not constant, it will be
necessary to refer to the somites as counted from the
posterior end, though in the early stages this method has
the disadvantage that the generative nuclei appear before
the last two somites are formed. The somite of the genera-
tive duct is the third from the posterior end, som,"*, the
somites in front of it being som.”°, som."*, etc,

Kach of the embryos, sections of which are represented in
figs. 13 and 14 on Pl. 8, possessed twenty-five pairs of
somites, that 1s two pairs fewer than the smaller number
that are accounted for in the adult female. Consequently
the last actually developed pair of somites in the two
embryos under consideration must be labelled som."*, and
the pair situated immediately in front of the last one, and
shown in section in the two figures above mentioned, as
"8. In the two embryos represented in section in these
figures, germinal nuclei occur in the splanchnic walls of
four pairs of somites situated immediately in front of the
last actually developed pair (som."®), but they are more
numerous in the twenty-third and twenty-fourth pairs than

som.

THE MALAYAN SPECIES OF ONYCHOPHORA. 65

they are in the twenty-first and twenty-second. They are
present in all stages of transformation, from the unmodified
nucleus of the mesoderm to that of somewhat enlarged
germinal nuclei. At first they are very slightly modified,
and situated among the single row of nuclei found in the
splanchnic walls of the somites. As they grow in size they
become pushed towards either side, chiefly to the endoder-
inal; that is, away from the cavity of the somite, the cells of
which arrange themselves in a layer round them (PI. 8, figs.
15a and 156). There seems to be no doubt that they
originate in the mesoderm ; and, in conclusion, I must state
that I have no hesitation whatever in saying that von Kennel
has made an error of observation in deriving the genital
organs from one pair of somites, and that Mr. Sedgwick
appears to be wrong in concluding that the germinal nuclei
are endodermal, unless in these respects the species in
question radically differ from the genus Noperipatus.

The Formation of the Ovary.—The celom of the
median portion of the four pairs of somites which take
part in the formation of the ovary, is not obliterated as in
the somites situated in front. From either side the somites
approach one another dorsally, and then fuse to form an
ovary which, at first, possesses two absolutely independent
cavities, and adheres to the ventral wall of the pericardium
(Pl. 8, figs. 15a and 156, and 160 and 16c). Later on
the septum which separates the two cavities of the ovary
becomes broken down near the anterior and posterior ends,
but is retained in the middle (Pl. 8, figs. 17a and 175).
As development proceeds the germinal nuclei, which at first
were few, increase in number, and give rise to two germinal
bands, which occupy the ventral wall of the ovary and almost
fillits cavity. The developing ovaare surrounded by follicle
cells, and are suspended in the body-cavity, in which for
want of space they are wedged against one another (Pl. 8,
fig. 176), an arrangement which should be contrasted with
that described by Gaffron in P. edwardsii (8, pl. xxi, figs.
9 and 10).

VOL. 45, part 1.—NeW SERIES. E

66 RICHARD EVANS.

The Development of the Genital Ducts.—In one of
two embryos slightly older than those of which sections are
shown in figs. 18 and 14, there were twenty-six pairs of
somites actually formed, and in the other twenty-seven. At
the posterior end, the former had two pairs of somites
possessing no germinal nuclei, while the latter had three
pairs devoid of them. Of these three pairs of somites which
develop no germinal nuclei, the anterior one (som."*) gives
rise to the genital ducts, the middle one (sum."') produces
the last renal organ, while the hindermost disappears in the
female, and gives rise to the accessory glands in the male.
It is not quite correct to say that the entire somite becomes
the genital duct, for the ventral portion becomes obliterated
by the coming together of its walls, of which the cells form
at a later period the lining of the blood-spaces which develop
in that region. An appendicular outgrowth, which is never
separated from the median ccelom, forms in the genital somite,
as in any other; but instead of opening to the exterior at
the base of the legs, the two tubes debouch together in
the median line. The inner ends of the two ducts thus
formed come together in the female and unite, subsequently
opening by a common pore into the cavity of the ovary. In
the male they do not communicate with each other at any
time, but each duct acquires a separate opening into the
testis of its own side.

Further Modifications of the Oviducts.—As regards
the external ends of the oviducts, the stage already described
almost corresponds to the structure in the adult. The ecto-
dermal ingrowth, such as it is, forms the extremely short
vagina of the adult, and no more. In one of my embryos
there is a distinct line of demarcation between the ectoder-
mal and mesodermal constituents. The vagina les almost
horizontally, and the ectodermal portion of its dorsal wall,
derived from the posterior hp of the external opening, is
decidedly shorter than that of the ventral wall, which pushes
its way forward as a tongue-shaped structure, situated in the
inedian line. ‘I'here seems to be no doubt that von Kennel

THE MALAYAN SPECIES OF ONYCHOPHORA. 67

is in error in deriving almost the whole of the oviducts from
the ectoderm ; for all the convolutions, which appear later
on in the development, are the result of growth in the inner
part of the oviducts and not in the outer.

The Receptaculum Ovorum.—The receptaculum
ovorum arises as a simple evagination on one side—usually
the dorsal side—of the oviducal wall (Pl. 8, fig. 17 ¢, re. 0.).
The evagination in question elongates, and the wall at its
free end becomes thinner, so as to form a membrane, which
closes its distal end (Pl. 9, figs. 26a, 266, 26c, 26d, and
26 e). The whole organ is merely a local growth which
appears relatively late, and for these reasons I am inclined
to reject the suggestion made by Mr. Sedgwick, and adopted
in the first part of my account of Hoperipatus, that the re-
ceptaculum ovorum is homologous with the renal end-sac (2).

The Receptaculum Seminis.—The development of the
receptaculum seminis has been described by Gaffron (8),
whose account is essentially correct. In Hoperipatus each
oviduct forms a loop (Pl. 9, fig. 26a, re. s.). The canal, at
the free end of the loop, begins to expand (PI. 9, fig. 26 },
re. s.), and the cavity of the receptaculum seminis forms by
the continued enlargement of this part (Pl. 9, figs. 26 c, 26 d,
and 26e,7e.s.). The lateral portions of the loop become the
narrow ducts of the receptaculum seminis and always remain
embedded in its wall, which is much thinner than was re-
presented by Gaffron, who described in the genus Peripa-
tus a thick middle layer, a layer which in Hoperipatus
is extremely thin. Gaffron also figures the openings of
the ducts into the receptaculum seminis as situated close
together, while in Eo peripatus they are placed on opposite
sides (Pl. 9, fig. 26e). At first there is no communica-
tion between the two portions of the main canal, save by
way of the loop, which gives rise to the receptaculum
seminis and its ducts; but later on the septum, at first
thick (Pl. 9, fig. 26 6), becomes thinner (Pl. 9, figs. 26c¢
and 26 d) and ultimately disappears (Pl. 9, fig. 26). At
first the lining of the canals, situated on either side of the

68 RICHARD EVANS.

loop, is not differentiated, but when the secondary commu-
nication has been formed the canals in question are lined re-
spectively by cells which are quite different im character.
The portion situated between the receptaculum seminis and
the ovary is lined by short cells with no definite cell outlines,
and with their nuclei placed either in the centre or near their
free end; but the portion situated on the other side of the
receptaculum seminis is lined with columnar cells, which
possess sharp cell outlines and nuclei placed at their base
(PI. 9, fig. 26e). This difference is probably prophetic of
the different functions which these lining cells have to per-
form later on in life; the columnar cells of the uterine part
having to provide the developing young with the enormous
amount of food material necessary to enable them to grow to
the unusual length of twenty-seven millimetres.

The Male Genital Ducts.—The male genital ducts de-
velop in the same way as the oviducts, up to the stage at
which they acquire an opening to the exterior, except that
their inner ends do not unite with each other (PI. 8, figs.
19a@and19b). After the formation of the external pore, the
short median portion, which alone is ectodermal, elongates
and forms a loop, which is usually placed on the left side of
the rectum, in the middle chamber of the body-cavity. Oc-
casionally, however, the loop in question may pass under the
left nerve-cord. In the former case the point of union of
the vasa deferentia to the common duct is placed in the
median chamber, and the right vas deferens does not pass
under the left nerve-cord; but in the latter case the point of
union of the two ducts is drawn to the left side of the cor-
responding nerve-cord, and the right vas deferens passes
under both cords, just in front of the external pore. It seems
from Balfour’s description of P. capensis (1) and von
Kennel’s account of P. edwardsii (4) that the condition least
prevalent in the genus Hoperipatus is almost universal in
these species. It would seem that there is no doubt that the
condition described by Balfour and von Kennel is the derived
one; the primitive condition being the one found in the

THE MALAYAN SPECIES OF ONYCHOPHORA. 69

majority of individuals of the genus Eoperipatus, in which
the common duct is placed in the median chamber and does
not pass under the nerve-cord.

The Second Somite from the Posterior End (som),
—In both male and female, the modifications which this pair
of somites undergo are similar to those passed through by
any pair of somites situated further forward, except that the
renal duct does not become elongated and coiled, and that
in the adult male it tends to disappear (PI. 8, figs. 205,
ren. or., and 21 b, ren. or.).

(6) The Development of the Male Accessory Glands.

The Last Pair of Somites (som").—In the female
the last pair of somites disappears almost immediately after
their formation; but in the male they are destined to
become the accessory glands. In an embryo slightly older
than the one shown in fig. 4 on Pl. 5 they are fully
formed and crescent-shaped. In a somewhat older embryo,
oblique sections of which are shown in figs. 20a and 20d
on Pl. 8, they are situated close to the posterior ends of
the nerve-cords, and curve round it in such a way that they
come in contact with the ectoderm of the ventral surface
just in front of the anal slit, though they do not yet open to
the exterior. In an embryo slightly older than the one shown
in fig. 5 on Pl. 5 the last pair of somites open into the
exterior, and have already assumed a tubular form, though
they are still short (PI. 8, figs. 21 a and 218, m.a.g.).

The above account proves that the male accessory glands
of Koperipatus are in part mesodermal, and that the cavity
of their inner moiety is coelomic. From this conclusion it
follows that the male accessory glands are homologous with
the renal organs. Therefore in Hoperipatus the salivary
glands, the renal organs, the genital ducts, and the male
accessory glands are all homologous organs, derived from
the mesoblastic somites, and put in communication with the
exterior by means of a short invagination of the ectoderm,

70 RICHARD EVANS.

VI. Tse DevELOPMENT OF THE NERVOUS SYSTEM AND
VENTRAL ORGANS.

The development of these two systems has been traced up
to a certain stage in describing the four embryos considered
in an earlier part of this paper. The common rudiment
of the nervous system and ventral organs was found to
consist of a thickening of the ectoderm which, develop-
ing from before backwards, became continuous from the
anterior to the posterior end. The internal layer of the
ectodermal thickening becomes separated off as the rudi-
ment of the nervous system, while the outer layer, at first
continuous, gives rise to the ventral organs. It is not until
the embryo has nearly reached the stage of development
shown in fig. 5 on Pl. 5 that any signs of breaking up
of the rudiment in question appear. The nervous system
does not become divided into separate ganglia, even in the
adult, except in so far as the shght swellings occurring
between each pair of appendages indicate such a division.
This undifferentiated condition of the nervous system renders
metameric comparison with the mesoblastic somites a matter
of no small difficulty. ‘he rudiment of the brain and first
pair of ventral organs is, at first, an undivided mass; but in
an embryo which has reached the state of development
shown in fig. 5 on Pl. 5 the lobes of the brain are
making their appearance, and the first pair of ventral organs
have been invaginated, though on the inner side they are
continuous with the brain ganglia lying above them. Fig.
23a on Pl. 9 shows a section of the brain which passes near
the mid-dorsal line and above both the base of the antennee
and the eyes. ‘lhe upper side of the figure corresponds to
the dorsal aspect, and passes through the archicerebral lobes
of the brain (a. lo.), while the ventral side shows the lateral
lobes (lo.', lo.*, and lo.*). Fig. 24 on Pl. 9 represents a some-
what oblique section of the frontal part of the brain, and it
will be noticed that, on the antero-frontal aspect, the archi-
cerebral lobes project forward as small prominences, which

THE MALAYAN SPECIES OF ONYCHOPHORA, 71

seem to correspond to the cephalic processes discovered by
von Kennel, and interpreted as the primitive cephalic antenne.
The lateral lobes of the brain, collectively, correspond to the
first somite, and there seems to be no reason whatever for
regarding them as separate ganglia representing metameres
which have disappeared. They are situated in front of the
ganglion which supplies the jaws, and which does not,
properly speaking, form a part of the brain. The connec-
tion between the ganglia of the jaws and the brain is a
ventral one (Pl. 9, figs. 22 b and 22c., lo.*), similar to that
occurring between any two pairs of ventral swellings of the
nerve-cords. ‘The nerves to the jaws are given off from
points situated near the dorsal ends of the fourth lobes,
one of which is shown on the left side in fig. 22 ¢, just
below the line marked lo.4.. As far as the structure of the
brain can be relied upon, it gives no indication of any meta-
meres having disappeared. The dorsal lobes seem to repre-
sent the archicerebrum ; the three lateral ones are merely
differentiations in the portions corresponding to the first
somite, while the fourth lobe supplies the jaws, and conse-
quently belongs to the region of the second somite.

The ventral organs arise in a manner which has been fully
described by von Kennel and by Sedgwick, and there is no
object in further describing them here; but it seems neces-
sary to discuss their relation to the nervous ganglia and the
mesoblastic somites. It would seem that the anterior one
corresponds to the first pair of somites and to the three-lobed
brain, and not to the cerebral somite and archicerebrum.
The second pair of ventral organs belong to the metamere
which carries the jaws, above which they are immediately
situated, though they have acquired a secondary relation to
the second and third lobes of the brain (Pl. 9, fig. 236,
v.o.’). Their relation to the Jaws seems more important than
to the brain, for the latter seems to have been brought about
by the secondary shifting of the mouth parts. On this view
of the second pair, the third pair of ventral organs must be
considered as having been divided into two halves, one of

72 KIGHARD EVANS.

which is drawn, during development, into the buccal cavity,
while the other is left outside. The position of the intra-
buccal half, above the point of union of the salivary glands
to form a common duct, tends to prove that it really belongs
to the somite of the oral papilla, to which also the salivary
elands are related.

The following tabular form may help to explain the
arrangement of the metameres which constitute the head of
Hoperipatus:

ys i Nervous ganglia |
Mba Toa oe) Ventral Organs. | Appendages. |
Somites. or lobes. |
|
|
. oe =~ =A |
.
Che archicerebral so- Archicerebral lobe — Precephalie
mite is vestigial. (dorsal lobe). | processes.

to the exterior and, the brain.
then disappears.

The first somite opens Three lateral lobes of First ventral organ. [Antenne

|
|
|

[he second somite Fourth lobe of theSecond ventral organ Jaws.
disappears without brain placedonthe placed above and on
opening to the ex- dorsal end of the the inner side of
terior. para-cesophageal the jaws.

cords. |

The third somite gives First ventral gan- Third ventral organ Oral papille.
rise to the salivary glion. divided into two

| gland. halves, one in the

buccal cavity and

one outside,

From the above table another arrangement of the ventral
organs suggests itself, and it must be admitted that there is
a good deal to be said in its favour. The view in question is
that the first ventral organ corresponds to the cerebral
somite, which has become vestigial; the second to the first
somite ; and that the usually described third one should be
considered as forming two complete ventral organs and not
two halves. On this view one of these would belong to the
somite of the jaws and the other to that of the oral papilla.

THE MALAYAN SPECIES OF ONYCHOPHORA. 73

However this may be, it makes no difference in the number
of parts which constitute the head of Hoperipatus; for on
either view the head is composed of an archicerebral portion,
followed by the segment of the antenne, of the jaws, and of
the oral papillee.

VII. ''ar DEVELOPMENT OF THE EYE.

The first rudiment of the eye appears as a small pit, situated
on the dorsal edge of the ectodermal thickening, which in
later stages gives rise to the brain (PI. 7, fig. 1l a, e.). By
degrees the depression in question becomes deeper, and in
section presents the appearance of a fold of the ectoderm.
The canal putting its cavity in communication with the
exterior finally disappears. In fig. 25 a, on Pl. 9, the
detailed structure of the rudiment of the eye at this stage
is shown. The position and arrangement of the nuclei is
interesting, those of the outer layer being arranged near
the cavity of the depression, while those of the inner layer
are separated from the cavity by a layer of cytoplasm, which
presents the appearance of being divided into areas. This
layer of cytoplasm later on gives rise to the rods existing in
the adult eye. Soon after the obliteration of the canal,
which puts the cavity of the eye rudiment in communication
with the exterior, the outer layer is not in any way distinct
from the overlying ectoderm, nor is the inner layer sharply
marked from the underlying brain rudiment (PI. 9, fig. 25d).
The cavity of the eye is still small and quite empty, and the
nuclei situated on the inner side are beginning to arrange
themselves in layers with their long axes radially directed.
In the same figure the eye is shown as having been to some
extent constricted from the underlying brain rudiment. In
fig. 25 c¢ on the same plate this constriction has been
carried much further, and the eye has been completely sepa-
rated from the ectoderm. The cavity of the eye is much
larger though it is still empty. Pigment is making its appear-
ance in the layer of cytoplasm situated on its inner side, and
distinctly marked by radiating lines, Simultaneously with

74, RICHARD EVANS.

the constriction of the wide stalk shown in Fig. 25 b a cord
of white matter has developed inside it. The cord in ques-
tion passes from the white matter of the brain to the eye,
and seems to spread itself over the inner half, the nuclei of
which are arranged in many layers (PI. 9, fig. 25 c). There
seems to be no doubt that this cord, that is the optic nerve,
is developed in situ from the wide stalk of earher stages
(Pl. 9, fig. 25 6), much in the same way as the nervous strands
which connect the ventral organs to the nerve-cords are
formed, but with one important difference, namely, that the
cells which form the latter are scarcely modified as compared
with those of the grey matter of the nerve-cord ; while the
cells which give rise to the former, that is to the optic nerve,
undergo changes similar to those brought about in the ele-
ments which produce the white matter of the brain and
nerve-cords. The cells which give rise to the white matter
are developed from the ordinary cells of the sensory rudi-
ment, that is the common rudiment of the nervous system,
eyes, and ventral organs. The cells of the grey matter have
small and highly granular nuclei, and a very small amount
of cytoplasm; but the cells which give rise to the white
matter have large nuclei, with very fine chromatin granules
and a large amount of cytoplasm, in which there are no
distinct cell outlines. The cells producing the white matter
were only observed in embryos. I am not aware of their
having been described before.

Conclusion.—There seems to be no doubt as to the cor-
rectness of Mr. Sedgwick’s account of the development of
the eye in P. capensis, as a cerebral eye, and the same
statement is equally true of the formation of that of KH.

weldoni.

VIII. TuHe ENDopvERM.

The history of the younger stages in the development of
the endoderm has been given in the early part of this paper
(pp. 46—55), but owing to the importance of the question it
is necessary to recapitulate, and to add some more facts re-

THE MALAYAN SPECIES OF ONYCHOPHORA. 75

garding the changes which take place during the later stages
in the development.

The endodermal elements are derived from the lips of the
blastopore, and travel inwards along the outer layer of the
yolk, which at first is devoid of nuclei (Pl. 6,. fig. 9 ¢, en.).
In the second embryo illustrated in figs. 10 a—j on Pl. 6
the endoderm forms a complete layer, in which the nuclei,
especially towards the dorsal aspect, are placed at a con-
siderable distance from one another. The central yolk is
still free of nuclei. In the third embryo, sections of which
are represented in figs. 11 a—d on Pl. 7, the endodermal
elements have invaded the central mass, and changed its
entire character into one resembling the peripheral layer of
the second embryo. In the fourth embryo, which is more
advanced than the third, the endodermal elements, leay-
ing behind them a number of yolk masses in the centre,
have re-entered the peripheral layer. That such is the
case is almost certain, from the irregular disposition of the
endodermal nuclei, and from the absence of degenerating
nuclei either in the previous or the present embryo. In the
next embryo, the seventh in the uterus counted from the
ovary, the endoderm has reconstructed itself. Blood-spaces
have already appeared, and the intestine is surrounded by a
layer of mesoderm. ‘lhe endodermal nuclei are placed rather
far from one another, and close to the mesodermal covering of
the intestine. The cytoplasm of the endodermal layer is sparse,
vacuolated, and seems to be undivided; and the whole layer
is devoid of any kind of food material. In an older embryo,
the ninth in the uterus, the general arrangement is the same
as in the seventh. The endodermal nuclei are situated at the
base, but are closer to one another ; and the cytoplasm, which
is still undivided, contains a number of small round bodies—
presumably food material. In the next embryo, the tenth in
the uterus, the general arrangement resembles that of the
previous stages in the development. The nuclei are still
basal, but the cytoplasm is distinctly divided by cell outlines ;
and yolk-bodies are much. more numerous and larger than they

76 RICHARD EVANS.

were in the last-mentioned embryo. The endodermal cells
are distinctly columnar and rather long, features which are
gradually becoming more marked stage by stage. In the
eleventh embryo found in the uterus the endoderm possesses
the same characteristics as in the last embryo, but all of
them are much more highly developed. ‘The cells are longer
and possess more marked outlines, and the food-bodies are
larger and more numerous. In the twelfth embryo occurring
in the uterus all the features which characterised the eleventh
are present; but the cells are still longer, and the food-bodies
have increased both in size and number. In the last two
embryos the cavity of the gut is partially filled with a kind
of débris. The globules occurring in the endodermal cells
of the older embryos found in the uterus are undoubtedly
the same as those found in the same position in the adult.
It will be noticed that there isin the above account of the
development of the endoderm in Eoperipatus no mention
of a histolytic process, such as has been described by Dr.
Willey in P. nove-britannie (7). If such a breaking
down of the endoderm took plate in Hoperipatus, it is
not likely that it would have been missed, as the nine
embryos whose endoderm has been described above repre-
sent all stages of development, ranging from an early
gastrula to an embryo coloured almost like the mother. As
faras P. nove-britanniz is concerned Dr. Willey seems
very decided; but when the subject is critically examined it
does not seem so certain that the process of histolysis de-
scribed by Dr. Willey does really take place in nature. In
the first place, his specimens were preserved in formal,
without opening them, consequently the preserving fluid had
to penetrate not only the body-wall of the mother, but also
that of an embryo almost ready for birth. To say the least
it is very doubtful whether formal is capable of doing this.
In the second place, the endodermal layer of the older
embryo found ‘in the uterus should not be compared, as
regards difficulty of preservation, with ordinary tissue, such
us the coelomic end-sac of the renal organs, but with such

THE MALAYAN SPECIES OF ONYCHOPHORA. ae

structures as the ova of Eoperipatus and Peripatoides,
cells which are full of food-yolk, and consequently most
difficult to preserve in a good condition. How difficult food
material of any kind is to preserve is too well known to need
any further explanation in the present paper. For these
reasons it seems that we are fully justified in questioning the
accuracy of Dr. Willey’s conclusion. It is much more likely
that the older embryos that he took out of the uterus of P.
nove-britanniz were not well preserved, than that there is
a periodic histolysis of the endodermal cells. Dr. Willey’s
“strands of protoplasm, beset with eosinophile globules of
varying sizes,” seem to be nothing but the broken cell walls
of badly preserved specimens.

Conclusion.

In addition to those whom I mentioned at the close of the
first part of my account of the Malayan species of Onycho-
phora, my thanks are due to Mr. P. J. Bayzand, the able
artist in the Department of Comparative Anatomy at Oxford,
for the trouble he has taken with the drawings on PI. 5,
and especially to Professor Poulton for reading over the
proof-sheets.

THe DEPARTMENT OF CoMPARATIVE ANATOMY,

THE Musrum, Oxrorp;
March 19th, 1901.

List or REFERENCES.

1. Batrour, I. M.—“The Anatomy and Development of P. capensis.”
Posthumous memoir, edited by H. N. Moseley and A. Sedgwick.
‘Quart, Journ. Mier. Sci.,’ vol. xxiv, pp. 218—259, Plates 13—20.

2. vans, R.—*On Two New Species of Onychophora from the Malay
Peninsula,” ‘ Quart. Journ. Mier. Sci.,’ vol. xliv, pp. 473—538, Plates
32—37.

3. Garrron, Kip.—“ Beitrage zur Anatomie und Histologie von Peripata,”
Parts I and II in Schinder’s ‘ Zoologische Beitrage,’ vol. i, 1885, pp.
33 and 145, Taf. vii—xii und xxi—xxiil.

78 RICHARD EVANS.

4. Kennet, J.— Entwicklungsgeschichte von Peripatus Edwardsii,
Blanch., und P. torquatus, n. sp.,” Theili, ‘ Arbeiten a. d. zool.-zoot.
Inst. Wurzburg,’ vol. vii, p. 95, Taf. v—xi; Theil ii, ibid., vol. viui,
Taf. i--vi.

5. Sepewick, A.—‘* The Development of the Cape Species of Peripatus,”
Part I, ‘ Quart. Journ. Mier. Sci.,’ vol. xxv, pp. 449—468, Plates 31,
32. Part II, ibid., vol. xxvi, pp. 175—212, Plates 12, 14. Part
ILI, ibid., vol. xxvii, pp. 467—550, Plates 34—37. Part IV, ibid.,
vol. xxviii, pp. 873—396, Plates 26—29.

8. SueLpon, L.—‘‘On the Development of P. nove-zealandia.” Part
I, ‘Quart. Journ. Micr. Sci.,’ vol. xxviii, pp. 205—237, Plates 12—16.
Part II, ibid., vol. xxix, pp. 283—294, Plates 25, 26.

7. Wiitey, A.—‘* The Anatomy and Development of Peripatus nove-
britanniz,” * Zool. Results based on Material from New Britain,’
etc., Part I, pp. 1—52, plates i—iv.

EXPLANATION OF PLATES 5—9,

Illustrating Mr. Richard Evans’s paper on “ The Develop-
ment of the Malayan Species of Onychophora.”

List of Reference Letters.

a.lo. Archicerebral lobe of the brain. az.gr. Anal groove. axt. Antenne.
b.p. Blastopore. b7, Brain. c.y. Central yolk. e. Kye. ec. Ketoderm. ex.
Eindoderm. ez. iz. Endodermal invagination. ez. 2. Endodermal nucleus,
ex. y. Hixternal yolk. ge. 2. Germinal nucleus. ge. 7. Germinal ridge. A¢.
Heart. jaw. Rudiment of the jaw. /o., Zo.?, /o.3, and do.4. The first, second,
third, and fourth lobes of the brain. m.a.g. Male accessory glands. m.d. .
Mid-dorsal line. mesod. Mesoderm. myot. Myotome. x.c. Nerve-cord.
ov. Ovum. ov. ec. Ovarian cavity. ovid. (= som."*). The antepenultimate
somite or oviduct. pa. com. Para-cesophageal commissure. po. 6. p. Posterior
end of the blastopore. pe. Pericardium. pr. g. Primitive groove. proctod.
Proctodeum. re.o. Receptaculum ovorum. ze. s. Receptaculum seminis.
ren.f. Renal funnel. vez.o. Renal outgrowth. rez. op. Renal opening. sa. gl.
Salivary gland. som.', som.?,...... som.™', som.™. Mesoblastic — somites.
stomod. Stomodeum. ‘tes. Testis. vas. def. Vas deferens. v. 0., v. 0.3, v. 0.3,
etc, Ventral organs. v. 0.52, The buccal portion of the third ventral organ.
v.02. The external portion of the third ventral organ,

THE MALAYAN SPECIES OF ONYCHOPHORA., 79

All the figures on Plate 5 were carefully drawn by Mr. P. J. Bayzand
under the author’s supervision. The remaining figures were drawn by the
author himself with the aid of the camera lucida.

PLATE 5.

Fig. 1 (x 120).—This figure represents a young embryo of Eoperipatus
weldoni. The blastopore is slit-like, and possesses irregular outlines.
Externally there is no sign of segmentation. The actual length of the
embryo was 1+] mm.

Fig. 2 (x 120).—This figure represents a young embryo of Eoperipatus
weldoni. ‘The blastopore has been divided into two portions by the fusion
of the lips. The anterior portion is small and slit-like, and, owing to the
external yolk, can only be seen in sections. The posterior portion is a big
hole with no very definite anterior lip, but the lateral and posterior lips are
considerably thickened and quite definite. The primitive groove and streak
have appeared, and two somites of the body are externally visible. The
actual length of the embryo was ‘9 mm.

Fic. 3 (xX 120).—This figure represents a young embryo of Eoperipatus
weldoni. The rudiments of nearly all the appendages are externally visible.
The antenne have three rings, but no other appendage is ringed. The actual
length of the embryo in its folded condition was 1°4 mm.

Fie, 4 (x 120).—This figure represents an older embryo than the one
shown in the preceding figure. The rings on the antenna have multiplied.
The legs as well as the body-wall are already ringed. The lips are appearing,
and the oral papille present a depression at their outer ends. The actual
length of the embryo in its folded condition was 1°3 mm.

Fie. 5 (x 27°5).—This figure represents an embryo which is considerably
older than the one shown in the previous figure. The elongation and the
increase in the number of rings on the antenne are well marked. The ring-
like markings on the body and the appendages are clearly visible. The
posterior end of the embryo exhibits a tendency to pass from the dorsal
aspect, a position occupied by it in the embryos shown in Figs. 38 and 4, and
to become situated at the side. The actual length of the embryo in its
folded condition was 31 mm.

Fig. 6 (Xx 10).—This figure represents an embryo still more advanced than
the one shown in the previous figure. The most marked change of form
that has taken place consists in the partial straightening of the posterior end
after slipping from the position occupied by it in the embryos shown in Figs.
3 and 4, a change which was just commencing in the embryo shown in Fig. 5.
The actual length of the embryo in its folded condition was 7 mm.

Fie. 7 (x 5).—This figure represents an embryo in which the body is
still more straightened than the one shown in the previous figure is. Not

80 RICHARD EVANS.

only is the posterior end free of almost any twisting, but the anterior end has
so far unfolded itself that only the antenne and oral papille are situated in
front of the curvature. The actual length of the embryo in its folded condi-
tion was 17 mm.

Fie. 8 (x 3).—This figure represents the oldest embryo in the uterus, an
embryo which is coloured almost like the adult. The body is quite straight.
The actual length of the embryo was 27 mm.

PLATE 6.

Fics. 9a—9 d(x 120).—These figures represent four transverve sections of
an embryo slightly older than that shown in Fig. 1, but not as old as the one
shown in Fig. 2. There are no nuclei in the yolk, a feature which should be
specially noticed.

Fig. 9a.—This figure represents a section passing in front of the anterior
end of the blastopore. The ectodermal layer is already thickening to
form the rudiments of the brain.

Vig. 94.—This figure represents a section passing through the anterior
end of the blastopore (d.p.), which is a wide groove devoid of nuclei
and situated on the ventral surface.

Fig. 9¢.—This figure represents a section passing through the posterior
edge of the blastopore, and shows the thickened lips on either side of
it. The blastopore in the region situated between the sections shown in
Figs, 94 and 9¢ is being obliterated by the growing across of cells from
the blastoporic lips, which are not well marked in the region in question.
The only somite as yet formed is cut across in the present figure, and
on its inner side are situated a few endodermal nuclei (ez.) derived from
the lips of the blastopore.

Fig. 9d.—This figure represents a section passing behind the blastopore.
The rudiments of the second and third somites are present as groups
of nuclei. The endodermal nuclei are more numerous in this region
than they are in the front part of the embryo.

Fies. 10 a—10 & (x 120).—These figures represent a number of sections
selected from a series, cut transversely, of an embryo about the same age as
the one shown in Fig. 2, in which only the first two somites are visible
externally. The blastopore is divided into two parts, an anterior and a
posterior. The central cavity of the embryo is occupied by a mass of yolk,
which protrudes from the blastoporie openings, and to some extent spreads
itself over the ventral surface of the embryo. The ectoderm of the dorsal
surface consists of a thin layer of protoplasm with flattened nuclei arranged
tangentially to the surface. On the lateral aspects, and on the ventral
towards the posterior end, the ectoderm is thickened to form the undiffer-
entiated rudiment of the appendages and nervous system. ‘The endodermal

THE MALAYAN SPECIES OF ONYCHOPHORA. 81

layer is quite distinct from the central mass of yolk, though it contains food
material in the form of large round yolk-bodies, as well as numerous small
refringent bodies, which stain like the yolk.

Fig. 10 a.—This figure represents the twelfth transverse section in the
series, counting from the anterior end. It passes through the first
pair of somites (som.1), On the left, above the somite, there is a small
cavity which does not exist on the other side, where there are only a
few nuclei, which may possibly represent a rudiment in which a cavity
similar to the one already existing on the left side will appear later in
the development. The centre of the section is occupied by the endo-
derm (ez.) of the anterior end of the embryo, The thickened portion
of the ectoderm, situated on the latero-ventral aspect, is the rudiment
of the brain.

Fig. 10 4.—This figure represents the third transverse section behind the
one shown in Fig. 10 a. It passes through the anterior limit of the

central yolk (¢. y.) and the posterior limit of the endodermal wall of the
anterior end of the embryo. The small cavity found in Fig. 6 @ has
disappeared, a fact which proves that the cavity in question does not
represent the anterior portion of the second somite.

Fig. 10 ¢c.—This figure represents the tenth section behind the one shown
in Fig. 104. It passes through the posterior end of the first somite
(som.'), the middle of the second somite (som.?), and, on one side, the
anterior edge of the third somite (som.’), which passes forward above
the second somite. The central yolk (c. y.) is quite distinct from the
endodermal layer, and contains no nuclei. At the bottom of the
section, at a point which is situated quite close to tle anterior edge of
the blastopore, the endoderm is being recruited from the undifferen-
tiated cells (ex. iz.).

Fig. 10 d.—This figure represents the eighth section behind the one
shownin Fig. 10c. It passes through the anterior end of the blastopore
(ant. b.p.) and the second and third pairs of somites (som.?, som.*). It
shows the yolk (ev. y.) protruding from the blastopore and spreading
over the lower aspect of the section. The central yolk (¢. y.) and
endoderm (e7.) have the same appearance as in Fic. 6 ¢, the former
being absolutely devoid of nuclei. This section shows the ingrowth
at the edge of the blastopore to form the endoderm (ex. iz.). The
mesoblastic somites are quite distinct and separate from this ingrowth.

Fig. 10 e.—This figure represents the eighth section behind the one
shown in Fig. 10d. It passes through the region where the lips of the
blastopore have fused (md. b.p.), the third somite (som.*) on one side
and the fourth (som.*) on the other having been cut.

Fig. 10 £—This figure represents the ninth section behind the one shown
in Fig. 10¢. It passes through the posterior moiety of the blastopore

VoL. 45, PART 1.—NEW SERIES. F

82 RICHARD EVANS.

(po. b.p.), the fourth somite (som.*) on one side and the fifth on the
other (som.°) appearing in it. The yolk protrudes out of the blasto-
pore (ea. y.). The central yolk (c. y.) is absolutely free of nuclei, while
the endoderm (ez.) can be traced to the edge of the blastopore
(po. b.p.). The mesoblastic somites have no connection with the in-
vagination, which gives rise to the endoderm.

Fig. 10 g.—This figure represents the twentieth segment behind the one
shown in Fig. 10. It passes through the anterior end of the primitive
eroove. On the left side the rudiments of the seventh (som.’) and
eighth (som.’) somites are shown; on the other side, owing to the
obliquity of the section, the mesoderm band (mesod.) is represented.
The external yolk spreads in a backward direction along the primitive
groove and ventral surface.

Fig. 10 2.—This figure represents the fourth section behind the one
shown in Fig. 10g. It passes through the primitive groove (pr. g.) and
the individual mesodermal bands at their thickest part (mesod.). The
central yolk is still present and devoid of nuclei.

Fig. 10 4.—This figure represents the eighth section behind the one
shown in Fig. 10 4. It passes through the endoderm (ez.) just at the
posterior limit of the central yolk.

PLATE 7.

Fics. 11 a—1l d (x 120).—These figures represent four transverse
sections of the embryo shown in Fig. 3, in which rudiments of all the legs
have appeared.

Fig. 11 a.—This figure represents the thirty-first section, from the
anterior end, of a transverse series. It passes through the first somite
(som.1), the inner wall of which has thickened considerably to form the
rudiments of the muscles of the stomodssum (stomod.), the anteriorly
directed loop of which appears in the section represented. The brain
(dr.) has thickened considerably.

Fig. 11 4.—This figure represents the twenty-sixth section behind the
one shown in Fig. lla. It is slightly oblique, consequently it passes
through the second somite (som.?) on the right, and the third (som.3) on
the left. The rudiment of the jaw is shown on the right side, while
the anterior edge of the rudiment of the oral papilla appears on the
left side. It shows the central yolk divided into masses, each of which
possesses a nucleus, a condition which should be compared with that
shown in Figs. 10 a—10 &, in which the central yolk is devoid of nuclei.
The endodermal layer is distinct in both cases.

Fic. 11 c.—This figure represents the twelfth section behind the one
shown in Fig, 118. It passes through the third somite, that is the

THE MALAYAN SPECIES OF ONYCHOPHORA. 83

somite of the oral papillae. Even at this early stage the somite in
question has been divided into two parts, one of which is situated in
the rudiment of the oral papilla; the other, with its cavity almost
obliterated, lies in a dorso-lateral position.

Fig. 11 d.—This figure represents the twenty-eighth section behind the
one shown in Fig. ll ¢. It shows a somite being constricted through
the development of the rudiment of the muscle on its latero-ventral
aspect.

Figs. 12a—12/ (x 120).—These figures represent a selected number of
sections, from a transverse series, of an embryo of the same age as the one
shown in Fig. 4. The rudiments of all the appendages, twenty-five in
number, are present. The antenne consist of several rings more than they
do in the embryo shown in Fig. 3, sections of which are represented in Figs.
lla—ll1 d.

Fig. 12 a4.—This figure represents a section across the antenna, near their
base. The brain lobes are cut just in front of their union in the
middle line. The most noticeable feature is the divided condition of
the first somite (som.1). The two portions of the somite extend the
whole length of the antenne.

Fig. 12 6.—This figure represents a section across the base of the antennz
and through the brain lobes. It cuts the anterior wall of the for-
wardly directed loop of the stomodzeum, and shows the two subdivisions
of the first somite.

Vig. 12 c—This figure represents a section across the stomodeal region.
It shows the cavity of the first somite divided into two parts (som.').

Fig. 12 d.—This figure represents a section across the stomodeeal region.
It shows the first somite divided into two parts, a dorsal and a lateral.
The lateral portion opens into the exterior by means of a canal, the
external opening of which is situated internally to the rudiment of the
lip and in front of the rudiment of the jaw, the former being much
enlarged in this region. The Figs. 12 a—12d show the brain (47.) as
an enormous thickening of the ectoderm.

Fig. 12e—This figure represents a slightly oblique section across the
region of the rudiments which develop in the adult into the rings
which surround the mouth. The left half of the figure is anterior to
the right. It passes also through the canal which puts the first somite
in communication with the exterior. The endoderm (ez.) appears as a
ring round the stomodeum, which is developed by an invagination of
the ectoderm and which is surrounded by a mesodermal sheath derived
from the inner walls of the first pair of somites.

Vig. 12,—This figure represents a section from the region of the rudi-
ment of the jaws. The section shows the rudimentary mouth-folds

84 RICHARD EVANS.

(Jip) as thin ectodermal outgrowths situated above the developing
appendages. The second somite (som.?) is divided into median and
appendicular portions. The appendicular portion grows downwards
into the rudiment of the jaws, but is becoming obliterated through the
proliferation of its ventral wall to form the muscles of the jaws. The
central yolk is devoid of nuclei.

Fig. 12 7.—This figure represents a section from the region of the oral
papille. The section shows the third somite divided into median
and appendicular portions. On the left side the appendicular portion
grows towards the ventral aspect, and is situated on the outer side of
the rudiment of the nerve-cord. On the same side, at the tip of the
oral papilla, is represented the rudiment of the slime-gland.

lig. 12 4.—This figure represents a section from the region of the ninth
somite, which is divided into median and appendicular portions. ‘The
appendicular portion on the right side grows towards the ventral
surface, but it does not yet open to the exterior.

Fig. 12 7.—This figure represents a section from the region of the eleventh
somite, which is undivided on the left side.

Vig. 12 4.—This figure represents a section from the region of the penul-
timate somite, which is provided with an immense cavity. The section
shows the inner end of the proctodzum cut across.

PLATE 8.

Vic. 138 (x 170).—This figure represents a transverse section, passing
through the fourth somite (som."%) from the posterior end of the embryo
shown in Fig. 8. The last two pairs of somites in the embryo in question are
still solid. Note especially the germinal nuclei (ge. 2.) situated in the
splanchnic walls of the somites. The left somite in the figure has been cut
through the middle, and shows a group of germinal nuclei which have as yet
scarecly assumed the structural and staining characters of the nuclei in
question. On the right side of the figure the section does not pass through
the middle of the somite, but slightly in front of it, and at the dorsal corner it
actually cuts through the wall of the somite situated in front. The outline of
the endoderm, which possesses irregularly shaped nuclei, is quite sharp and
distinct from that of the mesodermal somite.

Fic. 14 (x 170).—This figure represents a transverse section, passing
through the fourth somite from the posterior end (som."-8), of an embryo
slightly younger than the one shown in Fig. 4. The last two pairs of somites
in the embryo in question have not yet developed a celomic cavity. Note
especially the germ nuclei situated in the splanchnic walls of the somites, and
also note that, as regards size and structure, they represent all stages
transformation, from the ordinary mesodermal nuclei to that of fairly advance
germinal nuclei. The somite shown on the right side of the figure exhibits a

THE MALAYAN SPECIES OF ONYCHOPHORA. 85

rudimentary myotome, and consequently a developing outgrowth, which later
on becomes separated off as the appendicular portion of the somite.

Figs. 15 a4 and 154 (x 170).—These figures represent transverse sections
of an embryo about the same age as that shown in Fig. 5. The generative
portions of the somites situated on either side have almost come together on
the dorsal aspect of the mid-gut. The splanchnic walls, which, owing toa
slight change of position, have become ventral, have considerably thickened,
and the germinal nuclei are gradually becoming excluded from the ccelomie
cavity, that is the ovarian cavity.

Fig. 15 a.—This figure shows a section passing through the fourth somite
from the posterior end (som.”*).

Fig. 15 4.—This figure shows a section passing through the fifth somite
from the posterior end (som."-*).

‘Fics. 16a, 164, and 16¢ (XxX 170).—These figures represent transverse
sections of an embryo slightly older than that shown in Fig. 5.

Fig. 16 a.—This figure shows a section passing through the third pair of
somites from the posterior end (som."*), that is the somite of the
genital ducts. ‘The somites in question, which are devoid of germinal
nuclei, have not yet entered into communication on the dorsal aspect,
nor have they at their outer ends fused with the ectoderm of the
ventral surface. Note the large blood-spaces situated on either side
and ventral to the proctodeum.

Fig. 16 6.—This figure shows a section passing througli the fourth pair of
somites from the posterior end (som. ™%), Note the germinal ridges
(ge. r.) placed in the ventral wall of the ovary, and the double ovarian
cavity situated above the ridges. The whole structure is fused to—if

~ it does not form a part of—the ventral wall of the pericardium (p.c.).

Vig. 16 c.—This figure shows a section passing through the fifth pair of
somites from the posterior end (som."*). It presents the same
characters as the somite shown in the previous figure, except that the
germinal ridges are much more strongly developed.

Fics. 17 a, 17 6, and 17 ¢ (x 170).—These figures represent sections of an
embryo which is considerably older than the one sections of which are illus-
trated in Figs. 16 a, 4, and ec.

Fig. 17 a.—This figure shows a section passing through the third pair of
somites from the posterior end (som."-*). ‘Tle somites have fused on
the dorsal aspect to form a kind of median chamber, the walls of which
are devoid of germinal nuclei.

Fig. 17 6.—This figure shows a section passing through the fourth pair of
somites from the posterior end (som.*). It illustrates the germinal
ridges (ge. 7.) situated on the ventral wall of the ovary, and separated
from the dorsal wall by the slit-like ovarian cavity (ov. c.), and the

86 RICHARD EVANS.

developing ova surrounded by follicle cells derived from the ventral
wall of the ovary, and hanging freely in the heemoccele.

Fig. 17 ¢.—This figure shows a portion of the oviduct in the section suc-
ceeding that illustrated in Fig. 17a. Note the rudiment of the recep-
taculum ovorum (rec. 0.) developed as a simple outgrowth from the
wall of the oviduct.

Fics. 18 a, 18 4,18 c,and 18 d(x 170),—These figures represent four stages
in the growth of the ovarian ovum.

Fig. 18 a.—This figure shows an ovum in which the eytoplasm possesses
a beautifully alveolar structure. The nucleus, placed near the centre
and circular in outline, is provided with a nucleolus which is alveolar
in character.

Fig. 18 4.—This figure shows an ovum very similar in structure to the
one shown in the previous figure, but of greater dimensions in all its
parts.

Fig. 18 ¢.—This figure shows an ovum in which the cytoplasm is occupied
by a large number of globules which are almost uniform in size, and
are situated in the alveoli shown in Figs. 18 @ and 18 4.

Fig. 18 d.—This figure shows a transverse section of an almost fully
crown ovarian ovum. ‘Ihe globules shown in the foregoing figure
have fused to form compound systems. The nucleus has wandered
from the centre towards the periphery, and consequently has become
irregular in outline. Note also the disposition of the yolk-bodies with
regard to the nucleus.

Fics. 19a, 19 6, and 19¢ (x 170).—These figures represent sections of an
embryo slightly older than that sections of which are illustrated in Figs. 16 a,
b, and ec.

Fig. 19 a—This figure shows a transverse section through the third pair
of somites from the posterior end (som."*). The vasa deferentia open
to the exterior by a short common duct, a condition which is per-
manent in the female (vas. def).

Fig. 19 4.—This figure shows a cross-section of the left vas deferens
near its inner end, and a section through the funnel of the right vas
deferens.

Fig. 19¢.—This figure shows a section through the developing testes,
showing the germinal nuclei filling the internal cavity, and surrounded
by an epithelial layer. Both testes are fused to the ventral wall of the
pericardium, and are situated close together, much in the same way
as the developing ovaries are in Fig. 16c.

Vics. 20 a and 20 4 (x 170).—These figures represent sections of a male
embryo of the same age as the one shown in Fig, 5.

THE MALAYAN SPECIES OF ONYCHOPHORA. 87

Fig. 20a.—This figure shows an oblique section which passes through
the posterior end of the anal groove (az. gr.), and on the left cuts the
rudiment of the last appendage along its whole length. Note the last
somite (m. a. g., som.”).

Fig. 20 6.—This figure shows a section in a similar direction and parallel
to the one illustrated in Fig. 20a. Note the duct of the male accessory
gland (m.a.g.) and the renal organ (rez. or.), derived from the last
and penultimate somites respectively.

Figs. 21 a and 214 (x 120).—These figures represent sections of a male
embryo of the same age as the one illustrated in Fig. 6 on Plate 5,

Fig. 21a.—This figure shows a section of the male accessory gland
(m.a.g.) and of the last ventral organ (vex. 0.), which is vestigial in
character.

Fig. 21 6.—This figure shows a section of the last renal organ (rez. or.),
the male accessory gland (m. a. g.), and the last fully developed
ventral organ (ven. 0.).

PLATE 9.

Fics. 22a, 224, and 22c (x 50).—These figures represent the thirty-
second, the fifty-ninth, and the eighty-first sections respectively of a transverse
series of the brain of an embryo which was slightly older than the one shown
in Fig. 5, but younger than that illustrated in Fig. 6 on Plate 5.

Fig. 22 a.—This figure shows a section passing through the archicerebral
(a. Jo.) and the first (20.1) or antennal lobe of the brain, as well as the
first pair of ventral organs (v. 0.') near their posterior limit. On the
right side the eye is shown; but on the left the section passes in front of
that organ. On the ventral aspect the lips situated on either side of
the mouth, with the tongue passing down between them, are repre-
sented.

Fig. 22 4.—This figure shows a section passing through the archicerebral
(a. Jo.) and the third (Zo.5) lobe of the brain; the second (v. 0.7) and the
divided third (v. 0.5% and v. 0.**) ventral organs. On the right side it
passes along the anterior border of the para-cesophageal cord; on the
left it passes in front of that cord.

Fig. 22 ¢c.—This figure shows a section passing through the third (0.3)
and fourth (Zo.4) lobes of the brain, as well as the fourth ventral organ
(v. 0.4), which corresponds to the anterior pair of walking appendages.
On the right side it passes behind the para-cesophageal cords; on the
left the para-cesophageal cord is cut along its whole length. Note
that on the left side the fourth lobe (/o.4) of the brain, which lies on
the dorsal half of the para-cesophageal cord, is pierced by the fibres of
the mandibular nerve,

88 RICHARD EVANS.

Fics. 23 a and 234 (x 120).—These figures represent two oblique sections
of the head of an embryo.of the same age as the one sections of which are
shown in Figs. 22 a, 6, and e.

Fig. 23 a.—This figure shows a section passing dorsally near the median
plane, and ventrally above the antenna and the eye. It shows the
archicerebral lobe (a. /o.) above, and the three anterior brain lobes
below (/o.', Jo.?, Jo.3). In another section, quite close to the one repre-
sented, the continuation of the second and third lobes towards the
dorsal aspect, shown in the figure as one lobe, was distinctly divided
into two. The fourth lobe of the brain does not appear in the
section.

Fig. 23 4.—This figure shows a section passing through the head in such
a way as to cut the first, second, and third lobes (Jo.1, Jo.?, Jo.) of the
brain on one side, and the fourth (/o.4) on the other. It passes
through the first ventral organ (v. 0.) of the left side, the second one
(v. 0.2) of the right side, and the anterior portion of the third (v. 0.34),
which is median in position and produced by the fusion of rudiments
from either side.

Fie. 24 (x 50).—This figure represents an oblique section of the antero-
dorsal aspect of the brain of an embryo of the same age as the two previous
ones. It shows the small but distinct archicerebral lobe (a. Jo.) situated on
the left side of the median line; on the right side the section passes nearer to
the ventral aspect, and consequently cuts the first or antennal lobe of the
brain (/o.!). ;

Ties. 25a, 25 6, and 25¢ (xX 170).—These figures represent three stages in
the development of the eye, which is derived from the ectodermal thickening
which gives rise to the brain. During all the stages of development the two
organs are connected to each other, the optic nerve being formed in situ.

Figs. 26 a, 26, 26 c, 26d, and 26 e.—The first four of these figures are dia-
grammatic representations of four stages in the development of the recep-
taculum ovorum and receptaculum seminis ; but the last shows the structure
and arrangement of these organs in the adult. The figures were obtained
from reconstructions made from series of sections, and have been modified so
far as to show the receptaculum ovorum lying in the same plane as the recep-
taculum seminis, which seems never to be the case in nature.

A.S Huth, Lith® London.

P Bayzand ad nat del.

f
Aunt SourrMicr’ See Vol SENS GLE.

A.S.Huth, Litht London.
REvans ad.nat-del

REvans ad nat del

ec.

Oo
5,

an

Luart BournMior Seu Vol 15s P, 7

corre

ieee
eoberh

AS Huth, Lith’ London,

Quart. Sourn. Muer Sa, WALENSIL 8

R.Evans ad nat del

Luart Soiree Nei sus.

adnat de} :
AS Hath, Lith’ London

With Ten Plates, Royal Ato, 5s.
CONTRIBUTIONS TO THE KNOWLEDGE OF RHABDOPLEURA
AND AMPHIOXUS.
By E. RAY LANKESTER, M.A., LL.D., F.R.S.
London: J. & A. Cuurcuitt, 7, Great Marlborough Street.

Quarterly Journal of Microscopical

Science.

The SUBSCRIPTION is £2 for the Volume of Four Numbers;
for this sum (prepaid) the JournaL is sent Post Free to any part
of the world.

BACK NUMBERS of the Journat, which remain in print, are
now sold at an uniform price of 10/-.

The issue of Suprrement Numpers being found inconvenient,
and there being often in the Kditor’s hands an accumulation of
valuable material, it has been decided to publish this Journal at
such intervals as may seem desirable, rather than delay the appear-
ance of Memoirs for a regular quarterly publication.

The title remains unaltered, though more than Four Numbers
may be published in the course of a year.

Kach Number is sold at 10/-, and Four Numbers make up a
Volume.

London: J. & A. CHURCHILL, 7 Great Marlborough Street.

TO CORRESPONDENTS.

Authors of original papers published in the Quarterly Journal
of Microscopical Science receive twenty-five copies of their
communication gratis.

All expenses of publication and illustration are paid by the
publishers.

As a rule lithographic plates, and not woodcuts, are used in
illustration.

Drawings for woodcuts should nor be inserted in the MS.,
but sent in a separate envelope to the Editor.

Contributors to this Journal requiring extra copies of their
communications at their own expense can have them by applying
to the Printers,

Messrs. Aptarp & Son, 224, Bartholomew Close, E.C., on
the following terms :

For every four pages or less—

25 copies é : 5/-

D0.” sss . 6/-

EOD 25 . : 7/-
Plates, 2/- per 25 if uncoloured; if coloured, at the same rate for

every colour.

Prepayment by P.O. Order is requested.
ALL COMMUNICATIONS FOR THE EDITORS TO BE ADDRESSED TO THE CARE
or Messrs. J. & A. Cuurcuitn, 7, Great MarLeorouau Srreet,

Lonpon, W.

THE MARINE BIOLOCIGAL ASSOCIATION

OF THE

UNITED KINGDOM.
20:
Patron—HIS MAJESTY THE KING.
President—Prof. E. RAY LANKESTER, LL.D., F.RS.

BOE:

THE ASSOCIATION WAS FOUNDED “ TO ESTABLISH AND MAINTAIN LABORATORIES ON
THE COAST OF THE UNITED KINGDOM, WHERE ACCURATE RESEARCHES MAY BE CARRIED
ON, LEADING TO THE IMPROVEMENT OF ZOOLOGICAL AND BOTANICAL SCIENCE, AND TO
AN INCREASE OF OUR KNOWLEDGE AS REGARDS THE FOOD, LIFE CONDITIONS, AND HABITS
oF BRI’vISH FOOD-FISHES AND MOLLUSCS.”

The Laboratory at Plymouth
was opened in 1888. Since that time investigations, practical and scientific, have
been constantly pursued by naturalists appointed by the Association, as well as by
those from England and abroad who have carried on independent researches.

Naturalists desiring to work at the Laboratory

should communicate with the Director, who will supply all information as to
terms, &e,

Works published by the Association

include the following :—‘ A Treatise on the Common Sole,’ J. T. Cunningham, M.A.,
4to, 25/-. ‘The Natural History of the Marketable Marine Fishes of the British
Islands,’ J. T. Cunningham, M.A., 7/6 net (published for the Association by
Messrs. Macmillan & Co.).

The Journal of the Marine Biological Association

is issued half-yearly, and is now in its fifth volume (price 3/6 each number). ae

In addition to these publications, the results of work done in the Laboratory |
are recorded in the ‘ Quarterly Journal of Microscopical Science,’ and in other
scientific journals, British and foreign.

Specimens of Marine Animals and Plants,

both living and preserved, according to the best methods, are supplied to the
principal British Laboratories and Museums. Detailed price lists will be forwarded = |
on application. ‘ae

* ageterrnr ©

TERMS OF MEMBERSHIP.

ANNUAL MEMBERS . : ; - £1 1 Operannun.
Lire MEMBERS . : : ; . 15 15 O Composition Fee.
FOUNDERS . LOO OmaG a *

GOVERNORS (Life Members of Council) 500 0 0 = BS

Members have the following rights and privileges:—They elect annually the
Officers and Council; they receive the Journal free by post; they are admitted to
view the Laboratory at any time, and may introduce friends with them; they have the
first claim to rent a table in the Laboratory for research, with use of tanks, boats, &c.;
and have access to the Library at Plymouth. Special privileges are granted to Governors,
Founders, and Life Members.

Persons desirous of becoming members, or of obtaining any information with
regard to the Association, should communicate with—

The DIRECTOR,
The Laboratory,
_ Plymouth.

New Series, No. 178 (Vol. 45, Part 2). Price 10s.

NOVEMBER, 1901.

"152

THE

QUARTERLY JOURNAL

OF

MICROSCOPICAL SCIENCE.

EDITED BY

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

HONORARY FELLOW OF EXETER COLLEGE, OXFORD; CORRESPONDENT OF THE INSTITUTE OF FRANCE
AND OF THE IMPERIAL ACADEMY OF SCIENCES OF ST. PETERSBURG, AND OF THE ACADEMY
OF SCIENCES OF PHILADELPHIA; FORKIGN MEMBER OF THE ROYAL BOHEMIAN
SOCIETY OF SCIENCES, AND OF THE ACADEMY OF THE LINCEI OF ROME;
ASSOCIATE OF THE ROYAL ACADEMY OF BELGIUM, HONORARY MEMBER
OF THE NEW YORK ACADEMY OF SCIENCES, AND OF THE
CAMBRIDGE PHILOSOPHICAL SOCLETY, AND OF THE ROYAL
PHYSICAL SOCIETY OF EDINBURGH 5 HONORARY
MEMBER OF THE BIOLOGICAL SOCIETY
OF PARIS;
DIRECTOR OF THE NATURAL HISTORY DEPARTMENTS OF THE BRITISH MUSEUM; LATE FULLERIAN
PROFESSOR OF PHYSIOLOGY IN THE ROYAT INSTITULION OF GREAT BRITAIN.

WITH THE CO-OPERATION OF

ADAM SEDGWICK, M.A., F.RS.,
FELLOW AND TUTOR OF TRINITY COLLEGE, CAMBRIDGE;

W.F. he WHLDON, M.A., F-B.S.,

LINACRE PROFESSOR OF COMPARATIVE ANATOMY AND FELLOW OF MERTON COLLEGE, OXFORD;
LATE FELLOW OF ST. JOUN’S COLLEGK, CAMBRIDGE;

AND

SYDNEY J. HICKSON, M.A., F.R.S.,

BEYER PROFESSOR OF ZOOLOGY IN THE OWENS COLLEGE, MANCHESTER.

WITH LITHOGRAPHIC PLATES AND ENGRAVINGS ON WOOD.

K
“LONDON:
J. & A. CHURCHILL, 7, GREAT MARLBOROUGH STREET.
1901.

| ET ee ee ee Ee ee Ee ee

Adlard and Son,] : [Bartholomew Close.

CONTENTS OF No. 178.—New Series.

MEMOIRS:

PAGE
The Lateral Sensory Canals, the Eye-Muscles, and the Peripheral
Distribution of certain of the Cranial Nerves of Mustelus levis.
By Epvwarp Puexrs Autis, jun. (With Plates 10—12) . Shes sul

The Anatomy of Scalibregma inflatum, Rathke. By J. H. Asu-
wortH, D.Sc. (With Plates 13—15) . 2 ‘ ; : » 237

On the Pelvic Girdle and Fin of Eusthenopteron. By Epwin 8.
Goopricu, M.A., Fellow of Merton Coilege, Oxford. (With

Plate 16) dll

MUSTELUS LAVIS. 87

The Lateral Sensory Canals, the Eye-Muscles,
and the Peripheral Distribution of certain of
the Cranial Nerves of Mustelus levis.

By

Edward Phelps Allis, jun.
With Plates 1O—12.

I nave recently had occasion to examine three series of
sections of the head of embryos of Mustelus levis, one of
them being an embryo 12°2 cm. long. It was not my inten-
tion, when these sections were prepared, to make any extended
study either of the lateral canals or of the cranial nerves of
the fish ; the investigation I had proposed relating entirely to
the innervation of the sensory organs of the ampulle. I have
long had a very decided impression, opposed to that of most
workers on the subject, that these ampullary organs must be
genetically related to the terminal buds of ganoids and
teleosts rather than to the pit organs of those fishes; and I
thought that I should easily be able to get some positive
evidence of this in the general course and position of the
nerves that innervate them in advanced selachian embryos.
This positive evidence I have wholly failed to get, for the
very simple reason that, in the main nerve trunks, I could not
distinguish in my sections the ampullary fibres from the
lateral canal ones. Disappointed in this at the very begin-
ning of the investigation, I nevertheless decided to quite
carefully trace the lateral canals and the nerves that imner-
vate them and the ampullz, as far back as my sections went,
that is, nearly to the level of the first gill slit. Careful con-
sideration of these observations has fully convinced me,

VOL. 45, PART 2.—NEW SERIES. H

io 6)

8 EDWARD PHELPS ALLIS, JUN.

though indirectly, that the ampullary organs do represent
the terminal buds of ganoids and teleosts, and not the pit
organs. As, in this research, I was also led to trace the
other cranial nerves of the region under consideration, and as
my observations differ in certain respects from, and complete
in others, the results of earlier writers on the subject, I have
thought best to fully describe, not only the lateral canals of
the head and the ampullary canals, but also the facial, tri-
geminal, and eye-muscle nerves, notwithstanding the fact
that there will necessarily be, in these descriptions, a certain
amount of repetition of well-known facts.

The embryos used for the investigation were kindly sent
me by the Naples Zoological Station, and varied from 36 mm.
to 12:2 cm. in length. As I have no one at present in my
laboratory who could properly section these embryos for me,
I appealed for help to Prof. G. B. Howes, of the Royal College
of Science, London, and he most kindly undertook to have
them sectioned, under his personal supervision, by his pupil,
Mr. H. H. Swinnerton. Sections of 36 mm. and 55 mm.
embryos were first prepared, but these embryos were found
to be much too young for the purpose, and one of the largest
ones I had—a 12:2 em. one—was selected. This last size
proved an excellent one for the purpose in view, and the
following descriptions relate entirely to it unless otherwise
stated. Two different specimens of this age were sectioned,
one from the anterior end of the snout back nearly to the
first gill slit, and the other from the hind edge of the eye
back a certain distance beyond the spiracle. This second
specimen was sectioned in the hope that I might in it deter-
mine the ultimate distribution of the dorsal branch of the
elossopharyngeus, which I had been unable to follow in the
first specimen. I unfortunately could not follow it in the
second specimen either, the tissues being slightly broken at
a place where the nerve apparently enters the lateral edge of
the cranial extension of the trunk muscles, and the nerve
there lost in the displaced muscle fibres.

All of the embryos sectioned were, at Prof. Howes’

MUSTELUS LAVIS. 89

suggestion, double-stained in Khrlich’s hematoxylin and
Griible’s orange, the combination which he and Mr.
Swinnerton had found so successful in their recent work
on Sphenodon (88). The two large embryos were cut, the
first one 7 in thickness, and the other 10, in thickness.
The first series contained something over 3000 sections.

After the full course of the lateral canals of the head had
been first traced in the sections of the 55 mm. embryo, and
then in those of the 12:2 cm. one, I had them traced by dis-
section in two or three others of my larger remaining
embryos. This was at first undertaken simply as an aid and
guide in the preparation of the simple outline drawings
intended for illustration. Mr. Nomura, my assistant, under-
took this work, and he soon found that he could, with some
care and trouble, trace, not only the main lateral canals, but
also the ampullary canals, which latter it had been wholly
impracticable to trace in sections. The drawings made from
these dissections accordingly show the general course of all
the ampullary tubes, and the exact, or closely approximate,
number and position of their surface openings. They do not
show all the tubules of the lateral canals, and none of the
numerous surface openings, or pores, of these canals are
even indicated in the general drawings. It was found that
these tubules and pores could not be accurately made out
without much more work than the subject seemed to
warrant. The drawings accordingly only show, accurately,
those tubules that project to one side or the other of the
canals, but few of the many tubules that run directly out-
ward from the canals to the external surface being even
indicated. Fig. 7 shows the exact number and position
of the pores and tubules in a part of the suborbital canal
where they are particularly numerous, and this, with the
general drawings, will give a sufficiently good idea of their
arrangement elsewhere.

The methods employed in these dissections were, to
examine the undissected head by slanting lamplight, which
brought out the pores; to scrape off first the delicate

90 EDWARD PHELPS ALLIS, JUN.

epidermal layer of the ectoderm, and then the entire ecto-
derm, thus first exposing the tubules and then the canals ;
and finally to “skin” a head in a single piece, and examine
it in glycerine by transmitted light. Fig. 7 is made from
such a preparation.

Lateral and Ampullary Sensory Systems.

The manner in which the lateral canals develop in sela-
chians is evidently quite different, in certain respects, from
that that pertains in ganoids and teleosts. What it is I can-
not make out from my three series of sections, and I do not
find in the works at my disposal any description of it that
seems complete or satisfactory.

Balfour (8) leads one to suppose that there is in selachians
simply an abbreviation of the process that gives origin to the
canals in ganoids and teleosts. This abbreviation consists in
that it is the inner row of cells (Schleimschicht) alone of the
epiblast that is concerned in the involution that forms the
canal, the epidermis remaining always as a flat and even
layer above it. The canal, thus formed, then becomes de-
tached from, and sinks beneath, the epiblast, remaining
attached to it at certain points by cords of tissue which
represent the primary tubes of the canal. Opposite the outer
ends of these primary tubes the epidermis then becomes
perforated, thus giving rise to the primary pores.

The ampullary tubules are said by Balfour to arise in
exactly the same manner as the lateral canals.

Balfour says that the lateral canal of the body, developed
as above set forth, first appears at the hind end of the lateral
sensory line, and extends forward from there. Mitrophanow
(44, pp. 208-9) confirms the latter part of this statement,
but the involution of the canal as he describes it seems to
involve the entire ectoderm, and not simply its deeper, inner
layer. Moreover, it would seem from his figures as if the
section of canal said to be thus first formed was nothing more
nor less than what Clapp (10, p. 239) refers to as “a curious

MUSTELUS LAVIS. 91

fold of the epidermis, the so-called ‘ pocket,’ which covers the
growing end of the line.” Whether or not this ‘* pocket” re-
presents the first beginnings of the lateral canal, it is evident
that a canal that thus first appeared at the hind end of the
lateral sensory line must, appearing as it does before the line
reaches the hind end of the body, be continually being pushed
bodily backward, following and accompanying the growing
end of the line. There is thus here such an important depar-
ture from what is found in Amia, that I should hesitate to
accept the account as correct.

My own observations, limited to the four series of sections
that I possess, tend to confirm Balfour’s statement that the
lateral canals are formed by an involution of the deeper layer
only of the ectoderm. The covered gutter, rather than canal,
that is thus first formed then becomes, by a pressing together
of its walls, a sharp and apparently solid ridge projecting
inward from the inner surface of the ectoderm. This is the
condition found nearly everywhere in my 36 mm. embryo.
Where the cord is deepest it receives a branch from the
underlying and related lateral nerve, but there is as yet no
perceptible indication in my sections of a definite sensory
organ related to this nerve. Between each two consecutive
points where these nerves thus join the cord the cord becomes
less deep, and in the regions where it is the flattest there isa
small pit-like depression on the outer surface. This depres-
sion looks in certain sections like a shallow groove, while in
others it has vertical sides, and seems to cut clean through
the outer layer of ectoderm down to that deeper layer that is
alone directly related to the canal itself. This depression, or
pit, quite certainly represents a future pore, the pore thus
apparently appearing while the cord that represents the canal
is still everywhere attached to the ectoderm.

In my 55 mm. specimen the lateral canals were mostly
found as cords of tissue lying in the mesoderm, beneath the
ectoderm, and connected with the latter at intervals by
smaller cords, which represent the tubules of the adult. In
the large main cords there was a small central lumen, either

92 EDWARD PHELPS ALLIS, JUN.

formed or in process of formation, and this lumen was always
most fully developed opposite the points from which the
cords representing the future tubules arose. In certain
places it extended outward a short distance in these latter
cords, but in no place did it reach the outer surface of the
head. On this outer surface of the head there was, however,
almost invariably, opposite the outer ends of the cords that
represent the primary tubes, a slight slit-like depression, the
appearance being that of a pre-existing opening that had
been closed by the pressing together of its walls.

Even at this age, 55 mm., there was no_ perceptible
indication, in my sections, of the sensory organs in the main
canals, and the cords that represent the future tubules had
already begun to branch, and formed in certain regions a
somewhat complicated system. How these branching systems
arise could not be traced in my material, but it would seem
as if they must arise by the repeated dichotomous subdivision
of a single primary cord, exactly as the branching tubules of
Amia arise from a single primary tube (2).

The ampulle in my 55 mm. embryo were nearly all repre-
sented by small teat-like processes that arose from the inner
surface of the ectoderm, and projected into the underlying
tissues. Some of these processes seemed solid, while others
contained a small central lumen which sometimes led to the
outer surface, the process then appearing as a sharp fold of
the entire ectoderm. A small nerve was easily traced to the
inner end of each process. While no attempt was made to
trace the complete and definite distribution of these little
processes, it was easily to be seen that in certain places they
had exactly the relations to the lateral canals that the
surface pores of the ampullary tubes have in the 12:2 em.
embryo. This seemed to me to indicate that it must be the
pore in the adult, and not the ampulla, that indicates the
place of origin of the structure. Here, then, from the
primary distribution of these organs, as indicated by their
surface pores, was perhaps a manner of determining whether
they arose from pit-organs or from terminal buds.

MUSTELUS LAVIS. 93

To confirm my conclusion that the surface ampullary pore
represents, approximately, the place of origin of the am-
pullary organ, I sought in my younger embryos for the lines
of organs that should represent certain regular, constant,
and well-marked lines of ampullary pores in my larger
embryos. It will be sufficient to describe asingle one of them.

In my 12:2 cm. embryo there is on each side of the dorsal
surface of the head, and slightly anterior to the external
opening of the endolymphatic duct, a regular curved line of
ampullary pores. The tubes leading from these pores run
forward to a sub-group of the superficial ophthalmic group of
ampulle, these ampulle lying on the dorsal surface of the
nasal capsule. In the 55 mm. embryo there was, in exactly
the place occupied by these pores in the older embryo, a line
of surface sense organs which greatly resemble, in certain
respects, the pit-organs in larve of Amia, while in others
they greatly differ from those organs. My material was not
adapted to a histological study of them, but it may be said
that the organs were represented by a series of processes
arising from the inner surface of the ectoderm, each process
enclosing a little space which may or may not be in direct
communication with the exterior. The processes all turn
anteriorly, parallel to the overlying ectoderm, and a small
nerve enters each process at its deeper or anterior end.
These short processes of this 55 mm. embryo thus unques-
tionably represent the complete ampullee of the 12°2 cm. one.
The long ampullary tube that is found in the latter embryo
must then be formed by an exceedingly rapid growth of the
short process of the younger one, that process being, so to
speak, stretched out mto a long tube between the fixed point
represented by its surface opening and another relatively
fixed one, represented by the point where the sensory nerve
enters the process. The tube apparently offers less resistance
to this stretching process than the nerve does.

Further evidence that the ampullary pore does not usually
travel far away from its place of origin is found in the fact
that certain lines of these pores are frequently found on the

94 EDWARD PHELPS ALLIS, JUN.

side of a lateral canal opposed to that on which the ampull
themselves lie, the ampullary tubes passing internal and not
external to the lateral canal. As the lateral canals in Mustelus
are certainly already present as well-developed cords before
the ampullary tubes are developed, it is evident that the
latter tubes, in the cases above referred to, would have had to
cut through the canal to attain their adult position, if the pore
of the tube travelled from its place of origin in anything
resembling the manner that the lateral pores of Amia do (2).
An extreme case is shown in Garman’s (21) figure of Raia,
where long ampullary tubes open along the edge of the body
lateral to the lateral canals, the tubes passing internal to the
canals. The ampulle related to these tubes and pores must
certainly have had their place of origin in the immediate
neighbourhood of these pores of the adult, and from there
they must have travelled to their adult position by pushing,
or being pulled, through the tissues internal to the canals.

‘The topographical position of a group of ampulle in the
adult selachian is thus not necessarily any indication what-
ever of the point of origin of the several ampulle that form
the group, the ampullary organs differing radically in this
from the organs of the lateral canals. They also apparently
differ from the latter organs in that they have a later and
relatively much more rapid development. ‘his is shown in
the rapid development of the ampullary tubes just above
referred to, between the ages represented by the 55mm. and
12:2 cm. embryos. In my 36 mm. embryo I could not even
find any positive indication of this hne.

In Amia (2), and probably in all teleosts also (67), the
sense organs of the lateral system lie at first below the
outer surface of the ectoderm, along a cord of cells that is
differentiated in the deeper layer or layers of the ectoderm.
Immediately superficial to the central point of each organ
there is a large and specialised cell, which later becomes a
vacuole. As the-sense organ pushes through the overlying
cells to its final exposed relation to the outer surface, this
vacuole must, at a certain stage, become a small pit-like

MUS'TELUS LAVIS. 95

depression leading to the outer surface of the head. At
and immediately before this stage this organ would strongly
resemble in general appearance the ampullary pits in the
temporal region of my 55 mm. Mustelus. It would also
present somewhat the appearance of the pit-like depressions
that I have assumed to represent the future pores of the
canal lines in 36 mm. embryos of Mustelus, but my material
was not suited to the determination of the homologies here
involved. All that I could make out was that the canal organs
seem never, in Mustelus, to become exposed on the outer
surface of the head, as they do in Amia. On the contrary,
they seem to always remain in the deeper layer of the
ectoderm, where they arise, and then to split off from that
layer, enclosed in, and as part of, a long and nearly solid
cord, which later becomes a canal. The ampullary organs,
in somewhat marked distinction to the canal organs, may
become exposed at the bottom of a little pit apparently formed
by the bursting, so to speak, of the little vacuole that forms
above the central portion of the organ. The general form and
appearance of the several organs in my embryos thus give no
definite indication as to whether the ampulle are developed
from pit-organs or not. ‘They, however, seem to indicate a
difference between ampullary and lateral canal organs.

No organs in any way resembling the terminal buds of
ganoids and teleosts were anywhere observed on the outer
surface of any of my specimens, but in the large number of
sections examined, and especially as I was not particularly
looking for these organs at the time, it is certainly possible
that there may have been some, and that they escaped my
notice. ‘I'here were, however, in the 12:2 cm. embryo certain
other sensory organs found both on the outer surface of the
head and on the body. They closely resemble the organs
that in my 55 mm. embryo represent certain of the super-
ficial ophthalmic group of ampulle, but differ from those
organs in that their central lumen in every case leads directly
to the outer surface by a pit-like depression. The distribution
of these organs will be given in describing the ampulle.

96 EDWARD PHELPS ALLIS, JUN.

Infra-orbital Canal.

‘he general course and position of all the lateral canals of
the adult Mustelus has been well given by Garman (21).
That author makes use of a special nomenclature which I
shall adopt only when wishing to designate the various
sections of the several canals, elsewhere making use of the
nomenclature now ordinarily employed by other authors. The
term infra-orbital canal will be used to designate that part
of the so-called main infra-orbital canal of my earlier works
that is innervated by the buccalis and oticus facialis. In
Mustelus the section so innervated extends from the anterior
end of the canal back to the supratemporal cross-commissure.
The point at which this commissure arises from the main line
in different fishes is, as will be later explained, apparently
not a fixed one.

The infra-orbital canal of Mustelus, in my 12°2 cm. embryo,
and also in the adult, begins near the lateral edge of the
anterior end of the ventral surface of the snout, and there
communicates directly with the supra-orbital canal. From this
point the canal first runs mesially a very short distance, and
then turns backward in a short curve. It then continues
backward and somewhat mesially, and reaches the middle
line of the head somewhat in front of the transverse level of
the nasal aperture, curving gradually mesially shortly before
reaching this point. There it anastomoses completely with
its fellow of the opposite side, the two canals united
turning sharply backward in the median line. At about
the transverse level of the middle of the nasal aperture the
two anastomosed canals separate, each canal turning sharply
laterally, and then curving slightly forward until it reaches
the very edge of the nasal aperture. There it turns back-
ward and slightly laterally at a sharp angle, and, curving
gradually more and more laterally, passes posterior to the
nasal aperture toward the lateral edge of the snout. Before
reaching that edge, and not far from it, it makes a double

MUSTELUS LAVIS. Sf

bend. It first turns laterally and slightly forward, and
connects with the distal end of the supra-orbital canal,
this bend being short and being apparent in sections, but
not in dissections. It then turns sharply backward, in the
line of, and apparently as a direct continuation of, the
supra-orbital canal, and continues a short distance almost
directly backward, parallel to the lateral edge of this part
of the head. It then turns laterally and forward in a short
rounded angle, the hyomandibular canal arising at this bend
almost as a direct continuation backward of the infra-orbital
canal anterior to the bend. Running laterally and forward
a short distance the infra-orbital canal reaches the lateral
edge of the head, where it turns upward and forward on
to the lateral surface of the head, and, continuing in this
same direction, soon reaches a point approximately ventral
to the anterior edge of the eye. ‘There it curves gradually
backward in a short bend, and then runs backward below
the eye and upward behind it, between it and the spiracular
opening, thus encircling about one half the orbit. Dorsal
to the spiracular opening the canal turns upward and
forward, and, approximately dorsal to the hind edge of the
eye, anastomoses with the hind end of the supra-orbital canal.
It then turns sharply backward, upward, and mesially, and
so continues to the point where it joins the lateral end of
the supratemporal commissure. ‘There the canal turns
almost directly backward, and continues backward as the
lateral line of the body.

Along this infra-orbital section of the main lateral canal,
that is, from its anterior end to the point where the supra-
temporal cross-commissure is given off, there were, in my
embryo, 180 tubules of varying size and length, all leading
directly and independently from the canal, and opening on
the external surface by one or more surface pores. Along
this same length of canal there were 110 sense organs, all
innervated by branches of the buccalis and oticus facialis.
The tubules in certain parts of the canal lay regularly one
between each two successive sense organs, this being markedly

98 EDWARD PHELPS ALLIS, JUN.

the arrangement in the anterior portion of the line, that is,
in the anterior portion of that part of the line that is called
by Garman the prenasal. Following this anterior portion,
and extending to the hind end of the short median section,
there were groups of two or three tubules between each two
successive organs. Posterior to this median section the
tubules were less regularly arranged in reference to the
organs, and often had no apparent regular relation whatever
to them. The tubules vary greatly in length in different
parts of the line, being particularly long along the orbito-
nasal section of Garman’s descriptions, and immediately
ventral and dorsal to the spiracle. Opposite the spiracle
they are aborted or wholly wanting. Along the orbito-nasal
section of canal the tubules branch repeatedly, and open on
the outer surface by numerous pores, as shown in fig. 7.
‘The long tubules ventral and dorsal to the spiracle, on the
contrary, branch but little, each branch usually having but
a single pore at its outer end.

The first or most distal organ of the infra-orbital canal
lies near its anterior end, the canal there connecting with
the supra-orbital canal between organs 34 and 35 of that
line. The 24th organ of the line les at the point where
the canals of opposite sides meet and anastomose in the
middle line. Beyond organ 29 the two canals separate again.
‘There are thus 25 organs anterior to the median section of
the canal, and 6 organs in that section. These 6 organs
lie on the dorsal or latero-dorsal wall of the canal, always
lateral to the median line, and opposite, or nearly opposite,
an organ belonging to the line of the opposite side of the head.
Between each successive pair of these organs there was a
single tubule on each side, that is, a pair of tubules, excepting
only between organs 26 and 27. ‘The point between these
two latter organs is, morphologically, the middle point of the
median section of the canal, and there were here two tubules
on each side and a single median tubule. ‘This was the only
median tubule found in the entire lateral system, but there
was, perhaps, such a tubule at the middle point of the supra-

MUSTELUS LAVIS. 99

temporal cross-commissure—a point that could not be defi-
nitely determined, the sections there being slightly broken.

From the hind end of the median section of the canal to the
point where the canal joins the anterior end of the supra-
orbital canal there are 16 sense organs, the two canals
anastomosing immediately beyond organ 45 infra-orbital.
The organs along the entire line up to this point, that is,
organs | to 45, are all innervated by branches of a single
large branch of the buccalis facialis, but organs 1 to 265
form a sub-group somewhat separate, in their manner of
innervation, from organs 26 to 45.

Posterior to the point where the canal anastomoses with
the distal end of the supra-orbital canal, and back to the point
where the hyomandibular canal is given off, there are nine
sense organs, 46 to 54, and their innervation by two separate
branches of the buccalis indicates that they form a separate
sub-group, or two such groups, of infra-orbital organs. This
section of canal and the one that contains organs 30 to 45,
together form the nasal canal of Garman’s descriptions, that
section of canal thus not being a morphological unit. Organs
55 to 69 lie in the two arms of the suborbital bend of the canal,
that is, in that part of the canal that lies between the point
where the hyomandibular canal is given off, and a point near
the anterior end of the ventral edge of the eye. ‘This section
of canal thus includes the orbito-nasal of Garman and a part
of his sub-orbital, not corresponding exactly to any of his
sections. ‘hese organs, in their innervation, form a distinct
sub-group of the infra-orbital line, as do also organs 70 to
74, organs 75 to 78, organs 79 to 82, and organs 83 to 86, all
of which lie in the sub-orbital part of the canal. These 41
organs together, that is, organs 46 to 86 inclusive, form what
may be called a large sub-orbital group, in which there are
six well-marked sub-groups.

Posterior to organ 86 the remaining organs of the line, back
to the point where the supratemporal cross-commissure is
given off, are all innervated by three branches that arise
close together from a posterior prolongation of the ganglion

100 EDWARD PHELPS ALLIS, JUN.

of the nervus buccalis. This posterior prolongation of the
bucealis ganglion lies close against the side wall of the
orbital part of the skull, and extends backward a short
distance behind the hind edge of the trigemino-facial
foramen. The three nerves that arise from it run upward
and backward along the side wall of the skull, which here
belongs to the interorbital wall, and come into such intimate
relations with each other that it is impossible to determine
whether there is or is not an interchange of fibres between
them. One of the three nerves that again appear after this
intimate juxtaposition separates into two branches, both of
which pierce the overhanging cartilaginous roof of the orbit,
near its hind end, and reach its dorsal surface,the roof of the
orbit here being formed by a projecting part of the post-
orbital process. One of these two branches there innervates
organs 103 to 110 infra-orbital, those organs lying in that
section of canal that is called by Garman the occipital, and
that is included between the points where the supra-orbital
canal and the supratemporal cross-commissure anastomose
with the main infra-orbital canal. The other branch inner-
vates organs 97 to 102 infra-orbital, which organs are
postorbital, or, more properly, as will be later shown,
postfrontal in position. A second one of the three principal
nerves innervates organs 92 to 96, which are postorbital
in position; the third nerve innervating organs 87 to 91,
which are also post-orbital in position, organ 87 lying
between the third and fourth tubules ventral to the spiracle.
Several of the branches of the second nerve pierce the
overhanging cartilaginous roof of the hind end of the orbit
to reach the organs they innervate. The branches of these
two latter nerves all pass outward immediately anterior to
the dorsal end of the superior postspiracular ligament,
which will be later described, and also immediately anterior
to the levator maxille superioris muscle.

Organs 87 to 110 infra-orbital thus form a single large
group, sub-divided into four sub-groups. Along the canal,
between organs 89 to 96, there were no tubules leading to the

MUSTELUS LAVIS. 101

outer surface. There were, however, four short tubules
leading outward from the canal between certain of the organs
and ending blindly, these tubules thus being in process of
abortion. This plainly indicates that other tubules, related
to the other organs, have here wholly aborted. ‘This partial
or complete abortion of these seven tubules is most un-
questionably due to the near presence of the spiracular canal,
though why, for this reason only, they should have become
aborted instead of being retained as short tubules opening
along the anterior edge of the spiracle, is not evident.
Wright (70) describes in the so-called anterior diverti-
culum of the spiracular cleft of Mustelus what he considers,
under some reserve, as a sense organ belonging to the category
of lateral sensory organs, but said to be of hypoblastic instead
of epiblastic origin. It is said by him to be innervated by
fibres derived from the pretrematic branch of the facialis. My
observations lead me to believe that the organ here referred
to is innervated, on the contrary, by a branch of that one
of the three branches above described that innervates organs
87 to 91 infra-orbital. As the observations on which this
conclusion is based require some explanation, it will be dis-
cussed in a later section after the spiracular cleft and certain
related structures have first been described. The fact that
the innervation of this spiracular organ shows that it is
quite intimately related to, if it does not actually belong
to, the infra-orbital line should, however, be mentioned here.

Supratemporal Cross-commissure and Lateral
Canal of the Body.

The supratemporal cross-commissure arises, on either side,
from the hind end of the otic section of the main infra-orbital
canal, and curving slightly forward crosses the middle line
of the head, then curves backward, and joins the infra-orbital
canal of the opposite side. The commissure passes immediately
posterior to the pores of the endolymphatic ducts, those pores

102 EDWARD PHELPS ALLIS, JUN.

lying relatively close together near the middle line of the top
of the head. This position of these pores seems a singular one,
if each endolymphatic duct represents the persistent, primary
communication of the ear capsule with the exterior, and if
the ear capsule itself represents a section of the main infra-
orbital sensory line that has been cut out between the facialis
and glossopharyngeal sections of that line, and enclosed
exactly as the short separate sections of the lateral canals
are. The position of the pore, and its long tube, would seem
to be much more easily explained on the assumption that the ©
ear, if it be developed from an organ of the main infra-orbital
lateral line, had been developed after the manner of the
ampullee rather than after the manner of the lateral canals.

‘here were, in the half of the supratemporal commissure, in
the specimen examined in sections, eleven sense organs and
twelve primary tubules. Whether these tubules were all -
arranged one between each two successive organs could not
be definitely determined, but such was probably the case. The
mesial tubule was not median in position, and there was no
sense organ apparent between it and the mesial tubule of the
opposite side. In the dissected specimen from which the
drawings were made there were but eight tubules in the one
half of the entire commissure, the mesial one not being median
in position. The tubules all run directly backward from the
commissure.

The organs of the commissure are all innervated by
branches of a large nerve that arises from the nervus
lineze lateralis while that nerve is still traversing the canal
by which it issues, with the nervus vagus, from the cranial
cavity. This supratemporal branch of the nervus linee late-
ralis runs upward from the nervus, through a special canal in
the cartilage of the skull, and issues on the dorsal surface
of the skull posterior to the commissure and internal to the
cranial extension of the- trunk muscles. There it turns
forward, and passing internal to and beyond the commissure
reaches the anterior edge of the trunk muscles. ‘There it
turns backward superficial to those muscles, and breaking

MUSTELUS LAVIS. 103

up into several branches innervates the organs of the com-
missure. ‘The nerve is thus pushed forward, out of a direct
course, by the trunk muscles, as they push forward upon
the skull.

Posterior to the commissure the lateral canal of the body
begins, but the innervation of its sensory organs could not
be determined in either of my series of sections, the nerves
that innervate them traversing the superficial layers of the
adjacent muscles, and being lost owing to slight breaks in
the sections. A certain number of the most anterior organs
seemed to be innervated by a dorsal branch of the glosso-
pharyngeus. ‘This branch was an important one, and branches
from the lateral canal organs ran toward the point where it
was broken, and lost, as it passes upward over the dorso-
lateral corner of the skull. Another nerve, which arises
as a branch of the supratemporal branch of the nervus
line lateralis, close to its base, is also broken and lost
in the muscle-fibres here, and it might be it and not the dorsal
branch of the glossopharyngeus that innervates the anterior
organs of the lateral canal.

If certain of the anterior organs of the lateral canal are
innervated by the dorsal branch of the glossopharyngeus,
as seems probable, it is to be noted that they certainly lie
posterior to the supratemporal commissure. In Amia the
glossopharyngeal organs lie anterior to the commissure,
and Ewart assumes (18) that organs so innervated are pro-
bably found in a similar position in most elasmobranchs.
As Ewart definitely finds, in Lamargus, a section of the main
infra-orbital canal that lies immediately anterior to the com-
missure innervated by branches of the nervus linez lateralis,
it is evident that the commissure of Lemargus and that of
Mustelus do not arise from similar points of the main line.
If the commissure can thus, in principle, shift forward and
backward in its point of origin from the main canal, it is
evident that it must be used with some reserve in seeking
to establish the homologies of the related bones in the skulls
of teleosts and ganoids.

VOL. 45, PART 2.—NEW SERIES, I

104. EDWARD PHELPS ALLIS, JUN.

Supra-orbital Canal.

‘‘he supra-orbital canal begins near the lateral edge of the
ventral surface of the head, postero-lateral to the nasal
aperture, and slightly anterior to the transverse plane of
the anterior edge of the mouth cavity. At this morpho-
logically anterior, or distal, end of the canal it is in direct
communication with the infra-orbital canal in the region
opposite organs 45 to 50 of that line. As the infra-orbital
organs including and anterior to organ 45 form a group
separate and distinct, in their innervation, from the organs
including and following organ 46, it is probable that the
anastomosis of the two canals takes place between organs
45 and 46,

Starting from this point the supra-orbital canal runs for-
ward and mesially in a gentle curve, parallel to the lateral
edge of the head, and, near the most anterior point of its
course, meets and anastomoses with the anterior end of the
infra-orbital canal, the two canals meeting, at approximately
a right angle, between organs 34 and 35 supra-orbital.

Anterior to this point the supra-orbital canal runs forward,
or forward and laterally, for a very short distance, and
reaches the edge of the snout, where it turns upward and
reaches the lateral surface of the head. There it at first runs
upward, backward, and laterally, in a nearly straight line,
until it reaches a point somewhat anterior to the anterior edge
of the eye. There it turns forward mesially and upward in
a short curve, continues in that direction for a short distance,
and then turns upward and backward ina second short curve,
beyond which it runs backward, dorso-mesial to the eye, in
a line approximately parallel to the mid-dorsal line of the
head. Dorso-mesial to the hind edge of the eye it anasto-
moses by its hind end with the infra-orbital canal between
organs 102 and 103 of that line, that is, between the otic and
postfrontal groups of infra-orbital organs.

The canal at its distal end les yentro-lateral to the hind

MUSTELUS L@&VIS,. 105

end of the cartilaginous nasal capsule, about midway between
it and the ectoderm, and the transverse sections that here cut
the region where the canal anastomoses with the infra-orbital
eanal cut also the hind end of the nasal epithelium. The
canal at this point lies considerably lateral to the free, ventro-
lateral edge of the cartilage of the nasal capsule, but as it
runs forward from here it approaches this free edge of the
cartilage, and opposite the hind edge of the external nasal
aperture les but slightly lateral to it.

Opposite the nasal aperture the canal passes mesial to the
vertical plane of the free ventro-lateral edge of the cartilage
of the capsule, and there les in a depression on the lateral
portion of the internal, or dorsal, surface of what I take to be
the anterior process of Gegenbaur’s (28, p. 99) descriptions
of the “nasenfliigel” cartilage of the fish. In my specimen
this nasal-flap cartilage was wholly separate from and inde-
pendent of the cartilage of the nasal capsule. It, however,
closely approached anteriorly the ventral edge of the mesial
wall of the capsule, and posteriorly similarly approached the
ventral edge of a part of the lateral wall. ‘The anterior pro-
cess of the cartilage is strongly curved in transverse section,
the back of the curve being presented dorso-laterally, and
the cartilage being so placed that its lateral edge, which is
directed ventro-laterally, lies ventral to, or even ventro-
lateral to the ventral edge of the lateral wall of the capsule.
Near this lateral edge of the nasal-flap cartilage there is, on
its dorsal surface, a slight longitudinal ridge. ‘he portion
lateral to this ridge is slightly concave, is presented dorso-
laterally, and lodges the supra-orbital canal as it passes along
the region of the cartilage. The canal thus here has a
definite relation to the nasal-flap cartilage.

Anterior to the nasal-flap cartilage the canal lies, for a
time, ventro-lateral to the rounded, bulging, lateral surface
of the anterior end of the cartilage of the nasal capsule.
Continuing its course beyond the capsule it there has no rela-
tion to any underlying skeletal structure, until it reaches the
point where it turns upward on to the dorsal surface of the

106 EDWARD PHELPS ALLIS, JUN.

head. There it lies latero-ventral to the anterior, united ends
of the dorso-lateral and ventro-mesial rostral cartilaginous
bars of the skull.

At this point the canal turns backward, and here lies lateral
to the dorso-lateral rostral bar of cartilage, lying, however,
at such a considerable distance from it that the bar would
seem to have no direct supporting relation to the canal. The
canal, moreover, here lies latero-superficial to the bar, and
not directly superficial to it.

When the canal, in its backward course, now reaches the
transverse plane of the anterior end of the nasal capsule, it
approaches the roof (morphologically floor) of the capsule,
and here sinks to such an extent below the external surface
of the head that, when it curves forward, in front of the eye,
the dorsal arm of the bend lies at first directly dorsal to the
ventral one, at a normal distance from the surface. After
the canal has again turned backward it soon comes to lie
directly superficial to the rounded dorso-lateral surface of the
neighbouring part of the skull, and continuing backward
passes on to the dorsal surface of the projecting cartilaginous
roof of the orbit.

There were in all, in the entire length of the supra-orbital
canal, 93 sense organs. The number of tubules leading from
the canal was much larger, many of them having undergone
subdivision. In certain parts of the line they were, however,
still found, one tubule between each two successive organs.
Directly on the top of the snout, about midway between the
anterior end of the snout and the point where the canal takes
its double bend in front of the eye, there was a large and com-
plicated group of tubules. These tubules all issued from the
canal on its mesial side, and ran mesially or mesially and
backward. Ata varying but short distance from their bases
they were all connected by short communicating branches,
these branches together having somewhat the appearance of
a single curved connecting canal. Beyond these communi-
cating branches certain of the tubules branched dichoto-
mously in the plane of the ectoderm. In the plane per-

MUSTELUS LAEVIS. 107

pendicular to the ectoderm there were numerous short
tubules leading to the outer surface, and there opening by
pores. The whole group of tubules formed a large semi-
circular structure on the mesial edge of the canal. Both
anterior and posterior to it, along the entire length of the
canal, on both the dorsal and ventral surfaces of the head,
there were, excepting at the bend in front of the eye, no
tubules projecting mesially from the canal. All the other
tubules projected either laterally or directly outward toward
the external surface. This restriction of the tubules to the
lateral and dorsal aspects of the canal would seem, from
Garman’s figures, to be a marked characteristic of the canals
of the head in the rays, though what its special significance
is is not evident.

The 93 sensory organs found in the supra-orbital canal of
my embryo were all innervated by branches of the ramus
ophthalmicus superficialis, and were subdivided, by the
manner of their innervation, into six groups. ‘he first
group includes organs | to 16, organ 16 lying approximately
in the plane of the anterior end of the nasal sac. The
second group includes organs 17 to 34, organ 34 lying imme-
diately posterior to the point where the canal anastomoses
with the anterior end of the infra-orbital canal. The third
group includes organs 35 to 58, organ 58 lying near the
point where the canal turns upward in front of the eye.
The fourth group includes organs 59 to 64, all of which le
in the bend of the canal as it turns upward in front of the eye.
The organs of this group are related to those of the third
group more as a sub-group than as a separate and distinct
one. The next, or fifth group, includes organs 65 to 86,
organ 86 lying in the plane of the foramen of the superficial
ophthalmic nerve. The sixth and last group includes organs
87 to 98, the organs of this group all lying posterior to the
superficial ophthalmic foramen, and being innervated by a
single nerve which leaves the ramus ophthalmicus super-
ficialis shortly after it leaves its foramen, and runs backwards
to the organs of the group.

108 EDWARD PHELPS ALLIS, JUN.

Hyomandibular Canal.

The hyomandibular canal, using the name adopted by
Ewart (18) in his descriptions of Lemargus, begins at the
point where the infra-orbital canal bends laterally and
forward, just before it passes from the ventral to the lateral
surface of the head. It is there in direct communication
with the infra-orbital canal between organs 54 and 59, that
is, between the second and third groups of the organs of the
line. From there the canal runs almost directly backward,
lying close to and parallel to the lateral edge of the ventral
surface of the head. Posteriorly it gradually approaches
that edge, and at a certain distance beyond the transverse
plane of the spiracle reaches it, and there turns upward and
backward around it, and appears on the lateral surface of the
head. Continuing in this upward and backward direction a
relatively short distance, it reaches a point something more
than half the distance backward from the spiracle to the
first gill-slit, and there ends.

The exact number of organs in the canal could not be
determined, the organs and related nerves in the posterior
part of the canal being so shghtly developed that they could
not be recognised with certainty. The primary tubes, most
of which had undergone subdivision, would seem to indicate
that there were about forty organs in the canal. The organs
are quite unquestionably all innervated by branches of a hyo-
mandibular part or division of the mandibularis externus
facialis, but this could be definitely established for the first
twenty-four organs only, counting backward from the anas-
tomosis with the infra-orbital canal. The remaining organs
and their innervating nerves could not be traced.

Mandibular Canal.

‘The mandibular canal is relatively short, lies parallel to
the hind edge of the mouth, and extends from near the
middle line of the head backward and laterally to the level
of the outer edge of the upper labial fold. It lies, in the

MUSTELUS LAVIS. 109

greater part of its course, directly superficial to the mandi-
bular part of the adductor mandibule muscle, near, and
nearly parallel to, its anterior edge. At its mesial end the
canal extends beyond the adductor muscle, and there lies
superficial to the cartilage of the mandible. The canal has
very closely the length and direction of the anterior edge of
the mandibular part of the adductor muscle. The canal does
not approach any of the other canals, either anteriorly or
posteriorly, and the canals of opposite sides are separated at
their anterior ends by a considerable interval.

The number of organs contained in the canal could not be
determined, neither the organs nor the nerves innervating
them being sufficiently developed to be traced with certainty.
There were twenty-two tubules along the line, counting the
anterior and posterior terminal ones. As most of these
tubules seemed to be primary ones that had not undergone
subdivision, there should be about twenty-one organs in the
line.

The organs are all innervated by branches of an anterior
or mandibular division of the mandibularis externus facialis,
a large branch of the nerve innervating also the mandibular
group of ampulle.

Ampulle and Surface Sensory Organs.

There are in Mustelus four groups of ampulle on each side of
the head. These four groups correspond in general position
to the superficial ophthalmic, inner buccal, outer buccal, and
mandibular groups of HEwart’s (18) descriptions of sela-
chians, but the group that has the position of an inner
buccal group is innervated by branches of the ramus oph-
thalmicus superficialis, instead of by branches of the buccalis
facialis, as will be fully described in describing the nerves.
he group in Mustelus is accordingly not the homologue of
the inner buccal group of Ewart’s descriptions, and I shall
describe it as the deep ophthalmic group. No group corre-
sponding to Ewart’s hyoid group was found, but it may,
perhaps, be represented by a line of surface organs.

110 EDWARD PHELPS ALLIS, JUN.

The superficial ophthalmic group is a long one, lying in
the upper part of the snout, and not far from the middle line
of the head. ‘The ampulle of the group all lie mesial to the
dorso-lateral rostral bar of cartilage, the posterior ones lying
directly superficial to the nasal capsule. The ampulle are
grouped by their tubules in three somewhat separate sub-
groups. The tubules leading from the anterior sub-group
are relatively short, and radiate forward, laterally and
postero-laterally, their external openings all lying on the top
of the snout, the anterior ones mesial to the anterior portion
of the rostral part of the supra-orbital lateral canal, and the
posterior ones lying mesial to the large rostral group of
supra-orbital tubules and pores. In the specimen from
which the drawings were made there were forty-three
ampulle in this sub-group on one side of the head, and
forty-nine on the other, the number of ampulle bemg
determined by the number of related surface pores, which
alone were counted.

The tubules leading from the second sub-group of super-
ficial ophthalmic ampulle are long, and run at first almost
directly laterally, forming a broad band, which lies at first
immediately posterior to the rostral group of tubules and
pores of the supra-orbital canal. The ampullary tubules
then turn backward and laterally, and pass superficially
across the supra-orbital canal immediately anterior to the
point where that canal bends backward in front of the eye.
The ampullary tubules here lie internal to the short mesio-
anteriorly directed section of the supra-orbital canal, and
internal to the tubules that arise from that section of canal,
the appearance being that of a band of ampullary tubules
that had here first pressed the supra-orbital canal inward
out of its normal relations to the external surface of the
head, and then pulled it out of a straight course and given it
the antorbital bend which seems otherwise not easily
accounted for. The external openings of these ampullary
tubules form a group of pores the anterior ones of which lie
mesial or directly superficial to the supra-orbital canal, while

MUSTELUS LAVIS. tal

the posterior ones le lateral to that canal, extending nearly
to the anterior edge of the eye, certain of the tubules here
lying superficial to the tubules of a sub-group of the outer
buccal ampullz. There were twenty-five pores in this group
on one side of the head, and twenty-one on the other.

The relations of the tubules of this sub-group of ampullee
to the supra-orbital lateral canal is exactly that that would
necessarily arise in every place where the ampulla and its
pore lie on opposite sides of a lateral canal if the ampullary
tubules were formed by the pores travelling from the place
where the ampullary organ was first enclosed by involution,
after the manner of development of the canals and tubes of
the lateral system in Amia (2). I aceordingly turned to the
sections of my 55 mm. embryo to see if I could determine
whether these particular ampulle lay postero-lateral or
antero-mesial to the supra-orbital canal, and I found, as I
had expected, a number of the teat-like processes that here
represent the ampulle in exactly the position that the pores
of the group in question occupy in the older embryo. These
teat-like processes were directed anteriorly, as they should be,
and in my opinion unquestionably represent the ampulle of
the older embryo, though this can certainly not be positively
asserted until they have been followed through certain of
the intermediate stages. Why these particular ampulle
should have pushed forward external instead of internal to
the lateral canal is not evident.

The tubules of the posterior sub-group of superficial
ophthalmic ampullz are, in part, still longer than those in
the second sub-group. They all at first run postero-laterally
from their ampulle, the longer ones gradually turning
directly backward. Most of them have their external
openings in a line slightly mesial to that part of the supra-
orbital canal that lies directly dorso-mesial to the eye, but
some of them lie scattered between that line and the mid-
dorsal line of the head, and six of them form a curved line
which les shghtly anterior to the supratemporal cross-
commissure, and slightly anterior also to the transverse

112 EDWARD PHELPS ALLIS, JUN.

plane of the external openings of the endolymphatic ducts.
These last six pores formed, in all my specimens, one of the
most distinctly evident lines of pores on the entire head of
the fish, and this line would seem to occupy the position
ascribed by Garman (21) to the supratemporal commissure of
Chlamy doselachus—a position, however, also held by two sur-
face sense organs to be later described. ‘There were twenty-
six pores in this sub-group on one side of the head of the
specimen used for the drawings, and twenty-nine on the other.

All of the ampulle of the entire group of superficial
ophthalmic ampulle are innervated by branches of the ramus
ophthalmicus superficialis, in a manner that will be related
in describing that nerve.

The deep ophthalmic group of ampullz hes about half way
between the nasal aperture and the anterior end of the head,
there lying im the region included between the prenasal
section of the infra-orbital lateral canal, and the rostral and
subrostral sections of the supra-orbital canal. In transverse
sections that pass through the posterior portion of this group
of ampulle, they are seen to form a curved line, extending
from the internal surface of the prenasal section of the
infra-orbital canal dorsally, and laterally toward the rostral
section of the supra-orbital canal. ‘This curved line lies
slightly ventro-lateral to a line connecting the ventro-mesial
and dorso-lateral rostral bars of cartilage, the most lateral
ampulle lying lateral to the dorso-lateral bar. Anteriorly
the ampulle form, in sections, an irregular group that lies in
nearly these same relations to the lateral canals and rostral
bars. Posteriorly the anterior end of the nasal capsule
presses into the group, and separates it, in sections, into two
parts, one part lying ventro-mesial to the nasal capsule, and
the other lying on its dorsal surface. Most of the tubules of
the group radiate downward from the ampulle, running in
every direction toward the ventral surface of the snout ; but
some of them first extend downward and laterally, then turn
upward and backward as they approach the lateral edge of the
snout, and thus reach its dorso-lateral surface. ‘hese latter

MUSTELUS LAVIS. 113

tubules form a sub-group somewhat distinct from the
others.

Those pores of this group that lie on the ventral surface of
the snout all lie anterior or mesial to the nasal aperture, and
there is a line of them on each side of the subrostral section
of the supra-orbital canal, and on each side of the prenasal
section of the infra-orbital canal, certain of the tubules here
crossing the lines of each of the two lateral canals, internal
to them, to reach their opposite sides. ‘The pores that are
thus arranged in liue on either side of the two canals are, in
certain places, placed markedly one between each two
successive primary tubules of the related canal, canal tubules
and ampullary pores thus alternating. Certain of the
ampullary tubules extend backward into the tissues that
cover the cartilages that lie in the ventral wall of the nasal
sac, and in sections, but not in the dissections, certain of
them even had their openings in the very edge of the nasal
aperture. ‘The tubules that reach the dorso-lateral surface
of the snout turn backward, and the related pores form a
band that extends backward to the anterior edge of that part
of the infra-orbital canal that bends forward below the eye.
There were in this entire group of ampulle, on the one side
of the head on which they were counted, 194 pores. ‘The
ampulle are all innervated by branches of the ramus
ophthalmicus superficialis, as will be fully described in
describing that nerve.

The buccal group of ampulle lies ventral to the anterior
edge of the eye, in the region internal to that section of the
infra-orbital canal that lies between the suborbital bend in
the canal and the point where the hyomandibular canal is
given off. The tubules of the group may be separated into
five sub-groups. ‘The tubules of one of these five groups run
at first forward and upward, then curve gradually forward
mesially and downward, and have their external openings on
the dorso-lateral surface of the snout, lateral to the rostral
part of the supra-orbital lateral canal, between that canal
and the pores of that sub-group of the deep ophthalmic

114 EDWAKD PHELPS ALLIS, JUN.

ampulle that form a band along the lateral edge of the dorsal
surface of the snout. There were forty-three pores in the
sub-group.

The tubules of a second sub-group run upward, forward,
and mesially along the anterior edge of the eye, and there
spread, some continuing their earlier course, while others
turn backward above the eye. The pores of the sub-group
all lie between the antero-dorsal edge of the eye and that
part of the supra-orbital canal that lies posterior to the point
where the canal bends forward in front of the eye. ‘There
were thirty-seven pores in the group.

The tubules of a third sub-group all open on the ventral
surface of the head, the tubules running in large part
mesially, but in part almost directly forward, and in part
almost directly backward. The few tubules that run directly
forward have their external openings along either side of the
extreme distal end of the supra-orbital canal, that is, along
both sides of the posterior end of what Garman calls the sab-
rostral canal. Those tubules that run posteriorly open in a
group of pores that lie mesial to the anterior end of the
hyomandibular canal. Those tubules that run mesially spread,
and their external openings form a scattered group of pores
the larger part of which le between the nasal section of the
infra-orbital canal and the front edge of the mouth, a smaller
part lying between the same section of canal and the nasal
aperture. Certain of these pores form a line along each edge
of the nasal section of the infra-orbital canal, while certain
others form a marked line along the very edge of the upper
lip. There were 107 pores in this sub-group.

The tubules of a fourth sub-group run backward and,
excepting a few scattered ones, form a wide band of tubules
which extends backward ventral to the suborbital part of
the infra-orbital canal, occupying the entire lateral surface of
the head ventral to that canal. The external openings of the
few scattered tubules, and those also of certain of the longer
ones, lie on the lateral surface of the head posterior to that
section of the infra-orbital canal that hes between the sub-

MUSTELUS LAEVIS. 115

orbital bend in the canal and the point where the hyoman-
dibular canal is given off, the pores being most numerous
near the canal, and diminishing irregularly in number from
there backward. The long tubules run at first directly back-
ward, all having a nearly parallel course, the longest ones
extending beyond the level of the hind edge of the eye, and
there, toward their hind ends, turning upward and backward.
Their external openings form a regular, curved, and well-
marked line which extends from the hind edge of the infra-
orbital canal, immediately below the spiracle, at first back-
ward and but slightly downward, then, having nearly reached
the level of the hind end of the hyomandibular canal, turns
downward and slightly forward, anterior to the hind end of
the latter canal, and ends near the ventro-lateral edge of the
head. There were forty-six pores in the sub-group.

The tubules of the fifth sub-group form a continuation, on
the ventral surface of the head, of the broad band formed
by the fourth sub-group. Their external openings form a
long curved line, which begins on a level with, and consider-
ably lateral to, the hind edge of the gape of the mouth, and
from there runs backward and laterally, on the ventral sur-
face of the head, in a line lying mesial to and somewhat
parallel to the hyomandibular canal. There were twenty-
seven pores in the sub-group,

The sensory organs of the entire group are all innervated
by two branches of the buccalis facialis in a manner that will
be fully described in describing that nerve. The innervation
of these ampullee by two branches of the buccalis may perhaps
indicate that they represent the united inner and outer buccal
groups of Hwart’s descriptions of other selachians.

The mandibular group of ampulle is a small one lying
slightly posterior to the lateral third of the mouth opening,
between it and the mandibular canal. In my 12-2 em. speci-
men these ampulle were still in an undeveloped condition.
They had already sunk beneath the ectoderm and were united
in a close group, but none of them had as yet acquired the
pocketed form characteristic of the ampulle in all the other

116 EDWARD PHELPS ALLIS, JUN.

groups. The tubules of the group radiate from the ampulle,
all running forward, forward and mesially, and backward
and laterally, in the general direction of the hind edge of the
mouth opening. The organs of the ampull are all inner-
vated by a single short branch of the ramus mandibularis
externus facialis.

The so-called pit organs, or sensory follicles, of other
descriptions of Mustelus I did not seriously attempt to fully
or carefully trace. Certain of them that are very evident
are shown in the drawings, and I doubt there being many
others. A very distinct line of them extends in a curved
line, as shown in fig. 3, across the entire ventral surface of
the head somewhat posterior to the mouth. At each lateral
end this line turns upward on to the lateral surface of the
head, passes posterior to all the ampullary pores of the region,
and posterior also to the hind end of the hyomandibular
lateral canal; it then turns forward dorsal to and approxi-
mately parallel to the ampullary pores of the region, and
reaches the hind edge of the spiracle, where it ends. Its
innervation [ could not determine, nor could I satisfy myself
as to whether the line was a line of sensory pit organs, or
simply a line of undeveloped ampulle similar to those found
in my younger embryos. If the organs represent undeveloped
ampullz, it would seem as if they, or at least a part of them,
must represent the hyoid group of ampulle of other fishes,
that group not otherwise being represented in Mustelus. Tf,
on the contrary, the organs are pit organs, of a lateral sen-
sory type, the ventral part of this line of organs recalls
markedly in its position the gular line of pit organs of Amia,
and may represent that line. In Chlamydoselachus the
entire line would seem quite certainly represented in the
combined gular and spiracular lines of Garman’s (21) descrip-
tions, these lines being said by him to be open grooves. If
they are so represented it would seem almost certain that
the angular canal of Chlamydoselachus, which is clearly the
hyomandibular canal of Mustelus, must be the horizontal
cheek line of pit organs of Amia, and that the other lines on

MUSTELUS LEVIS. 117

the cheek of Chlamydoselachus, that is, the spiracular, cular,
jugal,and oral, represent the preoperculo-mandibular canal and
the vertical cheek line, mandibular line, and gular line of pit
organs of Amia. The accord is much too evident not to warrant
the supposition. The organs of the line of Mustelus may,
accordingly, represent a condition of surface sense organs on
the border line between terminal buds and ampullary organs
on the one side, and pit organs and canal organs on the other.

Four other pit organs were always found in all my larger
specimens, two on each side of the top of the head, slightly
anterior to, and on either side of, the endolymphatic pore.
These organs, on each side, lie in the line produced of the
five lateral ones of the curved supratemporal line of six
ampullary pores, and they may, perhaps, represent two
ampulle of that line that have retained their embryonic
place and condition. The fact that the mesial one of the
six ampullary pores does not lie in this same line seems,
however, to indicate that we have here to do with a different
class or group of organs, and that the two pit organs more
probably represent one of the head lines of pit organs of
Amia. In Chlamydoselachus, according to Garman, the
supratemporal cross-commissure of the lateral canals has
the position relative to the endolymphatic pores of these
four surface organs, and not that of the cross-commissure
of Mustelus and other selachians. This would be easily
accounted for on the assumption that the two organs in
Mustelus are lateral sensory ones, and became, in Chlamy-
doselachus, enclosed in a canal.

Other pit organs are found in Mustelus, irregularly ar-
ranged on the body of the fish posterior to the supra-
temporal commissure and dorsal to the main lateral line.
One line of these organs lies directly superficial to the lateral
canal, a position that seems to preclude its being in any way
directly related to the organs of that canal. Certain of the
organs undoubtedly form the line of ‘pit organs” that
Ewart (18, p. 81) describes in Mustelus, and which he says
are innervated by a branch of the nervus linez lateralis vagi.

118 EDWARD PHELPS ALLIS, JUN.

Review and Comparison.

The lateral sensory canals of selachians seem, at first sight,
to have a distribution that admits of but little detailed
comparison with the canals of the bony fishes. This differ-
ence in detail is, however, largely in appearance only, and
not real.

Imagine the snout of Mustelus pressed backward into and
on to the dorso-anterior surface of the head, and the mouth
pulled forward until it comes to lie at the anterior end of
the snout. What is actually a part of the ventral surface
of the snout would then become a part of its dorsal surface,
and an arrangement of the lateral canals would arise such
as is shown in lateral view in the adjoining cut. This cut

Hie. I:

Side view of an assumed projection of the lateral canals of Mustelus
on to the head of Amia.

represents, in fact, the projection, so to speak, of the canals
of Mustelus on to the head of Amia. <A front view of such
a head would be quite accurately represented in Garman’s
front view of Mustelus canis (21, pl. viii).

In this imaginary head the infra-orbital canal begins on
the top of the snout, posterior to the single nasal aperture,
and is there in direct communication with the supra-orbita]

MUSTELUS LAEVIS. 119

canal, at a bend in that canal. It then runs forward and
mesially, and, anastomosing with the corresponding canal
of the opposite side, forms a median longitudinal section
of canal which lies directly between the nasal apertures on
the dorsal surface of the anterior end of the snout. ‘The
canal then turns laterally and backward, ventral to. and
around the nasal aperture, and, posterior to the aperture,
anastomoses with the distal end of the supra-orbital canal.
It then makes a double turn, first upward and then back-
ward, giving off the hyomandibular canal at the first bend,
and, having encircled the inferior and posterior edges of the
eye, reaches and anastomoses with the hind end of the supra-
orbital canal dorso-posterior to the eye. It then turns back-
ward, and reaches and anastomoses with the lateral end of
the supratemporal cross-commissure. Posterior to this
point it becomes the lateral canal of the body.

The supra-orbital canal begins between the eye and the
nasal aperture, there being in direct communication with the
infra-orbital canal, at a bend in that canal and at a point
that lies between two distinct groups of the organs of the
line. The canal first runs forward and mesially on the top
of the snout, but, dorso-postero-mesial to the nasal aperture,
turns sharply backward, anastomosing at or near this bend
with the anterior end of the infra-orbital canal. It then
continues backward dorsal to the eye, until it reaches and
anastomoses with the infra-orbital canal between its otic and
postfrontal sections.

If the canals of this imaginary or projected head of
Mustelus be compared with the canals of Amia, it will be
seen that that section of the infra-orbital canal of Mustelus
that contains the first group of organs of that line, Nos.
1 to 45, and that les between the two anastomoses of the
infra-orbital canal with the supra-orbital one, corresponds,
in many respects, with that part of the infra-orbital canal
of Amia that encloses the first four infra-orbital organs of
that fish (2, p. 514). Those four infra-orbital organs form,
im Amia, a distinct and separate group, and lie in a part

VOL. 45, PART 2.—NEW SERIES, K

120 EDWARD PHELPS ALLIS, JUN.

of the canal that I considered as an anterior section or
commissure of the line. In Amia there are two nasal
apertures on each side of the head. Between these two
apertures the backwardly-directed distal end of the supra-
orbital canal is directed toward and approaches somewhat
the hind end of the so-called anterior, commissural section
of the infra-orbital canal. If two separate nasal apertures
had been developed in Mustelus, and the posterior one had
travelled backward toward the anterior edge of the eye,
it would necessarily, since the nose develops earlier than the
lateral canals, have passed backward between the growing
distal end of the supra-orbital canal and the point where
that canal is later to anastomose with the infra-orbital canal.
This would thus have here given closely the arrangement
found in Amia. It would also probably closely give the
arrangement found in Polypterus (6), but I have not yet
worked out the details of the innervation in that fish. In
Batrachus tau a very similar arrangement is also found
(10), but in this latter fish the anterior organ of the supra-
orbital line is not enclosed in, nor does it lie in the line
of, the antero-laterally directed distal end of the supra-orbital
canal.

In the projection of the canals of Mustelus the anterior end
of the infra-orbital canal anastomoses with the supra-orbital
canal at or near the point where that canal bends sharply
backward, postero-mesial to the nasal aperture. It then
joins and anastomoses with its fellow of the opposite side, and
here forms a short, median, longitudinal section, which lies on
the antero-dorsal surface of the snout, between the nasal
apertures of opposite sides of the head. In both Amia and
Polypterus the arrangement is here quite different, and there
are also differences in the arrangements presented by these two
fishes. They can, however, all be easily derived one from
the other, and in Conger conger I have lately found exactly
the arrangement here shown in the projection of Mustelus.
There is thus certainly a full homology in this part of the
lateral system of these several fishes. The same is also true

MUSTELUS LAVIS. 171

of the canals in Scomber and Gadus, but still other differences
here exist. In Scomber, that part of the infra-orbital line
that corresponds to the anterior, commissural section here in
question, has become fused (4) with the anterior end of the
supra-orbital canal, and forms a direct anterior prolongation of
that canal. In Gadus (12) the same section of canal forms
a short and direct anterior prolongation of the infra-orbital
canal, containing, as in Scomber, but a single sensory organ.
In both Scomber and Gadus the single nasal aperture of
the one, and the two apertures close together of the other, lie
relatively close to the eye, posterior to the anterior ends
of both the supra-orbital and antorbital canals, and there is
no indication of the loop, wholly or partly encircling the
nostrils, that is found in Mustelus, Amia, Polypterus, and
Batrachus.

In other teleosts and ganoids the descriptions of this part
of the lateral canals are either not clear or not sufficiently
detailed or complete to warrant an attempt to establish
their homologies. This applies even to Herrick’s (82) care-
ful work on Menidia, in which I am unable to trace the
relations of the several surface sensory organs he describes
in this region, either to each other or to the lateral canals, or
even to decide whether they are lateral sensory pit organs, or
terminal buds.

In Lemargus (18) the only points in which the canals
here differ from those of Mustelus are that the anterior end
of the infra-orbital canal does not anastomose with the supra-
orbital, and that the distal, turned-back end of the supra-
orbital canal anastomoses with the infra-orbital canal at a
point that seems to lie morphologically posterior to the one
where the anastomosis takes place in Mustelus.

In Raia (19) the two anterior sections of Ewart’s descrip-
tions of the infra-orbital canal, sections 10° and 107, each form,
at their anterior end, a transverse anastomosis with the cor-
responding canal of the opposite side of the head. Between
these two anastomoses the canals of opposite sides lie parallel
to each other, and not far apart. These canals thus differ

122 EDWARD PHELPS ALLIS, JUN.

from the arrangement found in Lemargus in that there is
an anterior anastomosis, indicated but not found in Lemar-
gus, and that in this last fish the posterior, internasal anasto-
mosis is formed by the juxtaposition and fusion of longitu-
dinal sections of the canals of opposite sides, as in Mustelus,
and not by the fusion, end to end, of transverse sections of
canals. Another difference between the two fishes is, that,
in Lemargus, the anastomosis of the distal end of the supra-
orbital canal with the infra-orbital canal, lies at a point on this
latter canal, that is, morphologically, anterior to the one at
which the hyomandibular canal is given off, while in Raia
it lies morphologically posterior to that point. Mustelus here
agrees with Lemargus.

In Chimera (11) the anterior, antorbital section of the
infra-orbital canal is probably the one that is innervated by
the anterior division of the outer buccal nerve of Cole’s de-
scriptions.

In Mustelus, posterior to the anterior section of the infra-
orbital canal above discussed, there are three principal groups
of infra-orbital organs, each of which is separated into two or
more, more or less distinct sub-groups. One of these three
principal groups is suborbital in position, and includes organs
46 to 78. The other two together are postorbital and otic in
position, and include organs 79 to 110.

Organs 46 to 78, which form the first of these three groups,
are innervated by the first three branches that arise from the
buccalis beyond its ganglion. ‘They arise consecutively from
the nerve, beginning at some little distance beyond its gan-
glion, and the third and more anterior branch is so large that
it may be called a division of the nerve rather than a branch
of it. The two posterior branches innervate, the one organs
70 to 74, and the other organs 75 to 78. The anterior branch,
or division, of the nerve breaks up into several smaller
branches and one large one, the smaller branches being
destined to supply organs 46 to 69, and the large one to
supply the buccal group of ampulle.

Organs 79 to 86 form the second one of the three groups,

MUSTELUS LAVIS. 1233

and they are innervated by branches of two nerves that arise
separately and independently, but close together, from the
ganglion of the buccalis, close to the base of the nerve itself.
The most anterior branch innervates organs 79 to 82; the
other innervating organs 83 to 86.

Organs 87 to 110 form the third group, and they are
separated, by their innervation, into four sub-groups, two of
which are postorbital in position, while the other two occupy
positions corresponding, in their topographical relations, the
one to the postfrontal part of the infra-orbital canal of Amia,
and the other to the otic and glossopharyngeal parts together,
of the main infra-orbital canal of the same fish. The post-
orbital organs of Mustelus include organs 87 to 96 ; the post-
frontal ones, organs 97 to 102; and the otic ones organs 103
to 110. The organs of the entire group are innervated by
three nerves that arise close together from a posterior pro-
longation of the buccalis ganglion, and that have already
been described.. That branch of one of these three nerves
that innervates organs 103 to 110 is certainly the homologue
of the nerve described by me in Amia as the ramus oticus
facialis. What the exact homologues are, in Amia, of the
other nerves of Mustelus that are here concerned is not easy
to tell. The nerve that, in Mustelus, innervates organs 97 to
102 would seem to be the exact homologue of the nerve that,
in Amia, innervates the single postfrontal organ, No. 14 inf.,
of that fish. The nerve that innervates this latter organ in
Amia always arises close to the base of the ramus oticus, if
not from the base itself of that nerve, and in one specimen of
Amia I found it arising as a branch of the oticus after that
nerve had issued from its cranial canal on to the roof of the
skull (2, p. 515). But it is a branch of the oticus, in Amia,
that innervates the spiracular organ, while, in Mustelus, it is
a branch of the nerve that innervates organs 8/7 to 91. It
may then be that the three nerves of Mustelus that together
innervate organs 87 to 110 and the spiracular organ are the
homologue of the otic nerve of Amia plus that branch of the
buccalis that innervates the postfrontal organ No, 14 infra-

124 EDWARD PHELPS ALLIS, JUN.

orbital. In that case the two nerves of Mustelus that inner-
vate organs 79 to 86 would be the homologues of the nerves
that in Amia innervate organs 11 to 13. It seems to me,
however, much more probable that these latter nerves in
Amia have their homologues, in Mustelus, in the nerves that
innervate organs 87 to 96, the nerve that innervates the
spiracular organ in the two fishes not having exactly the
same course and possibly not being exactly the same nerve.
However this may be, it is evident that either organs 87 to
110, or organs 79 to 110, must represent organs 11 to 16 of
Amia, and that the nerves that together innervate the organs
in the two fishes must be homologous. These several organs
in Mustelus, Nos. 87 to 110, or 79 to 110, as the case may be,
and organs 11 to 16 in Amia, thus form, in each fish, a group
of organs that is quite distinct and separate, in its innervation,
from the remaining more anterior infra-orbital organs. In
Amia these organs also develop somewhat separately and
independently from the rest of the line (2). As this same
grouping of these organs is found in Scomber also (7), it
would seem as if it might be considered as a general rule that
there are, posterior to the anterior, antorbital group of organs,
two separate and distinct groups of buccal organs in the main
infra-orbital line. The anterior one of these two groups, plus
the antorbital group of organs, must then correspond to the
organs innervated by the outer buccal nerve of Cole’s descrip-
tions of Chimera, and the other group to the organs inner-
vated by his inner buccal nerve. They must also correspord
to the two groups innervated by the inner and outer buccal
nerves of Cole’s descriptions of Gadus (12), and to those
innervated by the buccalis and oticus facialis of Herrick’s
(82) descriptions of Menidia. But, in Cole’s application of
these names, inner and outer buccal, to the nerves in Gadus,
adopted by Herrick (83) in his descriptions of the same fish,
it is the outer buccal that innervates the posterior group of
organs, and the inner one that innervates the anterior one,
the reverse of the conditions described by Cole in Chimera.
That this is due to an error in homologising the nerves in

MUSTELUS LAVIS. 125

Gadus with those in Chimera is evident, and, if perpetuated,
it might easily lead to some confusion.

It is, moreover, to be remarked that organs 7 and 8 in
Gadus, although innervated by terminal branches of the
nerve that innervates organs 9 and 10, are separated from
those organs by a considerable interval, and that in this,
and also in their general position, they seem to belong
more to the anterior group of organs than to the posterior
one. :

In Batrachus tau Clapp (10) shows a line of surface
organs lying along the ventral and posterior margins of the
eye, the line being continued backward by two similar sur-
face organs that lie in the otic region. This line, or group,
of surface organs is unquestionably the homologue of the
organs innervated by the so-called inner buccal and otic
nerves together of Chimeera, that is, to the postorbital, post-
frontal, and otic organs of Amia. ‘The so-called maxillary
line of organs in Batrachus would then be the homologue of
the suborbital organs of Amia, that is, of organs 5 to 10; and
the antorbital line of surface organs in Batrachus would be
the homologue of the antorbital commissure of Amia. It is
evident that if the anterior ends of the postorbital and
maxillary (suborbital) sensory lines of Batrachus were to be
continued forward until they met the supra-orbital line, an
arrangement almost exactly resembling that found in
Chimera would arise.

Regarding the hyomandibular line of selachians it is
impossible to judge whether this line is simply the pre-
opercular line of Amia, changed slightly in position at its
anterior end, or a line formed by some combination of the
pre-opercular canal line and the horizontal cheek pit line of
Amia. Amia, in which there is such an abrupt and marked
difference of level between the anterior end of the pre-
opercular canal and the hind end of the mandibular one,
might be considered as representing a stage intermediate
between the arrangements found in Gadus or Polypterus (6),
and the one found in Mustelus, the hyomandibular line in the

126 EDWARD PHELPS ALLIS, JUN.

latter fish being, in that case, the exact homologue of the pre-
opercular canal line of Amia.

One feature of this hyomandibular canal of selachians is
worthy of notice, though just what its morphological signifi-
cance may be I cannot tell. The so-called proximal end of
the canal, that is, the end at which it anastomoses with the
infra-orbital canal, leaves, or joins, that canal anterior to the
spiracle, while the nerve that innervates it descends posterior
to the spiracular canal. This relation of the canal to the
spiracle is much more marked in Polyodon (18, figs. 3 and
12), in which fish the branches destined to innervate the
organs in the dorsal part of the canal, if there be any, must,
to reach the organs, curve forward and upward around the
ventral surface of the spiracular cleft. This relation is a
singular one, and seems to call for some special explanation,
which may, perhaps, be found in the assumption that a part
of the lme is represented in the horizontal cheek line of
pit organs of Amia. In Polypterus (6) the dorsal end of
the pre-opercular canal lies at the hind end of the spiracle,
in what would seem to be its natural relation to that
opening.

In Chlamydoselachus (21) there is a very close general
agreement of nearly all the canal lines with the lines in the
projection of Mustelus. The so-called occipital canal of
Garman’s descriptions of the former fish is the otic canal of
my descriptions of the latter. The aural of the former fish
is probably represented in Mustelus, as already stated, by the
two supratemporal pit organs. The orbital, suborbital, and
orbito-nasal canals of Chlamydoselachus are the postorbital
and suborbital parts of the infra-orbital canal of Mustelus.
The nasal and prenasal of Chlamydoselachus, including
between them the region of the non-existent median, are tlie
antorbital section of the infra-orbital of Mustelus. The sub-
rostral, rostral, and cranial of Chlamydoselachus are the
supra-orbital of Mustelus. The angular of Chlamydoselachus
is the hyomandibular of Mustelus, and the jugal, oral,
spiracular, and gular of Chlamydoselachus are, together,

MUSTELUS LAVIS. aT |

represented in Mustelus by the mandibular canal and the long
cheek line of surface organs.

From the lateral lines of Chlamydoselachus, thus represented
in part by canals and in part by open grooves, Garman derives
the canals of the Batoidei.

We thus probably have, on the head of selachians, all the
lateral canal and pit lines of Amia excepting the middle and
posterior head lines of pit organs of my descriptions of the
latter fish, and no others. Both the supra-orbital and the
facialis part of the main infra-orbital canal lines of such
selachians as Mustelus come to a legitimate termination at
both ends by what seems to be an end anastomosis with
another canal. One should not, accordingly, expect to find
a line of unenclosed canal organs, such as certain of the pit
organs of Amia and teleosts certainly are, continuing the line
of the canal at either end. There should accordingly be, in
Mustelus, no pit organs on the top of the snout, and probably,
though not certainly, no anterior head line of pit organs.
The middle and posterior dorsal head lines of Amia then
alone remain to be accounted for of all the pit lines of the
head of the latter fish, the mandibular, gular, and cheek
lines of Amia being probably all represented in Mustelus, in
the manner just above set forth.

Is it, then, at all probable that the numerous ampullew of
Mustelus can possibly be represented in Amia by the few pit
lines on the head of that fish that are not almost definitely and
positively otherwise accounted for? And is not the primary
distribution of the ampulla, as indicated by their pores, wholly
opposed to any such assumption? The general distribution
of these pores, and their particular distribution in certain
places, closely resembles that of the terminal buds on the
head of larve of Amia, and it is noteworthy that, in those
places in which certain of the pores of a single group of am-
pulle are arranged in line along each side of a lateral canal,
the branches of the nerves that innervate the organs of those
ampulle must straddle the lateral canal, and hence its related
nerve, exactly as the trigeminal nerves that imnervate the

128 EDWARD PHELPS ALLIS, JUN.

terminal buds in Amia, straddle certain of the lateral canals
and their associated nerves (8).

The ampulle of Mustelus thus seem, in everything except-
ing their special form and apparent innervation by the
facialis, to correspond much more closely to the terminal buds
of Amia than to the pit organs of that fish. As to this
apparent innervation of the ampullary organs by branches of
the lateral canal nerves, it is quite possible that the fibres
destined to the ampullary organs have a different central
origin from that of the lateral fibres, as I shall attempt to
show, after describing the nerves of Mustelus, by a com-
parison of the ophthalmic nerves of fishes. The fibres that
innervate the ampullary organs would then represent a stage
intermediate between the terminal bud, or communis, fibres
of Amia, and the lateral canal ones. That such intermediate
stages should exist is wholly natural, if all the several forms
ot special sensory cutaneous organs are derived from a
single primitive form, represented most closely by the
terminal buds of ganoids and teleosts, as Wiedersheim says
(64).

In teleosts so-called pit organs, irregularly distributed
over the head, are frequently described. These organs, in
my opinion, must often be the homologues of the terminal
buds of Amia, and not of the pit organs of that fish. In this
I differ from Cole and Herrick, but Herrick, it is especially
to be noted, says (82, p. 37) that the pit organs of Menidia
are not situated in pits, being “strictly naked papille
projecting above the surface of the skin.” He further says,
in the same work (p. 86), that he inclines to regard the
so-called accessory lateral line organs on the trunk of
Menidia, “like the buds on the top of the head, as belonging
to the communis system.”

Nervus Oculomotorius.

The nervus oculomotorius of my 12°2 cm. Mustelus, as it
issues from its foramen, lies directly mesial to the rectus

MUSTELUS LAEVIS. 129

internus, and somewhat anterior to the surface of origin of
that muscle, as Tiesing (62) shows it in his figures. ‘Three
branches arising close together, if not as parts of a single
branch, are immediately sent to the ventral portion of the
rectus internus, and then a fourth branch to the dorsal
portion of the same muscle. A fifth and wholly independent
branch is then sent upward and backward to the rectus
superior. The main nerve, as it gives off these five branches,
is here running backward between the skull and the rectus
internus, the muscle being so pressed against the nerve that
its mesial surface is grooved to receive it. The muscle runs
forward immediately dorsal to the eye-stalk. The nerve
issues from the skull slightly anterior to the eye-stalk, and
dorsal to it, and runs backward above it.

The eye-stalk at this age is of cartilage, but this cartilage
is not directly continuous with the cartilage of the skull, an
intervening space of procartilaginous tissue separating them.
This procartilaginous part of the stalk curves downward, and
is continuous with the ventral edge of a defect or perforation
in the cartilaginous part of the side wall of the skull, the
defect of the cartilage being, however, entirely closed by the
perichondrial membranes. The external perichondrial mem-
brane extends outward on to the base of the stalk. ‘he
defect ‘lies almost directly anterior to, and not far from, the
external opening of the canalis transversus. What the
significance of the defect is I could not determine, but it
might receive an explanation under the assumption that the
eye-stalk is a remnant of a visceral arch, as Gegenbaur has
suggested.

As the oculomotorius gives off the five branches above
described it passes backward beyond the eye-stalk, and
beyond that structure the remaining portion of the nerve,
which is now its ventral division only, turns downward
between what look like two independent heads of the rectus
internus. ‘The ventral and much larger head of this muscle
arises from the side wall of the skull posterior to this point,
the surface of origin of the muscle lying immediately ventral

130 EDWARD PHELPS ALLIS, JUN.

to the profundus foramen, and extending backward toward
the dorsal edge of the trigemino-facial foramen. The surface
of origin of the muscle lies immediately dorsal to, and in
contact with, the surface of insertion of the ventro-anterior
head of the rectus externus, and immediately anterior to the
surface of insertion of the dorso-posterior head of the same
muscle. The dorsal and smaller head of the internus
separates from the ventral head shortly posterior to the poimt
where the latter receives its innervating branches from the
oculomotorius. ‘l'urning dorsally and backward it comes into
contact with the ventral edge and inner surface of the rectus
superior, and there, gradually diminishing in size, disappears
without ever reaching the cartilage of the skull. |

My specimen thus See a coded of the rectus internus
different from any of those described either by Schwalbe
(57), Tiesing (62), or Corning (14), and it will be again
referred to in describing the eye muscles. The vented
division of the oculomotorius passes outward between these
two divisions of the internus muscle, lying dorsal to the one
and ventral to the other. It lies wholly in front of, and hence
ventro-anterior to, the rectus superior. oe

The oculomotorius is accompanied in its course between
these two heads of the internus muscle by the’ ramus
ophthalmicus profundus, but the latter nerve here runs in
the opposite direction to the oculomotorius, that is, forward
instead of backward from its foramen. Both nerves lie
anterior to the rectus superior, and, as they cross, are closely
pressed against each other, as Schwalbe has said, but with-
out any observable interchange of fibres.

After the oculomotorius has passed the profundus, crossing
ventral to it, it continues downward and backward, reaches
the dorsal surface of the rectus inferior, and then soon
passes downward and forward around the hind edge of that
muscle. There it continues forward for a time,’ closely
applied to the ventral surface of the inferior muscle, and here 7
sends a large branch to that ‘muscle.

Where. the oculomotorius turns downward and forward

MUSTELUS LAVIS. 131

around the hind edge of the rectus inferior it holds that
muscle tightly in a short sharp bend. From the nerve, at
this bend, a small and delicate branch is sent outward to the
eyeball, near the point where’ the ophthalmic branch of the
anterior carotid artery pierces it, and was there lost. This
nerve would seem to be the ciliaris brevis, which must accord-
ingly arise from the ciliary ganglion, and from there run
backward along and closely against the oculomotorius to the
point where it has its apparent origin from the latter nerve.

Beyond the branch to the rectus inferior the remaining
portion of the oculomotorius, now simply the branch destined
to innervate the obliquus inferior, comes into intimate
relations both with the ciliary ganglion and the ophthalmic
branch of the anterior carotid artery. The nerve here passes
ventro-lateral to and closely against the artery. The ciliary
ganglion is, in transverse sections that cut through about
the middle of its length, a crescent-shaped mass that projects
downward from the oculomotorius, the hollow of the crescent
directed laterally, and the ventral horn of the crescent
curving around the mesial edge of the truncus buccalis-
maxillo-mandibularis. The anterior end of the ganglion lies
mesial to the ophthalmic artery at the point where that artery
curves downward to join the anterior carotid. The oculomo-
torius here lies lateral to the artery, the artery thus here
separating the nerve and ganglion. ‘The middle and posterior
parts of the ganglion lie mesio-postero-ventral to the artery,
and the ganglion is here intimately connected with the
oculomotorius. -The ganglion lies in the peri-orbital sinus,
which the artery and nerve both traverse.

What the character of the cells of the ciliary ganglion is I
could not determine, the several tissues here passing one into
the other without distinctly apparent limits. Certain of the
ganglion cells are, however, small, while others are large and
resemble the ganglion cells of the spino-cranial ganglia. There
are thus probably both spinal and sympathetic elements in
the ganglion.

Near the anterior end of the ganglion two delicate branches

132 EDWARD PHELPS ALLIS, JUN.

are given off, and, running forward and upward, join the
ramus ophthalmicus profundus anterior to the point where a
first large branch arises from that nerve. They accordingly
represent the radix longa, and their point of origin froin the
profundus would seem to indicate that they cannot be
sympathetic nerves, unless they have an intra-cranial sym-
pathetic origin. No extra-cranial sympathetic strand going
to the ganglion could be recognised, but it is here too diffi-
cult to recognise small nerve strands for me to assert that
none existed. The radix brevis is represented by fibres that
bind the ciliary ganglion and oculomotorius together.

The two nerves that together represent the radix longa lie,
in their forward and upward course, posterior to the nervus
opticus, anterior to the rectus inferior, and ventral to the
rectus internus. Just before they join the profundus a
line of tissue runs from them toward the optic nerve.
Whether this line of tissue is wholly fibrous or partly
nervous I could not determine, and it will be further re-
ferred to in describing the ramus ophthalmicus profundus.

Anterior to the ciliary ganglion the oculomotorious con-
tinues forward to the obliquus inferior, here having its
well-known relations to the other structures in the orbit.

Nervus Trochlearis.

The nervus trochlearis issues from the skull as two branches,
as Tiesing found it, its foramen lying directly internal to the
ramus ophthalmicus superficialis. The nerve runs forward
and downward around the lower edge of the ophthalmicus
superficialis, here being closely pressed against that nerve,
but, contrary to Schwalbe’s statement (57, p. 186), without
any apparent interchange of fibres. Lying always dorsal to
all the other structures in the orbit, it goes to the obliquus
superior muscle which it, and it alone, innervates.

Nervus Abducens.

The nervus abducens issues through the large trigemino-
facial foramen, lying along the ventral surface of the tri-

MUSTELUS LA@VIS. 1338

gemino-facial ganglion, and separated from it by membrane
only. Running forward and laterally in this position to the
anterior edge of the foramen, it there turns upward, mesial to
the truncus buccalis-maxillo-mandibularis, and immediately
enters the rectus externus.

Tiesing shows the abducens in the adult issuing by a
separate and independent foramen, the foramen lying dorsal
to the trigemino-facial foramen, which seems a singular
position for it.

Kye-Muscles.

The rectus superior arose in my large embryo mainly
from the dorsal edge of the trigemino-facial foramen, but
apparently also in part from the tough membrane that closes
the foramen around the nerves that issue through it. A part
of the muscle here lies between the dorsal edge of the fora-
men and the dorsal surface of the issuing nerves. Its surface
of origin thus lies posterior to the profundus foramen, and the
latter nerve lies, as it issues from its foramen, anterior, or
even slightly antero-dorsal, to the muscle; but it is evident
that this relation of muscle and nerve is not morphologically
different from that said by Tiesing to exist in the adult, the
nerve there lying antero-ventral to the muscle. The muscle
of the embryo would simply have to travel at its origin
slightly dorsally to produce the conditions found by Tiesing
in the adult.

The rectus inferior arises from the side wall of the skull
immediately posterior to the profundus foramen, and partly
surrounding the hind edge of that foramen. The surface of
origiu of the muscle lies dorso-posterior to that of the rectus
internus, and as the muscle runs outward, downward, and
forward, it passes between the two heads or bundles of the
latter muscle.

The rectus internus is represented by two bundles of
fibres, the ventral and larger one of which arises from the
side wall of the skull between the profundus and trigemino-
facial foramina, lying ventral to the one and anterior to the

134 EDWARD PHELPS ALLIS, JUN.

dorsal edge of the other. This bundle lies ventral to both
the oculomotorius and ophthalmicus profundus nerves, near
their exits from the cranium, and also ventral to the proximal
end of the rectus inferior muscle. ‘The other, smaller and
dorsal bundle of the muscle separates, proximally, from the
ventral bundle, and running backward and dorsally, dorsal
to the inferior division of the oculomotorius, to the ophthal-
micus profundus, and also to the proximal end of the rectus
inferior, comes into contact with the ventral edge and inner
surface of the rectus superior. There it gradually diminishes
in size, and disappears without ever having reached the car-
tilage of the skull. The dorsal bundle thus lies always dorsal
to the inferior division of the oculomotorius and to the pro-
fundus nerve, while the ventral bundle lies ventral to both
those nerves at and near their exits from the skull. Farther
forward this ventral bundle of the muscle lies dorsal to the
profundus, the profundus passing downward along the lateral
surface, and then forward across the ventral edge of the
muscle.

The only part of the internus of my embryo that has an
origin on the skull thus differs from the muscle as shown by
Tiesing in the adult (62, figs. 1 and 5) in that it has a
different origin, markedly different relations to the oculomo-
torius, and somewhat different relations to the ophthalmicus
profundus. The muscle of the adult is shown by Tiesing
arising from the skull dorso-anterior to the profundus fora-
men, and from there running forward dorsal to both the
inferior division of the oculomotorius and to the profundus.
In Tiesing’s figure the ocumolotorius is not shown perforating
the rectus superior, although it is said in the text to perforate
that muscle.

Corning (14) finds the inferior division of the oculomo-
torius of the adult Mustelus perforating the rectus internus,
as Schwalbe did before him, but he does not give the rela-
tions of the profundus nerve to the muscle. Schwalbe says
that this latter nerve also perforates the muscle, near its hind
edge, and then runs forward below it,

MUSTELUS LAEVIS. 135

The relations of the muscle to the oculomotorius, as given
by Schwalbe and confirmed by Corning, would arise if the
dorsal head of the muscle of my embryo should acquire an
origin on the skull. The relations of the muscle to the pro-
fundus nerve, as given in Tiesing’s figure, would arise if the
muscle of my embryo should simply shift forward somewhat,
and then upward, at its origin. Its relations to the oculo-
motorius, as given by Tiesing, are so different from what I
find that his results would seem abnormal, if not in part in-
correct. The arrangement shown by him would arise if the
ventral bundle of my embryo should entirely disappear, the
adult muscle being represented by the dorsal bundle only ;
or by supposing that the entire muscle of my embryo had
travelled forward at its origin beyond the oculomotorius, and
then backward above it.

My embryo thus probably shows a nerve in the so-called
process of traversing one of the eye-muscles, and it is
especially to be noted that this takes place in the manner I
assumed in my work on Amia (3, p. 522), across the ends of
the muscle-fibres, and not by cutting through them midway
of their length. The nerve, however, here traverses the
muscle, or more properly becomes surrounded by the muscle,
before the muscle has entirely acquired its origin on the
skull, instead of afterwards, as I assumed, the fibres of the
muscle growing inward, from their embryonic anlage, toward
the skull on both sides of the nerve, the nerve evidently
barring their passage, and rendering a part of the muscle
wholly functionless for a considerable length of time. That
the condition shown in my embryo is not abnormal or unusual
is Shown by its being found on both sides of the head in both
my other embryos.

The rectus externus arises by two wholly separate heads.
The ventro-anterior head arises from the side wall of the skull
immediately anterior to the trigemino-facial foramen, imme-
diately dorsal to the orbital opening of the canalis transversus,
and immediately ventral to, and in contact with, the surface
of origin of the ventral head of the rectus internus. The

VOL, 45, PART 2,—NEW SERIES, L

136 EDWARD PHELPS ALLIS, JUN.

dorso-posterior head arises from the side wall of the skull
immediately dorsal to, and contiguous with, the dorsal edge
of the trigemino-facial foramen. It there lies ventral to the
surface of origin of the rectus inferior, and anterior to that
of the rectus superior. The muscle did not have a straight
course in my sections, but this may have been due to the
action of reagents. When near the eyeball it turned for-
ward, then directly outward to the eyeball, and then back-
ward around it.

The obliqui superior and inferior arise close together,
at the anterior end of the orbit. The superior muscle arises
from the orbital wall immediately dorsal to the orbital
opening of a canal by which the ophthalmicus profundus
traverses the antorbital process. ‘The inferior muscle arises
immediately posterior to that opening. ‘The latter muscle
here lies immediately internal to the anterior end of the
muscle Add of Vetter’s nomenclature, and immediately
dorsal or dorso-anterior to the orbital opening of a large
canal that extends from the front end of the orbit into the
hind end of the nasal capsule.

This latter canal transmits a large branch of what Gegen-
baur (23, p. 77) considered, in adult selachians, as a peri-
orbital lymph sinus. Parker (47) has, however, since
described, in the adult Mustelus antarcticus, a large
venous orbital sinus that would seem to be the same vessel that
Gegenbaur described as a lymph sinus. In my embryos this
sinus is certainly a part of the venous system, and if both
lymph and venous sinuses are here found in the adult, they
must both have been cut off from this blood sinus of embryos.
That this is what probably takes place seems shown by the
fact that, in Amia, the peri-orbital sinus is certainly a lymph
one, and that it has the same position and relations to the
other orbital structures that the blood-sinus of my embryo of
Mustelus has. From the anterior end of this sinus, in
Mustelus, the large branch above referred to arises, and,
traversing the canal in question, enters the hind end of the
nasal capsule. From there it sends a branch downward

MUSTELUS LAiVIS. 157
through the ventral opening of the capsule, the branch then
turning mesially along the ventral surface of the skull, and,
at the middle line of the head, anastomosing with its fellow
of the opposite side. This large branch of the peri-orbital
sinus is thus certainly the anterior facial vein of Parker’s
descriptions of the adult. As it traverses its canal in my
embryos it is not accompanied by any other structure, so far
as I could find, certainly not by any branch of the ramus
ophthalmicus profundus. The canal is thus not the exact
equivalent of the orbito-nasal canal of Gegenbaur’s descrip-
tions. But if it should, in the adult, become fused with the
profundus canal, to be later described, and that canal should
become wholly shut off, by cartilage, from the cranial cavity,
a canal would arise which would seem to be the equivalent
of Gegenbaur’s canal. The canal, in my embryo, is quite
certainly the homologue of the orbito-nasal canal of my
descriptions of Amia (8, p. 514), and that canal is thus
probably not the exact homologue of the orbito-nasal canal
of Gegenbaur’s descriptions of selachians. The name orbito-
nasal is, however, properly applicable to it. It is evidently,
in origin, wholly separate from and independent of the pro-
fundus canal.

Another branch of the peri-orbital sinus of my embryos
traverses the canalis transversus, in exactly the manner that
Gegenbaur says that a branch of his lymph sinus traverses
the same canal in the adult (238, p. 77). This branch is not
described by Parker in his descriptions of the blood-vessels
of Mustelus antarcticus. If, nevertheless, it be venous in
the adult Mustelus levis, as it certainly is in embryos, or
even if it is simply derived from a sinus that is venous in
embryos of the age of my oldest ones, the canal it traverses
must be quite differently judged, in any comparison with
teleosts, from what it would be if the vessel were a lymph
one of independent origin. ‘The homology proposed by
Gegenbaur and Sagemehl (56) of this canalis transversus of
selachians with the eye-muscle canal of teleosts and ganoids
is, in fact, wholly based on the supposition that the canal in

138 EDWARD PHELPS ALLIS, JUN.

teleosts and ganoids lodges or transmits, as the canal im
selachians was asserted to, a lymph vessel which arises as a
branch of a peri-orbital lymph sinus. This assumption I
found to be incorrect in so far as Amia is concerned (8), the
eye-muscle canal of that fish being traversed by a vein, and
not by a lymph sinus, and a separate and wholly distinct
transverse canal transmitting a lymph vessel from orbit to
orbit. I hence concluded that the eye-muscle canal of Amia
and teleosts was not derived from the canalis transversus of
selachians. This opinion must certainly now be altered, for
if the canal in selachians transmits a vessel that is primarily
venous, it is evident that it must form a part of the eye-
muscle canal of Amia, which canal also transmits a venous
vessel.

One other foramen in the orbital region of Mustelus is to
be noted. It lies near the anterior end of the orbit, dorso-
posterior to the orbital opening of the orbito-nasal canal, and
anterior to all the foramina of the nerves, the nervus opticus
included. It is, in all probability, traversed by a branch of
the peri-orbital sinus, though I could not positively establish
this. A branch of the sinus penetrates the membranes that
line the outer surface of the cartilage, and is seen, for a
certain number of sections, as a vessel lying between the
outer and inner lining membranes of the skull. The two
edges of the foramen had then been pressed toward each
other by contraction or displacement in manipulation, and I
could not see that the vessel entered the cranial cavity. In
this latter cavity two blood-vessels approached the mem-
brane that lined the inner surface of the opening in the
cartilage, and there united. These two vessels seemed to be
simply two parts of a single vessel that approached the fora-
men and then bent sharply away from it, without having any
connection whatever with the external vessel. If, however,
such a connection existed, the intra-cranial vessels would be
branches of the branch of the peri-orbital sinus, and the vessel
would probably be the anterior cerebral vein of Parker’s
descriptions of the adult Mustelus antarcticus (47).

MUSTELUS LAVIS. 139

In my 55 mm. embryo this cerebral branch of the peri-
orbital sinus traverses the skull and then runs forward, as a
large vessel, along the lateral surface of the brain, between
it and the skull. Mr. Nomura here found a foramen in the
embryo he dissected.

Review and Comparison of Hye Muscles.

Corning (14), in his recently published and excellent work
on the eye-muscles and their innervation in the several
classes of vertebrates, to which reference has already several
times been made, comes to the conclusion expressed in the
following sentence :—“ Ich halte also ander Homologie der vom
Oculomotorius innervirten Muskeln fest.” This is directly
opposed to my own published opinion regarding these
muscles (3), and I cannot see that Corning in any way satis-
factorily explains the rather troublesome facts that I brought
forward in support of it. He is obliged, in fact, in order to
establish his proposition, to make certain assumptions that
have less support in fact than my assumptions had. To
explain the unusual innervation of the rectus inferior in
Petromyzon, by a branch of the abducens, he assumes that
the fibres destined to that muscle become detached from the
oculomotorius and attached to the abducens, intra-cranially.
And yet, although this interchange of fibres must surely be
extra- and not intra-cerebral, he did not establish it in the
specimen he examined, and it is in no way indicated in any
of the numerous works on the fish. To explain the important
differences in the innervation of the rectus internus, by a
branch of the superior division of the oculomotorius in certain
classes of vertebrates, and by a branch of the inferior
division of the same nerve in others, Corning assumes that it
would be sufficient, in order to produce the arrangement
found in the ones, for the opticus and oculomotorius of the
others to change slightly, at their exits from the cranium,
their relations to the recti muscles. Here he wholly over-
looks the very important fact that, in any such shifting
about of the points of exit of the two nerves in question, that

140 EDWARD PHELPS ALLIS, JUN.

branch of the oculomotorius that innervates the rectus
internus would still necessarily always remain on the same
side of the opticus, ventral or dorsal as the case may be. He
chooses Esox to represent one of the two manners of innerva-
tion, and Carcharias to represent the other. Take his
figure 4, which shows the muscles and nerves of Hsox, and
imagine the rectus internus to shift, at its origin, forward in
front of the opticus and then backward above it. This is
simply what Corning proposes, but in the reverse direction,
and it is the only way in which the muscle can be brought
dorsal to the opticus without assuming that the nerve cuts
through the muscle or the muscle through the nerve. The
muscle itself, in thus shifting at its origin, would evidently
acquire the selachian position, as Corning asserts, but the
nerve that innervates the muscle would most certainly not.
It would still remain ventral to the opticus.

In Esox the internal and inferior recti are not innervated,
as I assumed, from earlier descriptions, that they were in all
teleosts, and as they are in Amia. Corning calls attention to
this, and Herrick (82), before him, had called attention to
a similar difference in Menidia. I myself had already found
_this to be true of Scomber and certain other teleosts, and had
called attention to it im a work on Scomber sent to press now
nearly three years ago. In all these fishes the arrangement
seems practically to be that the superior division of the
oculomotorius innervates the rectus superior alone. A branch
is then given off which separates into two parts, one for the
rectus internus and the other for the rectus inferior, both
branches lying morphologically ventral to the optic nerve,
but the branch to the rectus internus passing dorsal to the
rectus inferior, instead of postero-ventral to it as it does
in Amia. The remainder of the oculomotorius then runs
downward and forward around the hind edge of the rectus
inferior, and supplies the obliquus inferior. Herrick con-
siders this latter’part of the nerve as the first branch of the
oculomotorius, the remainder of that nerve later separating
into dorsal and ventral portions, the former for the rectus

MUSTELUS LAVIS. 141

superior and the latter for the recti internus and inferior. This
latter portion, which Herrick considers as the continuation of
the main nerve itself, runs forward dorsal to both the muscles
it innervates, and it is to it that the ciliary ganglion is said
to be related in Menidia, as I also find it in Scomber.

Herrick says (82, p. 237) that the “ deviations” in the inner-
vation of the internal and inferior recti in Menidia, from that
given by me in Amia, “can be easily explained mechanically
by the great size of the eyes, and the consequent crowding
of the recti muscles far backward.” Corning explains the
differences between Ksox and the chick, the chick agreeing
practically with Amia, by assuming (p. 136) that the rectus
inferior of Hsox acquires an origin dorsal to that branch of
the oculomotorius that innervates the rectus internus. He
does not say that the muscle traverses or is traversed by the
nerve in this process, but it is evident that one of these two
things must happen, for no simple shifting of the muscles at
their origins could derive one arrangement from the other.
This apples to Menidia as well as to Ksox, and Herrick
doubtless recognised it, for although he would explain the
change in the relations of nerve and muscle by a different
principle, as explained below, he assumes that it could be
easily accounted for by a principle said to have been invoked
by me; his words being (p. 237) that “he (Allis) invokes a
principle to account for the diverse relations of nerve and
muscle in elasmobranchs, which, if applied more broadly,
might weaken the phylogenetic value of some of his other
cases.” ‘This statement of Herrick’s is, however, clearly
based on a misconception of the principle referred to as
invoked by me, which principle was contained in the state-
ment that the muscles I was discussing in elasmobranchs
“at their origins, either traverse or are traversed by the
issuing nerve.” I expressly eliminated the probability
of a nerve cutting through a muscle or a muscle through
a nerve “‘ midway in its length” (8, p. 522).

This is, in fact, the vital point of the whole discussion.
Does a nerve ever cut through one of the eye-muscles “ mid-

142 EDWARD PHELPS ALLIS, JUN.

way in its length”? Or does one of the muscles, either at
its origin or elsewhere, ever traverse a nerve “ midway in its
length”? I assumed, in my discussion of the subject, that
neither of these things occur, and I know of no single fact
that proves the contrary. I also assumed, in my discussion
of these nerves and muscles, and it was definitely implied
though not definitely stated, that an eye-muscle nerve once
laid down on one side of any of the eye-muscles, or on one
side of any of the nerves that innervate those muscles, or
that traverse the orbit, would always be laid down in the
same relation to that structure. I then further assumed
that the nerve once laid down never changed this embryonic
relation to another nerve, but that it might and did change
its relation to a muscle in the manner set forth in the
principle above referred to as invoked by me. In this
assumption that the nerves are always laid down in the same
relations to the eye-muscles i was quite unquestionably in
error, as my present work on Mustelus shows. For it is
evident that since the nerves that innervate the eye-muscles
or that traverse the orbit are relatively well developed before
the eye-muscles have acquired their attachments on the skull,
they might be so placed, because of correlation to other
parts, as to obstruct a muscle as it sought its cranial attach-
ment, and hence deflect it, or even split it, as the obstructing
nerves split the rectus internus in Mustelus. ‘This is, how-
ever, in reality simply an application, in embryonic stages,
of the principle I invoked for adults; that is, the muscle
here, in principle, traverses, at the end that is to become its
end of origin on the skull, a nerve that it encounters. This,
it will be readily seen, is a totally different thing from the
assumption that the nerve cuts through the muscle fibres;
and also totally different from another principle frequently in-
voked in this connection, and which isvery definitely expressed
by Herrick (82, p. 237) in the following sentence :—“ If this
be true (that the eye-muscle nerves grow directly out from
the brain), I see no reason why a given motor nerve should
not grow out either above or below some other structure

MUSTELUS LEVIS. 143

depending upon the peripheral relations of its end organ
with reference to that structure.” This seems to me too
dangerous a principle to be applied without definite facts to
support it in each particular case, and such facts do not here
exist. Pushed to a legitimate extreme, a nerve might even
be assumed to change its relations to such an obstacle as a
visceral cleft. In any event a pure assumption is here
invoked by Herrick to explain certain facts that I had sought
to explain by what was, at the time, an equally pure assump-
tion; that is, that the internal recti muscles of vertebrates
are not all homologous structures. Recent work relating to
this especial subject supports, however, my assumption. For
both Hoffmann (84) and Sewertzoff (58) say that the rectus
superior and rectus internus of Acanthius and Torpedo re-
spectively arise from the dorsal wall of the head cavity from
which all the muscles innervated by the oculomotorius arise ;
while Rex says (58, p. 255) that in the duck the rectus
internus and the rectus inferior arise from a single embryonic
muscle-mass that has its origin from the ventral wall of the
same cavity, the rectus superior alone arising from the dorsal
wall of the cavity (p.238). The bundle of fibres that I found
in one specimen of Amia hanging festoon-like between the
rectus internus and rectus inferior muscles thus receives an
evident explanation, and, inversely, tends to confirm for Amia
the origin ascribed by Rex to the corresponding muscles in
the duck. ‘This thus seeming to be established, the conditions
presented in Mustelus seem to be such as might have neces-
sitated or led to the development of another and different
internus. The internus in this fish is so obstructed in its
effort to obtain an attachment on the skull, that a con-
siderable part of it remains for a long time wholly func-
tionless. Had it been still more obstructed the muscle
might not have been able to maintain its separate and
independent existence, and a new and wholiy different in-
ternus would have been developed.

Herrick calls attention (32, pp. 2834—236) to two apparent
errors in my diagrammatic representation of the eye-muscles

144 EDWARD PHELPS ALLIS, JUN.

and nerves of vertebrates, and he considers them of sufficient
importance to ‘‘suggest that his (my) entire phylogenetic
scheme should be received with some reserve.” One of these
apparent errors is that I have shown the ophthalmic nerves
in all my figures arising “from a common stem which hes
ventrally of the III nerve.” The nerves in my diagrams are
all intended to be shown cut some distance after they issue
from their foramina, and I purposely avoided attempting to
show their relations at their exits. The data on which my
diagram was based were too meagre and too conflicting to
warrant my attempting to show anything that could possibly
be omitted. The other apparent error relates to the position
of that branch of the oculomotorius that innervates the rectus
superior in Petromyzon. I showed this branch in my figure
lying over instead of under the ramus ophthalmicus.~ In the
text (p. 523) I called attention to the fact that Fiirbringer says
that it runs under the ramus ophthalmicus, but I assumed
this to be an error. As Corning says (14, p. 129) that he
has confirmed the correctness of Fiirbringer’s statements, this
simply adds another and important variation to be accounted
for in the innervation of the eye-muscles.

In my work on Scomber, still in press, I suggested that
the differences in the innervation of the rectus internus and
rectus inferior in that fish, and in Amia, might be explained
by the assumption that the internus of the one was the
inferior muscle of the other, ganoids and teleosts thus
representing different lines of descent from my proto-urodele
type. As, however, Workman (68) has recently shown that
the eye-muscles of Amiurus melas are innervated much as
they are in Amia, while in Pomatomus they are innervated
exactly as they are in Menidia, it is evident that it is
useless to speculate on the subject until further facts have
been accumulated.

Trigemino-facial Complex.
The trigemino-facial ganglion is partly intra-cranial and
partly extra-cranial in position. The apparent roots by which

MUSTELUS LAVIS. 145

it arises from the brain could not be satisfactorily determined
in my 12:2 cm. embryo, because of the crowding together of
the parts here concerned, and the different bundles of fibres
that compose the several roots could not be separately traced
through the ganglion, my sections not having been prepared
for this especial research. In my 55 mm. embryo the
apparent roots could, on the contrary, be easily distinguished,
and the different components of the several peripheral nerves
could be determined in a general, but certainly not in a
complete or absolutely accurate way. From a consideration
of the two embryos the following approximately correct
results were arrived at.

The so-called anterior root of the trigeminus is formed by
two somewhat distinct rootlets—a dorsal one, having a deeper
origin in the medulla, and a ventral one, having a more
superficial origin. The dorsal rootlet is certainly largely
formed by a bundle of fibres that have a course corre-
sponding exactly to that ascribed by Haller (28) to the
inner motor trigeminus root of his descriptions of Scyllium.
The ventral rootlet certainly arises largely from the ‘ dorso-
laterale Langsbahn” of Haller’s descriptions, that is, from
the ascending fifth tract of English authors. Haller says
(p. 437) that in Scyllium all the sensory fibres of Trigeminus |
are derived from this latter tract, and this, if true of Scyllium,
must doubtless also be true of Mustelus. Both rootlets in
Mustelus enter the anterior portion of a large intracranial
ganglion. Different regions, rather than separate and in-
dependent parts of this ganglion, can be distinguished, and a
considerable part of the fibres of the ventral rootlet are seen
to enter a ventro-lateral part of the ganglion, which, opposite
the anterior half of the ganglion, is separate and independent
from the rest of the ganglion, but posteriorly is completely
fused with it. From this lobe or process of the ganglion the
ramus ophthalmicus profundus arises, that nerve hence, very
probably, being composed entirely of general cutaneous
fibres derived from the ascending fifth tract. The ramus
profundus of Mustelus can thus in no way be the homologue

146 EDWARD PHELPS ALLIS, JUN.

of such a nerve as the so-called ramus profundus of Haller’s
descriptions of Scyllium, in which fish the nerve is said
(p. 438) to be exclusively motor, and to be derived entirely
from the inner motor trigeminus root and the anterior upper
motor trigeminus nucleus. That Haller has here certainly
mistaken a superficialis ophthalmic nerve for a profundus
one will be later shown. That he has also made some
further mistake is evident, for the superficial ophthalmic
nerve of Scyllium is certainly in no part a motor one.

From the antero-dorsal corner of the intracranial gan-
glion of Mustelus a bundle of fibres, which Haller calls in
Scyllium the ventral ‘‘ Wurzelportion” of the ramus ophthal-
micus superficialis, is sent forward and outward to join on its
ventral aspect what Haller calls the upper portion of the
posterior root of the trigeminus. This bundle of fibres, in
Mustelus, has its apparent origin from the intracranial
ganglion opposite the point where the inner motor trigeminus
root of Haller’s nomenclature (my dorsal rootlet) joins it.
The fibres of the bundle, however, do not come from that
rootlet, but come upward along the lateral surface of the
ganglion, from that part of the ganglion that has its origin
in relation to the dorsal portion of the fibres that arise from
the ascending fifth tract. The bundle is thus seen, from its
origin alone, to certainly be composed, in large part, of
general cutaneous fibres, and its peripheral distribution, to
be later given, shows that it is quite unquestionably entirely
so composed. Motor fibres certainly do not exist in it.

From the posterior part of the intracranial ganglion
formed on this anterior trigeminal root the truncus maxillo-
mandibularis trigemini has its origin. Issuing from the
skull by the trigemino-facial foramen, this truncus, or root
of the truncus, there becomes more or less confused with the
root or stem of the facialis, and also with the hyoidean part
of the posterior trigeminal root. A large extracranial
ganglion is formed on these several stems.

The so-called posterior root of the trigeminus, or Tri-
geminus II, is formed in Mustelus, as it is in Scyllium, of

MUSTELUS LAEVIS. 147

two portions,—a dorsal anda ventralone. There is not here,
however, a strict agreement with the conditions described
by Haller in Scylium. The dorsal root, which arises in the
so-called lobus trigemini, separates into three parts, two of
which run downward close against the lateral surface
of the brain, while the third one turns outward and
forward, enters the ramus ophthalmicus superficialis, and
forms the lobus trigemini component of that nerve. The two
other portions, running almost directly downward, straddle
the nearly horizontal ventral portion of the root, which
latter portion arises from the “ tiussere sensorische Oblongata-
gebeit”’ of Haller’s descriptions, that is, from the tuber-
culum acusticum. One portion, or bundle, of the fibres
of this ventral root continues directly forward, and,
joining that bundle of the dorsal root that goes to form part
of the ramus ophthalmicus superficialis, forms the tuberculum
acusticum component of that nerve. The remaining portion,
or bundle, of the ventral root turns sharply downward with
the two descending bundles of the dorsal root, the three
bundles becoming confused in the sections, and together
forming a large stem, which issues from the skull through
the trigemino-facial foramen, a part of it going to form the
ramus buccalis and ramus oticus facialis, and a_ part
becoming lost in the facial, or hyoidean, part of the extra-
cranial trigemino-facial ganglionic complex.

On this so-called posterior trigeminal root, thus formed of
two rootlets, there is no intracranial ganglion whatever, and
not even any ganglion cells, excepting just as the roots issue
from the skull. Externally to the skull, on both the oph-
thalmic and buccal parts of the root, there is a large,
separate, and independent ganglion; but no separate gan-
ghon is found related to the hyoidean branch of the root.
The two rootlets together form the so-called lateral sensory
root or roots of all those authors that do not treat them as a
part of the trigeminus. It is, however, to be especially
noted that both the ophthalmic and buccalis-hyoidean parts
of the root receive fibres from two quite distinct regions of

148 EDWARD PHELPS ALLIS, JUN.

the brain—the lobus trigemini, and the tuberculum acusticum.
The assertion that the lobus trigemini is a part of the
tuberculum acusticum, somewhat differentiated as a separate
lobus, does not seem sufficient in itself to account for this
origin in Selachians, and also in Acipenser (89), of the
so-called lateral fibres from two distinct centres, while in
ganoids and teleosts they arise from a single centre, the
tuberculum acusticum (40). While I have wholly failed in
my attempt to trace in Mustelus the fibres from the
lateral canal organs to one of these two centres, and those
from the ampullary organs to the other, it is evidently a
legitimate supposition that such may perhaps be their
central origin, and this has already been suggested by
Strong (60, p. 194). If it be admitted, the natural con-
clusion must be, as Strong has said, that the ampullary
organs represent the terminal buds of Amia and other fishes,
and not organs similar to the pit organs found on the head of
Amia. That the ampulle may be derived from, or be directly
related to, certain of the so-called pit-organs of current
descriptions of teleosts is, however, quite probable, and will
be again referred to.

The facial root of the trigemino-facial complex of Mustelus
arises in my 55 mm. embryo by two rootlets, one of which
has its central origin in the lobus vagi, while the other
contains motor fibres, Mustelus thus apparently fully agreeing
with Scyllium. On this facial root there is no important
intra-cranial ganglion, though ganglion cells begin on its
dorsal surface just before it leaves the skull. It issues
through the postero-ventral part of the large trigemino-facial
foramen, and a large extra-cranial ganglion immediately
forms on it, this ganglion being somewhat separate from the
extra-cranial ganglion on the truncus maxillo-mandibularis,
and lying postero-ventral to it.

The trigemino-facial nerves of my embryos of Mustelus
thus issue from-the skull by the three well-known trunks or
stems, and each stem has its own special foramen. All three
of these foramina pierce the cranial wall in the orbital region,

MUSTELUS LAEVIS. 149

one lying near the roof of the orbit, one near its floor, and
the other intermediate between the two. Schwalbe (57, p.
183) says that the ophthalmicus superficialis and ophthal-
micus profundus issued by a single foramen in the adult
specimen of Mustelus levis that he examined, the tri-
gemino-facial nerves thus issuing from the cranial cavity
by two instead of by three foramina. ‘liesing (62), on the
contrary, shows, in the adult, three separate foramina exactly
as I find them in embryos.

The dorsal one of the three trigemino-facial foramina of
my embryos is the most anterior one of the three, and it
transmits the so-called ramus ophthalmicus superficialis, this
nerve containing a large component derived from the tuber-
culum acusticum, another large component derived from the
lobus trigemini, and a third and smaller one derived from the
so-called anterior or first root of the nervus trigemini. ‘The
first two of these three components together form what is
now usually described as the ramus ophthalmicus superficialis
facialis, the third one being the ramus ophthalmicus super-
ficialis trigemini. The latter nerve is certainly largely com-
posed of general cutaneous fibres, though it may perhaps also
contain a certain number of communis fibres, for although
Haller says (p. 437) that in Scyllium the seusory part of
Trigemini I is formed entirely of fibres derived from the
ascending trigeminus tract, this tract is said to have pre-
viously received a bundle of fibres coming from the lobus
vagi (p. 431). If, then, the ampulle of selachians represent
the terminal buds of Amia and teleosts, these fibres from
the lobus vagi must be so-called visceral-sensory communis
fibres, which have naturally not suffered the modification
that I assume that the terminal bud fibres are undergoing.

The middle foramen is the next anterior one of the three,
and it transmits the ramus ophthalmicus profundus, derived
from a ventro-lateral and somewhat separate and independent
part of the intra-cranial trigeminus ganglion.

The third trigemino-facial foramen is much the larger one
of the three. It les ventral and posterior to the other two,

150 EDWARD PHELPS ALLIS, JUN.

and transmits the united stems of the hyoideo-mandibularis
facialis, maxillo-mandibularis trigemini, and buccalis facialis.

In one of my embryos, dissected by Mr. Nomura, a fourth
trigemino-facial foramen was found. It lay slightly in front
of the ophthalmicus superficialis foramen, was small, and
transmitted a small bundle of fibres which immediately joined
the ramus ophthalmicus superficialis. The relative size of
the two foramina by which the ophthalmicus superficialis thus
issued from the skull in this specimen would lead one to suppose
that the smaller foramen transmitted the general cutaneous
component of the nerve, the larger one transmitting the united
lobus trigemini and tuberculum acusticum components.

Ramus Ophthalmicus Superficialis.

The ramus ophthalmicus superficialis, after it issues from
its foramen, immediately becomes ganglionic, as already stated,
and is a broad, flattened nerve, which hes against and runs
forward along the side wall of the skull, immediately ventral
to the projecting and overhanging roof of the orbit. The
nerve is here formed of two perfectly distinct parts—a large
ganglionated dorsal portion, and a small non-ganglionated
ventral one. The dorsal portion contains the fibres destined
to supply the sensory organs of the lateral canals and those
of the ampulle, and may be referred to as the special sensory
portion. The ventral portion contains the general sensory
fibres of the nerve.

The descriptions of this nerve and of the others that follow
are based entirely on the conditions found in the 12:2 em.
embryo examined by sections, and hence do not agree exactly
with the figures.

Immediately outside its foramen the ophthalmicus super-
ficialis gives off two branches, one arising from the dorsal,
ganglionic portion of the nerve, and the other from its ventral,
non-ganglionated portion, this latter branch running upward,
as a flat nerve, along the mesial surface of the special sensory
portion of the nerve. Dorsal to the main nerve the two
branches anastomose to a greater or less extent, and together

MUSTELUS LAVIS. 151

perforate the overhanging orbital roof. There they turn back-
ward along the dorsal surface of the roof and separate, the
special sensory nerve, which is purely a lateral sensory one,
innervating the posterior sub-group of the supra-orbital
organs,—that is, organs 87 to 93 of that line, while the
general sensory branch goes to the general tissues. This pair
of nerves together form the first one of the so-called frontal
branches of the main nerve, the portiominor of the nerve thus,
contrary to Schwalbe’s statement (57, p. 184), taking part in
the formation of these frontal nerves. Schwalbe says that
the portiominor only gives off one branch during its passage
through the orbit, that branch being the communicating
branch to the trochlearis—a branch that I do not find.

Shghtly anterior to this first pair of frontal branches a
second pair is given off, one a general sensory branch, and
the other a lateral sensory one destined to innervate supra-
orbital organ 86. Both of these branches have a forward
instead of a backward course. Anterior to them two or three
general sensory branches and five lateral sensory ones are
successively given off, the latter branches subdividing and
innervating organs 76 to 85 supra-orbital. These five lateral
sensory branches all pierce the orbital roof by separate
foramina, the general sensory branches either accompanying
them or traversing separate foramina of theirown. All these
branches have a forward and upward course.

Beyond the root of the most anterior one of the five lateral
sensory branches above referred to, that is, anterior to the
branch that innervates supra-orbital organs 76 and 77, and
near the front end of the orbit, the ophthalmicus superficialis
itself pierces the orbital roof, and reaches its dorsal surface.
In the slightly younger specimens examined by dissection
the nerve here traversed a notch in the orbital roof, as shown
in fig. 4, instead of piercing it. Shortly before leaving the
orbit it passes dorsal to the nervus trochlearis, and there
comes into intimate juxtaposition with that nerve, but, as
already stated, without any indication whatever, that I could
observe, of any interchange of fibres.

VOL. 45, PART 2,—NEW SERIES. M

152 EDWARD PHELPS ALLIS, JUN.

As, or shortly after, the ophthalmicus passes upward out
of the orbit a lateral sensory branch is sent to organs 74 and
75, and two others to organs 65 to 73, organ 65 being the
middle one of three organs that lie in that short section of the
supra-orbital canal that les between the two bends in the
canal, slightly in front of the eye. At the same time that
these three lateral branches are given off the larger part of
the general sensory component of the nerve passes gradually
upward, as two nerve strands, along the mesial surface of
the remainder of the nerve. ‘These two nerve strands leave
the main nerve separately, and then, farther forward, unite
to form a single nerve, which here lies immediately dorsal to
the remaining and much larger part of the nerve. This
latter, deeper part of the nerve now consists almost entirely
of special sensory fibres, but there is still a small bundle of
general sensory fibres in its ventral portion.

Having issued from the orbit the two strands now formed
of the originally single nerve he at first along the sloping
surface of the lateral wall of the brain case, immediately
dorsal to the anterior end of the roof of the orbit, and then
pass dorsal to the antorbital process. Anterior to that
process a large general sensory branch is given off from the
ventral aspect of the main nerve. Running forward slightly
below the main nerve it soon turns outward toward that part
of the supra-orbital canal that hes anterior to the double
bend in front of the eye, and accompanies it in its forward
course. Slightly anterior to the point where this general
sensory branch is given off a large lateral canal branch is
sent downward and backward, at first along the lateral
surface of the brain case, and then along the lateral surface
of the nasal capsule. From it numerous branches, some of
them quite long, run forward and downward along the side
wall of the brain case and nasal capsule, and innervate organs
1 to 16 supra-orbital, the section of canal thus innervated
lying on the ventral surface of the head, in the region lateral
to the nasal aperture. This large lateral branch leaves the
main nerve directly dorsal to the foramen by which the

MUSTELUS LAVIS. 153

ophthalmicus profundus issues from the brain case, after
having re-entered it near the anterior end of the orbit. It
is evidently the homologue of the branch s. of.? of Ewart’s
(18) descriptions of other selachians.

Slightly anterior to the large lateral branch to organs 1 to
16, a lateral branch is given off to supply organs 59 to 64
supra-orbital, which organs lie in or near the distal one of the
two bends of the supra-orbital canal immediately in front of
the eye.

Anterior to this point the main nerve, running forward,
comes to lie in the angle between the side wall of the
brain case and the dorsal surface of the bulging side wall
of the nasal capsule. The ramus ophthalmicus profundus,
having issued a second time from the cranium, here lies
slightly ventro-lateral to the main ophthalmicus superficialis,
the large general sensory branch of the latter nerve lying
slightly dorso-mesial to it. As the main nerve continues for-
ward, in this position, it soon reaches a point where the
cartilaginous side wall of the brain case is entirely replaced
by membrane, this perforation of the cartilaginous skull being
the much enlarged prefrontal opening of Gegenbaur’s descrip-
tions of the adult. The nerve here hes lateral to the now
membranous side wall of the brain case, on the dorsal surface
of the cartilaginous nasal capsule, slightly lateral to its highest
point. In this position it runs forward, and soon gives off,
from its dorsal aspect, two large ampullary sensory branches
which run at first forward, and somewhat dorsally and
mesially. The more superficial of these two branches here
soon breaks up and supplies the organs of that sub-group of
the superfizial ophthalmic group of ampul'z that has its sur-
face pores on the dorsal surface of the head mesial to the
orbit. The ampulle themselves, of this group, lie, in
12:2cm. embryos, along the lateral surface of the membranous
anterior end of the brain case, and immediately dorsal to the
nasal capsule. The tubules of the lateral ampullz of this sub-
group lie internal to, and separated by membrane from, the
tubules of that sub-group of the ophthalmic ampulle that has

154 EDWARD PHELPS ALLIS, JUN.

its surface pores in, and posterior to, the ventral one of the
two bends of the supra-orbital canal immediately in front of
the eye. The organs of this latter sub-group of ampulle, and
also those of the other, more anterior sub-group of the super-
ficial ophthalmic ampullz, are innervated by the deeper one
of the two ampullary branches above referred to. This branch,
running forward dorsally and mesially, reaches a position
dorso-mesial to the base of the dorso-lateral rostral bar of
cartilage, and then, sinking deeper, comes to lie mesial to
the bar, where it breaks up and supplies the organs of its two
sub-groups of ampulle.

In that part of the course of the main ophthalmic nerve
where these two ampullary branches are given off the lateral
canal fibres of the nerve have been gradually collecting in
two regions—a dorsal and a ventral one. The dorsal fibres
soon separate as a separate branch, and from it ten or more
branches are given off, some of them arising from the main
nerve, from the bundle of fibres destined to form the branch,
but before these fibres have entirely separated as a separate
branch from the main nerve. Branching, these several lateral
sensory branches supply organs 35 to 58 supra-orbital, all of
which lie proximal to the point where the canal anastomoses
with the anterior end of the infra-orbital canal, the main
branch ending in two branches which supply, the one organs
37 and 38, and the other organs 389 and 40. Organs 38 and
39 lie one directly dorsal to the other in the anterior bend
of the canal, where it passes upward from the ventral to the
dorsal surface of the snout. A general sensory nerve, derived
from the main nerve, accompanies this lateral nerve, and be-
yond its anterior end breaks up into a number of terminal,
general sensory branches.

The remaining, ventral, and larger part of the main
ophthalmic nerve contains the lateral sensory fibres destined
to organs 17 to 34 supra-orbital, the ampullary fibres destined
to the deep ophthalmic group of ampulle, and certain general
sensory fibres destined mainly to supply the prenasal parts of
the snout. Organs 17 to 34 all lie in the subrostral part of

MUSTELUS LAVIS. 155

the supra-orbital canal, distal to the point where the canal
anastomoses with the anterior end of the infra-orbital canal.
The fibres destined to supply the organs leave the main nerve
in several branches. all of which run forward to their destina-
tion, neither any of them, nor the main nerve itself, having a
recurrent course, such as Ewart shows in his fig. 2. Ewart
has here, I think, combined the profundus and certain of the
terminal branches of the ophthalmicus superficialis, for the
profundus turns backward in the manner shown by him for
the superficialis, and Mr. Nomura, in his first dissection, con-
sidered it as the distal end of the latter nerve.

This ventral part of the main ophthalmic nerve, in the
region where the lateral sensory branch to organs 35 to 58 is
given off, comes into intimate juxtaposition with the ramus
ophthalmicus profundus, and so remains through about 100
sections. ‘There was here, apparently, no interchange of
fibres, but the relations of the two nerves are much too
intimate for me to assert that it does not take place. One
lateral sensory branch, the one to organs 55 and 56 supra-
orbital, also here comes into such close relations with the pro-
fundus that it is impossible to say that it does not join it,
instead of joining the ophthalmicus superficialis. Neverthe-
less, I should certainly have decided that it did not join the
profundus were it not that Cole finds, in Chimera, a lateral
branch that has its apparent origin from the profundus, and
that supplies two supra-orbital organs lying in a part of the
canal that is strikingly similar to that in which my organs 55
and 56 are found. The exact innervation of these two organs
is morphologically too important not to demand further in-
vestigation.

Near the anterior end of the nasal capsule the profundus
nerve separates from the superficial ophthalmic one, turns
downward around the anterior end of the nasal capsule, and
then backward along its ventral surface. The main super-
ficial ophthalmic nerve now acquires a position close to, and
ventral or ventro-lateral to, the dorso-lateral rostral bar of
cartilage, the separate, general sensory branch acquiring

156 EDWARD PHELPS ALLIS, JUN.

position dorsal or dorso-lateral to that bar. The main nerve,
continuing forward, here breaks up into many branches, some
destined to supply the supra-orbital canal organs 17 to 34, as
already stated, certain others being general sensory ones,
but much the larger part of the nerve going to supply the
deep ophthalmic group of ampulle.

Ramus Ophthalmicus Profundus.

The ramus ophthalmicus profundus issues from the skull by
a special foramen, as already stated, the foramen lying some-
what posterior to, and nearly in line with, the foramen for the
nervus oculomotorius. As the nerve issues from its foramen
it lies immediately anterior to the surface of insertion of the
rectus inferior muscle, immediately dorsal to the ventral head,
or bundle, of the rectus internus, at or near its insertion, and
ventro-anterior to the rectus superior. Running forward and
slightly outward from there it passes dorsal to and then lateral
to the inferior branch of the oculomotorius, which latter
nerve here runs backward. The profundus here lies posterior
and ventral to that branch of the oculomotorius that inner-
vates the rectus superior muscle, and posterior to the point
of origin of the several branches to the rectus internus, thus
crossing the oculomotorius distal to those branches. The pro-
fundus thus here lies in its well-known position—between the
superior and inferior divisions of the oculomotorius, ventral
to the one and dorsal to the other, Running outward and
forward the nerve passes through the interval between the
two bundles of the rectus internus, already described, here
accompanying the inferior branch of the oculomotorius, but
running in the opposite direction to that nerve. While in
this interval a branch arises from the lateral surface of the
profundus, and there is, at this point, a small group of gan-
glion cells which lie on the external surface of the profundus
and along the base of the branch in question. These cells
have the size and general appearance of the ganglion cells of
the cerebro: spinal ganglia, and not that of those of the ciliary
ganglion. They would seem, accordingly, to represent a

MUSTELUS LAVIS. 154

detached and extra-cranial portion of the profundus ganglion.
That they represent the entire profundus ganglion seems im-
probable, though they occupy exactly the position in which
that ganglion would naturally be looked for. That the cells
represent one of the two sympathetic ganglia described by
Onodi (46) in the adult Mustelus, I greatly doubt; but I
find nothing that corresponds to those ganglia, and I cannot
even determine, from his descriptions, where they should be
looked for.

The branch that arises from the profundus at the place
where the small ganglion is found runs at first backward and
but slightly outward along the dorsal edge of the inferior
branch of the oculomotorius, here passing with the latter
nerve across the anterior edge, and then backward along the
dorso-lateral surface of the rectus inferior. While the two
nerves are in this latter position the branch of the profundus
leaves the oculomotorius, and running outward and backward
reaches an outer membranous envelope of the eyeball. In
or against this membrane it runs backward and separates
into two parts, both of which can be traced backward and
outward to the posterior surface of the eyeball, where, still
in or against the membrane, they lie against the antero-
lateral surface of the rectus externus, between that muscle
and the sclerotic. The dorsal and smaller branch of the
nerve here gradually disappears in the membrane, and quite
certainly never pierces the sclerotic. The other and larger
branch continues further outward around the eyeball, and
quite probably pierces the sclerotic, though this could not be
definitely established, the sections here being imperfect.
This nerve, in its origin and in its relation to the oculomo-
torius, corresponds exactly to the posterior ciliary nerve, or
ciliaris longus, of Schwalbe’s descriptions, and Schwalbe says
that his nerve pierced the sclerotic. ‘Tiesing says that the
oculomotorius, as it passes over the eye-stalk, gives off a
delicate ciliary nerve, which unites with a branch of the pro-
fundus and goes to the bulbus. The conditions that I find
differ somewhat from those described by either of these

158 EDWARD PHELPS ALLIS, JUN.

authors, in that one branch of my nerve quite certainly does
not pierce the sclerotic. This branch is thus quite probably
not a ciliary nerve, and would seem to correspond approxi-
mately to the lachrymal nerve of human anatomy.

As it gives off this first branch the ophthalmicus profundus
issues from the interval between the two bundles of the rectus
internus, and then continues its forward course, passing
downward and forward across the postero-lateral surface of
the internus muscle, and then ventral to that muscle. When
it reaches the ventral edge of the muscle it gives off a
delicate branch which separates into two parts, and possibly
into three, for a line of tissue here continues the course of the
branch, and I am unable to tell whether it is partly nervous
or wholly of fibrous or connective tissue. The two parts of
which Iam sure form the radix longa, and run downward
and backward to the ciliary ganglion. The other one, if it
be a nerve, must be a part of the ciliaris longus, for it runs
downward and forward to join the optic nerve. —

Anterior to the point where the radix longa is thus given
off, the ophthalmicus profundus continues its forward course,
lying lateral or ventro-lateral to the ventral edge of the
rectus internus. In this position it passes dorsal to the optic
nerve, and immediately beyond that nerve gives off a large
branch which runs forward and outward to the antero-ventral
aspect of the eyeball. There it pierces the outer membranous
envelope of the eyeball, aud continuing forward between that
membrane and the sclerotic diminishes gradually in size, and
finally disappears without ever piercing the sclerotic, so far
as I could determine. It lies, in this terminal part of its
course, slightly lateral to the ventral edge of the rectus
internus, this position thus seeming similar to that occupied
by the ciliaris brevis of Herrick’s descriptions of Menidia
(32, p. 224), just as, or before, that nerve pierces the sclerotic.
If the nerve in Mustelus does not pierce the sclerotic, as
my observations would seem to indicate, it cannot be the
homologue of Herrick’s nerve, for it \vould not be a ciliary
nerve. If it be not such a nerve it would seem to corre-

MUSTELUS LAEVIS. 159

spond somewhat to the infra-trochlear branch of the nasal
nerve of man.

Near the anterior end of the orbit the ophthalmicus
profundus passes between the obliquus superior and obliquus
inferior muscles, near their insertions, lying in its accustomed
relations, ventral to the one and dorsal to the other. Beyond
these muscles the nerve passes along the mesial edge of the
muscle Add}, and then, when it reaches the level of the hind
end of the nasal chamber, pierces the cartilaginous side wall
of the brain case, considerably dorsal to the nasal chamber,
and enters the cranial cavity. There it runs forward along
the inner surface of the side wall of the brain case, there
lying directly internal to the antorbital process. Somewhat
anterior to that process it again pierces the cartilaginous side
wall of the brain case, and, issuing a second time from the
cranial cavity, reaches the position slightly ventro-lateral to
the ophthalmicus superficialis already described in describing
the latter nerve. In this position it passes on to the dorsal
surface of the nasal capsule, and toward the anterior end of
that capsule, after giving off a large branch, comes into
intimate juxtaposition with the ventral surface of the
ophthalmicus superficialis. Leaving that nerve, it gives off
a small branch, and then separates into two large branches,
each of which again separates into two branches. All five of
the terminal branches thus formed of the nerve turn down-
ward over the anterior end of the nasal capsule, and then
backward along its ventral surface. Runuing backward,
ventral to the nasal capsule, the branches of the mesial one
of the two large branches of the nerve gradually disappear.
The largest branch of the other lateral and larger portion of
the nerve soon pierces the ventral wall of the nasal capsule,
and reaches the mner surface of the capsule. There it
continues backward toward the nasal aperture, breaking up
into several branches, some of which pierce the cartilage
again, passing outward through it to its external surface.

The large branch given off just before the main nerve
comes into intimate juxtaposition with the ophthalmicus

160 EDWARD PHELPS ALLIS, JUN.

superficialis, runs forward, laterally and downward, along the
lateral surface of the anterior portion of the nasal capsule,
and owing to breaks in the sections at this point could not
be definitely traced. It ran down toward and seemed to be
distributed to the region adjoining the antero-lateral end of
the nasal-flap cartilage, where the supra-orbital lateral canal
comes into close relation with that cartilage.

The only branches of the profundus nerve that were
sufficiently large to be distinguished and traced thus go
either to the eyeball or to the nasal capsule. ‘The nerve,
however, diminishes in size between these several branches,
and certain of its fibres may have a different and general
distribution. That the nerve has anything whatever to do
with the innervation of the ampullary tubes, as Ewart states
is probable in Lemargus (17, p. 527), I greatly doubt.
Certain of the terminal branches of the nerve would seem to
correspond to the branch said by Cole (11) to be dis-
tributed, in Chimera, to the “ outer surface of the inner
wall of the nostril.” This nerve Cole considered as the
probable homologue of the motor division of the profundus
of the Cyclostomes. This will be later further discussed.

Whether the distribution of the only apparent branches of
the profundus to the eyeball and the nasal capsule has any
morphological significance or not, or what that significance
may be, I cannot venture to assert. The nerve is evidently
the nasal or naso-ciliary nerve of the higher vertebrates, as
is generally asserted and accepted, and the dorsal and
proximal one of its two orbital branches would correspond
somewhat to the lachrymal nerve cf man, the ventral and
distal one corresponding somewhat to the infra-trochlear nerve.

Truncus Buccalis-maxillo-mandibularis.

The truncus maxillo-mandibularis or trigeminal part of the
large buccalis-maxillo-mandibularis nerve trunk arises from
the antero-ventral part of the extra-cranial portion of the
trigemino-facial ganglion, the ramus buccalis arising from
the antero-dorsal part of the same ganglion. That part of

MUSTELUS LEVIS. 161

the ganglionic complex from which the buccalis arises forms
a distinctly separate ganglion, which lies partly embedded in
the underlying antero-ventral part of the entire complex.
Both nerves, forming a single trunk, run at first laterally,
forward, and downward, the truncus maxillo-mandibularis
separating immediately into two strands, one of which is the
ramus maxillaris trigemini, and the other the ramus mandi-
bularis trigemini. The ramus mandibularis immediately
passes laterally nnd forward, ventral to the buccalis, and thus
soon acquires a position postero-lateral, and in part ventral
to the latter nerve. ‘The ramus maxillaris remains along the
antero-mesial edge of the buccalis, also lying in part ventral
to it, the buccalis forming, in sections, a large blunt wedge
between the two trigeminal nerves. The three nerves
together here form a single large and flattened trunk, which
lies on the dorsal surface of the so-called basal plate of the
skull, which plate, projecting laterally and somewhat down-
ward, forms the floor of the orbit. The nerve trunk lies
immediately internal to the large peri-orbital sinus, and in
its forward course it passes internal, or ventral, to all the
structures in the orbit, excepting only to two arteries, the
anterior and posterior carotids of Parker’s descriptions. It
passes dorsal to the anterior carotid, immediately postero-
lateral to the point where its ophthalmic branch is given off,
and then lies antero-lateral to the artery. The posterior
carotid pierces the floor of the orbit, and passes from its
ventral to its dorsal surface, ventral to the hind end of the
extra-cranial trigemino-facial ganglion. It then reaches the
lateral aspect of the ganglion, and as it traverses the orbit,
running forward and laterally, lies ventral to the lateral edge
of the truncus buccalis-maxillo-mandibularis, the truncus
thus lying dorsal to the artery.

In this orbital part of the course of the truncus the
buccalis is, as already stated, a somewhat triangular strand
wedged in between the mandibularis and maxillaris. Soon
after the truncus leaves its ganglion, certain general sensory
branches arise from the ramus maxillaris, and run outward

162 EDWARD PHELPS ALLIS, JUN.

and forward, ventral to the buccalis, and then upward along
the lateral surface of that nerve. They thus come to he
along the dorso-lateral edge of the buccalis, between it and
the mandibularis, and from that position they later run to
the regions they innervate. The ramus mandibularis thus
here begins to become detached from the remainder of the
truncus, and shortly before reaching the point where the
palato-quadrate cartilage articulates with the antorbital part
of the skull, it runs outward over the dorsal edge of the
palato-quadrate cartilage, and then downward and backward,
at first lying along the lateral, external surface of the
adductor mandibule muscle, but gradually becoming im-
bedded in the fibres of that muscle.

The remaining portion of the truncus now represents the
united maxillaris trigemini and buccalis facialis. It turns
more directly forward, and passes lateral to the articular
process of the palato-quadrate, there lying at first along
the dorsal edge of the dorso-anterior corner of the adductor
mandibule muscle, but soon turning downward and forward
to reach the lateral surface of the muscle Add of Vetter’s
descriptions (68). Here both of the nerves immediately
give off numerous branches, which, united more or less
completely, seem to form a large and important outer
division of the truncus. ‘This division, however, immediately
breaks up into numerous branches, some destined to innervate
the sense organs of the infra-orbital canal, others to innervate
the general tissues, and still other and more important ones
to innervate the sense organs of the buccal group of ampulle.

The remaining portion of the truncus has now reached
the lateral edge of the head, and here turns mesially, forward
and downward on to its ventral surface, still lying along the
external surface of the muscle Add, that surface here
being presented ventro-laterally. Continuing its course in
this position, it soon passes beyond the level of the anterior
end of the palato-quadrate cartilage, and reaches the level of
the hind end of the nasal capsule. It there still lies ventro-
lateral to, and along the external surface of, the muscle

MUSTELUS LAEVIS. 163

Addp, but that muscle here separates into two heads, one of
which has its origin from the mesial edge of the ventral
opening of the cartilaginous nasal capsule, near its hind
end, while the other arises, dorsal to the hind end of the
nasal capsule, from the dorsal and ventral surfaces of the
projecting ventro-lateral end of the antorbital process.

Anterior to this point the truncus follows the ventral head
of the muscle Addj, lying ventral to its antero-lateral edge,
and, with it, crossing the hind end of the ventral opening
of the cartilaginous nasal capsule, posterior to the external
nasal aperture. Anterior to the muscle Add, what is left
of each of the two main nerves turns directly forward, and
here les near the lateral edge of the horizontal plate-like
extension of the ventral edge of the internasal cartilage.
In this position the two nerves pass mesial to the external
nasal aperture, and, anterior to that aperture, lie at first
along the ventro-mesial aspect of the rounded anterior end
of the nasal capsule, in the angle between the capsule and
the ventro-median rostral bar of cartilage, and then, anterior
to the capsule, continue forward toward the end of the snout,
remaining always along the lateral aspect of the ventro-
median rostral bar.

The general course of this truncus buccalis-maxillo-man-
dibularis having now been described, the branches of the
three nerves that form it can be given.

Ramus Bueccalis Facialis.

The first five branches of the buccalis facialis arise from
the ganglion of the nerve, and not from the nerve itself.
The first three branches are given oft, close together, from
a dorso-posterior prolongation of the ganglion of the nerve,
and have already been fully described when describing the
lateral organs they innervate. They run upward and back-
ward in the orbit, along the side wall of the skull, and
separate into four branches, one of which pierces the over-
hanging orbital roof to innervate organs 103 to 110 infra-
orbital, while the other three turn outward and innervate

164 EDWARD PHELPS ALLIS, JUN.

organs 87 to 102 and the sensory organ of the spiracle, in
the manner already described. The branch that innervates
organs 103 to 110 is the ramus oticus facialis of current
descriptions. ‘The three branches that together innervate
organs 87 to 102 probably represent the inner buccal of
Cole’s descriptions of Chimera, though it may be that that
nerve of Chimera also includes the next two following
branches of the buccalis of Mustelus. ‘These next two
following branches of the buccalis are given off, close to-
gether, from the lateral aspect of the ganglion of the nerve,
not far from the three preceding ones, but from a somewhat
distinct and different part of the ganglion. Hach nerve is
joined at once by a general sensory branch, which arises
from the base of a large nerve that has its origin from the
postero-ventral end of the trigeminal part of the extra-
cranial part of the trigemino-facial ganglionic complex.
This large nerve is partly motor, and will accordingly be
described as a branch of the ramus mandibularis trigemini.
Its motor fibres are destined to innervate the levator maxilla
superioris and spiracle muscles. Hach of the lateral sensory
nerves here under consideration, with its accompanying
general sensory branch, runs at first backward and laterally,
ventral to the peri-orbital sinus, then turns outward through the
sinus, and then forward and outward, the lateral components
separating from the general sensory ones and going to inner-
vate, respectively, organs 83 to 86, and 79 to 82, infra-orbital.

The next two branches of the buccalis arise from the
buccalis itself, and not from the ganglion of the nerve.
They are given off not far apart, from the lateral aspect of
the nerve, just as the ramus mandibularis separates from
the still united buccalis facialis and maxillaris trigemini.
There are thus no branches given off by the buccalis between
its point of origin from its ganglion, and the point where
the ramus mandibularis definitely separates from it. The
two branches given off at the latter point run forward a
considerable distance, nearly parallel to, and not far from
the buccalis. The proximal one then first turns laterally,

MUSTELUS LAVIS. 165

then laterally and backward, and separates into two parts—
one of which continues backward and innervates organs
77 and 78 infra-orbital, the other turning forward and
innervating organs 75 and 76. The distal one of the two
branches continues still farther forward, and, branching,
innervates organs 70 to 74. Hach of these two buccalis
branches is closely accompanied, through a part of its
course, by a general sensory branch of the ramus maxillaris.

Anterior to these last two branches the buccalis turns
downward, with the maxillaris trigemini, on to the lateral
surface of the muscle Addjj, and there immediately gives
off a large branch which, breaking up into several smaller
branches, is distributed entirely to the sensory organs of
the large buccal group of ampullz. The several branches
of this large ampullary branch are grouped into two por-
tions, rather than forming parts of two main branches. ‘The
branches of the larger one of these two portions go to
ampullz that he dorsal to a membrane that extends inward
from the internal surface of that section of the infra-orbital
canal that lies between the suborbital bend in the canal and
the point where the hyomandibular canal is given off, the
branches of the other portion going to ampullz that lie
ventro-mesial to the same membrane. he membrane and
nerves thus here indicate two sub-groups of this large group
of ampulle, and they may, possibly, correspond to the inner
and outer buccal groups of Ewart’s descriptions. It is,
however, to be remembered that neither of the sub-groups
of Mustelus occupies the position of Hwart’s inner buccal
group, that position seeming to correspond exactly to that
occupied by my deep group of ophthalmic ampulle.

As this large ampullary nerve is given off by the buccalis,
the remaining fibres of the nerve, which are now entirely
lateral sensory ones, separate into two principal portions.
One of these portions separates almost immediately into
several branches, all of which run outward, forward, and
downward, and supply organs 46 to 69 infra-orbital, these
organs all lying in the three arms of the double suborbital

166 EDWARD PHELPS ALLIS, JUN.

bend of the canal; that is, organ 69 hes in the circum-
orbital part of the canal, shortly posterior to the point
where the canal turns downward and backward ventral to
the anterior edge of the eye, and organ 46 lies immediately
proximal to the point where the infra-orbital canal anasto-
moses with the distal end of the supra-orbital canal. This large
lateral sensory branch, the ampullary branch, and certain
general sensory branches, together form a large group of
nerves, which have the appearance of being a somewhat
separate branch or division of the truncus buccalis-maxillaris.

The second and remaining portion of the buccalis forms
an anterior or terminal division of that nerve. It runs
forward along the external surface of Add with the
remaining portion of the ramus maxillaris, the two nerves
lying close together, but as two wholly separate strands.
The buccalis here hes lateral to the maxillaris, and soon
separates into two parts, one destined to innervate organs
26 to 45 infra-orbital, which lie between the point where the
canal anastomoses with the distal end of the supra-orbital
canal, and the middle point of the short median section
of the canal; and the other one destined to innervate organs
1 to 25. The branch destined to supply these latter organs
forms the terminal section of the entire nerve, and in its
forward course it crosses ventral—that is, superficial to the
ramus maxillaris, and reaches its mesial side.

This anterior division of the buccalis thus innervates the
organs that lie in that part of the infra-orbital canal that
corresponds to the anterior, or antorbital, cross-commissure
of my descriptions of Amia. The nerve accordingly corre-
sponds to that somewhat separate and terminal part of
the buecalis that in Amia innervates organs | to 4 infra-
orbital (2, p. 514). In Chimera, Gadus, and Scomber the
corresponding fibres of the buccalis also form a separate
branch or division of the nerve, and I have already had occa-
sion to call especial attention to it, not only in the work on
Scomber that is still in press, but also in one of my published
works (4, p. 366). This branch or division of the buccalis,

MUSTELUS LAVIS. GZ

together with certain branches of the maxillaris trigemini, 1s
quite unquestionably the nerve to which Howes has recently
made reference (38, footnote, p. 51) as a nerve for which the
late Professor Huxley intended proposing the name palato-
nasal, or hyporhinal. Professor Howes has very kindly
sent me! a full copy of the unpublished manuscript in which
this nerve is discussed by Professor Huxley, and I consider
the subject so important that I shall quote quite fully what
Huxley says about it. I shall also later quote what Huxley
has to say, in the same manuscript, about the chorda tympani.

This unpublished memoir of Huxley’s was to have been
entitled “‘ Vertebrate Head; with Special Reference to Tri-
geminal and Facial Nerves and Trabecule.”

Having described the branches of the facial nerve in man
and in Amphibia, Professor Huxley then takes up the facial
nerve in the common Ray. The posterior division of this
nerve in this fish is first described, and I shall quote his
description in full in connection with my discussion of the
chorda tympani. He then says :—“ II. The anterior division
runs forward in the dorsal aspect of the oral mucous mem-
brane, on the inner side and in front of the spiracle. From
its outer side it gives off numerous branches, some of which
are distributed to the muscles of the spiracle. and some to the
spiracular branchia ; some run outwards and ventrally along
the anterior face of the spiracular cartilage and the liga-
mentous fibres which connect it with the palato-quadrate car-
tilage, and can be traced as far as the articulation of the
mandible, while yet others proceed to the palatine mucous
membrane extending to the posterior margin of the palato-
quadrate cartilage.

“Tt is clear that this nerve answers to some extent to the
palato-nasal nerve of the amphibia
posterior palatine nerve, but it gives off no fibres having
the distribution of those which pass between the nasal sacs
on the dorsal side of the vomers, and which I suppose to be
the homologue of the nerve of Cotunnius.

' «Life and Letters of T. H. Huxley,’ vol. 11, App. 1.
VOL, 45, PART 2.—NEW SERIES, N

answers, in fact, to the

168 EDWARD PHELPS ALLIS, JUN.

“But there is a very large nerve, a branch of the tri-
geminal, which in the Ray arises almost directly from the
short common trunk, but is buried up for a short distance
with the second division of the fifth, and then passes forwards
and inwards along the base of the skull, and on the inner side
of the nasal capsule, to the snout, where it divides into a leash
of branches for the sensory tubes of that region. From the
outer side ef this nerve, while it lies in the orbit, branches
are given off to the posterior face of the olfactory capsule.

“This nerve is identified by Stannius with what he terms
the ramus buccalis of the maxillary nerve in osseous fishes.
But it appears to me to have nothing to do with this nerve,
which is otherwise represented in the Rays.

“The branch of the trigeminal under discussion appears
to exist in all the Plagiostomes, and considering the many
embryonic characters retained by these fishes, I am led to
believe that it represents in a distinct form a nerve which in
most fishes and in the higher vertebrates has coalesced with
the anterior palatine division of the seventh, and has in that
manner given rise to the Videan nerve.” This branch of the
trigeminus in the Ray is later always referred to by Professor
Huxley as the palato-nasal.

Huxley then states his conclusions :—(1) That the spiracle
of the Ray ‘‘answers to the tympano-Kustachian passage of
the higher Vertebrata,” and (2) ‘‘that the palato-quadrate
cartilage in the Ray and other corresponding structures in
the embryo of the higher Vertebrates is a mere appendage
of the mandibular arch, and not a distinct visceral arch, as I
was formerly inclined to suppose.”

He then says :—“ But if the pterygoid cartilage is only an
outgrowth of the mandibular arch, it follows that the space
between the mandibular and the trabecula on each side repre-
sents a trabeculo-mandibular visceral cleft, and this ought to
be supplied by a nerve dividing into anterior and posterior
divisions as the portio dura does. Itappears to me that there
is no difficulty in finding the posterior division or nerve dis-
tributed to the (morphologically) anterior face of the man-

MUSTELUS L2XVIS. 169

dibular arch. The second and third divisions of the tri-
geminal, in fact, fulfil all the requirements of such a nerve,
the second being distributed to the outer face of the ptery-
goid appendage of the mandibular arch, just as the palatine
nerves are distributed to its inner face, and the mandibular,
or third division of the fifth, ranning along the outer face of
Meckel’s cartilage, in correspondence with the distribution of
the long descending branches of the palatine in the Ray on
its inner face. If this be the posterior nerve of the trabeculo-
mandibular visceral cleft, where is its anterior nerve? On
this point it may be well to speak cautiously, but certain facts
are highly suggestive. The palato-nasal branch of the fifth
in the Ray virtually runs along the ventral outer, a morpho-
logically posterior edge of the trabecula—which, as I think,
cannot be doubted is the homologue of a visceral arch; on
the other hand, the orbito-nasal nerve (first or ophthalmic
division of the fifth) runs along the dorsal, a morphologically
anterior edge of the trabecula.

“Tt seems to me highly probable, therefore, that the
palato-nasal nerve of the fifth (or nerve of Cotunnius) is the
anterior division of the nerve of the trabeculo-mandibular
cleft, which should run along the morphologically posterior
face of the trabecula, and that the orbito-nasal nerve is the
nerve of the anterior face of the trabecula, which found an
actual aperture only in the marsipobranchii (and pharyngo-
branchii ?), where it limits the naso-palatine canal laterally.”

It is impossible to recognise in Mustelus exactly what
fibres are included in this palato-nasal nerve of the Ray, and
it seems to me that Huxley himself did not have a correct
or definite idea of exactly how the nerve is composed. That
the nerve contains the terminal portion of the buccalis, as
that nerve was described by Stannius, and as it has since
been described by other authors, is evident, and the lateral
sensory component of the nerve is certainly not coalesced with
the anterior palatine division of the seventh in osseous fishes,
as Huxley supposed. That this terminal part of the buccalis
belongs to a premandibular arch is, however, extremely pro-

170 EDWARD PHELPS ALLIS, JUN.

bable, and is later more fully discussed; andit is to Huxley’s
very definite opinion on this subject that I have wished to
‘all especial attention. The lateralis component of this nerve
of fishes would not, according to recent ideas, be retained in
the higher Vertebrates ; but the ampullary component quite
probably would be retained, and, with such general cutaneous
fibres as may belong to the nerve in fishes, it would form the
nerve in the higher Vertebrates.

Ramus Maxillaris Trigemini.

The first two general sensory branches of the nervus tri-
geminus arise from the very base of the large ramus ad
muse. lev. max. sup., and would accordingly seem to belong
to the ramus mandibularis trigemini rather than to the
ramus maxillaris. They, however, both accompany branches
of the ramus buccalis, and are hence here considered as
branches of the maxillaris trigemini. They have both already
been referred to in describing the buccalis. One of them
accompanies the buccalis branch to organs 83 to 86, and the
other the buccalis branch to organs 79 to 82. Both branches
run at first backward and outward, near the floor of the orbit,
and then forward and outward to the regions they innervate.

Anterior to these two branches a branch is sent out with
the buccalis nerve to organs 75 to 78, and another with the
buccalis nerve to organs 70 to 74. ‘These two branches both
arise from the latero-ventral part of the maxillaris, and both
run at first forward, downward, and laterally, closely applied
to the ventral surface of the buecalis, and then forward and
laterally, passing outward along the lateral surface of the
buccalis, between it and the ramus mandibularis. Having
thus reached the dorsal surface of the large truncus, between
its buccalis and mandibularis components, they there have
their apparent origins from the truncus, and join the buccalis
nerves they accompany.

Beyond these four branches the maxillaris turns downward
on to the lateral surface of the muscle Add, and as it

MUSTELUS La&vis. 171

acquires this position a large branch becomes detached from
the dorsal surface of the nerve, and runs forward and
laterally directly internal to that portion of the buccalis that
goes to the infra-orbital organs 46 to 69 and to the buccal
group of ampulle. Two small branches arise from this
branch before it becomes entirely detached from the main
trunk, and running forward, dorsal to the buccalis and
ventral to the peri-orbital sinus, reach the ventral surface of
the antorbital process of the skull, being closely accom-
panied in their course by a branch of the posterior carotid
artery. ‘The main branch, after it becomes detached from
the maxillaris, sends several branches to the tissues antero-
ventral to the eye, certain of them accompanying certain of
the buccal branches that go to infra-orbital organs 46 to 69.
The remaining portion of the main branch then separates
into two parts, both of which run forward along the lateral
surface of Addf, and beyond the anterior edge of the dorsal
head of that muscle lie, one along the ventro-lateral corner
of the nasal capsule, and the other in the angle between that
capsule and the side wall of the brain case. The ventral
branch here hes directly internal to and accompanies the
subrostral section of the supra-orbital canal, a maxillaris
branch of the trigeminus thus here being associated with the
ophthalmic lateral canal.

The large branch just above described is given off from the
dorsal surface of the maxillaris. At about the same time two
branches become detached, one from the ventral surface of
each of the two branches into which the buccalis separates
after having given off the branches to the buccal group of
ampulle. These two branches, although having their appa-
rent origins from the two branches of the buccalis, are
maxillaris and not buccalis nerves, but they become detached
from the maxillaris and intimately associated with the buccalis
a very considerable distance posterior to this point, while
the main nerves are still on the floor of the orbit. Like the
other maxillary branches that leave the nerve in that part of
its course, they pass along the ventral aspect of the buccalis

172 EDWARD PHELPS ALLIS, JUN.

in their forward and outward course from their real to their
apparent points of origin.

The lateral one of these two maxillaris branches has the
most posterior origin from the mavxillaris. After it has
become detached from that portion of the buccalis that it
accompanies, it runs laterally and downward until it reaches
the inner surface of a membrane that here bounds the inner
surface of the region occupied by the buccal ampullee and
ampullary canals, where it turns backward, downward, and
mesially along the inner surface of that membrane. In this
position, having first sent a small branch forward, it con-
tinues backward, downward, and mesially until it reaches
the anterior end of the integumental furrow that separates
the maxillary labial fold from the under surface of the head.
It then continues backward along that edge, lying ventro-
mesial to it, and gradually disappears.

The other one of the two branches also soon turns down-
ward and backward, sending two small branches forward at
the bend. It remains at first close to the latero-ventral
surface of Add, but later passes into the maxillary labial
fold, which it innervates. Whether or not it continues
backward until it reaches and enters the mandibular labial
fold also could not be determined.

One or both of these two branches evidently correspond
to the so-called maxillary branch of the superior maxillary
nerve of Amia and teleosts.

Anterior to these two branches the main maxillaris nerve
acquires a position directly internal to the buccalis, and there
gives off a large branch from its ventral edge, and another
from its dorsal edge. The ventral branch is joined by the an-
terior branches of the branch just above described, with which
it fuses. ‘The nerve thus formed runs downward, mesially
and but slightly forward, passes beyond the ventro-mesial
edge of the muscle Addj3, and reaches the ventral surface
of the extreme anterior end of the palato-quadrate cartilage.
There it crosses the middle line of the head as a stout nerve,
aud hence doubtless here anastomoses with its fellow of the

MUSTELUS LAEVIS. 173

opposite side, though this could not be definitely established,
my sections being of one half of the head only.

The dorsal branch given off by the maxillaris at the same
time as the above-described ventral one passes outward and for-
ward dorsal, and hence internal, to the buccalis, and separates
into two parts which accompany the two branches of the
buccalis sent, the one to organs 44 and 45. infra-orbital,
and the other to organs 42 and 45, these four organs lying
in the bend of the infra-orbital canal, just distal to the
anastomosis with the distal end of the supra-orbital canal.
These two branches of the maxillaris continue forward and
mesially beyond the distal ends of the two buccal branches
they at first accompany, and reach the ventral surface of the
nasal capsule, in the region adjoining that part of the infra-
orbital canal that les posterior to the nasal aperture. A
branch is here sent, from the mesial one of the two nerves,
to accompany the branch of the buccalis that goes to organs
40 and 41 infra-orbital, the remaining portions of both nerves
continuing forward, internal to the infra-orbital canal, toward
the hind end of the nasal aperture. The branch that at first
accompanied the buccal nerve to organs 42 and 43 infra-
orbital, here turns dorsally through one of the slit-like
foramina or imperfections in the ventral floor of the nasal
capsule, and, entering that capsule, sends one branch forward
and mesially, and another forward and laterally, along the
ventral surface of the nasal sac. How far forward these
branches extended could not be determined.

The main ramus maxillaris, after giving off the branches
above described, follows the course of the infra-orbital canal,
and passes, with it, mesial to the nasal aperture, here lying
always dorso-mesial to the canal and mesial to the buccalis.
Several branches here arise from the maxillaris, some running
forward and mesially, and others forward and laterally, the
latter passing internal, that is, dorsal, to the buccalis. One
of these latter branches turns upward into the nasal capsule,
and could be traced forward a certain distance, ventral to
the mesial portion of the nasal sac. It was then lost, but

174 EDWARD PHELPS ALLIS, JUN.

an important branch which would seem to very probably
represent its most anterior section was found turning back-
ward around the anterior end of the sac.

The remainder of the maxillaris then continues forward,
lying at first ventro-lateral to the median section of the
infra-orbital canal, and then directly internal, and hence
dorsal, to the more anterior part of the same canal. It is
there distributed to the general tissues of the region.

Ramus Mandibularis Trigemini.

The ramus mandibularis or maxillaris inferior trigemini
arises from the trigeminal part of the trigemino-facial
ganglion as part of a short truncus trigemini which im-
mediately separates into maxillaris and mandibularis portions.
The ramus mandibularis immediately runs laterally downward
and forward, ventral to the ramus buccalis, and acquires a
position postero-lateral to the latter nerve, there forming the
postero-lateral portion of the large, flat, truncus buccalis-
maxillo-mandibularis.

At or before the ramus mandibularis thus separates from
the ramus maxillaris, and while the truncus trigemini is still
ganglionic, a large branch is given off from its posterior
aspect. ‘Turning laterally upward and backward, it passes
dorsal to the posterior carotid artery, immediately anterior
(distal) to the orbital branch of the artery, and then passes
upward along the lateral aspect of the facial part of the
trigemino-facial ganglion, here being partly imbedded in the
eanelion. It passes dorsal to the ramus palatinus facialis,
accompanied by the orbital branch of the posterior carotid.
Dorso-posterior to the trigemino-facial ganglion it lies ventro-
lateral, and then lateral or antero-lateral to the truncus
hyoideo-mandibularis facialis, accompanying the truncus in
its course, and lying, with it, at first along the mesial surface
of the hind end of the peri-orbital sinus, and then ventral
to that posterior prolongation of the sinus that goes back-
ward to join the jugular vein. This posterior prolongation

MUS'TELUS LEVIS. P75

of the peri-orbital sinus is called by Ridewood (54), in
Scyllium, the post-orbital blood-sinus. Here the trigeminal
nerve leaves the truncus facialis and turns laterally and
backward across, and dorsal to, the anterior portion of the
anterior diverticulum of Wright’s (70) descriptions of the
spiracular cleft of the fish. It has here already separated
into three parts—one larger and two smaller ones,—all of
which pass outward anterior to the superior postspiracular
ligament of Ridewood’s descriptions of Seyllium, and anterior
also to the posterior or auditory diverticulum of the spiracular
cleft. One of the smaller parts of the nerve was here lost
in the sections. ‘The other two parts both reach the internal
surface of the levator maxille superioris muscle of Vetter’s
descriptions, which they seem to penetrate and innervate.
A part of the fibres of the nerve are, however, quite un-
questionably distributed to the general tissues of the region,
for that the nerve contains general sensory fibres is clearly
shown by the two branches that arise from near its base, and
that accompany the two buccalis branches.

The levator maxille superioris muscle of my embryo is
apparently exactly such a muscle as Tiesing describes in the
adult. The muscles of my embryo that represent the three
muscles of the adult considered by Tiesing as parts of Csd,,
differ somewhat from those muscles as described by him,
but I should state that I am unable to clearly comprehend
his descriptions and figure of them. The muscle called by
Tiesing the levator palpebre nictitantis I find exactly as
he describes it, excepting that, as this muscle crosses the
two other muscles, dorsal to ithe spiracle, its postero-ventral
edge is imbedded in those muscles to such an extent that the
ventral half of its outer and inner surfaces is covered by
them. In transverse sections the levator muscle is here of
oval section, inclining towards the skull, the other two
muscles together forming, in sections, a deep gutter which
encloses the lower half of the levator. The other two
muscles together form a T-shaped muscle, more or less
fused with the levator maxille superioris. ‘lhe 'T’ is placed

176 EDWARD PHELPS ALLIS, JUN.

upside down, its leg representing a part of the muscle that
first runs inward, and then upward and backward along the
outer surface of the hind edge of the levator maxille
superioris, to have its origin with that muscle from the
skull, as Tiesing describes for his muscle Csdyy. The pos-
terior arm of the inverted ‘l' runs backward above the
spiracle, and has its origin on the fascia there as Tiesing
describes. The anterior arm of the inverted T' is the re-
tractor palpebre superioris of Tiesing. Nerve fibres could
not be definitely traced to these muscles, but they would
seem to be most unquestionably innervated by branches of
the same nerve that innervates the levator maxillee superioris,
as Tiesing states. The posterior head of the deeper muscle
might, however, easily receive fibres from that branch of the
facialis that innervates Csd,.

Beyond this Jarge and largely motor branch no branch is
given off by the ramus mandibularis until after it separates,
as already described, from the truncus buccalis-maxillaris
just before that truncus passes outward from the floor of
the orbit on to the lateral surface of the. muscle Add.
Running laterally across the dorsal edge of the palato-
quadrate cartilage, the ramus mandibularis reaches the
dorsal edge of the antero-dorsal corner of the adductor
mandibule. There the larger part of the nerve turns down-
ward and backward along the outer surface of the adductor,
lying in a groove between that muscle and the muscle
Addj3.

As the ramus mandibularis here turns downward and back-
ward it gives off a large branch, which, separating into two
parts, sends one branch forward and downward into the
anterior portion of the muscle Add, and another back-
ward and downward into the posterior portion of the same
muscle. This is, accordingly, the motor nerve of the muscle
Add, and it is certainly in my embryo a branch of the ramus
mandibularis, and not a branch of the ramus maxillaris. Both
Stannius (59) and Tiesing (62) find this muscle innervated by
a branch of the ramus mandibularis in the selachians examined

MUSTELUS LAVIS. divas

by them, while Vetter (63) and Jackson and Clark (quoted
from Tiesing, p. 96) both find it innervated by a branch of
the ramus manillaris. In Galeus I (8, p. 573) found the
muscle innervated by the terminal portions of a small double
nerve that arose from the base of the truncus trigemini, if not
even from the ganglion of that nerve. Proximal branches of
this double nerve innervated the levator maxille superioris
and the larger spiracle muscle. While I cannot affirm the
accuracy of these statements regarding Galeus, taken from
notes and not confirmed, they led me, at the time that I was
considering this subject, to conclude that the muscles Csd,,
Lins, and Add might, in part, represent parts of the con-
strictor superficialis dorsalis of one or more pre-oral arches
(3, p. 974). This conclusion, in so far as it relates to the
muscle Add, is greatly strengthened by what I now further
find in Mustelus.

The muscle Add, the levator labii superioris of Tiesing,
arises in all my embryos by two perfectly distinct heads,
instead of by a single large one, as Tiesing describes it in the
adult. The larger one of the two heads arises mainly from
the ventral surface of the antorbital process, and from the
outer surface of the bulging hind end of the nasal capsule
ventral to that process, but a part of the fibres of the muscle
pass upward around the hind edge of the ventral projection
of the antorbital process, and arise from its dorsal surface.
The other head of the muscle arises from the mesial edge of
the ventral aperture of the cartilaginous nasal capsule, slightly
anterior to its hind end. The fibres of the two parts fuse
completely, and when the muscle reaches the level of the
antero-dorsal corner of the adductor mandibulw, it les ventral
to the latter muscle. Posterior to this pomt a tendon begins
to form on the rounded dorsal surface of the mesial portion of
the muscle (fig. 8), and then, gradually, as this part of the
muscle diminishes in size, acquires a position at its ventro-
mesial edge (fig. 9). There it separates into three parts.
One of these three parts retains its position along the ventro-
mesial edge of the united Addj} and adductor mandibule

178 EDWARD PHELPS ALLIS, JUN.

muscles, and continues backward to jom, or form part of,
the aponeurosis that separates the superior and mandibular
parts of the adductor. A second part has its msertion on the
anterior end of the posterior upper labial cartilage. The
third part continues backward parallel to, and shehtly dorso-
mesial to, the latter cartilage, and has its imsertion on the
dorso-posterior end of the mandibular labial cartilage. This
insertion of these tendons of Addj3 on these two labial car-
tilages certainly indicates that a part, at least, of the muscle
belongs to a pre-oral arch, if the labial cartilages do, as I
beheve is very generally accepted. The nerve that imnervates
these parts of Addj3 must then belong to the same arch, and
must be a motor component of the post-trematic nerve of the
arch. Hoffmann, it should be noted, says (84, p. 266) that the
inuscle Addfi arises from the upper part of the wall of the
“ Kieferbogenhéhle,” an origin which would seem to preclude
its belonging to a pre-oral arch.

After giving off this motor branch to the muscle Add
the ramus mandibularis sends a relatively small but still im-
portant branch downward and backward along the external
surface of the adductor. When it reaches, approximately,
the level of the anterior end of the posterior upper labial car-
tilage it turns forward a short distance, then mesially, and
then downward, laterally, and backward into the maxillary
labial fold, there lying dorsal to the labial cartilage. If the
cartilage belongs to a pre-oral arch, this branch of the man-
dibularis thus has the relations to the mouth, considered as a
cleft, of a ramus pre-trematicus, the ramus mandibularis itself
being the related ramus post-trematicus. Considering it as
such a branch it would be the ramus branchialis posterior of the
posterior labial arch, the motor nerve to the muscle Add/3 being
the motor component of the ramus branchialis anterior of the
same arch. Those two branches of the ramus mavxillaris
superior trigemini that are also distributed to the labial fold
of the fish would then either be the sensory components of
the ramus branchiahs anterior of the same arch, or perhaps
be, respectively, the pre- and post-trematic branches of the

MUSTELUS LAVIS. 179

next preceding or anterior labial arch. It should here be
stated that there are in my embryos, contrary to Gegenbaur’s
statement (23, p. 214) for the adult, two separate and distinct
upper labial cartilages, both lying in, or in relation to, the
labial fold of the fish. The anterior cartilage has no direct
relations to any of the muscles of the fish.*

There are thus two pre-oral arches indicated in embryos of
Mustelus. In each arch there are what are considered by
Gegenbaur as remnants of the cartilages of the arch, and I
now further find not only muscles definitely related to one of
the arches, but also nerves that certainly might be considered
as the pre- and post-trematic branches of each arch. The
nervus trigeminus would then be a nerve formed by the fusion
of at least three segmental nerves, and the ramus maxillaris
superior trigemini would probably contain the pharyngeal
elements of all these nerves, in addition to containing, in its
proximal portion, certain of their pre- and post-trematic
elements.

The ramus mandibularis of Mustelus, after giving off the
branch to the labial fold above described, passes downward
and backward on to the lateral surface of the adductor man-
dibule, lying at first between that muscle and the posterior
portion of the muscle Addj3, and then becoming gradually
imbedded in, and entirely enveloped in, the single muscle
formed by the fusion of these two muscles. Continuing back-
ward and mesially in this position it separates into two nearly
equal portions. One of these two portions is probably wholly
motor, and is the ramus ad musculum adductoris mandibule.
It remains always in the interior of the muscle-mass, and
running backward and downward sends branches both to the
superior and inferior divisions of the muscle. ‘Two branches
were sent through the muscle to its dorso-mesial corner, and

* Gegenbaur in his references to the labial cartilages of Mustelus, and also
elsewhere, refers in the text to his fig. 2, Plate XLII, as representing
Mustelus, but this figure, in the description of the plates, is said to be of
Galeus—an evident typographical error, as Marshall and Spencer have already
pointed out,

180 EDWARD PHELPS ALLIS, JUN.

were there lost. Whether or not they issued from the muscle
and supplied other tissues could not be definitely deter-
mined,

The other portion of the ramus mandibularis runs down-
ward, backward, and mesially, until it passes the transverse
level of the hind corner of the gape of the mouth, where it
issues on the ventral surface of the adductor muscle, and,
turning forward and mesially, immediately passes from the
outer surface of the superior division of the muscle on to the
outer surface of the inferior, or mandibular division. This
nerve, in this part of its course, 1s certainly shown in the
nerve marked V; in Tiesing’s figure 9,in which figure the
nerve would seem to be cut. But how this ventral part of
this nerve unites with the dorsal part it is impossible to tell
from the figure; and the mandibularis externus facialis is
apparently not shown at all, certainly not in its proximal
portion. I am accordingly unable to here make any com-
parison with his work. i

When the ramus mandibularis reaches the outer surface of
the inferior division of the adductor, it sends one or more
small branches along the external surface of the united
adductor muscles, and a large branch downward and mesially,
internal to the mandibularis externus facialis, along the
external surface of the inferior division of the adductor. This
last branch sends two branches forward along the external
surface of the adductor, and itself, close to the mesial edge of
that muscle, pierces a well-differentiated intermandibularis
muscle, which Tiesing neither describes nor shows in his
fioures. As it traverses this intermandibularis muscle, the
nerve turns dorsally and backward around the ventro-mesial
edge of the mandibular cartilage, and reaches the dorsal, or
internal, surface of the intermandibularis muscle, near its
lateral edge. ‘There it continues backward, sends certain
branches into the muscle, and, for a certain distance,
diminishes in size. Then it begins to increase in size again,
and becomes the terminal part of the ramus hyoideus facialis.
These two nerves thus run directly into each other, and anas-

MUSTELUS LAVIS. 181

tomose completely, exactly as the branch rv. g@hi of the manil-
laris inferior trigemini and the ramus hyoideus facialis of
Amia do (8). The relations of the nerves to the muscles are
not exactly the same in Mustelus and Amia, but the differ-
ences are probably not of morphological importance.

The mesial one of the two branches sent forward along the
external surface of the adductor by this ramus ad muse. inter-
mand. passed beyond the adductor on to the ventral surface
of the intermandibularis, and there, at a certain pomt in my
sections, was lost while it still had a relatively large section.
Whether it penetrated the intermandibularis muscle, as the
main branch did, or not, could not be determined. The other
and lateral one of the two branches continues forward beyond
the anterior edge of the adductor muscle, approaching and
accompanying the main mandibularis nerve in the terminal
part of its course.

The main ramus mandibularis, after giving off the large
branch above described, continues its forward and mesial
course, lying immediately dorso-lateral to the mandibularis
externus facialis, both nerves soon passing beyond the anterior
edge of the adductor inuscle and reaching the ventral surface
of the mandibular cartilage, near its lateral edge. The ramus
mandibularis here hes directly lateral to the mandibularis
externus facialis. Somewhat mesial to the latter nerve, and
following an approximately parallel course, is the anterior
branch of the ramus ad = musc. intermand. just above
described.

When the ramus mandibularis reaches the transverse level
of the hind end of the mandibular labial cartilage, a branch
is sent outward from it into the mandibular labial fold. The
main nerve then separates into three or four branches, all of
which continue forward, and here, at first, lie along the ventral
surface of the mandibular labial cartilage. At the level of
the anterior end of that cartilage the nerve passes internal to
the mandibular group of ampulle, those ampulle here taking
up the larger part of the mandibularis externus facialis, a
small branch only continuing forward to supply the anterior

182 KDWARD PHELPS ALLIS, JUN.

organs of the mandibular canal. ‘The mandibularis trigemini,
beyond this point, continues forward, im two principal
branches and other smaller ones, toward the tip of the jaw,
lying always superficial to the mandibular cartilage.

Nervus facialis.

The truncus hyoideo-mandibularis facialis and the ramus
palatinus facialis arise separately from the facial part of the
extra-cranial trigemino-facial ganglion, the one from its dorso-
posterior portion, and the other from its ventral portion.

Ramus palatinus facialis.—The ramus palatinus runs
at first downward and almost directly laterally, alone the
shelving but nearly horizontally cartilaginous floor of the
orbit. In this part of its course it passes dorso-lateral to the
hind end of the orbital part of the posterior carotid artery,
posterior (proximal) to its orbital branch, and then mesio-
ventral to the anterior carotid artery.

Before reaching the lateral edge of the floor of the orbit
it sends three branches backward, two of them arising from
it close to its base. The proximal branch separates into two
parts. One of these parts runs backward in the angle
between the side wall of the skull and the floor of the orbit,
and, leaving the orbit at its hind end, turns outward posterior
to the anterior diverticulum of Wright’s descriptions of the
spiracular cleft. There branches are sent to the mucous
lining of the cleft, the main branch itself joing and becom-
ing confounded with the superior postspiracular heament as
that ligament passes downward behind the spiracular cleft.
There is here, perhaps, an anastomosis with a small branch of
the glossopharyngeus. This branch of the palatinus facialis
thus seeks, and acquires in its terminal portion, a post-
trematic distribution, and is, accordingly, a recurrent branch
of the ramus anterior of the nervus facialis, similar to certain
branches of the rami anteriores of the vagus nerves shown in
my drawing 59 of Amia (8).

The other part of the proximal branch of the palatinus

MUSTELUS L2ZVIS. 183

runs laterally and backward, and is accompanied by the
other two of the three orbital branches of the palatinus just
above referred to, these latter two branches soon uniting to
form a single nerve. Branches, which form anastomoses
among themselves, are then given off by the two nerves, so
formed, certain of them accompanying the pseudobranchial
artery to the spiracular pseudobranch, while others turn
downward and forward to the roof of the pharynx.

These first three branches of the ramus palatinus thus, to-
gether, probably represent the nervus prespiracularis of Tie-
sing’s descriptions of the adult Mustelus. In Wright’s (70) de-
scriptions of embryos smaller than my own they are probably
what he refers to in the words, “ Behind the orbit a few
twigs, which run backward to the filaments of the mandibular
pseudobranch.” Wright further says that the main truncus
facialis, which is said to arise from the facial ganglion
separately and independently of these twigs, “forks between
the two diverticula of the hyomandibular cleft into the pre-
and post-trematic branches.” The few twigs that run back-
ward to the pseudobranch do not, accordingly, form, in his
interpretation of the nerves, the prespiracular division of the
facial nerve, this conclusion being based, I believe, on an error
of observation, his preetrematic nerve quite probably being, as
is shown below, either a branch of the oticus facialis, or a
hgament mistaken for a nerve.

After giving off these first three branches, the ramus
palatinus reaches the lateral edge of the cartilaginous floor of
the orbit, and there turns forward, lateral to, and close to
that edge, giving off at the same time three branches. One
of these three branches runs forward, parallel to, and close
to, but shehtly ventral to, the main nerve itself, the other
two branches running backward and downward. ‘The
anteriorly directed branch hes near the dorsal corner of the
lateral surface of the pharynx, and gives branches to the
tissues there. The main nerve hes slightly dorso-mesial to
this branch, and sends branches to the tissues of the roof of
the pharynx. Both nerves can be traced forward to the

VoL. 45, PART 2.—NEW SERIES, 0)

184 EDWARD PHELPS ALLIS, JUN.

transverse level of the anterior end of the palato-quadrate
cartilage, where they become small and are lost. The two
backwardly directed branches run downward to the dorso-
lateral corner of the pharynx. One of them there continues
backward, giving branches to the adjacent tissues. The
other, larger branch runs downward until it reaches the
superior surface of the mandibular cartilage, that surface
being here directed dorsally and mesially. There it turns
forward and continues in that direction, lying immediately
ventro-mesial to the lower edge of the band of mandibular
teeth. In that position it gradually diminishes, and finally
disappears, being traceable with certainty, only to, approxi-
mately, the transverse level of the angle of the mouth. It
lies, in this distal part of its course, in the deeper part of a
fold of the tissues of the mouth cavity, and its relations to
this fold, as well as its general course and position, show,
beyond question, that it is the homologue of the nerve identi-
fied by Green (27) as the chorda tympani in the several
selachians examined by him. This identification of the chorda,
by Green, is said by him to be based on the homologies pro-
posed for that nerve by Herrick, in his Menidia paper, and is
further discussed below.

In Raia and Spinax, Stannius (59) says that the nervus
palatinus is represented by three branches, a delicate posterior
one, and two stouter anterior ones. One of the anterior
branches is said by Stannius to be the true ramus palatinus.
The other is considered by Herrick (82) and Green (27) as the
homologue of the chorda tympani. The posterior branch is
said by Stannius to arise from the ramus palatinus by two
roots, these roots later uniting to form a single nerve, which
later receives an anastomosing branch from the nerve above
identified as the chorda tympani. The three branches that I
find, in Mustelus, arismg from the orbital part of the nervus
palatinus thus together represent the posterior branch of
Stannius’ descriptions plus the communicating branch from the
so-called chorda tympani, this latter nerve in Mustelus being
wholly separate and independent of the other three. The

MUSTELUS LAVIS. ; 185

three branches of Mustelus accordingly represent the ramus
pretrematicus facialis sensu stricto of Herrick and Green.

Haller says (28, p. 414-16) that, in Scyllium, the nervus
facialis is represented by two principal branches and a third
smaller one. The two principal branches are said to be a
ramus facialis and a ramus hyoideus, the third one being said
to be, perhaps, a ramus palatinus. All three of the nerves
are said to arise independently from the so-called facial
ganglion, and a separate, independent, and more distal
ganglion is said to be found on the ramus facialis. The
ramus facialis becomes juxtaposited to the truncus maxillaris
trigemini, and this juxtaposition is said to represent an early
stage in that ultimate fusion of the two nerves that is said by
this author to be found in ganoids, and still more complete
in teleosts. Reference is here made to Goronowitsch’s (25)
work on Lota, and it is evident, from this reference, that
Haller considers his ramus facialis, either in whole or in part,
as the homologue of Goronowitsch’s posterior division of the
palatinus facialis. Haller does not say whether his so-called
ramus facialis is, in Scyllium, a prespiracular or a_post-
spiracular nerve, but his reference to Gegenbaur’s work on
Hexanchus (to which I am unable to sefer), and Gorono-
witsch’s reference to the same work (24, p. 485), leaves little
doubt that it is prespiracular. This pretrematic part of the
facialis nerve of Scyllium must then have a quite different
course, and quite different relations to the trigeminus, to those
found in Mustelus. It would seem as if it must contain the
nerve identified by Green as the chorda tympani in the
selachians examined by him, that nerve, and its homologue
in Mustelus, accordingly being quite certainly represented,
in Amia, in the maxillaris internus trigemini of my descrip-
tions (8).

In Chimera Cole (11) comes to the conclusion that the so-
called ramus pretrematicus facialis of his descriptions is the
chorda tympani of the fish, and Herrick (82, p. 165) accepts this
conclusion as probably correct. Herrick also accepts (p. 167) as
probably correct my conclusion (8, p. 749) that the chorda is

186 EDWARD PHELPS ALLIS, JUN.

probably represented in Amia by certain branches that I
described as the ramus mandibularis internus trigemini. But
these nerves of Amia are probably not represented, in
Chimera, by the so-called ramus pretrematicus of Cole’s
descriptions. It seems to me much more probable that they
are represented in one or both of two other nerves described
by Cole. One is the so-called pharyngeal, or visceral branch
of the maxillary branch of the trigeminus; the other a nerve
said to be formed by the fusion of three branches of the
mandibular branch of the trigeminus. The maxillary branch
is said by Cole to curve “round under a large muscle attached
to the angle of the upper jaw.” The three mandibular
branches are said to “ curve round and under a corresponding
muscle to that which the similar branch of the maxillary
curved round.” What the muscles to which these nerves are
thus related may be is not stated, but the general course of
the nerves, so far as given, resembles strongly that of the
mandibularis internus trigemini of Amia. If these two nerves
in Chimera are, one or both, the homologue of the nerve in
Amia, they must also be the chorda tympani of the fish if the
nerve in Amia is. Cole’s so-called prebranchial nerve could
not then be the chorda, and it might even be found to be a
postbranchial nerve, similar to the mandibularis internus
facialis of Amia, with which Cole himself homologises it,
maintaining, in order to establish his homologies, that the
nerve of Amia is a prespiracular nerve (12, p. 205). That
this nerve of Amia is not a prespiracular one I have already
had occasion to assert (4).

In seeking the homologue of the chorda tympani in Amia
I concluded (8, p. 638), as certain other authors had before
me, that it must be a prespiracular nerve. This opinion was
said to be based on the course of the “ nerve in man through
the upper portion of the tympanic cavity, and then downward
anterior to that cavity.’ Cole, however, says (11, p. 660)
that “the chords tympani of mammals passes morphologically
under the tympanum.” Cole then contends that the nerve,
though thus topographically a post-branchial one, is morpho-

MUSTELUS LAVIS. 187
logically a prebranchial one. Herrick (82, p. 160), following
me, concludes that the chorda is a prespiracular nerve because
of its course “through the tympanic cavity and above and in
front of the Eustachian tube.” Dixon, however, one of the
latest workers on the subject, certainly shows the nerve in a
post-Hustachian position in a five weeks’ old human embryo
(15, Fig. 15). The relation of the nerve to the Eustachian
tube in this figure cannot be mistaken, for the tube lies
between the facialis and the trigeminus nerves, and the chorda
arises from the facialis considerably beyond the tube and has
not a recurrent course. In an embryo seven weeks old
Dixon shows the nerve in what would seem to be a pre-
Eustachian position (Fig. 9). Whether this difference in the
two figures is due to some error on Dixon’s part, to some error
in my interpretation of his figures, or to an actual change of
position between these ages, I am unable to judge. That
Dixon himself considers the nerve as a prespiracular one, in
the adult man, is definitely stated in one of his later works
(16, p. 479) ; and that the chorda changes its relations to the
cleft seems indicated in Broman’s statement (9, p. 568) that
im his human embryo No. III, ‘ Die hintere Spitze der ersten
inneren Visceralfurche, die sich im vorigen Stadium gleich
hinter der Chorda tympani, lateral yon dieser bis an die
Koérperwand hinausstreckte, befindet sich jetzt eben an der
medialen Seite der Chorda.” This apparent change in the
relations of the two structures might be caused simply by a
change in the direction of the first visceral cleft, though this
seems improbable. In Broman’s Fig. 8, Pl. A, I should
certainly say that the chorda had a recurrent prebranchial
course. In his Fig. 2, Pl. C, the first visceral cleft seems to
he between the recurrent chorda and the trunk of the facialis,
the chorda thus here being postbranchial in position. These
apparent changes in the relation of the nerve to the cleft may,
possibly, be in some way related to the difference in relation
of the chorda to the “Ohrcolumella” in Amphibia and
Reptila, noted by Gaupp. According to that author (22,
p- 159) the nervus hyomandibularis facialis of Amphibia runs

188 EDWARD PHELPS ALLIS, JUN.

backward over the columella and then gives off its ramus
mandibularis internus, which Gaupp considers the un-
questioned homologue of the chorda, “der ohne jede weitere
Beziehung zur Ohrcolumella aussen um das Zungenbeinhorn
herum zu seinem Endgebiete am Unterkiefer verlauft.” In
Reptilia, on the contrary, the mandibularis internus is said to
take a recurrent course and to pass forward over the columella
again to reach its destination, thus having exactly the oppo-
site relations to that structure that it is said to have in
Amphibia.

While I am unable to ever properly follow this discussion,
in so far as it relates to mammals, it is evident that Dixon’s
figure 15 shows a striking resemblance to the conditions found
in Amia, the chorda tympani being considered as postspiracu-
lar in position, and as represented in the mandibularis
internus facialis of the fish, and the mandibularis internus
trigemini of Amia being considered as the homologue of the
lingual nerve of man. It is of course possible that the
mandibularis internus trigemini of Amia may contain both
facial and trigeminal fibres, and hence represent not only the
lingual nerve but also the chorda, if that nerve be a pre-
spiracular one.

In short, if the chorda be a prespiracular nerve I look for
its homologue in the mandibularis internus trigemini of my
descriptions of Amia. If it is a postspiracular nerve it must
be represented in the mandibularis internus facialis of the
same fish.

In this connection it seems proper to give quite fully the
late Prof. Huxley’s opinion of this nerve, as set forth in the
unpublished manuscript to which I have already made
reference. In this memoir Prof. Huxley first describes the
chorda in man as a branch of the posterior division of the
portio dura or seventh nerve. He then says that, “In the
frog, the portio dura arises in close connection with the
auditory nerve from the medulla oblongata, and as it passes
outwards becomes so intimately united with the Gasserian
ganglion that it has been commonly described as a root of the

MUSTELUS LAVIS. 189

trigeminal. In the tadpole, however, and in all the Urodela,
the portio dura forms a ganglion which lies close to the
Gasserian, but perfectly distinct from it, and which answers
to the ganglion geniculatum.

“Tn all the Urodela a commissural trunk connects the
portio dura with the Gasserian ganglion; this I apprehend
represents the nervus petrosus superficialis minor.

“From the ganglion of the seventh, whether it be closely
united with that of the fifth or not, two nerves proceed—a
posterior and an anterior.

“1, The posterior nerve is the larger. It passes outwards
in front of the auditory capsule in the frog, or beneath its
anterior end in the Salamander (but being in each case
morphologically anterior to the labyrinth), and then turns
outwards. In the frog it enters the tympanic cavity, and
keeping close beneath the tegmen tympani and in immediate
apposition with the outer wall of the auditory capsule, it
passes above the level of the fenestra ovalis and stapes, over
the columella auris; and finally turning downwards appears
in the region between the hinder edge of the suspensorium
and the hyoidean arch. ‘The course of this part of the nerve
is therefore precisely similar to that of the trunk of the
posterior or facial division of the seventh nerve in man, while
it is contained in the Fallopian canal. Moreover, it gives off
an anastomotic branch to the glossopharyngeal, and supplies
the depressor of the mandible and hyoidean muscles. And it
gives off a slender branch which passes directly downwards
and forwards to the inner side of the articulation of the
mandible with the suspensorium, and runs along it to the
symphysis, anastomosing with the parallel mandibular branch
of the third division of the trigeminal. It is clear that the
nerve is the homologue of the chorda tympani. In all
respects there is a complete correspondence between the
posterior division of the seventh nerve in the frog and the
facial nerve in man.....

“Tn osseous fishes the portio dura is very generally closely
united with the trigeminal, but in the Plagiostomes the nerves

190 EDWARD PHELPS ALLIS, JUN.

are quite distinct as in the Urodela. Ina common ray, or
skate, for example (Raia clavata, R. batis), the portio dura
arises from the medulla oblongata in close contiguity with
the trigeminal in front and with the auditory nerve behind.
Tt then turns outwards through a distinct canal, in front of
the membranous labyrinth, and reaching the exterior of the
auditory capsule divides into two branches, a very stout
posterior, and a small anterior division, Stannius (l.c., p. 57)
has observed ganglion corpuscles at the bifurcation so that it
answers to the ganglion geniculatum.

“T, The posterior division passes outwards and backwards
behind and beneath the attachment of the spiracular carti-
lage, and behind the spiracle, just as, in the frog, it passes
behind and beneath the otic process of the suspensorium and
behind the tympano-Eustachian canal. It then turns over
the hyomandibular cartilage, about the middle of its length,
just as, in the frog, the nerve passes over the columella auris,
and then divides into branches, the greater part of which are
distributed to that aggregation of cutaneous sensory tubes
which lies behind the angle of the jaw; others go to the
muscles of the vicinity, and two slender branches run for-
wards in the inner side of the mandibular cartilage and
represent the chorda tympani, which, as in the Amphibia,
goes direct to its destination.”

Huxley thus considered the chorda tympani as a post-
spiracular nerve, and he definitely identified it in the ramus
mandibularis internus facialis of Amphibia and fishes. The
homology of the chorda tympani with the ramus mandibularis
internus facialis of the Anura, which is said to have its
homologue in the ramus alveolaris facialis of the Urodela, is
now very generally accepted (Strong, Gaupp). This branch
of the facialis nerve cannot then be the homologue of the
similarly named nerve of fishes, unless that nerve also is the
homologue of the chorda. This subject thus clearly needs
further investigation, and first of all it is absolutely indis-
pensable to know the definite relations of the chorda of
mammals to the spiracular cleft.

MUSTELUS LAVIS. 191

Spiracular Cleft.

The spiracular cleft of a 5 cm. embryo of Mustelus, with
its two diverticula, has been described by Wright (70).
Ridewood (54) has since made short reference to these diver-
ticula in the adult Mustelus, but, although referring to
several earlier authors, he apparently was not acquainted
with Wright’s paper. In my 12:2 cm. embryo I find both of
these diverticula, and also what seems to be a third one; and
the structures deserve a fuller description than Wright has
given.

What Wright describes as the ventral portion of the
anterior diverticulum is, in my embryo, a relatively large
pocket projecting dorso-mesially from the antero-dorsal sur-
face of the spiracular cleft, near its inner end. From the
dorso-antero-mesial portion of this pocket two small apertures,
lying one directly posterior to the other, lead, the one forward
and the other forward and upward, into wholly separate diver-
ticula.

The anterior aperture leads into a small diverticulum which
runs almost directly forward through sixteen sections of my
embryo. It is at first of oval section, but before it ends it
becomes a simple slit, which inclines mesially and upward
toward the skull. The diverticulum has at its extreme
anterior end two short but separate and distinct prolonga-
tions, or lobes, which extend through two sections only. The
diverticulum may be called, because of its position, the ventro-
mesial diverticulum of the fish. It isnot described by Wright.
It is lined with an epithelium which, near the posterior end of
the diverticulum, seems to resemble exactly that of the spira-
cular cleft. Toward the anterior end of the diverticulum the
lining membrane has somewhat the character of the “ modified
epithelium ” that Wright describes in his so-called anterior
diverticulum.

The posterior aperture of the spiracular pocket of my
embryo leads into what Wright has described as the distal

192 EDWARD PHELPS ALLIS, JUN.

portion of the entire anterior diverticulum of the fish. The
aperture represents, in fact, that middle point of the entire
diverticulum of Wright’s descriptions where, according to
that author, the lumen of the diverticulum becomes suddenly
constricted. From here the diverticulum in my embryo led
upward and forward until it acquired a position slightly
dorso-lateral to the ventro-mesial diverticulum. There it
continues forward until it reaches the anterior end of the
ventro-mesial diverticulum, where it first takes a sharp turn
dorso-laterally, then another forward again, and ends. Its
dorso-anterior portion, distal to the first sharp bend, is some-
what enlarged, as Wright has said, and is lined with what he
describes as a “modified epithelium.” The entire diver-
ticulum may be called the dorso-mesial diverticulum of the
fish.

The ventral opening of the pocket from which these two
mesial diverticula arise lies mesial to the spiracular pseudo-
branch. Lateral to that pseudobranch, and from the opposite
wall of the cleft, the auditory diverticulum of Wright’s
descriptions has its origin. The pseudobranch, which is, in
position, a mandibular hemibranch of the spiracular cleft,
thus lies between the regions where these two diverticula
have their origins from the cleft.

The auditory diverticulum runs upward and mesially, at
first immediately posterior to the levator maxille superioris,
and then internal to the dorsal portion of that muscle, and
reaches the side wall of the skull. There it turns upward
and laterally along the sloping side wall of the skull, and, as
Ridewood says, broadens in an antero-posterior direction ; its
upper end thus being T-shaped, and here lying external to
and below the external semicircular canal of the ear. This
auditory diverticulum has, in sections of Mustelus, most
decidedly the appearance of the antero-dorsal end of the first
branchial (hyobranchial) cleft of sections of larvee of Amia.
In this latter fish this antero-dorsal end of the first branchial
cleft is a dorsal pocket, or cranial extension, of the cleft, and
it almost reaches the side wall of the skull in the otic region

MUSTELUS LAVIS. 193

The conclusion thus seems inevitable that it has its serial
homologue in the auditory diverticulum of Mustelus, that
diverticulum then being nothing more nor less than the dorsal
end, or dorsal pocket, of the spiracular cleft. The truncus
facialis in Mustelus lies always ventro-mesial to this diver-
ticulum, and the diverticulum nowhere presents a modified
epithelium such as is found in both the mesial diverticula.
On the dorsal portion of the orbital wall of the skull,
ventro-mesial to the foramen by which the ramus oticus
facials pierces the overhanging post-orbital process, and
directly external to the external semicircular canal of the
ear, a large and strong ligament has its origin. It is the
superior postspiracular ligament of Ridewood’s descriptions,
and has already been several times referred to. Running
downward mesially and backward, external to the postorbital
blood sinus, it reaches the dorsal surface of the dorso-mesial
diverticulum of the spiracular cleft. There it spreads out,
and seems to separate into three somewhat separate parts.
One part continues ventro-mesially and is lost in a mass of
dense connective or fibrous tissue that lies antero-mesial to
the dorso-anterior end of the diverticulum. A somewhat
separate part of this fibrous tissue directly surrounds the
distal end of the diverticulum, and into it a second part of
the ligament has its insertion. The third and larger portion
of the ligament continues downward, backward and laterally,
lateral to the spiracular pocket from which the two mesial
diverticula arise, and then posterior to the spiracular cleft
itself. At its distal end it separates into two parts, both of
which are inserted on the distal end of the hyomandibular,
near its anterior edge, or in the articular ligaments that bind
the hyomandibular and palato-quadrate cartilages together.
It passes, as already stated, between the truncus hyoideo-
mandibularis facialis and that trigeminal nerve that mner-
vates the levator maxille superioris. The truncus facialis
lies between the ligament and the anterior edge of the car-
tilaginous hyomandibular, and this may give an explanation of
the varying and perplexing relations of the facial nerve to the

194 EDWARD PHELPS ALLIS, JUN.

hyomandibular in different fishes. The chondrification of a
ligament is of frequent occurrence, or, if the ligament itself
does not chondrify, it may determine the direction of a car-
tilaginous process of an element on which it has its insertion.
If the hyomandibular of Mustelus were to acquire a higher
articulation on the side wall of the skull it would become
parallel to, and lie immediately posterior to, the superior post-
spiracular ligament. If then this hgament were to chondrify,
or if it were simply to determine the direction of growth of a
cartilaginous process from the ventral end of the hyoman-
dibular, the truncus facialis would become enclosed in the
cartilage, and the relations of element and nerve found in
Amia and teleosts would arise. If, furthermore, the anlage
of the selachian hyomandibular and that of its related liga-
ment thus lying parallel to each other, the primary chondrifi-
cation of the region were to take place in the ligament, and
the tissues that represent the hyomandibular were to acquire
a ligamentous instead of a cartilaginous character, the rela-
tions of the nerve to the element shown by Van Wiyhe (65)
in Acipenser, Spatularia, and Polypterus would arise. And
in Polypterus I find a hgament extending, posterior to the
main facial nerve, from the dorsal to the ventral end of the
hyomandibular, enclosing the nerve between itself and the
hind edge of the hyomandibular, exactly as might have been
expected. The hyomandibulars of these several fishes would
then not be strictly homologous structures, and one would
not have to assume, in order to explain the facts, that the
facial nerve cuts through the hyomandibular, being some-
times found in front of that element, sometimes behind it, and
in still others intermediate in position between these two
extremes.

Wright says, as already stated, that the truncus hyoideo-
mandibularis facialis of his embryo of Mustelus “ forked,”
between the diverticula of the hyomandibular cleft, into the
pre- and post-trematic branches of the nerve. The pre-
trematic branch is said to “appear” to “end in the mucous
membrane of the dorsal and anterior wall of the hyoman-

MUSTELUS LA&AVIS. 195

dibular cleft, without there being any modification of the
epithelium at the point in question.” The modified epithelium
of the anterior diverticulum of his descriptions of the cleft,
considered by Wright as a neuro-epithelium, is said to receive
some fibres from this pretrematic branch.

In none of my embryos did the truncus facialis here
separate into pre- and post-trematic portions, and it seems
to me almost beyond question that Wright mistook the
superior postspiracular hgament of the fish, together with
a nerve described below, for a pre-trematic branch of the
facialis, a part of the lgament, together with the nerve,
having, in embryos, almost exactly the position and dis-
tribution of the so-called nerve described by him. The
modified epithelium of Wright’s anterior diverticulum, my
dorso-mesial one, received, in all my embryos, a nerve which
made its first appearance in sections as a large nerve lying in
the middle of the postspiracular ligament at the point where
that hgament begins to spread and break up into its three parts.
The section of the nerve as it here first appeared in following
the series of sections backward was large, and lay close to
the truncus facialis, but it was distinctly and definitely
separated from that truncus by a small slip of the hgament.
Traced posteriorly the nerve ran downward and backward
to, and ended in, the modified epithelium of the dorso-mesial
spiracular diverticulum. Fibres of it here quite certainly
also went to the modified epithelium of the closely adjacent
ventro-mesial diverticulum, but this could not be definitely
traced because of the dense fibrous tissues that here surround
both diverticula. In the section next anterior to the one in
which this spiracular nerve thus suddenly appears, there
was no indication whatever of it, excepting only a shght
discoloration of the mesial edge of the postspiracular hga-
ment. I was, therefore, at first inclined to consider the
nerve as a branch of the truncus facialis, as Wright had
done before me. Later I became convinced that it could
not be a branch of the truncus facialis—that I can posi-
tively affirm,—and I was then first inclined to consider it

196 EDWARD PHELPS ALLIS, JUN.

as a part of the ligament and not a nerve. Still later I
found that the slight discoloration of the mesial edge of the
ligament, noticed in the section next anterior to the one im
which the nerve so suddenly and distinctly appeared, could
be traced upward and forward to a point where the ligament
crossed, and lay closely against that branch of the ramus
oticus that innervates organs 87 to 91 inf. Here, then, was
a rational explanation of it, and as I could not definitely
trace the nerve in sections, | had Mr. Nomura try to find it
by dissection. He succeeded, in two embryos, in finding the
dorsal end of the nerve, where it turned downward along the
ligament, but he could not trace it to its termination in the
diverticulum of the spiracle. As I readily find its ventral
end in sections, I am fully convinced that the nerve is a
branch of the oticus facialis that has this course and origin—
a conclusion practically confirmed by Hoffmann’s (86) work
on Acanthias. Such being the case, and there being no
special sensory tissue in the auditory diverticulum of Mus-
telus, it is plainly evident that the sensory diverticula of
Amia and Lepidosteus cannot be the homologues of the
auditory diverticulum alone, of Mustelus, as Wright was
inclined to believe, unless it be assumed that the spiracular
sensory organs in the three fishes are not homologous. It
seems much more reasonable to assume either that the three
diverticula of Mustelus have fused to form the single diverti-
culum of Amia, and of Lepidosteus, or that the auditory
diverticulum of Mustelus has entirely disappeared in Amia,
but is still represented in Lepidosteus in that dorso-lateral
branch of the diverticulum of the fish that Wright considered
as the blind outer end of the spiracular cleft itself. In
Wright’s descriptions of the cleft in Lepidosteus he says,
“The cleft itself is directed upwards, outwards, and shghtly
backwards, but it will be readily observed from fig. 5 that
it is separated from the most anterior filaments of the pseudo-
branch by the whole thickness of the hyomandibular adductor
muscles.” In my copy of Wright’s paper the plates are
lacking, so I cannot compare this statement with the figure

MUSTELUS LEVIS. 197

referred to; but it seems singular that the pseudobranch
should have acquired a position posterior to the adductor
hyomandibularis instead of lying in front of it, where it
would naturally he, and where it is found in Amia.

The auditory diverticulum of Mustelus being simply the
dorsal end or pocket of the spiracular cleft, do the two
anterior or mesial diverticula of this fish represent similar
pockets of more anterior clefts ?

Hoffmann (36, pp. 350 and 366) describes, in Acanthias,
what he calls the “dorsale Spritzlochanhang.” This pocket
of the spiracular cleft is said to receive at its suinmit a
ventral branch of what Hoffmann designates as the “ Ramus
accessorius ” of the facial nerve, that ramus being said to be,
in all probability, the ramus oticus facialis of other authors.
This pocket in Acanthias must therefore be what { have
described as the dorso-mesial diverticulum of Mustelus.
Hoffmann considers this pocket in Acanthias as a remnant
of that branchial cleft that should be found between the
mandibular arch and Gegenbaur’s second labial arch. He
then states that a still more anterior cleft, which must have
originally existed between the first and second labial arches,
has disappeared entirely, without even leaving a remnant.
If Hoffmann’s interpretation of these pockets 1s correct, may
not the ventro-mesial diverticulum of Mustelus be such a
remnant ?

Hoffmann and Wright both consider those tissues of the
spiracular diverticula of selachians that are innervated by the
so-called ramus accessorius of the one, and the ramus pretre-
maticus VII of the other, as of undoubted hypoblastic origin,
Both authors accordingly consider the nerve that is distributed
to these tissues as a ventral branch of a dorsal cranial nerve.
If the sensory organs here concerned are of hypoblastic
origin, they would seem necessarily to be homologous with
the sensory buds described by Alcock (1) on the diaphragm
of Ammoccetes. These organs of Ammocecetes are, however,
only found in the glossopharyngeal and vagus arches, as
I understand Alcock, and they are innervated by a nerve

198 EDWARD PHELPS ALLIS, JUN.

that arises from the ventral division of the nerve of the arch,
wholly separate and apart from the ramus dorsalis of the
segment. The innervation of the organ in Mustelus by a
branch of the ramus oticus, usually considered as a ramus
dorsalis, and its position in what Hoffmann considers as a
remnant of the mandibular cleft, are accordingly both de-
cidedly opposed to the assumption that the organ finds its
homologue in the hypoblastic organs of Ammoceetes. Hoff-
mann avoids this difficulty, in so far as it relates to the
innervation, by the very legitimate conclusion that both the
ramus oticus and the ramus buccalis, the nerve from which
the oticus arises, are ventral and not dorsal nerves.

In my work on Amia (2) I was led to suggest that the
spiracular organ of that fish was simply an infraorbital lateral
canal organ, of ectodermal origin, that had wandered into the
spiracular cleft as the external opening of that cleft was
closed. If the same assumption be made regarding Mustelus,
it is evident that, if Hoffmann’s conclusions are correct, the
cleft into which the organ wanders must be the mandibular
and not the spiracular cleft. It is also probable, under
the same assumption, that another infraorbital organ similarly
wanders into the next anterior, or labial cleft. That these
assumptions are not improbable is evident from Hoffmann’s
account (86, p. 339) of the similar wandering of what he
calls rudimentary “ Hautsinnesorgane” from the outer sur-
face into the branchial clefts related to the glossopharyn-
geus and vagus nerves. These rudimentary cutaneous sensory
organs are said to be formed in relation to the ganglion of the
ramus ventralis of the nerve concerned, and they are said to
abort completely soon after entering the related cleft. From
Hoffmann’s account of them they seem not to be considered
by him as organs homologous with the organs of the
lateral canals, but it seems probable that they find their exact
serial homologues in the so-called spiracular organs of
Mustelus, excepting only in their being innervated by a
branch of a ramus ventralis instead of by a branch of a so-
called ramus dorsalis. This difficulty is, however, wholly

MUSTELUS LAVIS. 199

overcome if one accepts Hoffmann’s conclusion, already stated,
that the ramus buccalis facialis is a ramus ventralis and not
a ramus dorsalis. While I can express no opinion as to this,
it seems to me highly probable that the ramus buccalis may
be a serial homologue of the ramus mandibularis externus
facialis, and if it be accepted as such it would seem to lead
legitimately to the conclusion that the separation of the
buccalis lateral canal of fishes, by its innervation, into three
somewhat separate sections, is due to the fact that each
one of these sections has a descending course, one in relation
to the mandibular arch and the other two in similar relations
to the next two anterior arches. The preopercular canal of
ganoids and teleosts would then be a similar descending line
related to the hyoid arch, and the descending line of epithelial
pits described by Alcock (1, p. 145) in the glossopharyngeal
arch of one specimen of Ammoccetes would be a similar line
related to that arch. The several ventral pit lines of Ammo-
coetes would then represent lines corresponding, in the several
arches to which they are related, to the mandibular part of
the canal line of the hyoid arch of Amia and teleosts, these
lines not existing in the premandibular arches.

One of the branches of the ramus pretrematicus facialis of
Mustelus runs backward and then outward posterior to the
dorso-mesial diverticulum of the spiracular cleft, and there
turns downward posterior to the latter cleft. The other two
branches, and also the so-called chorda tympani, lie anatomi-
cally anterior to the dorso-mesial diverticulum of the spiracu-
lar cleft. If, then, that diverticulum represents a remnant of
the mandibular cleft, the nerves here concerned must actually
lie anterior to that cleft. How they could have acquired that
position, if they are regular pretrematic branches of the nerve
of the hyoid arch, it seems difficult to imagine, unless it be
assumed that the persistent remnants of the mandibular and
more anterior clefts represent only the outer, external portions
of the clefts concerned, the fusion of the clefts with the hyoid
cleft, and with each other, taking place at their external ends
only, the internal ends wholly aborting. Hoffmann’s descrip-

VOL. 45, PART 2,—NEW SERIES, p

200 EDWARD PHELPS ALLIS, JUN.

tion of the development of the diverticulum in Acanthias
seems, in fact, to indicate that this is what takes place. That
the internal ends of the clefts could have persisted, and that
the prebranchial facial nerves could have in any way slipped
forward over the aborted top of the cleft, seems excluded by
the fact that both a ligament and a branch of the oticus
facialis extend downward to the dorsal end of the cleft,
directly across the path that the nerves concerned must have
taken in so slipping forward.

Trunecus hyoideo-mandibularis facialis.

The truncus hyoideo-mandibularis facialis, after its origin
from the dorso-posterior surface of the trigemino-facial gan-
elion, runs backward and upward along the side wall of the
skull, immediately anterior to the hind end of the peri-orbital
sinus. It is accompanied, in this part of its course, by the
r. ad. musc. lev. max. sup., as already fully described. Soon
after this trigeminal nerve turns outward, anterior to the
superior postspiracular ligament, the truncus facialis turns
outward posterior to that hgament, passing, as the trigeminal
nerve does, ventral to the postorbital blood-sinus. The
truncus facials here runs dorsal to the posterior portion of
the dorso-mesial diverticulum of the spiracular cleft, and
antero-mesial to the dorsal end of the auditory diverticulum
of the same cleft. Its relations to the spiracular structures
are thus exactly those given by Ridewood (54) in his descrip-
tions of the corresponding parts of Scylium. The truncus
then turns downward and laterally, passes immediately
posterior to the spiracular cleft, and reaches a position lateral
to the anterior edge of the distal end of the hyomandibular.

Shghtly distal to the pot where the truncus thus turns
downward and laterally, a small ganglion forms on its hind
edge. From, or in connection with, this ganglion four
branches arise. Three of them run outward along the
anterior edge of the muscle Csd, passing posterior to the
auditory diverticulum of the spiracular cleft, and postero-

MUSTELUS LAVIS. 201

ventral to the levator maxillze superioris. Having reached
the outer surface of Csd, they there turn backward, and then
break up and spread out on the outer surface of the muscle,
near its ventral end. They are apparently entirely sensory,
but they evidently correspond to the branches described and
figured by Vetter in Acanthias, one of which branches is said
to penetrate Csd, from its outer surface, and innervate it.
The fourth branch, in Mustelus, arises from the hind end of
the small ganghon, runs outward to the inner surface of the
muscle Csd,, and there separates into two parts, one of which
penetrates the muscle, and running backward is lost in it,
while the other turns backward along the inner surface of
the muscle, and there gradually disappears. This fourth
branch thus probably contains all the motor fibres destined
to the muscle Csd,. What the ganglion from which the four
branches arise may be I cannot determine, nor do I find any
description of anything that seems to correspond to it in any
of the works at my disposal. The ganglion cells have the
size and general appearance of those of the ciliary ganglion,
which may indicate that 1t is a sympathetic ganglion.
Beyond this ganglion and the branches that arise from or
in connection with it no branch is given off by the truncus
facialis until it reaches the level of the distal end of the
hyomandibular, where the nerve separates into several
branches. One of these branches is destined to innervate
the sense organs of the hyomandibular lateral canal, and
another those of the mandibular lateral canal and the mandi-
bular group of ampulle. These two branches arise from the
truncus close together, if not as two parts of a single branch,
and they together constitute the ramus mandibularis externus
facialis of the fish. The branch destined to the organs of
the hyomandibular canal separates into two parts, both of
which run at first laterally and shghtly downward, but soon
separate, one turning forward and the other backward, both
going exclusively to the sense organs of their canal, no
related ampullee bemg found. The other branch or part of
the externus runs downward, forward, and mesially, along

202 EDWARD PHELPS ALLIS, JUN.

the external surface of the adductor mandibule muscle,
passes into the mandible, lying superficial to the ramus
mandibularis trigemini, and ends in the lateral and ampullary
sense organs it innervates.

Beyond the point where the ramus mandibularis externus
is given off by the truncus hyoideo-mandibularis the remaining
portion of the truncus separates into three principal branches,
and one or more smaller ones. The smaller ones go to the
general tissues of the region. One of the three larger and
principal parts runs at first downward and backward, and
acquires a position on the external surface of the posterior
portion of the adductor mandibule. It then turns forward
and mesially, and passes from the outer surface of the
adductor on to that of the muscle Csv,, where it gradually
disappears. A second one of the three principal parts runs
directly backward, immediately lateral to the dorsal end of
the ceratohyal, and near the hind edge of that element turns
downward, internal to the muscle Csv,, and around the hind
end of the mandibular cartilage. It then runs forward and
mesially between the superficial and deeper layers of Csvo,
to both of which it sends branches, apparently mnervating
them. Anteriorly it les immediately internal to the inter-
mandibularis muscle, and there anastomoses completely with
that branch of the mandibularis trigemini that pierces the
intermandibularis from its outer surface. ‘his anastomosis
is, as already stated, the undoubted homologue of the one
found in Amia of the ramus hyoideus facialis, with the branch
r.ghi of the maxillaris inferior trigemini.

These two principal branches of the truncus facialis of
Mustelus thus form the ramus hyoideus of Stannius’ descrip-
tions (59, p. 65).

The third and remaining part of the truncus facialis of
Mustelus turns downward, forward, and mesially between the
ceratohyal and the mandibular cartilage, and closely accom-
panies the pseudobranchial artery as far as that artery
extends. Continuing forward beyond the artery, internal to
the mandibular cartilage, and near its ventro-mesial edge, it

MUSTELUS LAVIS. 203

there gradually disappears, numerous branches being sent
toward the adjacent mucous lining membrane of the mouth
cavity. In the terminal part of its course it lies quite close
to the terminal portion of that branch of the ramus palatinus
that Green considers as the homologue of the chorda tympani.
This part of the truncus facialis of Mustelus is thus the ramus
mandibularis internus s. profundus of Stannius, and it is the
nerve that Ruge (55) and many others homologise with the
chorda tympani of man. Whether it be or be not that nerve
depends on whether the chorda is a post- or pre-spiracular
nerve.

Review and Comparison of the Ophthalmic Nerves.

There is, as 1s only too well known, much confusion in the
descriptions of, and more particularly in the nomenclature
relating to, the ophthalmic nerves of fishes. This is often so
misleading that I venture to give, before discussing the
nerves, an explanation of the manner in which it seems to
have arisen.

Stannius (59, p. 54) included all the ophthalmic nerves of
fishes under the name ramus primus n. trigemini s. ophthal-
micus, a name which, shortened to 'Trigeminus I, is even still
retained by certain authors. Stannius says that the nerve
thus defined by him is always wholly sensory, excepting in
the Cyclostomata, and that it is formed by the union of two
bundles of fibres, one derived from an anterior trigeminal
root and the other from a posterior one. In certain teleosts,
he says that the ramus ophthalmicus is found as a single
trunk which runs forward, near the roof of the orbit, dorsal
to all the muscles of the eyeball. In other teleosts, and in
all the Plagiostomata, the nerve is said to be represented by
two branches which sooner or later unite, more or less com-
pletely, to form a single trunk. In these latter teleosts both
branches of the nerve are said to have the course of the
single trunk in the first-mentioned ones. while in all the

204 EDWARD PHELPS ALLIS, JUN.

Plagiostomata one branch only of the nerve has that course,
the other running forward across the orbit ventral to the
rectus superior and obliquus superior muscles ; its relations to
the rectus internus not being especially mentioned. ‘This
latter branch of the nerve in the Plagiostomata is called by
Stannius the ramus ophthalmicus profundus, and as he does
not apply this term, profundus, to the deeper one of the two
branches in those teleosts in which two ophthalmic branches
are said to be found, it seems evident that he intended the
term to be limited to a nerve that held a position anatomically
different from that of an ophthalmicus superficialis. It seems,
however, equally evident that he considered the two strands
of the superficial nerve in teleosts as the homologues of the
two more completely separated branches in the Plagiostomata.
He says that the superficial strand in teleosts, and the ramus
superficialis in the Plagiostomata, are both derived from a
posterior, broad-fibred root, while the deeper strand in the
former, and the ramus profundus in the latter, arise from the
anterior trigeminal root.

Schwalbe (57, p. 182) concluded that the ophthalmicus
superficialis of Stannius’s descriptions of selachians was formed
of two distinct and different components, the larger one of
which always lay superficial to the other. He accordingly
called the two components the portio major and portio minor
of the superficialis nerve, and says that they derive their
fibres partly from the dorsal part of the posterior root of the
trigeminus, and partly from the anterior, more ventral root of
that nerve. These two bundles of fibres he calls the “ radix
dorsalis (posterior) and ventralis (anterior) ophthalmici.”
The portio major of the ophthalmic nerve is said to receive
all the fibres that traverse the radix dorsalis, and to be com-
posed of those fibres alone. The portio minor receives one
half of the fibres that traverse the radix ventralis, the other
half of those fibres going to form the ramus ophthalmicus
profundus. The portio minor and the ramus profundus are
thus formed of similar fibres, and those fibres are said to be
derived from the anterior root of the trigeminus. The ramus

MUSTELUS LAEVIS. 205

profundus is said to either pierce the rectus superior muscle,
or to he wholly ventral to it, and then to run forward ventral
not only to the obliquus superior, as stated by Stannius, but
ventral also to the rectus internus. Schwalbe’s descriptions
and figures also show that the ramus profundus lies between
the superior and inferior divisions of the nervus oculomotorius,
ventral to the former and dorsal to the latter, but he does not
call attention to this important fact. In his references to
teleosts and ganoids he makes no application of the terms
portio major and portio minor.

Balfour in his well-known work on the development of
elasmobranch fishes, to which I am unable to directly refer,
and which was published shortly before Schwalbe’s work
above referred to, described, according to Marshall and
Spencer (43), the ophthalmic nerves both in the adult and in
embryos of Scylhum. In those descriptions he is said to
have entirely overlooked the true profundus nerve, giving the
name ramus ophthalmicus profundus to what Schwalbe
describes as the portio minor of the superficial nerve.
Balfour, although he was thus, as Marshall and Spencer state,
the first to clearly recognise the double nature of the super-
ficial ophthalmic nerve of elasmobranchs, was also probably
the first to make that misapplication of the name profundus
that has simece caused so much confusion in the descriptions
and discussions of these nerves. Balfour evidently did not
so apply this term either thoughtlessly or carelessly, for
Marshall and Spencer state that he sought to explain the
apparently exceptional position of the nerve in Scyllium by
the assumption that it had shifted upward from the deeper
position in which it is usually found in selachians ; this
explanation adding, in my opinion, a misconception of the
nerves to a misapplication of their names.

Marshall, in his own embryological work on Scyllium (42),
and also in a second work on the same fish published in con-
nection with Spencer (48), concluded that the portio major
and portio minor of Schwalbe’s descriptions were, respec-
tively, the ophthalmic branches of the facial and trigeminal

206 EDWARD PHELPS ALLIS, JUN.

nerves (p. 89). The ramus ophthalmicus profundus is said
by him to be of a totally different nature from either of these
superficial ophthalmic nerves, and is said to be definitely
characterised by its own particular course and position, from
which, in his opinion, it never shifts. This opinion, definitely
and succinctly stated, seems to have since been either entirely
overlooked, or, perhaps, regarded as not worthy even of con-
sideration.

Van Wijhe (66) confirmed Marshall’s conclusion that the
portio major and portio minor of Schwalbe belonged, respec-
tively, to the facial and trigeminal nerves, and he moreover
found in stages K and L of embryos of Scyllium and Pris-
tiurus, what he considers as a third component, or portio, of
the superficial ophthalmic nerve. — It is said by him to arise
from his ciliary (profundus) ganglion, and, because of its
origin and distribution, he calls it the portio ophthalmici pro-
fundi of the ramus ophthalmicus superficialis. This portio
profundi is said (p. 21) to abort in shghtly older stages of
Scyllium and Pristiurus, but in Polypterus and Lepidosteus it
is said to persist throughout life. It is fully described by
him, in these latter fishes, m another work (65). In this
latter work (p. 275) van Wijhe also identifies and describes,
in the adult Lepidosteus, the other two portiones (portio tri-
gemini and portio facialis) described by him in the superficial
ophthalmic nerves of embryos of Scyllium and Pristiurus. In
the adult Polypterus, his descriptions, in the same work, lead
one to infer that he found all three portiones in that fish also,
but the reference is not clear, and Pollard (51, p. 395),
somewhat later, could not definitely establish the existence of
a portio trigemini. In Polypterus both van Wijhe and
Pollard describe a ramus ophthalmicus profundus which is
said to have the usual selachian relations to the rectus
superior and obliquus superior muscles, but not to the rectus
internus, the nerve lying dorsal, instead of ventral, to that
muscle, the internus of Polypterus, like that of Amia and
teleosts, hence perhaps not being the homologue of the simi-
larly named muscle in selachians,

MUSTELUS LA!VIS. 207

In the same work in which van Wijhe describes the nerves
in Polypterus and Lepidosteus, he also describes the nerves in
Acipenser. In this latter fish, doubtless misled by Balfour,
he first describes an ophthalmicus superficialis trigemini and
an ophthalmicus profundus trigemini which have the mutual
relations of the superficialis and profundus nerves of Bal-
four’s descriptions of Scyllium. Later, but in the same
work (p. 250), van Wijhe says, in a footnote, that the
anatomical position of this so-called profundus nerve pre-
cludes its being a ramus profundus, in the sense in which that
term is applied in selachians, and he accordingly coneludes
that the two ophthalmic nerves of his descriptions of Aci-
penser are very probably simply the portio major and _ portio
minor of Schwalbe’s nomenclature, and hence both integral
parts of the ramus ophthalmicus superficialis. There is
accordingly, according to his descriptions, no true ramus
ophthalmicus profundus in Acipenser, and there is, moreover,
no portio ophthalmici profundi described as such. In here
first giving to the two ophthalmic nerves of Acipenser the
names superficialis and profundus, van Wijhe had probably a
more direct influence than Balfour on that misapplication of
the name profundus that has since been so frequently
repeated, and that is so confusing and misleading if one is
not continually on one’s guard.

Instances of this misapplication that I am sure of are found
in Goronowitsch’s descriptions of Acipenser (24), Lota (25),
and Salmo (26), and in Wright’s descriptions of Amiurus
(69) ; and it is exceedingly probable that the same error
occurs also in Pollard’s descriptions of Siluroids (52). The
misapplications relating to Lota and Amiurus have already
been recognised and signalled by Herrick and Workmann
respectively.

This use of the term ophthalmicus profundus for a nerve
that has the anatomical position of an ophthalmicus super-
ficialis always implies, whether intentionally or not, that a
nerve having the position of a selachian profundus has moved
upward, from that position, to the position of a selachian

208 EDWARD PHELPS ALLIS, JUN.

superficialis, and there fused completely with the portio
minor of the ramus superficialis. Such an ascent of the pro-
fundus nerve is, | believe, very generally assumed to have
taken place in all teleosts and in most ganoids. In Amphibia
and the higher animals it is, on the contrary, as generally
assumed that the portio minor of the superficial nerve has
moved downward and fused with the ramus profundus.
Wiedersheim (64, p. 285) so definitely asserts, and Strong
(60, p. 193) accepts as probably correct, a similar statement
attributed to Wilder in a work I have not at my disposal.
Balfour also definitely accepted this principle in his statement,
already referred to, that the profundus nerve of Scyllium had
shifted upward to the position of a portio minor of the
ophthalmicus superficialis.

The assumed ascent of one, or descent of the other, of
these two ophthalmic nerves seems usually to be made in the
sense of a simple juxtaposition and subsequent fusion of the
two already developed nerves, but it is evident that this
could not take place without the enclosing of the trochlearis
and the superior division of the oculomotorius in the single
nerve so formed. While this might be assumed, from exist-
ine descriptions, to have taken place in certain fishes and
other animals, my work leads me to believe that im every such
instance it will be found that the motor nerves are simply
juxtaposited to one or the other of the two ophthalmic nerves,
and not enclosed between the two nerves united. An appa-
rent exception to this might be considered as being presented
in a single specimen of Carcharias that I examined (5). In
that specimen a part of the trochlearis certainly perforated
and traversed the ophthalmicus superficialis. An important
and apparently normal ophthalmicus profundus was, however,
found, in this same specimen, in its typical place and relations
to the other nerves and structures of the orbit. It might,
nevertheless, be said that this perforation of the superficial
nerve by a part of the trochlearis here represented some
intermediate stage-in the fusion of the superficialis and pro-
fundus nerves. ‘his I do not believe, and while the condition

MUSTELUS--LEVIS.- - -- 209

presented by Carcharias certainly needs further investigation,
its explanation may perhaps be found in the frequent assertion
that the trochlearis is a branch of the ophthalmicus trigemini.

That in the ascent or descent of one or the other of the
ophthalmic nerves one of them could cut through the troch-
learis and oculomotorius, temporarily severing their fibres, or
that it could cut through the intervening eye muscles distal
to the point of attachment of the associated nerve, is not I
think assumed by anyone.

If then a simple juxtaposition of the two nerves, already
developed in their typical positions, be eliminated from the
discussion, it is evident that any assumed fusion of them pre-
supposes, either that one or the other of the two nerves 1s not
primarily developed in its typical position, or that the two
nerves develop before the structures that typically separate
them have been sufficiently developed to interfere with their
juxtaposition and subsequent fusion. ‘I'he whole question of
the development of nerves is thus here involved, and it is not
my intention to in any way discuss it. Certain statements
regarding the development of the nerves here especially
under consideration, and bearing directly upon their relations
to other structures in the orbit, deserve however to be given.

Dixon (15) believes, with His, that the permanent fibrous
nerve grows outward, either directly from the brain, or from
the related ganglion. Of the ophthalmic nerve in man, he
says (p. 33) that the first-formed fibres of the nerve are
deflected from their primary direction by reason of some cor-
relation to other growing tissues, and that they become the
nasal branch of the ophthalmic nerve of the adult. Later
fibres, not similarly obstructed, continue in the primary direc-
tion of the ophthalmic nerve and form the frontal branch of
the nerve of the adult. We thus have, under this theory of
the development of nerves, a statement which certainly in-
volves, in principle, the change of a nerve from the position of
a superficialis to that of a profundus. Because of its position
Dixon considers the frontal nerve as the probable homologue
of the portio trigemini of the ophthalmicus superficialis of

210 EDWARD PHELPS ALLIS, JUN.

selachians, the nasal branch being quite unquestionably consi-
dered as the homologue of the ophthalmicus profundus, though
I do not find this statement definitely made. With equal, or even
ereater reason, the frontal branch of the nerve in man might
be considered as the homologue of the portio ophthalmici pro-
fundi of ganoids, and hence more probably the homologue of
branch I of the ramus ophthalmicus profundus of Ewart’s
descriptions of Lemargus (17) than of the ramus ophthalmicus
superficialis trigemini of that fish. Under either assumption,
Dixon’s facts and conclusions are opposed to the assumption that
the superficials and profundus trigemini of selachians are both
represented in the nasalis, or naso-ciliaris, of higher animals.

Platt (49) says that, in Necturus, the ramus ophthalmicus
profundus is split off, as a line of cells, from the under surface
of the same line of thickened epidermis that later gives similar
origin to the ophthalmicus superficialis faciahs. In my work
on Amia I assumed (3, p. 635), im referring to an earlier
work by Platt (48), that the two nerves thus described by her
were the portio minor and portio major of the ophthalmicus
superficialis, and that they retained, throughout life, the posi-
tions of those nerves. _Platt’s later work shows conclusively
that the profundus of her descriptions is the ramus nasalis of
Schwalbe’s descriptions of other Urodela, in which animals
it is said by the latter author to have exactly the relations, to
other orbital structures, of the ramus ophthalmicus profundus
of selachians. We thus have, under a theory of the develop-
ment of nerves opposed to that of His, a nerve that can be
followed, if Platt is correct, from the position of an ophthal-
micus superficialis to that of an ophthalmicus profundus, This
profundus nerve is said to differ from all the other sensory
nerves of the animal in that it “is formed from the ectoderm
in the same manner as are the cranial ganglia” (p. 488), and
that neural crest cells participate in its formation throughout
its entire length (p. 535). This, it will be seen below, would
seem to exclude the possibility of the nerve containing any of
the elements of the portio minor of selachians, a descent of the
latter nerve thus not here taking place.

MUSTELUS LA&VIs. ola

Goronowitsch (26) says that, in Salmo fario, the two
nerves called by him the nervus ophthalmicus superficialis tri-
gemini and nervus ophthalmicus profundus trigemini are both
developed from ectoderm cells proliferated from the sam
supra-orbital region of the ectoderm. As the profundus nerve
of these descriptions holds the position, throughout life, of a
ramus ophthalmicus superficialis, it might be here assumed
that this nerve contained the elements of the similarly named
nerve in Platt’s descriptions of Necturus, and that in Salmo
its descent to the typical profundus position had been simply
arrested, the nerve remaining where originally laid down. The
profundus nerve, as first laid down in Salmo, is, however, said
to be directly developed from ectoderm cells, and not from
cells that could in any way be attributed to the neural crest,
thus differing so markedly in this from the ophthalmicus pro-
fundus of Necturus, as described by Platt, that the two nerves
cannot be considered as homologous. A marked difference
in the manner of development of this nerve in Salmo and that
of the ophthalmicus profundus of the chick is also noted by
Goronowitsch, who says (p. 31): “ Bei den Végeln zeigt die
Entwicklung des Ophthalmicus Stammes eine Abweichung,
inden der Nervenstamm primiir im Mesoderme angelegt wird.
Der Ophthalmicus der V6gel bekommt auch, wie von mir be-
schrieben, ectodermales Bildungsmaterial durch eine Reihe
von Zellenvermehrungs-Heerden welche lings der Verlaufs-
richtung des Nerven zerstreut sind.” What the exact
anatomical position of this ophthalmic nerve of the chick may
be I cannot find, but I presume it holds the position of a ramus
naso-ciliaris, which is I believe always that of a true profundus.
Marshall, however, says (41, p. 29), that it “passes under the
rectus superior but dorsal of the other eye muscles and of the
optic nerve,”—a somewhat anomalous position.

Neal (45), in Squalus, finds the ramus ophthalmicus pro-
fundus developed from two processes formed at the ventral
end of a descending line of neural crest cells. One of these
processes grows posteriorly, unites with the anlage of the
ganglion of the nervus trigeminus, and forms the basal por-

212 EDWARD PHELPS ALLIS, JUN.

tion of the future profundus nerve. The other process grows
anteriorly and forms the distal portion of the same nerve.
The descending line of neural crest cells is said to be Platt’s
thalamicus nerve, and it is said to later entirely disappear.
The processes that form the profundus nerve do not begin to
develop until the descending line of cells that represents the
thalamicus has reached the dorsal surface of the optic vesicle,
the profundus nerve thus being first laid down in its adult
position. The oculomotorius and trochlearis are later de-
veloped as fibrillar processes from neuromasts in the medulla,
and take at once their adult relations to the profundus. The
nerve that Neal identifies as the ramus ophthalmicus super-
ficialis trigemini develops later than the profundus, and is

first described in 17mm. embryos, where it “

appears as a
fibrillar nerve with peripheral nuclei extending from the Gas-
serian ganglion just dorsal to the poimt of exit of the fibres of
the r. ophth. profundus V, and passing anteriorly close to the
ectoderm below the r. ophthalmicus superficialis VII” (p.
233). The two trigeminal ophthalmic nerves are thus first laid
down in their adult positions, and there is no question what-
ever of the profundus cutting through the oculomotorius or
trochlearis, or of its cutting through the as yet undeveloped
eye muscles. The late development of the ophthalmicus
superficialis corresponds with Dixon’s account of the late
development of the frontal branch of the ophthalmic nerve in
man, and strongly suggests their bemg homologous nerves.
The ramus ophthalmicus superficialis facialis is said by Neal to
develop “in close connection with the skin along what in
the head corresponds with the dorso-lateral line of the trunk.”

There is thus nothing in the development of the ophthalmic
nerves, so far as known, to definitely indicate either that the
portio trigemini of the ophthalmicus superficialis of selachians
ever descends, in other verebrates, to the position of a pro-
fundus, or that the latter nerve ever ascends to, or simply
remains in, the position of a superficialis.

Leaving aside, now, all question of the possibility of a
simple juxtaposition and subsequent fusion of the profundus

MUSTELUS LAVIS. 213

and superficialis nerves, the central origin and composition of
these nerves can be considered.

In Acipenser, Goronowitsch (24) says that the nerve called
by him the ophthalmicus profundus trigemini, and which 1s,
as stated above, simply the ramus ophthalmicus superficialis
trigemini in the nomenclature of most other authors, arises
from the ventro-anterior trigeminal root of the fish, which
root is called by him the root of the nerve Trigeminus I. The
ramus ophthalmicus superficialis trigemini, or ophthalmicus
superficialis facialis of other authors, is said to arise from
Trigeminus II, and this latter nerve is said to arise from the
medulla by two roots, one dorsal and the other ventral. The
ventral root is said (p. 477) to be a thick-fibred, motor one,
arising from a so-called dorso-lateral tract of the medulla.
The dorsal one is said to be a fine-fibred, sensory one, arising
from the lobus trigemini. Strong (60, p. 168) concluded that
the ventral root of Trigeminus II, thus described by Gorono-
witsch, could not be motor, and he ascribed to it a lateral sen-
sory character, the dorso-lateral tract of Goronowitsch being
homologised with the tuberculum acusticum of his own and
certain earlier descriptions of the medulla. The dorsal root
of Trigeminus IT, Strong considers as a part of the fasciculus
communis system, and not as a lateral line nerve (p. 192).
Kingsbury (40) agrees with Strong that the ventral root of
Goronowitsch’s descriptions must be a lateral sensory one, and
he homologises it with a root called by him VIIb in Amia;
that is, with the entire lateral sensory root of the latter fish.
He considers the lobus trigemini of Goronowitsch’s descrip-
tions as simply a somewhat separate part of the tuberculum
acusticum, and says that both it and the root that arises from
it, Trigeminus II dorsalis, are “absent as such” in Amia,
Lepidosteus and teleosts. The fasciculus communis compo-
nent of the V—VII complex 1s, according to him, represented
entirely in the dorsal root of the facialis of Goronowitsch’s
descriptions, he thus differimg from Strong in the root to
which this component is assigned. Goronowitsch, in a later
work (28, p. 12), refers to Strong’s conclusion that the ventral

214 EDWARD PHELPS ALLIS, JUN.

root of Trigeminus II could not be a motor one, but, after
renewed investigation, he reaffirms his belief that it is such.
Johnston (39) follows Strong and Kingsbury in assigning
this ventral root to the lateral sensory system, the dorsal root
of Trigeminus IT being said by him to also belong to the same
system, he thus here agreeing with Kingsbury as against
Strong. Herrick (82), still later, also asserts that the ventral
root is sensory, and not motor, and such is unquestionably
the case.

Goronowitsch does not give, in Acipenser, the separate
peripheral distribution of the fibres arising by each of the two
roots of Trigeminus II, but Strong assumes that the fibres of
the two roots are completely mingled, and that they must
accordingly be found in all the so-called lateral sensory
branches that have their origin from this root (60, p. 179).
This would seem to be practically established by what I find
in Mustelus. Strong says that the branches said by Gorono-
witsch to be derived from the nerve Trigeminus II, his
(Strong’s) facialis, have, in Acipenser, the same course and
position as the lateral sensory branches of the facialis in the
tadpole. They also have the same course and position as the
lateral sensory facialis branches in Amia and Scomber (3 and
7), excepting only the branch said by Goronowitsch to join
and accompany the ramus hyoideus facialis. This lateral
branch of Acipenser seems to be represented in Amia by
that part only of the mandibularis externus facialis that goes
to the preopercular lateral canal, or perhaps even to the dorsal
part alone of that canal. These latter fibres of the mandi-
bularis externus of Amia arise directly from the truncus
hyoideo-mandibularis, as apparent branches of that nerve,
while the remaining, more distal, branches all arise from the
‘amus mandibularis externus after it separates from the ramus
hyoideus, to which nerve the mandibularis externus has, in its
further course, no relations whatever. The absence, in Aci-
penser, of this distal and independent part of the mandi-
bularis externus nerve of other fishes was noticed by van
Wijhe, who says (65, p. 237) that his failure to find it was

MUSTELUS LAVIS. Pil

probably due to its being so small that it was overlooked, or
missed, in dissection. It is, however, to be noticed that
neither he nor Collinge (13) describe, in Acipenser, either a
preopercular or mandibular section of the lateral canals. This
thus certainly calls for further investigation.

According to Goronowitsch (p. 481) neither the ophthal-
micus superficialis nor the ophthalmicus profundus of his
descriptions of Acipenser receives a communicating branch
from the facials roots. If, then, Kingsbury is correct in his
assertion that the communis component of the V—VII com-
plex of Acipenser is derived entirely from the dorsal root of
the facialis of Goronowitsch, it is evident that the communis
fibres, if they exist as such in either of the two ophthalmic
nerves of the fish, must be derived from those fibres that are
said by Goronowitsch (p. 479) to connect the ganglion of
Trigeminus IT with the nervus facialis. As the communis
elements that could possibly enter into the ophthalmic nerves
by this route would necessarily be limited, they certainly
cannot represent the nervous supply of the nerve-sacs of the
snout of the animal, those organs thus quite certainly being
innervated by fibres derived from one of the two roots that
Kingsbury and Johnston both consider as lateral sensory
ones; that is, the nerve-sacs of Acipenser and the ampullee
of selachians are quite unquestionably innervated in the same
manner, and are hence homologous organs.

In Amia, both Kingsbury (40) and myself (8), in works
published at nearly the same time, find two of the trigemino-
facial roots arising close to the root of the nervus acusticus.
Kingsbury calls them VIIb and VIlaa, and says that VIIb
enters the tuberculum acusticum, and is composed of coarse
fibres identical with those of the lateral line nerve; while
Vilaa arises “from the fasciculus communis system which
disappears with the exit of this root” (p. 7). The spinal Vth
tract of the fish is said by Kingsbury (p. 23) to probably
furnish all the general sensory elements of the trigeminal
nerve. ‘The root here called VIIaa by Kingsbury was con-
sidered by me (I. ¢., p. 596) as the antero-dorsal root of the

VOL, 45, PART 2,—NEW SERIES, Q

216 EDWARD PHELPS ALLIS, JUN.

facialis, and was said to have its central origin “ at a high
level in the brain, probably from the fasciculus communis of
Osborne and Strong.” The ramus ophthalmicus superficialis
trigemini of my description of Amia is certainly largely, and
perhaps entirely, composed of fibres derived from this com-
munis root. The ramus ophthalmicus superficialis facialis 1s
certainly largely, and probably entirely, composed of fibres
derived from the tuberculum acusticum. The portio ophthal-
mici profundi is probably exclusively composed of spinal fifth
fibres. Compared with Acipenser the ramus ophthalmicus
superficialis of Amia thus contains a large communis compo-
nent, not found as such in Acipenser, while the nerve of
Acipenser contains a large lobus trigemini component not
found as such in Amia. The inference is evident that the
two components are homologous, and as the communis com-
ponent of the nerve of Amia is largely, and perhaps exclu-
sively, destined to the innervation of terminal buds, and as
there are neither nerve-sacs nor ampullz in Amaia, these latter
organs of Acipenser and selachians must be the homologues
of the terminal buds of Amia. That the ampullee of selachians
are derived from the terminal buds of ganoids and teleosts
was Strong’s impression, rather than opimion, for he says
(60, p. 202), in discussing selachians, that “it would seem
likely that those fibres in the lateral lme nerves of the head
derived from the lobus trigemini are devoted to the innerva-
tion of the ampulle. If this were true, as further research 1s
necessary to show, the ampullee would represent the end buds
of other fishes.”

The descriptions of other fishes do not throw much further
light upon this subject, but they are certainly in accord with,
rather than opposed to, my conclusions. ‘The varying use and
misuse of descriptive terms, and certain probable errors in the
descriptions, make most of the comparisons very difficult and
of but little value. The probable homologies can, however,
be indicated.

In Lota Goronowitsch (25) says that Trigeminus IT arises
by two stems, a dorso-median and a ventro-lateral one. The

MUSTELUS LAVIS. Ai Wy

dorso-median stem is said to arise mainly from a_ special
centre of grey substance, which Goronowitsch says (pp. 25-27)
is peculiar to teleosts,.and which he considers as the homo-
logue of the lobus trigemini of Acipenser. Certain fibres
found in the dorsal root of Trigeminus IT of Acipenser, which
root arises from the lobus trigemini of the fish, are, however,
said by Goronowitsch not to be found in the dorso-median
stem of Lota. What they are is not evident. The ventro-
lateral stem of Lota is said to be formed largely of motor
fibres, but partly of fibres derived from the same centre as
the dorso-median stem. he motor fibres here referred to
are, as in Acipenser, unquestionably lateral sensory ones, as
Herrick has already pointed out (32, p. 210), and it might be
assumed that the dorso-median stem was composed of com-
munis fibres, because of Goronowitsch’s conclusion that the
centre from which the stem arises is the homologue of the
lobus trigemini of Acipenser. ‘lhe composition of the two
ophthalnic nerves of the fish are, however, decidedly opposed
to this assumption. The dorso-median stem of Trigeminus I,
as I understand Goronowitsch, alone sends fibres to his ramus
ophthalmicus superficialis, and as that nerve is largely a
lateral sensory one, the fibres it receives through this stem
must be lateral sensory ones, if there be no error here. ‘The
nerve is not exclusively a lateral sensory one, since it is said
to send certain branches to the skin, but this is not important
in this connection. ‘he ramus ophthalmicus profundus of
Goronowitsch’s descriptions, which is simply a deeper strand
of the superficialis nerve, is said to be formed of two bundles
of fibres, one derived from T'rigeminus I, and the other from
the so-called dorsal root of the facialis. The former bundle
is the spinal fifth component of the ophthalmic nerve, the
latter bemg the communis component. ‘The communis com-
ponent of the ophthalmic nerve thus being accounted for, the
special centre of grey substance from which the dorso-median
stem of Goronowitsch has its origin, cannot be the homologue
of the lobus trigemini of Acipenser if, as I conclude, that
lobus is a centre of communis fibres. As the so-called pro-

218 EDWARD PHELPS ALLIS, JUN.

fundus nerve, which receives no fibres from the lateral sen-
sory centres, is said to send a branch to the sensory canal in
the antorbital bone of the fish, it 1s evident that the subject
needs further investigation. The spinal fifth component of
the so-called profundus nerve is evidently the homologue of
the portio ophthalmici profundi of my descriptions of Amia.

In Amiurus the fasciculus communis system, and especially
its pre-auditory portion, that is, root VIlaa, is said by Kings-
bury (40) to be enormously developed, giving origin to the
lobus trigemini of the fish (p. 30). From this root and system
the deeper ophthalmic nerve of Wright’s descriptions is said
by Kingsbury to derive most of its fibres (p. 14). By Work-
mann (68) this same nerve is said to receive general cutaneous
and communis fibres in approximately equal numbers, and
the nerve is said to be, in position, an ophthalmicus super-
ficialis trigemini, and not a profundus nerve. The two com-
ponents that form the nerve are said to arise, one from the
sensory trigeminus root, and the other from the communis
root of the facialis, the former thus being the probable homo-
logue of the portio ophthalmici profundi of my descriptions
of Amia, and the latter the homologue of the ophthalmicus
superficialis trigemini.

In Menidia Herrick (80, p. 428) says that the “r. oph. sup.
Vil and the r. oph. sup. V are fused throughout their entire
course, but each can be easily distinguished and separately
followed by difference in the calibre of the fibres.” The
ophthalmicus superficialis facialis is a lateral line nerve, and is
said to arise from the dorsal one of two lateral line roots of the
V—VII complex, a root which must accordingly be the homo-
logue of the dorso-median stem of Trigeminus II of Gorono-
witsch’s descriptions of Lota. No communis fibres are here
said to accompany those nerves.

In a later work (81) Herrick says that some fasciculus
communis fibres “probably run forward with the ophthal-

micus superficialis,”

while in his still later and complete work
(32) these communis fibres are said to be so completely

united with the ophthalmicus trigemini that it is impossible

MUSTELUS L&VIS. 219

to separate them. This trigeminal ophthalmic nerve, thus
formed, is accordingly quite probably the homologue of the
portio ophthalmici of Amia plus the ophthalmicus superficialis
trigemini of the same fish. “A very small r. profundus
V” is said by Herrick to be found in Menidia, being said to
be there represented by certain general cutaneous fibres that
accompany the radix ciliaris longa to the ciliary ganglion.
While I should certainly consider these general cutaneous
fibres as an integral part of the radix longa, and not as
representing a ramus ophthalmicus profundus, as will be later
more fully discussed, it is important here to notice that the
fibres are general cutaneous ones, this thus indirectly sup-
porting my conclusion that the portio ophthalmici profundi of
Amia, and that nerve alone, is the general cutaneous com-
ponent of the ophthalmic nerves of teleosts.

In Polypterus both van Wijhe (65) and Pollard (51)
describe a portio ophthalmici profundi, and also a ramus
ophthalmicus profundus that has the position of a selachian
profundus; but, in marked distinction with Amia and Lepi-
dosteus, no portio trigemini of the ophthalmicus superficialis
is definitely given by van Wijhe, and Pollard says that he
wholly failed to find that nerve. There is thus in this fish,
as in Acipenser, an apparent absence of a communis com-
ponent in the superficial ophthalmic nerve. In Acipenser
this apparent absence was accounted for under the assump-
tion that the dorsal one of the two so-called lateral sensory
roots of the fish, the Trigeminus IT dorsalis of Goronowitsch,
is a somewhat modified communis root, and that it is con-
cerned in the innervation of the sensory organs of the
nerve-sacs, those organs being derived from terminal buds.
What the apparent absence of this component in Polypterus
is due to is not evident from existing descriptions of that
fish.

Turning now to selachians, Haller (28) has recently very
thoroughly investigated the central origin and composition of
the cranial nerves in Scyllium. In this fish this author says
that Trigeminus IT is mainly, if not exclusively, sensory, and

220 EDWAKD PHELPS ALLIS, JUN.

that it arises by two roots, an upper one from the “ lobus n.
trigemini,” and a lower one from the ventral portion of the
so-called outer sensory ‘“ Oblongatagebiet” (p. 436). The
lobus trigemini is said by him to be simply an “ Abschnitt ”
of the outer sensory oblongatagebiet, and that part of the
latter “ gebiet” from which the ventral root of Trigeminus II
arises, is said to be the homologue of the dorso-lateral tract
of Goronowitsch (p. 423).

Kingsbury (40, p. 27), before the publication of Haller’s
work, was led to conclude that the lobus trigemini of Acipenser
was the probable homologue of the similarly named structure
in sharks, and, moreover, that further imvestigation would
probably prove this lobus to be a modified portion of the
acusticum system.

According to Haller no fibres from the inner sensory
oblongatagebeit (Lobi vagales) enter either of the trigeminal
roots of Scyllium, that fish thus agreeing with Acipenser in
the total or practical absence of communis fibres, as such, in
the so-called trigeminal roots.

The two roots of Trigeminus IT of Scylium thus seem to
be the exact homologues, in so far as the central regions from
which they take their origins are concerned, of the two roots
of the same nerve in Acipenser, and it may be assumed that
the several peripheral branches that arise from the two roots
in the two fishes have a similar distribution, though this
‘annot be definitely confirmed from Haller’s descriptions. A
bundle of fibres from each of the two roots of Scylhum is
said by Haller to be sent to the so-called second trigeminus
ganghon, from which the ramus ophthalmicus superficialis,
and that nerve alone, has its origin. The remaiming fibres of
the two roots are said to enter the Gasserian ganglion. From
the hind edge of the latter ganghon, and even partly as a
part of the dorsal root of Trigeminus II, four branches, said
to be lateral branches of the Trigeminus, are said to always
arise. The three distal ones of these four branches always
unite to form the ramus oticus, this last nerve being thus
formed of three separate branches, as it is in Mustelus. The

MUSTELUS LA&VIS. 921

other and most proximal branch is the one that has its ap-
parent origin partly from the dorsal root of Trigeminus II. It
runs backward, and joins the so-called ramus hyoideus facials,
thus doubtless forming the lateral and ampullary sensory
component of the post-trematic branch of the facialis. There
thus remains only the ramus buccalis to be accounted for of
all the lateral sensory branches of the facialis, and this nerve
is, unfortunately, not especially described, and not even
indicated in any of Haller’s figures. It may form part of the
so-called ramus maxillaris superior trigemini, and it would
seem as if it must receive fibres from both of the roots of
Trgeminus IJ. That the nerve exists as an important nerve
is evident from Marshall and Spencer’s descriptions of it im
embryos.

The ramus ophthalmicus superficialis of Haller’s descriptions
of Scyllium is thus formed of two components, a dorsal and a
ventral one, both of which arise from what are considered by
several authors as lateral sensory tracts of the brain. As in
Haller’s descriptions, there are but two ophthalmic nerves, a
ramus superficialis, and a ramus profundus, the natural
conclusion would be that the two components described by
him in the ramus superficialis must necessarily be the united
portio major and portio minor of the nerve. A comparison
with Schwalbe’s descriptions of this same fish (57, p. 187)
lead one, however, to strongly suspect that Haller has simply
repeated Balfour’s mistake, and given the name profundus to
what is in reality the portio minor of the superficial nerve.
Haller’s fig. 52 practically confirms this. The subject is then
still further complicated by Haller’s statement (p. 496) that
in Salmo the fibres corresponding to those that form the
‘amus profundus of Scyllium arise from the posterior instead
of from the anterior root of the trigeminus, thus belonging to
his Trigeminus IT instead of to Trigeminus I. Haller attempts
to explain this uncomfortable fact in the statement that “ bei
den Salmoiden und wohl auch anderen alteren Vertretern der
Teleostier eine Rostralwirtsverschiebung vom Boden des
chordalen Hirnes stattfand und dass damit auch ein grosser

222 EDWARD PHELPS ALLIS, JUN.

Theil des erster Trigeminus nach rostralwarts wanderte,
wahrend der R. ophthalmicus profundus, fixirt durch seinen
gesonderten Austritt aus dem Cranium, zuriickblieb und nun
nuit dem zweiten Trigeminus vom Centralorgan abgeht.” It
seems much more probable that Haller has made some mistake
in homologising his fibre tracts, in addition to wholly over-
looking the true ramus profundus.

Scylhum thus probably presents a portio major of the
‘amus ophthalmicus superficialis, formed by two bundles of
fibres derived, the one from the lobus trigemini, and the other
from the tuberculum acusticum, and a large portio minor of
the same nerve, derived from the anterior root of the
trigemino-facial complex. These two nerves issue from the
skull by separate foramina (Schwalbe), but soon unite to
form a single nerve. The portio minor is evidently the
homologue of the general sensory component of my descrip-
tions of Mustelus, and hence is what I am led to consider as
the portio ophthalmici profundi of the superficial nerve. It
is relatively very large in Scyllium, the true ramus profundus
being, according to Schwalbe’s descriptions, correspondingly
small. The two components of the portio major are, according
to my conclusions, destined, the one to innervate the sensory
organs of the lateral canals, and the other the sensory organs
of the ampulle. The former is the homologue of the portio
facials of the ophthalmic nerve of Amia, and the other the
homologue of the portio trigemini. Haller says that his
so-called profundus nerve is exclusively motor. This is too
manifestly an error to need discussion, but 1t well shows how
lable one is to error, in the present state of our knowledge
of the subject, if one limits one’s attention entirely to the
central origin of a nerve.

In Lemargus the ramus ophthalmicus superficialis trigemini
of Ewart’s descriptions is said by him to be found as a separate
slender nerve, which springs either from the trunk of the
trigeminus, or from its mandibular branch. Of it Ewart
says (18, p. 76) that it “neither innervates sensory nor
ampullary canals. It may, however, supply some of the

MUSTELUS LAVIS. 293

taste buds found in the roof of the mouth of certain fishes.
If these taste buds are modified lateral sense organs, the
nerves supplying them are likely to be made up of supra-
branchial fibres.” Kwart further says (p. 45) of this so-called
trigeminal ophthalmic nerve in elasmobranchs in general,
“Jn sharks and rays this nerve supplies the eyelids and the
skin over the anterior part of the cranium, but it also sends
fibres to the snout. More or less distinct in sharks, the super-
ficial of the ophthalmic of the trigeminal in rays consists of a
very few fibres which, on leaving the trigeminal, at once more
or less completely unite with the superficial ophthalmic of the
facial.” Kwart’s statement that branches of this trigeminal
herve may imnervate taste buds in the roof of the mouth is
clearly an error, and the nerve is, in all probability, the
homologue of the general sensory component of the ophthal-
mic nerve of Mustelus; that is, according to my conclusions,
the homologue of the portio profundi of the ophthalmicus
superficialis of Amia, and not of the portio trigemini of that
fish. In Lemargus there is, however, a branch of the oph-
thalmicus profundus called branch I by Ewart (17), which
also has the course of an ophthalmicus superficialis, the portio
profundi of the superficial ophthalmic nerve of this fish thus
apparently being represented by two separate branches.

In Acanthias Hoffmann says (85, p. 287) that the portio
minor s. trigemini of the ramus ophthalmicus supertficialis
is a branch of the ramus ophthalmicus profundus, and that
it is connected with the nervus trochlearis by a communicating
branch. He further says (p. 294), as I understand hin, that,
after a certain stage of development, that “ Stiick des Troch-
learis, welches den Verbindungsfaden mit dem Ophthalmicus
profundus bildet,” becomes the portio minor of the ophthalmic
nerve. On page 287 he says that the nervus profundus arises
as an independent nerve, wholly separate and distinct from
a single, large anlage from which the nervus trochlearis,
portio minor s. trigemini, ramus maxillo-mandibularis, and
communicating branch from the trochlearis to the trigemimi
all arise. In a later work (86, p. 357) he says that “der

224, EDWARD PHELPS ALLTS, JUN.

Trochlearis und die Portio trigemini rami ophthalmici super-
ficialis ein und derselbe Nerv ist.’ While it is evidently
unfair to Hoffmann to quote these several statements apart
from his general discussion of the subject, I must contess
to being unable to form a clear idea either of the relation-
ships of the nerves here concerned, or of their manner of
development. If the portio minor s. trigemini is a branch
of the ramus ophthalmicus profundus, it is evidently the
homologue of the portio ophthalmici profundi of Aimia, as
IT am seeking to establish, If it arises as a branch of a
nerve complex that is wholly separate and independent of
the nervus profundus, it would seem as if it could not be
the homologue of the nerve of Amia, and it may be that
serially homologous branches of two independent nerves—the
trigeminus and profundus—are here concerned, the two
branches being found separate and distinct in the two branches
of Lemargus just above referred to. The so-called thalamic
nerve of Hoffmann’s descriptions, the nerve marked X in
certain of his figures (84, fig. 38), may then be one of these
two branches.

In Squalus Acanthias Neal says (45, p. 233) that the fibres
that form the rami ophthalmicus superficialis and ophthal-
micus profundus trigemini arise from the posterior root or
portio major of the crigeminus. ‘This so-called posterior
trigeminus root (portio major) is simply a part of the anterior
trigeminus root of Haller’s nomenclature, and as the anterior
root (portio minor) of Neal’s descriptions is said by him to be
largely motor, and destined entirely to the mandibular arch,
the posterior root must be the spmal fifth component of the
nervus. The ophthalmicus superficial trigemini of this fish
is then evidently the homologue of the portio ophthalmici
profundi of Amia, and not of the portio trigemini of that
fish. Neal shows the nerve X of Hoffmann’s descriptions in
one of his figures (fig. K, p. 234), but I do not find that
he describes it.

As some slight further evidence of my conclusion that the
terminal-bud communis fibres that form the so-called ramus

MUSTELUS LAVIS. 225

ophthalmicus superficialis trigemini of Amia are represented
in Acipenser and selachians by those fibres of the ramus
ophthalmicus superficialis of the latter fishes that have their
origin in the lobus trigemini, it may be noted that Gorono-
witsch says (25, p. 8) that a few fibres of the dorsal or lobus
trigemini root of Trigeminus II of Acipenser can be traced
back into the lobi vagales. As no other communis fibres are
found, as such, in the ophthalmic nerves of the fish, are these
few fibres the remnant of an aborting nerve, or are they a
few fibres, still unchanged, of a nerve that is undergoing
modification ? My impression is decidedly in favour of the
latter one of these two suppositions. That communis fibres,
as such, once existed in these ophthalmic nerves, and that
some of them have actually undergone modification, must be
admitted if it be accepted that the sensory organs of the
lateral canals of all fishes pass through a stage represented
by the terminal bud (64, p. 298).

It should here be noted that Herrick concludes (82, p. 169)
that his own work, and Johnston’s also, favours the belief that
the acustico-lateral has been differentiated from the general
cutaneous system, and hence not from the communis system.

The ramus ophthalmicus profundus now remains to be
considered. This nerve is almost invariably said to be a
purely sensory one, excepting only in the Myxinoids, where
it is said to have motor branches. Haller’s assertion that the
profundus of Scyllium is entirely motor (28, p. 438) has
already been referred to; but apart from the fact that this
statement is probably an error, his profundus nerve is quite
unquestionably simply the portio minor of the ophthalmicus
superficialis, and has been already discussed.

In Chimera, Cole says (11, p. 645) that one of the branches
of the ophthalmicus profundus is possibly motor, and that its
origin and distribution make it exceedingly probable that it
corresponds to the motor division of the profundus found in
the Cyclostomata. The nerve thus referred to in Chimera
hes dorsal to the opticus. ‘The motor nerve with which it is
compared les ventral to the opticus in both Myxine and

226 EDWARD PHELPS ALLIS, JUN.

Petromyzon (20). In Bdellostoma the “ Stamm des Ophthal-
micus ” is said by Fiirbringer (p. 31) to run forward over the
opticus, and his reference to Miiller (a work I have not at my
disposal) leads one to suppose that this stem contains motor
fibres. ‘This statement, attributed to Miller, certainly deserves
to be controlled, and if found to be correct, the homologies of
the muscles so innervated well deserve to be carefully estab-
lished. That they can be the homologues of muscles that are
innervated in Myxine by a nerve that lies ventral to the
opticus I wholly doubt, and yet this is always assumed to
be the case. Pollard even cites this particular case in support
of his propositions (52, p. 397) that “The topographical posi-

2? ala

tion and course of nerves is not of great unportance ;
that “the fundamental grounds for determining the homology
of nerves are, (1) origin from homologous nerve-cells, (2)
terminal distribution to definite structures. The course of
the fibres is of less importance.” And so impressed is this
author with this very easy manner of accounting for the
apparently anomalous positions of certain nerves, that he
does not even attempt to show either that the cells of origin
or the structures innervated are homologous. ‘The reasoning
is that the structures innervated must be homologous, because
otherwise the nerves would not be. Much more rational
assumptions would be, either that there is some error in the
descriptions of Bdellostoma, or that the muscles innervated
are not homologous. As to Myxine, it would seem as if
certain motor and sensory fibres of the truncus maxillaris
trigemini were simply juxtaposited to the ramus ophthalmicus,
as that nerve issues from the skull, and that they later take
their proper course ventral to the opticus, leaving the true
ophthalmic fibres to course above that nerve. Such a juxta-
position is shown by Herrick (29) in Amblystoma, and is said
to occur in Cryptobranchus also. In these two cases there
is certainly no question of anything more than a simple
juxtaposition of the two nerves concerned, and I believe
that a similar explanation will be found to hold in all other
similar cases. Pollard himself even says (p. 400) of this

MUSTELUS LALIS. 997

motor nerve in Myxine, that although “it is usually con-
sidered to be an ophthalmic branch .... it is better to
consider it as a special branch, and not a portion of the
ophthalmic.” He then, by the name that he gives it—the pre-
maxillary nerve—definitely homologises it with the similarly
named nerve in his descriptions of Siluroids, in which fishes
it is said to be a branch of the maxillaris trigemini. He
cannot, however, rid himself of the idea that the nerve is the
homologue of the ophthalmic nerve of Bdellostoma.

In Chimeera, Cole says (p. 645), as has been already stated,
that two sense organs of the supra-orbital lateral canal are
innervated by a branch that has its apparent origin from the
ramus ophthalmicus profundus. In Mustelus I find one
lateral sensory branch so closely associated with the pro-
fundus that I could not definitely determine whether it fused
with that nerve, or later left it to join the main ophthalmicus
lateralis. A possible association of lateral sensory fibres with the
profundus nerve is thus here indicated, and if the profundus
nerve is actually thrown down from the same line of ectoderm
that later gives origin to the superficialis, as Platt states to
be the case in Necturus, it would certainly not be wholly
improbable that certain lateral fibres might be dragged down
with it, and so apparently form part of it. I, however, agree
most decidedly with Cole (p. 659) that the subject needs
further investigation. A totally different case is presented in
Platt’s statement that m Necturus four organs of the infra-
orbital line “are supphed by nerve twigs composed in equal
parts of fibres coming from the buccalis, and from the oph-
thalmicus profundus” (49, p. 530). That profundus fibres
could descend to the buecalis or bucealis fibres ascend to the
profundus is clearly impossible, the early developed nervus
opticus intervening. Platt’s explanation is contained in her
statement (p. 540) “that the attachment of the superficial
receptive cell to one fibre of transmission is not constant. A
shorter part when offered is at once accepted.” This, as
already stated, my work does not lead me to accept, and
Goronowitsch seems to hold a similar opinion, for he says (26,

228 EDWARD PHELPS ALLIS, JUN.

p. 48) that “ Die Verbindung zwischen dem Nerv und seinem
Endorgan ist . . . . in der Ontogenie eine Primiare.”
Herrick finds, in Menidia, what he considers as a “ramus
ophthalmicus profundus fused for its entire length with the
radix longa” (82, p. 209). The fibres that represent the
ramus profundus, thus identified, are said to be general
cutaneous ones. As all the other fibres that enter the radix
longa of the fish are said to be sympathetic ones, it seems as
if Herrick must arrive at his conclusion by the assumption
that the term radix longa should strictly be applied only to a
sympathetic nerve. This seems to me an error, for Thane
says (61, p. 238) that the “long or sensory root” of the
ciliary ganglion of man arises from the nasal branch of the
ophthalmic trunk, and is wholly separate from the “ middle
or sympathetic root” of the ganglion. The so-called long
root thus here contains no sympathetic fibres. Moreover,
Schwalbe says (57) that, in the dog, the ramus naso-ciliaris
sends a radix longa to the ciliary ganglion, and that ganglion
is, according to Holtzmann (87), partly spinal and partly
sympathetic in character. ‘he radix longa of the animal
must, accordingly, very probably contain general cutaneous
as well as sympathetic fibres, and yet there is a distinct and
separate ramus naso-ciliaris, the homologue of the ramus pro-
fundus of fishes. I should accordingly look upon the general
cutaneous fibres of the radix longa of Menidia as an integral
part of that root, and not as a remnant of the ramus pro-
fundus, that ramus being wholly wanting, as it is in other
teleosts, so far as known. This is,
shown to be the case, by the arrangement found in Scomber,

moreover, practically

where there is (7) a radix longa, which arises from a separate
and independent profundus ganglion, and is later joined by
sympathetic strands from a large sympathetic ganglion asso-
ciated with the trigeminal ganglion.

SUMMARY.

There are, in fishes, several ophthalmic nerves between
which it is necessary to carefully distinguish. While their

MUSTELUS LAVIS. 229

exact inter-homologies cannot as yet be definitely affirmed,
certain very probable homologies can certainly be arrived at.

The portio ophthalmici profundi of bony ganoids, and the
ramus ophthalmicus profundus and ramus ophthalmicus
superficialis trigemimi of elasmobranchs, are general sensory
cutaneous nerves, and probably contain all of those general
sensory elements that belong to the ophthalmic nerves of
vertebrates.

The portio ophthalmici profundi of ganoids is the homo-
logue either of the ramus ophthalmicus superficialis trigemini
of elasmobranchs, or of a frontal branch of the ramus pro-
fundus that has, in certain selachians (Lamargus), the posi-
tion of a superficial ophthalmic nerve ; or it is the homologue
of both those nerves of elasmobranchs. In the higher
animals this portio profundi becomes the frontal branch of
the ophthalmic nerve.

The ramus ophthalmicus profundus of elasmobranchs and
of Polypterus, and that nerve alone, is the homologue of the
ramus nasalis or naso-ciliaris of higher animals. This nerve
and the portio ophthalmici profundi vary inversely in 1m-
portance.

The ramus ophthalmicus superficialis trigemini of Aimia is
a communis nerve, and is probably the homologue of that
part of the ophthalmicus superficialis facialis of Acipenser
and elasmobranchs that is derived from the so-called dorsal
root of Trigeminus II. This latter root arises from the lobus
trigemini of the latter fishes, is considered by certain authors
as a lateral sensory root, and is called by them the dorsal
root of the ramus opthalmicus superficialis facialis.

The ramus ophthalmicus superficialis facialis of Amia is a
lateral sensory nerve, and has its probable homologue in that
part only of the ophthalmicus superticialis facialis of Acipenser
and elasmobranchs that is derived from the so-called ventral
root of Trigeminus II. This root arises, in all fishes, from the
tuberculum acusticum.

The so-called ramus ophthalmicus superficialis trigemini of
teleosts 1s the ramus ophthalmicus superficialis trigemini of

230 EDWARD PHELPS ALLIS, JUN.

Amia plus the portio ophthalmici profundi of that fish, these
two components being found in varying proportions.

The ramus ophthalmicus superficialis of selachians probably
always contains the same three components, or portiones, that
are found in the superficial ophthalmic nerve of Amia, but
the portio ophthalmici profundi, called in selachians the
portio trigemini, is usually small and may be reduced to a few
fibres only (Rays, Ewart). The portio trigemini of Amia is
represented in selachians, as stated above, by those fibres of
the so-called portio facialis that arise from the lobus trigemini ;
that is, by the fibres that form the so-called dorsal root of
Trigeminus IT.

The terminal buds of ganoids and teleosts, the nerve-sacs
of Acipenser, and the ampulle of selachians, are in all proba-
bility homologous structures.

LITERATURE.

1. Aucocx, R.—‘ The Peripheral Distribution of the Cranial Nerves of
Ammoccetes. I. The Branchial Nerves, and the Innervation of the
Lateral Line System,’’ ‘Journ. of Anat. and Phys.,’ vol. xxxiil;
N.S., vol. xiii, pt. i, pp. 131—153, October, 1898.

2. Auuis, Epwarp Pueces, jun.—‘ The Anatomy and Development of the
Lateral Line System in Amia ecalva,” ‘Journ. of Morph.,’ vol. ii,
No. 3, April, 1889.

3. Atiis, Epwarp PHeLps, jun.— The Cranial Muscles and Cranial and
First Spinal Nerves in Amia calva,” ‘Journ. of Morph.,’ vol. xii,
No. 3, March, 1897.

4. Auuis, Epwarp Puetps, jun.—‘‘A Reply to certain of Cole’s Criti-
cisms of my workon Amia calva,” ‘ Anat. Anz.,’ Bd. xv, Nos. 19, 20,
pp. 364—379, February 24th, 1899.

5. Atuis, Epwarp Puexrs, jun.—‘‘ An abnormal Musculus obliquus
superior in Carcharias,” ‘ Anat. Anz.,’ Bd. xvi, No. 24, December 19th,
1899.

6. Auuis, Epwarp Pue ps, jun.—* The Lateral Sensory Canals of Poly p-
terus bichir,” ‘ Anat. Anz.,’ Bd. xvii, No. 23, June 21st, 1900.

7. Auuis, Epwarp Puetps, jun.—‘ The Skull and the Cranial and First
Spinal Muscles and Nerves of Secomber scomber, (In press.)

12.

13.

14.

15.

16.

Fie

18.

19.

20.

21.

22.

MUSTELUS LAVIS. 231

. Batrour, F. M.—‘ Handbuch der vergleichenden Embryologie,’

uebersetzt von Dr. B. Vetter, 1880.

. Broman, Ivar.—‘ Die Entwickelingsgeschichte der Gehorknéchelchen

beim Menschen,’ Anat. Hefte,’ Heft xxxvil, pp. 507—670, 1899.

. Crave, Cornetia M.—“ The Lateral Line System of Batrachus tau,”

‘Journ. of Morph.,’ vol. xv, No. 2, November, 1898.

. Cotz, F. J.—“On the Cranial Nerves of Chimera monstrosa

(Linn.); with a Discussion of the Lateral Line System, and of the
Morphology of the Chorda tympani,” ‘Trans. Roy. Soc. Edinburgh,’
vol. xxxviil, pt. 8, No. 19, 1896.

Coz, F. J.—“ Observations on the Structure and Morphology of tke
Cranial Nerves and Lateral Sense Organs of Fishes, with special
reference to the genus Gadus,” ‘Trans. Linn. Soe. Londoii,’ vol. vii,
pt. v, October, 1898.

Cottince, W. E.—* The Sensory Canal System of Fishes. Part I,
Ganoidei,” ‘ Quart. Journ. of Mier. Science,’ N. S., No. 144, vol. xxxvi,
pt. 4, August, 1894.

Cornine, H. K.— Ueber die vergleichende Anatomie der Augenmusku-
latur,” ‘ Morph. Jahrb.,’? Bd. xxix, Heft 1, October 23rd, 1900.

Dixon, A. Francis.—‘‘ On the Development of the Branches of the Fifth
Cranial Nerve in Man,” ‘ Sci. Trans. Roy. Dub. Soc.,’ vol. vi (series ii),
pp. 19—76, May, 1896.

Dixon, A. Francis.—“ The Sensory Distribution of the I’acial Nerve in
Man,” ‘ Journ. Anat. and Phys.,’ vol. xxxiii, pt. 3, April, 1899.

Ewart, J. C.—“On the Cranial Nerves of Elasmobranch Fishes,”
‘Prelim. com. Proc. Roy. Soc.,’ vol. xiv, March 7th, 1889.

Ewart, J. C.—“‘ The Lateral Sense Organs of Elasmobranchs. I. The
Sensory Canals of Lemargus,’ ‘Trans. Roy. Soc. Edinburgh,’
vol. xxxvii, pt. 1, Nos. 5, 6, 1891.

Ewart, J. C., and Mitcnetz, J. C.—* The Lateral Sense Organs of
Elasmobranchs. II. The Sensory Canals of the Common Skate
(Raia batis),” ‘Trans. Roy. Soc. Edinburgh,’ vol. xxxvil, pt. 1,
Nos. 5, 6, 1891.

Firserncer, Paut.—* Untersuchungen zur vergleichenden Anatomie
der Muskulatur des Kopfskelets der Cyclostomen,” ‘ Jen. Zeitschr. f.
Naturw.,’ Bd. ix, Heft 1, pp. 1—98, January 30th, 1875.

Garman, Samunt.— On the Lateral Canal System of the Selachia and
Holocephala,” ‘ Bull. Mus. Comp. Zool. Harvard College,’ vol. xvii,
No. 2, September, 1888.

Gaupr, E.—* Ueber das Primordialcranium von Lacerta agilis,” ‘ Anat.
Anz. Erganzhfr. z.,’ Bd. xiv, 1898.

VOL. 45, PART 2,—NEW SERIES. R

232 EDWARD PHELPS ALLIS, JUN.

23. GrcENBAUR, Cart.—‘ Untersuchungen zur vergleichenden Anatomie
der Wirbelthiere, Heft 3, “Das Kopfskelet der Selachier: ein
Beitrag zur Erkenntniss der Genese des Kopfskeletes der Wirhel-
thiere,” Leipzig, 1872.

24. GoronowitscH, N.—“ Das Gehirn und die Cranialnerven von Acipenser
ruthenus: ein Beitrag zur Morphologie des Wirbelthierkopfes,”
‘Morph. Jahrb.,’ Bd. xiii.

25. GoronowitscH, N.—‘ Das Trigemino-facialis-Complex von Lota
vulgaris,” ‘ Festschr. f. Carl Gegenbaur,’ Bd. ili, 1896.

26. GoronowitscH, N.—‘‘ Untersuchungen tiber die erste Anlage der
Kranialuerven bei Salmo fario,” ‘ Nouveaux Mém. dela Soc. Imp. des
Naturalistes de Moscou,’ vol. xvi, liv. 1, pp. 1—54, 1898.

27. Green, H. A.—“On the Homologies of the Chorda Tympani in
Selachians,” ‘Journ. of Comp. Neur.,’ vol. x, No. 14, December, 1900.

28. Hatter, B.—“ Vom Bau des Wirbelthiergehirns.” Theil I, ‘‘ Salmo
und Scyllium,” ‘ Morph. Jahrb.,’ Bd. xxvi, Hefte 3, 4, December 2nd,
1898.

29. Herrick, C. J—“The Cranial Nerves of Amblystoma puncta-
tum,” ‘Journ. Comp. Neur.,’ vol. iv, December, 1894.

30. Herrick, C. J.—‘ The Cranial Nerve Components of Teleosts,” ‘ Anat.
Anz.,’ Bd. xiii, No. 16, pp. 425—431, May 22nd, 1897.

31. Herrick, C. J.—‘‘'The Cranial Nerves of the Bony Fishes,” ‘Journ,
Comp. Neur.,’ vol. viii, No. 3, August Ist, 1898.

32. Herrick, C. J.—‘ The Cranial and First Spinal Nerves of Menidia: a
Contribution upon the Nerve Components of the Bony Fishes,’
Dissertation, State Hospitals Press, Utica, N.Y., 1899.

33. Herrick, C. J.—‘A Contribution upon the Cranial Nerves of the
Codfish,” ‘Journ. Comp. Neur.,’ vol. x, No. 3, October, 1900.

34. Horrmann, C. K.—‘‘ Beitrage zur Entwickelungsgeschichte der
Selachii,” ‘Morph. Jahrb.,’ Bd. xxiv, Heft 2, pp. 209—286, August
18th, 1896.

35. Horrmann, C. K.—lIbid., Bd. xxv, Heft 2, pp. 250—3804, July 2nd,
1897.

36. Horrmany, C. K.—Ibid., Bd. xxvii, Heft 3, pp. 325—414, July 7th,
1899.

37. Hottzmann, H.—“ Untersuchungen itiber Ciliarganglion und Ciliar-
nerven,” ‘ Morph. Arb.,’ Bd. vi, Heft 1, pp. 114—140, 1896.

38. Howes, G. B., and Swinnerton, H. H.—‘‘ On the Development of the
Skeleton of the Tuatara, Sphenodon punctatus; with Remarks
on the Egg, on the Hatching, and on the Hatched Young,” ‘ Trans,
Zool, Soc. of London,’ vol. xvi, pt. 1, February, 1901,

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

MUSTELUS LA&AVIS. 233

Jounstoy, J. B.—“Hind Brain and Cranial Nerves of Acipenser,”
‘Anat. Anz.,’ Bd. xiv, Nos. 22, 23, pp. 580—602, July 11th, 1898.
Kryeszury, B. F.—* The Structure and Morphology of the Oblongata

in Fishes,” ‘Journ. Comp. Neurol.,’ vol. vii, No. 1, April, 1897.
MarsnatL, A. Mitnes.—‘‘ The Development of the Cranial Nerves in

the Chick,” ‘Journ. Micr. Sci.,’ vol. xviii, N. S., pp. 10—40.
Marsuat., A. MiItnes.—‘‘ On the Head Cavities aud Associated Nerves

of Elasmobranchs,” ‘ Journ. Mier. Sci.,’ vol. xxi, N. S., pp. 72—97.
Marsuatt, A. Mines, and Spencer, W. Batpwin.—“ Observations on

the Cranial Nerves of Scyllium,” ‘Studies from the Biol. Lab. of the
Owens College,’ vol. i, 1886.

Mrrropnanow, Paut. “Etude embryogénique sur les Sélachiens,”
‘Arch. de Zool. Exp. et Gén.,’ 3e sér., vol. i, pp. 161—220, 1893.
Neat, H. V.—‘ The Segmentation of the Nervous System in Squalus

acanthias,” ‘Bull. Mus. Comp. Zool. Harvard College,’ vol. xxxi,
No. 7, pp. 148—292, May, 1898.

Onop1, A.—“ Das Ganglionciliare,”’ ‘ Anat. Anz.,’ Bd. xix, Nos. 5, 6,
March 6th, 1901.

Parker, T. JEFFERY.—“On the Blood-vessels of Mustelus an-
tarecticus: a Contribution to the Morphology of the Vascular
System in the Vertebrata,” ‘ Phil. Trans.,’ 1886.

Pratt, J. B—*‘Ontogenetische Differenzirung des Ektoderms in
Necturus,” Studie I, ‘Arch. f. mikr. Anat.,’ Bd. xliii, Heft, 4, pp. 911
—966, June 30th, 1894.

Pratt, J. B.—‘Ontogenetic Differentiations of the Ectoderm in
Necturus.”” Study II, “On the Development of the Peripheral
Nervous System,” ‘Quart. Journ. Mier. Sci.,’ No. 152, pp. 485—547,
February, 1896.

PuessEN, Jos. von, and Rasinovicz, J.—‘ Die Kopfnerven von
Salamandra maculata in vorgeriickten Hmbryonalstadium,’
Minchen, J. F. Lehmann, 1891.

PonuarD, H. B.—“‘On the Anatomy and Phylogenetic Position of
Polypterus,” ‘ Zool. Jahrb.,’ Bd. v, Hefte 3, 4, October 20th, 1892.

Pouttarp, H. B.—“ The Oral Cirri of Siluroids and the Origin of the
Head in Vertebrates,” ‘Zool. Jahrb.,’ Bd. viii, Heft 3, pp. 379—493,
May 11th, 1895.

Rex, H.—“ Zur Entwicklung der Augenmuskeln der Ente,” ‘ Arch. f.
mikr. Anat.,’ Bd. lvii, Heft 2, January 7th, 1901.

Ripewoop, W. G.—‘On the Spiracle and Associated Structures in
Elasmobranch Fishes,” ‘ Anat. Anz.,’ Bd. 2, No. 14, pp. 425—433,
December, 1895.

934 EDWARD PHELPS ALLIS, JUN.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64,

65.

66.

67.

68.

69.

70.

Rucer, Grore.—‘‘ Ueber das peripherische Gebiet. des Nervus Facialis
bei Wirbelthieren,” ‘ Festschr. f. Carl Gegenbaur,’ Leipzig, 1896.

SAGEMEHL, M.—“ Beitrige zur vergleichenden Anatomie der Fische.”
I, “Das Cranium von Amia calva, L.,” ‘Morph, Jahrb.,’ Bd. ix,
Heft 2, 1883.

ScuwaLBe, G.—“ Das Ganglion oculomotorii: ein Beitrag zur vergleich-
enden Anatomie der Kopfnerven,” ‘Jen. Zeitschr. f. Naturw.,’
Bd. xiii, Heft 2, July 16th, 1879.

Sewertzorr, A. N.—‘Studien zur Entwickelungsgeschichte des
Wirbeithierkopfes.” J, ‘ Die Metamerie des Kopfes des electrischen
Rochen,’ Moscow, 1899.

Stannius, H.—‘ Das peripherische Nervensystem der Fische: anato-
misch und physiologisch Untersucht,’ Rostock, 1849.

Strone, Otiver S.—‘ The Cranial Nerves of Amphibia: a Contribu-
tion to the Morphology of the Vertebrate Nervous System,” ‘ Journ.
of Morph.,’ vol. x, No. 1, January, 1895.

Tuave, G. D.—‘Quain’s Elements of Anatomy,’ vol. iii, pt. 2, “The
Nerves,” London, 1895.

Tiesinc, BertHoLp.—‘* Kin Beitrag zur Kenntnis der Augen-Kiefer und
Kiemenmusculatur der Haie und Rochen,” ‘Jen. Zeitschr. f. Naturw.,’
Bd. xxx, Heft 1, October 18th, 1895.

Verrer, Brnsamin.—‘ Untersuchungen zur vergleichenden Anatomie
der Kiemen- und Kiefermusculatur der Fische,” ‘Jen. Zeitschr. f.
Naturw.,’ Bd. vili, Heft 3, September, 1874.

Winpersuem, R.—‘Grundriss der vergleichenden Anatomie der
Wirbelthiere,’ Jena, 1893.

Van Wisur, J. W.—< Ueber das Visceralskelett und die Nerven des
Kopfes der Ganoiden und von Ceratodus,” ‘ Niederl. Archiv f. Zool.,
Bd. v, Heft 3, July, 1882.

Van WisueE, J. W.—‘ Ueber die Mesodermsegmente und die Entwick-
lung der Nerven des Selachierkopfes,’ Amsterdam, 1882.

Witson, Henry V.—“ The Embryology of the Sea Bass (Serranus
atrarius),” ‘ Bull. U.S. Fish Com.,’ vol. ix, 1891.

Workman, J. S.—* The Ophthalmic and Eye Muscle Nerves of the
Cat-fish (Amiurus),” ‘Journ. of Comp. Neur.,’ vol. x, No. 4,
December, 1900.

Wricut, R. Ramsay.—‘‘On the Nervous System and Sense Organs of
Amiurus,” ‘ Proe. Can. Inst.,’ vol. i, fase. 8, Toronto, 1884,

Wrieut, R. Ramsay.—‘On the Hyomandibular Clefts and Pseudo-

branchs of Lepidosteus and Amia,” ‘Toronto Journ. of Anat. and
Phys.,’ vol, xix, July, 1885,

MUSTELUS La&VisS. 935

DESCRIPTION OF PLATES 10—12,

Illustrating the paper by Mr. Edward Phelps Allis, jun.,
on “The Lateral Sensory Canals, the Eye Muscles, and
the Peripheral Distribution of the Cranial Nerves of
Mustelus levis.”

Fig. 1.—Side view of the head of a 10 cm. embryo of Mustelus levis,
the skin being removed so as to expose the lateral and ampullary canals. The
surface pores of the lateral canals are not shown at all; and very few of the
tubules of the canals, excepting those that project to one side or the other,
are even indicated. x 9.

Fic. 2.—Top view of the same. xX 9.

Fic. 3.—Bottom view of the same. x 9.

Fig. 4.—Side view of a deeper dissection of the same head. ‘The lateral
and ampullary canals, the eyeball, and the eye muscles have been removed,
the other muscles being all left in place. Part of the ampulla have been
removed, others being left in place. xX 9.

Fic. 5.—Bottom view of the same. x 9.

Fig. 6.—Side view of the skull of a 10 em. embryo of Mustelus levis.
x 5.

Fie. 7.—The suborbital section of the infra-orbital canal of a 10 em.
Mustelus levis, showing the tubules and pores of the canal. x 12.

Fie. 8.—Part of a transverse section of a 12:2 em. embryo of Mustelus
levis, showing the adductor mandibule and Addf muscles, the mandibular
cartilage, and the anterior upper labial cartilage.

Fic. 9.—Part of a more posterior section of the same embryo, showing also
the anterior end of the posterior upper labial cartilage, with a part of the
tendon of the muscle Add@ inserted on it.

Index Letters.

acafr. Koramen of anterior carotid artery. acz/r. Foramen of anterior cere-
bral vein. dddB. Muscle AddB of Vetter’s descriptions. dm. Adductor
mandibule. éa. Buccal group of ampulle or their pores and tubes. da..-°
Five sub-groups of buccal ampullary pores and tubes. 4/. Ramus buccealis
facialis. .ma. md. Truncus buccalis-maxillo-mandibularis. c¢. Canalis trans-
versus. Csd,. Muscle Csd, of Vetter’s descriptions. Csv,. Muscle CUsv, of

236 EDWARD PHELPS ALLIS, JUN.

Vetter’s descriptions. exd. Endclymphatic pore. es. Hyestalk. eye. Hye.
ge. Ciliary ganglion. g/. Nervus glossopharyngeus or its foramen. 4/.
Ramus hyoideus facialis. Ame. Hyomandibular lateral canal. HIMZD. Hyo-
mandibular cartilage. Amf. Truncus hyoideo-mandibularis facialis. Jm.
Intermandibularis. ioc. infra-orbital Jateral canal. 7/c. Lateral line canal.
M. Meckel’s cartilage. m. Mouth. mda. Mandibular group of ampulle or
their tubes and pores. mde. Mandibular lateral canal. md¢. Ramus man-
dibularis trigemini. mef. Ramus mandibularis externus facialis. m/e. Man-
dibular labial cartilage. ma¢. Ramus maxillaris trigemini.. 2a, Nasal aperture.
nc. Nasal capsule. o. Nervus opticus or its foramen. oem. Nervus oculomo-
torius or its foramen. ocmz. Inferior branch of nervus oculomotorius. xe.
Orbito-nasal canal. opad. Deep group of ophthalmic ampulla. opas. Super-
ficial group of ophthalmic ampulla. opas'-. Three sub-groups of pores and
tubes of superficial ophthalmic ampulle. opp. Ramus ophthalmicus profundus
or its foramen, oppe. Ophthalmicus profundus canal. Ops. Ramus ophthal-
micus superficialis or its foramen. Os. Obliquus superior. o/f. Ramus oticus
facialis. peafr. Foramen of posterior carotid artery. pf. Ramus palatinus
facialis. PQ. Palato-quadrate cartilage. soc. supra-orbital lateral canal.
spr. Spiracle. sso. Surface sense-orgaus. s¢c. Supra-temporal cross com-
missure. ¢dddB. Tendon of muscle AddB. /fr. Trigemino-facial foramen.
tr. Nervus trochlearis or its foramen. «lca. Anterior upper labial cartilage.
ulep. Posterior upper labial cartilage. «. Foramen which probably transmits
the general cutaneous component of the ramus ophthalmicus superficialis.

\\

0
=—=eSte
80

She ae

ea = Se

3 1 me) Ce
| s-=- = Ope"
' end.

=

>
. 2.2 2 293955 555555

~

West, Newman chromo.

Quart Fourn, Mor 5.00.45, HS GL

LAN-- ------- toc

Y
y

West Newman chromo

FRA PSsSN TTS

=
—— .

\ \----uoc

\
\

A= === = TUL
\

;

West, Newman: chrom

THE ANATOMY OF SCALIBREGMA INFLATUM.

237

The Anatomy of Scalibregma inflatum,

Rathke.
By
J. H. Ashworth, D.Sc.,

Lecturer on Invertebrate Zoology in the University of Edinburgh.

With. Plates 183—15.

ConTENTS.

1. Introduction

2. Historical Account .

3. Distribution and Habits :

4. External Characters : Segmentation, ees Sie Golany

5. Parapodia: Cirri, Glands of Cirri

6. Sete

7. Skin

8. Musculature

9. General Anatomy of the ieee tens
10. Coelom ;
11. Alimentary Canal
12. Vascular System
13. Gills : : , : ; ;
14. Central Nervous Syxtenn Brain, @sophageal Connectives, Nerve-

cord : :
15. Sense-organs: Prostomial Bpitheium, Nuchal Great Cit Sete .
16, Lateral Sense-organs
Occurrence of Lateral Sense-organs in icles Epica a
Morphology of Lateral Sense-organs of Polycheta

17. Nephridia : : ; 5 :
18. Reproductive Organs

19. The Family Sealibregmide .
20. Affinities of the Scalibregmide

21. Summary of Results

22.

Literature .

PAGE
238
239
240
242
246
249
252
253
254
256
256
259
262

263
269
270
274
276
280
284
286
297
300
303

238 J. H. ASHWORTH.-

1. InrRopucTIon.

My attention was drawn to our scanty knowledge of
Scalibregma when discussing the affinities of Arenicola
with other Polychetes (1900,! p. 544). It was found impos-
sible to make any definite statement regarding the affinities
of these two genera, owing to the small amount of informa-
tion available regarding the structure of Scalibregma.
This Polycheete has received little attention from zoologists,
perhaps on account of its comparative rarity and the some-
what small size of the majority of specimens im museum
collections. Most of the references to this animal in zoolo-
gical literature are mere records of its capture. ‘There are
only three or four memoirs which refer, quite briefly, to some
details of its structure, and only one, by Danielssen (1859,
p- 69), which contains a connected account of its internal
anatomy. ‘This memoir contains no mention of the nephridia,
although the accompanying figures show structures which
are obviously nephridia, but which Danielssen considered to
be ovaries. He also stated that Scalibregma is hermaph-
rodite, and that its nerve-cord is provided with typical
ganglionic swellings. Subsequent authors do not throw hght
on any of these matters. In these points there was such
marked difference between Scalibregma and Arenicola,
which in other respects seemed to have much in common,
that I was anxious to reinvestigate the anatomy of the
former as soon as specimens were available, chiefly with the
intention of comparing the nephridia, gonads, and nerve-cord
of these two genera. The material placed at my disposal has
enabled me to study the anatomy and histology of this
northern Annelid, and to determine some interesting points
connected with most of the systems of organs.

IT am grateful to the authorities of the United States
National Museum in Washington for the loan of a number
of specimens collected on the east coast of America, and to

1 The dates in parentheses form references to the literature quoted at the
end of this paper.

THE ANATOMY OF SCALIBREGMA INFLATUM. 239

Dr. Théel, of Stockholm, and Dr. Appelléf, of Bergen, for the
gift of several excellent specimens from the coast of Norway.

This work has been done in the Beyer Zoological Labora-
tories of the Owens College, Manchester, and in the Zoological
Laboratory of the University of Edinburgh.

2. Historicat Account.

Rathke (1843, p. 182) founded the genus and species
Scalibregma inflatum upon specimens obtained at Molde,
in Norway. He described the external characters in con-
siderable detail, directing attention to the form of the para-
podia in different regions of the body, and to the brown or
black structures upon them.

Three years later M. Sars (1846, p. 91) was fortunate in
securing a very large specimen (58°5 mm. long), which he
described under the name Oligobranchus roseus. He
has given a good account of most of the external characters
of the animal, but overlooked the black structures on the
parapodia. He considered this animal was allied to the
newly described genus Humenia, Oersted,*’and he also men-
tioned its general affinity with the Ariciide and the Areni-
colidee.

Danielssen (1859, p. 69) has given the only account of the
internal anatomy of Scalibregma. The form of the ali-
mentary canal, the circulatory system, the nervous system,
the paired segmentally arranged organs—interpreted by him
as ovaries,—and the structures he mistook for testes are
described in considerable detail and illustrated by clear
figures.

In 1873 Verrill (1873, p. 605) described the external
features of a new species, S. brevicauda, which had been
obtained off Newhaven, Connecticut, U.S.A., and Hansen
(1882, p. 34) found among the material of the North Atlantic
Expedition specimens which he referred to new species, S. (?)
abyssorum and 8. (?) parvum. Wirén (1887) made
scattered references to some points in the structure of the
alimentary canal, and the arrangement of the muscles and

240 J. H. ASHWORTH.

the four anterior diaphragms. The other references to
Scalibregma in zoological literature are mostly mere
records of its capture, chiefly in Norwegian waters.

38. DISTRIBUTION AND Hapsirts.

Scalibregma inflatum is recorded chiefly from the
North Atlantic and Arctic Oceans; but it is not restricted to
these northern seas. The “ Challenger” (see McIntosh, 1885,
p. 359) captured this species at two stations in southern seas,
viz. at station 141, between the Cape of Good Hope and
Marion Island, where numerous specimens were dredged
from a depth of 98 fathoms, and at station 169, off the east
coast of North Island of New Zealand, where a single speci-
men was obtained from a depth of 700 fathoms.

It is interesting to find that these southern specimens
correspond very closely with those obtained from European
seas. McIntosh states that the southern specimens are some-
what smaller than Kuropean examples, the largest one taken
by the ‘‘ Challenger” being 18 mm. long. ‘This is not a
character of any importance, as the size of Scalibregma
varies between wide limits. Most of the northern specimens
are little, if any, larger than those taken by the ‘ Challenger.”
Of eleven specimens sent to me from Bergen, six are be-
tween 12 and 15 mm. long, two are incomplete but would
probably be about 13 and 20 mm. long ; the other three are
26, 35, and 56 mm. long respectively, while eight of the ten
complete specimens from the east coast of the United States
are between 5 and 9 mm. long. McIntosh remarks that the
cills of southern forms are smaller than those of Norwegian
examples; but this, again, is a very variable character, de-
pending on the age and size of the specimen. We may
conclude, therefore, that the specimens of Scalibregma
obtained by the ‘ Challenger” are not distinguishable by
any essential and constant character from those taken in the
North Atlantic.

S.inflatum occurs in the Arctic Ocean as far eastward as

THE ANATOMY OF SCALIBREGMA INFLATUMN. 241

Cape Grebeni (the southern point of Waigatsch Island) and
the Sea of Kara (Théel, 1879, p. 51). It is found off the
western shores of Spitzbergen (Malmgren, 1867, p. 77; von
Marenzeller!), Nova Zembla (von Marenzeller,? Théel, 1879,
p- 51), and along the western coast of Norway as far south-
wards as the island of Floroé (Rathke, Sars, Danielssen,
Malmgren, Appellof?). Scalibregma is also recorded from
the south-western coast of Sweden. Sars! found several small
examples in Christiania Fjord, and Malm® soon afterwards
obtained specimens near Goteborg.

Scalibregma occurs on the north, east, and west coasts of
Scotland, being recorded from the Shetlands,® from St.
Andrews,’ and from Loch Maddy in the Hebrides by
McIntosh, and from near Millport in the Firth of Clyde
by Koélliker.* The last named is the most southerly European
station from which Scalibregma has been obtained.

On the western side of the Atlantic this Polychete has been
taken off the western shores of Greenland (McIntosh ®), and
at several stations off the eastern coast of North America,
between George’s Bank, off Nova Scotia, and Newhaven,

1 “Spitzbergische Anneliden,” ‘Archiv fiir Naturgeschichte,’ 55 Jahrg.,
p. 129, 1889.

2 «Die Coelenteraten, Echinodermen, und Wiirmer der K. K., Osterreich-
isch-Ungarischen Nordpol Expedition,” ‘ Denkschriften der Matem. Naturw.
Classe der Kaiserl. Akad. der Wissenschaften,’ xxxv, 1877.

3 « Faunistiske Undersogelser i Herlofjorden,” ‘Bergens Museums Aares-
beretning,’ p. 10, No. 11, 1894-5.

* *Bidrag til Kundskaben om Christianiafjordens Fauna,’ III, p. 46.
Christiania, 1873.

* *Zoologiska Observationer,’ p. 88, Kongl. Vet. o Vitt. Samballets i Gote-
borg Handlingar. Gdteborg, 1874.

° “Report on Annelida dredged off Shetland Islands by Mr. Gwynn
Jeffreys,” ‘ British Assoc. Report for 1868, p. 836. London, 1869.

7 “On the Annelids of St. Andrews,” ‘ British Assoc. Report for 1867,’
p. 92. London, 1868.

8 *Wurzbirger Naturwiss. Zeitschr.,’ p. 243. 1864.

® “Annelida obtained during Cruise of H.M.S. ‘ Valorous’ to Davis
Straits,’ ‘Trans. Linnean Soc., Zoology,’ second series, vol. i, p. 506.
London, 1879.

949 j. H. ASHWORTH.

Conn. (Verrill1). The specimens sent to me by the Smith-
sonian Institution were taken at four stations off this coast,
the most northerly one being off Nova Scotia, and the most
southerly in latitude 40° N.

Although Scalibregma is found in some places in con-
siderable numbers in shallow water (as in some of the fjords
of Nordland (see Danielssen, 1859, p. 25), it is more usually
obtained by dredging, and sometimes from considerable
depths. Those from the Smithsonian Institution were all
obtained at depths varying from 43 to 99°5 fathoms, and
the ‘“‘ Challenger ’’ specimens were dredged from 98 and 700
fathoms respectively.

Scalibregma burrows in sand, which is often more or less
intermixed with mud or clay, to a depth of one or two feet,
forming long passages which in some places, as in the fjords
of Nordland, are accessible at low water (Danielssen, p. 25).
In its mode of life it evidently strongly resembles the common
lugworm (Arenicola marina) of our coasts.

4, ExrernaL Cuaracters (13 Pl.).

The general aspect of Scalibregma inflatum may be
described as arenicoliform, but its shape varies considerably
in different individuals. The anterior end of the animal is
short, and resembles a truncated cone (fig. 1). The following
region of the body is inflated to a greater or less degree, the
inflation extending sometimes over only four or five seg-
ments, but more generally comprising about ten segments.
‘he swolleu portion is either globular or more or less cylin-
drical. In many cases, especially in the smaller specimens,
the body swells out abruptly about the fifth or sixth segment,
decreasing in diameter almost as suddenly at the end of the
inflated portion; but in most of the larger specimens there
is a much more gradual transition from the inflated portion
to the regions in front of and behind it, as shown in fig. 1.

1 “New England Annelida,” ‘Trans. Acad. of Arts and Sciences,’ vol. iv,
Part 2. Newhaven, 1882.

THE ANATOMY OF SCALIBREGMA INFLATUM. 243

Behind the swollen portion the body tapers gradually to the
posterior end.

The animal is strongly convex above but flattened below.
There is a depressed area along almost the whole length of
the mid-ventral line (marking the position of the nerve-cord),
which is divided by transverse grooves into a series of
squarish or hexagonal areas (fig. 2). In some specimens the
position of the cesophageal connectives is also marked ex-
ternally by two shallow depressions, the metastomial grooves,
which pass round the mouth and unite at the anterior end of
the mid-ventral groove.

The head or prostomium is distinct and well developed,
forming a somewhat quadrangular mass overhanging the
mouth, and bearing at each side anteriorly a short rounded
tentacular process (fig. 3).

Immediately behind the head there is an achetous peri-
stomial segment, composed of two annuli. The rest of the
body of the animal is divided into segments bearing para-
podia. Hach of the first three chetigerous segments is
composed of three annuli, the middle annulus bearing a
prominent pad on each side, from which the neuropodium
and the notopodium arise. All the following fully-formed
segments of the body are divided into four annuli, on the
third of which the parapodia are borne (figs. 1,2). In the
large specimen, 56 mm. long (fig. 1), sixty-one segments (in
addition to the peristomium and pygidium) may be dis-
tinguished. ‘lhe parapodia are clearly visible on all the seg-
ments up to and including the fifty-third ; those of the next
four segments are very small, the dorsal cirrus being the only
easily visible appendage. The last four segments of the
animal are divided from one another only by faint grooves,
and do not bear parapodia. Following these there is a very
short terminal portion or pygidium, which even in large
specimens is only about ‘3 mm. long. There are four pairs
of branched shrubby gills, situated immediately behind the
notopodia of the second, third, fourth, and fifth chetigerous
segments (for further description of the gills see p. 262).

244, J. H. ASHWORTH.

Apertures.

The mouth is a wide transverse ventral sht between the
peristomium and the first cheetigerous segment. It is overhung
by the prostomium, and is bordered anteriorly and poste-
riorly by papille (fig. 2). The pharynx, when fully pro-
truded, is smooth and globular.

The anus is terminal, and surrounded by four slender
anal cirri, two on each side, situated somewhat ventro-
laterally (fig. 6). In one specimen there are five cirri, there
being three on one side and two on the other, but this is
evidently abnormal. In full-grown specimens the cirri are
are about ‘8 to 1:0 mm. long, and ‘05 to ‘06 mm. thick.
There is a small protuberance on each side of the mid-ventral
line of the pygidium, from which the two cirri arise.

The nephridiopores are exceedingly minute and diffi-
cult to see. The first nephridium opens on the fourth
cheetigerous annulus, but as this nephridium is very small its
opening can usually be found only in sections. The second
nephridium is a little larger, but its opening is almost equally
difficult to find. The apertures of the succeeding nephridia,
while being small, are however visible in cleared preparations,
and occasionally in surface view of favourable spirit prepara-
tions. Hach nephridiopore is a minute oval aperture, situated
ventral and anterior to the neuropodium of the segment on
which it opens (fig. 5). The aperture is close to the anterior
border of the cheetigerous annulus, and it is often obscured
by lying in the groove which separates this annulus from
the preceding one. In the most favourable specimens in my
possession the largest nephridiopores are only about ‘06 mm.
in diameter. Towards the posterior end of the animal,
where the nephridia become smaller, the nephridiopores
become correspondingly more difficult to see in surface view,
although they may be distinguished in sections in each seg-
ment almost to the posterior end of the worm. Ina specimen
13°6 mm. long the last visible nephridiopore is only 1 mm,
from the posterior end of the animal,

THE ANATOMY OF SCALIBREGMA INFLATUM. 245

At each side of the prostomium there is a narrow longi-
tudinal slit (fig. 3), which leads downwards, backwards, and
inwards into the nuchal organ (for description see p. 269).

There are small structures on the body-wall situated in
each segment about midway between the notopodium and
neuropodium. At first sight they are liable to be mistaken
for apertures, and Levinsen (1883, p. 133) suggested that
they were sexual openings. On further examination each of
these proves to be a depression, from the floor of which a
small elevation arises, the apex of which may be seen a little
below the level of the mouth of the pit (figs. 1, 4, 5, 8. O.).
These structures are best seen in the segments just behind
the branchial region, but on careful examination they may
be seen in all the chetigerous segments of the animal.
Sections prove that these are lateral sense organs, similar to
the “Seitenorgane” described by Hisig (1887) in the Capi-
tellide. A detailed description may be found on p. 270.

Size.

Scalibregma inflatum varies in size between wide
limits. Of the twenty-two complete specimens in my posses-
sion the smallest is 5 mm. long, and the largest 56 mm. long
and 10 mm. broad at its widest part. ‘The latter is one of
the largest specimens yet recorded (fig. 1). Sars’s specimen
was 58°5 mm. long and 5 mm. broad at its widest part; and
Rathke’s example was one inch and seven lines (about 40 mm.)
long and 5 mm, broad. In the specimen 56 mm. long there
are sixty-one segments, and a very short pygidium (about
*3mm.long). Parapodia are clearly visible on the first fifty-
three segments, but those of the succeeding segments are
very small, or absent altogether. It is interesting to note
that in Sars’s specimen there are also sixty to sixty-one seg-
ments. One of my specimens 35 mm. long contains ova which
appear to be almost ripe, so that the animal reaches maturity
when little more than half its maximum size.

Colour.
Sars has recorded the colour of his living specimen. The

246 J. H. ASHWORTH.

general colour of the body was vermilion red, the parapodia
being light yellow, and the gills blood-red. Rathke’s speci-
mens were greenish grey or dirty greenish yellow in colour.
My spirit specimens are a pale yellowish brown, due to the
large number of yellowish granules in the epidermal cells.

5. Parapopia (PI. 18),

Kach of the parapodia throughout the body is ciearly
divided into a notopodium and a neuropodium, which closely
resemble each other in shape and size. In the anterior four-
teen or fifteen segments the parapodia consist simply of two
blunt conical mammille, each bearing a bundle of sete,
Those of the first five chatigerous segments are situated upon
large elevations, each of which is borne chiefly by the cheeti-
gerous annulus, but also partly by the annulus before and
the one behind it. There are also elevations supporting the
parapodia of the next nine or ten segments, but they are
smaller than those just described (fig. 1). The parapodia of
the anterior portion of the body (as far back as the fourteenth
or fifteenth segment) are comparatively small, and the setal
prominences, which are bluntly conical, project only a little
way from the body-wall. In the succeeding segments the
parapodia gradually increase in size, and each is supported
upon a flattened base, the two rami of the parapodium and
the basal outgrowth forming a large lamella, projecting at
right angles to the body (fig. 8).

The notopodium of the sixteenth segment of most speci-
mens bears a small cirrus (Curr. D, fig. 1), and in one or two
examples a small dorsal cirrus is also present above the noto-
podium of the fifteenth segment. The parapodia of the fully
developed segments behind this bear both dorsal and ventral
cirri. The cirri of the middle part of the body are short,
blunt conical outgrowths, but further back they become
lamelliform or digitiform structures.

Near the posterior end of the animal the parapodia and
cirri are small, and on the last three or four segments, which

~J

THE ANATOMY OF SCALIBREGMA INFLATUM. 24,

are divided from each other only by shallow grooves, para-
podia have not yet been formed (fig. 6). Just in front of these
there are a few segments (about four) in which the parapodia
have only been recently formed, and in these the dorsal cirri
are considerably larger than the ventral ones,—in fact, the
latter have not yet appeared in some segments which possess
dorsal cirri of moderate size. The dorsal cirri are thus
formed before the ventral ones. The notopodium and a few
of its sete are formed before the neuropodium appears.

The cirri are sensory structures, and from their earliest
appearance are supplied with stout branches from the lateral
nerves given off in each segment from the ventral nerve-cord
(Pl. 14, fig. 16).

On each cirrus, a little behind its tip, there is a distinct
darker area, which is somewhat oval, reniform, or pyriform
in shape (fig. 8). In most specimens this area is very
obvious, on account of its brown or black colour, but in some
its colour is much lighter. This is not a structure separate
from and standing out from the cirrus, as the description by
Rathke (pp. 185, 186) would lead one to believe. The
darker appearance of this portion of the cirrus is due to the
presence within it of a collection of special gland-cells, the
dark-coloured glandular mass being visible through the semi-
transparent walls of the cirrus. Rathke examined and
reported on these dark masses in considerable detail, and
rightly inferred that they are similar in structure to the
black or brown spots on the notopodia of Nereis dumerilii,
which he had described on a previous page as glandular
(Hautdriisen). Sars does not mention them, but Danielssen
(p. 75) re-examined them, and came to the conclusion that
they are testes, as they are composed of a large number of
somewhat coiled tubes, filled with minute elongate bodies,
which he took to be “ zoosperms.”? Rathke’s interpretation
is the correct one; these dark bodies are parapodial glands,
the secretion of the cells of which is in the form of minute
slender rods (see p. 248). McIntosh (1885, p. 360) remarks
that these curved bodies in the parapodia of Scalibregma

VOL. 45, PART 2,—NEW SERIES. S

248 J. H. ASHWORTH.

are probably homologous with those described by Kélliker
and Greef in Ephesia (Spherodorum).

The cirri arise as solid outgrowths of the epidermis. From
their earliest appearance they contain gland-cells, which at
first are similar to the ordinary flask-shaped or club-shaped
glandular cells found in the epidermis of the squarish ele-
vations of the skin seen in the anterior third of the animal
(see p. 252 and fig. 12). The gland-cells of the cirri are at
first pear-shaped or club-shaped, and only about 10—12
long. They stain deeply with hematoxylin, and each has a
well-marked nucleus. ‘The secretion is at this time of a
finely granular nature. As the cirri increase in size the
eland-cells elongate, and when the latter become 30—40 pu
in length, their secretion is then clearly seen to be in the
form of exceedingly thin rod-like bodies. ‘The cells continue
to elongate, and in worms only about 15 mm. long the gland-
cells are rather more than ‘1 mm. long (fig. 10). The
ereater part of each cell is occupied by a bundle of fine rods,
but in favourable specimens the nucleus may be seen towards
the rounded inner end of the cell. There is a small amount
of connective tissue around the bases of these rod-secreting
cells. In the largest specimens at my disposal, 35 mm. and
56 mm. long respectively, the gland-cells are twisted, and so
closely packed together that it is almost impossible to
determine the limits of the individual cells. They form a
compact, deeply staining mass, situated a little behind the
tip of the cirrus (fig. 8).

The very fine-pointed ends of the gland-cells open on the
free surface of the epidermis. On examination of medium-
sized specimens it 1s seen that most of the glands of the
notopodial cirrus open on its dorsal side, while those of the
ventral cirrus open chiefly on the ventral wall (fig. 9).

The rods are at first short, and there are comparatively
few in each cell, but later they are much more elongate, and
present in large numbers in each cell. In the largest speci-
men (56 mm. long) the fully formed rods are 40—50 w in
length, and about 3 w in width at their widest pomt. They

THE ANATOMY OF SCALIBREGMA INFLATUM. 249

are spindle-shaped, and taper gradually from about the
middle, where they are thickest, to their very fine-pointed
ends. They are sometimes straight, but more usually are
somewhat curved, sinuous, or twisted (fig. 11). These pecu-
har glands are not strictly confined to the cirri. In four of
the specimens which have been cut into sections there is,
just below each of the neuropodia of two or three of the
anterior segments (ranging from the third to the sixth), a
collection (or sometimes two) of deeply staining cells in the
epidermis. Hach of these cells contains a bundle of rods
exactly like those above described from the parapodial
glands. In two or three cases there is a small bundle of
these rod-forming cells, either in or immediately below the
epidermis, near the terminal portion of the first or second
nephridium. Claparéde (1868, p. 15) has noticed the con-
nection of similar rod-containing cells with the excretory
pores in certain Hesionids.

Rod-secreting glands similar to those of Scalibregma
are known to occur in the skin and subepidermal tissues of
a large number of Polychetse. Claparéde has described
almost indentical structures (“ bacilliparous follicles”’) in the
err of Phyllodoce, sp. (1863, pl. xi, figs. 19, 20), in
papille on the neuropodia of Aricia foetida (1868, pl. xx,
figs. 2B, 2 C), in Nereis cirratulus, especially in the
parapodia and their appendices (1868, pl. xxiv, fig. 1 Z).
A very useful series of figures of these glands, some original
and others collected from various authors, is given by Hisig

(1887, pl. xxxvii).

6. Sera (PI. 13, fig. 9, and Pl. 15, figs. 25, 26).

Both Rathke and Sars described the sete of Scalibregma
as simple, fine, capillary bristles, and they quite overlooked
the peculiar furcate sete which are present in both divisions
of the parapodia throughout the body. Hansen (1882, p. 34)
first observed these curious sete in the ventral fascicles of
S. (?) abyssorum, 8. (?) parvum, and in 8. inflatum,

250 J. H. ASHWORTH.

The sete of S.inflatum are lodged in sacs in the para-
podia, and the tips of the bristles project beyond the promi-
nent lips of these setal sacs. On first examining a para-
podium only the simple capillary sete are seen, but after
rendering the tissues more transparent by treatment with
warm potash solution the furcate sete become visible. The
simple sete project a long way beyond the mouths of the
setal sacs, being exposed for quite half their length, while
the furcate sete are almost entirely enclosed, only their fine
tips protruding from the mouths of the sacs (fig. 9).

The setee of a parapodium may be divided into four groups,
there being one group or row of simple sete and one of
forked sete in each notopodium and neuropodium. The
simple sete of the notopodium and neuropodium form two
straight fascicles, projecting from the parapodium in almost
parallel or in very slightly diverging lines; but the furcate
sete lie in two bundles, which are usually placed so that while
their proximal ends are adjacent their tips are widely diver-
gent, those of the notopodium being directed dorso-laterally,
and those of the neuropodium ventro-laterally. The bundles
of furcate and simple sete form two almost vertical and
parallel rows in each ramus of the parapodium. The row of
forked sete is usually the more anterior.

The simple sete are fine capillary structures, attaining a
length of about 1°7 mm. in the largest specimen (56 mm.
long). They are about 8 w in diameter at their inner ends,
where they are thickest, and taper gradually to a very fine
point. They are marked in their proximal portion by very
minute longitudinal ridges and furrows. ‘Those setve which
have not been worn by use bear exceedingly minute hair-like
processes on their distal third (fig. 26).

The length of the furcate sete is generally about three
fourths that of the simple sete of the same parapodium. In
the specimen, 56 mm. long, they reach a length of 1:2—1°3
mm. ‘They are considerably stouter than the simple sete,
being 15—18 w in thickness at their inner ends. They taper
gradually to the base of the fork, their diameter at this point

THE ANATOMY OF SCALIBREGMA INFLATUM. 251

being only 6—8 wz (fig. 25). The prongs of the fork are
sometimes straight, but more usually curved, their very fine
tips pointing away from each other. The two rami are not
quite equal; in the largest specimens they are 50—65 pw and
65—75 mw in length respectively. The proximal portion of
the edge of each prong bears a number of minute curved
pointed processes.

In the large worm (56 mm. long) there are fifty to sixty
simple sete, and about twenty to twenty-four furcate sete in
each ramus of the parapodium of the anterior half of the
animal.

On clearing the posterior end of another specimen by
treatment with warm potash solution the very small sete
present in the newly formed parapodia are seen. Hach of
the notopodia and neuropodia in this region bears only one
or two simple sete, accompanied by one furcate seta. Both
kinds of setze are therefore present in the parapodia through-
out life.

Furcate setz were first discovered by Malmgren (1867, p.
187) in EKumenia crassa, and were shortly afterwards
observed by McIntosh (1868, p. 419, and pl. xvi, fig. 5) in
HKumenia (Lipobranchius) jeffreysii. Théel (1879, p.
49, and pl.1ii, fig. 47a) figures them in Humenialongisetosa,
and Hansen (1882, p. 54, and pl. v, figs. 16—19) in Seali-
bregma inflatum, 8. (?) parvum, and S. (?) abyssorum;
but the figures of these authors do not show the minute
barbules on the inner side of each prong. McIntosh (1885,
pl. xxi A, fig. 21) saw the barbules on both prongs of the
forked sete of his southern specimens of 8. inflatum, and
figured similar sete from Humenia reticulata (1885, p.
360, and pl. xxii A, fig. 20), and S. Joseph (1894, p. 106,
and pl. v, fig. 133) has observed them in Sclerocheilus
minutus,

The furcate bristles of Humenia glabra described by
Ehlers (1887, p. 170, and pl. xlv, fig. 4) are remarkable for
the great inequality in length of the prongs, one being
nearly three times the length of the other. Ehlers (1887, p.

ooo J. H. ASHWORTH.

127, and pl. xxxviii, fig. 6) has also figured from Nephthys
inermis a forked seta, which is similar to those of Scali-
bregma, except that the two prongs of the fork are equal in
length in the former.

Furcate setz are also known to occur in the Ariciide,
having been described by McIntosh (1879, p. 504, and pl. Ixv,
fio. 7) in Aricia greenlandica, by 8. Joseph (1894, pl. v,
fig. 116, and 1897, pl. xxi, fig. 172) in A. latreillii and
A.levigata, and by Cunningham and Ramage (1888, pl.
xxxvill, fig. 7u; pl. xl, fig. 8p) in Scoloplos armiger and
Theodisca mammillata. The sete of these worms are,
however, evidently quite different to those of Scalibregma,
the fork of the former not being fixed quite so symmetrically
upon the shaft, and the tips of the prongs are not fine and
pointed, but slightly thickened.

Furcate sete, agreeing in essential characters with those
of Scalibregma—that is, possessing unequal barbuled and
finely pointed prongs, are practically confined to the genera
Kumenia, Lipobranchius, and Sclerocheilus, which on
other grounds have been placed with the genus Scalibregma
in the family Scalibregmide.

7. SKIN.

In the anterior and inflated portions of the animal the
annuli bear longitudinal grooves on their dorsal and lateral
regions, which subdivide the skin into a series of squarish or
oval elevations (fig. 5). These are due chiefly to the fact
that the epithelial cells which form them are elongated,
columnar cells, while those of the grooves are much shorter,
almost flattened cells. Many of the cells of the papille are
club-shaped, mucus-forming cells, which stain deeply with
hematoxylin (for other glands of the skin and cirri see pp.
247—249). These cells do not occur in the intervals between
the elevations (fig. 12). In some specimens in which there is
an excessive amount of inflation of the body, the skin of the
inflated region is almost transparent. Behind this region

THE ANATOMY OF SCALIBREGMA INFLATUM. 205

the elevations become less marked, and in the posterior half
of the animal the skin is subdivided only by the circular
grooves which separate the annuli.

Around the bases of the notopodia and neuropodia of the
segments immediately behind the branchial region there are
sometimes epidermal elevations of considerable size. These
are best developed in old specimens (see figs. 1, 5).

Sections of old specimens show that the yellow-brown
colour of the skin is due to the presence of numerous insoluble
yellow granules in the epidermal cells. These granules are
light yellow when viewed singly, but appear brown in the
aggregate.

There is only a small amount of connective tissue between
the epidermis and the underlying musculature.

8. MuscuLaturRE.

Immediately beneath the epidermis there is a layer of cir-
cular muscles, beneath which are the longitudinal muscle
bands which project into the ccelom (fig. 16). The circular
muscles in old specimens usually form a continuous sheet
beneath the epidermis, but in younger ones are sometimes
subdivided into hollow hoops, of which there are two (oc-
casionally only one) in each annulus (fig. 12).

The longitudinal muscles are interrupted along three lines,
viz. on each side at the level of the insertions of the oblique
muscles, and mid-ventrally by the nerve-cord (fig. 16). They
are thus divisible into three groups, of which the two ventral
he between the nerve-cord and the insertions of the oblique
muscles, the other forming an uninterrupted series extending
over the dorsal and lateral regions of the body-wall. The
ventral bands are rather more strongly developed, especially
in young specimens. The longitudinal muscles are covered by
a very thin coelomic epithelium.

The oblique muscles are present throughout the cheti-
gerous segments of the body. They are short, thin, narrow
bands arising at the sides of the nerve-cord and inserted into

254 J. H. ASHWORTH.

the body-wall immediately dorsal to the level of the notopodial
setal sacs (figs. 14, 16). The nephridia are usually almost
hidden from sight beneath these oblique muscles.

The parapodial muscles are moderately well developed
(fig. 9). Each bundle of sete is moved by (1) a number
(about five to eight) of slender protractor muscle strands
attached to the base of the setigerous sac and to the body-wall
near the level of the mouth of the sac, and (2) a few short
strands which pass from the base of the notopodial setal sac,
and are inserted into the base of the neuropodial sac. By con-
traction of the latter muscles the bases of the setal sacs ap-
proximate, and at the same time the distal ends of the two
groups of the sete are caused to diverge.

Into the inner end of each lateral sense organ a special
retractor muscle is inserted. The other end of this muscle is
attached to the base of the notopodial setal sac (fig. 9).

The position and arrangement of the four anterior dia-
phragms and the occurrence of small strands of connective
tissue, representing septa, accompanying the segmental
vessels in the post-branchial region of the body are described
below (p. 205).

There is on each side a short muscle band arising from the
lateral body-wall and inserted into the inner and lower end of
the corresponding nuchal organ. On contraction this muscle
serves to retract the nuchal organ, and also, to a smal]
extent, the prostomium. ‘I'he latter is well supplied with
muscles (fig. 15).

There are several strong muscle strands passing from the
buccal mass to the neighbouring body-wall. These are the
retractors of the proboscis (fig. 14).

Along almost the whole length of the stomach and intes-
tine there is an incomplete ventral mesentery, consisting of
numerous separate muscle strands passing from the ventral
wall of the stomach to the body-wall close to the nerve-cord.

9, GENERAL ANATOMY OF THE INTERNAL ORGANS.

Fig. 14 shows the appearance of the animal when opened

THE ANATOMY OF SCALIBREGMA INFLATUM. 290

by a dorsal incision. The ccelom is spacious, especially in the
inflated portion of the animal. It is subdivided anteriorly
by four diaphragms or septa placed transversely at the level
of the posterior end of each of the four branchiferous
segments. Hach diaphragm is inserted at the level of the
hinder border of the annulus which immediately follows the
chetigerous annulus. The second, third, and fourth dia-
phragms are perforated by the minute funnels of the first
three nephridia.

Behind the branchial region the ccelom is not subdivided by
transverse partitions, but is continuous to the posterior end of
the animal. In the post-branchial portion of the body each
of the segmentally arranged blood-vessels is accompanied by a
small strand of connective tissue, which near the nephrostome
spreads out shghtly and is attached to the body-wall a little
above the level of the neuropodium. The gonads are
developed near the nephrostome on the surface of the expanded
portion of this strand (fig. 21). These narrow bands are
the equivalents of the septa of the branchial region and of
other Annelids, suchas Arenicola grubiiand A. ecaudata,
in which the transition from the narrow bands to complete
septa is well seen (Gamble and Ashworth, 1900, pl. 25, figs.
4A, AD),

The stomach and intestine are loosely bound to the mid-
ventral body-wall by numerous thin strands of muscular
tissue, which form an imperfect ventral mesentery. Just asin
Arenicola (Gamble and Ashworth, 1898, p. 14) the stomach
is probably swung backwards and forwards by the movement
of the body, thus bringing about a thorough mixing of the
sand, etc., with the secretion of the cesophageal pouches and
of the stomach ; the muscle strands forming the incomplete
ventral mesentery allow a certain amplitude of swing, as the
drawing of the dissection shows. In the specimen, the pro-
boscis of which is strongly retracted, the stomach is probably
drawn backwards to its most posterior position, as is shown
by the backward trend of the blood-vessels.

The intestine is probably moveable in a similar manner, but

256 J. H. ASHWORTH.

to a less extent, as the blood-vessels which pass between the
subintestinal vessels and the nephridia and body-wall are
capable of considerable extension without injury (see fig. 14).

From the level of the fourth diaphragm to that of the four-
teenth seta there are six rather long median blood-vessels
running from the ventral vessel to the mid-ventral wall of the
stomach, and paired segmental vessels pass right and left from
the ventral vessel to each parapodium (and corresponding
nephridium) up to the fourteenth. From the fifteenth seta to
the end of the animal there are two segmental vessels, an
afferent and an efferent, on each side.

The nephridia are to a large extent hidden beneath the
numerous oblique muscle bands, and even when exposed by
dissection are difficult to see, on account of their small size ;
they are usually only about -25 mm. in diameter.

10. Catom.

The ccelom is spacious, especially in the inflated region of
the animal. It is subdivided by septa only in the branchial
region. In the rest of the body the septa are very small,
being represented by a thin strand of tissue running along-
side the afferent nephridial vessel.

he coelomic fluid is, as far as can be judged from spirit
specimens, very similar to that of Arenicola (Gamble and
Ashworth, 1898, p. 29, and pl. 5, fig. 24). It contains the
reproductive cells in various stages of growth, and coelomic
cells, some fusiform about 30 « long, and others spherical or
amoeboid.

The reproductive cells collect principally in the space
between the oblique muscles and the ventral body-wall,
especially in ripe females, in which this space is crowded
with ova. ;

11. AtimenrTARY CANAL.

Danielssen (1859, p. 69) described the general form of the
alimentary canal, pointing out its various divisions, and draw-

THE ANATOMY OF SCALIBREGMA INFLATUM. SY)

ing attention to the nature and probable functions of the
cesophageal pouches, and to some details of the structure of
other parts of the digestive tract, e.g. he observed the
ciliated epithelium lining the cesophagus. Wirén (1887, pp.
30, 37) referred to some points in the histology of the ceso-
phagus and stomach.

The mouth (fig. 2) is a transverse slit, situated ventrally
between the peristomial and the first cheetigerous segments,
through which the smooth, spherical, eversible pharynx or
“proboscis”? may be extruded. The mouth is bordered in
front and behind by papilliform elevations of the skin. The
pharynx, when fully protruded, is a smooth, globular structure,
not provided with spines or any other armature. When it is
withdrawn the anterior part of the alimentary canal—the
part lying in front of the first diaphragm—forms a spherical
mass, from which muscle strands pass to the neighbouring
body-wall (fig. 14).

The cesophagusis a narrow cylindrical tube about 8—9
mm. long (in the specimen 56 mm. long), bearing just in
front of the fourth diaphragm a pair of hollow glandular
pouches, which in this specimen are about 2 mm. long and
1:8 mm. wide. Each is a somewhat heart-shaped sac, at-
tached to the wall of the cesophagus by its apex, its free
wider end being bi- or tri-lobed. The two pouches are
united in the middle line, either directly or by a small median
sac, into which both lateral pouches open. They discharge
their secretion into the cesophagus through a small duct
leading from the median sac,

About the level of the sixth to eighth set the cesophagus
passes somewhat suddenly into the much wider stomach,
which even in spirit specimens still bears traces of the
bright orange-yellow colour which Danielssen noticed in
fresh specimens. In all the specimens examined the walls of
the stomach are folded, but whether these folds are natural
it is impossible to state with certainty. The walls of the
stomach and intestine are marked by a number of parallel
lines which pass round the tube from the ventral side to the

258 J. H. ASHWORTH.

dorsal vessel; these are the blood-vessels or sinuses (see
below) similar to those described in Arenicola.

About the level of the fourteenth to sixteenth setz the
stomach passes gradually into the intestine, which is a
cylindrical tube narrowing slightly towards the anus.

The ventral wall of the anus is shghtly notched in the
middle line, and on each side of the notch is a protuberance
from which the two anal cirri arise (fig. 6).

As pointed out above (p. 255), it is moderately certain that
the stomach, and to a less extent the anterior part of the
intestine, are swung backwards and forwards during diges-
tion. In addition to this the passage of the contained sand,
etc., 1s aided by the strong peristaltic movements of the
anterior part of the intestine which have been observed by
Danielssen (1859, p. 70).

The alimentary canal of most of the specimens was dis-
tended with fine sand and débris in which quartz grains,
spicules, frustules of diatoms, and Foraminifera were clearly
recognisable.

Histology.—The cesophagus is lined throughout by cili-
ated columnar cells. There are no gland-cells in this part of
the alimentary canal. ‘he ciliated cells are supported by a
thin layer of muscle-fibres. The walls of the cesophageal
pouches are raised internally into a number of folds, which
are at first mere ridges, but increase in size with the growth
of the animal. Hach fold consists of two layers of epithelial
cells, between which is a blood-sinus, slightly enlarged, near
the inner edge of each fold (fig. 23).

On the external surface of the pouches there is a network,
apparently a blood-sinus, with which the sinuses of the folds
are continuous. ‘The cells lining the cavity of the cesopha-
geal pouches are cubical or flattened, and are not ciliated.
The protoplasm of these cells usually contains an enormous
number of minute spherical granules (or cavities from which
the granules have been dissolved), which give rise to the
glandular secretion (fig. 24). The latter may often be seen
in masses of considerable size in sections of the hinder

THE ANATOMY OF SCALIBREGMA INFLATUM. 259

portion of the cesophagus and the anterior portion of the
stomach,

The stomach is lined by columnar cells which are strongly
ciliated. Among these there are numerous glandular cells
which are swollen with granules of secretion and stain deeply
with hematoxylin. There is an exceedingly small amount of
muscular tissue in the walls of the stomach.

The columnar or cubical cells which line the intestine are
supported by a thin muscular layer. In the cells of the
dorsal and lateral walls of the intestine of large specimens
there are very numerous yellow granules, probably chlorogo-
genous. There is a well-marked ventral groove, the cells of
which are columnar and bear long cilia, running along the
whole length of the intestine to the anus (fig. 16). I have
traced this groove forwards as far as the level of the fifteenth
or sixteenth sete. The function of this groove is probably
the same as in Arenicola, viz. to carry backwards along the
intestine the digested substances which have been extracted
from the sand. In some specimens food particles may be
seen in the groove surrounded by a thin covering of mucus.
Towards the posterior end of the intestine the whole of its
inner wall appears to be ciliated, and the cilia seem to be
especially strongly developed in the last few segments.
There are two cords situated in the ventral wall of the
intestine below the ciliated groove. These, which are best
developed in old specimens, are apparently nervous (see

p. 268).

12. Vascutar System (fig. 14).

Danielssen (1859, p. 70) has described and figured some of
the principal parts of the vascular system, but as his account
is not complete, and is incorrect in some respects, I propose
to describe the vascular system as seen in the dissection of
my largest specimen (56 mm, long).

The dorsal vessel arises near the anus, and runs along the
whole length of the alimentary canal, breaking up into capil-
laries on the pharynx. It is closely adherent to the gut, and

260 J. H. ASHWORTH.

receives a large number of fine vessels (or lacune; see below,
p- 261) from the walls of the stomach and intestine. Near
the anterior end of the stomach the dorsal vessel presents a
well-marked enlargement, which is apparently constant, as it
is present in the other specimens examined. In my largest
example this swelling, which, following Danielssen, we may
call the blood-reservoir, is 7 mm. long and 1:2 mm. thick in
its widest part. Anterior to this the dorsal vessel resumes
its normal diameter for a length of about 4—5 mm. and then
abruptly dilates into a conical bulb, which Danielssen named
the heart, about 1°5 mm. in diameter. ‘I'he vessel then nar-
rows to its previous size, and gives off four pairs of stout
afferent vessels which run along the corresponding dia-
phragms to the gills. On reaching the pharynx the dorsal
vessel divides into two branches, which soon break up
into smaller vessels supplying the pharynx, buccal mass,
brain, etc.

The ventral vessel arises near the mouth, by the union of
small vessels from the prostomium and peristomium. It runs
along the whole length of the animal just above the nerve-
cord. Soon after its origin it receives four pairs of efferent
branchial vessels, and thus becomes almost at once a thick
trunk. In each of the post-branchial segments the ventral
vessel gives off a pair of slender vessels supplying the
nephridia, setal sacs, and neighbouring tissues. Besides
these paired branches the ventral vessel gives off to the
stomach six median vessels, the first of which is situated
just behind the fourth diaphragm, and the last at the level
of the thirteenth sete. In the posterior portion of the
animal, behind the twentieth segment, the ventral vessel
bears a large number of short, blind, usually curved out-
erowths, which are covered with a layer of cells, probably
chlorogogenous, and corresponding to the similar tissue
clothing the biind outgrowths of the ventral vessel of
Arenicola marina (Gamble and Ashworth, 1898, pl. 2,
fig. 5).

Along the whole length of the intestine there is a pair of

THE ANATOMY OF SCALIBREGMA INFLATUM. 261

subintestinal sinuses situated one on each side of the ventral
groove of the intestine (figs. 14—16). These may be traced
from behind forwards as far as the level of the fifteenth sete,
then they taper rapidly and disappear. Anterior to this
point the stomach receives blood only from the six median
vessels above referred to. In each segment, from the fifteenth
to the end of the body, a pair of vessels collecting blood from
the nephridia and setal sacs opens into the subintestinal
sinuses.

On the walls of the stomach and first part of the intestine
there are numerous fine blood-streams, which carry blood
from the ventral portion of the gut into the dorsal vessel.
These are not distinguishable on the posterior part of the
intestine, as this portion of the gut is surrounded by a sinus,
by means of which blood is conveyed from the subintestinal
sinuses to the dorsal vessel. The whole of the blood in the
walls of the stomach and intestine is contained in sinuses;
the intestine, as seen in section, appears to be quite enclosed
in a blood-sinus. ‘The subintestinal sinuses are somewhat
specialised parts of the general sinus. ‘The dorsal vessel is
not distinct from the sinus in the posterior part of the animal,
but from the level of the twelfth setz (i.e. a point a little
behind the blood-reservoir) it is distinct, and has a wall of
its own.

In Arenicola the blood in the walls of the stomach and
intestine is apparently contained in vessels in young speci-
mens, but in sinuses in old specimens (Gamble and Ash-
worth, 1900, p. 460) ; but even in the latter it is sometimes
difficult to determine whether the gastric plexus is formed of
vessels or sinuses. In Scalibregma the blood in the walls of
the stomach and intestine is certainly contained in sinuses,
which in the posterior part of the intestine are large.

The body-wall and nerve-cord are very sparingly supplied
with blood. No vessels are distinguishable in the body-wall,
except in immediate proximity to the setal sacs, and these
vessels are few and small.

The walls of the heart and blood-reservoir are very thin,

262 J. H. ASHWORTH.

and their structure is difficult to determine. The walls are
composed of a layer of peritoneal epithelium within which a
very thin sheet of muscle-fibres may be distinguished. In
some sections an exceedingly delicate endothelium appears
to be present, but this is difficult to distinguish with cer-
tainty. There is no trace of heart body such as is present in
the dorsal vessel of some other Polychetes.

Danielssen states that the blood-reservoir and the heart
are contractile, alternately expanding and contracting with
considerable force, driving the blood forward to the gills.
The blood-plasma is red and the corpuscles are few in
numbers. They are spherical or ellipsoidal cells, 6—9 w in
diameter, and have prominent nuclei. It is very difficult to
ascertain where they are formed, but apparently some arise
from the cells lining the wall of the dorsal vessel, especially
in the region of the heart and blood-reservoir. In one speci-
men there is a mass of corpuscles in the ventral vessel imme-
diately behind the fourth diaphragm. 'These corpuscles re-
semble in appearance and in reaction to stains the cells lining
the wall of the vessel in their immediate neighbourhood.
Possibly corpuscles are formed at various points in the vessels.

io. (GiLis:

The four pairs of gills are shrubby, much-branched out-
growths of the body-wall situated immediately above and
behind the notopodia of the second, third, fourth, and fifth
chetigerous segments (fig. 1).

Rathke describes the gills as being oe on the fourth
to seventh segments, and figures three chetigerous segments
anterior to the first pair of gills; but other authors describe
the gills as being situated on the second to fifth segments.
It seems unlikely that Rathke’s specimen, while agreeing
very closely with other specimens obtained from the same
locality, should differ from these only in the position of the
gills. It seems. probable that in this respect Rathke’s ac-
count is incorrect.

The first gill is considerably smaller than the other three,

THE ANATOMY OF SCALIBREGMA INFLATUM. 263

which are nearly equal in size. The branches of each gill
spring from a single stem, which is short and stout, and
soon divides into two main branches, one of which is directed
dorsally and the other ventrally (fig. 4). Each of these
usually divides again into two, and these branch freely,
sometimes dichotomously, or often dividing into three.

In the living animal the gills are red and the fine branches
reddish yellow, due to the contained blood (Sars). The
gills are hollow, each containing a prolongation of the
ceelom. ‘heir walls are composed of single layers of epi-
thelial cells, within which is a delicate coelomic epithelium
surrounding the axial cavity. Between these two layers is a
thin sheet of muscle-fibres, upon the presence of which the
contractility of the gill depends.

The gills are supplied with blood by four pairs of afferent
vessels given off from the dorsal vessel, and they return
blood by a corresponding number of efferent trunks to the
ventral vessel (fig. 14). The position of the vessels and the
circulation of the blood in the gills is difficult to make out
from my material, as the gills are almost bloodless in all the
specimens.

14. Centra Nervous System.

Danielssen (p. 72) has given a brief account of the nervous
system. He figures (pl. i, fig. 5) the nerve-cord as a double
chain, upon which there are ganglia in the middle of each
segment, each giving off a pair of nerves to the body-wall.
From the cesophageal connectives three fine nerves are given
off on each side. The brain, which consists of two masses
connected by a transverse commissure, also gives off three
nerves on each side, which run forwards.

I cannot agree with Danielssen on several of these points,
and especially on the ganglionation of the nerve-cord. I find
that the cord is of almost uniform thickness, there being no
ganelia visible either in dissections or in horizontal sections.

The central nervous system closely resembles that of

voL. 45, PART 2.—NEW SERIES. T

264. J. H. ASHWORTH.

Arenicola, especially that of A. claparedii (Gamble and
Ashworth, 1900, p. 469), with which it agrees even in many
of its details.

The Brain.

The brain is lodged in the middle portion of the pro-
stomium (fig. 15). It is somewhat A-shaped, the single
anterior lobe being in contact with the anterior face and
dorsal wall of the prostomium, and the two posterior lobes
lying in contact with the inner sides of the two nuchal
grooves. In some specimens the anterior brain-lobe is not
in close contact with the dorsal prostomial epithelinm along
its whole length, but in the posterior half is separated from
the epithelium by a thin sheet of muscle-fibres. The brain
is placed in a slightly slanting position, its anterior lobe
being situated more dorsally than the posterior lobes.

The anterior lobe is almost entire, the only trace of division
being a very shght groove along its ventral surface ; but the
two posterior lobes of the brain are separated from each
other by a considerable space lined by ccelomic epithelium,
and containing muscle-fibres and blood-vessels.

The dorsal and lateral portions of the anterior brain-lobe
consist chiefly of small oval or pyriform cells, some with
small deeply-staining nuclei, others with vesicular nuclei,
with one or two small dark nucleoh. A few larger cells are
found here and there. ‘The ventral part of this lobe of the
brain consists chiefly of a delicate neuropile.

The anterior brain-lobe gives off a pair of moderately stout
nerves to the hollow tentacles (N. Tent., fig. 15). The nerve
spreads out just beneath the epidermis of the base of the
tentacle, gradually thinning out towards the tip. The stout
cesophageal connectives arise from the brain a little further
back, i.e. about the middle of its length. The tentacular
nerves receive fibres from the dorsal and ventral part of the
anterior brain-lobe, and there is a considerable mass of cells
immediately below and to the outer side of the origin of each
of these nerves. The connectives also receive fibres from

THE ANATOMY OF SCALIBREGMA INFLATUM. 265

the dorsal and ventral portions of the brain, and there is a
group of larger nerve-cells just below their point of origin.
The posterior brain-lobes consist of nearly equal parts of
cellular and fibrous elements. The fibrous matter is covered
internally by a thin layer of cells, but externally has a thick
coating of ganglion cells, which are closely applied to the
nuchal epithelium. This mass of cells forms a large gan-
ghonic centre. The posterior brain-lobes are broad in front
where they are fused with the anterior part of the brain, and
in this region there are numerous comparatively large nerve-
cells, especially on the inner faces of these posterior lobes
abutting on the coelomic cavity. On tracing these lobes
backwards along the inner side of the nuchal organ, it is seen
that the cells decrease rapidly in quantity, and each lobe is
continued as a fibrous tract or nerve, which is accompanied
by only a very thin covering of cellular elements (N. Nue.,
fig. 15). This divides into two or three nerves near the
posterior end of the nuchal organ. ‘The nerves lie between
the epithelium and the sheath of the organ.

There is a little neurilemma on the dorsal and ventral faces
of the brain, from which strands pass inwards, supporting the
nervous elements.

The above is a description of the brain of moderately
young specimens 13 to 14°3 mm. long. The average measure-
ments of the brain of five such specimens are ‘23 mm. long,
°22 mm. broad, and 16 mm. deep. As the animal increases
in size the brain not only grows in bulk but undergoes con-
siderable changes in appearance. Ina specimen of 56 mm.
long the brain is °35 mm. long, *5 mm. wide, and ‘35 mm.
deep. The fibrous portion of the brain in this specimen is
proportionately larger and much more complex, and the
neuroglia is better developed than in smaller specimens.
The nerve-cells, some of which are 30 w in diameter, are
aggregated into definite groups, separated by masses of
fibrous tissue. As in younger specimens the fibrous elements
are chiefly internal, and are covered by the cells. At the
point of origin of the connectives the fibrous matter is ex-

266 J. H. ASHWORTH.

ceedingly abundant, and here, too, on each side are a few large
unipolar cells with prominent nuclei. There seems to be a
definite nerve-tract arising from this mass of cells and passing
into the connective. The fibres of the connective appear to
be derived almost entirely from the anterior and middle
portions of the brain, only a very small proportion being
derived from the posterior lobe. In the posterior lobe there
is also a number of larger cells, but the fibrous and cellular
elements are in almost the same proportion as in younger
specimens.

The Gsophageal Connectives.

The connectives arise from the brain, slightly anterior to
the middle of its length. ‘They run just beneath the epi-
dermis, and at first fall nearly vertically downwards near the
middle line, then diverge sharply, pass round the mouth, and
unite just anterior to the level of the second chetigerous
annulus to form the nerve-cord. The course of the connec-
tives is marked externally by the metastomial grooves, which
are well seen only in comparatively few specimens.

The connective of each side gives off nerves to at least two,
and sometimes three, annuli, through which it passes, and also
a nerve through which it runs along the sides of the mouth
to the eversible part of the pharynx. The former nerves are
situated just beneath the epidermis; the latter nerve, which
also supplies the upper lip, may be traced by the aid of its
distinct sheath for some distance along the dorso-lateral
region of the pharynx between the epithelial and muscular
layers, and is probably in connection with the stomatogastric
system.

The connectives are composed chiefly of fibrous matter,
but there is a thin coating of cells on the external face, and
at the point of union of the two connectives there are several
larger nerve-cells. The connective is enclosed in a sheath of
neurilemma, which is better developed in old specimens, and
in the latter sends ingrowths which partially subdivide the
connective into two or three,

THE ANATOMY OF SCALIBREGMA INFLATUM. 267

The Nerve-cord,

The most striking point in Danielssen’s description of the
nerve-cord of Scalibregma is the ganglionation. I am
unable to find any trace of the segmentation or gangliona-
tion of the nerve-cord either in dissections or in sections
taken in various planes. Ganglion cells occur, apparently
evenly distributed along the whole length of the cord on its
lateral and ventral faces, as in Arenicola (Gamble and
Ashworth, 1900, p. 480). In most species of Arenicola,
however, an indication of the segmentation of the cord is
afforded by the presence of giant-cells placed at regular in-
tervals along the cord near the posterior boundary of each
segment. In Scalibregma there are no such landmarks,
giant-cells and giant-fibres are entirely absent.

The nerve-cord of Scalibregma is not ccelomic in posi-
tion in any part of its course. It is situated in the body-wall
outside the layer of circular muscles, and in close contact
with the epidermis (fig. 16).

The cord gives off a pair of nerves situated in each inter-
annular groove in the basal portion of the epidermis (fig. 12).
The nerves which lie in the groove immediately behind each
cheetigerous annulus are larger than the rest. Besides these
there is a pair of moderately large nerves given off opposite
the middle of each chetigerous annulus, which also run be-
tween the epidermis and the circular muscles. Hach of these
nerves (fig. 16) gives off (1) a branch passing into the base
of the ventral cirrus, and spreading out beneath the epi-
dermis ; (2) a branch to the lateral sense organ ; (5) a branch
to the dorsal cirrus. The nerve then continues dorsally
along the annulus, gradually tapering, and becoming very
difficult to trace. ‘The nerve to each cirrus comes into close
contact at one point with, and sends fibres to, the corre-
sponding setal sac near its mouth.

In the posterior portion of the animal the nerve-cord lies in
very close relation to the epidermis, which is here very thin.

268 J. H. ASHWORTH.

Near its termination in the tail segment the cord gives off a
pair of comparatively large nerves supplying the anal cirri.

In transverse section the nerve-cord is oval in shape, being
flattened from above downwards (fig. 17). In some speci-
mens it is very much flattened in the posterior region of the
animal, whereas in others it is not so compressed.

Ganglion cells occur along the whole length of the cord,
being placed on the ventral face and at the ventro-lateral
angles. ‘These cells are small and subequal, although here
and there a few larger cells may be seen generally situated
near the median line in the fissure between the two fibrous
tracts. The fibrous matter of the cord is partially sub-
divided into two by a median vertical sheet of neuroglial
tissue, the fibrille of which form, in transverse section, a
network, more obvious in the ventral portion of the cord.
In horizontal sections the neuroglial fibrille form wavy
strands resembling the neuro-fibrille, but the former are
generally more deeply staining than the latter.

In older specimens there is a proportionately greater
amount of fibrous matter in the cord, and the cells are
restricted almost entirely to the ventral face, and most of
them are situated in the small fissure between the fibrous
tracts, there being very few at the ventro-lateral angles of
the cord. In such specimens (80—56 mm. long) the neuri-
lemma sheath and neuroglial network are more highly de-
veloped than in younger specimens.

The brain and nerve-cord, and especially the latter, are
poorly supplied with blood.

There are two cords (fig. 16, Int. N.), best developed in
old specimens, running along almost the whole length of the
intestine. ‘They are situated in the ventral wall just below
the ciliated groove. They are composed chiefly of fibrous
elements, but cells are present at frequent intervals. From
their appearance and structure they seem to be nervous, but
I have been unable to find any connection between them and
any other part- of the nervous system. ‘The cords become
gradually smaller as they approach the posterior end, and
finally coalesce. They may be traced as far as the anus.

THE ANATOMY OF SCALIBREGMA INFLATUM. 269

15. Sense Orcans.

The sense organs are (1) the epithelium of the prostomium
and tentacles, (2) the nuchal organ, (3) the lateral sense
organs, (4) the dorsal, ventral, and anal cirri. Probably also
the long capillary sete should be added to this list. There
are no eyes or otocysts.

Prostomial Epithelium.

The epithelium of the anterior and dorsal faces of the pro-
stomium and its tentacles consists of columnar cells, among
which slender fusiform sense organs inay be distinguished.
The latter are generally seen in small groups, and their
slender tips are level with or project slightly beyond the
outer surface of the cuticle. The bases of these cells are in
intimate relation to either the cells of the brain itself or the
fibres of the nerves which supply the two tentacles.

The Nuchal Organ (figs. 3, 15).

On each side of the prostomium there is a narrow longi-
tudinal slit which leads inwards and downwards into the
blindly-ending nuchal organ. In sections the inner ends of
the two nuchal organs are seen lying close together near the
middle line, below and behind the brain. The character of
the epithelium lining the organ varies considerably. Near
the mouth of the depression the epithelial cells are short,
columnar, or cubical, and stain lightly, but towards the inner
end they rapidly increase in length, and here they are long,
narrow, columnar, and deeply-staining, and many of them in
the terminal portion are ciliated. In some specimens there
is quite a sharp line of demarcation between the cubical and
elongate cells. In favourable sections sense-cells may be
seen among the columnar cells in the middle and inner por-
tions of the organ. From the inner ends of these fusiform
sense-cells slender fibrils may be traced to the adjacent
nerve, which is in continuity with the posterior lobe of the

270 J. H. ASHWORTH.

brain. The epithelium lining the inner or terminal portion
of the organ is strongly folded, and suggests that this section
of the organ is to a certain extent eversible. Possibly the
small papilla noticed by Sars in 8. inflatum (1846, fig. 21),
and by Hansen in 8. (?) parvum (1882, p. 34, and pl. v,
fig. 8), may be the nuchal organ only partially withdrawn.
The retraction of the nuchal organ is effected by a small
muscle arising from the body-wall at a point about midway
between the notopodial and neuropodial sacs of the first
chetigerous segment, and inserted into the inner end of the

organ (Nuc. Retr., fig. 15).

Dorsal, Ventral, and Anal Cirri; Sete.

As pointed out in the section dealing with the ventral
nerve-cord, there is a pair of nerves given off in each seg-
ment, supplying among other structures the parapodial cirri
and the setal sacs. The nerves to the cirri spread out be-
neath the external epithelium. The nerves which supply the
cirri send fibres to the setal sacs near their mouths; these
may be traced for a short distance along the sacs towards
their inner ends, but owing to their small size they are soon
lost from view. ‘They probably end among the bases of the
seta, as Retzius! has shown for Arenicola.

The anal cirri are abundantly supplied with nerves by a
pair of trunks given off near the termination of the cord.
Hach cirrus is an epidermal outgrowth, along the whole
length of which there is an axial nervous strand. The nerve
is surrounded only by a single layer of epidermal cells.

16. THe LateraL SENSE ORGANS.

These are the most interesting sense organs of Scali-
bregma. Levinsen (1885, p. 133) noticed the prominent
lips guarding the depression into which the sense organ is
withdrawn, but he mistook the structure for an aperture
“probably sexual.” Théel (1879, p. 49) observed a papilla

1 * Biolog. Foren. Forhandi.,’ Band ii, Hefte 4—6, p. 85, 1891.

THK ANATOMY OF SCALIBREGMA INFLATUM. PAGAL

between the two rami of each of the parapodia of Humenia
longisetosa, but he did not recognise the nature of these
papillee.

Our knowledge of lateral sense organs of this kind is due
almost entirely to Hisig, who has described their relations
and structure in his monograph of the Capitellide (1887).

The lateral sense organs of Scalibregma occur in each
chetigerous segment throughout the whole length of the
body, midway between the notopodium and neuropodium
(fig. 1, S. O.). The sense organs on the first and second
chetigerous segments are small rounded eminences, very
difficult to distinguish in surface view, even with a moderate
magnification. Those of the succeeding segments are, in
preserved specimens, generally sunk and hidden in a depres-
sion bordered by prominent lips of epidermis (figs. 4, 5).
The essential portion of the organ is hable to be overlooked,
and the depression, by reason of its prominent lips, may then
be readily mistaken for an aperture. ‘The sense organ itself
is a papilla arising from the bottom of the above-named
depression, its free, oval, curved surface bearing a very
narrow, dark, almost flat area, running dorso-ventrally. ‘The
sense hairs arise from this darker area or ‘hair field”
(Hisig), which, in the largest organs of a specimen 56 mm.
long, is only about 50 w long and 10—15y broad, In this
specimen the surface of the free pole of the largest papille
is about 250 « long and 90 « broad.

The sense organs are best developed in the region of the
body just behind the gills. In the posterior third of the
animal they gradually decrease in size, and in the last five or
six segments they are difficult to find even in sections.
About the sixth or eighth segment from the posterior end the
organ is recognisable as a minute oval elevation, measuring
about 15 « along its longer diameter (fig. 7).

The structure of these organs can be best studied in thin
transverse sections of specimens about 15 mm. long (figs. 28,
29). The organs have attained almost the same stage of
growth even in specimens only 5—7 mm. long.

272 J. H. ASHWORTH.

The sense hairs form a dense tuft, covering the flattened
area in the middle of the free surface of the papilla. They
are moderately stiff hairs, attaining a length of about 40—
50 x, but they are exceedingly delicate, being less than 1 pu
thick at their bases. ‘here are a hundred or more hairs in
each of the sense organs in the anterior and middle regions
of the body. The whole papilla is covered by the thin cuticle
continuous with that covering the general epidermis, but over
the hair field the cuticle is exceedingly thin, and is pierced
by the sense hairs. Beneath the cuticle, over the greater
part of the surface of the sensory papilla, there is a layer of
columnar or cubical epidermal cells, but in the hair field there
is a striking departure from this arrangement. Here, below
the cuticle, are long, exceedingly thin columnar cells, closely
and regularly arranged. These rods are in most specimens
12—15 yw long (but in the largest specimen, 56 mm. long,
they attain a length of 20—25 ,) and about 1 w wide. They
stain darkly, but not quite homogeneously, there being a
more deeply-staming, elongated, flattened nucleus near the
distal end of each rod. Hach rod bears only one or two
hairs. The rods are continued inwards as delicate fibrils,
many of which may be traced into continuity with the delicate
drawn-out ends of pyriform or fusiform ganglion cells, which
occupy the axis of the sensory papilla. Many of these gan-
glion cells are clearly bipolar, the outwardly directed process
being, as described above, in connection with the base of a
rod, the inward process passing into the nervous mass
formed by the spreading out of the spinal nerve in the basal
portion of the sensory outgrowth. In older specimens espe-
cially, the ganglion cells are nearly all obviously bipolar.
These ganglion cells are few in number, there being only
about eight or ten in each sense organ. ‘They are usually
about 15—20 pw long,and about 8—10 mw wide, and their large
nuclei are 6—8 yp in diameter. Occasionally, especially in
large specimens, cells 30 « long with nuclei 8—12 min dia-
meter may be seen.

There are other ganglion cells generally aggregated into

THE ANATOMY OF SCALIBREGMA INFLATUM. 273

a small mass near the base of the papilla, but these cells are
rather smaller and more spherical than those described
above.

At the base of the papilla, around the nervous axis, there
are numerous very deeply staining nuclei-like bodies, about
4—}5 » in diameter. According to Hisig (1887, p. 505) these
are to be regarded as nuclei of multipolar ganglion cells
which have been deprived of their cellular substance. These
nuclei are situated upon a network of fine fibres, probably
nervous, since they are in close relation to the fibrils of the
branch of the spinal nerve supplying the sense organ.
These fibres probably represent the protoplasmic part of the
cells of which the deeply staining bodies are the nuclei.
These and the basal ganglion cells are more numerous on
the ventral side of the axis of the organ than on the dorsal
side. '‘l'his is probably accounted for by the fact that the
branch of the spinal nerve enters the papilla on the ventral
side. ‘The nerve, soon after entering the sensory elevation,
turns nearly through a right angle, and then runs along the
axis of the papilla, its ultimate branches terminating among
the ganglion cells and bases of the rods (fig. 30).

Inserted into the base of the sensory papilla there is a
retractor muscle, the fibres of which spread out fanwise on
the basal part of the nervous substance of the sense
organ (figs. 16,29). In some cases the muscle-fibrils extend
inwards into the papilla as far as the ganglion cells. In
favourable specimens the intimate relation of these muscle-
fibrils and the nerve-fibrils may be clearly seen, and it
appears probable that there is an arrangement similar to
that shown by Eisig for the Capitellide (1887, p. 505), viz.
that fine processes of the ganglion cells end in the fibrils of
the retractor muscle. ‘he muscle is attached to the inner
end of the notopodial setal sac.

The position of the sense organ between the two projecting
setal sacs affords it considerable protection, and additional
protection is given to the sensory area by its withdrawal into
a depression of the epidermis by means of the special

274, J. H. ASHWORTH.

retractor muscle. Some such arrangement is necessary to
prevent injury to the delicate sense hairs when the animal is
burrowing in the sand.

The sense organs differ widely in structure according to
their age and the size of the specimen from which they were
taken. Very young organs may be seen in sections of the
last few segments of an animal.

Sense organs are clearly distinguishable in about the third
segment in front of the tail segment (fig. 27). The rods
are exceedingly small and difficult to see; they occupy an
area equivalent to that of one or two epidermal cells. The
sensory area is only 10 4 long. Below the rods are two or
three small ganglion cells about 8—10 pw long, and below
these are about twenty deeply staining nuclei. In the next
anterior segment the rods are rather more obvious, being
5-—6 w long, and in the segment further forward the sense
hairs are clearly visible, and have attained a length of about
5—6 p. In these posterior segments the rods, ganglion cells,
and nuclei are closely compressed, and their relations are
difficult to determine; but further forwards, as the sense
organs increase in size and the various structures become
better differentiated, their connections with each other may
be more readily seen. The foregoing description on pp. 272,
273, is taken from fully developed sense organs of specimens
13—15 mm. long. In older specimens there are still further
changes (fig. 30). The axial portion of the organ becomes
more fibrous, the ganglion cells undergo little change, but
there are very many more of the deeply staining nuclei at
the base of the organ than in younger specimens. ‘The rods
also stain more homogeneously, their nuclei being almost
invisible, their position being indicated by a slightly darker
area in each rod.

Occurrence of Lateral Sense Organs in other
Polycheta.

Lateral sensé organs are proved to occur in only a very
few Polycheta. They are found in the Capitellide (with

THE ANATOMY OF SCALIBREGMA INFLATUM. 275

the exception of Capitella), and their relations and
structure have been exhaustively investigated by Hisig
(1887, p. 494). HE. Meyer! has described the small and
simple “Seitenorgane” of Polyophthalmus pictus. hese
are, I believe, the only hitherto published accounts of the
lateral sense organs of Polychetes.

These sense organs of Scalibregma closely resemble
those of Capitellide, except that in the latter there are no
large ganglion cells beneath the rods. I have also found
similar sense organs in specimens of Humenia crassa and
Lipobranchius jeffreysii, which reached me when this
paper was almost completed. In the two latter genera, the
depression in which the sense organ is lodged is easily visible
on each side midway between the notopodium and neuro-
podium of each chetigerous segment. In several cases the
sensory papilla and the “hair field”? can be clearly seen
within the depression.

From an examination of published figures of Polycheeta it
appears probable that ‘“‘Seitenorgane”’ are rather more
widely distributed than is generally supposed, for there are
certain structures shown in these figures which strongly
remind one of lateral sense organs, both by their position
and appearance. It was stated above (p. 270) that the aper-
tures of the depressions containing the sense organs of
Scalibregma have been mistaken by earlier observers for
sexual openings. It is probable that certain apertures de-
scribed as occurring in a corresponding position in other
genera may eventually be shown to be depressions lodging
sense organs. 8S. Joseph (1898, p. 371, and pl. xxi, figs.
187, 188) noticed in Ophelia neglecta an oval pore
opening into a goblet-shaped depression situated between
the two rami of nearly all the parapodia. Kiikenthal? has
also figured a structure in a corresponding position in the

1 «Zur Anatomie und Histologie von Polyophthalmus pictus,”
‘Archiv fiir mikros. Anat.,’ Band xxi, p. 791. Bonn, 1882.

2 “Uber das Nervensystem der Opheliaceen,” ‘Jenaische Zeitschr. f,
Naturw.,’ Band xx, p. 510, and Taf. xxiii, fig. 24. Jena, 1887,

276 J. H. ASHWORTH.

parapodia of Ophelia limacina, and Rathke (1843, p. 202,
and Tab. x, fig. 15; p. 203, and Tab. xi, fig. 14) has de-
scribed small apertures similarly situated in Ammotry pane
cestroides and in A.(= Ophelia) limacina. He believes
these to be ovipores.

The papilla figured by Théel (1879, pl. iii, fig. 46°)
between the notopodium and ueuropodium of Kumenia
longisetosa is almost certainly a sense organ. ‘There are,
in some specimens of Arenicola cristata (Gamble and
Ashworth, 1900, p. 443, and pl. 24, fig. 33), small papille,
or sometimes depressions, corresponding in position to that
of the above-described sense organs, but whether these
structures in Arenicola are sensory could not be deter-
mined, owing to the defective preservation of the specimens
examined.

The Morphology of the Lateral Sense Organs of
Polycheia.

The morphology of the lateral sense organs of Capitellidze
has been exhaustively treated by Professor Hisig (1887) in
his classical monograph of this family of Polychetes. In the
spring of last year, while working in the Zoological Station
in Naples, I had the privilege of discussing this question
with Professor Eisig, and I am grateful to him for so care-
fully explaining to me his views upon this subject. He
believes that these sense organs are modified cirri, and bases
his conclusions on the following arguments. (1) The known
sensory nature of cirri, as indicated by the presence on the
cirri of some Polycheetes of fine stiff hairs. (2) If a gradual
shortening of a cirrus took place, the free nerve endings
would become more and more aggregated at the free pole of
the papilla, thus producing an organ of the same shape and
general structure as a ‘‘Seitenorgan.” (3) Hach of the
lateral sense organs of Capitellide, being situated imme-
diately dorsal to the neuropodium, is considered by Hisig as
equivalent to the dorsal cirrus of the neuropodium. He
turns for confirmation and support to the parapodia of the

THE ANATOMY OF SCALIBREGMA INFLATUM. 21

Glyceridz. In this family the parapodium is not so obviously
biramous as in many others, and he considers the whole
parapodium of the Glycerids is really a neuropodium (the
notopodium being absent) equivalent to that of Capitellids.
In this case the dorsal cirrus present on the parapodium of
Glycerids would occupy a position corresponding to that of
the sense organ on the neuropodium of Capitellids. But, im
my opinion, the evidence afforded by the study of the com-
parative morphology of the parapodia of Glycera and other
Polychetes is against this argument and the conclusions
drawn from it by Hisig.

The parapodia of several of the Glyceridz are, at first
sight, single outgrowths, in some species the division into
two being only feebly marked. Each parapodium is, how-
ever, essentially biramous, as is shown by (1) the bifid tip of
the parapodium ; (2) the sete are imp!anted in if in two more
or less distinct divergent bundles. These points may be at
once verified by reference to the figures of parapodia of
several species of Glycera which Kisig has collected and
placed on the last plate of his monograph (see Taf. xxxvil, and
note especially G. capitata, fig. 31). The chief differences
between the parapodia of the various species of Glycera
are traceable to the varying amount of compression and
approximation of the parts of the parapodium. In some
cases the two rami are so closely approximated that the
mouths of the two setal sacs are almost confluent, but even in
these cases, on tracing the sete to their inner ends, it is
usually seen that they fall into two distinct and separate
groups—a notopodial and a neuropodial. There are usually
two acicula in each parapodium, one in the dorsal and the
other in the ventral ramus. These acicula are points of
insertion of the muscles which move the parapodium and
sete, and the presence of two indicates the essentially bifid
character of the parapodium by pointing to the fact that
there are two sets of muscles for moving the set, one for
each bundle (see McIntosh, 1885, pl. xlii, figs. 5, 6, 8, 10).

It is, therefore, most probable that the parapodium of the

278 J. H. ASHWORTH.

Glyceridg is truly biramous, each ramus bearing a group of
sete. Moreover, the notopodium and neuropodium of
Glycerids are equivalent to the correspondingly named
structures in other Polychetes, the sole difference being that
in the former the typical parts of the parapodium have
become more closely approximated and merged into one
another than is usual.

This may be further emphasised by a comparison of the
parapodia of Glycera and Scalibregma. There is no
difficulty in homologising the parts common to both. The
parapodium of each of these genera bears two bristle bundles
—a notopodial and a neuropodial—and also a dorsal and
ventral cirrus. In addition there is in each parapodium of
Scalibregma a structure not represented in Glycerids,—the
lateral sense organ, which is situated between the two rami
in a position corresponding to that of the “ Seitenorgane” of
the Capitellid parapodium.

Hisig’s assumption that the whole parapodium of Glycerids
is equivalent to the neuropodium of Capitellids does not
appear therefore to hold good. The facts cited above go far
towards proving that the Glycerid parapodium contains a
notopodium and a neuropodium morphologically equivalent
to, but less distinct than, those of Capitellids. If this be
admitted, then the assumed homology of the “‘ Seitenorgane ”’
of Capitellids with the dorsal cirrus of Glycerids falls, as the
two structures are not in the same morphological position,
and have not the same relationship to the respective rami of
the parapodia, for the dorsal cirrus is an appendage on the
dorsal side of the notopodium, whereas Hisig contends
that the “ Seitenorgan ”
side of the neuropodium.

represents the cirrus on the dorsal

After a careful examination of the lateral sense organs of
Scalibregma, I have come to the conclusion that they are
not very intimately related to either ramus of the para-
podium, they occupy a position between the two rami.
Moreover, a study of the excellent figures which Hisig has
given of these organs in the Capitellide shows that they are

THE ANATOMY OF SCALIBREGMA INFLATUM. 279

not related to the neuropodium, and, indeed, if they are
connected with one or other division of the parapodium, their
relation is rather with the dorsal ramus. than with the
ventral. The sense organs in several of the Capitellide are
considerably nearer the notopodium than the neuropodium,
especially in the thoracic region (see, for example, EHisig,
1887, Taf. ii, fig. 8; Taf. xxiv, fig. 3; and Taf. xxvii, fig. 16).
The sense organ may be separated from the neuropodium by a
considerable interval, containing some other structure, e. g.
see Hisig’s Taf. xvi, fig. 83, which gives a lateral view of the
abdomen of Dasybranchus, each of the sense organs of
which is situated some distance from the corresponding
neuropodium, and a gill is interposed between the neuro-
podium and the sense organ. The retractor muscle of the
sense organ is, in some genera, derived from the same
group of muscles as the protractors of the notopodial sete, as
in Scalibregma (see p. 273, and PI. 13, fig. 9). In these
cases the retractor of the “ Seitenorgane ” is attached to the
inner end of the notopodial setal sac (see Hisig’s figure of
Notomastus lineatus, Taf. x, fig. 10). The muscles of
the neuropodial sete are never in any way connected with
the lateral sense organs.

There is evidently, therefore, little reason for regarding
the lateral sense organs of Capitellids even as closely related
to the neuropodium, as in those cases in which the sense
organs are to some extent associated with one of the rami of
the parapodium, the association is invariably with the noto-
podium and not with the neuropodium. I conclude, however,
that the sense organ is not to be regarded as an appendage
of, or as intimately associated with, either ramus of the
parapodium ; it is on neutral ground between the two rami.

Tam unable to suggest any alternative to Hisig’s hypothesis
to explain the origin of lateral sense organs in Polycheetes. It
may be pointed out that these organs have arisen in a well-
protected position near the path of a large nerve—the an-
nular nerve. In course of time the epidermis of this pro-
tected area has become much more sensitive than the less

VOL, 45, PART 2.—NEW SERIES. U

280 J. H. ASHWORTH.

favourably situated cells around, and the former has gra-
dually received a larger nerve supply. The definite sense
organ has resulted from a gradual conversion of a number of
these protected epidermal cells, perhaps at first into fusiform
sense cells, similar to those met with on the cirri or on the
general body surface of some Polychetes. These became
further differentiated forming rod-like sense elements, and
the nervous apparatus connected with their bases became in-
creasingly complex. As shown above (p. 274), when the
sense organ appears in the newly-formed segments near the
posterior end of Scalibregma, it is distinguishable only by
reason of the differentiation of very few epidermal cells into
rod-like sense cells, and the presence beneath them of certain
nerve-cells; it is never cirriform at any period of its growth.

17. NEPHRIDIA.

The character of the nephridia of Scalibregma is prac-
tically unknown. All our information regarding these organs.
is contained in Danielssen’s account (1859), in which they are
described as the female reproductive organs. In his speci-
men, which was a large one, ke found forty to forty-two
pairs of tubular yellowish bodies, one pair in each segment
of the animal “from the sixth to the anal segment.’ Those
lying in the sixth to the thirteenth were larger than the
others, being about four lines (8 mm.) long. These sac-like
bodies were ciliated internally and filled with an enormous
number of roundish cells, many of which contained yellowish-
green granules. Danielssen believed these organs to be
ovaries from which all the ova had been discharged into the
ccelom, as at that time of the year (June) the ccelomic fluid
contained an enormous number of ova.

The nephridia are almost hidden from view by the oblique
muscles which are present in each chetigerous segment
throughout the body (fig. 14). The nephridia of Scali-
bregma are not large sac-like organs as in Arenicola, but
slender loops, each formed by a tube bent once upon itself,

THE ANATOMY OF SCALIBREGMA INFLATUM. 281

The nephridium may, for purposes of description, be divided
into four regions: (1) the funnel, (2) the fine straight tube
which leads directly backwards from the funnel into (3) the
loop formed by a U-shaped tube, the two limbs of which are
parallel and close together, the second limb leading into (4)
the short terminal tube of the nephridium (fig. 18). The only
part of the nephridium visible without dissection is the loop,
which is generally seen through the intervals between the
oblique muscle bands.

The funnel is very small; even in large worms it is seldom
more than ‘4 mm. in diameter. The opening is generally
directed ventrally. Both lips are simple, and bear no pro-
cesses. The dorsal lip is larger than the ventral one, and
overhangs the aperture like a hood. Occasionally in large
nephridia the dorsal lip is slightly folded. The funnel is
difficult to see because it is partially hidden by several
blindly ending processes of the afferent nephridial vessel,
which are bound to the funnel by a strand of connective
tissue which represents the septum in the post-diaphragmatic
segments (see p. 255). The funnels of the first three
nephridia are situated on the anterior faces of the second,
third, and fourth diaphragms respectively.

The funnel leads into a short ciliated tube, which generally
runs directly backwards. ‘This portion, even in the largest
nephridia, is only about 1 mm. long, and its lumen is small
(about ‘04 mm. in diameter).

The two limbs of the loop of the nephridium, which form
the excretory part of the organ, are closely applied together.
Hach is ciliated. The lumen of the straight tube of the
nephridium undergoes a gradual enlargement as this tube
merges into the first limb of the loop, the lumen of the latter
portion being about ‘08 mm. wide in large specimens. A
little more than halfway along the first limb of the loop
there is a distinct narrowing of the lumen, which throughout
the second limb is only about ‘04 mm. in diameter, so that
the two limbs of the loop may be easily distinguished in
section by the relative sizes of their cavities (see fig. 16).

282 J. H. ASHWORTH.

The loop is nearly 4 mm. long in the largest nephridia. In
young nephridia the lumina of the two limbs of the loop are
about equal in size.

‘’he terminal portion of the nephridium is generally bent
almost at right angles to the loop. It is seldom longer than
‘5mm. It is not distinguished from the adjacent portion of
the loop by any external structural character, except in a few
cases in which there is a slight dilation of the terminal tube
just before reaching the nephridiopore (fig. 18).

The first nephridinm is very small, being only 1°56 mm.
long ina large worm. The sixth is usually the largest, being
4to5mm.long. In most of the nephridia of large speci-
mens and in the first ten or fifteen nephridia of specimens
about 15 mm. long the loop is the most obvious part of the
nephridium, the straight tube being only one half to one
fourth its length; but the young nephridia found in the pos-
terior segments have very short loops, considerably shorter
than the straight tube of the same nephridium (fig. 21).

The nephridiopores are small oval apertures. The first is
situated just below and slightly anterior to the fourth neuro-
podium, but this is so minute that it can usually be seen only
in sections. The others may usually be found either in sur-
face view of favourable spirit specimens, or in specimens
cleared in oil (fig. 5). The pores are about -04—:06 mm.
along their longer diameter (see also p. 244).

Histology.—The lips of the nephrostome are lined by a
single layer of elongate columnar cells with well-marked
nuclei. The cilia are better developed on the dorsal lip
(fig. 19). The straight tube behind the nephrostome is lined
by almost cubical ciliated cells, the nuclei of which are small
and spherical, and lie close to the lumen of the tube. The
cells of the loop are larger than those of the preceding
portions of the nephridium. In surface view they appear
pentagonal or hexagonal, and are closely fitted together at
their borders. Their protoplasm contains (in preserved
specimens) numerous cavities which in life were probably
filled with excretory substances (fig. 20). In some speci-

THE ANATOMY OF SCALIBREGMA INFLATUM. 288

mens there are numerous small masses of excretory granules
distributed throughout the cells of the loop. These granules
appear dark brown or black when seen in masses, but yellow
or light brown when examined singly. In some specimens
the deposits are in the form of yellowish needle-shaped
crystals. The nuclei of these cells are very small, and
situated very close to the lumen. The terminal tube closely
resembles the adjacent part of the loop, except that there are
fewer concretions in the former. On approaching the
nephridiopore the wall of the tube becomes thinner (fig. 20),
and in some specimens the cells of this part are not ciliated.
There is no muscular tissue in the walls of any part of the
nephridium.

In some specimens Coccidian parasites, which in section
strongly resemble ova, are embedded in the cells of the loop.

The nephridium is covered by a very thin layer of peri-
toneal epithelium, and the blood-vessels seen on the funnel
only he between the peritoneum and the bases of the excre-
tory cells.

Blood Supply.—The nephridia are supplied with blood
by branches of the afferent branchial vessels which are given
off from the dorsal vessel (as in the case of the first three
nephridia) or by the segmentally arranged branches of the
ventral vessel (fig. 14). The latter vessels usually bifurcate
near the setal sacs, one branch passing to the nephrostome
and the other to the body-wall; the latter vessel usually bears
blind outgrowths which partially obscure the nephrostome.

The nephridia receive only a small amount of blood, the
whole of which apparently goes to the funnel (and to the
rudimentary septum on which the gonads are formed). I
have not found vessels on any other part of the nephridium.

The first three nephridia return blood to the second, third,
and fourth efferent branchial vessels. ‘he nephridia of the
fifteenth and following segments return blood to the sub-
intestinal vessels.

It is interesting to compare Danielssen’s observations with
the foregoing. He evidently saw only the loops of the

284 J. H. ASHWORTH.

nephridia, the two limbs of which are indicated in some of
the anterior nephridia of his figure (pl. 1, fig. 1); in these
nephridia the anterior limit is drawn broader than the pos-
terior, as is usually the case in large specimens. He appa-
rently overlooked the first two nephridia, which are small
and difficult to see, as they are partially hidden by the dia-
phragms. He describes the first nephridium as situated in
the sixth segment and immediately behind the last dia-
phragm. ‘This is really the third nephridium, as may be seen
from a glance at fig. 14. The yellow granules observed by
Danielssen in the cells of the nephridia were probably excre-
tory.

18. RepropuctTIvE ORGANS.

Danielssen (1859, pp. 73—76) described Scalibregma as
hermaphrodite. He examined the tubular nephridia, and
concluded they were ovaries from which all the ova had
been discharged into the celom. The parapodial glands
present on the dorsal and ventral cirri from the sixteenth to
within a few segments of the posterior end of the animal
were mistaken by him for testes, and the minute rods which
they contain were supposed to be spermatozoa. Secali-
bregma is not hermaphrodite, it is dicecious. As pointed
out above (p. 280), the paired segmental tubular organs are
nephridia, and the structures in the cirri are (p. 247) modified
epidermal glands which, instead of producing a liquid mucous
secretion, give rise to rod-like bodies which may be dis-
charged on the surface of the cirri.

Tor a considerable time I was unable to locate the gonads,
as they are extremely small. The nephrostome is connected
to the body-wall by a thin sheet of tissue, which probably
represents the septum of the segment. On the surface of
this strand, and especially in the region of the smaller lip of
the funnel, there is a collection of loosely arranged cells pro-
duced by proliferation of the peritoneal cells covering the
base of the funnel and the neighbouring portion of the

THE ANATOMY OF SCALIBREGMA INFLATUM. 285

septum. These cells are the very young reproductive cells.
They are found only in the region of the smaller lip of the
nephrostome, and for a little distance along the outer side of
the straight tube of the nephridium where this organ is in
contact with the rudimentary septum. Gonads are present
on all the fully grown nephridia, but those on the first and
second nephridia are very small. The gonads connected
with the first three nephridia are situated on the smaller lips
of the funnels and on the neighbouring portions of the an-
terior faces of the last three diaphragms.

‘The nuclei of the reproductive cells are large and deeply
staining.

It is not possible, I think, to distinguish the sex of a
specimen by examination of the cells of the gonad, because
they are shed into the ccelom when so small that they have
not become sufficiently differentiated to be recognisable as
either young ova or young spermagonia. It is only after an
examination of the more mature sexual products usually
found in the ceelom that the sex of the specimen can be
determined.

The reproductive cells leave the gonad when about 10—
12. in diameter. They increase in size in the ccelom, and
by the time they have reached 15 in diameter their nature
may be determined, as in males division of the cells now
takes place, and in females the cells are recognisable as
young ova. The ovum continues to grow in size, and its
protoplasm—which up to this point has been clear and homo-
geneous — gradually becomes loaded with spherical yolk
granules about 1 in diameter, which stain deeply with
hematoxylin. The largest ova in my specimens are ‘12 mm.
in diameter; these are probably almost ripe. The nucleus is
excentric and vesicular, about 30 in diameter, and has a
prominent deeply staining nucleolus. The peripheral layer
of protoplasm is almost free from yolk granules, but these
granules are moderately uniformly distributed throughout
the other parts of the protoplasm. ‘lhe vitelline membrane
of these ova is thin.

286 J. H. ASHWORTH.

The spermatozoa develop exactly as in Arenicola (Gamble
and Ashworth, 1898, p. 32, and pl. 5, figs. 29—34). The
spermagonia fall into the ccelom, and after several divisions
give rise to spherical or disc-shaped hollow masses of sper-
matids. The central cavity of each mass contains a small
coagulum, the remains of the blastophore.

By what means the genital products escape is difficult to
say. The nephridia are much too small to serve as oviducts,
as the width of the lumen of the narrow tube immediately
following the nephrostome is only about one third the dia-
meter of a ripe ovum. It is possible that the spermatozoa
escape by means of the nephridia. he escape of genital
products has been seen by Danielssen, who observed that
one of his specimens extruded eggs through a small rupture
in the body-wall which appeared when the animal was
strongly contracted.

19. THe Famity SCALIBREGMID2.

The family of the Scalibregmidz was established by Malm-
gren (1867, p. 186) to contain Scalibregma inflatum,
Rathke, and Humenia crassa, Oersted, there being only
one species of each genus known at that time. Since then
other species and genera have been described as allied to the
foregoing, and have been included in the family, although in
some cases their characters do not agree with those of the
original genera in one or more important respects. The
classification of this family is therefore at present in con-
siderable confusion, and this is further increased by the
almost inextricable entanglement of the three principal
genera—Scalibregma, Humenia, and Lipobranchius.
The position is rendered more difficult by the fact that some
of the species of these genera have been only briefly de-
scribed, and are known only from the external characters of
a single specimen, and this sometimes a mutilated one.

The original description of Scalibregma by Rathke
(1843, p. 182), and of Humenia by Oersted (1844, p. 99),

THE ANATOMY OF SCALIBREGMA INFLATUM. 287

mentions the presence of gills as one of the diagnostic cha-
racters ; and Levinsen (1883, p. 133) includes this among the
characters of the family Scalibregmide, but qualifies the
statement by adding that gills are present only in older
worms, or may also be present in younger individuals.
Hansen (1882, pp. 34, 35) described a gill-less worm
obtained by the Norwegian North Atlantic Expedition as
Scalibregma (?) parvum, and McIntosh (Kumenia
jeffreysii, 1868, p. 419; and KE. reticulata, 1885, p. 360),
Théel (E. longisetosa, 1879, p. 49), and Ehlers (EK. glabra,
1887, p. 169) have referred other gill-less specimens to the
genus Eumenia. Levinsen (1883) renamed Théel’s speci-
men Scalibregma longisetosum; but this is not satis-
factory, as an abranchiate species is thus placed in a genus
one of the distinctive characters of which is the presence of
gills. Cunningham and Ramage (1888, p. 655) recognised
that specimens similar to those named by McIntosh Eu-
menia jeffreysii did not belong to the genus Humenia;
they considered them to constitute a new genus—Lipo-
branchius. ‘The absence of gills in his specimens (which
were 30)—37 mm. long) was carefully considered by 'Théel
(1879) before naming them Humenia longisetosa. Pos-
sibly with a view of accounting for the absence of gills in
these specimens, or at any rate of minimising the value of
these organs as diagnostic, he states that the gills of Eu-
menia crassa do not begin to grow until the animal has
attained a length of 40—50 mm. The absence of gills in E.
Jongisetosa might, therefore, be due to the fact that the
specimens were young ones in which gills would have ap-
peared later. But this seems scarcely probable, and Théel’s
statement regarding the formation of the gills of EK. crassa
at a comparatively late stage of the animal’s growth also
seems extraordinary, as in many branclhiate Polychetes the
gills are formed in early life when the animal is quite small,
e.g. a specimen of Arenicola marina 4 mm. long already
bears the full number (thirteen pairs) of gills. 'lhéel’s state-
ment, moreover, does not agree with that of at least one

288 J. H. ASHWORTH.

author, viz. Johnston (1865, p. 221), who describes a speci-
men of E. crassa which bears gills and is only one inch and
a quarter (about 31 mm.) long. Iam indebted to Dr. Théel
for two specimens of H. crassa, 29 and 35 mm. long respec-
tively, each of which bears four pairs of gills. Another
specimen in my possession, obtained off the south coast of
Nova Scotia, is 31 mm. long, and also bears the same number
of gills. It would, I think, be better to consider the gills as
one of the diagnostic characters of the genus Humenia, and
to consider that Théel’s specimens, named by him EH. longi-
setosa, do not really belong to this genus, from which it
differs in other respects (see p. 292).

S. Joseph (1894, p. 103) has evidently resolved to maintain
the branchiate character of the genera Scalibregma and
Eumenia, for he divides the tamily of the Scalibregmidz
into two sections :

(1) Those with gills—Scalibregma, Humenia.

(2) Those without gills—Sclerocheilus, Lipobran-
chius.

In the last-named genus he would place L. jeffreysii,
Cunn. and Ram. (= Humenia jeffreysii, McIntosh), and
two other abranchiate species of Humenia, viz. KE. reti-
culata, McIntosh, and HE. glabra, Ehlers. ‘he genus Lipo-
branchius in his hands thus becomes a somewhat hetero-
geneous assembly, the members of which agree in general
shape and absence of gills, but differ in other characters
quite as important, e.g. the prostomium and parapodia. 5S.
Joseph’s classification is therefore not entirely satisfactory,
and requires some modification.

Before proceeding further it will be advisable to review
the chief characters of each of the genera already known,
beginning with the best known genera and species.

‘The genera Scalibregma and Humenia are very similar,
and it is not easy to find many characters by which they may
be distinguished.

Taking the species S. inflatum and KH. crassa as typical
of the respective genera, we may say that the former is more

THE ANATOMY OF SCALIBREGMA INFLATUM. 289

or less arenicoliform, while the latter is maggot-like and
tapers only slightly at each end. The most obvious external
difference is found on examining the parapodia. In S. in-
flatum the parapodia from the fifteenth segment to the
posterior end project outwards some distance from the body,
forming vertical lamine bearing dorsal and ventral cirri.
Hach of the parapodia of E. crassa is formed by two mam-
millee which arise separately from the body-wall and bear
the sete; cirri are absent. There are also anal cirri in 8.
inflatum, but none in EK. crassa.

The gills of these two Annelids are closely similar. In
Oersted’s description of H. crassa it is stated that gills are
found in the six anterior segments, and Johnston states that
they are “confined to the first six segments.”

Théel (1879) finds only four pairs of gills. In each of the
three specimens in my possession (two from the coast of
Sweden and one obtained off Nova Scotia) I find only four
pairs of gills situated in the second to fifth chetigerous seg-
ments, thus exactly agreeing in number and position with
those of Scalibregma, ‘The first gill of my specimens of
H. crassa is small, and the fourth is the largest. It is
scarcely probable that any other gills would have been
formed in these specimens, which are about 30 mm. long and
practically mature, as determined by an examination of the
sexual products in the ccelomic fluid. here is, therefore,
some difference of opinion with regard to the number of gills
present in HKumenia crassa. It will be noted that John-
ston does not definitely state that there are six pairs. I
have been unable to procure a specimen bearing more than
four pairs of gills, and though inclining to the opinion that
this is the normal number of branchia, I cannot settle this
question definitely until a much larger number of specimens
is available for examination,!

1 Cunningham and Ramage (1888, p. 655, and pl. 42, fig. 18) describe and
figure a specimen named E, crassa, dredged in the Firth of Forth, which

differs in several respects from specimens described by other authors. The
parapodia of their specimen are lamelliform, project prominently from the

290 J. H. ASHWORTH.

The prostomium of Scalibregma inflatum may be
described as almost T-shaped, the two horizontal limbs of
the letter representing the tentacular processes. The head
of Eumenia crassa is more deeply divided in the middle
line, and its lateral angles are rounded, and not prolonged
into tentacular processes (fig. 13).

These two Polychetes agree in the position and relations of
the nuchal grooves and peristomium ; in the appearance and
structure of the skin; in possessing two kinds of setz, capil-
lary and furcate, in each of the rami of the parapodia ; in the
presence of a sense organ between the rami of each para-
podium.

Internally there is also a strong resemblance between these
two worms. (My specimens of Humenia crassa were
unfortunately not in good condition internally, and I was
unable to determine some of the finer structural details.)
They agree in the following respects :—the general form and
relations of the alimentary canal, the situation of the four
anterior diaphragms, the musculature, the non-ganglionated
nerve-cord, the nephridia, each with simple funnel leading
into a slender tube bent once upon itself, a considerable part
of the proximal limb of the loop being wider than the distal
limb.

Scalibregma inflatum and HKumenia crassa differ
only (so far as their anatomy is at present known) in the

body-wall (see fig. 18 8), and bear large flattened leaf-like dorsal and ventral
cirri. All other authors agree that the parapodia of E. crassa are without
cirri. In the Firth of Forth specimen there are six pairs of gills stated to
occur in front of the notopodium of the first six chetigerous somites. In
other recorded specimens of this Polychete the gills are situated behind the
corresponding notopodia. The prostomium of the Forth specimen terminates
in two diverging tentacles, thus differing from others (see above, p. 290).
‘These points taken in conjunction with the fact that the specimen described
by Cunningham and Ramage is an elongate worm gradually tapering
from about the eighth or tenth segment to the posterior end, whereas all
other recorded specimens are maggot-like, shows that this specimen, if it be
B®. crassa, is quite different from any other example of this animal recorded
from the time of Oersted (1844) up to the present.

THE ANATOMY OF SCALIBREGMA INFLATUM. 291

shape of the prostomium, the character of the parapodia, and
the presence in the former genus of parapodial and anal
GiEst:

Lipobranchius jeffreysii agrees closely with Hu-
menia crassa, except that in the former there are no gills.
The absence of gills cannot be ascribed to the youth of the
specimens, for three of those in my possession are 32, 38, and
40 mm. long respectively, and are almost mature, having ova
‘O09 mm. in diameter in the coelomic fluid. These cannot,
therefore, be regarded as immature specimens of KH. crassa
upon which gills would afterwards have been formed, and
moreover gills are present in specimens of E, crassa before
they reach this size; there are four pairs of gills in a specimen
29 mm. long. Lipobranchius jeffreysii agrees with EK.
crassa in the shape of its prostomium, and the character of
its parapodia, which are without cirri; in fact, given a speci-
men from which the first six segments have been removed, it
would be a matter of some difficulty to determine to which
genus the specimen in question belonged.

Among the specimens of Scalibregma sent to me from
the United States National Museum there are five small
worms 4°6 to 8 mm. long, which have no gills, but otherwise
are indistinguishable from Scalibregma. I find that these
are not, as I at first supposed, young specimens ; some at least
are almost sexually mature, e. g. the specimen 8 mm. long
contains large ova (‘1—'11 mm. in diameter). These, there-
fore, cannot well be regarded as young specimens of Scali-
bregma; had they been such their gills would already have
been quite obvious structures, for Scalibregma acquires
its gills at an early age ; a specimen 5 mm. long already bears
the full number (four pairs) of well-developed gills. If,
therefore, the branchiz are to be regarded as one of the
diagnostic characters of the genus Scalibregma, these
abranchiate specimens do not belong to the genus.
pose to call them Pseudoscalibregma,

Several Polychtes have been described which agree with
Scalibregma in general characters, but are without gills;

I pro-

292 J. H. ASHWORTH.

these might also be placed in the genus Pseudoscali-
bregma. S. parvum, Hansen, 8S. (Humenia) longise-
tosum, Théel, and Eumenia reticulata, McIntosh, prob-
ably belong here.

S. parvum agrees very closely with the abranchiate spe-
cimens of Scalibregma in my possession in the shape of
the prostomium and in the parapodia, which from the twelfth
segment to the posterior end of the animal form projecting
laminze, each bearing a dorsal and a ventral cirrus.

Théel’s Humenia longisetosa differs markedly from E.
crassa in several respects, e. g. in the former the prosto-
mium is prolonged at each side into a well-marked tentacular
process; and though the first eleven pairs of parapodia are
small, those of the twelfth and following segments bear leaf-
like cirri. Levinsen (1883, p. 183) and von Marenzeller
(1892, pp. 401, 426) have realised that Théel’s specimens
are more closely allied to Scalibregma than to Humenia,
and have accordingly renamed them Scalibregma lon-
gisetosum,. ‘The latter author suggests that the specimens
(which are about 12 mm. long’) described by Hansen as 8.
parvum are merely young forms of Théel’s species, and
from a comparison of the descriptions and figures of the pro-
stomia and parapodia I consider the evidence supports this
view very strongly. S. parvum and 8. longisetosum
are, however, abranchiate, and should be separated on that
account from the genus Scalibregma.

Kumenia reticulata, McIntosh, is evidently nearly re-
lated to the foregoing, for its prostomium is prolonged into
tentacles, and its parapodia from the fifteenth segment
onwards to the posterior end of the body form projecting
lamellae, which McIntosh compared to those of 'Théel’s speci-
mens. McIntosh (1885, p. 361) remarks that “one of the
specimens presented the aspect, dorsally, of Scalibregma
without the branchie.”

EKumenia glabra, Ehlers, differs so considerably from
any of the species mentioned above, that it is doubtful
whether it should be included in any of the hitherto de-

THE ANATOMY OF SCALIBREGMA INFLATUM. 293

scribed genera.! Its prostomium is drawn out into two well-
marked tentacular processes, but its skin appears to be
smooth, and (in the figures) bears no signs of secondary
annulation, and the parapodia do not project prominently
from the body. The posterior end of the animal forms a thin
tail, upon the end of which the anus opens. It is distin-
guished from Eumenia and Lipobranchius by its prosto-
mium and skin, and from Scalibregma by its parapodia
and skin.

There still remain for consideration two members of the
family, viz. Sclerocheilus minutus, Grube, and Lipo-
branchius intermedius, S. Joseph. The former was
discovered and briefly and somewhat incorrectly described
by Grube? in 1863, but this description has been revised and
extended by S. Joseph (1894, p. 103). The animal is small,
only about 5—20 mm. long, and is found living in oyster-
shells. The head is of moderate size, and bears two blunt
tentacular processes. ‘There are eversible nuchal organs at
the sides of the prostomium. The parapodia contain capil-
lary and furcate sete, and those of the second segment (the
peristomium being acheetous) also contain stout acicular sete
which are curved near the tip. Behind the twenty-second
segment each parapodium bears a small digitiform cirrus
below the neuropodium. This is not a gill, as it contains no
vessels ; it is evidently sensory, as indicated by the presence of -
fine stiff hairs upon it. The skin is sculptured as in Scali-
bregma. ‘here are no gills. There are five® (or rarely
six) cirri around the anal aperture. The alimentary canal
resembles that of the other members of the family except
that there are no cesophageal pouches. ‘The brain and nerve-
cord agree with those of the other members of the family,

1 Ehlers (1887, p. 170) evidently doubted whether this animal should be
included in the genus Eumenia, for he says, “Ich stelle diese Art vorlaufig
in die Gattung Eumenia.”

* “ Beschreibung neuer oder weng bekannter Anneliden, Sechster Beitrag,
‘Archiv fiir Naturgeschichte, Jahrg. xxix, Band i, p. 50. Berlin, 1863.

3 Grube describes and figures four anal cirri,

2

294. J. H. ASHWORTH.

and the nephridia are formed on exactly the same plan as
those of Scalibregma, each nephridium being a delicate
tube, the excretory portion of which is once bent upon itself.
S. Joseph believes that the nephridia act as genital ducts,
but this seems improbable, if not impossible, as the lumen of
the nephridium is too small to permit the passage of a ripe
ovum (judging from S. Joseph’s figures, pl. v, figs. 137,
142). The animal derives its name from two plate-like pig-
mented structures on the head, which Grube believed to be
“ horny” and protective, but S. Joseph describes them as
eyes.!

This animal is most closely allied to the gill-less forms
of Scalibregma (Pseudoscalibregma), with which it
agrees in general shape, in the characters of the prostomium,
furcate sete, nephridia, and nervous system, but differs from
them in possessing strong sete in the second segment, the
presence of eyes (?), and the absence of dorsal cirri.

According to 8. Joseph (1894, p. 113), his new species
Lipobranchius intermedius is very similar in almost all
respects to Sclerocheilus minutus, with the exception
that the former bears no eyes upon the head and no eirri
upon the parapodia. It seems to me that this animal is not
a Lipobranchius; it differs from that genus in at least two
important respects, viz. the shape of the prostomium and the
‘ possession of strong acicular sete in the first parapodium.
This animal is more nearly allied to Sclerocheilus than to
any other member of the family of Scalibregmide. It may
for the present be named Asclerocheilus intermedius, a
name which indicates its relationship to Sclerocheilus,
and at the same time reminds us that the pigmented plates,
the distinctive character of the latter genus, are absent from
the former.

It would have been better had the preparation of the
following table of characters and classification been post-

1S. Joseph (p. 105) states that these pigmented areas occur on the dorsal

surface of the head; while Grube describes and figures them (p. 50, and Taf.
v, fig. 34) on the ventral face of the prostomium near the mouth.

THE ANATOMY OF SCALIBREGMA INFLATUM. 295

poned until I had been able to examine many more speci-
mens of some of the species therein mentioned. There
seems little prospect of obtaining further material for some
considerable time, and I have therefore appended the table,
although it cannot be regarded as a final one, as it is defi-
cient in several respects. It is based largely on the results
recorded in the former part of this paper, supplemented by
the diagnosis of the authors responsible for the various
species.

The Scalibregmide form a moderately compact family, the
characters of which may be thus stated.

ScCALIBREGMID®.

Limnivorous Polycheta, arenicoliform or maggot-like in
shape. Gills, if present, confined to the first five (or six,
according to Johnston) segments. Prostomium small, in
some drawn out at its antero-lateral angles into short pro-
cesses; in others its two lobes are blunt and rounded;
bordered laterally by the two nuchal grooves through which
the eversible nuchal organs may be protruded.

Parapodia consist of almost identical notopodia and neuro-
podia, each bearing sete of two kinds, viz. simple capillary
setee, and furcate setee with unequal, barbuled, pointed limbs.
Between the two rami of each parapodium there is in most
species a small sense organ, which may be withdrawn into a
shallow depression of the epidermis. ‘lhe segments are sub-
divided into annuli, and the skin, especially in the anterior
part of the body, is usually raised into squarish or oval eleva-
tions. Internally there are four transverse diaphragms
situated at the posterior end of the first, second, third, and
fourth chetigerous segments. The eversible pharynx is
smooth, there being no armature whatever. ‘The heart is
median and is an enlargement of the dorsal vessel. There is
a pair of nephridia in each segment except in a few of the
anterior ones. Hach nephridium is a narrow ciliated tube,

VoL, 45, PART 2,—NEW SERIES, x

296. J. H. ASHWORTH.

the excretory part of which is bent once upon itself; the
funnel is small. Dicecious, gonads microscopic.

Found in the temperate seas of the Northern and Southern
Hemispheres, and in the colder seas of Northern Hurope.

I. Body arenicoliform, prostomium more or less T-shaped,
the antero-lateral angles being drawn out to form short ten-
tacular processes.

A. The parapodia of the segments behind the twelfth or
fifteenth project prominently at right angles to the body,
each forming a laminate appendage bearing a dorsal and a
ventral cirrus. Body often swollen anteriorly.

1. Scalibregma.—Gulls present on the anterior seg-
ments.

S. inflatum, Rathke (Oligobranchus roseus,
Sars). Four pairs of gills in chetigerous segments
2—5. Four anal cirri.

S. brevicauda, Verrill. Four pairs of gills on seg-
ments 2—5. No anal cirri described.

S. (?) abyssorum, Hansen. Anterior part only
known, from one specimen. ‘Three pairs of gills on
segments 2—4, ‘I'he prostomium bears very short
tentacles.

2. Pseudoscalibregma.—No gills. No anal cirri de-
scribed.

P. longisetosum. (Humenia_ longisetosa,
Théel.) The eleventh or twelfth and following
parapodia bear cirri.

P. parvum. (Scalibregma [?] parvum, Hansen.)
Cirri present on the parapodia of the twelfth and
following segments. ‘This may be a young form of
the preceding species.

P. reticulatum. (Humenia reticulata, McIn-
tosh.) Cirri on the fifteenth and following seg-
ments.

B. The parapodia do not form laminate appendages, and
are without dorsal cirri. Ventral cirri, if present at all, are
digitiform and confined to the posterior region. Hach para-

THE ANATOMY OF SCALIBREGMA INFLATUM. 297

podium is composed of two simple rounded elevations, in
which the setz are lodged.

3. Sclerocheilus.—Two triangular masses of pigment
on the prostomium. Strong, curved, acicular sete are found
in the first chetigerous segment. The parapodia in the
posterior segments of the animal bear ventral cirri. Four
anal cirri present.

S. minutus, Grube.

4. Asclerocheilus.—Pigment masses absent. Curved
acicular sete. finer than those of Sclerocheilus are present
in the first three chetigerous segments. No ventral cirri.

A. intermedius. (Lipobranchius interme-
dius, S. Joseph.)

II. Body of animal maggot-shaped. Prostomium dis-
tinctly divided anteriorly into two by a median groove ; each
half of the head is blunt and rounded; there are no tenta-
cular processes. The parapodia do not form projecting
lamella, and do not bear cirri. Hach parapodium is com-
posed of two simple rounded elevations which arise sepa-
rately from the body-wall. Anal cirri absent.

5. Humenia.—Four (or six) pairs of gills present on
chetigerous segments 2—6 (or on the first six according to
Johnston).

K. crassa, Oersted.

6. Lipobranchius.—Gills absent.

L. jeffreysii, Cunningham and Ramage. (EHu-
menia jeffreysii, McIntosh.)

20. AFFINITIES OF THE SCALIBREGMIDS.

The Scalibregmidz, as references in the previous portion
of this paper have shown, resemble the Arenicolide in many
of their structural features, and they also agree in some
points with the Opheliide. These three families have several
points im common, as they are limnivorous, and present
certain of the peculiarities characteristic of such Polychetes,

298 J. H. ASHWORTH.

They have a spacious ccelom, subdivided anteriorly by dia-
phragms, and non-septate in the middle part of the animal ;
the alimentary canal consists of an eversible pharynx, fol-
lowed by an cesophagus bearing a pair of lateral glandular
outgrowths, a dilated stomach with glandular walls, and a
straight intestine, usually with a ventral groove ; the blood-
vessels of the middle region of these animals are so arranged
as to leave the stomach considerable freedom of movement,
all the blood-vessels to the stomach passing to its ventral
wall, and being arranged so that they can accommodate them-
selves to the backward and forward motion of this part of the
gut.

The Scalibregmide agree with the Arenicolide in the
above-named characters, and in the general shape of the
body, the subdivision of the segments into annuli, the sculp-
turing of the skin, the small-lobed prostomium (which, in
some specimens of Scalibregma, is quite comparable to
that of Arenicola claparedii), and the presence (in Scali-
bregma and Humenia) of gills of a similar type. The
brain and non-ganglionated nerve-chain of Scalibregma
resemble those of the marina section of the genus Areni-
cola. There are also points of difference between these two
families which are of considerable importance. In the
Scalibregmide the two rami of the parapodia are practically
identical, but in the Arenicolide the notopodium is a
conical elevation, and the neuropodium a cushion-like out-
growth. In members of the latter family the neuropodium
bears crotchets only, and the notopodium bears capillary
sete; in the Scalibregmidee both rami of the parapodia bear
two kinds of sete, capillary and furcate, the latter being
characteristic of the family. In some of the Scalibregmidz
the parapodia form laminate appendages bearing dorsal and
ventral cirri, which are absent in Arenicola (cirri are very
rarely seen in the posterior region of American specimens of
A.cristata). The gills of Scalibregma and Humenia
are confined to the first five (or six) segments; they are
never present in the first seven segments of Arenicola,

THE ANATOMY OF SCALIBREGMA INFLATUM. 299

The dorsal vessel of Scalibregma is dilated just behind
the fourth diaphragm to form the heart, which is therefore
a median structure, thus differing entirely from the hearts of
Arenicola, which are paired, and not directly connected
with the dorsal vessel. The nephridia of Scalibregmide are
minute but numerous, and the simple microscopic funnel
leads into a slender U-shaped excretory tube. The nephridia
of Arenicola are fewer in number, and are wide sacs, each
with a large funnel fringed with ciliated vascular processes.
Several of the Scalibregmidze bear complex, segmental,
lateral sense organs, which are not found in Arenicola.

The Scalibregmidz have only a few features in common
with the Opheliidz. Besides the points mentioned above as
common to the three limnivorous families, they agree in the
great development of the muscles of the ventral body-wall
(especially in Humenia and Lipobranchius), the nerve-
cord without ganglia, the dorsal heart, and the principal
features of the circulatory system. ‘The resemblances may
be best seen on comparing the Ophelid Ammotrypane
cestroides with Humenia and Scalibregma. ‘he
Ophelidee and Scalibregmide differ in their nephridia, those
of Opheliids being comparatively few and sac-like; in their
prostomia, that of Ophelids is a single conical outgrowth; in
their parapodia and sete. The Scalibregmide have little in
common with any other family of Polychetes.

We may therefore say that the Scalibregmide agree in
several respects with the Arenicolide and Opheliidz, and it
is difficult to say that they are more related to one of these
than to the other, though, on the whole, there are rather
more features in which the Scalibregmidz agree with the
Arenicolidee (e. g. shape, secondary annulation, sculpturing
of skin, character of the gills when present, prostomium,
brain, and nerve-cord) than with the Opheliide. The Scali-
bregmide, however, are clearly distinguished from these
families by the presence of the peculiar furcate sete in the
parapodia, and by their numerous delicate nephridia, among
other characters.

300 J. H. ASHWORTH.

21. Summary or RESULTS.

1. Specimens of Scalibregma inflatum from the nor-
thern seas of Kurope and America are not distinguishable by
any essential and constant character from those obtained by
the ‘‘ Challenger ” in southern seas.

2. The parapodia of the segments posterior to the fifteenth
or sixteenth bear dorsal and ventral cirri which contain large
unicellular glands, the secretion of which is in the form of
elongate fusiform rods. The fine-pointed tips of the glands
pass between the epidermal cells and open on the free
surface. The notopodium, its cirrus, and some of its sete
are formed in advance of the neuropodium and its corre-
sponding parts.

3. The sete of each ramus of the parapodium are of two
kinds: (1) simple capillary bristles, the distal third of which
bears (in unworn examples) a large number of minute hair-
like processes; (2) rather stouter, shorter sete, furcate
distally, the two unequal pointed limbs bearing on their inner
faces a number of curved barbules. Both these kinds of
setze are found in the earliest recognisable parapodia. Fur-
cate setz of this type are practically restricted to the family
Scalibregmidee.

4, The dorsal vessel is dilated at two points to form the
blood-reservoir and the heart. There is no cardiac body in
the heart.

5. The brain consists of an anterior lobe in relation to the
prostomial epithelium, and two posterior lobes, each applied
to the inner side of the corresponding nuchal organ. The
ganglion cells are found chiefly on the dorsal and ventro-
lateral faces of the brain. The anterior lobe gives off a pair
of nerves to the tentacles ; the cesophageal connectives arise
from the middle region of the brain; the posterior lobes give
off nerves which run along the sensory epithelium of the
nuchal organs. In old specimens the fibrous part of the
brain becomes proportionately larger and more complex, and
the ganglion cells become aggregated into groups.

THE ANATOMY OF SCALIBREGMA INFLATUM. 301

6. The nerve-cord is situated close to the epidermis, and is
non-ganglionated. The ganglion cells are distributed along
the whole length of the cord on its lateral and ventral faces.
The cord gives off in each segment a pair of nerves lying in
each interannular groove, and a pair lying in the cheti-
gerous segment. The latter supplies the cirri, sense organs,
and seta] sacs.

7. A pair of lateral sense organs is present in each cheti-
gerous segment. Hach sense organ is a small eminence
rising from the base of a shallow depression bordered by
prominent lips of epidermis. From a darker area in the
centre of the papilla the delicate sense hairs arise. They are
implanted in exceedingly slender columnar cells, closely and
regularly arranged. These cells are in connection at their
inner ends with pyriform or fusiform ganglion cells, which
occupy the axis of the sensory papilla. Around and below
these are numerous deeply-staining nuclei, which are prob-
ably, as Kisig showed, the nuclei of multipolar ganglion cells,
the protoplasm of which forms the fine network upon which
the nuclei are situated. The sense organ receives a mode-
rately stout branch from the spinal nerve, which runs along
the middle of the chetigerous annulus. The sense organ
may be withdrawn into the depression in the epidermis by
the contraction of a retractor muscle attached to its base.
In very young sense organs, such as are found in the last two
or three segments of a specimen about 15 mm. long, the rods
which occupy the space of only one or two epidermal cells
are exceedingly small, and do not bear sense hairs. ‘There
are only two or three small ganglion cells at their bases,
accompanied by about twenty of the deeply staining nuclei
mentioned above. In the next segment anterior to this the
rods are more obvious, and one segment further forward the
sense hairs may be seen on their distal ends. In old sense
organs the axial part of the organ is more fibrous, and the
deeply staining nuclei are very numerous.

8. Similar sense organs are present in Humenia crassa
and Lipobranchius jeffreysii.

302 J. H. ASHWORTH.

9. These organs are similar to those described by Hisig in
the Capitellide, except that in the latter there are no large
ganglion cells beneath the rods. he sense organ is not, as
Hisig supposed, morphologically equivalent to a neuropodial
dorsal cirrus. It does not form part of the neuropodium ; it
occupies a position between the two parapodial rami, but it
may be connected by means of its retractor muscle to the
base of the notopodial setal sac. Hisig believes that the
sense organ is homologous with the dorsal cirrus of the
Glyceride, and that the parapodium of Gycerids is a neuro-
podium only (the notopodium being absent) equivalent to the
neuropodium of Capitellide. This view cannot be supported ;
the parapodium of the Glyceride is essentially biramous, its
division into notopodium and neuropodium being less obvious
than in many Polychetes, owing to the close approximation
of the two rami. (For further details of the discussion
see p. 276.)

10. Hach nephridium is a delicate ciliated tube opening
into the ccelom by a minute simple nephrostome. The ex-
cretory part of the tube is bent once upon itself. ‘There is a
pair of nephridia in each chetigerous segment except the
first three.

11. Scalibregma inflatum is diewcious, and not herma-
phrodite, as described by Danielssen. ‘The gonads are formed
by proliferation of the cells covering the septum by which
the nephrostome is attached to the body-wall. The genital
cells fall from the gonad at a very early stage, and complete
their growth in the coelomic fluid. In their structure and
stages of growth the ova and spermatozoa closely resemble
those of Arenicola. Humenia crassa and Lipobran-
chius jeffreysii are also dicecious, and their genital pro-
ducts are similar to those of Scalibregma.

12. The prostomium is an important character in the
classification of the Scalibregmide. It affords, along with
the nature of the parapodia, the most reliable means of deter-
mining whether a given specimen belongs to the Scal
bregma—or to the Humenia—section of the family.

THE ANATOMY OF SCALIBREGMA INFLATUM. 303

13. The Scalibregmidz resemble the Arenicolide and
Opheliudze in several respects, but several of these characters
may be largely due to the limnivorous mode of life of the
members of these three families. The following characters
are common to them :—the spacious ccelom non-septate in the
middle region of the body; the eversible pharynx followed
by an cesophagus bearing a pair of glandular outgrowths; a
dilated stomach with glandular walls and a straight intestine
with a ciliated ventral groove; the blood-vessels of the
middle region of the gut are arranged so as to allow the
swinging movement of the stomach.

The Scalibregmidz agree with the Arenicolide also in ine
annulation and sculpturing of the body-wall, the prostomium,
the brain, and non-ganglionated nerve-cord. They differ in
their parapodia, sete, the position of the gills, the heart, and
the nephridia.

The Scalibregmidee resemble the Opheliide in their mus-
culature, the non-ganglionated nerve-cord, and the circula-
tory system; but they differ in their prostomia, nephridia,
parapodia, and sete.

The Scalibregmide, although allied to some extent to the
Arenicolid, and to a less degree to the Opheliide, form a
separate and compact family, one of the most characteristic
features of which is the presence of the peculiar furcate sete
in the parapodia.,

22. LITERATURE.

1843. Ratuxe, H.—“ Beitrage zur Fauna Norwegens,’’ ‘Nova Acta Aca-
demiz Cesare Leopoldino-Caroline, Nature Curiosorum,’ tome
xx, p. 182. Breslau, Bonn, 1843.

1844. Orrstep, A. S.— Zur Classification der Annulaten, mit Beschreibung
einiger neuer oder unzulanglich bekannter Gattungen und Arten,”’
‘Archiv fir Naturgeschichte,’ Jahrgang x, Bandi, p. 99. Berlin,
1844.

304

1846.

1859.

1865.

1867.

1868.

1868.

1873.

1879.

1882.

1883.

1885.

1887.

1887.

1887.

1888

J. H. ASHWORTH.

Sars, M.—‘Fauna littoralis Norvegiz,’ Heft 1, p. 91: “ Beschrei-
bung des Oligobranchus roseus, einer neuen Form der Riicken-
kiemenwirmer.” Christiania, 1846.

Dantetssen, D. C.—“ Beretning om en Zoologisk Reise i Sommeren,
1858, Anatomisk-physiologisk Undersogelse af Scalibregma
inflatum,” ‘ Kongl. Norske Videnskselsk. Skrifter,’ Band iv, Heft
2. Trondhjem, 1859.

Jounston, G.—“ Catalogue of the British Non-parasitical Worms in
the Collection of the British Museum.” London, 1865.

Matmeren, A. J.—* Annulata Polycheeta,” ‘Kongl. Vetenskaps-
Akademiens Forhandlingar,’ No. 4, p. 186. Helsingfors, 1867.

CLaPpAREDE, E.—‘‘ Les Annélides Chétopodes du Golfe de Naples.”
Geneve et Basle, 1868.

McInrosu, W. C.—‘ On the Structure of the British Nemerteans and
some New British Annelids,”’ ‘'Transactions of the Royal Society of
Edinburgh,’ p. 419, vol. xxv. Edinburgh, 1868.

Verritt, A. K.—* Report upon the Invertebrate Animals of Vineyard
Sound and the Adjacent Waters,” ‘United States Commission of
Fish and Fisheries,’ p. 605. Washington, 1873.

Tuten. Hj.—‘ Les annélides polychetes des mers de la Nouvelle-
Zemble,” ‘Kongliga Svenska Vetenskaps-Akademiens Handlingar,’
Ny Foljd, Bandet xvi, No. 3. Stockholm, 1879.

Hansen, G. A.—‘ The Norwegian North-Atlantic Expedition,’ Part
vii, Annelida, p. 84. Christiania, 1882.

Levinsen, G. M. R.—“<Systematisk-geografisk Oversigt over de
nordiske Annulata, ete.,”’ ‘ Vidensk. Meddel fra den Naturh. Foren
i Kjobenhavn,’ p. 133. Kjobenhavn, 1883.

McIntosu, W. C.—‘‘‘Challenger’ Report,” vol. xii, Annelida
Poiycheta. London, 1885.

Euers, E.—“ Reports of the Results of Dredging in the ‘ Blake,’ ”
‘Report on the Annelids ;’ ‘Memoirs of the Museum of Compara-
tive Zoology at Harvard College,’ voi. xv, p. 169. Cambridge,
U.S.A., 1887.

Kisic, H.—‘ Monographie der Capitelliden des Golfes von Neapel.’
Berlin, 1887.

Wixty, A.—“ Beitrage zur Anatomie und Histologie der limnivoren
Anneliden,” ‘Kongliga Svenska Vetenskaps-Akademiens Hand-
lingar,’ Ny Foljd, Bandet xxii, No. 1. Stockholm, 1887.

Cunnincuam, J. T., and Ramace, G, A.—‘‘ The Polychaeta Sedentaria
of the Firth of Forth,” ‘Transactions of the Royal Society of Edin-
burgh,’ p. 655, vol. xxxili. Edinburgh, 1888.

THE ANATOMY OF SCALIBREGMA INFLATUM. 305

1892. Marenzetier, KH. von.—‘ Polychaten von Ostspitzbergen,” ‘ Zoolo-
gisch. Jalirb.,’ Band xvi, p. 397. Jena, 1892.

1894, Saint-JosEepu, pE.—‘‘ Les Annélides Polychétes des Cotes de Dinard,”
‘Annales des Sciences Naturelles: Zoologie,’ Series vii, tome 17,
p. 103. Paris, 1894.

1898. Sarnt-JOsEPH, DE.—“ Les Annélides Polychétes des Cotes de France,”
‘Aunales des Sciences Naturelles: Zoologie,’ Series viii, tome 5,
p- 209. Paris, 1898.

1898. Gamexy, F. W., anp Asuworrn, J. H.—‘* The Habits and Structure
of Arenicola marina,” ‘Quarterly Journal of Microscopical Sci-
ence,’ vol. xli, part 1, p. 1. London, 1898.

1900. Gamsug, F. W., anp Asuwortn, J. H:i—“ The Anatomy and Classifi-
cation of the Arenicolide,” ‘ Quarterly Journal of Microscopical
Science,’ vol. xliii, part 38, p. 419. London, 1900.

EXPLANATION OF PLATES 13—15,

Illustrating Dr. J. H. Ashworth’s memoir on “ ‘The Ana-
tomy of Scalibregma inflatum, Rathke.

List of Reference Lelters.

An, Anus. Ant. Cr. Anterior cornu of brain. 7. R. Blood reservoir.
Br. Gill. Br. Aff. Branchial afferent vessel. Lr. Lif. Branchial efferent
vessel. Cirr. dn. Anal cirri. Cirr. D. Dorsal cirrus of parapodium. Cirr.
G/. Gland situated in the cirrus. Cirr. VY. Ventral cirrus of parapodium.
Cel. Celom. Cel. Epith. Celomie epithelium. Cut. Cuticle. Dphm.'—,
Diaphragms. D.V. Dorsal blood-vessel. Zp. Epidermis. Zp. Gl. Gland-
cells of epidermis. Zp. Pap. Epidermal papille. /. Fibrous part of nerve-
cord. Gang. C. Ganglion cells. Gex. C. Genital cells. Gr. V. Ventral
groove of intestine. H¢. Heart. Jn¢. Intestine. Jvé. N. Nerves (? see
p. 268) of intestine. Jvé. S. Intestinal sinus. A/o. Mouth. J. Long.
Longitudinal muscles of body-wall. J/. O4/. Oblique muscles. MW. PA,
Retractor muscles of pharynx. J/, Protr. Protractor muscles of the setal
sacs. WV. Nucleus. WN. Avnul. Annular nerve situated in the interannular
groove. N.C. Nerve-cord. NV. Chet. Annul. Annular nerve of chetigerous
annulus. Neur. S. Neuropodial seta. Ng/. Neuroglia. Ngl. Sk. Neuro-
glial sheath. N//m. Neurilemma. Nm. Neuropodium. J. A/. C. Nucleus
of multipolar ganglion cell. MW. Nuc. Nerve to nuchal organ. NV. O, Ex.

306 J. H. ASHWORTH.

ternal opening of nephridium. Moém. Notopodium. Noé. S. Notopodial
sete. Nph. Nephridium. Nps. Nephrostome. NV. Tent. Nerve to pro-
stomial tentacle. Nuc. Gr. Nuchal groove. Nuc. Retr. Retractor muscle of
nuchal organ. (#. (sophagus. G@. Gl. Gsophageal glands or pouches.
Per. Peristomium. PA. Pharynx. Post. Cr. Posterior cornu of brain.
Prost. Prostomium. Prost. Tent. Prostomial tentacle. 2. Rods of lateral
sense organ. §. Aff. V. Segmental afferent vessel (from ventral vessel).
S. Cap. Capillary seta. S. Hf V. Segmental efferent vessel (to subintestinal
vessel). Sep¢. Septum. 8. Furc. Furcate seta. 8. H. Sense hairs. S. 0.
Lateral sense organ. JS. O. Retr. Retractor muscle of sense organ. Sp. NV.
Spinal nerve. Stom. Stomach. Swbiné. V. Subintestinal vessel (sinus). 7’.
Tail segment. V. Mfes. Ventral mesentery (imperfect). V.V. Ventral
blood-vessei. J, If, III, IY... LX. Somites beginning with the first
chetigerous,

PLATE 138.

All figures, except Fig. 18, are drawn from specimens of Scalibregma
inflatum.

Fic. 1.—The large Norwegian specimen, 56 mm. long, seen from the left
side to show the external features, prostomium, parapodia, cirri, sete, gills,
segmentation and annulation, the sense organs, etc. The nephridiopores
(NW. O.), the first of which opens on the fourth chatigerous annulus, are very
small, and are not well seen in this drawing (see Fig. 5). There are sixty-
one segments in the specimen. X 45.

Fic. 2.—Ventral view of a very regular American specimen which, if com-
plete, would have been about 20 mm. long. The prostomium, peristomium
and first seven chetigerous segments are seen. The mouth (Jo.), bordered
by epidermal papilla, and the secondary annulation of the skin are seen. The
nerve-cord runs along the middle line of the median depressed area. xX 22.

Fic. 3.—Dorsal view of the anterior end of the same specimen to show the
prostomium, nuchal grooves, peristomium, and the first and second cheeti-
gerous segments, the latter bearing the first pair of gills. x 22.

Fic. 4.—The first gill of the specimen drawn in Fig. 1, along with a per-
tion of the second chatigerous and succeeding annuli. Only the dorsal half
of the gill is fully drawn; the ventral half is cut down to the bases of the two
main branches. x 16.

Fie. 5.—A portion of the tenth chetigerous segment of the left side of the
same specimen. Note the four annuli, the skin of which is subdivided into
squarish elevations, the epidermal papilla, the prominent lips of the setal
sacs, the sense organ, and the nephridiopore (V.0.). x 16.

Fic. 6.—Ventral view of the posterior end of a specimen 13 mm. long,
showing the pygidium or tail segment (7'), the newly formed body segments

THE ANATOMY OF SCALIBREGMA INFLATUM. 307

anterior to this, and the cirri and secondary annulation of the older segments.
The dorsal cirri are formed earlier than, and are larger than, the ventral cirri.
The slightly raised area in the median line marks the position of the nerve-
cord (which is non-ganglionated; the appearance of ganglionation presented
by the specimen is due to the contraction of the body-wall). The anal cirri
on one side have been cut off close to their bases. x 80.

Fic, 7.—Posterior aspect of a parapodium from the specimen drawn on the
preceding figure. The parapodium was situated three segments in{ front of
the oldest segment shown in Fig. 6. x 80.

Fig. 8.—The thirty-fifth parapodium of the specimen drawn in Fig. 1.
The dark area (Cirr. Gl.) in each cirrus marks the position of the gland, which
is seen by transparency through the epidermis of the cirrus. The sense organ
is situated_in and hidden by the small papilla seen between the bases of the
notopodium and neuropodium. xX 20.

Fic. 9.—The thirtieth parapodium of a specimen 14 mm, long, which was
stained, cleared, and compressed in order to bring the muscles into the same
plane as the other structures. The typical parts of a parapodium are shown
—the dorsal and ventral cirri with their large gland-cells, the notopodium
and neuropodium each with simple and furcate sete, the sense organ and its
retractor muscle, and the protractor muscles of the setal sacs. x SO.

Fie. 10.—Five gland-cells from one of the glands shown in the preceding
figure. The pointed ends of these unicellular glands pierce the epidermis and
open on the free surface. Mach gland contains a large number of rod-like
bodies. The nucleus of one of the cells is seen near its inner end. x 500.

Fic. 11.—Rods from the parapodial glands of the specimen shown in Fig. 1.
Compare their size with that of the rods from a much younger specimen shown
in the preceding figure. x 500.

Fic. 12.—A thick longitudinal section (25 » thick) through an annulus of
a specimen 14 mm. long, to show the unicellular glands of the skin, the cir-
cular and longitudinal muscles, and the annular nerves. WN. Chet. Annul. is
a section of the nerve of the chetigerous annulus, which is larger than the
nerve (NV. Annul.) supplying the following annulus. x 270.

Fig. 13.—Dorsal view of the anterior end of a specimen of Eumenia
crassa 29 mm. long, to show the prostomium, nuchal grooves, peristomium,
and the first and second chetigerous segments, the latter bearing two small
gills. xX 9.

PLATE 14.

All the figures relate to Scalibregma inflatum.

Fie. 14.—Dissection of the anterior portion of the specimen drawn in
Fig. 1, to show the general arrangement of the internal organs. The principal
features shown are the four anterior diaphragms, the alimentary canal, and

308 J. H. ASHWORTH,

cesophageal pouches, the vascular system and the nephridia (see also p. 254).
The neuropodia and the afferent nephridial vessels are drawn only in the five
segments immediately behind the last diaphragm, and the oblique muscles,
which are present in all the chetigerous segments, are shown only in the last
three segments drawn on the right side. The incomplete ventral mesentery,
which binds the ventral wall of the stomach to the body-wall near the nerve-
cord, is omitted. In order to prevent confusion, the course of the blood.
vessels running on the left side of the body-wall is not fully shown behind
segment 15. Some of the folds in the wall of the stomach are probably arti-
ficial. xX 3.

Fie. 15.—A section passing almost horizontally through the head, pro-
stomium, and first and second cheetigerous somites, to show the brain with its
anterior and posterior cornua lying in the prostomium, the nuchal organ, and
its retractor muscle, ete. The ganglion cells which cover the brain are in
close relation to the epithelium of the prostomium and of the nuchal organ.
The section has not passed through the whole length of the anterior lobe of
the brain, only its posterior portion is seen here. X 80.

Fic. 16.—Trausverse section of the specimen shown in Fig. 1 passing
through the twenty-fifth chetigerous annulus. ‘The various parts of the
parapodium and the sense organ are seen on the right (cf. Fig. 9). In the
ventral divisions of the ccelom sections of the small tubular nephridia are
seen; on the left a nephrostome has been cut through. At the base of the
ventral groove in the intestine are two cords (Zv¢. V.) seen in section. From
their structure and general appearance they appear to be nervous. x 24,

Fie. 17.—Transverse section of the nerve-cord and surrounding structures
from a specimen 14 mm. long. The fibrous part of the cord is partially sub-
divided by a neuroglial ingrowth. The ganglion cells are situated chiefly on
the ventral side of the cord. The origin of a spinal nerve is seen on the
right. Note also the nuclei of the longitudinal muscle-fibres. x 200.

Fie. 18.—A nephridium from the twentieth segment of the specimen seen
in Fig. 1. The lumen of the nephridium is shown, as seen in optical section
Attached to the nephrostome is the rudimentary septum bearing the genital
cells. x 40.

Fic. 19.—Section of a nephrostome from the thirtieth segment of the same
specimen. ‘The dorsal lip is seen on the right, it is more strongly ciliated
than the ventral lip. On the left is the septum bearing the genital cells.
The blood-vessel lies between the ccelomic and the ciliated epithelium.
x 250.

Fie. 20.—Section of a small nephridium at the junction of the excretory
and terminal portions, The latter is on the right; its cells are cubical or
even slightly flattened, while those of the excretory portion are columnar,

aud have vacuoles which in life probably contained excretory products,
X 250.

THE ANATOMY OF SCALIBREGMA INFLATUM. 309

Fre. 21.—A very small nephridium drawn in situ on the body-wall after
the preparation had been stained and cleared. Note the septum accompany-
ing the blood-vessel to the nephrostome. On the right of the nephrostome
the nuclei of genital cells are seen. The excretory part of the nephridium—
the loop—is at this stage very short. The nephridium is closely invested by
a delicate cceelomic epithelium, the nuclei of which are seen at intervals.
x 300.

Fic. 22.—Section of an ovum from a specimen 35 mm. long. The peri-
pheral layer of protoplasm is almost free from yolk granules. x 150.

Fie. 23.—Section of a portion of the wall of the cesophageal pouches, to
show the blood-sinus enclosed between two epithelial lamella. x 50.

Fic. 24.—Some of the cells of the wall, showing the cavities in which the
granules of secretion are usually found. They have been dissolved from these
cells leaving the cavities empty. x 300.

PUADE 153

Fic. 25.—Two furcate sete. A in full view; Bin profile. The portion
shown in the figure represents only the distal twelfth of each seta. x 800.

Fic. 26.—The tip of a capillary seta, to show the hair-like processes. The
portion figured is only J; of the seta. x 800.

Fie. 27.—Section of a very young sense organ. The organ was situated
‘25 mm. from the posterior end of a specimen 13 mm. long. There was ouly
one chetigerous segment behind this one, and in both these segments the
notopodial sete only were formed, as shown. ‘This is the earliest, recognisable
sense organ in the specimen. Note the minute rods (2.), the three ganglion
cells, and the nuclei of the multipolar ganglion cells (V.J/.C.). For further
explanation see p. 274. x 600.

Fie, 28.—A rather thick longitudinal section of a sense organ about 3 mm.
from the posterior end of the specimen 13 mm. long. The sense-hairs are
now seen. This sense organ contains an exceptional number of large ganglion
cells. x 500.

Fic. 29.—Transverse section of a parapodium of the same specimen
situated 2 mm. from the posterior end. The retractor muscle of the sense
organ is seen. See also p.272. x 500.

Fic. 30.—Transverse section of an old sense organ from the twenty-fifth
segment of a specimen 35 mm. long. ‘lhe fibrous part of the organ is pro-
portionately larger. ‘I'he ganglion cells are situated nearer to the rods in this
organ than in most others. Note the stout nerve supplying the organ enter-
ing on the ventral side, and turning almost through a right angle into the axis
of the organ, See p. 274. x 200,

Giri a-pe ‘ * 7 TY
'
i
v
,
‘
:
2
‘
; ,
:
}
iy

= =
_

-

ae er

a 7

iy - ae
. r
met ®
; : j
£ i
‘
5
a ax

ro ad - :
——
—_ =

- : 7 —

a Vr.

b : rf
wy = a“
ris i
j
¢
i
we
‘ m
= “e

com

a on - - . eo Se ae eet Se ow

ae $uart Sourn Mion Set, VAGUS ATE

Fig.l.

pail

CorD.

ir

Con Arw.--f-

J HAshworth del : AS. Hoth, Lith’ London

yee
—~ a =) 2
vue ee
. ai |
ue ;
a
a
7 | ar

) | s | Dart Sermons VASE

a

: “MProt.

---- PR

Sas SMe sie

eee
”*>-Nphg2?

"== M0dz.

“A

aie
HYNIY
ee -- ML ong.

Hi
) ae
) S—--M Gre
Zip. Gl
Cot. Nyt. Sh Ep Sp.W.
JH Ashworth de]

A.S.Huth Lith® London.

J.H.Ashworth del.

Quart. Sourn Morr Si. VAASNE PLAS.

A.S.Huth, Lith’ London.

7

.

PELVIC GIRDLE AND FIN OF EUSTHENOPYERON. 3811

On the Pelvic Girdle and Fin of
Eusthenopteron.

By

Edwin 8. Goodrich, M.A.,
Fellow of Merton College, Oxford.

With Plate 16.

TuroucH the kindness of Mr. A. Smith Woodward,! I
have recently had the opportunity of looking through the
fossil fish acquired by the British Museum since the Cata-
logue was published. Amongst these was found a specimen
of Kusthenopteron foordi, Whit., showing the endo-
skeleton of the pelvic girdle and fin, of which I here give a
description. ‘he interest attaching to this fossil is con-
siderable, since, of all the numerous extinct fish usually
included in the group “ Crossopterygii,” it is the first and
only one in which the parts of the skeleton of the pelvic girdle
and its fin have been found complete and in their natural
relations.”

The specimen (P. 6794) of which both the slab and the
counterslab have been preserved, comes from the Upper
Devonian of Canada. In it can be made out the skeleton of
the pelvic girdle and fin of the right side, in a fairly com-
plete and well-preserved condition, as represented in PI. 16,
fig. 1, natural size.

1 To Mr. Smith Woodward I am also indebted for constant help when
working in his Department.

2 The skeleton of the pelvic fin of Megalichthys has to some extent been
made known by Cope, Miall, and Wellburn (2, 5, and 9), and the essential
structure of that of Kusthenopteron has been briefly described by Traquair (7).

VOL, 40, pARY 2,—NEW SERIES. y

312 EDWIN 8S. GOODRICH.

It will be seen at once that the skeleton of this fin closely
resembles that of the pectoral fin of the same species already
described and figured by Whiteaves and Smith Woodward
(10 and 12).1. In the pelvic fin (figs. 1 and F) we find an axis
consisting of three segments or mesomeres, and three pre-
axial endo-skeletal rays or parameres. Of these the first two
and largest are borne by the first and second mesomeres
respectively, whilst the last is in a rudimentary condition,

Fic. A.—Ventral view of the pelvic girdle and fins of Heptanchus cine-
reus, Gm.; in this and the succeeding figures the complete skeleton is
exposed on one side only. For the lettering see the Explanation of
Plate 16.

being represented by a small rounded piece at the distal end
of the outer branch of the slightly bifurcated terminal
mesomere. A similar semi-lunar piece fits on to the axial
branch of the third mesomere. The whole skeleton of the
fin is formed, then, of an axis consisting of three large
segments, and a small terminal piece, and of two well-

1 This resemblance was pointed out by Traquair (%7), who writes “ A very
similar arrangement is found in the pelvic fin . . . ; here J find at least two

mesomeres, each bearing a paramere, there being, I think, also a probability
of the presence of a third or distal mesomere.”

PELVIC GIRDLE AND FIN OF EUSTHENOPTERON, 315
defined pre-axial rays, and probably a vestigial third
ray.

The chief difference between the pectoral and the pelvic
fin is, that whereas in the former there are post-axial
expansions on the first, third, and fourth axial segments,
in the pelvic fin no such post-axial process is visible in our
specimen (figs. 1 and 2).

The pelvic fin projects from the body as a free lobe of
considerable size, and is covered with scales similar to those
on the trunk. Round this lobe the extensive web of the fin
is supported by jointed bony dermal fin rays. On the pre-

Fic. B.—Ventral view of the pelvic girdle and fins of Chimera
monstrosa, L.
axial side the dermal rays are, as usual, stronger than on the
post-axial side. The pre-axial edge of the fin is straighter
and considerably longer than the post-axial, so that the fin is
not symmetrical about its skeletal axis either internally or
externally.

The pelvic girdle is represented on the right side by a
somewhat triangular elongated bone. It is pointed in front,
and widens out behind into a broad plate. The outer edge
is almost straight, whilst the inner edge is sharply curved
where the bone widens out. ‘l'o the posterior edge is
articulated the axis of the pelvic fin. The whole girdle
consists of two such bones, which in the living animal no

314. EDWIN S. GOODRICH.

doubt lay parallel to the ventral surface, with their sharp
ends pointing forwards and converging towards the middle
line. ‘he posterior expansions would also extend towards
the mid-ventral line (fig. F).

The structure of the skeleton of the pelvic girdle and fin
is of great importance as a taxonomic character. But before
attempting to discuss the value of these organs in deciding
the position occupied by Eusthenopteron in the scheme of
classification, it will be well to briefly compare the various
types of pelvic supports in the Fish series.

Fic. C.—Ventral view of the pelvic girdle and fins of Acipenser sturio, L.

Amongst the Elasmobranchs we find embedded in the
body-wall a median-ventral cartilaginous plate, to which the
pelvic fins are attached by a moveable joint (Fig. A). In the
Holocephali there is on each side an elongated cartilage
supporting the pelvic fin; these cartilages are joined to-
gether in the middle line by a ligament (Fig. B). A some-
what similar, but shorter, pair of cartilages is found in the
extinct Pleuracanthide (Fig. D). A specimen in the Oxford
Museum (Fig. E) shows particularly well their ligamentous
union in the middle line. Coming now to the ‘eleostomi,

PELVIC GIRDLE AND FIN OF KUSTHENOPYERON. 315

we find in the Chondrostei somewhat ill-defined, more or
less triangular, cartilaginous plates stretching from the
base of the pelvic fins towards the middle line (Fig. C).
Amia lepidosteus and all the “Teleostei,” in fact all

Fic. D.—Ventral view of the pelvic girdle and fin of a female Pleura-
canthus Oelbergensis, Fr. (from Fritsch).

the Actinopterygii, possess paired bony ventral plates sup-

porting the pelvic fins. These plates may be joined together

in front by a median cartilage, as in Gadus (Fig. J), or they

Fie. E.—Ventral view of the pelvic girdle of Pleuracanthus Gandryi,
Brogn.

may be merely united by ligament (Figs. H, I). In the

Dipnoi, on the contrary, the pelvic support is represented by

a median cartilage with two diverging branches, to which are

articulated the fins (Fig. K). Of the so-called Crossopterygi,

316 EDWIN S. GOODRICH.

the only forms in which the pelvic girdles are thoroughly
well known are the Coelacanthide and the Polypteride.! In
both these families the pelvic bones closely resemble those
of the Actinopterygn (Figs. G, L).

Concerning the morphology of these pelvic supports there
is considerable confusion. Whilst the older anatomists be-
lieved them, I think quite rightly, to be homologous, and the
representatives of the pelvic girdle of other fish, some modern
authors would have us believe that they are of quite different

Fic. F.—Ventral view of the pelvic girdle and fins of Eusthenopteron
Foordi, Whit., restored.
nature in the various orders of Pisces. They hold that
whilst, for instance, in the Selachii, Holocephali, and Dipnoi,
a true pelvic girdle is present, the supports in the Crossop-
terygii and Actinopterygii are, on the contrary, derived from
the fin skeleton itself.
Let us see what difficulties such views lead us into.
Whether we hold, with the advocates of the fin-fold theory,
1 Paired bony plates seem to have been present in Megalichthys (5, 9), and
Specimens 21,547 and P.6513 of the British Museum Collection show traces
of similar structures in Osteolepis and Glyptolepis,

PELVIC GIRDLE AND FIN OF EUSTHENOPTERON. 317

that the pelvic girdle originated as an ingrowth of the base of
the primitive fin skeleton, or whether, following Gegenbaur,
we consider it to have been derived from a gill arch, it will
be admitted that the girdle was primarily differentiated as a
right and a left support and fulcrum for the fin, and as a
point of attachment for the muscles whereby the fin is
moved. The girdle plate must have been from the first

G iS

lic. G.—Ventral view of the pelvic girdle and fins of Polypterus bichir,
Geoffr.

embedded in the ventral body-wall from which sprang the

free lobe of the pelvic fin.

Such appears to have been the structure of the paired pelvic
girdle of the Pleuracanthide (Fritsch [8] and Figs. D and
K), and such it is essentially at the present day in the Holo-
cephali (Fig. B). The development of the pelvic girdle in the
Selachii (Balfour [1], Mollier [6]) warrants the view that the
median cartilage there found has been formed by the fusion
of two originally separate halves. Presumably in this way
has also originated the median cartilage of the Dipnoi.

318 EDWIN S. GOODRICH.

It is not until we reach the Teleostomes that difficulties
arise. Davidofft held that the girdle proper is represented
in the Bony Ganoids and Teleosts by the cartilage at the
anterior ends of the long pelvic bones, which themselves
would be homologous with the metapterygium (basiptery-
gium) of the Selachian fin.

Weidersheim (11), considering the bones as metapterygial,
believes the girdle to be appearing in Polypterus as small

Fic. H.—Ventral view of the pelvic girdle and fins of Amia calva, Bon.
(Partly from Davidoff.)
paired or median cartilages (Fig. G) at the tip of the pelvic
bones. A somewhat similar little piece of cartilage, occasion-
ally found at the anterior extremity of the pelvic supports of
the Chondrostei, is supposed to have the same significance.
Rautenfeld (8), on the other hand, compares the basal sup-
port in the Ganoids to the propterygium of the Selachu.
Wiedersheim’s view, that the minute paired or median

1 Davidoff, ‘* Beitrage z. verg]. Anat. der Hinteren Gliedmasse der Fische.”
‘Morph. Jahrb..’ vols. v, vi, and ix.

PELVIC GIRDLE AND FIN OF EUSTHENUOPTERON. 319

cartilages often found in front of basal supports, represent
the first origin of the girdle, has not met with much favour
for many reasons. More especially the obvious objection
may be urged against it, that the pelvic girdle is already
fully developed in more primitive forms (Klasmobranchs).
Gegenbaur (4), whilst adopting the theory of the homology of
the basal supports with the metapterygium of the Selachian
fin, considers that the small anterior cartilages of the Ganoids
represent the last vestiges of the pelvic girdle, which has

Fic, 1.—Ventral view of the pelvic girdle and fins of Lepidosteus osseus,
L. (Partly from Davidoff.)

undergone degeneration, and may have entirely disappeared
in other Teleostomes.

Now Gegenbaur’s view seems to be no less open to
objection than Wiedersheim’s. For if we are to believe that
the girdle has disappeared and been functionally replaced by
bones derived from an ingrowth of the already differentiated
fin skeleton, we may well ask, what plausible reason can be
given for the substitution in the place of the girdle supports
of these new structures of very similar shape, and of perhaps

VOL. 49, PART 2,—-NEW SERIES. Z

320 EDWIN S. GOODRICH.

even larger size? During this important change what has
become of the muscles attached to the girdle for the moving
of the fin? Have they disappeared also, and been replaced
by others, or have they shifted their base of attachment on
to the basals? What evidence is there that the moveable
joint, where the base of the fin skeleton articulates with
the girdle, firmly embedded in the body-wall, has not always
been where it now is, but has been carried forwards at the
tip of the basal bones and lost its primitive function ? What
evidence is there that this primary articulation between the
moveable fin skeleton and the fixed pelvic girdle, has been

Fic. J.—Ventral view of the pelvic girdle and fins of Gadus morrhua, L.

replaced by a new joint between two different regions of the
fin skeleton itself ?

Moreover, is it credible that such a fundamental alteration
in the relations of the internal skeleton should have taken
place without a corresponding change in external shape ?
Here the evidence afforded by the structure of Kusthenop-
teron may be called in. On the one hand there seems to be
no reasonable doubt that the pelvic bones of this fish are
homologous with those (so-called basals) of Polypterus,
Coelacanthus, or Amia (Figs. G, L,and H). On the other hand,
it will, I think, be allowed that the moveable joint whereby

PELVIC GIRDLE AND FIN OF EUST'HENOPTERON. 321

the segmented axis of the fin of Eusthenopteron articulates
with the pelvic bone is strictly homologous with the corre-
sponding articulation at the base of the pelvic fin of Ceratodus
or Pleuracanthus.' We are, then, inevitably led to the con-
clusion that the pelvic supports, whether paired or unpaired,

Fic. K.—Ventral view of the pelvic girdle and fins of Ceratodus Forsteri,
Kr, (Partly from Davidoff.)

are homologous throughout the fish series. These structures

are similarly situated, fulfil the same functions, and are, as

far as we know, developed in the same way in all fish. In

1 The persistence of the same articulation between the girdle and fin
skeleton is obvious in the case of the pectoral limb of fishes.

pp EDWIN 8S. GOODRICH.

some cases, however, as in Dipnoi and Teleosts, they are well
differentiated ; in other cases, as in the Chondostrei, they are
ill-defined, and probably in a more or less degenerate con-
dition, not clearly marked off from the true fin skeleton.

To conclude, we may briefly mention the evidence afforded
by the structure of the skeleton of the pelvic fin as to the
systematic position of Husthenopteron. Unfortunately we
know hardly anything about the structure of the fin skeleton
of other extinct “Crossopterygi.” But from our know-
ledge of the Dipnoi, it may be inferred with some degree of
certainty that the skeleton ot the elongated lobed fins of
such forms as Glyptolepis and Osteolepis was built on the

Fie. L.—Ventral view of the pelvic girdle and fins of Holophagus gulo,
Huxley.

biserial archipterygial plan (distichopterygia). It is there-
fore of considerable interest to note that although in the
shape of the outline of the fin-web, and in the disposition
and structure of the dermal rays, the pelvic limb of Eus-
thenopteron approximates to that of the more highly
specialised ‘l'eleostomes (Actinopterygii); yet its internal
skeleton is probably to be interpreted as a modification of
the biserial archipterygium, with a distinct axis, im which
the post-axial endo-skeletal rays have been lost. Further,
the skeleton of the pectoral and of the pelvic fin of Kusthe-
nopteron still exhibit that close resemblance to each other
which is so marked a characteristic of the Dipnoan fins, and
presumably also of the more primitive forms from which they
have been derived.

In contrast to this we find in Polypterus and the Actinop-

PELVIC GIRDLE AND FIN OF EUSTHENOPTERON. 328

terygii a great and increasing modification in structure of
the fins. All trace of an axis is soon lost in the pelvic limb,
whilst at the same time the pelvic girdle bones in these fish
and the Celacanthide, assume that peculiar elongated and
flattened shape, widening out in front, which is so charac-
teristic.

Finally it may be pointed out that, whilst Husthenopteron
is undoubtedly closely allied to the Rhizodontide, judging
from the skeleton of the pelvic fin, it appears to be very far
removed from Polypterus, which probably belongs to the
‘Actinopterygian line of development.

List oF REFERENCES.

Bateour, I’.—‘ Comparative Embryology,’ vol. ii.

. Cork, E. D.—‘ Proc. W. 8. Nat. Museum,’ vol. xiv.

. Fritscu, A.—‘ Fauna der Gaskolile,’ Prague, vol. ii, 1889.

. GeGenBaur. C.—‘ Vergl. Anatomie der Wirbelthiere,’ vol. i, Leipzig,
1898.

5. Mraut, L. C._—‘ Quart. Journ. Geol. Soc.,’ vol. xl, p. 347.

6. Mo.tuier, S.—‘ Die paarigen Extremitaten,” ‘ Anat. Hefte,’ vol. i, 1893.

7. Traquatr, R. H.—* Devonian Fishes of Canada,” ‘ Geol. Mag.,’ vol. vii,

1890.
8. Raurenretp, KE. V.—‘* Ueber das Skelet der hinteren Gliedmassen,”’
‘Inaug. Dissert. Dorpat,’ 1882.

9. WetiBury, FE. D.—* On the Genus Megalichthys,” ‘ Proc. Yorks. Geol.
and Polyt. Society,’ vol. xiv, 1900.
10. Whitraves, J. F.—‘“‘ The Fossil Fishes of the Devonian Rocks of
Canada,” ‘ Trans. Roy. Soc. Canada,’ vol. iv, 1887, and vol. vi, 1888.
11. Wiepersuemm, R.—‘ Das Gliedmassenskelet der Wirbelthiere,’’ Jena,
1892.

12. Woopwanrn, A. 8.—‘ British Museum Catalogue of Fossil Fishes,’ pt. 2,
1891.

Pon

VOL. 45, PART 2.—NEW SERIES. AA

324 EDWIN S. GOODRICH.

EXPLANATION OF PLATE 16,

Illustrating Mr. Edwin 8. Goodrich’s paper ‘On the Pelvic
Girdle and Fin of Eusthenopteron.”

List oF Rererence Lerrers in Piate 16 Aanp TEXtT-FIGURES
A—Ti:

a. Axial mesomere. 6. Basipterygium; basal mesomere in fig. 2. ec.
Cartilage. d.7. Dermal fin ray. @.p. Lateral process. m.c. Median carti-
lage. p. Pelvic cartilage. p.f. Pelvic fin. pr.7. Preaxial endoskeletal ray.
p.s. Posterior expansion, pé.7. Postaxial endoskeletal ray. s. Scale. a.
Postaxial process.

Fie. 1.—Outer view of the right half of the pelvic girdle and of the right
fin of Eusthenopteron Foordi (Brit. Museum, No. P. 6794).

Fic. 2.—Outer view of the left pectoral fin of Husthenopteron Foordi
(Brit. Museum, 6796), copied from A. Smith Woodward.

4INS GY 18.

Wy

Y
Cb

If

) ; Tt VA A,
LuantFournJAMon:

fig. I.

== HN

A.5. Huth Lith? London

* ! :

«.. " s » : 7
4
mt ee :
a 4
: - <
e -
7 ‘
7 i ‘
5 . ‘ .
‘
: ; st
ore
‘ .
Dri * , :
= a Tek
) "
‘
- .
> ars .
i - . .
_

With Ten Plates, Royal 4to, 5s.
CONTRIBUTIONS TO THE KNOWLEDGE OF RHABDOPLEURA
AND AMPHIOXUS.
By E. RAY LANKESTER, M.A., LL.D., F.R.S.
London: J. & A. Cares: 7, Great pees ouldy Street.

Quarterly Journal of Microscopical

Science.

The SUBSCRIPTION is £2 for the Volume of Four Numbers;
for this sum (prepaid) the JourNAL is sent Post Free to any part
of the world.

BACK NUMBERS of the Journat, which remain in print, are
now sold at an uniform price of 10/-.

The issue of Supptument Numbers being found inconvenient,
and there being often in the Editor’s hands an acenmulation of
valuable material, it has been decided to publish this Journal at
such intervals as may seem desirable, rather than delay the appear-
ance of Memoirs for a regular quarterly publication.

The title remains unaltered, though more than Four Numbers
may be published in the course of a year.

Hach Number is sold at 10/-, and Four Numbers make up a
Volume.

London: J. & A. CHURCHILL, 7 Great Marlborough Street.

TO CORRESPONDENTS.

Authors of original papers published i in the Quarterly Journal
of Microscopical ° ‘Science receive tw enty-five copies of their
communication gratis.

All expenses of publication and illustration are paid by the
publishers.

As a rule lithographic plates, and not woodcuts, are used in
illustration.

Drawings for woodcuts should nor be inserted in the MS.,
but sent in a separate envelope to the Editor.

Contributors to this Journal requiring extra copies of their
communications at their own expense can have them by applying
to the Printers,

Messrs. ApLarD & Son, 224, Bartholomew Close, E.C., on
the following terms :

For every four pages or less—

25 copies : : 5/-
QO 55 ; : : : 6/-
BOOES i058 7/-

Plates, 2/- per 25 if uncoloured ; if colour -ed, at the same rate for
every colour.
Prepayment by P.O. Order is requested.
ALL COMMUNICATIONS FOR THE EDITORS TO BE ADDRESSED TO THE CARB
oF Mussrs-J. & A. Courcuitt, 7, Great Mariporoucs Srreet,
Lonpon, W.

THE MARINE BIOLOGICAL ASSOCIATION

OF THE

UNITED KINGDOM.
20%
Patron—HIS MAJESTY THE KING.
President—Prof. E. RAY LANKESTER, LL.D., F.R.S.

tO)

THE ASSOCIATION WAS FOUNDED “ TO ESTABLISH AND MAINTAIN LABORATORIES ON
THE COAST OF THE UNITED KINGDOM, WHERE ACCURATE RESEARCHES MAY BE CARRIED
ON, LEADING TO THE IMPROVEMENT OF ZOOLOGICAL AND BOTANICAL SCIENCE, AND TO
AN INCREASE OF OUR KNOWLEDGE AS REGARDS THE FOOD, LIFE CONDITIONS, AND HABITS
OF BRITISH FOOD-FISHES AND MOLLUSCS.”

The Laboratory at Plymouth
was opened in 1888. Since that time investigations, practical and scientific, have
been constantly pursued by naturalists appointed by the Association, as well as by
those from England and abroad who have carried on independent researches.

Naturalists desiring to work at the Laboratory

should communicate with the Director, who will supply all information as to
terms, &c.

Works published by the Association

include the following :—‘ A Treatise on the Common Sole,’ J. T. Cunningham, M.A.,
4to, 25/-. ‘The Natural History of the Marketable Marine Fishes of the British
Islands,’ J. I. Cunningham, M.A., 7/6 net (published for the Association by
Messrs. Macmillan & Co.).

The Journal of the Marine Biological Association

is issued half-yearly, and is now in its fifth volume (price 3/6 each number).

In addition to these publications, the results of work done in the Laboratory
are recorded in the ‘ Quarterly Journal of Microscopical Science,’ and in other
scientific journals, British and foreign.

Specimens of Marine Animals and Plants,

both living and preserved, according to the best methods, are supplied to the
principal British Laboratories and Museums. Detailed price lists will be forwarded
on application.

TERMS OF MEMBERSHIP.

ANNUAL MEMBERS ; : > £1 LT Operannum:

Lirz MEMBERS . : : : . 15 15 O Composition Fee.
FOUNDERS . = LOO O30 a ry
GOVERNORS (Life Members of Council) 500 O O 4 “5

Members have the following rights and privileges:—They elect annually the
Officers and Council; they receive the Journal free by post; they are admitted to
view the Laboratory at any time, and may introduce friends with them; they have the
first claim to rent a table in the Laboratory for research, with use of tanks, boats, &e. ;
and have access to the Library at Plymouth. Special privileges are granted to Governors,
Founders, and Life Members.

Persons desirous of becoming members, or of obtaining any information with
regard to the Association, should communicate with—

The DIRECTOR,
The Laboratory,
Plymouth.

MAR 7 1902

|New Series, No. 179 (Vol. 45, Part 8). Price 10s.

‘) ca FEBRUARY, 1902.

THE

QUARTERLY JOURNAL

, . OF

MICROSCOPICAL SCIENCE.

EDITED BY

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

HONORARY FELLOW OF EXETER COLLEGE, OXFORD; CORRESPONDENT OF THE INSTITUTE OF FRANCE
AND OF THE IMPERIAL ACADEMY OF SCIENCES OF ST. PETERSBURG, AND OF THE ACADEMY
OF SCIENCES OF PHILADELPHIA; FOREIGN

MEMBER OF THE ROYAL BOHEMIAN
SOCIETY OF SCIENCES,

AND OF THE ACADEMY OF THE LINCEI OF ROME;
ASSOCIATE OF THE ROYAL ACADEMY OF BELGIUM; HONORARY MEMBER
OF THE NEW YORK ACADEMY OF SCIENCES, AND OF THE
CAMBRIDGE PHILOSOPHICAL SOCIETY, AND OF THE ROYAL
PHYSICAL SOCIETY OF EDINBURGH } HONORARY

MEMBER OF THE BIOLOGICAL SOCIETY

OF PARIS;
HISTORY DEPARTMENTS OF THE BRITISH MUSEUM; LATE FULLERIAN
PROFESSOR OF PHYSIOLOGY IN THE ROYAL INSTITUTION OF GREAT BRITAIN.

DIRECTOR OF THE NATURAL

WITH THE CO-OPERATION OF

ADAM SEDGWICK, M.A.; F.RS.,

FELLOW AND TUTOR OF TRINITY COLLEGE, CAMBRIDGE$

W. F. R. WELDON, M.A., F.R.S.,

LINACRE PROFESSOR OF COMPARATIVE ANATOMY AND FELLOW OF MERTON COLLEGE, OXFORD}
LATE FELLOW OF ST. JOHN’S COLLEGK, CAMBRIDGE}

AND

SYDNEY J. HICKSON, M.A., F.R.S.,

BEYER PROFESSOR OF ZOOLOGY IN THE OWENS COLLEGE, MANCHESTER.

WITH LITHOGRAPHIC PLATES AND ENGRAVINGS ON WOOD.

“LONDON: |
J. & A. CHURCHILL, 7, GREAT MARLBOROUGH STREET. |
1902. |

Adiard and Son,] {Bartholomew Close.

“i | we

CONTENTS OF No. 179.—New Series.

MEMOIRS:

Dendrocometes paradoxus. Part 1—Conjugation. By SypNEY

J. Hicxsoy, M.A., D.&c., F.R.S., Beyer Professor of Zoology in

the Owens College, Manchester ; assisted by Mr. J. T. Wans-
wortH. (With’ Plates 17 and 18) ;

On the Oviparous Species of Onychophora. By ARTHUR Deer D. Se.,
F.L.S., Professor of Biology in the Canterbury College, University
of New Zealand. (With ‘Plates 19— 22)

A New aud Annectant rs of Chilopod. By R.I. Pochix, (With
"Plate 23) 2 : :

The Trypanosoma Brucii, ie Geena fount in Nace: or Tse. tse
Fly Disease. By J. R. Braprorp, F.R.S., and H. G. Priumer,
ELS. (from the Laboratory of the Brown Institution). (With
“Plates 24 and 25) . 5 :

Notes on Actinotrocha. By K. Rost Mzxox, idee Pro-
fessor, Presidency College, Madras. (With Plate 26)

Review of Mr. Iwaji Ikeda’s Observations on the Development,
Structure, and Metamorphosis of Actinotrocha

PAGE

325

363

417

449

473

4.85

DENDROCOMETES PARADOXUS. 325

Dendrocometes paradoxus.

Part I.—Conjugation.
By

Sydney J. Hickson, M.A., D.Se., F.R.S.,
Beyer Professor of Zoology in the Owens College, Manchester

Assisted by
Mr. J. T. Wadsworth.

With Plates 17 and 18.

INTRODUCTION.

THERE is no species of the class Acinetaria that can be
obtained with greater facility at all times of the year than
Dendrocometes paradoxus (Stein). This is due to the
fact that it is found with unfailing regularity on the gills of
our commonest fresh-water Crustacean, Gammarus pulex.
~ In addition to the fact that it is readily obtained, however,
it presents us with the further advantage of being attached
to a soft and tolerably transparent gill, which can be easily
seen with the naked eye, and manipulated without difficulty
on the slide and in the paraffin bath. The difficulties that
have usually to be overcome in preserving and staining the
Infusoria whole or in sections are, in this form, largely
obviated, and it is not very difficult, after the preliminary exa-
mination of the specimens attached to any one gill, to imbed
and cut them into a series of sections in any plane that may
be desired. Possessing these advantages, it was obvious
that a careful examination of the changes of the nuclei

VOL, 45, PART 5,—NEW SERIES. Z

326 SYDNEY J. HICKSON.

during conjugation and gemmation might yield results of
interest and importance.

I have had preparations of Dendrocometes under observa-
tion for a considerable number of years, but it was not until
1899 that I obtained specimens which clearly showed the
initotic figures of the division of the micronuclei, and then I
determined to investigate the matter with greater thorough-
ness. Imay explain that during the last two years I have
entrusted the mounting of the preparations, as well as the
staining and cutting of the sections, to Mr. J. I’. Wadsworth,
and that he has saved me an immense amount of time and
labour in pointing out to me the slides that exhibited interest-
ing features, and in keeping a catalogue of the preparations.
The discoveries therefore that are here recorded were, in the
first instance, made by him, and I have to acknowledge here
his unfailing assistance and perseverance in the investigation.

The structure of Dendrocometes has been investigated by
Biitschli (1), Wrzesniewski (27), Plate (21), Maupas (19),
Schneider (24), and Sand (28) ; but, notwithstanding their
excellent work, many points of interest and importance re-
main to be illustrated and described.

Dendrocometes paradoxus is found attached to the
gills of Gammarus at all times of the year, but in the summer
months there are usually fewer specimens on each gill than
in the spring and autumn. We have found that by keeping
about twelve or fourteen Gammarus in a shallow pie-dish
containing an inch or two of water, with a little mud and
waterweed at the bottom of it, the number of Dendrocometes
on the gills increases, and that in a fortmight’s time a con-
siderable number of pairs may usually be found in a state of
conjugation.

Dendrocometes appears to have a wide geographical dis-
tribution. I have found it myself on Gammarus at Oxford,
Cambridge, in the neighbourhood of Manchester, and on
specimens of Gammarus sent to me by Mr. Bolton from Bir-
mingham. It was observed by Prof. Lankester on Gam-
marus in the ponds at Hampstead. On the continent of

DENDROCOMETES PARADOXUS. S20

Europe it also appears to be generally distributed. It
certainly occurs in Belgium (Sand), France (Maupas), Ger-
many (Butschli, Plate, etc.), Russia (Wrz.). It also occurs
in North America (Biitschli). It is probable that the Den-
drocometes which occur on the gills of Gammarus
putaneus (Lachmann) may be a distinct species, but, apart
from this, Dendrocometes only occurs on the gills of Gam-
marus pulex, although exceptionally a few specimens
may be found attached to the hairs on the legs of the same
host (Sand).

Dendrocometes paradoxus has a planoconvex-shaped
body, and is provided with three, four, or five arms (PI. 17,
fig. 1). The morphology of these arms is a matter of some
dispute, but it is not an unreasonable view to regard them as
homologous with a bundle of Acinetarian tentacles. They
capture, kill, and partially digest the prey in the same manner
as the tentacles and suckers of other Acinetaria. The body
contains a single large meganucleus, and a variable number,
but usually three, micronuclei. ‘There is a single large con-
tractile vacuole. Reproduction is effected by an interesting
process of internal gemmation, first described by Biitschli
(1). The single planoconvex bud that is formed is fre-
quently called the “embryo,” but I think it is more appro-
priate to call it the “gemmula.” It escapes from the
parents without arms, but provided with a girdle of three
bands of cilia, and swims away.

From time to time, or perhaps, under certain conditions
only, nearly the whole substance of an individual escapes from
the pellicullar sheath and swims away from the gill ina form
which cannot be readily distinguished from.a gemmula.

With this brief introductory account of the structure and
reproduction of the animal I pass on to the phenomena of
conjugation, to which I have paid special attention.

Part I—Conjugation.

The phenomena of conjugation in Dendrocometes may be
briefly stated as follows ;

328 SYDNEY J. HICKSON.

Two individuals in proximity on a gill of Gammarus send
out simultaneously blunt lobe-like processes, which may be
called the conjugative processes. ‘These meet but do not
completely fuse, a distinct membrane delimiting the process
of each individual throughout the conjugation. ‘This mem-
brane does not prevent the fusion of the meganuclei nor of
the conjugative micronuclei in the latter stages, nor does it
prevent a certain amount of mixture of the cytoplasm of the
conjugating individuals.

Stage A (PI. 17, fig. 2).—In the initial stage one or both the
meeanuclei may exhibit pseudopodial processes indicating that
they have some power of amceboid movement. The micro-
nuclei are a little but sometimes very little larger than they
were before conjugation. (In this respect, as in so many
others, there is considerable variability. The reader will
notice that the individuals drawn in fig. 1, in which the
conjugative processes have not yet met the micronuclei, are
actually larger than they are in the individuals drawn in
fig. 2, which are actually conjugating.)

Stage B (fig. 3).—The micronuclei increase considerably
in size during this stage, the chromatin being resolved into
a delicate skein. ‘The meganuclei also increase in size,
become spindle-shaped, and show an arrangement of the
chromatin into roughly parallel lines. From this stage
onwards until Stage K (fig. 13) is reached, the meganuclei
increase in size without showing any material change in
structure ; and as the interest of the phenomena now centres
in the micronuclei, further reference to the behaviour of the
meganuclei will be for the present omitted.

Stage C (figs. 4, 5)—When the micronuclei have reached
their full size the chromatin collects in the form of numerous
minute chromosomes in an equatorial plane. At the same
time extremely delicate, faintly staining threads (linin
threads) roughly parallel with one another, forming a tub-
shaped spindle, make their appearance. Neither in this nor
in any other stage of mitosis is there any sign of the
presence of centrosomes or similar bodies. I have found it

DENDROCOME'TES PARADOXUS. 329

impossible to count the number of chromosomes with any
degree of accuracy in this or in any other stage of division.
The mode of division of the chromosomes is also very
difficult to determine, but I have seen V-shaped appearances,
Pl. 18, fig. 15, very similar to those figured by Prowazek (22)
in Bursaria. I believe with him that these appearances
point to longitudinal division.

The chromosomes separate into two parties, which travel to
the opposite poles of the spindle (fig. 5), where they apparently
fuse to form a solid irregular lump of chromatin. ‘I'he
spindle then elongates enormously, so that the two chromatin
bodies are sometimes separated from each other by a
distance equal to three fourths of the full diameter of the
Dendrocometes. ‘The spindle then becomes detached from
the chromatin and dissolves in the cytoplasm (fig. 6). The
stages in the division of the micronuclei are usually syn-
chronous in the two individuals, but sometimes one set of
figures is a little in advance of the other, as shown in fig. 5.

By this division six micronuclei are formed in_ those
individuals which began the process with three.

Stage D (fig. 7)—One of the five micronuclei of each
individual passes down the conjugative process to a position
very close to the membrane, where it enlarges and again
forms a mitotic figure. The other micronuclei degenerate
and disappear.

The nuclei which are found close to the membrane give
rise by their division to the germ nuclei, to use the term
employed by Wilson (26). ‘The division is always in a plane
parallel with the membrane (fig. 8).

Stage H (fig. 9).—The germ nuclei take up such a position
in contact with the membrane that each germ nucleus of an
individual is exactly opposite one of the other individual.
These nuclei consist of a clear vacuole containing a single
coarse skein of chromatin. The spindle entirely disappears.

Stage F.—The germ nuclei then fuse in a manner shown
in Pl. 17, fig. 10, and Pl. 18, fig. 12, giving rise to the
cleavage nuclei of the two individuals.

330 SYDNEY J. HICKSON.

Attention may here be directed to two important points of
comparison with ciliate Infusoria, The difference between
the migratory or male germ nucleus and the stationary or
female germ nucleus is in Dendrocometes reduced to a
minimum. It is possible that in all cases one germ nucleus
traverses the membrane and the other does not, so that the
distinction remains, but the two nuclei are as nearly neuters
as can be. In the second place, the fusion of the germ
nuclei takes place during a resting and not in a mitotic state.

According to the researches of Maupas, Hertwig, and
others, the germ nuclei of the ciliate Infusoria fuse when in
the form of spindles or mitotic figures.

Stage G (Pl. 17, fig. 11)—One of the cleavage nuclei
passes into each of the conjugating individuals and prepares
to divide again by mitosis. The early stages of this division
probably occur very soon after the fusion of the germ nuclei,
as the figures may be seen sometimes quite close to the
membrane (cf. Pl. 18, fig. 18). This stage may be dis-
tinguished from Stage D, which it somewhat resembles, by
the fact that the axes of the spindles are not parallel.

Stage H (PI. 17, fig. 12).—The cleavage nucleus divides
into two nuclei which take up a position in close proximity to
the meganucleus.

Stage J (Pl. 18, fig. 1) —The nuclei formed by the division
of the cleavage nucleus again divide, and almost immediately
one of the four becomes a little larger than the other
three.

Stage K (Pl. 17, fig. 13)—The largest of the four
nuclei of the last stage becomes the new meganucleus, the
other three the new micronuclei. There is some evidence to
show that occasionally two of the three smaller nuclei again
divide in this stage, giving rise to a condition in which there
are six nuclei in all, as seen in the preparations from which
Pl. 18, fig. 8, was drawn. In some cases, too, it appears
that two nuclei enlarge to give rise to new meganuclear
structures, as seen in Pl. 18, fig. 19. Variations of this
kind at these stages have added very much to the ordinary

DENDROCOMETES PARADOXUS. 331

difficulties of the investigation, but the establishment of the
fact that variations of such an important character do occur
is, In my Opinion, a result of considerable interest.

At some time during the last three stages (H, J, K) the old
meganucleus becomes very large, and is bent on itself in the
form of a loop or horseshoe. One extremity of this figure
passes into the conjugative process, and approaching the
limiting membrane traverses it and fuses with the correspond-
ing extremity of the meganucleus of the other individual.
The exact phase at which this meganuclear conjugation takes
place seems to vary considerably; all that can be said at
present is that, so far as my experience goes, it usually
occurs between Stages J and K. The number of cases of
actual contact of meganuclei observed by Mr. Wadsworth
and me is small, and this may be interpreted to mean either
that the period of meganuclear conjugation is very short or
that it does not always occur. Until some satisfactory
method is invented of watching the nuclear phenomena of the
conjugation of living Dendrocometes, it is impossible to
prove that the meganuclear conjugation never fails. Jam
inclined to believe that it always occurs. Similarly I have no
proof to offer of the length of time occupied by this process ;
but I am inclined to believe, on the circumstantial evidence
at my disposal, that it is very brief.

Soon after the meganuclei have conjugated they separate
and begin to degenerate.

he usual phenomena of nuclear change during conjuga-
tion in Dendrocometes may be represented by the following
diagram, in which the circles above M represent the stages in
the meganucleus, and the black dots above m, m, m, the stages
in the history of the micronuclei. Conjugation of the mega-
nuclei usually occurs in Stage J, as explained above, and in
Stage K these bodies disintegrate.

In the following diagrams I have endeavoured to interpret
certain phenomena which appear to be variations of the more
usual stages. In Diagram B, which starts with Stage G,
the important variation is that two of the micronuclei formed

302 SYDNEY J. HICKSON.

by the second division of the cleavage nucleus divide again
(see p. 330).
In Diagram C we have the same variation as in Diagram B,

Diagram A.

Mc RO eO@ me ao)

Fe Se ch ie a ieee

Stage K

with the additional peculiarity that two of the nuclei give
rise to new meganuclei in Stage K, and I have added in
Stage L the suggestion that these two new meganuclei

DENDROCOMETES PARADOXUS. 335

fuse together to form the single meganucleus of the new
individual.
The phenomenon of conjugation in the Acinetaria has been

Diagram le).

ogee

known for a great many years. It has been observed in
several species of Acineta by Claparéde and Lachmann (4),
Fraipont, D’Udekem, Keppen (15), and others; in

Diagram Ce

©

Metacineta (Claparéde and Lachmann and Lieberkiihn), in
Podophrya, 'Tokophrya, Stylocometes, and in Dendrocometes.

In Dendrocometes it was observed by Wrzesniowski (27), but
more fully described by Plate (21).

354. SYDNEY J. HICKSON.

As regards the external features very little can be added to
Plate’s description. As he observes, conjugation usually occurs
when the gill of the host is unusually crowded with the Dendro-
cometes. ‘The two conjugating individuals are in most cases
similar in all essential respects. Occasionally, however, a dif-
ference may be observed between the two conjugates. Plate
states that sometimes one of the individuals which is clearer
than the other withdraws its tentacles during the process. Sand
(23) says, “ Chez Dendrocometes la conjugaison a lieu souvent
entre un individu amaigre et un animalcule bien nourri.”
Mr. Wadsworth and I have frequently observed differences
between the conjugating individuals, and Mr. Wadsworth
has observed the retraction of the arms of certain individuals
during conjugation.

Plate observed occasionally a conjugation of three indi-
viduals. This I am able to confirm, but the occurrence is so
rare that no series of nuclear changes in them have been
followed.

As Plate denied the existence of micronuclei, he failed
to see any of the stages of their division and conjugation
which are described in this paper. In one of his figures
he shows the points of the meganuclei in the bases of the
conjugative processes, but he did not observe the fusion
of these bodies. He gives a good figure to illustrate the
fragmentation of the meganucleus at the close of conjugation.
In Nchneider’s (24) figures of Dendrocometes in the act of
conjugating the meganuclei are shown to be approaching
much more closely than in Plate’s figure, and in the same
author’s figures of the closely allied genus Stylocometes the
meganuclei are actually shown to be in contact. ‘The state-
ment made by Plate and Schneider that the new meganucleus
is formed by a regeneration of the fragments of the old 1s
not correct. '

The mixing of the cytoplasm in the conjugative process is
affirmed, and I believe correctly, by Plate, Sand, and others.
I have myself observed a flow of protoplasm passing back-
wards and forwards through the membrane with each

DENDROCOMETES PARADOXUS. 339
contraction of the vacuole for several hours in one pair, but
the period during which this occurs is limited, and during
the greater part of the forty-eight or more hours during
which the pair remain in conjugation no interchange of
particles can be seen.

An important feature of the conjugation of this genus
is that the ordinary vital processes of the individuals are
not materially affected during the act. The arms remain
fully extended—they catch and swallow particles of food,
which are digested in the ordinary course. The contractile
vacuoles continue their pulsations for some time after the
process has commenced.

Conjugative Processes.—These processes are formed
specially for conjugative purposes in Dendrocometes. In
development and in structure they differ from ordinary arms,
and I do not consider that they are rightly considered to be
homologous with arms. This view does not agree with that
expressed by Biitschli (2), who, on the strength of the
observations of A. Schneider on Stylocometes, regards them
as rudimentary arms. In Stylocometes, according to that
author, an arm of each of two neighbouring individuals
becomes abnormally thick and elongated to form the conjuga-
tive process, and that it is a true arm is proved by the fact
that it contains a canal. I have had no opportunity of
examining Stylocometes, but I do not feel that Schneider’s
account of it is convincing. Until his account is confirmed,
therefore, I must agree with Plate that the conjugative pro-
cesses in Dendrocometes are not homologous with the arms.

Biitschli (2) states that in Dendrocometes one individual
sends out a process which fuses with the body-wall of
another, but that such an occurrence is rare. I have never
seen such a phenomenon. In nearly all cases the conjugative
processes begin and grow simultaneously, so that they are
throughout approximately equal. In the preparation from
which fig. 1 was drawn, the individual to the left has a
decidedly longer process than the individual on the right, but
such a difference as this is exceptional. It is impossible to

336 SYDNEY J. HICKSON.

state what stimulus there is that causes two individuals
to begin the sexual proceeding. It may be that some
excitement may be caused by the touching or brushing of the
arms, but nothing of the kind has been observed. Judging
from the development of the conjugative processes alone,
which is all that we have to guide us in the initial stages, it
seems probable that the sexual stimulus affects the two
individuals simultaneously, and that there is no differentiation
of sex.

The Micronuclei.

It is an interesting fact that, notwithstanding the careful
and elaborate investigations made by several observers on
the micronuclei of the Ciliata, there is at present no satis-
factory account of these structures in the Acinetaria.

That micronuclei do occur in all Acinetaria is not yet proved,
but nevertheless it is extremely probable that they are as
constant a feature of the anatomy of this group of animals as
they appear to be of the Cilata.

In Biitschli’s great work on the Infusoria (p. 1873) the
following notes will be found on the subject. Biitschli him-
self in 1867 discovered a micronucleus in a species of Sphae-
rophrya. Maupas proved with certainty the existence of a
micronucleus in Tokophrya limbata, Acineta tuberosa,
Podophrya fixa, and Podophryacyclopum. He pro-
bably, but not with certainty, found them also in Acineta
Jolyi, whilst certain bodies which may have been micro-
nuclei were seen in Hphelota gemmipara. In Toko-
phrya limbata Modbius also proved the existence of micro-
nuclei.

In a recent paper on the anatomy of a new species of
Ephelota, Ishikawa writes with some uncertainty about the
existence of micronuclei. he most satisfactory figures of
the micronuclei of Acinetaria are those given by Keppen (18),
but as his monograph is written in the Russian language I
am unable to read it.

The authors of the ‘'T'raité technique de Zoologie’ (5) accept

DENDROCOMETES PARADOXUS. 337

the view that there is a micronucleus in the body of the
Acinetarian.

In the recently published monograph of the Acinetaria,
however, Sand (28) denies the existence of true micronuclei.
He believes, however, in the existence of a body lying close
to the meganucleus, which he calls the ‘* centrosome.”

“Un corps un peu plus colorable que le noyau, creusée
exprés pour le recevoir. . . . Quand ce corps est vu au
centre du noyau, il parait entouré d’une petite zone claire
formée par le cytoplasma qui le sépare du noyau.”. He
further states that no Acinetarian ever contains two of these
bodies, and that it is absolutely homogeneous, both in a state
of rest and of division. He has proved the existence of this
body in sixteen species. I have very little doubt that the
“ centrosome ” of Sand is not a “ pseudo-micronucleus,” but
a true micronucleus. Maupas (20) says that his series of
stages of the micronuclei of Podophrya fixa is incomplete,
and he would probably himself admit that the figures he has
published of them are not satisfactory. ‘T'o account for his
unsatisfactory results in the Acinetaria, Maupas ventures
upon the statement that the micronuclei of these Infusoria
are smaller than they are in the Ciliata, and stain very
lightly. He adds that the study of the nuclear phenomena
of the Acinetaria is beset with so many difficulties that he
considers it to be “une des recherches les plus pénibles
qu’un micrographe puisse entreprendre.”

In Dendrocometes, however, these excuses cannot be put
forward, as the micronuclei are even larger than they are in
many Ciliata, and their affinities for certain stains are ex-
ceedingly powerful.

The presence of micronuclei in this genus has, neverthe-
less, been denied.

Plate (21) described a number of small bodies in the cyto-
plasm which, on account of their affinities for safranin, he
called the ‘“ Tinktinkérnchen,’”’ but he denied the existence of
““ Nebenkerne.”’ Maupas replied to Plate’s paper by stating
emphatically that micronuclei do occur in Dendrocometes

338 SYDNEY J. HICKSON.

and referring to Plate’s own figure, pl. vi, fig. 17, as afford-
ing a convincing proof of it. Neither Butschli (1), Wrzesni-
owski (27), nor Schneider (24) appear to have observed them,
and Sand (23) only mentions one “centrosome” as being
present in this genus.

The number of micronuclei in Dendrocometes varies, but,
as stated above, the usual number is three. In individuals
that are neither conjugating nor preparing for gemmation it
is sometimes difficult to count the micronuclei, as they are
very small, and are difficult to distinguish from other bodies
in the cytoplasm which have affinities for the stains that are
used. _ However, a certain number of individuals with
moderately clear cytoplasm can nearly always be found, and
in these it is not difficult to recognise the micronuclei in sec-
tions stained by the iron-hematoxylin and iron-brazilin
methods (13). There are sometimes two micronuclei, and
sometimes four, and but rarely five. I do not believe that
there are ever more than five or less than two.

In the state of rest the micronucleus consists of a small,
irregular granule of chromatin, enclosed in a clear zone,
which is invariably perfectly spherical in shape. There is no
evidence of a definite membrane surrounding the clear zone.
The zone is from 2—4 u in diameter. It is very difficult to
form an opinion of the nature of the substance composing the
clear zone. When the uucleus is ina state of rest this zone
resists the action of all the stains I have tried, and no lines
nor granules of any kind or description can be seen in it. It
is possible that it may bea mere artefact, due to the shrinkage
of the chromatin during the process of preservation, but the
regularity of its shape and its relation to the chromatin
granule do not support this view. In my opinion it really
represents the “‘achromatin ” elements of the nucleus.

When preparing for division the micronuclei increase con-
siderably in size, the solid granule of chromatin becoming
converted into a ‘coarse skein (PI. 18, fig. 14m). Later the
skein breaks up into a much finer tangle (PI. 17, fig. 3), which
eradually fills up nearly the whole of the clear zone. In

DENDROCOMETES PARADOXUS. 339

the next stage there is a differentiation of the skein into a
number of lines, which stain very faintly in iron-hemato-
xylin, but a little more deeply with iron-brazilin, and a band
of deeply staining rods and granules having the usual
chromatin reactions. At this stage the micronucleus is
frequently 10 « in diameter (PI. 17, figs. 4,11,12). The exact
determination of the nature of the equatorial band of chro-
matin granules is a matter of considerable difficulty. With
the highest powers of the microscope the granules or minute
rods appear to be connected together in the manner shown
in Pl. 18, fig. 15, but still I do not feel so far convinced of
this connection as to deny the proposition that they are
isolated chromosomes. With lower powers of the microscope
they have every appearance of being a band of rod-shaped
chromosomes (Pl. 17, fig. 12). Whatever the future may
reveal regarding these bodies, I think it is clear that the
chromosome elements are numerous,—too numerous, in fact,
to count with any degree of accuracy.

The micronucleus next becomes somewhat oval in shape
(Pl. 18, fig. 16), and the band divides into two bands. The
faintly staining lines, which we may call the linin threads,
are arranged in a roughly parallel manner forming a some-
what tub-shaped spindle. They do not come toa point at
each end of the figure, and there is never any dot or granule
that can suggest the presence of a centrosome. In the next
stage (Pl. 18, fig. 17) the two bands of chromosomes have
passed to the extremities of the figure, and soon become
ageregated together to form a single irregular lump. In
some cases there is a clear vacuole present at one of the
poles (the right pole in PI. 18, fig. 17), but I have seen it so
rarely that I am ata loss to understand its meaning. In
the next stage (PI. 18, fig. 18) the spindle becomes very
much more elongated, so that the whole figure may be
25—30 uw in length; the chromatin is in the form of a single
spherical lump surrounded by a clear zone at the points o
the spindle. The spindle next becomes detached from the
new micronuclei (PI. 17, fig. 6), and gradually dissolves in the

340 SYDNEY J. HICKSON.

cytoplasm. Throughout this mitosis the clear outline
(membrana limitans ?) is never lost.

The mitosis of the micronuclei of Dendrocometes has con-
siderable resemblance to the mitosis of the micronuclei of
Parameecium, as described by Hertwig (8), of Bursaria, as
described by Prowazek (22), and of other Cihiata.

The Meganucleus.

In specimens of Dendrocometes that are neither con-
jugating nor preparing for gemmation, the meganucleus is
usually a spherical body 0°02 mm. in diameter. It is not,
however, constant in shape, many examples being found that
are oval or even spindle-shaped.

It is usually situated in the centre of the animal’s body,
but it is often more or less excentric. On examination in
section with an oil immersion lens, there may be seen a
distinct meshwork of darkly staining lines which appear to
support a series of minute rounded chromatin granules (PI. 18,
fig. 14). In many cases the lines connecting these granules
cannot be easily seen when the appearance is as shown in
Pl, 18, fig. 10. In the meshes of the darkly staining chromatin
there is a homogeneous substance, which in many iron-
hematoxylin preparations is quite colourless, but stains
faintly yellow with brazilin.

Plate states that in the nucleus there are a number of
nucleoli: ‘ Bei einem Thier meiner Priaparate ist der ganze
Kern dicht erfiillt von solchen Binnenkérperchen deren jedes
von einem hellen Hof umgeben ist.” I have frequently seen
in my preparations appearances similar to this described
and figured by Plate, but my interpretation of them is
different. In the first place I must state very emphatically
that in my opinion there is never in the meganucleus of
Dendrocometes any body or bodies which correspond with the
nucleoli of Metazoan cells. The clear space round the “ Binnen-
kGrperchen” of Plate is in my opinion due entirely to the
refraction of the light in passing through the preparation ;
and this opinion has been thoroughly tested by comparison

DENDROCOMETES PARADOXUS. 341

of whole mount preparations such as Plate studied, with the
thinnest sections of the meganucleus. My opinion is that
the substance in the meshes of the chromatin is quite homo-
geneous. The periphery of the meganucleus is not in my
opinion surrounded by a definite membrana limitans, notwith-
standing Plate’s statement that ‘ein Kernmembran ist deutlich
erkennbar.”

When the meganucleus is spherical or oval in shape and
situated at the centre of the body, or in other words, when
it isat rest, no definite membrana limitans can be seen (PI1.-18
fig. 10); but in the elongated and dividing meganuclei the
chromatin network at the periphery is so arranged that a
limiting membrane seems to surround the whole nucleus:
The limiting membrane is not a definite and peculiar struc.
ture of the nucleus, but a temporary arrangement of the
substance of the chromatin-bearing network at the periphery
during the nuclear movements.

During gemmation the meganucleus undergoes a simple con-
striction, and is divided into two parts, one of which is retained
by the parent, and the other by the young bud (PI. 18, fig. 14).
There is no reason to believe that these two parts are exactly
equal in size. In many preparations the part retained by
the bud is apparently smaller than that retained by the
parent, but, as [ have no means of measuring the capacity
of these bodies, I cannot make any positive statement on
the subject. There is, however, no arrangement of the
chromatin in rods or bars during this division which would
suggest equivalent chromatin division. Nor have I been
able to find after very careful search anything of the nature
of anterior fibres, centrosomes, asters, or other charac-
teristic features of karyokinesis. ‘'here can be no doubt
whatever that when the maganucleus divides the process
is purely amitotic.

The Fusion of the Meganuclei during Conjuga-

ion.—Whatever difficulties there may be in finding an
explanation of the fact, there can be no doubt that the
meganuclei do, during .conjugation, meet and become

VOL. 45, PART 3,—NEW SERIES. AA

342 SYDNEY J. HICKSON.

continuous. The statement of this fact was made in my
preliminary communication (Hickson, 1900, 12). Iam not
the first, however, to maintain that the meganuclei of the
Tnfusoria fuse. In 1867 Stein made the following statement
concerning the Acinetaria (Stein, ‘Der Organismus,’ vol. 1,
ps 139), LeG7a

“Die conjugation verliiuft auch bei denjenigen Acineten,
bei welchen sie bisher genauer studirt wurde, im Wesent-
lichen auf dieselbe Weise, wie bei der gleichartigen
Conjugation der Vorticellen; es verschmelzen zuerst die
Korper der beiden Acineten zu einem einzigen, und dann
fliessen auch deren Nuclei in einem gemeinsamen Nucleus
zusammen.”

Bitschli (2) gives a figure (pl. Ixxii, 96) of two attached
Vorticellids in which the meganuclei are in junction, but
considers that this is doubtfully a case of conjugation.
Schneider (24), in Stylocometes, figures the junction of the
meganuclei in two individuals that are conjugating, but
suggests that this also may be a case of fission. Biitschli
may have been right as regards his Vorticellids, but such a
method of fission as Schneider suggests for Stylocometes is
extremely improbable. In a recent paper by Prowazek (22)
a number of new and excellent figures are given of the
nuclear phenomena during the conjugation of Bursaria, and
it seems probable from these that in this ciliate Infusorian
there is a junction of the meganuclei before they disintegrate.
Unfortunately Prowazek’s description is not very clear, and
he does not attach much importance to the phenomenon.

In my preparations of Dendrocometes I have at least three
cases in which the meganuclei actually touch, but a consider-
able number in which they approach one another very closely
in the conjugative processes. That the junction is not
merely casual contact, but actual organic connection, is
proved by the preparation which is represented in PI. 18,
fig. 11. Here there is no sign of any boundary between the
two nuclei, and the chromatin granules are fixed in such a
manner as to suggest very forcibly that during life they were

DENDROCOMETES PARADOXUS. 343

flowing from one side into the other. Apart from this
evidence, however, attention may be called to the fact, which
is evident not only from my own preparations (PI. 18, figs.
6—8), but also from Schneider’s figure of Stylocometes, that
when the points of the meganuclei pass down the conjugative
processes they converge to the same spot on the membrane.
This shows, I think, that there is some force at work which
is bringing them together. Sand refers to this in Stylo-
cometes when he says, ‘‘Les deux noyaux s’approchent et se
placent dans le bras dilaté vis-a-vis l’un de l’autre séparés par
une couche de plasma.”’ But he adds, “ Pourquoi, dira-t-on,
les noyaux vont-ils se placer dans le pont qui réunit les deux
Stylocometes ou les deux Dendrocometes? C’est peut-étre
pour diriger les échanges et les mouvements qui ont lieu dans
ce pont.”

The organic junction of the two meganuclei lasts a very
little while, I believe, and it is probably followed immediately
by their disintegration. Hach meganucleus breaks up into a
number of irregular lumps, in each of which there are at first
several granules of unaltered chromatin. A large piece of
darkly staining substance is frequently present in these
lumps, but in many of them the central parts are simply
vacuolated (PI. 18, figs. 15 and 19).

In the next stage the cytoplasm is filled with numberless
vacuoles, granules, and lumps (PI. 18, fig. 19) of endless forms,
sizes, and colourable property. In such a preparation as that
from which the figure was drawn, it is almost impossible to
distinguish the various remnants of the old meganuclei from
food bodies and micronuclei.

The mode of formation of the new meganucleus at the
close of conjugation is of great importance and interest.

I must confess that in the earlier stages of this investiga-
tion I had some hesitation in believing that the new
meganucleus is formed from a product of the segmentation
nucleus. The descriptions, and more particularly the figures,
of the nuclei of conjugation in Ciliata by Maupas (20),
Hertwig (8), Hoyer (14), and others, although unanimous

344 SYDNEY J. HICKSON.

are not convincing. In none of these papers are the stages
of the enlargement of the micronuclear element to form the
characteristic meganuclear body very complete, and it seemed
to me that there was just a possibility that these authorities
were mistaken, and that the new meganucleus arises indepen-
dently in the cytoplasm, or from one or more of the old
meganuclear fragments. This hesitation in accepting the
orthodox view was due to the fact, that in the earliest stages
Thad then found of the formation of the new meganucleus
there was no chromatin in its centre. It was, moreover, very
much larger than any of the other micronuclei, and the con-
nection between it and a micronucleus could not be traced.

The subsequent discovery of the intermediate stages, how-
ever, removed my doubts, and now I feel that it is quite an
established fact that the new meganucleus is formed from
one of the four nuclei produced by the second division of the
germ nucleus. Plate, Schneider, and Sand, who maintain
that the new meganucleus is formed by the reconstitution of
one or more fragments of the old meganucleus, are in error,
and I believe that the views expressed by Maupas and
Hertwig as to the origin of the new meganucleus in Cihata
are correct.

The principal stages in the formation of the new mega-
nucleus are shown in Pl. 18, figs. 2—5. In the first stage
one of the four micronuclei (figs. 1 and 2) increases in size in
a manner very similar to that in which the micronuclei swell
up just before mitosis in bud formation, or in the earlier
stages of conjugation. In the preparation, the enlarged
micronucleus was 5 x in diameter, and the others 4. This
enlargement is caused by a considerable increase in the clear
substance, and by the resolution of the chromatin into a
coarse skein. In the next stage (fig. 3) the nucleus has still
further increased in size to about 8 mu in diameter; the
chromatin has become more diffused, and does not stain so
deeply. It is probable that the change in staining power
indicates some slight change in constitution, but there is
no evidence as to the nature of this change. The greater

DENDROCOMETES PARADOXUS. 345

part of this modified chromatin is arranged in the form of —
a thick ring at the periphery. There are, however, some
strands stretching across the nucleus, and a considerable
number of rows of granules extending from the edge into the
cytoplasm. I think there can be little doubt that at this stage
either the whole or the greater part of the chromatin, in its
modified form, passes into the surrounding cytoplasm, leaving
the new meganucleus perfectly clear and homogeneous.

The elimination of chromatin from nuclei is a phenomenon
of rare occurrence in animal and vegetable cells. In the
maturation of the ovum of many animals a considerable
amount of chromatin is ejected into the cytoplasm. Wilson
(26) says, ‘In these cases (Asterias, Polycherus, Thalassema,
Nereis) only a small fraction of the chromatin substance is
preserved to form the chromosomes, the remainder degene-
rating in the cytoplasm. Some years ago I described the
fragmentation of the germinal vesicle of the Stylasterid
Allopora (this Journal, vol. xxix) and the distribution of
its chromatin in the cytoplasm. A similar phenomenon
occurs in Distichopora (10). I have recently devoted a
considerable amount of attention to the ovum of Alcyonium,
and in this case, too, the whole of the chromatin appears to
be ejected into the cytoplasm before fertilisation takes place.
In certain insects, judging from the figures given by
Henking (9) and others, the amount of chromatin that
takes part in the formation of the first polar figure is a
very small fraction of the chromatin originally present in
the germinal vesicle (cf. Cuenot, 3a).

That the elimination of chromatin is not confined to the
nuclei of egg cells is clear from the discovery of Boveri’s,
that in those blastomeres in the early stages of development
of Ascaris which are destined to produce somatic cells, “a
portion of the chromatin is cast out into the cytoplasm,
where it degenerates, and only in the germ cells is the
sum ‘total of the chromatin retained” (quoted from
Wilson, 26). In all these cases of the elimination of
chromatin from the nuclei of ova and blastomeres there

346 SYDNEY J. HICKSON.

appears to be no recovery in the amount of chromatin before
the next division occurs. In the history of the formation of
the new meganucleus of Dendrocometes, however numerous,
granules of chromatin subsequently appear at the periphery
(Pl. 18, figs. 4, 7, 8), and later they invade the clearer
central parts (Pl. 18, figs. 5, 15), to build up the
characteristic chromatin network of the functional mega-
nucleus. The exact meaning of this elimination and recovery
of chromatin at this stage is a mystery, but taken in con-
junction with the other phenomena of conjugation, 1b may
be regarded as a part of the general process of protoplasmic
reconstitution of the organism, which is the essential feature
of the sexual act.

GENERAL CONSIDERATIONS.

Maupas, in his famous work on the conjugation of the
Infusoria, expressed the opinion that conjugation is essentially
an affair of the micronuclei; and I think that the prevailing
opinion held by zoologists who have taken a special interest
in this matter, is in general agreement with this view.
Biitschli’s opinion, as expressed in ‘Infusoria’ (p. 1648), is
that the meganucleus is of the nature of a somatic nucleus
(Gewebekerne), which becomes gradually exhausted (allmah-
lich abgenutzt wird) during somatic life, whilst the micro-
nucleus is of the nature of the sexual nuclei of Metazoa, and
does not become exhausted by the vital processes (keine
solche Abnutzung erfahrt). Wilson (26) expresses very
fairly the prevalent view in this sentence: “ During conjuga-
tion the macronucleus degenerates and disappears, and the
micronucleus alone is concerned in the essential part of the
process.”

With the general proposition that the meganucleus is of
the same essential nature as the nucleus of the somatic cells
of the Metazoa, and that the micronucleus is essentially a
sexual nucleus, | am in agreement; but there are serious
objections to be raised to the further proposition, that the only

DENDROCOMETES PARADOXUS. 347

essential process in connection with conjugation is that in
which the micronuclei are concerned.

Plate expressed a view that during conjugation there is a
recovery of some essential substance of the nucleus from the
cytoplasm (see Biitschli, 2) ; but, as Biitschli rightly points
out, it is difficult to understand upon what grounds Plate’s
view is based.

Without going further into a review of the opinions of
various writers on conjugation, it may be sufficient to state
here the problem which is still in need of solution. Is the
interchange of molecules of the cytoplasm of the two con-
jugates during conjugation an essential part of the process ?

This question cannot be answered by direct evidence at
present. Whatever interchange of molecules of the cytoplasm
there may be during conjugation, no method of observation
has yet been discovered by which the course of the molecules
of one individual can be traced into the body of the other.
It is otherwise with the micronuclear nucleoplasm, the peculiar
structure and staining properties of which enable us to trace
with certainty the course of one micronucleus (or germ nucleus)
into the body of the other. The direct evidence which we
have in the case of the micronuclear fusion is absent in the
case of the cytoplasm. It is not reasonable to conclude from
this alone that the cytoplasm plays no part in the process of
conjugation, nor that conjugation is simply “ une affaire de
micronucleus.” It appears to me that there is some indirect
evidence, however, on this point which is worthy of consider-
able attention. If the micronuclei alone were concerned in
the process, the act of conjugation need be of very short
duration. In fact,if the germ nuclei were prepared for their
transposition a mere momentary contact would be sufficient.
It might also be conceived that such a momentary conjugation
would be of advantage to the species in lessening the dis-
advantageous conditions of the conjugating phase, particularly
in the free-swimming Ciliata. In all cases, however, the
conjugation is a lengthy process, lasting from twelve to forty-
eight hours or more. It is inconceivable that the state of

348 SYDNEY J. HICKSON.

insensate, helpless, defenceless syzygy would remain so long
if there were nothing else of essential importance done but
the interchange of germ nuclei. ‘The fact suggests that
there is during the process some interchange of the molecules
of the cytoplasm, and, indeed, that the interchange or mixing
of the molecules is thorough, and not partial or local in
character. If there is during conjugation an interchange of
the molecules of the cytoplasm such as has been suggested,
it is probable that some protoplasmic streaming movement
would be noticed between the two individuals. The observa-
tions on the changes or movements of the cytoplasm during
the process are, however, very limited. Maupas observed that
numerous granules (zooamylum) appear in the cytoplasm
during the conjugation of certain Ciliata, which he supposed
to be connected in some way with the active metabolism that
is going on; but I cannot find in his writings any reference
to a streaming movement taking place between the two
individuals. But Maupas, hke Butschli and many others,! it
must be remembered, regarded the micronuclear phenomena
as the only essential phenomena of the process, and did not
expect to find any such flow of cytoplasm.

In Dendrocometes a flow of cytoplasm between the two
conjugates does certainly take place. ‘l'his was observed by
Plate and is confirmed by my own observations. Sand (28,
p. LOO) goes so faras to say that conjugation is essentially a
process of plastogamy, and that there is not the least mixing
of the nucleoplasm of the two individuals. But Sand’s view
is, | believe, as far wrong in the one extreme as the older view
is in the other.

Whether a similar streaming movement of the cytoplasm
between the conjugates can actually be observed in the
group of the Ciliata or not, is a question upon which I have
no evidence to offer. But whether it can or cannot be
observed under the microscope, the intimate contact of the
two cytoplasms renders an invisible interchange of molecules

Delage and Hérouard say, ‘ Les phénomenes intérieurs de la conjugaison
sont SURTOUT nucléaires.”’

DENDROCOMETES PARADOXUS. 349

possible, and the “onus probandi” really rests upon those
who maintain that two globules of protoplasm, such as these,
can remain in junction for twenty-four hours without becom-
ing intimately mixed.

A third point of indirect evidence bearing upon this
question is afforded by the behaviour of the meganucleus of
Dendrocometes during conjugation. If we regard the
meganucleus asa somatic nucleus—that is to say, as a nucleus
which is functionally connected with all the vital functions
except the sexual functions of the body, and as a nucleus
therefore which controls or is controlled by the greater part
of the cytoplasm of an animal cell such as Dendrocometes is,—
then the presence of the meganuclei in the conjugative
processes during the interchange of the molecules of the
cytoplasm is not a matter for surprise. ‘Pourquoi, dira-t-on,”
says Sand, “les noyaux vont-ils se placer dans le pont qui
réunit les deux Stylocometes et les deux Dendrocometes ?
C’est peut-étre pour diriger les échanges et les mouvements
qui ont lieu dans ce pont.” Iam prepared, however, to go
further than Sand, and regard the presence of the meganuclei
in the conjugative processes (le pont) not only as evidence
of their relation to the interchanges taking place in the
cytoplasm, but as evidence of the necessity of the inter-
change of molecules of the substance of the meganucleus
itself. During conjugation there is, in my opinion, a mixing
or a shuffling of the molecules of all the essential plasms of
the body, namely, of the micro-nucleoplasm, of the mega-
nucleoplasm, and of the cytoplasm.

Concerning the conjugation of the meganuclear elements
two or three obvious objections appear. It might be urged
that the rarity of recorded observations of the fusion of the
meganuclei in the Heterokaryota, the disintegration of the
meganucleus during conjugation, and the origin of the new
meganucleus from the micronuclei, are facts which prove that
the junction of the meganuclei during conjugation in Dendro-
cometes is a matter of no essential importance.

It may be pointed out that in the majority of the Cilata

350 SYDNEY J. HICKSON.

the meganucleus undergoes fragmentation at an earlier stage
than it does in Dendrocometes, and consequently any con-
jugation that takes place between meganuclear fragments
might be very easily overlooked. ‘The fact that in his most
recent publication Prowazek figures (figs. 27 and 30) the
extension of a fragment of the meganucleus of one con-
jugating Stylonychia into the body of the other, supports the
suggestion that it may occur elsewhere. The second objection
is fatal to the view I am putting forward, if it is true, that the
meganucleus dies when it fragments. It is, however, really
of the nature of an assumption to say that the meganucleus
dies at the close of conjugation,

Entz, Balbiani, Gruber, Maupas, Hoyer, and Prowazek are
agreed in the statement that the fragments of the meganucleus
are absorbed by the cytoplasm. In some species (Chilodon
cucullulus, Colpidium colpoda, etc.) the meganucleus
does not even fragment, it simply gradually diminishes in
volume and disappears. On the other hand, Biitschh (2,
p. 1617) is of opinion that in Colpidium and Stylonychia the
fragments of the meganucleus areejected by the anus after con-
jugation. Having very carefully examined the process in Den-
drocometes, and found no evidence of the rejection of any part
of the meganucleus during or after conjugation, I am disposed
to agree with those who believe that the old meganucleus is,
as a rule, absorbed by the protoplasm. It is quite possible,
however, that with the absorption of the greater part of the
meganucleus there may bea rejection, in some species, of the
remainder.

‘'he expression “absorption ” or “solution,” as applied
to the meganucleus at this stage, is very lable to mislead.
We may hold the view that the meganucleoplasm is killed,
converted into some proteid food substance, and then assimi-
lated by the surrounding cytoplasm, and we may use the
word ‘absorption ” to express this meaning. Or we may
hold the view that the meganucleoplasm becomes more fluid
in consistency, and is diffused in a chemically unaltered, or
very slightly altered, condition through the cytoplasm, and

DENDROCOMETES PARADOXUS. 351

we may use the word “ solution” to express this meaning ;
but we have no evidence that it is either of these processes
that actually takes place. All the information we have is
that at a certain stage in conjugation certain structures, which
by their form and reactions to certain stains we recognise to
be meganucleoplasm, become indistinguishable from ordinary
living cytoplasm. ‘There is evidence of a certain change in
chemical constitution, and perhaps this is only a very slight
change, and there is evidence of a certain change in con-
sistency. There is really no evidencé that any substance
actually dies. Theoretically, there is no inconsistency in the
view that after the disappearance of the old meganucleus, its
nucleoplasm is still living in a modified form diffused through
the cytoplasm. The new meganucleus of the Dendrocometes
individual is an enlarged and modified nucleus derived from
one of the four micronuclei which are produced by the second
division of the segmentation nucleus, as described above, or,
to put the matter in few words, the meganucleus is derived
from a micronucleus. ‘he important changes which occur
during the transition from a structure we call a micronucleus
to a structure we call a meganucleus are these :—I1st. A con-
siderable increase in size (from 4 4 in diameter to over 12 wu in
diameter in Dendrocometes). 2nd. A considerable increase
in the amount of chromatin. From whence is this increase
in substance derived? It must come either directly as
formed nucleoplasm, or indirectly as food material from
which nucleoplasm can be formed, from the surrounding
cytoplasm. ‘The evidence as to which of these two alterna-
tives is correct is not conclusive, but there is no sign of such
metabolic activity as might be expected if the material brought
to the new meganucleus is unformed food material, and con-
sequently it is very probable that the increase in size is due
to formed nucleoplasm transfused from the cytoplasm to the
new meganucleus. If this is the case, then the phenomenon
of the conjugation of the meganuclei receives an explanation.

This view appears to me to receive considerable support
from the observation made by Biitschli that the posterior

352 SYDNKY J. HICKSON.

fragment of the meganucleus of Huplotes charon does not
die, but fuses with the new meganucleus. A similar observa-
tion was made by Maupas on Euplotes patella.

The investigation of the conjugation of Dendrocometes
described in this paper throws no new light on the important
question of the initial stimulus to syzygy. It is well known
that Maupas was able to induce conjugation in several
species of Ciliata by a judicious withdrawal of food material
after a certain number of binary fissions; that he was of
opinion that in natural conditions it is the exhaustion of the
food supply which affords the main stimulus to the epi-
demics of conjugation. ‘The views of Maupas have recently
received some support from the experiments of Prowazek
(22), who was able to induce conjugation by hunger in
Stylonychia pustulata. On the other hand, Joukowsky
(17) failed to induce conjugation by hunger in Pleurotricha
after experimenting for eight months and reaching the four-
hundred-and-fifty-eighth generation.

I tried the experiment several times of isolating a number
of Gammarus bearing the Dendrocometes in filtered water
for six days or a week, and obtained in some cases sufficient
evidence that the Acinetarians were affected by hunger; but
there were on an average neither more nor less pairs in
conjugation than in the Dendrocometes of the control
experiment. Starvation cannot be extended for more than a
week in this case, as the hosts soon die in the filtered water,
and their macerating bodies afford ample food again for the
epizoites.

Dendrocometes itself is peculiar among Infusoria in that
it appears to be capable of feeding all through the process
of conjugation. Mr. Wadsworth and I have observed the
arms of conjugates catch food and pass it down into the body
protoplasm. Judging from the food granules as seen in
sections, the onset and progress of conjugation appear quite
indifferent to the condition of hunger or satiety.

The following notes will illustrate this point :

(The letter F in the third and fourth columns signifies that

DENDROCOMETES PARADOXUS. 353

the conjugate contains food vacuoles, and the letter S that its
cytoplasm is clear or moderately clear.)

Number of Stage of Conjugate Conjugate
Slide. Conjugation. iM B.
146 B dike Re

83, 84 B F. S.
10 C E. S.
129 C 8. 8.
127 C S. 8.
126 iH 8. S.
122 1D) Bt S.
121 EK lth iY:
25 J S. S.
P| J S. S.
120 J i: S.
dh J F. F.

The General Morphology of the Heterokaryote
Body.—The investigations of Maupas (20), Biitschli (2)
and Keppen (15), notwithstanding the writings of Plate
(21), and more recently of Sand (23), have placed beyond
all reasonable doubt the zoological affinity of the classes
Acinetaria and Ciliata. In these two classes alone there are
two kinds of nuclei in each independent organism. In all
other Protozoa, with the exception perhaps of a few forms
like Pelomyxa, in which there are only scattered granules
of chromatin, there is only one kind of nucleus. This
fundamental distinction of the Cillata and Acinetaria justifies
us in placing them together in a subdivision of the Proto-
zoa, which may be called the Heterokaryota (Hickson, 11).

There may be some difficulty in giving an absolute defini-
tion of what is a nucleus. It will be agreed, however, that
every structure in a protoplasmic mass that contains

1 Cuenot (8) has recently discovered that in a Gregarine belonging to the
genus Diplocystis, which is parasitic in the common cricket, the two forms of
nuclei occur. ‘lhe micronucleus, however, does not become visible until the
onset of sporulation, but it then divides by a mitotic process to give rise to
the nuclei of the spores, while the meganucleus disappears.

354 SYDNEY J. HICKSON.

chromatin and that divides by mitosis is a nucleus. It is
frequently very difficult, however, to distinguish true nuclear
chromatin from substances in the cytoplasm that are not
chromatin, and there are many examples of nuclei known to
science that do not divide by mitosis. It may be taken,
however, as a further axiom of histology that every structure
originating as a daughter nucleus by mitosis of a pre-existing
nucleus is itself a nucleus.

Both the meganucleus and the micronucleus of the Hetero-
karyote body, therefore, are true nuclei; the former on the
ground that it originates from the nucleus formed by the
mitotic division of a micronucleus, notwithstanding the fact
that it always divides amitotically,! and the latter on the
ground that it divides by mitosis. These two nuclei, how-
ever, differ from each other in several important particulars.
The meganucleus is very much larger in bulk during the
somatic life of the individual than the micronucleus. In
fission or gemmation it divides amitotically. It does not
divide during conjugation, but during or at the close of this
process it ceases to exist as a definite entity.

The micronucleus, on the other hand, is very much smaller
than the meganucleus during somatic life. In fission and
gemmation it divides by mitosis. It does divide, again by
mitosis, during conjugation, and one of the products of its
division gives rise to the germ-nuclei. It is not necessary to
discuss further in this place the relations of these two nuclei.
The reasons set forth by Biitschl with masterly ability in his
great work on the Infusoria, for considering the meganucleus
to be the “somatic” nucleus, and the micronucleus as the
“ sexual” nucleus, are sufficient for my purpose. If, how-
ever, we accept the view that in the body of the Heterokaryote
there is one (or occasionally more than one) somatic nucleus
and one or more than one sexual nuclei, we are led to the
further inquiry whether there is also a distinction between
the somatic cytoplasm and the sexual cytoplasm.

1 Apparent exceptions to this rule are afforded by the meganuclei of
Opalina and Kentrochona,

DENDROCOMETES PARADOXUS, 355

There is no evidence of a positive character to show that
this is the case, but the absence of any visible boundary line
between the sexual cytoplasm and the surrounding somatic
cytoplasm is not a definite proof that the distinction does not
occur. Many instances could be quoted, both from animal
and vegetable tissues, in which each nucleus of a plasmodium
has its own sphere of influence in the surrounding protoplasm,
even when no cell boundaries can be distinguished. It is,
indeed, contrary to our general knowledge and usual concep-
tions of cell structures that any nucleus should be entirely
independent of the cytoplasm that immediately surrounds it,
just as it is that any nucleus should exist entirely free from
any cell protoplasm.

There is one feature of the sexual cells of the Metazoa
which at this point in the argument I should like to call
attention to. When ova and spermatozoa are ripe, that is to
say, when they are ready to perform the only function they
possess, they are entirely free from surrounding cell structures.
There is no reason to believe that in any case I can call to
mind the individual ovum or spermatozoon is in protoplasmic
continuity or even contact with other cells. There are no
other cells of the animal body, except the white blood-
corpuscles, of which the same statement can be made, and it
is a feature of some interest and importance that in the
Metazoa these cells are in their mature condition independent
entities. Now in the Heterokaryota the sexual cytoplasm
must be in contact with, and in all probability is in continuity
with, the somatic cytoplasin at the time of maturity, and
even after the fertilisation has been effected. In this respect
then, there is an essential difference between the Metazoa and
the Heterokaryota. In the Metazoa a conjugation of the
somatic cells and of the somatic nuclei could have no possible
effect upon the sexual cells, either before or after fertilisation.
In the Heterokaryota, on the other hand, whatever effect the
conjugation of the meganuclei and the somatic cytoplasm
may have, it must be felt by the sexual nuclei and the sexual
cytoplasm with which they are in contact. This consideration

356 SYDNEY J. HICKSON.

throws some light on the phenomenon of the conjugation of
the meganuclei in the Infusoria, a phenomenon which has no
parallel in the Metazoa.

In the recent discovery of the phenomenon called “ Xenia”
by the botanists in plants, we find a parallel although not
strictly homologous case. The ripe ovum of the angiosperm
is not an isolated cell. Its germinal cytoplasm is continuous
with the general cytoplasm of the embryo sac, in just the
same way as, according to my views, the germinal cytoplasm
of a Dendrocometes is in continuity with the somatic cyto-
plasm. It is quite possible, therefore, that anything which
influences the polar nuclei or the general cytoplasm of the
embryo sac would influence also the ovum (oosphere) before
or after fertilisation is effected. Nawaschin and Guignard
(7) have shown that in Lilium and some other Angiosperms
the second nucleus of the pollen grain does pass down the
tube, and conjugates with one of the polar nuclei to form the
mother nucleus of the endosperm nuclei. The second nucleus
of the pollen grain and the polar nuclei of the embryo may
be compared with the meganuclei of the Heterokaryote body.
Like these nuclei they conjugate at the time of the true sexual
conjugation of the germinal nuclei, and, moreover, they do
not by subsequent division give rise to the nuclei of the new
individual. It is true that there are important differences
between the two cases. In the plant the conjugation of these
nuclei is not temporary as it is in Dendrocometes, but
permanent, and the product of the conjugation gives rise toa
considerable progeny of well-defined nuclei in the endosperm
before their history is closed. But such differences as these
are not surprising in organisms so widely separated as the
Infusorian and the Angiosperm plant. Detailed comparison
of the two phenomena would probably not be profitable, and
might, indeed, be misleading. All that the comparison can
do for us at present is to confirm the impression that the
temporary fusion of the meganuclei of Dendrocometes that
has just been described is an important and essential part of
the process of conjugation, and not an exceptional or accidental

DENDROCOMETES PARADOXUS. Shiv

juxtaposition of the nuclei in the individual cases examined.
It may also lead to the discovery of other cases of the
conjugation of meganuclei in the Acinetaria and in the
Ciliata.

As a general result of these considerations, it seems to me
that we must either abandon the use of the expression
“unicellular organisms” in our definition of the Protozoa,
or else very largely extend the meaning of the term “ cell.”
In the recent text-books published by Sedgwick and by
Shipley and Macbride, the former course is adopted; but
Lang, in his ‘ Lehrbuch der vergleichenden Anatomie,’ 2nd
edition, 1901, says, “Die einfachsten Organismen, die
einfachsten Thiere (Protozoa), und die einfachsten Pflanzen
(Protophyta) sind weiter nichts als selbstandig und unabhangig
lebende Zellen.”

The body of a Paramcecium or of a Dendrocometes is no
more a single independent cell than is the embryo sac of
an Angiosperm plant.

If we are prepared to extend the use of the term cell
so as to include all structures that are bounded by an
undivided cell wall or cell boundary, then the expression
“unicellular”? may still be applied to the Protozoa; but, in
my opinion, the inconvenience of such a course would far
exceed the advantages it might present.

List oF THE PRINCIPAL PAPERS REFERRED TO IN THIS
Memorr.

1. O.Bitscutt.— Ueber den Dendrocometes paradoxus,” ete., ‘Zeits. f. w.
Zool.,’ vol. xxvili., 1877, p. 49.

2. O. Bitscuii.—* Infusoria” in Bronn’s ‘ Thierreich,’ 1889.
3. L. Cuznor.—‘ Evolution des Grégarines célomique du Grillon domes-
tique,” “C. R.,’ exxv, p. 52.
3a. L. Cuenor.—‘ L’Epuration nucléaire au début de l’Ontogenése,’ t. c.,
p. 190.

4. BE. Cuaparipe et J. LacumMany.—‘ Etudes sur les Infusoires,’ Geneva,
1857—1860.

VOL. 45, PART 3.—NEW SERIES. BB

358 SYDNEY J. HICKSON.

5.

fo)

@

10.

i:

12.

13.

14.

15.

16.

17.

18.

19.

20.

2.

22.

Yves Detace et E. Hirovarp.—‘ Traité de Zoologie concrete,’ tome
i, Paris, 1896.
Y. Detace.—“ Embryons sans noyau maternel,” ‘C. R.,’ exxvii, p. 528.

. L. Guienarp.—* Sur les Anthérozoides et la double Copulation sexuelle

chez les Végétaux angiospermes,” ‘C. R.,’ exxviii, p. 864.

R. Hertwie.—‘‘ Ueber die Konjugation der Infusorien,’ ‘ Abh. der
bayr. Akad. der Wiss.,’ Cl. 1], Bd. xvii.

O. Henxkine.—“ Untersuchungen tuber die ersten Hntwickelungs-
vorginge in der Hiern der Insekten,” ‘ Zeits. f. w. Zool.,’ liv, 1.

S. J. Hicxson.—* The Early Stages in the Development of Distichopora,
with a short essay on the Fragmentation of the Nucleus,” ‘Q. J.
Micr. Sci.,’ 1893, vol. xxxv, 1.

S. J. Hicxson.—“ The Reproduction and Life History of the Protozoa,”
‘Trans. Manch. Microscop. Soc.,’ 1900.

S. J. Hickson.—“‘ The Nuclei of Dendrocometes,” ‘Reports of the
British Assoc.,’ 1900.

S. J. Hickson.—“ Staining with Brazilin,”’ ‘Q. J. Micr. Sci.,’ 1901,
vol. xliv, 3.

H. Hovrr.—“ Ueber des Verhalten der Kerne bei der Conjugation des
Infusors Colpidium Colpoda,” ‘ Arch. mikr. Anat.,’ 54.

— Kerppren.—“ Remarks on the Infusoria Tentaculifera,’ ‘Mem. Soc.
Nat.,’ Odessa, 1888, xiii, 2 (in the Russian language).

C. Isuikawa.—‘ Ueber eine in Misaki vorkommende Art von Ephelota
und uber ihre Sporenbildung,” ‘Japan Coll. Sci. Imp. Univ.,’ vol. x,
p. 119 (1896-8).

D. Joukowsky.—‘“ Beitrage zur Frage nach den Bedeutungen der
Vermelrung und des Eintritts des Conjugation bei den Ciliaten,”
‘Verh. naturh. med. Ver. Heidelberg,’ n. F., vi.

HE. Maupas.—“ Sur la Podophrya fixa,” ‘Arch. Zool. expér.,’ vol. v.
1876, p. 401.

E. Mauras.—“ Contribution a l'étude des Acinétiens,’ ‘Arch. Zool.
expér.,’ vol. ix, 1881, p. 299.

EK. Mavras.—* La Rajeunissement karyogamique chez les Ciliés,”
‘Arch. Zool. exper.,’ II série, vol. vii, 1889.

L. Prate.— Untersuchungen einiger an den Kiemenblattern des Gam-
marus pulex, lebenden Ektoparasiten,” ‘ Zeits. f. w. Zool.,’ vol. xliii,
p. 175, 1886.

S. ProwazeK.— Protozoenstudien,” ‘Arb. aus Zool. Instit. Wien,’
tom. x1, 8, 1899.

DENDROCOMETES PARADOXUS. 359

23. R. Sayv.—‘ Etude monographique sur le groupe des Infusoires tenta-
culiféres,’ Brussels, 1901.

24. A. ScHNEmDER.—‘ Fragments sur les Infusoires. Tablettes zoologiques,’
vol. i.

25. H. J. Wepper.—“ Xenia,” ‘ U.S. Dept.of Agriculture,’ Bulletin 22.

26. I. B. Witson.— The Cell in Development and Inheritance,’ 2nd edition,
1900.

27. A. WxztsniowskI.—‘‘ Beitrage zur Naturgeschichte der Infusorien,”
Z. f. w. Z., vol. xxix, 1877, p. 255.

EXPLANATION OF PLATES 17 & 18,

Illustrating Mr. Sydney J. Hickson’s paper on ‘ Dendro-
cometes paradoxus.”

PLATE 17.

The figures in this plate, with the exception of 1, 2, and 5, are constructed
from a series of drawings of the actual sections of the Dendrocometes. The
micronuclei do not all occur in the same plane as represented, and it is very
rarely that the whole of the meganucleus can be seen in one section. ‘The
structures have been represented as nearly as possible in their true relative
positions. Throughout, M. refers to the meganucleus ; m., the micronuclei; A.,
the arms; P., the conjugating process; L. M., limiting membrane. The number
at the end of the description of each figure refers to the permanent preparation
from which the figure was drawn, ‘lhese preparations are preserved in the
Zoological Laboratory at the Owens College, and may be inspected by qualified
zoologists.

Fic. 1.—Two individuals of Dendrocometes about to conjugate. Hach one
is protruding a conjugating process (P. P.) and these ultimately meet. In these
two individuals the arms are of approximately the same size and degree of
branching. Whole mount, No. 14.

Fic. 2.—Two individuals which have just joined together in conjugation
(Stage A). One of them (to the right) has one short arm and two very
rudimentary arms (A. 2, A.3). The other has one large branched arm and two
shorter simple arms (see p. 328). The micronuclei are very small, and the
meganuclei have undergone very little change (see p. 331). Whole mount,
No. 29.

360 SYDNEY J. HICKSON.

Fic. 8. Stage B.—The meganuclei have become spindle-shaped and are
slightly enlarged. The micronuclei (three in each individual) are considerably
enlarged, the chromatin forming a loose meshwork. Section No. 139.

In this and the following figures details of the arms are omitted.

Fie. 4. Stage B (later).—The micronuclei now show chromosomes arranged
in equatorial planes and linin fibrils running through them to the poles. In
this preparation only two micronuclei can be seen in each individual.
Sections No. 129.

Fie. 5. Stage B (close)—The micronuclei are in later stages of their
mitosis. In the individual on the right the chromosomes of the micronuclei
lave separated into two parties travelling toward the poles. In the individuals
on the left the chromosomes have reached the poles and fused into a compact
mass. Whole mount, No. 14.

Fic. 6. Stage C._—There are now six micronuclei in each individual, and the
undissolved remnants of someof thespindles (Sp.) may be seen in the cytoplasm.
The meganuclei have been omitted from the drawings to render the positions
of the nuclei clear. Section No. 139.

Fic. 7. Stage D.—Five of the micronuclei in each individual are now under-
going degeneration, but one in each (m.c.) travels down the conjugative
process and approaches the membrane of separation. The meganucleus is
omitted in the figure from the individual on the right. Sections No. 126.

Fic. 8. Stage D.—The conjugative micronuclei (me., me.) are now dividing
by mitosis in the conjugative process of each individual. Sections No. 134.

Fie. 9. Stage E.—The conjugative micronuclei have now divided into two
separate nucleii—the ‘‘ germ nuclei.” Sections No. 138.

Fic. 10. Stage F.—The germ nuclei of the two individuals have now fused
orare fusing (upper one). Sections No. 140.

Fig. 11. Stage G.—The cleavage nuclei (S.m.) formed by the fusion of the
halves of the conjugative nucleus in the last stage now travel towards the
centre of each individual and again show mitosis. During the preceding
stages the meganuclei have been gradually enlarging and have now reached a
considerable size. Degenerate remnants of the other micronuclei may still be
seen in the cytoplasm. Sections No. 141.

Fic. 12. Stage H—The segmentation nuclei having divided once more
show mitosis. Sections No. 107.

For illustrations of Stage J see Plate 17, figs. 1, 7, 8.

Fic. 13. Stage K.—The segmentation nuclei have now divided into four
nuclei in each individual, three of which become reduced in size and the
chromatin concentrated into a single granule, and one ,in each becomes
enlarged to form the new meganucleus (n. M.). The old meganucleus is
beginning to disintegrate. Sections No. 25.

DENDROCOMETES PARADOXUS. 361

PLATE 18.

1. Section through one of a pair of conjugates after the second division
of the segmentation nuclei (Stage J), showing one nucleus (n. M.) slightly
larger (5 w) than the other three (4 ~). This larger nucleus becomes the new
meganucleus. No. 122.

2. The new meganucleus from the last figure enlarged to show that the
chromatin is at this stage in the form of a coarse skein lying in the centre
of aclear space.

3. The new meganucleus ata later stage. It is now about $ pm in diameter.
A considerable quantity of the chromatin has now collected at the periphery,
and some of it appears to be escaping into the cytoplasm. No. 145.

4, New meganucleus at a still later stage, 10 pw in diameter. Darkly
staining granules of chromatin are now seen at the periphery, one or two
within the periphery, but the central parts stain very faintly indeed. Series
131.

5. New meganucleus at a still later stage, 12 p in diameter, containing
numerous evenly distributed granules of chromatin. Slide 131.

6. Section through a pair of conjugates showing the approach of the old
meganuclei (M.) to each other at the bar of junction. No. 52. (The micro-
nuclei were not clearly stained in this preparation, and are consequently
entirely omitted from the drawing.)

7. Section through a pair of conjugates (Stage J), showing one meganucleus
at the limiting membrane, the other pointing towards it but not reaching it.
There is one new meganucleus and three micronuclei represented in each. (In
the preparation, owing to an unfortunate tear, only one micronucleus can be
actually seen in the lower conjugate.) No. 141.

8. Section through a pair of conjugates (Stage J), showing the approach of
the old meganuclei to each other at the limiting membrane. In each of
these there is one new meganucleus and five micronuclei (an exceptional
condition). In the cytoplasm of the lower conjugate there may be seen three
granules of chromatin (?). These may be the remnants of the polar nuclei.

Slide 131.

9. A small portion of the old meganucleus of one of the conjugates of
the last preparation more highly magnified, showing the chromatin arranged
in irregular parallel lines with thickened nodes and lumps.

10. Section through a resting meganucleus, stained by iron-hematoxylin.
No membrana limitans can be seen. No. 255.

11. Section through the conjugative processes of a pair of Dendrocometes,
showing the organic fusion of the two meganuclei in Stage J. No. 98S.

12, Section through the conjugative processes of a pair of Dendrocometes

362 SIDNEY J. HICKSON.

(Stage F), showing the fusion of the germ nuclei (G.m.). On the left the two
nuclei have not completely joined; the chromatin is in the form of a coarse
skein with thickened nodes. On the right the pair have fused, and the
chromatin has assumed an irregular asterid form. No. 140.

13. Section through the conjugative processes after the fusion of the germ
nuclei to form the segmentation nuclei (S.m.). The segmentation nuclei
become mitotic soon after their formation, but the axes of the figures are not
parallel with the membrane nor with one another. In this case the upper figure
is seen in longitudinal section and the lower in transverse section. No. 117.

14, Section through a Dendrocometes at an early stage in the forma-
tion of a gemmula to show the normal mode of division of the meganucleus.
B. B. Bands of concentric modified cytoplasm which form the peritrichous
bands of cilia of the gemmula. No. 34 s.

15. Section through one of a pair of. conjugates in Stage K, showing the
fragments of the old meganucleus and the new meganucleus. m., One of the
micronuclei? No. 92.

16—18. Three stages in the division of the micronuclei. 16. Imme-
diately after the division of the chromosomes. 17. The chromosomes separated
to the poles of the figure. 18. The chromosomes collected into a granule of
chromatin at each of the poles, and the achromatin in the form of an elongated
spindle. 16 and 17, No. 129. 18, composition drawing from several prepara-
tions.

19. Section through a Dendrocometes at the close of conjugation. It
shows the rare condition of two new meganuclei. The old meganucleus has
almost completely disintegrated. No. 74 E.

Note.—In Stage K, when the old meganucleus has fragmented, it is
extremely difficult to distinguish the micronuclei from fragments of the old
meganucleus. I have therefore made no attempt to reconstruct Fig. 15 and
Fig. 19, so as to show all the micronuclei in their correct relative positions.
These two figures were drawn with the assistance of the camera lucida from
one section only of each series of sections.

Lf f
7 . 1 7
s
2 ee
‘ ' ae f
\ \
\
x. -
z ~ “S
a ‘ ‘
a - Pa
s he - J 7
: 7 :
: : j
be
\\
s
a
a
X
5. “Sera
i ’ ‘
a >. = e ss a
- = - - .

Ede Se Sacre

A'S Huth, Lith! London.

Hickson del.

3.5

Quant Fourn. Micr Sa Vo 45 NSPLIS.

=
=
—~ TM. '

dM, AS Huth, Lith? London.

S.J Hickson del

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 363

On the Oviparous Species of Onychophora.
By

Arthur Dendy, D.Sc., F.L.S.,
Professor of Biology in the Canterbury College, University of New Zealand.

With Plates 19—992.

ConTENTS.
PAGE
1. IntRopuctTion . : 5 : : . 863
II. THe Genus OOPERIPATUS . : : ; . 3868
a. Diagnosis : ‘ : : : . 368
4. External Characters : , : . 93869
c. Internal Anatomy . : : : 372
d. YHgesand Development . ; : ; . 375
e. Distribution and (cology ; : . 386
J. Phylogeny A : 5 : 387
III. Description oF Species AND SYNONYMY : : 5 othe
]. Ooperipatus oviparus : : + 1893
2. Ooperipatus viridimaculatus : ; . 399
3. Ooperipatus insignis ; : . 403
IV. Summary or ReEsvuts ; ’ ; ; . 408
V. List oF LITERATURE REFERRED TO . , : . 410
VI. Descrretion oF Figures : ; : : ee lb}

I, Inrropvction,

THE present memoir is an attempt to bring together and
extend our information with regard to a very remarkable
group of Onychophora, the species of which are characterised
by their egg-laying habit and by a corresponding change in

364 ARTHUR DENDY.

the structure of the female organs. ‘The subject is one upon
which I have been engaged at intervals for more than twelve
years, and our knowledge of which for various reasons has
progressed very slowly. Owing in part, at any rate, to an
unfortunate confusion in nomenclature, for which I can
scarcely hold myself entirely responsible, my earlier observa-
tions were at first met with scepticism! and hostile criticism,
(Fletcher, 6) and the scarcity of material and difficult nature
of the investigation were equally discouraging. The recent
discovery of a new egg-laying species in New Zealand has,
however, stimulated further inquiry, and though I cannot
even now make the work anything like complete, I think
the time has come when a general account of the subject with
the necessary illustrations may be found useful. As previous
observations on these species have been published in scattered
periodicals, I prefix to this memoir a short historical notice.

In December, 1888, I found two specimens of Peripatus in
a fern-tree gully at Warburton, on the Upper Yarra, Victoria.
These specimens I described in a letter to ‘Nature’ (1), pub-
lished on February 14th, 1889.? Peripatus had previously
been known from Victoria only by a single specimen
discovered at Warragul by Mr. R. ‘I’. Baker, and exhibited
by Mr. J. J. Fletcher at a meeting of the Linnean Society of
New South Wales on July 27th, 1887. Mr. Fletcher (1)
considered the specimen to be “in all probability an example
of P. Leuckartii, Sanger.” Owing, however, to the
peculiar colour-markings of the Warburton specimens, I came
to the conclusion that these belonged to a new species, which
I, however, refrained from naming.

In reply to my letter to ‘ Nature’ Mr. Sedgwick (8) wrote
to the same journal (February 28th, 1889), and expressed
a doubt as to the distinctness of the Victorian species.

The examination of additional specimens unfortunately

1 Compare Sedgwick (4, p. 10). ‘* All species are viviparous. It has
lately been stated that one of the Australian species is normally oviparous, but
this has not been proved.”

2 Also in ‘The Victorian Naturalist,’ January, 18 9

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 365

convinced me that Messrs. Fletcher and Sedgwick’s sugges-
tions as to the specific identity of the Victorian species with P.
Leuckartii were correct, and on several subsequent occa-
sions (2, 3, 4, 5, 6, 7, 8, 9) I referred to the Victorian species
under that name.

In 1891 I made the somewhat surprising discovery that
the Victorian Peripatus, unlike all other known species, lays
eggs, and I therefore announced that P. Leuckartii was
oviparous (6, 7, 8, 9). For making this statement I was very
severely criticised by Mr. J. J. Fletcher (6), who certainly
showed conclusively that the common New South Wales
Peripatus is viviparous, like the great majority of species, a
fact which, through the kindness of my friend Mr. T. Steel,
I was subsequently enabled to verify for myself. The
unfortunate controversy on this subject, which has probably
done much towards preventing zoologists from appreciating
the true facts of the case, was really due to the confusion
between two species (the nomenclature of one of which is not
yet by any means definitely settled), so that perhaps it was
hardly worth while to say so much about it. As, however, I
replied fully (11) to Mr. Fletcher’s criticisms at the time,
I need say no more about them in this place.

Meanwhile, in 1890, I had described (5), under the name
Peripatus insignis, a second Victorian species, dis-
tinguished by the presence of only fourteen pairs of walking
legs, and this species was afterwards found by Professor
Baldwin Spencer (2) in ‘l'asmania, which would appear to be
its headquarters.

In my Presidential Address to the Biological Section of
the Australasian Association for the Advancement of Science,
at Brisbane in January, 1895, I pointed ont a further element
of doubt which enters into the nomenclature of the Australian
species of Peripatus. Professor Baldwin Spencer had
obtained in London a translation of Singer’s original
diagnosis of Peripatus Leuckartii, of which he had
kindly given me a copy. Concerning this I made the
following remarks in my address :

366 ARTHUR DENDY.

“The diagnosis commences: ‘Found in New Holland,
north-west from Sydney. Fifteen pairs of legs—one pair
without claws, fourteen with.’ If this be correct, then the
common Australian species usually accepted as P. Leuckartii
is certainly not the species described by Sanger under that
name, for I can certify that it has fifteen pairs of legs, all of
which bear claws. There appear to me to be two possibilities
in the case: (1) Sanger has failed to observe the claws on
one of the pairs of legs, or (2) there were really only fourteen
pairs of claw-bearing legs in his specimen, and he counted the
oral papille as a pair without claws. It is difficult to say
which of these alternatives is more likely to be correct, but
it seems just possible that my P. insignis may be the real
Leuckartii, with only fourteen pairs of claw-bearing legs.
The only way to settle the question definitely would be by
an appeal to Singer’s original specimen, which is stated to
have been in the possession of Professor Leuckart.”

Up to the present time no such appeal to Sanger’s original
specimen has, so far as I am aware, been made.

After my address was written, when passing through Sydney
I called upon Mr. Fletcher and discussed the question of nomen-
clature with him, and found that he had independently arrived
at conclusions similar to those contained in my manuscript.
It was then arranged that we should each contribute a paper
on the subject at the next meeting of the Linnean Society of
New South Wales, and that in my contribution I should
confine myself to the egg-laying species of Victoria, which
we agreed should receive a name. My description of P.
oviparus (16) was published in accordance with the above
arrangement.

Before finishing his paper on the subject (7), however, Mr.
Fletcher received specimens from Western Australia which
caused him to modify his views; and, after a lengthy discussion,
he comes to the conclusion that the most satisfactory arrange-
ment would be to consider all known Australian specimens of
Peripatus as referable to one comprehensive species with
four varieties, viz. :

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 367

1. P. Leuckartii, var. typica = P. insignis, Dendy.

2. P. Leuckartii, var. occidentalis, var. nov. (for the
West Australian specimens).

3. P. Leuckartii, var. orientalis, for the New South
Wales and presumably the Queensland specimens (the P.
Leuckartii, auctorum).

4, “The Victorian Peripatus to be dealt with by Dr.
Dendy,” referring to Peripatus oviparus, my description
of which is placed after Mr. Fletcher’s paper.

Although I myself pointed out that P. insignis might be
identical with P. Leuckartii, the latter possibly having only
fourteen pairs of claw-bearing legs, instead of fifteen, as
usually believed, yet I do not by any means consider that the
evidence is sufficient to justify the rearrangement pro-
posed by Mr. Fletcher. Indeed, it seems to me very
improbable that Singer should ever have had Peripatus
insignis in his possession. In any case the original
accounts by Leuckart and Singer are so inadequate and
apparently contradictory that the species (without re-exami-
nation of the type) cannot be certainly identified, and there-
fore we are fully justified in following the usual custom, and
applying the specific name Leuckartii to the common
New South Wales species, while retaining the name in-
signis for the very different southern species with fourteen
pairs of legs. This question will be more fully discussed
later on.

So far, however, from agreeing with Mr. Fletcher that all
the Australian forms are varieties of the same species, I
maintain that there are in Australia two genera of Onychophora
each with at least two species. I consider, in fact, that
P. oviparus may be regarded as the type of a new genus,
for which I have (19) proposed the name Ooperipatus,
and in which I also include P. insignis and the species
from New Zealand lately (17) described by me under the
name P. viridimaculatus. All these three species differ
from the common Australian Peripatus (P. Leuckartii,
auctorum) in the possession by the female of a prominent

368 ARTHUR DENDY.

ovipositor.!| P.oviparus and viridimaculatus certainly
lay eggs with thick sculptured shells. In P. insignis the
same habit may be inferred from the presence of the
ovipositor.

It is this extremely interesting group of egg-laying species
which forms the subject of the present memoir. So far as I
know no figures have yet been published of any of these three
species, and it is this deficiency in particular which I wish to
make good. Inasmuch as they all three agree very closely
as regards internal anatomy with one another, and, with the
exception of the reproductive organs, with the already well-
known species of Peripatus, I have not felt it necessary
to enter into details which would be merely repetition. It is
the female reproductive organs which are of chief interest,
and upon these attention has been mainly concentrated.

I have to thank many kind correspondents both in
Australia and New Zealand for assistance in collecting
material, and I am especially indebted to my friends Miss
Ferrar, of Christchurch, New Zealand, and Mr. C. C.
Brittlebank, of Myrniong, Victoria, for the large amount of
trouble which they have taken in order to satisfy my desire
to have coloured drawings of the three species.

II. THe Genus Oopzripatus, Denny (19).
a. Diagnosis.

Oviparous Onychophora. Eggs with thick sculptured
chorion. Genital aperture in the female at the end of a
prominent ovipositor which les between the legs of the last
pair; in the male the aperture is only slightly prominent
between the legs of the last pair. Male with crural glands.
Legs with three spinous pads. ‘Transverse ridges of the

1 It seems probable that the female from Queensland described by
Sedgwick (2) as P. Leuckartii was really a specimen of Ooperipatus
oviparus, for Sedgwick says “ the genital papilla of the female is remarkably
prominent, and bears at its free end a longitudinally disposed slit.”

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 369

integument interrupted in the mid-dorsal line by a narrow
longitudinal groove, from the floor of which pigment is
absent. Papille in an approximately single row on each
transverse ridge.

b. External Characters.

Shape and Size.—In shape all three species do not differ
conspicuously either amongst themselves or from other
Australasian Onychophora. It must be remembered that
we have in Australasia species of Onychophora with
fourteen [O. insignis, O. viridimaculatus], fifteen
[P. Leuckartii, P. occidentalis, P. nove-zealandia,
O. oviparus], and sixteen [P. Suteri] pairs of walking
legs respectively, and, as might be expected, the size
of the adult appears to be proportional to the number
of pairs of legs, O. insignis and O. viridimaculatus
being, so far as my experience extends, usually some-
what smaller than P. Leuckartii, P. nove-zealandia,
and O. oviparus; while P. Suteri! is considerably
larger than any of the others. The “anal cone”
(figs. 10, 31) is only about as long as the legs of the last
pair. It is less well developed than represented by Sedgwick
(2, fig. 23) in P. novee-zealandiz, upon which Bouvier’s (2)
diagram of the genus Peripatoides appears to be founded.

Appendages.—The number of claw-bearing walking legs
is, as has just been stated, either fourteen or fifteen in the
species of Ooperipatus hitherto described, but in view of
past experience it would be unwise to generalise from these
three species, and it is quite possible that species with more
or even fewer pairs of legs may be forthcoming. Each leg is
provided with three spinous pads, of which the middle one
is the broadest, while the proximal one is very narrow. The
foot is provided with the usual pair of horny claws and
dorsally with three conspicuous primary papille—anterior,
posterior, and median.

The oral papillae have the usual structure, consisting each

1 Dendy (18).

370 ARTHUR DENDY.

of a proximal portion with little pigment and no papille, but
pleated transversely and capable of extension, like an
accordion; and a distal pigmented portion, almost hemi-
spherical in form, and bearing papille.

The jaws, also, are quite normal in structure. The outer
blade has an accessory tooth at the base in O. oviparus
(figs. 7, 8), but this tooth is extremely small and irregular,
while in O. insignis and O. viridimaculatus (figs. 29, 34)
there is no accessory tooth. The inner blade (figs. 7, 30, 35)
has from about five to about eight accessory teeth.

The ringed antenne have no special characteristics as
compared with other species. They are broadly rounded,
or even perhaps slightly club-shaped at the extremity, and
very similar, as is the case also with the oral papille, to
those which Sedgewick has figured for P. nove-zealandie.

External Apertures.—The mouth with its tumid lips,
tongue, and jaws; the terminal anus; the apertures of the
slime-glands and the nephridial apertures (fig. 28) resemble
closely the corresponding parts in other Onychophora. In
the fourth and fifth pairs of legs the nephridial aperture is
shifted outwards into about the middle of the narrow
proximal spinous pad, which is thus divided transversely
into two parts, separated by the nephridial papilla.

The genital aperture of the female is a longitudinal slit
at the extremity of a very protrusible ovipositor (figs. 9, 10,
27, 31) placed between the legs of the last pair. The
ovipositor when retracted is still a conspicuous organ ; its
shape is subcylindrical, bluntly rounded, and may be
slightly enlarged at the extremity, and covered, at any rate
distally, with small spinose papille of a yellowish colour. It
is capable of being extended until two or three times the
length of the legs, and in this condition I have seen it in a
drowned specimen (fig. 10) of O. oviparus, in which it was
4 mm.long. It is of about the same size and shape in all
three species.

The male genital aperture les between the legs of the last
pair, and is only slightly prominent.

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 371

Crural glands are apparently developed in the male only,
and their apertures are more or less conspicuous on the under
surfaces of more or fewer of the legs in all three species
(fig. 28). They are present in most if not all of the legs, and
T include under this name the glands of the last pair of legs,
whose apertures lie at either side of the genital orifice. The
apertures of the erural glands are borne on white papille
which are retractile, and may be considerably protruded in
the form of thin-walled vesicles. Behind the male orifice is
yet another pair of white papille, probably bearing the
apertures of accessory glands.

Tracheal pits occur scattered over the surface of the
body, as seen in sections of P. oviparus (fig. 5, 7. P.). In
this species (and probably in the others) there is a tracheal
pit immediately in front of the mouth, and a pair of very
large ones opening in the buccal cavity just behind and close
to the base of the inner jaw on each side, and running back-
wards for some distance, at first just outside the lateral nerve-
cords (fig. 5, B. Tr.) and then above them and just inside the
salivary glands. ‘These buccal tracheal pits have a thick
chitinous lining, and may be traced back in a series of
transverse sections very nearly to the level of the second
pair of walking legs. ‘They give off along their course and
from their extremities an immense number of very fine
tracheal tubes. When the jaws are removed, these enor-
mously elongated tracheal pits may be pulled out in connec-
tion with them, and the chitinous lining of the pit appears to
pass over into the chitinous covering of the smallest accessory
tooth of the inner jaw.

Integument.—The structure of the integument closely
resembles that described by Gaffron (1) for P. Edwardsii.
As usual, it is furrowed by narrow transverse grooves, and
produced into papille of varying size on the intervening
ridges. There is approximately a single row of papille on
each ridge. The pigment is lodged in the polygonal,
nucleated, epidermic cells, outside each of which the cuticle
forms a small, transparent, sharp-pointed, scale-like spine.

372 ARTHUR DENDY.

These small spines are prominent upon the papille, each of
which may bear in addition a single spine at its apex.

On the spinous pads of the feet the spines are long and
slender, and covered near the base with numerous minute
secondary spines.

Along the mid-dorsal line there runs a very narrow longi-
tudinal groove (figs. 5, 6, D. F.), from the floor of which
pigment is absent, giving rise to a narrow white line which
may be concealed by overarching of the lips of the groove in
contracted specimens.

The predominant colour of the pigmented epidermic cells,
when seen by transmitted light, is a beautiful indigo-blue,
which may be replaced by green, tawny orange, brown, or
nearly black, giving rise to a variety of patterns, which will
be described when dealing with the specific characters.

The eyes appear externally as a pair of small, pearl-like,
hemispherical protuberances in the usual situation, one just
behind and on the outer side of the base of each antenna. In
one specimen of O. oviparus the eyes appeared of a bright
red colour.

c. Internal Anatomy.

General.—The internal anatomy of Ooperipatus (as
exemplified by O. oviparus and O. viridimaculatus)
agrees so closely in most respects with the well-known
Peripatus type, that it seems almost superfluous to do more
than refer to the general dissections shown in figs. 4 and 27.

The slime-glands are enormously developed, so that when
unravelled they are much longer than the entire animal.
When the animal is starved for some time, the tubular slime-
reservoir becomes greatly distended by the accumulation of
the secretion, which the animal has been unable to make use
of in capturing ‘its prey (fig. 27). I have elsewhere (4)
pointed out that in O. insignis the secretion of the slime-
glands contains very numerous corpuscles, but as the liquid
rapidly hardens into an enamel-like mass on exposure to the

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 373

air, it is difficult to say whether or not they are amceboid, as
is the case with the similar corpuscles found in the liquid
discharged from the nephridia (in O. oviparus).

Reproductive Organs of the Male.—The male organs
of reproduction are, owing to the small size of the animal and
the brittle nature of the organs themselves, extremely difficult
to dissect in spirit-preserved material. I have, however, made
some observations on the subject in the case of O. viridi-
maculatus, and the following notes are perhaps worth
recording.

The unpaired portion of the male duct (vas deferens) is
extremely long, considerably exceeding the entire length of
the animal when at rest. It has very much the same
structure as described by Gaffron (1) in the case of
P. Edwardsii. Its middle portion contains along spermato-
phore, while the terminal portion forms a long muscular
ductus ejaculatorius. The spermatophore is in general
similar to that of P. Edwardsii. It appears to contain no
ripe spermatozoa, but numerous spherical sperm mother-cells.
It is irregularly swollen out at intervals, but the sperm
mother-cells occur throughout almost the entire length. The
outer cover of the spermatophore is a thick, homogeneous,
transparent, apparently chitinous sheath, and there appear to
be no spherical globules, such as occur on the surface of the
spermatophore in P. Edwardsii.

Reproductive Organs of the Female.—Of greater
interest are the internal reproductive organs of the female,
which I have been able to study in freshly-killed specimens,
and of which I have already (15) given a short account in
the case of O. oviparus. The ovary (O. oviparus and
O. viridimaculatus) is as usual placed far back in the
body cavity above the alimentary canal (figs. 4, 6, 12, 27, 31).
It consists of right and left halves united with one another in
front and behind (fig. 12), and attached by a mesoarium to
the pericardial septum in the mid-dorsal line. It contains a
large number of eggs, varying enormously in size according
to the amount of yolk which they have received, the structure

VOL. 45, PART 3,—NEW SERIES. coe

374 ARTHUR DENDY.

of which will be considered in the next section. The wall of
the ovarian tubes is extremely thin and delicate, and in such
a state of collapse that the tubular character is extremely
difficult to demonstrate. The eggs project into the body-
cavity from the outer surface of the ovary in all stages of
intra-ovarian development, the larger ones being attached by
short epithelial pedicles (fig. 13) and containing much food
yolk, and the entire structure, when exposed by dissection,
appears as a mass of eggs held together by a thin, transparent
membrane, but readily separable into right and left halves
except in front and behind.

The oviducts, owing to their great length, are in their
natural position much convoluted in the hinder part of the
body-cavity (fig. 81), but do not extend forwards much, if at
all, beyond the ovary itself. They have a common origin
from the posterior end of the ovary (figs. 4, 12, 27), with
which they are directly continuous. There are no receptacula
ovorum, but the receptacula seminis (figs. 4, 11, 12, 27) are
well developed, and each opens as usual by two short ducts
(fig. 11) into the oviduct of its own side at only a very short
distance from the ovary. ‘The receptacula seminis may
contain spermatozoa, and it is difficult to believe that the
latter enter the body of the female through the integument,
as has been suggested for P. capensis.' Both the extreme
toughness of the integument and the presence of receptacula
seminis seem to argue against such a hypothesis.

In O. oviparus (fig. 4) I was able to recognise a division
of the oviduct into three parts, though by no means sharply
defined. All three parts are narrow, except where swollen out
by the contained eggs. The first is very short, and extends
from the commencement of the oviduct in front to the recep-
taculum seminis behind. Its wallis much folded, and provided
with little irregular protuberances (figs. 11, 12, Pr. O.) on
the side opposite to the receptaculum. ‘I'he middle and last
portions of the oviduct are of about equal length; the former
has very thick glandular walls and the latter very thin

1 Compare Sedgwick (4).

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 375

membranous walls, though whether the difference is simply
due to stretching by the contained eggs may be regarded
as an open question. At their hinder ends the oviducts
unite in a thick-walled, muscular, triangular sac (figs. 4,
9, 10, Tr. S.), whose posterior angle is continued into the
ovipositor. Over this sac the nerve-cords pass, enlarging
upon its dorsal surface to form a pair of especially large
ganglia. The ovipositor isa thick-walled muscular organ,
with an outer layer of more or less longitudinally and an
inner layer of more or less circularly or obliquely arranged
muscle-fibres. Eggs were found in the middle and last
portions of the oviducts, but much more abundantly in the
last. Their number, of course, varies; thus in one specimen
there were three eggs in each oviduct ; ina second, six in one
and seven in the other; in a third, eight in one and nine in
the other.

In O. viridimaculatus the structure of the female repro-
ductive organs (figs. 27, 31) is closely similar, but I have not
detected any differentiation corresponding to that between the
second and third parts of the oviduct in O. oviparus, while
the short proximal division of the oviduct in front of the
receptaculum seminis is devoid of distinct protuberances,
though its wall is folded. The number of eggs produced in
this species appears to be smaller than in O. oviparus, for
of the two freshly killed specimens dissected, one (fig. 31)
contained only a single egg in the right oviduct and none
in the left, while the second (fig. 27) contained only two in
the right and one in the left oviduct; the former specimen,
however, had two very large ovarian eggs, apparently almost
ready to enter the oviducts (fig. 51).

d. Eggs and Development.

The Question of Oviparity.—Notwithstanding the
scepticism of certain writers, who have apparently never
thoroughly investigated the species in question, I do not
think that any impartial observer could hesitate for long in

376 ARTHUR DENDY.

pronouncing both O. oviparus and O. viridimaculatus to
be genuinely oviparous, and, by analogy, O. insignis may
pretty safely be included in the same category. As, how-
ever, the eggs have not yet been actually observed in the
case of the latter species, it will be as well to reserve our
final judgment in this case, and the following remarks, of
course, apply only to the two former.

In the case of O. oviparus, I have already (6, 7, 8, 9, 10,
11, 12, 15) entered pretty fully into the question of the egg-
laying habit ; some years ago eggs were laid by this species
in my vivarium at Melbourne, one of which, after an extra-
ordinarily long period of development (seventeen months),
finally hatched out. In the case of O. viridimaculatus,
only one deposited egg has yet been observed, and that was
found by me last autumn in rotten wood in which a specimen
of O. viridimaculatus had been packed for transmission,
the animal itself having unfortunately died on the journey.

Equally strong evidence is afforded by the fact that no
specimen of either species has yet been found to contain
recognisably developed embryos, while nearly all the females
that have been dissected have contained large thick-shelled
egos. In other species, such as P. Leuckartii (the common
New South Wales species), as is well known, the adult
females invariably contain developing embryos enclosed only
in very thin transparent membranes.

The idea that the deposition of the eggs by Ooperipatus
is merely an abnormal phenomenon, such as occasionally occurs
in Peripatus, has been practically refuted by Steel (1), who
observes of P. Leuckartii (NewSouth Wales) that “ pregnant
females somewhat readily extrude the young when distressed
by close confinement or uncomfortable conditions. Frequently
soft adventitious eggs are laid. ‘These bear no resemblance
to those described by Dendy from P. oviparus, but are
quite smooth and have a very flaccid thin envelope. They
very soon break up into a drop of turbid liquid. My supposi-
tion is that they are merely ova which have escaped fertilisa-
tion, and are thus making their natural exit from the body.”

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 377

The presence of the ovipositor in Ooperipatus alone
amongst Onychophora is in itself strongly indicative of the
oviparous habit, while the elaborate and closely similar
structure of the egg-shell in both species in which it has been
observed also indicates that the habit is constant and normal.

The Kegs.—tThe ovarian eggs (O.oviparus), formed pre-
sumably from the epithelial cells of the ovarian tubes, range
in size from about 0:037 mm. to about 1°4 in diameter. In the
young ovarian ege (fig. 14), as seen in sections stained with
borax carmine, the vitelline membrane, if present, is very thin,
the cytoplasm uniformly and very finely granular, and the
nucleus very large in proportion to the entire egg. There is a
very distinct nuclear membrane, and a single large and remark-
ably well-defined spherical nucleolus, with finely granular,
darkly-staining contents. The nucleolus is placed excentri-
cally, and around it the nucleoplasm stains very lightly, and
sometimes at any rate has the appearance of being vacuolated
(fig. 14). Outside this light area the nucleoplasm stains fairly
darkly with borax carmine. In older ovarian eggs (fig. 15)
the vitelline membrane is much thicker and the cell-body has
increased in size much more rapidly than the nucleus, owing
to the deposition in the former of large quantities of food-
yolk, which appears first in the form of minute highly
refringent granules thickly scattered through the cytoplasm
(fig. 15). In still older ovarian eggs these granules appear
to be arranged around spherical globules of some clear trans-
parent substance (fig. 16). Some of these globules occupy
the interior of irregular polygonal corpuscles slightly larger
than themselves (fig. 16).

In order to study further the structure of the yolk, a
quantity was removed from an egg which had passed into
the oviduct. ‘This material, which was obtained from a
specimen of O. oviparus which had been preserved for a
long time in alcohol, was partly treated with osmic acid, and
partly stained with eosin and examined in oil of cloves. The
yolk (fig. 17) was found to consist chiefly of the clear,
transparent globules above mentioned, each enclosed in a

378 ARTHUR DENDY.

polygonal corpuscle of almost homogeneous structure. The
polygonal shape of the corpuscles is apparently due to mutual
pressure. Their diameter is about 0:016 mm. Both the
globular body and the enclosing corpuscle appear to be
stained fairly darkly by alcoholic solution of eosin, but osmic
acid (2 per cent. solution) has little effect on either, and in
sections treated by the borax-carmine and acid-alcohol
method they appear quite unstained. Whether the highly
refractive yolk granules observed in the ovarian eggs and
shown in fig. 15 have any relation to the corpuscles and their
contained globules, it is impossible at present to say with any
degree of certainty, but it seems not unlikely that the yolk
is first deposited in the finely granular form, and subsequently
converted into the comparatively large globules with their
enveloping corpuscles. The appearances shown in fig. 16
support this hypothesis.

The Egg Envelopes.—The ovarian egg, as we have
already seen, is enclosed in a distinct, transparent, appa-
rently structureless vitelline membrane, which attains only a
moderate thickness, and persists inside the chorion or egg-
shell for a very long time, probably throughout the entire
development, for I found it around an embryo in an egg
which had been laid for at least eight and a half months.
This vitelline membrane is, I believe, homologous with the
thin membrane which alone surrounds the embryo of Peri-
patus Leuckartii (New South Wales), and which Steel (1)
has shown to persist until birth. Willey (1) considers this
membrane in the New South Wales species to be a chorion
and not a vitelline membrane, but I can see no reason for
maintaining this view. ‘The “shell” described by Miss
Sheldon (2) in the ovarian eggs of P. nove-zealandia,
P. Balfouri and P. capensis, is also probably a vitelline
membrane, but the shell which the same writer [Sheldon (1) ]
describes as lying outside the vitelline membrane in the
uterine egg of P. nove-zealandiz is no doubt correctly
regarded as a chorion.

In Ooperipatus oviparus and O. viridimaculatus

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 379

the chorion is very strongly developed and has an elaborate
structure. It is formed while the egg is passing down the
oviduct, evidently partly, if not entirely, by a secretion of the
walls of the latter, and is, I believe, homologous with the
coating of the spermatophore in the male and with the
chorion of the uterine egg in P. nove-zealandiz.

The structure and mode of formation of the chorion are,
however, extremely difficult to understand, and I propose to
describe it separately in the two species in which it has been
observed.

In O. oviparus I have observed three well-marked stages
in the development of the chorion, which we may term A, B,
and C, respectively.

Stage A.—This stage I have observed only in a specimen
from Cooran, Queensland, collected by Professor W. Baldwin
Spencer, in October, 1891. In surface view the chorion
exhibits a finely and uniformly punctate appearance, except
over certain areas which appear clear and _ transparent
(fig. 18). These clear areas are rounded in outline, and
distributed at fairly regular intervals over the chorion. In
a specimen stained with eosin the clear areas seem to be
crossed by very irregular networks of darkly stained threads,
and in many of them a darkly stained nucleus-like body is
visible (fig. 18). Possibly each of the clear areas marks
the position of an epithelial cell, derived either from the
ovarian follicle or from the wall of the oviduct, but the
scanty material at my disposal has not enabled me to decide
this question. In optical section (fig. 19) the chorion is seen
to be very thick (about 0°02 mm.), except in places corre-
sponding to the clear areas, where it thins out very greatly.
It is also seen to be marked by fine vertical striz, possibly
canaliculi, whose cross sections undoubtedly give rise to the
punctate appearance seen in surface view. The fact that this
specimen was collected in the spring probably accounts for
the incomplete development of the chorion; the eggs would
probably not be deposited till much later in the season, as
I shall show subsequently.

380 ARTHUR DENDY.

Stage B.—This stage was found in an animal which had
lain for a long time in spirit previous to dissection, and the
chorion had assumed a brown colour. Where thin places
occurred in the Queensland specimen we now find thick ones,
forming rounded protuberances very regularly scattered over
the surface (figs. 20, 21, 22). The chorion now consists of
two distinct layers, an inner much thicker one exhibiting
radial striation and evidently corresponding to the chorion
as already described in Stage A, and an outer one which is
a new formation—or at any rate was only beginning to be
formed in the preceding stage. This outer layer is developed
chiefly, if not entirely, over the thin clear areas in the inner
layer, which thus come to protrude externally in the form of
low, rounded knobs. ‘The outer layer is for the most part
clear and structureless, and has a chitinoid appearance, but
in this particular specimen a group of highly refringent
granules is very conspicuous in the middle of each protuber-
ance. This clear outer layer of the chorion probably increases
in thickness as the egg passes down the oviduct.?

Stage C.—This stage occurs in eggs which have been
deposited, and may be regarded as the mature condition.
The two layers are still clearly recognisable (figs. 25, 26).
The inner one exhibits the usual radial striation, but it no
longer shows the thin areas of the first stage, which seems to
indicate that unequal thickening has taken place in it. The
outer layer is clear and transparent, and the refringent
granules of the preceding stage are no longer visible.
The extent to which it has become thickened varies in
different cases; fig. 26 shows, in optical section, a specimen
in which the protuberances remain comparatively low, though
the outer layer is clearly recognisable even between them.
Fig. 25, on the other hand, is drawn from a specimen in
which the protuberances were remarkably strongly developed.
After the deposition of the egg, the outer layer of the chorion

1 It seems possible that the nucleus-like bodies observed in Stage A may

belong to cells which are concerned in the formation of the outer layer of the
chorion.

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 381

undergoes a certain amount of wrinkling, probably in drying,
and the wrinkling gives rise to the extremely characteristic
surface pattern or sculpture to which I called attention (cf. 7)
long ago, and samples of which are shown in figs. 23 and 24. It
will be seen that the pattern is not quite identical in all cases,
the differences, doubtless, depending upon the thickness of the
outer layer and the particular conditions under which drying
takes place.

The views expressed above as to the formation of the
chorion in Ooperipatus oviparus differ to some extent
from my previously expressed ideas on the subject, but
the question is an extremely difficult one, and cannot even
yet be regarded as by any means finally settled. We
may, however, now regard it as certain that the chorion
consists of two layers, an inner thick and radially striated
layer, and an outer clear layer which is thickened over
certain areas to form mound-like protuberances. The fully
formed chorion is a tough, flexible membrane, varying in
thickness up to about 0°04 mm. (exclusive of protuberances).
When newly laid it has a pale yellow colour, but preserved
specimens may become much darker, and the deposited
egos darkened greatly in course of time.

In Ooperipatus viridimaculatus the chorion certainly
has a structure very similar to that exhibited by O.
oviparus. I have observed stages corresponding to B and
C of the former, B in the oviduct and C after deposition.
In the former I was unable to detect the radial striation
in the very thick inner portion, but the rounded protuber-
ances on the outside were quite distinct. In the latter
(figs. 32, 33) the radial structure was indicated by the
strong tendency to split up radially which the chorion
manifested on compression. ‘The rounded protuberances
on the outer surface had become to some extent wrinkled
(fig. 32), but not to the same degree as seems to be usual in
O. oviparus. It must be borne in mind, however, that only
a single egg of O. viridimaculatus was observed after
deposition, and that may possibly have been prematurely

382 ARTHUR DENDY.

laid, owing to the unfavourable conditions which resulted
in the death of the parent.

Number of Eggs produced; Size and Shape.—The
number of eggs produced by the different species of
Onychophora appears, as in other animals, to be inversely
proportional to the amount of yolk, and consequently to the
size of the egg. Thus in the viviparous P. Leukartii
(New South Wales) the ovarian egg is comparatively small
and with little yolk, the largest ovarian egg which I have
seen measuring only about 0°37 mm. in greatest diameter,!
while I have myself found as many as thirty embryos in
the uteri of one female, and Mr. Fletcher (7) has found as
many as fifty-three. In P. nove-zealandie the egg,
according to Miss Sheldon, measures as much as 1°5 mm, in
greatest diameter, and in this species, which is also viviparous,
Captain Hutton (1) found from four to twenty-six embryos
in each female, and Miss Sheldon from seven to eighteen.
In O. oviparus the eggs are a little larger, measuring (in
the oviduct and when laid) about 1:9 by 1°5 mm., while the
number found in the two oviducts varied from six to seven-
teen. In O. viridimaculatus the eggs are perhaps a shade
longer and somewhat narrower in proportion to their length,
and the number produced appears to be still smaller, the
three females dissected containing respectively only one,
three, and seven eggs in the two oviducts.

The eggs are deposited in the winter or in the late autumn,
which perhaps accounts for the fact that in the only specimen
of O. oviparus captured and dissected at the end of July,
the oviducts were found to be empty.

Only one batch of deposited eggs of O. oviparus and a
single deposited egg of O. viridimaculatus have, as yet,
so far as I am aware, been obtained. ‘The former was laid
(at any rate fourteen of them) between May 18th and
July 31st, 1891, in a vivarium in Melbourne, in which there
were three females and one male animal. Fourteen eggs

1 As I have carefully examined the ovarian eggs in only one specimen, killed
in January, this result needs confirmation.

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 383

were found beneath and in the crevices of bits of rotten
wood placed in the vivarium, and a fifteenth, though per-
haps laid about the same time, was not found till the
vivarium was completely turned out on September 16th.

The single deposited egg of O. viridimaculatus was
found on April 18th, 1900, amongst rotten wood in which an
adult animal had been forwarded to me from Lake T'e Anan,
New Zealand. The package was unfortunately delayed in
transit, and the animal itself was dead when it arrived, so
that the deposition of the egg may have been hastened by
abnormal conditions.

Development.—Owing to the great scarcity of material,
it is impossible to say much under this head, and the remarks
which I have to offer, except as regards the newly laid egg,
are based exclusively upon O. oviparus. ‘T'wo facts, how-
ever, stand out clearly, the first being that the newly
deposited egg contains no embryo recognisable by ordinary
means, and the second that the development progresses
extremely slowly, and may occupy as much as seventeen
months from the time of laying to the time of hatching.

The recently laid egg is filled with a milky fluid containing
very many yolk granules, the structure of which has already
been described (fig. 17).

The first embryo, already far advanced in development,
was removed on November 380th from the egg found on
September 16th, so that it was at least ten weeks old from
the time of oviposition, and probably a good deal older.
Within the chorion the embryo was surrounded by the thin,
transparent, vitelline membrane, which fitted closely on to it
and was very difficult to remove. ‘The total length of the
embryo, exclusive of the antenne, is about 4 mm. It is
spirally coiled, making rather more than one complete turn
of the spiral, so that the posterior extremity lies against the
side of the neck, pointing in an opposite direction from the
head, and the ventral surface occupies the inside of the
curve. ‘lhe form of the body is already very lke that
of the adult, except that the head is larger in proportion,

384 ARTHUR DENDY.

The head when viewed from above appears broadly and
squarely truncated between the bases of the antenne, one
of which has been accidentally broken off, the other pointing
forwards. The two cerebral ganglia and the eyes (with
pigment) are clearly recognisable. The antenne are well-
developed and annulated. Owing to the position of the
embryo (which is now mounted in Canada balsam) it is
possible to see only one of the oral papille, and the jaws are
completely hidden from view. There are perhaps nine pairs
of walking legs, but it is impossible to count the exact
number. They appear as blunt outgrowths of the body,
ventro-lateral in position. No claws are yet visible, and little
or no pigment is yet developed in the integument.

The second embryo obtained was at least eight and a half
inonths old from the time of egg-laying, and was already a
perfect young animal, differing externally from the adult only
in its smaller size and less deeply pigmented skin. It was
removed from the egg on April 14th, 1892, when only three
eggs remained in the hatching box, the others having been
taken away, as they showed signs of going bad.’ One of the
remaining three had been showing dark pigment inside for
some days past. This egg I removed and dissected. I found
the shell of a much darker (yellow) colour than when laid, a
good deal crumpled on the surface, and very soft, as though
beginning to decay away. ‘The contained embryo was
removed and found to be in excellent condition. It was
enclosed in the usual vitellme membrane within the chorion.
As in the previous case it was tightly coiled up. When
uncoiled it measured about 5 mm. in length (exclusive of the
antenne) and 1 mm.in breadth. All the appendages were
developed, namely, antennee, oral papilla, jaws, and fifteen
pairs of claw-bearing legs. The eyes were conspicuous at
the bases of the antenne, and the antennze themselves
showed each about twenty deeply pigmented annuli. The

1 The majority of the eggs unfortunately perished, shrivelling up and being
altacked by a mould; which is hardly to be wondered at when we consider the
difficulty of keeping the conditions suitable for so long a period.

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 385

remainder of the body was nearly white, but very distinct
isolated pigment patches (chiefly indigo-blue, with a few
specks of orange) appeared scattered pretty abundantly over
the legs and back. The mouth was surrounded by the very
characteristic, thick, transversely furrowed lip. The dermal
papille were very obvious, and exhibited the characteristic
spines, the cuticle being very strongly developed. The claws
on the feet were very distinct. The alimentary canal was
full of granular food-yolk.

The embryo just described was between eight and a half
and eleven months old from the time of egg-laying, and was
apparently just about ready to hatch. It is therefore very
remarkable that another embryo of the same batch remained
in the egg for another eight and a half months (or there-
abouts) before hatching.

About the end of 1892 only a single egg remained in the
hatching box, all the others having either gone bad or been
used for investigation. On January 38rd, 1893, not having
opened the box for some days, I found this egg, so far as I
could tell, in its former position, lying on a small piece of
rotten wood, which rested on the glass bottom of the hatching
box. ‘The shell, however, was split on one side, and the young
animal had made its escape. It was found lying dead on the
glass, 25 mm. from the egg-shell. ‘he animal was itself
only about 5 mm. in length, so that, even assuming that it
had moved in a perfectly straight line, it must have crawled
for a distance of five times its own length, off the rotten
wood and along the glass to the position in which it was
found.

To the naked eye the young animal appeared of a pale
greenish colour. It cannot have been dead for very many
days, but decomposition had already set in, and the body
adhered to the glass upon which it lay, so that it could not
be removed without considerable injury. I mounted it in
Canada balsam, however, and even in its present condition it
shows under the microscope such characteristic features as
claws and jaw-blades and indigo-coloured pigment. The

386 ARTHUR DENDY.

ruptured chorion also still shows a portion of the character-
istic pattern on the outside.

It appears, therefore, that the only egg yet known to hatch
out, did so after being laid for about seventeen ‘months.
There is no reason, however, to believe that under natural
conditions the process of development takes as long as this,
and the apparently fully-formed embryo obtaimed at about
eight and a half months indicates a normally shorter period.
Development was possibly retarded by the eggs being kept
in a room which was unusually cool in summer; probably, also,
conditions of moisture and the softening by decay of the
tough chorion, have a good deal to do with the date at which
the young animal escapes.

e. Distribution and Cicology.

The genus Ooperipatus, so far as we know at present,
is confined to the eastern portion of Australia (including
Tasmania for O. insignis) and New Zealand, so that its
distribution forms a striking parallel to that of many
vegetable types, and may be similarly accounted for by
supposing it to have spread southwards at a time when New
Zealand was connected with North-eastern Australia, probably
not later than the cretaceous period. O.oviparus has been
found in Victoria, New South Wales, and Queensland,
O. viridimaculatus in the South and possibly also the
North Island of New Zealand, and O. insignis in Victoria
and ‘Tasmania.

Like other members of the group, the species are all
thoroughly cryptozoic, hiding away beneath stones or logs, or
in the crevices of decaying tree trunks. O. oviparus Is
usually found beneath stones or fallen logs, either on the
ground, or attached to the under surface of the stone or log
which covers it. O. viridimaculatus, on the other hand,
is, according to my experience, found in the interior of rotten
tree trunks, which have to be broken to pieces in order to
obtain specimens, and I think the same will usually be found

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 387

to be true of O. insignis. All species seem to be very rare,
though pretty widely distributed. Thus I have never heard
of O. oviparus or O. insignis being found in anything
like the quantity in which Steel (1) has found Peripatus
Leuckartii in New South Wales, nor is O. viridimacu-
latus anything like so common in New Zealand as is
P. nove-zealandie.

The nature of the locality in which the logs or stones used
as cover may occur varies greatly. Thus O. oviparus may
be found under logs in a fern-tree gully, or under stones on a
bare hill-side,! while probably the largest collection of this
species? yet made was obtained at an altitude of 5000 feet
or more, on Mount Kosciusko,® and amounted to thirty-five
specimens.

f. Phylogeny.

The genus Ooperipatus evidently stands in the closest
phylogenetic relationship to the other Australasian genus
of Onychophora—Peripatoides, which occurs side by
side with it both in Australia and New Zealand. The genus
Peripatoides was founded by Pocock (1) in 1894, for the
reception of the Australasian species of Onychophora, and
characterised as having the legs furnished with only three
spinous pads, and the generative aperture between the legs of
the last pair and well in advance of the anus. These characters
are also found in Ooperipatus, which is distinguished from
Peripatoides, however, by its egg-laying habit and the
presence of the ovipositor in the female.

Of Peripatoides I distinguish four species, as follows:

1. P. Leuckartii, auctorum,* with 15 (fifteen) pairs of
claw-bearing legs and an accessory tooth on the outer blade
of the jaw. Male withcrural glands. Characteristic of New
South Wales.

1 Compare Dendy (2).

? Steel (2) has identified the Kosciusko specimens as P. oviparus; they
were described by Fletcher (3) as P. Leuckartii.

3 Compare Fletcher (3).

4 =P. Leuckartii, var. orientalis, Fletcher (7).

388 ARTHUR DENDY.

2. P. occidentalis,! Fletcher, with 15 (fifteen) pairs
of claw-bearing legs and no accessory tooth on the outer
blade of the jaw. Male with crural glands. Characteristic
of Western Australia.

3. P. nove-zealandixw, Hutton, with 15 (fifteen) pairs
of claw-bearing legs and no accessory tooth on the outer
blade of the jaw. Male without crural glands. Characteristic
of New Zealand.

4. P. Suteri,? Dendy, with 16 (sixteen) pairs of
claw-bearing legs and no accessory tooth on the outer blade
of the jaw. North Island of New Zealand.

In the presence of crural glands in the male, Ooperipatus
is most closely related to the Australian species of Peripa-
toides, but in the structure of its eggs it stands nearer to
P. nove-zealandiw. We may take P. Leuckartii and
P. nove-zealandie as typical of the Australian and New
Zealand sections of their genus respectively. P. Leuckartii
is viviparous. ‘lhe ovarian eggs, so far as my experience goes,
are comparatively small, and I can find no trace of a chorion,
but only a vitelline membrane. P. nove-zealandia,
though viviparous, has eggs nearly as large as those of
Ooperipatus, and according to Miss Sheldon, there is a
second membrane (chorion) outside the vitelline membrane.
Thus P. nove-zealandiz in the structure of its eggs to
some extent resembles Ooperipatus, and this fact may help
us to solve the vexed question whether the oviparous or
viviparous habit is the more primitive.

This question has been recently discussed by Willey (1), and
he has come to the conclusion that “the oviparity of P.
oviparus, Dendy, is an acquired habit, and not in any way
to be confused with the primitive deposition of alecithal ova.”
his may be very true, but Iam inclined to think that even
if the structure of the eggs is secondary, yet the viviparous
habit must be regarded as still more recently acquired.

P. nove-zealandia, with its large, heavily yolked eggs

1 =P. Leuckartii, var. occidentalis, Fletcher (7).
2—= P. nove-zealandie, var. Suteri, Dendy (13).

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 389

and chorion, may be either progressing towards a condition of
oviparity or away from it. Organs are not usually developed
ahead of their uses, and the chorion may be regarded as a
vestigial structure inherited from oviparous ancestors, in which
a chorion formed an essential protection to the deposited egg
during the lengthy period of development. In P. Leuckartii
and other viviparous species the chorion may have completely
disappeared.

It is hardly likely that the oviparous habit should have
been independently acquired by two or more different species
in Australasia, but in no other part of the world. Yet,
unless we regard P. nove-zealandiz as representing an
ancestral form of Ooperipatus, which is hardly admissible,
if only on account of the absence of the crural glands, this
is the conclusion to which we should be forced if we accept
the viviparous as being more primitive than the oviparous
habit. If, on the contrary, we regard the oviparous habit
as being the more primitive, it is not difficult to arrange the
Australasian Onychophora in a phylogenetic series, which may
be tentatively represented as follows :

P. Suteri P. nove-zealandie P.Leuckartii P. occidentalis
rs a \ Zz ‘
= \S ee
SSK Le

X yh
\ /
NS,

N

Ooperipatus

Of the three species of Ooperipatus,two (O.insignis and
O. viridimaculatus) are very nearly related to one another.
Each has only fourteen pairs of walking legs and no accessory
tooth on the outer blade of the jaw. O. oviparus, on the
other hand, has fifteen pairs of walking legs and an accessory
tooth, and is evidently very closely related to the New South
Wales P. Leuckartii, though by no means far removed
systematically from its congeners.

The occurrence of such closely allied species in Kastern

VOL, 45, PART 3,—NEW SERIES. DD

390 ARTHUR DENDY.

Australia and New Zealand respectively is very remarkable.
They must have had a common origin, and the question arises,
“Did the ancestral form enter Australasia from the north or
from the south?” Hither view might be maintained. It may
have spread northwards from an Antarctic continent or archi-
pelago, as some of the Australasian plants are supposed to
have done; or it may have come from the north at a time when
a more or less close connection existed between New Zealand
and North-Hastern Australia. Personally I am inclined to
favour the latter hypothesis, which seems to be supported by
the distribution of Ooperipatus and certain other animals
and plants at the present day. It is generally believed that
New Zealand has been disconnected from the Australian
continent since at least the close of the Cretaceous epoch, and
the distribution of Ooperipatus, therefore, indicates a very
remote antiquity for the egg-laying habit. In fact, we may
believe that the oviparous species have persisted with little
modification since the Cretaceous period. It is, of course, just
possible that Ooperipatus spread from New Zealand to
Australia, or vice versa, across the '’asman Sea in com-
paratively recent times, but this appears to me highly im-
probable.

Wherever the ancestral form may have originally come from,
it appears not only to have maintained itself successfully in
Australasia, but to have given rise to several new species
which have lost the primitive oviparous habit; in Australia
P. Leuckartii and P. occidentalis, and in New Zealand
P. nove-zealandiz and P. Suteri may be looked upon as
descended from a common oviparous ancestor.

Thus the distribution of the Australasian species seems
perfectly consistent with the view that the habit of laying
large and heavily yolked eggs is more primitive than that of
retaining the young in the uterus throughout the period of
their development. ‘his view is also strongly supported by
the testimony of Sedgwick and Nclater, derived from the
study of the development of viviparous species.

Sedgwick, dealing with the development of P. capensis,

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 891

lays stress upon the large size! of the egg combined with
the almost complete absence of yolk. He assumes “ that the
ovum of the Cape species has only recently lost its yolk, and
that it may be compared to an ovum of the New Zealand form
[P.nove-zealandiz] from which the yolk has been almost
completely dissolved out by some reagent. As a matter of
fact [he adds] it is impossible, with our present methods, to
effect this complete solution of yolk and leave its proto-
plasmic framework; but what we cannot effect has been done
by nature in the most complete manner, leaving an ovum
which is little more than a loose protoplasmic spongework,
excepting at one point where the protoplasm is more dense.”

Willey (1), in discussing this question, observes that “ the
view that the egg of Peripatus capensis exhibits a stage
in the process of acquiring yolk, instead of being a stage in
the loss of yolk, could be maintained with equal force.”
Surely, however, it is hardly likely that the protoplasm would
acquire a vesicular structure in anticipation of the formation
of yolk; in order to justify us in accepting this view we
should be obliged to show that the vesicular structure has
some value altogether apart from yolk formation.

Writing on the development of a South American species
(P. Imthurni) Sclater (1) also comes to the conclusion that
the alecithal condition of the egg is secondary.

In this species the ovum appears to be much smaller than
in P. capensis. The segmentation is complete, and there
1s no appearance of sponginess, such as occurs in P. capen-
sis; “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. . . . Now,
it seems to me that the loss of yolk has had precisely the
same effect in 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 (5) the formation of what
I have termed the embryonic vesicle, which appears to me

1 Sedewick gives the length of the youngest ovum found in the oviduct as

0°4 mm,

392 ARTHUR DENDY.

to be exactly analogous to the blastodermic vesicle of
mammals.”

IT am aware that the loss of yolk by the mammalian ovum
has been questioned, but it is possible to doubt anything,
and the subject need hardly be discussed in this place.

The small number of legs in all the Australasian species—
smaller in O. insignis and O. viridimaculatus than in
any other known species except the South African P. brevis
—imight be used as an argument against their primitive
nature. Bouvier, indeed, maintains (1) with some reason
that the species with more numerous legs are more primitive
than those with a smaller number, and that “in many
respects the species belonging to Oceania [from Hastern
Australia to New Zealand] mark the present limit of the
evolution of the Onychophora.” ‘In any case,” he adds,
“it appears to be quite certain that Central America and
the Caribbean region have been the centre of origin and
migration of the species of Peripatus.”

Bouvier’s knowledge of the Australasian species appears,
however, to have been limited. Thus he adheres! to the
old idea of fifteen pairs of legs being characteristic of the
“ Oceanian ” forms; and, speaking of the jaws, he states that
“here the accessory tooth disappears on the outer blade
also,’ as if this were a character of all the Australasian
species. He had, when he wrote, apparently never heard of
Peripatus Suteri, P. insignis, or P. oviparus, and in
another paper (8) he refers to P. nove-zealandiz as
having fourteen pairs of “ pattes.” Nevertheless it is not
necessary at present to dispute Bouvier’s general conclusions
as to the course of the evolution of the Onychophora, viz.
that as they spread from their original home in America,
“their limbs atrophied in succession posteriorly, and, at the
same time, their number became more and more constant.
The proximal spinulous arches followed, up to a certain
point, the same regressive course; the nephridial papille of
two pairs of limbs advanced by degrees towards the following

1 Tt must be remembered that this was written in 1900.

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 393

arch; the wrinkles in the skin became more complicated,
then interrupted on the dorsal median line; and lastly, the
dentiform armature of the jaws underwent successive re-
duction.”

In accordance with these views, it is obvious that in most
respects the genus Ooperipatus is by no means primitive,
but it by no means follows that it may not, all the same,
have preserved a primitive oviparous habit which has been
lost in other forms. It is a singular coincidence that the
only egg-laying Mammalia which now exist are confined to
the same zoological region.

In concluding this discussion, I may say that, personally,
I am strongly inclined to believe that the ancestors of at
any rate all the Australasian Onychophora were oviparous,
or, in other words, that Ooperipatus represents in this
respect the ancestral form of Peripatoides. It also seems
highly probable that the egg-laying habit was at one time
universal throughout the group Onychophora, while the
formation of an elaborate sculptured chorion may possibly
be looked upon as another indication of relationship between
the Onychophora and the Insects.

III. Description OF SPECIES AND SYNONYMY.

1, Ooperipatus oviparus, Dendy. (Figs. 1, 4—26.)
Synonymy.

1887. ? Peripatus Leuckartii (“in all probability”),
Fletcher. ‘Proc. Linn. Soc. N.S. Wales,’ vol. ii,
series 2, p. 450.

1888. ? Peripatus Leuckartii, Sedgwick. ‘Quarterly
Journal of Microscopical Science,’ vol. xxviii, p. 463.

1889. (Probably new species,) Dendy. ‘ Nature,’ vol. xxxix,
p. 366.

1889. (Probably not new species,) Sedgwick. ‘Nature,’
vol. xxxix, p. 413.

1889. Peripatus Leuckartii, Dendy. ‘ Proc. Royal Soc.
Victoria,’ vol. 11 (new series), p. 51.

304.

1890,

1390;

soe

1893.

1895.

1898.

1895.

18¢

“2

6

LOO

ARTHUR DENDY.

Peripatus Leuckartii, Dendy. ‘ Handbook of Mel-
bourne (A.A.A.8.),’ p. 95; ‘Victorian Naturalist,’
vol. vi, p. 173; ‘ Annals and Magazine of Natural
History,’ series 6, vol. vi, pp. 121—128; ‘ Proc.
Royal Soc. Victoria,’ vol. iii (new series), p. 44.

Peripatus Leuckartii, Fletcher (specimens from
Mt. Kosciusko). ‘Proc. Linn. Soc. N. 8. Wales,’
vol. v, series 2, p. 469; ‘Annals and Magazine of
Natural History,’ series 6, vol. vi, p. 352.

Peripatus Leuckartii, Dendy. ‘Proc. Royal Soe.
Victoria,’ vol. iv (new series), p. 31; ‘ Zoologischer
Anzeiger,’ xiv, p. 461; ‘ Nature,’ vol. xliv, p. 468;
‘ Victorian Naturalist,’ vol. viii, p. 67.

2. Peripatus Leuckartii, Spencer. ‘Victorian Natural-

ist,’ vol. ix, p. 30.

2. (Common Victorian species of Peripatus,) Dendy.

‘Report of the Australasian Association for the
Advancement of Science, Hobart,’ 1892, p. 375.

92. (Larger Victorian Peripatus,) Dendy. ‘ Proc. Linn.

Soc. N. S. Wales,’ vol. vii, series 2, p. 267; ‘ Proc.
Royal Soc. Victoria,’ vol. v, p. 27.

(Larger Victorian Peripatus,) Dendy. ‘ Nature,’ vol.
xlvii, p. 508, and ‘Proc. Royal Soc. Victoria,’ vol.
vi (new series), p. 118.

(Oviparous Victorian form,) Dendy. ‘ Report of the
Australasian Association for the Advancement of
Science, Brisbane,’ 1895, p. 108.

Peripatus oviparus, Dendy. ‘Zoologischer Anzeiger,’
xvill, p. 264, and ‘Proc. Linn. Soc. N. S. Wales,’
vol. x, series 2, p. 195.

(Variety of Peripatus Leuckartii,) Fletcher. ‘ Proc.
Linn. Soc. N. 8. Wales,’ vol. x, series 2, pp. 183, 194,

Peripatus oviparus, Steel. ‘Proc. Linn. Soe.
N.S. Wales,’ vol. x, series 2, pp. 98, 99, 102.

Peripatus oviparus, Steel. ‘Proc. linn: - Soci
N. 8. Wales,’ vol. xxu, p..124.

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 3995

1898. Peripatus oviparus, Willey. ‘ Anatomy and Deve-
lopment of Peripatus nove-britannie,’ pp.
30, 38.

1900. Ooperipatus oviparus, Dendy. ‘ Zoologischer

Anzeiger, xxi, p. 510.

Description.—There are fifteen pairs of claw-bearing
legs, exch (including the last pair) with three pale-coloured
spinous pads on its ventral surface. The proximal pad is
much narrower but at the same time longer than the others,
and the distal one is much the smallest. On the fourth and
fifth legs the proximal pad is transversely divided into three
parts, the middle part being very small and bearing the
aperture of the nephridium. The foot bears three large
primary papille, anterior, posterior, and dorsal, overhanging
the pair of claws.

The inner jaw-blade (fig. 7) has one large tooth and about
eight small ones; the number of the latter is probably not
constant, and some of them are very minute. ‘The outer
blade (figs. 7, 8) has one large tooth, and one small, blunt,
and feebly developed accessory tooth.

In the adult female, the ovipositor (fig, 9) is very con-
spicuous, even in a state of retraction, as an ovoid body of a
pale yellow or orange colour, lying between the legs of the
last pair. In specimens ordinarily contracted in spirit it is.
about as large as each of the legs between which it lies, but
it is capable of great extension (fig. 10). Its surface is
ornamented with minute spine-bearing papille, and the
genital aperture is a slit at the apex, placed parallel with the
long axis of the animal. I have seen no indications of crural
glands or accessory genital olands in the female.

In the male, the genital aperture lies between the legs of
the last pair, but is not specially prominent. Between it and
the anus are a pair of apertures of accessory glands lying
close to the middle line. Crural glands probably occur in all
the legs except the first pair; their apertures may or may not
be visible, according to the state of contraction. Those of

596 ARTHUR DENDY.

the fifteenth pair of legs are rendered conspicuous by a pair
of white papillee lying in the angles between the leg and body
at each side of the genital aperture; the others may also be
indicated by white papille.

The predominant colours of the skin are red and indigo-
blue, the former passing into yellow and the latter into black
in some specimens. The pattern and relative amounts of the
different colours vary greatly, but as far as my own personal
experience goes the following appears to be the fundamental
arrangement from which other types may be derived (compare
fig. 1). On the dorsal surface is a series of segmentally
arranged diamond-shaped patches, in which the red colour
predominates. Hach patch is made up of two triangles, whose
bases face one another on each side of the narrow mid-dorsal
longitudinal groove, while their apices le over the legs, and
at about one third of the distance from the mid-dorsal line to
the insertion of the legs. The separation of the diamonds
from one another is by no means complete, so that there are
really two continuous bands of red, one on each side of the
mid-dorsal line, with their outer margins deeply indented
between the legs.

The edges of the mid-dorsal groove are often darkly
pigmented, and may give rise to an apparently single median
dark line when the lips of the groove are closed together by
contraction. There is commonly, also, a dark edging to the
red diamonds on their outer margin, forming a zigzag longi-
tudinal stripe. This typical pattern may be almost, if not
quite, obliterated by the replacement of the red pigment by
dark indigo-blue, but even in very dark specimens it may still
be represented by a row of small, pale yellow or red spots,
each of which occupies the apex of one of the red triangles
in typical specimens. ‘The ventral surface is mottled in the
different colours, but is paler than the dorsal, and exhibits no
characteristic pattern. ‘There isin the middle line a row of
still paler areas, placed one between the legs of each pair but
the last. Patches of dark indigo-blue are usually present on
the under surfaces of the legs, near to their bases,

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 397

A number of variations of pattern have been described by
myself (2) in specimens from Ballarat, Victoria, and by
Fletcher (8) in specimens from Mount Kosciusko, which Steel
(2) has shown to belong to this species.

A good-sized female specimen, when crawling, measured
39 mm. in length, exclusive of the antenne. Full-grown
females preserved in alcohol and contracted in the usual
manner (not extended by drowning’) measure about 20 mm. in
length (exclusive of antenne) by 4°5 mm. in greatest breadth
(exclusive of legs). The presence in the body of the very
large eggs may make a female appear much broader than
would otherwise be the case (compare fig. 6).

The males seem to be commonly somewhat smaller than
the females, but not very much. Further information is
wanted on this head, however, and also on the proportions in
which the sexes occur. As to the latter question I may say
that there does not appear to be any great difference between
the numbers of the two. I have not kept any record myself,
but Mr. Fletcher (8), amongst thirty-five specimens from
Mount Kosciusko, found eighteen males and seventeen
females.

Discussion of Relationships.—That Ooperipatus
oviparus is in many respects very closely related to the
common viviparous New South Wales species (Peripatus
Leuckartii) there cannot be the slightest doubt. Indeed,
apart from the oviparous habit and, in correlation therewith,
the presence of an ovipositor in the female and of a chorion
outside the vitelline membrane of the egg, I know of no
characters by which it could be distinguished, for the pattern
of the skin is in both species too variable to be satisfactory,
and the two may resemble one another very closely in this
respect.

In the case of female specimens, the presence of the
conspicuous ovipositor affords an easy means of recognition,
and I am glad to be able to quote in this connection the
testimony of Mr. T. Steel, who has perhaps examined a larger
number of individual specimens of Australian Onychophora

398 ARTHUR DENDY.

than any other observer, having collected in a single summer
579 adult specimens of the viviparous New South Wales
species (P. Leuckartii), of which 390 were females.! Writing
in 1897, Mr. Steel observes (2) : “I desire to place on record
the occurrence in New South Wales of P. oviparus, Dendy,
the Victorian form of Peripatus. While collecting in
January of this year, between Exeter and Bundanoon (Moss
Vale district), on turning over a log I noticed a Peripatus,
which from its attitude and general appearance specially
attracted my attention. This proved to be a female specimen
of the above species, and, so far as I am aware, this is the
first occasion on which its occurrence in this colony has been
definitely recorded. The lozenge-shaped pattern which
characterises most of the specimens found in Victoria is well
displayed ; and the fact of the ovipositor being fully extruded
in the specimen, which I now exhibit, is sufficient guarantee
of its identity. When visiting the Australian Museum a few
days ago, I had an opportunity of examining the specimens
of Peripatus preserved there, and I was interested in
noticing that those collected by Mr. Helms in 1889 at Mount
Kosciusko belong to the same species. All of the females in
the museum collection from that locality which I examined
have the ovipositor plainly visible, and in many of them it is
fully extruded.”

The difficulty of distinguishing O. oviparus from P.
Leuckartii (except by the ovipositor) was forcibly illus-
trated in the case of the Queensland specimens collected by
Professor Baldwin Spencer at Cooran. ‘‘ After long search-
ing,” says Professor Spencer (1), “ I came across Peripatus
Leuckartii, very dark purple in colour, and evidently similar
to the typical form and without the curious diamond-shaped
markings characteristic of the Victorian form. Though
searching hard, I only found nine specimens altogether, and
all these close to’Cooran.” I myself also believed these
specimens (one of which I still have in my possession) to
belong to the viviparous species P. Leuckartii, and have
l Steeles

eee)

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 399

referred to them as such in earlier publications, being
misled partly by the very dark colour, and partly by the
fact that the chorion was incompletely developed, the
specimens being collected very early in the season (October) ;
but the ovipositor in my specimen is very conspicuous and
fully protruded to a length of 4 mm. (fig. 10), the specimen
(if I remember rightly) having been killed by drowning,
while microscopic investigation of the eggs showed the chorion
to be really present although incomplete (figs. 18, 19).

The specific distinction between the males of O. oviparus
and P. Leuckartii is a matter of much greater difficulty,
and as yet I fail to see how they can be distinguished
otherwise than by the company in which they are found. As
the females differ, so far as we know at present, only in the
structure of the reproductive organs, this is not surprising.

Localities.

Victoria.—Warragul (coll. Baker, probably this species)
Warburton (coll. Dendy) ; Ballarat (coll. Nye and Avery) ;
Macedon (coll. Hogg and Dendy) ; Mount Baw Baw (coll.
Frost); Walhalla (coll. Hoge); Pyalong (coll. Lucas, male
only).

New South Wales.—Mount Kosciusko (coll. Helms) ;
Moss Vale District (coll. Steel).

Queensland.—Cvoran (coll. Spencer).

2. Ooperipatus viridimaculatus, Dendy. (Figs. 2,
27—33.)
Synonymy.

1900. Peripatus viridimaculatus, Dendy. ‘Nature,’ vol.
lxi, p. 444; ‘Trans. and Proc. N. Z. Inst.,’ vol. xxxui,
p- 436.

1900. ? Peripatus viridimaculatus (probably), Fletcher.
‘Proc. Linn. Soc. N.S.W..,’ vol. xxv, p. 116.

1900. Ooperipatus viridimaculatus, Dendy. ‘ Zoolo-
gischer Anzeiger,’ xxi, p. 510.

4.00 ARTHUR DENDY.

Description.—There are fourteen pairs of claw-bearing
legs, each with three spinous pads on its ventral surface.
The proximal pad is much narrower than the others, and con-
tains a considerable portion of the orange pigment, while
the two others are dark indigo-blue in colour. The proxi-
mal pad also sometimes shows a tendency to break up
transversely into papilla, so that it is altogether much less
conspicuous than the other two, and in the last pair of legs
may be scarcely recognisable. ‘The median pad is the
largest. On the fourth and fifth legs the proximal pad is
trausversely divided into three parts, the middle part being
small and bearing the nephridial aperture. The foot bears
three large primary papille—anterior, posterior, and dorsal
overhanging the pair of claws.

‘The inner jaw-blade (fig. 30) has one large tooth and about
seven small ones. The outer blade (fig. 29) has no accessory
tooth.

The ovipositor of the female (figs. 27, 31) resembles that of
O. oviparus, lying between the legs of the last pair and
bearing the genital aperture as a longitudinal slit at its apex.
There are no external indications of crural or other accessory
olands in the female.

In the male the genital aperture also lies between the legs
of the last pair, and just behind it, at the base of the short
broad anal cone, les a pair of small white papille, doubtless
indicating the apertures of accessory glands. Whitish
papille, indicating crural glands, occur on all the legs of the
last nine pairs, i. e. from the sixth to the fourteenth inclusive,
As usual they lie just distal to the nephridial aperture,
except in the last pair of legs, where there is apparently no
nephridial aperture, and the papillz lie close to the genital
aperture on either side. All these crural papille are very
conspicuous.

The general coloration of the dorsal surface (fig. 2), when
only slightly magnified and examined as an opaque object,
appears to be dark grey mottled with orange, with a dark
median band, fifteen pairs of green spots arranged seg-

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 401

mentally over the appendages from the oral papillz to the
last pair of legs, and a black or nearly black triangular patch
between each two successive green spots on each side. ‘I'he
ventral surface under the same conditions appears to be
mottled grey or violet, with pale areas between the legs.

Microscopic examination of the skin, after it has been
rendered transparent in Canada balsam, shows that three
very distinct pigments take part in the production of the
pattern, viz. indigo-blue, dull brownish orange, and bright
emerald green. In the mid-dorsal line is the usual very
narrow unpigmented groove or fissure; this is bordered on
either side by a dark narrow band, partly of indigo-blue and
partly of orange brown. Outside this comes a broader band,
chiefly of orange. ‘Then comes the row of irregularly shaped
green spots alternating with triangular dark areas, the latter
consisting almost exclusively of dark indigo-blue and the
former of bright emerald green, the two colours being
separated by outward continuations of the orange bands.
Outside the row of green spots comes a zone in which dark
indigo-blue and orange-brown papille are about equally
intermingled, the orange increasing immediately below the
intervals between the legs. Ventrally only the orange and
indigo-blue occur, and both lighter in tint than on the dorsal
surface. The legs are mottled with orange-brown and indigo-
blue, the two distal spinous pads and the large primary
papille of the feet being indigo-blue. ‘The antenne are for
the most part dark indigo-blue, but about every fourth
annulus is orange. ‘The ovipositor is pale dull orange.

Of course, as in other species, the colour and pattern
may vary, but the above may be taken as typical, and I
have never noticed any specimen without the green spots.

The length of an adult female when fully extended and
crawling was 31 mm. and the breadth 3 mm., exclusive of
antenne and legs, the fully extended antenne being about
5 mm. long. The males appear to be a little smaller. The
two sexes seem to occur in about equal numbers. Out of
thirty specimens which I collected myself at Lake 'l’e Anau

4.02 ARTHUR DENDY.

nine were female, twelve male, and nine young, while three
specimens sent to me subsequently from the same locality by
Mr. Donald Ross were all female.

Discussion of Relationship, etc.—Whether or not
the specimens recorded by Fletcher (8) as probably belonging
to P. viridimaculatus are really referable to this species
must remain doubtful until we have further information about
them. ‘The only published account of them known to me is
the following, which I quote in full:—“ Mr. Fletcher exhibited
five specimens (male 2, female 3) of a Peripatus with four-
teen pairs of walking legs, the males with white papille on
the legs of the posterior nine pairs, from the North Island of
New Zealand. ‘The specimens were obtained by Mr. C. 'T.
Musson near Te Aroha in the early part of last January. They
will probably prove to be referable to the species for which
Professor Dendy (‘ Nature,’ March 8th, 1900, p. 444) has
recently proposed the name P. viridimaculatus, founded
on specimens collected at the head of Lake ‘le Anau in the
South Island. he (spirit) specimens exhibited, however, do
not in their present condition seem to show the ‘ fifteen pairs
of green spots arranged segmentally,’ which Dr. Dendy
describes as characteristically present in the specimens from
the South Island.” Whether these specimens belong to O.
viridimaculatus or not, it is indeed a most remarkable
coincidence that a fourteen-legged species should have been
discovered in the North Island of New Zealand certainly
within a few days of the time when I discovered one in the
South! The green pigment in my specimens does not appear
to be soluble in spirit.

Extremely interesting is the close relationship of O.
viridimaculatus to the Tasmanian and Victorian species
O. insignis. Indeed it would be difficult, if not impossible,
to distinguish the two species except by the characteristic
colour. markings. ‘I have seen no trace in O. insignis of
the bright emerald-green pigment which forms such a
characteristic feature in O. viridimaculatus, and the
absence of the green spots, if confirmed, may necessitate

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 405

the identification of the Te Aroha specimens with tlie
Victorian and ‘Tasmanian species.

Localities.

New Zealand.—South Island: Clinton Valley, head of
Lake Te Anau (coll. Dendy aud Ross). ? North Island:
Te Aroha (coll. Musson).

3. Ooperipatus insignis, Dendy. (Figs. 3, 34, 35.)
Synonymy.

1890. Peripatus insignis, Dendy. ‘ Victorian Naturalist,’
vol. vi, p. 1738; ‘Annals and Magazine of Natural
History,’ series 6, vol. vi, p. 121; ‘Proc. Royal Soc.
Victoria,’ vol. 1i1 (new series), p. 44.

1894. Peripatus insignis, Spencer. ‘Proc. Royal Soc.
Victoria,’ vol. vii (new series), p. 31.

1895. Peripatus insignis, Dendy. ‘Report of the Aus-
tralasian Association for the Advancement of
Science,’ Brisbane, 1895, p. 109.

1895. Peripatus lLeuckartii, var. typica, Fletcher.
“Proc.. Linu. Soc. N.S.W..,’ vol. x, sertes 9, p. 185.

1900. Ooperipatus insignis, Dendy. ‘ Zoologischer
Anzeiger,’ xxiil, p. 510.

Description.—There are fourteen pairs of claw-bearing
legs, each with three spinous pads on its ventral surface, the
proximal pad being, as usual, narrow and interrupted in the
fourth and fifth legs by the nephridial papillae. On some of
the legs, and especially on those of the last pair, the proximal
pad is ill-defined, and shows a tendency to break up trans-
versely into papilla. The colour of the pads is pale yellowish,
shading into indigo-blue, without any conspicuous difference
between the proximal pad and the other two, though the
distal one is the darkest of the three. The foot bears three
dark-coloured primary papille—anterior, posterior, and
dorsal, overhanging the pair of claws.

4.04, ARTHUR DENDY.

The inner jaw-blade (fig. 35) has one large tooth and about
five small ones. ‘he outer blade (fig. 34) has no accessory
tooth.

The ovipositor of the female resembles that of the other
species of the genus, and is conspicuous between the legs of
the last pair as a subcylindrical protuberance of a pale
yellow colour, bearing the genital aperture as a longitudinal
shit at its apex. There are no crural gland papille in the
female. In the male the genital aperture occupies the
usual position, and there are indications of a pair of white
papillee (perhaps fused together) just behind it on the short
anal cone. Crural gland papille may occur on the legs of
the last nine pairs.

The general coloration of the dorsal surface to the naked
eye appears dark, sometimes almost black, speckled with pale
orange or yellow. Closer examination of typical specimens
shows a very characteristic pattern, which may be described
” The general ground colour
is dark indigo-blue,! often almost black, and this is chequered
by more or less regularly arranged patches of pale dull
orange or yellow in transverse and longitudinal rows.

in general terms as “ chequered.

Though irregular, the patches may generally be described as
roughly rectangular in outline, forming a kind of chessboard
pattern with the dark ground colour (fig. 3). A longitudinal
row of the lighter patches runs down the middle; these are
small, and each is divided longitudinally in the middle line by
a narrow dark band, which in turn is interrupted by the very
narrow, unpigmented, mid-dorsal groove, visible under the
microscope. ‘The dark band widens out between the successive
light patches of the row, forming alternating dark patches.
Then on each side comes a row of larger lhght patches
alternating with those of the middle row. Then another row
of light patches alternating with the last, and lying in the
same transverse-rows as the median light patches, over the
legs. In addition to these patches individual light-coloured

' May be greenish indigo in preserved specimens.

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 405

papilla are scattered in the dark areas, and dark ones in
the light areas, the papille being arranged as usual on
transverse ridges of the skin. Sometimes the light-coloured
primary papille appear to be arranged to some extent in
irregular longitudinal rows, and sometimes the chessboard
pattern is almost obliterated, leaving the longitudinal rows of
light-coloured papillz scattered over a nearly uniform back-
ground. ‘The dorsal surface of the legs and feet is dark
indigo-blue, with two or three orange or yellow papille on
the legs.

The ground colour of the ventral surface is pale yellowish.
Over this are scattered a number of papillae, mostly of an
indigo-blue colour, but some dull orange, arranged in trans-
verse rows on the ridges of the skin. In the mid-ventral
line, between the legs of each pair except the last (where the
genital aperture is situated), are the usual pale areas of skin
devoid of papille.

The antennz are dark indigo-blue, sometimes ringed with
orange. The characteristic chessboard pattern of the dorsal
surface appears to be very constant, and I have seen it in
Tasmanian as well as in Victorian specimens.

I have no measurements of living specimens, and the
females in my possession do not appear to be fully grown.
After preservation in spirit in the ordinary manner, my
largest male specimen (Tasmanian) measures about 11 mm
in length by 2°5 mm. in greatest breadth, exclusive of appen-
dages. ‘The specimens collected by Professor Spencer in
Tasmania were killed by drowning, and therefore presumably
in an extended condition, and he gives the measurements of
three of the largest as respectively 23, 17,and 15 mm. in length
(exclusive of tentacles), and 4,3, and3 mm. in breadth. It is
possible, as has been suggested, that the Tasmanian specimens
may be normally larger than those of the mainland, but the
evidence is not sufficient to enable us to form a definite
conclusion on this point. The three Tasmanian specimens in
my possession are all male, and I have a male specimen from
Victoria of about the same size. Probably the females are a

VOL, 45, PART 3.—NEW SERIES, EE

4.06 ARTHUR DENDY.

little larger; I have a specimen from Victoria, apparently
killed by drowning, which is about 15 mm. long.

Discussion of Relationships.—The close relationship
which this species bears to the New Zealand O. viridi-
maculatus has already been pointed out. Itis obviously quite
distinct, as shown by the absence of the accessory tooth on
the outer jaw-blade, and by the presence of only fourteen
pairs of walking legs, both from O. oviparus and from the
common New South Wales Peripatus, which I still term
P. Leuckartii. Since, however, it has been suggested with
some show of reason that my O. insignis may be really
Singer’s original P. Leuckartii, andas Mr. Fletcher (7) has
definitely adopted this view of the case, it becomes necessary
to briefly discuss the question of nomenclature.

(1) As to the number of appendages in the original
P. Leuckartii we have no absolutely certain information,
for it appears that Leuckart and Sanger gave conflicting
accounts. Leuckart (1) first said there were ‘“ 16 Beinpaaren.”
Mr. Fletcher observes, “‘ But in regard to the Australian
Peripatus, it seems evident that Professor Leuckart inten-
tionally included the oral papillae among the sixteen pairs,
but without indicating the fact.”

Sanger, working apparently on the same specimen, states,
on the other hand, that there are “fifteen pairs of legs, one pair
without claws, fourteen with. This character is also found in
P. brevis, described by Blanchard.’’! Mr. Fletcher interprets
this to mean that Sanger includes the oral papille in the
fifteen pairs, and that the original P. Leuckartii had only
fourteen pairs of claw-bearing walking legs. I do not think
that this interpretation is necessary. Mr. Fletcher himself (7)
has recorded the occurrence of a specimen of the common
New South Wales Peripatus in which two pairs of legs had
the claws missing, and it was quite possible that Leuckart’s
specimen was also abnormal in this respect (especially if it
had been subjected to much handling before coming into

1] quote from a translation of Singer’s paper (1) with which Prof,
Spencer kindly furnished me.

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 407

Singer’s possession), the fact perhaps not being considered
worth noticing in the summary which Sanger gives, and which
(quoted by Leuckart) has been generally accepted as bearing
the ordinary interpretation. This summary (according to
the translation obtained in London by Professor Spencer)
runs as follows:—“ Fifteen pairs of legs; sexual organs
between the last pair; the ‘ sole’ consists of three segments,
one long and curved, and two short and straight. New
Holland, Australia.”” In any case Singer’s statement as to
the number of legs appears to be completely neutralised by
Leuckart’s. As to the argument derived from Singer’s
comparison of his specimen with P. brevis, it seems probable
that Sanger knew very little about P. brevis, for, as Mr.
Fletcher shows, he erroneously attributes the description
thereof to Blanchard instead of to De Blainville. Moreover
against this comparison we may surely set Leuckart’s
comparison of P. nove-zealandiz, Hutton. Leuckart
says (as quoted by Fletcher, Hutton’s Abhandlung ‘On
Peripatus nove-zealandiz,’ ‘Ann. Mag. Nat. Hist.’ [4],
xvill, Nov., 1876, pp. 361—369, pl. xvii) “macht uns mit
einer Form bekannt, die 15 Beinpaare besitzt, wie der von
Sanger (J. B., 1870,S. 410) beschriebene P. Leuckartii, der
unserm Verf. freilich unbekannt geblieben ist, obwohl seine
neue Art vielleicht damit zusammenfillt. Jedenfalls ist nicht
der P. nove-zealandiz, sondern der P. Leuckartii die
erste Art des Gen. Peripatus, die aus Australien kommt.”
Surely if Leuckart, the owner of the specimen, thought it
was so similar to P. nove-zealandiz with fifteen pairs of
legs that it might be identical, it is hardly likely that it after
all had only fourteen pairs. There can hardly have been any
misunderstanding of Hutton’s original description, for Hutton
says “fifteen pairs of ambulatory legs, and a pair of oral
papillee.”

(2) The locality of Sanger’s specimen, ‘found in New
Holland, north-west from Sydney,” renders it extremely
improbable that it can have been O. insignis, which is rare
even in the south, and has never yet been recorded from

408 ARTHUR DENDY.

New South Wales, where the common viviparous species has
been obtained in hundreds.

(3) If Mr. Fletcher’s conclusion that Sanger’s specimen
was a female be correct, then it can hardly have been
O. insignis, for no mention is made of the very characteristic
ovipositor of that species. If, on the other hand, it was
a male, it is hardly likely that Singer would altogether have
overlooked the papille of the crural glands.’ Therefore it
was probably a female of the common New South Wales
viviparous species, with fifteen pairs of walking legs.

In short, it appears to me only reasonable to adhere to the
old view that the common viviparous New South Wales
species is P. Leuckartii, and to retain my specific name
insignis for the fourteen-legged oviparous Victorian and
Tasmanian species, at any rate until such time as a re-
examination of Siinger’s type may prove this view to be

erroneous.

Localities.

Victoria.—Macedon (coll. Hogg) ; Sassafras Gully ; Fern-
tree Gully ; Gembrook.

Tasmania.—Dee Bridge (coll. Spencer); Mount Welling-
ton (coll. Morton).

TV. Summary or ReEsvtts.

The principal conclusions arrived at in this memoir may

be briefly summarised as follows :
1. The genus Ooperipatus includes a number of ovi-
parous Onychophora characteristic of Hastern Australia,

1 It may be argued that if my view be correct, Sanger has made a much
more serious omission in not mentioning the oral papillae. He may well have
considered, however, that the oral papillae, exhibiting no specific characters,
did not require specific notice. It is hardly likely that he would count the
oral papille as legs, for he distinctly says that the legs have ‘‘soles”
(meaning spinous pads), and these, of course, as well as the claws, are wanting
in the oral papillae.

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 409

Tasmania, and New Zealand; distinguished by laying large,
heavily yolked eggs with a thick sculptured chorion, and by
the presence in the female of a conspicuous muscular ovi-
positor.

2. The egg at the time of laying contains no recognisably
developed embryo, and development takes place afterwards
with extreme slowness.

3. The oviparous habit is very ancient, dating back at
least to the Cretaceous epoch, as indicated by the geographi-
cal distribution of the species. The conclusions of Sedgwick
and Sclater as to the loss of yolk in the eggs of certain vivi-
parous species are thereby supported.

4, Three species of Ooperipatus are at present known,
viz. O. oviparus, O. viridimaculatus, and O. insignis.
Some slight doubt attaches to the last, because the eggs have
not yet been observed, but the females have the conspicuous
ovipositor.

do. The genus Ooperipatus is very closely related to
Pocock’s Peripatoides, and may be regarded as represent-
ing an ancestral form from which the viviparous Australasian
species are descended.

6. Except as regards the egg-laying habit and structures
associated therewith, the genus Ooperipatus is, according
to the views of Bouvier, very far from primitive in its
characters, the number of walking legs being reduced to
fifteen or fourteen, the spinous pads being only three in
number, and the transverse ridges of the integument being
interrupted in the mid-dorsal line by a narrow unpigmented
groove.

7. There is no sufficient reason for supposing that Ooperi-
patus insignis, Dendy, is identical with Peripatus
Leuckartii, Sanger, which last name must be retained for
the common viviparous species of New South Wales.

Curistcuurcu, N.Z.; December, 1900.

4.10 ARTHUR DENDY.

V. LIvERATURE.

Bouvier, M. EH. L.
1. “On the Geographical Distribution and the Evolution of Peripatus”
(Preliminary note, translated). ‘Ann. and Mag. Nat. Hist.,’ series
7, Vol. li, p. on.

2. ‘‘ Quelques Observations sur les Onychophores (Peripatus) de la Col-
lection du Musée Britannique.” ‘ Quarterly Journal of Microscopical
Science,’ April, 1900, p. 367.

3. “Sur les Caractéres externes des Péripates.” ‘Proc. International
Congress of Zoology, Cambridge,’ 1898.

Denby, A.

1. “ Peripatus in Victoria.” ‘Victorian Naturalist,’ January, 1889, aud
‘Nature,’ February 14th, 1899.

2. “ Observations on the Australian Species of Peripatus.”’ ‘ Proc. Royal
Soc. Victoria’ for 1889, p. 50.

8. “Zoology: Invertebrata.” ‘Handbook of Victoria,’ Australasian
Association for the Advancement of Science, Melbourne, 1890.

4. “On the Presence of Corpuscles in the Liquid discharged from the
Apertures of the Nephridia and Oral Papille of Peripatus.” ‘ Proc.
Royal Soc. Victoria’ for 1890, p. 44.

5. “ Preliminary Account of a New Australian Peripatus.” ‘ Victorian
Naturalist,’ vol. vi (1890), p. 173, and ‘ Aun. and Mag. Nat. Hist.,’
series 6, vol. vi, p. 121.

6, ‘On the Oviparity of Peripatus Leuckartii.” ‘Proc. Royal Soe.
Victoria ’ for 1891, p. 31.

7. ‘The Reproduction of Peripatus Leuckartii,” Sanger. ‘ Zoologischer
Anzeiger,’ No, 380, 1891.

8. “ An Oviparous Species of Peripatus.” ‘ Nature,’ September 17th, 1891.

9. ‘Mode of Reproduction of Peripatus Leuckartii.” ‘ Victorian
Naturalist,’ September, 1891, p. 67.

10. “Further Observations on the Kegs of Peripatus.” ‘Proc. Austral-
asian Association for the Advancement of Science,’ Hobart, 1892.

11. “Further Notes on the Oviparity of the Larger Victorian Peripatus,
generally known as P. Leuckartii. ‘Proc. Linn, Soc, N.S.W..,’
series 2, vol. vii, p. 267, and ‘ Proc. Royal Soc. Victoria’ for 1892,
p. 27.

12. “The Hatching of a Peripatus Egg.” ‘ Proc. Royal Soc. Victoria
for 1898, p. 118, and ‘ Nature,’ vol. xlvii, p. 508.

13.

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 411

** Note on a New Variety of Peripatus nove-zealandiz,” Hutton.
‘Trans. aud Proc. New Zealand Institute’ for 1894, p. 190.

14. “The Cryptozoic Fauna of Australasia.” Presidential Address,
Section D, Australasian Association for the Advancement of Science,
Brisbane, 1895.

15. “Preliminary Notes on the Reproductive Organs of Peripatus
oviparus.” ‘ Zoologischer Anzeiger,” 1895, p. 264.

16. “ Description of Peripatus oviparus. ‘ Proc. Linn. Soc. N.S.W.’
for 1895, p. 195.

17. “A New Peripatus from New Zealand.” ‘ Nature,’ March 8th, 1900,
p. 444.

18. ‘‘ Note on Peripatus viridimaculatus.” ‘Trans. and Proc. N.Z.
Inst.,’ vol. xxxil, p. 436.

19. ‘Preliminary Note on a proposed New Genus of Onychophora.’

‘Zoologischer Anzeiger,’ 1900, p. 509.

Fuercuer, J. J.

1.

2.

8.

GAFFE
1

“Note on Specimen of Peripatus.” ‘Proc. Linn. Soc. N.S.W,’
for 1887, p. 450.

* Note on Peripatus.” ‘Proc. Linn. Soc. N.S.W.’ for 1888,
p. 892.

. “ Additional Notes on Peripatus Leuckartii.”’ ‘Proc. Linn. Soe.

N.S.W.’ for 1890, p. 469.

. “Note on Peripatus Leuckartil.” ‘Proc. Linn. Soc. N.S.W- for

1891, p. 577.

. “Note on Peripatus Leuckartii.” ‘Proc. Linn. Soc. N.S.W.’ for

1892, p. 40.

. “A Viviparous Australian Peripatus (P. Leuckartii, Saeng.).”

‘Proc. Linn. Soc. N.S.W.’ for 1892, p. 179.

. “On the Specific Identity of the Australian Peripatus, usually supposed

to be P. Leuckarti, Saenger.” ‘Proc. Linn. Soc. N.S.W.’ for
1895, p. 172.

* Note on a New Zealand Peripatus.” ‘Proc. Linn. Soc. N.S.W.’
for 1900, p. 116.

RON, EK.
“ Beitrage zur Anatomie und Histologie von Peripatus.” ‘ Zoologische
Beitrage,’ Breslau, 1883.

Hurton, F. W.

ifs

“On Peripatus nove-zealandia.’ ‘Annals and Magazine of
Natural History,’ November, 1876.

412 ARTHUR DENDY.

Levckant, R.
1. “ Note onan Australian Peripatus.” ‘Archiv fir Naturgesch.,’ Jalirg.
xxvii (1862), Bd. ii, p. 235 (quoted from Fletcher, 7).

2. “Notice of Sanger’s Paper on P. Leuckartii.” ‘Archiv fir Natur-
gesch.,’ Jahrg. xxxvil, Bd. ii, p, 406.

3. ‘Notice of Hutton’s Paper on Peripatus nove-zealandie.”
‘Archiv fiir Naturgesch.,’ Jalrg. xlii (1887), Bd. 1, p. 509 (quoted
from Fletcher, 7).

Pocock, ave l.
1. ‘Contributions to our Knowledge of the Arthropod Fauna of the West
Indies,” Part iii. ‘Journal of the Linnean Society of London,’
* Zoology,” vol, xxiv, p. 473.

SANGER.
1. “ Description of a Peripatus from Australia.” ‘Transactions of the
Russian Assembly of Naturalists held at Moscow in 1867,’ Moscow,
1869.

ScuatErR, W. L.
1. “On the Early Stages of the Development of a South American Species
of Peripatus.” ‘Quart. Journ. Micro. Sci.,’ vol. xxvill, 1888,

p. 343.

SEDGWICK, A.
1. “The Development of the Cape Species of Peripatus.” ‘Quarterly
Journal of Microscopical Science,’ vols. xxv—xxviil.

“ A Monograph on the Species and Distribution of the Genus Peripatus
(Guilding).” ‘Quarterly Journal of Microscopical Science,’
vol. xxvili (1888), p. 431.
3. “ Peripatus in Australia.” ‘ Nature,’ February 281h, 1889.
4. “Peripatus.” ‘The Cambridge Natural History,’ vol. v, 1895.
SHELDON, LiLian.
1. On the Development of Peripatus nove-zealandizx,” Parts |
and 2. ‘Quart. Journ. Micro. Sci.,’ vol. xxvii (1888) and vol. xxix
(1889).
2. “The Maturation of the Ovum in the Cape and New Zealand Species of
Peripatus.” ‘Quart. Journ. Micro. Sci.,’ vol. xxx (1890).
Seencer, W. B. .
1. “A Trip to Queensland inSearch of Ceratodus.” ‘ Victorian Naturalist,’
vol, ix, p. 16 (1892).
2. “ Note on the Presence of Peripatus insignis in Tasmania.” ‘ Proc.
Royal Soc. Victoria’ for 1894, p. 31.

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 413

STLEL, I’.
1. “Observations on Peripatus.” ‘Proc. Linn. Soc. N.S.W.’ for 1896,
p. 94.
2. “ Note on Peripatus.” ‘Proc. Linn. Soc. N.S.W.’ for 1897, p. 124.
Wiutey, A.
1. ‘The Anatomy and Development of Peripatus nove-britannia.’

VI. EXPLANATION OF PLATES 19—22,

Mlustrating Dr. Dendy’s memoir ‘ On the Oviparous Species
of Onychophora.”

Explanation of Lettering.

4. Anus. 8. 7r. Buccal tracheal pit. ©. G. Cerebral ganglion. Ch.
Chorion. Cr. P. Crural papilla (bearing aperture of crural gland), D. F.
Mid-dorsal longitudinal groove or fissure. #. Egg in oviduct. Zzé. Intestine.
Intg. Iutegument. Z. Od. Left Oviduct. 4. B. Nucleoid bodies in clear
areas of immature chorion. WV. @. Nerve-cord. Ne. 4. Nephridial aperture.
Ni. Nucleolus. Nu. Nucleus. Od. Oviduct. Oes. Gisophagus. O. Ov.
Ovarian ova. Ovp. Ovipositor. Ovy. Ovary. Ped. Peduncle of ovarian
follicle. Pd. Pharynx. Pr. O. Protuberances of oviduct opposite to
receptaculum seminis. 22. Rectum. 2&. Od. Right oviduct. &. S. Recep-
culum seminis. S. D. Salivary duct. S. G. Salivary gland. S¢. D. Slime-
duct and reservoir. S/. G. Slime-gland. Zr. P. Tracheal pit. 77 &.
Triangular sac at base of ovipositor, formed by the union of the two oviducts,
V. M. Vitelline membrane. Y. Yolk.

PLATE 19.

Fie. 1.—Ooperipatus oviparus, Dendy. A Victorian specimen showing
the typical pattern. From a drawing by C. C. Brittlebank, Esq. x 6
diameters.

Fig. 2.—Ooperipatus viridimaculatus, Dendy. Specimen from Lake
Te Anau, showing the typical pattern. From a drawing by Miss M. M.
Ferrar. xX 43 diameters.

Fic. 3.—Ooperipatus insignis, Dendy. A Victorian specimen showin
the typical pattern. From a drawing by C. C. Brittlebank, Esq. x 9
diameters.

414, ARTHUR DENDY.

PLATE 20.
Figs. 4—11. Ooperipatus oviparus.

Fic. 4.—General dissection from the dorsal surface, slightly diagrammatic.
Fie. 5.—Transverse section of male specimen through the region of the
. . . . 7
pharynx, showing the buccal tracheal pits, union of the salivary ducts, dorsal
longitudinal furrow. Drawn under Zeiss A, ocular 2, camera outline.

Fie. 6.—'Transverse section of female specimen, showing two eggs in the
oviducts. Drawn under Zeiss A (bottom lens removed), ocular 2, camera
outline.

Fie. 7.—Outer and inner jaw-blades. Drawn under Zeiss C, ocular 2,
camera outline.

Fic. 8.—Another jaw-blade. Drawn under Zeiss C, ocular 2, camera
outline.

Fie. 9.—Lower parts of oviducts, with ovipositor.
Fic. 10.—Posterior portion of specimen from Cooran, Queensland, dis-

sected from ventral surface, showing lower parts of oviducts, ovipositor
protruded. x 10.

Fic. 1].—Upper portion of oviduct, showing receptaculum seminis and
irregular protuberances. Drawn under Zeiss A, ocular 2, camera outline.

PLATE 21.
Figs. 12—26. Ooperipatus oviparus.

Ftc. 12.—Ovary, showing separation into right and left halves.

Fic. 13.—Portion of ovary, showing ovarian ova in different stages of
development. Drawn under Zeiss A, ocular 1, camera outline.

Fic. 14.—Young ovarian ova, from section stained with borax carmine.
Drawn under Zeiss D, ocular 1, camera outline.

Fic. 15.—Older ovarian ovum, from section stained with borax carmine,
showing vitelline membrane and formation of yolk granules. Drawn under
Zeiss D, ocular 1, camera outline.

Fie. 16.—Yolk corpuscles and granules from still older ovarian ovum from
section stained with borax carmine. Drawn under Zeiss D, ocular 1,
camera outline.

Fig. 17.—Yolk corpuscles from thick-shelled egg from oviduct; stained
with eosin and mounted in oil of cloves. Drawn under Zeiss D, ocular 1,
camera outline.

Fic. tS. Part of immature chorion from egg of Queensland specimen.
Stained with eosin. Drawn under Zeiss D, ocular 2, camera outline.

ON THE OVIPAROUS SPECIES OF ONYCHOPHORA. 415

Fic. 19.—Optical section of the same specimen. Drawn under the same
conditions.

Fie. 20.—Part of chorion of egg from oviduct, from a specimen kept long
in spirit. Drawn under Zeiss A, ocular 2, camera outline.

Fic. 21.—Part of surface of same, showing the regularly arranged protuber-
ances containing granules. Drawn under Zeiss D, ocular 2, camera outline.

Fic. 22.—Perspective view of same, drawn under same conditions.

Fic. 23.—Surface of chorion of deposited egg, showing sculptured pattern.
Drawn under Zeiss D, ocular 2, camera outline.

Fic. 24.—Surface of chorion of another deposited egg, showing variation
n sculptured pattern. Drawn under Zeiss D, ocular 2, camera outline.

Fie. 25.—Optical section of chorion of deposited egg with high protuber-
ances (same egg as Fig. 24). Drawn under Zeiss D, ocular 2, camera outline.

Fic. 26.—Optical section of chorion of deposited egg with low protuber-
ances, showing very distinctly the two layers of which the chorion is composed.
Drawn under Zeiss D, ocular 2, camera outline.

PLATE 22.
Figs. 27—33. Ooperipatus viridimaculatus.
Fic. 27.—General dissection from the ventral surface. x 5.

Fic. 28.—Fifth and sixth legs of male specimen, seen from below, showing
apertures of nephridia and crural gland, spinous pads, ete.

Fic. 29.—Outer jaw-blade. Drawn under Zeiss C, ocular 2, camera
outline.

Fie. 30.—Inner jaw-blade. Drawn under Zeiss C, ocular 2, camera
outline.

lc. 31.—Reproductive organs of female exposed from the ventral surface,
but not unravelled.

Fic, 32.—Surface view of portion of chorion of deposited egg, showing
sculptured pattern and fissures due to splitting. Drawn under Zeiss C,
ocular 3, camera outline.

Fic. 33.—Optical section of same, drawn under same conditions.

Figs. 34, 35. Ooperipatus insignis.
Fig. 34.—Victorian specimen, outer jaw-blade. Drawn under Zeiss C,
ocular 2, camera outline.

Fig. 35.—Inner jaw-blade of same. Drawn under same conditious.

>
7
b

i # it até sis sat let vile tail at wt

ae 7 +r <<

ae) vr iesata it

of

“ ide ide ate hi ‘
‘) evita fF nr hi UTI ES

« I so) a iw poo 2 * =:

3 rh Awe. oo a nee

Livi) ta ir: vat i Set
La aie. b pene

valle anoint adh tae

A ieteie Wa eribnce } Siew"

me ORE GFA Tae. ||

had Tea Th erie ds Tig?

hie Dept use

7 ; ; \j Tear
, a Mest! is

i
j a ‘
j
‘ i gp
t
‘
i
‘
‘, ; $ he oy ' .
wai Mla slope el
rile

cine awh) oid le euee we yeh Be

SUD RADU 141A & ete 7
Wopuoy syyr7 i le eee rg hi.ta Loy)

Ld i V

2
STULMSU STL Lad a6 7
y (iL S$ ign di.tadoy) SaALDLA0 suzn hi.tado

av]

—

eof. VWI FA VW VAAD

( i t vi LAS | A) L®©

Pole Oi a MU
}

4

\\

i #*

~~ =

Iuawrt. Foun Muor Seu, Vet 45 NSP, 20

PrO-¥ pom,

Arthur Dendy del

OOPERIPATUS OVIPARUS .— ANATOMY.

=
i
-

ee ee ae ee x —— = ¥ ~~ a

Arxths

* Dendy dal

OOPERIPATUS OVIPARUS.— OVARY AND EGGS.

anticny
NUH

“$y

Arthur Dendy del

OOPERIPATUS

VIRIDIMACULATUS (Fig?27-33) AND O.INSIGNIS (Figs 34,35)-ANATOMY.

AS Huth, Lith’ London.

ae

A NEW AND ANNECTANT TYPE OF CHILOPOD. 417

A New and Annectant Type of Chilopod.
By

R. I. Pocock.

With Plate 23.

INTRODUCTION.

THE continent of Australia and its adjacent islands have
hitherto proved singularly disappointing in the production
of interesting types of Chilopoda. Setting aside Cerma-
tobius, which was discovered some fifteen years back in the
island of Adenara, none of the genera known up to the
present time from the Australian region throw special light
upon the phylogeny of the families of this class; none can
be regarded as archaic forms that have found refuge in this
southern tract of land. Collections made in various districts
of the country have yielded specimens of Scutigera,
Henicops, Lithobius, Scolopendra, Cryptops, Geo-
philus—genera which may be met with on a single hillside
in almost any part of Southern Europe; of Cormocephalus
and Ethmostigmus, which in an almost equal degree are
characteristic of tropical Africa or the Oriental region. The
entire fauna, in short, bears the stamp of comparatively recent
origin by immigration from South Eastern Asia, with perhaps
an infusion from the Ethiopian region or Madagascar.

In 1892, however, I received from Mr. G. M. Thomson, of
Dunedin, New Zealand, a small consignment of Myriapods
from the summit of Mount Rumney, Hobart, ‘Tasmania, which
completely falsified the opinion that the Antipodes hold nothing

418 R. I. POCOCK.

peculiar or primitive in the way of centipede-life and are
wholly given over to the occupation of widely distributed
and well-known forms. The collection in question contained
a couple of centipedes representing a species which proves to
be comparable in interest from a morphological standpoint to
either of its compatriots, Ceratodus or Ornithorhynchus,
inasmuch as it unmistakably represents an archaic type,
which has survived in this isolated corner of the world—a
type which possesses the two-fold interest of exhibiting
certain unique structural peculiarities of its own, coupled
with others that serve to link together three of the best-
known and most diversified sections of the class; and also
of showing the true, but previously unknown and unsus-
pected, nature of the connection between the metamerism of
the Scolopendromorpha and that of the Lithobiomorpha.

This new centipede is the subject-matter of the following
essay. The account of it may be conveniently divided into
four headings :—(1) A description of the external structural
features of the animal; (2) a comparison between it and the
other existing orders; (3) its significance in testifying to the
transformation of the Scolopendroid into the Lithobioid
type; (4) the classification of the Chilopoda.

Part I.—Drscriprtion OF THE GENUS AND SPECIES.

Fam. Craterostigmide.

Gen.—Craterostigmus, noy.

The cephalite or head shield (Pl. 28, figs. 1, 3, 8) is
longer than wide, leaving the toxicognaths largely uncovered
on the dorsal side, as in some genera of Geophilomorpha (i.e.
Mecistocephalus) ; its frontal sulcus is distinct, and its
preantennal area sharply recurved to form a long subfrontal
area, as in other pleurostigmous Chilopods. The eyes are
represented by a single pair of ocelli, as in the Lithobio-
morphous genera Henicops and Cermatobius. The an-
tenn are basally approximated, and project forwards from

A NEW AND ANNECTANT TYPE OF CHILOPOD. 419

the fore margin of the cephalite, and consist of about the
same number of segments as is found in the primitive
Scolopendromorpha. The labral sclerites are distinct,
and the labral border is deeply and subquadrately excised in
the middle, the sides of the excision being membranous and
furnished with bristles, while its anterior border is armed
with five teeth (Pl. 23, fig. 9).

Gnathites.—The first pair, or mandibles, are furnished
beneath (posteriorly) with a thick fringe of hair, and above
(anteriorly) with a large membranous lobe beset with short
hairs ; between this lobe and the hairy fringe appear nine
pectiniform, horny teeth, arranged in slightly overlapping
series of three each (PI. 23, figs. LO—12). The second pair or
maxille much resemble those of, e.g. Scolopendra, the
basal segment (coxa or stipes) having its proximal margin
widely rounded, the first segment of the ectocorm (= first
malar segment of the external maxilla of Latzel) short, and
the second segment (= second malar segment) a little longer
than broad and evenly convex along the margin; the entocorm
(= internal maxilla) consists of a single short segment. The
basal segments of the right and left sides meet in a much
longer sutural union than is seen in Scolopendra (Pl. 23,
fig. 13).

The palpognaths (= labial palpi, second pair of maxille,
or first pair of maxillipedes) are also much like those of Scolo-
pendra, but the ectocoxite (= outer segment of the stipes of
Latzel) is more strongly produced posteriorly. As in the
genus just mentioned, the distal segment is swollen and
thickly fringed on its upper (anterior) margin, but the claw
is almost concealed in hairs, both from its anterior and
posterior aspects (Pl. 23, figs. 14, 14a).

The toxicognaths (second pair of maxillipedes, poison-
jaws or prehensors) present a combination of characters re-
calling those of Scolopendra in the presence of dentate
preecoxal and preeaxial femoral processes; those of Mecisto-
cephalus and some other species of Geophilomorpha in
the large-size and subquadrate shape of the coxal plate, the

4.20 Ro 1. POCOCK,

length of the femora, and the extent to which the terminal
segment or fang overlaps the cephalite anteriorly; those
of the Lithobiomorpha in the completeness of the penulti-
mate and antepenultimate segments (patella and tibia), which
postaxially intervene between the distal extremity of the
femur and the proximal extremity of the fang (tarsus)
(Pl. 28, figs. 1—3).

The pleural sclerites of the toxicognaths extend as far
forwards as the anterior border of the tergum of the
toxicognathic somite, but fall far short of the proximal
end of the femora (Pl. 23, figs. 1, 3, pl.).

The tergum of the toxicognaths (= basal plate) resembles
that of some Geophilomorpha (e.g. Mecistocephalus) in
size ; it is as wide as the cephalite, which it underlies in front,
but much narrower than the tergum of the first pediferous
somite, which it overlaps behind (PI. 23, figs. 1, 3, bp.).

The pediferous somites differ from those of all
Chilopoda hitherto known, in that the tergal plates are
numerically in excess of the sternal, the latter being
fifteen in number, as in the Lithobiomorpha, and the former
twenty-one, as in most Scolopendromorpha.

The terga are unequal in size, but not alternately larger
and smaller; the first is very large, and considerably over-
laps the second, the fourth overlaps the third in front and
the fifth behind, the seventh overlaps the sixth in front and
the eighth behind, the ninth meets the tenth without distinct
overlap, the twelfth overlaps the eleventh in front and the
thirteenth behind, the fifteenth overlaps the fourteenth in
front and the sixteenth behind, the remainder normally over-
lapping the anterior border of those that succeed them,
the overlapped area being smooth and defined by a transverse
groove ; the tergum of the last somite is completely coalesced
with the chitinised pleural area.

The pleural surfaces of all the pediferous somites, except
the thirteenth, fourteenth, and fifteenth, are furnished with a
distinct preecoxal sclerite (= epimeron of Latzel) ; the stigma-
tiferous somites are furnished in addition with a usually smaller,

A NEW AND ANNECTANT TYPE OF CHILOPOD. 421

stigmatic sclerite in front and a larger metastigmatic sclerite
behind, both beneath the edge of the tergum; a sclerite
representing the stigmatic in position is also found below the
edge of the third, sixth, ninth, eleventh, fourteenth, and
seventeenth terga ; and a sclerite corresponding to the meta-
stigmatic is present below the first, second, fifth, eighth,
tenth, thirteenth, and sixteenth terga; but below the third,
sixth, ninth, eleventh, fourteenth, and seventeenth terga it
is less developed, being either small and narrow, or repre-
sented by a weakly chitinised integumental ridge. ‘Towards
the posterior end of the body the preecoxal sclerite increases
in size, and encroaches upon the pleural membrane, until on
the last three somites the pleural area is covered with a
continuous chitinous plate, which is fused with the sternum,
and, in the case of the last somite, with the tergum also, as
already stated (Pl. 25, figs. 3—7).

The sterna are subequal and laterally emarginate for the
articulation of the legs; a pair of transversely elongated
preesternal sclerites extend in front of their anterior border,
as in the Scolopendromorpha ; on the last two somites these
sclerites fuse with the sterna (Pl. 23, figs. 5, 4, 6).

The stigmata, subcircular or suboval in shape, are six
in number on each side, and he above the third, fifth, eighth,
tenth, twelfth, and fourteenth pairs of legs, as in the genus
Lithobius, and beneath the fourth, seventh, twelfth,
fifteenth, eighteenth, and twentieth terga. With the ex-
ception of those on the penultimate somite, which lie far
back, the stigmata are situated more forwards than in
other Chilopoda, being slightly in front of the articulation
of the legs, and of the middle of the lateral margin of the
terga (Pl. 23, figs. 3, 4, 15).

The legs are fifteen in number on each side, as in the
Lithobiomorpha ; with the exception of the last pair they
consist of what is doubtless the primitive number of segments
in the Chilopoda, namely six, the distal segment being
undivided, as in the Geophilomorpha and the less specialised
genera of Scolopendromorpha and Lithobiomorpha; the

von. 45, PART 3.—NEW SERIES, nn)

422 R. I. POCOCK.

basal segments are relatively larger than in the Scolopendro-
morpha, but smaller than in the Lithobiomorpha, and they
show no progressive increase in size towards the posterior
end of the body, such as is characteristic of the members of
the latter order (Pl. 28, figs. 16, 17). The legs of the
fifteenth pair differ in certain particulars from those of all
Chilopods. ‘lhe basal segments or coxe are of inedium size,
freely articulated to the posterior end of the last somite,
without encroaching in any way upon its pleural area, and
like the trochanters of the two preceding pairs are armed
below with a long spike. The second segments or tro-
chanters, on the other hand, are reduced in size and com-
pletely fused with the proximal end of the femora, as in the
Scolopendromorpha. In the members of the latter order,
however, as well as in the Geophilomorpha, the coxal segment
is indistinguishably united to the enlarged precoxal pleural
sclerite. Again in the Lithobiomorpha and Scutigeromorpha,
which resemble Craterostigmus in the freedom of the coxa,
this segment occupies the whole of the pleural area of the
somite, and the trochanter remains distinct. As in all
Chilopoda, with the exception of some Geophilomorpha,
the tarsus of the fifteenth leg is bisegmented (Pl. 23, figs.
18; 19).

Projecting between the legs of the last pair is a sclerite,
formed by two valves united in the median dorsal line and
meeting ventrally, when closed, in a long slit representing
the combined genito-anal aperture. ‘his bivalvular sclerite
is probably the homologue of the dorsal plate of the anal
somite, which in some genera of Chilopoda, e. g. Scolopendra
and the male of Lithobius, is of larger size than that of the
genital somite, and partially or wholly supersedes it. No
trace of skeletal plates of the genital somite or of gonopods
is visible externally. The latter in all probability have
atrophied from enclosure within the genito-anal cavity,
embraced by the above-described sclerite (Pl. 28, figs. 5—
7519s).

A NEW AND ANNECTANT TYPE OF CHILOPOD. 423

Craterostigmus tasmanianus, sp. n.

Measurements in mm.—Total length from anterior
border of cephalite to posterior border of twenty-first tergite
57, width across coxal plate of toxicognaths 3°8; length of
cephalite 3:4, width 2:2; length of antenna 9°5, of posterior
leg 12; length of last tergite 2°4, width 2.

Colour yellowish brown, with the cephalite and toxico-
enaths darker reddish brown. Integument sparsely hairy
and punctured.

Cephalite parallel-sided, its posterior border convexly
rounded ; frontal area with its sides converging between the
eyes and the base of the antenne.

Eyes some distance behind the antenne; frontal sulcus
projecting posteriorly between the eyes, with strongly convex
backward curvature.

Antenne about one fourth the length of the body and
head, rather less than three times as long as the head;
the segments hirsute, especially towards the distal end of
the appendage, subcylindrical, longer than wide, the first
seoment strongly constricted in its basal half.

Preecoxal processes of toxicognaths long, armed apically
and externally with seven teeth. Inner side of femur and
femoral process armed with about five teeth, the proximal
of which lies far back just in front of the suture marking
the line of union of the trochanter and the femur.

Basal plate with posterior angles rounded.

Terga without longitudinal grooves, with posterior border
straight, posterior angles rounded and unthickened margins ;
the lust tergum, and to a lesser degree the last but one,
granular ; the last with a weak longitudinal lateral crest in
its anterior half, its posterior border sinuous, its middle third
slightly and convexly produced.

Sterna without grooves, those of the posterior somites
granular, the granules, as in the case of those on the terga,
formed by the elevations round the setiferous pores ; sterna

42.4, R. I. POCOCK.

area of last marked anteriorly on each side with an oblique,
shallow groove.

Legs shortish, hairy, armed with a single inferior tibial
and tarsal spine. Claw with two basal spinules, one inferior
and one posterior. Posterior legs long and slender, about one
third the length of the body and head, longer than the
antenne, without spines, protarsal and tarsal segments sub-
equal, segments subcylindrical, tarsal and protarsal subequal ;
femur, patella and tibia progressively decreasing in length.
Trochanter of the thirteenth and fourteenth armed below
with a horny spike, which is shortest on that of the thirteenth
coxa of fifteenth similarly armed.

The genito-anal sclerite is about half the length of the last
leg-bearing somite. When viewed from the dorsal or ventral
aspect its sides are seen to be convex, and to converge pos-
teriorly to a point. From the lateral aspect its upper edge,
which is compressed, is straight and horizontal ; its inferior
edge convex, the two meeting at an acute angle of about 45°.

Parr II].—Summary oF THE CHIEF CHARACTERISTICS OF
CRATEROSTIGMUS, AND A COMPARISON BETWEEN THIS GENUS
AND OTHER CHILOPODA.

The structural peculiarities of this new Chilopod and the
complex nature of its relationship to the other existing types
of this class, which are only imperfectly manifested in the
above-given description, may be more fully emphasised and
revealed in a clearer light by a brief summary : firstly, of the
characters in which it particularly resembles each of the three
most nearly allied groups; secondly, of those that it shares
with two of them to the exclusion of the third ; and lastly, of
those in which it differs from all. Adopting the terminology
I have elsewhere proposed, I shall speak of these groups as the
Lithobiomorpha, Scolopendromorpha, and Geophilomorpha,
leaving the explanation of the use of these names and the
rank assigned to them to be dealt with in a later part of this

essay.

A NEW AND ANNECTANT TYPE OF CHILOPOD. 420

With the fourth group, Scutigeromorpha, no comparison of
Craterostigmus will be necessary, since the only morpho-
logical points the two have in common, apart from those
shared by all Chilopods, are those mentioned in the statement
of the resemblances between Craterostigmus and_ the
Lithobiomorpha.

One or two of the features of Craterostigmus, namely, the
armature of the posterior legs, the structure of the mandibles,
etc., have been omitted from the list detailing the peculiarities
of this genus, as of doubtful taxonomic value, although the
structure of the mandible is likely enough to prove of high
importance in this particular.

1. Characters in which Craterostigmus resembles the
Lithobiomorpha and differs from the Scolopendro-
morpha and Geophilomorpha.—The completeness of the
penultimate and antepenultimate segments of the toxico-
gnaths, so that the femoral segment and fang are separated
from each other on the outer (post-axial) side of the appendage;
the presence of fifteen sternal plates and of fifteen pairs of loco-
motor appendages, with relatively large basal segments ; the
distinctness of the coxaof the fifteenth pair of legs; the presence
of six pats of stigmata on the somites represented by the
third, fifth, eighth, tenth, twelfth, and fourteenth pairs of legs ;
the presence of a single monomeniscous eye on each side of
the head, as in the genera Henicops and Cermatobius.

2. Characters in which Craterostigmus resembles the
Scolopendromorpha and differs from the Lithobio-
morpha and Geophilomorpha.—The co-existence of large
dentate preecoxal and femoral processes on the toxicognaths,
as in Scolopendra, Otostigmus, etc. The presence of
twenty-one tergal plates. The presence of a large and dis-
tinct metastigmatic pleural sclerite, and the increase in size of
the preecoxal sclerites towards the posterior end of the body.
The fusion of the trochanter (second segment) of the legs of
the fifteenth pair with the femur.

3. Characters in which Craterostigmus resembles all
or some Geophilomorpha, e.g. Mecistocephalus, and

42.6 R. IT. POCOOK,

differs from the Scolopendromorpha and Lithobio-
morpha.—The large size of the toxicognaths and the extent
to which they overlap the cephalite anteriorly and laterally ;
the presence of a distinct, subquadrate, basal plate, which is
much narrower than the tergum of the first leg-bearing somite
and intervenes between it and the cephalite ; the entirety of
the distal segment of the penultimate pair of legs.

4, Characters in which Craterostigmus resembles the
Scolopendromorphaand Geophilomorpha and differs
from the Lithobiomorpha.— The enlargement of the
tergum of the first leg-bearing somite; the size and com-
pleteness of the fusion of the two halves of the coxal plate of
the toxicognaths; the relative equality in size between the
coxe of the legs ; the presence of the prasternal sclerites
upon the ventral area of the somites.

5. Characters in which Craterostigmus resembles
some or allof the Lithobiomorpha and Scolopendro-
morpha and differs from all the Geophilomorpha.
—The number of antennal segments exceeding fourteen ;
the presence of eyes; the inequality in size of the terga;
the terga, sterna, and pairs of legs falling short of thirty-
one; the reduction in the number of stigmata, which fall
short of half the number of somites; the spine armature
of the legs.

6. Characters in which Craterostigmus resembles
the Lithobiomorpha and Geophilomorpha and
differs from the Scolopendromorpha,—tThe presence
of a distinct “basal plate’’ separating the cephalite from
the tergum of the somite bearing the first pair of legs;
the inability to withdraw the anal and genital somites within
the somite bearing the legs of the last pair.

7. Characters in which Craterostigmus differs
from all other known Chilopoda.—The fact that
the pleural sclerite of the toxicognath covers only the posterior
portion of the upper surface of the coxal plate on each side, and
falls far short of the proximal end of the femur; the numerical
excess of the terga over the sterna, the pairs of legs and the

A NEW AND ANNECTANT TYPE OF CHILOPOD. 427

stigmata; the fusion of the sternal and pleural sclerites of the
penultimate and antepenultimate somites, and the complete
coalescence of the tergal, pleural, sternal, and presternal
elements of the last somite to form a compactly chitinised
subcylindrical tube; the distinctness of the coxa of the
posterior leg, coupled with its articulation to the posterior
extremity of the somite and the fusion of the trochanter
with the femur; the representation of the external skeletal
elements of the genital and anal somites by a pair of valves
fused dorsally, but opening ventrally by a long slit-like
aperture, whence the genital and anal products escape to the
exterior.

The above-given comparisons demonstrate the impos-
sibility of associating Craterostigmus with either of the
groups of Chilopoda hitherto recognised, and substantiate
its claim to take a rank at least as high as that which is
assigned to the others individually, call them families,
sub-orders, orders,—what you will.

Part II].—Tse Testimony suPpPLIED BY CRATEROSTIGMUS
AS to THE Descenr oF tHE LITrHOBIOMORPHA FROM THE
ScOLOPENDROMORPHA.

Apart from the characteristics peculiar to itself, which
present us with new facts in Chilopod morphology ; apart,
too, from those that it shares with either one or more of the
previously known divisions of this class of Arthropods, the
chief point of interest vested in Craterostigmus lies in the
explanation it furnishes of the principal resemblances and
differences obtaining between the Lithobiomorphous and
Scolopendromorphous types of structure, and also in the
new and wholly unexpected light it throws upon the
metamerism of these two types, enabling us to picture
the process by which the one has been converted into the
other.

Except for the numerical difference, the somites of
Lithobius and Scolopendra exhibit certain obvious and

428 R. IT. POCOCK.

well-known resemblances with regard to the distribution
of the stigmata, and the alternate development of larger
and smaller tergal plates.

In both forms the first, third, fifth, seventh, eighth, tenth,
twelfth, and fourteenth terga are large, the second, fourth,
sixth, ninth, eleventh, thirteenth, and fifteenth small, and
stigmata are situated beneath the lateral margins of the
third, fifth, eighth, tenth, twelfth, and fourteenth. On the
other hand, the difference consists in the circumstance that
whereas in Lithobius the genital and anal somites succeed
the fifteenth, in Scolopendra no fewer than six somites
—repeating the characters of those that precede them,
the sixteenth, eighteenth, and twentieth being large and
stigmatiferous, the seventeenth, nineteenth, and twenty-
first small and astigmous—are intercalated between the
fifteenth and the genital.

These facts forcibly suggest the conclusion that the
fifteen somites of Lithobius are homologous to the
anterior fifteen somites of Scolopendra, and that the
reduction in the number of somites in the former is due
to the excalation of six somites between the fifteenth and
the genital, or to the failure to develop them from the
embryonic “caudal lobe,” the genital somite being pre-
sumably homologous in the two forms, and, in conjunction
with the anal somite, representing in the adult the posterior
portion of the “ caudal lobe ” in the embryo.

There is an attractive simplicity about this view of the
case which formerly induced me to hold it as a working
hypothesis. But the discovery of Craterostigmus puts
a quite different complexion on the whole question.

In the first place, the correspondence with respect to
numbers of stigmata and legs between Craterostigmus
and Lithobius, and the position of the stigmata above the
bases of the third, fifth, eighth, tenth, twelfth, and four-
teenth legs on each side, fully justifies the supposition that
the stigmata and legs are strictly homologous in the two
forms. Indeed, it is not easy to see what other opinion

A NEW AND ANNECTANT ‘TYPE OF CHILOPOD. 42,9

on this point is to be held. But the stigmata in Cra-
terostigmus lie beneath the fourth, seventh, twelfth,
fifteenth, eighteenth, and twentieth tergal plates; whence it
follows that these tergal plates are the dorsal elements of
the somites bearing the third, fifth, eighth, tenth, twelfth,
and fourteenth pairs of legs. That is to say, Cratero-
stigmus, as compared with Lithobius, has one extra
tergal plate without sternal or appendicular representative
between the first and third pediferous somites, one between
the third and fifth, two between the fifth and eighth, one
between the eighth and tenth, and one between the tenth
and twelfth, making in alla total of six. Which are these
supernumerary terga? the second or the third? the fifth
or the sixth? which two of the eighth, ninth, tenth, and
eleventh ? the thirteenth or fourteenth? the sixteenth or
seventeenth ?

The answers to these questions are not at first sight
easily given, owing to the difficulty of deciding which of
two terga overlying a sternum is its representative on the
dorsal area. Careful examination, however, of the nature
of the folds and the development of sclerites in the pleural
integument shows that the terga without sternal comple-
ments are the third, sixth, ninth, eleventh, fourteenth, and
seventeenth, these being the plates beneath which the repre-
sentatives of the metastigmatic pleural sclerite is scarcely
or not at all developed. Moreover the third, sixth, eleventh,
and fourteenth terga are overlapped posteriorly by the
terga that lie behind them; the ninth does not overlap
the anterior border of the tenth; and the seventeenth,
although overlapping the eighteenth in the normal manner,
does not overlie the precoxal sclerite of the eleventh
pediferous somite, but is, as it were, wedged in between the
tergum of this and of the following somite.

In the second place, the similarity between Craterostig-
mus and Scolopendra in the structure of the maxille,
palpognaths, and toxicognaths, in the number of antennal
segments, the equality in size of the cox, the fusion of the

430 R. I. POCOCK.

trochanter of the posterior legs with the femur, the presence
of preesternal sclerites, etc., renders unavoidable the conclusion
that the twenty-one terga in the two forms are homologous
plate for plate. Assuming that this is so, and that the
opinion given above touching the supernumerary terga 1s also
correct, it follows that the passage from the Scolopendroid
type, with twenty-one complete somites, to the Lithobioid
type, with fifteen complete somites, was effected by the ex-
calation of the third, sixth, ninth, eleventh, fourteenth, and
seventeenth somites of the former type. This new explanation
of the connection between the metamerism of the two types is
illustrated in the annexed diagrams (p. 433), and may be
tabulated as follows :

ScoLOPENDROID TPE. LitHoBioip TYPE.
1 major and astigmous represents . : . 1 major and astigmous or
stigmatiferous.
2 minor and astigmous represents . y . 2 minor and astigmous.
3 major and stigmatiferous. Excalated.
4 minor and astigmous represents . : . 3 major and stigmatiferous.
5 major and stigmatiferous represents . . 4 minor and astigmous.
6 minor and astigmous. Excalated.
7 major and astigmous or stigmatiferous
represents. ; ‘ : : . 5 major and stigmatiferous.
8 major and stigmatiferous represents . . 6 minor and astigmous.
9 minor and astigmous. Excalated.
10 major and stigmatiferous . : : . 7 major and astigmous.
11 minor and astigmous. Ixcalated.
12 major and stigmatiferous represents . . 8 major and stigmatiferous.
13 minor and astigmous represents ; . 9 minor and astigmous.
14 major and stigmatiferous Excalated.
15 minor and astigmous represents 4 . LO major and stigmatiferous.
16 major and stigmatiferous represents . . 11 minor and astigmous.
17 minor and astigmous. Excalated.
18 major and stigmatiferous represents . . 12 major and stigmatiferous.
19 minor and astigmous represents : . 13 minor and astigmous.
20 major and stigmatiferous represents . . 14 major and stigmatiferous.
21 minor and astigmous represents ‘ . 15 minor and astigmous.
22 genital represents. ; : : . 16 genital.
23 anal represents 5 ; ; ; . 17 anal.

If the theory here set forth be in accord with fact, it is

A NEW AND ANNECTANT TYPE OF CHILOPOD. 431

evident that the similarity, exact though it be, between the
fifteen leg-bearing somites of Lithobius and the anterior
fifteen leg-bearing somites of Scolopendra, with respect to
the alternation in size of tergal plates and the number and
disposition of stigmata, is the outcome of homoplasy and not
of homology.

Startling as this proposition may at first seem, the difficulties
in the way of its comprehension are not insuperable.

Accepting the comparison between the somites of the two,
as tabulated above, and assuming, as is justifiable, that the
stigmata occurring upon the first, third, and tenth somites of
the Lithobioid type and unrepresented on the corresponding
somites of the Scolopendroid type have not been independently
acquired in the former case, we are forced to conclude that
the Scolopendroid type itself has been derived from a form
possessing stigmata at least upon the first, third, fourth, fifth,
seventh, eighth, tenth, twelfth, fourteenth, fifteenth, six-
teenth, eighteenth, and twentieth somites, since stigmata are
found upon these somites either in the Scolopendroid type
or upon their hypothetical equivalents in the Lithobioid.

It is possible, however, to go a step further. Since all the
Geophilomorpha, without exception, and one, in this respect,
aberrant genus of Scolopendromorpha, namely Plutonium,
possess stigmata upon all the leg-bearing metameres except
the first and last, it is no strain upon probability to assume that
the primitive Scolopendromorph was furnished with stigmata
upon all the leg-bearing somites except the last. Further-
more, since stigmata are found only upon major somites, it
follows that all the somites were originally of that kind.
This contention is fortified by the consideration that simi-
larity in size and structure of the individual segments of a
series is one of the first laws of metamerism.

Hence, although it may be impossible to give a physio-
ogical explanation of the fact, no difficulty on morphological
grounds obstructs the acceptance of the opinion that a typical
Scolopendroid, like Rhysida, with stigmata only upon the
third, fifth, seventh, eighth, tenth, twelfth, fourteenth, six-

432 R. I. POCOCK.

EXPLANATION OF FIGURE OPPOSITE.

Diagram to illustrate the homology of the metameres in
Chilopoda of the Scolopendroid, Craterostigmoid, Litho-
bioid, and Scutigeroid types.

LETTERING AND NUMBERING FOR ALL FIGURES.

at, Antenna. c. Cephalite. 4. Basal plate = tergum of somite-bearing
toxicognath. ¢. Toxicognath. p. Palpognath. 1—21, 1—15, 1—8. Terga
of leg-bearing somites. 1'—21’, 1’—15'. Sterna of leg-bearing somites.
g. Genital somite. a. Anal somite.

A. Hypothetical primitive Scolopendroid with twenty-one subequal stigma-
tiferous somites, with gonopods on the genital somite (g.), a distinet tergal
plate (4) on the somite bearing the toxicognaths, and with the penultimate
and antepenultimate segments complete on the post-axial side of this appen-
dage.

B. Typical Scolopendroid, resembling Rhysida, derivable from A by the
restriction of the stigmata to the somites numbered 3, 5, 7, 8, 10, 12, 14, 16,
18, 20; by the incipient differentiation of the leg-bearing somites into major
stigmatiferous and minor astigmous; by the suppression of the gonopods ;
and by the obliteration of the tergum of the toxicognath and of the penulti-
mate and antepenultimate segments of this appendage on its post-axial side.
This figure also shows the juxtaposition of the two major somites numbered 7
and 8.

C. Diagram of Craterostigmus to show its derivation from A by the
restriction of the stigmata to the somites numbered on the dorsal side 4, 7,
12, 15, 18, 20, and by the excalation of the sternal area and appendages of
the somites numbered 3, 6, 9, 11, 14, 175 also by the coalescence of the
elements of the genital and anal somites to form a bivalvular capsule (a + 9).

D. Typical Lithobioid, resembling Henicops, derivable from C by the
completion of the excalation uf the half-somites in C numbered on the dorsal
side 3, 6, 9, 11, 14, and 17, but retaining the distinctness of the genital and
anal somites and the stigmata upon the first leg-bearing somite, asin A. This
figure also shows the alternation of major stigmatiferous and minor astigmous
somites, the juxtaposition of the two major somites, 7 and 8, and the change
of the major somites from the odd to the even numbers in the posterior half
of the body.

BE. Diagram of Scutigera to show its descent from a Lithobioid of the
Henicops type (D) by the dorsal migration and fusion of the stigmata; by
the enlargement of the stigma-bearing tergal plates, their backward extension
over, and almost complete obliteration of, the terga of the minor somites
represented ventrally by the sterna numbered 2’, 4’, 6’, 9’, 11’, 18’, and by the
fusion of the terga of the seventh and eighth somites to form a single plate
(4)—the two processes resulting in the presence of eight visible tergal plates
instead of fifteen. This figure also shows the widely separated antenne, the
polymeniscous eye, the large projecting preantennal area of the cephalite, the
long and pediform palpognath and partially rotated toxicognath, characters in
which Scutigera differs from all kuown Chilopods.

435

ANNECTANT TYPE OF CHTILOPOD.

AND

NEW

A

. 1 : \ ' '
= i) ' Vu ‘ u H \ \ 1
vy ' et el We f Z vi \ ME M el 1,2! 1 Vl 1,81 1
\ ——— — = — !
=: CO ie) in<9) SS aw,
d. : ~ (Sv : aie =
: fe) fe)
a = > —> <=
— N ‘ XN y ‘ ‘ ; ; @. 6/ ol AW 2 ] > 7 Saree > ha 7 = -
Sole RAS Sales etd Ati 19; iy lea el Sil Sie 27 al" Cline" te
ar , \ PK \ \ : i] ee D RES y OT ue 2 2
\ , Les - 45 ae is
’
7 7

Cl climes

fl vl Sl

Sl 9 Li

4

/ 81 6l 02 oa)

] \
H Silomnns 7 : § ‘ 1 \
99 4 tit ane ipa 19 /e oer or Ws el isn: pie si) 9) | 201 el | Gl i osm ie) fp
i oh Geet ee fog: Bi bt me ame N te ha ia pe
; ie ee lg RPGR as MMR! acta fey gee Oot tat AL! ig Ice OP tee ty a et \
, ; ! t } 1 1 ‘ ! ' ! 1 i ' ! \ i ' H ' i
iN i eye VS e ete Ochi: eb ye el i Siete) Bis el! oe) ie) \
: . : L 1

112 3

1

'
! \ 1 \ '

. ' \ i)

> I ' \ i \ AL eit \ H x 1 \ :

,o4 \ Wa \ enerT i ¥ ' \ u : \

i \ \ ya \ \ \

, i! \ UE ' \

6I | 02
\

i]
\ i
1
'

7
/ / / C74 4
Ms 1 hee Here / Tie 7 ¢ 7
' ' ' / , Wee, 77 es
7

Le ¢
4 SN \ \N \ a \ ' by y Js y pete i yer
N - \ \ vi 1 uf / os of
“ \ | Wwe Aw WW \! S 1/9 iH WE V 8 f £6 Ol 1 Me? ; | 4£\- Al A
» _ 2 ‘fs a 4 ue

i @ 1,601 01! Hi

‘

eid eae oer ane
‘2lrsii vt igi}

ie}
'

Lv]
1

esd eal li

ee

v
|

434, R. I. POCOCK.

teenth, eighteenth, and twentieth somites, and Craterostig-
mus with these organs restricted to the somites represented
by the fourth, seventh, twelfth, fifteenth, eighteenth, and
twentieth tergal plates, have descended along divergent lines
of evolution from a primitive Scolopendromorph with stigmata
upon all the leg-bearing somites except the last.1_ But since
the fourth, seventh, twelfth, fifteenth, eighteenth, and twen-
tieth tergal plates in Craterostigmus represent the dorsal
elements of the third, fifth, eighth, tenth, twelfth, and four-
teenth somites of a Lithobiomorph, the similarity between
the latter and a typical Scolopendroid, with respect to the
location of stigmata, may be regarded as merely coincidental
without any infringement of probability.

So, too, with regard to the resemblance in alternation of
larger and smaller somites. This alternation is no doubt
beneficial in the way of favouring flexibility and rapidity of
torsion. Its acquirement, therefore, may be regarded as
advantageous. But this does not explain why there should
be an exact correspondence in this particular, if the two
forms under consideration have been independently evolved
from a type in which all the somites were major. Why, it
may be asked, are the second, fourth, sixth, ninth, eleventh,
thirteenth, and fifteenth somites minor in their constitution
in both cases, and the others major? The explanation is to
be found in the situation of the stigmata. The invariable
absence of stigmata from minor somites in the Chilopoda
justifies the view that the presence of these apertures and of
the tracheal trunks that arise from them is incompatible with
somites of that nature. This is supported by what occurs
in the Geophilomorpha, where a large number of breathing
orifices appears necessary for respiration underground, and
where all the somites are majorand subequal. But functional
representatives of the minor somites of the Scolopendroid

1 The atrophy of the breathing orifice upon the first leg-bearing somite in
the Geophilomorpha, the Scolopendromorpha, and Craterostigmus is no
doubt traceable to the modification impressed upon this somite by the great
development of the muscularity of the toxicognathis.

A NEW AND ANNECTANT TYPE OF CHILOPOD. 435

and Lithobioid types have been acquired by a secondary
segmentation of the metameres which has given rise to
the intercalation of a series of small segments between the
genuine somites. Unless in the case of these centipedes
there were some factor inimical to the presence of stigmata,
and the trunks arising therefrom, upon minor somites, it 1s
difficult to explain why the somites do not alternate in size,
when the necessity for the alternation is attested by the
secondary development of the sub-segments above referred to,
and when the capacity for the segmentation of fresh somites
from the embryonic caudal lobe seems practically lmitless.

Postulating, then, the expediency of alternation in size of
somites, and the incompatibility of the presence of stigmata
upon those belonging to the minor category; seeing too
that the third, fifth, eighth, tenth, twelfth, and fourteenth,
both in the Lithobioid and the Scolopendroid types, have
retained their stigmata and their major nature; and that the
first, although astigmous in the Scolopendroid and often in
the Lithobioid, is also major, it is no matter for surprise
that the second, fourth, sixth, ninth, eleventh, thirteenth,
and fifteenth are minor in both cases.

But again, it may be asked, why does the regular alterna-
tion of major and minor somites cease with the somite which
in the two types is numerically the seventh from the anterior
end—an arrangement which results in the change of the
major somites from the odd to the even number, the major
being odd and the minor even in front of the seventh, while
the converse holds behind it ?

As for the change itself, we must remember that it is a
law of Chilopod metamerism that the number of leg-bearing
metameres is invariably odd. Hence, since the first leg-
bearing metamere belongs always to the major and the last
always to the minor category, both being odd-numbered,
the regular alternate succession of major and minor meta-
meres is an impossibility so long as the law is adhered to.
At one point at least in the series an alteration im the order
of sequence must be introduced, which will result in two

436 R. 1.) POCOCK:

major or two minor somites coming together. And surely
we should expect the alteration to be effected, as indeed it
is, by the juxtaposition of two major instead of two minor
metameres if, as has been suggested, all the somites belonged
originally to the major category.

That the change sets in in Lithobius close to the middle
of the body, and in Scolopendra in the anterior third, may
perhaps be explained by some such consideration as the
following. The juxtaposition of two major somites is no
doubt unfavourable to free flexibility; and since it is advan-
tageous for the greatest possible freedom of movement to lie
in the anterior and posterior ends of the body, which bear
the principal external organs set apart for the performance
of the vegetative and generative functions, the reason becomes
clear for the juxtaposition of two major somites to take place
as near as can be to the middle of the body in Lithobius,
where the body is relatively short, few-jointed, and less
flexible. In Scolopendra, on the other hand, with its longer,
more jointed, and therefore more flexible body, it is probably
immaterial where the juxtaposition of the major somites
exactly occurs, provided it is sufficiently distant from both ends
not to interfere with the full flexibility of the head and tail.

‘here appears, then, to be no insuperable or even serious
difficulties in the way of believing that the similarity in
segmentation between the Lithobiudez and typical members
of the Scolopendride is assignable to homoplasy. If, on the
other hand, it is attributable to the direct homology of the
somites concerned, the view that Craterostigmus exhibits
a stage of segmentation intermediate between those of the
other two must be wrong. In that case no genetic signifi-
cance can be attached to the presence of twenty-one tergal
plates in Craterostigmus and in Scolopendra. The
numerical likeness must be regarded as purely coincidental.
This view of the matter I find it impossible to adopt when
the other many and deep-seated structural details in which
the two forms approximate to one another are taken into
consideration,

A NEW AND ANNECTANT TYPE OF CHILOPOD. 437

Finally, it may be stated with confidence and without fear
of contradiction that the true nature of the connection, as
above explained, between the metamerism of the Lithobioid
and Scolopendroid types would never have been guessed, had
it not been for the fortunate survival of this intermediate
form with the six additional somites of the last-named type
in process of excalation.!

Part ITV.—CHARACTER AND CLASSIFICATION OF THE CHILO-
PODA.

It is beyond the scope of this article to give a detailed
account of the development of the various classifications of
the Chilopoda that have been proposed. ‘he main outlines
of the subject need only be sketched. Concerning the
division of the class into the four main sections typified by
Geophilus, Scolopendra, Lithobius, and Scutigera,
there has been a pretty general consentience ; but authorities
have been divided into two schools of opinion touching the
relationship of these sections to one another. The first,
maintaining the isolation of the Scutigeride, was set on foot
by Latreille in 1825 (‘Faune du Régne Anim.,’ p. 327), was
supported by Brandt (‘ Bull. Acad. St. Petersb.,’ vii, p. 311,
1840), and held its own practically unchallenged until the
publication of Newport’s last monograph in 1856, when
Brandt’s classification of the class into Schizotarsia for

1 It is perhaps necessary to point out that the line of argument here
adopted to explain the descent of the Lithobioid from the Scolopendroid type
is based upon the conviction that intercalation of somites never occurs. Were
this otherwise the process of transformation might be reversed, and the Scolo-
pendroid type derived from the Lithobioid ; but this would in no sense invali-
date the contention that Craterostigmus occupies an intermediate position
between the other two. Haase (‘ Zeits. ent. Breslau,’ pt. vili, 1881, pp. 93
—115) held an entirely different view in dealing with the phylogeny of the
Chilopoda. He regarded the richer metamerism of the Geophilide and Scolo-
pendride as a derived, and not as a primitive character. Both Geophilide and
Scolopendride have descended, according to him, from a primitive ‘ epimor-
phous” stock, which was itself derived from an earlier “ anamorphous ” type,
whence, along independent lines of development, arose the Lithobiide and
Scutigeride.

VOL, 45, PART 3.—NEW SERIES, Ga

438 R. I. POCOOK.

Scutigera, and Holotarsia for the Lithobiide, Scolo-
pendride, and Geophilidx, appeared modified only in
minor particulars. The second, uniting Scutigera with
Lithobius, to the exclusion of the other two, was founded
by Meinert in 1868 (‘ Nat. Tidskr.,’ p. 246), who arranged
the groups of Chilopoda as follows :—Geophili, Scolopendre,
and Lithobii, the last including the two tribes Lithobiini
and Scutigerini.

Although this system was rejected in 1880 by Latzel
(‘Die Myr. Osterr.-Ungar. Mon. 1’), who allowed the four
families instituted by Leach and Newport to stand unasso-
ciated, it was taken up and amplified by Haase, who, adopting
the four families and Meinert’s view of the near relation-
ship between Scutigera and Lithobius, went a step
further and classified the Chilopoda into the Anamorpha
and Epimorpha, the first containing the Scntigeride and
Lithobiidee, the second the Scolopendrides and Geophilide
(‘Schles. Chilopoda,’ Breslau, 1880, Inaug.-Dissert.).

This classification of Haase’s was adopted unchanged by
Meinert (‘Trans. Amer. Phil. Soc.,? xxin, p. 1638, 1886),
and with certain amplifications by Bollman (‘ Bull. U.S. Nat.
Mus.,’ xlvi, p. 163, 18935) and Silvestri ( Ann. Mus.,’ Genova
(2), xiv, pp. 622—634, 1895).

The variation proposed by Bollman consisted in the sub-
division of the “order”? Anamorpha into two sub-orders,
for the first of which, containing the Scutigeride, he used
Brandt’s term Schizotarsia; while for the second, including
the Lithobiide and Cermatobiide, he introduced the new
term “ Unguipalpi,” in allusion to the presence of a claw
upon the palpognaths.

Silvestri followed Bollman in splitting the Anamorpha,
and replaced the Schizotarsia and Unguipalpi with the new
names Anartiostigmata and Artiostigmata, to which an
ordinal value -was attached. He further divided the
Kpimorpha into two orders, namely, Oligostigmata for
the Scolopendride, and Pantastigmata for the Geophilide.

Although Haase’s classification has the unquestionable

A NEW AND ANNECTANT TYPE OF CHILOPOD. 439

merit of drawing attention to the undeniable relationship
subsisting between the Scutigeride and the Lithobiide on
the one hand, and the Scolopendride and Geophilide on
the other, it appears to me to be open to the serious
objection of obscuring the value of the great structural
differences which separate the Scutigeride not only from
the Lithobiide, but from the Scolopendride and Geophilide
as well, and of ignoring the many and important features
in which the Lithobiide resemble the Scolopendride.}

The characters which justify the association of the
Lithobudz with the Scutigeride are the following :—the
presence of fifteen pairs of legs in the adult, of seven
pairs in the newly hatched young, the rest being added
with successive moults ; the presence of stigmata upon the
same somites ; the completeness of the penultimate and ante-
penultimate segments of the toxicognaths ; the absence of all
trace of “sub-segments” on the somites; the large size of
the cox of the legs, and the freedom of those of the
posterior pairs, and the presence of powerful gonopods in the
female—characters which are not found in the other families.

But although these similarities attest the derivation of
the Scutigeride from Chilopoda of the Lithobioid type,
there can surely be no two opinions as to the extent to
which they are outweighed when placed in the balance
against the many and deep-seated characteristics in which
Scutigera differs from all other existing centipedes.

My views on this point were expressed in the classification

1 No doubt Haase found confirmation for his views in the discovery of the
genus Cermatobius—a genus differing from all the Lithobiide known to
him, and approaching the Scutigeride in its longer, “ funiculate” antenne, longer
and subdivided tarsi, absence of coxal pores, longer female gonopods, and
slightly more dorsal stigmata. There is no doubt, however, that Cermatobius
is very closely related to Lamyctes and its allies. For example, although
typically four or five in number in these forms, the coxal pores are reduced
to one on each coxa in Haasiella and in Henicops, as typified by
maculatus, the anterior tarsi consist of three segments, and the posterior
pairs of six—facts which largely discount the value of the absence of coxal
pores and the rich tarsal segmentation in Cermatobius.

4.4.0 Re 12 BOCOCK:

proposed in 1896 (‘ Biol. Centr. Amer.,’ Chilopoda) —a classi-
fication which, allowing for the alteration in terminology
and an elevation in rank of the groups, was that of
Latreille, Brandt, and Newport amplified.

Two sub-classes were recognised, the Artiostigma for the
order Scutigeromorpha, and the Anartiostigma for the Jitho-
biomorpha, Scolopendromorpha, and Geophilomorpha.

If, as I venture to think, this system expressed at the
time more accurately than Haase’s the importance of the
characters of Scutigera and the relationship between the
Lithobiide and the Scolopendride, the opinion that it
embodies receives the strongest support from the discovery
of Craterostigmus. In no sense does this genus tend to lessen
the structural interval between the Scutigeroid and Litho-
bioid types; but it largely depreciates the value of the
characters the two have in common—the characters, in short,
upon which Meinert and Haase relied when uniting them. Those
who in the future maintain the Anamorpha must base this
group upon the completeness of the penultimate and ante-
penultimate segments of the toxicognaths and the numbers
of stigmata and pairs of legs, if Craterostigmus be included ;
for this genus negatives the remaining characters of the
group, with exception of the number of legs at hatching,
a point about which few would venture to hazard a guess so
far as Craterostigmus is concerned.

In connection with the classification here submitted, it is
desirable to justify the changes in terminology that have
been introduced. Notostigma and Pleurostigma are preferred
to Artiostigma and Anartiostigma because the latter were
used ina much more restricted sense by Silvestri; and to
Schizotarsia and Holotarsia because these two lost their
applicability with the discovery of Cermatobius. Ungui-
palpi connotes a character which holds only when the
Lithobiide are associated with the Scutigeridz ; and Panta-
stigmata and Oligostigmata make no allowance for the
pantastigmous character of Plutonium, which neverthe-
less is closely related to the oligostigmous forms,

A NEW AND ANNECTANT TYPE OF CHILOPOD. 44]

Class Chilopoda.

Opisthogoneate tracheate Arthropods descended from a
primitive type in which the body consisted of a large number
of metameres similar in size and form, each with a broad
tergal and sternal plate, connected by a pleural membrane,
a pair of stigmata, and a pair of short hexarthrous, mon-
onychous appendages. At an early phylogenetic stage the
stigmata disappeared from the last three somites in con-
nection with the special functions these somites were set
apart to fulfil, the last or anal losing its appendages and being
degraded to the condition of a carrier of the anal orifice ;
the penultimate or genital being, like the anal, reduced in
size as a safeguard against damage, but retaining its
appendages in a dwarfed form as manipulators of the ova
or spermatophores; and the antepenultimate or last leg-
bearing somite having its appendages enlarged and directed
backwards to protect the genital somite, and, in later
stages, to become modified as tactile, offensive, or secondary
sexual organs. Anteriorly two pairs of appendages, retain-
ing a primitive biramous structure, were set apart as
gnathites ; and a single pair, lying in front of the latter,
consisting of not fewer than fourteen homonomous segments,
and widely separated at the base, as antenne. The dorsal
area of the three somites represented by the appendages
just mentioned was covered by a single plate, the cephalite
or head shield, which was furnished laterally with eyes, and
was curved downwards in front over the mouth to form a
labrum or upper lip. For seizing and holding prey, the
six-jointed appendages of the two metameres behind the
head were turned forwards towards the mouth, their basal
segments becoming enlarged and encroaching upon the
sternal area, each appendage at the same time under-
going a partial axial rotation, so as to fold in a horizontal
rather than a vertical plane. The first pair of these appen-
dages retained its primitive pediform structure, while in

449 R. I. POCOCK.

virtue of the development of a poison gland, accompanied
by an increase in the thickness of its segments, practically
the sole function of prehension was taken over by those of
the second pair.

Sub-class Pleurostigma.

Chilopods resembling the primitive type in the possession
of a system of tracheal tubes, the orifices of which open upon
the pleural area of more or fewer of the somites; and in
the presence of a distinct tergum and sternum on each leg-
bearing somite, the number of sterna never exceeding that
of the terga. Eyes are either preserved or lost; when
preserved they are monomeniscous, and variable in number
on each side from one to about forty. Correlated with the
relative simplicity in the structure of the metameres Just
mentioned is a marked specialisation of the head and the
appendages connected therewith, a specialisation which was.
initiated apparently by the retroversion of the preantennal
area of the cephalite resulting in the projection of the
antennee from the fore-margin of the head thus formed, and
in the backward movement of the mouth with the gnathites
of the first and second pairs into more intimate connection
with the palpognaths and toxicognaths. As concomitants
of this process the last-named appendages shortened and
thickened, were axially rotated so as to move and fold ina
horizontal plane, one segment of the palpognaths was sup-
pressed, and the basal segments of the toxicognaths became
firmly fused and incapable of independent movement; and
lastly, the gnathites of the second pair being underlain by
the two pairs that succeed them, failed to develop or lost
the sense-organ which is found in the other sub-class.

Order 1.—Geophilomorpha.

Chilopods retaining a large and indefinite number of meta-
meres In response to the demands of a subterranean existence

A NEW AND ANNKECTANT TYPE OF CHILOPOD. 443

requiring serpentine powers of movement. ‘l'o further subserve
this end, most of the somites are either partially or completely
divided into two anteriorly, a process which, when perfected,
gives rise to a series of sub-segments alternating with the
true somites, the former being represented by a pretergal
and a pair of presternal plates. Stigmata are retained on
all the somites, except the first and last, for purposes of
respiration under ground ; and the legs, which are short and
play but a subsidiary part in locomotion, preserve the primi-
tive number of segments, the basal of which remains small.
The eyes, as useless organs, have disappeared; and the
antenne are relatively short and consist invariably of fourteen
segments. The gnathites are specialised and variable in struc-
ture; the penultimate and antepenultimate segments of the
toxicognaths are reduced on the preaxial side of the appen-
dage to the state of arthrodial integumental folds between the
distal segment and the femur, which are firmly articulated
together on its post-axial side. The tergal plate of these
appendages always remains distinct and usually large, thus
separating the cephalite from the tergum of the first leg-
bearing somite. ‘The pleural sclerites are well developed,
and those of the last leg-bearing somite fuse with the coxa
of the appendage, the second segment of which remains free.
‘The anal and genital somites are incapable of retraction within
the somite which bears the last pair of legs, and the gonopods
typically persist in the male as small-jointed appendages, and
in the female also as jointed or unjointed, often lobate sclerites.
Many families—Geophilide, Oryidew, Gonibregmatide, etc.

Order 2.—Scolopendromorpha.

Chilopods departing from the primitive type and from the
Geophilomorpha in having the number of leg-bearing somites
fixed definitely to twenty-three or twenty-one, the reduction
in number in the latter case being effected by the excalation
of the two somites preceding that which bears the last pair
of legs; by the evanescence of the tergal plate of the toxico-
gnaths, which results in contact between the cephalite and the

444, R. I. POCOOK.

tergum of the first leg-bearing somite; in the ankylosis of
the second and third segments of the legs of the last pair; in
the incipient differentiation of the metameres into alternating
major and minor leg-bearing somites, the regularity of the
alternation, however, being interrupted at the seventh and
eighth somites, both of which belong to the major category ;
in the power to withdraw the anal and genital somites out of
harm’s way into the last leg-bearing somite, in the total dis-
appearance of the female gonopods, and the reduction of the
segments of the male gonopods to one. With the shortening
of the body and the lessening of the need of serpentine
mobility, no preetergal sclerites are definitely cut off from the
terga, though pairs of presternal sclerites are developed. The
eyes may be lost; when present they are invariably four in
number. The segments of the antenne never normally fall
below seventeen nor above about thirty. On the other hand,
the basal segments of the legs are feebly developed as
in the Geophilomorpha, and the toxicognaths are modified
and the pleural sclerite of the last leg-bearing somite en-
larged and fused with the basal segment of the appendage as
in the members of that order.
Several families—Scolopendriidee, Newportiidee, etc.

Order 58.—Craterostigmomorpha.

Chilopods derivable from a primitive form of Scolopendro-
morpha (which was furnished with twenty-one pedigerous
somites, nineteen pairs of stigmata, complete penultimate and
antepenultimate segments on the toxicognaths, distinct tergal
plate for these appendages, and undivided tarsal segment
on all but the last pair of legs, which had the first segment
free), by the excalation of the sternal area and limbs of the
third, sixth, ninth, eleventh, fourteenth, and seventeenth
somites, thus reducing the number of legs to fifteen
pairs; by the retention of six pairs of stigmata upon the
somites bearing the third, fifth, eighth, tenth, twelfth, and
fourteenth pairs of legs; by the enclosure of the anal and

A NEW AND ANNECTANT TYPE OF CHILOPOD. 4.45

genital orifices in a bivalvular sclerite, representing probably
the tergum of the anal somite, and by the complete fusion of
the sternum and pleura of the fourteenth, and of the sternum,
pleura, and tergum of the fifteenth leg-bearing somites.

One family—Craterostigmide (Craterostigmus).

Order 4.—Lithobiomorpha.

Chilopods derivable from the primitive form of Cratero-
stigmomorpha by the excalation of the third, sixth, ninth,
eleventh, and seventeenth tergal plates, which brings the terga
and sterna into numerical conformity ; by the differentiation of
the terga into major and minor in the interests of flexibility ;
by an increase in the size of the basal segment of the legs,
especially of the posterior pairs, and the segmentation of the
tarsi of all the legs, or at least of the thirteenth and four-
teenth pairs, to subserve rapidity of movement. In other
respects the members of this order stand nearer the primitive
type of the Chilopods than those hitherto considered. The
tereum of the first pedigerous somite is relatively small in
correlation with the weak muscularity of the toxicognaths,
the basal segments of which are less strongly fused; the
gonopods, present and jointed in both sexes in more
archaic types, are always well developed in the females, and
are supported upon a large ventral plate, which results from
the fusion of their basal segments with the sternum of the
genital somite; and lastly, in two families, the Henicopidx
and Cermatobiidx, the stigmata are preserved upon the
somite that bears the first pair of legs.

Three families—Lithobiide (Lithobius, ete.), Henico-
pidx (Henicops, etc.), Cermatobiide (Cermatobius).

Sub-class Notostigma.

Chilopods descended from the Cermatobioid type of Litho-
biomorpha by the replacement of the normal tracheal system
by a series of median dorsal pulmonary sacs furnished with
tubes dipping into the pericardial space, and opening each

44.6 R. I. POCOCK.

by a single stigma which results from the upward migration
and coalescence of the normal pair of stigmata upon the
first, third, fifth, eighth, tenth, twelfth, and fourteenth
somites—a change in the method of respiration which is
accompanied by the complete disappearance of the tergum
of the seventh somite, either by excalation or by fusion with
that of the eighth, and by the evanescence of all the minor
terga except that of the fifteenth leg-bearing somite.
Further specialisation is attested by the presence of a
sense-organ on the gnathites of the second pair, by the
polymeniscous eyes,! the duplication of the gonopods in the
male,” and the extreme length and perfected annulation of
the antenne and of the distal segments of the legs, those
of the fifteenth pair being clawless and simulating a couple
of feelers both in form and function. In virtue probably
of these extremely specialised features, mouth parts of a
markedly primitive type have been preserved; the long
and slender toxicognaths exhibiting only a partial axial
rotation, long penultimate segments, and disunited and in-
dependently moveable coxx, the palpognaths being also
incompletely rotated, long and pediform, with the same
number of segments as the primitive Chilopod limb, and
projecting freely at the sides of the head; the preantennal
area of the cephalite projects forwards and downwards in
front of the bases of the antenne, which remain widely
separated as they are in the early embryonic stages in this

1 There provisionally adopt the view already suggested by others, that the
faceted eye of Scutigera resulted from the packing together and mutual
pressure of a number of monomeniscous eyes like those of Lithobius. But
in view of the many primitive features appertaining to the head of Scutigera,
the possibility must be borne in mind of the derivation of a set of mono-
meniscous eyes from one of a polymeniscous type, as has doubtless occurred
in the case of the lateral eyes of Scorpio. Or, again, it may be that the
large convex eye of Scutigera, with its many facets, corresponds as a whole,
either in a derivative or originative sense, to the large single-lensed eye of
Cermatobius, the genus which, beyond all possibility of doubt, is most
nearly related to Scutigera in all other structural points.

2 Possibly the additional gonopods represent the missing appendages of the
anal somite.

A NEW AND ANNECTANT TYPE OF CHILOPOD. 447

class; and the second pair of gnathites are lengthened
especially with regard to the second (femoral) segment to
reach the mouth, which lies relatively far forwards.

Order Scutigeromorpha (characters as above).

One family—Scutigeride (Scutigera).

EXPLANATION OF PLATE 23;

Illustrating Mr. Pocock’s paper, “A New and Annectant
Type of Chilopod.”

(The figures, depicting the structural features of Craterostigmus tas-
manianus, were drawn by Mr. Pocock from the specimens of this centipede
which are preserved in the Natural History Museum.)

Fic. 1.—Dorsal view of anterior extremity. c. Cephalite or head shield.
o. Kye. sf. Frontal suture. a. Antenna. dp. ‘ Basal plate” = tergum of
toxicognath. p/. Pleural sclerite of toxicognath. ca. p. Coxal plate.
J/- Femur or third segment. p. Patella or fourth segment. 7¢. Tibia or fifth
segment. cl. claw or sixth segment of toxicognath. 1, 2,3. Terga of first,
second, and third somites.

Fic. 2.—Ventral view of toxicognaths, showing the precoxal processes of
the coxal plate (cz. p.), the second segment or trochanter (¢r.), which is fused
with the third segment or femur (/.), the preaxial teeth of the latter, and

the completeness of the fourth and fifth segments (p., ¢.) on the post-axial
side of the appendage.

Ftc. 3.—Lateral view of head and anterior extremity of body. Lettering
of head and toxicognath as in Fig. 1. 1—7. Terga of anterior seven somites,
numbers 3 and 6 being without sternal representatives. 1/—5’. Sterna of
anterior five leg-bearing somites. s. Stigmatic, and ms., metastigmatic
sclerite. pl.m. Membrane of pleural area. cw. Coxa or basal segment of
leg. per. precoxal sclerite. ps. presternal sclerite.

Fig. 4,—Lateral view of posterior extremity of body. 15—2l. Terga of

fifteenth to twenty-first somites. 10’—15’. Sterna of tenth to fifteenth leg-
bearing somites. The rest of the lettering as in Fig. 3.

Fic. 5.—Dorsal view of last leg-bearing somite with ano-genital capsule

(ag.) projecting from its posterior extremity. ¢. Tergal area of somite.
1 aud 3. First and base of third segment of leg of fifteenth pair.

448 Ro ie POCOCK

Fic. 6.—Ventral view of last leg-bearing somite and of ano-genital cap-
sule (ag.). 1—8. First, second, and base of third segment of leg of four-
teenth pair. 1’, 3’. First and base of third segment of leg of fifteenth pair.
s. Sternal area of last leg-bearing somite. ps. Fused presternal element.

Fic. 7.—Lateral view of the ano-genital capsule (ag.) and of the posterior
half of the last leg-bearing somite. /a¢. Lateral area of this somite. 1. Basal
segment of the leg.

Fig. 8.—Ventral view of cephalite or head plate. @. Basal segment of
antenna, sf. subfrontal or retroverted praeantennal area with labral sclerites.

Fig. 9.—Labrum. ¢. Teeth. mm. Membranous lateral portion of excision.

Kia. 10.—Lower or posterior aspect of mandible or gnathite of first pair.
/. Fringe of hairs at its distal extremity.

Fig. 11.—Upper or ‘anterior aspect of the same. / Fringe of hairs.
/. Pectiniform teeth. /. Membranous lobe.

Fic. 12.—Distal extremity of the last still more enlarged, showing the
series of nine pectiniform teeth arranged in groups of threes.

Fic. 13.—Inferior or posterior view of gnathites of the second pair
(maxille). 1. Basal segment bearing the bisegmented ectocorm (1, 2) and
the one-jointed ectocorm (éz/.).

Fic. 14.—Lower or posterior view of left palpognath. 1. Ectocoxite. I’.
Entocoxite. 2—5. Second to fifth segments, the second, or trochanter, fused
with the third, or femur.

Fig. 14 a.—Distal segment of the same from its anterior or upper side,
showing fringe of hair.

Fic. 15.—Lateral view of part of the third leg-bearing somite (fourth
somite from the dorsal side). ¢. Tergum. s. Stigmatic sclerite with stigma.
ms. metastigmatic sclerite. p/. Pleural membrane.

Fic. 16.—Leg of fourteenth pair. 1—6. Segments ; 2 with inferior spike,
5 and 6 with spines.

Fic. 17.—Basal portion of same. Jdaé. Part of lateral portion of the four-
teenth leg-bearing somite. 1. Basal segment or coxa of leg with subspherical
excrescence or nodule (#.). 2. Second segment or trochanter with spike (sp.).
3. Base of third segment or femur.

Fic. 18.—Leg of the fifteenth pair. 1—7. First to seventh segments, the
sixth and seventh representing the sixth of Fig. 16; 1 with inferior spike,
2 ankylosed to 3.

Fic. 19.—-Basal portion of same. /a?¢, Lateral area of part of fifteenth leg-
bearing somite. 1. Basal segment or coxa with (z.) nodular prominence and
(sp.) spike. 2. Remnant of the second segment or trochanter fused with the
base of the third segment or femur (3).

yi oe fs

as
>

, 4. : -

we ‘

r . =%
a t = ; {
- = - ; 7 { ra =
; 5 ww \

lat.----~-

A.S.Huth,Litht London.

THE TRYPANOSOMA BRUCII. 44.9

The Trypanosoma Brucii, the Organism found
in Nagana, or Tse-tse Fly Disease.
By
J. R. Bradford, F.R.S.,

and
H. G. Plimmer, F.L.S.

(From the Laboratory of the Brown Institution).

With Plates 24 and 25.

ContTENTS.
PAGE
I. Nomenclature ; - : ; : . 450
II. The Adult Form of the Trypanosoma Brucii ‘ . 450
a. Examined in the Living State. : ; : 7450
z. Examined after Fixation and Staining. : . 452

1. Methods : : : : ;
2. The Appearances of the Adult ‘'rypanosoma . 453
III. Reproduction of Adult Forms of Trypanosoma Brucil . 455

IV. Conjugation between the Adult Trypanosomata Brucii . 496

V. “Ameboid” and “ Plasmodial”” Forms of the Trypanosoma
Brucii : : 5 : , . 458

VI. The Distribution of the Trypanosoma Brucii, its Varia-
tions in Different Animals, and their Resistance to it . . 460

VII. The Differences between the Trypanosoma Brucii and the
Trypanosoma found in Sewer Rats : . 464
VIII. The Micronucleus : : ; : ‘ . 465
IX. The Life-history of the Trypanosoma Brucii ; . 466

Xe Explanation of the Plates ; ' : : . 468

4.50 J. R. BRADFORD AND H. G. PLIMMER.

I. NoMENCLATURE.

In our preliminary paper’ on the organism found in the
T'se-tse fly disease we suggested Trypanosoma Brucii as
a suitable and appropriate name for it. But after this
Laveran, in his papers on this organism, gave it the name
Herpetomonas Brucii, which apparently he considered
to be a more accurate designation biologically. However,
in a recent paper by Laveran and Mesnil,’ discussing these
terms from a morphological point of view, they show that
Trypanosoma is a more accurate term than Herpetomonas ;
in fact, that this latter term ought to disappear altogether,
and that the generic name Trypanosoma ought to be used to
designate all those flagellated organisms at present known,
which are parasitic in the blood of vertebrates. Hence we
shall retain the name we used in our paper above mentioned,
as being probably the more accurate one.

IJ]. Tae Aputt Form or tHe Trypanosoma Bructt.

A. Examined in the Living State.—The adult form
of the Trypanosoma can be easily studied in the fresh blood
of an affected animal, by placing a drop on a slide and
simply covering it with a cover-glass, and examining it with
an objective of suitable power. But, as these organisms are
extremely delicate, and are very easily broken up by slight
pressure, it is better to make a very thin cell, by running a
small drop of melted paraffin on to a slide in the form of a
ring or square, by means of a red-hot platinum loop, and to
place the drop of blood in the centre, covering it with a
cover-glass, which must be sealed with paraffin, in order to
prevent evaporation, if the blood is to be examined for any

1 «A Preliminary Note on the Morphology and Distribution of the Organism
found in the Tse-tse Fly Disease,” by H. G. Plimmer and J. Rose Brad-
ford, F.R.S., ‘ Proc. Roy. Soc.,’ vol. lxv, p. 274.

2 «Sur la structure du Trypanosome des grenouilles et sur l’extension du
genre Trypanosoma, Gruby; par Laveran et Mesnil, (Comptes rendus de
la Société de Biologie,’ tome liii, No. 28.

THE ''RYPANOSOMA BRUCII. 451

length of time. ‘I'his method obviates pressure, and the organ-
ism may be easily studied with even the highest powers.
The Trypanosoma will then be seen to consist of a worm-like
body, more or less homogeneous in structure, with at one enda
blunt, stiff extremity, and at the other a long, wavy flagellum.
With good illumination this flagellum can be seen to be
continuous with a wavy membrane which extends along
one surface of the long axis of the organism. For some
time the Trypanosoma will be seen to be in active
motion, which is caused by the rapid lashing movements of
the flagellum, by the wavy movements of the undulating
membrane, and by the contractions and relaxations of the
protoplasmic mass forming the body. In order to arrest
this movement a drop of 1 per cent. gelatine solution, or of
a weak solution of cherry-gum, should be mixed with the
blood, which will so reduce the movement that it will be
quite easy to study as much of the organism as can be made
out in unstained preparations. In order to see the structure
of the living organism, dark ground illumination, or mono-
chromatic light must be used. ‘The former is difficult to
use with an immersion lens, but can be very satisfactorily
used with a 2 mm. immersion objective, after the method of
Gebhardt.'. We have used spectral monochromatic light
(blue), but very good results can also be obtained more
easily with Gifford’s malachite-green screen. With these
methods of illumination the organism appears as a highly
refractive body, and near the middle is seen a darker mass
—the nucleus; and near the blunt, stiff end of the body is
seen a tiny dark dot, which we will here call the micro-
nucleus; the question of its proper nomenclature will be
discussed later. At the same end of the Trypanosoma, in
varying relations to the micronucleus, is seen a vacuole.
The protoplasm is now seen not to be uniform, but it appears
to have an alveolar structure, as described by Biitschli.
With regard to the movements of the organisms, the
commoner mode of progression would seem to be with the
1 « Zeitschrift fiir wissenschaftliche Mikroskopie,’ Band xv, p. 289.

452 J. R. BRADFORD AND H. G. PLIMMER.

flagellum forwards, but it is quite easy for them to move as
quickly with the blunt end forwards. The size and length
of the body vary very much with the period of the disease,
and with the kind of animal in which they are growing.
Measurements made on the full-sized living Trypanosoma
in rat?s blood give a length, including the flagellum, of
25—35 mu, and a width at the widest part of about 2 p.

pB. Examined after Fixation and Staining.—
1. Methods.—In order to study the real structure of, and
changes in, the Trypanosoma, recourse must be had to fixing
and staining, and for both of these processes the ordinary
methods are of no avail. ‘The films must be made as thin
and even as possible, and the method we have found most
effective is to place a small drop of the affected blood on
one corner of a cover-glass or slide, and to spread it very
carefully with a piece of goldbeater’s skin, which is held in
a pair of forceps. The edge of this must be cut quite
straight, and it should be of a little less width than the
cover-glass. It is very important that the film should be
very carefully made, otherwise both fixing and staining will
be imperfect. As regards fixation we have found that
osmic and acetic acid vapours (osmic acid 2 per cent. and
glacial acetic acid equal parts) give the very best specimens ;
but for general work a mixture of formalin ten parts and
absolute alcohol ninety parts, as suggested by Gulland,}
gives very good results. Fixation with this latter formula
is effected in from five to ten minutes, after which the
specimen must be well washed in running water, and then
dried before it is stained. The only stain which we have
found to be of any value in our work upon the Trypanosoma
is that which we indicated in our Preliminary Note; viz.
a modification of Romanowsky’s practical application of an
idea of Ehrlich’s, which consists in mixing an acid and
a basic stain in such proportions that a third body is formed,
which has an extraordinary affinity for chromatin. Our
method of applying this idea is as follows. ‘The stains used

1 * Scottish Med. and Surg. Journ.,’ 1899, p. 312.

THE TRYPANOSOMA BRUCII. 4.53

are methylen blue and erythrosin; the methylen blue used
must be the methylen blue med. pur. (Héchst). A 1 percent.
solution in distilled water is made, and to it °5 per cent.
potassium carbonate is added and dissolved ; the mixture is
then incubated at 37° for forty-eight hours; when cold it is
filtered, and is then ready for use. Instead of using eosin
(under which name many substances of different shades and
of different composition are sold) for the second stain, we
have found it far more certain and effective to use erythrosin
(the tetraiodide of fluorescein), which, as found in commerce,
is of more definite and constant composition and colour than
eosin; a ‘001 per cent. solution is made, to which ‘25 per
cent, formalin can be added in order to prevent the growth
of moulds. Of these two solutions we have found the
following to be the best mixture for demonstrating the
different structures in the Trypanosoma: 20 c.c. of distilled
water are put into each of two beakers, to one of which is
added twenty drops of the erythrosin solution, and to the
other six to eight drops of the methylen blue solution.
These solutions are then quickly mixed, and are immediately
poured into a flat dish, in which the slides or cover-glasses
to be stained have been placed. About twenty minutes
suffices for the differential staining, as shown in the plates,
to take place. The specimens are then washed in distilled
water till no more colour comes away, and are then allowed
to dry in the air. No heat must be used for drying, other-
wise the red colour will entirely disappear. They are then
mounted, preferably in turpentine colophonium, in which
the colours keep, according to our observations, for the
longest time.

Various other modifications of this method of staining have
been suggested by Nocht, Reuter, Laveran, Leishman, and
others, but although they all give the specific reaction, we
have not found any which give so brilliant and clear a picture
as that described.

2. The Appearances of the Adult Trypanosoma.—
When the preparation has been successful the appearances

VoL. 45, PART 3,—NEW SERIES. HH

454 J. R. BRADFORD AND H. G. PLIMMER.

seen are those shown in figs. 1—6. The body of the organism
is stained blue, but not uniformly, there being parts which are
stained faintly, and some hardly at all, which is in confirmation
of the alveolar structure which has been mentioned above.
The macronucleus is stained a pinkish-red, and is, in the
resting form, nearly homogeneous, with only traces of a
reticular structure; it is also sharply defined and extends
right across the short axis of the body. The micronucleus
is stained of a darker red colour, and it always stains more
intensely than any other part of the organism: it appears as
avery small round dot, and we have been unable, with the
highest powers, to make out any further structures in it. It
is generally surrounded by an unstained halo. The flagellum
is stained of a yellowish-pink colour, as well as the edge of
the undulating membrane, which can be seen to end in the
micronucleus. ‘The edge of this membrane is apparently
thickened and merges into the flagellum. In some organisms
the part of the body behind the micronucleus stains of a
darker blue than the rest of the body, and the darker part
may contain a row of faint dots arranged regularly in the
long axis of the organism. ‘These are seen especially in those
animals from which the spleen has been previously removed.
Besides these forms, the first of which constitutes by far
the greater part of those seen in the blood, there are other
adult forms which are seen at certain times and in certain
places. That shown in fig. 5 is seen in the blood of the rat
when the disease is well advanced, about a day or two before
death, and they are also found in the blood of spleenless rats
from the second day; they are also seen in the blood of the
mouse, and in greater numbers especially in the blood from
the lungs taken immediately after death. This organism is
much larger than the ordinary adult form, and is much wider,
often more than double the width, and is more irregular in
shape. ‘The protoplasm is quite homogeneous and much more
delicate, and it stains very faintly with the methylen blue.
The macronucleus is large and roundish in shape, and the
chromatin is more condensed at the circumference, so as to

THE TRYPANOSOMA BRUCII. 455

give it aringed shape ; and it never extends completely across
the short axis of the organism, as is the case in the ordinary
form. The chromatin is in the form of coarse threads or
rods, and there is almost invariably a clear unstained place in
the centre of the nucleus, which may contain a few scattered
dots of chromatin. The micronucleus is very distinct, as are
also the undulating membrane and flagellum. In this form
the thickened edge of the undulating membrane is most
distinct.

Another form, which is also found principally in the blood
from the lungs in the mouse, is that shown in figs. 6—8. In
this, besides the macro- and the micronucleus, there are a
number of dots staining very intensely, but of a much darker
colour than the micronucleus. In the one figured the macro-
nucleus is still visible, but we have found these forms quite
full of these dots, with no visible macronucleus. We thought
at the time of our first paper that this form was that following
conjugation, as a similar process occurs after conjugation in
organisms not far removed biologically from this one, but
this question will be discussed later.

III. Reeropuction or Aputt Forms or TRYPANOSOMA
Broucil.

In our former paper we stated that we thought there
were two modes of multiplication by division of the adult
Trypanosoma, viz. by longitudinal division and by trans-
verse division. We now think that there is only one mode,
that of longitudinal division, and we think that the appear-
ances which we interpreted as transverse divisions were those
of ameeboid forms and not those of division. Besides this
means of reproduction by division, there is that of the forma-
tion of a plasmodial mass, from which amceboid and adult
forms are given off, which will be mentioned later. This form
of reproduction is, we think, of the greatest importance and
interest.

We will take first the method of reproduction by division.

The process of longitudinal division down the long axis of

456 J. R. BRADFORD AND H. G. PLIMMER.

the organism is shown in figs. 11—17. The first step is a
eeneral increase in the size of the organism ; the micronucleus
first enlarges, and then the macronucleus, which becomes
also very distinct (this is well shown in figs. 11—13); the
micronucleus lengthens out and divides into two, to the
second of which a new undulating membrane, which is by
this time developed, is attached (figs. 12 and 17); the macro-
nucleus then divides (figs. 14—17), and a white line is seen
extending down the long axis of the body (figs. 14—16) ;
eventually the organisms separate into two fully formed
Trypanosomata. The order of the division is in general
this :—firstly the micronucleus, then the undulating mem-
brane, then the macronucleus, and lastly the protoplasm.

In fig. 17 is shown the division of one of the large, hyaline
variety of the organism, which has been described above. In
this kind the process is exactly similar to that in the ordinary
variety, only it is much more distinct, on account of the
larger size of the organism, and its clearer staining.

IV. ConsuGATION BETWEEN tHE ADULT ‘T'RYPANOSOMATA
Brvuctt.

In the blood of the rat, and of all the animals we have
examined (mouse, dog’, cat, rabbit, guinea-pig, horse, mule,
pig, spring-bok, goat), forms have been seen, such as those
shown in figs. 18—23, which we think can only be interpreted
as conjugations. The organisms can be seen in the living
blood joined together by their blunt ends, and moving about
in this state by their flagellated extremities, the joined parts
remaining motionless. In stained specimens they are found
as figured, and the only difference in these is the position of
the micronuclei. At first, after the bodies have joined, these
micronuclei are in their normal position at the usual distance
from the blunt end of each organism (fig. 18), then they are
seen to approximate to each other (fig. 19), and then they
get still nearer and nearer (figs. 20 and 21), until they
become fused into one micronucleus (fig. 22). Several other

THE TRYPANOSOMA BRUCII. 457

arrangements of the micronuclei in these conjugating
organisms can be seen, such as a dot in one Trypanosoma
and a rod in the other, or two dots in one and one in the
other. The arrangement of a dividing rod and a dot can
be seen in fig. 20; that of a rod and a dot in fig. 21.

Although it is difficult to definitely associate this junction
of two individuals with any immediate change in form, yet
we think, from the fusion of the micronuclei which takes
place, that it must bea true conjugation: and it is known
that this takes place in allied organisms, e.g. the Bodonine,
which separate afterwards and go on as before, but perhaps
with renewed energy for division.

After the fusion of the micronuclei we have not been able,
notwithstanding the examination of several hundred speci-
mens, to be sure of any constant further stages, but we imagine
that the most hkely occurrence is that the fused micronucleus
lengthens out into the rod form, and then divides into two,
one for each organism, and that then the organisms them-
selves separate. In that case many of the forms which we
have seen, in which there is this rod form of micronucleus,
or in which the two micronuclei are separated, may be those
which have passed the fusion stage and are going on to
separation. In many, but not all, of these conjugating
forms the macronucleus shows signs of activity, i.e. it
enlarges and becomes oval or circular, and the chromatin
becomes aggregated at its margin, and we have thought
that these forms may be those in which conjugation is
ending, and which are about to separate and to go on to
renewed longitudinal division.

In our former paper we stated that we thought the
dotted forms described above (figs. 6—8) might be those
immediately following conjugation. But we have no
evidence of this; and moreover, with very careful staining
and illumination, these dots are seen to be of a little
different tint to the micro- or macro-nucleus. Sometimes
they are of a quite dark purple colour, instead of the
red nuclear colour, and they are moreover of very varying

458 J. R. BRADFORD AND H. G. PLIMMER.

size. We have never observed conjugation taking place
between the large hyaline forms described above.

V. “ Ama@porp” anp®* PrasmopiAL” Forms OF THE
TRYPANOSOMA Bructt.

We do not wish or mean to attach any particular biological
meaning to these two terms, which we use here only for
convenience of description. Besides the forms described
above, there remain two other forms which practically seem
to be of the greatest importance of any, as the death of
many of the animals we have used seems to depend upon
their presence.

The first of these forms is that which for convenience
we have called “amceboid.” By this we mean a form
consisting of a small irregularly shaped mass of protoplasm,
with a micro- and macro-nucleus, and with or without a
flagellum ; they are all very mobile when examined in the
living state. This form is depicted in figs. 24—37. The
protoplasm is very delicate and uniform, and the micro-
und macro-nuclei are small, but very distinct; and this
form is only seen well in stained preparations when the
differential stain we have described above is used. But
they can be seen in fresh unstained preparations with very
careful critical illumination; in fact, our first observation
of them was ina fresh unstained preparation. ‘The flagella
may or may not be seen, but we believe that they are,
when free, all flagellated, but that in the preparation of
the film the flagella may become detached, as free flagella
attached to their micronuclei are seen in most preparations
in which these amoeboid forms can be found. ‘They are
found in the blood from the lungs of most of the animals
we have used, especially in that of the rat, mouse, cat,
and dog, where they are present in enormous numbers ;
they are also found in the glands in small numbers, and
in the bone marrow of animals from which the spleen has
been removed before inoculation. They may also be

THE TRYPANOSOMA BRUCII. 459

demonstrated in the blood of spleenless animals, also in
that of the more refractory animals, such as the rabbit
(as shown in figs. 9 and 10) and guinea-pig, when
repeated examination of the blood fails to reveal any of
the ordinary adult forms. They are also present in the
spleens of all the animals we have examined in greater
or less number. Another place where they are specially
found—and this is of importance from the clinical side—
is in the capillaries of the brain and medulla. Most of
our animals have died with symptoms referable to the
nervous system; either they become comatose and die in
a few hours, showing no symptoms till the coma appears,
or they have fits of an irregular character, or they become
paralysed in one or more limbs. Many of the cerebral
capillaries can in these cases be found to be quite plugged
up with these amceboid forms, as is shown in figs. 40
and 41. The similarity of these appearances with those
of the cerebral capillaries in pernicious malaria is very
striking. In order to demonstrate these forms either in
the lung, brain, or elsewhere, films must be made from
the organs directly after death, and treated with the same
stains as described for blood. We have not found yet
any method of fixation of pieces of tissues fine enough
to demonstrate them in sections of these organs.

These amoeboid forms can be seen in all stages of division,
and they appear to multiply much more rapidly than the
adult organism. Stages can also be seen intermediate
between these and the adult Trypanosoma. '‘he method
of division corresponds closely with that of the adult
organism. First the micronucleus becomes rod-shaped, and
then divides; the macronucleus enlarges at the same time
and divides a little after the micronucleus, and lastly, the
protoplasm divides. Many irregular forms are seen, espe-
cially inthe bone marrow, resulting from the irregular division
of these amceboid forms.

Besides these there is another form, which we have, also
for convenience, called a “ plasmodial” form. This is seen

4.60 J. R. BRADFORD AND H. G. PLIMMER.

in figs. 37—39. It seems to consist of a protoplasmic mass,
containing a very large number of micro- and macronuclei
irregularly embedded in its substance ; in the central part
there are no divisions to be seen in the mass of protoplasm
surrounding the nuclei, but at the marginal parts amceboid
forms can be seen in process of detachment, such as is shown
in fig. 837. These masses are best seen in the spleen of the
rat or mouse, but can also be observed in the blood of
animals from which the spleen has been removed. They can
be seen also in less quantity in the blood from the lungs of
most animals. In the bone marrow of spleenless animals
this form can also be demonstrated. At present, it does not
seem possible to say for certain whether this form is a true
“ plasmodium,” i.e.a fusion of several elements, or whether
it is merely an aggregation of amoeboid forms. Our opinion
is in favour of the former view. This form is of the greatest
importance, as its presence not only causes the great enlarge-
ment of the spleen observed in certain animals, bui is also a
means of very rapid multiplication of the organism.

VI. Tue Disrrisution or THE TRYPANOSOMA Brvcir; ITs
VARIATIONS IN DirrFeRENT ANIMALS, AND THEIR RESISTANCE
TO IT.

The distribution of the Trypanosoma in the different
animals we have used varies considerably. In the rat and
mouse—animals that die in the shortest time —the organism is
found in the blood from forty-eight hours after inoculation,
and they go on steadily increasing till at the time of death,
usually from six to nine days, there may be as many as
3,750,000 per cub. mil. These are nearly all of the adult
form, and may be seen in all stages of longitudinal division,
and in conjugation. The spleen becomes, towards the end of
the disease, enormously enlarged, measuring about 5 to 6 em.
by 1 to 2 cm., and itis found to contain an enormous quantity
of plasmodial material, as seen in fig. 38, some amoeboid forms
and a very few adult forms. The blood from the lungs con-

THE TRYPANOSOMA BRUCII. 461

tains avery large number of amceboid forms, a few plasmodia,
and a few adult forms. The liver, kidneys, and bone marrow
only contain adult forms in quantity corresponding with
those in the blood. The glands contain many ameeboid
forms, and a fair quantity of adult forms. The organism is
first found in the gland nearest to the point of inoculation, in
eighteen to twenty-four hours after inoculation. The cerebral
capillaries at the time of death are in great part blocked
with amoeboid forms (figs. 40 and 41). In the mouse the
appearances are quite similar, but the number in the blood
never reaches the number found in the rat. In the dog and
cat the spleen is enlarged and contains amoeboid and plas-
modial forms, but it is not so large relatively as that of the
rat. Their blood does not become infectious until the fourth
day after inoculation, and never contains such quantities of
the organisms as does rats’ blood. The glands in both cat
and dog are always affected, as in the rat, and contain both
amoeboid and adult forms; the bone marrow also contains
many adult and ameeboid forms. In the rabbit the organ-
isms are only found at irregular intervals in the blood, and
then only in very small numbers. The glands are not
enlarged at the time of death, which may be as long as three
months from the date of inoculation. The spleen is not
enlarged, and only contains a few amceboid forms; the bone
marrow is almost entirely free. In this animal the chief
obvious lesions are a swelling of the eyelids with a slowly
progressive panophthalmitis, and a soft cedema of the genitals.
These also occur, but less markedly, in the dog and cat.
It begins always as a plugging up of the lymphatics,
which contain numbers of the organisms, principally of the
amoeboid form, so the disease in these animals seems princi-
pally to affect the lymphatic system. ‘The goat shows very
little sign, and the organism is not found abundantly in the
blood, but the animal gets cedema of the genitals, and the
eyes become somewhat opaque. It dies in about two months
after moculation, with paralysis. The spleen is not enlarged ;
the nasal mucous membrane becomes swollen, and interferes

4.62 J. R. BRADFORD AND H. G. PLIMMER.

with breathing. In the horse and mule the organisms are
constantly found in the blood, and the spleen is enlarged
and contains amoeboid forms. The sheath becomes swollen,
and the abdomen cedematous, and they die in about six or
eight weeks with paralysis. ‘I'he Trypanosoma is longer and
thinner in these animals than in any we have used. The
guinea-pig shows the organism in the blood in varying
numbers and at varying periods; sometimes there are none
to be found for long periods, and again they may be nume-
rous for a time. ‘The spleen is not enlarged, and the bone
marrow contains no organisms. Amoeboid forms can be
found in the blood from the lungs. The guinea-pig may live
as long as eighteen weeks after inoculation. The pig shows
the organism in the blood very rarely and in very small
numbers, and dies with pulmonary symptoms. ‘The spleen 1s
not enlarged. The spring-bok dies in about four weeks
with nervous symptoms, and with the organisms present in
the blood.

It will be seen from the above that there is a good deal of
variation in the distribution of the ‘Trypanosoma, and in the
form in which it is mostly present in the different animals
The rat and mouse show the greatest number and die quick-
est ; the guinea-pig and rabbit show fewer organisms, with a
small number of the amoeboid forms, and take much longer
to die; the goat shows the least sign and is, in our experi-
ence, the most resistent animal. In those animals which live
longest the amoeboid and plasmodial forms are few in number
or absent. In certain animals there is a tendency to blocking
up of the lymphatics of certain regions (e.g. eye, sheath,
abdomen, and legs), and of lymphatic effusion around ; and
in others, a tendency to blockage of the brain capillaries
with amoeboid forms, which, in most animals, independent of
variations generally, is the proximal cause of death.

In animals (rabbit, cat, dog, rat) from which the spleen
has been removed the distribution of the organisms is
different. During life there can be seen in the blood tangles,
consisting of numbers of the adult organisms writhing about

THE TRYPANOSOMA BRUCII. 463

in close apposition to each other. ‘hen the outlines of the
individual organisms become indistinct and they apparently
fuse together, so as to form a plasmodium, consisting of a
mass of protoplasm containing micro- and macro-nuclei,
similar to that seen in the spleen and in blood from the lungs
of rats and mice. It is principally on account of this
appearance in the blood of spleenless animals that we think
it possible that the formation of this plasmodial form is due
to fusion of adult forms, and a similar process also occurs in
related organisms. The detachment of flagellated amoeboid
forms from these masses can be always seen. ‘The amoeboid
forms are always present in the bone marrow and glands of
spleenless animals.

None of the animals we have used have been found to be
immune against the Trypanosoma Brucii. They have all
died at varying periods, from five days—which is the shortest
time in our experiments—in the case of the rat, to eighteen
weeks—which is the longest period—in the case of the guinea-
pig. But there is probably in all animals some attempt at
resistance, and this, so far as we have seen, is by phago-
cytosis. We have only observed phagocytosis of the amoeboid
form of the organism, and never of the adult forms. We
have seen this occur in the peritoneal fluid of the guinea-pig,
in the spleen of the rat and mouse, and in the blood of the
spleenless dog and cat. We have, in figs. 44 and 45, shown
phagocytosis occurring in the blood of a spleenless cat, in
which animal it can, in our experience, be most easily
observed. This cat died in twelve days, whereas the average
time the normal cat lives after inoculation is about twenty-
five days. All the animals from which we have removed the
spleen previous to inoculation have died in a shorter time
than the normal animal, so that probably there is, in the
earlier stages of the disease, a good deal of phagocytic action
taking place in the spleen.

4.64, J. R. BRADFORD AND H. G. PLIMMER.

VIL. Tut DIFFERENCES BETWEEN THE T'rYPANOSOMA Bructt
AND THE ‘I'RYPANOSOMA FOUND IN Sewer Rats.

The distribution of the Trypanosoma Lewisi amongst
sewer rats seems to be very variable, the organism being
prevalent in some districts and not in others. For instance,
we first examined six sewer rats from the south of London
with negative results; then twelve from the north of London,
of which number five contained the Try panosoma Lewisi;
then twenty-four from the south of London again with nega-
tive results.

There are the following points of difference between the
sewer rat ‘l'rypanosoma and that found in Nagana. The
former, the Trypanosoma Lewisi, are a little shorter and
somewhat thinner than the Trypanosoma Brucii, and the
posterior end is much more pointed, as is shown in fig. 46,
The micronucleus is placed transversely as a rule, and is
larger. ‘lhe macronucleus is placed at the end of the body
of the organism, instead of in the middle, as in the ''ry pano-
soma Brucii. The protoplasm of the Trypanosoma Lewisi
is not so homogeneous as that of the Nagana l'ry panosoma.
We have not observed any such forms as we have described
as conjugations, nor any amceboid nor plasmodial forms in
the blood of the sewer rat; apparently in this organism
multiplication by longitudinal division only seems to be the
rule. If the Trypanosoma Lewisi be examined in the
living blood under critical illumination, three dark spots can
be seen in nearly all the organisms, apart from the nuclear
structures; one is very near the micronucleus, and the other
two between this and the macronucleus, on that side of the
body opposite the undulating membrane. These are not visible
in the stained specimens. The Trypanosoma Lewisi does
not protect the animal containing it in the smallest degree
from the Trypanosoma Brucii; sewer rats naturally
infected with the former were inoculated with the latter, and
died in the usual time,

THE TRYPANOSOMA BRUCII. 465

Laveran! states that the Trypanosoma of Nagana has
the same structure as that of the sewer rat Trypanosoma
but that it is a little longer. This latter fact is quite correct,
but it will be seen from the above that there are other very
noticeable differences of structure between them.

VIII. Tue Micronvucteus.

There seems at present to be much uncertainty as to
what the exact zoological title of this body should be, to
which we gave the above provisional name in our first
paper.

In the article by Laveran on the multiplication of the
sewer rat Trypanosoma, which we have mentioned above,
he uses the term blepharoplast for this body, as being
the most accurate one. This term, he says, was employed
by Webber to designate analogous bodies in vegetable
cells. This term is used to express a body having the
reactions of chromatin, to which an undulating membrane,
or flagellum, is attached, or from which they take rise.
Rabinowitch and Kempner? called this body a nucleolus,
and Wasielewski and Senn® gave it the name of Geissel-
wurzel (root of flagellum). Henneguy* has stated in an
earlier paper that the blepharoplast is of the nature of a
centrosome, i.e. a centrosomal structure to which a flagellum
is attached, or from which it arises.

From our observations on the Nagana ‘Trypanosoma
we are obliged at present to adhere to the provisional title
of micronucleus. We have seen it apparently coming off
from the macronucleus, as shown in fig. 25, and we have
seen it fuse with the corresponding body in another
organism, as we have described above under the heading

1 “Sur le mode de multiplication du Trypanosome du rat,”’ Laveran et
Mesnil, ‘Comptes rendus de la Société de Biologie,’ tome lii, No, 35.

2 «Zeitschr. f. Hygiene u. Infectionskr.,’ 1899.

3 Tbid., 1900.

4 *Arch. d’Anat. Microsce.,’ 1897.

4.66 .J. R. BRADFORD AND H. G. PLIMMER.

of conjugation. This seems to us a very definite reason
for keeping the title we have used, which would correspond
to that given to a similar body in organisms not far
removed from these, in which conjugation is well known.
It has this fact in favour of its being a centrosome, that
it nearly always divides before the macronucleus does in
the longitudinal division described above. It is also
present through all the other stages of the Trypanosoma,
in the amceboid and the plasmodial. Guignard’ has come
to the conclusion that the bodies generally called centro-
somes have very variable characters, and that the so-called
blepharoplasts can be grouped under this name. So that it
is, after all, a question of terms, to which various meanings
are given by different writers. The main point which
induces us at present to adhere to the term micronucleus
is the behaviour of this body during the process of con-
jugation. Laveran apparently has not observed this stage
in the Trypanosoma Brucii, hence his insistence on
the term blepharoplast, or centrosome,

IX. ‘ae Lire-sistory oF THE TRYPANOSOMA BRvuctl.

The life-history of this organism appears to be much
more complicated than that of the sewer rat Trypanosoma,
as described by Laveran and others, on account of the
presence of forms which we have described under the
headings of Conjugation, and Amceboid and Plasmodial
forms, which do not appear to be present in the sewer
rat when infected with the Trypanosoma Lewisi.

In our previous paper we placed those forms which are
shown in figs. 6—8, and which are described in section
II—2, as those following conjugation, supposing that
these forms broke up into the amceboid forms, and that
these again fused-to form the plasmodia. Since that time
we have observed these forms with great care, and we
have been unable to find any definite intermediate stages

1 «Ann. des se. nat. botan.,’ 1897.

THE TRYPANOSOMA BRUCII. 467

between them; and we are, moreover, very doubtful
whether the dots in these organisms are really chromatin,
as we have stated when describing them, since they stain
a much darker colour. The probability seems to be that
they are not. We are now of opinion that the following
would be the more probable sequence of the life-history
of the Trypanosoma Brucii.

1. Longitudinal division, which is very common, and can
be seen more or less in the blood of all the animals we have
used.

2. Conjugation, the essential of which is the fusion of
the micronuclei of the conjugating pair of organisms.
After this process the organisms probably separate and
go on as before, probably with renewed energies. This
is known to be the case in some nearly allied organisms.

3. We are here inclined to place tentatively the fusion
of the adult forms. ‘This process begins by the formation
of tangles, aud the organisms then lose their individual
forms, and fuse into a more or less homogeneous mass,
which under the best optical conditions appears to be an
irregular aggregation of protoplasm containing many micro-
and macro-nuclei.

These latter divide again and again, so that the mass
seems to consist of little else but micro- and macro-nuclei ;
then flagellated amceboid forms are given off from the
margin, and these grow and eventually lengthen out into
the adult organism. ‘There is a tendency for these masses
to be formed in certain structures only: in the rat, mouse,
cat, and dog they are found in the spleen, and in the blood
just before death. ‘They are also found in the blood from
the lungs, in the lymphatics of the eyes and genitals, and
in the cerebral capillaries.

A point in favour of the above opinion is the fact that
in animals from which the spleen has been previously
removed, these masses are found in all stages of formation
in the blood during life, and before the animal has begun
to get apparently ill; the spleen being removed, in which

468 J. R. BRADFORD AND H. G. PLIMMER.

this fusion process generally occurs, it takes place in the
blood itself. In these spleenless animals the bone marrow
is always affected, and always contains quantities of the
amoeboid form, which is not ordinarily the case in normal
animals.

As regards the life of the Trypanosoma outside the
body there is at present little to say. Laveran succeeded
in keeping them alive in blood kept at 0° C. for as long
as three months. We have kept them alive in thin films
for six days after removal from the body. We have also
kept a large quantity of blood containing them in an
atmosphere of oxogen, and have found that, although the
adult forms soon disappeared, with the formation of tangles,
plasmodial masses, and then amoeboid forms, the blood
was infective for three days at least, at which time our
experiment came to an end.

In the ordimary way the blood of animals loses its in-
fectivity in a few hours after death, as decomposition sets
in with great rapidity in this disease.

X. EXPLANATION OF PLATES 24 AND 25,

Illustrating Messrs. J. R. Bradford and H. G. Plimmer’s paper
“On the Trypanosoma Brucii, the Organism found
in Nagana or 'se-tse Fly Disease.”

All these figures were drawn under a Zeiss 38 mm. apochromatic objective
of N.A. 1°40, used with a Zeiss achromatic condenser of N.A. 1:0, and with
compensating oculars of various powers, which will be indicated when referring
to each figure, or set of figures, separately. The specimens were all stained
by the method described in the paper.

Fires. 1—4 (Oc. 8)—Ordinary adult organisms from rat’s blood on fourth
day, stained lightly to show the general shape and structure. a. micronucleus ;
b. vacuole; ¢. macronucleus.

There are three red corpuscles, drawn with Figs 1—9 from rat’s blood
under Oc. 8, to show the relative size of the organism.

THE TRYPANOSOMA BRUCII. A469

Fie. 5 (Oc. 12).—Large hyaline form, stained deeply to show the undulating
membrane and its connection with the micronucleus. From rat’s blood, eighth
day.

Fies. 6—8 (Fig. 6, Oc. 8; Figs. 7 and 8, Oc. 12)—Forms containing
granules, from blood from mouse’s lung just after death on seventh day. In
Fig. 6 the macronucleus is still distinctly visible, but in Figs. 7 and 8 its
position is only just indicated. What happens further to these forms has not
been traced.

Fics. 9 and 10 (Oc. 12).—Irregular ameeboid forms from blood of rabbit
on twenty-ninth day. In Fig. 9 the edge of the undulating membrane is
seen attached to the micronucleus, and then nearly surrounding the organism.
In Fig. 10 the flagellum and undulating membrane are quite developed and
distinct.

Fic. 11 (Oc. 12).—Full-sized organism from rat’s blood, sixth day, stained
deeply to show the undulating membrane and its attachment to the micro-
nucleus ; and also commencing enlargement of the macronucleus.

Fie, 12 (Oc. 12).—From rat’s blood, seventh day, showing the early stages
of longitudinal division. The micronucleus has divided, and a second undu-
lating membrane has developed. The macronucleus shows commencing
enlargement.

Fic. 13 (Oc. 12).—From same specimen, showing the division and separa-
tion of the micronucleus, and the great enlargement of the macronucleus.

Fie. 14 (Oc. 12).—From guinea-pig’s blood, thirty-ninth day, showing the
division of the micro- and macro-nuclei, and the line of division along the long
axis of the organism.

Fie. 15 (Oc. 12).—Krom same specimen as Figs. 12 and 18, showing com-
plete separation of micro- and macro-nuclei, with the line of division very well
marked,

Fie. 16 (Oc. 12).—From another rat’s blood, seventh day, showing two
complete organisms just at point of complete separation.

Fie. 17 (Oc. 12).—From mouse’s blood, sixth day, showing division of the
large hyaline form, in which the different structures are more clearly seen.

Fic. 18 (Oc. 12).—F rom rat’s blood, fourth day, showing the conjugation
of two adult organisms, with the micronuclei at some distance apart.

Fie. 19 (Oc. 12).—From spleenless dog’s blood, ninth day, showing con-
jugation with the micronuclei closely approximated.

Fre. 20 (Oc. 12).—From the same specimen, showing conjugation in which
one micronucleus is in the dumb-bell form which it has in the early stages of
longitudinal division of the organism, and the other is seen as a round dot.

Fie, 21 (Oc. 12).—From blood of a spleenless cat, showing conjugation in
which one micronucleus is in the form of a rod, and the other in the form of
a dot.

VOL. 45, PART 3,—NEW SERIES, TI

470 J. R. BRADFORD AND H. G. PLIMMER.

Fie. 22 (Oc. 12).—From same specimen, showing conjugation completed,
in which the two micronuclei are fused into one large dot.

Fig. 23 (Oc. 12).—From blood of rat, eighth day, showing an unusual form
of conjugation in which the central micronuclei seem to have disappeared,
whilst two others are being detached from the macronuclei.

In all these figures, from 18 to 23, the macronucleus seems to be in a state
of activity, as shown by its alteration in shape, the chromatin being principally
arranged peripherally.

Fies, 24—27 (Oc. 18).—Ameeboid forms from the bone marrow of a spleen-
less dog, twenty-one days. The flagella are not visible in any of these, but
probably were broken off in the process of preparation. In Figs. 25 and 27
division is taking place, and the macronuclei in all appear to be in a state of
activity. In Fig. 26 the organism is lengthening out, as if it were becoming
an ordinary adulf form.

Fires. 28—31 (Oc. 18).—From the bone marrow of a spleenless cat, show-
ing amoeboid forms. In these the flagella are visible, and all the organisms are
undergoing division. In Fig. 3] there appears to be a complicated process of
division, the smaller organism at the upper part appearing to have four micro-
nuclei and (wo macronuclei, as if preparing for some such form as is shown in
Fig. 31 a.

Fie. 31 a (Oc. 18).—From blood of a spleenless dog, eighth day, showing
four adult organisms being given off from an ameeboid mass.

Fries. 32, 33 (Oc. 18).—From bone marrow of a spleenless dog, showing in
Fig. 82 a simple ameeboid form, and in Fig. 33 one in which the micronucleus
is divided.

Fre. 34 (Oc. 18).—F rom blood from lung of mouse, seventh day, showing a
flagellated amceboid form full of granules, similar in colour to those seen in
certain adult organisms, as shown in Figs, 6—8.

Fie. 35 (Oc. 18).—From same specimen, showing a dividing flagellated
amoeboid form.

Fre. 36 (Oc. 18).—From the same specimen, showing an amoeboid form
lengthening out, as if becoming an ordinary adult form.

Fic. 87 (Oc. 18).—From blood from lung of rat, ninth day, showing a
plasmodial mass in the centre with micro- and macro-nuclei scattered about in
a mass of protoplasm, and at the outside a few amceboid forms being detached.

Fig. 88 (Oc. 12).—Smear from spleen of rat, eight days, showing the plas-
modium, with its numbers of micro- and macro-nuclei wedged in between the
splenic cells. A few red corpuscles are also drawn.

Fre. 39 (Oc. 12).—From a scraping from the lung of a mouse, seventh day,
showing a plasmodial mass with cells from the lungs which are accidentally
present as the result of the method of preparation,

HE TRYPANOSOMA BRUCII. 471

Fre. 40 (Oc. 18).—Small branched capillary from a smear from cortical part
of brain of rat, showing the lumen completely blocked with amceboid forms.

Fie. 41 (Oc. 18).—Small capillary from smear from medulla of rabbit,
showing the lumen more completely blocked with ameeboid forms.

The large red bodies in Figs. 40 and 41 are the nuclei of the capillary
walls,

Fie. 42 (Oc. 18).—Froi blood of spleenless cat, twenty-first day, showing
the formation of tangles; the central part is already quite like the plasmodium
shown in l’igs. 38 and 39, containing micro- and macro-nuclei.

Fic. 43 (Oc. 12).—From same specimen, showing a plasmodial mass with
flagella around margin; probably the last stage of the formation of a small
plasmodium.

Fras. 44, 45 (Oc. 18).—Leucocytes from blood of a spleenless cat, showing
phagocytosis. In Fig, 44 two entire amceboid forms with the micro- and
macro-nuclei can be seen, and at lower end a still undestroyed micronucleus,
and also two nearly colourless bodies, probably the last stage of destruction
of the organism. In Fig. 45 two ameeboid forms are seen in fair condition,
the one above having still its flagellum attached to it. ‘There is also a larger
body with a dot in it, which is probably a destroyed organism, with only the
micronucleus left.

Fie. 46 (Oc. 12).—Trypanosoma Lewisi from blood of sewer rat, to
show the gross differences between these and the Trypanosoma Brucii.

nA 7 aa 7 % aopigiilinad a) Te ay ee eels
& “ : . see : :
2 ne myn wr Aa 6 YS or tee ete f
a tes ci ey ie 7k > bk T B fiers Pi. a as iki £44 fered ee
* } i: ll et sor iad .|
a s % 3 J
nat : pes
: \ 7 . ie —
\ - re i (ee en
a 7 ~ Fe 7 ey Lk
La a - v 4
. = - 2s
| hie) arta 2

« an wb
_. a
7 : |. J. i
; : 7 a eh

i.

i Seeger ea ari Maat Pere

bis ant: #4) ine al Sele FTE eee

| 7
hw * . -
, ri « ae
m ; 27 x BoE . 2 cw
x “
- = Fi “- a]
i $ {
,
}
t
4
t
~~
.
”
~
-

Quant. Fourn. Mur Sea Vol 4ENSGA. 24.

AS Huth, Lith® London.

%

Quart Fourn.Nucr Sea Vol £2

ae at

ate SRE A

——

(oy)

NOTES ON AUTINOTROGHA, 47

Notes on Actinotrocha.
By
K. Ramunni Wenon,

Assistant Professor, Presidency College, Madras.

With Plate 26.

Tue following notes on Actinotrocha (a careful comparison
of which with Brachiopod larve seemed desirable) may be of
interest in connection with Masterman’s recent investigations
on this larval form. I wish to express here my deep indebt-
edness to my kind teacher, Professor Bourne, for affording
me every facility and help for working in his laboratory.

The larvee which were caught during the latter part of
October last belong to three different species, none of which
I have been able to identify. One of these, which is very
rare (owing, perhaps, to its being only an occasional immi-
grant into the surface water), has a short, thick body, short
tentacles, a small hood, and a strong perianal band, and, if
the conspicuous broad pigmented zone round the body may
be regarded as affording a clue, probably belongs to
Phoronis australis. The second species is much longer,
has longer tentacles, a well-developed hood, and a transparent
body-wall. My observations are based mainly on this species.
The third species is much smaller than the second, and but
for the fact that it metamorphoses directly into the young
Phoronis, would be regarded as a young stage of the second.
The larvee were preserved in corrosive sublimate, osmic acid,
and chromo-aceto-osmic mixture, cut in celloidin and stained
with hemalum.

474 K. RAMUNNI MENON.

The metamorphosis of the larva is usually completed in a
few minutes. Occasionally, however, the ventral diverticulum
is evaginated, and it is only after a considerable interval that
the alimentary canal is drawn into it—the animal then pre-
senting a curious external resemblance to Rhabdopleura, as
far as one can conceive of it from figures. The young
Phoronis fixes itself to the bottom of the glass vessel by
means of a sticky secretion, and also secretes a transparent
cuticular tube round the body. Observation of the living
animal shows that the annulation of the body, said to be due
to contraction in spirit, is quite natural. These annuli are
formed by the ectoderm being raised up into ridges. If the
water in which they are kept is not changed daily, the young
animals lose their heads—a fact in agreement with Cori’s
observation that the tentacle crown is often thrown off by
Phoronis, especially under unfavourable conditions.

Epistome.—This structure is usually regarded as a
remnant of the preoral lobe of the larva. In the young
Phoronis, soon after metamorphosis, the epistome is not
present. The preoral lobe and the tentacles of the larva are
apparently wholly taken into the alimentary canal during the
metamorphosis. In the case of slow metamorphosis these
persist for a comparatively long time, and their disappear-
ance must be due to disintegration. In the young Phoronis

the dorsal part of the region surrounding the mouth and

immediately within the ring of tentacles is raised up into a
round protuberance (fig. 1). This is the beginning of the
epistome. The collar cavity, which was encroached upon by
the masses of blood-corpuscles in the old Actinotrocha, does
not extend into this swelling, the interior of which is occupied
by a nucleated mass containing irregular spaces, and with
scattered muscle-fibres and blood-corpuscles. The surface
epithelium does not differ from that covering the rest of the
oral region. The papilla gradually becomes flattened out
into a broad flap (figs. 2—4), having a thick base attached
along the dorsal edge of the mouth, and a thin free edge.
The internal mass is absorbed ; and between the basement

NOTES ON ACTINOTROCHA. 475

tissues of the two surfaces are found small spaces separated
from one another by bridges of basement tissue with a few
nuclei. The epithelium covering the oral surface consists of
columnar ciliated cells with long nuclei, while over the aboral
surface it consists of cubical cells; along the free edge the
cells are longer and the nuclei more closely arranged. The
collar ccelom, which is now a definite space having a well-
marked ccelomic epithelium, extends only into the base of the
epistome, and is separated from the peripheral spaces by a
mesentery. Further extension of the celom into the epi-
stome must take place in later stages, as the cavity of the
epistome is stated to be ccelomic. In considering the
homology of the epistome it must also be remembered that
the cavity of the preoral lobe of the larva, of which the
epistome is considered to be a remnant, is, whatever its nature
may be, quite distinct from the collar cavity, from which it is
separated by a well-marked mesentery. The occurrence,
though rare, of old Actinotroche without any trace of the
preoral lobe is also interesting in this connection. Thus
there can be no reason to doubt that the epistome is not a
remnant of the preoral lobe, but is a new structure, developed
as an outgrowth of the collar region.

The most important points adduced by Masterman, in
establishing the Chordate affinities of Phoronis, have refer-
ence to the following structures in Actinotrocha :—(1) five
body-cavities ; (2) proboscis pores, collar pores, and trunk
nephridia; (3) two notochords; (4) subneural gland (hypo-
physis), which is homologised with the proboscis vesicle of
Balanoglossus; and (5) the tubular nerve-ganglion. These
structures will be dealt with in the order given.

(1) Body-cavities.—There are three principal cavities in
Actinotrocha. It will be convenient to adopt Masterman’s
names, and to call them the cavities of the preoral lobe,
collar, and trunk. These spaces are separated from one
another by mesenteries. The mesentery between the preoral
cavity and the collar cavity can be roughly compared to a
watch-glass with its edge attached to the body-wall, and with

4.76 K. RAMUNNI MENON.

its convex side directed towards the preoral cavity. The
cavity of the preoral lobe is traversed by protoplasmic
strands, some of which cross it from the upper wall to the
lower, while others run parallel to its free border. A definite
coelomic epithelium cannot be made out in this cavity, the
only structures clearly seen beneath the ectoderm being a
thin basement membrane (chondroid tissue), with a few
isolated patches of nuclei adhering to the inner face of the
membrane. Scattered blood-corpuscles and ccelomic cor-
puscles are occasionally met with in this cavity, the nature
of which it is thus difficult to decide without embryological
evidence.

Below the nerve-ganglion which les in the root of the
preoral cavity, just in front of the attachment of the
mesentery above mentioned, there is a small space, the
posterior wall of which is formed by this mesentery, which
thus separates it from the collar cavity, and the anterior wall
by another mesentery stretched between the upper and lower
walls of the preoral lobe, which thus separate it from the
rest of the preoral cavity. ‘This is the subneural sinus, and
the nerve-ganglion forms part of its roof. It is a completely
closed vesicle, and the epithelium forming its wall can be
traced along the roof of the sinus immediately below the
nerve-ganglion. The nerve-ganglion thus really lies outside
the cavity. here are two muscle strands developed in con-
nection with the wall of this cavity, and attached to the
dorsal body-wall near the ganglion. ‘The subneural sinus
does not communicate with the dorsal blood-vessel; nor does
it contain any blood-corpuscles. Masterman compares this
sinus to the central blood-sinus of the proboscis of the
Knteropneusta. If a comparison is to be made between it
and any structure in the Hnteropneusta, its complete epi-
thelial wall, and the muscular development in connection
with its wall, strangely support Harmer’s view that it is the
“ heart vesicle.”

The collar cavity is bounded posteriorly by the mesentery
between it and the trunk cavity, which is attached to the

NOTES ON ACTINOTROCHA. 477

body-wall some distance posteriorly to the bases of the
tentacles, and runs forwards and inwards to meet the wall of
the gut. There is a well-marked epithelium lining the
cavity. Dorsal and ventral mesenteries do not appear to be
present. ‘he muscle strands above mentioned run through
this cavity. Hach divides into two bands, one of which runs
through the collar cavity and meets the dorsal body-wall, and
runs posteriorly along one side of the median line; the other
runs ventrally, externally to the cesophagus, meets the
ventral body-wall and then runs posteriorly along one side
of the median line. he collar cavity, which is well developed
in the young larva, becomes, in the later stages, obliterated
by the development of a large mass of blood-corpuscles.
‘These are developed around the ventral diverticulum of the
stomach, and in the old Actinotroche often fill the collar
cavity. ‘There is no epithelium between these corpuscles,
which form a mass adhering to the wall of the stomach and
the collar cavity, and there cannot be any doubt that they lie
freely exposed to the collar cavity (fig.6). This explains
Caldwell’s statement, often disputed, that free communication
exists between the body-cavity in front of the septum (1. e.
the collar cavity) and the vascular system. ‘The blood-
corpuscles do not, however, float in the fluid in the collar
cavity, so that there is no communication between the collar
cavity and the vascular system in the sense that blood flows
into the collar cavity and mixes with the ccelomic fluid. The
mass of corpuscles is connected with the dorsal blood-vessel
which runs along the dorsal wall of the stomach posteriorly.
The mass of corpuscles breaks up during metamorphosis, and
the corpuscles are apparently withdrawn into the blood-
vessels. With regard to the origin of the corpuscles, it may
be said that they are formed from the splanchnopleure (fig. 5),
covering the stomach and its diverticulum, the cells of which
proliferate and become blood-corpuscles. Blood-corpuscles
are also found between the splanchnopleure and the stomach
wall, having been found, apparently, in the same manner.
It may be mentioned here that, according to Cori, blood-

478 K. RAMUNNI MENON.

corpuscles are formed in the adult Phoronis from the endo-
thelium of the blood-vessels.

‘To return to the body cavities—the trunk cavity is divided
into right and left halves by a complete ventral mesentery
and by a dorsal mesentery, which is only represented by a
small strand at the posterior end. The somatopleure is
separated from the ectoderm by a thick layer of chondroid
tissue (also developed in the collar, though not to the same
extent), from which it is detached in preparations, and often
thrown into irregular folds. At the anterior boundary of the
perianal band the somatopleure leaves the body-wall and
runs inwards and backwards to meet the gut wall, and become
continuous with the splanchnopleure. There is thus left a
space surrounding the posterior part of the intestine and
between the coelomic epithelium in front, the ectoderm on
the outside, and the endoderm on the inside. The ectoderm
forming part of the wall of this space is highly vacuolated,
the vacuoles being large and regularly arranged, and the cell
substance is reduced to an external and an internal limiting
layer and strands between the vacuoles. ‘The nuclei of these
cells are arranged in the external and internal layers, but
mainly in the former. The arrangement of the parts will
be made clear by fig. 7. The space under the perianal band
thus appears to have a lining of its own. If the coelomic wall
of this space were thrown into a fold projecting into the
trunk coelom we could get a section (without the internal
opening, which may be artificial) pretty much lke the one
shown in Masterman’s drawing of the posterior nephridia. 1
venture to suggest this as a possible explanation of the
posterior nephridia, which I have not found as such, especially
as Masterman himself seems to be doubtful about them.

(2) Nephridia.—Masterman describes for his Actino-
trocha proboscis pores, collar nephridia, and trunk nephridia.
I have carefully looked, both in the living larva and in
sections, for the cellular tubes and the pores which Masterman
homologises with the proboscis pores, but have found no
trace of them. ‘There is nothing to be added to what has

NOTES ON ACTINOTROCHA. 479

been already said about the posterior nephridia. The
Actinotrocha studied by Masterman seems worthy of rein-
vestigation, if for no other reason than to demonstrate the
proboscis pores and the posterior nephridia. There are two
nephridia in Actinotrocha, ‘They are placed on the mesentery
between the collar and trunk cavities, on the face of it turned
towards the collar cavity. Wach nephridium opens to the
exterior ventro-laterally behind the attachment of the
mesentery. From its external opening the duct of the
nephridium runs forwards, crosses the mesentery, and then
bends dorsally and runs forwards along the mesentery. The
nephridium opens at its anterior end into the collar cavity by
two funnels, one placed below the other (fig. 8). Numerous
ovoid cells are attached by long stalks to the edges of the
funnels. I have not been able to detect cilia in the lumen of
the nephridium. he coelomic epithelium covers the outer
surface of the nephridium. The appearance of the nephri-
dium as seen in the living larva is so well described in the
quotation which Masterman gives that it is unnecessary to
say anything more about it. It is probable that this type of
nephridium, which so strongly recalls the nephridium of
Amphioxus, was characteristic of the primitive Chordata.
Masterman compares these nephridia to the collar canals of
Balanoglossus. There is no doubt a general resemblance
between them, and the collar canals may be vestiges of such
nephridia; but the detailed structure differs in the two cases.
It may also be pointed out that if the small part of the
nephridium which lies behind the mesentery were better
developed, as it is in the adult Phoronis, the nephridium
would be considered as belonging to the trunk, with its
funnels opening into the cavity of the segment in front.

These nephridia become the nephridia of the adult Phoronis.
In the newly metamorphosed Phoronis the nephridia still
project into the irregular collar spaces, but their external
apertures have shifted dorso-laterally, owing, no doubt, to the
mechanical narrowing of the dorsal surface of the body
during metamorphosis.

480 K. RAMUNNI MENON.

(3) Notochords.—In the alimentary canal of Actino-
trocha, Masterman describes an cesophagus, a collar stomach
or pharynx, an cesophagus leading from this into the true
stomach, and an intestine. I have been able to make outi
only three divisions in the alimentary canal—a tube leading
from the mouth to the stomach, which obviously corresponds
to the esophagus (Masterman), a stomach placed partly in
the collar and partly in the trunk segments, and an intestine.
It seems to me that this cesophagus corresponds to the
stomodeum, pharynx, and cesophagus of Cephalodiscus,
described by the same author. However this may be, the
“esophagus” is often folded transversely (this is also the
case in young Phoronis) into pouches, and the subneural
gland isa diverticulum of its dorsal wall. The ‘ cesophagus”
opens into the large stomach anteriorly and dorsally. The
anterior part of the stomach grows forwards in the form of
a diverticulum, which is placed ventrally to the cesophagus.
This diverticulum is absent in the young forms. The
anterior end of the diverticulum may be bilobed, and the
two lobes, often unequally developed, le ventro-laterally or
laterally to the cesophagus. There can be no doubt that the
diverticulum with the lobes is the structure referred to as
“the notochord,” ‘hepatic diverticula,’ etc., although it
must be confessed that I have not seen vacuoles or brown
concretions, stated to occur in these structures. Indeed, I
have not been able to make out any difference between the
lining membrane of the diverticulum and that of the rest of
the stomach. he cells are columnar ciliated cells, and are
taller than the cells of the “ cesophagus,” and the nuclei
form an irregular layer between the base and the middle of
the cells. The protoplasm of the cells presents a granular
appearance and stains homogeneously, but not so deeply as
the protoplasm of the cesophageal cells, and does not differ in
any respect from the protoplasm of the lateral ridges of the
stomach. These ridges of protoplasm, containing numerous
scattered nuclei and large vacuoles with remnants of food
particles in them, are two longitudinal ridges projecting into

NOTES ON ACTINOTROCHA. 481

the lumen of the stomach and extending along the sides
from beyond the opening of the cesophagus to near the
posterior end. The cells along the dorsal median line of the
stomach, and the cells of the posterior part of the stomach
just before it opens into the intestine, have elongated nuclei
placed vertically, and resemble the cesophageal cells. There
does not appear to be any reason for regarding the diverti-
culum as anything but a part of the stomach. The stomach
could grow forwards only ventrally to the cesophagus, and
when the front end of this outgrowth meets the cesophagus,
which bends down to the mouth, it is obliged to grow past it
along its sides. The form and the position of the diverticulum
can thus be explained on mechanical grounds. Whether the
diverticulum has any special function ; whether, for instance,
its association with the large mass of blood-corpuscles is
purely an accident, I cannot say.

(4) The Subneural Gland.—This structure has given
rise to a good amount of discussion. According to Masterman
it arises as an ectodermal invagination in front of the mouth.
I have not seen this stage. In the earliest stage of the organ,
that I have observed, it is present as a shallow diverticulum
of the dorsal wall of the “ cesophagus,” which projects some-
what backwards into the collar cavity, just behind the
mesentery between the preoral and collar cavities. In the
latest stage I have seen the gland opens by a wide aperture
just within the mouth, and is directed forwards into the pre-
oral lobe, in which it lies between the ectoderm of the lower
wall and the cavity of the lobe, and is separated from this
cavity by the basement tissue and scattered nuclei which line
the preoral cavity. If I interpret these facts correctly, the
organ first appears in connection with the collar, and during
development shifts forwards into the preoral lobe; and this
is in harmony with the development of the “ Hicheldarm ” of
Balanoglossus. In Balanoglossus, the “ Hicheldarm ” during
development is described as lying in a space between this
ventral wall and the cavity of the preoral lobe; and it is
certainly a remarkable coincidence between this arrangement

4.82 K. RAMUNNI MENON.

and the condition already mentioned as occurring in Actino-
trocha. Ihave not observed any peculiar relation between
the organ under consideration and the subneural sinus.
There seems to be a general agreement that this structure
is homologous with the “notochord” of Rhabdopleura and
Cephalodiscus ; but whether these structures are homologous
with the ‘‘ Hicheldarm” of Balanoglossus remains undecided.
The facts mentioned above certainly support Harmer’s views
that they are homologous with the ‘‘ Hicheldarm.” The
ciliated epithelium of the organ and its position in relation
to the mouth, however, indicate in Actinotrocha some
function connected with the ingestion of food.

(5) The Tubular Nerve-ganelion.—The ganglion has
the structure described by Masterman. ‘lhe tube formed by
ectodermal invagination lies in the actual ganglion, and not
merely in front of it. The tube is lined by ciliated cells, and
its inner end is sacculated. This ganglion is compared to
the dorsal nerve-cord of the collar in Balanoglossus, which,
however, lies in the collar, while the ganglion of Actinotrocha
lies in the preoral lobe.

Sense-organ.—This organ lies in front of the ganglion,
and is absent in the young larva. The ectoderm in this
region becomes thickened and raised up into a conical
papilla, and there is a development of nervous tissue at its
base. The cavity of the preoral lobe is continued into the
base of the papilla. Possibly this sense-organ represents
the median dorsal tentacle of the larva displaced and retarded
in its development. Attention may be drawn here to the
compound eyes which Masterman describes as occurring on
the tips of the tentacles in Cephalodiscus. Whatever its
homology may be, the organ is now a sense-organ of some
kind. It is curious that this organ is not developed during
the earlier and more active stages of the larva, but makes its
appearance during the later stages, and is best developed
when the larva becomes slow in its habits and when the need
for such an organ appears least. Before metamorphosis, the
larva, with its ventral surface directed downwards, glides

NOTES ON ACTINOTROCHA. 483

along the bottom of the vessel in which it is kept, performing
a slow pirouette in its course. Probably the sense papilla
which is placed on the ventral surface of the hood, and which
would thus come in contact with the ground, is the organ by
means of which the larva feels its ground before settling
down for the rest of its life.

In conclusion it may be added that if Actinotrocha is
related to the Chordata at all, as the presence of three
divisions of the body with their corresponding cavities, of
collar nephridia, of a dorsal diverticulum of the anterior
part of the gut, and of a dorsal tubular nerve-ganglion
renders probable, the absence of such important struc-
tures as the gill-slits, and of the proboscis pores shows
that the relationship is to be traced through a form like

Rhabdopleura.

EXPLANATION OF PLATE 26,

Illustrating Mr. K. Ramunni Menon’s paper entitled “ Notes
on Actinotrocha.”

List of Abbreviations.

an. Anus. an. bd. Perianal band. J/.-c. Blood-corpuscles. 42.-v. blood-
vessel. c¢.c. Collar cavity. ¢. ep. Somatopleure. ec. ep'. Splanchnopleure
round cesophagus. c.ep'!, Splanchnopeure round stomach. @. d/.-v. Dorsal
blood-vessel. ep. Epistome. Ad. Preoral lobe. Ad. c¢. Preoral cavity. m.
Mouth. mes. Mesentery between preoral cavity and collar cavity. mes’.
Mesentery between collar cavity and trunk cavity. J. nep. Posterior nephri-
dium (Masterman). zep. Nephridium. qs. Cisophagus. rec?. Rectum.
tent. Tentacle. ¢r.c. Trunk cavity. ¢, w. body-wall of the trunk region.
vac. ect. Vacuolated ectoderm.

Fie. 1.—Horizontal section of the anterior part of a newly metamorphosed
Phoronis.

Fies. 2—4.—Sagittal section of the anterior end of young Phoronis, to show
the development of the epistome.

484, K. RAMUNNI MENON.

Fig. 5.—Transverse section of Actinotrocha in the collar region. The
collar wall is only shown in outline.

Fic. 6.—A similar section to show the blood-corpuscles in the collar eavity.

Fic. 7.—Part of a horizontal section of Actinotrocha to show the so-called
posterior nephridia.

Fie. 8.—Part of a sagittal section of Actinotrocha to show the nephridium.
The ventral wall of the collar and part of the ventral wall of the trunk are
drawn.

hT London.

a I
LS S93
| Pe
ty Pore { ’ il
§ fei | 4
SS a ae S Sp
wz |
A o =|
5 8 oy sé a
‘ — x ~ y
. «Farge Ss <i

Pe 7 I
Way ttn mo) pny
i anaaagumaoe .
h ares

9

76 VALENS PL,6
:

ay

Supe Sour Mur.

Pon
sass .

| e i
Wagan ; " a Se
. a ;

peer
oe

ray yunee! ' ry
we ° ‘ ar pees a

ee
i srg gterpuan eral Ligeaneity

Perea tt 7 get?
s IN SNTTUTNLMS

fr
Ce Sa iy state

YR.Menon del.

DEVELOPMENT, STRUCTURE, ETC., OF ACTINOTROCHA. 485

Review of Mr. Iwaji Ikeda’s Observations on
the Development, Structure, and Metamor-
phosis of Actinotrocha.!

By

A. J. Masterman,
Edinburgh.

Zootoaisrs are indebted to this publication of the Tokyé
University for an account of many valuable zoological re-
searches, and by nu means the least important is contained
in the volume now lying before us. Mr. Ikeda has been
enabled to study the living larva and the embryonic stages
at the Misaki marine station. The newly described species,
Phoronis ijimai, Oka, furnished the embryonic stages, and
for the structure of the larva no less than four species (?)
were examined. Of these perhaps the most remarkable is
type D, which is a large larva about 5 mm, in length and
with about forty-eight tentacles.

In the embryonic development we naturally look at the
parts dealing with such a much-disputed point as the origin
of the mesoblast. Anyone who has studied the literature of
the development of Phoronis will be able to gather a clear
account of the general features, and will be led to conclude
that the minor differences between the various workers are
largely specific. Speaking generally, a total equal segmenta-
tion, producing a blastula larva, invagination to form a
gastrula, and the persistence of the blastopore as the larval

‘ «Journal of the College of Science,’ Imperial University, Toky6, Japan,
vol. xiii, pt. 4, 1901.

VOL. 49, PART 3.—NEW SERIES. KK

486 A. J. MASTERMAN.

mouth may be accepted as general. But when we come to
the question of mesoblastic origin we are met with difficulties.
Ikeda finds, like Caldwell and myself (my work on the early
development, in this JournaL, was not in Ikeda’s hands till
after going to press), that certain bodies, which he calls
“plasmic corpuscles,” are present in the blastoccele cavity,
but are non-nucleated, and hence cannot be regarded as
mesoblast cells. At the commencement of invagination he
finds that certain hypoblast cells migrate singly into the
blastoccele space. These contribute later to the formation of
mesoblastic organs. After invagination has been largely
completed he finds further mesoblast cells are derived from
a pair of anterior diverticula of the hypoblast, lateral to the
blastopore. These agree exactly in appearance and position
with the similarly named structures of Caldwell. The
presence of these ‘‘ collar somites ” has been independently
verified by myself, and we are also in agreement that the
mesoblast cells produced from them do not at first contain
any cavity, aS maintained by Caldwell. The presence of
these collar somites has been so forcibly denied by both
Roule and Schultze that this corroboration is the more
gratifying.

A third source of mesoblast cells he finds in the ventral
eroove, which consists of cells not yet invaginated to form
hypoblast, lying along the mid-ventral line of the embryo.
These are in the nature of single cells set free into the
blastoccelic cavity. Lastly, he corroborates the presence of
a posterior invagination (Caldwell’s posterior diverticula),
but claims that this is merely an ectoblastic nephridial pit
which gives rise to the pair of nephridia. For these posterior
diverticula I searched in P. Buskii without success, but I
have since been enabled to find them in this species and in P.
hippocrepia. lam inclined to accept Ikeda’s view that they
are the “ anlage” of the nephridia, the development of which
I did not follow in P. Buskii. In passing we may say that
Ikeda does not recognise the presence of my “ posterior
somites,’ but he figures without comment certain indications

DEVELOPMENT, STRUCTURE, ETC., OF ACTINOTROCHA. 487

of their presence, e. g. his fig. 33. Perhaps I may add that
in P. hippocrepia I have found certain mesenchymatous
cells at the gastrula stage, and I think it is possible to regard
these as having the same relation to the mesoblast arising
from the invaginated hypoblast as similar mesenchymatous
cells have to the enteroccelic mesoblast of Echinodermata.

From this it will be clear that it is not difficult to gather a
fairly consistent account of the origin of the mesoblast from
the figures of Caldwell, Ikeda, and myself.

But the second part of Ikeda’s memoir, dealing with the
larval structure of Actinotrocha, is unquestionably the most
important and the most welcome at the present time. Since
my paper on this subject in this Journat Professor Roule
has, at international congresses, in the ‘ Zoologische Anzeiger,’
and in several French journals (‘Comptes rendus,’ ‘ Annales
des Sciences naturelles,’ etc.), spared no pains in reiterating
that the essential facts as described by me, such as the
presence of coelomic cavities, mesenteries, and so on, which
led me to regard Actinotrocha as a highly organised
coelomate larva, were mere figments of imagination, or were
due to this factor combined with a judicious admixture of
defective technique. On the other hand, he has insisted
that it is to be regarded as at the morphological level of
the trochophore. No one comparing our figures could possibly
reconcile the two series of results. Roule was perfectly right
in saying that if his descriptions are accurate mine must
be untrue, and we are perfectly justified in believing the
converse. It is therefore, from the point of view of
zoological progress, eminently satisfactory that my work has
received so complete a corroboration as the present.
Speaking candidly, we do not think that the author has
fully appreciated, or at least emphasised, the corroborative
nature of his work. With one or two important exceptions,
to be noted later, the whole of his anatomical drawings might
be used as they stand to illustrate my paper of 1897. We
can read through page after page of his description in
which the anatomical details are a virtual repetition

488 A. J. MASTERMAN.

of the work referred to, yet the agreements are so little
emphasised, and the disagreements so dwelt upon, that any
reader who had not studied my work could almost suppose
that the total result was rather a refutation than a corrobora-
tion of the latter. However, Ikeda’s figures are so accurately
and carefully drawn that it is possible to explain some of the
more important differences. The first kind of discrepancy
is simply due to a difference of interpretation. For example,
he shows clearly in several of his figures (such as the series
fic. 59, a, b, c, d) the subneural blood-sinus in even greater
detail than myself, but he prefers to label it as an artefact.
As the blood system of Actinotrocha is hemoccelic, and
hence consists of mere sinuses or spaces between the three
primary layers, it is naturally possible to regard it in its
entirety as an artefact. The same remark applies equally to
the whole vascular system of Cephalodiscus, the greater
part of that of Balanoglossus, and that of numerous larvee.
Indeed, there is little to urge against the assertion that the
hemoceelic body-cavity of arthropods is an artefact. On
similar grounds he fails to corroborate the presence of the
perianal blood-sinus (one of the most conspicuous characters
of the larva), and of other parts of the blood system. Similar
considerations apply to his refusal to recognise the presence
of the epidermic pit (‘ neuropore”) and of the subneural
gland. Both of these are clearly figured in his sections with
exactly the relations indicated by myself, but he prefers to
regard them as artefacts. As in all my fully developed
larvee, alive or dead, I have never failed to find all these
organs constant in position and in structure, I can hardly
agree that they are mere freaks of reagents.

Ikeda suggests that the epidermic pit in front of the
ganglion is produced by contraction of muscles drawing
the ganglion backwards (!), and similarly he accounts for
the evident presenee of the subneural gland by an artificial
bulging forwards and outwards of the part of the hood im-
mediately anterior to it (!). Surely the mechanical con-
ditions induced in each case would tend to exactly the

DEVELOPMENT, STRUCTURE, ETC., OF ACTINOTROCHA. 489

opposite result, i.e. that of straightening out these de-
pressions.

A. second group of discrepancies is apparently due to
specific variation. Amongst these we may instance Ikeda’s
failure (as in the case of Roule) to find two pleurochords.
In the North Sea larvee the anterior wall of the stomach
pushes forward as a diverticulum below the csophagus,
which opens into the antero-dorsal corner of the stomach.
If it simplifies matters at all to regard this diverticulum as
“ventral” instead of anterior, there seems no objection to
doing so. As development proceeds, this diverticulum
throws out two lateral pleurochords as I have described. In
the Japanese larve this stage is never reached, but the
diverticulum remains single with only one row of vacuoles ;
hence we must be prepared for other organs in like manner
being abbreviated or even undeveloped. For example, the
nephridia in these larve, as figured by Ikeda, are very small,
very simple, and very embryonic as compared with the large
branching tubules in the North Sea type. This considera-
tion may easily account for the failure of Ikeda to find the
internal openings of these organs in any of his larve. It is
possible that the absence of proboscis pores, of a post-oral
nerve-band, and of dorsal and ventral nerve-fibres may be
due to a similar cause. In future work upon the central
plexus of Actinotrocha [ shall have a further occasion to
refer to these matters.

A. third source of seeming discrepancy is due to an un-
intentional misunderstanding,
statements.

I had to notice the existence of small processes of the
trunk coelom which lie in the perianal sinus, and in doing so
was obliged to hold in view the possibility of their being the
first indications of the adult nephridia. Further investiga-
tion has convinced me that they have little morphological
significance, but both Roule and Ikeda have imputed to me
the definite statement that the trunk ccelom has definite
nephridia, apparently only for the purpose of contradicting

on the author’s part, of my

490 A. J. MASTERMAN.

it. Perhaps it would be as well to state once for all that the
trunk has no nephridia and possesses no normal openings till
after the metamorphosis.

In reference to the oral and “ pharyngeal” (? atrial) grooves
the author speaks of a “grave error” into which I fell by
not examining the living larve. With all due deference to
him I must state that the existence of these grooves first
appeared to me by experimental feeding of the living larva.
Indeed, the course of food and water currents indicates their
presence far more clearly than does their structural differen-
tiation. In feeding, the larva places the preoral hood well
down, and the food particles pass upwards along the ventral
surface of the collar (not of the hood, as wrongly quoted by
Ikeda, p. 536) in two slight depressions, indications of which
are also seen in sections (fig. 29). On the other hand, the
atrial water passes in a stream along the line of junction
between hood and collar and out at the dorsal edge of the
collar. These atrial grooves are also more in physiological
than morphological evidence, though my sections show that
there is a groove between hood and collar. (Woodcut 5, on
p. 301, should sufficiently explain this.) Lastly, the author
remarks, ‘ Masterman has made the statement that in its
natural attitude the hood had its length disposed parallel to the
principal body axis. However, if the larva be examined in
the living state, it will at once be discovered that its normal
disposition is horizontal” (p. 536). The actual statement
was, “The hood is in the position which perhaps may be
described as norma] ” (p. 289), which is not by any means
the same thing. Actinotrocha spends most of its time
with the hood flexed ventrally, but the extension of the hood
forwards is effected at pleasure and usually with great
frequency. The term “normal” was used in much the same
sense as one would regard the proboscis of the nemertine, when
extruded, as being in the normal position, or the extended
arm of the human subject. A little careful study of the
work which he contradicts would have enabled Ikeda to
avoid these seeming discrepancies. In addition to Ikeda’s

DEVELOPMENT, STRUCTURE, ETC., OF ACTINOTROCHA. 491

important verifications of the results of his predecessors, and
the few differences from my work which I have indicated
above, he has also added several interesting and new facts.
He describes two retractor muscles of the hood which run
from the main nerve-ganglion down to the dorsal body-wall.
In one larva he finds a pair of long trunk-retractors. We
may also note the description of a pair of globular glands
which open on the dorsal wall of the hood in one species of
larvee, but not in the others.

In the third part of his paper the author deals with the
metamorphosis. As one who has attempted many times to
follow in detail this important process, and each time has
been forced to wait for more material, I can speak of his
attempt with sympathy. At the same time it will be well to
accept with great caution all results, especially referring to
the mesodermic organs. The external changes have already
been carefully noted by many, and the changes of the
alimentary canal are also easy to follow. In describing
these latter we may note that Ikeda is able to correct the
obvious error of Roule that the intestine ruptures, and thata
fresh intestine is created from the cordon dorsal. (As is
to be expected, he finds that this organ is the dorsal blood-
vessel.) But when we come to the mesoblastic organs the
difficulties begin. According to Ikeda, an “adult” collar
cavity (or the supra-septal cavity) is formed round the base
of the tentacles apparently from part of the larval collar
cavity (at least the walls), and the rest of the larval cavity is
converted into the blood ring-sinus of the adult. The diffi-
culties relating to the nephridial openings into the trunk,
the adult nervous system, and the epistome are discussed,
but one would desire the author to pursue his studies on the
metamorphosis, and it is pretty certain he will find that
arguments from probabilities have no weight whatever with
a metamorphosing Phoronis. At least, this is the writer’s
experience, and it is the main reason which conduces to a
belief in the fate of the larval collar cavity as described by
the author. But these remarks are not in any way meant to

492 A. J. MASTERMAN.

We recommend for the Réntgen and Scientific
Institutes our

Special Rontgen

AND

Special Extra Rapid Schleussner
Plates,

which are unrivalled, and recognised as the best
by the Professors

Dr. A. Hoffa, N. Wolfer, Dr. Winkelmann,
Dr. R.. A. Koch, Dr: Relhn; Dr. H. C.. Vogel,
Dr. Max Wolf, H. Kayser, Dr. N. Thege,
Dr. F. Krause, Dr. H. Nitsche, Dr. Rieder,

Dr. Quincke, etc.
Only one trial of these Plates is required to prove

that they possess in every respect the good qualities
we claim for them.

eee

LPL PPP

PLEASE WRITE FOR SAMPLES TO THE MANUFACTURERS,
TROCKENPLATTENFABRIK

Dr. C. SCHLEUSSNER,

AKTIENGESELLSCHAFT,

FRANKFORT-ON-THE-MAIN.

PRICE LISTS AND SAMPLES FREE.

ce
dilb awe Aysmetesttey ne a
slg, 6 A

—_—s

DQ TETAS ise
zoinis 4

af 17 A fe

Ly a

mite ie

“i A: | .h é d
iow. sere
nin st |

‘ini ak

soy Aa oe

Vash \ 194 eo
>.

DD Adc

Pi )

HO? a ViA ia |

[4s CPRLK Bee
:
' a4

oa

a”
—_

-

With Ten Plates, Royal Ato, ds.
CONTRIBUTIONS TO THE KNOWLEDGE OF RHABDOPLEURA
AND AMPHIOXUS.
By E. RAY LANKESTER, M.A., LL.D., F.R.S.
London: J. & A. CHURCHILL, 7, Great Marlborough Street.

Quarterly Journal of Microscopical

Science.

The SUBSCRIPTION is £2 for the Volume of Four Numbers ;
for this sum (prepaid) the JourNAL is sent Post Free to any part
of the world.

BACK NUMBERS of the Journat, which remain in print, are
now sold at an uniform price of 10/-.

The issue of SuppLement Numpers being found inconvenient,
and there being often in the Hditor’s hands an accumulation of
valuable material, it has been decided to publish this Journal at
such intervals as may seem desirable, rather than delay the appear-
ance of Memoirs for a regular quarterly publication.

The title remains unaltered, though more than Four Numbers
may be published in the course of a year.

Each Number is sold at 10/-, and Four Numbers make up a
Volume.

London: Je & A. CHURCHILL, 7 Great ee Street.

TO CORRESPONDENTS.

Authors of original papers published i in the Quarterly Journal
of Microscopical ‘Science receive tw enty-five copies of their
communication gratis.

All expenses of publication and illustration are paid by the
publishers.

As a rule lithographic plates, and not woodcuts, are used in
illustration.

Drawings for woodcuts should nor be inserted in the MS.,
but sent in a separate envelope to the Editor.

Contributors to this Journal requiring ewvtra copies of their
communications at their own expense can have them by applying
to the Printers,

Messrs. ApLARD & Son, 224, Bartholomew Close, E.C., on
the following terms :

For every four pages or less—

25 copies : 5/-
50°, wef Secethae eae
100 7/-

Plates, 2/- per 25 if uncoloured ; if coloured, at the same rate for
every colour.
Prepayment by P.O. Order is requested.
AL COMMUNICATIONS FOR THE EDITORS TO BE ADDRESSED TO THE CARE
or Messrs. J. & A. CHURCHILL, 7, GREAT MARLBOROUGH STREET,
Lonpon, W.

THE MARINE BIOLOGICAL ASSOCIATION

OF THE
UNITED KINGDOM.

Patron—HIS MAJESTY THE KING.
President—Prof. E. RAY LANKESTER, LL.D, tn

205

THE ASSOCIATION WAS FOUNDED “‘ TO ESTABLISH AND MAINTAIN LABORATORIES ON
THE COAST OF THE UNITED KINGDOM, WHERE ACCURATE RESEARCHES MAY BE CARRIED
ON, LEADING TO THE IMPROVEMENT OF ZOOLOGICAL AND BOTANICAL SCIENCE, AND TO
AN INCREASE OF OUR KNOWLEDGE AS REGARDS THE FOOD, LIFE CONDITIONS, AND HABITS
OF BRITISH FOOD-FISHES AND MOLLUSCS.”

The Laboratory at Plymouth
was opened in 1888. Since that time investigations, practical and scientific, have
been constantly pursued by naturalists appointed by the Association, as well as by
those from England and abroad who have carried on independent researches.

Naturalists desiring to work at the Laboratory

should communicate with the Director, who will supply all information as to
terms, &c,

Works published by the Association

include the following :—‘ A ‘Treatise on the Common Sole,’ J. T. Cunningham, M.A.,,
4to, 25/-. ‘The Natural History of the Marketable Marine Fishes of the British
Islands,’ J. TI. Cunningham, M.A., 7/6 net (published for the Association by
Messrs. Macmillan & Co.).

The Journal of the Marine Biological Association

is issued half-yearly, and is now in its fifth volume (price 3/6 each number).

In addition to these publications, the results of work done in the Laboratory
are recorded in the ‘ Quarterly Journal of Microscopical Science,’ and in other
scientific journals, British and foreign.

Specimens of Marine Animals and Plants,

both living and preserved, according to the best methods, are supplied to the
principal British Laboratories and Museums. Detailed price lists will be forwarded
on application.

TERMS OF MEMBERSHIP.

AnnvuAL MEMBERS ., ; ; ee col 1220 peranunum,

Lirzr MEMBERS . t ‘ : . 15 15 O Composition Fee.
FOUNDERS . : : ; - 2 LOOE Oss0. “ “4
Governors (Life Members of Council) 500 0 O 5 =

Members have the following rights and privileges:—They elect annually the
Officers and Council; they receive the Journal free by post; they are admitted to
view the Laboratory at any time, and may introduce friends with them; they have the
first claim to rent a table in the Laboratory for research, with use of tanks, boats, &c.;
and have access to the Library at Plymouth. Special privileges are granted to Governors,
Founders, and Life Members.

Persons desirous of becoming members, or of obtaining any information with
regard to the Association, should communicate with—

The DIRECTOR,
The Laboratory,
Plymouth.

Arn

|

|
|
|
|

New Series, No. 180 (Vol. 45, Part 4). Price 10s.

MARCH, 1902.

ee

THE

QUARTERLY JOURNAL

OF

MICROSCOPICAL SCIENCE.

EDITED BY

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

CORRESPONDENT OF THE INSTITUTE OF FRANCE

HONORARY FELLOW OF EXETER COLLEGE, OXFORD;
AND OF THE IMPERIAL ACADEMY OF SCIENCES OF ST. PETERSBURG, AND OF THE ACADEMY

OF SCIENCES OF PHILADELPHIA} FORRIGN MEMBER OF THE ROYAL BOHEMIAN
SOCIRTY OF SCIENCES, AND OF THE ACADEMY OF THB LINCEI OF ROMB;
AND OF THE AMBRICAN ACADEMY OF ARTS AND SCIENCES OF BOSTON}
ASSOCIATE OF THE ROYAL ACADEMY OF BELGIUM; HONORARY MEMBER
OF THE NEW YORK ACADEMY OF SCIENCES, AND OF THE
CAMBRIDGE PHILOSOPHICAL SOCIRTY, AND OF THE ROYAL
PHYSICAL SOCIETY OF EDINBURGH ; HONORARY
MEMBER OF THE BIOLOGICAL SOCIETY
OF PARIB;
HISTORY DEPARTMENTS OF THE BRITISH MUSEUM; LATE FULLERIAN

DIRECTOR OF THE NATURAL
OGY IN THE ROYAL INSTITUTION OF GREAT BRITAIN.

PROFESSOR OF PHYSIOL

WITH THE CO-OPERATION OF

ADAM SEDGWICK, M.A., F.RS.,

FELLOW AND TUTOR OF TRINITY COLLEGE, CAMBRIDGE;

Ww. F. BR. WELDON, M.A, F.BS.,

OF COMPARATIVE ANATOMY AND FELLOW OF MERTON COLLEGE, OXFORD;

LINACRE PROFESSOR
LATER FELLOW OF ST. JOHN’S COLLEGE, CAMBRIDGE,

AND

SYDNEY J. HICKSON, M.A., F-.R.S.,

BEYER PROFESSOR OF ZOOLOGY IN THE OWENS COLLEGE, MANCHESTER.

WITH LITHOGRAPHIC PLATES AND ENGRAVINGS ON WOOD.

FLONDON:
J. & A. CHURCHILL, 7, GREAT MARLBOROUGH STREET.
1902.

Adlard and Son,] [Bartholomew Close.

CONTENTS OF No. 180.—New Series.

MEMOIRS:

On the Structure of the Excretory Organs of Amphioxus. Part I. By
Epwin 8. Goopricn, M.A., Fellow of Merton Collece, Oxford.
(With Plate27) :

A Contribution to the Morphology of ta nee Head Skeleiann
based upon a Study of the Developing Skull of the Three- spined
Stickleback (Gasterosteus aculeatus). By H. H.Swinnerron,
B.Sc., from the Zoological Laboratory, Royal College of Science,
London. (With Plates28—31 and 5 Text Illustrations)

The Development of Admetus pumilio, Koch: a Contribution 6
the Embryology of the Pedipalps. By L. H. Goven. (With
Plates 32 and 33) .

On the Teeth of Petromyzon ae iiscee. By ee Winasee
D.Sc., Assistant Professor of Zoology, University College, London.
(With Plate’34)

Typhlorhynchus nanus: a New Buatdoudle. By F, Fr, Laan
B.A. (With Plate 35) , : : :

PAGE

493

503
595
631

637

APR ¢ 02

STRUCTURE OF EXCRETORY ORGANS OF AMPHIOXUS. 493

On the Structure of the Excretory Organs of
Amphioxus.

Part, Fr

By
Edwin 8S. Goodrich, M.A.,
Fellow of Merton College, Oxford.

With Plate 27.

THE excretory organs of Amphioxus were independently
discovered by Weiss and Boveri in the year 1890 (1 and 18).
Weiss described a series of small tubules regularly dis-
tributed at the top of each secondary tongue-bar throughout
the region of the pharynx. ‘The tubules are situated, for the
most part, in the wall separating the dorsal ccelom from the
atrial cavity ; they lie, therefore, between the coelomic and
the atrial epithelium, generally separated from the latter by
a network of fine blood-vessels. ‘These kidney tubules open
into the atrium by a pore just opposite the dorsal end of the
secondary gill-bar. Weiss suspected the presence of an
internal opening, but could not find it. The physiological
significance of these organs he established by means of
feeding experiments with carmine and other colouring
matters.

In 1892 Boveri published a detailed and beautifully illus-
trated account of the excretory organs of Amphioxus (2).
In this paper such a clear and accurate description of the
appearance, general structure, and distribution of the
kidneys is given that little remains to be said on these

VOL, 45, PART 4,—NEW SERIES. LL

4.94, EDWIN S. GOODRICH.

subjects. It will be necessary only to deal in detail here
with certain points on which we differ, and concerning which
I am able to correct Boveri’s account in some important
particulars.

The following is a brief statement of Boveri’s description
of the excretory organs :—Hach consists of a tubule ciliated
internally, and opening into the atrium by a pore near the
top of the secondary gill-bar. In the mid-region of the
pharynx, where the canal is best developed, it extends
forwards to near the origin of the primary bar in front, and
backwards to the primary bar behind, down which it runs
for some little distance. Along its course the tubule is said
to open into the dorsal ccelom by means of a varying number
of small funnels; and spread over the ccelomic wall in the
neighbourhood of each funnel are many peculiar pin-shaped
cells called “ fadenzellen.” A long, slender process, starting
from each of the “ fadenzellen,” is attached to the lip of the
funnel. To quote from Boveri: “Jede Zelle léiuft namlich
in einen feinen, aber deutlich doppelt contourirten Faden
aus, der mit den tibrigen Auslaufern der gleichen Zellengruppe
zu einem Trichter hinzieht und in der Miindung desselben
eine Strecke weit verfolet werden kann;” and further:
“Die Faden ziehen frei durch die Leibeshdéhle schrag
abwirts in die Trichteréffnung hinein gegen die laterale
Wand des Canilchens und heften sich mit ihnen Enden an
die Zellen des Nierenepithels an” (2).

Some years ayo, being struck with the resemblance these
“ fadenzellen” bear to the solenocytes I had just dis-
covered in the nephridia of Polychate worms, I examined the
kidneys of Amphioxus, and came to the conclusion that the
similarity was only superficial, and that Boveri’s description
was essentially correct (5, Part III). This winter, however,
whilst occupying the British Association table at the Stazione
Zoologica in Naples, I determined to re-examine these
organs, and I am now able to definitely state that the
‘ fadenzellen’’? of Amphioxus are indeed solenocytes of
typical, though somewhat peculiar structure (7).

STRUCTURE OF EXCRETORY ORGANS OF AMPHIOXUS. 495

The methods pursued are of the most simple kind, and it
is within the power of any one with living material at hand to
easily verify my results. The Amphioxus is pinned out on
its back in a shallow dish of sea water. The atrium is ripped
up with a needle along the mid-ventral line, and the two
metapleural folds pinned aside. The exposed pharynx is
then also ripped up with a needle, and portions of the right
or left side of the pharynx can then be torn out with forceps

Figure of a portion of a section across the gill-bars showing the
excretory canal cut through. 4. Atrium. ©. Dorsal cclom.
po. Primary gill-bar. dv. Blood-vessel. ae. Atrial epithelium.
ce. Colomic epithelium. 7. Lumen of the excretory canal. 2.
Nucleus of asolenocyte. ¢. Tube of a solenocyte, Cam. and oil
immersion.

and placed on a slide,’ care being taken to lay the outer
side uppermost. When covered and examined under the
microscope the general structure of the kidney can be quite
easily seen; but the details which I am about to describe
can, unfortunately, only be made out after prolonged study
with the highest powers (,‘; oil immersion and No. 8 eyepiece,
for instance).

‘ Such pieces can be stained and mounted, or cut into sections,

496 EDWIN S. GOODRICH.

> are seen to con-

Examined in this way, the “ fadenzellen’
sist of a small cell-body containing a nucleus. The cell-body
is of somewhat irregular shape, being circular, triangular, or
elongated, and occasionally drawn out into a process, re-
minding one of the outer processes present on some Poly-
chete solenocytes (5). A neck-like region, sometimes
straight, sometimes curved, gradually narrows down and
joins the cell-body to the distal extremity of the ‘ thread,”
which is, in fact, a slender hollow tube of great length (Fig.
in text, p. 495, and PI. 27, figs. 1 and 4).

The longest tubes belong, of course, to those cells which
are situated furthest from the renal canal; they reach some-
times a length of 90 mw, or nearly =; mm. ‘The wall of the
tube does not appear to be as stiff as in the case of Poly-
chete solenocytes ; and under the pressure of the cover-
olass it is often much curved. In the living animal, however,
I believe the tubes are always straight. The proximal end
pierces the wall of the excretory duct, and projects a little
into the lumen of the canal (figs. 3 and 4). A long flagellum,
attached at its base to the cell placed at the end of the tube,
works rapidly down the tube and far into the excretory canal
(figs. 3 and 4).!

Boveri (influenced perhaps by the current dogma, which
affirms, in spite of all evidence to the contrary, firstly, that all
tubular excretory organs are of homologous nephridial nature ;

and secondly, that nephridia are derived from the ccelom,
and generally, if not always, open into it) described open
coelomic funnels in Amphioxus, as already mentioned above.
Neither in the living nor in sections of preserved specimens
have I been able to detect any direct communication between
the excretory canal and the ccoelom. The branches of the
tubule may be very numerous, of considerable length, and
may themselves divide, but they end blindly (fig. 1). It
is to these blind ends that the tubes of the solenocytes con-
verge, and here the wall of the canal is less loaded with

1 T was fortunately able to demonstrate the correctness of these observa-
tions to Prof. Boveri himself at the zoological station in Naples,

STRUCTURE OF EXCRETORY ORGANS OF AMPHIOXUS. 497

excretory granules, and thinner than elsewhere. The nuclei
also, which are so numerous in the other regions of the
tubule, are not present just in those parts where the
solenocytes traverse its wall (Fig. in text, p. 495).

Briefly to summarise the observations described above, it
may be stated that in Amphioxus there is a series of
excretory tubules opening into the atrium, but not into the
coelom, and provided at their blind internal ends with a large
number of solenocytes. These tubules are situated “ mor-
phologically ” outside the ccelom, being covered with ccelo-
mic epithelium ; the solenocytes alone push through into the
ccelomic cavity.

At each end of the pharynx the excretory organs dwindle
in size, as already pointed out by Boveri. The tubule in
these regions becomes shorter, the branches become reduced
in size, or are not developed at all, and the number of soleno-
cytes becomes much less. This is also the case throughout
the pharyngeal region of small specimens. Fig. 3 repre-
sents a small portion of the kidney of a young Amphioxus
31 mm. long. Not only are the solenocytes not so crowded,
but the average length of the tubes is less than in full-
grown individuals (fig. 1).

That the segmental kidneys of Amphioxus really fulfil an
excretory function has been amply demonstrated by Weiss
and Boveri; but the part played by the solenocytes them-
selves is less clear. Boveri, who remarked that their dis-
tribution over the wall of the coelom coincides with that of a
network of blood-vessels, concluded that the ‘“ fadenzellen ”
were concerned in the elimination of waste products from the
blood: ‘diese Zellen dem Chemismus der Excretion dienen.”
He could not find that they took up colouring matters—a
result which agrees with my own observations on the soleno-
cytes of Polychetes (5, Part II). It seems to me probable,
therefore, that, as I have already suggested for worms
(5, Part II), the solenocytes are concerned chiefly with the
elimination of fluid substances which can pass by osmosis
through the thin walls of the tube, weil adapted for such a

498 EDWIN S. GOODRICH.

purpose. ‘he flagellum would serve to propel the fluid into
the excretory canal and thence to the exterior. That a con-
siderable amount of fluid could pass through the tubes be-
comes evident when we remember that in a full-grown
Amphioxus there are, roughly speaking, 100 kidneys on each
side, or some 200 inall. Now each of these has, on an average,
about 500 solenocytes—to take a low estimate,—making the
number of solenocytes in the whole animal roughly 100,000.
The average length of the solenocyte tube may be taken at
so mm., or 50 u. There are, therefore, about 5 metres of this
thin-walled tube in each full-grown individual, representing
no inconsiderable area for osmotic exchange in an animal of
such small bulk.

Conclusion.—I do not propose in this paper to enter
into a detailed discussion of the homology and taxonomic
importance of the segmental kidneys of Amphioxus, but
the extraordinary resemblance they bear to the nephridia I
have described in the Nephthyide, Glyceridz, and Phyllo-
docidze must be insisted upon. For the purpose of com-
parison, figures are given of the inner end of the nephridium
of Phyllodoce Paretti (figs. 2 and 5), which, of all the
Polycheetes I have studied, most closely resembles Amphioxus
in the structure of its renal organ. The nephridium of this
beautiful Phyllodocid is large enough to be dissected out.
It is then seen to end in the colom in a bunch of blind
branches, which are provided with a number of solenocytes
arranged like the ribs of a fan. The tubes are in double
rows, while the cell-bodies of the two rows of solenocytes are
closely packed together, and wedged in alternately.! In
fig. 5 I have given a diagrammatic representation of the
extremity of a branch of the nephridium, to compare with
the similar diagrammatic figure of a small portion of the
kidney of Amphioxus (fig. 4). These figures bring out
clearly the wonderful likeness of the two organs.

1 I estimate the number of solenocytes in a Phyllodoce Paretti roughly
at 600,000, there being about 1500 to each nephridium, and some 200
segments.

STRUCTURE OF BXCRETORY ORGANS OF AMPHIOXUS. 499

It results from these observations, as I have already pointed
out in a preliminary paper (7), that in their segmental
arrangement, in their function, and in their histo-
logical structure, the excretory organs of Am-
phioxus and the nephridia of Phyllodoce are in all
essentials identical.

Before committing ourselves to new theories, something
must be known of the development of these organs; but,
considering how remarkably close is the agreement between
the two, it seems more than probable that they are homologous
structures. If two such excretory organs as the solenocyte-
bearing nephridia of Phyllodoce, and the solenocyte-bearing
kidneys of Amphioxus, could be shown to have been inde-
pendently involved, we should have to give up structural
resemblance as a guide to homology.! But there seems to be
no danger of our being driven to abandon the problem as
yet, and all we need assume is, not that the vertebrates have
been evolved from the Polychetes, but that the remote
common ancestor of these now highly differentiated phyla was
of more elaborate structure than most authors have been hither-
to inclined to suppose. We must assume that it possessed not
only paired ccelomic (genital) sacs and ccelomostomes (6 and
8), but also nephridia, whose blind internal end was provided
with solenocytes.2. We may conclude provisionally that
now, for the first time, true nephridia have been shown to
occur in the vertebrate phylum; and further, we may hope
to trace in the vertebrates the same two series of organs—
the nephridium and the coelomostome—which I have elsewhere

1 The only case which seems to me at all comparable is that of the
nematocysts in Ceelenterates, Planarians, and Molluscs.

2 This is all the more easy to believe since I have found these cells at the
blind inner end of the nephridium of the larva of Phoronis (it will be remem-
bered that Masterman observed cells similar to Boveri’s “ fadenzellen” in
Actinotrocha [9], and they have been described by Wagener [12]); and
“ flame-cells” very like solenocytes have been described in Nemertines by
Birger (4), in Molluscs by Meisenheimer (10), and in Rotifers by
Shephard (11). Also the “flame-cells” of Platyhelminths and Entoproctous
Polyzoa are probably of the same nature (6).

500 EDWIN 8S. GOODRICH.

endeavoured to prove exist in the majority of ccelomates
(6, and 5, Part III).

It follows, from the conclusion provisionally adopted above,
that the many theories which have been built on the assumed
homology between the separate segmental excretory organs
of Amphioxus and the renal organs of ccelomic origin of the
higher vertebrates must be allowed to drop for the present.
These latter organs (pronephros, mesonephros, and genital
ducts) have nothing to do with nephridia, and appear to
belong undoubtedly to the category of ccelomostomes (5,
Part III, and 8). Their homologues in Amphioxus may be
sought in the opening of the larval “ head cavities,” in the
“brown funnels” described by Lankester, and in the segmental
genital sacs as already suggested by Boveri (8). It is not
impossible, however, that true nephridia may yet be found at
some stage of development amongst the craniate vertebrates,
and more especially in the Cyclostomes.}

List oF REFERENCES.

1. Bovert, ‘I.—“ Ueber die Niere des Amphioxus,” ‘ Miinch. med. Wochen-
schrift,’ No. 26, 1890.

2. Boveri, 'l.—‘ Die Nierencanalchen des Amphioxus,” ‘Zool. Jahrb.,’
vol. v, 1892.

3. Bovert, T.—“ Ueber die Bildungstate der Geschlechtsdriisen, etc.,”
‘Anat. Anzeig.,’ vol. vil, 1892.

4. Bircer, O.—* Die Nemertinen,” ‘Fauna und Flora des Golfes von
Neapel,’ vol. xxii, 1895.

5. Goopricu, E. 8.—“On the Nephridia of the Polycheta,” pt. i, ‘Quart.
Journ, Mier. Sci.,’ vol. xl, 1897; pt. ii, ibid., vol. xli, 1898; pt. iii,
ibid., vol. xliii, 1900.

6. Goopricu, E. 8.—“On the Colom, Genital Ducts, and Nephridia,”
‘Quart. Journ. Mier. Sci.,’ vol. xxxvii, 1895.

7. Goopricn, EH. 8.—‘On the Excretory Organs of Amphioxus,” ‘ Proc.
Roy. Soce.,’ 1902.

1 | fully expect that nephridia will some day be found in the Enteropneusta
aud even in the Tunicata.

STRUCTURE OF EXCRETORY ORGANS OF AMPHIOXUS. 501

8. Lanxester, E. Ray.—“ The Enteroccela and the Coelomoceela,” ‘ Treatise
on Zoology,’ pt. 2, 1900.

9. Masterman, A. T.—‘ On the Diplochorda,”’ ‘ Quart. Journ. Mier. Sci.,’
vol. xl, 1897.

10. MerIsenuEIMeER, J.—“ Entwickl. von Dreissensia polymorpha, Pall,”
‘ Zeit. f. wiss. Zool.,’ vol. Ixix, 1901.

11. SuepHarp, J—‘On the Structure of the Vibratile Tags or Flame-cell
in Rotifera,” ‘ Proc. Roy. Soc. Victoria,’ vol. xi, 1899.

12. WacrenEeR, R.—“‘ Ueber den Bau der Actinotrocha branchiata,”’
‘Archiv f. Anat. und Phiys.,’ 1847.

13. Wuiss, E.—‘“ Excretory Tubules in Amphioxus lanceolatus,’ ‘Quart.
Journ. Mier. Sci.,’ vol. xxxi, 1890.

EXPLANATION OF PLATE 27,

Illustrating Epwin 8. Goopricu’s paper ‘On the Structure
of the Excretory Organs of Amphioxus.”

Fic. 1.—Enlarged view of an excretory organ of Amplhioxus, drawn from
the living.

Fie. 2.—Enlarged view of the terminal tuft of the nephridium of Phyllo-
doce Paretti, drawn from the living.

Fic. 8.—A small portion of the excretory canal of a young Amphioxus,
with its solenocytes, from the living. Cam. =; oil immersion, oc. 8.

Fies. 4 and 5.—Semi-diagrammatic views of portions of the excretory
organs of Amphioxus and Phyllodoce Paretti, from living and preserved
specimens.

ae

1

eo Loe

re
sede li Age “h3) y j
eh iol

BA) ea he GE
ae i:

{ iA Ls
: ory
T r
‘ =
}
- '
<
yas
nm
f " nl
7
!
il
I
\
. p :
a ae
Ag
y
~
»
Pea _
+)
= ‘

‘
’
,

2949087 0

“2DUDD Akozakore FT

uf

agtaouazos

alod qaqa sy

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 5038

A Contribution to the Morphology of the Teleos-
tean Head Skeleton, based upon a Study of
the Developing Skull of the Three-spined
Stickleback (Gasterosteus aculeatus).

2
av

H. H. Swinnerton, B.Sc.,
From the Zoological Laboratory, Royal College of Science, London.

With Plates 28—31 and 5 Text Illustrations.

ContTENTS.

PAGE
I. InrRopUcTORY . : : ; ; : . 503
Il. MatEertaL anD Meruops : : ; ; 4 505

Ill. Descriptive anp CoMPARATIVE:
The Cranium F : ; : 5 . 506
The Visceral Skeleton : F : : 5 oR:

LV. GenERAL ConSIDERATIONS :

The Relation of the Trabecule to the Parachordals . .- 560
The Primordial Cranium : , : : . 562
The Relation of the Visceral Skeleton to the Cranium . 568
The Systematic Position and Affinities of Gasterosteus . 574
V. Summary anp Conciusions P ‘ A F . 583

I. Inrropuctory.

During the years which have elapsed since W. K. Parker
(73) published his admirable ‘ Monograph on the Skull of the
Salmon,’ comparatively little has been done to advance our

504. H. H. SWINNERTON.

knowledge of the development of the Teleostean head skele-
ton; McMurrich (83) is the only investigator who has dealt
with the development of the whole head skeleton of any other
type with any approach to detail; Pouchet (78) has dealt
with isolated stages of Gobius, Syngnathus, Labrus, Atherina,
and Engraulis; Stohr (88) has confined himself to extending
or rectifying Parker’s observations; whilst Ganin (80) has
devoted two pages to these processes in the skull of Gaster-
osteus and Rhodeus. Several others have dealt with the
development of parts with more or less detail.

It was with a view towards supplying this deficiency that
the investigation, the results of which are here recorded, was
undertaken. Acting upon a suggestion originally made by
Professor G. B. Howes, the common three-spined stickleback
was chosen as the basis of my observations, partly because
of the ease with which the material could be obtained, and
since, while it offered a more specialised type than the
salmon, it seemed to form, according to Cope (70) and later
systematists, a suitable starting-point for the study of certain
specialised and peculiar groups, e.g. Lophobranchs.

Beyond the work of Ganin the only previous observa-
tions upon the skull of Gasterosteus which I have been
able to find are those of Huxley (58 and 59), chiefly in his
Croonian Lecture “On the Theory of the Vertebrate Skull.”
Figuring and describing the hyosuspensorial apparatus of
the adult, he gives three figures of the larval skull, which are
remarkable for their accuracy. One figure represents a
stage coincident with my Stage II, and the other two
correspond to my Stage IV. Amongst his original drawings,
now preserved in the Huxley laboratory, are several other
figures of the same stages, which were evidently studies for
parts of those which he published.

I tender my heartiest thanks to Professor G. B. Howes for
his ever-ready help in obtaining material, his guidance in
working, with the reading, and also to Mr. M. F. Woodward
and Mr. G. A. Boulenger for much valuable assistance.

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 505

II. Marerrat AnD Meruops.

Part of the material requisite for my investigation was
obtained by keeping adult sticklebacks in small aquaria
during their breeding season; the rest by hatching, under
the tap, spawn found in a stream near London. In this
way it seemed possible to obtain a definite idea of the size
of the various embryos and larvee at different ages. During
the earlier stages this was possible; but later on, owing to
the great variation in length of individuals belonging to
the same hatch, the attempt became useless.

The following table contains a list of all the specimens
examined, with details which may prove useful to future
workers.

Table of Specimens with Descriptions of the Stages.

No. Age. Length. Description of Stages.
l a—e | 5th day * Stace [L—No hyaline cartilage; no
2 a—e |Karly 6th * sign of the cranial roof; the sym-
day plectic not elongated; the palatine
3 6th day * process not yet attached to the eth-
4a—b6 | 6th day 3°5 mm. moid region,

5 a—ec | 6th day | 36—3°7 mm.
6 7th day *

7 a—b | 7th day 3°6 mm.

8 a—d | 7th day | 4-0—4°2 mm.

9 8th day * Srace I].—Much hyaline cartilage ;
10 Sth day 3°8 mm, epiphysial bar present, but no other
iil 9th day 5°0 mm. portion of the cranial roof; symplectic
12 9th day 5*4 mm, | elongated ; palatine process attached
13 a—b | 9th day 57 mm, to the ethmoid.

14 | 11th day 63mm. SracE I1I.—Supra-occipital portion of
15 | 11th day 6:4 mm. cranial roof is formed; posterior
16 11th day 6°6 mm. cranial fontanelle not divided into two
ay — 72 mm. iateral ones ; hyomandibular merely
18 = 7°3 mm. a rectangular plate. No ossification
Tks) i) 90mm. _|_ of cranial cartilage.

|
|
}

* Embryos that were fixed whilst in the egg.

506 H. H. SWINNERTON.

| No. Age. Length. Description of Stages.
90 | — 11:0 mm. = |Stace 1V.—Posterior cranial fontanelle
21 _ 140 mm. divided into two lateral ones ; hyo-
22 — 16:0 mm. mandibular the same shape as in the
| 23 _ 17:0 mm. adult ; all cartilage and dermal ossifi-
| 24 — 21:0 mm. cations present ; chondrocranium still
| 25 — 25°0 mm. largely cartilaginous.
26a—d| — Upto 50:0 mm. |Srace V.

For material used in investigating general questions
arising out of the study of the stickleback’s skull, I have
relied upon the teaching collection of the Royal College of
Science, on a series of microscopic preparations which
were given to the College by the family of the late Dr.
Pollard, and on the fish skeletons now being incorporated
in the collections of the British Museum of Natural History.
The first two of these were kindly placed at my disposal by
Professor Howes, and the last by Mr. Boulenger.

The methods pursued were the same as those used when
investigating the skeletogeny of Sphenodon in conjunction
with Professor Howes (01) ; and as they are described with
some detail in our memoir on that animal, I shall not repeat
them here.

For embryonic, larval, and young material I found
Perennyi’s fluid more useful for fixing purposes than either
picro-sulphuric or corrosive sublimate, because it was less
hable than these to bring about distortion.

For adults 4 per cent. formalin was used.

III. Descriprive AND COMPARATIVE.
The Cranium.

Stage I.—Up to the end of the fifth or even early on the
sixth day the only definite representative of the skeleton in
the head of the stickleback is the notochord, the cells of

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 507

which have already taken on that highly vacuolated structure
so characteristic of advanced chordal tissue (PI. 30, fig. 25).
On the former day, whilst the medullary groove is still open
in front, the notochord runs parallel to the egg-shell for the
whole of its length; on the latter a slight ventral and also
lateral displacement of the anterior or cranial end has taken
place. From this time onwards to the time of hatching, this
flexure increases no further, but the chorda merely moulds
itself to the shape of the brain floor (fig. 58, ch.’).

Balfour (78, p. 210) has described a similar but much more
strongly marked flexure in Hlasmobranchs, and has figured
it for embryos of the Torpedo, Pristiurus, and Scyllium.
In them it may take place through “an angle of 180°,” and
“is not directly caused by the cranial flexure.” Nor is
the slight flexure above described in the stickleback due
to this cause, for here, under normal conditions, I fail to
find any trace of a cranial flexure; the cause should rather
be sought for in this case in the rapid development and
consequent great increase in size of the brain, and in the
presence of a rigid zona radiata (fig. 56 **) which prevents
this organ from expanding dorsally.

Under certain circumstances, however, the cranial notochord
may exhibit an extremely well-marked ventral flexure. In
figs. 56 and 57 are represented, diagrammatically, medium
longitudinal sections through embryos of exactly the same
age, viz. three days before the time for hatching ; but whilst
fig. 56 represents one killed within the egg, the other is
taken from one which was killed after being artificially
released. In the former all the parts of the head were
packed closely together, and the head itself flattened out
upon the yolk (figs. 54 and 56); in the latter (figs. 55 and 57)
it is no longer flattened but distended, and has undergone a
well-marked ventral flexure, chiefly through the region of the
hind brain (b.h.). ‘The roof of the mid-brain (b.m.) con-
sequently comes to be the most anterior point of the body,
the axis of which is indicated by the arrow. This flexure is
still more conspicuously shown by the notochord, which,

508 H. H. SWINNERTON.

from being slightly flexed (as shown by the dotted outline,
fig. 57, ch.’), has now become bent back almost on itself (ch.).
When three days later the larva hatches out, all sign of
flexure has disappeared from the brain (fig. 58), and the
intercranial notochord itself has resumed approximately the
same shape as that which it exhibited within the egg (fig. 56)
on the sixth day.

In a typical Elasmobranch’s egg, that of Scyllium, the
embryo is at liberty to move about considerably, and there-
fore exists under conditions practically the same as those in
which the prematurely released or hatched stickleback exists.
Comparison of the structural peculiarities already referred to
in the Hlasmobranch embryo must accordingly be, not with the
imprisoned, but with the released stickleback.

As to the cause of this flexure exhibited by the free
embryo (fig. 57), it is natural to expect that an elastic
body like the brain filled with fluid should, when liberated
from the pressure of the zona radiata, lose its flattened
form and assume a distended one; but it is not so obvious
why such an alteration in external conditions should, on the
sixth day, produce a strongly marked flexure, and on the
ninth a slight straightening out (fig. 58).

It is certain from the facts already put forward that the
cause for these unexpected effects is one which can be held
in abeyance by mechanical means, such as the resistance of
the zona, which disappears with advancing development.
The coincidence of the time of appearance of skeletal elements
with the time of great flexure suggests that these may be one
factor in that cause. Another factor is perhaps to be found
in the great difference in the degree of development between
the dorsal and ventral portions of the brain; for whilst the
latter is very massive, the former is at present little more
than an epithelium (figs. 54, 55, b.h.). A comparison of
figs. 57 and 58 will show that on the sixth day the skeletal
elements (fr. and ch.), as compared with the brain, are
relatively much shorter than on the ninth day. The fact
that in the released embryo (fig. 57) the optic nerve (11),

MORPHOLOGY OF TELEOSTEAN HEAD SKELBYTON. 509

pituitary body (pé.),and infundibulum (inf.}, notwithstanding
the great distension of the head, have undergone no change
of position in a longitudinal direction relatively to the
trabeculae and notochord, proves that a close connection
exists between these and the floor of the brain. The presence
of short, comparatively inelastic elements side by side with
a longer and more highly elastic and distensible one, such as
the brain, supplies us with structural and mechanical relations
analogous to those found in the Brequet’s thermometer. In
the head of this sixth-day embryo, under a change of external
conditions, the long, weak, unrestrained dorsal portion of the
brain, from the optic nerve (11) to the line (*), distends more
than the short, strong, restrained ventral portion; and the
result is the well-marked bend. As development advances,
these opposite conditions are done away with. The tissues
of the upper surface become stronger, and that portion of the
ventral surface supported by the trabecule (t7.) and notochord
(ch.) more extensive. Consequently the influence exerted by
the latter is no longer one of restraint, but of extension ;
and therefore, at the time of hatching (fig. 58), the whole
head straightens out from the curvature enforced upon it by
the zona (**),

It would be digressing too far from the main line of the
present memoir to pursue this interesting topic further; suffice
it to point out that these facts, by showing the existence
of the necessary mechanical conditions, lend support to
Sewertzoff’s hypothesis concerning the influence of the state
of development of one organ upon that of another (99, p. 319) ;
that, as shown by the observations of Dendy and Howes (99),
Howes and Swinnerton (01), upon the eggs of Sphenodon,
other pressures than the mere passive resistance of the shell
exist; and that possibly prematurely released embryos may
undergo other distortions, the study of which would throw
hight upon the mechanics of development.

It has been hinted above that the first traces of the
head skeleton, other than the notochord, appear during the
latter part of the sixth day. In my youngest embryo which

VoL. 49, PART 4,—NEW SERIES. MM

510 H. H. SWINNERTON.

shows any sign of these they have already practically
assumed the condition shown in figs. 1 and 6 (PI. 28),' the
main difference being that the cartilaginous portions are not
yet so extensive.

On either side of the ventrally flexed notochord (fig. 1, ch.)
the parachordals have become differentiated as long plate-like
tracts of skeletal tissue, in each of which there are two
distinct regions of chondrification (p.ch., oc. e.) separated by
an intervening area of pro-cartilage, and corresponding to
Sewertzoff’s mesotic and occipital sections (99, p. 310). In
embryos of the seventh day the whole length of each tract
is well chondrified, but these two regions may still be easily
distinguished.

The mesotic section (p. ch.) extends nearly three quarters
of the length of the tract, and anteriorly its inner border is
separated from the notochord. From its outer border arises
a stout process which expands distally to form the auditory
capsule. At the same point a ridge starts which divides the
section into two parts: an anterior which supports the brain,
and a posterior or otic portion which helps to support the
auditory organ.

The occipital section (oc. e.) is much shorter, and shows
none of those signs of segmentation recorded by Sewertzoft
for Acanthias. Like the hinder part of the other section and
the intervening pro-cartilage, it hes in contact with the side
of the notochord.

In the salmon, according to Stohr, the auditory capsule
arises first as a plate of pro-cartilage, under the auditory
vesicle, joined to the parachordal by fibrous tissue; in
Kmbryo III this plate becomes chondrified ; in Embryo IV,
which corresponds to Parker’s Stage V, it becomes connected
anteriorly with the end of the parachordal by a strip of
cartilage, and thus assumes the condition above described
for the stickleback. Since the larve of Parker’s Stage V

In following the deseription of this stage it should be borne in mind that

the model from which figs. 1 and 6 were taken was made from an embryo of
the prematurely released type.

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. O11

had been hatched a fortnight, and correspond in other
respects to my Stage III, it is evident that there has been
some influence at work hastening the development of this
particular region. The stages during which the otic element
should be independent thus seem to have been completely
suppressed. On the other hand, such an early connection of
the auditory capsule with the basis cranii is a by no means
exceptional feature, for in Acanthias (Sewertzoff, 99, p. 284)
it arises as an outgrowth of the latter; and in Sphenodon
(Howes and Swinnerton, 01, pl. iii, figs. 1 and 3), whilst
this region is still pro-cartilaginous, it is quite continuous.

From the anterior dorsal border of the auditory capsule
there arises the pro-cartilaginous post-orbital process (fig. 6,
sb. p.) which represents the supra-orbital bar described by
Parker in the salmon (73, p. 129), and by Sewertzoff in
Carassius (99, p. 312), and is all that is present of the
sphenoid region. Of the separate alisphenoid plate described
by Sewertzoff in Acanthias (97, p. 413) there is no sign.

Turning now to the prechordal elements (fig. 1, tr.) of the
cranium, the trabeculz are already as well differentiated as in
-Stohr’s salmon, No. III. They are long and rod-like, and
chondrified for the greater part of their length. Anteriorly
each passes imperceptibly into an expanded plate of pro-
cartilage (e.) which shows signs of bemg united to its fellow
by tissue rich in nuclei. The anterior halves of the trabecule
run parallel to one another; the posterior diverge (pt. f.) to
enclose the infundibulum and pituitary body, and end in a
slight enlargement.

The relation of the hinder ends of the trabeculee (¢7.) to
the parachordals on the sixth day is somewhat variable. ‘The
commonest condition is that in which the trabecule pass con-
tinuously into the anterior end of the parachordals by means
of a tract of pro-cartilage. Two seventh-day embryos
(Nos. 7, 8) of this stage show a different condition ; in the
first both the trabecule are separated from the parachordals
by a comparatively wide space (as indicated by the unshaded
portion in fig. 1); in the second the trabecula of the one side

512 H. H. SWINNERTON.

shows the usual condition, that of the other the unusual.
In the other seventh-day embryos there is no discontinuity
between the chordal and prechordal portions of the cranium,
though the two regions can be easily recognised by their
difference in calibre.

Stage II]—As already described, the intercranial noto-
chord is no longer strongly flexed, but is now almost in a
line with that of the trunk, and has, indeed, undergone a
straightening out (fig. 58, ch.) in the process of hatching.

Comparison of a dorsal view of the cranium (fig. 2) at this
stage with that of the previous stage (fig. 1) gives, at first
sight, the impression that the notochord (ch.) has undergone
a reduction. ‘This, however, is only an appearance due to
the more rapid growth of the surrounding parts. By com-
paring median longitudinal sections (figs. 57, 58) of the
respective stages, it becomes evident that the notochord (ch.)
not only does not decrease, but undergoes an actual increase
in absolute length.

The parachordal tracts (fig. 2, p. ch.), which are no longer
divisible histologically into mesotic and occipital sections,
have grown considerably in all dimensions, quite idepen-
dently of the notochord (ch.), and their anterior ends (¢7.’)
are now situated far in front of the end of this. The space
(p. ch.’) situated between them, which received the freely
projecting half of the notochord in the first stage, has been
carried forward, and now receives only its extremity. ‘his
interparachordal space, as in the salmon (Parker, 73,
p. 129), is continuous in front with the pituitary fossa (pé. f.).
The anterior lateral process of the parachordal has also
undergone a similar forward shifting, by means of which it
now lies in a level with the extremity of the notochord. ‘The
oblique ridge, in the mesotic region, is less strongly marked,
and owing to the increased width of this it no longer extends
to the notochord, but runs along the middle line of the
parachordal plate to the occipital section. The otic portion
of the plate, which may be regarded as part of the auditory
capsule, occupies a lateral and not a postero-lateral pesition

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 5138

relative to the rest of the plate, and dips down somewhat
below the level of the flat basis cranii which the remainder of
the parachordal forms with its fellow.

The occipital section has now sent up on either side a stout
occipital arch (figs. 2 and 7, oc. a.). On the outer side and near
the base of this is a slight projection almost like a transverse
process (oc. a.’). In front of the arch lies the exit of the vagus
nerve (IX and X), whilst all the so-called occipital nerves
pass out entirely behind it.

The anteriorand external wall of theauditory capsule (av. c.)
has extended backwards, underneath, and along the outer
side of the auditory vesicle to the level of the occiput ; but in
the cranium figured has no connection with this. In that of
a slightly older larva, it is continuous both with the lateral
process of the occipital arch (oc. a.’) and with the postero-
lateral border of the mesotic region, thus forming a complete
boundary around the exit for the ninth and tenth nerves (IX,
X). In this larva, also, there are two fenestrae in the floor of
the auditory capsule (cp. fig. 2, fe.’) separated by a bridge of
cartilage which is incomplete in the younger one. This
bridge is formed in correlation with a slight elevation of the
floor of the auditory capsule. This elevation is more strongly
inarked posteriorly by the presence of a pillar of cartilage
(aw, c.’), around which the external semicircular canal runs.
The whole floor of the capsule fits the bottom of the auditory
labyrinth, and seems to have been moulded to it, so that where
this bulges the cartilage is thinned out, and fenestre (fe.’)
are left; where it is constricted a cartilaginous ridge is formed.

The anterior edge of the auditory capsule is greatly
thickened. The pro-cartilaginous post-orbital process (fig. 6,
sb. p.), which has elongated and become more completely
chondrified (fig. 7), arises from the dorsal end of this thicken-
ing; whilst a similar but more delicate process (0. pr.’), lying
between the exits of the fifth and seventh nerves, projects
from the ventral end. ‘I'his thickening is undoubtedly to be
compared with the alisphenoid region of the salmon, where it
arises as a “ growth of cartilage downwards from the (supra-

514 H. H. SWINNERTON.

orbital) band, and forwards and inwards from the ear-sac ”
(Parker, 73, p. 156).

The trabecule, now made up wholly of hyaline cartilage,
have fused with one another anteriorly to form a broad
ethmoid plate, which corresponds to Parker’s “subnasal
lamina” and MeMurrich’s “ rostral plate” (88, p. 625). The
brain, which in previous stages extended considerably beyond
this region (figs. 56 and 57, b. f.), has now relatively retired,
so that the prosencephalon lies upon this plate. From its
lateral borders arise two large processes (fig. 7, e. p. c.), which
by reason of their position and ultimate fate may be con-
veniently termed the parethmoid cornua, and the line
joining their posterior margins may be regarded as the
posterior boundary of the ethmoid plate. This can then be
described as rectangular. For reasons which will appear later
it will be useful to apply the name pre-ethmoid cornua
(e. pr.) to the anterior angles. Behind the ethmoid plate
the trabecule are united for a short distance, and are then
separated by a narrow slit, which widens gradually to form
the pituitary fossa (fig. 2, pt. f.).

The first trace of a cranial root has appeared in the form of
a transverse bar (ep. c.), consisting of a pair of chondrifica-
ticns which in older specimens are connected by pro-cartilage.
It lies in the dorsal fissure which separates the fore and mid-
brains (fig. 58, b. f. and b. m.). Immediately in front of it
lies the epiphysis; consequently it is to be regarded as the
homologue of the epiphysial bar described by Pollard (95,
p. 414) in the Siluroids, and by Sagemeh] in the Characinidee
(85, p. 41) and Cyprinidee (91, p. 511).

Stage II].—Apart from the absence of ossifications the
chondrocranium has now assumed practically the adult condi-
tion. The intra-cranial notochord has undergone no further
change beyond a slight increase in absolute length. The
interparachordal fossa has been carried some distance in front
of the notochord; and the parachordals themselves have
united across the intervening space and across the ends of
the notochord (c. p., dotted outline fig. 3) in such a way that

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 519

this projects below, but lies quite close against the basis
cranii.

Outside each mesotic ridge lies the floor of the auditory
capsule, which, owing to the disappearance of its fenestra
(fig. 2, fe.’), is no longer divisible into parachordal and external
portions. The outer border of the capsule (fig. 8) has
remained unaltered anteriorly, but posteriorly it has grown to
the extent of forming a roof over the auditory and cranial
cavities—a, process in which the occipital arch has taken very
little part. At the same time the ridge which separated the
two fenestrz has become more marked; whilst its posterior
pillar (ep. figs. 2, 8, aw. c.’) has grown up and fused with a
similar downgrowth from the roof. From the point at which
these meet, another pillar, over which the posterior semi-
circular canal runs, has grown back towards and united with
the occipital arch (fig. 8, dotted outline near fe.”). The
whole complex presents internally essentially the appearance
seen in the adult (PI. 29, fig. 20). The capsule as a whole is
so moulded to the shape of the auditory organ that the course
of the semicircular canals can be traced even externally. ‘Two
fenestrae (fig. 8, fe.”) have been left in the side walls.

The sphenoidal region is no longer a mere thickening of the
anterior auditory wall. ‘he post-orbital process (fig. 3, sb. p.)
above is at its maximum development, and the pro-otic process
(0. pr.) below is broad, plate-like, and fused with the outer
border of the parachordal. Thus it completes the anterior
boundary of the foramen for the seventh nerve. In front the
trabecule become increasingly flatter and broader, and are
fused along their contiguous margins in such a way as to
form a ridge above and a groove below.

‘he ethmoid region (e.) is still rectangular, but as a whole
it has elongated, thus separating the pre-ethmoid (e. pr.)
and parethmoid cornua (e. p. ¢.) from one another more dis-
tinctly. The latter are now expanded and wing-like, and pass
at their upper extremities into the supra-orbital bands (sb. b.).

In association with the withdrawal of the prosencephalon
from this region, a median process-lke upgrowth (e.m.) of

516 H. H. SWINNERTON.

the ethmoid plate—the nasal septum of the salmon (Parker,
73, p. 128)—hasappeared. At its upper end this mesethmoid
cartilage expands laterally and fuses with the upper ends
of the parethmoid cornua.

The ethmoid region of the larval salmon (25 mm.) presents
one or two important differences from that of the stickleback.
Looked at from beneath it is broad but not rectangular, for
whilst both cornua (PI. 31, fig. 47, e. pr., e. p. ¢.) are recognis-
able, the anterior border is not straight, but has grown out into
a triangular rostrum (7.). At the same time on the dorsal sur-
face the nasal septum has extended to the anterior end of this.

In the stickleback the epiphysial cartilage (Pl. 28, fig. 3,
ep. c.) has grown backwards, but the epiphysis is still closely
related to its anterior edge. Laterally it is continuous with the
hinder ends of the supra-orbital bands (sb. b.), and thus closes
in a space which may be called the anterior fontanelle (a. /.).
In the roof of a salmon 20 mm. long (fig. 61) this cartilage is
expanded laterally, and, except for its continuity behind with
the post-orbital process (sb. p.),and for the fact that the prosen-
cephalon overlies the dorsally excavated mesethmoid cartilage,
the whole roof presents the same condition as that above
described for the stickleback. In an older salmon (25 mm.)
the anterior fontanelle has become covered in by an inward
growth of the epiphysial cartilage and supra-orbital bands,
except for a smaller space over the epiphysis (ep. f.) and a larger
one (a. f.) at the anterior extremity which corresponds to that
marked m..c. by Parker (78, pl. iv, fig. 2, and pl. vii, fig. 4)
and to the median fontanelle described by Winslow (98, p. 186)
in the trout. This fontanelle is now bounded in front by
the more fully developed mesethmoid cartilage, now no longer
overlapped by the prosencephalon.

Through the kindness of Professor Howes I have been
enabled to examine the head of a larval Amia 19 mm. long.
The roof of this was in exactly the same condition as that
just described for the 25-mm. salmon, even to the presence
of the small foramen above the epiphysis. It is safe, therefore,
to assume that here also the tegmen cranii is formed, not by

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 517

back growth of the ethmoid region, but by inward extension
of the supra-orbital and epiphysial cartilages over a pre-
existing fontanelle.

In Engraulis and Atherina, Pouchet (78, pl. x, figs. 44, 49)
found both fontanelles. he anterior one is in precisely the
same condition as in Gasterosteus ; but the posterior one of
Engraulis is divided medianly by a strip of cartilage into
two lateral fontanelles. A transverse epiphysial bar is present.

I have found this bar in Siphonostoma also (fig. 48, ep. ¢.).

‘These facts make it evident, therefore, if we take develop-
mental stages into account, that the epiphysial cartilage, so
far from being confined to the Ostariophysi, as Sagemehl (91,
p. 576) thought, is an element widely distributed among
Teleosts, and probably derived by them from an ancestor
with no other cranial roof.

In the stickleback it never undergoes any further extension
forwards than is shown at this stage. If it is safe to assume
that, because primitive 'leleosts have a well-developed tegmen
eranil, this specialised little fish has lost it phylogenetically,
we have, in the epiphysial cartilage, an interesting example
of an old character which has survived all the vicissitudes of
a shortened ontogenetic record.

In the oldest individuals of this stage all the membrane
bones, except the mesethmoid, nasals, and parietals, have
appeared; but there is no marked ossification of the chondro-
cranium itself.

Stage 1V.—The notochord has undergone only a slight
increase in length, as a comparison of figs. 25 and 26 (PI. 30)
will show; but it has now begun to assume characters which
anticipate the adult condition. In front it has the same
proportions as at previous stages, whilst behind it has become
expanded and funnel-shaped (fig. 26, ch. s.) im preparation
for the articulation with the first vertebra. Internally it is
no longer filled with a few highly distended vacuolated cells,
but, especially at its anterior end, it consists of a great number
of collapsed cells of a similar type crowded together. Where
the extra cells arise is uncertain, for intercranially the chordal

518 H. H. SWINNERTON.

epithelium shows no signs of active growth even in younger
individuals of this stage. In the articular region, however,
this is rich in nuclei, and appears to be actively giving rise
to the substance of the thick fibrous sheath externally and to
cells internally ; an appearance which suggests that cells from
this region may have crowded forwards. Such an interpre-
tation is supported by the fact that those cell walls which are
attached to the chordal sheath all have their inner portions
inclined forwards, as though there had been a movement in the
same direction on the part of the central cells. Hxternally it is
completely enclosed in a casing of a bone—the basioccipital,—
whose relation to the notochord was accurately described by
Huxley, and compared by him with the urostyle (58, p. 440).

Apart from the appearance of numerous ossifications, the
changes which at this stage have taken place in the chondro-
cranium are few (Pl. 28, fig. 9). Owing to the suppression of
part of the supra-orbital bands, and of those parts of the
trabeculz which border on the pituitary fossa, the two parts
of the chondrocranium which are essentially trabecular and
parachordal in origin have become completely separated from
one another.

In the hinder or parachordal portion the interparachordal
fossa (fig. 12, p. ch.’) has been carried so far away in front of
the notochord that the plate formed by the median union
of the parachordal now furnishes a considerable portion of
the basis cranii. Those parts lying immediately on either
side of the fossa have now begun to undergo a movement
of depression, by which they have already come to le
slightly below the level of the basis cranii.

‘his is perhaps associated with a similar movement on the
part of the recti muscles. In the previous stage the inner ends
of these were inserted into one another and into the tissue which
fills the hinder part of the fossa. Owing to the point of their
insertion on the eyeball (fig. 58, im. e.) being ina plane situated
a little in front of this, but considerably above that of the basis
cranii, their course outwards over the anterior ends of the para-
chordals was obliquely upwards. In this stage, however, their

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 519

posterior insertion has come to lie under the hinder border
of the fossa (fig. 38, ime., o.pr.), whilst that on the eyeball is now
tending to pass down with the latter to the level of the basis
eranii. In consequence of this tendency, which finds its con-
summation in the adult (fig. 59, me.), the course of the muscles
is still more obliquely forwards and less obliquely upwards.

Passing to the anterior or trabecular portion of the chondro-
cranium, the ethmoid plate (fig. 9, e.) has greatly elongated.
The connections between the mesethmoid and the parethmoid
cornua have become more extended, and give the appearance,
when viewed dorsally (fig. 4), of a rudimentary tegmen cranii.
In front of the mesetlimoid the ethmoid still retains its original
plate-like character, but the pre-ethmoid cornua (e. pr.) have
now become slightly drawn out dorso-ventrally. Fig. 30, which
representsa section through the pre-ethmoid region (e.), passes
through the extreme upper end of this process (e. pr.) ; a
few sections further back the two parts become continuous.

Apart from its great length the ethmoid of Siphonostoma
(fig. 48, e.) is just the same as in the stickleback.

Concerning the regions in which, in the stickleback, sup-
pression has taken place, remnants of the supra-orbital
bands are still seen in the backward processes of the rudi-
mentary tegmen cranii; whilst those parts of the trabecule
which in the previous stage were fused together, now remain
as a posterior prolongation of the ethmoid plate. Of the
parts which have disappeared no trace remains; nor is any
cause forthcoming, unless it be that they are functionally
replaced by the frontals and parasphenoid respectively.

The epiphysial cartilage, now quite free of the ethmoid
region, has grown considerably posteriorly, and become
connected with the hinder cranial roof by a thin strip of
cartilage running through bone (¢. p.s.). This strip is the
homologue of the massive cartilage which lies between
the small lateral cranial fontanelles in the adult salmon,
and of the small process (fig. 62, e. p. s.) of the epi-
physial cartilage which projects back into the posterior
cranial fontanelle (p. f.) of the larva. The condition in

520 H. H. SWINNERTON.

the stickleback is now identical with that described and
figured by Walther (82, p. 10) for the young pike, and
Pouchet (78) for Engraulis.

All the ossifications present in the adult are now repre-
sented, but as yet are widely separated by cartilaginous areas.

The basioccipital (fig. 12, oc. b.), owing to the hinder
funnel-shaped enlargement of the notochord, has begun to
take on the usual centrum-like condition. At its extreme
anterior edge the ventral lamina cf bone leaves the cartilage
and projects freely forwards, as indicated by the yellow
pointed portion in fig. 12, and is better understood from the
longitudinal section of this region of the adult (PI. 30, fig. 37,
ey. c.). ‘This free portion lies between the posterior pro-
longations of the parasphenoid (fig. 36, p. s.), and is formed
by ossification in connective tissue.

The vagus and pneumogastric nerves pass out through the
exoccipital (figs. 9, 12, oc. e.) by a common foramen, as in the
previous stages; but in the individual figured the pneumo-
gastric nerve of one side passed out by a small foramen,
separated from that of the vagus by a narrow bridge of bone.
This is the beginning of a process which continues on both
sides, and ultimately carries the two foramina far from one
another.

Posteriorly the exoccipital, by extending into the sur-
rounding tissue, has enclosed the first spino-occipital nerve
(fig. 9, oc. n.’) completely, and the second (oc. n."’) partially.

The supra-occipital (fig, 4, oc. s.) extends along the strip
of cartilage which joins the posterior cranial roof to the
epiphysial cartilage, almost to the level of the post-orbital
processes. The ossification does not yet affect any part of
the epiphysial cartilage itself, but it extends laterally into
regions where no cartilage exists or has existed during
ontogeny; that is to say, into the connective tissue which
roofs over the fontanelles. It cannot therefore be regarded
as a simple cartilage bone; but it may be that this lateral
extension is indicative of a former greater extension of the
cartilage,

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 621

Of the otic bones (figs 9, 12, 0. ¢.,0. p., 0. sp., 0. pr.) the
opisthotic is absent, and the remainder are represented by
rounded areas of ossification, but are still widely separated
by cartilage. The circular form and wide separation of the
ossifications calls to mind the condition of the same set of
bones in the adult Amia (Sagemehl, 84, p. 202). The main
part of the pro-otic (0. pr.) completely surrounds the exit for
the seventh nerve, whilst its anterior border is deeply
indented for the passage of the second and third branches of
the fifth. The portion which les between these two nerves
is the portion hitherto called the pro-otic process. The front
part of the sphenotic is formed by ossification of the plate-like
extension of the auditory capsule—the sphenoidal region.

In the younger individuals of this stage, except for the
pro-otic, each of these ossifications is confined mainly to the
external surface of the cartilage. Under the central portions
of the epiotic, pterotic, and sphenotic the cartilage has
disappeared, and consequently this outer lamina of bone
comes to form part of the internal wall of the capsular and
cranial cavities. Apart from this, the condition exhibited at
this stage bears a close resemblance to that found in Amia
(Bridge, 77, p. 612), in which the pro-otic is the only bone of
this series visible internally.

In the anterior portions of the skull the expanded plate-like
portions of the parethmoid cornua have given rise to the
parethmoid bones (figs. 4, 9, e. p. b.), whilst a centre of ossifi-
cation, the mesethmoid (e. m.), has appeared on the dorsal
surface of the corresponding cartilage. The edges of the
latter ossification extend freely into the surrounding tissue,
and give the impression of a membrane bone whose
central portion has united with the cartilage, leaving the
edges quite free. Though no further evidence that this has
actually taken place is forthcoming in any of the examples
which I have examined, there can be little doubt that the
impression thus formed is a true one. In the salmon
(Parker, 78, pp. 98, 138) this bone exists even in the adult
as a membrane bone, perfectly free of the underlying

522 H. H. SWINNERTON.

cartilage. In the Characinide (Sagemehl, 85, p. 14) and
Cyprinide (ibid., 91, pp. 497, 499) every gradation exists
from the condition here shown in the young stickleback to
that found in the Scomber (fig. 44, e. m.), in which, as in the
majority of Teleosts, this bone has lost all sign of a
membranous origin, and consists entirely of ossified mes-
ethmoid cartilage.

All the cranial membrane bones are now present.

Adult.—In assuming the adult condition the chondro-
cranium undergoes but little change, beyond a continuation
of those ossificatory processes which had already set in at
the last stage. Consequently, the cartilaginous portion is now
confined mainly to regions where primary ossifications have
met and formed sutures, and to the ethmoid region, where
it is still moderately massive.

Its upper and under surfaces are sharply marked off from
one another by a plane in which lie the pre-ethmoid region,
the parethmoid, and a conspicuous ridge running along the
middle of the sphenotic and pterotic bones. The upper
surface does not present that strongly crested and grooved
appearance so characteristic of Teleosts, but is perfectly even
from end to end and from side to side. It thus presents a
condition comparable to that of Amia. The temporal fossa
is a Shallow uncovered depression, and, as in Amia, is the
only portion of the dorsal surface upon which trunk muscles
encroach. ‘The rest of the roof is covered by a thin skin.
A further similarity to the Amioid or Ganoid condition is
seen in the fact that the surfaces of all the roofing bones
are beautifully sculptured and polished (PI. 29, fig. 18), the
rugze on any one bone generally radiating from a central point.
This sculpturing is a feature by no means universal among
‘'eleosts, and those forms in which it does occur to any
appreciable extent are frequently among the lowhest of the
group to which they belong; e.g. Hrythrinus and Sarcodaces
(Sagemehl, 85, p. 34) among Characinide, Araparima among
Malacopterygii, and to a lesser degree Sphyrena among
higher Teleosts. Nevertheless the condition in such a form

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 523

as Uranoscopus warns us not to place too much rehance on
mere sculpturing.

The intercranial notochord (PI. 30, fig. 27) has continued
to increase somewhat in total length, and by reason of the
multiplication of cells its internal structure, especially at its
extremity, has become denser even than in the previous stage.
In the same region it has become slightly attenuated, but
posteriorly its diameter has more than quadrupled (fig. 27).
Possibly in association with this increased bulk the internal
tissue is much looser and the vacuoles are normal. ‘The
funnel-shaped expansion thus made gives to this region of
the basioccipital, when viewed from behind, that conically
concave appearance so characteristic of it in many 'T'eleosts.

It is thus seen that at no stage in development is there any
sign of a reduction or suppression of the intercranial noto-
chord; the real fact of the case being that during the whole
period of skeletogenesis the length of the notochord increases
only thrice, whilst that of the cranium itself increases approxi-
mately twenty times (cp. figs. 59 and 21, ch.). The natural
outcome of this disproportionate growth is the retirement of
the notochord or, more accurately, the advancement of the
skeletal elements, so that the pituitary and interparachordal
fossee are carried far away in front of the notochord. A
similar apparent retirement has been noted by Parker in
Lepidosteus (816, p. 458) and in Salmo (78, pp. 125, 102).
Since he gives no details, however, concerning the condition
in the adult, it is impossible to make any comparisons as to
the relative rate of growth in these forms.

‘he suture between the basioccipital and exoccipital
(fig. 21, oc. e.) differs from that between the other bones of
the chondrocranium, in being formed by the interdigitation of
bony lamelle. Betweenit and the pro-otic there still remains
(fig. 21) a small area of exposed cartilage. Medianly it is
overlapped by the forked end of the parasphenoid, for the
reception of which it has a deep channel formed by a down-
growth of the cartilage and bone on either side (PI. 30, fig. 36).
From the middle of the roof of this channel the basioccipital

524 H. H. SWINNERTON.

sends down a crest of bone (figs. 86 and 37) which lies
between and under the back processes of the parasphenoid
(fig. 36, ps.). Within the substance of the fore-part of this
crest is a cavity (figs. 36 and 37, ey.c.) which opens in front
and receives the hinder end of the external rectus muscle ;
this is the homologue of the anterior conical excavation of
the basioccipital of the pike (Huxley, 71, p. 133) and of many
other Teleosts.

The exoccipital is separated from its fellow in the middle
line, above by the supra-occipital, and below by the basi-
occipital (Pl. 29, fig. 24). The facet for articulation with
the first neural arch present in so many bony fishes is quite
unrepresented. The internal plate, which in the majority of
Teleosts (Klein, 84, p. 131) shuts out part or all of the
basioccipital from forming the floor of the cavum cranii, is
here not altogether absent, as Klein states (85, p. 133), but is
represented by a slight ingrowth extending anteriorly over
the basioccipital (PI. 30, fig. 36, oc. e.).

The ninth and tenth nerves no longer have a common exit,
but the process which commenced in Stage IV (fig. 12) has
given rise to two foramina widely separated from one another
(fig. 21, IX, X). In Amia, in which these are both present,
that for the ninth les just behind the pro-otic and bears no
relation to the exoccipital, that for the tenth between this
and the opisthotic (Bridge, 77, p. 4). By coming to include
the vagus foramen it evidently trespasses somewhat on the
region of the opisthotic. In the part which immediately
surrounds the foramen magnum ossification has continued to
extend into surrounding tissue, so that the second occipital
nerve (fig. 19, oc. ”.”) now has a foramen also. In this
region in Amia both Bridge (77, p. 611) and Sagemehl
(84, p. 194) recognised the homologues of two, or possibly
three, vertebral arches; Gegenbaur (87) shows that the
elements they refer to are present in some, but not in all
Teleosts. If the two occipital nerve foramina in the
stickleback are indicative of two originally distinct arches,
it seems strange that whilst the first true vertebral arch is

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 525

preformed in cartilage, these are not. Nor does the develop-
ment and structure of the basioccipital or its enclosed
notochord show any signs that centra have been absorbed.
It may be, of course, that this is another case in which
during ontogeny features have been suppressed which were
present in the ancestors of this fish.

The supra-occipital is an unusually large bone (fig. 24, oc. s.).
Anteriorly it seems to partially separate the frontals; in
reality it extends under them (fig. 20) almost to the level of
the outstanding post-orbital processes, and now embraces the
larger part of the epiphysial cartilage itself also. It is quite
absent in Amia. In the salmon, pike, Alepocephalus,
Characinide, and Cyprinide it falls far short of that of the
stickleback in forward and lateral extension. But whilst in
Characinide and Cyprinide, the majority of which have more
completely osseous chondrocrania than the stickleback, the
parietals are never separated by the supra-occipital ; in other
physostomes—excluding the siluroids, in which the parietals
are either absent or are fused with the supra-occipital—there
is every gradation from the primitive Amioid condition, with
its parietals side by side, to the condition in which they are
widely separated. The former is seen, for example, in
Osteoglossum, Araparima, and many Clupeoids; the tran-
sitional forms are seen in some Salmonoids, and the latter
condition is seen in the rest. In the case of Stenodus, a
Salmonoid, the separation or non-separation seems to be
merely amatter of age, forin young individuals the parietalsare
apart, in old ones they are in contact (Boulenger, 95, p. 299).

Allowing for the existence of the peculiar fenestrae between
the frontals and parietals in the Ostariophysi, the constancy
of the approximation of the parietals in them, on the one
hand ; and the inconstancy of the same feature, or the general
separation of these bones in the non-ostariophysous fishes, on
the other; are but superficial expressions of a fundamental
difference between the chondrocranial roofs of these two
great groups. In the former, the supra-occipital can never
extend far forwards, because in the adult the post-cranial

VOL, 45, PART 4,—NEW SERIES, NN

526 H. H. SWINNERTON.

fontanelle is very large (Sagemehl, 85, p. 41; 91, p. 512;
Pollard, 95, figures), and the posterior portion of the cranial
roof is not extensive. Hence the whole bears a close resem-
blanee to the cranial roof of the young salmon (fig. 61) or
stickleback (fig. 3). The approximation of the parietals is
therefore a necessity for actual roofing purposes. In the
non-ostariophysous Teleosteans, however, the posterior cranial
fontanelle, judged by the development of the salmon,
stickleback, pike, Hngraulis, and by the adult Alepocephalus,
always becomes divided by a more or less massive median
bridge of cartilage, and either greatly reduced or completely
suppressed by the growth of this and of the posterior cranial
roof (cp. fig. 4). Thus the supra-occipital ossification has
free scope to advance forward or laterally to any extent, and
consequently the necessity for the parietals as roofing
elements of the cranial cavity diminishes. For the same
reason these bones, as compared with the frontals, are
generally large in one group and small in the other. These
features in themselves are almost important enough to
separate the two groups, and while they lend great support
to the validity of the group Ostariophysi, they also support
Sagemehl’s suggestion that these fish might have been
derived from a type even lower than Amia (85, 91).

Concerning the origin of the supra-occipital, Sagemehl
(91, p. 523) sought to show that it was the dorsal ossification
belonging to the morphological first vertebral arch. Its
position does not, however, support this view, for both in
salmon and stickleback it arises, not between the upper ends
of the exoccipital, but further forward, between the epiotics.
In some cases (e. g. Pleuronectidee) it may be separated from
the foramen magnum not merely by the exoccipital but also
by the epiotics (Klein, 84, p. 131). Whatever its origin, as
Sagemehl pointed out (91, p. 521), it is the youngest bone in
the Teleostean skull, and has arisen wholly within this class.

In the presence of a vertical ridge, and in the relation of
this to the exits of the fifth and seventh nerves, the sphenotic
region closely resembles that of Alepocephalus,

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 527

The pro-otic (0. pr.) has now extended to the borders of the
interparachordal fossa, and has met its fellow in the middle
line for some distance behind that. In the pike, salmon, and
Alepocephalus it completely surrounds the exit of the fifth
nerve, but here that part which lies in front of this foramen
is suppressed, and is functionally replaced by overlapping
processes of the sphenotic, frontal, and parasphenoid.

The process of depression of those parts bounding the
interparachordal fossa laterally has continued, so that this
region now appears to bea mere downward process of the pro-
otic (fig. 20, 0. pr.”), with its cartilaginous extremity mortised
into the sides of the parasphenoid (fig. 35, ps., 0. pr”). This
appearance is enhanced by the fact that posteriorly each
process is continued into a ridge running along the under
surface of the hinder portion of the pro-otic (fig. 20). These
two ridges are continuous with those already described under
the basioccipital, and there is a channel thus formed which
runs a considerable length of the basis cranii, is closed
ventrally by the parasphenoid, and opens anteriorly into the
cavum crani by means of the interparachordal fossa. In the
salmon this canal is much more strongly developed, and
opens posteriorly. Klein (84, p. 155) gives a list of those
families with and those without an eye-muscle canal, but he
errs in including Gasterosteus among the latter. A study of
his lists and of many of the original skulls themselves strongly
suggests that the degree of development of this feature is not
of fundamental importance, but is largely dependent on the
shape of the head—on its depth or its depression.

A similar canal is described by Bridge and Sagemehl, and
with greater detail by Allis in Amia. Sagemehl regarded it
as a portion of the cavum crani, which has become roofed
over by two horizontal lamellz from the pro-otic (84, p. 207).
The same idea is conveyed by him in regard to this feature
in Cyprinide and Characinide, and also by Gegenbaur in
Alepocephalus. Hach of these authors describes the rool
as a bridge formed by lamelle projecting from the pro-otic.
According to Allis, “the eye-muscle canal in Amia and

528 H. H. SWINNERTON.

Lepidosteus, and hence probably in all fishes, is thus an
intercranial space opened secondarily to the orbits” (97, p. 10).
This can apply, however, only to the anterior part of this
canal in Teleosts. In Amia this canal is apparently com-
pletely closed behind and beneath by cartilage. In Alepo-
cephalus only the anterior region is thus closed, whilst the
hinder is, as in the stickleback, closed by the parasphenoid.
In larval Amia this canal is not present, but there is a well-
marked interparachordal fossa to which the eye muscles bear
the same relation as in the stickleback. It is probable,
therefore, that in this fish also a process of depression and
secondary growth goes on on either side of the fossa and
below the pro-otic ; but that, whereas in the other types the
fossa persists and transmits the eye muscles back again out
of the cranial cavity beneath the basis cranii, in Amia it
disappears, owing to continued cartilaginous growth. As far
back as the so-called pro-otic bridge, these muscles may be
said to run in an actual derivative of the cranial cavity ;
behind that they run in an extra-cranial space secondarily
enclosed.

This interparachordal fossa is a common feature among
Teleosts. Sagemehl (85, p. 66) speaks of it as the space
between the pro-otics, and expresses doubt as to the homology
of this with the pituitary fenestra of other vertebrates.
Development justifies his doubt, by showing that the fossa is
related to the parachordals and the fenestra to the hinder
ends of the trabeculae.

Though the Teleostean pro-otic bears the same relations to
the seventh nerve and the auditory organs as in the higher
vertebrates, those of its other parts which meet in the middle
line behind the interparachordal fossa have trespassed upon
the area occupied by the reptilian basisphenoid (Howes and
Swinnerton, 01, pl. iv, fig. 6). The interest of this is increased
by the fact that among reptiles there are traces of a paired
origin for the basisphenoid (ibid., p. 42). When allowance is
made for the fact that, owing to the union of the trabeculee
with the anterior end of the parachordals in 'Teleosts, and

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 529

with the ventral surface in reptiles, the pituitary fenestra has
a relatively different position in the two types, the resem-
blance becomes more striking. It is, therefore, possible to
regard this bone either as a pro-otic which has taken on the
functions of a reptilian basisphenoid; or a basisphenoid which,
like the exoccipital, has taken on relationships to the auditory
organ and a cranial nerve (cp. exoccipital).

In the ethmoid region the general shape of the whole is of
considerable importance. Looked at from beneath (fig. 21),
after the removal of the nasals (na.’), it has the appearance
of a rectangle, somewhat expanded anteriorly and posteriorly,
with its long axis parallel to that of the skull, and, except
for a sight median groove continuous with that formed by
the fusion of the trabecule behind, it is quite flat. The
expanded portions correspond to the pre-ethmoid and pareth-
moid cornua respectively. The former still serves as the point
of attachment for the palatine (fig. 51, pa.’ ; fig. 31, e. pr., pa.).
From the middle of its dorsal surface arises the mesethmoid
in the form of a high longitudinal ridge of cartilage (figs. 32,
42, 51, em.c.). Laterally and in front of this the ethmoid
retains its original plate-like character (cp. figs. 31, 32, 42,
dL,-e.).

In describing the same region in Alepocephalus, Gegenbaur
calls attention to the fact that there it is normally broad, but
its anterior portion is drawn out into a terminally expanded
rostrum (78, p. 4), against which the projecting anterior
portion of the palatine bone abuts (p. 11), the whole differing
from that of the pike only in length (p. 4). Hxcept for the
proportionately greater height of the mesethmoid ridge and
its more complete ossification, exactly the same might be said
of the ethmoid region of the stickleback (cp. Huxley, 71,
p- 182).

The significance of this becomes more obvious when the
corresponding region is examined in the salmon. During
the developmental stages it shows a general resemblance to
that above described, but does not possess a terminal expan-
sion for the attachment of the palatine (Parker, 73, pl. v,

530 H. H. SWINNERTON.

fis. 3,4). In the adult, when looked at from beneath, this
region (ibid., pl. vii, fig. 2) is triangular; and the nasal
septum is very massive, extending not merely to the anterior
end, but also to the lateral edges of the ethmoid plate or sub-
nasal lamina (ibid., pl. vil, figs. 5—8).

In the absence of well-marked laminal or lateral remnants
of the ethmoid plate; in the forward extension of the massive
mesethmoid cartilage to the anterior end; in the removal of
the pre-ethmoid cornu (figs. 44, 45, pa.’) from the end of the
rostrum, the typical Acanthopterygian ethmoid departs from
that of Alepocephalus, pike, and stickleback, and approaches
that of the salmon. here are, however, some Acanthoptery-
gians (e.g. Mesoprion) whose ethmoid is but a modification
of the other type.

Only three primary ossifications are present in this region,
viz. the two parethmoids and the mesethmoid.

The mesethmoid is no longer a mere superficial ossification,
but it has penetrated in the case of old individuals far into
the substance of the cartilage (figs. 20, 42, e. m.). Super-
ficially its central area is coated with a thin layer of peculiar
cartilaginous tissue (fig. 42), to which I shall return when
dealing with the premaxillee.

In Amia, Bridge (77, p. 615) describes two ossifications in
the antero-lateral angles of the subnasal lamina, and regards
them as the homologues of the “ paired endosteal ossifications
which are to be found at the distal end of the great pre-nasal
rostrum in the pike.’ He also suggests that they are homo-
logous with the “ septo-maxillary bone, described by Mr.
Parker as existing in the floor of the nasal capsules in the
frog.’ Sagemehl (84, p. 204) agrees with him in homolo-
eising this bone with that found in the pike; but though he
regards its homology with the septo-maxillary of the frog as
doubtful, he does not venture to give it a new name. In his
posthumous work on the Cyprinide two similarly situated
ossifications are described, but here, though they are quite
enclosed in cartilage, they are referred to as septo-maxillary

bones (91, p. 511). Allis (98, p. 446, et seq.) has recently

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 531

summarised all that is known concerning its presence or
absence in fishes, and describes it in considerable detail for
Amia. Throughout he refers to it as the septo-maxillary.
Thus it has come about that to a bone in fishes undoubtedly
endosteal, there has been applied a name given by Parker to
one whose precise nature in the frog he does not state, but
which he classes among the “ true dermal bones” in Lacerta
(78, p. 601). The same bone is present in Sphenodon
(Howes and Swinnerton, 01, p. 56), and is there undoubtedly
a true dermal bone.

A somewhat similar state of affairs exists for other bones.
In higher vertebrates the squamosal, post-frontal, and pre-
frontal are purely dermal bones; but the bones in the 'T'eleos-
tean skull to which these names are frequently applied, viz.
pterotic, sphenotic, and parethmoid, are cartilage bones with
a possible dermal origin. ‘That the original dermal elements
in the latter case are the homologues of the dermal elements
in the former is still an unsettled question ; but even supposing
them to be homologous, it has still to be decided whether
the ossification in cartilage which has lost all sign of its dermal
origin should be regarded as the homologue of the purely
dermal ossification.!. For these reasons the latter terminology
has been adopted throughout this memoir.

The particular bone we are now considering should, there-
fore, have given to it some name which involves no doubtful
homologues, but accurately expresses its topographical rela-
tions. Following the terminology which is applied to other
ossifications of the ethmoid cartilage, I would suggest the
name Pre-ethmoid for it, as expressive of its relation to the
pre-ethmoid cornu.

1 In cases where actual genetic relationship has been established between
dermal and cartilage bones, e.g. squamosal of Amia and pterotic of Teleosts,
such relationship could easily be expressed, and at the same time distinction
be maintained, by a more extended use of the prefixes dermo- and chondro—
thus, dermo-pterotic, chondro-pterotic.—the former being used for bones which
are quite free of the cartilage, and the latter for those which involve cartilage,
irrespective of the degree of ossification of this, or of the retention of dermal
characters (ep. Bridge, 77).

ae H. H. SWINNERYON.

The pre-ethmoid bone is absent in Gasterosteus.

The fact that it is present in Belone (fig. 50, e. pr. 6.), on
either side of the mesethmoid, suggests that it may be more
generally present than is supposed.

The parietals (fig. 18, par.) overlap the supra-occipital, the
epiotic, the pterotic, and the sphenotic, and are widely
separated from one another. Owing to the presence of
cartilage beneath it and between these bones (fig. 20) it is
completely shut out from taking any part in the roofing of
the craniuni.

The frontals (fig. 18, fr.) roof over the anterior and lateral
cranial fontanelles (cp. figs. 4, 5,18). Immediately in front
of the sphenotic each sends down a process which meets an
ascending process from the parasphenoid. ‘These two over-
Jap in such a way that the latter predominates externally.
A similar relation has been noticed by Klein (84, p. 145) in
Uranoscopus and Lophius.

The anterior end of each nasal (figs. 10, 19, na.’) sends
down a process which forms the anterior boundary of the
nasal aperture, lies directly on the ethmoid cartilage
(fig. 31, na.’), and wraps round on to the ventral surface of
this behind the pre-ethmoid cornu (fig. 22, na.’), and thus
gives the appearance, when viewed ventrally, of being an
anterior ossification of the ethmoid.

The two lateral wing-like processes of the parasphenoid
(fig. 21), by reason of their position in front of the exit of
the fifth nerve, cannot be regarded as the homologues of the
similar processes in Amia. Between them this bone is
pierced by two foramina for the carotid arteries.

The only representative of the sphenoidal region seen
during development is the anterior plate-like extension of
the auditory capsule, which passes above into the post-
orbital process, and below into the bridge between the exits
of the trigeminal and facial nerves. ‘The upper part of this
ossifies to form the sphenotic (0. sp.) ; the lower, the pro-otic
(o. pr.); whilst the post-orbital process, which in other
‘'eleosts forms part of the alisphenoid, remains unossified.

MORPHOLOGY OF TELEOSTHEAN HEAD SKELETON. 533

Orbitosphenoid, alisphenoid, basisphenoid are absent. Con-
sequently the cranial cavity opens anteriorly into the orbital
region by an extremely wide aperture.

The Visceral Skeleton.

Stage I.—The branchial apparatus (PI. 28, fig. 6) at this
stage consists of four pairs of simple rods of cartilage (br. 1—4),
each widely separated from its fellow. Lying in the ventral
space between the members of the first and second pairs is
another rod (br. b.), the homologue of the similarly situated
one in the embryo salmon—the copulare commune of Stohr
(83, p. 7). Unlike that, however, it does not extend in
between the hyoid arches.

The first pair of branchial arches is the most advanced, the
fourth least advanced, and the fifth is as yet quite unrecognis-
able. Above the roof of the buccal cavity, i. e. where the future
pharyngo-branchials will appear, there is no sign of skeletal
rudiments.

Thus this portion of the visceral skeleton presents, at this
stage, exactly the same condition as it does in the youngest
salmon described by Stohr (82, p. 2); and if, for convenience’
sake, we may take the development of the latter as a
standard of comparison, it has not attained to the same
stage as the parachordal tracts.

In the hyoid arch all the parts are represented except the
hypohyals and basihyal. The lower part consists of a stout
bar of cartilage (hy.) representing the future ceratohyal and
epihyal, which above is attached by the slender pro-carti-
laginous stylohyal to the upper part, and below is separated
from its fellow by a mass of undifferentiated tissue.

In the upper part the hyomandibular (hym.) and sym-
plectic (sym.) regions are not differentiated by an intervening
area of pro-cartilage, as in the salmon, but together form a
continuous chondrified plate. ‘To the middle of the hinder
border of this the stylohyal is attached. This point of con-
nection serves to denote the boundary between the hyo-

534 H. H. SWINNERTON.

mandibular and symplectic, which is not indicated in any
other way—not even by any sign of the sharp bend present
in the adult (fig. 23, hym., sym.). The axis of the plate pro-
jects obliquely forwards and downwards from the auditory
capsule, but not at so great an angle as in the later stages.
Perhaps it is 1n association with this that the line of attachment
to the auditory capsule, instead of being horizontal, is almost
vertical, and it is not marked by any change of structure
indicative of a future articulation, but hyomandibular and
capsule together form a continuous whole (fig. 41, au. c. hym.).

In the mandibular arch the quadrate (qu.) and Meckel’s carti-
lage (mk.) are already marked off by definite signs of chon-
drification from the dense tissue which hes between them.
Anteriorly and posteriorly the quadrate passes imperceptibly
into pro-cartilaginous processes. The former, the palatine
process (qu. pa.), is long and slender, and extends halfway
towards the free end of the adjoining trabecula; from its
fellow of the other side it is proportionately more widely
separated than at any later stage. The latter, the
metapterygoid process (qu. m.), is stouter, and lies with its
lower border close to, and parallel with, the upper border of
the symplectic.

Meckel’s cartilage, as regards its general shape, is com-
parable to that of Stohr’s Embryo No. II (83, p. 10), but as
regards completeness of chondrification it more closely
resembles his larva No. IV. In shape it is slender, and
meets its fellow in the middle line (figs. 1, 6, mk.) without the
intervention of pro-cartilage.

Thus the general developmental processes of the first two
visceral arches agree closely with those of the salmon, but in
the time at which these take place there is a considerable
difference. Here, as with the fore-part of the auditory
capsule, there has been some influence at work hastening
events. ‘l'hus, whilst the branchial arches at this stage are
comparable with those of Stéhr’s youngest embryo, these two
arches, by the absence of pro-cartilage between symplectic
and hyomandibular, and the elements of the mandible; and

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 9030

by the presence of a well-developed pro-cartilaginous palatine
process of the quadrate, exhibit the condition found in a
young salmon intermediate between Stohr’s Stages III and
IV, of which the latter had been hatched some days. It
seems, therefore, that the simple quadrate without palatine
process, described by Stohr (88, p. 9) in salmon, and Pollard
(94, p. 69) in Gobius and Blennius, is either very transitory
or entirely omitted in the development of the stickleback.
These parts, which have been accelerated in development,
serve aS a means of support to the mouth and operculum, or
as a means of attachment to the associated muscles. This
suggests that the acceleration may be due to these features
being of greater importance to this fish at an earlier stage
than to the salmon. In the stickleback there is practically
no alevin stage ; it hatches out on the ninth day, and by the
end of the eleventh its supply of yolk has been used up, and
it has begun to swim about freely and to use its mouth
for foraging. It seems as though in this fish those elements
develop most rapidly during the embryonic period which
are most necessary to meet the immediate exigences of
larval life.

Stage II.—The changes which have taken place in the
assumption of this stage are greater than any which take
place subsequently. All the elements in the adult pre-
formed in cartilage are already represented, and any further
development consists mainly in processes of remoulding
and ossifying.

In the branchial apparatus all the arches are present, but
are still unsegmented (figs. 7 and 14). The first three (br, 1—%)
are equally developed, the fourth (b7.*) complete but reduced,
and the fifth incomplete and still further reduced (br. 5).
Each complete arch has a flattened, expanded ventral end,
whilst its upper end is hook-shaped, and extends into the
tissue above the roof of the buccal cavity. Both these
features are lacking in the last arch.

The copulare commune (tig. 14, bv. b.1—~8), still unsegmented,
has extended forwards between the elements of the hyoid arch

536 H. H. SWINNERION.

and back to the level of the third branchial arch. From these
relationships it is evidently to be regarded as the representa-
tive of the basibranchial elements of the first three arches.
Behind it and between the lower ends of the fourth arch,
a small cartilaginous fourth basibranchial (br. 0.4) has
appeared.

The first representative of the pharyngo-branchial elements
is an oblong plate (br. p.3—4), but, owing to the absence of
any connections, it is impossible to decide definitely to which
arch it belongs.

Huxley (58, p. 408) described the general appearance and
some of the main features of the first two arches of this stage
and of the adult, but both his figures and descriptions make
it evident that, owing to the methods at his disposal, several
interesting and important details escaped his notice. It will
therefore be better to give here a detailed description of all
the parts, even at the risk of some slight repetition.

The hyoid arch is now quite complete. Medianly there is
present a basihyal (hy. b.), broader in front than behind, and
much stronger than the copulare commune with which,
between the hypohyalia, it is connected only by pro-cartilage.
This element, therefore, chondrifies quite independently of
the copulare, and it thus differs from that of the salmon in
its mode of origin, as described by Stohr (p. 7), who speaks
of it as segmenting off from this. My observations, however,
agree with those of Pouchet (78, p. 57) on Gobius.

The cartilage representing the combined ceratohyal and
epihyal (hy.) is still simple and bar-like ; but, proportionally
to the branchial arches, is much more massive than in the pre-
vious stage. Below, it fits into the concave outer surface of the
corresponding hypohyal (hy. h.), which, as in the salmon
(Stéhr, 83, p. 10), arises by differentiation of the pro-cartilage
surrounding its ventral ends. At its upper extremity it is
connected by a delicate pedicle of cartilage, the stylohyal,
with the remainder of the arch. This little element arises as
a separate chondrification in the pro-cartilaginous tract
described in this region for Stage I.

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 537

The hyomandibular and symplectic (hym., sym.) are quite
continuous, the only indication of boundary between them
being, as before, the point of attachment of the stylohyal.
Owing to the great elongation of the symplectic, a feature
which is still more strikingly seen in Syngnathus (MeMurrich,
83, and Pouchet, 78) and Siphonostoma (fig. 58, sym.), this
point is no longer in the middle of the tract of cartilage
formed by these two elements. The axis of this tract is now
gently curved downwards and forwards from the auditory
capsule, and in association with this the line of connection
between the hyomandibular and the capsule has become
almost horizontal. The hyomandibular as a whole is tri-
angular, with the foramen for the hyomandibular nerve close
to its front side.

The mandibular arch as a whole has essentially the same
shape as in Stage I, but all its parts are now well defined, and
consist of hyaline cartilage.

The quadrate is triangular, with the mandibular articula-
tion at its apex. Resting as it does on the symplectic, it
seems to have been carried forward by the elongation
of this, so that the articulation now hes under the anterior
border of the orbit, and thus shows a striking contrast to the
salmon (Parker, 78, pl. iv, fig. 1) or trout (Winslow, 98,
pl. iv, fig. 28) ; in the former it lies beneath and in the latter
behind the orbit. Consequently in these forms the mouth is
half the length of the whole skull, and the symplectic and
hyomandibular have preserved the same proportions as in the
stickleback at Stage I.

The metapterygoid process, when compared with the same
region of the salmon or trout (Text fig. 4, p. 571), is reduced
almost to vanishing point. A similar condition occurs in
Gobius and Atherina (Pouchet, 78) and Belone (Text fig. 2).
The vanishing point is reached in Syngnathus (McMurrich
and Pouchet) and Siphonostoma (Pl. 31, fig. 48), in
which this process is represented by the posterior upper
corner (qu. m.). This position is of course quite secondary,
for in the earliest stages the quadrate cartilage lies parallel

538 H. H. SWINNERTON.

to the symplectic (MeMurrich, 83, p. 631). A transitional
stage between that exhibited by the stickleback and that
found in the salmon, trout, and even also in Amia (fig. 61),
occurs in such a type as Zoarces (Text fig. 3). These facts
seem to point to a process of reduction, going on within the
teleostean series as we pass from lowly to more specialised
types. In the stickleback the metapterygoid process hes
for its whole length on the symplectic, with its extremity
inclined somewhat to the inside and situated on a level with
the attachment of the stylohyal.

The palatine process (qu. pa.) is also attenuated, but in
this respect it resembles the condition found in the salmon.
In stickleback and pipefish, owing to the forward position
of the mouth, this process is very short. Its extremity is
now connected with the ethmoid by means of an insignificant
tract of pro-cartilage. Though this connection is established,
there is no sign of any part of the process having arisen
independently, but chondrification has taken place from the
quadrate outwards. Stohr has said of the salmon that “der
pterygopalatintheil is demnach ein Auswuchs des Quadrat-
knorpels” (88, p. 11). The same is equally true of the
stickleback. In Syngnathus, McMurrich (83) has described
an independent cartilage, which he calls the ethmopalatine,
attached to the ethmoid in front and separated by a tract of
connective tissue from the “ pterygoid process” of the quad-
rate behind. The same element is present in Siphonostoma
at all the later stages I have examined, viz. in individuals
varying from 14 mm. to 26 mm. in length, but it is joined to
the upper end of the rod-like quadrate by a tract of pro-
cartilage which apparently never becomes chondrified.

Particular interest attaches to the relationship of the
extremity of the palatine process to the ethmoid. To fully
appreciate this it is necessary here to briefly describe this as
it is found in the’ larval Amia and pike of a shghtly later
stage.

In Amia (Pl. 31, fig. 60), as in the salmon and trout, the
extremity of the palatine process is expanded, but it differs

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 539

in that it gives a long continuous surface of attachment to the
ethmoid cartilage (e.) from the parethmoid cornu (e. p. ¢.)
to the pre-ethmoid cornu (e. pr.).

Walther (83, pl. 1, figs. 1, 7) gives two figures of the
Pike’s chondrocranium at two very early stages. In the one
the ethmoid is very short, and the expanded extremity of
the palatine process bears to it apparently the same relation-
ship as that just described for Amia. In the other, which
was older, the ethmoid is of considerable length, and exhibits
well-marked pre-ethmoid and parethmoid cornua; with each
of these the palatine process is connected, but not with the
intervening portion of the ethmoid. The anterior connection
is by far the stronger, whilst the hinder is so weak as to
suggest that it is secondary. Whether this is so, or whether
the two are formed by a breaking into two of the single
connection by fenestration, it is impossible to decide; the
fact remains that at this early stage two such connections
exist.

In the stickleback the attachment is solely to the
pre-ethmoid cornu (fig. 7, e. pr.). The same is true for
Syngnathus and Siphonostoma (fig. 48, qu. pa.), for the
ethmopalatine (pa.) must undoubtedly be regarded as the
homologue of the palatine process.

Thus in the development of these four distinct types of
fishes there are three different modes of attachment between
the palatine process and ethmoid, which by reason of their
early appearance in ontogeny seem to exhibit something of a
fundamental nature (v. pp.).

Meckel’s cartilage has now a recognisable coronoid and
angular processes, with an articular facet between, and as a
whole it is little more than one third the length of the hyo-
mandibulo-symplectic tract. That of the trout (Winslow,
98) is nearly twice the length of this. The shortness of the
palatine process, combined with the equally short mandible,
gives to the stickleback its characteristically small gape.

Thus all the peculiarities in the development of the
ethmoid, and of the hyoid and mandibular arches, which

540 H. H. SWINNERTON.

distinguish Gasterosteus, Syngnathus, and Siphonostoma,
from the salmon, trout, and pike, seem to be associated
with the production of a small mouth; and the differences
between the two types may be summed up by saying that in
the latter the symplectic and metapterygoid processes remain
short whilst the gape enlarges, and in the former these parts
elongate whilst the gape remains small.

Unlike the cranium, the visceral skeleton is not wholly
devoid of bones at this stage, for the dentary, maxilla, and
operculum, are already present. Of these the last two
remain throughout life as purely dermal elements ; but the
first, though now merely a delicate sheet of bone lying out-
side of Meckel’s cartilage, becomes later on closely related
to this.

Stage III.—lIn the branchial apparatus (PI. 28, fig. 15)
each of the original rod-like arches has now begun to break up
into segments. In the first three arches the flattened ventral
extremities form the hypobranchials; in the fourth it never
becomes separate. The epibranchial of the first ends freely ;
that of the second articulates partly with the corresponding
pharyngo-branchial, and partly with the side of the large
element behind. This latter also gives attachment to the
third and fourth epibranchials, and may be regarded as the
representative of the corresponding pharyngo-branchial. The
connection of the second epibranchial to this is probably
secondary, and indicative of a tendency to complete reduction
of the second pharyngo-branchial with transference of its
function to the one behind.

Both representatives of the pharyngo-branchial series are
now armed with teeth, and the hinder one has already
become partially ossified.

The copulare commune (fig. 15, br. b. 1~8) has extended
further backwards, and is now in contact with the basi-
branchial of the ‘fourth arch (br. b. 4), which has undergone
no change.

In the hyoid arch, the basihyal, and hypohyal, are prac-
tically the same as before ; but the combined ceratohyal and

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 541

epihyal (fig. 8, by.) has lost its bar-like character in the
upper half and become converted into a square plate, and
thus resembles the same element in Syngnathus (McMurrich,
83, p. 631).

The symplectic is unchanged, but the hyomandibular (fig. 8,
hym.) has grown rapidly in such a way as to transform the
triangle of the previous stage into a rectangle, whose long axis
is parallel to that of the skull, and whose attachment to the ear
capsule is now horizontal. At each end of this attachment
the cartilage has swollen somewhat to form an articular head,
whilst the intervening portion remains unchanged. Thus at
this stage the hyomandibular articulation is beginning to
depart from that simple elongated type which it exhibited in
the earlier stages, and which is retained in the adult state in
many 'l'eleosts, e. g. salmon ; and to take on that two-headed
type of articulation seen in the pike, and so commonly
possessed by Acanthopterous fishes. ‘he condition at this
stage might with fairness be compared to that described by
Gegenbaur for Alepocephalus (78, p. 14). There the double-
headed type is indicated, but the simple type is not yet quite
lost.

Owing to these growth processes, the hyomandibular nerve
foramen lies in the centre instead of near the anterior border
of the hyomandibular.

The quadrate (fig. 8, qu.) has also grown, but is still
triangular with the articular head for the mandible, and
the palatine, and the metapterygoid processes in the same
relative positions as before. The latter process is propor-
tionately smaller, but this is due to growth of quadrate rather
than to actual reduction of the process, for its extremity is still
opposite the insertion of the stylohyal. In young Gaste-
rosteus spinachia the extremity of the metapterygoid
is removed far in front of this point, owing to the continued
elongation of symplectic and ethmoid. ‘The mouth and
its associated parts are thus carried much further forwards
from the eye than in G. aculeatus.

The palatine process retains its uniformity of thickness,

VOL. 45, PART 4,—NEW SERIES, 00

542 H. H. SWINNERTON.

and is in direct contact with the pre-ethmoid cornu, whilst
still lacking relationship to the parethmoid (e. p.c.). These
features are much better seen now that the ethmoid region
has begun to elongate. Beyond the articulation with the
pre-ethmoid cornu the palatine sends a small projection, the
future maxillary process, forwards on to the upper end of the
maxilla.

A fusion of the quadrate cartilage with the hyoid arch,
such as that seen by Pollard (94, p. 19) in Blennius and
Gobius, is nowhere indicated.

Of the osseous elements belonging to these two arches,
all, except the palatine and pterygoid, have appeared im the
older larvee of the stage.

Stage IV.—As with the cranium, so with the visceral
skeleton, this stage is distinguished chiefly by the presence of
nearly all the osseous elements possessed by the adult, and of
most of the cartilaginous ones. Except when otherwise
stated, the ossification of a cartilaginous portion must be
taken to mean that only the surface is ossified, whilst the
cartilage inside remains intact.

In the branchial apparatus (fig. 16) the copulare commune
has become ossified in three places, and thus the same number
of basibranchials (br. b. 1—3) are recognisable. Apart from
this it has undergone no further change except, perhaps, for
a slight tendency on the part of the unossified cartilage to be
less hyaline than the rest. The first basibranchial (br. 6. 1)
sends obliquely backwards and downwards from its ventral
surface a strong osseous process tipped with cartilage for the
articulation of the urohyal. The second and third basi-
branchials (b7. b. 2-8) remain simple, and the fourth (br. b. 4)
is still separate, small, and cartilaginous. Kach lies in front
of the arch to which it belongs.

The hypobranchials (br. h.) of the first three arches and
the flattened ventral extremity of the fourth are still wholly
cartilaginous. All the cerato-branchials are ossified, and in
case of the fifth, except for a small portion at the lower end,
all the cartilage has disappeared, ‘Thus, though it was

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON, 543

the last to chondrify, it is the first to become completely
ossified.

Similarly all the epibranchials (br. e.) are ossified; the
first is connected by its inner pointed extremity with the
anterior end of the second pharyngo-branchial; the second,
whilst still remaining fully connected with this, has also
taken on a still more intimate relation to the compound
pharyngo-branchial behind ; the third is slightly changed in
shape by the formation of a small cartilaginous process
dorsally, but otherwise, with the fourth, it retains all the
relationships of the earlier stage. Both pharyngo-branchials
(br. p.) bear teeth and are completely ossified, though a little
cartilage remains at the anterior end.

The basihyal (hy. b.) has become osseous for the greater
part of its length, its anterior portion alone being wholly
cartilaginous. ‘The posterior portion sends down a keel
of bone, and abuts behind directly on the first basi-
branchial.

Kach hypohyal (fig. 13, hy. h.) has commenced to ossify
from two centres, a dorsal aud a ventral. The outer surface
of the ventral forms a concave facet into which the convex
inner end of the ceratohyal (hy. c.) fits ; that of the other part,
on the contrary, is strongly convex, and fits into a corre-
sponding facet in the upper surface of the same bone.

The ceratohyal occupies the whole of the rod-like, and the
lower third of the expanded portions of the original hyoid
cartilage. Hxternally the ossification extends into the
surrounding tissue, and thus partly encloses the upper
hypohyal; internally a similar but more laminar extension
(hy. ¢.’) overlaps a considerable part of the epihyal, which is
formed from the remainder of the cartilage.

As in the previous stage, the hyomandibular (fig. 13, hym.)
exhibits the most striking change in shape. ‘This time, how-
ever, the change is brought about by the rapid growth of the
anterior half of the lower border, whereby the stylohyal
(hy. 7.) and the metapterygoid (pg. m.) are carried a con-
siderable distance from the auditory capsule. Pollard (94,

544 H. H. SWINNERTON.

p- 19) describes a similar sequence of events in the develop-
ment of this region in Blennius and Gobius, but is probably
inaccurate in speaking of a shifting of the stylohyal attach-
ment; it is not the stylohyal which shifts, but the hyo-
mandibular which elongates. The articular border of the
latter has differentiated still more distinctly into two articular
heads, which fit into corresponding facets (fig. 12, hym.”), the
hinder one on the pterotic, the front one between the pro-otic
and sphenotic. The intervening portion of the border has
thus become practically functionless for suspensory purposes.
Both these heads, as well as the opercular process (fig. 13,
hym.’), remain cartilaginous. Looking at the bone internally
(fig. 13) the hyomandibular nerve foramen (hym. f.) occupies
the same position as it did in the last stage, but externally
(fig. 9) at first sight it seems to have disappeared ; in reality
it has been carried to the ventral edge (hym. f.) by over-
erowth of bone.

Though the cartilaginous core of the symplectic (figs. 13,
39, sym.) has undergone no change either in shape or thick-
ness, its osseous portion has sent out extensive lamin of bone
dorsally and ventrally ; consequently if no attention is paid
to the condition of the original cartilage an entirely false
idea of the true relationships of this bone will be gained.
Its distal end is still unossified, and les almost completely
enclosed by the quadrate above, and by a long bony process
of the quadrate below.

Under the symplectic, and at the level of the hinder end
of the process just mentioned, there is a small oblong
cartilage (figs. 39 and 40, w.) which is present only in
individuals Nos. 22 and 28, and is quite unrepresented
as far as I can ascertain in any individual larger or smaller
than these. I have been unable to determine its homologies.
That it is not a separate portion of the symplectic is proved
by its appearing’ long after this is completely enclosed in
bone. Nor can it be regarded as an articular cartilage of
any kind, for none of the surrounding bones, all of them
more or less rigidly fixed, present anything which can be

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 545

regarded as an articular surface. Its total disappearance
in older individuals also tells strongly against this view.

The ossification of the quadrate, which commenced around
the articular head, has now extended over a considerable
portion of the main body. From its lower edge, and just
behind the articulation, the bone sends out a long process-
like extension, triradiate in cross-section, and half the length
of the quadrate.

The metapterygoid process (qu. m.) still maintains its
relationship to the symplectic, and its extremity has become
the metapterygoid bone. Owing to the extension of bone
far beyond the end of the cartilage the process now seems
to end some distance above the level of the stylohyal, as
in the salmon ; therefore this feature, being secondary, is not
important.

The palatine bone (pa.), unlike the metapterygoid, is not,
at the end of the process, but around it, and mainly behind
its attachment to the ethmoid. The process (fig. 9) is, rela-
tively, much slenderer than at any previous stage, and its
extremity, owing to the continued growth of the maxillary pro-
cess (figs. 9, 13, pa. m.),is expanded. Behind this the cartilage
is not merely in contact, but in actual continuity with that of
the pre-ethmoid cornu (fig. 30, e. pr.). The palatine bone sur-
rounds this point, and extends back almost to the level of the
parethmoid. For a short distance, about the middle of its
length, this bone has completely replaced the enclosed
cartilage (as indicated by the yellow in fig. 9, pa.). Thus at
this stage the cartilage of the palatine process becomes
broken into an anterior portion, having the same relationships
as the ‘ethmo-palatine” of Syngnathus or Siphonostoma
(fig. 48, pa.); and a hinder portion, widely separated from
this and continuous posteriorly with the quadrate cartilage,
and therefore homologous with the “ pterygoid process” of
these forms. It is quite conceivable, therefore, that in these
highly specialised forms, the ontogenetic stage, having the
“ ethmo-palatine ” continuous with the “ pterygoid process ”
(as, for example, that shown in Stage II], fig. 8, qu. pa., of

546 H. H. SWINNERYON.

the stickleback), has been quite suppresssed, or is merely
represented in pro-cartilage. McMurrich compares this
‘“‘ethmo-palatine ” with that cartilage spoken of by Parker
as the “ pterygo-palatine.” Shortly before he wrote his
paper, however, Stéhr showed that Parker’s description was
erroneous, and that the so-called pterygo-palatine, 1.e. the
whole of the palatine process, was never separate from the
quadrate, though its slender connection with that rendered
it liable to be torn away. ‘lhis is just what Parker did.
Entering into the development of this region in greater
detail, he described a ledge of pro-cartilage near the inner end
of the maxilla, not far from the end of the trabecula. ‘This,
he says, may chondrify separately, and may fuse with the
extremity of the “ pterygo-palatine”’ or palatine process.
Thus Parker’s separate element does not exist, and even if it
did, it would be something entirely different from the
inconstant element described by Stohr. Pollard, working
upon the cartilages in the head of Siluroids, describes a
separate element which lies upon the upper end of the maxilla
in front, and may fuse with the ethmoid behind; this has,
therefore, all the essential relations of the maxillary process.
He calls it the pre-palatine and says, “ ‘I'his piece is well
known (Stohr and Parker) to arise in ‘l’eleostei independently
of the cartilaginous upper jaw ” (94, p. 356). From this it is
evident that he misunderstood the description of these two
authors. His ability to speak of the pre-palatine as arising
independently is wholly due to the course he pursued in
making his models, for he represented only those portions
which were still cartilaginous, and neglected those which
were osseous, even though they might possibly have replaced
cartilage developmentally. If the same had been done for
the ethmoid and palatine in fig. 9, the palatine bone would, of
course, have been omitted, and the real extremity of the
palatine process would have appeared as a cartilage, wholly
independent of the cartilaginous upper jaw, but continuous
with the ethmoid. ‘This, however, would give an absolutely
false idea of the parts. Mxaminavion of sections through the

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 547

head of Auchenaspis and Silurus show that his figures give
such an idea. The individuals he dealt with were as far
advanced in respect to ossification as the stage we are now
considering, and accordingly were unreliable for proving the
presence or absence of a primarily independent cartilage in
this region. If such an element existed, we should naturally
expect to find it in such lowly forms as Amia and Lepidosteus ;
but my observations on the larva of the former, and Parker’s
on the development of the latter, show that it is not forth-
coming. Whatever may be the significance of the small,
separate chondrification described by Stohr, it is certain that,
as far as present knowledge goes, no part of the cartilaginous
upper jaw of 'l'eleosts is primarily of independent origin, but
that at least as far as the connection with the ethmoid it is a
direct extension of the quadrate.

Standing out from each side, but quite separate from the
anterior end of the dentary, is a small rod of cartilage (figs. 9,
10, 28, 29, /.), which lies wholly within the lower labial fold
(l. f-). Brooks (88, p. 179) describes a similar ‘“ rod-like
body ” in the haddock, and homologises it with ‘the lower
labial cartilage in Hlasmobranchs.” There can be no doubt
that it is the same element that is present in Gasterosteus.
The cartilage forming it in the latter is not so hyaline as that
in the rest of the skull, but histologically it agrees closely
with that found in the tentacular skeleton of the Siluroids.
Though it lacks a basal plate, its basal portion is more
strongly developed (cp. figs. 28, 29, /.), and the whole bears a
close resemblance in position to the mental tentacle of
Silurus, and is probably homologous with that. This rod of
cartilage is probably of more frequent occurrence among
Teleosts than is usually supposed, for it is present in Perea,

Among the membrane bones, the pterygoids and suborbitals
have now appeared, and thus the full total of ossifications
present in the adult is made up.

The pterygoid bone (fig. 13, pg. en.) as looked at from
inside is triradiate ; its antero-dorsal ray is proximally tubu-
lar, completely enclosing the delicate palatine cartilage, aud

548 H. H. SWINNERTON.

distally spicular, lying along the inner surface of the palatine ;
its antero-ventral ray is the smallest, and lies on the anterior
edge of the quadrate, thus occupying the position of an
ectopterygoid (Parker’s pterygoid) ; its posterior ray, which
is the largest, lies along the inner side and upper border
of the quadrate, and thus functions as an entopterygoid
(Parker’s mesopterygoid). Possessing, as it does, all these
complicated relationships, there is some difficulty in ascertain-
ing its homologies. ‘The position of its largest portion, in
relation to the inside and upper border of the quadrate,
strengthens the probability that it represents the entoptery-
goid.

From these details it will be seen that Huxley (58, p. 409)
was correct in his delineation and description of the position
and origin of the hyomandibular, symplectic, operculum,
metapterygoid, and mandibular bones, but that he overlooked
the true palatine and mistook the pterygoid for this.

Adult.—There is little to record concerning the branchial
apparatus (fig. 22) beyond the appearance of a centre of
ossification in each hypobranchial ; the more complete ossifi-
cation of all the parts; the formation on the fourth epibran-
chial of a small ascending process which at its apex meets
the corresponding one on the third (br. e.3) ; and the shifting
of the insertion of the hypobranchials belonging to the third
and fourth arches from the sutures between the basibranchials
on to these themselves.

Cope (70) observed that the pharyngo-branchials were small
in accordance with the general tendency of the whole apparatus
to become weak ; he also pointed out (p. 457) that they bear
a close resemblance to those of Belone. The first, present in
so many Acanthopterygii as a styloid bone, frequently called
the suspensory pharyngeal, is quite absent in both these
fish.

At Stage IV. the ceratohyal was equal in size to the epihyal
(hy. e.), but now, owing to a considerable increase in
length, it seems to bear the latter lke an epiphysis at its
upper end.

MORPHOLOGY OF TELEOSTBAN HEAD SKELETON. 549

The hyomandibular and symplectic are widely separated
from one another by a broad area of cartilage. As a whole,
the former presents a striking contrast to the simple hyoman-
dibular of the salmon, and seems to be made up of four
radii connected by more or less extensive laminz of bone.
The two uppermost end in cartilaginous articular heads (hym.”)
which fit into facets described above. The edge of the
lamina between them takes no part in the articulation, so that
during the development of this region there is a complete
transition from the long simple articulation present in the
adult salmon to the double articulation with two rounded
heads. ‘The pike has the double-headed articulation; the
heads, however, are not rounded. In Belone this articulation
is intermediate between that of the salmon and the pike.
The anterior lower radius now passes obliquely forward, as in
pike and Belone.

The hyomandibular nerve has the same course as in the
previous stage. In the cod it passes in front of the main
radius, and thus seems to suggest that even within the
Teleostei this nerve does not bear a constant relationship

to the hyomandibular—a conclusion to which Pollard (94,

p. 24) was led by a comparison of its position in Ganoids.

The symplectic (sym.) is still nearly twice the length of
the hyomandibular—itself a long bone,—and thus presents a
striking contrast to this in other forms, for both in pike
and Alepocephalus it is shorter than the hyomandibular,
and in the salmon it is but an insignificant appendage of
this. Belone offers an intermediate condition, for in it the
bones are of equal length. In the stickleback its extremity
remains cartilaginous, and is almost completely enclosed by
the quadrate and its laminar outgrowth.

The metapterygoid is unaltered except for the fact that in
surface view the original shape and extent of the cartilage,
now largely ossified, is disguised by the great laminar
extensions of bone dorsally and ventrally. Nevertheless,
the reduced condition of that cartilage is more than indicated
by its width at the point where it enters the ossification. In

550 H. H. SWINNERTON.

the majority of 'Teleosts the metapterygoid abuts directly on
the quadrate, but here it is separated by a wide space, partly
occupied by the pterygoid bone, and this is, no doubt, largely
due to the forward shifting of the quadrate. In Belone the
two bones (fig. 49, gw. and p.g.m.) lie one against another ;
and, as the suture between them is cartilaginous, it gives an
indication of the original weakness of the metapterygoid
cartilage—an indication all the more trustworthy because the
laminar outgrowth of this bone is suturally connected with
the pterygoid and not with the quadrate, which has no such
extension dorsally. These relations seem to be the prevalent
ones, so that the length of the suture between the metaptery-
goid and quadrate may geuerally be taken as indicative of the
extent of the original cartilage. If so, then in pike (fig. 43),
Zanclus (fig. 52), Balistes (fig. 53), and other Plectognathi, the
Acronuride, Cyprinodontide, and the great majority of
Acanthopterygu, the cartilage was strongly developed; as
opposed to its marked reduction in Belone and Gasterosteus,
and its complete suppression in Syngnathns and Siphonostoma.
In the adult of the last two forms, as well as of Fistularia,
all the bones of this region, including the quadrate, have
sent out extensive lamine to form the side walls of the
tubular snout, so that only the study of ontogeny can show
the true distribution of the original cartilages.

The palatine bone (pa.), as already described, appears as
an ossification of the palatine cartilage around and behind
its connection with the ethmoid. As in Siluroids (Pollard, 94,
p. 355), the continuity between these two cartilages becomes
lost with advancing ossification; but the bone still articu-
lates with the pre-ethmoid cornu. The outer surface of its
anterior half appears in a lateral view of the head skeleton
(fig. 17, pa.) as a sculptured bone lying between the nasal and
the first bone of the suborbital series (so.1), and bears such a
close resemblance to the pre-orbital bone (antorbital, Sagemehl)
of Amia that one would be almost forced to acknowledge its
homology with that if one did not already know that Amia
possessed a palatine bone. The resemblance is all the more

MORPHOLOGY OF TELBOSTEAN HEAD SKELETON. dol

striking, because of the presence of several sensory papille
(fig. 31) on the skin which overlies it.

The cartilage which the pterygoid enclosed has disappeared,
not because of the encroachments of ossification, as in the
parethmoid (fig. 32, e.p.b.), but because of actual reduction,
as in the case of the supra-orbital band and the hinder portion
of the trabeculee.

The relation of the palatine and pterygoid bones and
adjoining cartilages to the ethmoid region in 'l'eleosts is of
great importance, aud a closer study of it will be found
extremely useful in the attempt to ascertain the inter-relation-
ships of these fishes.

For this purpose no fish presents so instructive a condition
and so convenient a starting-point as the pike (fig. 43). In
it the palatine is a long and massive bone, formed, as in the
stickleback, by ossification of the distal portion of the palatine
process. At its extremity it bifurcates slightly, and partially
encloses the pre-ethmoid cornu with its ossification. Behind
this bifurcation, and on the outer side, it presents a rough
facet, with projecting upper border (pa. m.) for the insertion
of the inner end of the maxilla. ‘he whole of its ventro-
internal surface is armed with numerous powerful teeth. For
a short distance behind this the original cartilage persists,
enclosed by the ectopterygoid externally and the entoptery-
goid internally. Opposite the parethmoid cornu this cartilage
enlarges considerably, and offers a well-developed concave
surface for articulation with, or more accurately, for sliding
on, the rounded ventral surface (pa.”) of the cornu. ‘Thus
in the adult, as in the larva, there are two points for attach-
ment of the palatine process to the ethmoid, a pre-palatine
and a post-palatine.

In Scomper (fig. 44) the same condition exists, but the
parts are more specialised. As in the pike, the palatine bone
(pa.) is massive and dentigerous, but here the anterior
extremity of the bone is prolonged into a well-developed,
curved maxillary process (pa. m.). Internally to the base of
this is a facet, which fits on to and forms a typical articula-

Do2 H. H. SWINNERTON.

tion with the pre-ethmoid cornu (pa.’). The latter, owing
to the formation of a rostrum, such as was described for the
young salmon, is situated some distance behind the extremity
of the ethmoid, and, because of the almost complete ossifica-
tion of this, is bounded by the vomer below and _ nearly
surrounded by the mesethmoid above. Behind the palatine
bone the cartilage (qu. pa.) persists, and, as in the pike, it
offers a well-formed concave surface, which slides with
a lateral motion on the rounded, cartilaginous ventral surface
of the parethmoid.

In Pagellus centrodontus (fig. 45) another modification
of the same type exists. Here, as before, the palatine (pa.) is
massive and dentigerous, but its maxillary process (pa. m.)
is much longer, more curved, more powerful. Bearing the
same relation to this as in Scomber is a facet for the
pre-palatine articulation. Here again, owing to the formation
of a rostrum even stronger than that of the Scomber, the
pre-ethmoid cornua are situated far back behind the end of
the ethmoid (e.), and lie mainly on the anterior portion of the
parethmoid bone, close to the suture between it and the vomer.
In this example, however, probably owing to the shortness of
the ethmoid, the post-palatine sliding surface is formed not
by cartilage, but by the hinder end of the palatine bone.

As already pointed out, the larval salmon has both pre-
and post-palatine articulations. They are also present in the
adult Salmo trutta, but owing to the ethmoid being wholly
cartilaginous, the facets are not so conspicuous as in Scomber
and Pagellus. ‘he pre-ethmoid cornu is much smaller, and
not so distinct as the other, but it nevertheless exists, and
was evidently seen, though not noticed by Parker, for he
represents it in his figure of the ventral surface (pl. 7,
fig. 2) of Salmo salar at the posterior end of the “curved
ridge.” Thus the palato-ethmoid region in the salmon, on
the whole, conforms to that of the pike, and differs mainly in
the fact that elongation of the ethmoid takes place in the
latter between the cornua, and in the former in front of the
pre-ethmoid. If this region in the salmon were more com-

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 555

pletely ossified there would be little to distinguish it from
that of Pagellus.

The condition in the Cyprinodonts is a little difficult to
make out, but is well seen in Characodon luitpoldi. By
careful examination of this, and by serial sections of this
region in a specimen of Haplochilus fasciatus, which Mr.
G. A. Boulenger kindly placed at my disposal, I am satisfied
that the condition in them differs from that in the pike only
in the shortness of the ethmoid, the more complete ossification
of parts, and in the absence of teeth on the palatine.

The Acanthopterygil, with one or two notable exceptions,
conform to one or other of the conditions just described, or
are modifications of these. For example, in Pagellus ery -
thrinus the ethmoid is perfectly ossified, the parethmoid cornu
has practically disappeared, and the pre-ethmoid is greatly
reduced ; nevertheless the shape of the palatine and its
maxillary process, and the form of the ethmoid with its well-
marked rostrum, leave no doubt that it belongs to the same
type as that shown by its kindred species (fig. 45). Again,
in Odax Richardsoni the post-palatine articulation appears
to be formed by the pterygoid bones.

It is no uncommon thing for the palatine to be strongly
attached to the vomer by ligament, and even to present a
flattened surface to that bone, e.g. Scizna, but this is not a
feature of primary importance, nor does it affect the presence
of the other articulations.

These facts prove that in Salmonidx, Hsocide, Cyprino-
dontidz, and Acanthopterygu (with exceptions to be dealt
with), one type of palato-ethmoidal relation exists, viz. one in
which the palatine process or its derivatives are attached to
the ethmoid at two points, the pre-ethmoid cornu and pareth-
moid cornu, by two facets, the pre-palatine and post-palatine.

In the stickleback, during development, a post-palatine
attachment is always absent, a pre-palatine always present
(figs. 8, 9, e. pr.). The palatine bone itself is small, edentu-
lous, and possesses an insignificant maxillary process. The
ethmoid region is of the same type as that of the pike in

554 H. H. SWINNERTON.

the absence of a well-marked rostrum ; hence the pre-palatine
articulation (fig. 51, pa.’) is situated at its anterior extremity,
and, as in the pike, the facet is not specialised.

Both in the young (Text fig. 2) and adult Belone (fig. 49)
precisely the same conditions prevail; indeed, in one small
individual the palatine and pre-ethmoid cartilages were even
fused together. The region of this is indicated by the dotted
line in fig. 49 (pa.’). The palatine (pa.) bone is simple in
shape, lacks a maxillary process, and is edentulous; poste-
riorly it sends a lamina of bone along the outer side of the
palatine cartilage (qu. pa.) to meet a similar process from the
quadrate (qu.). Thus the ectopterygoid is functionally
replaced and is absent. The cartilage itself, which never
breaks down though exposed dorsally, does not enlarge at
any point to form a post-palatine articulation. The ethmoid
is predominantly cartilaginous, and in the absence of a well-
developed rostrum is of the same type as that of the pike,
but is much shorter.

The condition in Exocoetus is identical.

While the almost wholly cartilaginous condition in the
Belone is the retention of a primitive condition, the greatly
ossified palato-ethmoid region of Syngnathus is at the
opposite extreme of specialisation (fig. 50). The palatine
bone (pa.) has the same characters as in the stickleback,
and like that is partially enclosed posteriorly by the single
pterygoid (c.) (McMurrich, 88). The ethmoid region when
compared with that of young Siphonostoma (fig. 48, e.) 1s
seen to owe its great length to elongation, not of the hinder
half, containing the mesethmoid cartilage, but to that of .the
front half, consisting purely of ethmoid plate. Nevertheless
the mesethmoid bone (fig. 50, e. m.) has apparently extended
quite to the anterior end (MeMurrich, 83), including the pre-
ethmoid cornu. The palatine bone (pa.) which is attached
to the pre-ethmoid cornu (pa.’) between the mesethmoid bone
and vomer is carried too far forward for it to bear any
relationship to the parethmoid bone (e. p. b.).

Fistularia differs from Syngnathus only in the fact that

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 55d

though the palatine is small it possesses a moderately
developed process and is beset with teeth, thereby showing
that the absence of these features in all the other forms
possessing this type of palato-ethmoidal relationship is by no
means an essential accompaniment of it.

Among Acanthopterygii the Zanclide and Acronurida,
whose position is by no means a settled question, are character-
ised by the possession of this type of palato-ethmoid relation.

In Zanclus cornutus (fig. 52) the small palatine (pa.) has
an insignificant maxillary process behind, which is the facet
for the pre-palatine articulation. he ethmoid region is of
the type usually associated with the possession of this single
articulation, and its shape as a whole calls to mind that of
young Siphonostoma (fig. 48) and is still largely cartilaginous
anteriorly.

The Acanthurus ceruleus (one of the Acronuride)
presents a condition so closely similar to that of Balistes
maculatus (fig. 53) that a description of the latter—a Plecto-
gnath—will do service for both. Its palatine has the same
relations as that of Zanclus. The ethmoid region is not so
long, but is much more massive. In my specimen the pre-
ethmoid cornu, with its well-defined articular surface (pa.’),
is still cartilaginous, and is bounded beneath by the vomer
and above by the mesethmoid bone. There is no trace either
of rostrum or of post-palatine articulation. Klein, when
describing the ethmoid region of Acanthurus (Teuthis ; 84,
pp. 187, 219) and Balistina (ibid., pp. 190, 257), called atten-
tion to the close resemblance between them.

Thus all the Scomberesocide, Hemibranchi, Lopho-
branchu, Plectognathi, and among Acanthopterygii, the
Zanclide and Acronuride, are characterised by the pos-
session of only one palato-etlhmoidal connection, viz. the
pre-palatine, and of an ethmoid region with its pre-ethmoid
cornua carried at the extremity.

The majority of these forms are characterised by the great
elongation of the ethmoid region and by the shortness of the
mandible, thus suggesting a possibility that the post-palatine

556 H. H. SWINNERTON.

connection may have been lost through the coincidence of
these two features; but a comparison of figs. 43—53 (each of
which has been drawn so that the whole skull would be the
same length as that indicated in fig. 43, from the tip of the
ethmoid to the arrow marked *) shows that this is not the case,
for both Belone (fig. 49) and Gasterosteus (fig. 51) have a
short ethmoid as compared with that of the pike and
Scomber; and the former at the same time possesses a long
mandible. If a long ethmoid were a condition necessary to
the absence of post-palatine articulation, then it should be
absent in the pike. Again, if a long gape and short ethmoid
were the essential accompaniments of a double attachment,
the Belone should have this. But as the pike possesses a
well-formed post-palatine attachment, and the Belone does
not, it becomes clear that the two types are wholly indepen-
dent of either length of ethmoid or size of gape.

When, in connection with these facts, we consider that both
the single and double types are quite distinct throughout deve-
lopment, that notwithstanding all the changes of form which
the head undergoes among Acanthopterygil, from the short-
snouted Bovicthys to the long-snouted Scomber, from the
deepened Drepane to the flattened Platycephalus, the double
articulation is retained, we cannot refrain from concluding that
the distinction we make between the two types is not a super-
ficial one, and that the single articulation is due, not to
the existence of a long-snouted Lophobranch or Plectognath,
but that these owe their origin to the existence of this type.
Whether this conclusion is right or wrong, future work alone
will decide ; meanwhile, the fact remains that two such types
exist, and consequently the necessity arises in descriptive
work for distinguishing between them concisely. Accord-
iugly, all forms possessing both pre- and _post-palatine
articulations, irrespective of the possession or non-possession
of a rostrum, may be described as Disartete; and all
possessed only of the pre-palatine articulation, and of an
ethmoid whose pre-ethmoid cornua are situated on each side
of its anterior extremity, may be described as Acrartete,

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 557

In the light of these facts, considerable interest attaches
to the condition found in Amia, which has been carefully
described by Allis. Speaking of the septo-maxillary (my
pre-ethmoid), he says, “The ventral surface of the lateral
edge of the bone forms the anterior end of a low, longi-
tudinal, condylar eminence, which extends backwards along
the ventral surface of the lateral edge of the ant-orbital
process of the chondrocranium, and gives articulation to the
palato-quadrate arch ” (98, p. 447). Here, then, is a type of
articulation which seems to combine the characters of the
disartete and the acrartete conditions, and yet differs from
these in the fact that the attachment of the palatal to the
ethmoidal region is neither broken into two, nor confined to
the pre-ethmoid. or the description of this type, the term
Panartete seems most suitable. From these considerations,
and from the developmental facts given above, it is evident
that this is a type of articulation which very probably gave
rise to the disartete condition, and may possibly have done
the same for the acrartete. ‘The settlement of this question
rests with a closer study of the Malacopterygi and
Lepidosteus: in the former, according to Gegenbaur’s
descriptions and figures of Alepocephalus, we may expect to
find the panartete condition passing into the disartete ; and
in the latter, according to Parker, a condition exists which
bears the same relations to the acrartete as does that of the
salmon to the disartete, for it possesses a rostrum. It is quite
conceivable that all these types of articulation were exhibited
also by the immediate Ganoid ancestors of the Teleosts.

In the lower jaw of the adult stickleback the labial
cartilage (/.) is still present.

The premaxilla alone bounds the gape above. At the pre-
vious stage the upper ends of the ascending portions ended
in amass of densely nucleated tissue, which lay chiefly on their
under side. ‘his mass has now become chondrified (PI. 30,
fig. 42, pmz.’), and passes into the bony tissue almost with-
out boundary line. This chondrified tissue closely resembles
that of the labial cartilage. In the earlier stages of develop-

VoL. 45, PART 4,—NEW SERIES, PP

558 H. H. SWINNERTON.

ment of Siphonostoma the reduced premaxillary processes
end in similar tissue, which, however, in the latter stages,
forms a separate cartilage homologous with Sagemehl’s
rostrale (85, p. 100). It seems highly probable, therefore,
that the cartilage we are considering in the stickleback also
has the same homologies. Whether the same is true of the
similar layer on the mesethmoid bone is an open question.
Whatever their morphological significance may be, func-
tionally these cartilages serve as a means for allowing the
premaxillz to slide on the ethmoid.

IV. GENERAL CONSIDERATIONS.

Looking back over the past thirty years of investigation
into Teleostean phylogeny, one of the many features which
calls for attention is the gradual shifting of the centre of
interest from the Klasmobranchs to the Marsipobranchs. At
the beginning of this period Haeckel was so impressed
by the many peculiarities of the latter group that he created
for its reception a division, the Monorrhina, equal in value
to another, the Amphirrhina, containing all the other craniata
(70, p. 507). Representing his views of the inter-relationships
of the latter in a diagram, he placed the Selachii at the point
of origin of the rest (p. 513). In these views he was followed
by Gegenbaur (72), who, whilst he acknowledged the existence
of several special features in the Selachu, considered that on
the whole the organisation of the ancestral form of other
fishes must have been similar to that presented by them (p. 22).
Referring especially to the cranium, he gave expression to
the opinion that the primordial cranium of higher vertebrates
was but a transitory reproduction of the cartilaginous shark’s
skull, and that it was impossible for the latter to be derived
from an originally bony condition.

The natural outcome of all this could only be an attempt
to harmonise the conditions exhibited by the Teleosts and
the closely allied Ganoids with those found in Selachi. Thus
Sagemehl (84), in dealing with the cranium of Amia, explains

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 559

the absence of a cartilaginous wall between the auditory cap-
sule and cavum cranii by assuming a tendency for fenestra-
tion to proceed from the peripheries of nerve foramina
(p. 207), and considers that many of the parts point to a type
most nearly approached by the Notidanide (p. 227).

Meanwhile, Huxley (76), by a comparative study of the
heads of Petromyzon and tadpole, was led to conclude that
there was “no sufficient foundation in the present state of
knowledge for regarding the Marsipobranch skull as one
which departs in any important respect from the general
vertebrate type” (p. 427).

Since then various workers, especially during the last
decade, have brought forward much further evidence to show
that the same may be said of the other organs, and that
living Selachii are more specialised than Gegenbaur supposed.
Foremost among these was Beard (90), who, basing his
remarks chiefly on the structure of the brain, pronephros,
lung, and swim-bladder, proposed to place the Marsipobranchs,
Ganoids, and Teleosts in a group separate from and equal to
one containing the Selachii, Dipnoi, and Amphibia. Allow-
ing for a transference of the Dipnoi from the latter, Howes
(91) was led to a similar conclusion by a study of the urino-
genital organs. Klautsch’s work on the development of the
vertebral column resulted in an entire agreement with Beard.
The most recent contributions to this subject are those of
Hatta (00) and Goette (01). The former, dealing with the
development of the pronephros of Petromyzon, says, “The
whole system of the pronephros of Cyclostomata, Teleostei,
and Amphibia is not perfectly homologous with the Selachian
pronephric system” (p. 407). The latter, working at the
development of the gills and branchial vessels in fishes, finds
that the gills in Cyclostomes (Enterobranchies) are not the
homologues of those in Teleostomes and Selachii (Dermato-
branchies), but represent a very old type ; and that as regards
the origin of their afferent and efferent branchial vessels the
Teleostomes and Selachii exhibit two divergent lines of
modification,

560 H. H. SWINNERION.

Thus it appears that whatever may be thought concerning
Amphibia and the Dipnoi, there is a general consensus of
opinion that the Teleosts and Ganoids are not so closely allied
to Selachii as was once thought to be the case. In order to
ascertain how far this is supported by the head skeleton, I pro-
pose now to briefly review what is known of this part in the
development and the adult of Teleostomes under the follow-
ing headings:—The relation of the trabecule to the para-
chordals; the primordial cranium; the relation of the
visceral skeleton to the cranium.

The Relation of the Trabecule to the
Parachordals.

In writing two interesting and instructive memoirs on the
development of the skull, Sewertzoff (97 and 99) has brought
together many scattered facts relating to this in the various
great classes of vertebrates, and has sought to reduce them to
some kind of order. Thus, for example, in dealing with the
position of the trabecule, he has introduced the term “ hori-
zontal” to describe their position when their hinder ends
fuse or arise in continuity with the extremity of the para-
chordals; and “ vertical,’ when fusion takes place with
their ventral surface and some distance behind their anterior
ends (99, p. 316). He also distinguishes two types of skull:
(1) that in which the trabecule arise separately from the
parachordals ; (2) that in which the trabecule arise in con-
tinuity with a plate which is or becomes the anterior part of
the parachordals. The fact that in the stickleback these
two types appear to be mere matters of individual variation
suggests that this distinction is not an important one. Again,
he tried to show that these two positions—“ horizontal ” and
“vertical’’—were intimately associated with the time of
appearance of the mesocephalic flexure, of the medullary
flexure, and of the trabecule.

In the stickleback, as in Selachu, the time of appearance
for the trabecule coincides with that of the greatest flexure.
Though the development of these parts takes place under

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 561

conditions which correspond to absence of cranial flexure,
yet in other Teleosts, e.g. Gobius, which have a capacious
egg, and consequently flexed head during development, the
horizontal position is assumed.

The study of the stickleback’s head (cp. fig. 57) shows
that flexure here is not so much at one point as at all points ;
that the anterior ends of the parachordals as well as the
closely associated trabeculee are involved in this curvature ;
and that the flexure here is much less than in the Selachian,
where the fore-brain lies practically under the hind brain,
and apparently does not involve the parachordals. These
facts lead to the inference that the position of the trabecule
in Ichthyopsida is related mainly to the degree of meso-
cephalic flexure and to the part which the parachordals take
init. ‘That is, so long as it is small—through not more than
90°,—there is nothing to prevent the trabeculz from uniting
with the ends of the parachordals ; but immediately it over-
steps that, and the fore-brain begins to pass backwards under
the hind brain and parachordals, there will be a tendency to
carry the “anlage” of the trabecule backwards also, and
consequently the upper ends of these will either have to take
a sharp bend forwards to unite with the extremities of the
parachordals, or have to fuse with their ventral surfaces.

It was long ago pointed out by Balfour that “the cranial
flexure is least marked in Cyclostomata, Teleostei, Ganoidei,
and Amphibia, while it is very pronounced in Hlasmobranchii,
Reptilia, Aves, and Mammalia” (81, p. 347). Consequently,
if the views just expressed contain any truth, the former
should exhibit the horizontal, and the latter the vertical
position for the trabecule.

In his earlier paper Sewertzoff (97) recognised that in those
cases in which the mesocephalic flexure of the brain is least
marked the trabecule are approximately horizontal, and where
it is greatly marked they are vertical; but he does not collate
these cases. He points out (99) that in both Petromyzon and
Acipenser the former position exists. From this and from
Parker’s figures and descriptions of Lepidosteus ; from those

562 H. H. SWINNERTON.

of Stéhr on Siredon and Triton (81, p. 85), Rana (82, p. 85),
and the salmon (838); those of Pouchet on Labrus, Gobius,
Atherina, Syngnathus (78) ; those of Miss Platt on Necturus
(96, p. 455) ; and from my own on Gasterosteus, it becomes
evident that the horizontal position is exhibited by precisely
those groups which, according to Balfour, have only a weak
cranial flexure.

Amongst those fishes whose development has been fully
worked out only the Elasmobranchs, viz. Acanthias (Sewert-
zoff, 99, p. 285), Prestiurus (p. 300), Raia and torpedo (ibid.,
97, p. 417), exhibit the vertical position for the trabecule.
The presence of a post-clinoid wall may be taken as a strong
indication of the former existence of the same type in many
of the other Klasmobranchs.

This may apply, but with diminished force, to the Chime-
roids and Dipnoi.

There can be little doubt that the horizontal condition is
the more primitive one, and that the vertical is secondary, due
to a strong cranial flexure, which in its turn is explained by
Balfour, probably correctly, “as being associated with an
embryonic as opposed to a larval development ;”’ and with
some advantage to be ‘ gained by a relatively early develop-
ment of the brain ” (81, p. 267).

In this respect, therefore, the Cyclostomata, Ganoidei,
Teleostei, and Amphibia have retained a more primitive con-
dition than the Selachi, and if their ancestral stock was in
any way related to these, it must have been to some form
whose foetal life was not yet sufficiently lengthened out to
permit the development of a strongly-marked flexure, and
the associated ventral position of the trabecule.

The similarity in the condition of the brain and the
trabeculz in Amniota to that in Selachii must be regarded as
due rather to convergence than to genetic affinity.

The Primordial Cranium.

As already mentioned, Gegenbaur’s work on the cranium

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 563

of Selachii led him to regard the chondrocranium of all the
higher Vertebrata as derived from a similar type ; a conclusion
to which he was guided mainly by a consideration of the carti-
laginous condition in the former as opposed to the osseous con-
dition in the latter. No doubt he was right as far as the mere
substance of which the cranium was composed was concerned,
but apparently it did not occur to him to inquire whether the
chondrocranium of a higher vertebrate with its numerous
fontanelles did in reality arise from a simple box-like Selachian
cranium. Nevertheless, since he published his unique memoir,
workers upon the piscine skull have largely assumed that such
was the case; that the fontanelles of the one have arisen by
fenestration of the other. ‘Thus, beyond what has already
been mentioned, Sagemehl in dealing with the Characinide,
which have anterior and posterior fontanelles in the roof of
the chondrocranium, regarded the former as the homologue
of the Selachian “ Preefontalliicke,’ but the latter as an
entirely new formation. More recently Pollard looked upon
a small cartilage in the large supra-cranial fontanelle of
Polypterus as the last remnant of an originally complete
cartilaginous roof, and started a comparison of its cranium
with that of Notidanus, with the ‘ obvious postulate” that
“the supra-cranial fontanelle be considered to have a complete
cartilaginous tegmen cranii of which only the remains are
now known ” (92, p. 402).

Gegenbaur, comparing Alepocephalus with Salmo and
Esox, concluded that mass of cartilage was no criterion of a
lowly position, but the form and ossification of the same, and
says: “ Aus diesem Allen geht hervor, dass das reiche Maass
von Knorpel in phylogenetischer Beziehung nicht verwerthet
werden kann. Es wird nur auf die ontogenetische
Anlage bezogen werden kénnen,' und wird dann als
eine Weiterentwickelung dieser Anlage gelten diirfen” (78,
p. 27). Again, Huxley, as early as 1858, says: ‘‘ There are
discrepancies in the structure of the skull itself, which would
forbid too close an approximation between the bony and the

1 The italics are not mine.

564 H. H. SWINNERTON.

unossified crania if their adult forms alone were examined
The study of the development of the ossified vertebrate skull,
however, eliminates the difficulty, and satisfactorily proves that
the adult crania of the lower Vertebrata are but special develop-
ments of conditions through which the embryonic crania of
the highest members of the sub-kingdom pass ” (p. 420).

Given a brain requiring protection, it is quite as reasonable
to suppose that specialisation for this purpose would, in one
case, show itself by an increase of cartilage, as in another by
the accession of bone; and that the presence of bone would
tend to preserve features belonging to the primitive cartila-
ginous cranium, which otherwise would be obliterated by
excessive growth processes. But suppositions and assumptions
must not displace facts.

Comparing the adult cranium of Amia with that of
Heptanchus, it is seen that in the latter, except for nerve
foramina, the ear cavity is completely shut off from the
cavum cranii; and that the only fontanelle present is the
“* Preefontalliicke,’ which lies at the anterior end of the
roof, and, according to Sagemehl, in front of the epiphysis.
In the former the cranial roof is massively cartilaginous,
continuous, and unfenestrated ; in the orbital region there is
a “considerable vacuity ” closed by membrane, and regarded
by Sagemehl as a much enlarged optic fenestra (84, p. 202) ;
and in the anditory region there is a gap, between the cavum
crani and the ear-capsule, which owes its existence, according
to the same authority, to the enlargement of the labyrinth ;
and, finally, the floor is occupied largely by a well-developed
pituitary fontanelle. Thus it becomes evident that, notwith-
standing the presence of a bony covering, the one fontanelle
present in the Selachian skull is absent in Amia; and those
fontanelles which are present in this are represented, at the
most, only by small nerve foramina in that.

In the advanced Amia larva (19 mm. long), as in the young
salmon (fig. 62) and Lepidosteus (Parker, 81), a large fon-
tanelle certainly is present in the roof, but it cannot be
homologised with the “ Preefontalliicke,” for it lies entirely

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 565

behind the epiphysis and between the sphenoidal and
auditory regions. Again, it is not continuous cartilage
which makes up the floor and the side walls, but fontanelles
proportionately larger than those in the adult. In Lepi-
dosteus and Salmo at least there is no sign at any later stage
of a breaking down of walls to form fenestra, but the mass
of the chondrocranium goes on increasing into old age, and
consequently, fontanelles may be reduced or even obliterated.

There is, in this connection, one feature about the develop-
ment of the stickleback to which particular attention may be
paid, viz. the fact that in it we have repeated ontogenetically
a breaking down of pre-existing cartilaginous palatine process,
suborbital bands, and trabecule, such as might have taken
place phylogenetically. Consequently it is reasonable to
expect that, if a fenestrating process ever did take place,
such lowly forms as Lepidosteus and Salmo should show
some developmental indications of it.

At still earlier stages Gasterosteus, Atherina, Hngraulis,
and Salmo also possess an anterior fontanelle homologous
with the “ Preefontalliicke,” and separated from the posterior
one by the epiphysial cartilage.

Thus the whole course of development points to an
ancestor whose cranium had two dorsal fontanelles separated
by a band of cartilage usually related to the epiphysis ; also
lateral fontanelles for transmission of optic, trigeminal, facial,
and auditory nerves ; and a pituitary fontanelle.

Further evidence from independent sources is to be sought
for in the Ostariophysi, in Polypterus, and the cartilaginous
Ganoids.

The first constitute a branch of the bony fish series probably
separate from that to which the forms we have so far con-
sidered belong, and they exhibit all those features just
mentioned as probably characteristic of the ancestor.
Sagemehl thought that the epiphysial bar had remained in
order to serve as a support for the blood-vessels passing to
the brain; but it could be said with equal plausibility that
this was the cause of its first appearance.

566 H. H. SWINNERTON.

In Polypterus Traquair describes a large supra-cranial
fontanelle (71, p. 167) bounded laterally by the upper edges
of the sphenoid. Pollard (92, p. 400) confirms him, but
further remarks on the presence of a small sheet of cartilage
in the middle of the fontanelle, which he says “ may be
regarded as indicating the former existence of a complete
cartilaginous tegmen cranii;”’ and in alater paper (95, p. 414)
he speaks of it as the homologue of the epiphysial cartilage,
aud of the space in front as homologous with Gegenbaur’s
‘“Preefontalliicke.” The true supra-cranial fontanelle, how-
ever, is the whole region enclosed by the supra-orbital band,
and is much larger than the one described by Traquair under
that name. The so-called sphenoid has not the same place
relations as the alisphenoid and orbito-sphenoid, for in Tele-
osts these bones are continuous above with the supra-orbital
band ; in Polypterus, however, its upper border lies some dis-
tance inside of, and is separated by a wide space from this.
Strictly speaking, therefore, the region between the trabecule
and the supra-orbital band is unoccupied either by cartilage or
bone, and in any case the existence of a wide space between it
and the sphenoid betokens the existence of a large, morpho-
logically lateral, fontanelle. ‘The auditory capsule has no
internal wall, and the cranial floor possesses an exceptionally
well-developed pituitary fontanelle.

Among cartilaginous Ganoids the cranium of the adult
sturgeon is so specialised in respect to the growth of carti-
lage that it is natural neither to expect nor to find any fonta-
nelles. Nevertheless these do exist during development.
The most instructive stage in this connection is that described
by Parker (82) as Stage III. In this the cranium possesses
a large fontanelle in the roof from end to end. The supra-
orbital bands project towards each other in the region lying
in front of the mid-brain, and thus indicate an epiphysial bar.
In the side walls the “ optic foramen is a fenestra” (158), and
there seems to be a separate fenestra for the exit of the fifth
nerve also. ‘The pituitary fontanelle, though small, is
present, and the auditory capsule lacks a cranial wall.

MORPHOLOGY OF TELEOSTBAN HEAD SKELETON. 567

Polyodon (Bridge, 78) retains this last feature, together with
“ Preefontalliicke ” and well-developed parietal foramina,
representing post-cranial fontanelle, into the adult.

All definite reference to the orbito-sphenoid has been
purposely omitted hitherto, but it should be remembered
that whilst Allis states that it is a cartilage bone in Amia
(97, p. 7), both Vrolik (78, p. 57) and Parker (72, p. 106)
look upon it as a distinct membrane bone; if the latter are
right, then to the already existing optic fontanelle must be
added the space occupied by this bone.

According to Sewertzoff (99) young Acanthias possesses a
large dorsal fontanelle, a large lateral one transmitting at
first trigeminal and facial as well as optic nerves, and widely
open auditory capsules. In Raia (Parker, 76, p. 216) the
dorsal fontanelle is divided by a transverse “ cartilaginous
beam.”

Thus the development of Acanthias, as well as the adult
anatomy and comparative ontogeny of ‘leleostomes, leads
back to an ancestor with a fenestrated cranium.

In the light of all this it is interesting to read Parker’s
account of this region in the Marsipobranchs (83), especially
in a moderately young lamprey. After mentioning the
presence of a large pituitary fontanelle, he goes on to say
the ‘side! walls of the chondrocranium of the lamprey are
well developed,” but “optic and trigeminal nerves pass out
of considerable fenestra, and not out of mere foramina.”
“ The orbito-sphenoidal region is wider than the alisphenoidal,
but the latter mounts up into the roof, and the two sides
meet round the middle and fore-part of the hind brain. The
occipital ring does not exist” (p. 414). He describes the
presence in the front of this posterior sphenoidal “tegmen ”’
of a “large pyriform fontanelle,” and also states that the
passage from the auditory capsule into the cavum crauil is
large, especially in Myxine.

Apart from the absence of occipital region and of ossifica-
tion, all that is required to complete the broad resemblance

1 The italies are not mine.

568 H. H. SWINNERTON.

between the cranium of a lamprey and that of a Teleost is
a close relation between the epiphysis and the sphenoidal
“tegmen,” which would then be the undoubted homologue
of the epiphysial cartilage. Failing this, it will suffice to
recall the fact that in the Cyprinoids and Characinoids this
cartilage also arises directly from the sphenoidal region.

Parker himself considered that other types of fish, especially
Lepidosteus, were descendants of primordial Marsipobranchs
(83, p. 404), and the facts just put forward lend considerable
support to this.

The possibility, of course, still remains that between this
‘primordial Marsipobranch ” and the bony fish a form like a
cartilaginous Ganoid, having bony scutes, as well as a box-like
cartilaginous cranium, may have intervened; and that as the
scutes became more closely related to this, the fontanelles
present in the embryo were preserved, and the necessity for
further growth in cartilage was obviated. It is equally
possible that on the bony fish side bony specialisation
appeared earlier, and was more rapid than in the Chondrostei,
and resulted in a preservation of the fontanelles (with the
possible exception of the roof) all through.

Again, the fontanelles always present during development
might conceivably be explained as due to the relatively large
size of the brain; but such an explanation would not apply
to the condition of the Marsipobranch cranium.

The Relation of the Visceral Skeleton to the
Cranium.

Great imterest attaches to a study of the relationships of
the so-called hyoid and mandibular arches, and such a study
may best be commenced by briefly comparing the conditions
found in the adult Heptanchus, as a primitive Selachian, and
Gasterosteus, as a typical Teleost.

In the latter the hyomandibular articulates on the auditory
capsule below and behind the post-orbital process, and the
palatine process of the pterygo-quadrate articulates with the
side of the ethmoid plate.

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 9569

In the former the pterygo-quadrate cartilage is attached
to the cranium at two points; the hinder one under the post-
orbital process, by means of the otic process, and an anterior
one between the exits of the second and fifth nerves, by means
of the palato-basal process. In front of this the cartilage

‘‘ oberkiefersfortsatz ’? under the

continues forwards as the
ethmoid region to unite with its fellow. The hyomandibular
is slender, and has no articulation with the auditory capsule,
and is fastened to this by ligament some distance behind the
post-orbital process.

One glance at these two types suffices to show that the
palato-quadrate of one possesses just those cranial attach-
ments which are absent in the other, and vice versa. ‘'T'o
derive either condition from the other is therefore impossible.

The position of the hyomandibular relatively to the
auditory region in Teleosts, together with other considerations
concerning the associated muscles and the variations in the
related nerve, led Pollard (94, p. 23) to conclude that the
elements of that name in these two great groups of fishes
were not homologous; and to seek for the Teleostean
hyomandibular in the otic process of Heptanchus, and for
that of the Elasmobranchs in the stylohyal of Teleosts. In
view of his observation that in young Silurus the two sets of
cartilages form a continuous whole, and that in Gasterosteus
the broad head of the hyomandibular is formed by backward
growth during development, it must be admitted that his
case, though startling, is strong, and if true, at once puts all
other Selachian types of suspensorium out of court as possible
conditions for the derivation of the Teleostean type (ep.
Gegenbaur, 72, p. 175).

In 1876 Huxley introduced three useful terms—autostyli,
hyostylic, amphistylic—to express the “manner in which the
mandibular arch is connected with the skull.” He regarded
the second of these as “ perfectly exemplfied by Ganoids,
Teleostei, and Plagiostomes,” and characterised by the fact
that the ‘ palato-quadrate cartilage is no longer continuous
with the chondrocranium, but is, at most, united with it

570 H. H. SWINNERTON.

by ligament” (p. 41). The condition presented by the
Teleosts cannot, however, be completely described by any one
of these terms, for whilst it is true that the hinder end of the
paulato-quadrate is suspended by a well-developed something,
which may yet be shown to have no homology with the
hyomandibular of Klasmobranchs, the fore-end is not merely
suspended by ligaments, but articulates directly with the
cranium, and may, as in Gasterosteus, Belone, and some
Siluroids, be continuous with the ethmoid cartilage until
a very advanced stage. This mode of suspension is therefore
hyostylic only posteriorly, but autostylic anteriorly.

In the first stage in the development of Gasterosteus the
palatine process does not reach the end of the trabecule. This
cannot be interpreted as an approximation to the Selachian
type, for the opposite processes are parallel. In early stages
of Acanthias the homologues of these processes have not met,
though the halves of the mandible are already united (Sewert-
zoff, 99, p. 289). Similarly for other embryonic Klasmobranchs
and for the adult Notidanus (Gegenbaur, 72, p. 187).

It has already been pointed out that a line of advancement
or retrogression can be recognised within the Teleostean
class, in the gradual reduction of the metapterygoid region
(Text-figs. 1—4). Reversing the order of procedure, and
passing from forms without metapterygoid process (fig. 1) to
those lowly forms in which it is strongly developed (fig. 4),
other structures begin to appear. Thus, in the larval Amia
(Pl. 31, fig. 60), the metapterygoid region, besides being
large and plate-like, sends from its upper border towards the
trabecule (tr.) a large process (p.), the pedicle. Below the
exits of the optic and trigeminal nerves the trabecule, which
are here pierced by the carotid artery (ca.), project slightly
towards it.

In Salmo trutta (Text-fig. 4) the same condition prevails,
and the pedicle is almost equally well developed. This is
figured, but not referred to by Winslow (98).

These facts suggest that at some former time an actual
connection or articulation must have existed between the

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 571

pedicle and the trabecule. Such a condition exists in
Lepidosteus (Text-fig. 5), in which, at all stages from the
embryo two thirds of an inch long to the full-grown adult
(Parker, 82), the pedicle is strongly developed, and forms an

Text-Ficgs. 1—5.—Diagrams illustrating the various modifications
exhibited by the quadrate and its processes in the larve of Bony
Fishes.

Fig. 1, Syngnathus (after Pouchet) ; Fig. 2, Belone (orig.) ; Fig. 3, Zoarces
vivipara (orig.); Fig. 4, Salmo trutta (modified from Winslow); Fig. 5,
Lepidosteus osseus (after Parker), e. Ethmoid. e.p.c. Parethmoid
cornu. e¢. pr. pre-ethmoid cornu. Aym. Hyomandibular. gw. Quadrate.
qu. pa. Palatine process of quadrate. gw. p. Pedicular process of same.
qu. m. Metapterygoid process of same.

articulation with the ‘ basipterygoid process.” This stands
out from the same region of the trabecule as that just
described for Amia. In the extinct Lepidotus also there is a
stout process of the metapterygoid which bears a large
facette, and which, according to Smith Woodward (95, p. 79),

Hie H. H. SWINNERTON.

may have articulated with a lateral element in the cranium.
I have no doubt that this is the homologue of the pedicle in
Lepidosteus.

In Osteoglossum, Bridge (95, p. 309) describes a similar
condition, but here it is a process of the parasphenoid, which
supplies a surface of articulation for the process from the
metapterygoid. He regards it as a “striking example of
parallelism in evolution.” May it not, however, be a case in
which the pedicle has been retained, whilst the basipterygoid
process has been lost and functionally replaced ?

Whether this is so or not, the fact remains that the lowly
Teleosts and Amia bring the bony fish type near to one
having a pedicular articulation. This leaves little doubt
that the condition found in Lepidosteus is not secondary,
as Bridge supposed, but primary ; and that in the ancestor
the quadrate cartilage was attached, not merely to the edge
of the ethmoid plate, but also to the posterior region of
the trabecule somewhere between the exits for the second
and fifth nerves.

Huxley’s comparison between the subocular arch of the
frog and the hyosuspensorial apparatus of the Teleosts (58)
onthe one hand, and the cranio-facial skeleton of Petromyzon
(76) on the other, implied a resemblance between the latter
two, which, as pointed out by Pollard, he ceased to recognise
later. Nevertheless, this resemblance does exist, and in no
respect is this more clearly shown by Petromyzon than in the
fusion of the arch, anteriorly with the ethmoid and _ pos-
teriorly with the trabecule, between the exit of the optic
nerve and below that of the trigeminus (Parker, 83, p. 401).
Nor is continuity of substance to be regarded as an important
difference, for, as shown above, it exists between the palatine
andethmoid in the latest developmental stages of some Teleosts,
and has been also observed by Parker in his earlier stages of
Lepidosteus (82, p. 490). Moreover, in my Stage III, the
hyomandibular forms a continuum with the auditory capsule.
Again, in Myxine all these points are marked by the presence
of a different kind of cartilage from the rest,

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 578

If Pollard was right in regarding the hyomandibular of
Teleosts as part of the same cartilage to which the quadrate
belongs, then this has a third connection through the auditory
capsule with the cranium, and increases the resemblance to
the Marsipobranch, for both Myxine and Bdellostoma possess
an otic attachment in addition to the palatal and pedicular
ones.

Turning to Klasmobranchs, and descending from the more
specialised to the less specialised forms, it is noticeable that,
apart from the hyomandibular, in the former, e.g. Scyllium,
the palato-quadrate is suspended only by ligament ; in inter-
mediate forms, e. g. Acanthias, a palato-basal process or
pedicle is present; and in the latter, e. g. Heptanchus, which
is undoubtedly a lowly type, suspension is by means of both
otic and pedicular processes. ‘There are some forms, e. g.
Cestracion, which do not conform to this, but the constant
recurrence of the pedicle amongst Selachii points back to an
ancestor which certainly possessed that feature. Whether
an otic articulation was also present must, in view of Huxley’s
(76, p. 44) and Sewertzoff’s (99, p. 299) opinions that it is
secondary, and Gegenbaur’s (72, p. 186) and Pollard’s
(94, p. 25) that it is primary, be left undecided until the
homology of the Teleostean hyomandibular is definitely
determined.

Sewertzoff considers that the palato-basal process was the
primary attachment for the mandibular arch (99, p. 299),
because at a very early period in the development of
Acanthias it is united with the trabecule by dense tissue.
The fact that in Petromyzon the facial skeleton commences as
an outgrowth—the pedicle—from the trabecule (Parker, 83,
p- 441), that in Lepidosteus the pedicle appears very early,
and that in T'eleosts the palatine process always grows out
from the quadrate (Stohr, Pollard, and above in Stage 1),
strongly supports this view, and points to a remote time when
neither palatal nor otic processes existed.

VOL, 45, PART 4,.—NEW SERIES. QQ

574 H. H. SWINNERTON.

Systematic Position and Affinities of
Gasterosteus.

The present views of the relationships of the Gasterosteidee
can be best stated by entering briefly into the history of their
growth.

On account of the presence of spines in the fins, and of the
union between the suborbital bone and the pre-operculum,
Cuvier and Valenciennes (29) classified this family with the
mail-cheeked Acanthopterygu, which included the Trighde
and Cottide. ‘Thirty years later Giinther (61), whilst still
retaining it among the Acanthopterygii, separated it from
the mail-cheeked forms; and because of its spinous anterior
dorsal fin and the abdominal position of the pelvics,
put it with the Fistularidee in a separate division, the
Acanthopterygii Gastrosteiformes, and indicated his ideas
of the affinities of this by placing the Mugiliformes and
the Centrisciformes on either side. In a paper, which
has since formed a basis for much recent systematic
work, Cope, because of the position of the pelvic fins
(70, p. 456), separated these three divisions from the
Acanthopterygil, creating for the Mugiliformes the distinct
order Percesoces, and for the other two, the Hemi-
branchii. The latter he regarded as annectant between the
former and the Lophobranchiu; looking upon the weak
branchial apparatus, the presence of interclavicles, the simple
post-temporals, the prolongation of the muzzle, and the
presence of ganoid plates on the body, as approximations to
the Lophobranch type; and upon the character of the dorsal
spines as an indication of relationship to Nematocentris, one
of the Atherinide. Later systematists have followed Cope
more or less closely. Gill (93) placed the Hemibranchii next
the Lophobranchu, and considered it as equivalent not
merely to the Acanthopterygii, but to an order—the Telo-
cephali—containing these and the Haplomi, Scomberesocide,
and Percesoces. Jordan and Evermann (96) acknowledge
that it is closely allied to the Lophobranchii, and that though

MORPHOLOGY OF 'TELEOSTEAN HEAD SKELETON. 575

well marked off from the Percesoces and other 'l'elocephali, it
descended from the ancestors of these groups. In this they
are followed by Kingsley (00), who adds to it the extinct
Dercetidze (Hoplopleuride) which “show relations towards
Belone.” ‘This year Boulenger (01, p. 378) remarks on
their close affinity with the Lophobranchii, more especially
through the Fistularide and the fossil Pseudosyngnathus.

It may at once be stated that there is nothing to support
Parker’s (68) suggestion of Siluroid affinity for Gasterosteus.

In order to see how far these views are supported by the
study of the head skeleton, we may now briefly compare that
of Gasterosteus with that of Syngnathus, a typical Lopho-
branch, and of Fistularia, atypical Hemibranch. MeMurrich’s
paper (88), supplemented and confirmed by my own observa-
tions, is my authority for details concerning the former.
Klein (84, 86) on the cranium, and Rutter (v. Jordan and
Kvermann) on the branchial skeleton, are the only
workers who, so far as my knowledge goes, have dealt with
the latter. To attempt to explain the why and the wherefore
of Klein’s tangle in describing the auditory region would be
of no use for my present purpose; I shall therefore give the
results of my own examination without reference to his.

Taking the various parts in the same order as in the former
part of this paper, the consideration of the hinder region
of the cranium comes first.

That of Gasterosteus is not so compressed dorsi-ventrally as
the other two, but all are alike in the absence of an opistholic
and basisphenoid, the even upper surface, the sculpturing of
the roofing bones, the simplicity of the post-temporal, the
essential shape of the ethmoid, and the great size of the
supra-occipital, which separates the parietals widely, and
appears to separate the hinder portion of the frontals. In
Gasterosteus the exoccipital extends forwards between the
pterotics and basioccipital to the pro-otic. In the others the
pterotic extends ventrally to the basioccipital, and also part
of the way into the large membranous space between this and
the pro-otic, thus separating the exoccipital widely, and the

076 H. H. SWINNERTON.

basioccipital partially, from the pro-otic. In Gasterosteus the
space just referred to is represented by an area of cartilage
(Pl. 29, fig. 21). In the sphenoidal region of Fistularia
the pro-otic completely encloses the foramen for the exit
of part of the fifth nerve, and forms the hinder boundary
of the other exit. The large alisphenoid forms the front
boundary for the rest. No eye-muscle canal is present,
consequently the parasphenoid les flat against the floor of
the cranium. Laterally it sends out processes up to the
sphenotic. In Syngnathus MeMurrich was doubtful about
the presence of an alisphenoid, but the enclosure of the tri-
geminal nerve exit by the pro-otic, and the absence of the
eye-muscle canal, furnish points of resemblance to Fistularia ;
whilst in the union of parasphenoid and frontal processes it
is similar to Gasterosteus.

Thus the last named is more normal in the occipital region
and more advanced in the sphenoidal than the other two, and,
on the whole, differs more from them both than they do from
one another.

In the anterior portion of the cranium, Fistularia and
Syngnathus present the same features as those given above
for the Gasterosteus, but it is greatly elongated, and almost
completely ossified. In the first this region is propor-
tionally much wider, because the narrow pre-ethmoid is
supplemented laterally by the nasals. Apart from this, and
the presence of teeth on the vomer, it bears a much closer
resemblance to the second than to the last, for though
Gasterosteus spinachia has an elongated ethmoid, this is
still almost wholly cartilaginous.

In the visceral skeleton all are alike in the tendency
towards a weakening of the branchial apparatus, in the great
forward slant of the hyomandibular, in the great elongation
of the symplectic, in the great reduction or complete sup-
pression of the metapterygoid cartilage, in the absence of an
ectopterygoid, and in the possession of the acrartete condition.

In Fistularia the reduction of the branchial skeleton has
advanced much further than in Syngnathus, for all the basi-

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 577

branchials and the fourth epibranchial are absent. The
pharyngobranchials of the second to fourth arches are
present, but, unlike those of Gasterosteus, the first two are
fused ; the third is free, and all are rod-like, and lie one
behind the other. In Syngnathus the first and second
basibranchials and the second hypobranchial alone are
present ; the fourth epibranchial has gone, but the edentulous
pharyngobranchials, though rod-like, occupy the same position
relatively to one another as in Gasterosteus.

In the hyoid arch the basihyal, though present during
development, is absent in the adult Syngnathus, but attains
a great length in Fistularia.

Of the bones immediately concerned in the gill-cover and
branchiostegal membrane, the operculum alone survives in
Syngnathus, but are all present, together with five branchio-
stegal rays, in the Fistularia. The interoperculum, which
MeMurrich failed to recognise as such, bears, by reason of its
position, a superficial resemblance to the gular plates of
Polypterus, and was mistaken for such by Parker (68).
Otherwise, as in Fistularia, it is quite normal in its relation-
ships.

In describing the other suspensorial bones, MceMurrich
mistook the pre-operculum for the infra-orbital, and con-
sequently went wrong in his identification of the rest. The
true infra-orbital, or rather the first bone of the suborbital
series (fig. 50, so.'), articulates with the parethmoid (e. p. b.)
above, and forms the lower border of the narial opening.
Ventrally it appears to divide into two lamine, lying on the
outer and inner sides respectively of the cheek muscles, and
is attached by its lower border to the combined symplectic
and pre-operculum (sym. + 0. pr.). In front of the latter lies
the greatly extended quadrate (qu.), of which only the small
part indicated by the dotted line originated by ossification of
cartilage. Along its upper and anterior borders lie three
bones, a, b, c, whose homologies are uncertain; b and ¢
together have all the relationships of the pterygoid in the
stickleback, but as 5 is developed in relation to the vestigial

578 H. H. SWINNERTON.

metapterygoid process, it must be the metapterygoid bone.
ais probably the nasal. The palatine is insignificant and
edentulous. In Fistularia the inner lamina of the suborbital
bone alone remains, the quadrate is much larger posteriorly,
and the pterygoid bone bears a close resemblance to that
of Gasterosteus. Between the hinder process of the last-
named bone and the suborbital is the undoubted metaptery-
goid, which thus occupies a similar position to, but is much
smaller than, a in Syngnathus.

Thus once more this fish exhibits a much closer resem-
blance to Syngnathus than to Gasterosteus. At first sight
one is inclined to think that this may be in some way
associated with the long snout. ‘That the formation of such
a feature does not, however, of necessity produce the same
characters, is well shown by the differences which existed m
the pterygoid and metapterygoid of the two fishes. Again,
the long-snouted Mormyroids do not possess the acrartete
condition, but one which might be most aptly described
as an elongated panartete condition.

Add to this the fact that forty years ago Kner and
Steindachner (68, p. 28) were driven to conclude that the
extinct Pseudosyngnathus was intermediate between Lopho-
branchs and Fistulariide, and to prophesy that these two
groups would some day have to be relegated to the same
order.

Again, consider the latest diagnoses of these two orders
(Jordan and Kverman, 96, pp. 741, 759). After the elimina-
tion of those characters common to both; those mentioned as
characteristic of one, though equally characteristic of the
other, e. g. bony plates; those based upon error, e. g. condition
of branchial apparatus in Syngnathus ; those not common td
all members of the order, e. g. for Hemibranchs, the elongated
anterior vertebra, which are not even indicated in Aulorhyn-
chus and Gasterosteus ; the only distinctive features left are
the tufted gills and single opercular bone for the one, as
opposed to the pectinate gills for the other.

‘To retain these two groups of fishes as distinct orders in

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 579

the face of these considerations would savour too much of
unnecessary conservatism.

On the other hand, when put together they form a com-
pact and natural group, clearly marked off from the Acan-
thopterygii by the reduced condition of the metapterygoid,
the shape of the ethmoid, and the abdominal position of the
pelvic fins ; and further from the Percesoces and Haplomi by
the acrartete condition, the elongated symplectic, the strong
tendency to the reduction of the branchial apparatus, and the
possession in the trunk of so-called interclavicles and of bony
scutes, which may either form a more or less complete arma-
ture, or be developed merely along dorsal, ventral, and lateral
lines.

There is some difficulty im determining what degree of im-
portance must be attached to this last feature. Sagemehl
(85, p. 3) has brought forward a number of valid arguments
against its reliability ; but when, in a natural group of fishes
such as this, all the members, including even the scaled
forms, possess bony scutes, generally arranged in definite
order, it can only be regarded as something more than a
superficial similarity. That a group of Teleosts should have
arisen in which the ganoid plates along some lines were re-
tained, instead of being converted into scales, is a possibility
which receives an air of probability from the fact that in
Gasterosteus the plates of the lateral series are provided
with that peg-and-socket arrangement (Parker, 68, p. 41) so
characteristic of ganoid plates.

Other features there are—such as the sculpturing of the
roofing bones in the less specialised members; the simple
post-temporal; the absence of opisthotic; the general ab-
sence of maxillary process and the teeth in the palatine; the
single pterygoid bone; the complete withdrawal of the
cranial cavity from the orbit, even in the lowest members ;
the general shape of the body; the elevated pectorals ; the
position of dorsal and ventral median fins relative to one
another—to which much importance cannot be attached
sinely, but which, when taken together, help substantially to

580 H. H. SWINNERTON.

strengthen the conviction that these forms should be referred
to a separate sub-order. I propose, therefore, that these fishes,
hitherto placed in the sub-orders Hemibranchii and Lopho-
branchii, should be put together in one sub-order, having the
name horacostei,! as expressive of the presence in all of a
more or less complete bony armature, and especially of infra-
clavicles.

The two old sub-orders might still be provisionally retained
as divisions until such forms as Solenostomaand Aulorhynchus
have been thoroughly examined ; but, in view of the facts
already put forward, the sticklebacks should certainly be
referred to a separate division, the Gasterosteoidei.

One feature of great interest about the Thoracostei is the
possession of undoubted highly specialised characters side by
side with primitive ones, which indicate that they branched
off from lowly physoclistous, or even physostomous fishes.
We may now briefly inquire into their relationships to other
fishes.

To the best of my knowledge, the nearest approach to this
order among living fishes is made by the Scomberesocidee.
Indeed, so close is this approach that on a consideration of
the head skeleton alone one would be almost obliged to place
Belone in the same sub-order with Gasterosteus. Give its
cranium an arched instead of a flattened roof, replace its ali-
sphenoid by overlapping frontal and parasphenoid processes,
shorten the premaxillee and mandible to a normal length, elon-
gate the symplectic still further, and it would be extremely
difficult to find any feature of importance in which the two
crania differed, for in the Belone all the roofing bones are
sculptured; in spite of its lowly affinities, its opisthotic is
absent ; the ethmoid, though more cartilaginous, is of the same
type; the branchial apparatus is an exact replica of that in
Gasterosteus in the number and nature of the basibranchials,

1 Since the above was written, Smith Woodward has published his much-
needed fourth volume of the ‘Catalogue of Fossil Fishes in the British
Museum.’ In this he has united the two old sub-orders under the one name
Jiemibranchi,

MORPHOLOGY OF THLEOSTEAN HEAD SKELETON, 581

in the number, shape, and proportional size of its pharyngo-
branchials, and in all other features except the fusion of the
vestigial elements of the fiftharch. Again, the hyomandibu-
lar is of the same shape, though its articulations are more
generalised; the metapterygoid is equally reduced; one
pterygoid line alone is present; the palatine is small, edentu-
lous, and lacks a maxillary process; finally it presents the
acrartete condition.

The similarity is so great that one may say with consider-
able truth that the little stickleback is but a slightly
specialised Belone.

In the trunk region, however, though the pectorals are
raised, the pelvics, abdominal, and the arrangement of the
other fins is the same as in Thoracostei; yet the complete
absence of bony plates and infra-clavicles gives some excuse
for not including the Scomberesoces in the new sub-order.

Boulenger (01) has recently placed them as a family of the
Percesoces, and the possession of a reduced metapterygoid in
Atherina supports this view. It is interesting to note that, at
least so far as the head skeleton is concerned, they depart
from the other members of that sub-order in just those
features in which they approach the Gasterosteoidei; for all
the other Percesoces | have examined possess a suspensory
pharyngeal, a well-developed opisthotic (Starks, 99), and the
acrartete condition.

If this is their true position, it would strongly tend to
show that the acrartete condition was derived from a lowly
disartete by loss of the post-palatine articulation. If it is
not, then the most convenient way would be to keep the
Scomberesoces in a separate sub-order as before, and to
speak of them and the closely allied Thoracostei collectively
as the Scomberesocine series. For my own part, I do not
like the word Scomberesoces, because it implies relationships
which do not exist. On the other hand, the term Synen-
tognathi of American writers is equally inapplicable.

Klein has shown (79, p. 120) that the opisthotic may be
absent in very diverse families of ‘l'eleosts, and that even in

582 H. H. SWINNERTON.

those forms in which it is constantly present its relationships
are extremely variable. Nevertheless it cannot be a mere
coincidence that all the forms belonging to the Scomber-
esocine series, including so lowly a form as_ Belone, should
be lacking in this bone. It must therefore be lacking also,
or much reduced, in their immediate ancestors and allies.

Turning now to the paleontological side of the question,
we find the Thoracostei well represented by forms differing
but little from those of the present day as early as the
Hocene.

There seems at present much more to be said against than
for Kingsley’s suggestion concerning the relationship between
these forms and the extinct Dercetiformes or Hoplopleurides
(Pictet, 54, p. 215). Smith Woodward, to whom I am
indebted for the opportunity of examining a number of the
original fossils, places Belonorhynchus, with considerable
show of reason, among the Chondrostei (95), and the re-
maining forms in two families, the Dercetidze and Kncho-
dontidz (01). Despite the presence of rows of bony scutes
along dorsal, lateral, and ventral lines, the condition of the
cranial roof and mandibular suspensorium alone make it
very improbable that they have anything to do with the
Thoracostei.

One other group of bony fishes—Zanclide, Acronuride
(Teuthidide), and Plectognathi—demands a reference be-
cause it possesses the acrartete condition. Jordan and
Evermann (96) place the first two families together with
the EHphippidee and Cheetodontide in one sub-order, the
Squamipinnes, and say that “The Teuthididz and Balis-
tide are as nearly related to each other as the Kphippide
are to the Cheetodontide ;” and again: “There can be no
doubt of the common origin of the Balistidee and Teuthididee,
and that the divergence is comparatively recent.” For this
reason they subordinate the Plectognathi to the Acantho-
pterygii, and place them as an offshoot of the Squamipinnes.

What has already been said concerning this ethmoid
and suspensorial region in 'l'euthis and Balistes tends to

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 585

support the statements (just quoted) of these authors very
strongly, but tells with equal strength against the relation-
ships of these forms to the Cheetodontide and Ephippide.
In Cheetodon plebeius (fig. 46), the ethmoid region is an
extremely specialised one of the Pagellus type, for though
the palatine lies against the parethmoid, it is united to that
exclusively by ligaments, and the articular surfaces have
aborted. The palatine, moreover, has the large maxillary
process so characteristic of the Acanthopterygu, and mobility
for the suspensory apparatus is gained, as in Cyprinoids, by
articulation between the palatine and pterygoid bones.
Judging by the character of the ethmoid region, and its
relation to the palatine alone, the relationships of the
Plectognathi, and the undoubtedly closely allied Zanclide
and Acronuride, must not be sought within the Acantho-
pterygii; nor, judging by the well-developed metapterygoid,
and presence of both pterygoids, must they be sought in the
Scomberesocine series, but somewhere lower down.

V. SuMMARY AND CONCLUSIONS.

1. The cranial flexure, together with other features in the
shape of the embryonic head skeleton in Teleosts, is probably
a mechanical effect due to differences in the degree of
distensibility between the dorsal and ventral surfaces of the
brain, and to the presence of skeletal structures in close
association with the latter (pp. 507—509).

2. The presence of an epiphysial bar, with consequent
division of the large dorsal, cranial fontanelle into an anterior
and a posterior portion, is a common feature among Teleosts
during development (pp. 516, 517).

3. The Ostariophysi differ from all other Teleosts in the
retention of this early developmental condition of the cranial
roof in the adult (pp. 525, 526).

4, The intra-cranial notochord, so far from undergoing re-
duction, never at any stage ceases to grow (pp. 513, 516, 523).

5. In Gasterosteus, during embryonic life, those skeletal

584, H. H. SWINNERTON. ,

elements immediately concerned in the support of the jaws
and operculum, and in the attachment of associated muscles,
seem to undergo a considerable acceleration in the rate of
development as compared with the rest of the head skeleton
(pp. 5384, 535).

6. Among Teleosts and the immediately related Ganoids,
three types of palato-ethmoidal relationship exist (pp. 538,
539, 551—557).

(a) The Panartete, in which the palatine cartilage or its
derivatives is attached to the ventral surface of the
ethmoid for the whole length of this, from the par-
ethmoid to the pre-ethmoid cornua, e.g, Amia, probably
presented also by many Malacopterygil (Isospondyli).

(b) The Disartete, in which the attachment is at the
parethmoid and pre-ethmoid cornua, but not at any
intermediate point, e.g. Hsox, also presented by the
Salmonidee, Cyprinodontidw, Acanthopterygu, and prob-
ably some Malacopterygu (Isospondylh).

(c) The Acrartete, in which the attachment is confined
solely to the pre-ethmoid cornua, e. g. Gasterosteus, and
also presented by the Thoracostei, Scomberesoces,
Plectognathi, Zanclidze, Acronuride, and in a modified
form by Lepidosteus.

7. The study of the adult anatomy and comparative
ontogeny of the head skeleton in Hlasmobranchs and Teleo-
stomes seems to point to a common ancestral stock for these
two great divergent branches of fishes. It presented among
other features the following :

A short embryonic life; weak cranial flexure; trabecule

united to the extreme anterior end of the parachordals.

A wholly cartilaginous cranium, possessing trabecular,
parachordal, and occipital portions (pp. 560—562).

A cranium having a large dorsal fontanelle, which may or
may not have been divided by a transverse epiphysial
bar. Also two lateral fontanelles for the passage of the
optic, and possibly also the trigeminal and facial nerves.
Also a ventral or pituitary fontanelle. Also a large

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON, 585

opening between the cavum cranii and auditory capsule
(pp. 562—568).

A quadrate cartilage supporting a lower moveable jaw,
formed by the union of two cartilages in the middle line,
and bearing dorsally two, possibly three processes ; an
anterior one, parallel to its fellow, and not united with
it, but with the ethmoid plate, so that a moveable upper
jaw did not exist; a middle one articulating with the
trabecule, in the region lying between the optic and
trigeminal nerves ; a posterior one articulating with the
auditory capsule (pp. 568—573).

A branchial apparatus consisting of at least five arches,
already segmented into four parts.

Balfour’s term Prolognathostomata (81, p. 271) would be
sufficiently expressive of such a type.

8. The manner of mandibular suspension in Teleosts is

insufficiently described by the term Hyostylic (pp. 569, 570).

9. The Lophobranchii and Hemibranchii should no longer
be kept in separate orders, for they together constitute a
natural. group, which may be designated the Thoracostei
(pp. 575—579).

10. The Scomberesoces, through the Gasterosteoidei,
approach more closely to the 'Thoracostei than do any other
living Physoclisti, and seem to form with them a compact series,
which may be provisionally spoken of as the Scomberesocine
series (pp. 580, 581).

11. As judged by the study of the ethmoid and suspensorial
regions, the Zanclide and Acronuride are closely allied to
the Plectognathi, but the affinities of these forms must not be
sought amongst living Physoclisti (pp. 582, 583).

BIBLIOGRAPHY.

97. Axis, EK. P., jun.—* The Morphology of the Petrosal Bone, and of the
Sphenoidal Region of the Skull of Amia calva,” ‘Zool. Bull.,’ vol. i,
1897, pp. 1—26.

586 H. H. SWINNERTON.

98.

78.

81.

90.

95.

Ol.

The

95.

83.

70.

29.

99.

80.

72.

78.

87.

93.

Ol.

Auuts, BE. P., jun.—‘ On the Morphology of certain of the Bones of the
Cheek and Snout of Amia calva,” ‘Journ. of Morphology,’ vol. xiv,
1898, pp. 425—466.

Barrour, F. M.—‘ A Monograph on the Development of Elasmobranch
Fishes,’ London, 1878.

Batrour, F. M.—‘ A Treatise on Comparative Embryology,’ London,
vol. ii, 1881.

Bearp, J.—< The Inter-relationships of the [chthyopsida,’” ¢ Anat. Anz.,’
Bd. v, 1890, pp. 146—159, and 179—188.

Boutencer, G. A.—‘ Remarks on some Cranial Characters of the
Salmonoids,” ‘ Proc. Zool. Soc.,’ 1895, pp. 299—302.

Boutrencer, G. A.—‘ Les Poissons du Bassin du Congo,’ Bruxelles,
1901.

Brings, ‘T. W.—“* The Cranial Osteology of Amia calva,” ‘ Journ. of
Anat. and Phys.,’ vol. xi, 1877, pp. 605—622.

Brive, T. W.—“ On Certain Features in the Skull of Osteoglossum
formosum,” ‘ Proc. Zool. Soc.,’ 1895, pp. 8302—310.

Brooks, H. J.—“The Osteology and Arthrology of the Haddock” (Gadus
eelefinus), ‘Proc. Roy. Dublin Soc.’ (N. 8.), vol. iv, 1883, pp.
166—196.

Corr, BH. D.—** A Contribution to the Ichthyology of the Lesser Antilles,”
‘Trans. Amer. Phil. Soc.,’ vol. xiv, pp. 445—483.

Cuvier and VALENCIENNES.—‘ Histoire naturelle des Poissons,’ Paris,
vol. iv, 1829—1849.

Drenpy, A., and Howrs, G. B.—‘'The Hatching of Tuatara Eggs,”
‘ Nature,’ 1899, p. 340.

Gann, M.—“ Ueber die Entwickelung des Kopfskeletts bei Knochen-
fischen (Rhodeus, Gasterosteus),” ‘ Zoologischer Anzeiger, Bd. ii,
1880, pp. 140, 141.

GrGENBAUR, O.—‘ Das Kopfskelet der Selachier,’ Leipzig, 1872.

GrcGENBAUR, C.—‘ Ueber das Kopfskelet von Alepocephalus ros-
tratus,” ‘Morph. Jahrb.,’ Bd. iv, 1878, Supplement, pp. 1—40.

GrGENBAUR, C.—“ Ueber die Occipitalregion und die ihr benachbarten
Wirbel der Fische,” ‘ Festschrift fiir Albert v. Kolliker,’ Leipzig,
1887, pp. 1—81.

Git, T.—“ Families and Sub-families of Fishes,” ‘Memoirs Nat. Acad.
of Sci.,’ Washington, vol. vi, 1893, pp. 127—188.

Gontrr, A.—** Ueber die Kiemen der Fische,” ‘ Zeitsch. wiss. Zool.,’
Bd. Ixix, 1901, pp. 583-—575.

61.

70.
oo.

91.

ol.

58.

59.

Zale

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 587

GintHer, A.—‘ Catalogue of the Acanthopterygian Fishes in the British
Museum,’ London, vol. iii (Systematic Synopsis), 1861.

HaEckeEL, E.—‘ Natiirliche Schopfungsgeschichte,’ Berlin, 1870.

Hatta, S.— Contributions to the Morphology of Cyclostomata. II. On
the Development of Pronephros and Segmental Duct in Petromyzon,”
‘Journ. of the College of Science,’ Tokyo, vol. xiii, 1900, pp. 318—419.

Howes, G. B.—‘‘ On some of the Hermaphrodite Genitalia of the Codfish
(Gadus morrhua),” ‘Journ. Linn. Soc.,’ vol. xxiii, 1891, pp.
539—557.

Howss, G. B., and Swiynerton, H. H.—“On the Development of the

Skeleton of Tuatura (Sphenodon punctatus),” ‘ Trans. Zool. Soce.,’
vol. xvi, 1901, pp. 1—78.

Houxtey, T. H.—“On the Theory of the Vertebrate Skull” (the
Croontan Lecture), ‘Proc. Roy. Soc.,’ 1858, pp. 881—457.

Huxtey, T. H.—‘ Observations on the Development of some Parts of
the Skeleton of Fishes,” ‘Quart. Journ. Micro. Sci.,’ vol. vil, 1859.
Huxtey, 'T. H.—‘ A Manual of the Anatomy of Vertebrated Animals,’

London, 1871.

76a. Huxtey, T. H.—* On Ceratodus Forsteri, with Observations on the

Classification of Fishes,” ‘ Proc. Zool. Soc.,’ 1876, pp. 22—59.

76 6. Huxtry, T. H.—‘On the Nature of the Cranio-facial Apparatus of

96.

oo.

79.

84.

85.

86.

93.

63.

Petromyzon,” ‘Journ, Anat. and Phys.,’ vol. x, 1876, pp. 412—429.
JorpaNn, D. S., and Evermany, B. W.—‘ The Fishes of North and
Middle America,’ Washington, 1896—1900, vols. i—iv.
Kineos ey, J. S.—‘ Text-book of Vertebrate Zoology,’ London, 1900.
Kien, —.— Beitrage zur Osteologie des Schidels der Knochenfische,”’
‘ Jahresheft des Vereins fiir Vater Naturkunde,’ Wurttemberg, Bd.
xxvili, 1872, pp. 66—124.
Kiem, —.—* Beitrage zur Bildung des Schadel der Knochenfische,”
ibid., Bd. xl, 1884, pp. 129—255.

Kuriy, ——‘ Beitrage zur Bildung des Schadel der Knochenfische,”

ibid., Bd. xli, 1885, pp. 107—261.

Kuri, ——“ Beitrage zur Bildung der Schadel der Knochenfische,”
ibid., Bd. xlii, 1886, pp. 205—299.

Kuiaatscu, H.— Beitrage zur Vergleichenden Anatomie der Wer-
belsaule,” ‘Morph. Jalrb.,’ Bd. xx, 1893, pp. 148—186.

Kner, R., and SrernpacuNnER, F.—‘‘ Neue Beitrage zur Kenntniss der
fossilen Fische Oesterreich,’ ‘ Denksch. Akad. der Wiss.,’ Wien,
1863, Bd. xxi, pp. 17—86.

588 H. H. SWINNERTON.

83. McMornricu, J. P.—“‘On the Osteology and Development of Syn-
enathus Peckianus,”’ ‘Quart. Journ. Micros. Sci.,’ vol. xxiii, 1883,
pp. 623—650.

68. Parxrr, W. K.—‘ Monograph on the Shoulder-girdle and Sternum,’
London, 1868.

73. Parker, W. K.—“<On the Structure and Development of the Skull in
the Salmon,” ‘ Phil. Trans. Roy. Soc.,’ vol. elxiii, 1873, pp. 95—145.

76. Parker, W. K.—“‘On the Structure and Development of the Skull in
Sharks and Skates,” ‘ Trans. Zool. Soc.,’ vol. x, 1878, pp. 190—234.

78. Parker, W. K.—“ On the Structure and Development of the Skull in
the Lacertilia,’ ‘Phil. Trans. Roy. Soce.,’ vol. clxx, 1879, pp.
595—640.

81a. Parker, W. K.—< On the Structure and Development of the Skull in
Sturgeons,” ‘ Phil. Trans. Roy. Soe.,’ 1882, pp. 139—185.

814. Parker, W. K.—* On the Development of the Skull in Lepidosteus
Osseus,” § Phil. ‘Trans. Roy. Soc.,’ 1882, pp. 443—492.

83. Parker, W. K.—“ On the Skeleton of the Marsipobranch Fishes. I and
II,” * Phil. Trans. Roy. Soe.,’ 1883, pp. 373 —457.

54. Picret, F. J.—‘ Traité de Paléontologie,’ 2nd edit., Paris, ts ii,
1854.

97. Piatt, J. B—** The Development of the Cartilaginous Skull and of the
Branehial and Hypoglossal Musculature in Necturus,’ ‘ Morph.
Jahrb.,’ Bd. xxv, 1896—1898, pp. 375—464.

92. Portarp, H. B.—‘On the Anatomy and Phylogenetic Position of
Polypterus,” ‘ Zool. Jahrb.,’ Bd. v, 1892, pp. 885—428.

94. PoLtarp, H. B.—< The ‘ Cirrhostomial’ Origin of the Head in Verte-
brates,” ‘ Anat. Anz.,’ Bd. ix, 1894, pp. 8349—359.

94. Pottarp, H. B.—‘‘ The Suspension of the Jaws in Fish,” ‘ Anat.
Anz.,’ Bd, vi, 1895, pp. 17—25.

95. Pottarp, H. B.—“ The Oral Cirri of Siluroids and the Origin of the
Head in Vertebrates,” ‘ Zool. Jahrb.,’ Bd. viii, 1895, pp. 879—421.

78. Poucnet, G.—* Du Développement du Squelette des Poissons osseux,”
‘Journ. de l’Anat. et de la Phys.,’ t. xiv, 1878, pp. 34—100, 140—153.

SAGEMEHL, M.—‘ Beitrige zur vergleichende Anatomie der Fische:’
84. J. “Das Cranium von Amia calva,” ‘Morph. Jahrb.,’ Bd. ix,
pp. 177—227.

85. III. “Das Cranium der Characiniden,” ‘ Morph. Jahrb.,’ Bd. x,
pp. 1—117.

91. IV. “Das Cranium der Cyprinoiden,” ‘Morph. Jahrb.,’ Bd. xvii,
pp. 489—595.

MORPHOLOGY OF TELEOSTEAN HEAD SKELBYTON. 589

97. Sewertzorr, A—“ Beitrag zur Entwickelungsgeschichte des Wirbel-
tierschadel,” ‘ Anat. Anz.,’ Bd. xiii, 1897, pp. 409—425.

99. Sewertzorr, A.—“ Die Entwickelung des Selachierschidel,” ‘ Festschrift
von Kupffer,’ Jena, 1899, pp. 281—320.

99. Starks, EH. C.—* The Osteological Characters of the Fishes of the Sub-
order Percesoces,’ ‘Proc. U.S. Nat. Mus. Washington,’ vol. xxii,
1899, pp. 1—10.

80. Stour, P. “Zur Entwickelungsgeschichte des Urodelenschadels,”
‘Zeits. f. wiss. Zool.,’ Bd. xxxiii, 1880, pp. 477—526.

82. Stour, P.—‘ Zur Entwickelungsgeschichte des Anurenschadels,” ibid.,
Bd. xxxvi, 1882, pp. 68—103.

83. Stour, P.—“Zur Kntwickelungsgeschichte des Kopfskelettes der Tele-
ostier,” ‘ Festschrift zur dritten Sekularfeier der A. J. Maximiliana,
Wurzburg,’ Leipzig, 1883.

70. Traquair, R. H.—* On the Cranial Osteology of Polypterus,” ‘Journ.
of Auat. aud Phiys.,’ series 2, vol. v, 1871, pp. 166—183.

73. Vrouk, J. A.—“* Verknocherung und die Knochen des Schadels der
Teleostier,” ‘ Niederland. Archiv f. Zoologie,’ Bd. i, 1873.

83. WattueR, J.— Die Entwickelung der Deckknochen am Kopfskelett des
Hechtes,” ‘Jenaische Zeitschrift f. Naturwiss.,’ Bd. xvi, 1883, pp.
57—87.

98. Winstow, G. M.—“ ‘The Chondro-cranium in the Ichthyopsida,”’ ‘ Tuft’s
College Studies,’ Mass., 1898.

95, O01. Woopwarp, A. 8.—‘ Catalogue of the lossil Fishes in the British
Museum,’ vols. iti and iv.

EXPLANATION OF PLATES 28—31,

Illustrating Mr. H. H. Swinnerton’s paper, “ A Contribution
to the Morphology of the Teleostean Head Skeleton,
based upon a study of the Developing Skull of the Three-
spined Stickleback (Gasterosteus aculeatus).”

List of Reference Letters.

a. 6. c., see page 577. a. f. Anterior cranial fontaneile. az, Angular. ar.
Articular. az.c. Auditory capsule, aw. c’. Pillar round which the horizontal

VOL. 45, PART 4,.—NEW SERIES. RR

590 H. H. SWINNERTON.

semicircular canal runs. 6. Brain. 6. /. Fore-brain. 6. 4. Hind brain. 3. m.
Mid-brain. 47. Branchial arch. dr. 6. 1—4, Basibranchials 1—4. dr. ¢.
J—5. Ceratobranchials 1—5. 67. e. 1—4. Epibranchials 1—4. dr. A. 1—3.
Hypobranchials l—3. 7. p. 2—4. Pharyngo-branchials 2—4. dr. 7. Branchi-
ostegal ray. ca. Foramen for internal carotid artery. ch. and ch’. Notochord.
d. Dentary. e. Ethmoid region. e. m. Mescthmoid. e. m. c. Mesethmoid
cartilage. ¢. p. 6. Parethmoid bone. e.p.c. Parethmoid cornu. e. pr.
Pre-ethmoid cornu. e. pr. b. Pre-ethmoid bone. ep. Epiphysis. ep. e.
Epiphysial cartilage. ep. Epiphysial foramen. ep. s. Connection between
epiphysis and supra-occipital region. ey. Hye. ey. c. Eye-muscle canal. fe’.
Fenestre in the floor of the auditory capsule. fe. Fenestra in the side wall
of the auditory capsule. fr. Frontal. jr’. Process from frontal. g. Gape.
hy. Hyoid arch. hy. 6. Basihyal. Ay.c. Ceratohyal. hy. c’. Process of the
ceratohyal. Ay. e. Epihyal. Ay. 4.1, 2. Hypohyal 1, 2. Ay. ¢. Stylohyal.
hym, Hyomandibular. Aym', Opercular process of the hyomandibular, Aym'’.
Articular processes of the hyomandibular. Aym!’”. Articular facets for the
hyomandibular. cf Infundibulum, ¢. Labial cartilage. 7/7. Labial fold.
m. e, External rectus muscle. mé. Meckel’s cartilage. ma. Maxilla. xa.
Nasal. xa’. Process of nasal bone. 0. ¢. Epiotic. o. p. Pterotic. 0. pr.
Pro-otic. 0. pr’. Pro-otic process. o. pr”. Anterior end of the parachordal.
0. sp. Sphenotic. oc. a. Occipital arch. oc. a’. Small process on the occipital
arch. oc. 6. Basioccipital. oc. c. Occipital chondrification. oc. e. Exoccipital.
oc. n’. First occipital nerve. oc. x’. Second occipital nerve. oc. s. Supra-
occipital. op. Operculum. op. ¢. Interoperculum., op. py. Pre-operculum.
op.s. Suboperculum. p. ch. Parachordal. p. ch’. Interparachordal fossa. p. f.
Posterior cranial fontanelle. p. ¢. Post-temporal. pa. Palatine. pa’. Pre-
palatine articular facet. pa”. Post-palatine articular facet. pa. m. Maxillary
process of the palatine. par. Parietal. pg. ec. Ectopterygoid. pg. ex. Ento-
pterveoid. pg. m. Metapterygoid. pma. Premaxilla. pm’. Cartilage on the
ascending process of the premaxilla. ps. Parasphenoid. ps’. Median process
of the parasphevoid. ps’’. Lateral process of the parasphenoid. pé. Pituitary
body. pé.f. Pituitary fossa, gu. Quadrate. qu’. Process of quadrate. gu. m.
Metapterygoid process of the quadrate. gu. pa. Palatine process of the
quadrate, gu. pd. Pedicular process of the quadrate. 7. Rostrum. sd. 4.
Supra-orbital band. sd. p. Post-orbital process. so. 1—3. Suborbitals 1—3,
sym, Symplectic. ¢g.'legmen cranil. ¢r. Trabecule. ¢+’. Point of union of
the trabeculae with the parachordals. vo. Vomer. «. Infra-symplectic
cartilage. I—X. Nerve foramina. * Boundary between the head and body.
** Curve of the egg-shell. Articulation between palatine and pterygoids.

Fic. 1. Stage 1—Dorsal view of the chondro-cranium of a larva, 3°6 mm.
long. x 70.

Fie. 2. Stage I1.—The same of a larva, 5'7mm. long. x 60.

MORPHOLOGY OF TEKELEOSTEAN HEAD SKELETON.

Fic. 3. Stage I11,—The same of a larva, 6°6 mm. long. x 50.

Fig. 4. Stage [V.—The same of a larva, 16°0 mm. long. x 16.

Fic. 5. Stage [V.—Dorsal view of the skull with the membrane bones.
ScoalG:

Fie. 6. Stage I.—Lateral view of the chondro-cranium. x 70.

Fic. 7. Stage I1.—Lateral view of the chondro-cranium. x 60.

Fie. 8. Stage IL].—Lateral view of the chondro-cranium. x 50.

Fic. 9. Stage [V.—Lateral view of the chondro-cranium. x 16.

Fie. 10. Stage 1V.—Lateral view of the skull with the membrane bones.
x 16.

Fig. 11. Stage 1V.—Posterior view of the skull with the membrane bones.

<6:

Fig. 12. Stage 1V.—Ventral view of the posterior region of the chondro-
cranium. x 16.

591

Fie. 13. Stage [1V.—Internal view of the mandibular and hyoid arches with
the mandible removed. x 16.

Fic. 14. Stage I1.—Ventral view of the hyoid and branchial arehes. x 60.

Fie, 15. Stage 111.—Ventral view of the hyoid and branchial arches. x 50.

Fie. 16. Stage 1V.—Ventral view of the hyoid and branchial arches. x 16.

Fie

s. 17—24 are all of the adult ;

Fic. 17.—Side view of the cranio-facial skeleton. x 33.
Fie. 18.—Dorsal view of the cranium. x 34.

Fic. 19.—Side view of the cranium. x 33.

Fie. 20.—Internal view of the cranium. x 33.

Fie. 21.—Ventral view of the cranium. x 33.

Fig. 22.—Ventral view of the hyoid and branchial arches. x 34,

Fic. 23.—LExternal view of the mandibular and hyoid arches with
the mandible removed. xX 4.

Fig. 24.—Posterior view of the cranium. xX 33.

Fig. 25.—Sagittal section of the intercranial notochord of a larva, 6°3 mm.

long.
Fig
Fic
Fig

x 35.

x 80.
. 26.—The same of a young stickleback, 21°0 mm. long. x 80,

. 27.—The same of a full-grown adult, 47-0 mm. long. x 80.

. 28.—Transverse section through the labial cartilage of the adult.

Fig. 29.—The same, but 30 sections behind. x 35.

Fic

. 30.—Transverse section through the pre-ethmoidal and pre-palatine
regions of larva belonging to Stage IV. x 140.

592 H. H. SWINNERTON.

Fic. 31.—Transverse section through the same region of the adult. x 30.
Fie. 32.—Transverse section through the parethmoid region, x 380.

Fic. 33.—Transverse section through the anterior ends of the parachordals
of a larva, 9°0 mm. long. x 140.

Fie. 34.—The same of a young stickleback, 16°0 mm. long. x 70.
Fig. 35.—The same of an adult. x 35.
Fie. 36.—Transverse section through the basi-occipital of the adult. x 27.

Fic. 37.—Median sagittal section through the basis cranii of the adult.
Hl

Fic. 38.—The same of a young stickleback, 14 mm. long. x 70.

Fic.'39.—Transverse section through the region of the infra-symplectic
cartilage of a young stickleback, 16 mm. jong. x 70.

Fic. 40.—Ventral view of the same region. X 37.

Fic. 41.—Transverse section through the auditory capsule and hyomandi-
bular of an embryo, 3°6 mm. long. x 400.

Fic. 42.—Sagittal section through the ascending process of the premaxilla
of the adult. x 30.

Fic. 43.—Lateral view of the ethmoid and quadrate regions of the adult
Esoxlucius. xX 4.

Fic. 44.—Lateral view of the ethmoid and palatine regions of the adult
Scomberscomber. xX 13.

Fic. 45.—The same of the adult Pagellus centrodontis. x 8
Tic. 46.—The same of the adult Chatodox plebius. x 4,
Fig. 47.—The same of a larval Salmo salar, 20 mm. long. xX 25.

Fic. 48.—Projection of the chondrocranium of a young Siphonostoma
typhle, 22 mm. long. x 15.

Fie. 49.—Lateral view of the ethmoid and quadrate regions of the adult
Belone acus. Not size.

Fig. 50.—The same of Syngnathus. x i}.

Fic. 51.—Lateral view of the ethmoid and palatine regions of the adult
Gasterosteus aculeatus. xX 5.

Fig. 52.—The same of Zancluscornutus. x 2.
Fig. 538.—The same of Balistes maculatus (young?). x 28,

Fic, 54,.—Diagrammatic transverse section through the head region of a
sixth-day embryo of Gasterosteus aculeatus; killed within the egg. x 80.

Fie. 55.—The same; killed after being released from the egg. x 80.

MORPHOLOGY OF TELEOSTEAN HEAD SKELETON. 593

Fie. 56.—Diagrammatic sagittal section through the head region of a
sixth-day embryo of Gasterosteus aculeatus; killed within the egg. x 80.

Fie. 57.—The same ; killed after being released from the ege. x 80,

Fie. 58.—Diagrammatic sagittal section through the head region of a just-
hatched larva of Gasterosteus aculeatus of the ninth day. The dotted
outline represents parts belonging to an eighth-day embryo which was killed
within the egg. x 80.

Fic. 59.—Diagrammatic sagittal section through the head region of an
adult of Gasterosteusaculeatus. x 6.

Fie. 60.—Dorso-lateral view of the trabecular and quadrate portions of the
head skeleton of a larval Amia calva, 19 mm. long. x 8.

Fic. 61.—Projection of the chondrocranial roof of a larval Salmo salar,
20 mm. long. x 20.

Fie, 62.—Projection of the chondrocranial roof of a larval Salmo salar,
25 mm.long. x 17.

Notge.—Figures 1—16, 40, 47, 60, were taken from wax models. In all
except Figs. 47 and 60, the uncoloured portions represent pro-cartilage ; those
coloured blue, cartilage; those coloured green, ossifying cartilage; those
coloured yellow, bone without cartilage.

yas
cm Ola sa* A Ver Eps

an | Jietele! UPI See os Aree ted
e) i) ¥4e2 Glin S rane
7 7

ie aan te cae’ bo tld eee ee

ca o+8
re, i) | 7 wy J ’
*
th \ re t a
~ ' : *
. : he ; ? a ines ]
" tr la ' vn a
¥ o 1 = “Bs
aS _ f a dhe
= “ wile ’f
Pl
ps
b gtk | y'}

A,S.Huth, Lith™ London,

Luart Fourn Mur Sea Vol 4EN SPA, 29

H.H.S. del.

A.S.Huth, Lith® London.

30

ateerereseaeet eer pare Eo ————————— —
eeetrest
aston =: Se ee eS SPS.
=
Sos wits
—————
——
a
er
———
ae

Quart FournMuor Sei Vol LENS I 30

Td
~ neers. NE

AS.Huth, lath” London.

-

neem tpn =i tect 8 three

AS Huth, Litht London.

7 . , ~~. oes
2 v8 bide tong
: ‘ ° e ; hey, 4 s
, 0

THE DEVELOPMENT OF ADMETUS PUMILIO. 595

The Development of Admetus pumilio, Koch:
a Contribution to the Embryology of the
Pedipalps.

By

L. H. Gough.

With Plates 32 and 33.

INTRODUCTION.

I BEGAN ny investigations of the development of Admetus
pumilio in November, 1899, During the previous year I
had been studying the development of spiders under Prof.
A. Goette, in Strassburg.

Through the kindness of Prof. R. Burckhardt I received
materials to examine, which he had obtained from Dr. Goelde,
in Para. ‘These materials consisted of two females of
Admetus pumilio, each of which carried a batch of
egos, which will be described under Stages II and V.

In January, 1900, I received two further batches of eggs,
this time through the kindness of Dr. G. Hagmann, also in
Para; unfortunately only one of these batches was of use ;
from it I procured Stages I, HI, and IV; the other had
been deserted by the mother animal, as Dr. Hagmann stated.
Cuts through embryos coming from this batch only show an
inner mass of yolk surrounded by a thick and dense layer of
bacilli. I was not able to procure any further material.

At this place I must also express my thanks to Prof.
F. Zschokke, Director of the Zoological Institute of the

596 be. He COUGH:

University of Basle, in whose laboratory this paper was
prepared, and to Prof. R. Burckhardt, to whom I am much
indebted for the help he rendered me during my studies, and
more especially whilst working at this paper.

I must also express thanks publicly to Dr. Goelde and
Dr. Hagmann for having procured me the materials, and to
my friend and fellow-student Oscar Huber, for some of the
drawings given in this paper.

TECHNICAL REMARKS.

It was at first very difficult to obtain good cuts through the
egos, because of the presence of the great amount of yolk
which they contained. When embedded in paraffin the yolk
always crumbled away before the knife of the microtome.

The following is the method I usually used :—'The embryos
were brought through successive degrees of alcohol into
absolute alcohol. Next they were left in celloidin for some
days. hey were then taken from the celloidin and put
directly into chloroform. This has the double effect of
making the celloidin firm, and of rendering the embryos
ready to be transferred into paraffin. After having been
treated thus, the embryos were cut as ordinary paraffin
objects. The cuts were attached to the object-glass with
water; when all the cuts were arranged on the glass, the
water was quickly removed with a piece of blotting-paper,
whereupon very fluid celloidin was poured over the cuts;
they were then allowed to dry, but being still enclosed in
paraffin the cuts could not shrink in any way. When quite
dry the coating of celloidin became very thin indeed.

To remove the paraffin I again always used chloroform.
Afterwards the cuts were treated in the usual manner. In
this way I managed to obtain several perfect series, the
yolk remaining in its place without being liable to crumble.

The staining was always done with eosin and hematoxylin,
and invariably after cutting, the outer cuticle of the embryos
being impermeable to the reagents used.

~I

THE DEVELOPMENT OF ADMETUS PUMILIO. 59

LITERATURE.

The object of this paper is to supply, as far as the materials
at my disposal permit, a gap in the literature of the embryology
of the Arachnids. Only three papers have as yet appeared
treating the development of the Pedipalps. Of these one
is on the development of the Thelyphonide, the other two
treating the development of the Phrynidz. Their titles
are—(1) Dr. Strubell (42), ‘Zur Entwicklungsgeschichte
der Pedipalpen’ (Vorlinfige Mittheilung); (2) Sophie
Pereyaslawzewa (37 and 388): (a) ‘Les premiers stades du
développement des Pedipalpes ;’ (b) ‘Les derniers stades du
développement des Pedipalpes ;’ (8) M. Laurie (81), ‘On the
Morphology of the Pedipalpi.” Of these three authors only
the two last mentioned treated of the Phrynide.

When we compare the ages of the embryos I intend to
treat about, we find that my first two are younger than
Pereyaslawzewa’s youngest, my third one is of the same
age as her second, and my fourth corresponds to her third.
My fifth stage is younger than both Pereyaslawzewa’s last
stage and Laurie’s embryos.

PRELIMINARY REMARKS.

I intend to divide the description of the embryos into five
parts, giving each stage in my possession a section to itself.
In each stage I will first consider the general superficial
appearance, and then the details found in the sections.

Stace I.

The earliest stage at my disposal was obtained out of a
batch of well-developed eggs, the rest of which had already
undergone reversion. ‘The egg in question had for some
reason or other, probably pressure, stopped growing very
soon after fertilisation.

It was perfectly spherical before being cut into sections,
and was enclosed in a loose outer membrane. Superficially
no differentiation whatever could be discovered,

598 L. H. GOUGH.

After having cut it I found it to contain about eight
nuclei; these were all situated near each other, not far from
the surface. The rest of the egg consists of yolk, the whole
being surrounded by a delicate inner membrane.

The position of the nuclei in a group near each other, under
the surface, makes it seem probable that the fertilisation of
the egg took place in the middle of the egg, and that the
first cleavages took place there too; the cells resulting from
these cleavages then wandering towards the surface, just as
they do in spiders. The pressure of the two membranes
also points to a resemblance with most of the other orders
of Arachnids, whose eggs are also regularly enclosed in two
membranes.

The yolk itself consists of at least two different kinds of
yolk elements. he first of these tinges very freely with
eosin, and is not influenced by hematoxylin. It consists of
larger and smaller spherules measuring between ‘09 mm.
and ‘02 mm. ; they form the greater bulk of the yolk. The
other yolk element 1s much more irregular in shape, and
helps to fill the spaces between the spherules of the element
first described, to which it adheres, thus becoming crescent-
shaped in cuts, and in reality forming convex-concave lenses.
It also differs in tinction from the first described kind of
yolk, staining both with hematoxylin and eosin to a light
purple.

The size of these yolk-bodies varies between ‘036 mm. and
‘(05 mm. The distribution of the yolk elements does not
seem to follow any definite rule, unless, perhaps, that the
yolk elements last described are more numerous in the
interior of the egg. I shall not again refer to the yolk or
to the membranes, excepting where I have to point out a
change in them.

Stace ITI.

The next stage at my disposal is only a little further
developed than the one just described, It is derived from

THE DEVELOPMENT OF ADMETUS PUMILIO. 599

a batch of eggs which are all at about the same stage of
development.

Surface views of these eggs show that the blastoderm is
already formed. It appears as a long white ribbon covering
about one third of the surface of the egg in leneth and one
sixth in breadth. At one end the white patch seems much
thicker, being more opaque in reflected light. The blasto-
derm is not of the same breadth throughout its length, but
is somewhat narrower in the middle, and rounds out at both
ends. The margins are not distinct or abrupt, but gradually
fade away until at last they become invisible.

It is not possible to distinguish any further details super-
ficially, the material having, unfortunately, been kept too
long in alcohol.

Sections through the egg at this stage show that the
blastoderm already consists of the three germinal layers.

The ectoderm consists of a continuous epithelium or layer
of cells coverimg the surface of the egg. The walls of
these cells are distinctly visible. The cells themselves are
rounded, their nuclei fairly large and somewhat oval ; they
measure ‘Ol mm. in length by ‘O07 mm. in breadth. ‘he
chromatin is not evenly distributed, but forming a number of
small particles, gives the nuclei a spotted appearance. ‘This
layer is sometimes as much as three cells deep.

The mesoderm lies directly under the ectoderm. ‘The cells
forming it are disconnected ; whether they form a continuous
layer in life or not Lam not able to decide, as the cells may
have shrunk in alcohol. Between the cells of the mesoderm
yolk particles are often to be seen, which makes it probable
that the mesoderm never consisted of a continuous layer.
The nuclei of the mesoderm cells are larger than those of
the ectoderm, and also rounder; they measure ‘014 mm. in
diameter.

The cells of the mesoderm seem to be engaged in rapid
reproduction, karyokinetic phenomena being very frequent,
and the centrisomata very often visible.

The entoderm consists of single cells, lying deeply

600 he 4. GOUGH:

embedded in the yolk. Generally only the nuclei of these
cells are to be seen. ‘I'he nuclei are of about the same size
as those forming the ectoderm, measuring ‘01 mm. by ‘012 mm.
Their shape is sometimes convex-concave. Their chromatin
is very evenly distributed, but still presents a shghtly
granular appearance.

These three germ layers correspond in appearance and
position very closely to the same three of spiders, as
described in the handbook of Korschelt and Heider (26).

The yolk only differs from the description given in the
first stage in the parts nearest to the blastoderm. Where
the yolk penetrates the mesoderm, and in all parts adjacent
to the same, the yolk spherules have become very small
indeed; which would lead one to think that the mesoderm
is causing the yolk to break up.

The outer membrane shows no difference from that
described in Stage I. The inner has, however, changed
since the first stage, having risen above the surface of the
ego in the region of the blastoderm, leaving a free space be-
tween itself and the ectoderm.

Srage ITI.

The third stage at my disposal is considerably more
advanced than the last. The reversion of the embryo has
already taken place, and the extremities have begun to grow.
I have only found one egg at this stage; it comes from the
same batch as the eggs described under Stage I and Stage
IV. Its growth had evidently been stopped by the pressure
of the surrounding eggs, which also caused the embryo to
develop in two disconnected halves, only held together
by the yolk. The embryo also showed several other signs
of not having developed normally.

My sections through this egg are, unfortunately, not all
that I could wish, but still they are good enough to make
out several details. For this reason I shall only attempt to
describe a few of the organs.

THE DEVELOPMENT OF ADMETUS PUMILIO. 601

The brain is already divided into its three parts, the com-
missures and the two kinds of nuclear tissue, the difference
of which will be explained in full in Stage IV. These two
tissues are chiefly characterised by the size, tinction, and
density of the nuclei they contain, the smaller being
closely packed, and stained darker. They measure ‘005 mm.,
the larger, lighter stained, and less dense nuclei, on the other
hand, measuring as much as ‘009 mm. to ‘01 mm.

The ventral ganglion-cord has not yet begun to appear.
Those parts of the abdomen and cephalothorax which are
bent upon each other at the time of the reversion of the
embryo are devoid of any cellular integument, although
lined by a thick cuticle. It is in this part that the ventral
ganglion-cord takes its origin in the next stage.

It is interesting to observe that, as in spiders, the legs are
full of yolk, only the tips being free. The yolk seems to get
pushed backwards during the growth of the legs pari passu
with the growth of the muscles and nerves. Neither the
median nor the lateral eyes have yet begun to appear, nor
are the lungs, heart, coxal gland, lateral organs, or alimentary
canal in any degree traceable.

It is in the yolk that the chief peculiarity is to be seen.
The neighbouring eggs have, through pressure, caused a
deep semicircular impression on one side of the embryo, and
at the same time caused it to develop in two halves. ‘The
pressure has also influenced the yolk, causing several of the
yolk-spherules to amalgamate and to form a mass without
any regular shape. The appearance of this fusion necessi-
tates the conclusion that the living yolk particles were of a
fatty nature, and that they were suspended in a watery lymph,
forming a kind of emulsion.

Srace IV.

The next stage which I possess is only a little more
mature than the one just described. The reversion of the
embryo has taken place long ago, the cephalothorax being

602 L. H. GOUGH.

separated from the abdomen. The extremities are much
further advanced than those of the third stage and, except at
their bases, are devoid of yolk. In surface views one sees
the Anlage of the brain as two white spots at the front end
of the cephalothorax. On the sides may be seen, at the
base of the fourth pair of extremities, a horseshoe-shaped
excrescence; this is the lateral organ, first described by
Laurie as occurring in Phrynus; it is always covered by
a shred of some dark substance.

The joints of the legs can already be distinguished, the
embryo at this stage greatly resembling the one given in the
drawing (figs. 1 and la) less the eyes, and with a much
smaller amount of brain.

There are six pairs of extremities to be observed. The
first pair, the cheliceree, now lie before and on both sides of
the mouth, and under the fold which carries the median
eyes; their bases he fairly wide apart. They are double-
jointed, the second joints pointing towards each other, not as
in adults, parallel and pointing downwards ; this becomes
still more conspicuous in the fifth stage.

The pedipalpi, too, begin near the same fold, to which,
indeed, their bases are attached. At their bases they give
off a branch, the endopodite, which forms a kind of mandible.
As yet no thorns are to be found on the ectopodite; they
first begin to appear in the fifth stage. The ectopodite is
much larger than the endopodite.

The third pair of extremities corresponds to the first pair
of walking legs in spiders; here they are developed into
long whip-like legs, with probably only a sensory function.
They lie doubled up upon themselves, surrounding the pedi-
palpi, and with their tips reaching as far as between the
chelicere. At the bases of this and the next posterior
extremity I have been able to find the coxal gland. The
next three pairs of extremities are the walking legs; they
are all of about the same length, and shorter than the whip-
legs. They are likewise doubled upon themselves, and their
free ends are tucked under the preceding legs, only the sixth

THE DEVELOPMENT OF ADMETUS PUMILIO. 608

pair making an exception, the tibia and tarsus of which run
along the under side of the other legs.

The whole embryo is enclosed in a loose outer skin, which
follows the contour of the whole body, extremities included,
without ever actually touching it, except, perhaps, in the
region of the lateral organ, and in parts of the dorsal side of
the abdomen. This cuticle is covered in parts with numerous
wart-like processes, as mentioned by Pereyaslawzewa (87)
and Laurie (31). On surface views nothing further is to be
seen; we will therefore proceed to the description of the
organs as seen in cuts.

(1) The Skins.—The outer cuticle just mentioned is seen
in cuts to consist of several exceedingly thin strata; it is
otherwise perfectly structureless. The embryo itself is
covered by a thin epidermis, which in most places consists
of a single layer of cells. It does not as yet extend over
the entire surface of the embryo, patches on the sides of the
abdomen being still void of this covering.

(2) The Coelom.—I have not been able to follow out the
development of the ccelom. In my younger embryos it was,
of course, not to be found. At this stage it has already
reached a high development. I consider two layers of cells
found under the epidermis, which send folds into the interior
of the yolk, to be the coclom, from their resemblance to the
ccelom of Spiders and Scorpions. The ccelom also enters the
bases of the legs. With the folds the Anlagen of the future
dorso-ventral muscles of the cephalothorax and abdomen also
enter the yolk. Between ccelom and epidermis are also
found the first Anlagen of the segmental muscles, which arise
from the somatic coelom layer.

(8) The Lateral Organ.—The lateral organ of the
Pedipalps has already been described by Laurie (81) and
Strubell (42); it has also been described by Bruce (9) and
Pereyaslawzewa (87).

In the stage which we are now considering the lateral
organ seems to be at the height of its development. It
appears as a horseshoe-shaped excrescence on the base of

604 fi. He GOUGH.

the coxa of the fourth extremity, which it covers on both
sides. It is always covered by shreds of some dark sub-
stance, which is probably secreted by it. Cuts show that
the lateral organ is a more complicated structure than it
seems to me to have been considered as yet.

I am not able to state anything as to the origin of this
structure, as I possess only embryos without the lateral
organ, or with it in later stages of development.

Cuts show that the lateral organ consists of an outer layer
of cells, forming the external wall of the sac; the interior of
the sac contains two cavities filled with yolk, which are
separated from each other by a second internal wall of cells,
running nearly parallel to the external wall. The cells
forming the outer wall of the lateral organ (fig. 5) are deeply
stained with eosin; they protrude into the processes of the
cuticle, at whose bases a dark substance is being deposited.
The nuclei of these cells are oblong, and take a deep stain
from hematoxylin.

The internal walls of cells resemble the external in its
histological elements, with the difference that its cell walls
are not visible, and that the cells have no distinct outline,
so that it is almost impossible to determine their boundaries.
On the side of this wall nearest to the body of the embryo
protoplasmic processes of these cells are seen enveloping
yolk particles.

The yolk in the one cavity does not resemble that in the
other. In the outer cavity, enclosed by the external and
internal wall, this yolk consists of very minute particles,
which appear to have had to pass the internal wall before
having reached the external cavity. ‘The yolk in the internal
cavity still in every respect resembles that filling the abdo-
men, and is in direct communication with it.

At the base of the lateral organ, where it is attached to
the base of the fourth extremity, the epidermis has begun to
grow inwards, forming a partition between the lateral organ
and the leg. ‘his partition has still an opening in the
middle, through which the inner cavity of the lateral organ

THE DEVELOPMENT OF ADMETUS PUMILIO 605

communicates with the yolk of the embryo. Later on this
opening closes, after which the lateral organ becomes func-
tionless and drops off.

The lateral organ of the Thelyphonide has been
described by Strubell (41), and there is no doubt that it is
identical with that of Admetus. In other orders of
Arachnids we only find the lateral organ in Solpugids and
Pseudo-scorpions.

Croneberg (10) described the lateral organ of Galeodes.
This also seems to be very similar to that of Admetus. I
have also been able to find a similar organ in a young adult
Chelifer ; it seems to me to have already become functionless,
and much resembles that described for Admetus in the fifth
stage. The possession of a lateral organ seems to me to
point out a nearer relationship between these three groups.

(4) The Coxal Gland.—There is in Admetus, as in
all Arachnids, a so-called coxal gland. I have found it,
like the authors already cited, Miss Pereyaslawzewa ex-
cepted, at the base of the third extremity; I am also
able to prove its existence, if only as a rudiment, at the
base of the fourth extremity. This last atrophies very
soon. At this stage the coxal gland consists of a per-
fectly straight tube, running from the under surface of the
coxa inwards, and lying immediately on the surface of the
brain itself. It is easily distinguishable from the brain by the
lighter tinction of its nuclei. It is everywhere surrounded by
connective tissue. Whether it is still in open communication
with the ccelom or no I am not able to decide. The occur-
rence of the coxal gland at this stage in the fourth extremity as
well as the third makes it appear probable that it was origi-
nally segmental, as in scorpions, and itis likely that it would
be discoverable in every segment in earlier stages of
growth.

(5) The Nervous System and Sense-organs.—
The nervous system at this stage consists of the cerebral
ganglia and of the ventral ganglion-cord ; this last extends
far into the abdomen.

VOL. 4), PART 4.—NEW SERIES. Sie

606 L. H. GOUGH.

Of sense-organs only the median eyes and the coxal sense-
organs have as yet made their appearance.

The preoral part of the brain consists, as in Stage III,
of two distinct tracts, distinguishable by size, colour, and
number of nuclei. The more anteric® part consists of closely
packed, small and darkly stained unclei, averaging ‘0034
mm. in size. Nowhere in the rest of the brain are the nuclei
so small, so densely packed, or se darkly stainable. We
shall call this part of the brain the accessory brain. In all
other parts of the brain the nuclei :.re less densely packed ;
they average ‘(01 mm. in diameter. The brains of adult
Admetus also show the same division.

These two parts of the brain, following Laurie (81), who
has briefly referred to the difference, are also said to be
distinguishable in Scorpions.

I am ina position to give more notes about the develop-
ment of the ventral ganglion-cord. ‘his last extends on the
one hand from the cerebral ganglia, into which it merges
imperceptibly, and runs along the ventral side of the cephalo-
thorax and ends in the abdomen. In the latter it merges
into the ectoderm so continuously that it is impossible to say ~
where the ventral ganglion-cord begins and the ectoderm of
the abdomen ends.

As all the different stages of the embryonic development
of the brain are to be found in the ventral ganglion-cord of
one embryo, it will be as well to give an account of it, begin-
ning, for the sake of simplicity, with the distal and least
differentiated part of the cord (see fig. 8).

If we examine the epidermis covering the median ventral
line of the abdomen at the part where it is thinnest, we find
that it consists of two layers of cells, surrounded by much
protoplasm. The cell walls are not visible. The nuclei are
oblong and tinge deeply with hematoxylin; the chromatin
not being equally distributed, but collected in small masses,
gives the nuclei a spotted appearance. ‘The micro-nucleus is
conspicuous in several of the nuclei.

‘he nuclei of the two layers le with their axes pomting in

THE DEVELOPMENT OF ADMETUS PUMILIO. 607

different directions, the rounder nuclei of the outer layer,
which measure ‘01 mm. in length and ‘008 mm. in breadth,
being mostly inclined at an angle of about 75° to the surface ;
the more oblong nuclei of the inner layer, measuring ‘01 mm.
in length and ‘004 mm. in breadth, lie parallel to the outer
surface of the embryo. ‘This inner layer is continued for-
wards, and forms a skin covering the inner surface of the
brain; it is of mesodermatic origin.

The first step towards the differentiation of the brain out
of the epidermis just described consists of a rapid thickening
of the outer or germ layer, through the multiplication of its
cells. It soon becomes six to seven cells thick, and the nuclei
have no longer any common direction. Through the thicken-
ing of the outer layer the surface of the embryo has here
become a little arched. After this the outer or germ layer
again becomes thinner, and only consists of one layer of cells.
The inner layer also thickens till it becomes two or three cells
deep; but as it does not take part in the construction of the
brain I shall refer to it again as little as possible.

After proceeding forwards a short distance, probably
corresponding to one abdominal segment, the cell-mass of the
germ layer becomes again reduced to a single layer of cells ;
then it thickens again as before.

When the germ layer has reached its former thickness we
observe several important changes. In the first place a new
element has begun to develop. The cells of the germ-layer
nearest to the surface have changed the direction of the axes
of their nuclei. The nuclei themselves have become much
elongated, measuring now ‘015 mm. in length by ‘004 mm. in
breadth ; they have also become poorer in chromatin. In this
manner a new outer integument, the definite epidermis, has
been formed.

The cells forming the germ-layer were till now equally
distributed throughout the whole of the layer; this now
ceases to be the case. A little space void of nuclei is formed
on the surface of the germ-layer. This is filled with plasm;
just above it there isa gap in the epidermis. For convenience,

608 fie CH. COUGH:

sake I shall call such superficial masses of un-nucleated
protoplasm “surface pits;” this name I give only
with reference to their appearance, and not to
their structure, and I desire that this may be dis-
tinctly borne in mind. The nuclei of the cells of the
germ-layer adjacent to this “surface pit” are beginning
to arrange themselves radially around the centre of the
hollow. Further on, again proceeding forwards, the germ-
layer again presents its former appearance. Still proceeding
towards the cephalothorax, we meet with another new
change. Just before the next thickening containing a
“surface pit” is reached, the mesodermatic inner skin
begins to lift itself from off the nuclei of the germ-layer.
The space between both is now filled with nerve-fibres, the
longitudinal fibres being the first to appear, as in Scorpions
(Brauer [7]). This “‘surface pit” only distinguishes itself
from the last by the presence of a slight depression on
its surface.

After the next “surface pit” has been reached, the
germ-layer loses its continuity and breaks up into a number
of groups of cells, each lying in front of its corresponding
“surface pit.” The space around these groups is filled
with nerve-fibres, forming the commissures; they seem to
run in every direction. A change has also taken place in the
epidermis ; its nuclei are no longer so elongated, having be-
come more rounded ; they now measure ‘01 mm. by ‘005 mm.

The first of the groups of brain-germ cells met with
measures ‘(08 mm. in length by ‘02 mm. in breadth, and lies
‘02 mm. from the surface of the embryo. It is composed of
nuclei lying about three deep and twelve broad, measuring
between ‘009 mm. and ‘01mm. They all lie with their long
axes pointing towards the surface. The nuclei of these
groups are all connected with the centre of the “surface
pit” by radial fibres, which all converge towards it.
The fibres probably belong to the neuroglia, and serve as a
supporting tissue.

The epidermis is discontinued just at this ‘surface

THE DEVELOPMENT OF ADMETUS PUMILIO. 609

pit,’ as it was in the last. The “ surface pit” also
resembles the last described in having a slight depression on
its surface. The cells of the epidermis also take an active
part in the building up of the groups, by forming a funnel-
like sheath around the radial fibres. The nuclei of the cells
forming this sheath measure about ‘018 mm. in length by
‘005 mm. in breadth. Karyokinetic figures are sometimes to
be found in them.

The next such collection of germ-cells is, perhaps, even
more typical. It lies just at that pomt where the abdomen
forms an angle to the cephalothorax (fig. 8). Although
more advanced than the last its dimensions are not larger.

The length and breadth are the same as those of the last
group, the length of the radial fibres being ‘02 mm., as in the
last case; the group, however, now lies ‘037 mm. from the
surface.

The cause of the group lying deeper than the last described
is an ingrowth of the epidermis in the shape of a narrow
tube. Its bore is ‘(007 mm., and its length ‘01 mm. The
walls of this tube are formed by a single layer of cells,
whose nuclei are but lightly stained, and which contain very
large micro-uuclei. One nucleus I measured was ‘01 mm.
long by ‘004 mm., its micro-nucleus, however, actually
measuring ‘004 by ‘003 mm.

The nuclei of the cells forming this tube are all arranged
with their long axes lying at right angles to the axis of the
tube. ‘he tube ends in an expansion, forming a nearly
spherical sac, whose diameter is about ‘01 mm. The walls of
this sac are formed by the ends of the radial supporting
fibres. The sac just described lies within the funnel-shaped
sheath, which encloses the radial supporting fibres. The
next group of cells so much resembles the last described
that it is not necessary to go into any details about it.

Still proceeding forwards, we now meet with a group
which half retains its original form and independence,
and is half connected with and converted to the ventral
ganglion-cord in its definite form. The original germ-group

610 L. H. GOUGH.

lies ‘1 mm. from the surface; its inner surface abuts on the
commissures. On the side of the group whicl is turned to-
wards the abdomen, the sheath covering the radial support-
ing-fibres and the germ-cells forming the group are as before,
but the cells forming the tube have multiplied rapidly, and
now reach as far as the cells forming the group. On the
anterior side all parts of the group have multiplied their cells
to such an extent that they build a solid mass of nuclei,
reaching from the surface to the commissure. In this mass
only the nuclei of the sheath can be distinguished from the
others. ‘The interior of the epidermis-tube is now filled with
supporting fibres.

From the interior surface of the nucleiferous mass are
seen in parts nuclei wandering through the commissural
part of the brain, the ‘‘ Punktsubstanz ;” these afterwards
form a cellular layer on the dorsal side of the commissures,
which had been hitherto only covered by the mesodermatic
inner skin mentioned at the beginning of this description.
That the whole of the post-oral ventral ganglion-cord is
formed in the manner just described is proved by the
presence of numerous remains of tubes, filled with tangled
masses of supporting fibre, in the brain-mass. Hach such
tube denotes the place where there was such a group of cells
during the earlier development of the cord. Later on the
tubes become entirely obliterated. These groups of cells,
with their radial supporting fibres, their sheaths and
tubes, possibly represent what Patten (34) compares to
ommatidia in his description of the development of the
brain of Arthropods. The comparison with sense-organs
is certainly very good. I find, however, more resem-
blance with the “Higelorgane der Seitenlinie,” as
described by Leydig for fishes, and by the Sarasins (40)
for the larvee of Amphibians. Judging by Patten’s figures,
however, I must conclude that he saw a phenomenon in part
different to that which I have just described. In his
drawing one sees that each segment of the brain is composed
of several sense-organs, with larger ones in the middle line.

THE DEVELOPMENT OF ADMETUS PUMILIO. 611

I have only been able to find one in each neuromere. This
probably corresponds to the larger sense-organs in Patten’s
figures. It is to be regretted that Patten did not publish
any drawing of a transverse section of such a sense-organ, or
a description of their development. Kishinouye (22) also
states that the nervous system of Limulus consists at one
period of peculiar cell-groups resembling ommatidia. Ac-
cording to the drawings he gives, he must have cut them at
right angles to the direction in which I have cut tkem in the
embryo described.

According to the description just given, we find that the
brain is derived in all its parts from the ectoderm, as
follows :

(1) Directly, as regards the germ groups.

(ii) Indirectly, through the epidermis, as regards the
sheaths surrounding the radial supporting-fibres
and the tubes.

In more mature portions of the brain the cell-elements
have reached a higher development. We find in the ventral
ganglion-cord the two usual kinds of nerve-cells, lying the
one amidst the other.

(i) Cells with smaller, generally oblong nuclei, measuring
‘01 mm. by ‘007 mm., without visible protoplasm; these form
the chief part of the brain behind the accessory brain. It is
impossible to distinguish these nuclei from those of the
germ-layer or from the neuroglia.

(ii) Cells with larger, generally round nuclei, measuring
‘015 mm. in diameter; these nuclei do not generally stain so
deeply with hematoxylin. They are usually surrounded by
protoplasm, and often have very distinct micro-nuclei. ‘They
are usually to be found near the commissural substance.
They much resemble the ganglion-cells of other animals.

The real meaning of the just-described process of develop-
ment becomes clearer when we compare the development of
the central nervous system of the Arachnids with that of
other classes of animals. Since the Arachnid brain has
reached such a high degree of perfection, we must look for

612 fH. "GOUGH.

analogies chiefly in classes of similar development of brain.
First we may look to the Vertebrates.

In the central nervous system of Vertebrates and Arach-
nids the grey and the white matter seem to have totally
different relative positions ; the white substance surrounding
the grey in Vertebrates, and being almost surrounded by it
in Arachnids. This seeming dissimilarity is soon explained.
Since the central nerve-tube of Vertebrates becomes folded
inwards during its development, it is clear that its innermost
walls, which abut on the canalis centralis, are that part of
the wall which was originally outermost. Thus we would in
both cases have the grey substance outermost, the white
substance lying under it.

His (16) describes the spinal cord of Vertebrates as arising
out of “Zwei Kernfrei Zonen, eine iiusserste und eine innerste,
und eine die kerne enthaltende Mittelzone.” He terms the
outer of the two zones without nuclei the “ Randschleier,”’
the inner the “Séulenschicht.” These three layers or
“Zonen” are all to be met with in Arachnids, though some-
what modified, having in part become discontinuous. The
columnal layer, ‘‘ Saulenschicht,” could be compared to the
“ surface pits;” but it, of course, no longer forms a con-
tinuous layer in Arachnids as it does in Vertebrates. ‘The
“‘ Mittelzone ” can be compared, as regards position and struc-
ture, to the nuclei groups in the developing ventral ganglion-
cord of Admetus, and the “ Randschleier,” which gives rise
to the white matter in Vertebrates, is evidently equivalent to
the commissural substance, or “ Punktsubstanz” of the brain
of Admetus, also as regards position and structure.

The ganglion cells of Admetus are almost always to be
found in the depth of the grey matter, just below the Punkt-
substanz or white matter, asis also the rule with Vertebrates.
In the latter all nerves proceed out of the white substance.
ven this has an analogy in Arachnids, as all the nerves I
met with in my cuts left the ventral ganglion-cord just
above the grey matter, at places where the nucleiferous
covering of the ‘‘ Punktsubstanz” was very thin. These

THE DEVELOPMENT OF ADMETUS PUMILIO. 613

were the chief points of resemblance. The chief points of
difference are—in Vertebrates the origin of the neuromeres
is evidently secondary; nerves issue from the spinal cord only
in those places where the vertebral sheath leaves them spaces
to pass through; in Arachnids the neuromeres appear, in
Admetus at least, to be primary in origin, and at the same
time to be segmental. ‘his does not point to any funda-
mental difference in the whole process ; on the contrary, it is
only to be expected that the segmentation which occurs in
Arthropods should be more constant and more early in
appearance than that of Vertebrates. In these the segmen-
tation is disappearing everywhere, and in all systems of
organs,

According to a theory of von Kupffer and others, the
sense-organs of the Vertebrates are derived froma hypothetic
primitive form of sense-organs, the so-called placods. The
placods remain in their least changed form as the “ Hiigel-
organe der Seitenlinie” of Leydig, and as papille of the
organ of taste. ‘These sense-organs much resemble those
just described as building part of the Anlage of the ventral
ganglion-cord of Admetus. A similar feature is that the
epidermis of the branchial region of Vertebrates gives rise
to the epibranchial ganglia,—that is to say, in Vertebrates, as
in Arachnids, primitive sense-organs take part in the build-
ing up of the ganglia.

The Sarasins (41) have also pointed out the resemblance of
small sense-organs of Helix Waltoni to the “ Hiigel-
organe”’ of Ichthophis. Here the sense-organs also seem
to take part in the building up of the ganglia. <A peculiar
coincidence is to be found when we compare the brain of
Helix Waltoni with that of Admetus. The Sarasins men-
tion that the most anterior portion of the brain of Helix
Waltoni consists of very small, darkly stained, and closely
packed nuclei, containing no Punktsubstanz and no ganglion-
cells; this portion of the brain of Helix Waltoni they
eall the accessory brain. The process of the development of
the accessory brain of Helix Waltoni resembles the same

614. L. H. GOUGH.

process in Admetus very closely and conspicuously. Accord-
ing to the Sarasins the accessory brain of Helix Waltoni
takes its origin out of the so-called cerebral tubes; by
analogy with Scorpions, it is perhaps possible to prove the
same for Admetus, though I can only find it by comparing
the papers of Brauer (7) and Laurie (29). Laurie says that
the median eyes of Scorpions arise out of an invagination,
which takes place just in front of the Anlage of the brain ;
he describes the cells forming it as resembling brain-cells,
but having smaller, darker stained and more densely packed
nuclei; in another paper (81) he states that the brain
of Phrynus resembles that of Scorpions inasmuch as
in both cases we find a division of the brain in a
part containing smaller, darker, and more densely packed
nuclei, and a part with lighter, larger, and less dense nuclei.
According to Brauer the median eyes arise on a small pro-
jection, which lies beyond the invagination mentioned by
Laurie, this invagination becoming deep, and giving rise to
an accessory part of the brain. Brauer, however, omits to
mention whether he found the nuclei composing this part to
differ from those forming other parts of the brain.

According to Kleinenberg (25) part of the nervous system
of Lopadorhynchus is also formed out of primitive sense-
organs, which afterwards become converted entirely into
ganglion-masses.

We are thus led to suppose that the central nervous system
of Arthropods to some extent corresponds in origin and
structure to that of Vertebrates; namely, as regards the
origin of the three layers (‘ Saulenschicht,” “ Mittelzone,”’
and “‘Randschleier” ‘surface pit,’ germ-group, and com-
missural substance), and as regards the conversion of sense-
organs into ganglia. ‘To some extent it corresponds to that
of Molluses, namely, as regards the origin of ganglia out of
sense-organs, and perhaps in possession of an accessory
brain. Lastly to that of Worms, namely, as regards the
origin of ganglia out of sense-organs.

The Median Hyes.—As yet only the median eyes have

THE DEVELOPMENT OF ADMETUS PUMILIO. 615

begun to make their appearance. Being unpigmented they are
not visible on surface views. The lateral eyes are first met
with in more mature stages. Just above the chelicera, on the
front of the cephalothorax, we meet on each side with a lappet-
like execrescence, the bases of which have already begun to
unite with each other where they meet in the median plane ;
thus they form a kind of mask in front of the embryo,
covering its mouth. The space beneath this mask remains
hollow till much later.

These folds are the Anlagen of the median eyes ; they con-
sist of two simple thickenings of the ectoderm, which have
probably been inverted in the way described by Brauer for
Scorpions.

The front wall of the eyes still only consists of a single row
of cells; in the posterior wall the cells have begun to
multiply. The space between the anterior and _ posterior
walls of the eyes contains no nuclei, being only filled with
cell-walls and plasm.

The eyes measure in length ‘1 mm.; they are ‘04 mm.
thick.

Besides the eyes there is only one other set of sense-
organs to be found. These are the segmental sense-organs
first described by Patten (34) as occurring at the base of the
legs of Spiders; they were also described by Brauer (7),
who found them in Scorpions. The description and drawings
given by Brauer accord with those which I could give for
Admetus,

(6) The Alimentary Canal.—The alimentary canal is
at this stage still very incomplete, only the most anterior part
existing as yet. It consists of a simple tube which just
pierces the brain. Its outer anterior end does not pro-
ject beyond the brain, the posterior end doing so for
about one third of its length. From the point where it
leaves the brain it is slightly bent in a ventral direction.
The cells forming the walls of the alimentary canal form
a thick layer at both ends of the tube. No muscles are
as yet observed in connection with the alimentary canal,

616 L. H. GOUGH.

In transverse sections it otherwise presents a similar appear-
ance to that of more advanced stages. ’

The cavity of the alimentary canal is Y-shaped; the cells
composing its walls are high and cylindrical. The nuclei
chiefly le at the end of the cell nearest to the cavity. A thin
cuticle has been secreted by these cells.

The whole is surrounded by a thin layer of mesoderm cells,
which form a skin round the tube. ‘he nuclei of these cells
have begun to elongate, and will probably form the ring-
shaped muscle found later on in this part.

(7) The Heart.—The circulatory organs are at present
represented by the heart alone. The origin of the heart
seems to me to be the same as in Scorpions (Brauer [7])
and other Arachnids. It is evidently formed by the ccelom
on both sides of the embryo meeting, leaving a space between
the walls on either side, which, although surrounded by
ccelomatic walls, does not belong to the ccelom itself. The
heart causes a slight ridge on the surface of the embryo.

(8) The Lungs.—At this stage the lung-books are just
beginning to make their appearance. They belong to the
first and second abdominal extremities. Laurie (31), however,
has located them on the first and third. They differ neither
in development nor in appearance from the lungs of other
Arachnids.

(9) The Muscles.—Typical muscles are as yet nowhere to
be found. In the extremities, and also in the abdomen, we
find the cells which give rise to the future muscles. These
are distinguishable by their long, granular nuclei. Though
already spindle-shaped, they do not as yet stain with eosin,
as the muscles in more advanced embryos always do.

(10) The Genital Organs.—The genital organs have not
as yet begun to make their appearance.

Srace V.

The two illustrations (figs. 1 and 1 a) of the embryo at this
stage will give a very good idea of the general superficial

THE DEVELOPMENT OF ADMETUS PUMILIO. 617

appearance of these embryos. They appear in almost every
respect similar to the embryos of the fourth stage, and the
description of those embryos only requires a few additions
to be perfect for the embryo of the fifth stage.

In the first place we must remark the median eyes. They
are now very deeply pigmented, except in a narrow line
which separated the two eyes.

The lateral eyes are now also visible; they resemble Y-
shaped marks on both sides of the median eyes. This is
caused by the pigment being deposited between the single
ocelli, which are three in number. The margin of the fold
carrying the median eyes is also more distinctly visible than
in the fourth stage.

1. The Skins.—No important change can be observed
in the skins since Stage IV.

2. The Ccelom.—The ccelom has undergone many changes
since Stage [V. It has, for instance, given rise to the
muscles, of which at least two sets are to be distinguished.
A dorso-ventral muscular system has followed the foldings of
the ccelom, and now consists of a paired muscle in each
segment, running from the carapace to the ventral side,
except perhaps in the first two segments of the cephalo-
thorax. The other set consists of the intersegmental muscles.
These run parallel to the surface, from segment to segment.
The walls of the coelom also take part in building up the
mid-gut. On the ventral side of the abdomen we also find
the genital cells lying inside the ccelom, as will be described
further on.

3. The Lateral Organ.—The lateral organ is still to be
found, but it seems to me to have ceased to have any function.
Superficially it seems to be in the same state as it was in the
fourth stage; itis only on cuts that the difference is remarked.

Sections show that the lumen of the lateral organ is no
longer directly connected with that of the rest of the body,
and with the base of the fourth extremity in particular.

The cells which at a less mature stage formed the walls of
the lateral organ have died off, and it is only here and

618 le A: (GOUGH:

there that traces of them are found. With the drying up of
the wall cells their plasma has been withdrawn from the
interior of the wart-like processes of the cuticle.

In the fourth stage the outer cavity of the lateral organ
was filled with very small yolk particles; these have now
conglomerated and form a solid mass, staining deep red with
eosin. The partition that had begun to form between the
base of the lateral organ and the base of the fourth ex-
tremity.is now complete.

A remarkable fact is that a substance resembling the
filling of the lateral organ is to be found in the adjacent
parts of the base of the fourth extremity.

The shreds of dark substance covering the lateral organ in
the fourth stage have become much thicker (they have been
removed in the drawing [fig. 1a], so as not to hide the
lateral organ). This makes it appear more probable that
they have been secreted by the lateral organ.

4, The Coxal Gland.—At this stage the coxal gland is
at the height of its development. Only the gland belonging
to the third extremity is still to be found. It occupies all
the space between the cerebral ganglia and the ventral gan-
glion-cord that is not occupied by the muscle-stomach.
Bulging over the sides of the ventral ganglion-cord, it now
lies in at least four segments, namely, the third, fourth, fifth,
and sixth, its opening lying at the base of the third extremity.

There are two different parts of the gland which can be
distinguished, namely, the mouth end and the gland proper.
The mouth end is composed of cells in every way resembling
those that form the epidermis. Its nuclei are oblong, and
stain deeply with hematoxylin. In this part I observed no
trace of cell-walls. The second part builds up the chief
bulk of the gland itself. It consists of an unbranched tube,
much convolved at its inner end. The tube runs nearly
straight from the surface up to near the brain; next it bends
backwards and runs in almost a straight line till it reaches
the sixth segment; then it begins to twist and turn so
much that it is impossible to follow it further. As I have

THE DEVELOPMENT OF ADMETUS PUMILIO. 619

nowhere been able to observe a branching of the tubes, I
suppose that it must be simple in its whole length, as in other
Arachnids,.

The cavity of the gland is about ‘01 mm. wide. The cells
composing the tube have very distinct cell-walls, the cell-
plasm being very clear and staining very slightly. The
nuclei are perfectly round, and do not stain very deeply ;
their diameter is*‘007 mm. They lie almost always on the
side of the cells nearest to the lumen of the gland, leaving,
however, a space between themselves and the cell-walls. In
this space the plasm seems at its thickest. he length of
the cells, measured from outside the gland to the cavity of
the gland, is ‘02 mm.

The tube forming the gland itself is enclosed in a sheath
of flat cells, of mesodermatic origin, as in other Arachnids.
The space between the windings of the tube is filled out by
connective tissue. he whole gland is likewise covered by
an outer skin, the elements of which somewhat resemble
those forming the sheath of the tabule.

5. The Central Nervous System.—'The nervous system
now consists of the cerebral ganglion, the ventral gan-
elion-cord, and the nerves.

When compared with the fourth stage, the ventral gan-
olion-cord appears much contracted. In the fourth stage it
was continued far into the abdomen, now we only meet with it
in the cephalothorax.

A great advance on the fourth stage is also to be observed
in the development of the ganglia. In Stage IV 1b was im-
possible to distinguish them from each other; now they are
very distinct. The whole brain has become very similar to
that of the adult.

The outer form is to be seen in the illustration (fig. 2),
made after a model constructed by me from a section-series.
I will begin by a description of the superficial appearance of
the brain.

A glance at the illustration will show that the brain con-
sists of two parts, a smaller dorsal and a larger ventral. It

620 L. H. GOUGH.

is pierced by the alimentary canal in the region where the
two parts are joined to each other.

Five pairs of nerves issue from the ventral or post-oral
part of the brain; these belong to the pedipalpi and the legs.
The nerves and ganglia belonging to the chelicerz are
preoral.

No other nerves as yet leave the dorsal portion of the
brain, though two swellings on each side of it, just behind
the ganglion of the chelicere, denote the optic ganglia.
They are not yet connected with the eyes by nerves, these
being very late to appear. The dorsal portion of the central
nervous system contains the accessory brain, the four
cerebral ganglia, and part of the ventral ganglion-cord, con-
taining the ganglia of the chelicere.

The ventral ganglion-cord consists of eighteen ganglia,
six belonging to the extremities. The other twelve are very
small; although lying in the cephalothorax they really
belong to the abdomen.

The elements forming the brain and ventral ganglion-cord
are histologically mostly the same as those described in the
fourth stage. We now remark, however, small masses of
darker stained fibres in the commissural parts of the central
nervous system, the origin and structure of which it is diffi-
cult to understand. They are not to be found in the adult
brain.

6. The Eyes.—The median eyes have become much
further developed since the fourth stage. This can best
be described in connection with the drawing (fig. 6). The
inedian eyes are still situated on a fold in front of the
mouth. The mouth opens into the cavity (c.) formed between
the fold carrying the eyes and the cephalothorax. We can
distinguish between three distinct layers of the fold—the
corneal, the retinal, and the subretinal layer. This last is
separated from the retinal by a fissure.

The stratum corneum (co.) consists of a layer of nuclei, two
or three deep. In front it has begun to deposit chitin, the
future lens (.).

THE DEVELOPMENT OF ADMELUS PUMILIO. 621

Under the cornea lies the retina (r.). It is deeply pigmented
in its anterior half. One can distinguish thicker and thinner
lines of pigment (p.) in front. The retine of the two eyes
are very distinctly separated from each other by a light un-
pigmented line (s.).

The post-retinal layer (pr.) resembles the epidermis else-
where.

I am not in a position to say how the eye has developed
out of the Anlage described in the fourth stage.

There are three lateral eyes on each side of the cephalo-
thorax. They lie in groups in a line with the median
eyes. The lateral eyes originate out of simple ectoderm
thickenings, as Laurie stated, as will be seen in the
drawing (fig. 7). Pigment is as yet only deposited in the
spaces between the single eyes. The lateral eyes of the adult
also consist of three facets.

7. The Alimentary Canal.—Since the fourth stage
the alimentary canal has made very rapid progress. It now
forms a nearly complete tube, only the foremost part of the
midgut remaining absent. ‘The alimentary canal now con-
sists of the following parts:—mouth, pharynx, cesophagus,
muscle-stomach, midgut, and rectum. ‘The mouth is situated
on a small protuberance, which projects into the subocular
cavity, and lies between the bases of the pedipalpi. Just behind
the mouth we find the pharynx; that has an I-shaped cavity,
the walls of which are lined with a thin chitinous membrane.
The cavity measures *1 mm. by ‘01 mm,

The cells forming the walls of the pharynx have very dis-
tinct cell-walls; the nuclei are small, oblong, and stain very
deeply with hematoxylin; they measure ‘007 mm. by ‘004
mm. Besides the lateral and dorsal muscles attached to this
part of the alimentary canal, described by Laurie, I must
draw attention to another set of muscles connected with the
pharynx. This set is a ring-muscle, which runs round the
pharynx ; it evidently acts as the antagonist of the lateral and
dorsal muscles, and serves to close the cavity when it has
been distracted by the other muscles.

VOL. 45, PART 4,—NEW SERIES. oD

622 iy H. COUGH,

The pharynx goes over into the cesophagus, when it enters
the brain-mass. Its hollow is now Y-shaped. The cells
forming its walls are similar to those of the pharynx. A
distinct cuticle is still to be seen. It is no longer enclosed
in a ring-muscle, but is only covered with connective tissue.
After having passed through the brain, the alimentary canal
again changes its character and becomes muscular once
more. Its cavity is much wider here, and is X-shaped in
transversal sections. The cells forming this part much
resemble those of the parts already described.

The musculature of the muscle-stomach can be divided into
two systems, similar to those of the pharynx. The first of
these consists of radial muscles, the longest of which rans
dorsally towards the carapace; the two others are much
shorter, and insert laterally in a cartilage, which also serves
to support the coxal gland.

The other system again consists of a ring-muscle, which is
much stronger than that in the region of the mouth.

As yet the muscles are all smooth, in the adult they are
striated. The dorsal and lateral muscles pierce the ring-
muscle and insert in the walls of the stomach itself.

The alimentary canal has a break in its continuity, just
behind the muscle-stomach; we next meet with it in the
abdomen.

Through in-foldings of the ccelom the yolk is divided into
several distinct masses. ‘The walls of the midgut are in part
formed by the cceelom. At its anterior end it is wide, open,
and funnel-shaped, and it tapers towards its posterior end. |
It is everywhere filled with yolk. On the interior side of the
funnel formed by the ccelom the entoderm cells have built up
an epithelium. ‘This epithelium seems to be separated by a
membrane from the walls composing the ccelom.

The cells forming the lining of the midgut have
distinct cell-walls; their nuclei stain lightly with hema-
toxylin; they are perfectly round, and measure ‘01 mm.
in diameter. A micro-nucleus is often to be found in
them.

THE DEVELOPMENT OF ADMETUS PUMILIO. 623

The end of the midgut is not in open communication with
the rectum, being closed by a plug of entoderm cells.

The rectum is formed by an invagination of the ectoderm.
At its exterior end its cells quite resemble those forming the
epidermis ; at its interior (proximal) end the cells are vacuo-
larised. The rectum, as also the most anterior parts of the
dilators, canal, is supplied with powerful muscles ; one set, the
alimentary, runs from its walls to the skin of the abdomen ; the
contractors being ring-shaped, as in the pharynx and muscle-
stomach.

As in Scorpions the Malpighian tubes are without doubt of
entodermatic origin, as they enter the alimentary canal near
the posterior end of the midgut. In this stage they are
already well developed; they are very long and run parallel
and juxta-apposed to the alimentary canal.

8. The Heart.—The heart of the Pedipalpi has been best
described by Pereyaslawzewa as yet.

The heart at this stage consists of a long tube, lying
dorsally, immediately beneath the skin, in the median plane
of the abdomen. Its walls are thick, but do not contain
many muscle elements.

At each segment the heart widens, and seems to me to
give off a pair of small arteries. A large artery leaves the
heart at its anterior end, this runs into the cephalothorax ;
following the outer surface of the embryo and reaching the
cerebral ganglion it suddenly bends downwards; soon after-
wards it divides into two branches, which run forwards on
both sides of the muscle-stomach till they reach the central
nervous system, when they terminate abruptly.

My embryos not being so advanced as Pereyaslawzewa’s
(37), I am not in a position to state anything about the other
arteries and veins which she has seen; at the same time I
consider her statement that the heart terminates ‘‘par une
artére post-abdominale” as at all events not perfectly
correct, since, as is hardly necessary to state, the Phryniscidee
have no post-abdomen, either as embryos or as adults.

Inside the heart the blood-cells are to be seen. These are

624. L. H. GOUGH.

large cells, staining red with eosin. They are usually
nearly spherical, measuring up to ‘036 mm. in diameter, and
seem to be surrounded by a thin membrane. The nuclei of
the blood-cells always lie on the surface of the cell, and
are oblong. Hach cell contains many nuclei, of which
one is always remarkable as being several times larger than
the others; this one measures ‘007 mm., and is often in a
state of mitosis. These cells are possibly the same as the
fat-cells seen by Kishenonye (28) in spiders, but are here
confined to the interior of the heart. Besides these larger
blood-cells, smaller ones are also to be seen in the cavity of
the heart. These have only one nucleus, and are poor in
plasma; they resemble those found in the heart in the fourth
stage. ‘There the larger, spherical blood-cells are missing.

9. The Lungs.—It is not necessary to follow the
development of the lungs, as it follows the same type as
most other Arachnids.

10. The Genital Organs.—I am not in a position to
state anything new about the genital ducts, Pereyaslawzewa
(38) having given a fuller account of them than can be made
out of my embryos. I have only been able to find the
genital germ-cells. These are the largest cells in the whole
embryo. ‘They are situated in the remains of the ccelom, on
the ventral side of the abdomen, in the region of the second,
third, fourth, fifth, sixth, and seventh abdominal segments.

The shape of the genital-cells is always more or less oval
(fig. 4), the nucleus resembling the cell in shape. The
genital cells are surrounded by a thin cell-wall; the plasma
stains red with eosin, and is granular in appearance. It
often contains as many as three vacuoles; one at all events
seems never to be wanting.

The nucleus stains only a little darker than the plasma,
and is little influenced by hematoxylin. The chromatophores
are distinctly visible in the shape of bands of darker stained
substance. ‘I'he nucleus also seems to be separated from the
cell-plasm by a thin membrane.

The micro-nucleus is very conspicuous; it is always per-

THE DEVELOPMENT OF ADMETUS PUMILIO. 625

fectly round, and stains very deeply with hematoxylin. It
is almost always surrounded by a clear space containing
little stainable matter. It is always found inside the
nucleus.

The average measurements of the genital cells are :—
Length of cell 07 mm., breadth of cell 03 mm., length of
nucleus ‘(03 mm., breadth of nucleus ‘02 mm., diameter of
micro-nucleus ‘004 mm.

11. The Muscles.—The two chief systems of muscles
have already been referred to under the heading Ccelom.
They are at this stage very well developed, and are composed
of smooth fibres only.

Pereyaslawzewa (88) declares that the muscles of the
cephalothorax are all striated, those of the abdomen being
all smooth. It is hard to understand why this should be
the case, and I think that the statement requires further
confirmation, especially as all the muscles are derived from
the same segmental sources, both in cephalothorax and
abdomen.

12. The Yolk.—It is now only necessary to state that
pari passu with the development of the central nervous
system, and with the contraction and withdrawal of the
ventral ganglion-cord into the cephalothorax, the chief bulk
of the yolk has been forced back into the abdomen, as is
the case with all Arachnids.

CoNCLUSIONS.

On the whole the development of the Pedipalps follows
the types prevalent among other Arachnids, sometimes
leaning more towards the one, sometimes more towards the
other class. It resembles—

(1) That of Spiders :

a. In the first cleavages (probably).

b. In the egg-envelopes.

c. In the general build of the blastoderm.

d. In the development outside the mother animal.

626 L. H. GOUGH.

e. In the development of the lungs, heart, alimentary
canal, and coxal gland.
(2) That of Solpugids and Pseudo-scorpions :
In the development of the lateral organ.
(3) That of Scorpions :
a. In the development of the central nervous system.
b. In the presence of an accessory brain.
c. In the development of the median and lateral eyes.
d. In the development of the lungs, heart, coxal
gland, and parts of the alimentary canal and
Malpighian tubes.
The mode of development of several of the organs is the
same in Spiders as in Scorpions,—for example, heart, lungs,
etc.

BIBLIOGRAPHY.

1. ADENSAMER.—“ Die Coxaldriise von Thelyphonus caudatus,”
‘ Zoolog. Anzeiger,’ xviii.

2. Bernarp.—‘ Terminal Organ of the Pedipalpi of Galeodes, and the
Discovery of an Homologous Organ on the Pedipalp of Phrynus,”
‘Ann, Nat. Hist.,’ xii.

3. BernarD.—‘‘ On the Spinning-glands of Phrynus, with an account of
the so-called Penis, and of the Morphology of the Operculum,” ‘J.
Linn. Soc.,’ xxv. -

4, Bernarp.—‘* The Comparative Morphology of the Galeodide,” ‘J. Linn.
Soc., XXv.

5. Bertkau, Zu Leprpinsky.—‘ Die Entwicklung der Coxaldriise von
Phalangium,” ‘ Zool. Anz.,’ 1892.

6. Braver. —“ Beitrage zur Kenntniss der Entwicklungsgeschichte des
Scorpions. I,” ‘ Zeitschr. f. wiss. Zool.,’ 1894.

7. Braver.— Beitrage zur Kenntniss der Kntwicklungsgeschichte des
Scorpions. II,” ‘ Zeitschr. f. wiss. Zool.,’ 1896.

8. Bruce.— Observations on the Embryology of Spiders,” ‘ Am, Nat.,’ xx.

9. Brucr.—‘ Nervous System of Insects and Spiders, with Remarks on
Phrynus,” ‘ Johns Hopkins Univ. C.’

10. CronEBERG.—“ Ueber ein Entwicklungsstadium des Galeodes,” ‘ Zool,
IANIZe5 Xs

11.
12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.
24.

25.

26.

27.

28.

29.

30.

31.

THE DEVELOPMENT OF ADMETUS PUMILIO. 627

Faussex.— Zur Embryologie von Phalangium,” ‘ Zool. Anz.,’ 1891.
Frorier.— Uber ein Ganglion des Hypoglossus und Wirbelanlagen in
der Occipitalregion,” ‘ Arch. f. Nat. und Entwicklungsgesch.,’ 1882.
GRIFFITHS.—“‘ On the Blood of Invertebrates: Arachnids,” ‘Proc. R.

Soc. Kdin.,’ xix.

HentscHELy.— Zur geographischen Verbreitung der Thelyphonen,”
‘Zool. Anz.,’ xxii.

HentscuEett.— Beitrage zur Kenntniss der Spinnenangen,” ‘ Zool.
Jahrb. f. Anat.,” xii.

His.—* Die Neuroblasten und deren Entstehung im embryonalen Mark,”
‘Arch. f. Anat, und Entwickelung.,’ 1899.

Jaworowsky.—“ Uber die Extremitaten, deren Driisen und Kopf-
segmentierung bei Trochosa singoriensis,” ‘Zool. Anz.,’ 1892.
Jaworowsky.—Uber die Extremititen bei den Embryonen von Arach-

niden und Insecten,” ‘ Zool. Anz.,’ 1891.

Jaworowsky.— Die Entwicklung der sogenannten Lungen bei den
Arachniden und speziell bei Trochosa singoriensis, nebst Anhang
iiber Crustaceen Kiemen,” ‘ Ztschr. f. wiss. Zool.,’ Ixviii.

Kinestey.—‘ The Embryology of Limulus,” ‘J. Morph.,’ vii.

Kisuinovye.—‘ On the Lateral Eyes of Spiders,” ‘Journ. Coll. Sci.,’
1892.

Kisutnovye.—‘‘ On the Development of Limulus longispina,” ‘J.
Coll. Sci.,’ 1892.

Kisninovyr.—*“ On the Development of Araneina,” ‘J. Coll. Sci.,’ 1891.

Kisurnovye.—‘‘ Note on the Ceelomic Cavity of the Spider,” ‘J. Coll.
Sci.,’ vi.

Kierenserc.—* Die Entwicklung von Lopadorhynchus,” ‘ Ztschr. f.
wiss. Z.,’ 1886.

Korscuett und Hemer.—< Entwicklungsgeschichte der Wirbellosen-
hiere:2

Lanxester, E. Ray.—‘ On Skeletotropic Tissues and Coxal Gland of
Limulus, Scorpio, and Mygale,” ‘Q. Journ. Micr. Sci.,’ 1891.

Lanxester, E, Ray.—‘ Limulus an Arachnid,” ‘Q. Journ. Mier. Sci.,’
1881.

Laurie.—* The Embryology of a Scorpion,” ‘Q. Journ. Micr. Sci.,’
1891.

Laurie.—On the Development of the Lung-books in Scorpio
fulvipes,” ‘ Zool. Anz.,’ 1892.

Lavrigs.—‘ On the Morphology of the Pedipalpi,” ‘J. Linn. Soc.,’ xxv.

628 L. H. GOUGH.

32.

33.

34.
35.

36.

37.

38.

39.

40.

41.

42.

43.

44.
45.

46.

47.

48.

a

Lepipinsky.—‘‘ Die Entwicklung der Coxaldriise bei Phalangium,’
‘ Zool. Anz.,’ 1892.

Locy.— Observations on the Development of Agalena nevia,”
‘Ann. Nat.,’ xx.

Parrrn.— Segmental Sense-organs of Arthropods,” ‘J. Morph.,’ 1889.

Parren.—“* On the Origin of Vertebrates from Arachnids,” ‘Q. J.
Micr. Sci.,’ 1891.

Patren.—“ On the Morphology and Physiology of the Brain and Sense-
organs of Limulus,” ‘Q. Journ. Micr. Sci.,’ 1894.

PEREYASLAWZEWA: “ Les premiers stades du développement des Pedi-
palpes,” ‘C. R. Ac. Sci.,’ exxv.

PEREYASLAWZEWA: ‘Les derniers stades du développement des Pedi-
palpes,” ‘C. R. Ac. Sci,’ exxv.

Purcenu.—‘ Uber den Bau der Phalangiden Augen.,” ‘ Ztschr. f. wiss.
Z.,’ lvili.

Sarasin.—“ Zur Entwicklungsgeschichte und Anatomie der ceylones-
ischen Blindwithle Lehthyophis glutinosa,” ‘ Erg. wiss. Forsch. in
Ceylon.’

Sarasin.—“ Die Entwicklung von Helix Waltoni,” ‘ Erg, wiss. Forsch.
in Ceylon.’

SrRUuBELL.—‘ Zur Entwicklungsgeschichte der Pedipalpen,”’ ‘Zool.
Anz.,’ 1892.

Tarnanr.— Die Genitalorgane der Thelyphonus,” ‘ Biolog. Centralbl.,’
1889.

Tarnani.— Zur Morphologie der Thelyphonus,” ‘ Zool. Anz.,’ xix.

‘(u1eLe.—“ Uber Sinnesorgane der Seitenlinie und das Nervensystem
der Mollusken,”’ ‘ Ztschr. f. wiss. Zool.,’ xlix.

Waener.—‘ Zur Entwicklungsgeschichte des Ixodes,” ‘Zool. Anz.,’
1892.

Wartase.—* On the Morphology of the Compound Eyes of Arthropods,”
*Q. J. Mier. Sci.,’ 1890.

Wartase.—* Neuroblasts in the Arthropod Embryo,” ‘J. Morph.,’ iv.

THE DEVELOPMENT OF ADMETUS PUMILIO. 629

EXPLANATION OF PLATES 32 & 33,

Illustrating Mr. L. H. Gough’s paper on “The Development
of Admetus pumilio: a Contribution to the Em-
bryology of the Pedipalps.”

[N.B.—It must be borne in mind that the descriptions in the text are made
from whole series of sections, but that the drawings were each made from
one cut only. |

PLATE 382.

Fies. 1 and 1 a.—Embryo of Stage V ; explanations in the text. Magnified
12 times.

Fie. 2.—Reconstruction of brain of embryo of Stage V. The parallel lines
are to demonstrate the position of the cavity caused by the alimentary canal,
otherwise as in text. Magnified 18 times.

Fic. 3.—Sagittal section of lateral organ of embryo of Stage 1V. Magnified
165 times. c. Cuticle. d. Deposit. e. w. External wall, 7. 2. Internal
wall. p. Partition. o.c. Outer cavity. 7. c. Inner cavity. y. Yolk.

Fie. 4.—Sagittal section of part of the abdomen of embryo of Stage V,
with genital cells. Magnified 220 times. v. Vacuole of germ-cell. mz.
Micronucleus of germ-cell. 2. Nucleus of germ-cell. c. p. Cell-plasm_ of
germ-cell. @.e. Fifth abdominal extremity. c. Remains of ecelom.

Fic. 5.—Sagittal section of a portion of the heart of embryo of Stage V.
Magnified 330 times. w. Walls of heart. d.c. Blood-cells. y. Yolk. e.
Kpidermis.

Fig. 6.—Sagittal section through cephalothorax of embryo of Stage V.
Magnified 60 times. c.g. Coxal gland. 0. ¢. g. Opening of coxal gland.
d. v. m. Dorso-ventral muscles. a. b. Accessory brain. 6. Brain.

Fic. 7.—Horizontal section of rectum and mid-gut of embryo of Stage V.
Magnified 220 times. e. Hpidermis. ve. Rectum. ex. Entoderm cells
forming mid-gut. m. Muscles forming sphincter and dilators. y. Yolk.

PLATE 33.

Fie. 8.—Sagittal section through embryo of Stage LV, showing Anlage of
part of ventral ganglion cord, and development of ganglia out of small sense-
organ-like structures. Magnified 280 times. s. Surface pit. ¢u. Tube.

630 L. H. GOUGH.

s. f. Supporting fibre. g. Ganglion cell. c.o.m. Punktsubstanz. 1—9.
Sense-organ-like structures in different stages of development.

Fires. 9 and 9 a.—Sagittal sections through lateral eyes of embryo of Stage
V. Magnified 420 times. 1, 2, 3. First, second, and third eye.

Fie, 10.—Horizontal section of median eye of embryo of Stage V. Mag-
nified 280 times. @. Lens. co. Stratum corneum. p. Pigment. 2. Retina.
F. Fissure. PH. Post-retinal layer. c.c. Subocular cavity.

AS Huth,Lith® London_

Wiis,

|
|

JW

SO

Gane Bat
@ @ aS

Boog
P Ose PaO

ee

AS. Hoth, Lith’ London

ON THE TEETH OF PETROMYZON AND MYXINE. 651

On the Teeth of Petromyzon and Myxine,
By
Ernest Warren, D.Sc.,

Assistant Professor of Zoology, University College, London,

With Plate 34.

THosE observers who regard the Cyclostomata as a degene-
rate offshoot from a gnathostome ancestor would naturally
desire to look upon the horny teeth as degenerate structures,
and they would endeavour to find in them remains of parts
homologous to those of an ordinary vertebrate tooth.

In 1889 Dr. Beard! published an investigation on the
teeth of the Marsipobranchu. ‘The teeth of Bdellostoma,
Myxine, and several species of Petromyzon were examined,
and the results arrived at seemed conclusive as to the
degenerate nature of the teeth. In the young tooth of the
Hag-fishes a more or less complete enamel epithelium, with
perhaps a trace of enamel, was identified ; also semi-calcified
odontoblasts were described, forming a conical mass beneath
the enamel epithelium. More recent observers, however,
have thrown doubts on these interpretations. In 1894 Prof.
Howes? expressed his opinion that the “odontoblasts ”
exhibited uo calcification, and in the same year Ayers ® stated
that he was unable to find a trace of enamel or dentine.

1 «The Nature of the Teeth of the Marsipobranch Fishes,’’ ‘ Zoologische
Jahrbiicher,’ iii, 1889. A bibliography is there given.

2 *Nature,’ Nov., 1894, Review of the Wood’s Holl Lectures.

3 Biological Lectures at Wood’s Holl, 1894.

§32 ERNEST WARREN.

In making a series of preparations illustrating a course of
lectures on odontology I was led to the Cyclostomes, and I
have cut sections through the heads of young specimens of
Myxine glutinosa (L.) (about five inches long), and of
Petromyzon marinus (L.). These sections have not
confirmed the views of Dr. Beard, and there can be no doubt
that the cone of “odontoblasts” is purely epidermal in
origin, and is, in fact, a successional tooth developing be-
neath the functional tooth.

The heads were stained with borax carmine, and treated
with acid alcohol. They were then slowly impregnated with
chloroform and paraffin, and ultimately were passed into
pure paraffin (58°). In the case of Myxine it was necessary
to paint the surface of the block with a solution of celloidin
and gum mastic before cutting every section. Good sections
were thus obtained, and they were counter-stained on the
slide with picro-nigrosin. With this stain anything of the
nature of horn becomes bright yellow, while connective-
tissue fibres become blue.

Petromyzon marinus (L.).—Fig. I (A, B, C) illustrates
the development of a horny tooth situated near the edge of tle
mouth. In fig. A the tooth follicle, if it may so be termed,
has already been formed by the downgrowth of the epidermis
into the corium. A small mesodermal papilla, like that of a
hair, is present. The first sign of the developing tooth
occurs a little above the mesodermal papilla. The cells at
this place proliferate, gradually expand, and become granular.
The cells of the epidermis above the developing tooth become
flattened by the pressure exerted.

In fig. B the young tooth is more definitely marked out ;
the cells of which it is composed are considerably larger
than before, and the granular nature of the protoplasm is
still more evident. With picro-nigrosin staining these cells
become faintly. yellowish. The mesodermal papilla has be-
come more or less hollow, and constitutes a kind of pulp-
cavity. The whole structure is strikingly like that of a
developing hair.

ON THE TEETH OF PETROMYZON AND MYXINE. 633

The cells of the young tooth gradually cornify, the process
proceeding from the apex downwards. Cornification may
continue for a considerable time around the base of the
tooth, and the nuclei which resist the process longer than
the rest of the cell may sometimes be seen embedded in the
horn near the growing edge. The tooth breaks through the
cells above, and soon projects above the surface, and like a
hair it is purely ectodermal in origin (fig. C). Even before
the tooth has broken through the surface a new tooth may
be seen developing below on the same site as was occupied
by the first.

A pulp-cavity cannot be found to all the teeth. Those
situated at some distance from the edge of the mouth become
differentiated simply out of a thickening of epidermis.

Myxine glutinosa (L.).—Fig. II represents the tooth of
Myxine in vertical section. Some little distance beneath the
horny tooth can be seen a cone of peculiar cells with large
nuclei. This cone must be the ‘odontoblast cone ” of Dr.
Beard. The large cells stain yellowish, just like the cells of the
developing tooth of Petromyzon. The granules tend to be
arranged in lines, and this gives the margin of the cone a
distinctly striated appearance. As in the case of Petromyzon,
the base of the tooth is firmly fixed in a deep groove of the epi-
dermis, and cornification continues here until the tooth is shed.

The main mass of the tooth projecting out of the epidermis
is formed by the cornification of the cone of large granular
cells with large nuclei, while the remainder is produced in
the horn-producing groove out of comparatively flat unaltered
epidermal cells. In thin sections through the horny sub-
stance of the functional tooth this difference in origin can be
made out. In the upper part of the tooth transformed large
nuclei, and also hollow spaces, can be observed. Towards
the base the transformed uuclei are small and flat. The
hollow spaces just mentioned gradually develop as the large
cells of the cone become converted into horn, and in a
partially cornified tooth they are very conspicuous.

The epidermal cells lining the pulp-cavity are distinctly

634 ERNEST WARREN.

columnar, and would correspond to the enamel epithelium
of an ordinary tooth.

In the case of the ctenoid lingual teeth the tooth-germs
have fused together, and the cones of large granular cells,
which will be converted into the succeeding horny tooth,
are continuous with one another, although small separate
pulp-cavities can be distinguished to each cone or originally
distinct tooth-germ.

The structures described do not lend support to the idea
that these horny teeth are degenerate calcified teeth, but if
they actually are degenerate they must be regarded as having
reverted to a condition that probably preceded the placoid
scale of an Hlasmobranch.

The placoid scale is of very great antiquity, being found
amongst the oldest organic remains, and living and fossil
forms give us little information as to its origin. The stages
in the evolution of a placoid scale must, therefore, be sur-
mised rather than actually observed.

A rough skin was undoubtedly of some use to the ancestors
of Elasmobranchs, and the simplest condition conceivable
would be where the general surface of the body and jaws
was covered with little horny warts. As the warts were
eradually converted into more pronounced structures they
would come to possess a pulp-cavity, in response to the need
of a supply of blood-vessels, etc., to the proliferating epi-
dermis. Such specialised warts where cornification was very
complete would result in structures like the horny teeth of
Petromyzon and Myxine.

Calcification of the outer portion of the pulp would add
strength to the horny scale, and there is no improbability
against supposing that such a variation arose. ‘The shape of
the resulting calcified mass would be moulded by the over-
lying Malpighian layer of the skin (enamel epithelium), just as
the shape of the future tooth is prefigured by the enamel
organ.

If the horny teeth or warts were useful to their possessors,
then a projecting cone of calcified substance would certainly

ON THE TEETH OF PETROMYZON AND MYXINE. 635

be more efficacious, and it would probably not be difficult for
natural selection to replace the horny tooth by its calcified
core. ‘T'his would be the primitive dentine.

We can imagine that at a later period the calcification
which first originated in the mesoderm of the pulp-cavity of
the wart afterwards extended into the Malpighian layer of
the epidermis, and this would constitute the primitive
enamel.

In support of such a view it should be remembered that
among the dentines of the teeth of living and fossil forms we
can meet with every transition from the most irregular
calcified mass of vaso-dentine to the highly organised, fine-
tubed varieties. Also among enamels there can be found
every transition from the thinnest varnish-like layer of
apparently homogeneous calcareous matter, to thick layers
of enamel, consisting of striated prisms and interprismatic
substance.

The teeth of the jaws in the primitive condition would
originate by a separate ingrowth (diagram 1) of epithelium
from the surface for every tooth. Such ingrowths are
frequently indicated externally by shght swellings, which
perhaps represent the horny tooth, which according to the
hypothesis advanced phylogenetically preceded the calcified
tooth. An advance on this condition can be seen in the case
of the Pike, where frequently an enamel-germ buds out
from an older enamel-germ instead of from the general
epidermis.

If the ingrowths of epidermis for the individual teeth
occurred close together or in contact along a single line
around the edge of the upper and lower jaws we should have
the beginning of the tooth-band (diagrams 2, 3, 4).

If, now, all the new enamel-germs for the successional
teeth were regularly budded off from the preceding germs,
instead of only a few, as in the Pike, we should arrive at the
condition well seen in the lower jaw of embryo Scyllium
catulus (diagrams 9, 6).

In an older embryo, and especially in the lower jaw, the

636 ERNEST WARREN.

distinction between the individual tooth-germs becomes lost
dorsally (diagram 7), but ventrally it is retained.

Ou the enamel-cups being pinched off on separate stalks
(“necks”) we should arrive at the typical condition of the
tooth-band seen in a reptile (diagram 8).

EXPLANATION OF PLATE 34,

Illustrating Dr. Ernest Warren’s paper “On the Teeth of
Petromyzon and Myxine.”

Fic. 1, A—C.—Vertical sections of developing teeth from the margin of the
mouth of Petromyzon marinus (L.). xX 280 diameters.

In € the successional tooth is beginning to cornify at its apex beneath the
functional tooth.

Fre. I1.—Vertical section through the median tooth of Myxine gluti-
nosa(l.). x 140 diameters.

Diagram 1.—'looth-germs budded off separately from the surface ; irregu-
larly scattered.

Diagram 2.—Tooth-germs arranged in a single row.

Diagram 8.—Tootlh-germs in contact.

Diagram 4.—Tooth-germs fused together to form a dental lamina (d. 2.) ;
ry = dental ridge.

Diagram 5.—Vertical section through tooth-germ of embryo Scyllium
ecatulus. B is a young tooth-germ being budded off from the pre-
ceding germ.

Diagram 6.—Three tooth-germs (1, 2, 3) and bud B. Above and below
there are indentations marking off several germs.

Diagram 7.—In a somewhat older embryo the distinction between the
individual germs tends to disappear on the upper surface.

Diagram 8.—Enamel cups pinched off on stalks (** necks”).

i

Ruant. Foun. Micr Save ENS P4.3h,

OS

x)
We)

7 Lig Ne
4p Gin \ a

Coca

0 Aa

AS.Huth,Lith® London

TYPHLORHYNCHUS NANUTS. 637

Typhlorhynchus nanus: a New Rhabdocele.
By

F. F. Laidlaw, B.A.

With Plate 35.

Tue small Rhabdoceele described below was found by Mr.
Goodrich at Naples, living on the body of the Polychete
worm Nephthys scolopendroides, Delle Chiaje. He sent
a number of specimens preserved with Lang’s reagent to
Dr. Gamble, who was good enough to hand them over to me
for examination. I am indebted to Mr. Goodrich for a
further series of specimens, some fixed with corrosive sub-
limate and acetic acid, others with Lang’s reagent; also for
figs. 2, 6a, b, drawn from life. I have, unfortunately, been
unable to study living specimens ; hence my account, especially
as regards the genital organs, is necessarily somewhat incom-
plete. The work was done in the Zoological Laboratory at
Owens College.

The external appearance and general characters
are shown in PI. 35, fig. 1.

The total length of the body varies from ‘7 mm. to 1 mm.
In size, therefore, it is quite comparable to a large infusorian
such as Stentor polymorphus, Mull. The variations in
length between different specimens depend largely on the
amount of extension or contraction of the body, especially of
that part of it constricted off to form the snout or proboscis.

The body is spindle-shaped (see Pl. 35, fig. 1), the anterior
end more pointed than the posterior. The front fifth of the

VOL. 45, PART 4.—NEW SERIES. UU

638 F, F. LAIDLAW.

body is constricted off to forma kind of proboscis or pre-
oral lobe (Pr.), which is non-retractile. Immediately behind
this constriction lies the “ mouth ” mid-ventrally (M.). There
are two genital openings, also mid-ventral, lying close together
at about a fifth of the total length of the body from its
hinder end; the male opening (3) is in front of the female
(2).

The surface of body is evenly ciliated throughout. In
some cases there is a disc-like flattening at the hinder end,
but this is only exceptionally present. On the sides of the
proboscis are papillee which resemble in appearance those
figured by von Graff for Proxenetes tuberculatus, v. Gr.
[2]; a few of these papillz are present at the hinder end of
the body.

The mouth leads through a well-developed pharynx (Ph.)
into the spacious gut cavity (Hn.), which sends forward a
median diverticulum (A.d.) into the proboscis, terminating
immediately behind the brain (5r.), which hes at about the
middle of the proboscis. On either side of this anterior
diverticulum lies a rounded testis (7e.). There is a well-
developed penis provided with a complicated armature (Pe.)
lying in front of the male opening. Into the penis open a
pair of vesiculee seminales (V.s.) and a number of unicellular
olands (Gl.). The female aperture leads into a bursa semi-
nalis (B. s.) provided with a chitinous appendage (Ch. a.).
The single ovary (Ov.) lies at the hind end of the body nearly
in the middle line ; it is somewhat curved, and the eggs are
progressively riper from behind forwards. In front of the
ovary is a structure which may be called the receptaculum
seminis (R.5s.).

The pigmented eye-spots found in many Rhabdocceles are
here absent, and there is no otolith.

In several respects this little creature differs from any
known forms; its nearest allies appear to be found amongst
the Mesostomide and Proboscide as defined by von
Graff [2], but as it cannot well be referred to any known
genus of either of these families I have found it necessary

TYPHLORHYNCHUS NANUS. 639

to create a new genus for its reception under the name of
Typhlorhynchus. The species may be called Typhlo-
rhynchus nanus.

Habits.—Typhlorhynchus nanug, as already stated, is
found on the body of Nephthys scolopendroides, to
which it attaches itself by its hinder end. From Mr. Good-
rich’s figures (figs. 2 a—c) it appears to use its proboscis as
a tactile organ. In no specimen that I have examined is
there any food substance in the gut space, but in the proto-
plasm of the gut wall are numerous fine food granules. Its
epizoic habit does not seem to have produced any marked
degeneration of the organs of the body. The loss of the eye
may be due to this, but a number of species belonging to the
genus Mesostomum which are free-living (e.g. M. Cuenotzi,
recently described by Dérler [5]) are also without eyes.

The other known parasitic Rhabdocceles belong to the
family Vorticide, and include a number of forms parasitic
in Kchinoderms, on the kidneys and gills of Molluses, ete.

Method.—Mr. Goodrich obtained specimens of Typhlo-
rhynchus by putting Nephthys scolopendroides into
sea water with about 10 per cent. of alcohol (70 per cent.),
when the parasite fell off in considerable numbers. No eggs
were found. They were then treated with corrosive sub-
limate and acetic acid or with Lang’s reagent. Those
treated with the latter yield on the whole the best results in
sections, the protoplasm being beautifully preserved. Those
hardened with the former reagent have shrunk to some extent,
but in them the gland cells and rhabdites stain more readily.
I have examined a number of series of sections both trans-
verse and longitudinal, as well as specimens mounted whole.
For the sections I obtained excellent results with brazilin ;
I also employed the iron-hematoxylin method of staining.
Whole preparations were stained with borax carmine.

Structure: Integument.—The body-wall is made up of
a layer of ciliated epithelium lymg upon a double muscle
layer. The epidermis is equally ciliated over the whole
surface of the body. A number of irregularly arranged

640 F, F. LAIDLAW.

papille formed by the bulging out of the epidermis occur on
the proboscis (fig. 5, 7’.), and some few at the hind end of the
body. The flattening observed at the hind end of several
specimens is undoubtedly connected with the mode of attach-
ment observed by Mr. Goodrich.

The average thickness of the epidermis is 5y. No
cell limits can be discerned init. In sections of specimens
preserved with Lang’s reagent small clusters of nuclei, four
or five in each cluster, are scattered at considerable inter-
vals through the epidermis. he outer limit of the epidermis
is seen as a delicate line, which under high power resolves
itself into a row of exceedingly fine dots, recalling exactly
the appearance figured by Bohmig for Monoophorum
striatum, Boh.; whilst the clusters of nuclei suggest the
“ Tastkérperchen ”’ of the same species (8, pl. xx, figs. 17—19),
except that there is no break in the cilia above them. In the
case of sections of specimens fixed with corrosive sublimate
and acetic acid numbers of small deeply stained rhabdites
are visible in the epidermis. Assuming that the clusters of
nuclei referred to above are to be regarded as connected
with sensory organs, the question arises as to the where-
abouts of the true epithelial nuclei. The nuclei in the
clusters are the only ones occurring in the epidermis so far
as my sections show.

The basal membrane is thin, about 1 mw thick or rather
less. The muscle layers consist of an outer circular and an
inner longitudinal layer, evenly developed all over the body
(fig. 7, Cir., Lon.). Special sphincter muscles, derived appa-
rently from the circular layer, he round all the three openings
in the body-wall (figs. 4—7, Sp.). But few gland-cells are
developed in connection with the surface of the body. A
few cells with granular deeply staining protoplasm lie here
and there immediately under the muscle layers of the body-
wall, and doubtlessly come into this category; they are
more numerous in the proboscis than elsewhere (fig. 5, Gi.).

Parenchyma.—tThe space between the various organs of
the body and the body-wall is occupied by the parenchyma,

TYPHLORHYNCHUS NANUS. 641

This consists of a reticulum of delicate fibrillar protoplasm
containing round finely granular nuclei; the protoplasm is
without cell limits, and the nuclei are few and widely scattered.
Most of the body organs, e.g. yolk glands, penis, ete., lie in
perfectly definite spaces in this parenchyma, but in two
cases, viz. the gut wall and the bursa seminalis, this is not
so. The lining of the gut space consists of protoplasm
without cell limits, of precisely the same character as that
of the parenchyma, and it is not marked off from the latter
in any way. ‘The only characters which serve to distinguish
the gut wall (endoderm) from the parenchyma are, firstly, the
nuclei, which in the endoderm are oval, and contain coarse
darkly staining chromatin granules (fig. 7, Nuc.), whilst those
of the parenchyma, as stated above, are round and _ finely
granular (fig. 7, P..N.); and secondly, the presence in the
endoderm protoplasm of numbers of fine granules, which are
probably food granules, but these only disappear gradually
in passing from the endoderm to the parenchyma.

In the case of the bursa seminalis the protoplasm forming
its walls, though denser and more hyaline than the general
parenchyma protoplasm, nevertheless merges quite imper-
ceptibly into it.

Owing to the spaciousness of the enteron and the size of
the yolk glands in the middle regions of the body, the
parenchyma in those regions is much reduced.

In the proboscis the parenchyma is densest immediately
under the body-wall; below this it is spongy and scarcely
distinguishable, especially towards the tip of the proboscis.

Alimentary Canal.—The mouth opens on the mid-ventral
line just behind the constriction at the base of the proboscis. As
already stated, there is a sphincter muscle arrangement around
the mouth opening, developed from the muscles of the body-
wall, most probably from the outer circular layer, but possibly
from both ; my sections do not bring this point out clearly.

‘he mouth opens into the pharyngeal pouch (“ Pharyngeal-
tacshe,” von Graff). This pouch is at first narrow, but as it
passes dorsally it widens into a chamber of small size, into

642 F. F. LAIDLAW.

which the lower part of the pharynx projects (fig. 3,
Ph. T., Ph. T;.), so that the roof of the chamber is formed by
the pharynx itself. The pouch agrees very closely with that
figured by von Graff [2] for Mesostomum Khrenbergili,
and, as in the latter, it is provided with a few muscle-fibres
attached to the walls at its widest part, running to the
body-wall. Neither the epithelium lining the wider part of
the pouch nor that on the exposed part of the pharynx is
ciliated, the cilia only extend to the narrow part of the
pouch. With this should be compared the condition found
in Mesostomum Ehrenbergii, v. Graff, in which the pouch
is ciliated throughout. In Mesostomum Cuenoti, recently
described by Dorler [5], on the other hand, only the roof of
the pouch, i.e. the projecting wall of the pharynx, is ciliated.
The pharynx itself has its principal axis a little elongated,
and running nearly dorsiventrally. It is pyriform, its narrow
end being ventral and the broader dorsal. The larger size
of the dorsal part is due to the greater size in that region
of the pharyngeal cells (fig. 3, Ph. C.).

The pharynx conforms to the type called by von Graff
‘pharynx rosulatus,” characterised by the numerous large
gland cells which in sections appear as coarsely granular
cells, and also by the arrangement of the muscle-fibres.
The latter consist of a double outer layer, forming, as it
were, a muscular capsule, and a double inner layer lying
immediately under the epithelium lining the lumen of the
pharynx. Between these two definite layers run numerous
radial fibres (fig. 8, R.M.). The outer layer consists first
of longitudinal fibres (O.Z.),—that is to say, of fibres running
parallel to the principal axis of the pharynx; and immedi-
ately below these, of stouter circular fibres (O. Cir.). The
inner layer consists of a number of fine circular fibres
(fig. 4, I. c.), lying immediately on the outside of the epi-
thelial lining of the lumen (PA. H.), and outside these of
longitudinal fibres (fig. 3, J. Z.), which like the inner circular
fibres are but feebly developed. The radial muscles are
most numerous in the ventral part of the pharynx.

TYPHLORHYNCHUS NANUS. 643

The epithelium lining the lumen (Ph. H.) is very much
reduced, excepting at the lower end of the pharynx, where
it is continuous with the epithelium on the exposed surface.
It is quite devoid of nuclei (cf. Mesostomum Cuenoti).

In addition to the intrinsic muscles of the pharynx a large
number of fibres run from the body-wall to become attached
to its anterior wall, apparently fusing with the outer longi-
tudinal fibres of the pharynx itself. It is the insertion of these
fibres into the body-wall which causes the constriction that
cuts off the proboscis from the rest of the body (fig. 4, A. ML).

There is no well-marked cesophagus; a few unicellular
glands lie in the neighbourhood of the upper end of the
pharynx.

The gut is spacious ; posteriorly it extends as far back as
the level of the female aperture. That part of it running
into the proboscis is best described as an anterior unpaired
diverticulum (fig. 1,A.d.). This diverticulum is often occluded
to a considerable extent by pseudopodial processes sent out
by the endoderm, but can always be distinguished in trans-
verse sections. ‘The characters of the endoderm have been
sufficiently described in dealing with the parenchyma.

Nervous System.—The brain lies in the proboscis at
about its middle, and consists of a quantity of ganglion cells
(fig. 5,G.) lying about a transversely elongated mass of
‘ Punktsubstanz,” apparently composed of exceedingly fine
fibrillee (fig. 5, br.).

The Punktsubstanz, or central mass, is divided into two
lobes by a slight median constriction, and from each of the
lobes a group of nerve-fibres runs outward for a short
distance towards the wall of the proboscis, and then turns
backwards (fig. 5, N.). A group of fibres also runs from the
ganglion cells lying in front of the central mass to each lobe
of the latter close to the middle line, and a similar group of
fibres passes into the central mass from ganglion cells lying
behind it. These posterior fibres do not form such compact
groups as the anterior pair, and enter the central mass more
laterally.

64.4. F. F. LAIDLAW.

The ganglion cells stain very darkly, and appear to be
bipolar ; the nucleus is finely granular and rather large, and
there is a small black nucleolus.

The arrangement of ganglion cells lying on the anterior
side of the central mass isremarkable. On either side of the
middle line a group of them extends forwards like a horn
from the main body of ganglion cells, and curves slightly
inwards at its anterior extremity, ending some distance from
the tip of the snout. The cells constituting these two horns
are identical in character with the other ganglion cells.
Numerous nerve processes run forwards from them to the
tip of the snout, others run down to join the groups of fibres
entering the brain anteriorly (fig. 1, H.).

The tip of the proboscis then appears to have a very rich
nerve supply, and we may conclude that the proboscis is’ the
chief seat of the tactile sense. Possibly with this is corre-
lated the presence on the proboscis of the papillz already
referred to, although such papille are not entirely confined
to it.

Organs of Reproduction.

(a) Female.—The ovary is single, and lies at the hinder
end of the body nearly in the middle line; it is short and
somewhat curved (fig. 1, Ov.), its length is about -12 mm., and
it contains from twelve to fifteen eggs, which are progressively
riper from behind forwards. The eggs are oblong, closely
pressed against each other, and have large nuclei. Their
protoplasm is finely granular. ‘he nuclei have a deeply
staining capsule ; immediately below this a number of coarse,
dark granules, and inside these a relatively clear space in
which hes a large black nucleolus (fig. 8, Ov.). As already
stated, the ovary lies in a space hollowed out in the
parenchyma, and is surrounded by a protoplasmic membrane
containing flattened nuclei.

In front of the ovary this membrane is apparently attached
to a short funnel-shaped structure which ends blindly in the

TYPHLORHYNCHUS NANUS. 645

parenchyma, and has its wider end next the ovary (fig. 8,
fi. s.). In the space enclosed by this body, which may be
called the receptaculum seminis, lies a small mass composed
of spermatozoa (fig. 8, 8.), which are thus immediately in
front of the ripest, most anterior egg.

There is a pair of unbranched yolk glands which extend
from the level of the front end of ovary as far forward as the
pharynx, lying close along the lateral wall of either side of
the body in a cavity of the parenchyma, which here is not
much developed.

These yolk glands are built up of elongated cells lying
parallel to one another and closely pressed together, with their
long axes roughly dorsiventral. Each cell has a nucleus at about
its middle ; each nucleus contains a nucleolus lying in a clear
space in the centre of the nucleus, surrounded by a finely
granular ring when seen in section. In each cell are two,
three, or several small black refringent bodies, which tend to
group themselves together in a ring. Amongst these black
bodies are found small refringent yellow globules of yolk
matter.

The female aperture is furnished with a sphincter muscle
arrangement (fig. 7, Sp.), and opens into a small chamber
(fig. 7, B. s.). The ciliated epidermis of the body-wall
extends to the walls of this chamber. Into it the bursa
seminalis (fig. 7, B.s.) opens dorsally through a narrow neck.
Above the neck, which is quite short, the cavity is transversely
widened, but narrow antero-posteriorly. Dorsally it extends
to a point just below the front end of the ovary, but does
not communicate with the latter directly. The cavity is
bordered by a chitinoid lining substance which stains deeply ;
beyond this lining the walls are built up of clear hyaline
protoplasm, which merges quite gradually with the proto-
plasm of the parenchyma. A small number of muscle-fibres
lie in the walls of the bursa, running from the neighbourhood
of the neck of the cavity to the body-wall, or im some cases to
the capsule of the penis (fig. 7, M.). The hinder wall of the

bursa at about its middle is produced to form a kind of spout

646 Ff. F. LAIDLAW.

or short tube, which is blocked up by a small chitinous plug
(Ch. A.). This latter agrees with the “ Chitinanhang” de-
scribed by von Graff for Hyporhynchus coronatus, v. G.
[2], and with a similar organ found in many other forms.

Owing to the way in which the walls of the bursa seminalis
merge into the parenchyma in whole preparations it is only
possible to determine the position of the lumen. In sections
some distinction, as pointed out above, can be drawn between
the tissue immediately surrounding the lumen and_ the
parenchyma proper. In no case that I have examined does
the bursa seminalis contain spermatozoa.

I have found considerable difficulty in interpreting the
funnel-shaped organ which I have called the receptaculum
seminis ; the explanation here put forward of this organ was
suggested to me by an examination of a figure of Byrso-
phlebs Graffii (Jens.) in Jensen’s work ‘ Turbellarier ved
Norges Vestkyst’ [1].

In Byrsophlebs Graffii the riper end of the single
ovary is posterior, and immediately behind it lies an organ
which Jensen calls the receptaculum seminis. This recepta-
culum opens to the exterior, and contains spermatozoa which
do not, however, according to Jensen in his account of this
species, reach it directly, but pass through the opening intoa
bursa copulatrix (the receptaculum and bursa having a
common opening), and from the bursa travel along a long
convoluted duct, called by him the ductus longus, into a
receptaculum. Now it is evident that, so far as its position
goes, the receptaculum bears the same relation to the ovary of
Byrsophlebs Graffii, as does the funnel-shaped organ to
the ovary of Typhlorhynchus nanus. mle moreover,
contain spermatozoa.

The organ called by Jensen bursa copulatrix would, then,
be homologous with the organ which I here call the bursa
seminalis. My reason for adhering to the latter name is that
it is used by von Graff to designate a comparable organ in
Hyporhynchus and other genera.

If, then, we may suppose that the receptaculum seminis in

TYPHLORHYNCHUS NANUS. 64.7

Typhlorhynchus has lost its opening to the exterior, we can
readily compare the female organs in this creature with those
of Byrsophlebs. The spermatozoa may reach the receptacu-
lum (funnel-shaped organ) by a duct similar to the ductus
Jongus of the latter; such a duct would scarcely be dis-
cernible save in the living state and when full of sperma-
tozoa.

It should be remarked that the walls of the bursa are much
folded, so that the lumen is quite irregular.

(b) Male: Testes.—There is a pair of compact spherical
testes, one on either side of the proboscis immediately behind
the brain (fig. 5, Ze.), each enclosed in a very delicate
membrane, which often is hardly distinguishable. In every
specimen that I have examined I have found apparently mature
spermatozoa lying for the most part on the dorsal and anterior
surface of either testis. The rest of the testis is composed
of sperm mother-cells in various stages of development.
I have not found it practicable to follow out the history of
the development of the spermatozoa. ‘The appearance of
the cells composing the testes agrees very closely with that
of the germ-cells of Plagiostoma Girardi figured by von
Graff (l.c., Taf. xvi, figs. 11—14). I have also compared
sections of the testes of Typhlorhynchus with some of
Mesostomum tetragonum, O. Sch., and find there, too, a
strone resemblance. Cells in a morula state are always
present (fig. 5, Mo.). ‘The position of the testes is hardly
paralleled amongst the Mesostomidz and Proboscide.

In Macrorhynchus Naegelii, as figured by von Graff
(loc. cit., af. xi, fig. 7), they extend as far forward as the
level of the brain and pharynx, but in no case do they lie in
the retractile proboscis. In most Mesostomide and Probos-
cide the testes lie in the middle region »f the body, and are
continuous with the vesicule seminales. In 'lyphlorhynchus,
however, I have not been able to find any communication
between them, although the two vesicule always contain
spermatozoa in my specimens. ‘These vesicule (fig. 1, V. s.)
are narrow tubes, with thin chitinous-looking walls, opening

648 F. F. LAIDLAW.

close together into the penis at their posterior end, and each
ending in a small swelling at their forward extremities.
They have a total length of about °15 mm.

The penis is about ‘1 mm. long, backwardly directed,
pyriform, with its apex curved ventrally. Close to the point
at which the vesicule seminales open into it there open also
a number of gland cells, which lie immediately ventral to the
penis, and pour a secretion into it. The penis itself consists
of an outer muscular capsule composed of muscle-fibres
running parallel to the long axis of the penis (fig. 7, Hx. M.).
There is an inner muscle layer also composed of a cylinder
of longitudinal fibres, attached at one end to the outer
capsule at its widest part, and by the other to the eversible
part of the penis (fig. 7, J. M.). The outer capsule is con-
tinuous with the lining of the cavity immediately within the
male aperture. The armature of the penis is very remark-
able, and quite unlike that of any previously described form.
Its appearance is well shown in Mr. Goodrich’s figures (figs.
6a, b). When evaginated the penis is mushroom-shaped,
with a convex head. From the margin of the disc of the
head extend two lobes, one on either side [4]. From the
centre of the head projects a long chitinous spine, whose
proximal end is sharply crooked and embedded in the penis
(fig. 6, Ch.). The convex surface of the head is covered with
meridionally arranged rows of short, slender, slightly curved
spines (C.S.). There appear to be some eighteen of these
rows, each with ten or twelve spines.

The large central chitinous spine is tubular at its proximal
end, but the tube distally appears to open out into a groove.
This spine is an impregnating organ homologous with the
“ Chitinrohr”’ described by von Graff [2] for Proboscide
and Mesostomide. ‘The spermatozoa probably pass into its
tube by an aperture at its proximal end (cf. Proxenetes
gracilis, v. Gr. [2,.Taf. viii, fig. 12]). A few muscle-fibres
are apparently attached to its base.

Affinities.—The character of the pharynx is sufficient to
indicate that the Rhabdoccele under consideration is allied to

TYPHLORHYNCHUS NANUS. 649

the families Mesostomide and Proboscidx, and there are
no features in the structure of Typhlorhynchus which
forbid us to refer it to one or other of these families. Which
of the two is to be selected depends chiefly on the importance
attached to the proboscis. This in Typhlorhynchus
differs sharply from a typical proboscis, such as is found in
Macrorhynchus or Gyrator, but not so greatly from
that of Pseudorhynchus. Inallthe genera referred by von
Graff [2] to this family, however, the proboscis is retractile
to some extent. Further, in none of them do the brain or
the testes occupy a position similar to that found in
Typhlorhynchus, and in Pseudorhynchus alone is the
proboscis invaded by the gut space.

On the other hand, Byrsophlebs amongst the Meso-
stomids is characterised by the presence of two genital
apertures, the male in front, the female behind—a character
that only occurs m this genus and in Typhlorhynchus
amongst the whole of the Rhabdoccela (s. str.), leaving out
of account the Prorhynchide. Further, as I think I have
shown, the female genital apparatus of Typhlorhynchus may
be compared in detail with that of Byrsophlebs.

A bursa seminalis provided with a chitinous appendage
very like that of Typhlorhynchus occurs in Hyporhyn-
chus amongst the Proboscide and Proxenetes amongst
the Mesostomids.

The penial apparatus, whilst differing greatly in detail
from any of those figured for these families by von Graff,
resembles them in a general way, especially in being
provided with a chitinous tube or spout (Chitinrohr—cf. von
Graff's figures of Hyporhynchus coronatus and Proxe-
netes gracilis).

It is, on the whole, I think, most convenient to place this
new genus amongst the Proboscide in the neighbourhood of
Pseudorhynchus. It differs sufficiently from other Proboscide
to warrant the creation of a sub-family to receive it. In some
respects, e.g. the female organs, it shows an approximation
to Byrsophlebs, and may be regarded as to some extent

650 F. F, LAIDLAW.

intermediate between the Mesostomidee and Proboscide. It
is particularly of interest as being the only member of either
of these families that has adopted an epizoic habit.

The character of the parenchyma should be specially
remarked. This, in the way in which it merges into the
endoderm, shows a distinct approach to the condition found
in the Alloioccela.

The genus Typhlorhynchus may be defined briefly as
follows :

Body provided with a non-retractile pre-oral lobe or
proboscis. Gut not clearly separated from the parenchyma,
provided anteriorly with a median diverticulum extending
into the pre oral lobe. Pharynx rosulate, no genital atrium,
male opening in front of female. Penis (when evaginated)
with meridionally arranged rows of spines, and in addition a
long chitinoid tube. The single pair of testes lie in the
pre-oral lobe ; ovary single, at hind end of body ; yolk glands
paired. Accessory female organs consist of—(1) a bursa
seminalis opening to exterior by the female aperture; (2) a
receptaculum seminis.

In conclusion, I wish to thank Dr. Gamble very sincerely
for the kind way in which he has assisted and advised me in
preparing this account.

LITERATURE.
1. Jensen, Oxar 8.-—‘ Turbellaria ad Litora Norvegia Occidentalia,
Bergen, 1878.
2. GRrarr, von.—‘ Monographie d. Turbellarien. I. Rhabdocceliden,’ 1882.
3. Boumie, L.—‘‘ Untersuch. iit. rhabdocoele Turbellarien,” ‘ Zeits. f. wiss.
Zool.,’ li, pp. 167—479, 1890-91.
4. Gamsie, I’, W.—“ British Marine Turbellaria,” this Journal, April, 1893.
5. Dorter, A.—* Neue u. wenig bekannte rhabdocoele Turbellarien,” ‘ Zeits,
f, wiss. Zool.,’ Ixviij, 1900, pp. 1—42.

T'YPHLORHYNCHUS NANUS. 651

EXPLANATION OF PLATE 35,

Illustrating Mr. F. F. Laidlaw’s paper “On Typhlorhyn-
chus nanus.

Fie.1.—Typhlorhynchus nanus, diagrammatic, seen from above. x 90.
Pr. Pre-oral lobe, snout, or proboscis. M. Mouth. P&#. Pharynx. Hx. Gut
space. 4.d. Anterior gut diverticulum. 87. Brain. H. Horn-shaped process.
Te. Testis. Pe. Penis. G/. Glands opening into the penis. G/,. Glands
lying about the male aperture. @. Male aperture. V.s. Vesicula seminalis.
Ov. Ovary. &.s. Receptaculum seminis. 2B. s. Bursa seminalis. Ch.a.
Chitinous appendage. Y.gl. Yolk gland. 2. Female aperture.

Fie. 2.—Various attitudes assumed in life. x about 10 times.

Fie. 3.—Long. sec. through the pharynx. x 400. Ph.ec. Pharynx cells.
O. L. Outer longitudinal muscles of pharyux. O. Cir. Outer circular muscles,
and J. Z., inner longitudinal muscles of the same. PA. #. Epithelium lining
the lumen of the pharynx. PA. 7. Pharyngeal pouch. PA. 7). Narrow part:
of pharyngeal pouch. £&. AZ. Radial muscles of pharynx. (The inner circular
muscles are seen as a row of dots immediately to the inside of the inner
longitudinal fibres.) Sp. Sphincter muscle of the ‘mouth ”’ aperture. 4. WZ.
Muscles running from the anterior side of the pharynx to the body-wall.
Ee. M. Other extrinsic muscles of the pharynx.

Fie. 4.—Trans. sec. through the pharynx region showing the gut diver-
ticulum (4.d.). J. c. Inner circular muscles of the pharynx. Other lettering
as In Fig. 4,

Fic. 5.— Horizontal section through the pre-oral lobe. G. Ganglion cells.
Gl. Integumentary gland cells. Jo. Cells in morula stage in testis. J.
Nerve. Other lettering as in Fig. 1.

Fic. 6 a.—Penial armature closed.

Fic. 6 d.—The same evaginated, ‘“ much enlarged, drawn from life.’ Ch.
Chitinous tube. C. 8. Chitinous spines. Z. Lobes.

Fie. 7.—Long. sec. through the penis and bursa seminalis (drawn from
two sections and combined). (i. Circular muscle-fibres of body-wall. Loz.
Longitudinal fibres of body-wall. Zz. Endoderm. Hx. WM, External muscular
capsule of the penis. J. M@. Inner muscular layer of the same. J.
Musele-fibres. 47, Penial armature. Sp. Sphincter muscle-fibres of female

aperture. B.s. Chamber into which the bursa seminalis opens. JB. s,.

652 EY Pa) BALD TAW=

Bursa seminalis. ue. Endoderm nucleus. P. NV. Nucleus of parenchyma.
Other lettering as in Fig. 1.

Fie. $.—Optical section through the bursa seminalis, receptaculum seminis,
and ovary of a specimen mounted whole. Ov. Ovum showing the large
nucleus. i. Mass of spermatozoa in the receptaculum. Other lettering as
in Fig. 1.

(The position of the ‘mouth opening

Fig. 1.)

” is drawu rather too far back in

t
: \
- ‘~~ oe -_ = ae a a
= ‘ - Le - oe. io
a nto ee a > = — 1
PSSST SSeS ae ea a a ee ea SE a =: HiZIS HIHIA IES EE SS ee $ivazssasssszsass
=r = See == = ses preeeeeeteteteeretteeerrertrettetrnt arse etree sears ee ee arene scam
= SEHIES EES a aes aaeS gS STSE SE Soe Sa SSeS SanE SS FoSTE STS eons Sane oo es resents saaeae ge eeee meee
2 Sa See

Ail,

SS

===

Fig? 2.6.E.S. Goodrich del
{SE .
Fig? 13-5.7.8. FR Laidlaw del,

AS. Huth, Lith? London.

INDEX

TOr VO ln As,

NEW SERIES.

Actinotrocha, note on, by Ramunni
Menon, 473

Actinotrocha, review of Skeda’s work
on, by Masterman, 485

Admetus pumilio, development
of, by L. H. Gough, 595

Allis on sensory canals, eye muscles,
and cranial nerves of Mustelus,
87

Amphioxus, excretory organs of, by
Goodrich, 493

Ashworth on the anatomy of Seali- |

bregma inflatum, 237

Bradford and Plimmer on Try pauo-
soma Brucil, 449

Chilopod, a new and annectant type
of, by R. I. Pocock, 417

Dendrocometes paradoxus, con-
jugation of, by Hickson and Wads-
worth, 325

Dendy on the oviparous species of

Onychophora, 363

Euperipatus Weldoni, develop-
ment of, by Richard Evans, 41

Eusthenopteron, pelvic girdle and fin
of, by Edwin 8. Goodrich, 311

VoL. 40, pART 4,—NEW SERIES.

| Evaus onthe development of Euperi-

patus Weldoni, 41
Gasterosteus aculeatus, deve-
lopment of the skull of, by Swinner-
ton, 503
Goodrich on the excretory organs of
Amphioxus, 493

_ Goodrich on the pelvic girdle and fin

of Kusthenopteron, 311
Gough on the development of Adme- -
tus pumilio, 595

Hickson and Wadsworth on the con-
of Dendrocometcs
paradoxus, 325

jugation

Kerr, Graham, on the development of
Lepidosiren paradoxa, witha
note upon stages in Protopterus
annectens, 1

Laidlaw Typhlorhynchus
nanus, anew Rhabdocele, 637
Lepidosiren paradoxa, develop-
ment of, Part LI, by Graham Kerr,

i

On

Nagana or 'I'setse fly disease, the
organism of, by Bradford and
Plimmer, 449

xX X&

654

Masterman, review of Skeda’s work
on Actinotrocha, 485

Menon on Actinotrocha, 473

Mustelus levis, sensory canals,
eye muscles, and nerves of, by
Allis, 87

Myxine and Petromyzon, teeth of, by
Warren, 631

Onychophora, Malayan species of, by
Richard Evans, 41

Onychophora, the oviparous species of,
by Arthur Dendy, 365

Petromyzon and Myxine, teeth of, by
Warren, 631

Plimmer and Bradford on Try pano-
soma Brucll, 449

Pocock on a new and annectant type
of Chilopod, 417

Protopterus annectens, noves on
development of, by Graham Kerr,
]

INDEX.

Rhabdocele, a new, by Laidlaw, 637 —

Scalibregma inflatum, anatomy
of, by J. W. Ashworth, 237

Skull, development of, in Gasterosteus ,
by Swinnerton, 503

Swinnerton, development of the skull
of the three-spined stickleback,
Gasterosteus aculeatus, 503

Teeth of Petromyzon and Myxine, by
Warren, 631

Trypanosoma Brucii, by Brad-
ford and Plimmer, 449

Typhlorhynchus nanus, a new

Rhabdoceele, by Laidlaw, 637

Wadsworth and Hickson on the con-
jugation of Dendrocometes
paradoxus, 325

Warren on the development of the
teeth of Petromyzon and Myxine
631

PRINTED BY ADLARD AND SON,
BARTHOLOMEW CLOSE, E.C.; 20, HANOVER SQUARE, W.;
AND DORKING.

We recommend for the Rontgen and Scientific

Institutes our

Special Rontgen

N

Special Extra Rapid Schleussner
Plates,

which are unrivalled, and recognised as the best
by the Professors
Dr. A. Hoffa, N. Wolfer, Dr. Winkelmann,
Dr. R. A. Koch, Dr. Rehn, Dr. H. C. Vogel,
Dr. Max Wolf, H. Kayser, Dr. N. Thege,
Dr. F. Krause, Dr. H. Nitsche, Dr. Rieder,
Dr. Quincke, etc.
Only one trial of these Plates is required to prove
that they possess in every respect the good qualities —
we claim for them.

PLEASE WRITE FOR SAMPLES TO THE MANUFACTURERS,

TROCKENPLATTENFABRIK

Dro. SCH LE USSN Ee

AKTIENGESELLSCHAFT,

FRANKFORT-ON-THE-MAIN.

PRICE LISTS AND SAMPLES FREE.

Rs a

Tua) 3

4. A v A yi “J

With Ten Plates, Royal Ato, 5s.
CONTRIBUTIONS TO THE KNOWLEDGE OF RHABDOPLEURA
AND AMPHIOXUS.
By E. RAY LANKESTER, M.A., LL.D., F.R.S.
London: J. & A. CHURCHILL, 7 Great Marlborough Street.

Quarterly Journal of Microscopical
Science.

The SUBSCRIPTION is £2 for the Volume of Four Numbers ;
for this sum (prepaid) the JournaL is sent Post Free to any part
of the world.

BACK NUMBERS of the Journat, which remain in print, are
now sold at an uniform price of 10/-.

The issue of SurppLumenr Numbers being found inconvenient,
and there being often in the Hditor’s hands an accumulation of
valuable material, it has been decided to publish this Journal at
such intervals as may seem desirable, rather than delay the appear-
ance of Memoirs for a regular quarterly publication.

The title remains unaltered, though more than Four Numbers
may be published in the course of a year.

Each Number is sold at 10/-, and Four Numbers make up a
Volume.

London: J. & A. CHURCHILL, 7 Great Marlborough Street.

TO CORRESPONDENTS.

Authors of original papers published in the Quarterly Journal
of Microscopical Science receive twenty-five copies of their
communication gratis.

All expenses of publication and illustration are paid by the
publishers.

As a rule lithographic plates, and not woodcuts, are used in
illustration.

Drawings for woodcuts should nor be inserted in the MS.,
but sent in a separate envelope to the Hditor.

Contributors to this Journal requiring extra copies of their
communications at their own expense can have them by applying
to the Printers,

Messrs. Aptarp & Son, 224, Bartholomew Close, E.C., on
the following terms:

For every four pages or less—

25 copies ; . : ‘ 3/-
On JR See ee eal
LOD. 5; : 7/-

Plates, 2/- per 25 if uncoloured ; if coloured, at the same rate for
every colour.
Prepayment by P.O. Order is requested.
ALL CoMMUNICATIONS FOR THE EDITORS TO BE ADDRESSED TO THE CARE
or Messrs. J. & A. CuurcHitt, 7 Great Marieorovues SrReEEt,
Lonpon, W.

THE MARINE BIOLOGICAL ASSOCIATION

OF THE

UNITED KINGDOM.
20:
Patron—HIS MAJESTY THE KING.
President—Prof. E. RAY LANKESTER, LL.D., F.RS.

70;

THE ASSOCIATION WAS FOUNDED “ TO ESTABLISH AND MAINTAIN LABORATORIES ON
THE COAST OF THE UNITED KINGDOM, WHERE ACCURATE RESEARCHES MAY BE CARRIED
ON, LEADING TO THE IMPROVEMENT OF ZOOLOGICAL AND BOTANICAL SCIENCE, AND TO
AN INCREASE OF OUR KNOWLEDGE AS REGARDS THE FOOD, LIFE CONDITIONS, AND HABITS
OF BRITISH FOOD-FISHES AND MOLLUSCS.”

The Laboratory at Plymouth
was opened in 1888. Since that time investigations, practical and scientific, have
been constantly pursued by naturalists appointed by the Association, as well as by
those from England and abroad who have carried on independent researches.

Naturalists desiring to work at the Laboratory

should communicate with the Director, who will supply all information as to
terms, &c.

Works published by the Association
include the following :—‘ A Treatise on the Common Sole,’ J. T. Cunningham, M.A,,
4to, 25/-. ‘The Natural History of the Marketable Marine Fishes of the British
Islands,’ J. T. Cunningham, M.A., 7/6 net (published for the Association by
Messrs. Macmillan & Co.).

The Journal of the Marine Biological Association
is issued half-yearly, and is now in its fifth volume (price 3/6 each number).
In addition to these publications, the results of work done in the Laboratory
are recorded in the ‘ Quarterly Journal of Microscopical Science,’ and in other
scientific journals, British and foreign.

Specimens of Marine Animals and Plants,
both living and preserved, according to the best methods, are supplied to the
principal British Laboratories and Museums. Detailed price lists will be forwarded
on application.

TERMS OF MEMBERSHIP.

AnnvuAL MEMBERS . : : £1 2 Oper annum.
LIFE MEMBERS . ; : : . 15 15 O Composition Fee.
FOUNDERS . LOOM OO 5 if

Governors (Life Members of Council) 500 O O a 23

Members have the sake rights and privileges:—They elect annually the
Officers and Council; they receive the Journal free by post; they are admitted to
view the Laboratory at any time, and may introduce friends with them ; they have the
first claim to rent a table in the Laboratory for research, with use of tanks, boats, &e.;
and have access tothe Library at Plymouth. Special privileges ure granted to Governors,
Founders, and Life Members.

Persons desirous of becoming members, or of obtaining any information with
regard to the Association, should communicate with—

The DIRECTOR,
The Laboratory,
Plymouth.

—_—_———

ERNST MAYR LIBRARY

iii

3 2044 1

Date Due
| . F yi

APR 29 1955