Skip to main content

Full text of "Quarterly journal of microscopical science"

See other formats 


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