_ nae a Sit 


5 
Mi 
4, 


en a are 


2 ENT ee 


> Cite 
son irda Spina ford t 
a ee, 


HARVARD UNIVERSITY. 


LIS RADY 


OF THE 


QUARTERLY JOURNAL 


OF 


MICROSCOPICAL SCIENCE. 


EDITED BY 


E. RAY LANKESTER, M.A., LL.D., F.B.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 ACADRMY 
OF SCIENCES OF PHILADELPHIA; FOREIGN MEMBER OF THE ROYAL BOHEMIAN 

SOCIRTY OF SCIENCES, AND OF THE ACADEMY OF THE LINCEI OF ROME, 

AND OF THE AMERICAN ACADEMY OF ARTS AND SCIENCES OF BOSTON j 

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., 


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


AND 


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


BEYER PROFESSOR OF ZOOLOGY IN THE OWENS COLLEGE, MANCHESTER, 


VOLUME 46.—New Surtzs. 
With Rithographic Plates and Engrabings on Wood. 


Jt LONDON: 
J. & A. CHURCHILL, 7, GREAT MARLBOROUGH STREET. 
1905. 


a - es i i 


a —— el 
oP Rae A) ae 
a + 


| UA) ee 


a 


———_ 
4 
om 


a 
ea 


7 
7 
- 


- 7 


j 


Lee i Me ° Ss 
@' . vie ° ° ; 
a as 
re % 


i ae 


7 7 ’ O%2 te 32 wt oe ene 
: f ; ; j 
le ae. Five as que ae Bat ade ice te oe ee | 
g ; . 
iar : ci a tro s*S Ty 
OS ee Bd Tae ee Po 4; 4 ' 
7 co adil pat Oa Ape he ‘ ee Me a Ace 


ed 7 Bs , 
2 wae Fra 0, aon — “e 
, Be od ae er ANI, Tisvtt = | 
’ aa al bs pee toe 42 Ree) Steen yet At : #4 


* 


} Rel: thet: 1 eee ee eS 


fist 4 cite ee ee ae 


-« ¢ 


: — - a _ 
9 _—_ > -9 


aS et ie ; 
HOS hath EE fe PA Oe 
et 


CONTENTS. 


CONTENTS OF No. 181, N.S., JULY, 1902. 


MEMOIRS: 
PAGE 
On a Free-swimming Hydroid, Pelagohydra mirabilis, n. gen. 
et sp. By Arruur Deypy, D.8c., F.L.S., Professor of Biology 
in the Canterbury College, University of New Zealand. (With 
Plates 1 and 2) . : : ‘ : : ‘ 1 


Studies in the Retina. Parts III, IV, and V, with Summary. By 
Henry M. Brrnarp, M.A.Cantab. (From the Biological 
Laboratories of the Royal College of Science, London.) (With 
Plates 3—5) : : : - : : = 25 

Notes on the Relations of the Kidneys in Haliotis tuberculata, 
ete. By H. J. Funurs, B.Sc., U.C.W., Aberystwyth. (With 
Plate 6) . : ‘ : : . ; See 


Notes on the Development of Paludina vivipara, with special 
reference to the Urino-genital Organs and Theories of Gasteropod 
Torsion. By Isapetta M. Drummonp. (With Plates 7—9) . 97 


Is Chemotaxis a Factor in the Fertilisation of the Eges of 
Animals? By A. H. Reernatp Buter, B.Sc., Ph.D., Demon- 
strator in Botany at the University of Birmingham : . 145 


CONTENTS OF No. 182, N.S., SEPTEMBER, 1902. 


MEMOIRS: 
Maturation of the Ovum in Echinus esculentus. By Tuomas 
H. Bryce, M.A., M.D. (With Plates 10O—12) . 5 eee 


Studies on the Arachnid Entosternite. By R. I. Pocock. (With 


Plates 13 and 14) 225 


1V CONTENTS. 


PAGE 
On the Morphology of the Cheilostomata. By Sipnzy F. Harmer, 
Se.D., F.R.S. (With Plates 15—18) . . : . 263 
On the Development of Sagitta; with Notes on the Anatomy of 
the Adult. By L. Doncaster. (With Plates 19—21) - gol 
CONTENTS OF No. 183, N.S., DECEMBER, 1902. 
MEMOIRS: 
On a Cestode from Cestracion. By Wittiam A. Haswett, M.A., 
D.Se., F.R.S. (With Plates 22—24) . , . 399 


The Development of Lepidosiren paradoxa.—Part UI. De- 
velopment of the Skin and its Derivatives. By J. Granam 
Kerr. (With Plates 25—28) . : : 3 . 47 

The Metamorphosis of Corystes Cassivelaunus (Pennant). 

By Ropert GURNEY, poe Des F.Z.S. (With Plates 


29—31) . : ; . ; . 461 
Artificial Parthenogenesis and Fertilisation: a Review. By 
Tuomas H. Bryce : : ; ; : . 479 


CONTENTS OF No. 184, N.S., FEBRUARY, 1903. 


MEMOIRS: 
The Movements and Reactions of Fresh-water Planarians; a Study 
in Animal Behaviour. By Raymonp Peart, Ph.D., Instructor 
in Zoology in the Sa of x ee Ann Arbor, Michigan, 
US Ae : 509 
On the Diplochorda. _ Part IV. On the one Gane of 
Cephalodiscus dodecalophus, Mcl. By A.'T. Masrermay, 
M.A., D.Se., Lecturer on Zoology, School of Medicine, Edin- 
burgh. (With Plates 32 and 38) : : ' « Waa 
On Hypurgon Skeati, a new Genus and Species of Compound 
Ascidians. By Icerna B. J. Sotzas, B.Se.Lond. (With Plates 
34 and 35) : : : : Z : . 729 
The Anatomy of Arenicola assimilis, Ehlers, and of a New 
Variety of the Species, with some Observations on the Post- 
larval Stages. By J. H. Asuworru, D.Sc. (With Plates 36 
and 37) . ! : ' ‘ : : : for 


TitLE, INDEX, AND CONTENTS. 


New Series, No. 181 (Vol. 46, Part 1). Price 10s. 
Wey 


JULY, 1902. 


Ae: | 
oy oe 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 
SOCIFTY OF SCIENCES, AND OF THE ACADEMY OF THE LINCEI OF ROME; 
AND OF THE 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 
MBMBER OF THE BIOLOGICAL SOCIETY 
OF PARIS; 
DIRECTOR OF THE NATURAL HISTORY DEPARTMENTS OF THE BRITISH MUSEUM; LATE FULLERIAN 

PROFESSOR OF PHYSIOLOGY IN THE ROYAJ, 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.R.S., 


LINACRE PROFESSOR OF COMPARATIVE ANATOMY AND FELLOW OF MERTON COLLEGE, OXFORD} 
LATE FELLOW OF ST. JOHN’S COLLEGE, CAMBRIDGE 35 


AND 


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


BEYER PROFESSOR OF ZOOLOGY IN fHE OWENS COLLEGE, MANCHESTER. | 
f 


WITH LITHOGRAPHIC PLATES AND ENGRAVINGS ON WOOD. 


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


Adlard and Son,] [Bartholomew Close. 


CONTENTS OF No. 181.—New Series. 


MEMOIRS: 


PAGE 
On a Free-swimming Hydroid, Pelagohydra mirabilis, n. gen. et 
sp. By ArtHuR Drypy, D.Sc., F.L.S., Professor of Biology in the 
Canterbury College, University of New Zealand. (With Plates 
1 and 2) ; : ‘ : : ; 5 ; : : ; 1 
Studies in the Retina. Parts III, 1V, and V, with Summary. By 
Henry M. Bernarp, M.A.Cantab. (From the Biological Labora- 
tories of the Royal College of Science, Iondon). (With Plates 
3—5) ae : 25 
Notes on the Relations of the Kihiore in Haliotis tahneentatee 
etc. By H. J. Frzurs, B.Sc., U.C.W., eee (With 
Plate 6) : : : 77 
Notes on the Development of Palauan vivipara, with fe 
reference to the Urino-genital Organs and Theories of Gasteropod 
Torsion. By Isapetta M. Drummonp. (With Plates 7—9) OF 
Is Chemotaxis a Factor in the Fertilisation of the Eggs of Animals ? 
By A. H. Reeinatp Buiter, B.Sc., Ph.D., Demonstrator in 
Botany at the University of Birmingham : : : : . 145 


On a Free-swimming Hydroid, Pelagohydra 
mirabilis, n. gen. et sp. 
By 


Arthur Dendy, D.Sc., F.L.S., 
Professor of Biology in the Canterbury College, University of New Zealand. 


With Plates 1 and 2. 


ConrTENTS. 
PAGE 
1. Introduction : d ‘ : ; ; 1 
2. Notes on the Living ucwal ; ; ; : : 3 
3. The Hydroid: 
(a) External Characters : : : 2 ; 4 
(b) Internal Anatomy. ‘ : : F ; 5 
(c) Histology . ; : . : : : 7 
4. The Medusoid ; 
(a) Structure. s : : . 5 kG 
(b) Development A . : t sae ls 
5, Discussion of Results, Heletioneling, Giga 5 ‘ eee!) 
6. Diagnosis of New Genus and Family : : ; % yell 
7. Description of Plates ; 5 ‘ : : » 92 


1. INTRODUCTION. 


The remarkable organism which forms the subject of the 
present memoir was picked up by myself on the sandy beach 
at Sumner, a small watering-place near Christchurch, on the 
east coast of the South Island of New Zealand. One evening 
in October last (1901), while walking on the shore, I saw 
lying at my feet a small gelatinous object which had 
evidently just been thrown up by the tide. On placing it in 

you. 46, PART 1.—NEW SERIES, A 


2 ARTHUR DENDY. 


a glass of sea water I soon saw that it was still alive, and 
that it exhibited very unusual features, differing widely from 
any pelagic organism with which I was acquainted. After 
studying it for some time with the aid of a pocket lens I 
took it up to my laboratory at Christchurch, and continued 
my examination of the living animal the same night. Being 
unwilling to risk the attempt to keep it alive until the next 
morning, I then killed it by the addition of osmic acid to the 
sea water, and preserved it in alcohol. It was unfortunate 
that the lateness of the hour prevented me from making a 
more exhaustive examination of the living organism, as 
more light might have been thereby thrown upon its move- 
ments and habits; but it seemed best to try and make sure of 
having it well preserved for minute investigation subse- 
quently, and in this I was fairly successful. The action of 
the osmic acid was, as might have been expected in the case 
of so large an organism, very unequal, some of the more 
superficial parts being much blackened, while the interior 
was apparently not affected at all, and consequently turned 
out to be not in so good a condition for minute histological 
investigation as I could have wished. Had I suspected how 
complicated and remarkable the structure of the interior 
really was I might have thought it best to cut the organism 
in half in order to allow the osmic acid to penetrate, but as 
it was it did not seem to me desirable in any way to mutilate 
the unique specimen at that early stage of the investigation. 
It was very soon obvious that the organism was an 
enormous free-swimming hydroid, from the greater part of 
the surface of which numerous little medusoids were being 
budded off in groups. Being about to pay a visit to 
England, however, I postponed the greater part of the 
investigation until after my arrival, when I resumed the 
work in the zoological laboratory of the Owens College. It 
affords me very great pleasure to express my thanks to 
Professor Hickson and his staff for the kind hospitality 
which I received at their hands, and for the valuable assist- 
ance rendered to me during the progress of my research, 


ON A FREE-SWIMMING HYDROID. 3 


2. Notes on THE Livinc ANIMAL. 


The free-swimming hydroid person of Pelagohydra 
mirabilis (fig. 1) is apparently a pelagic organism. The 
conditions under which it was found, its subsequent be- 
haviour when observed in sea water, and its peculiar organi- 
sation, all point to this conclusion. When placed ina glass 
of sea water in front of a candle (it was too dark to examine 
it by daylight) it floated near the surface with the narrow 
proboscis-like portion of the body, bearing the mouth at its 
extremity, hanging downwards from the much larger balloon- 
like structure, which I propose to call the “float.’? The 
latter, though near the surface, was totally submerged. 
Subsequently, when placed in a tin can for removal to the 
laboratory and kept in the dark, the animal sank to the 
bottom, though still alive. Probably, therefore, it has the 
power of rising and sinking in the water like other pelagic 
organisms, and it may be that it always sinks to some depth 
beneath the surface when it is dark. The general colour of 
the organism was a very pale bluish tint, and it was of 
course translucent. The proboscis, however, was pale pink, 
intensified round the margin of the mouth. The manubria 
of the medusoids were also pink. During life the hydroid 
exhibited some slight power of changing its shape, the float 
being at one time oval (slightly elongated vertically) and 
at another contracted into a sphere, while the proboscis 
exhibited considerable power both of contraction, under 
which condition it became slightly trumpet-shaped at the 
end, and of flexion. When both elongated, as shown in 
fig. 1, the float was nearly an inch in greater diameter and 
the proboscis rather more than half as long as the float. 

The long, slender, tentacular processes of the float occa- 
sionally exhibited spasmodic movements of flexion, like 
gigantic flagella, many of them simultaneously, or nearly so ; 
and from this I am led to conclude that the animal has the 
power of rowing itself through the water by means of these 
organs. 


4, ARTHUR DENDY. 


Whether the medusoids naturally separate from the 
hydroid I cannot say from direct observation. They ex- 
hibited slight twitching movements of contraction, however, 
while still attached to the parent, and the structure of the 
larger ones leaves no doubt that they ultimately become free- 
swimming. Moreover many of them became detached when 
the organismn was killed. 


3. Tue Hyproirp. 


(a2) External Characters.—The body of the hydroid is, 
as compared with the ordinary hydroid type—such as we see, 
for example, in Tubularia,—greatly modified in form and 
structure, and the modification is such as to bring about the 
necessary adaptation to the changed conditions of life. The 
usual stalk is entirely wanting, nor is there the slightest 
indication of its having ever existed. The aboral portion of 
the body is enormously swollen out, and quite evenly rounded 
off at the upper pole, forming the nearly spherical ‘ float.” 
To the lower pole of the float is attached the cylindrical “ pro- 
boscis,’ bearing the mouth at its extremity. ‘The line of 
junction between the float and the proboscis is well marked 
even externally, and corresponds to an even more pronounced 
internal demarcation between the two. 

The float carries numerous long tentacles, which are 
scattered without any definite arrangement at approximately 
equal distances from one another all over its surface. ‘These 
tentacles are cylindrical and bluntly rounded at the extremity, 
never distinctly knobbed. When fully extended they may be 
about as long as the float itself. For the most part they are, 
as usual amongst the Hydrozoa, unbranched, but two or three 
were observed each with a single branch (figs. 3, 5, B.7.), 
this condition being probably abnormal. 

The proboscis is differentiated transversely into two por- 
tions (fig. 2). ‘The upper part bears no tentacles, and 
exhibits an appearance of circular and longitudinal striation, 


ON A FREE-SWIMMING HYDROID. 5 


The lower part, next to the mouth, bears numerous tentacles 
of greatly varying size; these are arranged, not quite 
regularly, in transverse rows or whorls, and decrease in size 
from the uppermost whorl, which contains the largest, 
towards the mouth, around the margin of which the tentacles 
are very minute. There is altogether a good deal of irregu- 
larity about the size and shape of these tentacles, and here 
again one of them was found to be branched (fig. 2, B.T.), 
but a better idea of their form and arrangement will be 
gained from the illustration than from any description which 
I can give. 

Scattered all over the surface of the float, between the 
bases of the tentacles (figs. 4, etc.), are numerous little 
branching processes, which we may term “stolons.” They 
branch quite irregularly, their branches remaining short and 
keeping close to the surface of the hydroid. On these stolons 
are borne groups of very small medusoids in various stages 
of development, from minute buds to fully formed bells 
apparently just ready to separate. 

(b) Internal Anatomy.—The most striking feature of 
the internal anatomy is the presence of two large cavities, 
completely separated from one another by a thin horizontal 
septum, as shown in fig. 5. ‘This septum lies at the level of 
the junction between the proboscis and the float, and is 
slightly arched upwards. A preliminary examination reveals 
the fact that the lower and very much smaller chamber is 
the main gastral cavity, while the upper one is apparently 
excavated in the enormously developed mesogloea between 
the ectoderm and endoderm of the roof of the gastral cavity : 
this second and much larger chamber I propose to call the 
“cavity of the float.” Its real origin will be discussed 
presently. It is not a simple cavity, but is subdivided by 
what I propose to term the “supporting membranes.” 
On the inner surface of the wall of the float there is a net- 
work of canals, which give it a honeycombed appearance. 
These canals are lined by endoderm, and are in reality con- 
tinuations of the gastral cavity, into which they open at their 


6 ARTHUR DENDY. 


lower extremities. I shall speak of them as the “endo- 
dermal canals.” 

The Gastral Cavity.—The main gastral cavity, then, 
occupies only the interior of the proboscis, but is continued 
upwards into the float in the form of endodermal canals. 
The lining membrane of the main gastral cavity is thrown into 
numerous very prominent longitudinal folds, forming ridges 
which project inwards (figs. 5—9, L.G.R.), and whose edges, 
in the contracted specimen, are very sinuous (fig. 6). Ata 
short distance below the septum the gastral cavity widens 
out somewhat, and the ridges almost die away. At the 
junction of the septum with the outer wall of the gastral 
cavity a prominent annular fold projects into the latter 
(figs. 7, 8, A.F.). 

The Endodermal Canals.—Above the fold just men- 
tioned, around the margin of the septum, which is otherwise 
imperforate, lie the openings of the endodermal canals (figs. 
7, 8, Op. End.). From the network which these canals form 
on the inner surface of the wall of the float (figs. 5, 6, 
End. C.) short branches are given off outwards, which run 
into the stolons; but the canals themselves have apparently 
no communication with the tentacles (fig. 8). 

The Septum.—The septum which separates the main 
gastral cavity from the cavity of the float is a thin but firm 
membrane. As already stated, it is somewhat arched 
upwards. Its two surfaces are both smooth, but to the 
upper one are attached some of the supporting membranes 
in the chamber of the float (figs. 5—8). It is, as already 
stated, imperforate, except for the openings of the endo- 
dermal canals, and in this respect differs from either of the 
> in the gigantic Branchiocerianthus 
imperator, which in some respects certainly resembles our 
hydroid.’ 

The Cavity of the Float.—The cavity of the float is 
very spacious, but it is subdivided by numerous very thin, 


r ce = ™ 2 
two ‘diaphragms 


1 Vide Miyajima, ‘Journ. Coll. Sci. Imp. University of Tokio,’ vol. xiii, 
p. 258, ete. 


ON A FREE-SWIMMING HYDROID. ‘ 


transparent, membranous sheets, which radiate outwards 
from a more solid mass of tissue formed by their union nearly 
in the middle of the chamber, and which have their edges 
attached to the inner surface of the wall of the float and to 
the upper surface of the septum. ‘These remarkable struc- 
tures I have called the ‘supporting membranes.” The 
inner surface of the wall of the chamber exhibits a honey- 
combed appearance, being marked out into roundly poly- 
gonal areas by the projecting endodermal canals. In the 
centre of each depressed area between the endodermal 
canals a knob-like projection may frequently be seen; this 
is caused by the tissue which fills the cavity of the tentacle 
projecting inwards into the chamber of the float like a plug 
(figs. 6, 8, Ten. Pl.). These structures we may call the 
“tentacle plugs.” 

The Tentacles.—All the tentacles are filled with a highly 
vacuolated tissue, composed of sheets or strands of delicate 
membrane. In the case of the tentacles of the float this 
tissue may, as Just stated, project as a plug into the float 
cavity. In the proboscis the mesoglcea in the wall of the 
gastral cavity is, in the neighbourhood of the tentacle bases, 
much thickened and highly vacuolated, giving rise to cavities 
of considerable size, and this vacuolated tissue is continued 
into the tentacles (figs. 8,9). The exact nature and origin 
of the tissue which thus fills the interior of all the tentacles 
are, however, by no means easy to determine, and the 
question will be best dealt with under the next heading. 

(c) Histology.—Pelagohydra exhibits, for a hydroid, 
a remarkable amount of histological differentiation. For 
purposes of description it will be most convenient to sub- 
divide this part of our subject according to the different 
regions of the body, rather than to attempt to follow out each 
layer completely before passing on to the next. Indeed, as 
we shall see later, in some parts of the body the delimitation 
of the layers is by no means always obvious—at any rate, in 
the case of the endoderm. 

As already indicated, the histological preservation of the 


8 ARTHUR DENDY. 


internal tissues is not all that could be desired, and it is 
greatly to be hoped that an opportunity may arise of working 
out this subject more in detail with the aid of material 
specially treated for the purpose. It is also highly desirable 
that a detailed comparison should be made of the histological 
structure of Corymorpha, Monocaulus, and Branchio- 
cerianthus, which are evidently related to Pelagohydra, 
and, like it, of exceptional size. 

Wall of the Proboscis.—The ectoderm is a thick layer 
densely charged with small, darkly staining nuclei and 
thread-cells irregularly scattered throughout its substance 
(figs. 9, 10, Het.). 

In section it exhibits numerous fine radial lines running 
in at right angles from its outer surface, and perhaps 
indicating the boundaries of a single layer of large prismatic 
cells. On its inner aspect, immediately contiguous to the 
mesogloea, is a well-developed layer of longitudinal muscle- 
fibres. In transverse sections (fig. 10) we see that this layer, 
consisting of an approximately single row of fibres, is thrown 
into longitudinal folds, the mesogloea being produced out- 
wards in plate-like ridges between the folds. This arrange- 
ment, so well known in the mesenteries of the Actinians, no 
doubt serves to increase the extent of the muscular tissue. 

The ectoderm decreases in thickness from below up- 
wards, and the folding of the muscular layer is especially 
conspicuous just above the region of the tentacles, and dies 
away as it approaches the upper limit of the proboscis. 
Between the bases of the proboscis tentacles the ectoderm is 
extremely thick, but thins out greatly over the tentacles 
themselves. 

The endodermal lining of the proboscis wall is enormously 
thick, and throughout the greater part of its extent is thrown 
into prominent longitudinal folds or ridges in the manner 
already described (figs 5—9). The structure of these ridges 
(figs. 9—11) is very peculiar. The mesogloeal supporting 
lamella which divides the endoderm from the ectoderm is 
uot continued into them, and is indeed sharply marked off by 


ON A FREE-SWIMMING HYDROID. 9 


another layer of muscle-fibres, which we may consider to be 
endodermal in origin. These fibres are arranged in a 
circular manner at right angles to those of the ectoderm (fig. 
10), and the extent of the muscular layer is increased by 
horizontal folds, similar to the vertical folds of the ecto- 
dermal layer. These horizontal folds are, of course, recog- 
nisable only in vertical sections, while the vertical folds of 
the ectodermal musculature are conspicuous in transverse 
section (fig. 10). 

The free surfaces of the gastral ridges bounding the 
gastral cavity are covered with an epithelium of a very 
peculiar type (figs. 9—11). It consists of long, slender, 
columnar cells arranged at right angles to the surface. 
They have a finely granular cytoplasm and distinct nuclei, 
and appear in the sections to be collected into small groups, 
like bundles of cigars, from the inner ends of which delicate 
wavy fibres run obliquely towards the central plane of the 
ridge, and thence inwards side by side till they meet the 
mesogloeal supporting lamella, where they probably give rise 
to the circular musculature. The grouping of the epithelial 
cells into bundles is, I think, probably a post-mortem con- 
dition due to contraction in alcohol. I imagine that the cells 
are normally arranged so that each is continued inwards 
into a separate fibre. We may probably regard the endo- 
derm of the gastral ridges as glandular-muscular in function, 
for no doubt it secretes the digestive fluid. There are no 
thread-cells in the gastral ridges, nor, indeed, have I seen 
them in any part of the endoderm. On approaching the 
annular endodermal fold which marks the upper limit of the 
proboscis the gastral ridges gradually die away, and their 
epithelium gives place to that which lines the gastral face of 
the septum on the one hand, and the endodermal canals on 
the other (fig. 8). 

The mesogloeal supporting lamella of the proboscis wall 
may be regarded as being bounded on the outside by the 


1 Compare the structure of the endodermal villi with their muscle-fibres in 
Myriothela (Hardy, ‘ Quart. Journ. Mier. Sci.,’ vol. xxxil, p. 505). 


10 ARTHUR DENDY. 


ectodermal, and on the inside by the endodermal layer of 
muscle-fibres respectively. It is continued into the folds of 
the muscular layers, and also into the annular fold of endo- 
derm. It has the usual clear gelatinous appearance,! and 
though everywhere more or less distinct, attains its maximum 
development in the neighbourhood of the tentacle bases, 
where it appears to become immensely thickened, and at the 
same time broken up by large vacuoles into a network of 
irregular sheets (figs. 8,9). It may possibly be invaded in 
this region by cells migrating from the endoderm, as will be 
described later in the case of the supporting membranes of 
the float; but this point I have not been able to determine. 
Tentacles of the Proboscis.—The larger tentacles of 
the proboscis are identical in structure with those of the 
float, shown in transverse section in fig. 15. The outer wall 
of the tentacle is formed by a single layer of short columnar 
cells; it is highly vacuolated, and abundantly charged with 
thread-cells in all stages of development ; on its inner face is 
a well-developed single layer of longitudinal muscle-fibres. 
A more or less distinct layer of mesogloea comes next, crossed 
in places by slender strands (of protoplasm ?) extending 
inwards from the ectoderm, while the axis of the tentacle is 
occupied by an irregular network of sheets continuous with 
the vacuolated mesogloea of the proboscis wall. Here and 
there over the surfaces of these thin and apparently structure- 
less sheets are scattered very well-defined bodies, which may 
be either small isolated cells with small nuclei, or, as 1 am 
inclined to think, themselves large nuclei with conspicuous 
nucleoli, These bodies are flattened against, or perhaps in 
the thickness of, the septa which separate the enormous 
vacuoles from one another. When seen en face they are 
nearly round, and about 0°0125 mm. in diameter. Their 
protoplasm stains fairly deeply, especially that of the small 
enclosed body, and is scarcely at all granular. It is note- 
1 Tt seems probable that the fibrillated character of the mesogleea described 


by Allman and Miyajima (loe. cit.) in Branchiocerianthus may be due to 
the ectodermal and endodermal muscle-fibres attached to it. 


ON A FREE-SWIMMING HYDROID. 1] 


worthy that two of these large nuclei may be found lying 
close together, side by side, on the same side of one septum, 
which seems to indicate that each cavity in the axial tissue is 
not simply the enlarged vacuole of a single cell. Though 
abundant in the tentacles themselves, the large nuclei are, so 
far as my experience goes, not to be found in the vacuolated 
mesoglcea with which the axial tissue of the tentacle becomes 
continuous in the proboscis wall. 

Owing partly to the specimen being somewhat injured in 
the neighbourhood of the mouth (possibly by being washed 
about by the tide on the sand, with mouth downwards), I 
have been unable to make a satisfactory investigation of the 
minute structure of the smallest tentacles. It is evident, 
however, that these conform much more closely to the ordi- 
nary Tubularian type than do the large ones. This may be 
chiefly owing to their smaller diameter, which enables the 
membranous septa to stretch right across transversely and 
more or less parallel with one another, so as to divide the 
interior into approximately a single row of chambers, sur- 
rounded by a very thick layer of mesogloea inside the 
ectoderm. ‘Thus it would seem that the axis of the smallest 
tentacles is occupied by a single row of large vacuolated 
endoderm cells as usual. Whether even in the smallest 
tentacles these axial cells retain their connection with the 
endodermal lining of the gastral cavity is extremely doubtful. 
In the case of the large tentacles there is no trace of any 
connection remaining between the axial tissue and the endo- 
derm of the gastral cavity,’ and the origin of this tissue 
must remain doubtful. It has probably been originally 
derived from the endoderm, but it has become so modified in 
structure and so completely disconnected that perhaps only 
embryological research can decide the question. 

Wall of the Float.—The wall of the float forms but a 
comparatively thin shell, enclosing the central cavity with its 
remarkable system of supporting membranes. ‘The histo- 
logical characters of the ectoderm (fig. 12, Hct.) are very 


‘ Compare Miyajima’s remarks on Branchiocerianthus, loe. cit. 


12 ARTHUR DENDY. 


similar to those of the corresponding layer in the wall of the 
proboscis. It is, however, less distinctly muscular. In the 
immediate neighbourhood of the tentacles it retains the 
characters which it exhibits in the tentacles themselves, being 
comparatively thin, and having the muscle-fibres arranged 
radially in continuation with the longitudinal muscular layer 
of the tentacle. Elsewhere the ectoderm is thick and very 
densely crowded with thread-cells. 

The Endodermal Canals.—The lining epithelium of 
the endodermal canals, directly continuous with that of the 
gastral cavity proper, is differentiated into two very distinct 
portions, differing greatly in histological character. The 
canals are somewhat flattened against the wall of the float ; 
their own outer walls form part of the thickness of the latter 
(fig. 12), and are lined by a layer of large epithelial cells with 
rounded club-shaped ends projecting into the lumen. These 
cells have very large vacuoles and small round nuclei, and 
their very darkly staining granular contents are collected 
together in or near their swollen club-shaped ends (fig. 12, 
End O.). They also contain darkly staining spherical globules 
of various sizes. ‘The epithelium forming the inner walls of 
the endodermal canals, on the other hand, consists of a single 
layer of smaller cells, approximately cubical in shape, with 
small nuclei and only a small quantity of faintly staining, 
finely granular cytoplasm (fig. 12, End. I.). 

The Supporting Membranes of the Float.—The thin 
transparent sheets of membrane which subdivide the cavity of 
the float (figs. 5—8, 12, Sup. Mem.) appear to have a 
very remarkable structure and origin. Hach sheet consists 
of a thin structureless layer of mesogloea (fig. 13, Mes.), 
thickening at the angles where the sheets meet one another. 
Spread out on each surface of this mesoglceal sheet is a still 
thinner layer of finely granulated, frothy-looking protoplasm, 
containing rounded nuclei irregularly scattered through it 
(fig. 14). No cell boundaries can be distinguished in my 
preparations, but the protoplasm appears to form a vacuolated 
syncytium. It may occasionally be collected or drawn 


ON A FREE-SWIMMING HYDROID. 13 


together into a thick rounded blob or drop, containing many 
nuclei (fig. 13), but this condition appears to be of rare 
occurrence. Probably the nuclei multiply by division, as 
indicated in fig. 14, at a This peculiar tissue appears to 
originate, in part at any rate, from the inner walls of the 
endodermal canals.! The mesoglceal portion of these walls 
may be very thick, and occasionally little groups of cells (fig. 
12, End. Bud) may be seen growing into it from the endo- 
dermal lining of the canal. These cells have very finely 
granular contents and small nuclei. Irregular cavities (fig. 
12, D. F. C.) are apparently developed between them, 
and gradually enlarge until the nuclei become widely 
separated, while the mesoglea is reduced to thin sheets 
separating adjacent cavities from one another, and the proto- 
plasm of the endoderm cells becomes spread out over these 
sheets in the form of a granular syncytium. 

Sometimes, where a comparatively thin layer of mesogloea 
lies behind the endoderm of the inner wall of an endodermal 
canal, threads of finely granular protoplasm may be seen 
stretching at right angles through the mesogloea from the 
one surface (covered by the finely granular syncytium) to the 
other (covered by the endodermal cells of the canal wall). 

Thus it appears that the supporting membranes of the float 
originate in a peculiar manner from the endoderm. It is not 
certain, however, that they do not receive cells from the 
external ectoderm also, for thread-cells in various stages of 
development may sometimes be observed in places where the 
mesoglcea is thick, beneath the external ectoderm and doubt- 
less derived from the latter. This inward migration of the 
cnidoblasts can hardly be looked upon as normal, but if they 
are able to migrate inwards it seems equally possibly that 
other ectoderm cells may do the same, and possible eventually 
take part in the formation of the supporting membranes. 

1 Professor Ray Lankester has pointed out to me that a somewhat similar 
method of tissue formation has been observed in the ‘‘ laminar tissue’ of 


Amphioxus (vide Pouchet, “On the Laminar Tissue of Amphioxus,” 
‘Quart. Journ, Mier. Sci.,’ vol. xx, n. s., p. 421, pl. xxix), 


14. ARTHUR DENDY. 


The Septum.—The histological structure of the septum 
which divides the main gastral cavity from the cavity of 
the float is practically identical with that of the inner walls 
of the endodermal canals, with which it is directly continuous. 
On its lower face it is covered by a layer of lightly staining 
cells with small nuclei and finely granular contents, and this 
is separated by a moderately thick layer of mesogloea from 
the finely granular syncytium which covers its upper surface. 
Some of the supporting membranes of the float are attached 
to its upper surface, and probably originate from the septum 
in the same way as those already described originate from 
the inner walls of the endodermal canals. 

Tentacles of the Float.—The tentacles of the float are 
histologically identical with the large tentacles of the pro- 
boscis, as will be seen by comparison of fig. 15 with the 
description already given. ‘The peculiar manner in which the 
axial tissue seems to project into the cavity of the float in the 
form of a cushion or plug has already been referred to. In 
the projecting plug, however, when best developed, the 
network of tissue is made up chiefly of a finely granular 
frothy syncytium, with very little mesogloea and small nuclei. 
In the tentacle itself the granular material is hardly recog- 
nisable, the septa (fig. 15, S.W.7.) are very thin, and the 
nuclei (fig. 15, Nw.) much larger and of a different character, 
like those in the proboscis tentacles. Thus the “plug” 
seems to be to some extent transitional in character between 
the true axial tissue of the tentacle and the very much coarser 
reticulation formed by the supporting membranes in the 
interior of the float. It is not always recognisable as a 
distinct structure, however, and even where best developed 
it passes gradually into the axial tentacular tissue beyond, 
while its apparent histological differences may be in part due 
to the want of penetration of the osmic acid with which the 
specimen was hardened. 

The endodermal canals come very close to the bases of the 
tentacles, and we may be pretty certain that the axes of the 
latter are endodermal in origin, though, as in the case of the 


ON A FREE-SWIMMING HYDROID. 15 


proboscis tentacles, embryological research may be required 
before we can say exactly how they arise. 

The Stolons.—The stolons are simply branching hollow 
outgrowths of the wall of the float in the neighbourhood of the 
endodermal canals, which are prolonged into them to their 
extremities (figs. 8, 12, St.). The ectoderm (fig. 12, Het.) is 
composed of the usual large clear cells, rectangular in 
longitudinal section, with small nuclei pressed against their 
dividing walls. At its base lies a feebly developed layer of 
longitudinal muscle-fibres. Thread-cells are almost entirely 
wanting. The mesoglcea is thick, and traversed by slender 
threads crossing from ectoderm to endoderm. ‘The endo- 
derm (fig. 12, Hnd.) is simply a continuation of the endoderm 
which lines the outer walls of the endodermal canals, and, 
like the latter, is composed of large cells, often with rounded 
extremities projecting into the central lumen, with enormous 
vacuoles and darkly staining contents massed together either 
in the rounded end or elsewhere. ‘They have small nuclei, 
and in addition contain darkly staining spherical globules of 
various sizes. 

The Thread-cells.—The thread-cells (figs. 16, 17) are of 
large size. The actual nematocysts or capsules are approxi- 
mately ovoid in shape, but truncated at the somewhat 
narrower outer ends, and measure, when fully developed, 
about 0°0128 mm. in longer diameter. Each one is more or less 
enclosed in a delicate cnidoblast (fig. 17, enb.). When fully 
developed the thread-cells lie in the outer parts of the large 
ectoderm cells just beneath the surface, and the cnidoblast is 
prolonged inwards to the base of the cell in the form of a long 
thread—the cnidopod! (figs. 16, 17, Cup.). The enidopod is 
remarkably distinct and tough, so much so that when the 
ectoderm of a tentacle has been abraded, so that the large 
ectoderm cells have disappeared, the cnidopods may still 
remain projecting from the surface like hairs, with or with- 
out the thread-cells still attached to their extremities. 


1 Compare Allman, ‘Challenger Reports,’ “ Hydroida,” Part 2, p. xv, for 
the use of this term, 


16 ARTHUR DENDY. 


I have seen no thread-cells with the threads everted, and 
have not been able to make out any details with regard to 
the thread itself. No barbs were visible in my preparations. 
Smaller thread-cells, in various stages of development, lie in 
the deeper parts of the ectoderm. 


4, THe MEDUSOID. 


(a) Structure.—Although no free-swimming meduse 
have as yet been observed, there can be little doubt that 
they normally separate from the parent hydroid. As already 
pointed out, they exhibit movements of contraction while 
still attached, and separate very readily in the process of 
killing and preserving. Moreover none of the meduse, 
which were found attached to the hydroid in large numbers, 
were sexually mature, and the largest were only about 1 mm. 
in longer diameter of the bell. 

In the largest examples the bell is considerably deeper 
than wide, and nearly square, though with rounded angles, 
in cross-section (figs. 22—24). The mouth of the bell is 
still very narrow (fig. 23), probably expanding considerably 
later on. It is surrounded by the velum, around which the 
margin of the bell has grown out into four arms or lobes, 
arranged in the form of a cross, per-radially, corresponding 
to the angles of the bell. Each of these arms bears five 
tentacles arranged in a very peculiar manner—a pair of 
larger ones, a pair of smaller ones, and a very small odd one ; 
the largest being furthest from the mouth, the odd nearest to 
the mouth, and the remaining pair intermediate in position, 
as shown in fig. 23. All the tentacles are short, even in 
the living animal, and they are only very slightly if at all 
swollen at their extremities. It is possible that the number 
of tentacles increases as the medusa grows older, but their 
peculiar and definite arrangement seems to indicate that the 
full complement is already present. The tentacles are filled 
with solid endoderm formed in the usual manner, while the 
arms or lobes upon which they are borne are characterised 


ON A FREE-SWIMMING HYDROID. 17 


by an enormous thickening of the ectoderm, containing 
numerous thread-cells. 

At the aboral apex of the bell is a depression, where the 
exumbrellar ectoderm dips in to meet an outward extension 
of the endodermal lining of the gastral cavity. This marks 
the spot where the young medusa is attached to the stolon 
(fig. 22, Z.). 

The manubrium (figs. 22, 25, Man.) is large, but does not 
project beyond the mouth of the bell. Its surface is smooth, 
and there are no outgrowths at its extremity. 

The subumbrellar cavity is, in the middle, somewhat 
octagonal in transverse section (fig. 25), being produced into 
four shallow per-radial angles where the ectoderm is attached 
to the radial canals, and four deeper interradial angles where 
it is attached to the endodermal lamella. Immediately 
beneath the subumbrellar ectoderm cells is a layer of trans- 
verse (“circular”) muscular fibres, and the entire epithelium 
with its musculature is thrown into transverse folds, as shown 
in figs. 22—24. Towards the mouth of the bell the cross- 
section of the subumbrellar cavity becomes square, the inter- 
radial angles alone remaining. 

The gastral cavity immediately above the manubrium is 
cruciate in transverse section, the four arms of the cross 
being produced outwards into the radial canals, and the 
endoderm being greatly thickened between them to form four 
ridges. Inthe manubrium itself the gastral cavity is squarish 
or irregular in section, with a variable number of longitudinal 
endodermal ridges. 

The four radial canals present no features of special interest, 
nor does the thin endodermal lamella by which they remain 
connected. Near the margin of the bell they open into the 
circular canal (fig. 22, c. can.), enlarged per-radially in the 
tentacle-bearing arms and then produced to form the solid 
axes of the tentacles. 

No gonads are yet recognisable, but the ectoderm of the 
manubrium exhibits a thickening all round about the middle 

voL. 46, PART 1.—NEW SERIES. B 


18 ARTHUR DENDY. 


of its length, which probably indicates the position in which 
they will subsequently appear. 

There appear to be no sense-organs, and I have not satisfied 
myself as to the existence of a nerve-ring. In life there isa 
pink spot on the outside of the base of each tentacle group, 
and the manubrium also is more or less pink in colour. 

(6) Development.—The medusz are developed as hollow 
outgrowths or “buds” from the branching stolons already 
described, and each stolon may bear as many as half a dozen 
at the same time in various stages of development. As soon 
as one medusoid approaches maturity another bud (fig. 20 a) 
appears on the stolon close to its point of attachment, ready 
to replace the first when it falls off. 

The youngest buds observed are represented in figs. 20 4 
and 188; each is a single hollow outgrowth of the stolon, 
composed of ectoderm and endoderm, but the thick mesoglea 
of the stolon disappears almost if not quite completely in the 
bud (fig. 20). The ectoderm and endoderm also change 
their character, becoming much more compact and solid- 
looking, and staining much more darkly. 

In the next stage (fig. 18 c) the endocodon is formed from 
the ectoderm at the apex of the bud. There is, in the 
section represented in the figure, some appearance of invagi- 
nation, but if not at first solid the endocodon speedily 
becomes so. 

The endocodon grows inwards, and at the same time the 
endoderm invaginates as if pushed before it (figs. 18, 19), 
forming a deep cup. The bottom of this cup is then pushed 
outwards again through the endocodon to form the hollow, 
finger-like manubrium, which makes its appearance very 
early (fig. 20). 

Meanwhile the cells of the endocodon arrange themselves 
in a single layer over the outer surface of the manubrium, 
the inner surface of the future subumbrella, and the inner 
surface of the future velum (fig. 20). These layers are at 
first in close contact, but ultimately the subumbrellar cavity 
inakes its appearance between them. 


ON A FREE-SWIMMING HYDROID. 19 


While these changes have been going on the original 
gastral cavity of the bud becomes further subdivided by 
the union of its inner and outer walls interradially (fig. 21) 
to form the solid endodermal lamella, thus defining the four 
radial canals and the circular canal. The ectoderm becomes 
greatly thickened outside the circular canal, and the tentacles 
begin to grow out. 

Hitherto ectoderm and endoderm have everywhere re- 
mained in close contact (figs. 20, 21), but the transparent 
gelatous mesoglcea now appears and forces the layers apart 
(fig. 25). About the same time the subumbrellar cavity is 
developed and the velum is ruptured in the middle (fig. 
20, w.), giving rise to the mouth of the bell (fig. 23). 


5. Discussion or Resuits, RELATIONSHIPS, ETC. 


Pelagohydra mirabilis is a remarkably interesting 
organism from several points of view. In the first place it 
forms an excellent example of adaptation to changed con- 
ditions of life, showing us how a representative of a group 
whose members are normally attached, in the hydroid phase, 
to the ends of fixed stalks may become adapted to a free- 
swimming pelagic existence. In the second place it exhibits 
remarkable structural features, especially in the compli- 
cation of the gastral cavity with its endodermal canals, 
and the development of the float with its extraordinary 
supporting membranes. It also has very striking histo- 
logical peculiarities, showing in this respect a degree of 
differentiation perhaps unequalled in any other hydroid.! 

As a pelagic member of a typically non-pelagic group of 
animals we may compare it with Pelagonemertes amongst 
the Nemertines, 'Tomopteris amongst the Annelids, and 
Pelagothuria amongst the Holothurians, and it may 


1 The gigantic Branchiocerianthus imperator probably resembles 
Pelagohydra closely in histological features, but requires further investiga- 
tion (vide Miyajima, ‘ Journ. Coll. Sci. Imp. University of Tokio,’ vol. xiii, 
p. 235, ete.). 


20 ARTHUR DENDY. 


possibly throw some light upon the origin of that remarkable 
pelagic group of Hydrozoa the Siphonophora, although it 
will perhaps hardly bear close comparison with any known 
member of that order. 

That it is an aberrant Tubularian hydroid there can, I think, 
be no doubt, and its nearest relations appear to be the 
enigmatical Corymorpha and its allies! In the genus 
Corymorpha we also find that there is no true stalk, and 
the curious prolongation of the body by which the animal 
fixes itself in the sand or mud is, I believe, homologous with 
what I have termed the float in Pelagohydra. In Cory- 
morpha also we have a system of endodermal canals forming 
a network around a spongy central mass, and communicating 
at one end with the main gastral cavity. Then, again, in Cory- 
morpha curious processes are given off from the surface of 
the body in the neighbourhood of the endodermal canals, 
which may be homologous with the stolons of Pelago- 
hydra, or possibly with the tentacles of the float. Little is 
known, however, of the minute anatomy and histology of 
Corymorpha, and a careful investigation in comparison 
with Pelagohydra is greatly to be desired. There are, 
of course, sufficiently striking differences between the two 
forms, but these are of a more superficial character, and 
mainly to be accounted for by the difference in mode of life. 
Instead of a float we find in Corymorpha a kind of rooting 
process, and the tentacles are confined to one end of the 
elongated body, where they are arranged in a proximal and a 
distal set, the latter obviously representing the tentacles of 
the proboscis in Pelagohydra. The position of the stolons, 
between the two sets of tentacles, is totally different; and the 
medusz also are quite distinct, for in Steenstrupia, the 
medusa of Corymorpha, we find a single odd tentacle, 
representing one only of the four tentacle groups of the 
Corymorpha medusa. In both cases, however, the meduse 
are markedly quadriradiate, and essentially similar in in- 
ternal organisation; while in Amalthza, which appears to 

1 Allman, ‘ Tubularian Hydroids,’ p. 386, ete. 


ON A FREE-SWIMMING HYDROID. 21 


be closely related to Corymorpha, all four tentacles are 
developed. 

It is a very curious fact that two distinct genera of 
Tubularian hydroids agreeing in such striking anatomical 
peculiarities should have become adapted to two such differ- 
ent modes of life, the one swimming freely in the open ocean, 
and the other rooting itself in the sand at the bottom. It 
would indeed be difficult to find a better example of the 
powers of adaptation to divers conditions of life. So far as 
I am aware there is no other hydroid yet known which has 
become specially adapted to a pelagic mode of life. It is 
true that floating hydranths—Acaulis and Nemopsis— 
are known, but these have probably become detached from 
stalks, and are not structurally adapted to a free-swimming 
existence. 


6. Diagnosis or New Genus AND FAMILY. 


Genus Pelagohydra, n. gen.—Hydroid solitary, free- 
swimming; the proximal portion of the body modified to 
form a float, supported internally by a system of radiating 
membranes of endodermal origin ; the distal portion forming a 
flexible proboscis, with the mouth at its extremity. Gastral 
cavity continued from the proboscis into the float in the form 
of endodermal canals, from which arise branching stolons. 
Tentacles filiform, scattered over the surface of the float and 
in whorls around the mouth. Meduse developed on stolons 
between the tentacles of the float; quadriradiate, symmetri- 
eal, probably with gonads in the wall of the simple manu- 
brium ; tentacles in four per-radial groups of five (possibly 
more in the adult). 

The genus may be regarded as belonging to a distinct 
family, for which I propose the name Pelagohydridea, and 
for which the generic diagnosis may at present suffice. This 
family is, however, closely related to the “Corymorphinz” 
of Delage and Herouard;’ indeed, some zoologists might 

1 «Traité de Zoologie concréte :’ “ Les Ceelenterés,” p. 88, 


22 ARTHUR DENDY. 


prefer to modify and extend their conception of the Cory- 
morphine so as to include Pelagohydra (as the authors 
referred to include the Hybocodonide and Monocaulidez 
of Allman) in preference to making a new family for its 
reception. 


7. DESCRIPTION OF PLATES 1 & 2. 


Illustrating Professor Dendy’s memoir on “ Pelagohydra 
mirabilis.” 


EXPLANATION OF LETTERING. 


A. F. Annular fold of endoderm around the margin of the septum. B. 7. 
Branched tentacles. C.Can. Circular canal. Cnb. Cnidoblast. Cnp. 
Cnidopod. D. F.C. Developing float cavities. Hen. Endocodon of medusa 
bud. Hct. Ectoderm. End. Endoderm. End. Bud. Buds of endoderm 
growing into the mesoglcea from the inner walls of the endodermal canals. 
End. C. Endodermal canal. End. I. Endoderm of inner wall of endodermal 
canal. Hnd.Z. Endodermal lamella of medusa. Hnd.O. Kndoderm of 
outer wall of endodermal canal. EH. U. EH. Exumbrellar epithelium of medusa, 
Fl. Float. G.C. Man. Gastral cavity in manubrium. JZ. G. BR. Longi- 
tudinal gastral ridges of endoderm. Man. Manubrium. Med. Meduse in 
various stages of development. Mes. Mesoglea. M. F. Ect. Ectodermal 
muscle-fibres. MM. F. End. Endodermal muscle-fibres. Mo. Mouth. Nu. 
Nucleus. Op. Hnd. Openings of endodermal canals into gastral cavity. 
Pr. Proboscis. R.Can. Radial canals. Sep. Septum between the main 
gastral cavity and the cavity of the float. S.M. 7. Internal supporting 
membranes of the tentacles. Sé. Stolons. §.U.C. Subumbrellar cavity. 
S.U. EH. Subumbrellar epithelium of the medusa. S.U.M. Subumbrellar 
muscular layer of the medusa. Swp. Mem. Supporting membranes of the 
float. Syn. Vacuolated syncytium covering the supporting membranes of 
the float. Z.C. Thread-cells. Ten. Fl. Tentacles of float. Ten. Pr. 
Tentacles of proboscis. Th.A. Thin area of wall of float around tentacle 
base. w. The point where the ectoderm of the young medusa ruptures to 
form the opening in the velum. @. Nucleus in syncytium apparently 
dividing. y. Point of attachment of subumbrellar epithelio-muscular layer to 
endodermal lamella. z. The place where the medusa was attached to the 
stolon, 


ON A FREE-SWIMMING HYDROID. 23 


Figs. 1—17 inclusive refer to the hydroid stage of Pelagohydra 
mirabilis; Figs. 18—25 inclusive refer to the medusoid stage of the same. 

Fic. 1.—The free-swimming hydroid, from a sketch of the living animal. 
x 2. 

Fic. 2,—External view of a piece cut out of the preserved specimen, 
showing the arrangement of the proboscis tentacles, etc. x 7. 

Fic. 3.—Three adjacent tentacles of the float, showing variation in shape, 
from the preserved specimen. 

Fic. 4.—Portion of the surface of the float, much enlarged, showing the 
stolons with the developing meduse, lying between the bases of the tentacles. 

Fic. 5.—The preserved specimen after removal of a portion of the wall, 
showing the gastral cavity, septum, float cavity, supporting membranes of 
float, endodermal canals, ete. x 4. 


Fre. 6.—Internal view of the piece represented in Fig. 2, showing septum, 
longitudinal gastral ridges, endodermal canals, etc. x 7. 


‘te. 7.—Portion of the same turned so as to show the under surface of 
the septum, with the annular fold of endoderm and the openings of the 
endodermal canals into the main gastral cavity. x 7. 


Fie. 8.—Diagrammatic longitudinal section through a portion of the wall, 
showing the relations of the internal cavities, septum, endodermal canals, 
supporting membranes, tentacles, stolon, medusa buds, ete. 


Fic. 9.—Part of a transverse section of the wall of the proboscis, through 
the bases of the larger tentacles and the longitudinal gastral ridges of the 
endoderm. Drawn under Zeiss objective A, oc. 2, camera outlines. 


Fic. 10.—Portion of a transverse section similar to and near the last, to 
show especially the arrangement of the muscle-fibres. Drawn under Zeiss 
objective D, oc. 2, camera outlines. 


Fic. 11.—Portion of a transverse section of one of the longitudinal gastral 
ridges, showing the endodermal epithelial cells continued into muscle-fibres. 
Drawn under Zeiss objective F, oc. 2. 


Fic. 12.—Part of a transverse section through the wall of the float, 
showing an endodermal canal continued outwards into a stolon, and giving 
rise to supporting membranes of the float by means of groups of cells budded 
off from its lining epithelium. Drawn under Zeiss objective C, oc. 2, camera 
outlines (slightly diagrammatic). 

Fic. 13.—Part of a transverse section of a supporting membrane from the 
interior of the float, showing the mesoglceal layer covered on each side by a 
syncytium, here collected on one side into a rounded multinucleate mass of 
protoplasm. Drawn under Zeiss objective F, oc. 2, camera outlines. 

Fie. 14.—Surface view of one of the supporting membranes of the float, 


24. ARTHUR DENDY. 


showing syncytium and nuclei. Drawn under Zeiss objective F, oc. 2, camera 
outlines. 

Fic. 15.—Part of a transverse section of a tentacle from the float. Drawn 
under Zeiss objective D, oc. 3, camera outlines. 


Fic. 16.—Part of the ectoderm layer from a section similar to the last. 
Drawn under Zeiss objective F, oc. 2, camera outlines. 


Fic. 17.—Two thread-cells with their cnidoblasts and cnidopods, from one 
of the tentacles of the float. Drawn under Zeiss objective F, oc. 2. 


(In Figs. 18—21 inclusive, showing stages in the development of the 
meduse, the histology is, for the sake of clearness, rendered diagrammatic- 
ally ; the endoderm is shaded; the external ectoderm is unshaded, and the 
ectoderm of the endocodon and its derivatives is unshaded but has the nuclei 
represented by dots. All are drawn, with the aid of the camera lucida, under 
Zeiss objective D, oc. 2.) 

Fic. 18.—Two young medusa buds seen in longitudinal section,—B before 
the formation of the endocodon; C with the endocodon and manubrium 
developing. (Owing to slight obliquity of the sections, the cavity of the 
stolon is not shown.) 

Fic. 19.—Slightly older medusa bud in longitudinal section. 


Fic. 20.—Still older medusa bud in longitudinal section, with a very young 
bud also springing from the same stolon at A. 

Fie. 21.—Transverse section of a medusa bud a little older than the last, 
showing the radial canals, ete. 

Fig. 22.—Side view of one of the oldest meduse found. Drawn from 
spirit specimen under Zeiss objective A, oc. 1, as a transparent object. 

Fie. 23.—Oral view of similar specimen under similar conditions. The 
mouth of the bell is now visible in the middle of the velum, between the four 
tentacle-bearing arms. 

Fic. 24.—Aboral view of similar specimen under similar conditions, 
showing the four radial canals, subumbrellar musculature, etc. 

Fre. 25.—Transverse section of a medusa of about the same age. Drawn 
under Zeiss objective A, oc. 3, camera outlines. 


Norz.—The microscopical sections were all stained with borax carmine. 


STUDIES IN THE RETINA. 25 


Studies in the Retina. 


Parts III, IV, and V, with Summary.! 


By 


fienry M. Bernard, M.A.Cantab. 
(Irom the Biological Laboratories of the Royal College of Science, London.) 


With Plates 3—5. 


Parr Tt. 
The Migration of the Retinal Nuclei. 


In this third part I had hoped to have dealt further with 
the material absorbed by the rods from the pigmented 
epithelium ; two important points, however, demand im- 
mediate attention. In Part I, in referring to the migration 
of the nuclei, I slightly misquoted Borysiekiewitz’s observa- 
tions, and in Part II I left a serious gap in the description of 
the outer ends of the developing rods. They were shown in 
the figures (e. g. Pl. 31, fig. 29) as if truncated, just, indeed, 
as they appeared in the sections. This gap I am now in a 
position to fill (see Part IV), while Part V will describe the 
fate of the absorbed pigment. 

Referring to the migration of the nuclei from the middle 
nuclear to the outer nuclear layer in Part I (p. 44), astonish- 
ment was expressed that it had not been noticed before. 
The only observer who, so far as I am aware, had called 

1 For Parts I and IL see this Journat, vol. xliii, 1900, p. 23, and vol. xliv, 
1901, p. 443. 

VoL. 46, part 1.—New SERIES. Bie 


26 H. M. BERNARD. 


attention to the phenomenon is Borysiekiewitz.| This writer 
recorded two evidences of migration (‘“ Ortswechsel”) of 
nuclei in the human retina. Nuclei wandered outwards—(1) 
from the outer nuclear layer into the basal limbs of the cones, 
an observation which was not new; and (2) from the middle 
nuclear layer through the outer reticular layer. I then added 
that in this latter case “it was the characters of the migrated 
nuclei, exactly like those of the layer they had left, and not 
at all like those of the layer into which they had moved, 
which convinced him that migration must have taken place.” 
The similarity of these nuclei had been so often noticed by 
myself as a convincing proof that the muclei embedded in the 
outer reticular layer were passing through it, to become 
transformed into rod nuclei, that after reading Borysiekiewitz’s 
two treatises, and finding that he had alse noticed the 
migration, I inadvertently attributed to him an observation 
which, however, he does not make. He only indirectly 
indicates it in his quotation from Dogiel, who recognised a 
layer of “ subepithelial nerve-cells”’? in the outer nuclear 
layer, i. e. a layer of “cells” exactly similar to those on the 
other side of the outer reticular layer, called by Dogiel’ 
“the bipolar cells of the ganglion retine.”” Borysiekiewitz 
remarks that such cells are probably merely his migrated 
nuclei,® but rightly adds that they do not form a “ layer.” 

Borysiekiewitz was himself convinced of the migration of 
the nuclei by finding a tract (I. c., p. 37) in one of his pre- 
parations in which the middle nuclear layer changed from two 
rows into one row and then back again; but where it was in 
a single row, the missing nuclei were visible either in, or on 
the outer side of, the reticular layer (for a parallel case see 
fig. 19, with description). This valuable observation shows 

1 «Weitere Untersuchungen tiber den feineren Bau der Netzhaut,’ Wien, 
1894. 

2 * Archiv f. mikro. Anat.,’ 38, p. 317. 

8 Borysiekiewitz uses the word “ Korn,” which does not exactly mean 
nucleus, but in this connection it is practically the nuclei alone about which 
anything ean be definitely stated. The point will be dealt with in my next 
paper. 


STUDIES IN THE RETINA. 27 


that the migration of the nuclei from the middle nuclear 
layer into the outer, which I described for the Amphibia, occurs 
also in the human retina ; and indeed, I may add, it occurs in 
all the vertebrate eyes I have yet examined. 
Borysiekiewitz’s own theory of the essential structure of 
the retina, in the light of which this migration finds no ex- 
planation, is very different from mine. According to him, 
these nuclei are inside the “ Miiller’s fibres,” in the more fluid 
axial portions of which they can move. The outermost ends 
of these radial fibres are, according to his view, the rods and 
cones. So that the migration of the nuclei beyond the outer 
reticular layer is a kindred phenomenon with their movement 
beyond the membrana limitans externa into the basal limbs 
of the “ cones,” both being mere shiftings outward along the 
axes of the “ Miiller’s fibres.” The comparative study of the 
“ Miiller’s fibres,” which will be found in Part V, makes the 
acceptance of this description impossible. I may add that 
the only difference which Borysiekiewitz can see between the 
rods and cones of the human retina, is that the latter are 
those tips of the Miiller’s fibres into which nuclei have 
migrated beyond the membrana limitans externa. There is 
no observable difference in the lengths of their outer limbs. 
Confining ourselves to the migration of the nuclei, we may 
review the position of the argument as far as it was advanced 
in PartsTand II. Inthe Amphibia migration from the middle 
to the outer nuclear layer can not only be seen—(1) in the 
actual passage of nuclei through the outer reticular layer 
(Part I, Pl. 3, fig. 5, e, f, and Part IT, Pl. 31, figs. 23, 24, 25), 
and (2) in the exact similarity of those in the outer nuclear 
layer which are not yet rod nuclei, but still close up against 
the outer reticular layer with certain nuclei in the outermost 
edge of the middle nuclear layer (see Part IJ, Pl. 30, 
fig. 16, b), but is a necessary assumption in order to account 
for the number of new rods required by the growing retina. 
A short migration within the outer nuclear layer can be 
seen in the fact that the “cone” nuclei, as the cones assume 
the definitive rod-form, move outwards from near the outer 


28 H. M. BERNARD. 


reticular layer towards the membrana limitans externa. The 
only possible escapes from the assumption that when the 
original supply in the outer nuclear layer has been exhausted, 
fresh supplies migrate outwards from the middle nuclear 
layer are two: (1) if it could be shown that the layer of rods 
and cones with their nuclei grows only at the edges ; and (2) 
if it could be shown that the nuclei for the new rods are pro- 
duced by the division of those already composing. the outer 
nuclear layer. 

With regard to the former of these alternatives, it is 
certainly true that the retina as a whole does grow mainly at 
the edges. I am not, indeed, now inclined to lay very much 
stress upon the argument used in Part I, that if growth 
took place only at the sides, the eye would not keep its 
shape, for growth at the sides alone would, I thought, 
merely carry up those sides, and the eyes would be funnel- 
shaped rather than round cup-shaped. This argument would 
perhaps hold if no other facters were present which could 
help to keep the retina hemispherical. There is, however, 
another traceable factor, the full force of which I did not 
then see. I refer to the vitreous humour which, as a collec- 
tion of semi-fluid matter in the hollow of the eye, would, if 
the supply is kept up at any pressure, compel its flexible 
walls to adopt the normal shape. But though this is a 
possible factor in keeping the growing retina round, the 
argument which refers the persistence of its shape to its own 
orowth-processes can hardly be put altogether on one side, 
for it is a fact that cones, 1. e. new rods, can be seen forming 
over the whole Amphibian retina at all stages of its 
crowth, and even in the eyes of adults. 

Secondly, the suggestion that the nuclei for new rods 
might be supplied by the division of those already present 
can be met by a decided negative. In embryonic eyes 
(Mammalia), or in amphibian eyes before they are functional 
(see fig. 5), 1. e. before any rods are formed, and only the 
merest traces of vesicular protrusions are to be seen, 
divisions of nuclei occur over the whole retina in the outer- 


STUDIES IN THE RETINA. 29 


most layers. But as soon as and wherever vesicles are pro- 
duced and rods begin to be developed out of them, a 
process which always takes place first in the centre of the 
retina and spreads outward from the centre, no divisions 
normally take place. In order to ascertain this point, I 
have examined the retinas of tadpoles (toads and frogs) 
killed at almost all hours of day and night.! Nuclear 
divisions were very numerous in the tadpoles killed in the 
night, and sometimes in those killed in the daytime. A 
study of them makes it quite safe to affirm that nuclear 
division is normally confined to the edges of the retina, 
that is, to those parts where there are no traces, or 
only the faintest traces, of vesicular protrusions, although 
one may just occasionally be seen dividing a short way within 
the zone where the vesicular protrusions are beginning.” This 
result is obtained from so many amphibian retinas that I 
have no hesitation in affirming that after the rods have begun 
to develop, nuclear divisions are never found normally in the 
layer of rod nuclei. This is, of course, what we should have 
theoretically expected, that cells specialised for some active 
function are incapable of mitotic division. 

We are, then, debarred from finding the source of the nuclei 
for new rods in the nuclear layer itself. Hence the new nuclei 
required must come into the outer nuclear layer from without, 
i.e. from the middle nuclear layer, by migration through 
the outer reticular layer ; and these migrating nuclei, whether 
the retina grows mainly at the edges or not, must be many 
thousands, considering the great numbers of “‘ cones” found 
in the central regions in all stages of its growth. 

This, then, brings us face to face with the question, Whence 
does the middle nuclear layer obtain the large supply 
necessary to furnish the outer nuclear layer with so many? 
No one will suggest that the supply could be kept up from 
the “ ganglionic cell” layer, which in the central regions is 


1 Viz. at almost every hour of the night, from 4 p.m. to 6 a.m. 
2 They are occasionally found in young fish retinas, even within the already 
functioning central region ! 


30 Hi. M. BERNARD. 


seldom more than one deep. It is true that nuclei from this 
layer do pass through the thick inner reticular layer to the 
middle nuclear layer. Most sections will show, as has been 
often noted before and variously interpreted, one or two 
actually within the inner reticular layer. Further, all sections 
show a number of nuclei in the innermost rows of the middle 
nuclear layer so like the “ ganglionic cells” that they have 
been recently freely claimed as being ‘ ganglionic,” i. e. 
as of much the same functional activity as the nuclei of the 
innermost layer, which has always been the “ ganglionic 
cell” layer of authors. Although this resemblance need not 
necessarily have anything to do with the question of migra- 
tion, there cannot, to my mind, be any doubt but that the 
“ganglionic cell” layer is drawn upon by the middle nuclear 
layer, and may, indeed, for considerable tracts, be quite ex- 
hausted (compare fig. 22, gl. ina, 6, andc). But such 
a supply, at its best, would be insufficient to counterbalance 
the drain on the middle nuclear layer. Further, as in the 
case of the layer of rod nuclei, no mitotic divisions are found 
in the “ ganglionic cell” layer after the eye has once become 
functional. 

We have therefore to seek elsewhere for the supply of 
nuclei required by the middle layer to enable it to send so 
many outwards through the outer reticular layer to become 
the nuclei of the new rods. One would think the most 
probable source for these nuclei would be the division of 
those already composing the layer, but here again we are 
baffled, for divisions do not occur, at least near or in the 
places where they are wanted. Indeed, the primary object I 
had in view in examining retinas killed at all hours of the night 
was to ascertain whether it was not possible that, as no 
divisions were ever seen in this layer in retinas killed during 


1 [have seen a few traces of fragmentation which deserve attention, but 
hardly wide-spread enough to meet the present difficulty. See also Borysie- 
kiewitz’s figures (‘Untersuchungen tiber den feineren Bau der Netzhaut,’ 
p. 19, 1887), which represent “twin ganglion cells.” They certainly suggest 
divisions of these cells, but are capable of a different interpretation. 


STUDIES IN THE RETINA. 31 


the day, they might take place during the night, when 
the eye is at rest. This, however, as above stated, proves 
not to be the case ; no divisions occur in the middle nuclear 
layer, except near the edges of the retina, where it is not 
possible to speak of the middle layer because the two reticular 
layers which separate the retinal nuclei into zones only begin 
where the rods and cones are themselves commencing to 
form. 

We have, then, no other source, except this undifferentiated 
rim, for the enormous number of nuclei required by the middle 
nuclear layer in order to keep up the supply of rod nuclei 
required by the growing retina. A few, one here and there, 
as we have seen, may be obtained from the innermost, 
or the “ganglionic cell” layer, but none from divisions of 
those already present. The real supply must come, as stated, 
from the rim of the retina; and however startling the idea may 
at first appear, we have to assume a stream of nuclei from 
the undifferentiated edges of the retina towards 
the base of the cup. Further, as long as growth lasts, 
this streaming must be considerable, for in addition to the 
supply of nuclei for the formation of new rods, the thickness 
of the middle layer is kept up,' even though the layer itself 
has to expamd greatly as the eye grows larger. Indeed, it 
has not only to extend as the eye grows, but, as compared 
with the bulk of the layer in very young eyes, it may also 
greatly thicken. Sections of small retinas (of tadpoles) about 
0°5 mm. in diameter may show the middle layer in the central 
region only three nuclei deep, while eyes over | mm. in diameter 
may show it six nuclei deep; in the adult frog a layer four 
deep is very common. But I do not think that much import- 
ance can be laid upon these variations in thickness, as they 
are probably accidents of nourishment and growth. It is 
quite possible that at times the supply of fresh nuclei may be 
greater than the immediate demand, in which case the layer 
would temporarily thicken; or in times of bad nourishment 


1 Apparently in all eyes, except in the “‘ fovea centralis” of human and 
ape retinas. 


32 H. M. BERNARD. 


the supply might be less than the demand, and the layer in 
consequence thin away. The really important fact is that we 
find ourselves compelled to assume a migration of nuclei 
ona very large scale. Not only can it be shown that nuclei 
travel outwards through the outer reticular layer in great 
numbers to become rod nuclei, but that all the nuclei destined 
for this function have, at least after the original supply has 
been used up, to travel down from the edge of the retina 
along the middle nuclear layer to their ultimate destinations. 
In addition to these movements it can be shown that 
nuclei of the so-called “ ganglionic cell” layer occasionaily 
travel outwards through the thick inner reticular layer until 
in old eyes (g.l., fig. 21) they may be almost entirely used up. 

Our investigations into the growth-processes of the retinas 
of some score of frog- and toad-tadpoles having thus 
eliminated all other possible sources for the nuclei of new 
rods required by the central regions of the retina except this 
immigration from the rim, it remains for us to see what direct 
or indirect evidence there is for such an unexpected 
phenomenon, not as an occasional, but as a normal growth- 
process. It is hardly likely that such a movement could take 
place without showing visible traces,—without leaving its 
mark on the tectonics of the retina itself. We shall now see 
that this surmise is fully justified. 

The divisions take place in early growth-stages along the 
whole of the rim into the iris, but are most numerous in the 
angle between the iris and the cup of the retina. ‘To this 
angle, as the iris becomes differentiated, they are usually con- 
fined. They also take place chiefly, though not exclusively, in 
the outermost row of nuclei, in what I have elsewhere called 
the palisade layer. Inthe part where the divisions are active 
it is common to find the large, radially arranged, more or 
less spindle-shaped nuclei attached either to the internal or 
to the external limiting membrane by a frequently thick stain- 
ing cytoplasmic strand. The nuclei are usually so numerous as 
to obscure the sections, so that one cannot state that these 
strands, each with its suspended nucleus, run distinct and 


STUDIES IN THE RETINA. 33 


isolated from membrane to membrane. All that is really 
important at the present moment is to note that the nuclei in 
the region of active division are attached to one or other of 
the “ limiting membranes ”’ by definite strands which are only 
found in this undifferentiated rim. 

Now these strands can often be seen showing the following 
interesting arrangement :—On the axial side of this area of 
nuclear division, and just where the differentiation of the 
retina into zones is commencing, the nuclei, still for the most 
part having retained their spindle shapes, are seen to be 
arranged in slight curves (figs. 1 and 2) ; the two ends of the 
curves are attached by these strands to the limiting mem- 
branes, and their middle parts bulge outwards towards the axis 
of the eye. This curving might easily be passed over, and 
when seen it might be considered as a purely accidental phe- 
nomenon. It is far more probable, however, that it is normal, 
and due to the process we are discussing, viz. the tendency of 
the nuclei to travel from the rim towards the functional 
axial region of the retina. It is clear that the curving could 
be so explained. 

Again, comparisons of the different thicknesses of the middle 
nuclear layer at different parts of the retina and at different 
stages in its growth tell the same story of movement. We 
always find that the layer is thickest near the rim where the 
nuclei produced by division are crowding into it, and thinnest 
in and near the centre where the nuclei are presumably in 
most demand. Further, the variations in thickness of this 
layer in the central region at different stages of growth 
clearly show fluctuations in the numbers and changes in the 
positions of its component nuclei. 

If the nuclei from the undifferentiated rim have, then, this 
tendency to stream inwards towards the axis, it is clearly 
those occupying the middle ranks in the retina which wonld 
be the freest to move, and which therefore would travel 
fastest. Those of the innermost ranks will be more firmly 
attached to the internal limiting membrane, and may, 
perhaps, be further entangled by the developing nerve-layer, 

voL. 46, part 1.—NEW SERIES. C 


34 H. M. BERNARD. 


while externally the nuclei are for the most part functioning 
as rod nuclei. It would thus be only a band of nuclei dowr 
the middle which would be freest to travel towards the axis. 
This fact gives us a clue to the origin of the zonal arrange- 
ment of the retina into alternating nuclear and reticular layers. 

It is fairly clear that if a band of nuclei travelled along 
between two stationary layers such as the innermost and 
outermost layers in an amphibian retina, and if, when they 
started, they had cytoplasmic attachments to the limiting 
membranes, they would almost certainly leave traces of those 
attachments trailed along on each side of the stream, and the 
accumulations of the trailings would separate them from 
the stationary fringing layers. We get, indeed, in this 
somewhat startling and unexpected manner a perfectly in- 
telligible reason for the existence of the two reticular layers. 
It is unexpected because, considering that all these cyto- 
plasmic strands are living protoplasm, it would appear more 
natural if they had readjusted themselves in the retina, letting 
the nuclei pass on. ‘The evidence, however, shows clearly 
that this is not the case, and that they are to a large extent 
trailed along and assist in the formation, at least, of the 
inner reticular layer. I say “assist” because they appa- 
rently only form its cytoplasmic basis; other elements, as we 
shall see later on, contribute to the final result. 

In studying thin sections of retinas of tadpoles I had 
often been struck by the fact that from the extreme end of 
the inner reticular layer irregular threads went off and 
curved inwards towards the membrana limitans interna. This 
is more striking in some cases than in others. Figs. 1 and 2 
are sufficient to show what is meant, but in some cases I have 
seen it so marked that it looked as if the inner reticular 
layer took its origin, at each end of the section, from 
the internal limiting membrane, sometimes almost shutting 
off the layer of “ganglionic cells” from those of the un- 
differentiated rim. ‘This appearance greatly puzzled me 
until the discovery of the migration of the nuclei made 
it clear that these threads which joined the inner reticular 


STUDIES IN THE RETINA. 35 


layer to the internal membrane were the remains of the 
attachments of the originally spindle-shaped nuclei which 
had moved away down the middle nuclear layer. 

More conclusive still is the fact that every now and then 
a section is found in which the nuclei of the middle layer, 
especially near the rim, are actually caught trailing irregular 
tangles of cytoplasmic threads in the manner shown i figs. 1 
and 2. This can be seen with some frequency, though by 
no means always, because it is probably a matter of accident 
whether the particular retina happened, at the moment 
it was fixed, to be in the exact phase of its life activities 
which required such movements. For it is hardly likely that 
the inward streamings of nuclei are continuous; periods of 
rest would probably intervene. However seldom they occur 
there is no mistaking their significance. 

Still keeping the movement of the nuclei in view, it is 
worth while paying further attention to the inner reticular 
layer. We find that the early stages in its appearance 
show differences which, though at first disconcerting, are yet 
on the whole entirely confirmatory. The earliest stages which 
I have so far seen are shown in figs. 5, 6, and 7, which Tinterpret 
as follows:—The nuclei, which had been fairly evenly dis- 
tributed through the retina, and not tightly squeezed together 
(see fig. 7), gradually separate along the line which will be 
later occupied by the inner reticular layer, the larger half 
migrating outwards. A row, two or three deep, remains 
against the internal limiting membrane, although one or two 
even of these, in the axis of the eye, may escape outwards, 
leaving a gap in the innermost layer (figs. 5 and 6). The great 
mass of the nuclei gradually move, as stated, outwards, but the 
very outermost can at the most move but a few micromilli- 
metres, being arrested at once by the pigment epithelium. 
The rest, therefore, leaving a few stragglers, crowd up close 
behind, with the result that the irregular but conspicuous rent, 
just described, occurs in the previously uniform nuclear ranks. 
This rent in its early stages seems to be mainly occupied by 
rounded vesicles, at least in the retina from which fig. 7 was 


36 H. M. BERNARD. 


drawn, but later becomes filled with a rather loose tangle of 
staining matter, composed mainly of the cytoplasmic frame- 
work in which the nuclei were suspended. ‘These phases not 
only reveal the outward movement of the nuclei, but also show 
that it is not due to pressure such as might be exerted from 
a region of active division. ‘I'he nuclei in these sections 
can only have moved outwards under the action of some 
attraction. This fact, that the force bringing about these 
migrations is attractive, 1s important, though we cannot stop 
at the present moment to develop it further. 

A slightly later phase in the formation of the inner 
reticular layer can be seen in figs. 3 and 4. The irregulari- 
ties seen in the layer in its first appearance, as a reticulum 
filling up a split among the ranks of the nuclei, asjust described, 
have become more definable as tongues running out among 
the outwardly pressing nuclei. In one case (fig. 4) several 
tongues appear of nearly equal size, although the one which 
appeared to be nearly in the axis of the eye was the largest. 
In another case (fig. 3) this axial tongue was very much larger 
than any of the others. As will be seen from the direction 
of the arrows in this latter figure, I explain these phenomena 
as due to the migration of nuclei from the sides. ‘lhe 
attraction which first drew the nuclei from their original 
positions in the embryonic retina to press outwards has 
extended and drawn nuclei from the peripheral portions of 
the retina which have not yet begun to function. This com- 
bined centripetal and outward movement of the nuclei would 
naturally give the rudiments of the inner reticular layer the 
shapes which they assume in these sections. ‘hat this 
movement is taking place in the direction of the arrows 
may be gathered from the closeness with which the nuclei 
in these sections are packed in the central and more actively 
functioning region as compared with their straggling and 
loose arrangement elsewhere. 

We have to add to this evidence, each item of which seems 
fairly conclusive, the fact that the inner reticular layer grows 
thicker as it slowly reaches the adult condition, and not 


STUDIES IN THE RETINA. 37 


only thicker but very much more extensive without the 
appearance of any special formative cells which would 
account for it.t If, however, the nuclei of the middle layer, 
each with more or less cytoplasm trailing behind it, do 
actually move along from the rim of the retina towards the 
axis, we can account not only for the gradual thickening of 
the inner reticular layer as the eye grows, but also for its 
curious stratification, which is sometimes very striking. 
The layer reaches its definitive thickness when the eye has 
ceased to grow and no more nuclei are produced at the rim. 

So far, however, we have only considered the inner 
reticular layer, but there are two such layers, as there should 
be if the mechanics of their formation here sketched be 
correct. If correct, it supplies us also with an explanation 
of the fact that the two reticular layers are always co- 
extensive with the region of rod-formation, only appearing 
where the vesicles are being protruded. A slight difficulty, 
however, now arises. If these layers are produced by the 
nuclei travelling down the middle layer from the rim towards 
the centre, why is not the reticular layer on the outer side of 
the stream as thick as that on the inner? An answer may 
be suggested which is probably correct, although it would 
be difficult to bring any evidence for or against it. Great 
numbers of the nuclei travelling on the outer side are 
arrested as they go and pass into the layer of rod nuclei. 
These might be expected to take all the cytoplasm they 
could with them as the formative substance of the vesicular 
protrusions which they are destined to send out from the 
retina for the formation of their rods. Hence it is probable 
that the greater part of the cytoplasmic reticulum which 
would otherwise be accumulated here as a counterpart of 
the inner reticular layer is carried outwards and used up 


1 [ have never seen any indication of the rows of small, faintly outlined, 
formative cells such as Borysiekiewitz (|. ¢.) describes for the inner reticular 
layer in human retinas ; whenever nuclei do occur in the layer, in all the eyes 
I have examined, they are always quite distinct, and to be regarded as migrat- 
ing outwards from the “ ganglionic cell” layer. 


38 H. M. BERNARD. 


in the production of rods. This seems to be a_ possible 
explanation of the difficulty. Some traces of an outer 
reticular layer, however, there always are, and doubtless 
here again the retinal cytoplasm forms its basis, and only its 
basis. In a future paper I shall show that neither of the 
reticular layers is a homogeneous structure; the outer layer, 
indeed, presents several difficult problems. In fig. 3, 0.7., we 
already see signs of accumulations of deeply staining matter 
along the line of the future outer reticular layer. These 
accumulations, which we shall meet with again in Part V, 
are apparently in some way due to the functioning of the 
nuclei, for it is obvious they cannot, from their position, be 
due to any merely mechanical streaming movements. 

In the very existence of these two reticular layers, as well 
as in their stratified texture, in their attachments round the rim 
by threads to the membrana limitans interna, and in the shapes 
they assume during early growth, we find strong evidence of 
the migration of the nuclei, which is the subject we have 
specially in hand.!. We may sum up the arguments briefly : 
(1) the nuclei of the adult rods protrude a little beyond the 
membraua limitans externa ; (2) the nuclei of the cones, which 
(in Amphibia) are early stages in the formation of new rods, 
move gradually outwards from near the outer reticular layer 
towards the membrana limitans externa as their rods de- 
velop; (3) no nuclear division takes place in this layer where 
rods and cones are developing ; the nuclei for the further pro- 
duction of rods come through the outer reticular layer from 
the middle nuclear layer; (4) no nuclear division takes place 
in this middle layer anywhere near the axial portion of the 
retina, and the supply must be kept up by migration from 
the sides. A very few may come through the inner reticular 
layer from the layer of “ ganglionic cells,” but the bulk of 


1 We shall refer to some of the very discordant views which have been put 
out as to the origin and constitution of these layers when we come to deal 
with them in detail. In the meantime a useful summary may be found in the 
Literatur-Verzeichniss to Borysiekiewitz’s first paper, ‘ Untersuchungen 
iiber den feineren Bau der Netzhaut,’ 1887, notes 19—27. 


STUDIES IN THE RETINA. 39 


those required travel along the middle nuclear layer from the 
undifferentiated rim of the retina where nuclear division is 
active during growth. Thus a stream of nuclei travels 
inwards from this undifferentiated rim towards the axis along 
the middle nuclear layer; (5) this stream of nuclei lays 
the foundation for the two reticular layers of the retina; the 
cytoplasmic trailings of the nuclei being, as it were, swept to 
the sides of the stream, accumulate, but while the inner 
accumulation persists the outer is mostly used up, probably 
in the formation of the rod-vesicles. 

Before leaving the subject for the present, I should like to 
call attention to the conviction which I expressed in Part II, 
p. 452, that the retina is a syncytium, in the reticulum of 
which nuclei are suspended, and that it is almost impossible 
to speak of “cells” in connection with its component 
elements. ‘he streaming of the nuclei and the trailing 
behind them of cytoplasmic tangles, which  trailings 
accumulate as the eye grows, may, I think, be regarded as 
complete justification for this conviction. I had not for- 
gotten and do not forget the large “ ganglionic cells,” 
which appear to supply an easy refutation. On the contrary, 
it was a prolonged study of these same “ cells” which first 
led me to this conclusion, as I shall relate in detail in a 
future paper. 

Lastly, I should like to venture the suggestion that the 
principle here established for the retina may be of wide 
application, although I cannot hear of any other exemplifica- 
tion of it as yet known. The principle is this: an organ 
has to continue to grow after it has begun to func- 
tion. Assuming that nuclei or cells are incapable of mitotic 
division when once specialised for some highly complex 
function, we should be compelled to postulate an undifferen- 
tiated region which would persist as long as growth lasts. 
From this region, which would be the centre of active nuclear 
or cell division, the new elements required by the functioning 
and growing area would have to migrate, through longer 
or shorter distances according to the exigencies of the 


4A H. M. BERNARD. 


particular case. Further, as in the case of the retina, these 
migrations may have considerable, influence on the tectonics 
of the organs in which they can be established. 


Part IV. 


On the Vesicular Swellings at the Tips of the 
“Cones” and some Earlier Form-phases in Rod- 
production in the Amphibia. 


As was shown in Part I,! the tips of the young cones swelled 
into vesicles on reaching the pigment layer. Vesicles or parts 
of vesicles were figured (P]. 3, figs. 2,3, and 10), and these justi- 
fied the construction of the series of form-changes shown in 
the diagram (fig. 4), but they were only certainly seen in eyes 
fixed with boiling corrosive sublimate. Other figures on the 
same plate (e.g. fig. 12, and on Pl. 31, Part II,? fig. 29) 
showed no traces of any such vesicular tips, and in some 
cases it was difficult to understand why, if they had existed, 
they should vanish so completely from the sections. This 
point has now been settled, not, I regret to say, by the 
discovery of a new and more perfect fixative, but by a kind 
of good fortune. I brought down a few tadpoles from 
Table Mountain, Cape ‘own, killed and fixed them in 
Perenyi’s fluid at midnight, i.e. when the pigment would 
be retracted. The object was to see whether, owing to the 
brilliant sunlight of South Africa and the intense heat, the 
pigmentation in the retina showed any modifications on that 
seen in our indigenous tadpoles, and if so whether any 
correlated changes in the retina could be discovered. For 
the same reason I made special efforts to obtain baboon’s 
eyes (see Part V). 

One interesting difference was at once apparent. ‘I'he 
pigment in the South African tadpoles is far greater in 
1 This Journal, vol, xtiii, 1900, p. 28. 

? Ibid., vol. xliv, 1901, p. 443. 


STUDIES IN THE RETINA. 41 


quantity and of a very much darker brown. The colour is in 
striking contrast with the reddish brown which is most 
common over here. But this was not all; probably in corre- 
lation with this increase in mass and quality of the pigment 
the rods were also different (see figs. 8, 9, and 10), in that 
the longitudinal striation is so marked that it can be seen at 
once with a low power. Cross-sections of rods which seem 
to be somewhat thin, 4 u across, often tapering to 3 pa, show 
a thick, straggling, branching, and knotted strand running 
down the axis of each rod as the representative of the axial 
reticulum, and connected irregularly with the dark strix 
running down the wall. Here and there the greater part of 
the axis of the rod is taken up by a mass of dull grey homo- 
geneous matter, in which case the axial reticulum is appa- 
rently represented by clump§ at the sides, but it usually 
comes into view again on focussing up or down. ‘These grey 
masses in the rods are the remains of material absorbed from 
the pigment granules (see fig. 8, with description). 

Apparently correlated with these strongly developed 
staining strie is the fact that the rods, though very thin, are 
comparatively speaking tough; for, quite unlike those in our 
own species of Amphibia, which break up so easily and 
usually part at the junction of the inner and outer limbs, in 
these eyes, where the retina and choroid have parted, they 
are drawn intact out of the dense pigment. 

Turning to the cones, we fortunately find that their vesi- 
cular tips share in this greater toughness. Very many of 
these latter, it is true, have broken down and have been 
reduced to a granular mash which is very conspicuous, but 
places such as those figured (figs. 9 and 10) might be 
multiplied to any extent. ‘The vesicles are shrunken, and it 
is largely owing to the folds in their walls that they are 
visible. Some seem to have clear traces of rows of dots 
running down them which remind one of the rows of dots on 
the longitudinal strize of the rods. In optical section the 
wall of the shrunken vesicle could often be traced quite 
plainly into that of the cone (fig. 10). 


42 H. M. BERNARD. 


We conclude, then, that in all young amphibian eyes, in 
which the rod layer seems to consist mainly of cones ending 
at some distance from the pigment, the apparently vacant 
space between the truncated cone-tips and the pigment is, in 
life, filled up by a compact mass of swollen vesicles. ‘These 
vesicles are, however, so exquisitely delicate that the process 
of fixation and hardening destroys them almost completely. 
But I should add that now that I have seen the vesicles in 
the Table Mountain specimens, I have been able to discover, 
in sections of our native forms, several cone-tips running out 
into faint diverging threads. 

Another peculiarity in the retinas of these Table Mountain 
tadpoles deserves mention. In Part I, p. 34, [remarked in a 
note that the only long-necked elements which I could find in 
frogs’ retinas at all resembling the long-necked “cones” figured 
by van Genderen Stort (‘Quain’s Anatomy,’ 10th ed., vol. 11, 
part 8, p. 48) were those which appeared in each case as one of 
the so-called twin or double cones (see PI. 3, fig. 5). Besides 
these, the only elements with long inner limbs were Schwalbe’s 
rods,in which the refractive globule had, as a rule, already 
disappeared and the outer limbs had already become 
cylindrical (Part I, Pl. 3, fig. 4, 7). But in the retinas of 
these T'able Mountain tadpoles, cones with striking refrac- 
tive globules like those figured by van Genderen Stort are 
very plentiful, close down against the pigment layer. The 
greater toughness of the walls may account for the persis- 
tence of the shape in a phase where it is quite lost in our 
native forms (see the phases Part I, fig. 4, c, andr). The 
fact that the refractive globule is not so quickly absorbed 
may be referred to the great quantities of pigment to be 
dealt with (for the origin of this globule see Part II, p. 463). 

The transformation of these long-necked cones into rods is, 
in some cases, very easy to follow. The conical portion 
thickens and shortens, while the swollen vesicle at the tip 
becomes cylindrical and the refractive globule disappears. 
In fig. 9 elements like those on the right and left hand are 
very common; that on the right shows a division in the 


STUDIES IN THE RETINA. 43 


material filling the rod; the short innermost darker portion 
clearly corresponds with the old staining tip of the cone 
(cf. the sections indicated by asterisks). The small numerals 
1—7 show a continuous developmental series illustrating the 
transformation of long-necked cones into Schwalbe’s rods. 

It was stated in Part I that the growth-forms (c, cg, cs) of 
the new rods shown in PI. 3, fig. 4, of that paper were due 
to the fact that each new element on being protruded 
had to force its way between tightly packed cylindrical 
rods; obviously the new vesicle could only swell into a 
sac after getting through, that is, against the pigment. 
If this explanation be correct we should expect to find 
other form-phases where the rods were not so long. A 
much more direct transformation of cones into rods was, 
indeed, figured in the series of elements supplied by the 
axolotl, Part 1, Pl. 3, fig. 8.! In this animal the rods are 
very thick, and, compared with their thickness, very short. 
Now it is interesting to note that we have at the sides of 
tadpole retinas, where the rods get progressively shorter, a 
very similar process of direct transformation of cones into 
rods to that which we found in the axolotl. The distal 
portion of the protruded cone seems to be neatly rounded off 
(fig. 11), as if there had never been any swollen vesicle at 
its tip. It is further quite distinctly striated longitudinally.’ 
Here, then, we have the cones chauging directly into rods by 
the absorption of the refractive globule and the lengthening 
of the outer limb at the expense of the inner. If we compare 
this process closely with that occurring among the long rods 
in the central regions of the retina (figs. 9 and 10), we find 
that it differs in two points: (1) there is no long thin neck, 
and consequently (2) there would appear to be no con- 
spicuously swollen vesicle at the tip which would have ulti- 


1 To make that figure true to life the tips of the cones in @ and d should 
have been drawn with delicate vesicles, but all traces of such vesicles had 
been destroyed in the actual sections. 

2 Compare the dots seen on the distal vesicles shown in figs. 9 and 10, also 
the remarks on the striation of the cones in Part II, p. 455. 


44. H. M. BERNARD. 


mately to be brought into the typical cylindrical form. It is 
easy to see that both these differences are due solely to the 
fact that where the rods are long cylinders the protrusion 
has to force its way between them, and only swells out into a 
conspicuous vesicle after getting through. 

Tt will be seen from the study of these details how important 
it is to keep the compactness of the layer of rods very clearly 
before the mind. The rod layer, in fact, arises as the result 
of the thrusting out of great numbers of vesicles from the 
retina, the vesicles only gradually assuming the long, cylin- 
drical rod shape. The varying forms which the early stages 
of new rods assume when first protruded, and until they are 
finally developed, depend not only upon the forms, but 
also upon the lengths of those among which they 
have to force their way. We have now seen two of 
these different series of form-changes, and it will be best in 
this connection to record the observations made on still 
earlier stages of growth, when the new vesicles are protruded, 
not among rods, but among other vesicles which have not had 
time to become rods. We shall see that whereas, when the 
rods are formed, and their shapes fixed, new vesicles have 
to adapt themselves entirely to them; while the rods are 
still unformed and vesicular the protrusion of new vesicles 
is able to modify their shapes. In the changes described in 
Part I we saw that the protrusion of fresh cones altered the 
shapes only of other cones, helpmg to change long cones 
into Schwalbe’s rods, but that they had no apparent effect 
upon finished rods, 

The first appearance of rod-vesicles begins very early, as 
soon as ever the eye begins to function. They can be seen in 
various sizes in figs. 3 to 7, as round clear spaces against the 
pigment. At first they are scattered and confused, because 
all the nuclei do not secrete vesicles simultaneously (see 
figs. 16, 17,and 18). A little later a stage is reached when 
they are arranged side by side as large sacs mutually com- 
pressing one another (figs. 12 and 15). It is at this 
stage that our sections usually failus. So long as the vesicles 


STUDIES IN THE RETINA. 45 


are small and round their outlines are clear, being preserved, 
no doubt, by the fixing of the pigment cells on the one hand, 
and of the deep staining matter, which usually forms their 
proximal walls, on the other (see figs. 17 and 18, where the 
shaded vesicles represent deeply stained walls). As soon, how- 
ever, as they lengthen out, the walls become so delicate that 
they collapse under the violent processes of fixing, hardening, 
and preparing the sections. It is common to find in young 
eyes great empty spaces where the rod layer should be 
between the retina and the pigment, the spaces occasion- 
ally interrupted by single, short, thick, deeply staining 
rod-like structures, one here and there having survived. 
That elements of some form or other filled these gaps is 
absolutely certain ; indeed, the ragged remains of membranes 
can often be seen fringing the distal ends of the nuclei, and 
protruding a little from the membrana limitans externa. A 
creat many sections show nothing but this, and one is apt to 
become hopeless of ever seeing the vesicles which, in life, had 
been crowded together in those gaps. On one occasion I 
found one of these spaces occupied by a single large 
vesicle with a complete pigment cel!, which had left the pig- 
ment epithelium, inside it. In time traces of long vesicles 
become more frequent because they are supported and 
preserved by being in contact with other more formed and 
stronger elements (Part I, Pl. 3, fig. 16). It is when a 
number of very fragile vesicles are mutually supporting and 
squeezing one another that they disappear from our sections 
leaving hardly a trace behind. 

In sections of retinas killed at night I have succeeded at 
last in finding vesicles intact. ‘They are slightly mottled 
and dotted over with stain, and I conclude that they owe 
their preservation largely to this fact, viz. that their walls 
were strengthened by this staining matter, as appears to have 
been the case with the rods and the cone tips in the retinas 
above referred to from Table Mountain. Fig. 12 shows a 
eroup which have fortunately been preserved intact, and 
fio. 13, a—e, are elements from the same retina. 


46 H. M. BERNARD. 


On looking at these our attention is at once arrested by 
figs. a—c. We have apparently typical cones with their 
points thrust into terminal vesicles. A little reflection, how- 
ever, shows how such appearances could be easily produced 
as transitory phases. ‘To make it clear I give a diagrammatic 
series, fig, 14, a—d. a represents an unmodified vesicle 
protruded as a long oval. ‘The pressure caused by the 
gradual protrusion of new vesicles will be exerted upon a in 
the direction of the arrows shown in b, with the result 
that @ will take the form shown by 6 (ef. the middle vesicle, 
fig. 12). In the narrow neck of b staining matter accumu- 
lates. Continuation of the pressure further lengthens the 
neck, and at the same time the adding of new vesicles forces 
back the pigment cells.! 

In the stage c I have introduced a refractive globule, 
which we may assume to have come out of the distal 
vesicle as matter absorbed by it from the pigment, as ex- 
plained at length in Part II. At this stage it is again the 
turn of the element whose form-changes we are following to 
receive another discharge from the retina or, as argued in 
Part II, from its nucleus. This discharge drives out the 
staining matter which occupied the neck, so that it protrudes 
into” the distal vesicle. The three figures 13, a—c, show 
three distinct degrees of thrust, quite accidentally selected, 
the figures having been drawn in the order shown in the 
plate before I was at all clear as to their meaning. In a, 
only the narrow tip of the matter from the neck has been 
pushed into the sac; in ¢, the tip and a portion of the 
refractive globule, in this case the matter composing the tip 
itself has been disarranged against the distal end of the 
vesicle; in b, a larger portion of the staining matter still has 
been thrust outwards into the sac. ‘These curious “ cones,” 


1 This lengthening of the vesicles widens the distance between the pigment 
layer and the body of the retina. The width is greatest in the centre of the 
retina, and in very young eyes diminishes rapidly on either side. This is 
certainly due to the greater activity of vesicle formation, i.e. of rod-produc- 
tion in the area of most active functioning. 


STUDIES IN THE RETINA. 47 


then, are due to the driving out of the staining matter which 
had accumulated in the neck of the squeezed-up vesicle. 
There is no telescoping of the membrane into itself. It is 
simply another form of the phenomenon shown by asterisks 
in fig. 9, where the original contents of what has hitherto been 
thought to be the tip of the cone become the proximal portion 
of the contents of the Schwalbe’s rod, which arises as soon 
as the vesicle has assumed its cylindrical shape. ‘lhe vesicle 
assumes this latter shape apparently in both cases as it 
becomes more and more turgid with matter received on the 
one hand from the retina, and on the other from absorption 
of pigment. The stages cand d in fig. 13 require no con- 
necting links, d being the next stage produced by the filling 
up of the distal vesicle. Still younger and simpler stages of 
transformation of vesicles into rods are shown in fig. 15. 
They need no comment. 

Many interesting details of observations in relation to this 
part of the subject might be added, but the task of dealing 
with the retina of the Amphibia alone threatens to lengthen 
out so greatly that only points necessary to a clear under- 
standing of the essential morphology of the retinas dealt 
with can be mentioned. 

How necessary it is to understand the minute details of 
rod-formation I need hardly insist, that is if we are to make 
any progress with our researches into the mechanics of vision, 
for the rods are the specific structures which constitute 
the retina the specific organ of this sense. Believing, as I 
do, that all structures are produced both phylogenetically 
and ontogenetically only in response to physiological needs, 
I feel confident that in a case like this where the rods are 
produced in situ, and only when required, their processes 
of formation must throw light upon the mechanics of their 
functional activities. Some further details relating to these 
activities will be found in the next part. 


48 H. M. BERNARD. 


_ Parr V. 


On the Removal of the Absorbed Pigmentary 
Matter from the Rods: an Explanation of the 
“™Miller’s Fibres.” 


In Part IT I described in detail a set of phenomena which 
found their simplest interpretation in the assumption that 
the protoplasmic vesicles, known as the rods, protruded by 
the retina against the pigmented epithelium, absorbed the 
pigmented granules, and at times also the cytoplasm of the 
pigment cells. I propose in this paper to describe another 
set of phenomena which indicate the way in which the rods 
are freed from the excess of matter thus absorbed. 

My results differ somewhat widely from any hitherto pub- 
lished, and especially from those obtained by the now popular 
impregnation methods, and I ought, perhaps, to make some 
excuse for not testing those other methods myself. My 
answer, I fear, can only be an apology. I selected the 
purely comparative method deliberately as the only absolutely 
certain way of obtaining light on intricate morphological 
problems, but the method is slow and laborious, and I grudge 
the time necessary to become an adept in the use of others, 
the results of which have still to be interpreted. 

The pigmented matter was, as we saw, absorbed through 
the walls of the outer limbs, and some of it found its way 
through the transverse membranes into the inner limbs, 
where it helped to form the bodies known as the ellipsoids. 

Part of the absorbed matter, then, finds its way through 
the transverse membrane into the inner limb. Here, unless 
it can find some further method of escape, it must accumulate 
and cause the inner limb to swell. No such swelling of 
the inner limb takes place in the Amphibia, but it is a 
striking phenomenon in many fish. This is the explanation 
giant cones”? which are so startling when seen for 
the first time (see figs. 20, b, and 21). 


ce 


of their 


STUDIES IN THE RETINA. 49 


Although the rods and cones in the fish are not our special 
subjects in this paper, it will be necessary to enter into a few 
details with regard to them. In the very young the elements 
are seen to be nearly all of uniform size, with apparently the 
same form-phases in their production as we described for the 
Amphibia, viz. (a) small cones, (b) gradually lengthening 
cones, (c) Schwalbe’s rods, and (d) fully developed rods 
(fig. 20, a). These are the natural stages in the formation 
of new rods in the amphibian retina. But in the fish, 
after the earlier stages of growth have passed, we find a very 
striking change, which seems to begin somewhere near the 
central region! and spread gradually over a large part of 
the retina. The change is as follows :—The inner limbs of 
the earliest formed rods gradually swell, until, in large and — 
presumably old fish, they may be of monstrous proportions. 
So that though, while growth is still going on, there may be 
room for a few more young cones to protrude or for a few 
more of the inner limbs of the Schwalbe’s rods to shorten 
while the outer limbs lengthen, that time comes to an end, 
and the retinal elements, at least over the modified area, 
consist entirely of (a) rods with monstrous inner limbs, and 
(b) bunches of Schwalbe’s rods. The thin thread-like inner 
limbs of the latter find their ways between the swollen inner 
limbs of the “ giant cones,” while their numerous cylindrical 
outer limbs fill up the spaces between the comparatively 
speaking small outer limbs of the ‘‘ giant cones.” Fig. 20, 
a, b, shows comparisons between the conditions of the elements 
in young and old eyes in the viviparous blenny. Fig. 22, 
a, b, c, shows different parts of the same retina of a young 
plaice, a being near the centre, where it has functioned most 
actively. Similar results might have been shown from my 
sections of trout and stickleback. I had no sections of young 
cod for comparison with fig. 21, from an old fish, but we may 
judge of the original thickness of the elements by those 
which persist as Schwalbe’s rods, a few of which are shown. 

1 Without having exactly located the region, I believe it to be the postero- 
ventral half of the central region. 

vou. 46, part 1,—NEW 8ERIKS. D 


50 H. M. BERNARD. 


Here, then, we have the very phenomenon we anticipated in 
the event of the refractive matter which passed from the 
outer into the inner limbs not being able to escape from the 
latter, at least as fast as it accumulates. These inner limbs 
become swollen with refractive matter. That this is the true 
explanation of the “ giant cones” is rendered clear by a study 
of large fish like the cod. While in some, especially smaller 
fish, the matter filling the inner limbs is often difficult to 
define, in the sections I possess of the retina of an old speci- 
men of this fish ! the refractive inatter is quite recognisable. 
It is often seen in round homogeneous pellets just inside the 
transverse membrane, and usually continued some way up the 
axis of the inner limb. Round the periphery the contents 
are more granular. Here and there, however, the whole 
inner limb is one smooth, bright, homogeneous mass. These 
smooth, round, refractive pellets, which seem to accumulate 
above the transverse membrane, may be compared with the 
refractive globules in the cones of the frog. The refractive 
matter here, as elsewhere, is deeply coloured by plasma 
stains, such as eosin, but easily gives up nuclear stains. 

If further evidence were wanted that the material which 
swells the inner limb is the refractive matter absorbed from 
the pigment granules by the outer limbs, it is supplied by 
those cases in which the colour of the absorbed matter is the 
same as that of the pigment. Such cases may be purely 
individual differences, and depend upon chemical variations in 
the pigments, or, perhaps, may be due simply to a too rapid 
absorption. Certain it is that though the shape of the pig- 
ment granules is lost, the colour of the absorbed matter may 
now and then be hardly altered. Among my sections of the 
plaice,? for instance, there is one in which the strong reddish 
colour of the pigment only slowly vanishes. It pervades all 
the outer half of the swollen inner limbs, sometimes extending 
some way up the “rod fibres.” Other more striking instances 

1 Fixed with corrosive sublimate. 


2 Specially fixed and preserved for me by my lamented friend Mr. Martin 
Woodward at the Plymouth Marine Laboratory. 


STUDIES IN THE RETINA. 51 


of the persistence of the colour right into the retina will be 
given below. 

Before continuing to consider whether and how the matter 
can escape from these inner limbs, one or two points may be 
noted in passing. (1) The “ giant cones” of the eyes of fish 
are not the morphological equivalents of the cones in the eyes 
of the frog. The latter are almost the earliest stages of rod 
formation; the former are not only fully developed rods, but, at 
least so far as growth proportions go, the most highly deve- 
loped elements known in vertebrate eyes; their great size 
will have some bearing upon the question as to the length of 
the life of individual retinal elements when that question 
comes to be put. 

(2) A very large proportion of these “ giant cones” are 
double. Although the dividing line in the swollen inner 
limb may become very faint, and sometimes only traceable 
in tangential sections, the presence of two nuclei and of 
two outer limbs will always show whether any large “ cone” 
is double. I have already shown that the peculiarities of the 
double cones in the frog are due to the fact that two 
protrusions of the retina start side by side almost simul- 
taneously, and their forms are due to mutual pressure, both 
being subject at the same time to the general pressure 
which we have to assume to account for the ordinary cone 
phases of rod formation. In these great double “ giant 
cones” we merely have two rods very close together, and about 
the same age. The fusion of their inner limbs will take place 
sooner or later, as these inner limbs swell with matter. The 
only part of the phenomenon which requires investigating is 
why these rods of similar age should be so frequently in 
pairs, or, tracing it a stage further back, why it is that as 
new nuclei arrive to send out their protrusions to form new 
elements for the growing retina, they leave so many pairs of 
nuclei between which they do not or cannot force their 
way. 

(3) The remarkable change which takes place in the 
forms of the elements of the growing’ fish retina, from an 


Sp H. M. BERNARD. 


early stage with (a) small cones, (b) “Schwalbe’s rods,” 
and (c) fully formed rods, into a stage with only two kinds 
of elements, viz. (a) rods with enormously swollen inner 
limbs, and (b) “Schwalbe’s rods” with long thread-like 
inner limbs, fully justifies the appeal which throughout all 
these papers we have made to pressure in order to account 
for the form-phases of the elements of the bacillary layer. 

(4) The fact that in the eyes of all Vertebrates higher than 
fish refractive matter no longer accumulates in the inner 
limbs, at least so as to swell them to such disproportionate 
sizes, apparently justifies the conclusion that these accumula- 
tions are not helpful to the specific function of the retina. 

Returning to the subject in hand, we must now show how 
the refractive matter ultimately escapes from these swollen 
inner limbs of fish retinas. 

Reference to the sections of the cod leaves no doubt on 
this point; the very size of the ‘giant cones,’ and the 
coarseness of their connections, reveal what the smaller 
elements of other eyes could not so plainly show, at least 
until the facts have already been made clear. What I take 
to be a thick stream ascends from each of these “ giant 
cones,” and ends in a refractive clump against the outer 
reticular layer, the ‘ cone ” nuclei being sometimes elongated 
in the line of the stream. The terminal clumps form, as it 
were, conical expansions where the streams meet the tan- 
gentially arranged tissue of the outer reticular layer. Here, 
again, microscopic examination of the “stream,” and espe- 
cially of its large conical expansion, as seen in the cod, show 
at once the presence of the same matter as that in the inner 
limbs. It is not meant, of course, that this matter is alone 
present, for in what follows it will be seen that this refractive 
matter follows the threads and fibres of the cytoplasmic 
network of the retina. In this case the stream and its 
conical expansion doubtless have a cytoplasmic framework. 
These streams with their expansions occur in one form or 
another in most if not all eyes, at least as physiological 


stages, and are usually described as the “cone fibres” with 


STUDIES IN THE RETINA. 53 


their intra-retinal terminal swellings.) Fig. 21 shows them 
in the cod, fig. 20, a, b, in the blenny, fig. 22, a, b, c, in the 
plaice, fig. 24, a, in the trout. These swellings, which are 
also specially conspicuous in Ramon y Cajal’s figures from 
metal impregnation preparations, have hitherto found no 
explanation. They can now be accounted for as the points 
where the refractive matter escaping from the rods is 
temporarily arrested as it reaches the outer reticular layer. 
Confirmatory evidence can be seen in the fact that their size 
depends upon the functional activity of the part. Fig. 22, 
a, b, and c, shows three parts of the same retina. They show, 
as do all the fish eyes I have examined, that the retina is 
very unequally used up. The part a from the central region 
shows the largest swelling of the inner limbs of the knobs of 
the rod fibres, and the most marked using up of the inner 
and middle nuclear layers. 

Having brought the refractive matter thus far in the eyes 
of the fish, we may go back and consider some other eyes in 
which it escapes from the outer limbs without swelling the 
inner limbs to such monstrous proportions. We can only 
refer to two cases, for a full treatment of the subject would 
require a close comparative study of the retinas of all the 
animal groups other than fish. The two cases chosen are 
especially interesting because they present such striking 
contrasts: (1) the Amphibia, with the inner limbs of their 
adult rods quite small and insignificant as compared with the 
outer limbs; (2) the Primates, with their long, rather thick 
inner limbs and thin outer limbs not much, if at all, longer 
than the inner limbs (see fig. 31, a). 

(1) That the refractive matter escapes into the inner limbs 
in the Amphibia we know from the invariable presence of 
the ellipsoid. But the ellipsoid does not, as a rule, seem to 
grow, so that if refractive matter is always exuding through 
the transverse membrane, it must as rapidly be transformed 
and conveyed away through the retina. Certain it is that 

1 I need only refer to the familiar text-book diagrams, such as fig. 52, 
p. 46, of ‘ Quain’s Anatomy,’ 10th edition, vol. iii, part 3. 


5A H. M. BERNARD. 


the inner limb is never swollen up with refractive matter. 
This sparing of the inner limbs in the Amphibia may 
perhaps be correlated with the enormous size of the outer 
limbs, for, so far as I know, no other group of animals has 
them so large in proportion. In the frog they are immense 
cylindrical vesicles, sometimes as much as 60, long and 
9 to 10m in diameter. These, then, form very capacious 
reservoirs for the absorbed refractive matter, and, perhaps, 
seldom require, during any single period of activity, to 
overflow into the inner limb. In this way the matter may 
be dealt with by the outer limb itself, and, apart from the 
ellipsoid, escape directly into the retina along the wall 
of the inner limb without entering it. That the refractive 
matter escapes directly from the outer limbs into the retina 
along the walls of the inner limbs can sometimes be actually 
seen (fig. 25, a, b, c). These cases are all from the South 
African tadpoles referred to in Part III, which, owing to the 
immense quantity and dark colour of the pigment, are 
very instructive in this connection. 

(2) Equally decisive for our contention are my sections of a 
human retina (the healthy normal eye having been excised 
for a morbid growth on the eyelid). This eye had clearly 
not been much exposed to light before excision. We conse- 
quently find the outer limbs of the rods free from all 
refractive matter, and, like the inner limbs, almost clear 
vesicles but for the longitudinal fibrils and the granules 
taking nuclear stain. ‘I'he fibrils on the inner limbs are dotted 
like those of the outer limbs in the Amphibia (see Part II and 
figures). Naturally no thick refractive streams can be seen 
running up into the retina from these rods. In very strong 
contrast with this are fig. 31, a, b, from retinas of the South 
African chama baboon,’ which live in the full glare of the 

' Kindly fixed in Perenyi’s fluid and preserved for me by the well-known 
ophthalmic surgeon, Mr. EK. Treacher Collins. 

2 They were generously obtained specially for the purposes of these re- 
searches by Mr. J. C. Kous, Tafelberg Station, Cape Colony, through the kind 


intervention of our mutual friends, Mr. and Mrs. Mallinson, of the Hex 
River Valley. 


STUDIES IN THE RETINA. 55 


South African sun. Here the rods, and especially the large 
inner limbs, are mostly full of pigment, which can be seen 
streaming inwards into the retina, no longer forming single 
fibrils with terminal knobs, but great tangles of refractive 
matter, which, in the eye (31, a) with dark blackish pigment 
are dull and blackish, but in the eye (31,6) with bright 
yellowish-brown pigment are bright yellow-brown. I may 
say that after seeing how the pigmented matter streamed 
through the retina in the tadpoles brought from the slopes of 
Table Mountain, I was quite prepared to find something of 
the kind in the retina of the baboons, but was myself 
surprised to see how very obvious the escape of the absorbed 
pigmented matter into the retina is in these cases. The pig- 
ment is so dense that the colouring matter is not bleached in 
the rods, nor, indeed, does it undergo much loss of colour 
throughout its passage through the retina, as it usually does, 
say, In our indigenous Amphibia. 

We have so far, then, traced the matter absorbed by the 
rods into the retina as far as the region known as the outer 
reticular layer. This is in many respects one of the most 
difficult parts of the retina to understand. The matted and 
deeply pigmented strands just below this layer in the 
baboon’s eye, as well as the conical expansions of the 
ordinary ‘‘ rod fibres,” indicate that the absorbed pigmented 
matter is temporarily stopped by it. But the exact cause of 
the stoppage at this point I have not succeeded in unravelling. 
Krause thought that there was a tangential membrane at 
this place, his ‘‘ membrana fenestrata,’ ' and certainly the 
first time one sees the outermost layer of nuclei of the middle 
layer arranged tangentially in a compact row, as shown in 
figs. 20, a, and 22, c, it is difficult not to think that Krause was 
right; but a study of older eyes (figs. 20, b, 21, and 24, a), or 
even of the more used-up parts of younger eyes (fig. 22, a), 
will show that these nuclei do not belong to any fixed mor- 
phological structure in the retina such as a membrane, but 

1] have unfortunately never seen a copy of the book written by Krause 
under this title. 


56 H. M. BERNARD. 


that they are merely nuclei of the middle layer passing out- 
wards to become rod nuclei, and apparently flattened against 
the same tangentially arranged cytoplasmic tissue as that 
which detains the refractive matter in the manner described 
above. But the difficulty is not quite so simple as this, viz. 
that the stoppage of the nuclei going outwards is due to the 
presence of tangentially arranged tissue, or even to a mutual 
blocking of the way on the part of the nuclei moving outwards 
and of pigmented matter moving inwards. ‘That this latter 
is not the cause is clear, because we find the same stoppage 
of the refractive matter even when, as in old eyes, nearly all 
the middle nuclei have passed outwards (see fig. 21). That 
other subtler complications are present can be gathered from 
the fact that the “rod fibres” often expand so as to form 
chambers in the outer reticular layer, and clumps of matter, 
often taking nuclear stains, may be seen in various conditions 
within these chambers. The relations of these clumps of 
staining matter to the terminal expansions of the “‘ rod fibres” 
is not easy to ascertain ; it is clearly necessary to keep them 
distinct in our minds. Borysiekiewitz, who, I believe, is 
the first to figure these chambers,! took them for a new and 
hitherto undiscovered layer of cells, the “nuclei of which may 
sometimes be seen dividing.” ‘This description, however, does 
not apply to any eye I have yet examined, for I have found 
them in all stages of formation, sometimes in patches, some- 
times all along the retina (cf. figs. 20, 22, and 24). A com- 
parative study has convinced me that they are, as stated, 
merely expansions of the inner ends of the “rod fibres ” 
round some peculiar mass of staining matter. Similar masses 
occur in the cytoplasmic chambers between the rod nuclei 
and the outer reticular layer in the frog, as shown in fig. 25, 
bandd. I can regard them, therefore, only as form-phases 
expressive of some physiological activity, the significance of 
which, so far as I have been able to unravel it, will be 
explained in a later paper. 

But whatever is the real structure of the outer reticular 


' Figured in 1887, but only claimed as a new “ cell” layer in 1894. 


STUDIES IN THE RETINA. 57 


layer, we shall see from what follows that the refractive 
matter sooner or later finds it way through it. We shall, 
indeed, now proceed to show what very startling effects its 
passage may have on the remaining layers. 

Every sagittal section of a functional retina will show us 
the matter streaming through the middle nuclear layer and 
through the inner reticular layer, in which latter, however, the 
streams frequently lose themselves. Indeed, as must be 
apparent by this time to every student of the retina, | am 
putting an entirely new interpretation upon a very familiar 
phenomenon, viz. the ‘“ Miller’s fibres.” These, as is well 
known, have hitherto always been regarded as sustentacular, 
and are said to be formed out of distinct cells with recognis- 
able nuclei. Buta survey of many eyes and of eyes of the 
same kind at different ages, and of the same eye at different 
parts and in different physiological conditions, shows beyondall 
mistake that they are only streams of absorbed pigmentary 
matter finding its way through the retina. The current 
doctrine that they are sustentacular has been based solely 
upon their appearances when most developed. Well-de- 
veloped streams may be found at almost any age, inasmuch as 
their development depends upon the degree of functional 
activity of the retina; but according to my experience they 
are found in this condition most frequently in very old eyes, 
as we shall see in detail below. 

This, then, is the next point we have to demonstrate; the 
chief difficulty in the way of doing so is how to select from 
the abundance of the evidence only that which is the most 
conclusive. 

First of all, it is best at the outset to record the obser- 
vation that the refractive matter seems to be temporarily 
arrested by all cytoplasmic strands and membranes which are 
arranged tangentially, and only to form definite streams along 
strands arranged radially. Hence the rapidity with which 
the refractive matter passes through the retina depends upon 
the number of suitably disposed radial strands. From the 
rods to the outer reticular layer most of the strands are 


58 H. M. BERNARD. 


radial, e. g. “the rod and cone fibres,” hence accumulations 
of amorphous matter seldom take place in this layer. On 
reaching the outer reticular layer there occurs, as described, 
some temporary obstruction, the exact nature of which we 
have not attempted here to unravel. Through this reticular 
layer, however, the matter escapes. In young eyes with a 
plentiful cytoplasmic reticulum supporting the rows of nuclei, 
radial strands can be found in abundance to carry the matter 
through the middle layer to the inner reticular layer; but in 
older eyes, when the nuclei of the middle layer have been 
largely used up and the cytoplasmic reticulum is so reduced 
that but few radially disposed strands can be found, the re- 
fractive matter tends to accumulate often in large quantities 
(m.n. of the figures). In fig. 20, b (blenny), it is seen in small 
irregular patches ; in fig. 24, a (trout), in thick tangential 
strands just above the outer reticular layer; in others, 
again, in immense tangentially arranged sheets. In fig. 23, 
a and b (plaice), the accumulations are near the outer 
reticular layer; in fig. 21 (cod) near the inner reticular 
layer. Many more figures might have been given, but 
these must suffice. Fig. 23, a, which was from a very 
large old plaice,! should be compared with fig. 22, a, 
b, c, which are from a young plaice, six inches long. In 
the least used-up part of the retina (c) no traces of these 
accumulations can be seen; in b they are beginning ; in a they 
are already of considerable size, but in the very old fish they 
are enormous, and occur over most of the retine. Of the few 
traces which I have so far seen of accumulations of matter 
in the retinze of mammals one is shown in fig. 30, where a thick 
strand runs along on the inner side of the outer reticular 
layer of a mouse which had been exposed to the light of an 
arc lamp. As it tapered away it gave off typical ‘“ Miiller’s 
fibres”? in the way figured (see also p. 37 and fig. 26). 

Before going on to the inner reticular layer, one word as to 
the supposed nuclei of the “ Miiller’s fibres.” These are nothing 

‘ Specially selected for these researches by my friend the late Mr. Martin 
Woodward, while temporarily associated with the Irish Fisheries. 


STUDIES IN THE RETINA. 59 


but the ordinary nuclei of the middle layer, and are used up 
like the rest. The appearances which have led to the sup- 
position that they are nuclei of fixed morphological strands 
are due to the fact that single nuclei are not infrequently 
involved in these streams of matter, and, indeed, may at 
times apparently enter into some intimate physiological 
association with them. ‘'hey may often be seen drawn out, 
and even at times robbed of their chromatic substance (see 
fie. 25, c). That they are not the nuclei of preformed sus- 
tentacular fibres follows from the fact that a comparative 
study shows that no such preformed structures exist, and that 
the so-called “Miiller’s fibres” are mere expressions of 
functional activity, and great numbers, even when best 
developed, have no such involved nuclei (fig. 32, a). 

Coming to the inner reticular layer, this also, like the 
middle nuclear layer, undergoes changes with age (cf. 7.7,, figs. 
20, a and b, and 24, wand b). In very young eyes the reti- 
culum is close, and forms what is called the “ Punktsubstanz.” 
As soon as the eye begins to function, before which time 
there are no “Miiller’s fibres,” streams of refractive 
matter begin to pass through it as very thin radial threads. 
Under a high power these are seen to be a fine zigzag; they 
are clearly not independent strands, but some staining matter 
running along the threads of the inner reticular layer. Fur- 
ther, they may branch or end suddenly in thin, tangentially 
arranged layers, from which new radial strands arise to run 
further in. Again, itis evident that these thin radial strands, 
which every one would at once call the ‘‘ Miiller’s fibres,” are 
not fixed structures, from the fact that in the retina of an 
older animal of the same kind (cf. figs. 20, a and b, 24,@ and 
b)! they may have disappeared altogether, and instead there 
occur thicker streams finding their way in much coarser zig- 
zags(fig.21) along the strandsand between the much more open 


1 The specimens of the viviparous blenny were fixed in Blés’ fluid 
in the St. Andrews Marine Laboratory, and kindly given me by Mr. Wallace, 
who had prepared them for his own work, The trout were specially fixed for 
these researches by Dr. Kyle, also of St. Andrews. 


60 H. M. BERNARD. 


meshes of the now altered reticulum. This fact is an absolute 
demonstration that these“ Miiller’s fibres” are not independent 
preformed structures, but merely cytoplasmic threads of the 
retinal reticulum thickened with matter. When they run 
quite straight without any zigzag we must regard it as due 
to a gradually acquired radial rearrangement of the threads 
of the reticulum (cf. fig. 32, a and b). 

Then, again, apparently at any point in the inner reticular 
layer, these strands may end suddenly, and the staining 
matter which was travelling along them may disperse 
to right and left (fig. 28,c). Many of the different aspects 
of the inner reticular layer are due to the presence of 
this refractive matter accumulated in different ways along 
its strands. One phenomenon is particularly suggestive; I 
refer to the darker zones which are frequently seen in it 
running for longer or shorter tracts round the retina. ~ I 
have seen them frequently (see figs. 20, a, 22, b, c, 24, a). 
Borysiekiewitz has also called attention to them. These 
dark zones are, as it were, waves of absorbed matter, records 
of former periods of functional activity, passing through the 
retina. This is not evident microscopically when the reti- 
culum is a close “ Punktsubstanz,” and the matter finely and 
evenly dispersed, but becomes quite obvious when the reti- 
culum is coarse and open, for then the individual strands of 
the affected part can be seen specially thickened (see fig. 24, a). 

All these facts become so obvious to any one who will take 
the trouble to study the retina comparatively that I feel it 
almost unnecessary to discuss the details any further. One 
or two points, however, remain to be noted. Just as the 
streams end almost anywhere in the inner reticular layer, the 
matter dispersing along the tangential strands, so fresh ones 
may begin anywhere within the same layer. And this brings 
us to the next layer, the nerve-fibre layer, or, as it is more com- 
monly but less accurately called, the “ ganglionic cell layer.” 

The appearance of the strands which run radially from the 
inner reticular layer to the membrana limitans interna is well 
known ; they are the typical inner ends of the“ Miiller’s fibres.” 


STUDIES IN THE RETINA. 61 


Usually comparatively thin as they leave the inner reticular 
layer, they expand into a conical arrangement of strands or 
membranes until they look in some sections like an arcade of 
expanding columns supporting the internal limiting mem- 
brane with its subjacent reticulum. Under the arches of 
this arcade are found the strands of the optic nerve, and the 
“ oanglionic cells.’ My faith in the sustentacular 
character of the “ Miiller’s fibres ” was first shaken by finding 
that in many of my preparations the majority of these 
columns arise from the edge of, or from various depths 
within the inner reticular layer itself, and that those which 
did so did not apparently differ from those which came 
through the inner reticular layer from the outer layers. It is 
quite apparent that when they arise from the edge of the 
inner reticular layer, they are in a position to collect and 
carry away matter from that layer (figs. 20, 6, 21, 24, a, 
26, b, 27). It is common also to find them arising in one of 


so-callec 


the darker zones above referred to, and when once the 
suggestion is made that they are, as it were, draining the 
inner reticular layer, a flood of light is thrown upon all 
their various shapes, for the typical arcade form I have 
described, though frequently found, is not invariable. 
In my preparations of the retina of a large cod, for instance, 
the typical expanding columns are somewhat rare, so 
that the matter, not carried away fast enough, clogs any 
strands or membranes running tangentially ; see fig. 21, in 
which it coats the strands (n.s.) supporting the nerves. 

Solid accumulations of this matter are, however, not often 
found in the nerve-fibre layer, although the clotting of the 
nerve and other strands which partly occupy the layer may 
be very dense (see fig. 28, b, from an old rat). Something 
more like solid accumulations are found in certain old eyes; 
e.g. figs. 26, a, b, and 28, a, show the absorbed matter 
accumulating within the conical expansions of the “ Miiller’s 
fibres,’ sometimes causing them to change their forms and 
become nearly bell-shaped—the trumpet shapes shown in 
fig. 26, b, are apparently due to distortion of the sections. 


62 H. M. BERNARD. 


But these accumulations and the clottings of strands and 
membranes are not sufficient to account for the lifelong 
streamings of refractive matter into this layer, and we should 
have to assume that it escaped finally through the internal 
limiting membrane to join the vitreous humour, even if the 
microscope did not clearly show us that this is what actually 
takes place. 

In very few sections will the internal membrane be seen 
quite thin and clear; it is usually found thick and apparently 
laminated, and layers are frequently found flaking off into 
the hollow of the eye. hat these flakes are, at any rate in 
part, due to the matter which comes along the “ Miiller’s 
fibres” can be seen in the fact that in osmic acid preparations, 
in which these streams are usually blackened, the portions 
of the internal membrane which cover their conical expan- 
sions not infrequently show different degrees of blackening 
(see fig. 29, a). This shows that the refractive matter is 
certainly deposited on the internal limiting membrane. The 
question is, Does it pass through? It certainly passed into 
the retina through the external protoplasmic membrane, 
pushed out in the form of rods it traverses the whole thick- 
ness of the retina, and if it does not pass through the exactly 
similar protoplasmic membrane on the inside of the retina it 
ought to accumulate in large quantities. The only accumu- 
lations which we actually find in connection with this mem- 
brane are the above-mentioned lamin, which, as is well 
known, belong to the vitreous humour. Some sections, 
indeed, show the stained ‘ Miiller’s fibres,” looking like so 
many processes rooting the similarly stained remains of the 
vitreous humour into the retina. And here let me say 
that absolute microscopic demonstration of subtle physio- 
logical processes may not be possible as so many separate 
details, but when all the facts are taken together the evidence 
may become as convincing as if we could prove each detail 
separately. This particular detail, however, namely, that the 
refractive matter absorbed by the rods passes ultimately into 
the vitreous humour, admits of demonstration. 


STUDIES IN THE RETINA. 63 


This demonstration is afforded us by the fact that in the 
baboon’s eyes the pigmented matter retains its colour right 
through the retina, being only slightly less bright and 
refractive near the internal limiting membrane, where it is 
present in enormous quantities. In the youngest baboon’s 
retina the congealed vitreous humour was left in situ in the 
base of the retinal cup, and appears in the sections. Its 
layers nearest the retina are coloured like the 
pigmented matter on the retinal side of the internal 
limiting membrane.! 

Returning to our review of the passage of the matter 
absorbed by the rods through the retina, we have seen that 
if, instead of the matter having to travel along zigzag paths 
on the strands of the cytoplasmic reticulum, it found a sufficient 
number of radial strands running in continuous courses right 
through, the passage would be much simplified. All the accu- 
mulations of matter which we have described in the eyes of fish 
might be avoided. ‘he most perfect radial strands which I 
have ever seen running through the inner reticular layer occur 
in sections of a human retina” which, from the scarcity of the 
nuclei in both the nerve-fibre layer and the middle nuclear 
layer, and from the condition of the inner reticular layer, I 
take to be that of an old individual (see fig. 32, a). Itis hard 
to believe that such “ Miiller’s fibres” as these were not per- 
manent structural elements ; if they were they had become so 
only during life, and to meet special functional requirements, 
for in the normal healthy retina of a man of forty-eight, 
referred to above, hardly a single straight radial strand can be 
found through the whole inner reticular layer. Faint zigzag 
streams alone occur here and there (fig. 32, 6), but are not 
numerous. ‘hat there should be no pronounced “ Miiller’s 


' As this absorbed matter streams through all parts of the retina (except 
the blind spot), and during a lifetime of functioning, it is clearly a factor 
which no student of the vitreous humour can afford to ignore. It suggests, 
for instance, a new and very simple explanation of Stilling’s canal. 

2 Purchased many years ago from Messrs. Watson, of Holborn, in a series 
of slides to illustrate the structure of the eye. 


64 H. M. BERNARD. 


fibres,” i.e. streams of matter passing through the inner 
reticular layer, in this eye, is just what we should expect, in 
view of the fact that for some days prior to excision it had 
not been exposed to light; but it is surprising that there 
should be no traces of any permanent rearrangement of the 
cytoplasmic reticulum so as to form continuous radial lines. 
It is possible that this only takes place in very old eyes, when 
both nuclei and cytoplasmic framework, all but its radial 
strands, seem to be disappearing (cf. the general condition 
of the inner reticular layer in fig. 32, a, with that in 32, b). 

In the baboon’s retinas, through which an enormous 
quantity of matter can be seen to have been passing, and in 
which the large inner limbs are filled with the same matter, 
all of it the same colour as the pigment, the conditions are as 
follows :—In the youngest retina (three months) thick 
yellowish-brown streams in immense numbers pass radially 
through the compact middle nuclear layer, but when they 
reach the inner reticular layer by far the greater number 
disappear; the few which seem to run straight through 
that layer, on examination with a high power, are seen to have 
avery zigzag and interrupted course. On the inner side of 
this layer dense streams again form and run towards the inner 
limiting membrane, expanding and losing their intensity 
before reaching it. 

In this young baboon’s eye, then, there are no clear radial 
arrangements of the fibres of the inner reticular layer which 
could, even under the most strained interpretation, be re- 
garded as sustentacular. 

In an adult male baboon the same is true, only the pigment 
is blackish. We again see what was described above for 
other retinas, that the reticulum of the inner reticular layer 
has become much coarser than in the younger eye, and 
consequently the zigzag of those streams which run con- 
tinuously through is much more pronounced. 

In an “old, very large male” the streams are still fewer 
in the inner reticular layer, apparently because every strand 
is clotted with pigmented matter, as is also every strand and 


STUDIES IN THE RETINA. 65 


membrane between this layer and the internal limiting mem- 
brane where the dark brown of the pigmented matter is very 
dense. Individual streams can hardly be followed. 

In these eyes, then, again, and in spite of the quantity of 
the pigment absorbed, we find the same difficulty as we found 
in the cod (ef. fig. 21, 7.7.) in establishing direct radial paths 
for the escape of the absorbed matter through the cytoplasmic 
reticulum ; that such paths do occur and may be very highly 
specialised we know from the (presumably old) human retina 
shown in fig. 32,a. These, seen alone, certainly appear as 
if they were sustentacular. 

Other perfect radial tracks seem to occur normally in the 
Amphibia, for as soon as the eye begins to function, that 
is in quite young tadpoles, there arise distinct, smooth, 
nearly straight radial fibres through the inner reticular 
layer, and these become so tough in preservation that 
they can be isolated intact if a section 1s teased up or crushed 
on aslide.! Further, in tangential sections they often appear 
running through the inner reticular layer within a clear 
passage. It is possible that this clear passage may be delu- 
sive, and due to the fact that the adjacent parts of the 
reticulum are drained by them of any matter which would 
render their delicate cytoplasmic membranes or threads 
visible. Compare with these apparently clear courses of the 
“ Miiller’s fibres ” through the inner reticular layer, fig. 29, b, 
where “ Miiller’s fibres” of a rabbit are shown cut trans- 
versely, and the tangential threads or membranes of the 
inner reticular layer thickened with matter are seen to be 
running into them. 

But the important contrast comes later. The establish- 
ment of direct radial streams through the inner reticular 
layers in the young tadpole is quite natural, for we remember 
that in other eyes the nearest approach we found to a straight 
course was in young retinas (see fig. 24,b) when the inner 
reticular layer is a close “ Punktsubstanz.” But whereas the 

' Many of them with nuclei of the middle layer attached to them (see 
fir. 25, c). 

voL. 46, pART 1.—NEW SERIES. E 


66 H. M. BERNARD. 


streams in these fish retinas become more and more zigzag as 
the meshes of the inner reticular layer get larger and coarser,in 
the frog and toad, for some reason or other, the early straight 
paths appear to become fixed. Whether this can in any way 
be correlated with the other peculiarity pointed out in these 
amphibian retinas, viz. that the absorbed matter passes by, 
apparently without entering, the inner limbs of the rods, which 
consequently remain very small, we are not yet in a position 
to decide. It is, of course, quite possible that the physical 
condition of the absorbed matter not coming in contact with 
the staining matter in the inner limbs might be different, and 
consequently its action on the cytoplasmic framework of a 
retina might also be different. 
It need hardly be pointed out that if a group of such 
streams as those shown in fig. 31, a,b, flowing through the layer 
of rod nuclei were to combine in or just after leaving the outer 
reticular layer (0.7.), and then flow on as one thick stream 
through the middle nuclear and the inner reticular layers, we 
should have the most developed type of “ Miiller’s fibre,’ such 
as that shown in fig. 32, a. Itis these most developed streams, 
looking as ifthey were permanent structural elements inthe eye, 
which have alone been regarded as typical ‘‘ Miller’s fibres.” 
Had all the minor forms of the same streams received equal 
notice, the error could never have been made of ascribing to 
them any fixed morphological significance. Such a wider 
survey would also have saved Borysiekiewitz, to whose works 
on the retina I should like here to express my indebtedness, 
from his conclusion that the “ Miiller’s fibres” are tubes 
conveying the nerve-fibrils to the rod layer.! Only these most 
developed strands which seem to rise directly from the rods 
could possibly supply the necessary conditions, and, if this 
conclusion were correct, we ought to find such developed 
strands in all and throughout all the retinas of the whole of 
the Vertebrata. This, as we have seen, is very far from 
being the case. Equally mistaken, too, are the conclusions 
based upon the impregnation method. In Ramon y Cajal’s 


1 “Weitere Untersuchungen,’ Leipzig und Wien, 1894. 


STUDIES IN THE RETINA. 67 


well-known figures of “ fixed morphological elements” re- 
vealed by the method, we find not only the “dendrites,” but 
also the ‘‘ Miiller’s fibres ” in their most developed form, and 
the “rod fibres” with their terminal swellings all equally 
clearly shown. The interpretation which we have put upon 
the latter two makes it more than probable that a proportion 
at least of the “‘ dendrites”’ are also nothing but the parts of 
streams already so frequently alluded to in the foregoing 
pages. I say a proportion of the “dendrites” for reasons 
which will be made clear in another paper, in which I shall 
also show that the nerve-paths through the retina can be 
demonstrated by ordinary methods of staining, and that they 
have no connection whatever with the “ dendrites.” 


SuMMARY. 


As the results so far attained in the preceding five parts of 
these studies are largely hidden under a mass of minute histo- 
logical detail, it is better, at this stage, that a summary be 
given of the more important. In the next paper we shall deal 
with the question of the nerves, which naturally has a much 
wider bearing than any detail of retinal structure merely as 
such. 

The conclusion which of all others now arrived at is of 
widest significance from a general point of view, is that the 
retina can no longer be regarded as built up of so many 
separate “cells,” each with some definite and permanent 
morphological value. his view, which has always been taught 
hitherto, has recently to all appearance been strongly confirmed 
by means of the metal impregnation method. This appears 
to reveal several distinct types of cells mainly distinguishable 
by their positions and by the different forms assumed by the 
ramifications of their respective cytoplasms. It is now main- 
tained, indeed, that these cells, to which special names have 
been given, have distinct and definite functions, so that if one 


68 H. M. BERNARD. 


single one were absent, a blind spot would ensue as a neces- 
sary consequence. 

The results here published, obtained solely by comparisons 
not only of different eyes but of the same eye at different 
ages, involve a direct contradiction to this interpretation of 
the phenomena. If there ever were distinct cells com- 
posing the retina, their walls were early lost.1 The func- 
tional retina is a continuous cytoplasmic reticulum 
in which nuclei are suspended, and the nuclei are 
not stationary. (1) A large proportion of those which are 
present in the young retina move outwards when it begins 
to function to become the nuclei of the new rods required by 
growth. (2) Their places are supplied by others migrating 
inwards from the rim. (3) The outward movement continues 
as long as life lasts, for in old eyes the nuclei of both the 
innermost and the middle nuclear layers are found to have 
largely disappeared. Whether 3 is for the supply of 
new rods or for some regenerative process we have no 
means yet of deciding. ‘These migrations, and especially 
this using up of the nuclei, in a retina which is all the while 
functioning normally, shows clearly that some other value 
must be assigned to its structural elements than that which 
is needed by the neuron theory as applied to this organ. It 
is clear that these nuclei are not the nuclei of cells taking 
part in fixed morphological chains, every link of which 
is essential. ‘he nearest approach we obtain to anything 
like a permanent cell in the retina is the rod with its nucleus ; 
that it would be inaccurate to persist in using the term 
“visual cell” in this connection will be conclusively shown 
in my next paper. 

With reference to the retina itself as the specific organ of 
vision, by far the most important result obtained is the 
discovery of some new details relating to the origin and 
structure of the rods, that is of those structures which are 
peculiar to the retina as the visual organ. According to the 


1 What appears to be the gradual dissolution of cell walls may often be 
seen where the young retina is passing into the cells of the iris. 


STUDIES IN THE RETINA. 69 


usual description they are of the nature of cuticular forma- 
tions. ‘This is a very natural summing up of the facts— 
(1) that they are almost certainly the end organs of the 
nerves, and (2) that their tips are filled with refractive 
matter of the nature of keratin. But the parallel with 
cuticular cells, although justifiable, is not very close. As 
protoplasmic vesicles thrust out against the pigment cells 
they absorb the pigment granules and (unless the quantity 
absorbed be too great, and its colour too intense) clarify them 
somewhat as the stratum lucidum of the epidermis receives 
and clarifies the pigment brought to it through the skin. 
Here, however, the parallel ceases, for while the cells of the 
cuticle perish with the waste matter they receive, and ulti- 
mately fall away as horn-cells, the rods get rid of their refrac- 
tive contents, which stream away through the retina. 

The working out of the finer structural details of the rods, 
taken up where the subject was left by Max Schultze thirty 
years ago, need not be repeated here, but one or two of the 
more important corrections of the current doctrine may be 
mentioned. 

What are called the ‘‘cones” of the vertebrate eye, to 
which special functions distinct from those of the rods have 
been assigned, are not always analogous structures. 

In the Amphibia they are the early stages in the forma- 
tion of new rods, and their form-phases are due to the 
squeezing of new vesicles between the already existing rods. 

In the fish analogous stages appear in very young eyes, 
but in older eyes the inner limbs of the earlier formed rods 
swell to such monstrous sizes that the conditions of the rod 
layer are altered, and the protrusion of new vesicles can no 
longer result in the formation of the same cone stages. 
The rods with the swollen inner limbs have been regarded as 
“oiant cones,’ although presenting no analogy whatever 
with the cones in the frog. 

In the Primates, what are usually called the cones are, as 
in the fish, merely rods with swollen inner limbs. In the 
centre of clear vision, where the pigmentary matter is 


70 H. M. BERNARD. 


absorbed in large quantities, all the elements are permanently 
of this character, but away from the centre only one here 
and there has its inner limb enlarged. Borysiekiewitz refers 
this to the protrusion of the nucleus, but as the nucleus is 
not always protruded, I prefer to refer it to an extrusion of 
fluid from the retina. Not only does the early protrusion of 
fluid vesicles from the retina in the first stages of rod- 
formation make this probable, but also the fact that globules 
of fluid are continually escaping from the retina into the 
rods, as described and figured in Part II. 

The striation of the rods, which has long been known, has 
now been traced to its true cause, viz. the existence of 
strands, sometimes taking stain, in the walls of the rod 
vesicles, while the lumina of these vesicles are occupied by a 
staining network in connection with these strands. 

The refractive matter which fills the outer limbs 
of the rods is absorbed pigment, which is usually, 
but not always, clarified during the process of 
absorption. ‘he correlation of this with the results of the 
classical researches of Boll, Kuhne, Ewald, and others I am 
not in a position to work out, for reasons given in Part V. It 
must be left to time, on the one hand, to show where we 
mutually confirm one another, and, on the other, to eliminate 
our respective mistakes. Had I commented on all the results 
obtained by previous workers whenever they overlapped the 
subject in hand, these papers would have been lengthened 
out indefinitely ; as itis, the histological details given in them 
have had to be limited to a small selection of those available, 

‘he curious zone formation within the retinal syncytium 
has been traced largely to the above-mentioned lateral move- 
ment of the nuclei of the middle layer from the rim towards 
the centre. 

The “ Miiller’s fibres,” however startling they may appear at 
their highest development, are merely streams of the pigment 
matter which have been absorbed by the rods, and which, 
with many interesting variations of detail, pass inwards 
through the retina, eventually to join the vitreous humour. 


STUDIES IN THE RETINA. 71 


EXPLANATION OF PLATES 3—5, 


Illustrating Parts III, IV, and V of Mr. H. M. Bernard’s 
paper on “Studies in the Retina.” 


N.B.—The measurements of the different eyes can only be approximate, 
because the shape is not always kept in very thin sections. If should be 
further noticed that sometimes a slightly older eye may be smaller than one 
obviously younger, a fact to be attributed to the accidents of nutrition. 


In all the figures m./. = limiting membrane, g./. = ‘‘ ganglionic cell” 
layer, a7. = inner reticular layer, m.z. = middle nuclear layer, 0.7. = outer 
reticular layer, 0.z. = outer nuclear layer. 


Fic. ].—Frog tadpole (Perenyi’s fluid). Eye diameter 0°32 mm. Part 
of section showing the spindle-shaped nuclei of the undifferentiated rim of 
the retina, attached to either the inner or the outer limiting membrane, and 
arranged on the axial side in curves bulging towards the axis of the eye. 
The arrow indicates the direction of the nuclear stream. A few of the 
nuclei already in the middle layer, selected because attached by trailing cyto- 
plasm to the inner reticular layer. 

Fig. 2.—Toad tadpole (Lindsay-Johnson’s fluid). Eye diameter 0°528 mm. 
Part of section drawn with the camera lucida, to show the attachment of the 
inner reticular layer to the membrana limitans interna, this connection being 
apparently due to the nuclei trailing their cytoplasmic attachments behind 
them as they travel towards the axis of the eye. These nuclear attachments 
tend to accumulate on each side of the stream, but persist as an accumulation 
only on the inner side (see text, pp. 1O—13). 

Fic. 3.—Frog tadpole (picro-sulphuric and iron hematoxylin). Eye dia- 
meter 0°24 mm., to show a younger stage in the formation of the inner 
reticular layer. The nuclei, which in the central region are loosely arranged 
where this layer is beginning to form, are densely crowded, five to six deep, against 
the pigment, aud are appareutly pressing inwards from the undifferentiated 
rim where nuclear divisions (&.) are taking place. The arrows indicate the 
direction of the streaming; 0.7. indicates a line of dark staining matter where 
the future outer reticular layer will run. 


Fic. 4.—From the other eye of the same animal. In both these eyes yellow 
fluid was apparent in the pigment cells, and here and there also apparently in 
vesicles which appear among the pigment granules, and were probably pro- 
truded from the retina (cf. Fig. 18). 


72 H. M. BERNARD. 


Fics. 5 anp 6.—Frog tadpole (picro-sulphuric). Still younger eyes showing 
earlier stages in the formation of the inner reticular layer as a kind of splitting 
of the nuclear ranks into two divisions, those forming the larger division 
crowding outwards against the pigment, leaving a loose, matted, and staining 
reticulum in the space from which they have moved. In both sections it is 
noticeable that nuclei in the very centre have even gone from the innermost 
layer. In Fig. 6 the cornea (c.) is seen thinning and clearing of pigment over 
the axis of the eye, and a nuclear division (4.) is seen near the centre of the 
retina, the two facts together indicating that the eye was only just beginning to 
function. 

Fic. 7.—Frog tadpole (picro-sulphuric). Eye diameter 0°20 mm. Shows a 
still younger stage (i.e. smaller, and with larger cavity in the lens). The 
crowding outwards of the nuclei in the optic axis not yet appreciable ; the 
beginning of the split among the loosely arranged nuclei is, however, indicated 
by an accumulation of vacuolar reticulum along the line occupied later by the 
inner reticular layer; nuclei seem also to be breaking away from the innermost 
layer, that forming the later so-called “ ganglionic cells.” A slight curving of the 
lateral nuclei, like that shown in Fig. 1, is alsoseen. The yolk granules which 
obscure the section are not indicated either in this or in the last two figures. 

Fie. 8.—Frog tadpole, from Table Mountain (Perenyi). Hye diameter 
0°S mm.; cross-sections of rods showing deeply stained internal reticulum ; 
this changes its pattern when the focus is changed. The reticulum is some- 
times forced to the sides by a refractive greyish mass, which at times may 
have a brownish centre of the same colour as the pigment. In these cases 
the reticulum frequently comes again into view on changing the focus. 

Fic. 9.—From the same retina, showing the distal ends of the cones as 
vesicles, often torn, but nearly always leaving ragged proximal ends still 
attached to the conical tips of the staining portion; in other cases the vesicles 
are complete, and their distal ends are immersed in pigment ; they are shrunken 
and often beaded with rows of dots. On comparing the elements marked with 
an asterisk and numbered 1 to 7, we can trace the transformation of a cone 
into a Schwalbe’s rod. 

Fie. 10.—The same, in which the relations are shown more completely. 
In both these figures the continuation of the vesicle membrane into that 
enveloping the “cone”’ is quite distinct. On the right is a new element with 
no staining proximal portion yet visible (cf. Part I, Pl. 3, fig. 2, a). 

Fic. 11.—The same, showing the more direct transition of the ‘ cone ” into 
the rod, nearer the side of the retina where the elements are shorter (namely, 
20 p instead of 45 to 50 p, as they are in Fig. 10). In this case the distal 
portion of the cone or new rod was visibly striated, which was not the case on 
the very young and still slightly swollen Schwalbe’s rod shown in Fig. 9 on 
the extreme right. 

Fic. 12.—From the same retina as Fig. 1. A group of elements in the 


STUDIES IN THE RETINA. 73 


early vesicular stage, the large vesicles not destroyed by the reagents. The 
exact relations of the nuclei cannot be made out. The nuclei of the two 


youngest vesicles may, perhaps, be those shown in or on the outer reticular 
layer. 


Fic. 13.—A few elements from the same, selected to show some of the 
form-changes from vesicles to rods. Most remarkable are the ** cones’ shown 
on the left (a, 6,c). Their tips are quite clearly within vesicles. The pheno- 
menon is explained in the next figure (14). The rod on the extreme right 
shows one of the bright staining globules referred to in Part II. 


Fie. 14.—A diagram to explain the “ cones” shown in Figs. 13 and 15, in 
which their tips are thrust into terminal vesicles. The pressure of new pro- 
trusions acting in the direction of the arrows converts the vesicle a into d 
and ¢ with a progressively lengthening neck. The staining matter which 
accumulates in this neck (see the central element in Fig. 12) is then thrust 
outwards (1) by the outward movement of the nucleus, and (2) by a fresh 
discharge of material (d). These “cone” tips, therefore, have no other 
membrane than the vesicle into which they are thrust. They are therefore 
sometimes disintegrated, and without defined outline. 


Fic. 15.—Frog tadpole (Perenyi). Hye diameter 0°-4 mm. Still shorter 
and stouter elements showing the same phenomenon, taken at different dis- 
tances from the centre of the retina. The formation of short thick rods out 
of vesicles can be easily understood. 


Fic. 16.—From the retina of a frog tadpole (picro-sulphuric, safranin). Eye 
diameter 0°24 mm., showing early protrusion of vesicles against the pigment. 
Deeply staining yolk granules are shown here and there; they are left out of 
Figs. 5 and 6. ‘Iwo nuclei are shown in the figure not yet in contact with the 
membrana limitans externa. They are selected because they show vacuoles 
inside. Intra-nuclear vacuoles and vacuoles extruded within the retina may be 
seen in all the sections of young eyes; the earliest phase of the inner reticular 
layer looks, indeed, like an aggregation of such intra-retinal vacuoles (cf. Fig. 7). 


Fies. 17 anv 18.—From the retina of another frog tadpole of about the same 
size (from 0°20 mm. to 0°25 mm.). Portions of sections differently magnified, 
showing more vigorous protrusion of vesicles against the pigment. ‘Those 
quite in the pigment are often yellowish in colour, while those nearest to the 
nuclei are clear white, partially framed round with densely staining matter 
(iron-hematoxylin). In Fig. 18 the vesicles marked with asterisks were 
yellowish. In other slides the yellow fluid, which here appeared in vesicles, 
was certainly inside the pigment cells as well (cf. Fig. 4). 


Fic. 19,—Frog tadpole (Perenyi). Eye diameter 0°8 mm.; a group of 
nuclei crowded outwards. The membrana limitans externa was not distinguish- 
able; its probable position is indicated by the marks of interrogation. Where the 
crowding was seen, the nuclei of the middle nuclear layer was diminished from 


74 H. M. BERNARD. 


four deep to only three (cf. the observation of Borysiekiewitz described in the 
text, p. 2). 


Fic, 20.—From the viviparous blenny (Bles’ fluid). a. Section of young 
retina, showing the early stage, with ordinary cones, Schwalbe’s rods, and two 
adult rods forming a double rod. Note the layer of cells at 0.7. and the dark 
zones in 7.7. The nucleim.z. were about eight deep. 4. From old specimen of 
same. The rods in two kinds only—(1) double rods with enormous inner limbs 
(22 p by 144), “ giant cones,” and (2) Schwalbe’srods, ‘The layer of nuclei at 
o.r. has disappeared, and most of those from m.z. The 7.7, is coarse-meshed, 
and very few strands run radially through. c. From the same, to show the 
tight packing of the swollen inner limbs, or “ giant cones.” 

Fic. 21.—Section of the retina of a very large old cod (corrosive sublimate). 
A“ giant cone,” with inner limb 42 by 22 p, and filled with refractive globules. 
Crowds of Schwalbe’s rods packed in between the ‘ giant cones.” The rod 
fibres end in terminal conical expansions against the or. The nuclei (m.z.) are 
nearly all gone, while the m.v, layer itself is choked up with solid sheets of amor- 
phous, finely granular matter, some of which is seen streaming away in a zigzag 
through the ¢.7. ‘The nerve-strands (z.s.) are clotted with finely granular matter. 


Fic. 22.—a—e. Section of the retina of a small plaice, 6” (Flemming), to 
show three parts of the same section, a being nearest the centre. To illustrate 
the gradual using up of the ‘‘ ganglionic cells”’ (g/.) and of the nuclei of the 
middle layer (z.z.) ; the disappearance of the continuous layer at 0.7. (the nuclei 
of Krause’s “ Membrana fenestrata’’), and the increase in the rows of rod 
nuclei; the gradual accumulations of granular matter in the m.z., and the 
swelling of the inner limbs, 

Fie, 23.—a, 6. From a very old plaice. a. Radial. 4. Tangential, showing 
the enormous accumulation of amorphous matter, extending in this old fish 
almost completely round the retina. 

Fic. 24.—a. Sections of retinas of Loch Leven trout, ca. 7’” to 8” (Perenyi). 
b. Section of very young trout, var.? To show the change of the ¢.7. from 
early “ Punktsubstanz,” with fine zigzag “ Miiller’s fibres’ running radially 
through it, to the coarse-meshed older condition, and with the threads thickly 
coated with matter in the outer zone. Similar thickened strands are seen just 
within the 0.7. 


Fie. 25.—From tadpoles from Table Mountain, and characterised by great 
abundance of pigment. The inner limbs remain small, aud the “ Miiller’s 
fibres’ rise straight from the walls of the rods. In é.d.e., just above the 
rod nuclei, are seen the dark-staining bodies which seem to correspond 
with those seen within the expansions of the ‘‘ rod fibres ” seen at. 0.7. in some 
of the foregoing figures, e.g. 20, 6; 22, a, d (cf. text, p. 32). 

Vic. 26.—a, d. From the retina of an old cat (13 years) (lemming), to show 
the inner expanded ends of the “ Miiller’s fibres,” containing accumulations 


STUDIES IN THE RETINA. 75 


of amorphous matter. In 6, exactly similar bodies arise entirely 
from the inner edges of the zr. The trumpet shape is due to accidental 
compression of the section. 


Fic. 27.—From a very old dog, showing the same as Fig. 26, 4, i.e. the 
inner expanded ends of the so-called ‘‘ Miiller’s fibres” arising entirely from 
the ir. Matter obviously belonging to the vitreous humour is seen flaking off 


the membrana limitans interna. 


Fig. 28.—From a rat, 3 years old (Flemming). a. To show the flaking off of 
matter from the membrana limitans interna; the flakes show a delicate texture ; 
the ends of the “* Miiller’s fibres ” filled with amorphous matter, and rising from 
various depths in the z.7, 4. To show the thickening of all strands and mem- 
branes indifferently, with matter coming from the 2.7. e¢ shows ‘ Miller’s 
fibres” both losing themselves and starting again within the ¢.7. 


Fic. 29.—From a rabbit (Hermann). a. Froma tangential section, to show 
a portion of the membrana limitans interna divided into differently darkened 
areas representing the covers of the conical expansions of the ‘* Miiller’s fibres.” 
6. Two conical expansions with cross-sections showing the “ fibres ” as solid, and 
with the tangential strands of the ¢.7. running into them, 


Fic. 30.—From a mouse which had been exposed to an electric arc lamp. 
A thick strand of amorphous matter winds its way through the cytoplasmic 
reticulum just inside the o.7., eventually giving off, and bending up into 
typical ‘ Miller’s fibres.” 


Fie. 31.—From a chacma baboon. «@. From an old male, with dense 
black pigment, which not only half fills the large inner limbs of the rods, but 
streams inward as far as o.r, without changing colour. J. From a young 
specimen, 3 months old, with yellowish-brown pigment, which also streams as 
far as the 0.7. without changing its colour. 


Fie. 32.—From the human retina. @. From a purchased preparation ; appa- 
rently of an old retina, showing au enormous development of the ‘ Miller’s 
fibres,” a very coarse inner reticular layer, and very few nuclei in the mz. 
5. From a “normal healthy ” retina of a man 48 years old, which had appa- 
rently been little used for some time prior to excision. Typical ‘‘ Miiller’s 
fibres” wholly wanting. Faint streams, often zigzag, and parts of 
streams alone occur, many of them processes from the so-called “ ganglionic 
cells.” The m.z. was five nuclei deep; compare the number in a. But for 
this latter comparison to be of weight we ought to know how far from the 
centre of the retina the parts shown were severally taken. 


ee rey, See ere 
. j “ 

Pele bed i) > asi le Le, © Vilwee OO \Gho esse iii 
ST en, re a ar abd = dhe evaht FaNGt od 
; r gg abe al poe 

eid one » gs i 2 . 2 @)iiipe ene: : Be Ut « wee 

aT: (os) . : JL) poy Leen o@-eebd bo ddieey Foetal 

ee ee Se wi ho Gay of iit eaaibes! eo owhwthy bel alle 

yeti) bei Renate pet 


= i ,; ; (Aree pal ytiad Uh @ ma € Ge, 

ince pieve. wad ( O08 1000 alr! ee eer ee 
Hives Sas , ek p-ilereiiawimes ala’ wee iY wary 
ann Qa, uw : mee Gia Ji de i 4,1 i.1Gh8 


pedi fi ad : ; ows iain fre quant bik 
| i) cali @-< +) 5 hte ter seri) ¢ pent aod; 

Wolly € inet : WP « reay> (0) (leg (ineahl ae Ae 

ain Paste « ipa 4 — 9 2 ats cited off paeG 
sa tick a ee ee Hythe 


i> cas 416 mah 7° ‘ Oe - ¥ naerren Stee 


figs sr Stila at eee OAS 
ip? te « ; ‘ ¥ ue» cc Wei; 
oe i ty ! A Li 
) GW n= ; : y } -T 
7 ") (apa 
oa) i ini iii i ia} Y ate %e delay 
} BAe . ii: > 34,40) «4 lapel 
ee a 24 Gay eer Ra lw hi eae 
; stil CON qeatgandes Mrs ILe ied 
part wh cal eR (fiw wisest , aA Peli vis i) ret = 3éi 
al MM 240. @ & --)bo9oh) Cm ale ey, pepe (ancy be 
: OG 1A - st epee’ wrt, een Ss ent B3 
pie Cetin eae, Pemagh eee e Ra ly jukeod tires ee 


eth. -) Waeeryg? walaere.eeri = £295 )(43 Vem, 11 or I 
dey |» oem Oe OI! (ee “erien © EDR, 


ku Ags j ted CLL eerie 2; iil ne ly Fee GER 
ae 7 i ip Vgbag £409 Si ‘eo ih | sith cae spa fH pee 
_— Tih a i icCé a pee PU a ee in a) Pai hg peer 


» ld a igi iaG Sikes 


= ; : Pnwilin Qh be 


— 
- 


RELATIONS OF KIDNEYS IN HALIONMS TUBERCULATA. 77 


Notes on the Relations of the Kidneys in 
Haliotis tuberculata, etc. 


By 


Hi. J. Fleure, B.Sc., 
U.C.W., Aberystwyth ; Fellow of the University of Wales. 


With Plate 6. 


Noumerovs accounts of the structure and relations of the 
kidneys of Diotocard Gastropods have been written, chiefly 
by workers interested in the question of the homology of the 
Monotocard kidney, but in the various descriptions occur 
several contradictory statements, which naturally lessen the 
value of current theories on the subject. The present paper 
is the first-fruits of a study of the gastropod kidney, and it 
endeavours, by throwing new light on one or two disputed 
points, to help on the solution of this difficult problem. 
There are also brought forward certain suggestions concern- 
ing the kidney and reproductive organs of Monotocards which 
I feel convinced should be studied together; these sugges- 
tions are naturally extremely tentative, pending further 
work. In making this communication I wish to express my 
deep indebtedness to Professor Ainsworth Davis, whose 
advice and encouragement have alone made this research 
possible. 

Haliotis possesses structures right and left of the peri- 
cardium (8), which, notwithstanding various views concerning 
their relations, are generally allowed to be the representatives 
of the right (7 x) and left (7 1) kidney of the primitive Dioto- 


78 H. J. FLEURE. 


card. The structure on the left of the pericardium is a small 
sac, whose walls contain, in parts, lymphatic tissue; it is 
doubtfully renal in function, and is called the papillated sac. 
That on the right side is the functional kidney, and it 
possesses various large lobes, including a long anterior 
one (A. L.) stretching forward on the left flank of the great 
shell muscle. All workers also find two openings at the back 
of the mantle cavity : one (1) placed definitely on the left side 
of the rectum, the other (2) further towards the right. 

Von Jhering (1), Perrier (6), and Weemann (4) consider 
that the right opening (2) is the orifice of the functional 
kidney (7 R), and the left one (1) that of the reduced left 
kidney (71) or papillated sac. They state that the gonaduct 
(from 6 R) opens into the right kidney (7 Rr). Haller (9) 
found that the right kidney (7x) communicated with the 
other (7 1), and that both opened by the orifice (1) of the 
left kidney, the right orifice (2) belonging solely to the 
gonaduct. My observations on these points agree with those 
of Perrier, who also worked with Haliotis tuberculata. 
Haller used H. glabra, and it would be necessary to study 
that species before rejecting his statements. Perrier, Weg- 
mann, and Hrlanger (8) described a pericardial communica- 
tion for the papillated sac, but not for the functional kidney ; 
while Haller found a funnel opening from the pericardium 
into the right kidney, but was unable to discover an internal 
orifice to the papillated sac. The former result has been 
generally accepted, and it is quoted as evidence that the 
right kidney is really degenerating in the Rhipidoglossa, and 
that the kidney surviving in Monotocards is therefore the 
Diotocard left (71). The methods adopted by these authors 
were the usual ones of dissection, injection, and section 
cutting, but in this case there seems to be another line of 
investigation. It is well known that the Diotocard female 
liberates her ova in an irregular fashion a few at a time into 
the mantle cavity. Now, if the gonaduct communicates with 
the kidney, ova might be found at a certain season in the 
cavity of the latter; and, further, if this kidney has a peri- 


RELATIONS OF KIDNEYS IN HALIOTIS TUBERCULATA. 79 


cardial opening stray ova might even make their way through 
it. It therefore seemed to me advisable to examine specimens 
of Haliotis taken during the breeding season. 

Haliotis tuberculata may be obtained in fair numbers 
round the coasts of the Channel Islands, more especially on 
the rock-strewn shores of Guernsey and Sark, where it passes 
its sluggish life attached to the under-side of large boulders. 
It therefore lives with ventral surface uppermost, and is said 
by the fishermen often to die if removed from its attachment 
and left in the reverse position. It frequents the upper part 
of the Laminarian zone, and seems to feed largely on small 
aleze. ‘lhe breeding season in this locality I have found to 
extend from about the end of December to the middle 
of February, and the specimens used for this investigation 
were collected in Guernsey during the spring tides of that 
period. They were soaked in 5 per cent. formalin, and 
mostly examined within a few days of their capture. The 
specimens were carefully taken out of their shells, and before 
they were placed in water or dissected at all their pericardia 
were opened on the left side well away from the kidney wall. 
The contents of the pericardial fluid were then examined, 
and found to consist mostly of corpuscles, a few epithelial 
cells, and sundries. One specimen, however, yielded a 
pinkish fluid, in which floated several ova; while two or three 
others also yielded each a few ova in the same way. ‘lhe 
ova are very different in appearance from the other peri- 
cardial contents, and from the components of the various 
tissues abutting on the pericardium. ‘They seem to retain 
their ovarian covering and a short stalk for a considerable 
time, very few having been found without them. The 
nucleus is prominent, and there is a small granular accumula- 
tion usually near the short stalk (see fig. 4, b). 

An interesting feature in the female at the breeding time 
is a characteristic pink coloration more or less diffused over 
the whole body, but most noticeable on the covering of the 
hepatic czecum, on the pericardial wall, on the head above 
the tentacles, and on the floor of the mantle cavity. 


80 H.f Jt) BEEGRE 


After sampling the fluid contents of the pericardium the 
contents of the right kidney were examined, and proved in 
several cases to be roughly divided into two sorts of material 
—darker brown fluid with excreta, and lighter coloured fluid 
containing both excreta and ova. Just behind the opening 
into it of the oviduct the kidney is partially subdivided by 
an internal projection of its right wall. The two parts are 
respectively an anterior one containing ova as well as excreta, 
and a posterior one containing almost solely the latter 
(see fig. 4). 

From these finds of ova it seems justifiable to conclude— 

a. That the gonaduct opens into the right kidney (6R 
into 7R). 

b. That the right kidney has a pericardial pore (5). 

c. That the anterior part of the right kidney is becoming 
connected more particularly with the reproductive system. 

Injections of the right kidney were rather unsatisfactory, 
as might be expected considering its large size ; but injection 
from the pericardium, on the other hand, showed very dis- 
tinctly a pericardio-renal communication (5) near the anterior 
right-hand corner of the pericardial cavity (8). 

By careful dissection of uninjected specimens from the 
pericardial side, an opening was found high up on the right 
wall of the pericardium, near its anterior right-hand corner. 
A fellow-student, Miss A. Ritchie, kindly confirmed this for 
me in another specimen. ‘I'he opening as seen in dissection 
seemed fairly distinctly lipped, and is possibly imperfectly 
valvular; it is situated near the point where the duct-like 
portion of the kidney may be said to begin (5). 

Further along toward the external opening, the wall of the 
functional kidney comes near that of the papillated sac; I 
have not, however, been able to find any interrenal communi- 
cation such as Haller describes for Haliotis glabra. 

Despite numerous attempts, by dissection, by examination 
of contents, and by injection, I have not been able to find 
evidence of a pericardial communication with the papillated 
sac, a conclusion in accord with that of Haller, but not with 


RELATIONS OF KIDNEYS IN HALIOTIS TUBERCULATA. 81 


that of Perrier and Erlanger. The wall of the sac is very 
thin, especially where it overlies the branchial vein. When 
it abuts on the pericardium it is also thin-walled over a part 
of the area, but the remainder is thickened into a mass of 
lymphatic tissue. ‘There are vascular connections taking 
blood to the efferent branchial vein and thus to the auricle 
direct. Perhaps some one of these is what has been taken 
for a pericardial communication of the papillated sac. 

To sum up, therefore, I think that Haliotis tuberculata 
has two separate kidneys right and left of the pericardium, 
opening externally by separate apertures (1 and 2). I find, 
also, that the gonaduct opens into the right kidney, which is 
the functional excretory organ, while the left kidney is partly 
degenerating into lymphatic tissue, and is becoming con- 
nected with the efferent branchial vein by direct blood- 
channels. So far my results agree with Perrier’s. I find, 
further, in opposition to Perrier, and in agreement with 
Haller, that the right or functional kidney communicates with 
the pericardium (via 5), while the left one does not. The 
evidence adduced is, in part, of a different nature from 
that brought forward by the authors mentioned. 

The foregoing results, if correct, lessen the divergence 
hitherto supposed to exist in this respect between Haliotis 
and Patella. The limpet has two kidneys right (7 R, fig. 3) 
and left of the rectum. The right kidney is very extensive, 
and performs most of the excretory work; it has several lobes, 
including a subrectal one (s. R. L.), which abuts on the wall of 
the pericardium. It serves as an exit channel for the repro- 
ductive elements, but evidence I have collected recently 
seems to hint at liberation of ova, at any rate, by rupture 
as well. This matter, however, needs further investiga- 
tion. 

The small left kidney is situated between the rectum and 
the pericardium, its circulatory system connects it intimately 
with the auricle. It is not shown in the diagram. 

The right kidney communicates with the pericardium (8), 
the opening (2) being in the floor of the kidney’s subrectal 

voL. 46, PART 1.—NEW SERIES. F 


82 Ho 33 FLEUR: 


lobe (s. R. L.), but opinion varies as to a pericardial pore of the 
left kidney, the latest statement being that of Mr. E. S. 
Goodrich (11), who makes certain he has found it. 

Mr. Martin Woodward has recently published a valuable 
account (14) of Pleurotomaria Beyrichii, in which he 
says that the right (7 R) and left (71) kidneys of that animal 
are in many respects comparable with the corresponding parts 
in Hahotis. The efferent duct of the right kidney (7 R to 2) 
is prolonged forwards, and has thick glandular walls in the 
female, so that it is practically an oviduct. Woodward 
found a pericardial opening and canal (4) for the left kidney 
but not for the right, a result which, if confirmed, makes 
Pleurotomaria an exception to the general rule. 

The most primitive type of kidney in Diotocards is, how- 
ever, that of Cemoria described by Haller (9). Here both 
kidneys are well developed and functional, each communi- 
cating with the pericardium (vid 4 and 5), and each receiving 
genital products from the gonad of its side. If this type is 
truly primitive we can, as Haller has said, derive from it the 
excretory organs of Pleurotomaria, Haliotis,and the Trochide, 
from which series the Docoglossa and the Fissurellidee would 
be fairly early offshoots. Throughout, the left kidney and 
the left gonad degenerate, whilst the right kidney becomes 
both the functional excretory organ and the exit channel for 
the sex products. The right kidney retains, in most forms, 
its pericardial pore. 

Perhaps the most marked contrast between the Diotocards 
and their Monotocard descendants is the presence of 
numerous accessory genital organs in the latter and their 
complete absence from the former group. In the latter, also, 
the reproductive and excretory systems are entirely separate, 
We must therefore seek out hints of the coming change 
among the ancestral forms. 

Mr. Woodward has shown that the excretory duct of the 
right kidney of Pleurotomaria is practically transformed, in 
the female, into an oviduct (vide fig. 2). 

In Haliotis the large anterior lobe (A. L., fig. 4) of the right 


RELATIONS OF KIDNEYS IN HALIOTIS TUBERCULATA. 83 


kidney is, practically, an accessory genital organ in posse. 
The external opening (2) of the right kidney is evidently 
becoming a genital pore. 

Trochus and Turbo vary very much in this respect, but in 
some species the “anterior lobe ” 1s very sharply marked off 
from the rest of the kidney. There is the same conflict of 
opinion between Perrier and Haller about the reno-pericardial 
funnels in T'rochus as in Haliotis, and Haller finds also an 
interrenal communication in Trochus gibberosus. With 
regard to these reno-pericardial funnels it is noteworthy that 
in all Teenioglossa, even in the most primitive (figs. 6 and 8), 
and in Nerita there is a well-marked communication between 
kidney and pericardium on the right side of the latter—a 
fact which strongly supports the view here put forward, that 
the right reno-pericardial pore is retained, as a rule, in the 
Rhipidoglossa. 

If the right excretory pore (2) becomes monopolised by the 
genital system, the functional kidney must find an exit for its 
excretory products; and it seems probable that this exit is 
through the external opening (1) of the left kidney, which 
would thus be the homologue of the Monotocard excretory 
aperture. The probability of this is increased by the fact 
that in no Monotocards have traces of a pore or sac been 
found to the left of the kidney opening. ‘This view, however, 
entails the further supposition that the mght kidney, or 
rather its posterior part, comes to communicate with the left 
kidney, and Haller claims, as was mentioned above, that such 
a communication already exists in Haliotis glabra and 
Trochus gibberosus. Perrier contradicts Haller, though 
he, too, supposes that the two kidneys come to communicate ; 
he, however, almost certainly errs in stating that the Monoto- 
card kidney opening is the right one (2) of Diotocards, for 
this statement raises a serious difficulty as to the homology 
of the genital opening. 

Bouvier found such an interrenal communication in Am- 
pullaria (5), and Perrier justifiably uses this observation in 
support of his views above mentioned. He further supports 


84. Hi. J: FLEURE: 


his conclusions by observations on the “renal gland.” He 


brings forward much evidence in favour of considering this 
gland, so generally found in Tznioglossa, as a modified vestige 
of the left kidney which has become intimately connected 
with the “ pericardial gland.” This gland consists of tubules 
which are lined by ciliated epithelium and open into the renal 
cavity (see fig. 9, b). The absence of this renal gland from the 
primitive Paludina suggests the hope that further work will 
reveal traces of the old left kidney in a less modified condi- 
tion. The probability that further work will result in the 
discovery of the above-mentioned interrenal connection in 
other forms is increased by the fact that such connections 
are by no means unusual in Mollusca. They exist in several 
Lamellibranchs and in Cephalopods, and, without presuming 
to suggest that they are homologous throughout, their occur- 
rence diminishes the improbability of their occurrence in 
Gastropods. 

The remaining problem is the derivation of the accessory 
reproductive organs of the Tzenioglossa and of their descend- 
ants the Opisthobranchs and Pulmonates. 

The male has, typically, a large penis (vp) at the right side 
of the head ; this organ is retractile in more primitive forms 
(Paludina, fig. 8), but permanently extruded in more special- 
ised forms (Buccinum, fig. 10). A very similar structure is 
found in the hermaphrodite Opisthobranchs, though their 
hermaphroditism has been shown by Pelseneer (10) to be due 
to the development of a male gonad in the female. 

Mr. J. E. S. Moore (12) found an archaic form in Lake 
Tanganyika, which he named after its abode—Tangan- 
yikia rufofilosa. The female of this animal possesses a 
brood pouch (8. P., fig. 5, a) on the left side of the head in the 
position of the penis of a male Paludina. He found the same 
structure in Melania episcopalis, and both also possessed 
a groove connecting this pouch with the genital aperture. 
This strongly resembles the spermatic grooves of some 
Opisthobranchs, and similar grooves also exist in the females 
of some Teenioglossa (figs. 8 and 9), while in the males the 


RELATIONS OF KIDNEYS IN HALIOTIS TUBERCULATA. 809 


spermatic duct (s.D.) running to the penis along the floor of 
the mantle cavity (fig. 10) or situated internally (fig. 8) has 
probably been formed by the covering in of such a groove. A 
groove of this kind is found in the male T'anganyikia, but the 
penis is absent (fig. 5, b). 

From these facts Mr. Moore has argued that the common 
ancestor of T’enioglossa and Opisthobranchs possessed some 
accessory reproductive organ which had probably become 
separated from the genital duct and remained connected 
with the genital opening by means of a groove, which tended 
to become covered over. ‘This accessory reproductive organ 
was somewhat variable, as it is lost in all Teenioglossate 
females except the two named, though the groove is retained 
in a few. In Opisthobranchs, which are originally female, 
and in male Teenioglossa this organ becomes the penis, while 
the groove is very often covered over and thus transformed 
into a duct. The development of such an accessory repro- 
ductive organ de novo is a difficulty further enhanced by the 
presence of the groove, but it is still premature, perhaps, even 
to suggest that possibly its ancestor is the anterior lobe of 
the Diotocard right kidney. It is interesting to note that 
Typhobia horei has a penis (fig. 7 P) which is extruded 
apparently vid the genital aperture, and which is placed as an 
anterior dilatation on the reproductive duct. 

To sum up, it will be most appropriate to give a brief 
statement of the views of previous workers and of the chief 
points raised in this paper. 

The first theory is that of Professor Ray Lankester. He 
believed that the excretory aperture of Monotocards is the left 
kidney opening of Diotocards, but he thought also, from its 
position with regard to the rectum, that the Monotocard 
kidney (7) was the Diotocard left (71). Since these views 
were stated, the supposed absence of a pericardial pore of 
the right kidney has been used as evidence of the degeneracy 
of this organ, and, consequently, in favour of homologising 
the Monotocard kidney with the Diotocard left. Hrlanger’s 
work on the development of Paludina is also quoted in 


86 Hisst PLEURE: 


support of this homology, but the structure which he takes 
to be the forecast of the Diotocard right kidney is merely a 
problematic and very transient vestige. The right kidney, on 
Lankester’s view, would become part of the reproductive 
system. 

Haller thought that the reproductive organs developed a 
duct which became continuous with the right kidney duct and 
opened through what was previously the right kidney 
aperture. The left kidney, he thought, degenerated, but 
became connected with the right, so that the Monotocard 
excretory organ was mainly right kidney, but opened through 
the left kidney’s aperture. According to him, only the right 
kidney retained a pericardial pore. 

Perrier agrees with Haller’s conclusions except as regards 
the pericardial communication, which, he holds, persists only on 
the left side. He differs also in a point of great importance, 
for he says that the Monotocard kidney opening is the 
Diotocard right. His principal contribution to the discussion 
is the tracing of the fate of the left kidney. This, he found, 
became the renal gland, consisting of tubules lined by 
ciliated epithelium, and opening into the kidney cavity. This 
gland was, he said, typically associated with that ancient 


ce 


molluscan feature, the ‘pericardial gland.” The conclusions 
supported in this paper are— 

1. Lankester’s view that the renal aperture of Monotocardsis 
the left one of Diotocards. This opposes Perrier’s conclusion. 

2. Perrier’s and Haller’s view that the twokidneys in some 
Teenioglossate ancestors came to communicate inter se. 

3. Lankester’s and Haller’s view that the rght kidney 
opening becomes the genital aperture. ‘This opposes Perrier’s 
conclusion. 

4, Haller’s view that the right kidney retains its peri- 
cardial communication in most Diotocards. This opposes 
Perrier’s conclusion. The evidence adduced is partly new. 

5. Perrier’s and Haller’s view that the Monotocard kiduey 
is composed of the right kidney of Diotocards, together with 
the cavity of the left (whose walls form the renal gland). 


RELATIONS OF KIDNEYS IN HALIOTIS TUBERCULATA. 87 


This I would slightly modify by stating that it is the posterior 
part of the Diotocard right kidney which seems to me to 
become the functional part of the Monotocard kidney. 

6. I would also suggest, in a very tentative fashion, that 
perhaps the forecast of the accessory reproductive organ, 
which becomes penis or brood pouch in the Monotocards, 
was originally a dilatation on the reproductive duct. Perhaps 
even an earlier condition of this organ is what has been called 
the anterior lobe in the Diotocard kidney. 

After outlining these conclusions I saw Mr. Martin Wood- 
ward’s paper on Pleurotomaria Beyrichii. Ina short 
discussion at the end of his paper he speaks in favour of con- 
clusions identical with Nos. 1, 2,3, and 5 above, but from his 
observations on Pleurotomaria he sides with Perrier as 
regards 4. liven, however, had i not found a reno-pericardial 
communication for the right kidney in Haliotis, I think there 
would still have been a balance of evidence from Fissurella, 
Patella, perhaps from Trochus, and especially from Nerita, 
Nassopsis, Paludina, etc., in favour of the view that the 
Monotocard reno-pericardial opening is that of the Diotocard 
right side. 

More work on the Trochide and Neritide is necessary for 
any further advance towards certainty in the matters above 
discussed, but it is interesting to note that Nerita has a 
single kidney with a well-marked pericardial opening on the 
right side of the pericardium, while the external aperture is 
to the left of the rectum. The genital system, described by 
Haller, is quite separate from the excretory organ, and lies to 
the right of it. The kidney has no anterior lobe correspond- 
ing to that of Halotis. The chief interest of the Trochide, etc., 
arises from the theory put forward by Perrier, Bouvier, and 
others that these forms are very near the ancestral stock of 
the Monotocards. Mr. Woodward, however, does not seem 
to share this opinion, 

It has been a difficulty to represent, in diagrams I1—10 the 
true relations of the anal, excretory, and genital openings, 
and Professor Davis therefore suggested to me the addition 


&8 A: cd. REEURE 


of schematic cross-sections of the mantle cavities of various 
forms, showing the rectum and oviduct or right kidney duct 
cut through, and the relative position of the left kidney or 
the Monotocard kidney opening. 

It will be seen that the openings of right and left kidneys 
in Diotocards have the same positions, with reference to one 
another, as the genital and excretory apertures in Monoto- 
cards. The relation of these openings to the rectum, on the 
other hand, varies to some extent in different forms. 

The most striking feature of these diagrams (11—15) 1s 
the migration of these openings to the animal’s right side, 
and we also notice the disappearance of the right ctenidium 
and the folding over to the left side of the originally right 
leaf of the other ctenidium. 

In the primitive Gastropods, as in Cephalopods, the in- 
coming streams of water entered the gill cavity on either 
side and bathed the gills, after which they made their way out 
along the median line of the cavity, taking away the excreta 
from the openings on this line (fig. 11). Later on, the left 
kidney degenerated, being perhaps partly pressed out of 
existence, and the right kidney became the sole functional 
excretory organ. ‘This process, already begun in Haliotis, 
was correlated with the disappearance of the ctenidium from 
the right side, and to the shifting of anus, renal, and 
reproductive openings to this side. The respiratory stream in 
such forms (figs. 12—15) would now come in along the left 
side and go out past the anus, etc., along the right, in- 
current and excurrent streams being thus freed from mutual 
interference. This more perfect separation, Professor Davis 
thinks, was the advantage which led to natural selection of 
variations along the lines of the changes just mentioned. This 
clockwise shifting of apertures and ducts has in some cases 
been continued so far that one or more have become situated 
along the extreme edge or even, in some cases, on the floor 
of the mantle cavity, so that their original relations appear, 
in horizontal plan, to have been reversed. Paludina shows a 
further modification, for here the reual opening is situated 


RELATIONS OF KLIDNBYS IN HALIOTIS TUBERCULATA. 89 


between the anus and the genital aperture (fig. 15), but this 
peculiarity may be connected with the development of a ureter. 


APPENDIX. 


Since the above paper was written I have had an oppor- 
tunity of seeing Pelseneer’s recent publication ‘ Les 
Mollusques archaiques,’ and as he touches here and 
there the questions of kidney homologies, a short discussion 
of the work of this eminent scientist will add to the com- 
pleteness of this little paper. 

The chief new facts which he brings forward all support 
the conclusions I have ventured to set forth. 

1. He finds that, in the Trochide, the right kidney has a 
communication with the pericardium, and, as I have found 
the same feature in Haliotis, arguments for the degeneracy 
of the right kidney of Rhipidoglossa, based on the absence of 
its pericardial communication, are definitely demolished. 
Pelseneer also finds the distinction between anterior and 
posterior regions of the right kidney of Trochus which has 
been mentioned in this paper as regards Haliotis. 

2. He finds that Haller’s interpretation of the kidneys of 
Cemoria was based upon errors of observation. As he has 
carefully examined specimens both from the source to which 
Haller had recourse (the Vettor Pisani collection) and 
from the White Sea, there seems to be little doubt that 
Pelseneer is right. If so, Cemoria, in its excretory and 
genital systems, resembles the typical Fissurellid. Its right 
kidney is of enormous extent, while the left is quite tiny, 
and only the right gonad is found. 

This result is of great interest, as it is now possible to say 
that in all known Rhipidoglossa the left kidney is either ex- 
tremely reduced or has undergone a transformation into 
a “papillary sac,’ an alteration which has profoundly 
affected its minute structure and its circulatory arrange- 
ments. It is therefore still more difficult than before to 


90 H. J. PLEDRE: 


homologise the left kidney of these forms with the single 
kidney of Monotocards. 

Towards the end of his paper Pelseneer briefly discusses 
this question of the homology of the Monotocard kidney. 

He first of all sets aside Haller’s views, as every sub- 
sequent worker has differed from that writer regarding the 
relation of gonaduct and kidney. Proceeding next to discuss 
Perrier’s theory, he points out the following weak points : 

1. It necessitates the supposition that the rectum and right 
kidney have undergone relative translocation. This he finds 
difficult to imagine. 

2. If we accept Perrier’s further conclusion that the Mono- 
tocard nephridial gland is the remains of the papillary sac, 
we are forced to assume that in some type the kidneys came 
to communicate inter se. Having examined Ampullaria, 

Pelseneer denies that such a connection exists in that type. 
Like the late Mr. Woodward, I cannot see that these 
objections are really vital. Perrier’s theory of the nephridial 
gland seems to have much in favour of it, as also has the 
idea that while the Monotocard kidney is that of the Rhipido- 
glossan right side the opening is that which formerly be- 
longed to the left kidney, but these are not essential to the 
theory. If, pending further evidence, we leave aside these 
additions, the whole of Pelseneer’s second objection dis- 
appears. 

As I have said further above, the frequent existence of 
the required interrenal communication in the more primitive 
molluscan classes seems to me to minimise the difficulty. 

The question of translocation of the rectum is raised and 
discussed in rather a new light in this paper, and I venture 
to think that the conclusions reached very markedly diminish 
Pelseneer’s objection. The rectum has undoubtedly shifted 
a great deal to the right, such shifting being far more 
important for the cleanliness of the ctenidium, and therefore 
for the efficiency of the branchial cavity, even than the 
rightward shift of the kidney. 

On the whole it does not seem too much to say that, not- 


RELATIONS OF KIDNEYS IN HALIOTIS TUBERCULATA. 91 


withstanding the views he expresses, Pelseneer’s results in this 
matter tend to strengthen the theory supported in this paper. 

If this theory be adopted it is possible to say, now 
Pelseneer has settled the “ Cemoria”’ difficulty, that in all 
Anisopleura the (post-torsional) right is par excellence 
the excretory side, even in those forms which still have a 
median anus. ‘his opens up a possibility of very great and 
far-reaching interest as regards gastropod morphology. 

Pelseneer’s theory of the gastropod twist is now generally 
accepted as amended in the matter of terms by Amaudrut 
and others. 

According to this the far-off untwisted ancestor had a gut 
going straight from front to back. This underwent — 

1. A ventral flexure, giving the gut a cephalopod-like 
disposition. 

2. A lateral torsion through 180° in a counter-clockwise 
direction, affecting all the animal except the head and foot. 
Asaresult of this torsion the branchial cavity and anus, 
previously postero-ventral, became antero-dorsal. 

It is generally allowed that the pre-torsional position of 
the branchial cavity militated against its efficiency in a form 
possessing a creeping foot, for it would be pressed down 
against the top of the foot by the weight of the shell. 
Natural selection, therefore, led to the upward shifting of 
the cavity by survival of upward variations of its position. 
As far as I am aware, no one has yet shown how it is that 
the twist is counter-clockwise, and this has been an undoubted 
weakness in the theory; but the difficulty is, I think, 
removed by a consideration of the excretory organs and 
ctenidia on the lines suggested in this paper. 

Let it be granted that in all Gastropods which have under- 
gone the torsion the right is the excretory side of the 
branchial cavity, the left being more particularly devoted to 
respiration. 

Then it is at least possible that this differentiation was 
already established during or before the torsion. It is not 
sufficient to argue that this is unlikely because the two 


92 H. J; FUEEBURE: 


ctenidia of Cephalopods are equivalent, for the Cephalopods 
have increased the efficiency of this branchial cavity by a 
device of their own, which permits the retention of the 
kidney and anal openings in their ancient position. The 
early Gastropod had to adapt itself to a shore life, where the 
branchial cavity was not as easily rinsed as in the more 
pelagic ancestor, and where, therefore, the excretory pro- 
ducts tended to spread over and interfere with the efficiency 
of the ctenidia, a tendency very imperfectly counteracted by 
the appearance of the slit in the shell. It was therefore 
desirable that any possible separation of incoming and 
outgoing currents should be encouraged. The reduction of 
its excretory function by the post-torsional left (pre-torsional 
right) kidney promoted this kind of separation between the 
incoming current of that side and the median outgoing 
one, and thus made the ctenidium of this side the more 
efficient. 

Probably long before this the pre-torsional left side had be- 
come more especially connected with the genital function, for 
we find such a condition in practically all Cephalopods and 
Gastropods. As the genital products in the ancestral Gastro- 
pods were expelled through the kidney, the renal opening of 
this side was very important. This explaims why it is that 
when one kidney diminishes it is that of the pre-torsional 
right (post-torsional left) side. 

In such a form slight clockwise variations of the position 
of the branchial cavity would— 

(a) Place the more efficient ctenidium in a less suitable 
position nearer the median line where the blockage of the 
foot would be most felt. 

(b) Place this ctenidium at a lower level than the anus 
and excretory openings, and thus make it likely to get soiled 
and hampered by fecal and excretory matter falling on it, 
there being no such powerful outgoing clearing currents as 
in Cephalopods. 

Counter-clockwise variations would, on the other hand— 

(a) Place the more efficient ctenidium always in a better 


RELATIONS OF KIDNEYS IN HALIOTIS TUBERCULATA. 93 


position in the sense of being further out of the chink 
between foot and shell. 

(Lb) Place this ctenidium ata higher level than the anus 
and excretory openings, and thus assure it against damage 
from the outgoing current. 

It may be urged against the suggestion— 

1. That, as Boutan says, we must not consider the excre- 
tory, etc., organs in discussing the torsion, because in 
ontogeny (in Acmiea) the torsion is completed before the 
definite appearance of kidney rudiments. The torsion, how- 
ever, entails such a serious disturbance of organs that its 
appearance in ontogeny is peculiarly liable to be hastened, 
for the earlier it appears the smaller is the derangement. 

2. That it proves too much ; in other words, that accord- 
ing to it the pre-torsional left ctenidium should have dis- 
appeared before the torsion was complete. ‘This ctenidium 
has disappeared in most Gastropods, but its occasional 
persistence is not a serious difficulty. In the first place its 
diminution would be retarded by the fact that its possessor 
was adapting itself to a life on the shore, where the time 
for breathing dissolved oxygen would be limited, thus making 
even a less efficient breathing organ temporarily valuable. 
In the second place this ctenidium persists mainly in forms 
which have evolved on special lines : 

(a) Among the Fissurellide, where the deepening of the sht 
and further changes have shortened the path of the outgoing 
current, thus reducing the possibility of its interference with 
the incoming one. 

(b) In the Haliotide, which have certainly come off from a 
very primitive prosobranch stock. Here, too, a secondary 
downward tilting of the other side of the branchial cavity 
has given this ctenidium an improved position. 

(c) In modern Pleurotomariz. The primitiveness of these 
forms is well known, and their very deep-water habitat seems 
likely to make respiration more difficult and so encourage 
the retention of all available respiratory tissue. 

3. The other objection is that it accounts for little more 


94, Here ene 


than the first 90° of the torsion. This objection is to some 
extent valid, but I think the process of completion of the 
torsion is correlated with the evolution of the shell-muscle, 
which I am at present endeavouring to investigate. 

If the supposition that the pre-torsional left side performed 
most of the excretory function thus enables us to solve 
satisfactorily the mystery of the counter-clockwise torsion, 
this is surely a strong argument in its favour. I therefore 
venture to hold, even more strongly than before, the view 
supported in the earlier part of this paper, though it is not 
in harmony with the opinion of a zoologist of M. Pelseneer’s 
eminence and insight. 


REFERENCES. 


1. H. von Jneninc.—“ Zur Morphologie der Niere der sog. ‘ Mollusken,’” 
‘ Zeit. fiir wiss. Zool.,’ t. xxiv, 1877. 
2. B. R. Lanxesrer.—Article “ Mollusea,” ‘ Encye. Brit.,’ 9th edition. 
3. J. T. Cunnincuam.—‘ The Renal Organs of Patella,’ ‘Quart. Journ. 
Mier. Sci.,’ 1883. 
4. H. Weemann.—“ Contributions a |’Histoire naturelle des Haliolides,” 
‘Arch. de Zool. exp.,’ 2me série, t. il, 1884. 
5. B. L. Bouvrer.—‘ Etude sur l’Organisation des Ampullaires,’ Mém. 
publ. par la Société Philomathique, 1888. 
6. Remy Perrirr.— Le Rein des Gastropodes Prosobranches,’ ‘Thése 
présentée a la Faculté des Sciences de Paris, 1889. 
7. RK. von Ertancer.—‘ Zur Entwickelung von Paludina vivipara,” 
‘Morph. Jahrb.,’ Bd. xvii, 1891. 
8. R. von Ertancer.—‘‘ On the Paired Nephridia of Prosobranclis,” 
‘Quart. Journ. Mier. Sci.,’ June, 1892. 
9. Bera Hatter.—‘ Studien tiber Docoglosse und Rhipidoglosse Proso- 
branchier,’ Leipzig, 1894. 
10. P. Petsenrer.— L’Hermaphroditisme chez les Mollusques,” ‘Arch. de 
Biol.,’ t. xiv, 1895. 
11. E. 8S. Goopricu.— On the Reno-pericardial Canals in Patella,” ‘ Quart, 
Journ. Mier. Sci.,’ vol. xli, 1898. 
12. J. BK. S. Moorr.—‘‘ The Molluses of the Great African Lakes,” 
‘Quart. Journ. Mier. Sci.,’ vols. xli and xlu, 1898 and 1899. 


RELATIONS OF KIDNEYS IN HALIOTIS TUBERCULATA. 95 


13. Arnotp Lane.—‘ Lelirbuch, etc.,’ ‘ Mollusea,’ Jena, 1900. 


14. Martin F. Woopwaryv.—* The Anatomy of Pleurotomaria Bey- 
richii,”’ ‘Quart. Journ. Mier. Sci.,’ vol. xtiv, 1901. 


EXPLANATION OF PLATE 6, 


Illustrating H. J. Fleure’s paper ‘ Notes on the Relations of 
the Kidneys in Haliotis tuberculata, ete.” 


EXPLANATION OF REFERENCE LETTERS, ETC., IN DIAGRAMS. 


1. External aperture of left kidney of Diotocards and of kidney of Monoto- 
cards. 2. Hxternal aperture of right kidney of Diotoeards and of genital 
system of Monotocards. 3. Anus. 4. Left reno-pericardial opening. 5. Right 
reno-pericardial opening. 61. Left gonad. 6 R. Right gonad. 7 v. Left 
kidney. 7 R. Right kidney. 7. Monotocard kidney. 7 N. Renal gland. 
7 7. Tubules of renal gland. 8. Pericardium. a. 1. Anterior lobe of right 
kidney. B.P. Brood pouch of Tanganyikia rufofilosa. vp. Penis. 
Gr. Groove connecting genital aperture with brood pouch (sometimes vestige 
only). Ss. D. Spermatic duct on floor of mantle cavity, probably formed by 
covering in of groove. vu. Ureter of Paludina. s.r. L. Subrectal lobe of 
right kidney of Patella. mr. c. Renal cavity (Fig. 9, 4). 6. Rectum. 
ty. Tissue of pericardial gland, in which renal gland is embedded. ct. Cteni- 
dium. ct!. Ctenidium of right side in Haliotis (Fig. 11). 


DiorocarDs. 
Fic. 1.—lxcretory and genital organs of Cemoria noachina (after 
Haller). Contradicted and disproved by Pelseneer. 


Fic. 2.—Exeretory and genital organs of Pleurotomaria Beyrichii 
(after Woodward). 


Fie, 38.—The right kidney of Patella vulgata (after Lankester). 
Fie. 4, a—Kidneys, gonad, ete., of Haliotis tuberculata. 


Fire. 4, 6.—Ova of Haliotis, much magnified. 


Mownorocarps. 


Vig. 5, a.—Exeretory and genital organs of the female of Tanganyikia 
rufofilosa (after Moore). 


96 H. J. FLEURE. 


Fic. 5, 4.—Exeretory and genital organs of the male of Tanganyikia 
rufofilosa (after Moore). 


Fic. 6.—The same organs in Nassopsis nassa (after Moore). 

Fic. 7.—Male genital orgaus of Typhobia horei (after Moore). 

Fic. 8, a.—Excretory and genital duct in Paludina vivipara—male. 

Fic. 8, .—The same—female. 

Fic. 9, a.—The samc—female of Littorina litorea. 

Fic. 9, .—Section of the hematic gland (pericardial gland and renal gland) 
of Littorina (after Perrier). 


Fic. 10.—Excretory and genital ducts, ete., in Buccinum undatum. 


ScuEmatic Cross-Sections OF MANTLE CAVITIES. 
Fie. 1].—Haliotis. 
Fie. 12.—Acmeea. 
Fic. 13.—Trochus. 
Fic. 14.—Littorina female. 
Fie. 15.—Paludina female. 


In Figs. 5, 8, 9, a, and 10 it is supposed that the roof of the pallial chamber 
has been cut along near the middle line and reflexed. 


In Figs. 11—15 it is not suggested that the various apertures, ete., occur 
in one and the same transverse section. 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 97 


Notes on the Development of Paludina vivi- 
para, with special reference to the Urino- 
genital Organs and Theories of Gasteropod 
Torsion. 


By 


Isabelia M. Drummond. 
With Plates 7—9. 


THESE researches have been conducted in the Laboratory 
of Comparative Anatomy at Oxford, under the super- 
intendence of Professor Weldon. They were originally 
undertaken in order to confirm or correct the account of the 
ccelom and its derivatives given by von Erlanger, and the 
results obtained are set forth in Part I of this paper. These 
then led to a renewed study of the whole course of develop- 
ment with a view to obtaining evidence upon the theories of 
torsion recently put forward by Pelseneer, Amaudrut, and 
Boutan. Von Erlanger’s account deals only with the organo- 
geny, and leaves the question of torsion on one side; it is 
perhaps for this reason that the development of Paludina 
has lately been sometimes regarded as too much modified 
and abbreviated to give any clear evidence on this point. 
Doubtless great modification has taken place owing to the 
loss of the free larval life; nevertheless, bearing this in mind, 
it is possible to a large extent so to disentangle the different 
processes of development which, owing to abbreviation, here 
go on side by side, as to be able to compare the results 

VoL. 46, rART 1.—NEW SERIES, G 


98 ISABELLA M. DRUMMOND. 


obtained with those of the authors above mentioned. In 
Part II, therefore, I give a brief account of the development, 
aiming not so much at a description of the organogeny, 
except in one or two cases where my results differ from those 
of von Hrlanger, as at making clear the changes in position 
and relative proportions of the organs in successive stages. 

Before concluding this introductory note it is conve- 
nient to say a few words with regard to the plates. The 
outlines of the figures of whole embryos have all been drawn 
with a camera lucida from preparations of the whole animal, 
which is represented as transparent, the organs being shown 
by a dotted line. These have, in fact, been also traced, 
where possible, from whole preparations, but the tracings 
thus obtained have been added to after a careful study of 
sections. Figs. 11 to 17 explain themselves; they are for 
the most part transverse sections through the visceral hump, 
taken as far as possible through corresponding regions in 
successive stages, and all orientated the same way on the 
page, that is, as if the creeping sole of the foot were parallel 
with the bottom edge, in order to facilitate comparisons. 
Allare taken looking from behind forwards,—that is, the left 
side of the figure is also the animal’s left side. 


Part I.—The Urinogenital Organs. 


Von Erlanger (5 and 6), in his account of the developing 
coelom and its derivatives in Paludina, made known for the 
first time the existence of the rudimentary original left 
kidney, and showed conclusively that the existing kidney of 
the Prosobranchs corresponds to the definitive left kidney of 
other Gasteropods. Moreover, he brought the Prosobranchs 
much more closely into line with other Molluscs than had 
hitherto been the case, by describing the gonad as a deriva- 
tive of the pericardium, and as discharging its products 
through a duct which was probably the duct belonging to 
the rudimentary kidney. While entirely agreeing to this 
extent with his results, | have arrived at conclusions with 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 99 


regard to the manner of development which differ from those 
of von Erlanger in certain important points, and which bring 
Paludina even more closely into line with other Molluses in 
respect of their ccelom. 

According to von Erlanger, the pericardium, while still 
showing its two distinct chambers, forms two little evagina- 
tions, one on each side, which are the rudiments of a pair of 
kidneys. These, fromm the time of their first formation, lie 
against the ectoderm, which very soon forms the inward duct- 
like prolongations of the mantle cavity. Of these the right 
one coalesces with the original right kidney and forms its 
duct, while the left is arrested in its growth and the left 
kidney disappears. At a later stage a new outgrowth of 
the pericardium takes place in the same position as that 
which formed the original left kidney, becomes nipped off 
from the pericardial epithelium, and forms a little vesicle, 
which is the rudiment of the gonad. At the same time 
there is an ingrowth of the mantle cavity, which is pre- 
sumably the arrested duct of the kidney that has disappeared. 
This grows towards the gonad, and finally fuses with it to 
form its duct. This account of the origin of the urino- 
genital organs has since been confirmed by the more recent 
researches of Ténniges (17).! 

I have nothing to add to von Erlanger’s description of the 
early stages of development of the pericardium and kidneys, 
and of their relation to the mantle cavity. The series of 
somewhat oblique transverse sections of which one is repre- 
sented in fig. 1 shows just such a condition as von Erlanger 
describes. In this figure the pericardium is shown with its 
two chambers still separated, the right beimg very much the 
larger of the two; the first rudiment of the heart, as appear- 
ing at h.; and just to the left of this is seen the original right 
kidney (r. k.) with its lumen, hardly showing in this section, 
communicating with the cavity of the pericardium. In a 

' | have, unfortunately, not been able to obtain access to the original paper 


by Touniges in the ‘S.B. Ges. Bef. d. ges. Naturw., Marburg,’ for 1899, and 
have had to rely upon the abstract in the ‘ Zool, Centralb.’ 


100 ISABELLA M. DRUMMOND. 


corresponding position on the wall of the narrow left-hand 
chamber of the pericardium is seen the little left kidney (J. k.), 
much less developed than its fellow on the right. In the 
next section of the series the solid ends of the two horns of 
the mantle cavity are found abutting each against the 
kidney of its respective side. It is probably at about this 
stage of development that von Erlanger describes the first 
appearance of the retrogressive development of the primitive 
left kidney. It is, indeed, extremely rudimentary at this 
time, and might easily be overlooked, but I cannot find that 
it ever wholly disappears; rather it might be said that its 
growth is arrested for a time, but at a slightly later stage it 
again resumes its development. Still less can I find traces 
of real retrogression in the primitive left horn of the mantle 
cavity. It ceases to grow, or at least does not grow nearly 
as rapidly as the primitive right horn which is to form the 
kidney duct, but it always retains its original relation to the 
pericardium, with its solid end abutting against the primitive 
left ventral corner, and is never, as von Erlanger both 
describes and figures, separated from it by a considerable 
space (see his pl. xxi, figs. 12 and 13). 

As far as can be judged from the relative positions of the 
organs, the embryo from which his fig. 12, pl. xxi, is taken 
corresponds almost exactly to my fig. F in the drawings of 
the whole animal. A section across the visceral hump of an 
embryo of this stage is depicted in fig. 15, and shows how 
the little left kidney, far from having disappeared, as 
von Erlanger describes, is now larger than at the stage when 
he believed it to be most fully developed. This figure is 
wholly comparable with fig. 1, except for the change in the 
position of the organs due to torsion. The heart, now fully 
differentiated, is seen at h. in the same position relatively to 
the other organs as the similarly marked mass of cells in 
fig. 1, and the original right kidney, the definitive kidney of 
the adult, is cut across at k., with its duct adjacent to it at 
k.d. The original left portion of the pericardium is even 
more narrow relatively to the right than is the case in fig. 1, 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 101 


and in the extreme original left and ventral (now right and 
dorsal) corner of it is the left kidney (/.k.), contrasted with 
the right in being much less developed, but in every other 
respect perfectly comparable to it, and showing exactly the 
same relations to both the pericardium and the original left 
horn of the mantle cavity as it did before. Dorsally the 
pericardium is narrowed to a point beside the liver, and here 
a proliferation of cells is just beginning to take place, which 
is the rudiment of the gonad (g.). 

The same structures are seen further advanced and more 
highly magnified in figs. 2, 3, and 4, which are three nearly 
adjacent transverse sections through a later stage. The 
position of the organs in the body is quite similar to that 
already seen, but here only the extreme (original) left-hand 
corner of the pericardium (pce.) is cut through. The rudi- 
mentary kidney is seen at /.k., now showing a wide lumen, 
but having only a narrow communication with the pericardium, 
and the solid end of the duct is seen as before at l.m.c. 
Fig. 4 shows the rudiment of the gonad (g.), now a well- 
developed cord of cells, distinctly connected with the peri- 
cardial epithelium; while fig. 3, a section intermediate 
between figs. 2 and 4, just cuts through the edge of both 
kidney (I. k.) and gonad (g.), and shows their close proximity. 
This section is, however, chiefly interesting as showing the 
thickening of the coelomic epithelium which connects these 
two organs, and very soon becomes indistinguishable from 
the gonad. A clear understanding of the position of these 
rudiments is important, and will readily be obtained by a 
comparison of the above-mentioned figures, especially of 
figs. land 15. From these it will be seen that whereas the 
kidney is from the first on the original ventral side of the 
pericardium, the gonad is a dorsal proliferation, which from 
the time of its first formation les close against the liver, the 
proximity of gonad and kidney being merely due to the 
extreme narrowness of the pericardium in this region. A 
comparison of these figures with von Erlanger’s (fig. 5, 
pl. xxiii) seems to me to point to the conclusion that his 


102 ISABELLA M. DRUMMOND. 


gonad (g.) is in reality the rudimentary left kidney, and that 
he has missed the true origin of the gonad altogether. I 
confess that I cannot fully understand this figure, but as far 
as I can make out, the gonad should he in a direction at 
right angles to that in which it is shown, if it is to maintain 
the relation to the liver shown in fig. 6, and again in fig. 17, 
and this would bring it into about the right relation with the 
evagination of the pericardium marked g., if we regard this 
latter as the rudimentary kidney. This view, moreover, 
would account for the discrepancy which exists between 
von Erlanger’s account of the origin of the gonad as an 
evagination, and my own. ‘The divergence in our descriptions 
of the duct is not so easy to explain, but I feel sure that 
von Erlanger is not correct when he speaks of the gonad 
(the left kidney, according to the present view) growing 
towards the ingrowth of the mantle cavity, for these are, 
and have been from the first, in the closest connection with 
each other (see figs. 2 and 15, l.k. and 1. m.c.). 

Fig. 5 shows a further development. The pericardium is 
cut across at the extreme right of the figure, from this the 
left kidney passes downwards at J. k., and the communication 
with the pericardium is still shown at r.p.c.ap. On the left 
of the figure is seen the gonad (g.), still solid, but now fused 
with the wall of the kidney, so that the little connecting 
portion of thickened pericardial epithelium is no longer 
distinguishable. The duct is cut at J.m.c., and does not yet 
open into the kidney. The exact position of these organs in 
the body may be seen in fig. 17, which is a drawing of a 
closely adjacent section of the same series. ‘The lettering is 
identical, and the gonad, duct, and pericardium are all shown. 
It will be seen that essentially the same relations obtain as in 
earlier stages, the gonad following the liver, and keeping 
always on the inside of the coil. 

From this stage onwards very rapid growth of the gonad 
takes place, so that it soon reaches the extreme tip of the 
visceral hump, and then it takes part in every coil as it is 
formed. At the same time it becomes hollowed out, from 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 103 


the apex downwards, till its lumen is finally put into commu- 
nication with that of the kidney, the opening being very 
close to the reno-pericardial aperture. A reconstruction of 
these organs from a series of transverse sections through an 
embryo, with a well-coiled visceral hump, is shown in fig. 6. 
The gonad (g.) is a hollow tube widening considerably at the 
apex, in reality following the coils of the visceral hump, but 
shown here spread out. The left kidney (l.k.) forms, as it 
were, merely the proximal extremity of the gonad (g.), from 
which it is separated at this time by no sharply marked histo- 
logical differentiation. In this particular specimen the reno- 
pericardial aperture (7, pc. ap.) is retained even at this late 
stage, and I have occasionally found it in other embryos of 
about the same age; more often it appears to be closed, but 
it is difficult to tell for certain which is the normal condition, 
as the opening is small and might become artificially closed 
during preservation. In this case, however, the close proxi- 
mity of the reno-pericardial and the reno-gonadial apertures 
is well seen. Even at this late stage there is as yet no com- 
munication between the left kidney and its duct, but the 
walls are now even more closely fused than before, and it is 
obvious where the exact point of communication will be. 
Details are shown in figs. 7, 8, and 9, which represent three 
sections through the same embryo from which the reconstruc- 
tion was made. Fig. 7, taken across the line a a in fig. 6, 
shows the left kidney (/. k.) with its opening into the pericar- 
dium (r.pce.ap.), and its blind end lying against the duct 
(l.m.c.), and nearly opening into it. Fig. 8 is the next 
section, taken across the line } b, and again shows the new 
pericardial aperture. Finally, fig. 9 is a section across the 
widened extremity of the gonad at ¢ c, showing the position 
in the narrow space between the liver and the outer 
epithelium of the body. 

All the essential relations between the different parts of 
the genital apparatus are now established as in the adult, 
and I have not followed their development in later stages. 
It is interesting, however, in confirmation of the correctness 


104, ISABELUA M. DRUMMOND. 


of this account, to notice von Erlanger’s description of these 
organs when they have more nearly attained their adult 
condition, and are beginning to show the development of 
the actual genital cells. Such stages, he says, show “ dass 
bei beiden Geschlechtern ein wenn auch kurzes Stiick der 
Leitungswege der Geschlechtsprodukte aus der Keimdriisen- 
anlage selbst hervorgeht;” that is, this small region, ap- 
parently belonging to and originating from the genital 
organ itself, never, in either sex, gives rise to genital cells. 
It is, of course, situated just at the junction of the gonad 
and the duct, which, as he himself points out, “‘ findet in der 
Gegend statt, wo der Verbindungskanal zwischen Herz- 
beutel und Nieren sich findet.” Surely this must be the 
original left kidney, still distinguishable in the adult. 

To sum up, then, the original left kidney and its duct do 
not, as von Hrlanger believed, disappear. Their develop- 
ment is arrested for a time, but they are both clearly present 
at the time when the gonad is formed as a proliferation 
from the original left dorsal extremity of the pericardium, 
and from this time increase in importance. ‘The gonad is 
for a long time solid, and is connected with the kidney by a 
thickening of the pericardial wall on the left side. Ata 
later stage the gonad becomes hollowed out, and its lumen 
communicates with that of the original left kidney, pre- 
sumably by means of the pericardial thickening, which must 
also have become hollowed out. The genital products there- 
fore pass through the original left kidney, and are ejected 
by its duct. 

The theoretical bearing of these conclusions is obvious, in 
that they show how, even in the adult of one of the most 
highly organised of the Rhipidoglossa, an unexpectedly 
primitive condition of the ccelom and its derivatives still 
obtains. Zoologists have long been agreed that the ancestors 
of the Mollusca must have had paired gonads, which shed 
their products into the coelom, to be carried thence by the 
kidneys; that the ccelom is now represented by the peri- 
cardium, and that, though great modification has taken 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 105 


place, a remnant of the primitive condition is found in the 
frequent connection between gonad and kidney in existing 
forms. 

While there is almost perfect agreement with regard to 
the general features of the anatomy of the primitive form, 
however, there is considerable divergence of opinion as to 
the course which evolution has followed, and consequently 
various interpretations are put upon the structures of 
existing forms, while hitherto embryology has been almost 
silent, and evidence has had to be almost entirely drawn 
from the field of comparative anatomy. Pelseneer (12 and 
13) and Haller (10 and 11) are among the chief writers who 
deal with the ccelom and its derivatives among the Proso- 
branch Gasteropoda, and still uphold quite different views 
upon many points, though they seem agreed in maintaining 
that a gradual loosening of the connection between gonad 
and kidney has taken place throughout the group. While 
Pelseneer, however, only maintains that the point of com- 
munication tends to shift away from its primitive position 
by the reno-pericardial aperture nearer to the external 
opening of the ureter, as is the case in the Lamellibranchs, 
Haller regards the connection between gonad and kidney as 
altogether severed among the higher Rhipidoglossa, a 
portion of the coelom becoming specialised as the gonaduct. 
Connected with this is the different view which these two 
authors take of the homology of the existing kidney. Thus 
Haller regards the functional kidney as in all cases the 
right one (after torsion). Among primitive forms (e.g. 
Fissurella) this keeps its connection with the gonad, while 
that on the left side of the body loses its connection with the 
gonad and kidney, and is fast disappearing. Further stages 
of evolution are shown by Haliotis, Trochus, and Paludina, 
in all of which the left kidney has entirely disappeared, and 
the connection between the gonad and right kidney is lost, 
while the latter has more and more passed over to the left 
side of the body. Pelseneer, on the other hand, has demon- 
strated, though this is still denied by Haller, that in both 


106 ISABELLA M. DRUMMOND. 


the Haliotidee and Trochide a very small definitive left 
kidney is present, and that the large kidney is the definitive 
right, and still maintains its connection with the genital 
organ. 

It might seem that von Hrlanger (5 and 6, see also 7) had 
already sufficiently demonstrated from embryology that the 
homologies which Pelseneer believes to hold for Haliotis 
and 'Trochus are equally true for Paludina. To this, however, 
Haller (11) objects that in a highly organised form, such as 
Paludina, torsion is very likely abbreviated, and the organs may 
be formed in their definitive position. This view is, it seems 
to me, quite untenable from von Erlanger’s description, while 
a further study of the development of this form shows even 
more clearly that a complete rotation of the organs through 
180° actually takes place in the course of development,! and 
that the adult kidney arises on the right, and ends on the 
definitive left side of the body. I have, fortunately, been 
able to add further to this evidence by showing how the 
gonad still stands in close relation with the definitive right 
kidney, though this has altogether lost its excretory character, 
and that no such separate duct as Haller describes is ever 
formed. It seems, then, that there is every reason for 
believing that the definitive right kidney has persisted 
throughout the Prosobranchia as the genital duct, in some 
cases, as in Haliotis, performing also its renal functions, 
while in Paludina these latter are carried on altogether by 
the left kidney, the right functioning only as a gonaduct. 

With regard to the manner of communication between the 
two organs, Pelseneer and Haller are also in disagreement. 
In the Docoglossa, at least, Haller describes a ventral 
coelomic chamber through which the genital products must 
pass in order to reach the kidney ; while Pelseneer regards 
this so-called ccelom as merely a portion of the kidney itself, 
the gonad being in direct communication with this latter, and 
altogether separated from the ccelom, which is only repre- 
sented by the pericardium. Whether Haller believes in a 


For evidence upon this point, see Part IT of this paper. 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 107 


coelomic connection between gonad and kidney in the primi- 
tive Rhipidoglossa similar to that which he describes for the 
Docoglossa is not very clear. Most writers, however, have 
described the gonad as having become separated from the 
coelom altogether, and having acquired a new opening into 
the kidney. To this, and also to Pelseneer’s view that this 
opening occurs nearer to the external aperture in the higher 
forms than in the lower, von Erlanger’s description of 
the course of development in Paludina lent strong support. 
This, however, has completely failed, for the communication 
between gonad and kidney has been shown to be close to the 
reno-pericardial aperture in Paludina, as Pelseneer has de- 
scribed in Fissurella and other primitive forms, while traces 
of an original coelomic connection between the two are found 
in the thickened ridge of pericardial epithelium described 
above, which can hardly be otherwise interpreted than as 
representing a groove in the ccelomic floor along which, in 
more primitive forms, the genital products passed to the 
reno-pericardial aperture. That the latter still remains open 
even after the communication between gonad and kidney is 
established is no real hindrance to such an interpretation, for 
the solid nature of the rudiments of both the gonad and the 
coelomic connection shows that the ontogeny is abbreviated, 
and gives no exact picture of the phylogenetic events. he 
opening of this pericardial groove into the kidney must, it is 
true, represent at least a portion of the reno-pericardial aper- 
ture. Phylogenetically, we may believe, the edges of the 
groove drew together and a tube was formed, opening at one 
end into the gonad and at the other into pericardium and 
kidney at once through the reno-pericardial aperture. 
When, by abbreviation, this tube came to be formed in the 
course of development as a solid rudiment, it is easy to 
understand how the hollowing out and subsequent communi- 
cation with the kidney might lead to the appearance of a 
rupture of the kidney wall. If this interpretation be correct, 
we have in Paludina, which has always been regarded as one 
of the most specialised of the Rhipidoglossa, a condition of 


108 ISABELLA M. DRUMMOND. 


the urinogenital organs in every way comparable to that which 
obtains among the Amphinoma. 

Finally, with regard to the origin of the single asymmetrical 
gonad in the Gasteropoda, both Pelseneer and Haller seem 
agreed that this has been formed by the fusion of the 
originally separate gonads of both sides. Phylogenetically, 
of course, this may have been the case, but ontogenetically 
there is no trace of it, the existing gonad being formed 
exclusively from the extreme left-hand corner of the original 
left division of the pericardium. 

Summary.—The conclusions at which we have arrived 
are as follows: 

(a) The embryology of Paludina demonstrates that the 
functional kidney of the adult belongs morphologically to 
the definitive left side of the body, as von Erlanger has 
already pointed out. 

(b) The definitive right kidney is not lost, as von Erlanger 
describes, but persists as the genital duct. 

(c) An indication of the original ccelomic connection 
between gonad and kidney is found in the course of develop- 
ment of Paludina as a thickened ridge of pericardial epi- 
thelium, which finally becomes indistinguishable from the 
gonad, and, after it has acquired a Jumen, communicates with 
the definitive left kidney close to the reno-pericardial 
aperture. 

(d) The gonad arises as a solid proliferation of the 
morphologically dorsal wall of the pericardium. It arises 
from the original left side only, and shows no sign of a 
pared origin. 


Part II.—The Development of Paludina viewed in Connection 
with Theories of Torsion. 


(A) Description of Development. 


Stage A (fig. 10).—The youngest stage which I have 
examined is a bilaterally symmetrical, oval embryo, with 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 109 


well-developed velum, and already a slight swelling ventrally, 
which is the rudiment of the foot. The chief points in the 
anatomy are shown in fig. 10, which is a sagittal section 
through an embryo of this stage. The gut (st.) 1s a simple 
sac opening posteriorly by the anus (a.), and ending blindly 
anteriorly where it abuts against an insinking of the 
ectoderm (m.), the rudiment of the stomodeum. No very 
clear differentiation of parts is yet visible in the gut, but the 
ventral wall begins to show the vacuolated structure charac- 
teristic of the liver at a later stage. Dorsally to the stomo- 
deeal invagination the velum is seen cut twice (v.), and more 
posteriorly is seen the shell gland (s.g.), a deep sac, widely 
open to the exterior in other sections of the series. The 
mesoderm at this stage is represented simply by scattered 
cells. 

Stage B (fig. B), shows considerable advance upon the 
last. The foot (f.) has grown out to form a prominent pro- 
jection on the ventral surface. The shell gland is partially 
evaginated, and begins to form the visceral hump (v. h.), 
which, however, is still partly surrounded by a groove, 
deepest behind, and gradually disappearing anteriorly. The 
velar area has increased in size, and the tentacles (¢.) are dis- 
appearing. The stomodeum has now broken through into 
the archenteron, and considerable differentiation has taken 
place in the latter. The middle portion has swollen and 
forms the stomach (st.), which lies at the apex of the visceral 
hump (v. A.), and from which the rectum runs downwards and 
backwards to open in the middle line behind (a.) The liver 
(/.) 1s an oval structure, sloping downwards and forwards 
from the apex of the visceral hump, where it communicates 
mostly with the stomach. The opening into the stomach is 
still so wide, and the demarcation between the two organs so 
vague in this region, that it is difficult to determine their 
exact relations, but the liver appears to lie to the left, and 
ventrally behind, while in ventral views of the whole embryo 
it can clearly be seen to pass below the cesophagus and to the 
right side anteriorly. It seems then to be an outgrowth of 


110 ISABELLA M. DRUMMOND. 


the left ventral wall of the stomach. Below the rectum, and 
lying between it and the liver, is a httle dense mass of 
mesoderm cells, which is just beginning to be hollowed out 
on either side to form the rudiment of the pericardium. Its 
position is shown in the figure at p.c. The otocysts have 
appeared on either side at o.¢., and are still widely open to 
the surface of the body. 

Stage C (figs. C and C,, and fig. 11).—Considerable 
growth in length has taken place, and the different regions 
of the body are clearly marked out. The foot is now a 
prominent organ in the anterior ventral region, and a slight 
constriction of the body separates the foot and head from the 
now well-developed visceral hump. This latter is surrounded 
posteriorly by the mantle folds (m.f.), which form a 
prominent ridge dorsally to the anus. At this stage the 
first rudiments of the mantle cavity appear as two little 
depressions lying one on either side of the anus. These are 
best seen in a ventral view of the whole animal, or in section. 
Fig. C, is a ventral view of a slightly older embryo than 
fig. C, but the essential relations of the organs are precisely 
similar. Here the two depressions are seen at c.m.c. and 
r.m.c., the right one being considerably in advance of the 
left. The same depressions are seen in transverse section in 
fig. 11. Von Erlanger describes the first appearance of the 
mantle cavity as “eine kleine Grube”’ ventrally and just in 
front of the anus, but it is quite clear that at this stage there 
is a distinct rudiment on either side, the rectum passing 
down a ridge between them (« in fig. 11), to open directly on 
to the surface of the body. It is only at a later stage that 
the portion of the body immediately in front of the anus 
sinks in and unites the two original depressions, thereby in- 
cluding the anus within the mantle cavity. I have never 
been able to find a stage in which these two original depres- 
sions are symmetrical. If this stage closely corresponds, as 
I believe it does, with von Erlanger’s fig. 1, plate xxi, he 
has overlooked an important point in the external anatomy of 
the embryo. It is not, as he says, perfectly symmetrical 


THE DEVELOPMENT OF PALUDINA VIVIPARA. tit 


externally at this stage, for not only is the symmetry dis- 
turbed by the inequality of the rudiments of the mantle 
cavity just noticed, but the whole visceral hump appears as if 
slightly tilted. The apex lies somewhat to the left of the 
vertical plane, which would divide the head and foot 
symmetrically, while the mantle fold on the left of the body 
is at a lower level from that on the right. This tilting is 
difficult to represent in surface views, though by rolling the 
whole embryo about it is perfectly easy to see. It is, how- 
ever, sufficiently obvious in the transverse section through the 
hump (fig. 11), which is orientated on the page as it would be 
on the body, the line aa representing the vertical plane 
through head and foot. 

In the internal organs there is little to add to von Erlanger’s 
(5) account. ‘The stomach and liver together form, as before, 
the apex of the visceral hump; they are now well-defined, 
though still retaining their wide communication with each 
other. Posteriorly the liver lies distinctly to the left, while 
further forward it gradually becomes almost ventral, passing 
over to the right, as before, in front of its opening into the 
stomach (fig. C,). From the posterior end of the stomach 
the rectum runs almost vertically downwards to open in the 
position already noticed between the two rudiments of the 
mantle cavity. Just dorsal to these, and anterior to the 
rectum, are the two rudiments of the pericardium, still 
separated from each other, that on the right being a good 
deal the larger of the two, as we saw to be the case also 
with the mantle cavity (fig. C,, fig. 11). At the left, in 
fig. 11, a slight thickening of the pericardial epithelium is 
seen, which must be the rudiment of the left kidney ; the 
right kiduey is not yet formed. In the head region the first 
appearance of the radular sac is noticeable, and also the 
appearance of the ganglia of the central nervous system as 
thickenings of the ectoderm, as they have been already 
described by von Erlanger. 

Stage D (fig. D, fig. 12).—The foot now begins to show 
for the first time a tendency towards the formation of the 


112 ISABELLA M. DRUMMOND. 


characteristic creeping sole. The visceral hump is much 
developed and surrounded by a strongly marked mantle fold. 
Externally the most noticeable feature is a prominent bulge 
on the left side, which appears almost like the first formation 
of a coil (fig. 12). A transverse section across the visceral 
hump, shows, however, that it is simply a bulging out of the 
side of the body where the liver and stomach are located. A 
comparison of figs. 11 and 12, in fact, makes clear that this 
prominence on the left side corresponds to the original apex 
of the visceral hump, which has become still further tilted in 
the same manner as before, thus leading to a rearrangement 
of the organs when orientated with reference to the head and 
foot, while their mutual relations are retained. The liver 
now lies ventrally, as well as slightly to the left of the 
stomach. 

Both liver and stomach are entirely on the animal’s left, 
and the mantle cavity lies entirely on the right, while the : 
anus, also, is displaced from the middle line and has travelled 
towards the right. Though readily seen by a comparison of 
transverse sections, this tilting is not so obvious when the 
embryo is examined entire, as it was at an earlier stage. 
This seems in part due to the great growth of the whole 
posterior region of the visceral hump, which has caused the 
rectum to bend forward to the anus, and the mantle cavity 
to take up sucha position that the two original depressions lie 
almost vertically one above the other instead of almost horizon- 
tally, as in fig. 11, but chiefly is it the result of the rapid growth 
of the mantle downwards on the right-hand side of the animal’s 
body, so that the right and left edges are now on about the 
same horizontal level, and a kind of false external symmetry 
is established in this respect. Rapid growth of the mantle 
leads, of course, to rapid extension of the mantle cavity, and 
consequently we find this far in advance of the preceding 
stage. The two original depressions are now united below 
the rectum, and form two horns which abut each against 
the kidney of its respective side, and then join and widen 
out to form the mantle cavity; a slight extension of the 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 13S 


mantle cavity to the right of the right horn is first noticeable 
in this stage, but will be more fully noticed in the next. 
The kidneys have been formed as outpushings of the 
(morphological) ventral wall of the pericardium on either 
side; the wall of the little left kidney and the adjoining wall 
of the left horn of the mantle cavity are cut through in the 
“ventral part of the section in fig. 12 (J. k.), while the right 
horn is seen more dorsally at r.m.c. The ventral position of 
the left horn is also seen in fig. D at l.m.c.! The right 
division of the pericardium has become much enlarged, and 
now occupies a very considerable portion of the visceral hump, 
while the left division remains small, and les in the narrow 
region between the liver and the left horn of the mantle 
cavity. The rectum bends sharply downwards from the 
stomach and then runs forward ventrally on the right 
to the anus, which lies just anteriorly to the junction of the 
two horns of the mantle cavity, and is now included in the 
latter. 

Stage E (figs. E, E,, E,; figs. 13 and 14).—The foot 
has grown back into its definitive position, and is separated 
by a marked constriction from the head. The features of 
the visceral hump noticed in the last stage are now accentu- 
ated. The bulge on the left side has become much more 
prominent, and is a very characteristic feature, giving, even 
more than before, the appearance of sinistral coiling when 
looked at from above (fig. E,). A comparison between figs. 
12 and 13 shows that the essential relations are the same as 
in the last stage. The liver, now ventral to the stomach, has 
increased much in size; the pericardium, of which only a 
portion of the right-hand division is shown in the figures, has 
swollen, and is found extending for a considerable distance be- 
side the stomach to the right of the original right kidney (see 
fig. EK, and fig. 15) ; while close to the kidney on the right is 
found the first rudiment of the heart as a little solid ingrowth 
of mesoderm cells pushing the pericardial wall before it. 

1 For further description of the kidneys and their relation to the mantle 
cavity, see Part I of this paper. 

von. 46, pArt 1.—NEW SERIES. H 


114 ISABELLA M. DRUMMOND. 


Corresponding to this extension of the pericardium is a great 
development of the mantle cavity to the morphological night 
of the original right horn or kidney duct. This is shown in 
section with the pericardium lying above it in fig. 14, but is 
best seen in a view of the whole animal from the left, fig. K, ; 
here the left horn is seen ventrally at J. m. c., the former 
roof of the mantle cavity is almost vertical and forms the 
posterior boundary, while the kidney duct (k. d.) is dorsal. 
Dorsal and anterior to this is now the chief extension of the 
mantle cavity, which is already visible in a dorsal view of the 
whole embryo (fig. E,), and reaching behind to the mid-dorsal 
line, while it narrows slightly in front. Von Hrlanger’s 
description of an embryo of this stage differs considerably 
in respect of the mantle cavity from the above, but I am 
unable to reconcile his fig. 7 on pl. xxi with my own 
observations, for he both figures and describes (p. 358) the 
kidney duct as arising from the dorsal extremity which is 
now advancing over the mid-dorsal line, while a comparison 
of the series (of which two sections are shown in figs. 13 and 
14) with the living animal (from which the outline of the 
mantle cavity in fig. EH, was drawn) seems to me to show quite 
conclusively that the kidney duct arises in the position where 
I have marked it, and that the dorsal extension of the mantle 
cavity is a new development. If this be so, very rapid growth 
must have taken place in the region in front of a vertical 
line passing through the opening of the two kidney ducts, as 
is in fact the case, and this would lead to the increased 
ventral flexure which attains its maximum about this 
time. 

Very little remains to be said of the other organs. ‘The 
cesophagus is much lengthened, and has acquired a sharp 
downward bend before entering the stomach, which slopes 
obliquely upwards and backwards. ‘The rectum bends sharply 
downwards and then runs forward to open on the right side 
in the mantle cavity. The right kidney has become further 
developed and forms a simple sae, still fairly widely open to 
the pericardium, while the left remains in much the same 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 115 


condition as in the previous stage. The pericardial septum 
has disappeared. 

Stage F (fig. F and fig. 15).—AIl the essential features 
may be seen in a view of the left side of the entire animal 
(fig. F). Comparing this with the similar view of the pre- 
ceding stage we find that there has been very rapid growth 
in all parts of the body, especially in the “neck”? region 
between the visceral hump and the head. The bulge con- 
taining the stomach and liver now lies nearly ventrally, 
the mantle cavity has extended over into the left side of 
the body, and just posterior to it is seen the pericardium, 
with the heart now well developed, and showing the auricle 
and ventricle separated from each other by a deep constric- 
tion. The rectum lies higher on the right side than in the 
last stage, and runs along the roof of the mantle cavity to 
open more anteriorly. Fig. 15 is a transverse section across 
the visceral hump of an embryo of this stage, and shows 
essentially the same relation of the organs to each other as in 
previous stages. The pericardium lies dorsally to the liver 
and stomach, and contains the heart. Dorsal to the peri- 
cardium are the two kidneys, the morphologically right, a 
well-developed but still simple sac, being seen at k., its duct 
at k.d.; the opening of the kidney into the pericardium on 
the one hand and the duct on the other are neither of them 
shown in this section. ‘To the right of the section is seen the 
rudimentary left kidney (/. k.), which seems to resume its 
development about this time. Its opening into the peri- 
cardium is well seen, and just above it is the solid end of 
the original left horn of the mantle cavity (l. m.c.). At 
the extreme right (morphologically left) of the pericardium 
the first rudiment of the gonad can just be distinguished 
at g.} 

A further important feature is the development of the 
visceral connectives, which are first visible at this stage. 
They arise anteriorly from the pleural ganglia and run back 


1 For further description of the gonad and its connection with the rudi- 
mentary original left kidney, see Part I of this paper. 


116 ISABELLA M. DRUMMOND. 


on either side of the cesophagus to about the region where 
the anus opens, when they appear to lose themselves in the 
epithelium of the floor of the mantle cavity. The morpho- 
logically left connective is by far the stronger, and arises 
behind, just below the anus, while the morphologically right 
appears at the extremity of the right extension of the 
mantle cavity. The two connectives are thus separated by a 
considerable distance posteriorly, and they are not at present 
united by a commissure. 

Stage G (fig. G and fig. 16).—The original apex of the 
visceral hump now points ventrally, though it is still more 
prominent on the left side than on the right, which gives the 
appearance, when the animal is looked at from above, of the 
visceral hump being set crookedly upon the foot. The 
cesophagus is elongated, and bends sharply downwards to 
open into the stomach. The stomach itself is much enlarged 
and lies chiefly ventrally, but ascends somewhat to open into 
the rectum, which then bends dorsalwards and runs forward 
in the roof of the mantle cavity’ to the anus. ‘The mantle 
cavity now extends far down on the left side, especially 
posteriorly, and a portion of the kidney is visible in a view 
of the left side of the animal. The rudiments of the 
ctenidium are clearly formed as projections from the roof of 
the mantle cavity on the left. The kidney duct soon after 
leaving the kidney now passes below the rectum and runs 
backward on its right side to open into a sort of little pouch 
of the mantle cavity together with the genital duct, as the 
original left kidney duct may now be called. Von HErlanger’s 
fig. 2, pl. xxii, shows very well the disposition of the kidney 
duct at this stage. The old relations of the rectum to the 
two original horns of the mantle cavity are thus disturbed, 
apparently by a drawing together of the edges of the mantle 
cavity in this region, the space between the two ducts being 
obliterated. The whole, or very nearly the whole, of the 
definitive mantle cavity seems, therefore, to be formed by 
the great extension of the original right horn, as noticed in 
Stage EH. In other respects the mutual relations of peri- 


THE DEVELOPMENT OF PALUDINA VIVIPARA. ilu by 


cardium, kidneys, and ducts remain as in the last stage, but 
the gonad is now clearly formed as a thin cord of cells lying 
beside the liver. 

The visceral connectives are now completely formed, and 
are united by a commissure, which in this stage lies asym- 
metrically, that is, wholly below the original right portion of 
the mantle cavity. This is, however, a secondary condition, 
and probably due to a tendency to place itself in relation 
with the symmetry of the external form, for a stage inter- 
mediate between this and the last shows that the commissure 
is formed in part from the floor of the mantle cavity just at 
the entrance of the genital duct. 

Fig. 16 is a transverse section across the visceral hump 
in the region of the kidneys and heart. Liver and stomach 
lie ventrally, and above them is seen the pericardium, with 
the heart at the extreme left of the section. Dorsally to the 
pericardium is seen the kidney (k.) with its openings into 
pericardium (pc.) and duct (k. d.), both cut through. The 
kidney duct is beginning to pass ventrally to the rectum (rec.) 
as described above. The genital duct is cut through at 
1. m. c., and the wall of the left kidney at J. k. 

Stage H (figs. 17, 18, 19).—A definite coil is now being 
formed on the right side, about one complete turn of the spiral 
having been made, and the old crooked setting of the hump 
on the foot is nearly lost. In this stage the organs attain very 
nearly their adult condition in all essential points. The ali- 
mentary canal has increased much in length and become more 
coiled. The cesophagus bends down and towards the right to 
open into the stomach, which now stretches as a great sac 
below the liver, opening into it dorsally, and forming at the 
right-hand extremity a blind sac, which shows a tendency to 
follow the coiling of the liver (fig. 17). The rectum now 
opens out of the stomach quite ventrally and posteriorly, 
passes towards the left side of the body, then bends sharply 
upwards behind the pericardium, and runs dorsally along the 
roof of the mantle cavity, bending suddenly to the right just 
before it reaches the anus. The liver is greatly developed 


118 ISABELLA M. DRUMMOND. 


and lobed, and forms almost the whole of the coil, being 
followed only to a very slight extent by the stomach and 
gonad. The pericardium is swollen, and the original left 
side, which has hitherto been so narrow, widens out con- 
siderably. In connection with this we may notice the 
advanced condition of the primitive left kidney and the now 
well-developed cord of cells (g.) which represents the gonad. 
The definitive kidney (k.) is seen on the left side of the 
section, and is easily recognisable by its slightly staining 
and now folded walls. The kidney duct (k. d.) is cut across 
to the right of the kidney, just where it passes below the 
rectum (rec.) as described in the last stage. Both kidney 
and genital ducts have now lost their primitive condition as 
simple specialised portions of the mantle cavity, and run for- 
ward in the roof of the latter as well-defined ducts, parallel 
to and to the right of the rectum, opening somewhat behind 
the anus. 

The mantle cavity now extends very low on the left side, 
especially posteriorly, so that it is Just cut in the section 
represented in fig. 17, the difference between anterior and 
posterior regions being much more accentuated than was 
formerly the case. The relations of the visceral connectives, 
as noticed already, though in a less degree in the last stage, 
are deeply affected by the asymmetrical growth of the mantle 
cavity, and as posteriorly the mantle cavity appears to lie 
wholly on the left side and with its floor almost vertical, so 
in the region just anterior to the commissure the two con- 
nectives lie almost in one vertical plane. A discussion of 
these relations may conveniently be left till the whole 
question of torsion is taken into consideration, but the twist 
of the connectives now remaius to be described. As is well 
known, the two connectives are bilaterally symmetrical when 
they leave the pleural ganglia, almost immediately the right 
passes below, the left above the cesophagus, each to the 
opposite side of the body, after which they again resume 
their bilaterally symmetrical disposition, only that now on 
the right side is the original left, and that now on the left is 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 119 


the original right. Following this twist in a series of trans- 
verse sections a peculiar relation between the cesophagus and 
the connectives is noticeable. In the region of the anterior 
ganglia and for some distance behind them the cesophagus is 
compressed so as to render its outline oblong in section. At 
first bilaterally symmetrical, it soon becomes completely 
asymmetrical, the long axis, as seen in section, sloping at 
first upwards and to the left, then passing through the hori- 
zontal position to slope upwards and to the right; in other 
words, the cesophagus apparently follows the connectives in 
their twist, as shown in figs. 18 and 19. At the completion 
of the twist the cesophagus becomes round in section, and 
passes back for some considerable distance lying between the 
two connectives, which are now once more in the same hori- 
zontal plane; then the connectives appear to become partly 
twisted a second time to take up the position with regard 
to the mantle cavity posteriorly, which has been already 
described. Unfortunately, in this region the cesophagus is 
still circular in section, and we have no direct evidence as to 
whether or not it follows the connectives in the same way 
that it does more anteriorly. 

The relations between the connectives and cesophagus are 
seen to a greater or less extent in stages previous to this, but 
the description of them has been deferred till now as being 
easier of comprehension when they are present in such a 
marked degree. Already in Stage C, where we first noticed 
the tipping of the visceral hump, a slight apparent twist of 
the oesophagus was visible; this was more marked in the 
following stage, and increased in each succeeding stage up 
to the present. It is as well to notice that this condition of 
the cesophagus was plainly visible before the visceral con- 
nectives appeared at all. 

It has already been said that the animal has now attained 
in all essential particulars to the anatomy of the adult. A 
few points, however, remain to be noticed, being for the 
most part only further developments of processes already 
begun. Of these the most obvious is the coiling which takes 


120 ISABELLA M. DRUMMOND. 


place rapidly from this time. The tendency already seen in 
tle stomach to grow out into a third sac which follows the 
coiling of the liver becomes considerably accentuated, while 
tle gonad grows very rapidly and soon passes right up to 
the tip of the last coil. At the same time it loses its solid 
character, and, becoming hollowed out, acquires an opening 
into the duct, as described in the special part of this paper. 
Meanwhile the mantle cavity deepens, and the rectum grows 
forward to open near the anterior edge. It is during this 
growth that it acquires the characteristic disposition of the 
adult, passing from mid-dorsally behind, obliquely down- 
wards and to the right, a disposition which is doubtless con- 
nected with the sharp bend towards the right described in 
Stage H. The pericardium alone begins at this late stage to 
show new relations, for it widens and becomes very irregular 
in shape, spreading amongst the other organs of the body so 
as to form a kind of general body-cavity. 

Monstrosities.—Whilst collecting material for the study 
of the normal course of development a few monstrosities 
were found which presented some remarkable features. 
Although it seems impossible fully to understand the 
meaning of all the abnormal conditions found in_ these 
embryos, some of them seem to me to be of sufficient interest 
to justify the insertion here of a description of the main 
features of their organisation. 

1. The simplest of these abnormalities is a small embryo 
between stages C and D in degree of development, perfectly 
normal in every respect, but wholly reversed. The liver 
and stomach form a bulge on the right side, while the mantle 
cavity and rectum are on the left, and all the other organs 
correspond in every particular. This is, so far as I know, 
the first record of a normally dextral Prosobranch so 
organised. It is unfortunately too young to show definitely 
the manner of coiling. 

2. This embryo, shown in fig. M 1, is very remarkable. 
A camera tracing was made while it was still alive, and the 
organs put in partly from life and partly after preservation. 


THE DEVELOPMENT OF PALUDINA VIVIPARA. WI 


It was then sectioned in a plane transverse to the long axis 
of the visceral hump, and portions of these sections are por- 
trayed in figs. 20, 21, and 22. The most noticeable feature is 
the greatly developed visceral hump, which was held erect 
over its head, bending in a decided manner at the apex, as 
though forming the first turn of an exogastric coil. A 
further remarkable feature is the perfect bilateral symmetry 
of the whole embryo, though it is obviously at an advanced 
stage of development, for though it is impossible to compare 
it with any given stage in the normal course, the head is 
well developed, the foot has found its normal creeping sole, 
and a small operculum is already present. The general dis- 
position of the organs can be made out from the drawings of 
the whole animal. The stomach, it will be seen, forms the 
apex of the visceral hump, while just below it an enormously 
developed pericardium fills up for some distance the space 
between the descending cesophagus and rectum. Paired 
kidneys are seen at k and k’, but are better described in 
connection with the transverse sections. The same is the 
case for the great bulge in the lower posterior region of the 
visceral hump behind the rectum, which might be taken for 
the mantle cavity. Sections, however, show it to be merely 
the continuation of a great space which surrounds all the 
organs nearer the apex of the visceral hump, as seen in 
fie. 20, which is a transverse section in the region of the 
kidneys, and just below the pericardium. The mantle cavity 
is shown in figs. 20 and 21 as a narrow and symmetrical 
organ lying anterior (morphologically dorsal) to this great 
space, and apparently compressed by it. Towards the apex 
of the hump it forms two symmetrical horns which run back 
on either side towards the kidneys, but never fuse with them. 
Here a wholly inexplicable condition obtains. ‘The peri- 
cardium, as already mentioned, is greatly developed, and it 
is not surprising, in an embryo otherwise symmetrical, to 
find out that here also two symmetrical evaginations have 
been formed. These, which must be the kidneys, are, on the 
one hand, very widely open to the pericardium, while on the 


122 ISABELLA M. DRUMMOND. 


other they come nearly into contact with the horns of the 
mantle cavity, but no communication is formed. Instead, 
each kidney communicates with a sort of little vesicle, as 
seen in fig. 20, which at the same time shows the close 
approximation of the horn of the mantle cavity to the 
kidney on either side. Below the kidneys, that is, nearer to 
the anus, these two vesicles unite, and run back for a short 
distance as a single duct, which opens into the mantle 
cavity mid-dorsally, as shown in fig. 21. Shortly afterwards 
the anus opens on the same mid-dorsal ridge. 

One further feature remains to be noticed, namely, the 
absence of liver. The stomach is well developed, but the 
only trace to be found of anything which might be interpreted 
as liver is a pair of little outgrowths of the alimentary canal 
just in the region where the stomach and cesophagus unite. 
These are shown in fig. 22. If these may be so interpreted, 
then the liver shares also in the symmetry shown by the 
mantle cavity and kidneys. The visceral connectives, as 
would be expected, run back perfectly symmetrically beside 
the cesophagus. 

3. Other monstrosities occurred, but of less interest than 
the above. They all showed traces of symmetry in a greater 
or less degree, and some of them the same tendency to 
exogastric coiling, but in most cases the organs were 
deformed, and very much less clear of interpretation. Only 
one other, therefore, is shown here (fig. M ur). This one is 
remarkable in that a greater degree of symmetry than is 
usual is combined with a slight sinistral torsion, and a fairly 
well-marked development of the mantle cavity to the left of 
the original left horn, as seen in the figure. The symmetry 
of the body is, however, confused, in that the pericardium is 
more developed on the right side than on the left. This 
embryo is further remarkable in that it is the only one of 
the monstrosities possessed of a clearly defined liver, which 
even in this case is very small, and hangs like a sac from the 
stomach ventrally, and slightly on the right side of the 


body. 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 123 


(zs) Theoretical Considerations. 


I do not mean to attempt to give here anything like a 
complete historical summary of the many views which have 
been held on the subject of Gasteropod torsion and asym- 
metry. This has already been done more or less fully many 
times (see especially Simroth [16] and Boutan [2]), and I 
shall therefore confine myself merely to a very brief conside- 
ration of those views which lend themselves to criticism from 
an embryological standpoint, and upon which a study of the 
development even of a single form may throw some light. 
Such theories, therefore, as that put forward by Lang, which 
claims only to be phylogenetic, and for which confirmation is 
not sought from the facts of ontogeny, will be passed over 
altogether ; while theories of authors who, like Biitschli, seek 
to base their conclusions to a greater or less extent upon 
embryology may be of some interest in this connection, and 
will therefore be considered. At the same time it must be 
remembered that the remarks upon these theories are pro- 
fessedly based only on embryology, and need not necessarily 
invalidate their phylogenetic value, though they may weaken 
the author’s argument. 

For the sake of convenience the theories under considera- 
tion may be placed in two classes. In the first of these are 
placed those theories which maintain that the present condi- 
tion of the Prosobranchia has been brought about by a simple 
process of unequal growth, resulting in the forward movement 
of the palleal complex in a horizontal plane ; while the second 
comprises those more recent theories of Pelseneer, Amaudrut, 
and Boutan, which regard asymmetry primarily as the 
concomitant of a twist which causes the palleal complex to 
move in a vertical plane. 

Biitschli (4) was the first to put forward in an exact and 
careful way the point of view which is now common to all 
theories of the former class. He puts aside the older view 
of Spengel, which obviously runs counter to known embryo- 


124. ISABELLA M. DRUMMOND. 


logical facts, by pointing out how the anus must have lain at 
all stages in the palleal groove, and then proceeds to build 
up his own theory of unequal growth, and the resulting 
gradual approximation of mouth and palleal complex on the 
right side. That such an approximation does take place in 
the ontogenetic history is, of course, well known, and the 
manner in which it is brought about seemed to Biitschli 
equally obvious. According to him, at a time when the anus 
lies in the middle line posteriorly, a narrow zone on the 
right side of the animal ceases to grow altogether, while the 
corresponding zone on the left grows with great vigour, and 
thereby the anus appears to be pushed up the right side of 
the body, while in reality the distance between it and the 
mouth remains always the same. Meanwhile the foot on the 
one hand, and the mantle on the other, continue to grow 
symmetrically. ‘This process cannot by itself, however, bring 
about the crossing of the visceral connectives. For this 
Biitschli has to invoke the aid of the mantle cavity, which, 
he says, is formed rapidly at a time when the anus hes far 
forward on the right side of the body, and, by its growth 
backwards and to the left, carries the organs of the original 
right side of the body back with it and over the mid-dorsal 
line. All this he puts forward as ontogenetic fact, and 
therefore probable phylogenetic theory. 

Biitschli’s views have been adopted with more or less slight 
modification by many authors, and have recently been brought 
forward again with some additions by Plate (15). The great 
difficulty, to which no one could find a fully satisfactory 
solution, was the absence of any known cause of asymmetrical 
growth in a perfectly symmetrical body. Plate seeks an 
explanation in the asymmetry of the liver. Starting from 
the nearly symmetrical liver of the Chitons, and comparing 
it with the asymmetrical organ in the Gasteropoda, he 
describes how, in the primitive form, a rapid growth of the 
left liver must have taken place at the expense of the right, 
which would result in the formation of a hernia posteriorly 
on the left side. Thus the first rudiment of a coil is formed, 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 125 


which, for reasons connected with the equilibrium of the 
body, lies with its apex pointing towards the right (see his 
fies. F to H, pp. 185 and 187). This it is, he believes, which 
causes the approximation of the mouth and anus on the right 
side. As coiling proceeds this process would be accentuated, 
and so, apparently, the condition which obtains in the adult 
Gasteropod is reached without the aid of the late develop- 
ment of the mantle cavity relied upon by Biitschli. Though 
this is put forward merely as a phylogenetic theory, Plate 
believes that the facts of development will fully bear it out, 
and it is only from this point of view that we can deal with 
it here. 

Pelseneer (14) was the first to put aside the old point of 
view. ‘l’o him it seemed that embryologically two distinct 
processes took place, both of which had for their object, as it 
were, the approximation of mouth and anus. ‘The first of 
these, which he calls “torsion ventrale,”’ leaves the embryo 
still symmetrical, but with the alimentary canal bent sharply 
so that the anus hes far forward ventrally. he mouth and 
anus being prevented from approaching nearer alone this 
line on account of the outgrowth of the foot, the second 
process comes into play. This isa“ torsion verticale,” which 
takes place at right angles to the last, and has the result of 
all the organs contained in the shell undergoing a rotation 
through 180°, the ventral anus thereby becoming dorsal, the 
organs of the original right side being carried over to the 
left, and those of the original left to the right. 

More recently Amaudrut (1) has approached the same 
problems from the point of view of comparative anatomy, 
and, from a study of the cesophagus and adjacent organs of 
a number of Gasteropods, has come to the conclusion that 
the region between the head and the visceral hump has 
undergone a twist through 180°. This, of course, would fit 
in well with Pelseneer’s observations, for, if the whole visceral 
hump has undergone torsion with regard to the head, the 
oesophagus must needs be twisted. 

Finally, Boutan (2) has brought out a paper on the 


126 ISABELLA M. DRUMMOND. 


asymmetry of the Gasteropods, which upholds essentially the 
same view of torsion for the Prosobranchs as Pelseneer and 
Amaudrut had already enunciated, and which derives its 
chief value from the author’s claim to have actually observed 
the vertical torsion take place in the case of Acmecea. 

It is at once evident that the processes which, broadly 
speaking, characterise respectively the two classes into which 
we divided the theories under discussion, will not have 
entirely similar results. Both, indeed, alike have, as their 
chief results, the forward dorsal position of the anus and the 
crossing of the visceral connectives, for it was to account for 
these facts that the theories were originally framed; but, on 
the other hand, the twisting of the cesophagus, if true, could 
never have arisen from the processes which Biitschli 
describes, while the growth of the mantle must be conceived 
quite differently, according to which hypothesis is accepted. 
Tf Biitschli is correct, what was originally right remains on 
the right side throughout; while, according to the view of 
vertical torsion, the mantle, and therefore also the shell, share 
in the displacement of the palleal organs. The same holds 
good for most of the viscera, and is especially clearly illus- 
trated in those organs which lie dorsally or ventrally ; for 
while on the theory of unequal growth a lateral shifting 
might easily take place in the same manner as is the case for 
the palleal organs, a dorso-ventral displacement is only 
readily understood on such a theory as Pelseneer’s. Thus 
Plate has to account for the gonad, which is dorsal in 
the Chitons, having a ventral position in the Gasteropoda, by 
supposing that a lobe of the great liver of the left side grew 
dorsally to it and pressed it against the foot. 

Taking these considerations separately, and beginning with 
the last, the facts of embryology seem to me to show in a quite 
unequivocal manner that actual rotation of the organs has 
taken place round an axis coinciding with the cesophagus in 
its direction. It is for this purpose that the drawings of 
figs. 11 to 17 were made, and a comparison of these with one 
another, bearing in mind that they are all orientated on the 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 127 


page in the same way with regard to head and foot, shows 
clearly how stomach, liver, pericardium, kidneys, and mantle 
cavity have all rotated in a perfectly definite manner, while 
retaining unaltered their relations inter se, and explains 
both the original dorsal and later ventral position of the 
gonad, without the intervention of any dorsally growing liver 
lobe. It is, moreover, striking that the torsion of the visceral 
connectives and the apparent twist of the cesophagus noticed 
on p. 119, keep pace perfectly with this rotation, so that it is 
almost impossible not to connect the two phenomena. 

A comparison of figs. 16 and 17, on the other hand, causes 
some little difficulty, for although torsion seems to have taken 
place through an angle of very nearly 180° in fig. 16, there 
seems to be an apparent twist of about 90° further in fig. 17, 
and so it may seem that we have proved too much by this 
comparison of transverse sections. ‘I'wo facts are, however, 
noticeable. In the first place, whereas up to Stage G, a 
transverse section of which is shown in fig. 16, the corre- 
spondence is perfect between the degree of torsion shown by 
the cesophagus and connectives on the one hand, and the 
rotation of the organs on the other, this is no longer the 
case in Stage H, where the connectives are only twisted 
through 180°, while the organs in the posterior region of the 
visceral hump have apparently rotated through about 270°. 
It is indeed true that an accessory twist has been noticed 
for the connectives as well as for the other organsin Stage H, 
but this is clearly marked off from the regular twist corre- 
sponding to that of earlier stages, which takes place more 
anteriorly; and a study of the development of the mantle 
cavity, and its disposition in this stage, makes it clear that 
the two twists are quite unconnected. The visceral commis- 
sure is formed from the floor of the mantle cavity, and, as the 
original right portion of the latter is carried far down on the 
left side of the body, so also is the origin of the corresponding 
half of the visceral commissure; while, as the original left 
side of the mantle cavity remains feebly developed, it, and 
consequently also the original left portion of the commissure, 


128 ISABELLA M. DRUMMOND. 


lie more dorsally. The commissure thus comes to lie obliquely 
quite independently of any direct connection with the torsion 
of the body, and this irregularity is accentuated by a tendency 
on the part of the commissure to pass even more over to the 
left side of the body, and so place itself symmetrically with 
regard to the asymmetrical mantle cavity ; or, in other words, 
symmetrically between its two extreme points of origin. This 
accessory twist may, therefore, be left out of account for the 
present purpose, and we are justified in saying that, as far as 
true torsion is concerned, the connectives only show a twist 
of 180°. 

The second point to be noticed is that, whereas growth in 
the circumference of the visceral hump has been hitherto so 
slow as to be almost inconsiderable—for instance, between 
Stages D and G—though at the same time torsion has 
advanced rapidly, an enormous growth has taken place 
between Stages G and H. I believe that it is in these two 
facts that we find the solution of the difficulty. Between 
Stages G and H a kind of accessory or false torsion has taken 
place among the organs in the posterior region of the visceral 
hump, due merely to unequal growth amongst themselves, 
and not having as its concomitant further true torsion in the 
anterior region. If we seek further for the cause, I think we 
find it in the lenethening of the alimentary canal, and the 
great development of the stomach, which have together 
brought about the present relation between the stomach and 
liver, and also in the sudden rapid widening of the original 
left portion of the pericardium, which has hitherto been so 
narrow, a process which would have, as a natural result, the 
pushing of both kidneys more towards the definitive left side 
of the body than they were before. We have some evidence, 
then, for believing that the changes which are noticeable in 
the position of the organs in the body in successive stages 
are at first due to a rotation of the whole visceral hump upon 
the head through an angle of 180°, but that after this is 
complete a further apparent rotation affecting the posterior 
region of the body only is induced by unequal growth and 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 129 


consequent rearrangement of the organs within the visceral 
hump. ‘These two processes are, of course, quite distinct. 

It has already been necessary to touch upon the subject of 
the torsion of the cesophagus. Amandrut, as noticed above, 
has worked out with great care the twisting of the anterior 
aorta, the salivary glands, etc., about the cesophagus, and has 
come to the conclusion that this is due to an actual torsion of 
the region of the body between the head and visceral hump, 
but hitherto, so far as I know, no embryological evidence ha 
come to hand. I have already described (p. 119) the curious 
compression of the cesophagus, and the manner in which, in 
transverse sections, the long axis changes its direction in 
Paludina embryos. This is not direct proof, but it is difficult 
to find any other explanation of the occurrence, except that it 
is due to the cesophagus being forced to undergo an actual 
twist. In quite old embryos, and in the adult, the cesophagus 
is no longer compressed, and this appearance is quite lost. 

Now, granted that such a twist does take place, it follows, 
as Pelseneer points out, that originally i. e., in the untwisted 
forms, the shell, if coiled, must have been coiled exogas- 
trically. It is exceedingly difficult to get any direct evidence 
upon this point, for,as Plate remarks, we cannot rely on the shell 
of any of the primitive Prosobranchs, like Fissurella, as these 
have all undergone torsion, and, on either view, an exogastric 
shell, if present, must be secondary ; while, on the other hand, 
coiling does not begin to take place sufficiently early in the 
course of development to give us clear evidence either way. 
It is, however, worthy of remark that all the monstrosities 
that I came across which, for some reason or other, had 
remained untwisted, if they showed any tendency towards coil- 
ing of the visceral hump at all, had begun to coil exogas- 
trically. Of these the most highly developed is shown in 
fig. M u, and has been already described. ‘There were, 
however, two others that showed a distinct tendency in the 
same direction, while I did not find one with anything like 
an endogastric coil. This fact seemsto me highly significant. 

The last of these broadly marked points of difference 

vou. 46, PART 1.—NEW SERIES. I 


13 ISABKLLA M. DRUMMOND. 


between the two general theories under consideration relates 
to the symmetrical growth of the mantle, and this is the only 
one the evidence upon which seems still to point in favour of 
Biitschli’s view ; for, if the whole visceral hump has under- 
gone a rotation, we should expect to find signs in the 
innervation. But the definitive right side belongs to the 
original left, and vice versA; whereas it is well known that 
the right pleural ganglion gives rise to the mantle nerve of 
the definitive right side of the body, and the left pleural 
ganglion to that of the definitive left. How this can be 
explained upon Pelseneer’s view is not quite clear, unless it 
may be that the mantle nerve is embryologically a late out- 
growth of the pleural ganglion, and is altogether post- 
torsional; but if this be so we shall have to admit a 
discrepancy between embryology and phylogeny. ‘The case, 
however, is not altogether easy, even for the upholders of 
siitschl’s theory of the symmetrical growth of this region 
of the body, for the innervation is not wholly symmetrical. 
As Bouvier (8) remarks, “‘ la branchie n’est rien autre chose 
qu'une formation palléale, et les mémes nerfs qui Pinnervent 
se répandent en méme temps dans le manteau,” and, as is 
well known, the definitive left ctenidium is innervated from 
a ganelion belonging to the original right side of the body. 
The question, then, is one of some difficulty whichever view 
one takes, and of hardly greater difficulty in the one case 
than in the other. 

An examination of the broad features of the two great 
classes into which we divided the theories of Gasteropod 
torsion leaves, then, a balance of embryological evidence in 
favour of that class of which Pelseneer was the first exponent. 
It will be well now to examine the individual theories more 
in detail. Biitschli’s views have already undergone criticism 
at the hands of Pelseneer and others as not corresponding to 
the facts of ontogeny as one sees them. Amaudrut criticises 
him more particularly with regard to the position of the 
supra- and sub-intestinal gangha. The  supra-intestinal 
ganglion, he says, is usually situated further back than its 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 131 


fellow on the other side of the body. Now, as the former 
was originally on the right side of the body, it would fall 
into the zone where Biitschli supposes that growth ceases, 
while the subintestinal ganglion, on the other hand, would 
be originally in a zone of very active growth. This, Amaudrut 
argues, should lead to the snbintestinal ganglion being 
pushed the further back of the two, while precisely the 
opposite is actually the case. If, however, we suppose that 
instead of the left it was at first the right side which grew 
most actively the existing condition of affairs would be 
obtained; and as, after torsion, the zone of active growth 
would be transferred from the right to the left, we probably 
find an explanation of Biitschli’s error in his having exa- 
mined too late stages of development. ‘he argument 
concerning the position of the ganglia seems sound so far 
as it goes, but the alleged reason of Biitschli’s mistake is 
not so easy to accept, seeing that he starts from a form 
in which the anus lies in the middle line behind, and in 
which even the ventral flexure has apparently not yet 
begun. 

Biitschli himself relies for his evidence of unequal growth 
upon having demonstrated that in different stages of develop- 
ment the mouth and anus do, as a fact, remain exactly the 
same distance apart while the body is increasing in size and 
torsion is taking place. Now, it seems to me that the value 
of this argument depends largely upon the view we take of 
the development of the mantle cavity. If we regard it as 
von Hrlanger did, and as I beleve we must regard it, as 
the result of an outgrowth of the mantle rather than as an 
invagination of the surface of the body, Biitschli’s argument 
is entirely destroyed, for then it is not the mouth and anus 
that we must compare so much as the mouth and the back of 
the mantle cavity ; and the fact that the anus lies near the 
outer edge of the mantle cavity shows only that rapid growth 
of the rectum has been taking place in this region, and has 
caused a closer approximation of the anus to the mouth than 
would otherwise have been the case. Whichever view we 


132 ISABELLA M. DRUMMOND. 


may take of this matter, however, we are forced to admit 
that rapid development of the mantle cavity must mean rapid 
erowth in the neighbourhood of its formation ; and we have 
already seen in the descriptive part how a rapid formation of 
the mantle cavity takes place between Stages D and E to the 
original right of the anus, that is, in Biitschli’s zone of 
cessation of growth, and this long before torsion is complete. 
His view of the late formation of the mantle cavity and its 
effect upon the position of the organs seems to me no less in 
entire contradiction to embryological fact. The mantle cavity 
is not formed by any means so late as Biitschli would put it ; 
it takes its origin, in fact, almost simultaneously with the 
first appearance of torsion, and is, as we have seen, strongly 
developed before we reach Stage F. Finally, it is, of course, 
impossible that its formation should have the effect attributed 
to it by Biitschli unless we regard it altogether as an in- 
vagination, which we have not sufficient evidence for doing. 
All the above remarks apply, also, to Plate’s theory, since 
he accepted Biitschl’s in its main features. ‘The former’s 
suggestion, however, that the liver is the first cause of 
asymmetrical growth is an interesting one, and must be 
further examined. His working out of the later stages of 
development, the formation of the hernia, and the production 
of the coil find no support in embryology. A hernia, indeed, 
is formed on the left side in the development of Paludina, but 
this is of quite a different nature from that described by 
Plate, as a comparison between his figures and mine will 
show; for whereas in his theoretical form the hernia contains 
only the liver of one side, and is at the same time the 
beginning of the coil of the visceral hump, the bulging 
out of the side of the body in Paludina is, as the figures 
show, equivalent to the once symmetrical apex of the visceral 
hump, and contains the stomach as well as the liver. It 
is, moreover, in no way comparable to the formation of a true 
coil, which is formed by a distinct outgrowth of the liver at a 
later stage (cf. figs. 13 and 16 with fig. 17). The coil in 
Paludina does not begin till torsion is nearly complete, and 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 133 


therefore it cannot form an ontogenetic cause for the 
forward movement of the palleal complex. Phylogenetically 
the evidence is only negative, but at least embryology gives 
no support to this part of Plate’s theory. 

Putting aside, however, the question of the coiling of the 
visceral hunip, if we follow the progress of torsion from stage 
to stage we can, I think, in no way regard this as dependent 
upon the growth of the liver in each stage. Comparing 
Stages C and D, for instance, both as complete embryos and 
in sections (cf. especially figs. 11 and 12), we find the slight 
growth of the liver more than counterbalanced by the great 
development of the pericardium and mantle cavity, so that 
whereas in fig. 11 a line joining the junction of the stomach 
and liver with the ridge «x, which, for our present purposes, 
may be taken to represent the position of the rectum, divides 
the section into very nearly equal portions, a similar line in 
fig. 12 makes that portion which contains the liver considerably 
smafler than the other, which contains the great original 
right extensions of the pericardium and mantle cavity. In 
the next stage (fig. 13) this has been partly rectified by rapid 
growth of the liver, but in Stage F the inequality is again 
very marked, and, in fact, from Stages EH to G the growth of 
the liver is very slight, while torsion is rapid. 

While entirely repudiating the idea, however, of the liver 
acting, as it were, as the propelling force throughout ontogeny, 
it may yet be possible to agree with Plate in regarding it as 
the original disturber of symmetry ; and some support is lent 
to this view not only by the very early asymmetry of this 
organ, but also by the fact that it is not present, or is only 
very slightly developed, in the symmetrical monstrosities. 
But it should, at the same time, be noticed that the develop- 
ment of Paludina gives no more support to this part being 
acted by the liver than by the mantle cavity. The latter 
organ also is asymmetrical from the time of its first formation 
in normal forms ; while in Monstrosity III, which was shghtly 
twisted towards the left, the mantle cavity is considerably 
developed in a manner to correspond to the torsion, while the 


154 ISABELLA M. DRUMMOND. 


liver is comparatively insignificant. Moreover, in following 
the normal development from stage to stage, it is obvious 
that the great original right-hand extension of the mantle 
cavity much more nearly keeps pace with the torsion than 
is the case with the liver; and, finally, it may be noticed that 
this view would harmonise with Amaudrut’s reasoning con- 
cerning the supra- and sub-intestinal ganglia. It may well 
be, however, that the growth of these organs is to a large 
extent dependent the one upon the other, and that equilibrium 
is maintained by the asymmetrical growth of the liver on one 
side of the body being compensated by asymmetrical growth 
of the mantle cavity on the other. 

We turn now to the other side of the question, but Pel- 
seneer puts forward his theory in such a broad and general 
form that it is difficult to enter into any detailed discussion 
beyond the general considerations which have already been 
adduced in his favour. The chief objection that might be 
raised in this case is, perhaps, one which arises from the 
difficulty, in a course of development like that of Paludina, of 
distinguishing Pelseueer’s two processes of ventral and 
lateral torsion. And, indeed, they do go on so closely hand 
in hand that as a matter of fact the anus travels in an oblique 
and never in a vertical direction. Nevertheless there is 
clear evidence of a vertical rotation of the organs contained 
in the visceral hump, as has been already pointed out, and 
in each stage it is possible to separate from the results of this 
process a certain clearer development of the visceral hump 
and sharper bend of the alimentary canal from that which 
obtained in the previous stage, which must be the result of a 
process akin to Pelseneer’s ventral torsion, or, as Amaudrut 
has better styled it, ventral flexion. In the early stages the 


distinction between the two processes is very clear—as, for 
instance, in Stage C, where the ventral flexion is already 
strongly marked, while the lateral torsion is but just begun. 

With regard to the cause which Pelseneer seeks for these 
processes, however, the development of Paludina offers no 
conlirmation. The growth of the foot, he says, forms an 


THK DEVELOPMEN'’ OF PALUDINA VIVIPARA. 30 


obstacle to the close approach of the mouth and anus, and 
therefore vertical torsion takes the place of the ventral 
flexure. This is really no true explanation at all, for we are 
left in the dark as to how the foot brings about this new state 
of affairs, and we do not get much nearer if we say, with 
Boutan, that there is an antagonism of growth between the 
foot and the visceral hump. It seems to me that in the 
development of Paludina it is altogether out of place to speak 
of such an antagonism, for torsion begins at a time when the 
foot is still but a comparatively insignificant ventral pro- 
jection, and long before the formation of the creeping sole. 
Amaudrut also is unsatisfactory in this respect, for he 
attributes torsion ultimately to voluntary effort on the part 
of the animal to get its gills into a better situation. With 
the main part of Amaudrut’s paper embryology has nothing 
to do; there are, however, one or two points in which it seems 
to me he is mistaken, owing to a too exclusive regard for the 
anterior region of the body. His description of the shell and 
visceral hump as aiding the torsion by their weight is wholly 
inapplicable to embryology; while his account of the 
manner in which the peculiar shape of the mantle cavity and 
the disposition of the organs included in it are induced 
receives complete contradiction. ‘ A peu prés dans le méme 
plan transversal qui passe pas le ganglion viscéral postérieur,”’ 
he says, ‘‘se trouvent la partie terminal de la région tordue 
du tube digestif, le fond de la cavité respiratoire, la partie 
postérieure de la branchie et le coeur. Ce plan marquant en 
arriére la limite extréme de la torsion, les organes qui s’y 
trouvent ont du exécutés un mouvement de rotation d’environ 
180° pour se rendre dans leur position définitive.”” This holds 
good for all organs behind this position, but in front of it 
obviously torsion will be less. It is these facts of which he 
makes use to explain the apparent slope of the mantle cavity 
from the left towards the right anteriorly, and of the rectum 
from mid-dorsal, where theoretically it should be, to the 
right side of the body, the characteristic position for the 
Prosobranchia. In Stage H, however, the region of the 


136 ISABELLA M. DRUMMOND. 


twist of the cesophagus and visceral connectives lies alto- 
gether in the anterior region of the mantle cavity, and in 
front of the anus. All the region behind this, we must 
believe, has equally undergone a torsion of 180°, and yet here 
we have clearly marked that peculiar disposition of the 
mantle cavity which Amaudrut seeks to explain. It seems 
to me that this is not to be regarded as due to torsion at all, 
but to unequal growth ; already in Stage H, as we have seen, 
the main features were present, the mantle cavity reaching 
over the mid-dorsal line behind and being much less ad- 
vanced in front, while the rectum showed no tendency toa 
corresponding disposition. It is not, in fact, till Stage H 
that the position of the anus ceases to be where theoretically 
it should be according to the degree of torsion, and here the 
displacement is due to a sudden bend of the rectum to the 
right quite close to the anus. The final disposition of the 
rectum is due to a forward growth in the direction indicated 
by this bend after torsion is, complete, and may perhaps be 
due, phylogenetically, to the advantage gained by the animal 
in having the anus in a position as far removed from the gill 
as possible. 

Boutan’s explanation of the cause of torsion has already 
been mentioned, and it has been shown to be hardly 
applicable in the case of Paludina. His conception of this 
antagonism of growth between visceral hump and foot is, 
much more than Pelseneer’s, that of an ontogenetic cause, 
for he expresses the opinion most definitely that if this 
antagonism could be suppressed torsion would not take place. 
Now amongst the monstrosities already referred to it is 
certainly true that I did not find one with an abnormally 
small foot and yet a visceral hump which had undergone 
torsion, but, on the other hand, fig. M ir shows how both 
foot and hump may be very highly developed and yet no 
torsion take place. Surely the antagonism of growth, if such 
exist, must be much greater in this case than, for instance, 
in Stage C, where torsion has already begun. ‘This leads 
on to Boutan’s view of torsion as the cause of asymmetry. 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 137 


The normal Gasteropod larva, he says, is perfectly bilaterally 
symmetrical, and remains so till torsion takes place; but as 
soon as this begins the asymmetrical growth of the internal 
organs begins, and, if torsion could be averted, symmetry 
would be maintained. Certainly in this connection the sym- 
metry of Monstrosity II is striking, but that asymmetry and 
torsion are closely connected no one doubts, and whichever 
were the cause of the other, or if both were the outcome of 
some common cause, the result would be the same. In the 
normal course of development it has been repeatedly pointed 
out by other writers that asymmetry in some form or other is 
found before torsion begins. In Paludina torsion begins so 
early that it is difficult to be quite sure of this, but the liver, 
at least, is never wholly symmetrical, and the unequal 
development of the original rudiments of the mantle cavity 
takes place as nearly as possible simultaneously with the 
beginning of torsion. 

Once more, Boutan turns to antagonism between foot and 
visceral hump in order to explain the coiling of the latter. 
If, when the creeping sole of the foot is developed, this can 
stretch out without lateral displacement of the visceral hump, 
then the shell, he believes, will remain symmetrical ; but if 
not, then the hump will be pushed to one side or other, and 
the sense of the future coil will depend upon which side it is 
pushed towards. Such a view, it would seem, would be 
quite impossible to accept after even a cursory view of the 
facts of development. Jor if, as would seem to be the case, 
the coiling of the visceral hump is primarily the result of a 
definite process of growth in the liver (cf. figs. 16 and 17), 
this is altogether independent of the exact relation in which 
this organ finds itself to the foot. As a matter of fact, the 
dextral coil of Paludina begins before torsion is quite 
complete, and therefore, while the apex of the visceral hump 
is still to the left of the foot, by a strong growth of the liver 
towards the right, which, when torsion is complete and the 
visceral hump nearly symmetrical, points to the right and 
upwards. It is altogether inconceivable that any accident of 


138 ISABELLA M. DRUMMOND. 


erowth, independent of torsion, which should cause the apex 
of the visceral hump to remain upon the left side of the foot, 
should alter this growth of the liver, and cause the coil to 
become sinistral. 

With regard to the cause of asymmetry, another view 
remains to be mentioned here, namely, that of Grobben (8), 
who, while accepting Pelseneer’s main conclusions, finds 
himself unable to regard the antagonism between foot and 
visceral hump as a true ontogenetic cause of torsion. He 
grants Pelseneer’s view that the growth of the foot necessi- 
tates a vertical displacement if the anus is to continue to 
approach the mouth, and in order to explain how this 
is produced he has recourse to Plate’s suggestion of 
an unequal growth of the two originally symmetrical 
liver lobes. For reasons already stated when Plate’s 
theory was under discussion this is not altogether satis- 
factory, for the growth of the liver does not keep pace 
with torsion, and the chief development of it takes place 
after torsion is complete. As a phylogenetic cause it may 
have played its part, but probably not quite in the manner 
that is described by Plate. _ 

Thiele (18) also takes up a position somewhat intermediate 
between the two extreme points of view on torsion, for, 
while agreeing with Pelseneer that before torsion the shell 
must have been bent with its apex pointing forwards, as 
though forming the beginning of an exogastric coil, he 
approaches Lang more nearly than anyone else in his view 
of how torsion has been effected. He dismisses Plate’s view 
of the liver as the disturber of the original symmetry of the 
body, and believes this part to be played by the gonad, in 
which he also sees the cause of the coiling of the visceral 
hump. Thus, as, in the Gasteropods, the gonad is formed 
only on the left, the coil also lies with its apex in the left. The 
condition of affairs which is thus reached is shown in his 
fig. 3, p. 15, and this be believes to be a position of unstable 
equilibrium, consequently a sudden rotation is effected till the 
visceral hump comes to rest in the normal adult position. 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 139 


Thiele seeks very little support for his view in the facts of 
development, except in the rapid rotation observed by 
Boutan in Acmoea, and it is seen at a glance to be wholly 
inapplicable to the development of Paludina where rotation is 
gradual, and where both gonad and coil are only formed 
when torsion is far advanced. It brings us, therefore, no 
nearer to forming a conception of the ontogenetic course of 
torsion, and the development of Paludina gives no evidence 
to support it as a phylogenetic theory. 

Whatever view may be held with regard to phylogeny, in 
ontogeny it seems to me that we are ultimately thrown back 
upon the problems of heredity, and for the present we must 
agree with Guiart (9) when he says, “ Mais 4 ceux qui nous 
demanderont la cause mécanique de cette torsion, et qui nous 
reprocheront de ne pas lavoir trouvée chez l’embryon, nous 
répondrons simplement ceci. Il ne faut pas confondre 
ontogénie et phylogénie, la cause w’existe pas chez l’embryon, 
mais chez le mollusque primitif.’ From the nature of the 
case the evidence which ontogeny can give upon the phylo- 
genetic cause 1s merely negative. 


SUMMARY. 


To sum up, then, theories of Gasteropod torsion may be 
divided into two classes : 

a. Those which view the present position of the palleal 
complex as due to a forward movement along the right side 
of the body, which resulted from greater growth of the left 
side of the body than of the right. 

b. Those which view the present position of the palleal 
complex as due to a ventral flexion followed by a vertical 
rotation of the whole visceral hump upon the head. 

The evidence for the second of these views seems greater 
than that for the first, in that— 

1. A vertical displacement through 180° of all the organs 
contained in the visceral hump takes place in the course of 
ontogeny. 


14.0 ISABELLA M. DRUMMOND. 


2. ‘There is some evidence, both from comparative anatomy 
and embryology, for believing that the oesophagus has 
undergone an actual twist. 

3. Monstrosities which retain the palleal complex in a 
ventral position show a tendency to form an exogastric 
coil. 

The innervation of the mantle was shown to be equally 
difficult to explain on either hypothesis. 

Also, against the first view was urged the insufficiency of 
the evidence upon which Biitschlhi bases his conclusions with 
regard to zones of unequal growth. 

With regard to the phylogenetic cause of the vertical 
twist, embryology can only give negative evidence ; while in 
considering the ontogenetic cause we are thrown back upon 
unsolved problems of heredity, and must confess our 
ignorance, 

In conclusion, I wish to offer most hearty thanks to 
Professor Weldon, not only for having placed freely at my 
disposal all the resources of the laboratory, but also for most 
kind personal aid at all stages of the work. My thanks are 
also due to Mr. Richard Evans for much help in the 
technique. 


List or WorKS REFERRED TO IN THE TExt. 

1. Amauprut, A.—“ La Partie antérieure du Tube digestif chez les 
Mollusques Gastéropodes,” ‘ Ann. des Sci. Nat. Zool. (8), vil. 

2. Bouran, L.—‘‘ La Cause principale de l’Asymétrie des Mollusques 
Gast éropodes,” ‘Arch. de Zool. exp.’ (3), vil. 

3. Bouvier, K. l.—-“ Systéme nerveuse, Morphologie et Classification des 
Gast éropodes Prosobranches,” ‘ Ann. des Sci. Nat. Zool.’ (7), iii. 

4. Burscnu, O.—* Bemerkungen ueber die wahrscheinliche Herleitung 
der Asymmetrie der Gastropoden,” ‘ Morph. Jalirb.,’ xii. 

5. Wroancer, von R.—* Zur Entwicklungsgeschichte von Paludina 
vivipara,” ¢ Morph. Jahrb.,’ xvii. 

6. Extancer, von R.—“*Zur Entwicklungsgeschichte von Paludina 
vivipara (vorlaufige Mittheilungen),’’ ‘Zool. Auz.,’ xiv. 

7. Wruancer, von R.-—“On the Paired Nephridia of Prosobranchs,” 
*Q. Journ, Mier. Sei.,’ xxxiii. 


THE DEVELOPMENT OF PALUDINA VIVIPARA. 141 


8. Grospen, K —‘ Kinige Betrachtungen ueber die phylogenetische Ent- 
stehung der Drehung und der asymmetrischen Aufrollung bei den 
Gastropoden,” ‘ Arb. Zool. Inst. Wien,’ xii. 

9. Gurart, J.—‘‘ Gast éropodes Opisthobranches,” ‘Mém. de la Soc. Zool. de 
France,’ xiv. 

10. Hater, B.—“* Betrachtungen ueber die Phylogenese der Gonade und 
deren Miindungsverhaltnisse bei niederen Prosobranchiern,” ‘ Zool. 
Anz.,’ XXxill. 

11. Hatter, B.—‘ Studien ueber Docoglosse, ete.,’ Leipzig, 1894. 

12. PrtsenerrR, P.—“ Les Reins, les Glandes génitales et leurs Conduits 
dans les Mollusques,” ‘Zool. Anz.,’ xix. 


13. PELSENEER, P.—‘ Recherches morphologigues et phylogénétiques sur les 
Mollusques archaiques,’’ ‘Mém. cour. et Mém. des Savants étrangers 
de Acad. R. de Belg.,” lvii. 


14. Petsenerr, P.— Recherches sur divers Opisthobranches,” ‘Mém. cour. 
et Mém. des Savants étrangers de l’Acad. R. de Belg.,’ lili. 


15. Prare, L.— Bemerkungen ueber die Phylogenie und die Entstehung der 
Asymmetrie der Mollusken,” ‘ Zool. Jahrb.,’ ix. 


16. Simrotn, H.—‘ Bronn’s Klassen und Ordnungen des Thierreichs,’’ B. iii. 
5 > 


17. Tonnices, C.—“Zur Organbildung von Paludina vivipara,” ‘S. B. Ges. 
Bef. d. ges. Naturw., Marburg,’ 1899. (Abstr. in ‘Zool. Centralb.,’ vi.) 


18. ‘TureLe, J.—‘‘ Ueber die Ausbildung der Korperform der Gastropoden,” 
‘Arch. f. Naturgeschichte,’ Ixvii, 1901. 


DESCRIPTION OF PLATES 7—9, 


Illustrating Isabella M. Drummond’s paper, ‘Notes on the 
Development of Paludina vivipara, with special 
reference to the Urinogenital Organs and Theories of 
Gasteropod Torsion.” 


Significance of Reference Letters. 


a. Anus. aa. Line representing a median ventral plane through head and 
foot. ect. Outer epithelium of the body. £ Foot. yg. Gonad. 4%. Heart. 
k. Kidney. 7. &, Original right kidney. 7. 4. Original left kidney. 4%. d. 
Kidney duct. /, Liver. m. Mouth. m.c. Mantle cavity: 7. m. e. Original 
right horn; Z. m.c. Original left horn. m./. Mantle fold. a@s. (Esophagus, 


142 ISABELLA M. DRUMMOND. 


op. Operculum, of. Otocyst. pe. Pericardium: 7. p. e. Original right 
division ; 2. p. ¢. Original left division. p. g. Pedal ganglion. p. x. Pedal 
nerves. 7. ap. Renal opening into the mantle cavity. vec. Rectum. 7.4. ap. 
Reno-gonadial aperture. 7. pe. ap. Reno-pericardial aperture of the original 
left kidney. 7.8, Radula sac. s.g. Shell gland. st. Stomach. sad. e. Sub- 
cesophageal connective. sep. c. Supra-cesophageal connective. 7. Tentacle. 
v. Velum. ves. and ves’. Vesicles attached to kidneys. v. 2. Visceral hump. 
a. Ridge between the two original horns of the mantle cavity on which the 
anus opens. 


PLATE 7. 


Fic. 1.—Slightly oblique transverse section through the visceral hump of 
an embryo between the age of that shown in Fig. C and that in Fig. D. 
Both kidneys and the first. rudiment of the heart are shown. xX 330. 


l'te. 2.—Oblique section through the extreme (original) left portion of the 
pericardium of an embryo rather older than that shown in Fig. G. The 
sharp curve of the surface of the body shown at ecé. indicates that the 
visceral hump is just beginning to coil, x 380. 

ie. 3.—Another section of the same series, and showing the same region, 
but rather more posterior. Two sections intervene between Figs. 2 and 3. 
x 330. 

Fig. 4.—The next section posterior to Fig. 3, and showing the same region. 
x 330. The figure should be rotated about 9 degrees to the left to compare 
with Fig. 3. 

Fic. 5.—A transverse section through the same region as the above, but of 
an older embryo in which about one complete turn of the spiral coil of the 
visceral hump is complete. It is one of the same series as Fig. 17. x 3880. 


Fira. 6.—An ideal longitudinal section of the whole genital apparatus of an 
advanced embryo, with about two turns of the spiral complete, reconstructed 
from a series of tiansverse sections. The gonad is represented spread out 
instead of coiled. x 140. 

l'ie. 7. —A transverse section across the region aa. of Fig. 6. x 380. 

Fig. 8.—A transverse section across the region 6d. of Fig. 6, showing also 
the close proximity of the genital organs to the liver. x 330. 

lic. 9.—A transverse section across the region cc. of Fig.6. x 380. 

Fie. 10.—Sagittal section through an embryo belonging to Stage A. x 
330. 

Fie. 11.—Transverse section through the posterior region of the visceral 
hump of an embryo belonging to Stage C. x 119. 


THE DEVELOPMENT OF PALUPINA VIVIPARA. 148 


Fig. 12.—Transverse section through the posterior region of the visceral 
hump of an embryo belonging to Stage D. x 119. 

Vie. 13.—Trans verse section through the posterior region of the viscera} 
hump of an embryo belonging to Stage E. x 119, 

Fic. 14.—Another section from the same series but passing through the 
anterior region of the visceral hump, and showing the creat anterior or right 
extension of the mantle cavity. x 119. The section of the overhanging 
foot, /, is introduced here in its relative position, but is omitted in Fig. 13. 

Fie. 15.—'Transverse section through the posterior region of the visceral 
hump of an embryo belonging to Stage F. x 119. 

Fre, 16.—Transverse section through the posterior region of the visceral 
hump and foot of an embryo belonging to Stage G. x 119. 

Fie. 17.—Transverse section through the posterior region of the visceral 
hump and foot of an embryo belonging to Stage H. x 119. 

Fies. 18 and 19.—Transverse section through the “neck” region of the 
same stage, showing the twist of the visceral connectives. x 119. 

Fic. 20.—Transverse section through the visceral hump of the monstrosit 

to) 
shown in Fig. 10, Plate 8, in the region of the kidneys. x 87. 


Fic. 21.—Another section of the same series passing through the opening 
of the kidney duct into the mantle cavity. x 87. 


ie. 22.—Transverse section through the stomach of the same embryo just 
behind the opening into it of the cesophagus. x 80. 


I'ic. B.—View of an embryo belonging to Stage B from the right side. x 


Fie. C.—View of the right side of an embryo belonging to Stage C. x 
70. 

Fig. C,.—View of the ventral surface of an embryo slightly older than the 
last: x70: 

Fic. D.—View of the left side of an embryo belonging to Stage D. x 70. 

Fie. f.—View of the left side of an embryo belonging to Stage NH. x 70. 

Fic. f,.—View of the right side of the same. x 70. 

Fre. Kj.—Dorsal view of the same. x 70. 

Fic. F.—View of the left side of an embryo belonging to Stage F. x 70. 

Fie. G.—View of the left side of an embryo belonging to Stage G. x 70. 

Fie. M 1r.—View of the left side of Monstrosity Il. x 70. 

Fic. M 111.—View of the left side of Monstrosity II. x 70. 


sS 


te] ’ ie CST! GnKe 


sath ‘ 9 
‘ mis ey Sites itp olin ere TS | rea! 


7 PS ee LUT ieee? wat) ty -z ik 


ipod £2 @ gre a eal?) 6a? Get AD 


am | Pc ria at, ba | sl OCIA 1 Gt 
=: ' Tl Ni: ae ih a> ot1@ @, pane . 
3 oe an ee 


Ps " ag Livse é e 
i ay i yt’ ! r= 
oI 24F ie : / 
y, - i i , - i Et Ww 
(hy 7 Yi] al ; ; Titel <0 
’ S '; i rsy~ Ola he 


? ih) Bees 6 LG 0! 4s ©) PU gin fl Ly = % 146 we 9 ae * 


a gui = sine) Am. 
Pei i406 el et) Al pe coinage! y pre i ae oe 


¢ 


2. e) é | ai ae A TG <4 é ce bey ‘ al Dp 1? tal 3 


OA te Tt fume a Louie dss ieee hs dha Pal 

PB ot) Oy ee 

‘ | cyt't ig) Gal ee Bw hive Rypmess tral fully 

oar i wi! VP gules yews G0 luis tal isl fe Bee ree 

= Oy & Sed Wet 1h oie Pees ” 2 (Oh, * 

ati. 3 aie 4 tame ‘ 7 

“ 1 > > ai apni of “i wa be ia wil, i iiee 7. eet 
a oS ot ined ascii sie. ill oil (ethene 

A ee eet fat 

Ee eS diet 


FERTILISATION OF THE EGGS OF ANIMALS. 145 


Is Chemotaxis a Factor in the Fertilisation of 
the Eggs of Animals? 


By 


A. H. Reginald Buller, B.Se., Ph.D., 
Lecturer in Botany at the University of Birmingham. 


ContTENTS. 

PAGE 
I. IntRopucTION . : : : : ; . 145 
Il. Some Frnrrintisation Prorgiems : ; ‘ 5 Us) 
WT. Marertan : 3 : : ; : loi! 

1V. Remarks wpon THE FEoecs aNnbD SPERMATOZOA OF THE 
EcHINOIDEA . E ; : : 5 alte, 
V. Tue Cuemoractic QursTIon . : ‘ ; = los 
VI. Tue Movements or SPERMATOZOA UPON SURFACES . 7 59 
VII. Tue Direction or PENETRATION OF THE GELATINOUS Coat. 167 
VILL. Toe ArracuMENT oF Spermatozoa To THE Eee 4 . 174 
IX. Summary or THE Cuter Resutts : : ; 5 YS 


I. IntRopuction. 


Te well-known researches of Pfeffer! have demonstrated 
the importance of the part played by chemotactic stimuli in 
causing the spermatozoa of liverworts, mosses, ferns, etc., to 
approach the oospheres. Among the yet higher plants— 
Gymnosperms and Angiosperms—the chemotropism* of 

1 TLocomotorische Richtungsbewegungen durch chemische  Reize,” 
* Untersuchungen aus d. Bot. Inst. zu Tubingen,’ 1884, Bd. i, p. 363. 

2 Molisch, “* Ueber die Ursachen der Wachstumsrichtungen bei Pollen- 
schliiuchen,” ‘Sitzungsber. der Kais. Acad. d. Wiss. in Wicn,’ 1889 and 1893. 
Also Lidforss, “ Ueber den Chemotropismus der Pollenschlauche,”’ ‘ Ber. d. 
D. Bot. Gesell.’ 1895, Bd. xvii, p. 236. 

VoL, 46, part 1.—NEW SERIES. K 


146 A. H. REGINALD BULLER. 


pollen-tubes takes the place of the chemotaxis of spermatozoa. 
We may therefore say that actual contact of the sexual 
elements of all plants from the liverworts onwards is brought 
about by chemical stimuli. 

Tn all the above-mentioned groups of plants the oospheres 
ure fertilised in their place of origin without being set free. 
The chemical stimulus, so far as is known, does not arise 
directly from the oospheres. The spermatozoa of the vas- 
cular cryptogams are attracted into the archegonia by a 
substance liberated from the cell-sap of the neck-canal-cells. 
The pollen-tubes are guided on their tortuous way to the 
oospheres by substances excreted by various tissues of the 
ovary and ovules. 

On the other hand, the ova of animals are fertilised after 
being set free from their place of origin, namely, the ovary. 
Fertilisation takes place in the case of terrestrial animals, 
e.@. mammals, reptiles, birds, and insects, in the oviduct, or, 
as happens with many aquatic animals, e.g. Echinoderms, 
many fishes, and amphibia, after the eggs have been deposited 
in water. If, therefore, chemotaxis plays a role in bringing 
the spermatozoa of animals into contact with the ova, the 
source of stimulation must be looked for in a substance 
excreted from the eggs. 

It appears to be the general opinion among zoologists that 
chemotaxis is actually a factor in the fertilisation of animal 
egos. Thus Bergh! says that during an artificial fertilisation 
experiment, e.g. in the case of the sea-urchin, “ the sperma- 
tozoa collect around the ripe eggs, probably attracted by a 
special substance.”’ 

Wilson,” in his latest edition of ‘The Cell,’ in dealing 
with the union of the germ-cells, remarks: “ There is clear 
evidence of a definite attraction between the germ-cells, 
which is in some cases so marked (for example, in the polyp 
Renilla) that when spermatozoa and ova are mixed in a small 

1 Bergh, * Vorlesungen uber allgemeine Embryologie,’ 1895, p. 43. 

2 Wilson, ‘The Cell in Development and Inheritance,’ 2nd ed., 1900, 
p. 196. 


FERTILISATION OF THE EGGS OF ANIMALS. 147 


vessel, each ovum becomes in a few moments surrounded by 
a dense fringe of spermatozoa, attached to its periphery by 
their heads, and by their movements actually causing the 
ovum to move about. The nature of the attraction is not 
positively known, but Pfeffer’s researches on the spermatozoa 
of plants leave little doubt that it is of a chemical nature. 

‘The experiments indicate that the specific attrac- 
tion between the germ-cells of the same species is owing to 
the presence of specific chemical substances in each case.” 
Here it may be at once remarked that the collection of 
spermatozoa attached by their heads to the eggs in artificial 
fertilisation experiments is no proof whatever that the 
spermatozoa have been attracted from a distance to the ege 
by a substance excreted from the latter. All that we can say 
in such a case, without further observation, is that the eggs 
retain the spermatozoa after these have come in contact with 
them. 

Verworn! goes so far as to say : “The splendid and methodi- 
cal researches of Pfeffer upon chemotropism had their origin 
in observations upon the spermatozoa of forms in which 
chemotropic relations to the egg-cell were discovered. Such 
relations, as we now know, have analogies in almost the 
whole of living nature and for the fertilisation of the eggs of 
animals by spermatozoa, just as for the eggs of plants, form 
an indispensable condition. The spermatozoon seeks the egg 
and is guided on the right course everywhere in the living 
world by a chemotropic action, which the metabolic products 
of the eggs exercise upon the free-swimming spermatozoa. 
That from the innumerable hosts of spermatozoa of the most 
diverse animals which in many places cloud the sea, each 
species finds its right and specific ege, a phenomenon which 
would otherwise excite astonishment, is in the great majority 
of cases a direct result of chemotropism, and easily explains 
itself on the ground that each spermatozoon is chemotropi- 
cally attracted by the specific substances which characterise 
the eggs of the species concerned.” It is one of the objects 

1 Verworn, ‘ Physiologie,’ 1895, p. 425. 


148 A. H. REGINALD BULLER. 


of this paper to show that such sweeping generalisations with 
regard to animals are so far entirely without experimental 
justification. 

With mosses, ferns, etc., there is as yet no proof that the 
egos attract the spermatozoa to them in the manner in which 
Wilson, Verworn, and others believe to be the case with 
animals. As was pointed out, the eggs of these plants are 
fertilised in their place of origin. ‘This permits of the 
surrounding cells, neck-canal-cells, and ventral canal-cell 
taking upon themselves the function of chemically attracting! 
the male sexual element to the female. ‘The eggs may not 
do more than simply retain the spermatozoa after contact has 
taken place. Since the eggs of animals are fertilised after 
liberation from their place of origin, there 1s no chance of 
such a division of labour as occurs with plants. In the 
analogy made by zoologists between ferns and animals there 
is thus a weak point. Credit is given to the reproductive 
egg of animals for an excretory function, which has not been 
demonstrated in the case of plants. 

There is one group of Algee—the Fucaceee—which are 
unique among plants in that their eggs, like those of the 
Echinoidea, are fertilised after extrusion into water. The 
egos of the Fucacez differ, however, from those of most 
animals, in being perfectly naked during fertilisation, and 
in containing chlorophyll which assimilates? in the light. 

1 The neck-canal-cells and ventral canal-cell secrete in the cell-sap of their 
vacuoles an attractive substance or substances (probably a salt or salts of 
malic acid). When the archegonium bursts these cells burst too, and die, 
thus liberating their cell-sap, which diffuses slowly out of, and from, the 
mouth of the archegonial tube. Pfeffer, loc. cit. 

2 This fact I was able to prove by means of Engelmann’s method, using, 
however, the spermatozoa of a sea-urchin instead of bacteria. A vast number 
of spermatozoa were added to a preparation containing a few eggs of 
Cystocyra barbata (one of the Fucacee). ‘The spermatozoa not in the 
neighbourhood of the eggs came to rest in five minutes. Those around the 
eggs continued in motion for more than an hour. ‘The movement also took 
place around non-nucleated fragments of eggs. When the light was cut off 
from the eggs the movement quickly ceased, to return again when light was 
once more admitted, 


FERTILISATION OF THE EGGS OF ANIMALS. 149 


According to Strasburger! they excrete a substance which 
attracts the spermatozoa from a distance equal to two 
diameters of an ege. On the other hand, the observations 
and experiments by Bordet? upon the fertilisation of the eggs 
of several species of Fucus led him to entirely negative 
conclusions as regards a chemotactic attraction, while he 
found that the spermatozoa were highly sensitive to contact. 
According to this observer it is simply the ability of the 
spermatozoa to adhere to surfaces by the tip of one of their 
two cilia, which leads to their collection upon an egg, while 
their meeting with it is simply a matter of chance. A few obser- 
vations of my own at Naples upon the fertilisation of Cysto- 
cyra barbata (one of the Fucacez) did not reveal to me any 
certain attraction of the spermatozoa from a distance, but the 
collection of the spermatozoa upon the eggs in consequence of 
their ability to cling to surfaces was clearly seen. Neverthe- 
less, in view of the positive statement of Strasburger, a careful 
reinvestigation of the question seems to me desirable. 

The other cases® of supposed attraction of spermatozoa to 
the egg-cells of plants all await a critical study. 

In the cases of Clamydomonas and of Ulothrix,' Pfeffer 
has observed that the meeting of the swarm-spores, which 
afterwards copulate, is purely a matter of chance. He also 
found that the spermatozoa of a bull? were not attracted by 
meat extract. 

At present, to the best of my knowledge, not a single case 
is known where chemotaxis plays a role inthe fertilisation of 
the eges of animals. 

Dewitz® has shown that the spermatozoa of certain insects 

1 Strasburger, ‘ Das bot. Prakticum,’ 2 Aufl., 1887, p. 402. 

2 Bordet, “ Contribution al’ Etude de l’Irritabilité des Spermatozoides chez 
les Fucacées,” ‘ Bull. de l’Acad. Belgique,’ 3e sér., tome xxvii, 1894, p. 889. 

3 See Pfeffer, loc. cit., pp. 446—449. 

4 Loe. cit., p. 447. 

5 Loc. cit., p. 449. 

6 Dewitz, ‘‘ Ueber Gesetzmiassigkeit in der Ortsverinderung der Sperma- 


tozoen und in der Vereinigung derselben mit dem Ki,” ‘Arch. f. die gesammte 
Physiologie,” Bd. xxxviil, 1886, p. 558. 


150 A. H. REGINALD BULLER. 


find their way to and through the micropyles of the eggs 
owing to the remarkable fact that on coming to a surface 
they remain in contact with it and continue to move in circles. 
This characteristic, which will be discussed more fully after- 
wards, I have found also shared by the spermatozoa of repre- 
sentatives of every group of the Echinodermata. 

Massart! made a careful investigation of the fertilisation of 
frogs’ eges. He came to the conclusion that the spermatozoa 
come in contact with the gelatinous coat by accident, and 
cling to it owing to a special sensibility to contact. He found 
that they bore through it radially. He believed that this is 
explained on the supposition that the spermatozoa seek to 
bore from the more watery outer layers to the less watery 
inner layers in consequence of a sensibility to the differences 
of saturation. 

My own investigations, undertaken at Naples, were made 
to determine the nature of the forces which bring the 
spermatozoa and eggs of the Hchimoidea in contact, especial 
attention being paid to the chemotactic question. ‘lhe work 
was taken up after a fairly extended study of the chemotaxis? 
of the spermatozoa of ferns. 


II. Somn Ferrinisation PROBLEMS. 


In the case of such eggs as those of the Echinoidea, which 
are surrounded by a thick gelatinous coat, some of the 
physiological questions that may be asked in regard to the 
manner in which the spermatozoa meet and fuse with them 
are as follows: 

1. Does a spermatozoon meet the gelatinous coat (zona 
pellucida) by accident, or is it attracted to it by some 

' Massart, (1) “Sur VIrritabilité des Spermatozoides de la Grenouille,” 
‘ Bull. de Acad. roy. de Belgique,’ 3me sér., t. xv, No.5, 1888; (2) “Sur la 
Pénétration des Spermatozoides dans uf de la Grenouille,” ‘ Bull. de PAcad, 
roy. de Belgique,’ 3me, sér., t. xvii, No. 8, 1889. 

* Buller, “Contributions to our Knowledge of the Physiology of the Sper- 
matozoa of Ferns,” ‘ Ann. of Botany,’ vol. xiv, 1900, p, 543. 


FERTILISATION OF THE EGGS OF ANIMALS. boil 


chemotactic substance which is excreted by the hving egg 
and diffuses through the gelatinous coat into the surrounding 
water ? 

2. After a spermatozoon has come in contact with the outer 
surface of the gelatinous coat, is it retained there mechani- 
eally or in consequence of a tactile stimulus exerted upon it 
by the surface ? 

3. Does the spermatozoon bore through the gelatinous coat 
radially ? Ifso, why? 

4, After reaching the outer surface of the hving egg (1. e. 
the protoplasm), what is the nature of the forces which lead 
the spermatozoon to unite with it ? 

5. Closely connected with the latter is the further question: 
by what means is the progress of a spermatozoon from the 
surface to the interior of an egg brought about ? 


TIT. Marerrat. 
The following species of Echinodermata were made use of : 


(Echinus microtuberculatus, Bly. 

Class Hchinoidea |Spherechinus granularis, Ag. 
Reeulares parbaieke pustulosa, Gray. 

\Strongylocentrotus lividus, Brdt. 
Irregulares Hchinocardium cordatus, Gray. 
Asterias glacialis, O.F.M. 
ate ais sepositus, Mill. Tr. 
Ophioderma longicauda, Mill. Tr. 
(Ophioglypha lacertosa, Lyman. 
Class Holothuroidea Holothuria Stellate, D.Ch. 


Class Crinoidea . Antedon rosacea, Norman. - 


Class Asteroidea 


Class Ophinroidea 


Observations upon the motility, especially in regard to 
surfaces, were made upon the spermatozoa of all. the above 
species. The experiments and observations upon fertilisa- 
tion were restricted to the first three Echinoidea, namely, 
Echinus, Spherechinus, and Arbacia. 


152 A. H. REGINALD BULLER. 


TV. RemarKs UPON THE EGGs AND SPERMATOZOA OF 'THE 
ECHINOIDEA. 


The eggs of the Echinoidea (as is also the case with all 
the Hchinodermata) are surrounded by a thick, very trans- 
parent, gelatinous coat, the zona pellucida, through which 
the spermatozoa have to make their way before they reach 
the living protoplasm of the egg. 

The following measurements from Echinus will give some 


idea of the relative sizes of the living eggs, the gelatinous 


Fic. 1.—Eeg of Echinus microtubereculatus. x 170. a. 
Outline of protoplasm. 4. Outline of zona pellucida after five 
minutes in sea-water. ce. Outline of zona pellucida after several 
hours in sea-water. s. Spermatozoon. 


coat, and the spermatozoa :—Diameter of living egg alone 
O-ll mm.; diameter of living egg and gelatinous coat 
0:18 mm.; thickness of gelatinous coat 0°036 mm.; length 
of a spermatozoon 0°051 mm. 

Kach of the measurements just given is the average of ten 
measurements. The eggs were measured almost directly 
after being placed in water. 

The jelly increases in thickness after deposition of the 
egg in sea-water. After twenty-four hours it is found to 
have nearly doubled in thickness, and to have become 
0°057 mm. wide. The width of the jelly, which in the fresh 


FERTILISATION OF THE EGGS OF ANIMALS. 5 


ege is less, is, then, in the eggs which have stood twenty- 
four hours in water, slightly more than the length of the 
spermatozoa (Fig. 1). 

The presence of the gelatinous coat 1s quite unessential to 
the union of spermatozoa and eggs, for if by shaking it be 
removed, fertilisation will still take place with the greatest 
ease. 
~ When liberated at the top, the eggs gradually sink to the 
bottom of a beaker containing still sea-water. ‘They thus 
appear to be heavier than their normal medium. Very small 
currents are, however, sufficient to keep the eggs floating. 
Probably in the sea, where the eggs are liberated, such 
currents are always present. In that case, too, the currents 
are of considerable importance in mixing the eggs and 
spermatozoa. 


V. Tae Cremoractic Question. 


After repeated trials with unilateral illumination, I was 
unable to detect the least sensitiveness of the spermatozoa of 
Arbacia or Echinus for heliotactic stimuli. It was there- 
fore not possible by application of such a stimulus to allow 
the spermatozoa to stream in the direction of the eggs (which 
may be done in the case of Fucus), and to observe whether, 
when passing, they deviate toward them. 

A large number of artificial fertilisation experiments were 
undertaken. ‘l’o one side of a drop, either open or under a 
raised coverglass, a small drop bearing spermatozoa was 
added. The spermatozoa spread quickly in all directions, 
and in the course of their wanderings came in contact with 
the eggs, bringing about fertilisation. Within a few minutes 
this was made evident by the raising of the vitelline mem- 
brane. I failed, however, to observe any attraction of the 
spermatozoa toward the eggs from a distance, or any collec- 
tion of the spermatozoa around the eggs outside the gela- 
tinous coat. On the other hand, spermatozoa were frequently 


154 A. H. REGINALD BULLER. 


seen to pass by an egg so as almost to touch it, apparently 
without being in any way influenced by its presence. 
Nothing was seen which in any way reminded me of the 
chemotactic phenomena either of bacteria or of the sperma- 
tozoa of ferns. 

It is undoubtedly true that the spermatozoa collect rapidly 
in the gelatinous coat of an egg. This is, however, due to 
the fact that the spermatozoa which strike the outer surface 
immediately bore into the interior. It will subsequently be 
shown more fully that the phenomenon takes place equally 
well when the jelly encloses (1) a ripe egg; (2) an egg not 
having undergone maturation; and (3) an egg which has 
been killed with osmic acid, and then washed. There is thus 
not the shghtest necessity to account for the collection of the 
spermatozoa in the gelatinous coat by any chemotactic sub- 
stance which diffuses through the jelly into the sea-water, 
and so attracts spermatozoa towards the ege. 

If a substance causing attraction is really excreted by the 
eges one should be able to collect it. On this assumption 
the following experiments were made. 

A freshly obtained female Arbacia was cut open. The 
egos, which were then extruded by the animal in dense 
masses from the oviducts, were collected in about 100 c.c. sea- 
water contained in a crystallising dish, As soon as the eggs 
had settled to the bottom, for the purpose of washing them, 
the water was nearly all removed by means of a pipette. 
Another 100 ¢.c. was then added, and so much again removed 
after the eggs had settled that the latter, very thickly placed 
together, formed a layer in about 1—5 mm. of the sea-water. 
Snfficient oxygen could thus be obtained for respiration. 
The eges were left in the water from two to twelve hours, 
usnally about six. At the end of this period the water was 
filtered, and the ege@s thus removed. Capillary glass tubes, 
about 12 mm, long and 0:1—0°3 mm. internal diameter, and 
closed at one end, were then half filled with the water by 
means of an air-pump. The tubes were then introduced into a 
large open drop of sea-water, in which fresh, highly motile 


FERTILISATION OF THE EGGS OF ANIMALS. 155 


spermatozoa were swimming. If the eggs excrete an 
attracting substance it was argued that it should be present 
in the tubes, and the spermatozoa should collect there. 

In order to make certain that the eggs had remained 
normal during their stay in the sea-water, just before filtra- 
tion some of the eggs were tested by artificial fertilisation. 
In all the experiments upon which reliance has been placed 
for results this took place in the normal manner, 1. e. the 
vitellne membrane became raised. The first stages of 
segmentation were also often watched, and found in the 
majority of cases to follow the usual plan. The eggs were 
also tested soon after they had been placed in water. If, as 
rarely happened, they did not become fertilised readily they 
were rejected. ‘he experiments were repeated with four 
sets (g¢ and 9?) of Arbacia, three of Spherechinus, and 
two of Hchinus. 

No attraction of the spermatozoa into a tube could be 
observed. Except for a surface-contact phenomenon to be 
further discussed, they went in and out with indifference. 
Apparently, therefore, the water which had contained the 
eggs exercised no directive stimulus on the spermatozoa 
whatever. 

I then attempted to find some substance which would give 
a chemotactic stimulus to the spermatozoa. The substances 
tested were such as are known to give a directive chemical 
stimulus to many protozoa, the spermatozoa of ferns, pollen- 
tubes, ete. The following solutions were tried by the capillary 
tube method: distilled water; meat extract 1 per cent. ; 
potassium nitrate 10 per cent., 2 per cent.; sodium chloride 
5:8 per cent., 2°9 per cent., 0°58 per cent.; potassium malate 
1 per cent., O'l per cent.; asparagin | per cent.; glycerine 
5 c.c. per cent.; grape sugar 18 per cent., 9 per cent., 4°5 
per cent., 2°25 per cent. ; peptone | per cent.; alcohol 50 per 
cent., 25 per cent., 10 per cent.; diastase (from Merk) 1 per 
cent. ; oxalicacid 0°9 per cent., 0:09 per cent., 0:009 per cent. ; 
nitric acid (concentrated) 1 per cent., 0-1 per cent., 0°01 per 
cent. 


156 A. H. REGINALD BULLER. 


No definite chemotactic reaction—neither attraction nor 
repulsion—was observed in any case. Into tubes containing 
the weaker solutions the spermatozoa went in and out with 
apparent indifference. On coming into contact with highly 
concentrated neutral substances (potassium nitrate 10 per 
cent., sodium chloride 5°8 per cent., grape sugar 18 per cent.) 
the spermatozoa came to rest from loss of water. They are 
evidently not able to avoid solutions by a negative tonotactic 
reaction. On coming into contact with strong acid solutions 
(oxalic acid 0°9 per cent., 0°09 per cent., nitric acid 1 per 
cent., 0°1 per cent.) the spermatozoa were killed, and thus 
formed slight collections. They were thus not able to avoid 
acids by means of a negative chemotactic reaction. 

Having obtained, so far as chemotaxis is concerned, only 
negative results by means of the capillary tube method, 
another method was employed, which Massart ! found effective 
in determining the tonotactic sensibility of a number of 
marine micro-organisms. 

Two large and equal drops were made in a moist chamber ; 
one was of sea-water containing spermatozoa, the other 
distilled water. The drops were then joined by a narrow 
bridge, so that diffusion between them could take place. 
The experiment was watched between one and two hours. 
Some spermatozoa gradually entered the fresh water. No 
collection, however, took place in either drop. There were 
thus no signs of attraction or repulsion. The spermatozoa 
that entered the too diluted water were killed. 

In another experiment a small drop of sea-water was dried 
upon a glass slide. A large oval drop of sea-water contain- 
ing motile spermatozoa was then placed near and spread so 
that one end just covered the crystals from the dried drop. 
The crystals began to dissolve rapidly, locally concentrating 
the sea-water. As diffusion from the concentrated end of 
the drop took place the spermatozoa in the neighbourhood 

1 Massart, ‘* Recherches sur les Organismes Inférieurs. IT. La Sensibilité 


a la Concentration chez les Etres Unicellulaires Marins,” ‘ Bull. del Acad. 
roy. de Belgique,’ 3me sér., tome xxii, No. 8, 1891, p. 148. 


FERTILISATION OF THE EGGS OF ANIMALS. 157 


were killed or brought to rest. They were thus not repelled 
by concentrated sea-water. 

When to one end of a large oval drop containing sperma- 
tozoa some crystals of sodium chloride or potassium nitrate 
were added, similar results were obtained. ‘The sperma- 
tozoa allowed themselves to be surprised by the advancing 
salt and were accordingly killed. 

The drop experiments were, then, not more successful than 
those with capillary tubes. No evidence of chemotactic 
reaction could be obtained by either method. 

Massart! found by experiment that two species of 
Spirillum A and C, the flagellate Heteromita rostrata 
and two species of Ciliata fled from solutions both more or 
less concentrated than sea-water, i.e. they always sought a 
zone with the concentration equal to their normal medium. 
Oxytricha gibba also fled from solutions more highly 
coucentrated than sea-water, but failed to avoid those less 
concentrated. In the case of his Spirillum B he obtained 
an organism which did not flee from solutions either more or 
less concentrated than sea-water, and which suffered in the 
experiments accordingly. Both the drop and capillary tube 
experiments described above appear to indicate that the sper- 
matozoa of the Echinoidea, like Massart’s Spirillum B, are 
quite insensible to tonotactic stimuli. 

Finally, I attempted to determine whether the spermatozoa 
are attracted or repelled by oxygen. Fresh spermatozoa 
were removed from a testis and placed fairly thickly together 
in a drop, so that the latter was very slightly milky. A cover- 
glass 0°15 mm. thick and 18 mm. square, supported on pieces 
of another cover-glass also 0°15 mm. thick, was then placed 
upon the drop in such fashion as to include a small bubble 
of air near the middle. Under these circumstances one sees 
neither attraction nor repulsion from the bubble, even when 
the experiment has continued some time and when the 
oxygen supply must be getting low. 

When, however, the drop is made very milky by spreading a 

' Massart, loc. cit., pp. 151—154. 


158 A. H. REGINALD BULLER. 


little of the thick white sperm-fluid in it by means of a pipette, 
a peculiar effect may be observed as a result of the presence 
of an air-bubble. The spermatozoa, in incredible numbers 
and constantly colliding, first swarm equally well all over 
the preparation. After about five minutes one sees macro- 
scopically, when looking at the slide upon the microscope 
stage, a black zone arise about 1 mm. from the edge of the 
air-bubble. On examination with the microscope one sees 
that there are fewer spermatozoa there than anywhere else. 
Three zones (Fig. 2) may then be made out around the air- 
bubble 2: a,an inner zone crowded with actively motile sper- 
matozoa; b, a much thinner zone (that appearing macro- 


Figs. 2 and 3. 


scopically black) in which there are comparatively very few 
spermatozoa; and c, the zone outside b (which extends over 
the rest of the preparation nearly to the edge of the cover- 
glass) where the spermatozoa are crowded, so far as I could 
judge, about as thickly as in zone a, but have all come to rest 
from want of oxygen. As one watches the preparation one 
sees (fig. 3) that the spermatozoa gradually leave the zone a 
and collect on the inner edge of the zone c, upon reaching 
which they cease to move. <A ring of thickly placed, dead 
spermatozoa thus arises. A similar collection of dead sper- 
matozoa also takes place about 1 mm. from the edge of the 
cover-glass. The explanation of this curious phenomenon, 


FERTILISATION OF THE EGGS OF ANIMALS. 159 


which was observed many times for both Arbacia and 
Echinus, may be as follows:—The spermatozoa in the 
zone c after using up all the oxygen at once come to rest. 
The zones a and b are by diffusion supplied with oxygen 
from the bubble of air. The spermatozoa in these zones are 
thus able to continue in movement. This they do in any 
direction until reaching the inner edge of the zone ¢; when 
the oxygen has all been used up they can move no longer, 
and come to rest there, forming a ring. It is not, however, 
clear to me why the zone b should not be as crowded 
as the zone a. In any case there is no gathering of 
motile spermatozoa in any zone around the air-bubble. The 
phenomenon here described has then a totally different 
appearance to that figured by Massart! for the oxytaxis of 
Spirillum and of Anophrys, and is also quite unlike the 
collection of Spirillum undula around air-bubbles, which 
I myself have had frequent oceasion to observe. I therefore 
fail to see in the phenomenon any evidence that the sper- 
matozoa are attracted by oxygen. 

From the above section of my paper it will be noticed that 
I have been unable to obtain any evidence, either direct or 
indirect, that the spermatozoa are attracted chemotactically 
to the eggs, and that, further, no success has attended my 
efforts to find any substance to which the spermatozoa are 
chemotactically sensible. 


VI. THe Movements or SrerMAtrozoA UPON SuRFACES. 


When swimming in a drop of sea-water and not in contact 
with its surface the spermatozoa of the Echinoidea swim 
spirally. The spirals may be so steep that the spermatozoa 
appear to swim in almost a straight line, and they then move 
relatively rapidly across the field of the microscope. On the 
other hand, the incline of the spiral may be so gentle that 
the spermatozoa appear to be swimming almost in circles. In 


' Massart, loc. cit., p. 157. 


160 A. H. REGINALD BULLER. 


this case the progress across the microscopic field is very 
slow indeed. Between these extremes there 1s every grada- 
tion. An approximation to the first case appears, however, 
to be the rule with the most active spermatozoa. 

When a spermatozoon comes in contact with a glass surface 
it is influenced by it in a curious manner. The spermatozoon 
remains in contact with the surface, and, unless it becomes 
immediately fixed to the glass, begins to make characteristic 
circular revolutions upon it. If the cover-glass be supported 
by pieces of another cover-glass, and the upper surface of the 
drop in contact with it be carefully focussed, it is seen that 
all the spermatozoa which are not attached by their heads 
but are moving there, are revolving, from the observer’s 
point of view, in a clockwise direction. If the lower surface 
of the drop in contact with the slide be examined a reverse 
rotation—the counter-clockwise—is seen to be the rule. In 
both cases, therefore, if the surfaces be regarded from the 
point of view of the spermatozoa the rotation is always in 
one direction—namely, the counter-clockwise. 

The clinging to surfaces and rotation upon them by the 
spermatozoa in the manner explained is not limited to glass. 
It takes place quite as well upon the surface of a drop 
bounded by air, and it is easily seen upon the outer surface 
of the gelatinous layer of eges of Echinus which have 
just been placed in sea-water. I have also sometimes 
watched it (in the case of Echinus) upon the protoplasm 
of the eggs and upon the vitellme membrane, where it was 
made possible by the looseness of the zona pellucida. ‘he 
nature of the surface for the phenomenon does not, there- 
fore, appear to be of essential importance. 

The following rule was found to hold good: whenever the 
spermatozoa of the Echinoidea come in contact with a sur- 
face they either become fixed to it at once or, more often, 
they rotate upon it, and in the latter case, looking from them 
to the surface in question, in a counter-clockwise direction. 

The phenomenon is most easily seen in open drops con- 
taining not too many spermatozoa. The drops I employed 


FERTILISATION OF THE EGGS OF ANIMALS. 161 


were about 10 mm. diameter and 1—2 mm. high. The 
upper and under surfaces (in which the direction of rotation 
appears reversed) are best examined with a magnification of 
from 70—240 diameters. In the case of Echinus the rota- 
tion is well seen under these conditions for ten minutes after 
the beginning of the experiment, fairly well for fifteen 
minutes, and ceases to be seen in about twenty minutes, when 
the spermatozoa have nearly all come to rest. 

The rule given above as regards rotation upon surfaces 
was found to hold good not only for the Hchinoidea, but 
for every group of the Echinodermata. Open drops were 
employed in the test. ‘he names of the species examined 
have already been given. 

It has already been mentioned that Dewitz! discovered that 
the spermatozoa of certain insects revolve upon surfaces. 
This has been confirmed by Ballowitz.2 Dewitz also found 
that the direction of revolution was counter-clockwise. It 
is a remarkable fact that the spermatozoa of groups so far 
apart as the Insecta and Kchinodermata should thus be 
affected by surfaces in the same manner. It is not improb- 
able that a similar phenomenon will be found for the 
spermatozoa of yet other animals. 

Immediately after my work had been brought to a close a 
preliminary paper was published by Dungern,’ entitled “ Die 
Ursachen der Specietiit bei der Befruchtung.’ In it the 
author communicated his discovery of the counter-clockwise 
rotation upon surfaces by the spermatozoa of Spherechinus 
and Arbacia. He failed, however, to observe any rotation 
of the spermatozoa of Echinus, but this, he states, took 
place after the addition of certain (unnamed) “ stimulating 
substances.’ Since it was in ordinary sea-water and with 
the spermatozoa of Echinus that I (independently) first 
became aware of the rotation, and since with this species 


1 Dewitz, loc. cit. 

* Ballowitz, ‘‘ Untersuchungen iiber die Structur der Spermatozoen, etc.,” 
Zeitschr. f. Zoologie,’ Bd. 1, 1890, pp. 392, 393. 

% Dungern, ‘ Centralbl. fiir Physiologie,’ April, 1901, Heft 1. 


von. 46, PART 1,.—NrW SERIES. 17 


162 A. H. REGINALD BULLER. 


T have made scores of observations upon the phenomenon in 
question, I am at a loss to explain his negative results. The 
detailed paper which he has promised will doubtless clear the 
matter up. At present, however, I am unable to accept his 
account of the ‘ stimulating substances.” 

From the published figures and from my own observations 
the spermatozoa of the Echinoidea appear to be radial struc- 
tures. The fact, however, that they rotate as a rule only in 
one direction shows that they cannot really be thus con- 
structed. The researches of Ballowitz! upon bird and insect 
spermatozoa have demonstrated the complexity of their form, 
which is often in part spiral. Perhaps a minute investigation 
would also lead to similar results in the case of the Kchinoidea. 

Careful observation of the rotating spermatozoa reveals the 
fact that a fewspermatozoa, probably not more than 1 percent., 
revolve in a direction contrary to that of the great majority. 

All that can be seen of a spermatozoon during its rotation 
upon the under side of a cover-glass is its head. This moves 
rapidly round in a circle. The tail is quite invisible. The 
centre of the circle remains fairly constant in position. 
Sometimes, however, the spermatozoon makes a wider curve 
for a while, and then begins to make circles with the same 
diameter as before around a new centre. The width of the 
circles varies. It appears, however, in the case of Echinus 
to be slightly less than 0°05 mm., the length of a sperma- 
tozoon. The rate of rotation of the actively moving 
spermatozoa in any preparation is fairly constant, as may 
easily be observed by direct comparison of rotating indi- 
viduals. In the case of Spherechinus a normally rotating 
spermatozoon was observed to make 109 circles around one 
point in 90 seconds, which gives a rate of slightly more 
than one revolution a second. ‘The diameter of the circle is 
about the same as in the case of Hchinus. Assuming it to be 
0-04 mm., the rate of movement of the head is calculated to 
be approximately 0:12 mm. per second, or 7 mm. per minute. 

If a supported cover-glass preparation containing a drop 

1 Ballowitz, loc. cit., p. 317. 


FERTILISATION OF THE EGGS OF ANIMALS. 163 


with freshly obtained spermatozoa be made as quickly as 
possible and examined at once, one notices that a large 
number of spermatozoa are already rotating in the manner 
just described upon the upper and _ lower — surfaces. 
One also notices, however, that a considerable number 
of spermatozoa have become attached by the ends of 
their conical heads. The heads then generally continue to 
move about their tips, usually executing circles apparently 
either clockwise or counter-clockwise. 

If a single spermatozoon not fixed by the head and rotating 
in circles upon glass be watched, it will often be seen to make a 
number of consecutive revolutions and then suddenly stop in its 
course and become fixed to the glass by the point of its head. 
In this way the number of spermatozoa attached to the glass 
surfaces gradually increases. 

Although the spermatozoa of the Hchinoidea appear to 
become most easily fixed to surfaces by the tips of their 
conical heads, yet fixation may take place in several other 
ways. Thus, I have seen spermatozoa attached (1) by the 
middle of the side of the head, (2) by the middle-piece, 
proof of such fixation being rotation of the head around 
these respective parts. The spermatozoa can also become 
attached by the hinder half of the tail, for I have observed 
cases in which the head and fore part of the tail have made 
excursions from the glass, returning, however, to their 
original position, while the hind part of the tail remained in 
one place, immovable upon the glass surface. In yet other 
cases One sees a spermatozoon apparently attached to the 
glass by its whole length, slight waves of movement pro- 
ceeding down the tail from just below the head. When 
spermatozoa have entirely come to rest upon a surface they 
are very frequently seen to be attached to the glass by their 
whole length, a fact which Ballowitz! has also observed in the 
case of the spermatozoa of Insects. From these various 
observations it appears that a spermatozoon may become 
attached to a surface in almost every possible manner. 

! Ballowitz, loc. cit., p. 393, footnote. 


164 A. H. REGINALD BULLER. 


One may not infrequently see a spermatozoon attached 
lengthwise to the under surface of a cover-glass move in a 
circle by means of a series of jerks, between every two of 
which it comes to rest. In these cases one often clearly sees 
the whole spermatozoon and not merely the head. One may 
then observe that the tail is curved and that circulation 
takes place in the direction of the curve. Fig. 4 represents 
what apparently happens, the spermatozoon being drawn in 
the resting stages. It is quite evident that rotation does not 
take place upon the tip of the tail, but that the whole tail at 
each jerk takes up a new position. 

Observations upon a few (possibly nearly exhausted or 
immature) spermatozoa lying just beneath, and apparently 
attached to the surface of the cover-glass, showed that, as 
in the case of insects, the head does not alter its shape and 
is not concerned in locomotion, and that waves of movement 
arise in the fore-part of the tail and proceed toward its end. 
The driving force thus appears to lie close beneath the 
head. 

With regard to the manner in which the rapid and con- 
tinuous revolutions upon surfaces take place, if isolated 
observations upon a number of slowly moving spermatozoa 
will allow one to draw conclusions, perhaps the explanation 
may be as follows:—When a spermatozoon revolving in a 
spiral comes into accidental contact with a glass surface by 
the tail, at least the hinder end of this is unable to leave the 
olass owing to its adhesiveness, but can more or less easily be 
dragged along it. The fore part of the tail, by means of its 
automatic movement, causes the head and itself to make 
constant excursions to and from the glass surface. The 
tail is probably slightly curved, and the direction of motion 
of the spermatozoon is thus constrained to be circular. 
The head must frequently come in contact with the glass 
by its tip, which is specially adhesive and lable to become 
fixed in one place. Ballowitz holds that the circles upon 
surfaces for insects are probably simply the modified spiral 
turns of the free-swimming spermatozoa. This view, which 


FERTILISATION OF THE EGGS OF ANIMALS. 165 


gives a purely mechanical explanation, is probably also correct 
for the spermatozoa of the Echinodermata. I therefore agree 
with Ballowitz' that Dewitz was in error in assuming that the 
movement of spermatozoa in circles upon surfaces is due to 
a special stimulus arising from the latter. 

The Flagellata and Ciliata on meeting with an obstacle 
receive a mechanical stimulus,” to which they respond by 
reversing their direction of movement for a while. They thus 
avoid obstacles. It is evident from what has gone before 
that the spermatozoa of the Hchinodermata do not react in 
this manner. 

One may now inquire what significance the relations of sper- 


matozoa to surfaces has for fertilisation. Special observations 
to obtain an answer were made upon the eggs of Hchinus. 
It was found that if spermatozoa be added to eggs which 
have just been placed in water, spermatozoa at once collect 
upon the outer surface of the gelatinous coat, and a number 
can be seen there making the characteristic circles. Others 
are seen to penetrate the jelly immediately on coming in con- 
tact with it. It is these spermatozoa which do not rotate 
upon the gelatinous coat which reach the living egg first. Of 
those spermatozoa which do rotate it was seen that they 

' Ballowitz, loc. cit., p. 393. 

> Jennings, “On the Movements and Motor Reflexes of the Flagellata and 
Ciliata,” ‘Amer. Journ. of Physiology,’ vol. iil, Jan., 1900, p. 229. 


166 A. H. REGINALD BULLER. 


often stop in their rotation, and that the head becomes fixed 
in the jelly. Sometimes the spermatozoon then succeeds in 
boring its way through, and may then reach the living 
protoplasm. In most cases, however, the head of a sper- 
matozoon which has rotated a number of times gets stuck 
in the outer layer of the jelly, and no successful penetration 
occurs. 

The gelatinous coat of an egg which has only been in 
water a few minutes is much more difficult for the spermatozoa 
to penetrate than that of an egg which has been in water 
several hours. Comparative experiments easily demonstrated 
this point, the difference being really striking. The jelly, as 
already mentioned, swells in water, and gradually nearly 
doubles its original breadth. At the same time it becomes 
softer. When spermatozoa are added after the swelling has 
taken place, scarcely a single spermatozoon is seen to rotate 
upon the eggs; on the contrary, they nearly all succeed 
in fixing their heads in the jelly, and the majority penetrate 
almost up to the living egg. 

Dewitz! believes that the rotation of the spermatozoa of 
Blatta upon surfaces is of prime importance in enabling the 
spermatozoa to find their way into the micropyles of the eggs. 
This may well be the fact. In the case of the Echinoidea, 
however, there are no micropyles, and the gelatinous coat is 
everywhere penetrable. Further, rotation upon the eggs 
appears not to be the rule. It seems to me, there- 
fore, that the fact that the spermatozoa will rotate upon 
resistant surfaces has no special biological significance in 
respect to fertilisation. On the other hand, the ability of the 
spermatozoa to cling to surfaces and to get stuck to them by 
the pointed end of their heads is of great importance in 
causing them not to leave the gelatinous coat of an egg after 
having come in contact with it, and in penetrating the same. 


! Dewitz, loc. cit. 


FERTILISATION OF THE EGGS OF ANIMALS. 167 


VII. Tse Drrecrion or PENETRATION OF THE GELATINOUS Coat, 


Fol! observed for Asterias that when a spermatozoon had 
come in contact with the gelatinous coat it placed itself 
perpendicular to it, and then penetrated radially to the egg, 
meeting and fusing when half-way through with a curious 
“cone Wexudation.”’ Fol concluded that the gelatinous 
coat is an apparatus for catching the spermatozoa when they 
come in contact with it, and attributed the radial structure to 
lines of more or less resistance, which serve to guide the 
spermatozoa directly to the egg. 

Selenka* investigated the development of the gelatinous 
coat, and found that at first it is penetrated by fine radial 
protoplasmic filaments, each in its own canal. Later the fila- 
ments become withdrawn, but the canals remain until after 
fertilisation. Selenka also observed that the spermatozoa 
penetrate the jelly “always ina radial direction,’ and stated 
that this is due to the spermatozoa making their way through 
the canals. This explanation appears to be very plausible. 
Before, however, accepting it as being sufficient, we shall do 
well to bear in mind the observations of Kupffer and Benecke 
upon the fertilisation of the eggs of Petromyzon. 

According to the last-named authors the spermatozoa of 
Petromyzon penetrate a thick gelatinous dome covering 
one end of an egg in a radial direction. This observation, 
upon which special stress was laid, confirmed the statements 
previously made by August Miller. Although the inner 
shell-layer was found to contain radial canals, Kupffer and 
Benecke could discover no trace of such in the outer shell- 
layer, while they described and figured the dome as being 

1 Fol, ‘ Recherches sur la Fécondation, ete.,’ 1879. 

* Selenka, ‘ Zoologische Studien, Befruchtung des Kies von Toxopneustes 
variegatus,’ 1878, p. 2. 

= ioe: cit., p. 5. 

* Kupffer and Benecke, ‘Der Vorgang der Befruchtung am Wi der Neu- 
naugen,’ Kouigsberg, 1878, p. 11. 


168 A. H. REGINALD BULLER. 


quite hyaline. The radial path of the spermatozoon is so 
striking that the authors believed it necessary to postulate 
some attraction! of the egg for the spermatozoon from a 
distance. Concerning the nature of the forces, however, no 
suggestion was made. 

In my own investigations special attention was paid to the 
direction of penetration of the spermatozoa through the gela- 
tinous coat of the eggs of Echinus. In this case, at least, 
it cannot be stated that penetration is always in a radial 
direction. A great many spermatozoa penetrate obliquely. 
It appeared to me, however, after having made a large 
number of observations for determining the point, that on 
the whole there is a tendency for the spermatozoa to make 
their way from the outside to the inside of the gelatinous 
coat. his tendency is best seen after the eggs have been 
from three to six hours in sea-water and the jelly has become 
considerably swollen. One then observes, upon adding 
spermatozoa, that on the whole, although many penetrate 
obliquely, the spermatozoa pass in a radial manner through 
the jelly to the egg. It is easy to observe spermatozoa which 
take an almost perfectly radial course. The path of many of 
them is seen to incline to a radius by an angle equal to 
between 10° and 30°. Others may be observed to start 
fairly radially, soon turn aside, and continue obliquely 
striking the eggs thus obliquely, or occasionally even making 
their way out again in a tangential direction. A considerable 
number of spermatozoa, after entering, stick fast in the jelly. 
The heads of these are then seen to be very variously oriented 
with respect to a radius. 

Having come to the conclusion that the spermatozoa do 
pass more or less radially through the gelatinous coat, my 
next inquiry was concerning the cause. It was found that 
the radial penetration could be equally well observed in (1) 
a ripe egg; (2) a full-sized egg which had not undergone 
maturation, the nucleus being still very large and uncon- 
tracted ; and in (5) a ripe egg which had been killed with 


1 Kupffer and Benecke, loc. cit., figs. 1, 7, and 8. 


FERTILISATION OF THE EGGS OF ANIMALS. 169 


osmic acid and then washed. In the last case the osmic acid 
turned the eggs brown. ‘The eggs so killed were put in 
100 c.c. sea-water for half an hour and stirred round at 
intervals. They were then caught in a pipette, placed in a 
drop on a slide, and spermatozoa added. he radial penetra- 
tion was quite as clear as in the living eggs. 

From the foregoing observations it seems evident that the 
radial penetration is not brought about by any special attrac- 
tion by the living egg, for it takes place equally well with a 
dead egg. Nor do the facts point to any chemotactic attract- 
ing substance as causing the phenomenon, for from a dead 
egg no excretion can take place. Selenka’s suggestion that 
the spermatozoa take a radial course because they make their 
way through canals, which during the development of the 
egg contained protoplasmic connections, also does not seem 
to me a satisfactory explanation. ‘I'he radial structure of the 
gelatinous coat after an egg has been a few hours in water 
is extremely faint, and, so far as one can directly observe, 
absent at the periphery where the spermatozoa start on their 
course. Selenka' admitted that the canals were finer than 
the width of the head of a spermatozoon. Surely with the 
swelling of the jelly these canals must be practically filled up. 
I have, as already stated, very frequently seen spermatozoa 
penetrate the gelatinous coat obliquely, often very obliquely. 
In these cases the spermatozoa could not be making their 
way through Selenka’s canals. Hence we may conclude that 
the canals, if such there are, are not necessary for penetra- 
tion. The thick gelatinous dome of a Petromyzon eg, 
and also, according to Massart,? the jelly around the ovum of 
the frog, are penetrated radially without the presence of any 
canals whatever. These various facts point to the conclusion 
that the penetration of the gelatinous coat in a more or less 
radial direction by the spermatozoa is not due to canals, but 
to some other cause. 

The above reflections led me to make experiments to find 


1 Selenka, loc. cit., p. 5. 
* Massart, “Sur la Pénétration, ete.,” loc. cit., p. 217. 


170 A. H. REGINALD BULLER. 


out how the spermatozoa behave toward jelly from other 
sources than that from the eggs of the Echinoidea. 

When the oosporangia of Cystocyra barbata (one of 
the Fucacez) are liberated into sea-water, the outer coat 
rapidly swells and gelatinises. Spermatozoa from Arbacia 
were added to a preparation containing some of the oospor- 
angia, Ata certain stage in the gelatinisation the sperma- 
tozoa entered the jelly in large numbers, thus becoming 
densely crowded together in it. 

A similar gathering was observed when the seed-coat of 
Linum usatissimum was placed in water contaiming 
spermatozoa. The outer cell-walls rapidly swell and become 
gelatinous. The spermatozoa, when the jelly had reached a 
certain consistency, collected in it in large numbers. 

It was also found that if the gelatinous coat of an 
Hchinus egg be separated by shaking, and spermatozoa be 
allowed access to the coat after several hours’ isolation, the 
number of spermatozoa which will gather in it is very 
considerable. 

The conclusion that is to be drawn from the above experi- 
ments appears to be that the spermatozoa are so constructed 
that they will bore their way into any jelly of a certain 
consistency without any aid from canals, chemotactic sub- 
stances, or influences from living protoplasm. 

Massart, as already mentioned, explains the radial pene- 
tration in the frog by supposing that the spermatozoa seek 
to pass from the more watery to the less watery layers of 
jelly, owing to a sensibility to these differences in saturation. 
Although this theory is plausible, it does not appear to me 
to be convincing. It does not sufficiently explain why the 
head of a spermatozoon is at first pushed into the jelly ina 
radial direction. After the head has been pushed in, whether 
this be radially or somewhat obliquely, the spermatozoon of 
Wchinus usually takes a fairly straight course with respect 
to the axis of the head. Evidence of picking and choosing 
between the gelatinous layers thus appears to be wanting. 
After an egg of Echinus has been in water for several hours 


FERTILISATION OF THE EGGS OF ANIMALS. RAL 


it is doubtful whether the outer layers of jelly are the more 
watery and the inner the less so. In fact, from the ease with 
which the spermatozoa rotate the egg inside the gelatinous 
coat (vide infra), one might well suppose that the innermost 
layers are the more watery. ‘The view that the resistance of 
the jelly decreases inwards has, indeed, already been upheld 
by Selenka! for the eggs of Asterias. For the Echinoidea 
and Asteroidea, therefore, the necessary basis of fact for an 
application of Massart’s theory seems to be wanting. 

There appear to me to be yet two possible explanations of 
the penetration: (1) It is due to reaction to a stereotactic 
stimulus ; (2) it is purely mechanical. 

1. Stereotropism has long been observed. Very many 
organisms, both animals and plants, in sea- and fresh-water, 
grow perpendicularly to their substratum, owing to the 
influence which the position of the latter has upon their 
direction of growth. In the same manner as for geotropism, 
heliotropism, chemotropism, etc., we have a corresponding 
tactic phenomenon, so also may it be with stereotropism. It 
is possible to imagine a free-swimming organism which, upon 
coming in contact with a surface, receives from it a stimulus 
which causes it to alter its movements in such a manner as to 
attempt to make its way more or less perpendicularly to the 
same, and through the substance concerned. Although such 
a stereotactic sensitiveness would neatly explain the radial 
penetration for the Echinoidea, Petromyzon, and the frog, 
yet conclusive observations in its support appear to me to be 
lacking. 

2. Owing to the extreme difficulty or impossibility of seeing 
exactly what the movements of a spermatozoon upon a gela- 
tinous surface are, the mechanical explanation must at present 
remain tentative and almost purely hypothetical. When a 
spermatozoon, swimming spirally, comes in contact with the 
outer surface of the gelatinous coat, the tip of the conical 
head, which reaches it first, possibly owing to the force of 
contact, possibly to adhesiveness, may well be supposed to 


' Selenka, loc. cit., IL ‘ Die Befruchtung, Das Spermatozoon.’ 


172 A. H. REGINALD BULLER. 


immediately fix itself in the jelly. This is, indeed, what 
appears to take place under the microscope. The tail of the 
spermatozoon then probably adheres to the outer surface of 
the egg-coat, and is dragged round and round on it about 
the conical head, which is gradually pushed forward through 
the jelly. It may well be these revolutions (the modified 
spiral of the usual mode of swimming) which cause a 
spermatozoon to bore through the jelly more or less perpen- 
dicularly to the surface. The fact that the head is of such a 
shape that when once embedded in the gelatinous coat it can 
be easily pushed forward, but offers considerable resistance 
to moving either backwards or sideways, together with the 
particular consistency of the jelly, may well account for the 
steady progress forward of the whole spermatozoon in one 
direction. It seems to me probable that some such explana- 
tion as the foregoing will be sufficient to explain all that 
takes place during the penetration of the gelatinous coat. 

It may here be remarked that since the presence of a 
gelatinous coat doubles the diameter of an egg its presence 
multiplies the chances of contact with its exterior surface 
by a spermatozoon four times. Since the more active sper- 
matozoa, after coming in contact with the jelly, are con- 
ducted by it to the living protoplasm of the egg, the chances 
of fertilisation by them is, by the presence of the gelatinous 
coat, also increased four times. Since, however, a consider- 
able percentage of the weaker spermatozoa get stuck in the 
gelatinous coat after entrance, thus not reaching the living 
ego, our estimate of the increased chances of fertilisation 
must undergo a large reduction. The first function of the 
jelly, which surrounds so many eggs, appears to be that of 
protection, making them distasteful to larger, and unassail- 
able by smaller, enemies. For the purpose of fertilisation its 
consistency must be such as to allow easy penetration by the 
spermatozoa. 

The ease with which spermatozoa enter and become fixed 
in gelatinous substances will explain a phenomenon which at 
first puzzled me. It was observed that, when a capillary 


FERTILISATION OF THE EGGS OF ANIMALS. 175 


tube containing sea water, in which eggs had previously 
been deposited, was placed in a drop contaiming sper- 
matozoa, the spermatozoa were not attracted into the tube. 
On the other hand, it frequently happened that the sper- 
matozoa gathered very thickly into small balls just inside 
and outside of a tube. The balls were sometimes 0°01 to 
0:05 mm. in diameter. It was apparent that, since the balls 
were only formed at the mouth of a tube, the cause of their 
formation was to be sought in the filtered sea-water. The 
phenomenon was found to take place after six successive filtra- 
tions. A drop of sea-water in which eggs had been deposited 
was placed upon a slide and a drop containing spermatozoa 
near it. On joining the drops a large number of small balls 
were formed in a very few seconds. When very numerous 
spermatozoa were present the balls became 0:1 mm. in 
diameter, containing many thousands of spermatozoa packed 
together in a dense mass. The following appears to be the 
explanation of the phenomenon:—From the ovary there 
come out with the eggs a large number of very small bits of 
jelly, which are so small that they will (like spermatozoa) 
pass through ordinary filter paper, and so transparent that 
one cannot directly see them. A few spermatozoa become 
attached to each piece of jelly, the presence of which may be 
inferred from the manner in which the small group of sper- 
matozoa move about. Owing to the length of a spermatozoon, 
although its head, may be embedded in a jelly particle, the 
tail may remain partly free. The little collections of sper- 
matozoa thus move about hither and thither in no particular 
direction. When two such groups come by accident into 
contact they fuse. Certain of the spermatozoa adhere to 
both little masses of jelly and lock them together. The 
fused mass combines with other simple and fused masses, and 
soon. It is by this curious synthetic process that, in a very 
few seconds, there may be formed a ball as large or larger 
than an Echinus egg and containing thousands and 
thousands of spermatozoa, looking black under the micro- 
scope, and easily seen in a drop of water with the unaided eye. 


174 A. H. REGINALD BULLER. 


VII. Tae ATTACHMENT OF SPERMATOZOA TO THE Kaa. 


As soon as a spermatozoon has penetrated the gelatinous 
coat it usually becomes fixed by the head to the periphery of 
the living egg. Sometimes it executes circles for a while 
upon the protoplasm, and occasionally even re-enters the 
jelly and makes its way through this in a radial direction, 
thus leaving the egg entirely. 

When a great number of spermatozoa are allowed access 
to an egg which has been some hours in sea-water, so many 
immediately penetrate and become attached by their heads 
that they set the egg in rotation. The rotation may be in 
any direction,! and often continues for about a minute, 
ceasing with the formation of the vitellne membrane. The 
rate of revolution varies according to the number of sper- 
matozoa attached to the egg. A rapidly moving egg of 
Arbacia was observed to make ten revolutions in thirty 
seconds. ‘The gelatinous coat during rotation scarcely moves 
at all, the living egg revolving quite independently within it. | 
The spermatozoa often move the egg with such force as to’ 
separate it from its gelatinous coat. One then observes that, 
except for those attached by their heads to the egg, there is 
no collection of spermatozoa around the latter. This fact is 
in accordance with the supposition that no chemotactic sub- 
stance is excreted by the egg. Numerous spermatozoa enter 
the isolated gelatinous coat. 

The spermatozoon attaches itself to the egg by its most 
adhesive part, i. e. the tip of the head. The question arises 
whether the attachment is purely mechanical. It may be 
that the outer surface of the protoplasm is such as to be 
best adapted for retaining a spermatozoon by adhesion as 

' For the eggs of the Fucaces the rotation appears to be constantly in a 
clockwise direction. Thus Farmer and Williams (‘ Phil. Trans. Roy. Soe.,’ 
vol. 190, 1898, p. 633) state for Halidrys “the movement is always in a 


clockwise direction.” I have also found this true for Cystocyra barbata. 
The fact as yet has not been explained. 


FERTILISATION OF THE EGGS OF ANIMALS. 175 


soon as this comes in contact with it by the tip of its head. 
On the other hand, it is possible, and even probable, that un- 
known stimuli here play a part. The advance of the sper- 
matozoon into the egg after leaving the periphery is, like the 
formation of the vitelline membrane, doubtless due to a 
stimulus given the egg by the spermatozoon. With regard 
to the exact nature of the stimulus and of the protoplasmic 
movements which appear to be its reaction we are as yet 
without any explanation. 


TX. Summary oF THE Cuter Resottrs. 


The chief conclusions arrived at during the research upon 
the fertilisation of the eggs of the HEchinoidea were as 
follows : 

1. The meeting of the spermatozoa with the outer surface 
of the gelatinous coat (zona pellucida) is a matter of chance, 
and not due to chemotaxis. 

2. The passage of the spermatozoa through the gelatinous 
coat (observed chiefly in Hchinus) is more or less ina radial 
direction as regards the egg. The direction taken is not due 
to any chemotactic substance being excreted from the egg. 
The phenomenon is possibly due to stereotaxis, but a purely 
mechanical explanation seems to the author more probable. 

3. The spermatozoa are probably not chemotactically 
sensitive. They do not respond to tonotactic or heliotactic 
stimuli. 

4, On coming in contact with a surface bounding their 
medium the spermatozoa cling to it, and usually continue for a 
time to revolve upon it in (from their point of view) a counter- 
clockwise direction. ‘l'his statement applies to every group 
of the Echinodermata. 

5. The spermatozoa easily become attached to glass and 
other surfaces by the tips of their conical heads. This 
phenomenon doubtless plays a réle in causing the spermatozoa 
to bore through the gelatinous coat after having come in con- 


176 A. H. REGINALD BULLER. 


tact with its outer surface, and also in their becoming 
attached to the living egg. 

6. The vast number of eggs, and still vaster number of 
spermatozoa produced, together with the motility of the 
latter and the action of sea-currents, quite suffices to bring 
the male sexual cells into contact with the zona pellucida. 

7. Many writers have supposed that chemotaxis is a 
constant factor in the fertilisation of animal eggs. This 
generalisation, which has been made by arguing from the 
attraction of the spermatozoa to the eggs of certain plants, 
is as yet entirely without experimental justification. From 
my own results with the Echinoidea, which are in accordance 
with those obtained by Massart in the case of the frog, and 
with the work of Dewitz upon the fertilisation of the eggs of 
certain insects, I have been led to suppose that chemotaxis, 
at least for a great number of animal species, plays no 
role whatever in bringing the sexual elements together. 


The work for the above paper was done at the Stazione 
Zoologica, Naples, during the months of March and April of 
each of the years 1900 and 1901. It gives me much pleasure 
to thank the Committee of the British Association for grant- 
ing me the use of the table, and also to acknowledge my 
indebtedness to the staff of the Stazione Zoologica for 
supplying me with material and apparatus during the 
research. 


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. 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 Suprrement Numbers being found inconvenient, 
and there being often in the Editor’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 Editor. 

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

Messrs. Apiarp & Son, 224, Bartholomew Close, EH.C., on 
the following terms: 

For every four pages or less— 


25 copies : : ; : 5/- 
DOR 255 : ; : : 6/- 
1005-65; ie 


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 Mzssrs. J. & A. Cuurcuitt, 7 Great MaRLBorouGcs SrREET, 
Lonpon, W. 


| 
THE MARINE BIOLOGICAL ASSOCIATION 


OF THE 
UNITED KINGDOM. 


Patron—HIS MAJESTY THE KING. 


President—Prof. E. RAY LANKESTER, LL.D., F.R.S. 
10: 


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. 


ANNUAL MEMBERS , ; - . £1 1 Oper annum. 
Lirr MemBErs . : : A . 15 15 O Composition Fee. 
FOUNDERS . ; 5 00! NORaO * a 


GOVERNORS (Life Members of Council) 500 O O + 6 


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 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. 


bis 1IUE 


as 2X) SEPTEMBER, 1902. 


ECB 


QUARTERLY JOURNAL 


OF 


EDITED BY 


Py) [ T 7 RN rT\ y if /) uJ 
KH. RAY LANKESTER, M.A., LL.D., F.R.S., 
HONORARY FELLOW OF EXETER COLLEGE, OX¥YORD; 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 
AND OF THE AMERICAN ACADEMY OF ARTS AND SCIENCES OF BOSTON 5 
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; 
DIKECTOK OF THE NATUKAL HISTOKY DEPARTMENTS OF THE BRITISH MUSEUM, LATE FULLERIAN 
PROFESSOR OF PHYSIOLOGY IN THE KOYAT 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.R.S., 


LINACKK PKOFESSOR OF COMPARATIVE ANATOMY AND FELLOW OF MERTON COLLEGE, OXFORD; 
LATE FELLOW OF ST. JOHN'S COLLEGE, CAMBRIDGE; 


AND 


SRDNEY JL. HICKSON, (MA? BUR.S:, 


BEYER PROFESSOR OF ZOOLOGY IN THE OWENS COLLEGE, MANCHESTER. 


WITH LITHOGRAPHIC PLATES AND ENGRAVINGS ON WOOD. 


. TLONDON: 


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


1902. 


Adlard and Son,] {London and Dorking. 


|New Series, No. 182 (Vol. 46, Part 2). Price 10s. 


MICROSCOPICAL SCIENCE. 


i 


| 


CONTENTS OF No. 182.—New Series. 


MEMOIRS: 

PAGE 

Maturation of the Ovum in Echinus esculentus. By THomas 
| H. Bryce, M.A., M.D. (With Plates 10—12) . : : ee Is 

Studies on the Arachnid Entosternite. ai R. I, Pocock. (With 
Plates 13 and 14). : : : : ; P F . 225 

On the Morphology of the Cheilostomata. By Stpyry F. Harmer, 
Se.D., F.R.S. (With Plates 15—] Byiow : : ‘ - 263 

On the Development of Sagitta; with Notes on the ae of the 
| Adult. By L. Doncaster. (With Plates 19—21) , Peet)! 


OC 1902 


MATORATION OF OVUM IN RCHINUS ESCULENTUS. bias 


Maturation of the Ovum in Echinus 
esculentus. 


By 


Thomas H. Bryce, M.A., M.D. 
With Plates 10—12. 


INTRODUCTION. 


Tur subject of the maturation of the sexual cells is a 
thorny terrain. It can be attacked only by the highest 
powers of the microscope, and the facts can only be 
reached by a process of patient mental reconstruction of the 
various phases. Historically the subject has been overlaid 
by some brilliant but premature hypotheses, which, how- 
ever much they may have stimulated research, have also 
tended to foster prepossessions. The necessary stimulus 
for research has been supplied by the hypothesis that 
the chromatin of the nucleus is the hereditary sub- 
stance, or, at least, the bearer from one generation to 
another of hereditary qualities. But apart from the interest 
connected with problems of heredity, and the meaning of 
fertilisation, the study of the intricate details of the process 
of maturation goes to the bottom of all our knowledge of 
cellular morphology. The study of the maturation phe- 
nomena in Kchinus was, in the first place, taken up merely 
with the motive of seeing some of the actual phases in the 
most readily obtainable material. But it was soon dis- 
covered that although the outward phases had been fre- 
quently studied, most of the finer details of the process, as 
seen in Kchinus, were undescribed, and therefore it was con- 
sidered worth while to make a study of the whole process. 

yon, 46, PART 2.—NEW SERIES. M 


178 THOMAS H. BRYCE. 


In view of the hopelessly diverging results for different 
forms obtained by different observers, an interest in the 
behaviour of the chromatin during maturation has declined 
of recent years, and the question of the centrosome has 
occupied more attention. Results which came out led me to 
certain conclusions, which, to my mind, tended to clear up in 
some measure the confusion at present prevailing. As 
the research was proceeding, Strasburger’s work, ‘ Reduk- 
tionstheilung, Spindelbildung, Centrosomen, und Cilien- 
bilden in Pflanzenreich’ (1900), came into my hands. In 
that work conclusions in the matter of the reducing divisions 
identical with my own, and foreshadowed in several 
previous botanical memoirs, are brought, by new com- 
parative investigations, to a focus, and are made a means 
of harmonising the apparently contradictory results in the 
case of plants. This obviously increased the importance of 
my own results, and inspired me to follow out, in spite of the 
large amount of labour involved, the whole series of 
phenomena, in order to obtain as complete a demonstration 
of the facts as possible. 


Maturation IN ECHINUS ESCULENTUS, L. 


Previous Observations on Maturation in 
Echinoderms. 


The Echinoderm ovum has been the classical material 
for all observations on the hving egg. The earliest 
observations on the maturation of the sea-urchin ege 
were made by Derbes in 1847. Agassiz, in 1864, described 
the polar bodies in both Toxopneustes and Asteracan- 
thion. Between 1872 and 1882 Van Beneden examined 
the phenomena in Asterias, Hertwig in Toxopneustus 
lividus and Asteracanthion, Giard in Psammechinus, 
Fol in Asterias glacialis, Greeff in Asterias rubens, 
and Flemming in Sperechinus brevispinosus, Echinus 
miliaris, and Toxopneustes. Since then the favourite 
material for the examination of the phenomena in the 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 179 


living egg has been Asterias. Notwithstanding this, 
little is known as to the finer details of maturation. 
Hartmann (1902) has, since this paper was written, pub- 
lished an account of the changes in this egg up to the forma- 
tion of the first polar spindle. The early observations were 
made on the entire egg—either in the living state, or 
fixed and cleared. The polar bodies in Hchinus are 
normally thrown off within the ovary, and when the naked 
eggs are shed into the sea water they remain entangled 
in the connective tissue of that organ. Sometimes it may 
happen that a partially immature ovary may be manipu- 
lated and some ova caught in the maturation stages. In 
the starfish, on the other hand, the eggs commence to 
show the phases when placed in sea water, and they can be 
watched. Again, by shaking immature sea-urchin eggs 
the stages can be induced artificially. Boveri (1890) has 
figured a few stages after the formation of the first polar 
spindle in Echinus microtuberculatus, but either the 
chromosomes, which are very minute in Hchinus sphera, 
are still more minute in Echinus microtuberculatus, and 
cannot be further analysed, or he has not seen the figures 
which I have made out by my methods. Further, the 
number of chromosomes is different. Matthews (1895) 
examined maturation in Asterias Forbesii. He was able 
to obtain only one ovary showing the stages up to the forma- 
tion of the first polar spindle, but supplemented his observa- 
tions by stages obtained by shaking the eggs. He describes 
the behaviour of the centrosomes, but gives no details as to 
the chromatin. Wilson, in his atlas of ‘ Fertilisation and 
Karyokinesis,’ shows a single photograph of a second polar 
spindle in Toxopneustes, and Boveri has drawn a single 
figure of the second polar spindle in his recent work pub- 
lished in 1901. Haecker (1893) also gives a diagrammatic 
drawing of the first polar spindle, but gives no description 
of maturation. In none of these figures is the finer constitu- 
tion of the chromatin elements represented. Cuénot and 
other observers have written on oogenesis in Echinoderms, 


180 THOMAS H. BRYCE. 


but their observations were confined strictly to the ovary 
and the formation of its epithelium, and to certain points in 
the characters of the nucleolus. Various observers have 
treated specially of Echinoderm spermatogenesis 
(Jensen, Pictet, and others),and many have studied the 
morphology of the spermatozoon, but of these Field (1895) 
brings the latest account. Owing to the excessive minute- 
ness of the chromosomes he seems to have confined himself 
to counting them in the different phases. Haecker (1893) 
published observations carried out on the living egg on the 
germinal vesicle and nucleolus of Echinoderms, but does not 
give any detail regarding maturation. 


Personal Observations. 


Methods.—My material was obtained from animals 
freshly out of the water! Small pieces of close on 
seventy ovaries were fixed, embedded in paraffin, and a 
few dozen sections cut from each. These were all care- 
fully examined, and when maturation was found to be 
proceeding some hundreds of sections were cut and gone 
over, and the details built up from these. The fixative fluids 
used were Flemming’s strong solution, and Hermann’s platinic 
chloride and osmic acid mixture ; almost identical results 
were obtained by both, and the small pieces of ovary—about 
a cubic centimetre or a little more—were well fixed through- 
out. At a later stage of the research, by way of control, 
pieces of ovary were fixed in Boveri’s picric and acetic acid 
mixture, and sublimo-acetic acid, as well as Lindsay John- 
stone’s fluid. ‘he picro-acetic material was unsatisfactory, 
but the sublimate gave good results in some respects. The 
chromatin was, however, much better differentiated by the 
osmic acid mixtures, especially by Hermann’s fluid; while in 

1 The material was obtained at the Marine Biological Station at Millport 
in late March and early April. At the end of April and beginning of May 
the ovaries are mature throughout. In January maturation has already 


begun, and from that time onward the relative proportion of mature to 
immature ova gradually increases, 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 181 


the sublimate material the centrosome gave quite a different 
picture, as will be seen in the sequel. 

Staining.—Osmic acid preparations being proverbially 
refractory to most staining reagents I have confined myself 
almost entirely to Heidenhain’s iron hematoxylin method, 
but have used by way of control other stains. To facilitate 
staining I have always allowed my preparations to stand for 
some time in old turpentine to remove the osmic acid. ‘The 
best results were obtained by iron hematoxylin alone, the 
picture presenting the vivid black chromosomes on a blue- 
grey field. Heidenhain’s preliminary stain with Bordeaux 
red rather confuses the picture of the chromatin, and the 
only other contrast stain used was a very weak coloration 
by alcoholic solution of fuchsin 8. 

It is necessary here to refer to the recent criticism by 
Boveri (1901) of the iron hematoxylin stain. He shows, as 
every one knows who has used the method, that different 
degrees of washing out yield different results, and refers to 
the fact that structures may appear which owe their existence 
to a purely mechanical cause and not to any difference in 
chemical composition. Thus a part which is not readily 
accessible, on account of its position, to the differentiating 
fluid retains the stain while the parts in the neighbourhood 
are decolourised; further, the fluid having a concentric 
effect in washing out, the superficial parts are decolourised 
while the central parts retain the black stain. ‘Thus he 
explains the different accounts given of the structure of the 
centrosome, and points out that by strong extraction of the 
colour even the chromosomes may be apparently diminished 
in size owing to their peripheral parts being decolourised. 
This is weighty criticism in view of a number of the 
appearances I shall have to describe, for he combats the 
generally accepted view that the true appearances are 
obtained by strong washing out, and believes that in regard 
to the centrosomes the opposite is true. As to the chromo- 
somes | may forestall criticism of my results by stating, first, 
that I have obtained similar appearances both with the osmic 


182 THOMAS H. BRYCE. 


acid and the sublimate mixtures, and by other stains besides 
the iron hematoxylin, though the greatest vividness of 
differentiation has been obtained by a combination of Her- 
mann’s or Flemming’s fluid with iron hematoxylin, and, 
second, that the proof that I am dealing with realities and 
not illusions is to be found in the fact that the appearances 
described for the chromosomes represent a complete and 
unbroken series of the steps or stages of a process that can 
be explained only by reference to the completed story. 

The drawings were made by aid of the Abbe drawing 
apparatus of Zeiss, the finer detail being filled in free- 
hand. ‘The combination used was in every case Zeiss 
2 mm. 1:40 numerical aperture, apochromatic objective, 
with either eight or twelve compensating eye-piece. The 
illuminating apparatus employed was a Zeiss 1 mm. 
numerical aperture, achromatic condenser. ‘The sections 
were cut in paraffin, and were of varying thickness. ‘he 
object in most cases being to obtain the masses of chromatin 
entire, comparatively thick sections were taken, six to seven 
microns. Thinner sections down to three microns were 
employed to determine certain points regarding the achro- 
matic structures. 

In dealing with the subject I shall first describe the 
changes in the ovum leading up to the disappearance of the 
germinal vesicle, and after that treat in separate sections of the 
behaviour of the achromatic and of the chromatic structures. 

My earliest preparations are from the growth period. Out 
of a large number of young oocytes of the first order I have 
only seen two or three in mitotic division, and these only in 
the spireme stage, so that I cannot speak as to the number 
of chromosomes in these divisions. ‘The young ovum shows 
a delicate reticular protoplasmic structure (fig. 1). The 
nucleus is already large and vesicular, with a distinct 
nuclear membrane, and a deeply staining eccentric nucleolus. 
This being intensely black, contrasts strongly with the 
granular and irregular nuclear network, which refuses to 
take on the chromatin stain, and remains pink in prepara- 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 188 


tions stained either with “fuchsin S” or “ Bordeaux red.” 
There is frequently a second smaller deeply staining circular 
body in the nucleus, but it has no regularity in position and 
is not invariably present. I cannot in any of my prepara- 
tions see the double nature of the threads described by 
Haecker. Close to the nucleus, very frequently on the side 
of that body towards which the nucleolus lies, there is some- 
times at this stage a body which presents much the appear- 
ance of the centrosome of a resting cell. It consists of 
either a single granule or pair of granules, sometimes a 
group of smaller granules enclosed in a circular area. While 
this may represent a centrosome it is impossible to dis- 
tinguish it from similar bodies with central granules that 
may be found in other parts of the cell, which are un- 
doubtedly cell inclusions, and therefore no structure can with 
certainty be identified as a centrosome. 

Structure of the Protoplasm.—Wilson (1899) has 
shown that in the young ovum the protoplasm is granular, 
and that as the ovum grows in size an alveolar structure 1s 
assumed. In the youngest ova of my fixed material the 
protoplasm presents a granular appearance which is certainly 
not alveolar, and can hardly be termed reticular (fig. 1). In 
the fully grown egg the appearances vary according to the 
stain. In fig. 3 the cytoplasm is represented as showing a 
reticulum which is composed of separate minute granules ; 
the meshes of this reticulum bound alveolar spaces. ‘These 
alveoli are on the whole rounded, and in this particular 
specimen, from which the iron hematoxylin was very 
thoroughly washed out and{replaced by a slight counter- 
stain by fuchsin, they were faintly red with a slightly darker 
periphery. In fig. 2, on the other hand, the appearances are 
different. The iron hematoxylin has not been so com- 
pletely washed out, and the alveoli have retained the dark 
stain, showing up as rounded dark points separated by an 
unstained reticulum. Sometimes the centre of the alveolus 
is occupied by a black dot, as if the centre had not been de- 
colourised. hus my preparations fully bear out Wilson’s 


184. THOMAS H. BRYCK. 


latest conclusion (1899) regarding the structure of the sea- 
urchin egg—namely, that the condition of the cytoplasm 
conforms to Biitschli’s description. It has the same physical 
characters as an emulsion; that is, there is a fluid framework 
in which the microsomes are suspended, and the alveoli are 
filled with a fluid of different physical characters. When 
the alveoli are wholly destained all that is seen is the micro- 
somic network, whereas when they are stained the alveoli 
stand out as the yolk granules embedded in the cytoplasm. 
Wilson shows, however, that the cytoplasm at certain periods 
may have a fibrillar structure, but to this point I shall return 
later. 

The changes which the nucleus undergoes during the 
vrowth of the oocyte, until it becomes the fully developed 
germinal vesicle, are very complicated and uncertain. 
Many irregular figures suggest that the germinal vesicle 
may undergo changes of shape. ‘They may well be arte- 
facts. I shall only refer to certain facts regarding the 
chemical reaction of the nucleus, which seem to be fully 
vouched for in my preparations. It has been shown by 
a number of observers that the staining reactions of the 
uucleus vary at different times. At one time the chromatin 
network will take the specific stains deeply, while at other 
times it remains unstained (Riickert, 1892). My experience 
tends to support these statements, though one must 
admit that very different effects are produced by different 
degrees of coloration with iron hematoxylin. ‘The effect 
depends on the degree of extraction of the colour, but it is 
quite certain that at certain stages of the nucleus the 
network very readily parts with the black stain, and is left 
as an irregular granular reticulum of a blue-grey colour, or 
of a red tint, in preparations stained for contrast with rubin. 
The nucleolus, on the other hand, is exceedingly tenacious of 
the stain, and appears as an intensely black spot (fig. 2). 
Again, at a stage I consider to be of later date, the network 
shows a basis of delicate linin threads, with deeply stained 
chromatin particles arranged on the thread, giving it a very 


MATURATION OF OVUM IN ECHINUS KSCULENTUS. 185 


irregular or feathery structure, while the nucleolus generally 
is less deeply stained and vacuolated (fig. 2). Finally, when 
the nucleus is fully grown and maturation imminent, we find 
the contrast is exactly the opposite of that described for the 
young nucleus. The network is intensely black, consisting 
of particles of chromatin arranged in a very intricate and 
irregular fashion, while the nucleolus parts with the stain 
very readily, and is left as an almost colourless, apparently 
empty vesicle. Soon after the resolution of the nuclear 
membrane it disappears from view. Very similar changes 
are described in many other forms,—for instance, in the 
Turbellarians, according to Francotte (1897) ; in Polycheerus, 
according to Gardiner (1898) ; and according to Gathy (1900), 
in Tubifex (an Annelid) the nucleolus loses its capacity for 
staining with iron hematoxylin at the end of the growth period. 

It is difficult to resist the conclusion that the chromatin 
substance is at first confined to the nucleolus, and later leaves 
it to form the chromatic basis of the nuclear network as a 
whole, and therefore also of the future chromosomes. The 
fate of the nucleolus in Hchinoderm ege’s has been variously 
interpreted. Derbes (in 1847) thought it was directly 
converted into the pronucleus of the mature egg, and 
Hertwig (1877) took the same view. Fol (1877) and 
Flemming (1882), however, proved that the chromatin 
itself became the future nucleus, after it was provided 
with a new nuclear membrane. Recently Carnoy and 
Le Brun (1899) have maintained the view that in the 
amphibian egg where there is no chief nucleolus, but a large 
number of smaller ones, certain of these become converted 
into the future chromosomes, thus reverting to the older 
view of Schultze (1887). It seems certainly true, as said 
above, that the chemical substance which is lodged in the 
nucleolus in the early ovum becomes later distributed into 
the germinal vesicle, and so indirectly goes to form the 
chromosomes. Hartmann (1902), for Asterias glacialis, 
describes the chromosomes passing directly out of the nucleo- 
lus, the remainder of the nuclear reticulum being rejected. 


186 THOMAS H. BRYCE. 


Strasburger regards the body as a storehouse of reserve 
substances, which pass into the cell during division to 
form the “kinoplasm,”’ which goes to form the spindle, 
the Hautschicht, membranes, and cilia. We shall see later 
that the phenomena observed in the sea-urchin egg 
may combine these two views. But in contradiction to 
both is Haecker’s view. His observations on the living 
ege@ of the sea-urchin reveal to him the nucleolus as a 
pulsating organ in which, periodically through the whole 
growth period, small vacuoles appear; these run into a 
single central vacuole, which increases and then diminishes 
in size. When the largest central vacuole appears the 
nucleolus removes itself to the periphery of the nucleus, and 
meantime the vacuole comes into relation with the outer 
layers of the nucleolus, as if to bring its contents into 
relation with the nuclear sap; and further, an indrawing of 
the wall of the germinal vesicle itself suggested that there 
was a communication between the cytoplasm and nucleolus. 
From these and other observations Haecker regards the 
nucleolus as a secretory organ, collecting the by-products of 
nuclear activity—not as a storehouse, or “ nuclein labora- 
torium ” (Fick, 1899). So far as my observations go, they 
tend to support the idea of the nucleolus being a storehouse 
or laboratory of nuclein. 

Centrosome.—There has been a great deal of discussion 
as to this enigmatical structure in the sea-urchinegg. Vary- 
ing accounts have emanated from Boveri, Wilson, Fol, 
Biitschli, Reinke, Hill, Kostanecki, and Erlanger. Boveri 
says, “Das Seeigel-Ei ist von allen objecten die von mir 
bekannt sind, dasjenige, welches einer sicheren Darstellung 
der Centrosomen die grossten Schwierigkeiten bereitet.” 
This quotation is taken from his recent work, ‘On the 
Nature of Centrosomes.’ He reconciles more or less the 
different accounts, and suggests a nomenclature which I 
shall adopt as being the latest and most authoritative. 

The centrosome is composed of a special and peculiar 
substance, the centroplasma, which, according to the perfec- 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 187 


tion of fixation and the manner of staining, presents different 
appearances. ‘his accounts for the different forms under 
which the body has appeared. It stains best with iron 
hematoxylin, and destains concentrically. When destaining 
has been carried far, it shows as a discoidal area surrounded 
by a clear halo, and has a very fine alveolar structure. This 
is the form in which I have observed it in all my osmic acid 
preparations, except that I do not see the halo, and when 
counterstained with rubin it has a red colour, which in- 
definitely fades away into the bluish-grey astral rays and 
spindle fibres. This rounded body, as division proceeds, 
becomes enlarged, then lens-shaped, and ultimately flattens 
into a plate which lies along the side of the nucleus. In 
polar view this is dumbbell-shaped; the enlarged ends are 
the daughter centrosomes which become surrounded with new 
radiations. In the maturation stages my preparations are 
not numerous enough to enable me to follow in detail the 
behaviour of the centrosomes, and I have not been fortunate 
enough to see the division of the body in the first maturation 
spindle. The centrosome of Boveri corresponds to the 
centrosphere of Wilson. In another set of preparations less 
destained, Boveri described the centrosome as a smaller body, 
showing in its centre a darkly staining particle, the centriole, 
which corresponds to Wilson’s centrosome. ‘his, as division 
proceeds, divides into two, and goes through the usually 
described evolutions. In_ picro-acetic and sublimo-acetic 
preparations I have seen such a centriole, but have been un- 
able to trace its division. Again, when destaining has been 
stopped early the whole centroplasm is black. This I have 
also seen in picro-acetic and sublimo-acetic material. The 
rays, according to Boveri, stop at the margin of his centro- 
some, and do not enter it so as to be inserted into the centriole. 
This seems to be the case, and in my cleavage preparations 
fixed with Lindsay Johnstone’s fluid the central reticular 
body is sometimes seen to have completely dropped out, so 
that the astral rays are seen to end abruptly, leaving an 
absolutely round empty space occupied in the other eggs by 


188 THOMAS H. BRYCE. 


the alveolar or reticular centroplasm. I do not presume to 
give an opinion on the much vexed question of the persistence 
of the centrosome as a special cell organ, but one thing seems 
clear, that the centroplasm is a focus of protoplasmic 
activity, and is ultimately to be explained on physiological 
and not on mechanical grounds. 

Changes in the Germinal Vesicle Preparatory to 
Division.—When the germinal vesicle has reached its full 
growth the nucleolus loses its staining capacity to chromatin 
stains, the nuclear network takes an intense stain, and the 
cytoplasm to its very outer edge is seen to have an alveolar 
structure. In many cases, presumably in stages close to the 
onset of maturation, the nuclear membrane is puckered. The 
germinal vesicle then moves towards the surface, and, as long 
ago described by Hertwig (1877) for Asteracanthion, at the 
spot nearest the surface a protoplasmic process projects into its 
interior (fig. 4). In osmic acid preparations the nuclear mem- 
brane is seen to be indented and folded before the process ; 
this, as it projects inwards, spreads out in every direction 
from the neck, so that at the margins of the process are seen 
sections of peninsule andislands. In sublimate material the 
nuclear membrane is not so sharply differentiated, and the 
inward folding of it is not so clearly seen. I have no doubt 
from my sections, such as shown in fig. 4, that this is a true 
invagination of the germinal vesicle by the cytoplasm. At 
the neck of the invagination the alveolar walls of the cyto- 
plasm are drawn inwards towards the centre, but in the pro- 
cess itself no very distinct fibrillar structure is at first to be 
seen. 

Hartmann represents at this stage a very distinct aster 
between the invaginated wall of the vesicle and the surface 
of the ovum. I have not seen such an aster in my sections ; 
the wall of the vesicle is always very close to the surface of 
the egg, leaving no room for such a formation, and the aster 
seems to form within the process. Sometimes the radiations 
from the neck of the invagination are much better marked 
than in the ovum represented, and in the process itself there 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 189 


is a distinct suggestion of radiation, which is, however, very 
difficult actually to define. The appearances suggest that 
the centrosome or kinetic centre les in the neck of the 
invagination ; but at this stage there is not, so far as the 
study from sections can determine, any wide-spreading astral 
formation as in Asterias. Soon it is seen that the whole mass is 
made up of what seem to be looping fibres. Possibly the folded 
and puckered nuclear membrane contributes to this appear- 
ance. In fig. 5 is represented a stage in which the chromatic 
reticulum has become finer, and at the neck of the process is 
an irregular mass which is destined to form the future chromo- 
somes. All this time the germinal vesicle remains close to 
the periphery of the egg. 

The next stage I can determine is the one represented in 
fig. 6. The nuclear membrane has now entirely disappeared, 
and in the irregular mass of looping fibres there are seen two 
asters. In each is a circular finely reticular area, the centro- 
plasm, and from the periphery pass out in every direction 
very delicate interdigitating fibres. Between the two asters 
the fibres are drawn out to form an irregular spindle arrange- 
ment. Round this area the greater part of the nuclear 
reticulum, which does not form chromosomes, but was related 
to the vegetative stage of the germinal vesicle, is seen merg- 
ing with the cytoplasm, but still retaining its reticular 
character. Betwéen the spindle and the surface the chromo- 
somal chromatin mass is seen. 

This description corresponds with Hertwig’s original 
account, and also in the main with Hartmann’s recent repre- 
sentation of the facts, but differs from Fol’s in that he describes 
no process projecting into the vesicle. It also differs from 
Mathews’ description of what occurs in Asterias Forbesii. 
He describes the two centrosomes probably passing out of the 
germinal vesicle at the nearest point to the surface of the 
egg by the rupture of the nuclear membrane at that point. 
They then pass some distance from the nucleus, and are seen 
to have round them a faint halo of ‘archoplasm.’’? This 
latter becomes distinct, radiations are developed, the whole 


190 THOMAS H. BRYCE. 


archoplasmic area divides, and the two parts being drawn 
asunder, a spindle is spun out between them, which moves 
tangentially over the nucleus. As it grows the spindle-fibres 
project into the vesicle, the nuclear membrane is dissolved, 
and the spindle then rotates to become the first polar 
amphiaster. In Haecker’s text-book (p. 123) the process in 
the living egg is described in much the same fashion. A 
clear area is developed between the remains of the germinal 
vesicle and the surface, surrounded by a radiation which soon 
forms a double star, which is the beginning of the amphiaster. 

The invagination of the germinal vesicle in the egg of 
Echinus is probably secondary. It may be related to the fact 
that the wall of the vesicle comes exceedingly close to the sur- 
face of the egg. The mounting of the vesicle to the surface 
is a fact which, so far as I know, has not been satisfactorily 
explained. Haecker (1893) suggested that it is due to the 
action of gravity causing a movement in the elements of the 
ego after the force connected with the exchange of material 
between nucleus and cytoplasm, which keeps the vesicle in 
the centre of the egg, ceases with full growth. My prepara- 
tions do not throw any light on the point. 

Fig. 8 represents a somewhat oblique section of the 
germinal vesicle at a later stage. It shows the two asters 
arranged tangentially to the surface of the egg, but between 
them, and extending towards the surface, is a finely reticular 
mass, out of which the delicate wavy and interdigitating rays 
of the asters are evidently spun. Hmbedded in this reticulum 
are seen the chromatin segments. At a later stage (fig. 9) 
all these are drawn into the area between the asters, which is 
seen now as a finely alveolar or reticular plate. Round this 
central plate is a complicated reticulum of fibres crossing and 
intercrossing, but on the whole radiating from the central 
plate. In this reticular zone is also seen, at a little distance 
from the plate, one centrosome surrounded by rays, obviously 
part of the general reticulum. In the adjoining section a 
second aster was present on the side of the plate removed 
from the surface ofthe ovum. At this stage one hardly ever 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 191 


sees a preparation in which both asters are cut in the same 
section. Griffin, in Thalassema, has described a disappear- 
ance of the spindle spun out between the centrosomes, and 
the development of a central mass very hke that which I 
have described between the asters. In most instances, 
however, the conditions are more like those described by 
Matthews in Asterias Forbesii. 

It is evident that when the nuclear membrane disappears, 
and the rejected chromatin passes into the cytoplasm, a pro- 
found effect is produced on the organisation of the ege. 
Whereas, with the germinal vesicle still intact, the alveolar 
structure can be traced to the surface of the egg, we now 
find that round the transformed nucleus, and projecting into 
the centre of the egg, is a fibrillar mass, which is sharply 
differentiated from the alveolar yolk. From this central 
area there also extends round the surface of the egg a layer 
of differentiated protoplasm. The central mass and surface 
layer have each a definite fate. The one is differentiated 
into the spindle and asters, while the surface layer is, I 
believe, associated with the formation of the membrane 
thrown off by the egg at the moment of fertilisation. The 
central mass of the yolk is unchanged in appearance, and the 
question is whether this reticular mass of protoplasm is 
differentiated from the cytoplasm, or is derived from the 
rejected nuclear reticulum. I am inclined to think that it 
is in large measure formed from, or under the influence 
of, the discarded nuclear material. This would be in 
harmony with the results of Carnoy and Le Brun (1899) 
in Triton. Another evidence of the excitement produced in 
the egg at this stage may perhaps be seen in the accessory 
asters formed, which, so far as I can see, have no relation to 
the formation of the definite asters of the spindle. 

From experiments by R. Hertwig (1896) and Morgan 
(1896) it seems that under special artificial chemical stimulus 
the cytoplasm may be excited to form asters, and even, in 
Hertwig’s experiments, amphiasters. Reinke (1894) also 
found that in the peritoneal cells of the larval salamander 


192 THOMAS H. BRYCE. 


three grades of asters are formed—primary, secondary, and 
tertiary. The last contribute to the secondary, and these 
again to the primary or definitive asters. Carnoy described 
accessory asters during the formation of the second polar 
body in Ascaris, and Meade (1897) showed that a great 
number of such asters were formed before the formation of 
the first polar spindle in Cheetopterus (an Annelid), which he 
thought contributed to the formation of the spindle asters. 
Watase (in Macrobdella) found as many as thirteen asters in 
the cytoplasm, with centres varying in size from the smallest 
microsome to the true centrosome. Griffin (1899) also de- 
scribes the formation of accessory asters in Thalassema. 
These experiments and observations are held to afford strong 
evidence of the free formation of the centrosome, in which 
case both that body and its aster would be the expression 
rather than the cause of cell activities. 

The secondary asters in Hchinus at this stage are pos- 
sibly produced in the cytoplasm under the influence of the 
nuclear material let loose on the disappearance of the nuclear 
membrane. 

All this tallies better with Strasburger’s views of the kino- 
plasm than with any other theory. He thinks of protoplasm 
as of two kinds, trophoplasm and kinoplasm: the former is 
vegetative in function and alveolar in structure ; the latter 
presides over the activities of the cell, forms centrosomes, 
mid-bodies, asters, and spindles, constitutes a peripheral layer 
from which membranes and cilia are derived, and is fibrillar 
in structure. This differentiation of the protoplasm takes 
place when mitosis sets in. Further, he thinks the nucleolus 
is a storehouse of reserve material, out of which, on need, the 
substance of the kinoplasm is drawn. 

I have shown that the nucleolus at first seems to contain 
all the chromatin substance which later is found in the 
nuclear reticulum, the larger portion of which is rejected, to 
form in turn, if I be right, directly or indirectly a 
reticular zone, out of which the asters and spindle are 


spu n, 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 193 


My conception of the meaning of the changes in the ovum 
does not, however, involve an acceptance of either Stras- 
burger’s kinoplasmic or of Boveri’s archoplasmic theory. It 
inclines rather to the view that the same ground substance, 
under the influence of the chemical changes underlying vital 
activities, may take on different forms in response to varying 
physiological needs, and further, that whereas, during the 
vegetative period, the main centre of these chemical activities 
hes in the nucieolus, in the division period that centre is trans- 
ferred to the centrosome, which is the expression of activities 
resulting in the vital phenomena of division. 

Fig. 15 represents a later stage. The spindle is not yet 
complete, but the two asters are situated radially, and the 
reticular mass, though still showing in some parts a radial 
distribution from the central plate, is becoming more and 
more focussed on the centrosomes. The spindle, in most 
forms, is said to be fully formed before this radial position is 
assumed, and the whole spindle is said to rotate through 
90 degrees. A very good example of this is seen in the egg 
of the mouse, as described by Sobotta (1895). 

In my preparations the spindle, as is the case also in 
Thalassema (Griffin), is late in being completed, and the asters 
seem to move independently through the cytoplasm, the fibres 
arranging themselves round the centre of activity until the 
definitive position is reached. 

The conditions described for the formation of the polar 
spindle are not unlike those accompanying the formation of 
the multipolar spindles described in the pollen and the spore- 
mother cells in many plants by Farmer, Belajeff, Osterhout, 
Mottier, Nemec, and Byxbee. According to the description 
of these authors there is a filar zone round the nucleus, out of 
which the multipolar figure is spun, the poles of which draw 
together to form the definitive bipolar spindle. I have seen 
one or two four-poled first-maturation spindles, but I cannot 
make out that in the reticular zone there are more than two 
asters which have any relation to the future spindles, and 
such four-poled spindles would thus merely indicate the 

vou. 46, part 2,—NEW SERIES, N 


194 THOMAS H. BRYCE. 


tendency to the formation of multiple centres of activity, or 
putting it in terms of the centrosome, to the formation of four 
centrosomes instead of two. 

The whole process leading up to the formation of the first 
polar amphiaster is very complicated in Echinus, and 
extremely difficult to trace in sections. It is as difficult 
to be sure of the phases in the living Hchinus egg, which 
are moreover difficult to get as the process normally 
takes place within the ovary. ‘The process does not 
seem to me to be so simple as it has been described for 
Asterias ; indeed, it is in many respects like what Mead has 
described for Cheetopterus. Dr. Teacher has recently studied 
the phases in the living egg, and has kindly let me see his 
results, which supplement my own. He has seen frequently 
a stage which I have described as follows—“ near the surface 
of the ovum is a clear granular area having the appearance 
of ground glass, surrounded by a darker ring merging into 
the alveolar-looking cytoplasm. This ring, at its circumfer- 
ence, is distinctly irregular, and suggests delicate radiations 
from the central granular area.” This is obviously the 
stage represented by the section depicted in fig. 9, and 
corresponds to the transformed germinal vesicle. At this 
stage I could not make out distinct asters in the living egg, 
and in the sections the astral rays which lie within this area 
are of great delicacy. Dr. Teacher has seen in this phase 
many specimens with a number of asters in the cytoplasm 
around the transformed germinal vesicle, and has made out 
at the same time, within the area itself, astral formations 
which he believed to be the definitive asters of the spindle. 
While, therefore, I have in the foregoing description traced 
the centrosomes as if they were persistent centres travelling 
through the cytoplasm, I cannot exclude the possibility of 
their free formation as described by Mead. 

Fig. 16 represents the spindle now completed. The 
chromosomes are being drawn into the equatorial plate. 
Fig. 17 shows the now completed spindle in metaphase. 
It is relatively bulky, with blunt and rounded ends, and 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 195 


frequently, when the section is through the side of it, the 
centrosome is not cut at either end. There is no central 
spindle. The chromosomes extend through the whole 
equatorial plate, and the peripheral rays of the asters are 
seen interdigitating opposite the equator. The fibres of the 
spindle itself are somewhat uneven, and in relation to the 
chromosomes there are darker bundles apparently of several 
fibres spun together. As metakinesis proceeds the waviness 
of the fibres becomes more distinct. The central centrosome 
becomes flattened, but I cannot determine the manner in 
which a division takes place. The outer centrosome 
diminishes in size. Its central astral rays shorten, fig. 19, 
and ultimately disappear as the apex of the spindle is 
protruded. The lateral rays are obliterated progressively 
until the point of the spindle stands clear. The first appear- 
ance of the protrusion of the polar body is a tiny elevation 
into which the end of the spindle is directed. Later, when 
the spindle has risen to the height of the equatorial plane, 
there is seen a depression on the surface of the egg where the 
constriction takes place, and in which afterwards the polar 
body lies. The rise of the spindle and its protrusion are very 
difficult to explain. It remains approximately of the same 
length throughout, and I do not see any special development 
of the central aster over the polar one. Wilson (1900) finds 
evidence in the protrusion of the spindle in favour of 
Diiner’s (1895) theory that the divergence of the poles of the 
spindle in mitosis is due to the progressive elongation of the 
central spindle. In Hchinoderm ova, neither before nor after 
fertilisation, is there a central spindle spun out between the 
centrosomes, but it is probable, according to Wilson, that the 
difference is only a secondary one, and that the spindle 
consists in part of continuous fibres, and the waviness of the 
spindle-fibres in the metakinesis would speak for the pushing 
hypothesis. In any event, I cannot see how any hypothesis 
founded on mechanical principles, such as illustrated in 
Heidenhain’s model, can explain the peculiar circumstances 
of the polar mitosis, 


196 THOMAS H. BRYCE. 


Fig. 23 shows the earliest phase of the second division 
which I have had the opportunity of observing. The asters 
are already separate, and a bunch of fibres from each is pro- 
jected towards the chromosomes, which are immediately 
drawn into the equator of the spindle. Thus no resting stage 
intervenes between the two divisions. The whole figure is 
still surrounded by the remains of the reticular or kino- 
plasmic zone. The spindle when fully formed is slighter than 
the first polar spindle. The central centrosome and aster 
progressively increase in size until the condition is found as 
in fig. 30. The astral rays are thick and fairly straight and 
widely spreading. The behaviour of the outer centrosome 
and the manner of protrusion of the polar body is exactly as 
I have described for the first polar body (figs. 27—29). 

Fig. 33 shows the condition of the nucleus long ago de- 
scribed by Hertwig after the extrusion of the second polar 
body. The first stage in the reconstitution of the nucleus is 
the formation of several small vesicles, which run together to 
form a single vesicle which is the mature nucleus. ‘The 
description given of the process in the living egg is, that 
several small vesicles appear approximately in the middle of 
the radiations remaining in the egg. In the sections this is 
clearly seen not to be the case, but the vesicles surround the 
centrosome, and the astral rays are broken up into bundles 
passing out between them. Later these all disappear, and a 
single vesicle is left without any trace of centrosome or radia- 
tion in its neighbourhood. 

Fig. 34 shows an interesting abnormality of the second 
polar body. It is here very distinctly a small cell, and pre- 
cisely the same phenomena are seen in the reconstruction of 
the nucleus as in the egg. 

I must now refer to a series of figures which accompany 
the constriction of the spindle in both maturation divisions. 
Associated with the disappearance of the spindle is formed 
the body called by Flemming the “ zwischenkorper.” This 
plays a considerable rdle in spermatogenesis, but is figured 
also in a considerable number of the descriptions of polar- 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 197 


body extrusion. I have seen it in various forms. In fig. 30 
are seen round the constricting spindle a series of points 
which afterwards, as seen in fig. 23, condense to form a ring 
round the remains of the spindle. It seems to persist for 
some time, fig. 35, and then disappears. 

A final point still remains to be described. When the 
matured nucleus retires towards the centre of the egg all 
remains of the reticular or kinoplasmic zone have dis- 
appeared, and the nucleus lies surrounded by the alveolar 
yolk, while round the periphery of the egg the kinoplasmic 
girdle has narrowed down into a delicate layer of differenti- 
ated protoplasm. In sublimate material this is seen as a 
distinct layer, in which large microsomes are arranged 
regularly side by side. In the osmic-acid material the dis- 
tinction is less sharp, but there is generally a difference in 
the characters of the surface layer. I think that possibly 
this layer has to do with the formation of the membrane 
thrown off when the selected spermatozoon enters the egg, 
and, as has been said, I refer it to the kinoplasmic zone 
which is differentiated on the breaking down of the germinal 
vesicle. 

History of the Chromatin.—As has been described 
the greater part of the nuclear reticulum is rejected, and 
gives rise probably to the reticular zone round the trans- 
formed germinal vesicle. Close to the base of the neck of 
the invading cytoplasm is found an irregular mass of chro- 
matin, just as Matthews describes for Asterias, which is pre- 
sumably the chromatin destined to form the future chromo- 
somes, figs. 5—7. ‘This condensation of chromatin at one point 
perhaps corresponds to Moore’s (1895) synaptic phase, though 
only a part, not the whole of the chromatin, as in spermato- 
genesis, is involved in the condensation. Hmerging from this 
condensed mass are seen in figs. 5 and 6 a series of separate 
elements as to the number of which I am not certain, but I do 
not think there are more than at a later stage. ‘The following 
stages, figs. 8 and 9, involve the collection of this mass of 
chromatin elements into the central plate before described. 


198 THOMAS BH. BRYCE. 


Often one sees the chromatin collected to one side of this 
plate; sometimes the separate elements are widely scattered ; 
in many instances, as in fig. 12, there are chain-like clusters, 
which suggest that a thread is being broken up into 
segments, and in practically every ovum at this stage one 
sees compound masses which are breaking down into the 
separate elements which enter the equatorial plate of the 
spindle. Hartmann, as already mentioned, has quite recently 
described the chromosomes as arising directly from the 
nucleolus. They arise as isolated rods, clumps, or threads 
having the chromatin particles arranged in series in them. 
The nature of my material makes it impossible for me either 
to deny or affirm the direct origin of the chromosomes from 
the nucleolus, but the appearances I have described are 
not otherwise at variance with those described by Hart- 
mann. 

I have between fifty and sixty sections of this stage, and 
the relatively large number indicate that the prophase is 
protracted. From the very first these always present the 
same form. Fig. 14 shows a fragment of the thread com- 
posed of spheres, united by a less deeply staining subtance. 
When separation is complete sixteen tetradal chromosomes 
of nearly uniform appearance are found. When seen from 
the side they have a dumb-bell shape, when seen en face 
they are obviously the tetradal groups of authors. I have 
counted the chromosomes at the various stages again and 
again, and have always reached the number fifteen or sixteen. 
Fifteen is an improbable number, and I feel sure that the 
proper figure is sixteen. I have never succeeded in making 
the number eighteen, which would be double the number 
(nine) found by Boveri (1890) in Echinus microtubercu- 
latus. R. Hertwig (1896) made the number of chromosomes 
emerging from the germ nucleus, in his experiments on the 
development of unfertilised sea urchin eggs, sixteen or 
eighteen, which would agree with my results. Field 
(1893) in EKchinoderm spermatogenesis counted twenty-six to 
thirty-two chromosomes in the spermatogonia, sixteen to 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 199 


eighteen in the spermatocytes, and eight or nine in the 
spermatids. He confesses to great uncertainty in regard to 
these figures on account of the minuteness of the chromo- 
somes, and the last figure is quite out of harmony both with 
Rh. Hertwig’s counts and my own. 

Careful analysis of this tetrad body shows that it is com- 
posed of two short stout rods placed side by side (figs. 8 
—10, 12, and 23). The ends have the form of little spheres, 
and looking back to fig. 14 one may conclude that they corre- 
spond to the spheres seen in that thread united in pairs, but 
there is no transverse cleavage of the thread between the 
four spheres. A complete tetrad, consisting of four indepen- 
dent round bodies as figured for Ascaris, or the mole-cricket, 
does not occur in Hehinus. Further, one cannot regard the 
two rods as separate and independent at this stage; they 
are bound together closely, and the figure is really a com- 
pound chromosome. 

According to the above interpretation the tetrads thus arise 
by a single longitudinal split of an original thread or threads. 
At no time are there any ring or other irregular figures, as 
described in so many other cases. The possibility. is not 
excluded, that the groups might result from conjugation of 
the dyadal bodies in pairs, as described by Wilcox (1895) 
in the grasshopper, and Calkins (1895) in the earthworm. 
In two instances only out of a large number of prophase 
stages have I seen a figure other than those described. In 
one section, just before the spindle is formed (fig. 15) there is 
a double comma form, which appears in all other sections at 
a later stage. 

As the compound chromosomes are gathered into the 
equatorial plate they lie irregularly, and in the metakinesis 
they do not seem to be resolved simultaneously, for in all my 
sections of this stage, about sixty in number, figures in 
different phases are seen, and as the chromosomes lie 
throughout the whole equatorial plate, and not only round 
the periphery of the spindle, various irregular bodies are 
seen which are portions only of whole chromosomes. ‘lhe 


200 THOMAS H. BRYCE. 


relatively large number of sections obtained in this stage 
indicates that it is of long duration. 

The varied figures drawn in figs. 18 and 21 are capable of 
only one satisfactory explanation, keeping in view that the 
end result is always the same. The little rods come to be 
placed radially on the spindle. Their central ends move 
apart to form a T-shaped figure. The cross-piece of the T 
representing the separating limbs opening out on the 
spindle, the stem of the T the outward directed, and still 
united portions of the chromosomes. As separation proceeds 
the stem of the T is pulled down until the figure is like two 
commas placed end to end. It is obvious that this evolution 
will open out the chromosome along the plane of the original 
longitudinal split from within outwards, as is seen in a series 
of drawings (fig. 21) of the chromosomes in profile view, but 
when observed en face (same figure) it is equally clear that a 
second longitudinal split has simultaneously been effected 
along a new plane, from without inwards, giving the double 
V-shaped figures represented in figs. 18, 19 and 21. 

If we describe the appearance in terms of the minute 
terminal spheres of each rod, we see that the spheres come to 
lie in a row exactly as Wheeler (1897) describes in Myzo- 
stoma glabrum. The equatoria] bodies then divide (figs. 
16—18 and 21), but the terminal spheres of each rod remain 
undivided, and are drawn away from the equatorial spheres, 
so that the whole chromosome is lengthened out very greatly, 
and the apical spheres are carried far away from the equa- 
torial, delicate, less deeply staining threads uniting them 
together. The equatorial spheres, after remaining long in 
contact in the equator, then part, and give rise to V-shaped 
figures with a single apical and two equatorial spheres, one 
at the end of each limb. These figures then shorten up by 
the contraction of the elongated thread, and in the final 
anaphase condense into short stumpy masses (figs. 19 and 21). 

These, when analysed, show that the apical sphere has also 
divided, and we have produced small tetradal bodies exactly 
like those in the prophases of the division, but of smaller 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 201 


size. In reality, just like the earlier bodies, they are short, 
somewhat curved rods, with dilated extremities placed side 
by side. Those at the outer pole pass into the first polar 
body, and those remaining in the egg persist, enlarge 
somewhat, and pass otherwise unchanged into the second 
polar spindle. Sometimes during the metakinesis the second 
longitudinal split is not so evident, and then long drawn 
out threads are seen, the double nature of which is difficult 
to make out. Ultimately, however, the two halves separate 
in the anaphase exactly as in other cases. Exactly similar 
figures have been described by quite a number of observers 
in other forms, for instance, and especially distinctly, by 
Klinckowstrém (1897), Francotte (1897), Van der Stricht 
(1898), Griffin (1899), Gathy (1900), but, as I shall describe 
in the sequel, their interpretation has been different, and 
leads to very ditferent theoretical conclusions. The transi- 
tion between the first polar andthe second polar spindle is very 
rapid, so that the number of sections found in this stage is 
relatively few. 

The little compound chromosomes are drawn into the 
equatorial plate of the second spindle (figs. 23 and 24), 
and there different appearances are seen, according to the 
plane of the section. In fig. 24 we have apparently little 
tetrads, which are really the lobed ends of the small, slightly 
curved chromosomes. In fig. 25 again the rods are seen 
lying back to back. These rods I have every reason to 
believe, from the various figures I have drawn, open out just 
as in the first spindle, only there is no second longitudinal 
split, and therefore the division is homotypical. A single 
preparation rather suggests that the rods may sometimes 
be simply separated along the plane of cleavage. It may 
well be that both methods are adopted, according to whether 
the body lies radially or tangentially to the spindle. The 
result 1s the same; the separation is effected in the plane 
of cleavage established in the anaphase of the first division. 

Similar figures in the second division have been described 
by the authors above mentioned, and in other instances, 


202 THOMAS H. BRYCE. 


also, the short, slightly curved rods have somewhat the 
appearance of tetradal groups. When the daughter chromo- 
somes have separated they pass to the poles of the spindle. 
Those at the external pole pass out with the second polar 
body, and remain as short, stout, distinctly bilobed bodies in 
many instances, after the second polar body is cut off 
(fig. 31). Those remaining in the ovum, however, at once 
begin to lengthen, and in the telophase are seen (fig. 
30) as long, bent rods. ‘These are gathered into the 
series of vesicles already described. Within each of these 
vesicles are seen elongated, curved rods, and round the walls 
there are tiny particles of chromatin, forming an incomplete 
membrane (fig. 33). Later, when the vesicles are fused, the 
nucleus is seen to be bounded nearly all round by semicir- 
cular loops of chromatin, and in the centre the reticulum is 
becoming restored (fig. 31). At a later stage (fig. 32) the 
reticulum takes on the form of irregular feathery strands, 
beset with chromatin granules of varying size, accumulated 
here and there to form irregular net-knots of chromatin. 
All trace of the separate chromosomes is absolutely lost in 
this network. 

The phenomena attending fertilisation and cleavage are so 
well known that I do not intend to enter on that subject, but 
I wish to refer to the behaviour of the chromatin threads in 
the metaphase of the cleavage division. ‘he primary rods 
segment into about thirty-two chromosomes. I have counted 
them in cross sections of the spindle a good many times, and 
generally reach that figure, which would make my count of 
the chromosomes in the maturation stages fall in exactly 
with the general law. 

Each chromosome when divided forms first a V-shaped 
figure. This mounts on the spindle so that a loop is formed 
with its apex directed outwards, and the ends of this loop 
are drawn out to the poles of the spindle, the threads 
lengthening as they go. Finally, the daughter chromosomes 
separate by the breaking apart of the thread at the point 
which corresponded to the apex of the loop. 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 208 


This is exactly the manner in which I have described the 
short, stout chromosomes of both maturation divisions as 
opening out on the spindle. The difference between the two 
types consists only in the stoutness of the chromatin rods in 
the polar mitoses, the occurrence of a second longitudinal 
spht in the first division, and consequently the absence of 
the usual longitudinal cleavage in the second division. 

Summary of Results (Text-figs. 2 and 3, pp. 218, 214). — 
The chromatin thread or threads, derived only from a portion 
of the mass of chromatin in the germinal vesicle, are found 
split longitudinally and segmented into sixteen bodies—half 
the number of the chromatin rods in the nuclei of the 
cleavage divisions. These bodies consist of two short rods 
placed side by side, and each rod is composed of two spheres 
united by a less deeply stained portion of the thread. The 
two rods are intimately associated so as to form a tetrad- 
like mass, and the whole figure is to be considered a com- 
pound chromosome. 

After a relatively long prophase each of these is resolved 
in the first polar metaphase, in such a manner that while the 
body is opened up along the original cleavage plane, another 
longitudinal cleft is effected, which is completed in the 
anaphase, and the final result is another compound chromo- 
some exactly like the original from which it sprang except in 
size. Hach of the sixteen double rods which remain in the 
ovum after the extrusion of the first polar body is resolved 
in the second polar spindle into its two elements without 
further cleavage taking place. 

In the telophase of the second division the elements which 
remain in the ovum after the extrusion of the second polar 
body elongate into rods which become bent on themselves, 
while those in the second polar body remain condensed as 
small bilobed rods. 

The maturation phases differ from the ordinary cleavage 
mitoses in respect of (a) the thickening and condensation of 
the chromatin rods, (b) the second longitudinal splitting 
which occurs in the first metakinesis, and (c) the absence of 


204 THOMAS H. BRYCE. 


longitudinal cleavage in the second metakinesis. The second 
mitosis thus merely distributes the granddaughter chromo- 
somes formed by the second longitudinal splitting in the first 
mitosis. 

There is thus no “ reducing division.” The only reduction 
which occurs is effected in the germinal vesicle, and the 
chromatin destined to form the chromosomes of the polar 
divisions is diminished in bulk merely. 

Critical Analysis of Results and Comparison with 
those of other Observers.—In describing the achromatic 
structures I have sufficiently indicated how the appearances 
I have described in my material are to be compared with 
those described by other observers. With regard to the 
chromatin elements I may now give a further analysis. 

Glancing over the whole field of research on the subject 
the first thing that strikes an observer is the remarkable 
unity of the process, even in detail, over a very large range of 
forms. ‘he figures represented for the great majority of 
both the higher plants and the Metazoa show resemblances 
so close that one cannot imagine they are produced in one 
way in one form and in another way in another form. 
Interpretation and theoretical conclusions may differ, the 
process is identical throughout. 

It has been insisted that the solution of the problem of 
reduction lies in the determination of the origin of the 
tetrads, but as these in typical form occur in a relatively 
small number of cases, it seems that the solution rather lies 
in a closer analysis of the heterotypical division, such as has 
lately been done for plants by Strasburger. 

Heterotypical division was first described by Flemming, 
in 1887, as a form of mitosis occurring in the spermatocytes 
of the salamander, and in all cases in which tetrads are not 
formed a heterotypical division in some sort ushers in the 
first maturation division with its reduced number of chromo- 
somes, and this is true of plants as well as animals. The 
distinctive features of this division as originally stated are: 

1. The spireme stage is not so compact as in other kinds 


P 


MATURATION OF OVUM IN ECHINUS ESCULENTUS, 205 


of cells. 2. Thesister threads round which the segments split 
are fused by their ends up to the metakinesis. 3. The 
monaster stage is short lived, and shows a radial arrange- 
ment only indistinctly on account of the twisted position of 
the threads. 4. The end stage of the metakinesis is very 
prolonged, and has a very special character, in consequence 
of the fusion of the ends of the threads. 5. A temporary 
and not understood second longitudinal cleavage of the 
threads appears in the anaphase. The outstanding feature 
of the heterotype was considered at first as being the in- 
complete separation of the two halves of the longitudinally 
split rods resulting in the formation of ring chromosomes, 
but the figures may assume very various forms according as 
the loop is bent, or drawn out so as to obliterate the hollow 
of it. Again, the rings or their derivatives may be attached 
to the spindle in different fashion, so that in their resolution 
different irregular figures emerge. This is shown in the 
series of diagrams given in the paper of Farmer and Moore, 
who first clearly pointed out the essential resemblance of 
the heterotype in plants and animals. The feature described 
by Flemming, namely, the second longitudinal cleavage 
found in the anaphase, seems until recently to have had very 
little significance attributed to it. 

The simplest idea of heterotypical division is that the two 
halves of the ring-shaped chromosomes are drawn out into 
U- or V-shaped daughter loops. This simple explanation will 
not explain many of the figures observed. Farmer (1895), 
in a study of the phenomena in the lilies, described a 
double cleavage taking place simultaneously in different 
planes as the compound chromosome is resolved into its 
daughter elements. In 1896, along with Moore, he gave an 
explanation of the phases, which only involved one split, the 
second being merely apparent. ‘The idea elaborated was 
that the elliptical ring was bent on itself, applied to the 
spindle at its apex, and then drawn out to the poles from 
the point of bending. ‘The original ends were ultimately 
broken across at the equator. Moore, in his work on 


206 THOMAS H. BRYCE. 


Elasmobranch spermatogenesis, adopted this explanation 
of his figures. Gregoire (1899), describing the stages 
in lilies, and Strasburger in his recent work, from a 
careful examination of the prophase in a large number of 
plant forms, absolutely decide against the idea of the bend- 
ing up of the ring. In Kchinus, where no ring is seen at 
any time, the explanation is easier and more direct, and the 
facts decide conclusively against such an interpretation. 
Strasburger, in addition to examining a large series of cases, 
reviews the results of other observers, and comes to the 
general result that all the processes can be referred to one 
type, namely, (1) As the result of the primary longitudinal 
cleavage of the chromatin thread, two rods, wavy or 
curved, are formed. These straighten and_ ultimately 
shorten down into stout rods. In shortening, various adhe- 
sions and twistings may take place, so as to form rings or 
twisted threads. (2) According to the position assumed by 
these various figures on the spindle, the character of the 
resulting metaphase figures depends. (a) If the chromo- 
somesare placed radially in the form of two rods side by side, 
they are drawn apart in the plane of the first cleavage, and 
at the same time a second slit is effected from the free end 
inwards. The result is the formation of V-shaped daughter 
chromosomes, which in the anaphase break apart at the apex 
to complete the second longitudinal cleft. (b) If placed 
tangentially the result depends on the point of attachment 
of the “ zugfasern,” but invariably as the limbs are drawn 
apart, a second longitudinal cleavage reveals itself, and two 
daughter V’s are formed. ‘The first type (a) is exactly what 
I have described in Hchinus; the second form (b) is exactly 
that described in amphibians. 

Flemming (in 1887), and Meves (1896) in Salamander, 
McGregor (1899) in Amphiuma, Kingsbury (1899) in Desmo- 
gnathus, give us conclusions in the main identical. All 
describe and figure a second longitudinal cleavage of 
the chromosomes in the dyaster stage, and this cleavage 
is preparatory to the second division, ‘The nucleus is 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 207 


partly reconstructed between the divisions, and the longi- 
tudinal cleft is lost sight of, to reappear in the second 
division either by re-establishment of the old or by a 
new longitudinal splitting. Kingsbury was able to trace 
the longitudinal cleft directly into the second division, 
owing to the fact that the nucleus is not so far recon- 
structed. More recently (July, 1900) Janssens described 
the phases in Triton, and took the further step of inter- 
preting the process in exactly the same terms as 
Gregoire in lilies. Flemming (1887), in his first paper, 
described tetrads, but regarded them as abnormal. Vom 
Rath (1893) redescribed these bodies as normal appearances, 
but his results were not maintained by Meves (1896), who 
failed to find the least evidence of tetrads in amphibian 
spermatogenesis, though he, in a short paper, described 
tetradal figures as an abnormality in the early oocytes. 
Amphibian oogenesis has been attacked by Fick (1893) 
and Born (1894), and Carnoy and Le Brun (1899). Practi- 
cally identical figures are given by all three, but the later 
authors give much more complete details, and offer a new 
interpretation. They describe the chromosomes as condens- 
ing after some intermediate phases into short rods. These 
are complex structures formed by the fusion of a considerable 
number of separate elements. ‘These short rods, or rather 
blocks, place themselves in the equatorial plane of the spindle 
in a circle round its periphery, and orient themselves so as to 
be placed with their thicker and larger ends on the spindle, 
the other end being directed outwards. Once installed in 
this position, the chromosomes go through varied movements, 
during which they submit to a double longitudinal splitting. 
The one is effected in the equatorial plane, the other in the 
axis of the spindle and perpendicular to the first. The 
equatorial division shows itself first and begins in the large 
part attached to the spindle, rising insensibly into the stalk. 
The second occurs later, and begins at the summit of the 
stalk, descending by degrees till a kind of tetrad is formed. 
Further complicated changes are described, which result in 


208 THOMAS H. BRYCE. 


double V’s originating in the wings of the “ equatorial 
crown.” These are separated and carried to the poles. A 
partial reconstruction takes place in the anaphase, but V’s 
again appear in the telophase, and the double V’s remaining 
in the ovum are separated from one another in the second 
polar spindle. The whole description shows a very compli- 
cated process, and exactly what I have found in Echinus in 
a much simpler form, because as there are no V’s or twisted 
threads to complicate the picture, I may say that only the 
initial stages described by Carnoy and Le Brun are found in 
Kchinus. 

This is the only positive evidence of the occurrence of a 
simultaneous double split of the compound chromosome of 
the heterotypical division in animals. It will be seen that 
my interpretation agrees in the main with that of Carnoy and 
Le Brun, and Janssens; and further, that the same idea has 
enabled Strasburger to reduce the heterotype in the higher 
plants to one common plan. 

Now Kchinus falls exactly into line in every essential 
respect with another considerable series of cases recently 
described. 

1. Prostheecrzeus, Klinckowstrém, 1897. 

2. Various Polyclads, Francotte, 1897. 

3. Thysanozoon, Van der Stricht, 1898. 

4. Thalassema, Griffin, 1899. 

5. Zirphea, Griffin, 1899. 

6. Tubifex and Clepsine, Gathy, 1900. 

In every one of these the figures belong to the same type, 
except that Van der Stricht and Griffin describe rings in the 
prophases. The last observer has not given details, because 
the chromosomes were too minute for analysis. 

The first four authors all explain their results according to 
the diagram given below (‘Text-fig. 1). ‘The double rods 
resulting from the compression of the ring are placed with 
the longitudinal cleft in the plane of the equator of the 
spindle, and are drawn apart by their middle points to 
form U-shaped or V-shaped figures, and the breaking 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 209 


apart of the U’s or V’s at the apex in the anaphase is 
held to be a transverse division of the original long 
chromosomes. Griffin alone says that the possibility of a 
second longitudinal cleavage is not absolutely excluded, 
but as the T-shaped figure is rare, he held to the other 
explanation. It is obvious that it is extremely unlikely 
that such exactly similar appearances should arise in different 
ways. I believe that the demonstration I have given of the 
nature of the process as it is seen in Hchinus might, if 
applied to them, reconcile all these instances with what is 
known to occur in Amphibia and the higher plants. A certain 
part of the contradiction in results would thus be removed, 


thy 


E 4. 5. 6. 
TEXT-FIG. facet Fee showing ie suczessive stages in the resolution 

of the chromosomes in the ‘‘ heterotypical divisions ” according to 

an interpretation which makes the apical break of the V atrans- 

verse cleavage. (After Wilson.) 
and instead of these cases being held to prove a reducing 
division in Weissmanu’s sense, they would, as does Hchinus, 
disprove it. 

The figures given by Linville (1899) for certain pulmonate 
Gastropods are very similar to those described in this group, 
but the origin of the figures is not completely worked out. 
He decided for a longitudinal division in the first division, 
and the elements are doubled in the anaphase, while these 
again are distributed in the second division. 

On the other hand, in Helix pomatia, Bolles Lee (1897) 
describes appearances which lead him to conclusions different 
from most other observers. He finds transverse divisions in 


VOL. 46, PART 2.—NEW SERIES. O 


210 THOMAS H. BRYCE. 


both mitoses, but no reduction in the number of chromosomes. 
He holds that there is both quantitative and qualitative 
reduction. His figures have a strong family resemblance to 
those in Hchinus, but the chromosomes are very lumpy and 
solid, and do not show the compound character of their 
prototypes which I have described. Bonin and Collin, in a 
recent paper on the ‘‘ Mitoses in the Spermatogenesis of 
Geophilus linearis (Koch),” also interpret the appear- 
ances as due to two successive transverse divisions. 

This is a very good example of the extraordinary variety 
in the manner of interpretation of closely similar appearances, 
which is evidence of the great difficulty of reaching any 
degree of certainty in cases where the chromosomes are 
small and numerous. 

I come now to another series of cases in which the so-called 
tetrads play a large part. <A figure consisting of four sepa- 
rate spherical bodies is very rare, occurring only in Ascaris 
and the Insecta, and can be explained in two different ways. 
First, in Ascaris, 1t seems, from the researches of Boveri 
(1897), Hertwig (1890), and Brauer (1893), that the primary 
chromatin rods split twice longitudinally, preparatory to two 
rapidly following divisions which succeed one another with- 
out a pause. ‘wo groups of four rods are formed, which 
condense into two tetrads. In the first maturation spindle 
two of these are linked together as dyads, and pass to the 
poles of the spindle. ‘he dyads retained in the ovum are 
resolved into monads in the second maturation division. 
Whilst there is a mass reduction there is thus no reducing 
division, no dissimilar distribution of the “ids” of the 
original spireme thread. 

Second, Henking (1891) described in Pyrrhocoris tetrad 
groups which arose in another way, by a single longi- 
tudinal and transverse cleavage of the spireme thread, and 
interpreted the first division as a reducing, the second as an 
equation division. Vom Rath (1892) followed this account by 
a description of the process in Gryllotalpa, the mole-cricket, 
in which he figured the halves of the split rods remaining 


MATURATION OF OVUM IN ECHINUS ESCULENTUs. 211 


united to form rings. The chromatin material was then con- 
densed on to four parts of the rings, which broke up to form 
typical tetrads. These were distributed as dyads in the first 
polar, and monads in the second polar spindle. Vom Rath 
held that each of these bodies represented a single chromo- 
some, and that both divisions were “reducing.” ‘There is thus 
not an “equation division,” but a dissimilar distribution of 
the “ids” of the spireme thread. In neither of these cases 
is the first maturation mitosis of the heterotypical form. 
Vom Rath’s results were partly corroborated, partly modified, 
by Wilcox (1896) for the spermatogenesis of Caloptenus 
femur-rubrum and Cicada tibicen. ‘The difference 
between the two interpretations is that Wilcox found the 
tetrad formed by conjugation of dyads, and reduction con- 
sisted, therefore, not in the unequal distribution of sister 
“ids” lying next each other in the spireme thread, but of 
any ‘‘ids”’ indifferently from any part of the spireme thread. 
Paulmier (1899), in Anasa, described the formation of the 
tetrads more in the fashion described for the Copepods by 
Rickert and Haecker, except that there is no spireme stage, 
and his first maturation division is unequal, owing to the 
manner in which the tetrad groups are placed on the spindle, 
separation taking place in the original transverse plane. ‘The 
second division is an equal division, the separation being 
effected along the original longitudinal plane of the tetrad. 
It is to be noticed that his tetrads are not composed of four 
separate elements, but are conipound bodies, the elements of 
which are condensed into a homogeneous mass. Riickert 
(1894) and Haecker (1892-3) examiued a considerable series 
of Copepods. They found the earlier stages to differ in the 
various forms, but the end result was always the same, 
namely, a condensation of the elements into tetrad groups. 
The early stages differed according as the split of the primary 
rod was complete or incomplete at one or both ends, the 
result being the formation of double rods, angles, or rings. 
Among the Copepods the case of Cyclops brevicornis 
(Haecker, 1895) requires special mention. ‘lhe splitting was 


Le, THOMAS H. BRYCE. 


here complete, and double rods were formed, which divided 
transversely, and then united again by their ends, so that the 
original tetrad figure was replaced by a pair of rods lying 
side by side. Rickert and Haecker, in explanation of their 


ce 


results, adopted the “apparent reduction” hypothesis,— that 
is, the reduction is only apparent in the first division, and 
is realised in the second, by a suppression of a second 
longitudinal splitting. It is to be particularly noticed that 
in Cyclops Haecker makes the chromosomes of the second 
division bivalent. 

It is interesting to note that, except in insects, the type, 
which Haecker calls the plant type, has claimed most of 
recently described cases. In this type, as before explained, 
the typical tetrad formation is absent, and Haecker homo- 
logises the rings described in the prophases with the tetrad 
groups by making them equivalent to the four elements of 
these bodies. Griffin has attempted to establish the same 
analogy by imagining his cross-figures as derived from a 
crushed ring, the four limbs of which represent the four 
bodies in the tetrad. 

Kchinus will not fall into any of Haecker’s types. The 
special value of the observations in Kchinus seems to me that 
the heterotypical division is present without previous ring 
formation on the one hand, and on the other distinct tetrads 
are formed which certainly submit to a second longitudinal 
division. 

I shall now endeavour to explain my results in terms of the 
tetrad, but I must first of all refer to Boveri’s and to Wilson’s 
figures of the second polar spindle in Hchinus micro- 
tuberculatus and Toxopneustes. ‘They both show obvious 
dyads in the equatorial plate, exactly as seen in Ascaris and 
Gryllotalpa, and this was the interpretation I was at first 
inclined to give to the appearances, until | convinced myself 
of the compound nature of the bodies, which at once trans- 
ferred Echinus from the group represented by the insects to 
that represented by the more recently described case of the 
Turbellarians, that is to Haecker’s plant type, though, as I 


MATURATION OF OVUM IN KCHINUS ESCULENTUS. 213 


have said above, it differs from that type in certain important 
particulars, and agrees closely with the Cyclops type. 

Kach half of the compound chromosome or tetrad is a short 
rod, showing at its ends small spherical bodies. If these 
spheres are to be interpreted as separate elements of the tetrad, 
there being absolutely no trace of a second longitudinal 

a|a 


If 
ala 


one adopted Haecker’s idea of a suppression of the last trans- 


division, I cannot represent the figure as in Ascaris — 


verse segmentation of the spireme thread, the figure could 
ala 

| 

Re 

b|b 


sphere to be the equivalent of a single chromosome. Follow- 


with perfect propriety be represented 


assuming’ each 


ing up this formula through the first and second divisions, 
it would work out as follows : 


PROPHASE. METAPHASE. ANAPHASE. TELOPHASE 


Re —=—=—=—7Y¥--_—oooo 
c) 


(ay () (ANA) 
fe) (B) 
XA) (AKA) 
C10 EEO,O 
(8) 
o) (A) (A) XA) 


I. 2. oe 3 AN 5, 


Trext-ric. 2.—Scheme of first maturation division. 


1. Double-rod prophase figure or tetrad ; first longitudinal split. 

2. Double-rod figure placed radially on spindle; opening out of daughter 
chromosomes in plane of first cleavage. Profile view. 

2a. Same in face view showing beginning of second longitudinal split. 

3. Elongation of chromosomes. 

4, Separation of daughter chromosomes. Tach is much contracted, and 
the second longitudinal split has further extended, so as to give rise to a V. 

5. Completion of second longitudinal split, converting the V's into double- 
rod figures, which are the granddaughter chromosomes. 


214 THOMAS H. BRYCE. 


PROPHASE. METAPHASE. ANAPHASE. 
(B) 
6) © 
(aya) (A) 
o€ 
(8X8) (B) Q 
8) 
He ian 3. 4. 


Text-F1G. 8.—Scheme of second maturation division, 

1. Double-rod chromosome = granddaughter chromosomes produced in 
anaphase of first division, 

2. Opening out of longitudinal split established in first division. 

3. Separation of granddaughter chromosomes. 

4. Large half-circle represents ovum; within it a granddaughter chromo- 
some elongated info a cuived iod. Smaller circles represent. polar bodies ; in 
1 a double-rod chromosome resulting from the second longitudinal split. of the 
daughter chromosomes of the first division, in 2 a single bilobed body, the 
granddaughter chromosome. 

I have given reasons for my belief that the sphere-lke 
portions of the rod can be identified through the hetero- 
typical division, and that each submits to a division in the 

a 
process ; and if we presume to call the first figure | if must 


necessarily follow that the elongated loops in the telophase of 
the second polar spindle, and the bilobed rod in the second 


a 
polar body, must also be labelled | , and the final result is that 
b 


the apparent reduction is not confined to the first division, but 
is maintained throughout,—in other words, that the chromo- 
somes are coupled in pairs, and go through their evolutions as 
linked chromosomes. Now, returning to the case of Cyclops 
brevicornis, Haecker regards each half of the double rod of 
the first metaphase as the result of a fusion of two elements 
end to end—so that each is bivalent, though they go through 
their evolutions as if they were univalent rods. Thus, if each 
half of my tetrad figure were bivalent, our results would up to 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 215 


this point agree. In the second polar metaphase I 
find double-rod figures, which are the granddaughter 
each being bivalent like the 


chromosomes lying side by side 
daughter chromosome from which it sprung. Haecker, in the 
second metaphase, finds half the number of elements seen in 
the anaphase of the first spindle, and he accounts for his 
pseudo-tetrad figures by supposing that the previous anaphase 
figures become linked together. According to his account, 
there is no second longitudinal splitting apparent, and there- 
fore the elements are daughter, not granddaughter chromo- 
somes joined together. Each of the daughter chromosomes 
is bivalent, so that when united, a complicated redistribution 
of the elements is brought about according to the formula— 


ac a b 
= or =< 
bd ed 


and separation being effected in the plane of the last trans- 
verse segmentation of the spireme thread, there is a true 
reducing division. Haecker suggests several possibilities in 
explanation of the figures, that just given being his choice; 
it does not seem very convincing to me, and the figures lack 
the inevitable sequence which is apparent im Hechinus.! 
Returning to my own results, we can only on theoretical 
grounds assume that each sphere represents a separate 
chromosome; but the idea certainly provides a_ plausible 
explanation, though, of course, such an interpretation de- 
prives the process of any significance such as Weissmann and 


1 While this paper has been passing through the press Lerat has published 
in the ‘ Anatomischer Anzeiger’ a preliminary note on the first maturation 
division in Cyclops strenuus. He does not adopt the explanations of 
Riickert and Haecker, but brings the first mitosis into Strasburger’s scheme 
of the heterotypical division, in which the longitudinal division of the 
daughter chromosomes is a fundamental character. 


216 THOMAS H, BRYCE. 


others have attributed to it. If we look upon the tetrad as 
a single chromosome longitudinally divided, then we cannot 
get beyond the statement that during maturation the chro- 
matin substance, which is retained to form the chromosomes, 
condenses into masses, which are half the number of the 
segments characteristic of the cleavage nuclei, and that these 
masses adopt a special form in the prophases of the first 
division, preparatory to the occurrence of a double longi- 
tudinal splitting. 

It would be tedious and unprofitable to attempt a complete 
recapitulation of all the cases described. In many forms, 
owing to the difficulty of obtaining an absolutely complete 
series of stages, the evidence is incomplete, while in others 
the minuteness of the chromosomes is a barrier to finer 
analysis. J must, however, refer briefly to the facts as repre- 
sented by the botanists regarding the heterotypical division. 

In the earlier days of investigation into the mitosis occur- 
ring in the pollen mother-cells of higher plants, Strasburger 
(1888) and Guignard (1891) described a longitudinal splitting 
at the beginning of each division, and in regard to reduction 
of the chromosomes they did not find that the pollen mitoses 
differed from the process in vegetative cells. 

Belajeff (1894) was the first to point out that the V-shaped 
figures of the heterotype were not due to the rods or rings 
being curved progressively in their ascent to the poles of the 
spindle. 

Farmer (1895), as I have already said, described an apparent 
double longitudinal cleavage simultaneously progressing, but in 
his paper in conjunction with Moore he elaborated the idea that 
the double rod of the prophase was produced by bending of 
the ring on itself and the fusion of the two halves. In the 
metaphase the rods were separated along the plane of fusion, 
so that only a single longitudinal cleavage was involved, and 
the separating elements were the original daughter chromo- 
somes. He held that there was a longitudinal splitting of 
the chromosomes in the second division. 

Strasburger (1895) gave an explanation involving two 


MATURATION OF OVUM IN FECHINUS ESCULENTUS. 217 


longitudinal cleavages, the second split completing itself in 
the anaphase preparatory to the second division. 

Dixon (1895) gave a somewhat different explanation, 
involving a longitudinal split taking place for the first time 
in the metaphase. 

Miss Sargant (1895) described two longitudinal splits in 
the primary chromatin thread, but adopted an idea of the 
heterotype, which was essentially similar to that of Farmer’s 
second interpretation. 

Ishikawa (1897) in Allium, and Calkins (1897) in Pteris, 
described for the first time tetrads in plants. According to 
the description of the former observer, these tetrads were 
resolved in the heterotypical division in such a fashion that 
when the daughter chromosomes broke at their apex a trans- 
verse cleavage was completed. 

Strasburger and Mottier (1897), under the influence of the 
idea of the bending of the ring and its subsequent resolution 
along the same plane, admitted the possibility that the 
separation of the V figures occurring in the prophases of the 
second division was a transverse splitting, but a few months 
later these authors thought they had discovered a longitudinal 
division during the prophase of the second division. 

Belajeff (1898) pronounced for a transverse division in 
Weissmann’s sense, but Guignard (1899), for Najas major, 
returned to the interpretation proposed by Strasburger in 
1895 for the lilies. 

Grégoire (1899) made a re-examination of the phenomena 
in the Liliaceze, and concluded for a double longitudinal 
cleavage, the daughter V’s in the first division being separated 
without further cleavage in the second. 

The difficult point in the heterotype in higher plants—a 
single longitudinal split being admitted in the prophases—is 
to account for the V-shaped forms and their varieties. 
Strasburger, after changing his ground several times, returns 
in his last pronouncement (1900) to his ideas of 1895 with 
some modifications. He finds, as I have already said, the V 
figures arise in two ways, according to the position assumed 


218 THOMAS H. BRYCE. 


by the prophase figures on the spindle, but in every case a 
second longitudinal cleavage takes place. 

An examination of his plates shows how remarkably closely 
the figures I have drawn for Echinus resemble, even in their 
details, those he has figured for the plant forms, and the 
general statement of his conclusion is as remarkably in con- 
formity with my own. He says (p. 81), “ The special peculi- 
arity of the first nuclear division of the spore and pollen 
mother cells, which follows the numerical reduction of the 
chromosomes, consists in this, that the daughter chromo- 
somes, which arise by a longitudinal splitting of the mother 
chromosomes, are inclined to a premature separation, and 
that they directly suffer a second longitudinal cleavage. 

“The second nuclear division, which follows on the redue- 
tion of the chromosomes, has only the mission of distributing 
to the granddaughter cells the granddaughter chromosomes 
already produced in the first division. 

“The two divisions differ from ordinary mitotic division 
only in the double longitudinal splitting in the first mitosis, 
and the condition thus created for the second division.” 
Again (p. 99), “The pith of the heterotypical division lies in 
the two longitudinal clefts, not in the form of the chromo- 
somes.” 

“The cause of the two cleavages of the chromosomes so 
rapidly following one another, which again conditions the 
rapid sequence of the two nuclear divisions, must le in the 
process of reduction which precedes the maturation division.” 


LITERATURE. 
AGassiz, 1864.—‘],. Agassiz’ Contrib. to Nat. Hist. ULS.A.,’ vol. v, Cam-= 
bridge, Mass. 
Batrour, F. M., 1878.—‘ Quart. Journ. Mier. Sci.,’ vol. xviii. 
Barrour, I’. M., 1880.—‘ Comparative Embryology,’ vol. i. 
VAN BENEDEN, J. P., 1841.—‘ Ann. de Se. Nat. Ser.,’ tome ii, 15, p. 126. 
VAN BenepeEN, Ep., 1875.—‘ Bull. Ac. Roy. de Belgique,’ xi. 


van BeneDEN, Ep., 1883.—‘ Arch. Biol.,’ tome iv. 


bo 
tt 
No) 


MATURATION OF OVUM IN ECHINUS ESCUI ENTUS 


Bonin, P., et Cottiy, R., 1902.—‘ Anat. Anzeig.,’ Bd. xx. 

Born, 1894.—‘ Arch. mikr. Anat.,’ Bd. xiii. 

Bovert, Tu., 1887.—‘ Zell Studien,’ Heft i, separate issue. 

Bovert, Tn., 1888.—‘ Zell Studien,’ Heft ii, separate issue. 

Bovert, Tu., 1890.—‘ Zell Studien,’ Heft ii, separate issue. 

Bovert, Tu., 1895 (1).—‘ Arch. f. Entwicklungsmekanik,’ Bd. ii. 

Bovert, Tu., 1895 (2).—‘ Verh. d. Phys.-Med. Gesell. zu Wiirzburg,’ n. F., 
XXiX. 

Bovert, T'u., 1901.—*‘ Zell Studien,’ Heft iv, separate issue. 

Birscuu, O., 1876.—‘ Abhandl. Senkenb. Naturf. Ges.,’ Bd. ix. 

Borscutt, O., 1894.—*‘ Protoplasm,’ Eng. trans., Adam and Charles: Black. 

Brtasere, 1894.—‘ Flora Erganzungsband,’ 1894. 

Berwaserr, 1898.—‘ Ber. d. deutsch. botan. Gesell.,’ Bd. xvi, Heft ii. 

BraveEr, 1893 (1)—‘ Arch. mikr. Anat.,’ Bd. xlit. 

Braver, 1893 (2).—‘ Areh. mikr. Anat.,’ Bd. xliii. 

ByxsBer, Evirn, 1899.—‘ Proc. Californian Acad. of Se.,’ vol. ii, No. 2. 

Catxkins, 1895.—‘ Journal of Morphology,’ xi. 

Carxins, 1897.—‘ Bull. Torrey Bot. Club,’ xxiv. 

Carnoy, 1885.—‘ La Cellule,’ i. 

Carnoy, 1886.—‘ La Cellule,’ iii. 

Carnoy and Lr Brun, 1899.— La Cellule,’ xii, xiv, xvi. 

Cutnot, L.—‘ Zool. Anzeiger,’ xv, No. 387. 

Derpes, 1847.—‘ Ann. Se. Nat.,’ vill, p. 83. 

Dixon, H. H., 1895.—‘ Proc. Royal Irish Acad.,’ vol. ili. 

Diner, 1895.—‘ Jen. Zeitschrift,’ xxix, 2. 

yon ErtancerR, 1898.—‘ Biol. Centralblatt,’ xvili, i. 

Farmer, 1895.—‘ Flora,’ 1895. 

Farmer and Moorr, 1896.—‘ Anat. Anzeiger,’ Bd. xi, 3. 

Fick, R., 1893.—‘ Zeitschrift f. wiss. Zool.,’ lvi. 

Fick, R., 1899.—‘ Anat. Anz.,’ Ereinzungsheft zu, Bd. xvi. 

Fretp, G. W., 1895.—‘ Journal of Morphology,’ xi. 

FLemmine, W., 1882.—‘ Arch. mikr. Anat.,’ Bd. xx. 

Fremmine, W., 1887.—‘ Arch. mikr. Anat.,’ Bd. xxix. 

Fou, 1877.—‘ Arch. des Sciences phys. et nat.,’ Genéve, 1877. 

For, 1891.—‘ Arch. des Sciences phys. et nat.,’ April, 1891. 

Francorte, P., 1897.—‘ Mém. Cours de l’Acad. Se. de Belgique,’ 1897. 

GaxpinER, E. C., 1898.—‘ Journal of Morphology,’ xvi. 


220 THOMAS H. BRYCE. 


Gatuy, 1900.—‘ La Cellule,’ tome xvii. 

Grarp, 1877.—‘ Bull. Sci. du Nord et de la Belg.,’ Bd. viii. 

Granp, 1877.—‘ Revue Scientif.,’ Bd. xx. 

Grarp, 1890.—‘ Bull. Sci. de la France et de la Belg.,’ xxii. 

Greerr, Van, 1876.—‘Sitz. der Ges. z. Beforderung de ges. Naturwiss. 
Marburg,’ No. 5. 

GREGOIRE, 1899.—‘ La Cellule,’ tome xvi. 

Grirrin, B. B., 1899.—‘ Journal of Morphology,’ xv. 

Guicnarp, 1899.—‘ Comp. rend. de |’Acad. des Sciences,’ 1899. 

Harcker, 1892 (1).—‘ Zool. Jahrbuch,’ v. 

Haecker, 1892 (2).—‘ Ber. Naturf. ges. Freiburg,’ vi. 

Haecxker, 1893.—‘ Arch. mikr. Anat.,’ xli. 

Haecker, 1895 (1).—‘ Arch. mikr. Anat.,’ xlv. 

Harcxer, 1895 (2).—‘ Annals of Botany,’ ix. 

Harcxer, 1895 (3).—‘ Arch. mikr. Anat.,’ xlvi. 

Harcker, 1899 (1).—‘ Praxis und Theorie der Zellen und Befructungslehre,’ 
Fischer, Jena. 

Haxcker, 1899 (2).—Art. “ Reifungserscheinungen,’’ Merkel and Bonnet’s 
‘ Ergebnisse,’ vill. 

Hartmann, Max, 1902.—*‘ Zool. Jahrb.,’ Abth. f. Anat. u. Ontog. d. Thiere, 
Bde xv, yA: 

Herta, V., 1893.—‘ Archiv de Biologie,’ xiii. 

Henxine, 1890-92.—‘ Zeitschrift f. wiss. Zool.,’ xlix, ], li. 

Hertwie, O., 1875.—* Morpholog. Jahrbuch,’ Bd. i. 

Herrwie, O., 1877.—‘ Morpholoe. Jahrbuch,’ Bde. ii, iii. 

Herrwie, O., 1878.—‘ Morpholog. Jahrbuch,’ Bd. iv. 

Herrwie, O., 1884.—‘ Jen. Zeitschrift, xvii. 

Hertwie, O., 1890.—‘ Arch. f. mikr. Anat.,’ Bd. xxxvi. 

Hertwie, O., 1895.—‘ The Cell, etec.,’ English translation. 

Herrwie, R., 1896.—* Festschft. fiir Gegenbauer.’ 

Hin, M. O., 1895.—‘ Quart. Journ. Mier. Sci.,’ vol. xxxviii. 

Tsnikawa, 1897.—‘ Journ. Coll. of Sci.,’ Imper. Univ., Tokio. 

Janssens, J. A., 1900.—‘ Anat. Anzeiger,’ xvii, Nos, 24, 25. 

JENSEN, 1883.—‘ Arch. de Biologie,’ tome iv. 

Kinospury, B. F., 1899.—* Zool. Bull.,’ ii, 5. 

Kuinckowstrém, 1897.—‘ Arch. f. mikr. Anat.,’ xlviii. 

Kostaneckt, 1896.—‘ Anat.,’ Heft vii, 2. 


MATURATION OF OVUM IN ECHINUS ESCULENTUS., 221 


Lez, A. Botuxs, 1897.—‘ La Cellule,’ tome xiii. 

Linvittz, 1899.—‘ Bull. Mus. Comp. Zool. Harvard Coll.,’ vol. xxxv, No. 8. 
Logs, J., 1899.—* Amer. Journal of Phys.,’ iil, 3. 

McGrecor, J. H., 1899.—* Journal of Morphology,’ xv (supplement). 
Mark, E. D., 1881.—‘ Bull. Mus. Comp. Zool. Harvard Coll.,’ vol. vi. 
Maruews, A. P., with Ep. Writson, 1895.—‘ Journal of Morphology,’ x. 
Meap, A. O., 1895.—‘ Journal of Morphology,’ x. 

Meves, F., 1896.—‘ Arch. f. mikr. Anat.,’ xlviii. 

Montcomery, I’. H., 1898.—‘ Zool. Jahrbuch,’ 1898. 

Moors, J. E. S., 1895.—‘ Quart. Journ. Micr. Anat.,’ xxxviil. 
Moore and Farmer, 1896.—‘ Anat. Anzeiger,’ Bd. xi. 

Moxean, T. H., 1893.—‘ Anat. Anzeiger,’ ix. 

Morean, I’. H., 1895.— Arch. f. Entwicklungsmek.,’ Bd. ii. 
Morean, T. H., 1896.—‘ Arch. f. Mntwicklungsmek.,’ Bd. viii. 
Mortier, D. M., 1897.—‘ Jahrb. wiss. Bot.,’ xxxi. 

Nemec, B., 1899.—‘ Jahrb. wiss. Bot.,’ xxviil. 

Ostexuout, W. J. V., 1897.—‘ Jahrb. wiss. Bot.,’ xxxi. 

Pautminr, F. C., 1899.— Journal of Morphology,’ xv. 

Picrer, 1891.—‘ Mittheil. Zool. Stat. Neapel,’ Bd. x. 

Puatner, G., 1886.—‘* Arch. f. mikr. Anat.,’ xxvi, xxvii. 

Pratner, G., 1889.—‘ Biol. Centralblatt,’ viii. 

Ras, 1885.—‘ Morpholog. Jahrbuch,’ x. 

vom Ratu, O., 1892.—‘ Arch. f. mikr. Anat.,’ xl. 

vom Ratu, O., 1893.—‘ Zeitschrift f. wiss. Zool.,’ vii. 

vom Ratu, O., 1895.—‘ Arch. f. mikr. Anat.,’ xlvi. 

Reinke, F., 1894.—‘ Arch. f. mikr. Anat.,’ xliii and xliv. 

Reinke, F., 1895.—‘ Sitz-Ber. Akad. d. Wiss.,’ Berlin, 1895. 
Rickert, J., 1892.—* Anat. Auz.,’ vii and viii. 

Rickert, J., 1893.—Merkel and Bonnet’s ‘ Ergebnisse,’ iii. 
Rickert, J., 1895.—* Arch. f. mikr. Anat.,’ xlv. 

SaBAscHNIKorr, M., 1897.—‘ Bull. Soc. Nat.,’ Moscow. 

Sareant, Wrner, 1895.—* Quart. Journ. of Mier. Science,’ 1895. 
Sareant, Eruen, 1896-7.—‘ Annals of Botany,’ vols. x and xi. 
SELENKA, 1883.—‘ Studien tiber Entwick.,’ II, Wiesbaden, 1883. 
Soporta, J., 1895.—*‘ Arch. f. mikr. Anat.,’ xlv. 

STRASBURGER, 1876,—*‘ Zellbildung und Zelltheilung,’ Jena, 1875. 
STRASBURGER, 1877.—‘ Jen. Zeitschrilt,’ xi. 


222 THOMAS H. BRYCE. 


STRASBURGER, 1894.—‘ Annals of Botany,’ vol. viii, 1894. 
STRASBURGER, 1895.—‘ Jahrb. f. wiss. Bot.,’ xxviii. 
STRASBURGER, 1897.—‘ Jahrb. f. wiss. Bot.,’ xxx and xxxi. 
STRASBURGER and Mortigsr, 1897.—‘ Ber. d. Bot. Ges.,’ xv, 6. 


SrraspurGER, 1900.—‘ Ueber Reduktionstheilung, Spindelbildung, Centro- 
somen und Cilienbildner im Pflanzenreich,’ Jena, Fischer, 1900. 


VAN DER Stricut, O., 1898.—‘ Arch. Biol.,’ xv. 

Scuuutrze, O., 1887.—‘ Zeitschrift f. wiss. Zool.,’ xlv. 

Warase, 1892-3.—‘ Journ. of Morphology,’ viti, ‘‘ Wood’s Holl Lectures.” 
Wetssmann, 1893.—‘ The Germplasm,’ English translation. 

Wueeter, W. M., 1897.—‘ Arch. de Biologie,’ xv. 

Wuirmay, C. O., 1878.—‘ Quart. Journ. Mier. Sci.,’ vol. xviii. 

Wiucox, HE. V., 1895.—‘ Bull. Mus. Comp. Zool. Harvard Coll.,’ xxvii and xxix. 
Witson, Ep. B., 1895 (1).—*‘ Atlas of Fertilization, ete.,” New York. 
Witson, Kp. B., 1895 (2).—‘ Journal of Morphology,’ xv. 

Witson, Ep. B., 1897.—‘ Science,’ vol. v, No. 114. 

Wutson, Kp. B., 1899.—‘ Journal of Morphology,’ xv (supplement). 
Witson, Ep. B., 1900.—* The Cell, ete.,’ Macmillan, 2nd edit., 1900. 
Zogsa, 1895,—‘ Anat. Anzeiger,’ xi, 10. 


EXPLANATION OF PLATES 10—12, 


Illustrating Dr. Bryce’s paper on ‘‘ Maturation of the Ovum 
in Echinus esculentus.” 


All the figures were drawn by aid of the camera lucida with Zeiss 2 mm, 
apochromatic objective of 1°40 mm. aperture (homogeneous immersion), and 
6, 8, or 12 compensating ocular. The magnification, tested by the stage 
micrometer, of Figs. 1 to 6 is 850, of the remainder ]200, except Tig. 20, 
which is magnified 1500 times. The drawings represent as realistically as 
possible what can be seen in the individual section, and, except in Fig. 18, no 
attempt has been made to make the drawing diagrammatic by incorporating 
data from two or more sections of the same ovum. A series of photo-micro- 
graphs of the stages, kindly taken for me by Dr. J. H. Teacher, was of con- 
siderable assistance for comparison with, and verification of, the drawings. 


MATURATION OF OVUM IN ECHINUS ESCULENTUS. 223 


PLATE 10. 
Fic. 1.—EHarly oocyte. 


Fie. 2.—A stage in the growth of the germinal vesicle, showing also the 
appearance of the cytoplasm when the preparation is not strongly washed out. 


Fre. 8.—Another phase in the growth of the germinal vesicle, showing also 
the appearance of the cytoplasm in a preparation from which the iron hema- 
toxylin has been strongly washed out. 


Fig. 4.—Section through invagination into germinal vesicle. 

Fic. 5.—Disappearance of nuclear membrane ; rejection of greater part of 
chromatin network; isolated mass of chromatin from which chromosomes are 
derived, 

Fic. 6.—Later stage, in which two asters have appeared. 

Kies. 7a and 7 ).—Two adjoining sections through the same germinal 
vesicle, showing the mass of chromatin from which chromosomes will be 

’ 5 
formed. 


Iie. 8.—The two definitive asters ; chromosomes in form of bilobed bodies 
or tetrads. 


Vig, 9.—Another section, showing the chromosomes within a central plate, 
surrounded by a crown of radiations fading imperceptibly into the eytoplasm ; 
one aster in view, a second lay in the adjoining section. Sometimes a third 
aster is seen at this stage. 


POSER: UL: 


Fics. 1O—14 represent the different forms assumed in various sections 
by the chromosomes before the formation of the first spindle. The dotted lines 
indicate the outline of the central plate. 


Fic. 15.—Asters now in radial position; early phase of first division 
spindle. 


Kies. 16, 17.—Two phases of metaphase of first. division. 

Fic. 18.—Same stage semi-diagrammatic. The various metaphase figures 
in two adjoining sections of the same ovum are each carefully drawn, but 
represented in the same plane. 

l'tc. 19.—Anaphase of first maturation division. 

ic. 20.—Slightly oblique section of constricted spindle in formation of 
first polar body. 

te. 21.—Various chromosomal figures seen during progress of first division, 
in profile and face views. 

Fig. 22.—The chromosomes retained in the ovum after the extension of 
the first polar body, as seen in an oblique section of central aster. 


224 THOMAS H. BRYCE, 


Fig. 23.—First polar body and centrosomes of second division before 
formation of spindle. 

Figs. 24—26.—First polar body and second polar spindle in metaphase, 
showing the chromosomes in various aspects. 


PLATE 12. 


Fics. 27 and 28.—Two stages in the anaphase of second polar division. 

Fig. 29.—Constriction of spindle in formation of second polar body. 

Vre. 30.—First polar body and second polar spindle in telophase. Section 
is oblique, so that second polar body is not seen. The point of constriction 
of the spindle shows one form in which the mid-body occurs. 

Fic. 31,—The two polar bodies and the reconstituted nucleus. 

Fic. 32.—The completely re-formed and matured nucleus, which has retired 
to near the centre of the egg. 

Fie. 33.—An early phase in the reconstruction of the nucleus; a number 
of vesicles are seen which run together to form the single vesicle seen in 
Fig. 31. The remains of the spindle and mid-body are also seen. 

Fie. 84.—Abnormality of second polar body. It is a small cell in which the 
nucleus is being reconstructed exactly as in the ovum. 


STUDIES ON THE ARACHNID ENTOSTERNITE. Zon 


Studies on the Arachnid Entosternite. 
By 


R. I. Pocock. 
With Plates 13 and 14. 


THE investigations here recorded were set on foot in the 
first instance with the purpose of settling certain contra- 
dictions as to matters of fact in the extant descriptions and 
published figures of the entosternites of various Arachnids, 
which a preliminary dip into the literature revealed. It was 
necessary to ascertain whether these discrepancies were 
attributable to a natural variability in the organ, to specific 
or generic differences between the species dissected, or to 
errors of observation on the part of the dissectors. In some 
cases, too, there was an entire lack of agreement on the part 
of observers in the interpretation of the facts recorded ; and 
the suggested homologies between the constituent parts of 
the entosternites of various species did not, on a priori 
grounds, appear to be in all cases satisfactory. I was 
anxious, moreover, to test the respective claims to recogni- 
tion of the two theories of the origin of the entosternite 
that have been put forward. 

I have made no examination of this organ in the Palpi- 
gradi, Pseudoscorpiones, Podogona, Opiliones, or Acari, 
and have nothing to add to what has already been said about 
the entosternite of these orders. 

The contents of this essay, which deals exclusively with 
the remaining existing orders, may be tabulated as follows : 

VoL. 46, PARY 2.—NEW SERIES. P 


226 R. I. POCOCK. 
I. The structure of the entosternite in the Xiphosure, 
Scorpiones, Pedipalpi, Aranez, and Solifugee. 
1. The entosternite of the Xiphosure, p. 226. 


2 45 a = Scorpiones, p. 227. 
3. 3 me 5 Pedipalpi, p. 251. 
A. as ea 5 Aranee, p. 233. 


5. The “so-called”? entosternite of the Solifuge, 


II. The comparative morphology of the entosternite, 
p. 239. 
III. Theories of the origin of the entosternite, p. 247. 


I. Srrucrure oF THE ENTOSTERNITE IN THE XIPHOSUR, 
Scorpiones, Pepipaner, ARANEX, AND SOLIFUGR. 


1. The Entosternite of the Xiphosure. 


The form and structure of the entosternite in the American 
Limulus is well known, thanks to the figures and descriptions 
of it published by Ray Lankester (5, 6) and Benham (2). 

It is a longitudinally oblong plate, with a pair of stout 
anterior bars, or cornua, forming the pharyngeal notch, two 
pairs of long and slender apophyses behind the anterior bars, 
diverging nearly at right angles from the main body of the 
plate, and a stout but short apophysis springing transversely 
from its postero-lateral angle on each side. There is also an 
irregular-shaped posterior median process, as well as a pair 
of short apophyses projecting subvertically beneath the 
latter. In the Moluccan species Tl’. gigas (= moluccanus), 
as was shown by Van der Hoeven, there is only a single 
long apophysis projecting from the lateral border in front. 
This represents the anterior of the two that are found in this 
position in X. polyphemus. ‘This peculiarity obtains also, 
I find, in the other Asiatic species, Tachypleus triden- 
tatus (=longispina) and Carcinoscorpius rotundi- 
cauda, thus confirming the opinion I have already put 
forward (‘Ann. Mag. Nat. Hist.,’ April, 1902) that these 


STUDIES ON 'THE ARACHNID ENTOSTERNITE. 227 


three forms belong to a group distinct from and more 
specialised than that of polyphemus, as is clearly shown 
by the structure of the genital operculum, etc. Also the 
pair of posterior ventral apophyses found in polyphemus 
are missing in the two young specimens of tridentatus 
and rotundicauda I have examined. In all species of 
Limulus the upper side of the entosternite is furnished 
laterally behind the middle with a short muscle-bearing 
excrescence, suggesting a suppressed or undeveloped apo- 
physis.! 


2. The Entosternite of Scorpions, 


The variations that affect the entosternite of Scorpions 
are principally correlated with the compression, antero- 
posterior or lateral as the case may be, of the exoskeletal 
metasternite. In Palamnzeus thorelli the ‘ body” of 
the entosternite consists of an irregularly transversely oblong 
plate. From its anterior angles rise the anterior cornua, 
which give off muscles to the appendages along their outer 
edge, and present a frayed or ragged appearance when cleared 
of these tissues (see fig. 20, Pl. 14). On its underside the 
plate dips down on each side of the nerve-cord, and passing 
and fusing beneath it forms a complete and rigid ventral 
ring through which the nerves pass backwards into the 
mesosoma. ‘The lower portion of this ring gives off in front 


1 Tn his paper on the anatomy of Limulus polyphemus (‘ Trans. Linn. 
Soc.,’ xxviii, 1873), Owen states (p. 469) that the entosternite of this species 
is furnished with a pair of “ sclerous processes” which diverge from “ near 
the fore-part of the dorsal surface,” and reference is made to fig. 5 on pl. 
xxXviii, which is an acknowledged copy of Van der Hoeven’s figure of the 
entosternite of L. moluccanus. Yet the original figures with which Owen’s © 
paper is illustrated are all, apparently, taken from examples of the American 
form (L. polyphemus), Hence it is difficult to account for his overlooking 
the presence of the two pairs of processes in this species. It may also be 
remarked in passing, though the lapsus is of no great moment, that the 
statement in the foot-note to p. 462 that ‘the species which he [ Van der 
Hoeven | dissected was the rapier-tailed Molucca crab (Limulus rotundi- 


cauda, Latr.) ” is an error. 


228 BR. L.) BOCOCK: 


a median process which divides into a pair of diverging 
tendinous apophyses. From the sides of the neural ring 
externally spring muscle-supporting processes. ‘The muscles 
rising from the posterior of these processes are extended 
laterally and dorsally to become attached to the sides and 
roof of the body-cavity, forming, with associated connective 
tissue, a great muscular sheet or “diaphragm”? which sepa- 
rates the cavity of the prosoma from that of the mesosoma. 
Inferiorly this partition is completed by muscles which run 
from the posterior side of the lower edge of the neural canal 
to the floor of the body-cavity. In the middle line above the 
“Pody” of the entosternite tlhe connective tissue of this 
muscular sheet is perforated by two foramina; the inferior 
gives passage to the alimentary canal, the superior to the 
aorta. 

In addition to the muscles already mentioned, three pairs 
of dorso-ventral muscles arise from the entosternite. Those 
of the posterior pair are attached to the underside of the 
tergite of the genital somite behind the diaphragm, and to 
the posterior side of the entosternite in front of it. Hence 
in their passage from above downwards they pass through 
the diaphragm. ‘he median pair extends from the aortic 
foramen in front of the diaphragm to the posterior border 
of the upper surface of the “body” of the entosternite. 
Just in front of their inferior points of attachment spring 
those of the anterior pair, which, rising vertically, meet im 
the middle line above the aorta, before attaching themselves 
to the underside of the carapace. 

As might be expected, the entosternite of this species 
agrees in all essentials with that of Palamneeus indus 
(= Buthus cyaneus) as described and figured by Lankester 
(6) and Beck (1). In a general way the entosternite of all 
Scorpions is formed on this plan. In minor particulars, 
however, there is considerable structural variation. In 
species with the metasternite antero-posteriorly compressed, 
the body of the “ entosternite’’ becomes shorter as compared 
with its length, as shown in the figure of that of Lurus 


STUDIES ON THE ARACHNID ENTOSTERNITIE. 229 


dufoureius, one of the Vejovidee. In this species the lateral 
crests which arise from the anterior cornua are better 
developed than in Palamnezeus, and the homologue of the 
solid lateral process of the latter is less solidified and rigid. 
Moreover the anterior process from the neural arch appears 
to be undeveloped (PI. 14, fig. 21). 

In Hadruroides charcasus, a member of the same 
family as Iurus, but with the sternum showing a markedly 
greater degree of antero-posterior compression, the “ body ” 
of the entosternite is relatively much shorter than in the last- 
named genus, though in other respects the entosternites of 
the two are very similar. 

The process of reduction in the size of the entosternite by 
longitudinal compression is carried to an extreme in the 
Bothriuridz (olim ‘Telegonidz), where the sternum is 
reduced to a transversely linear sclerite wedged in between 
the genital operculum and the cox of the appendages of 
the fourth pair (= second walking leg). In Bothriurus 
bonariensis (see Pl. 14, fig. 22) the portion of the body of 
the entosternite which forms the roof of the neural canal 
is reduced to a narrow transverse bar. This modification 
seems to have been accompanied by the disappearance of 
the anterior pair of dorso-ventral muscles ; those of the second 
pair pass up to the aortic foramen without fusing with the 
diaphragm. As in the genera of Vejovidee examined, 
the subneural process is apparently absent in the Both- 
riuride. The structure of the entosternite in this family 
bears out the view I have elsewhere expressed that these 
scorpions are a specialised offshoot of the Vejovide. 

In the Buthidz, which are characterised by a triangularly 
compressed sternum, the entosternite shows unmistakable 
signs of lateral compression, the “ body” being reduced to 
a longitudinal bar, from the posterior extremity of which, 
and rather between than behind those of the first pair, rise, 
in juxtaposition, the dorso-ventral muscles of the second pair. 
‘The lateral crests are well developed, as in the Vejovide, and 
the subneural arch is furnished with a median process 


230 R. I. POCOCK. 


ending in two fan-shaped apophyses similar to those of 
Palamneeus, but stouter. 

These characteristics are illustrated in the figure of the 
entosternite of Centruroides margaritatus (PI. 14, 
fig. 24), which may be taken as fairly typical of the ento- 
sternite of Buthidz in general. 

Schimkewitsch (10) gives a figure of the entosternite of 
Androctonus bicolor, which is quite unlike this plate in 
any member of the Buthidz I have examined. Presumably 
the form he names A. bicolor is the thick-tailed, dark- 
coloured species from Transcaspia which Olivier called 
crassicauda. In examples of this species the entosternite 
closely resembles that of Centruroides margaritatus 
(see Pl. 14, fig. 24), having the same narrow median longi- 
tudinal ridge and large lateral crests and the same narrow, 
nay, even narrower bar, with broad, fan-shaped apophyses 
running forwards from the subneural arch. Yet Schimke- 
witsch represents the supra-neural arch as a transversely 
oblong plate as wide in proportion to its length as in Hadru-. 
roides, furnished with lobate lateral projections, and a very 
broad subneural process with unexpanding apophyses. An 
entosternite of this description should belong to some species 
with a broad and short pentagonal entosternite. 

Speaking of the entosternite of the scorpions, Bernard 
(3) says that its points of attachment ‘to its parent 
cuticle correspond with the points of origin of the ento- 
sternite of Galeodes,’—that is to say, to the integument 
immediately above the preaxial surface of the coxa of the 
fourth prosomatic appendage, or, as he elsewhere (4) ex- 
presses it, between the third and fourth segments. In 
Palamneeus thorelli, the species examined by Bernard, I 
find that the anterior bar of the entosternite has a fibro- 
muscular attachment to the in-projecting anterior rim of the 
coxa of the fourth appendage (second leg). But I could not 
satisfy myself that there was any union with the adjacent 
integument,—certainly there was none such as to justify the 
speaking of the integument as the “parent” of this bar 


STUDIES ON TH ARACHNID ENTOSTERNITE. 231 


of the entosternite. Bernard also homologises the anterior 
bars of the scorpion’s entosternite with the second pair of 
ventral apophyses which are affixed to the sternum opposite 
the base of the third prosomatic appendage (first walking 
leg) in the spiders. Since, however, these bars in the 
scorpion give attachment diagonally to great muscles which 
supply the second and third appendages (chela and first leg), 
it seems far more likely that they represent the anterior bars 
of the entostermte of the spiders, Thelyphonus, etc., which 
are similarly continuous with the muscles supplying these 


appendages (PI. 14, fig. 20). 


9 


3. The Entosternite of the Pedipalpi (Thelypho- 
nide, Phrynide). 


In the Pedipalpi two types of entosternite are found, one 
characteristic of the Urotricha (‘helyphonidz), the other of 
the Amblypygi (Phrynidze). In the Thelyphonide the main 
portion of the plate is longitudinally oblong in shape. It is 
perforated mesially by two foramina, an anterior large and 
oval, and a posterior relatively small and circular. The two 
are separated by a transverse bridge; a similar bridge 
separates the anterior foramen from the pharyngeal notch. 
Near the edges of the upper surface of the right and left 
bars forming the external framework of the pharyngeal 
notch and of the foramina, rise five pairs of tendinous pro- 
cesses which are affixed by muscular fibres to the underside 
of the carapace. The first rises at the extremity of the 
anterior cornu, the second just in front of the anterior 
bridge. The latter apophysis is bifid and projects inwards, 
backwards, and upwards towards the central depression of 
the carapace. The others take a more lateral direction. The 
third rises close to the second and a little behind the 
anterior bridge ; the fourth just behind the middle of the 
large foramen ; the fifth on a level with the smaller foramen. 
Below the latter may be seen a bifid tendinous crest running 
downwards and outwards. Behind this point the entosternite 


232 R. I. POCOCK. 


is laterally constricted, then expands into a subcircular softer 
plate, to the ragged edge of which are fastened many 
muscles passing to the pregenital somite and to the 
appendages of the sixth pair. Sometimes at least there is, on 
the upper side of this plate, a pair of short processes which 
serially repeat apparently the longer tendons of the anterior 
part of the entosternite. From the underside in front arise 
two pairs of processes, the first passing from the anterior ex- 
tremity of the cornua to the coxe of the chelw, the second 
to the prosternum from a point on a level with the anterior 
bridge (P). 13, figs: 2,9). 

The figure and description of the entosternite of Thely- 
phonus published by Laurie (8) do not agree with this 
organ in the species examined by myself. ‘The anterior three 
pairs of dorsal processes and the two pairs of ventral pro- 
cesses, as well as the lateral crest, are omitted from the figure 
and unmentioned in the text; and I find no process pro- 
jecting from the sides of the posterior lobe such as he 
represents and describes. So, too, is the figure twice 
published by Schimkewitsch (9, 10) from a sketch by 
T'arnani and copied by Bernard (8) unlike, in certain particn- 
lars, the entosternites of the Thelyphonidz I have dissected, 
although resembling them in general form and in the 
number of the processes. For example, the bifid process 
numbered Sa.’ in my drawing is represented as rising from 
the side of the anterior bridge slightly behind the level 
whence the processes numbered 2¢g. diverge ; and the pro- 
cesses numbered 5¢g. spring further back in line with the 
posterior bridge, not just behind the middle of the larger 
foramen, as shown in my drawing. The lateral crest, too, 
was apparently unnoticed. Considering the uniformity 
in the structure of the prosoma throughout the family Thely- 
phonide, it seems hardly probable that these discrepancies 
are due to specific differences between the specimens examined. 
I find a practically complete uniformity of structure in the 
entosternite in species of the genera Thelyphonus, 
Hypoctonus, and Mastigoproctus. 


STUDIES ON THE ARACHNID ENTOSTERNITE. 233 


The entosternite of the Amblypygi is very different from 
that of the Urotricha. The pharyngeal notch is semi- 
circular and the anterior cornua large. Hach bears a pair 
of dorsally directed apophyses near the apex, also one on the 
underside, which dips down beneath the pharynx, and one 
above, at the base, which projects upwards and inwards. 
‘he body of the plate itself is wide, narrowed posteriorly, and 
solid, 1.e. without foramina. Near its lateral border on each 
side arise four apophyses which extend upwards and outwards 
to be inserted by means of muscular bundles to the under sur- 
face of the carapace. The first is very slender; the second 
and third are approximated at the base; the fourth is the 
stoutest. ‘These four spring from a common ridge beneath 
which the edge of the entosternite runs ont externally into a 
short angular crest to contribute support to the great 
appendicular muscles (Pl. 15, fig. 3). 

It would be unfair to criticise the figure of the entosternite 
of “ Phrynus”’ given by Bernard (8), because “ the prepara- 
tion was accidentally destroyed before the drawing was 
completed.” Four pairs of dorsal apophyses are represented, 
but I cannot satisfactorily homologise them with the six 
pairs shown in the figure here published (PI. 13, fig. 3). It is 
stated, moreover (op. cit., p. 20), that this plate has “ only 
one attachment to the ventral surface, and that is to the 
intersegmental membrane between the second and third pairs 
of limbs corresponding with the first pair of apodemes 
forming the entosternite in Mygale.” It is true that there 
is only one pair of ventral processes, and that they represent 
the similarly situated processes in Mygale. They are not 
attached, however, in the position Bernard states, but to the 
coxa of the second appendage, the point of their insertion 
appearing as a horny subcircular patch on the soft membrane 
below the mouth, when these appendages are pulled apart 
and examined from the front. 


4. The Entosternite in the Aranee. 


In typical members of the Aranez the entosternite closely 


234, R. I. POCOCK. 


resembles that of the Amblypygi in general form and in many 
structural details. In Ephebopus murinus and other 
mygalomorphons spiders of the family Aviculariidz it is a 
longitudinally oval imperforated plate, with large anterior 
cornua bounding the pharyngeal notch. ‘he upper side is 
furnished with four pairs of dorsally directed tendinous 
processes, arising, as in Phrynus, from a common ridge. 
Below this the edge of the plate runs outexternally into angular 
processes which afford attachment to some muscles of the 
legs. From the underside four pairs of processes pass down- 
wards to meet the sternum; the first pair arising from the 
anterior cornua and running to the bases of the appendages 
of the second pair, immediately behind the prosternum ; the 
second, third, and fourth radiating from a common median 
excrescence, behind the pharyngeal notch, and reaching the 
sternum opposite the third, fourth, and fifth appendages. 
Remnants of a similar apophysis corresponding to the sixth 
appendage, but failing to reach the sternum, are traceable 
near the posterior end of the entostermite. The clief 
difference between this entosternite and that of the Ambly- 
pygi lies in the presence of the ventral apophyses corre- 
sponding to the four posterior pairs of prosomatic append- 
ages. 

The figure of the entosternite of a Mygale given by 
Bernard (3, figs. 3 and 5), and taken from a specimen in 
the College of Surgeons’ Museum, is diagrammatic. It is to 
be noticed, however, that the ventral processes of the first 
pair are correctly represented as fused in the middle line. 
‘The entosternites of the species examined by Lankester (5, 6)! 
and Wasmann (11) agree closely with that of Ephebopus 
murinus. 

In the great majority of the Mygalomorphe the ento- 
sternite is in the main like that of Hphebopus, retaining 
the four dorsal and the four ventral apophyses, the points 
of attachment of the latter being visible on the external 


1 Tn the figure published in the second of the two works enumerated above 
the dorsal side is by an oversight represented as the ventral, and vice versa, 


STUDIES ON THE ARACHNID EN'TOSTERNITE. 230 


surface of the sternum as the so-called sigilla. In many 
-genera there is a tendency for the posterior pair to increase in 
size and shift their point of insertion from a submarginal to a 
subcentral position. ‘This is particularly noticeable in the 
so-called “trap-door”’ spiders, where the muscles and append- 
ages of the prosoma are specialised for fossorial work. 

In a few genera, e. g. Atypus, Hriodon,and Actinopus, 
all the four pairs of ventral apophyses have moved from the 
margin of the sternum towards its centre, the convergence 
reaching an extreme in Actinopus, where their points of 
attachment meet in the middle line, forming the well-known 
rosette or star-shaped sternal impression characteristic of 
this genus. ‘he union of these four apophyses on each 
side with one another and with their fellows of the opposite 
side results in the formation of a solid plate beneath the 
nerve mass, which is thus enclosed, as it were, in a basket, 
the lateral nerves to the limbs passing out through the 
spaces between the upright portion of the apophyses. 
From the middle of the anterior border of this ventral 
plate a short median process runs forward, forming the 
median unpaired lobe of the rosette-like impression on the 
outer side of the sternum. 

In Atypus the four apophyses retain their primitive dis- 
tinctness, and are arranged on the underside of the 
entosternite in the form of a circle, following the curvature 
of the pharyngeal notch. A fibrous strand runs forward 
from the anterior apophysis to the prosternum. 

A reduction in the number of ventral apophyses takes place 
in the typical genera of the Ctenizine, e. g. Pachy- 
lomerus, Stasimopus, and of the Idiopine, e. g. 
Acanthodon and Heligmomerus. In Pachylomerus 
the first and fourth apophyses persist, the second and third 
disappear. Stasimopus resembles Pachylomerus in this 
particular, but differs in that the apophyses of the anterior 
pair fuse across the middle line to form a complete collar 
round the nerve mass. In Acanthodon the first apo- 
physis is retained as in all the Mygalomorphe, and the second 


236 R. I. POCOCK. 


and third also as slender pillars with a marginal attachment 
to the sternum, but the fourth pair has vanished. In some _ 
of the genera of this group, e. g. Stasimopus and Pachy- 
lomerus, an additional apophysis is found on the dorsal side 
arising from the crest just behind the second apophysis 
from the anterior end, and directed inwards. Indications of 
a similar tendon are also observable behind the next suc- 
ceeding apophysis, and in Acanthodon similar supple- 
mentary tendons are observable behind the posterior two 
pairs of apophyses. 

In an immature specimen of Liphistius I find the four 
normal dorsal apophyses of exceptional thickness, and repre- 
sentatives of the two supernumerary apophyses that occur 
in Stasimopus, well developed. The entosternite in this 
specimen, however, perhaps on account of its immaturity, has 
no ventral apophyses extending to the sternum, although the 
muscular scars are visible at the sides of this plate. ‘This 
absence of ventral apophyses is full of interest, on account of 
its repetition in the Arachnomorphee, with which Liphistius 
has other features in common (PI. 15, fig. 7). 

On the structure of the entosternite in the Arachnomor- 
phous spiders (olim Dipneumones) my observations have not 
been far extended. A few examples of genera belonging to 
widely separated families have been examined, however, 
without the discovery of any very marked differences in the 
structure of this plate. In all there are four pairs of dorsal 
apophyses corresponding exactly to those of the Mygalo- 
morph and Liphistius, and in all, except Filistata, an 
additional pair arising, one on each side, between the normal 
second and third pairs, and directed obliquely inwards and 
backwards. This represents, no doubt, the muscle, sometimes 
with a tendinous base, which arises in the same position in 
some of the Mygalomorphe, e. g. Pachylomerus. In many 
strong-legged species, such as Lycosa, Ctenus, and Hresus, 
it is noticeable that the dorsal tendons are broad and divided 
distally into two branches. The extension of this split to the 
root of the tendon would give rise in each instance to two 


STUDIES ON THE ARACHNID ENTOSTERNITE. 237 


complete and distinct apophyses. ‘This cleavage appears to 
have taken place, probably once, possibly twice, in the case 
of the entosternite of Phrynus. 

In none of the Arachnomorphe have I found ventral 
apophyses extending to the sternum, such as are found in 
all the Mygalomorphe. The underside of the entosternite 
of Lycosa ingens, however, is furnished in its anterior 
half with a high median crest, from which five short and 
slender tendons arise on each side. These tendons appear 
to be homologous with the five inferior tendons seen in 
Kphebopus (fig. 15) and other genera of Aviculariidee. In 
the latter, however, only three pairs spring from a common 
centre, the first lying far forward; the fifth, often obsolete, 
far backwards. 

The diagrammatic transverse section of the entosternite, 
with its associated muscles, of the Aranew, figured by 
Schimkewitsch (10), shows on each side two dorso-ventral 
muscles, a lateral muscle, and two that pass to the legs, an 
external or elevator of the trochanter (second segment), and 
an internal (the depressor of the coxa), which passes ventrally 
to an entapophysis between the sternum and the base of the 


leg,—all rising from distinct apophyses. The last-named 
muscle is, I believe, the ventral portion of the tergo-sternal 
muscle, and the pair of dorsal muscles on each side repre- 


sent the tergal portion cleft to the base. 


d. The “So-called” Hntosternite of the Solifuge. 
The “so-called ” 
structure and attachment, is quite unlike the true entoster- 


entosternite of the Solifugee, both in its 


nite of other Arachnids. It consists of a pair of stout, rigid 
chitinous pillars united, with or without articulation, to the 
narrow prosternal plate, wedged in between the coxx of the 
third pair of appendages. From the prosternum it extends 
transversely along the narrow strip of integument joining the 
juxtaposed coxve of the third and fourth pairs of appendages. 
Internally the two pillars, running obliquely or subvertically 


238 R. I. POCOCEK: 


backwards from their base, converge and meet, without actual 
fusion, in the middle line, expanding beneath the alimentary 
canal to form a somewhat saddle-shaped enlargement for 
muscular attachment. This structure is supported beneath 
by a pair of slender chitinous rods which rise, one on each 
side, from a point on the ventral integument of the fourth 
somite close to the inner extremity of the tracheal stigma. 
In front of this entosternite there is a pair of fibrous 
nodules, each of which forms a centre for the attachment of 
five tendons, one passing backwards to be attached to a 
forwardly directed process from the expanded portion of the 
entosternite, a second passing vertically downwards towards 
the point of attachment of the entosternite, a third passing 
forwards, a fourth obliquely downwards and outwards, and 
the fifth downwards and inwards to the base of the rostrum. 
Bernard (3 and 4) says the entosternite of Galeodes “rises 
as a pair of infoldings of the cuticle between the third and 
fourth segments,” and his drawings represent the two pillars 
as attached some distance from the middle line to the external 
portion of the membrane between the coxe of third and 
fourth appendages, no connection with the prosternal plate 
being shown or mentioned. The only addition to be made to 
his description relates to the attachment of the entosternite 
to the prosternal plate as mentioned above. 

This attachment is of two kinds. Inthe case of a specimen 
of Solpugasagittaria the continuity of the entosternite with 
the prosternum and the intercoxal integument is_ plainly 
indicated after clarification in caustic potash and immersion 
in glycerine. I can find neither articulation nor sutural line, 
to attest its obliteration, between the two. On the contrary, 
the strengthening strands of thick chitin which traverse the 
entosternite pass without interruption into those of the pro- 
sternum, the two forming a rigid and continuous whole. A 
similar state of things is shown in the figure of the skeletal 
elements of the prosoma of Galeodes given by Schimkewitsch 
(10). 

The treatment mentioned above applied to the entosternite 


STUDIES ON THE ARACHNID ENTOSTERNITE. 239 


of Galeodes arabs revealed, however, a joint between the 
posterior extremities of the bars of the V-shaped prosternum 
and the diverging pillars of the entosternite. 

The question as to which of these two arrangements is the 
more primitive must remain unanswered until the origin of 
the Galeodean entosternite is settled. If, as Bernard main- 
tains, 1f is nothing but an exoskeletal entapophysis, the con- 
dition of unbroken continuity with the exoskeleton, as 
manifested in Solpuga, must be regarded as the original, and 
the jointed condition seen in Galeodes the derivative. If, 
on the other hand, the entosternite in this order proves to 
be an entoskeletal element like that of other Arachnids, its 
fusion with the exoskeleton must be a secondarily acquired 
characteristic, and its separation therefrom a_ primitive 
feature (Pl. 14, figs. 26, 27). 


Il. Tue Comparative Morpuonocy or THE ENTrosTERNITE. 


The evidence favouring the hypothesis that the prosoma of 
the primitive Arachnid was furnished with a broad segmented 
sternal area separating the post-oral appendages of the right 
side from those of the left needs no recapitulation. It may be 
claimed that the possession of a wide sternal area by the 
Amblypygous Pedipalpi and all the typical Aranez is an 
archaic feature; and that, in this particular at least, these 
orders are less speciaiised than the Scorpiones, Solifuge, 
Pseudo-scorpiones, and Opiliones, where the encroachment of 
the coxee of the appendages, aided in the case of the first 
and last-named orders by antero-posterior compression accom- 
panying the forward movement of the generative aperture, 
has more or less obliterated the sternal sclerites in the middle 
line. Although modified to a very appreciable extent in 
the direction of sternal suppression, the prosoma of the 
Uropygous Pedipalpi is more primitive than that of the 
four orders last named, more primitive even than that of 
the Amblypygi and Aranez in the lesser constriction of its 
posterior somite to share in the formation of the waist. 


240 R. I. POOCOCK. 


What is true of the Pedipalpi and Aranez is also true of 
the Xiphosure. In Limulus there is a relatively wide sternal 
area extending from the mouth to the posterior extremity of 
the prosoma, and strengthened by a pair of strong meta- 
sternal sclerites behind and a weakly chitinised promeso- 
sternal plate in front. 

Correlated with this primitive development of the sternal 
area we should expect to find entosternites of a more archaic 
type in the Pedipalpi, Aranez, and Xiphosurz than in the 
other orders of Arachnids. This expectation is justified by 
the unmistakably metameric arrangement of the constituent 
elements of the entosternite exhibited in these three orders. 
On almost any theory of the origin of this plate, segmental 
repetition of its parts must be postulated as a primitive 
feature. It is obvious that this characteristic is manifested 
in a far greater degree in the entosternite of the three orders 
named than in that of the Scorpiones, Pseudo-scorpiones, or 
Opiliones. 

A satisfactory settlement of the homologies of the several 
parts of the various types of entosternites already described is 
a matter of no little difficulty on account of the variation in 
number of the apophyses that arise from them. In the 
Avanee the dorsal apophyses range in number from four to 
six. In the Phrynide there are six; in the Thelyphonidee 
five, with indications of an additional pair on the posterior 
lobe of the entosternite. 

In the Araneze the apophyses in question are referable to 
two categories, those that are directed obliquely inwards 
towards the central apodeme of the carapace, and those that 
arise subvertically to be inserted serially along the area 
between its middle line and lateral border. The latter 
are invariably present, and invariably four in number on each 
side ; the former are either absent or represented by one or 
two pairs. In the two types of entosternite presented by the 
Pedipalpi there is a single pair of apophyses directed 
inwards and backwards, arising in each case close to the 
base of the anterior cornua. 


STUDIES ON THE ARACHNID ENTOSTERNITE. 241 


As explained below (p. 249), there are good reasons for sup- 
posing that the four apophyses of constant occurrence in the 
Araneze represent the tergal elements of the tergo-sternal 
muscles of the second, third, fourth, and fifth somites of the 
prosoma, those of the sixth somite being undeveloped as an 
accompaniment of the compression this somite has suffered. 

Seeing how nearly related in many particulars the Aranez 
are to the Pedipalpi, it can hardly be doubted that the four 
apophyses in question in the Aranez are homologous to the 
four that project laterally in the Thelyphonide, the fifth 
pair which is suppressed in the Spiders being retained in a 
rudimentary state by the Pedipalpi. In the case of Phrynus 
the question is complicated by the presence of an additional 
apophysis on each side, making a total of five. The posterior 
of these might be held to represent the apophysis which is 
missing in the Aranez and rudimentary in Thelyphonus; 
but since the last prosomatic somite in Phrynus is compressed 
in its sternal portion almost to the sarne extent as in the 
Aranez, and since even in helyphonus, where this somite 
retains its more primitive condition, the apophysis is scarcely 
developed, it seems more probable that this apophysis is also 
undeveloped in Phrynus, and that the fifth apophysis in 
this genus actually corresponds to the fourth in the Spiders 
and Thelyphonus. The likelihood of the truth of this view 
is enhanced by the basal juxtaposition of the third and fourth 
apophyses in Phrynus, which forcibly suggests their common 
origin from a tendon representing the third apophysis of the 
Spiders and Thely phonus secondarily subdivided into two. 
The possibility of the subdivision of these apophyses 1s clearly 
shown in the case of many Aranez, such as Ctenus and 
Lycosa where they are deeply cleft, in Acanthodon where 
those of the third and fourth pairs are double down to the reot, 
and in Thelyphonus where the apophysis rising from the 
upper side of the anterior cornu gives off a secondary branch 
towards the middle line. 

Cleavage of primary single tendons may account for the 
presence of the one or two pairs of supplementary tendons 


VOL. 46, PAkT 2.—-NEW SERIKS. Q 


242 Riiis POCOCK. 


which run obliquely inwards and backwards towards the 
centre of the carapace. ‘The constancy in position of these 
tendons in the Spiders suggests their homology throughout 
the order, and their origin from the second or the second and 
third of the larger normal apophyses. Only one such 
apophysis is developed in Thelyphonus, and this arises a 
little in front of the second marginal tendon, not behind it 
as in the Spiders. Interesting, therefore, it is to observe that 


in Phrynus 
Thelyphonus and the Spiders—the apparent homologue of 
this tendon lies a little farther back than in Thelyphonus,a 
little farther forwards than in the spiders. It is also noticeable 
as a peculiarity in Phrynus that in the third tendon, which, 


a genus in many respects intermediate between 


for reasons already given, may be regarded as reduplicated, 
the extra branch takes the same direction as its twin. One 
other small structural feature bears out the homologies here 
suggested. This is the presence in many spiders of a trans- 
verse thickening of the entosternite just in front of the fourth 
marginal apophysis. The posterior bridge of the entosternite 
in Thelyphonus has exactly the same relations. Similarly 
the posterior margin of the pharyngeal notch in the Spiders 
is generally thickened, so as to suggest its correspondence with 
the anterior bridge in Thelyphonus. As for the ventral 
apophyses, there cannot be much doubt that the pair passing 
from the anterior cornua to the basal sezments of the second 
appendages in Phrynus and Thelyphonus are the homo- 
logues of each other and of the anterior pair, which have the 
same origin and are affixed to the sternum close to the base 
of these appendages in the Mygalomorphe. So, too, must 
the second apophysis attached to the sides of the pro- 
sternum in Thelyphonus represent the second apophysis 
attached to the sternum opposite the base of the third 
appendages (first leg) in the Mygalomorphe. ‘Thus it 
is possible to bring into complete accord the apophyses 
developed on the dorsal and ventral sides of the entosternites 
of the three orders here considered. 

Scarcity of material for comparison seems to have prevented 


STUDIES ON THE ARACHNID ENTOSTERNITE. 243 


Bernard’s recognition of the homologies existing between the 
parts of this plate in different orders, homologies which are, 
at all events, fairly obvious in the case of the Pedipalpi and 
Avanee. ‘To quote his own words (4), ‘In the Spiders .. . 
the entosternite consists of four pairs of apodemes which 
meet in the centre, the second pair of which correspond with 
the entosternite of Galeodes and Scorpio, whilst the first 
pair is perhaps represented in Galeodes by the fibrous plate 
above described. In Phrynus the entosternite is difficult to 
unravel; it may perhaps represent only the first pair of 
apodemes of the spiders with secondary attachment of dorso- 
ventral muscles. In Thelyphonus we have a long 
fenestrated entosternite which may correspond with that of 
the Spiders ; the component apodemes not, however, meeting 
in a point.” 

I venture to think that the new facts and theories concerning 
the entosternite of the Pedipalpi and Araneze put forward in 
this essay will show that the homologies are by no means 
so vague and difficult to unravelas the passage quoted would 
lead one to suppose. 

Schimkewitsch (10) terms the apophysis that rises from 
the upper side of the anterior bar in Thelyphonus a dorso- 
ventral outgrowth, and those numbered Itg., 2tg., 5tg., and 4tq. 
in fio. 2, Pl. 13, as lateral outgrowths, homologising the latter 
apparently with the lateral crest, and the former with one of the 
dorsal apophyses in the Araneze. I cannot think this interpre- 
tation correct in view of what obtains in the entosternite of 
Phrynus. The Pedipalpi and the Aranez are so very closely 
related that the conclusion as to the homology between the 
apophyses of the entosternite appears to me inescapable. 

The entosternite of Limulus forcibly recalls that of 
Thelyphonus. ‘The anterior bars correspond in the two. 
Following these in L. polyphemus come the two long, 
slender apophyses running out towards the bases of the third 
and fourth appendages, and representing, I believe, the dorsal 
portions of the tergo-sternal muscles of the corresponding 
somites. A comparison between these and the second and 


244, kh. I. POOOCK. 


third pairs of lateral processes seen in Thelyphonus and 
Spiders is obvious, and is fortified by the evidence favouring 
the view that in the spiders at least these processes corre- 
spond with the second and third pairs of post-oral appendages. 
It is only necessary to homologise the muscle-bearing stump 
in Limulus with the fourth lateral process in Thelyphonus 
and the Spiders, and the strong postero-lateral apophyses in 
the entosternite of Limulus with the vestigial processes on the 
posterior lobe of the entosternite in Thelyphonus to com- 
plete the parallel. On the underside similarity between the 
entosternite of Limulus and the Arachnomorphous Spiders is 
to be found in the absence of ventral apophyses, with the 
exception that in L. polyphemus asingle abbreviated pair is 
present at the posterior end of the plate exactly as in some 
of the Mygalomorphous Spiders, e.g. Mphebopus. 

The entosternite of Scorpions has been so affected by the 
compression of the prosoma that it isnot easy to bring it into 
exact line with those of the orders hitherto considered where 
a more primitive condition persists. ‘That the anterior bars 
framing the pharyngeal notch are comparable to those of 
Limulus, the Pedipalpi, and Aranez hardly admits of a doubt 
(Lankester, 5, 6, and 7). Similarly the lateral tendinous 
crest supporting the leg-muscles, and so well developed in the 
Buthide and Vejovide, forcibly recalls that of the Arane. 
But the dorsal apophyses which form so conspicuous a feature 
in the entosternite of the Pedipalpi and the Araneze remain 
undeveloped. ‘They are represented by the two pairs of 
dorso-ventral muscles which he in front of the diaphragm, 
those of the third pair which perforate this partition being 
usually regarded as the tergo-sternal muscles of the genital 
or first somite of the mesosoma. 

‘’o which three of the four or five pairs of dorsal apophyses 
present in the Araneze and Pedipalpi do these two pairs in the 
scorpion correspond? Probably, I think, to the fourth and 
fifth pairs,—that is to say, to those that belong to the fifth and 
sixth segments of the prosoma. ‘This homology is suggested 
by their position at the posterior extremity of the prosoma, 


STUDIES ON THE ARACHNID ENTOSTERNITE. 245 


and by the fact that the somites in question have retained 
their sternal elements, and are therefore, in that particular at 
least, less modified than the somites in front, in which the 
sterna have disappeared. If, then, we suppose that the 
pharyngeal notch in the scorpions has been prolonged back- 
wards almost as far as the position of the posterior transverse 
bridge in Thelyphonus, or as the corresponding thickened 
ridge in the Aranez; that the anterior three pairs of 
apophyses have been suppressed upon the two cornua ; that the 
lateral tendinous crests represent the similar crests in the 
Aranee and Pedipalpi, those of Palamnzeus in particular 
recalling those of Thely phonus,—it is evident that the ento- 
sternites in the orders now under discussion are more alike in 
reality than appears at first sight on the surface. Furthermore, 
if we suppose that representatives of the muscles which radiate 
from the margin of the posterior lobe of the entosternite in 
Thelyphonus extended dorsally and laterally to meet 
the walls and roof of the prosomatic space, leaving a channel 
in the middle line for the transmission of the aorta and the 
alimentary canal, it is possible to bring even the diaphragm 
into harmony with parts already existing in the Thely- 
phonus. In short, strip away the apophyses lettered Ity., 
2tg., 3tg., and sa.’ in the figure of the entosternite of 
Thelyphonus, remove the anterior bridge (a. b.), and fill up the 
posterior foramen, and the homology of the remainder of 
the plate with the supra-neural portion of the entosternite in 
the Scorpions becomes obvious. The annexation by the 
entosternite of the tergo-sternal muscles of the genital somite 
probably took place before the upgrowth of the posterior flap 
shut off the cavity of the prosoma from that of the mesosoma ; 
and this consideration points to the formation of the dia- 
phragm after the suppression of the pregenital somite and 
after the forward movement of the ventral area of the genital 
somite, which brought its tergo-sternal muscle into contact 
with the entosternite. 

‘The origin of the dorsal portion of the diaphragm in this 
way from a pair of upgrowing muscular flaps embracing the 


246 R. I. POCOCK. 


alimentary canal and aorta is attested by the persistence of 
the divisional line between its right and left halves. They 
are merely united by a strip of connective tissue, perforated 
above and below by the aortic and alimentary foramina, 
which must be regarded as the sole remnants of the open 
space which originally separated the right and left portions 
of the flap from one another. 

The neural ring in the Scorpions has its counterpart in 
Actinopus, even to the development of an anterior median 
process. It may have arisen in the same way by the fibrous 
solidification and subsequent subneural fusion of the ventral 
moiety of a pair of tergo-sternal muscles. If so, the view 
that only one such pair of muscles is involved in its con- 
struction, and that that pair belonged to the sixth somite of 
the prosoma, is suggested by the absence of lateral perfora- 
tions in the sides of the neural arch for the exit of nerves to 
the appendages, and by the situation of the ring behind the 
point whence the nerves to the appendages of the sixth pair 
diverge. Equally well, however, may the sides of the canal 
have arisen from the downward growth of the lateral portion 
of the underside of the posterior portion of the entosternite. 

Lankester (6) suggests that the lateral process marked er. 
in the figure of the entosternite of Palamneus (Pl. 14, 
fig. 20) corresponds to the antero-lateral processes of the 
entosternite in Limulus. More likely, I think, is it that this 
process is the thickened and solidified representative of the 
posterior part of the crest developed (PI. 14, figs. 21, 24) in 
Centruroides and Lurus, and finds its homologue in the 
similar crests in Thelyphonus, and not in the dorsal 
apophyses, to which I believe the two processes in Limulus 
are comparable. Nor can I agree with the opinion of 
Schimkewitsch (10), that the lateral processes he finds 
on the entosternite of the Scorpion named Androctonus 
bicolor (see p. 231) are the homologues of the processes I 
have numbered Itg., 2tg., and 3tg. in Thelyphonus. 


STUDIES ON THE ARACHNID ENTOSTERNITE. 247 


II]. Taeorres oF tHE ORIGIN OF THE ENTOSTERNITE. 


As long ago as 1881 Lankester (5), when describing the 
entosternite of Limulus, said it may be regarded as an en- 
largement and interlacing of the respective tendons of the 
muscles which are attached to it. By implication a similar 
origin was predicated of the entosternites of Scorpions and 
Spiders. This opinion was accepted by Schimkewitsch (9, 
10) in the case of the entosternite of Scorpions, Spiders, and 
other air-breathing Arachnids, and for that of Limulus by 
Bernard (8), who, however, regarded it solely as a derivative 
of the ventral longitudinal muscle-bands. Bernard’s views as 
to the origin of the entosternite in the terrestrial Arachnids, 
which he considers to be in no way related to Limulus and 
its extinct allies, will be referred to later on. 

It appears to me that the evidence in favour of Lankester’s 
view of the mode of production of this plate in both groups 
of Arachnids is overwhelming; a comparative study of the 
entosternites in this class precludes, to my mind, any other 
hypothesis as to their source. 

What muscles, then, have taken the largest share in their 
formation ? 

There is reason to believe that the prosoma was originally 
supplied with five pairs of tergo-sternal (dorso-ventral) 
muscles serially repeating those of the opisthosoma, and 
passing vertically from the under surface of the carapace to 
be inserted ventrally on the sternum close to the points of 
articulation of the post-oral appendages. ‘There were also a 
dorsal and a ventral pair of longitudinal muscles traversing 
the prosoma from end to end (see Lankester, 7).!_ With the 


1 «The simple musculature in the ancestor consisted of—(1) a pair of dorsal 
longitudinal muscles passing from tergite to tergite of each successive seg- 
meut ; (2) a similar series of paired longitudinal ventral muscles ; (3) a pair 
of dorso-ventral muscles passing from tergite to sternite in each segment ; (4) 
a set of muscles moving the coxa of each limb in its socket. The confluence 
of the prostomium and the six anterior tergites to form a prosomatic carapace, 
as well as the specialisation of the six pairs of appendages of the prosoma, 
was common to the ancestors of both Limulus and Scorpio. This modi- 


248 R. I. POCOCK. 


welding together of the external skeletal elements to form a 
compact inexpansible whole, the function of these muscles as 
dilators and contractors of the prosoma would cease, leaving 
them available for other purposes if required. 

The fusion of tergites to form a carapace, accompanied no 
doubt by the partial or complete cessation of function of the 
longitudinal and vertical muscles, took place, as may be seen, in 
the T'rilobites, before the five pairs of post-oral appendages of 
the prosoma were set apart as the exclusive organs of locomo- 
tion and prehension. This specialisation, demanding an in- 
crease in the size and strength of the limbs in question, would 
be advantageously accompanied by an increase in the area for 
the attachment of the enlarged and subdivided muscles that 
control them. This area might be supplied, in the first in- 
stance, by the fibrous solidification of the central portion of 
the tergo-sternal muscles, aided perhaps by that of the 
adjacent portion of a longitudinal muscle on each side pass- 
ing from the anterior to the posterior extremity of the 
prosoma above the nerves radiating to the appendages. 

Does the structure of the most primitive types of ento- 
sternite known to us furnish justification for the opinion that 
they have originated in the manner here suggested? A 
good deal may be said, I think, in favour of an affirmative 
answer to this question. 

The points of attachment of the tergo-sternal muscles of 
the opisthosoma are generally apparent enough externally, 
both on its dorsal and ventral walls. On the prosoma they 
are usually much less apparent. In the Mygalomorphous 
spiders, however, the sternum is typically marked with four 
pairs of “scars,” the so-called “ sigilla,” one on each side 
close to the proximal end of the coxa of the second, third, 
fourth, and fifth appendages (i.e. the palpi and first three 
pairs of legs). ‘The position and nature of these scars at 


fication of form and specialisation of body regions entailed a corresponding 
modification of the muscular system. ‘The dorsal and ventral longitudinal 
muscles of the prosoma were suppressed. ‘The muscles of the prosomatic 
limbs acquired large size and became subdivided.” 


STUDIES ON THE ARACHNID ENTOSTERNITE. 249 


once suggest their correspondence with the similarly placed 
scars upon the sterna of the opisthosoma in, e.g., Phrynus, 
which admittedly indicate the ventral attachments of the 
tergo-sternal muscles. Dissection, however, shows that the 
scars on the sternum are the points of insertion, not of 
muscles, but of the tendinous processes which project down- 
wards from the lower surface of the entosternite. ‘These 
tendinous processes are, I believe, the solidified 
ventral moieties of the primitive prosomatic tergo- 
sternal muscles. Apart from other considerations, their 
muscular origin is attested by their representation in the 
Arachnomorphe (e.g. Ctenus) by muscles passing from 
the lower surface of the entosternite aud affixed by a fibrous 
strand to corresponding points on the sternum. 

The dorsal moieties of these same muscles are represented, 
I think, by the paired tendinous apophyses springing from 
the upper side of the entosternite and passing into muscular 
fibres which fasten them to the lower side of the carapace. 
There is never a sternal scar near the base of the sixth append- 
age in the Mygalomorphe, and no apophyses, either dorsal 
or ventral, of any appreciable length, corresponding to 
this limb, on the entosternite. Short chitinous ridges are 
observable, however, on this plate in the appropriate 
positions. ‘These considerations suggest the suppression of 
the fifth pair of apophyses as a concomitant, no doubt, of the 
constriction as the last somite of the prosoma. 

The acceptance of this view of the nature of the dorsal and 
ventral processes of the entosternite carries the conclusion 
that a large part of this plate in the Aranez results from the 
tendinous solidification of five pairs of tergo-sternal muscles. 

Evidence that a share in the formation of the plate has 
been contributed by at least one pair of longitudinal muscles 
is supplied by the following facts. Apart from the appen- 
dicular muscles, which originally took a transverse direction, 
both the anterior and posterior extremities of the ento- 
sternite afford support to muscles; those from the anterior 
bars passing forwards to the front wall of the prosoma, 


250 R. I. POCOCK. 


those from the posterior extremity running backwards 
into the opisthosoma in continuity with the longitudinal 
muscles of this region. ‘This is well shown in Limulus 
(Benham, 2); in Hpeira by Schimkewitsch, and in the 
specimen of Atypus figured in Pl. 14, fig. 17. Again, in the 
‘Thelyphonide, the anterior two thirds of the entosternite, 
apart from the dorso-lateral apophyses, consist of a pair 
of stout parallel longitudinal bars, united by an anterior and 
a posterior transverse bridge, the posterior lobe alone con- 
sisting of an undivided subcireular plate. 

he solidification of these muscles was no doubt brought 
about to afford a firm support for the muscles of the proso- 
matic appendages. ‘lo resist the action of these muscles, 
which would tend to draw the bars asunder, tendinous bridges 
were developed across the middle line serving to hold the 
entire structure in place. As a later development in the 
Aranee and the Amplypygi, the intervals between the bridges 
were filled in and the projecting marginal angular crests so 
characteristic of the entosternite of the Araneze were formed to 
increase the attachment-area of the leg-muscles. Latent 
potentiality for transverse fusion between the originally 
separated right and left halves of the entosternite may be 
inferred from the fusion of this nature that has actually 
taken place between the ventral apophyses in Actinopus. 

It is possible that the longitudinal muscles have not 
played so important a part im the formation of the ento- 
sternite as here suggested. The entosternite may be almost 
equally well derived on theoretical grounds from dorso- 
ventral and crural muscles alone, the anterior bars forming 
the pharyngeal notch resulting from the fusion of the tendons 
of the dorso-ventral muscle of the second somite with those of 
the appendages of the second and third somites, and the longi- 
tudinal direction of the bars being assignable to the forward 
movement of the second appendage pulling the tendons in a 
semicircle round the pharynx and brain, the notch thus 
formed becoming deeper and deeper to accommodate itself 
to the concomitant backward movement of these organs. 


STUDIES ON THE ARACHNID ENTOSTERNITE. Dil 


The opinion put forward by Lankester in 1881 (5) that the 
entosternite was formed by the enlargement and interlacing 
of muscular tendons was modified in 1885 (7) by the further 
suggestion that the prosomatic and smaller posterior (mesoso- 
matic) entosternites are merely the original subepidermic 
connective tissue of the sternal surface of the segments in 
which they occur, which has become thickened and carti- 
laginoid, and, in the case of the prosoma, has been at the 
same time floated off, as it were, from the sternal surface, 
taking up a position deeper, that is to say nearer the 
axis of the animal than that which it originally occupied. 
“And, again, in both Limulus and Scorpio the proso- 
matic entosternite or plastron represents the mid-sternal 
area of several segments fused, probably in both cases of 
all the prosomatic somites, though possibly in Scorpio the 
first segment is not included.” ‘The position of the ento- 
sternite above the nerve-cord is explained on the hypothesis 
that the detachment of the mass of connective tissue from 
the sternal surface occurred at a period when the nerve cords 
were still quite lateral in position, their union taking place 
after the flotation. 

This hypothesis assigns an immense antiquity to the 
entosternite, an antiquity dating back probably to the 
Trilobitic stage of Arachnid phylogeny, possibly earlier still. 
But supposing that the entosternite owes its origin to the 
detachment of subneural fibrous thickenings of connective 
tissue, a later phylogenetic stage can be ascribed to it by 
assuming its derivation from paired thickenings which floated 
off on each side of the united nerve-cords, and subsequently 
fused with one another both transversely and longitudinally 
to form a gate-like framework beneath the digestive tract. 
May be the fenestration of the entosternite of Thelyphonus 
is a survival of this early stage. But whether the entoster- 
nite had its origin in subneural thickenings of this nature, 
and, if so, the manner and purpose of their assumption of 
their present position, or whether it was derived from the 
fibrous solidification of muscular and connective tissue in the 


2a R. I. POCOCK. 


mid-regiou of the prosoma, as I am inclined to believe, are 
questions which must for the present be left unsettled. The 
evidence we possess that at least the dorsal and ventral 
processes of the entosternite in the spiders are modified 
muscular tendons seems to make it unnecessary to look else- 
where for the source of the formation of the entire plate. 

This conception of the origin of the entosternite from 
muscular and connective tissue differs entirely from that held 
by Bernard (8, 4), who would derive this plate in all 
terrestrial Arachnids from integumental apodemes and 
segmental constrictions. It may be inferred from what he 
says about the entosternite of Mygale that he regards its 
four pairs of dorsal and ventral processes as the remains 
of integumental infoldings marking the line of union of 
the originally separated tergal and sternal elements of the 
prosoma.’ He adds, “The shape of the whole fused mass 
has been, no doubt, much altered by the action of muscles, 
but its essential nature as a fusion of metamerically recurrent 
apodemes cannot be mistaken” (8, p. 20). In his paper on the 
morphology of the Galeodide (4, p. 327) he says that in the 
Pedipalpi and Araneee ‘four pairs of dorso-ventral muscles 
have been retained, more or less modified, as the dorsal 
attachments of the entosternite, and are now largely fibrous ; 
they suspend the entosternite and separate the alimentary 
diverticula in the typical manner.” Hence may be inferred 
the admission that part at all events of the dorsal processes 
of the entosternite have been derived from dorso-ventral 
muscles. If part, why not the whole of the dorsal process ? 
And if the dorsal process represents the part of the muscle 
above the entosternite, what reason is there for refusing to 

' Bernard states that the original distinctness of the terga of the carapace 
in the Arance is shown by the furrows on its dorsal surface. ‘These furrows 
are in reality the external indications of the radial arrangement of the great 
dorsal appendicular muscles, and mark the lines of attachment of the muscles 
rising dorsally between them from the entosternite. If the grooves indicated 
the union of tergal plates, such plates should be more clearly defined in the 
embryo, but so far as I am aware the carapace of spiders at no period of its 
development slows division into separated tergal plates. 


STUDIES ON THE ARACHNID ENTOSTERNITR. 20a 


regard the ventral processes as the representatives of the 
ventral moieties of these same muscles ? 

Bernard’s hypothesis involves the assumption of a degree 
of dislocation and rearrangement of muscles and integu- 
mental apophyses for which it is difficult to find justification. 
I can discover no evidence that the sternal scars of the 
Mygalomorphe and the ventral processes of the entosternite 
which rise from them are the remains of integumental dis- 
sepiments. If this were the case we should expect to find an 
intimate unseverable union between the sternum and the pro- 
cesses in question. No such union exists. ‘The expanded ex- 
tremities of the processes may be readily detached, the extent 
of their union with the sternum being quite compatible with 
the theory of their muscular origin, but hardly reconcil- 
able with that of their derivation from ectodermic ingrowths. 

The basis of Bernard’s hypothesis is to be found, firstly, 
in the structure and relations of the so-called entosternite of 
the Solifuge, which is shown by its histology and union 
with the integument to be an ectodermic entapophysis ; 
secondly, in the assumption that this skeletal piece is the 
morphological equivalent of the entosternites of other 
Arachnids; thirdly, in the conception that the Solifuge 
retain a more archaic type of prosoma than that of the other 
orders of this class. 

Assuming the truth of the propositions here stated or 
implied, the conclusion Bernard draws as to the ectodermic 
origin of the entosternite in the Spiders, etc., necessarily 
follows. But from what is known of the structure and 
development of the entosternite in the two orders there is 
little doubt that the first proposition is true, and that the 
second is untrue. As for the third, it is a matter of opinion 
depending upon the standpoint from which the morphology 
of the Arachnida is regarded. 

The available evidence is, in my opinion, decidedly in 
favour of the view that the “ entosternite’”’ of the Solifuge 
must be regarded as a post-oral entosclerite comparable to 
the crescentic pre-oral entosclerite of the Scorpions. But 


254 R. I. POCOCK. 


if there are any who see in it the homologue of the 
Scorpion’s entosternite, they will remember that chitin has 
been shown (Lankester, p. 6) to be present in the ento- 
sternites of Scorpio, Mygale, and Limulus, and will realise 
the possibility of the formation of the rigid and horny 
Galeodean entostermite by increased development of its 
chitin, followed or accompanied by fusion with the exo- 
skeleton of the second post-oral somite. 

Briefly, then, of the three suppositions that may be enter- 
tained with regard to the “entosternite” of the Solifuge, 
each points to its being a specialised, not a primitive 
structure. (1) If it is an entosclerite, as Bernard and 
Schimkewitsch maintain, it is not the homologue of the 
entosternite of other Arachnids, which is shown by its 
morphology and development to be an entochondrite, pro- 
duced by the condensation of connective tissue and the fusion 
of muscular fibres and tendons. In this case it has fune- 
tionally replaced the true entosternite, and is a_ recent 
specialisation, not a primitive structure. (2) If it is an 
endochondrite and the homologue of the entosternite of other 
Arachnids, its structural similarity to, and fusion with, the 
exoskeleton also attest igh specialisation. (5) If it has 
resulted from the union of the true entosternite with a pair 
of exoskeletal ingrowths—if, say, the expanded portion 
supporting the alimentary canal corresponds with the true 
entosternite, and the pillars diverging therefrom to the exo- 
skeletal elements,—the absence of all trace of union between 
the two, the complete continuity of their tissues, again 
indicate great specialisation. 

The evidence in favour of the truth of the first supposition 
is almost strong enough to enforce its unquestioned accept- 
ance. But whichever of the three prove consonant with fact 
and be ultimately adopted, the Solifugze must be regarded 
as the most specialised type of Arachnid known, so far as the 
a conclusion which is 


organ under discussion is concerned 
perfectly in accord with many, nay most, of the structural 
features of this order. 


STUDIES ON THE ARACHNID ENTOSTERNITE. A355) 


The one hypothesis of all others which, in my opinion, 
has least in its favour is that in the Solifugee we find a 
primitive type of prosoma and a primitive type of entoster- 
nite clearly attesting the exoskeletal origin of this plate 
in all orders of Arachnida; and the conclusions deduced 
from these disputable premisses that the true entosternites 
have been derived from chitinous integumental apodemes is 
contradicted by their structure in the adult and their meso- 
blastic origin in the embryo. 

Lastly, according to Bernard (8), the ‘‘ diaphragm’ of 
Scorpio, “like that of Galeodes, is the homologue of the 
great constriction between the sixth and seventh segments 
forming the ‘waist’ of other Arachnids, . . . . a diaphragm or 
waist being typical of Arachnids.” It is not at all clear how 
a partition lke the Scorpion’s diaphragm, composed of mus- 
cular and connective tissue and without exoskeletal elements, 
can be the homologue of an exoskeletal constriction. Ana- 
logous structures, structures with the same physiological 
significance, they no doubt are; but homologous, surely 
not. 

According to Bernard’s theory of waist formation, I pre- 
sume the condition initiated in Thelyphonus and eul- 
minating in the Spiders, a condition which results from the 
constriction and reduction of the pregenital somite, preceded 
the condition now met with in the Scorpions and Solifuge. 
In that case the “diaphragm” in these two orders must 
represent the pregenital somite, insunk and overgrown, plus 
the dorsal and lateral arthrodial membranes which connected 
this somite withthe prosoma in front and the mesosoma behind. 
This double partition then became united into one, the dorsal 
area of the pregenital somite disappeared, setting free the 
aorta and the alimentary tract, which were previously confined 
with the nerve-cord in anarrow canal, and enabling them to rise 
and take up respectively their original positions in the dorsal 
and central regions of the body-cavity, cleaving the partition 
as they rose. Only by entertaining some such conception as 
this is it possible to hold that the “ waist ” of the Spiders and 


256 R. I. POCOCK. 


Pedipalpi is the homologue of the “diaphragm” in the 
Scorpiones and Solifugee. 

This theory of the formation of the diaphragm seems to me 
scarcely more plausible if the pregenital somite, which was 
not recognised as such when Bernard wrote, is left out of 
consideration ;! and the following quotation shows that I have 
given no exaggerated rendering of liis hypothesis. He says (4), 
‘Between the sixth and seventh segments . . . there is in the 
Galeodide .. . a strong intersegmental constriction. Inter- 
nally this constriction has given rise to a very striking 
diaphragm. It forms a very complete wall between the 
interior of the cephalothorax and that of the abdomen, and 
is pierced by the dorsal vessel, the alimentary canal, the 
nerve-cords, and the tracheee. Close examination shows that 
the diaphragm is due to a strong indrawing of the interseg- 
mental membrane between the above-mentioned segments, so 
that it is composed partly of a chitinous infolding and partly 
of muscle-bands. It is clear thatif the opposite two internal 
faces of such a deep segmental constriction fuse together, they 
form a diaphragm ; if they remain unfused they form a waist. 
In the Galeodidze we seem to have an unspecialised arrange- 
ment, the intersegmental infolding being fused only in its 
deeper parts, forming the diaphragm, while the outer parts 
of the fold remain open, making an approach to a waist.” 

Such an infolding, in its unspecialised state, must have 
disturbed the position of the dorsal blood-vessel, forcing it 
down towards the central axis of the body on to the alimen- 
tary canal; but, as a matter of fact, the blood-vessel and 
alimentary canal show no trace of any disturbance of their 
primitive positions, the former perforating the diaphragm 
high up beneath the dorsal integument, the latter traversing 
its centre, while below, in line, is seen the canal for the nerve- 


1 Bernard has pointed out to me that his figure of the section of the 
“waist” of a Spider published on pl. xxxiii, fig. 6, of his paper ‘‘ On the 
Morphology of the Galeodide,” shows the presence of a pair of dorso- 
ventral muscles, and thus confirms the view that the waist is a genuine 
somite. 


STUDIES ON THE ARACHNID ENTOSTERNITE. PIS 


cord, to say nothing of the large tracheal apertures on each 
side. Hence if the diaphragm originated from an integu- 
mental infolding it has secondarily encircled the three median 
organs above mentioned, and isa highly specialised structure. 
But, as a matter of fact, in the “ diaphragm” of Galeodes 
[ can find no evidence of such a derivation. It appears to 
be formed of muscular and connective tissue like that of the 
Scorpions, and to have had an internal origin quite apart 
from the integument. The infolding of the integument 
Bernard speaks of appears to be quite superficial, and to 
occur only at the periphery of the diaphragm. 

Whether this diaphragm has been developed independently 
of the diaphragm of Scorpions, to which it is similar in its 
structure and position, or whether the two are to be regarded 
as a heritage from a common ancestor, are matters of quite 
another kind. The absence of such an organ in the Pedi- 
palpi, Araneze, Pseudo-scorpiones, Opiliones, and Acari, 
coupled with the wide structural divergences between the 
Scorpions and Solifuge, points to the independence of the 
origin of the diaphragm in these two orders in response to 
similar physiological needs. i 


Notrt.—I have elsewhere suggested (see ‘Ann. Mag. Nat. 
Hist.,’ 1893) that the value of the structural characters of the 
orders of terrestrial Arachnida may be expressed by grouping 
them into four divisions of superordinal rank: the first to 
contain the Scorpions; the second the Pedipalpi and the 
Aranez; the third the Solifuge; the fourth the Pseudo- 
scorpiones, Opiliones, and Acari. A study of the entosternites 
confirms this classification in a remarkable and unexpected 
degree, especially as regards the isolation of the Scorpiones 
and Solifugee, and the association of the Aranez with the 


Pedipalpi. 


BIBLIOGRAPHY. 


1. Beck, K. J—‘ Description of the Muscular and Endoskeletal Systems of 
Scorpio,” ‘ Trans. Zool. Soe. London,’ xi, 1885. 


voL. 46, PART 2.—NEW SERIES. R 


258 Bi: Ts "POCOCK. 


2. Bennam, W. B.—< Description of the Muscular and Endoskeletal Systems 
of Limulus,” ‘ Trans. Zool. Soe. London,’ xi, 1885. 
8. Bernarp, H. M.—‘‘The Endosternite of Scorpio, etc.,’ ‘Ann. Mag. 
Nat. Hist.,’ xiii, p. 18, etc., 1894. 
4. Bernarp, H. M.—‘ The Comparative Morphology of the Galeodidee,” 
‘Trans. Linn. Soe. Zool.,’ vi, 1896. 
5. Lanxester, 2. R.—* Limulus in Arachnid,” ‘ Quart. Journ. Micro. Soe.,’ 
xxi, 188. 
6. Lanxester, KE. R.—‘‘On the Skeletotrophic Tissues, ete., of Limulus, 
Scorpio, and Mygale,” ‘ Quart, Journ. Micro. Soe.,’ xxiv, 1884. 
7. Lanxester, h. R.—“‘ Comparison of the Muscular and Endoskeletal 
Systems of Limulus and Scorpio,” ‘Trans. Zool. Soe. London,’ xi, 
1885. 
8. Laurte, M.—“On the Morphology of the Pedipalpi,” ‘Journ. Linn. 
Soe. Zool.,’ xxiv, 1894. 
9. ScurmmKewitscn, W.—“ Sur la Structure . . . . del’ Endosternite, ete.,” 
‘Zool. Anz.,’ 1893. 
10. ScuimKewitscu, W.—“ Ueber Bau und Entwicklung des Entosternites 
der Arachniden,” ‘ Zool. Jahrb. Anat.,’ viii, 1894. 
11. Wasmanny, —.—‘ Anatomie der Spinnen,” ‘ Abh. Ver. Hamburg,’ 1846. 


EXPLANATION OF PLATES 13 & 14, 


Illustrating Mr. Pocock’s paper, “ Studies on the Arachnid 
Entosternite.” 


With the exception of Fig. 1 the figures have been drawn by the author 
from entosternites dissected and preserved with others, not here figured, in 
the Natural History Museum, where they are available for examination. In 
the figures representing the dorsal and ventral views of the entosternite the 
anterior extremity is uppermost; in those showing the lateral surface the 
anterior extremity lies to the right. In the case of the so-called entosternite 
of the Solifugee, however, the distal or anterior extremity is directed down- 
wards. ‘ 


LETTERING COMMON TO MOST OF THE FIGURES, AND INDICATING 
SuccEstep Homo.nocies. 


A.P. Anterior process or cornu of right side forming part of the framework 
of the pharyngeal notch (PA. N.).  1tg., 2ty., 3tg.. 4tg., 54g. Dorsal processes 


STUDIES ON THE ARACHNID ENTOSTERNITE. 259 


or muscles representing the dorsal moieties of the tergo-sternai muscles of the 
somites bearing the first, second, third, fourth, and fifth post-oral appendages. 
Ist., Qst., 3s¢., 4s¢. Ventral processes representing the ventral moieties of the 
same muscles. Cr. Lateral crest developed mainly to support some of the 
muscles of the appendages. 


PLATE 13. 


Fie. 1.—Dorsal view of entosternite of the American Limulus (X. poly- 
phemus), showing the two pairs of long slender antero-lateral processes 
(2¢g. and 3¢y.), the stunted muscle-bearing process (4/g.), and the large pos- 
tero-lateral process (5/g.). (After Lankester.) 


Fic. 2.—Dorsal view of the entosternite of a Thelyphonid (Mastigo- 
proctus giganteus) showing the dorsal processes (1/g. to 5¢y.), which are 
considered to represent the dorsal moieties of the tergo-sternal muscles of 
the second to the sixth somites of the prosoma; sa.', supernumerary or addi- 
tional apophysis, which has perhaps arisen from the fission of the apophysis 
numbered 2¢g.; a.d., anterior bridge, and p. d., posterior bridge, bounding 
the large foramen in front and behind; the smaller foramen is shown behind 
the posterior bridge; /.4., lateral bar, showing perhaps the origin of this 
portion of the entosternite from a great longitudinal muscle, or from paired 
subepidermal ventral entochondrites ; P. F., posterior plate, with frayed edge 
indicating the attachment of radiating muscles, the suggested homologue of 
the dorsal portion of the “diaphragm” in the Scorpions (see Pl. 14, figs. 21, 
22); er., lateral crest. 


Fie. 3.—Dorsal view of the entosternite of a Phrynid (Damon John- 
stoni), showing the duplication of the apophyses numbered 1¢g.and 3/g.; er., 
lateral crest, to which leg-muscles are attached; Isf., anterior ventral 
apophysis, the suggested representative of the sternal moiety of the tergo- 
sternal muscle of the second segment of the prosoma. 


Fic. 4.— Dorsal view of entosternite of one of the Mygalomorphe (Ephe- 
bopus murinus), with same lettering as in the last figure, showing the 
absence of supernumerary apophyses and the presence of a thickened ridge 
(p. b.), the suggested homologue of the posterior bridge in the entosternite of 
Thelyphonus. 


Fic. 5.—Dorsal view of the entosternite of Stasimopus Schénlandi, a 
Mygalomorphous spider of the family Ctenizide, showing the presence of two 
supernumerary apoplhyses (sa.! and sa.*) and the fusion of the anterior ventral 
apophyses (1s¢.) to form a neural collar. 

Fig. 6.—Dorsal view of entosternite of Actinopus Wallacei, a Mygalo- 
morphous spider of the family Actinopodide, showing the fusion of the four 
ventral apophyses to form a subneural arch with perforated walls for the exit 


260 Ri Ts) POCOCK: 


of the appendicular nerves ; 1s¢., anterior ventral apophysis; m., median pro- 
cess from the floor of the subneural arch. 


Fie. 7.—Dorsal view of entosternite of a young Liphistius, sp. ?, show- 
ing the thick dorsal apophyses (1ég. to 4¢g.), the two pairs of supernumerary 
apophyses (sa.! and sa.?), and the absence of anterior ventral apophyses. 


Fic. 8.—Dorsal view of the entosternite of Nephila femoralis, an 
Arachnomorphous spider of the family Argiopide, with the single pair of 
supernumerary apophyses (sa.'), and without anterior ventral apophiyses, as in 
Liphistius. 


Fic. 9.—Lateral view of entosternite of a Thelyphonid (Mastigoproctus 
giganteus) Ist. and Qs/., first and second ventral apophyses ; other lettering 
as in Fig, 2. 

Fic. 10.—Ventral view of entosternite of Pachylomerus nidulans, a 
Mygalomorphous spider of the family Ctenizidee, showing the persistence of 
the first (1s¢.) and fourth (4s¢.) ventral apophyses, and the disappearance of 
the second and third. 


Fic. 11.—Lateral view of the same, showing the four dorsal apophyses 
(lig. to 4¢g.) ; p. m., posterior median crest. 


Fic. 12.—Lateral view of entosternite of Actinopus Wallacei (see 
Fig. 6), showing the four dorsal (léy. to 4¢g.) and four ventral (1s¢. to 
4st.) apophyses, the latter meeting in the middle line beneath the nervous 
mass, leaving lateral spaces for the exit of nerves; m., median process from 
subneural arch. 

Fic. 13.—Ventral view of entosternite of Ephebopus murinus (see 
Fig. 4), showing the persistence of the four ventral apophyses (1sé. to 4s¢.), 
with indications of the fifth (5s¢.). 


PLATE 14. 


Fie. 14.—Lateral view of the entosternite of Ephebopus murinus, 
showing the four dorsal (1¢g. to 44g.) and the four sternal (1s¢. to 4s¢.) apo- 
physes. This figure clearly indicates the correspondence between the dorsal 
and ventral apophyses, which suggests their origin from tergo-sternal muscles 
(see Pl. 13, figs. 4 and 13). 

Fic. 15.—Ventral view of entosternite of Acanthodon opifex, a 
Mygalomorphous spider of the sub-family Idiopinse, showing the persistence 
of the first, second, and third pairs of apophyses (1s¢. to 3s¢.), those of the 
fourth pair (4s¢.) being rudimentary. 

Fie. 16.—Lateral view of the same, showing the duplication of the third 
and fourth dorsal apophyses, and the absence of the fourth ventral apophysis. 


Fic. 17.—Mygalomorphous spider of the genus Atypus, dissected from the 


STUDIES ON THE ARACHNID ENTOSTERNITE. 261 


dorsal side with entosternite in situ to show the muscles radiating to the 
post-oral appendages II to VI, the anterior longitudinal muscles which pass 
from an entosclerite above the rostrum to the extremities of the anterior 
cornua, and the longitudinal muscles which pass backward into the pregenital 
somite. 

Fic. 18.—Lateral view of entosternite of the same, with the four dorsal (1ég. 
to 4¢7.) and four ventral (1s¢. to 4s/.) apophyses, and the tendon running 
forwards from the first ventral apophysis to the prosternum. 


Fic. 19.—Ventral view of the same, showing the arrangement of the four 
ventral apophyses in a circle round the pharyngeal notch. 


¥ic. 20.—Hntosternite of Palamnaeus Thorelli, withthe posterior flap or 
diaphragm removed, showing the fibro-muscular attachment of the anterior 
cornu to the coxa (ex.) of the fourth appendage ; cr., lateral processes repre- 
senting the muscle-bearing crest seen in Thelyphonus; 4¢g., 5¢y., anterior 
and posterior pair of dorso-ventral muscles, the suggested homologues of the 
apophyses numbered 4é7. and 5/7. in the entosternite of Thelyphonus 
(PI. 13, fig. 2), and representing in all probability the tergo-sternal muscles of 
the fifth and sixth somites of the prosoma; V.P., median process of subneural 
arch dividing into a pair of apophyses ; Da., median dorsal portion of ‘ body ” 
of entosternite forming the roof of the neural canal. 

Fic. 21.—Entosternite of one of the Vejovide (Lurus dufoureius), with 
posterior flap (P. #.) or diaphragm attached, to show its correspondence in 
origin with the lateral crest (Cr.), and its median perforations for the gut (4/.C.) 
and aorta (doc.), between which lies the channel for the lodgment of the aorta, 
formed by the dorso-ventral muscle of the second pair (5/y.) and a strip of 
conuective tissue, which binds the right and left portions of the diaphragm 
together. 


Fic. 22.—Entosternite of Bothriurus bonariensis, showing the re- 
duction of the median dorsal portion of the ‘‘ body,” forming the roof of the 
neural canal, to a narrow transverse bar (Da.). 

Fic. 23.—Anterior view of entosternite of the same, showing the neural 
canal (WV. C.), dorsal arch (Da.), and subneural arch (Sq.). 


Fic. 24.—Entosternite of Centruroides margaritatus, with most of the 
diaphragm removed, showing the lateral compression of the “ body ” or dorsal 
arch (Da.) of the neural canal, the juxtaposition of the second pair of dorso- 
ventral muscles (5ég.), and the tips of the apophyses of the median process 
of the subneural arch (V.P.). 

Fig. 25.—Entosternite of the same, with its dorsal portion removed to show 
the cut ends of the lateral walls (Za.) of the neural canal, the floor (Sz.) of 
the latter and the median process terminating in two expanded fan-shaped 
apophyses (V.P.). 

Fic. 26.—The so-called entosternite of Solpuga sagittaria, cleaned 


262 R. I. POCOCK. 


with caustic potash, to show the continuity of its supporting chitinous strands 
with those of the prosternal plate (prs.), which is wedged in between the 
cox of the appendages of the third pair and its attachment to the sternal 
membrane (m.), between the third and fourth somites of the prosoma; Js., left 
bar of the entosternite, which expands at the free extremity to form with its 
fellow of the opposite side a supporting channel (4/.) for the alimentary canal. 

Vic. 27.—Lower extremity of the two pillars of the entosternite of 
Galeodes arabs, to show this articulation («.) with the prosternal plate 
(prs.), its union with the sternal membrane (m.); cox., coxa of appendage of 
third pair. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 263 


On the Morphology of the Cheilostomata. 
By 


Sidney F. Harmer, Sc.D., F.R.S., 


Fellow of King’s College, Cambridge; Superintendent of the University 
Museum of Zoology. 


With Plates 15—18. 


THE observations on which the present paper is based were 
commenced with the examination of a Flustra-like Cheilo- 
stome, found at Port Jackson by Mr. T. Whitelegge, who 
sent it to me for description, believing it to belong to a new 
genus. Although exhibiting several remarkable features, I 
think it may be placed in Huthyris, Hincks, and I propose 
for it the name KH. clathrata, in allusion to the bars which 
support its “frontal wall.’ The species possesses a large 
“compensation-sac,” a name which is due to Jullien 
(1888, 1), although the structure had to some extent been 
described by Busk and Waters. Jullien’s results have usually 
been discredited by later writers. he study of a number 
of Cheilostomatous genera has, however, not only led 
me to confirm the accuracy of Jullien’s statements, but 
has resulted in various conclusions which I believe to be of 
importance for the proper understanding of the Cheilostomata. 
A summary of my principal results has been communicated 
to the Cambridge Philosophical Society (1901). 

The present paper is divided into the following parts : 

I. Methods employed. 
Il. List of the species specially studied. 


264. SIDNEY F. HARMER. 


III. Description of Euthyris clathrata, n. sp., of 
EK. obtecta, Hincks, and of Huthyroides 
episcopalis, n. gen. 

IV. The morphology of the compensation-sac and of 
the operculum. 

A. Flustrina. 
B. Cribrilinide. 
c. Lepralioid genera. 
p. Microporelloid genera. 
&. Microporoid genera. 
V. The primary zocecium or ancestrula. 
V1. Classification of the Cheilostomata. 
VII. Summary. 
VIII. Literature. 
IX. Explanation of Plates. 


I. MrerHops EMPLOYED. 


‘he choice of species for investigation has been primarily 
dependent on the material available for the purpose in the 
collection of the University Museum of Zoology at Cambridge. 
Spirit material has almost exclusively been used, and I have 
in the main selected species in which the calcareous matter 
was not developed to so great an extent as to destroy the 
transparency of the object. The growing edges of healthy 
colonies have furnished the most satisfactory results. The 
material was in almost all cases stained, without decalcifica- 
tion, in diluted borax carmine for a prolonged period (five to 
seven days, or even more). After being placed in absolute 
alcohol containing picric acid, the fragments were mounted 
whole in Canada balsam. 

All my more recent preparations have been mounted by a 
method to which my attention was called by Mr. H. D. 
Geldart, of Norwich. This consists in transferring the speci- 
mens directly from absolute alcohol to a solution of dried 
Canada balsam in absolute alcohol. In preparing this solu- 
tion, the milky mixture which is at first produced becomes a 


THE MORPHOLOGY OF TH CHEILOSTOMATA. 265 


complete solution in the course of a few days, particularly if 
the bottle be left on the top of a water-bath at about 60° C. 
The cloudiness which appears on first mounting a preparation 
soon disappears from a slide left on the water-bath. his 
method cannot be too strongly recommended for certain 
Polyzoa, particularly for the more delicate Ctenostome 
genera, which are distorted almost beyond recognition by 
the use of oil of cloves. 

In the case of some of the more densely calcified species I 
have found great advantages in the use of the method 
recommended in my paper on Steganoporella (1900, 
p- 240) of removing the basal wall, by means of a scalpel, 
from stained colonies (not decalcified) embedded in paraffin. 
I have also made use of thin slices, cut by hand, of uncal- 
cified material embedded in paraffin. I am convinced of the 
great importance of studying the Cheilostomata in uncalcified 
Canada balsam preparations. Most of my slides have been 
examined with a binocular microscope and a quarter-inch 
objective. 


Il. List oF THE SPECIES SPECIALLY STUDIED. 


1. Euthyris clathrata, n. sp. : . Port Jackson. 
2. Fe obtecta, Hincks : . Torres Straits. 
3. Huthyroides  episcopalis, Busk 
(n. gen.) °. : ; : : . Victoria. 
4. Flustra pisciformis, Busk : . Bass’s Strait (Challenger Coll.). 
5 A papyrea, Pall. . : . Naples. 
6. 3 cribriformis, Busk  . . Singapore. 
7. Farciminaria hexagona, Busk . Amboina (Challenger Coll.). 
8. Dimetopia spicata, Busk. : . Victoria. 
9. Bicellaria grandis, Busk, var. pro- 
ducta, MacGillivray . : ; . Victoria. 
10. Bugula neritina, L. : : . Naples. 
11. Membraniporella nitida, Johnst. . S. Devon. 
12. Cribrilina philomela, Busk . . Marion Is. (Challenger Coll.). 
13. “ radiata, Moll . : . Naples. 
14. Umbonula verrucosa, sper. . SS. Devon. 
15. 3 pavonella, Alder. . North Sea. 


16. Lepralia pallasiana, Moll : . Naples. 


266 SIDNEY F. HARMER., 


17. Lepraliadorsiporosa, Busk . . ‘Torres Straits. 
18. 33 sincera, Smitt . : . Davis Straits. 
19. BA haddoni, n.sp. . ; . Torres Straits. 
20. Schizoporella linearis, Hassall . Naples. 
21: "3 sanguinea, Norm. . Naples. 
2p Se australis, Haswell . ‘Torres Straits. 
23. Urceolipora nana, MacGill. (= 

Calymmophora lucida, Busk) . Victoria. 
24. Smittia trispinosa, Jolhnst., var. 

arborea, Levins. . Greenland. 

25. g, reticulata, J. MaeCulivray Naples. 
26. Catenaria lafontii, Aud. . : . Naples. 


97. Vittaticella cornuta, Busk . . Victoria, 
28. Catenicella alata, Wyv. Thoms. . Victoria. 


29. BS plagiostoma, Busk, var. 
setigera, MacGill. . Victoria. 

30, 3 hastata, Busk . . Victoria (Challenger Coll.). 
3]. - lorica, Busk : . Victoria. 
See i wilsoni, MacGillivray . Victoria. 
33. Calwellia gracilis, Maplestone . Victoria. 
34. e (Onchopora) sinelairil, 

Busk . : : . S.of Kerguelen Is. (Challenger 

Coll., Stat. 153). 

35. , (Urceolipora) dentata, 

MacGillivray . ; . Victoria. 
36. Ichthyaria oculata, Busk : . S.E. of Buenos Aires (Chal- 


lenger Coll., Stat. 320). 
37. Onchoporella bombycina, Busk 


(not El]. and Sol.) . : : . New Zealand. 
38. Microporella malusii, Aud. . . Naples. 
39. * ciliata, Pall. . . Naples. 
40, Micropora, sp. . : . Torres Straits. 


41. Steganoporella Sele Hakmer Torres Stratis. 


Ill. Kuthyris, Hincks, and Huthyroides, n. gen. 


Kuthyris clathrata, n. sp. Pl. 16, tigs. 18—31. 


Zoarium tlustrine in habit, of stiff, corneous texture, 
composed of narrow, parallel-sided, frequently bifureating 
branches, with truncated ends. Zoccia opening on one 
surface only, the orifices arranged with great regularity in 


THE MORPHOLOGY OF THE CHELLOSTOMATA. 267 


oblique rows passing entirely across the branch in two inter- 
secting directions. Orifices apparently connected by a 
continuous, brown, transparent epitheca, a short distance 
below which the frontal surface of each zocecium is strength- 
ened by a system of irregular calcareous bars, which tend to 
radiate from a point in the middle of the base-line of the 
operculum towards the proximal and lateral sides of the 
zocecium. Basal side of the branch similarly covered by an 
epitheca, which each zocecium reaches along a longitudinal 
line narrower and shorter than itself. Opercula large, 
dimorphic, the ordinary form about as long as broad (250 to 
270 mn), the others with a broader base (290 to 320 u). Both 
kinds of opercula are strengthened by a conspicuous Q-shaped 
sclerite. The distal margin of the vestibule is provided with 
a chitinous lip, which is overlapped during retraction by the 
large lateral flanges of the operculum. Ovicells not found, 
and probably absent. 

The material was discovered by Mr. I’. Whitelegge under 
rock ledges, at low-tide line, Watson’s Bay, Port Jackson, 
and in Middle Harbour, Port Jackson. Although part was 
in spirit, its condition was not sufficiently good to make a 
complete anatomical investigation possible. 

The genus Huthyris was founded by Hincks! for a new 
species, H. obtecta, from North Australia. The generic 
name, introduced “to suggest the idea of higher structure ” 
in the operculum, is particularly appropriate to EK. clathrata, 
in which the operculum is specially complicated. The two 


1¢Ann. Mag. Nat. Hist.’ (5), x, 1882, p. 164. In 1871 Quenstedt (‘ Petre- 
faktenkunde Deutschlands,’ Abth. I, Bd. ii, p. 442), in discussing the struc- 
ture of a Brachiopod, used the following words :—‘ Man konnte sie daher 
wohl Kuthyris aber nicht Athyris nennen.” I have not been able to ascer- 
tain that Quenstedt made any further use of the word Huthyris, and he 
does not even refer the species he is discussing to that genus. It appears, 
therefore, that he was not in reality proposing a new genus (and he certainly 
did not define it), but was merely making a verbal criticism of the name 
Athyris. Although Kuthyris (Quenstedt) is mentioned by Zittel in his 
well-known ‘ Handbuch d. Paleont.’ (i, p. 684) as a synonym of Spirigera, 
it does not seem to me that it has any valid claim to recognition. 


268 SIDNEY F. HARMER. 


species agree in their habit, in their dimorphic opercula 
(probably associated with the absence of ovicells), and in the 
highly developed chitinous epitheca which overspreads the 
entire zoarium. I. clathrata differs from KH. obtecta in 
having its frontal calcareous wall composed of irregular bars, 
instead of a simple, perforated, calcareous film, and in the 
fact that there is no large space between the frontal epitheca 
and the calcareous walls. 

I feel doubtful whether Huthyris woosteri, MacGillivray,} 
is rightly referred to this genus; but, on the other hand, 
Carbasea moseleyi, Busk (1884, p. 56), perhaps belongs 
to it. 

The largest colony of E. clathrata measures about 18°5 
cm. or 7+ inches in length. ‘The branches are 2 to 4 mm. 
wide, averaging about 5 mm. near their free ends, but lessen- 
ing towards the base of the colony, which appears to have 
been attached by a narrow base without rootlets. A branch 
4mm. wide has about thirteen orifices in each oblique row. 
The colour is brown in the older parts, yellowish near the 
ends of the branches. The zoarium frequently bifurcates, 
showing some tendency to form a unilateral cyme. The 
terminal divisions (PI. 16, fig. 18) may reach a length of 3°5 
em. without bifurcating, but the ordinary length of the 
divisions is not more than 1°5 cm. The frontal surface is 
somewhat convex, the opposite surface flatter. The calca- 
reous walls of the zocecia are arranged as follows :—The 
lateral and terminal walls are everywhere complete, and are 
perforated by numerous pores. At the proximal end (fig. 22) 
the two lateral walls pass continuously into one another in a 
regular curve, which forms the base of the zocecium, and is 
placed some distance within the basal epitheca (fig. 26, b. ep.). 
At the distal end the lateral walls approach one another 
basally, and are separately inserted into the basal epitheca, 
forming a linear mark (figs. 22, 27) constantly shorter than 
the zocecium, but varying in length; this is connected with 
the similar part of the next zocecium by a chitinous “ mesen- 

1 * Proc. R. Soc. Vict.’ (N. §.), ili, 1891, p. 77. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 269 


tery” (fig. 22, m.). The effect of the arrangement indicated 
in fig. 27 is to keep the epitheca stretched as a flat membrane 
at some distance from the basal caleareous walls of the 
zocecia. The cavity beneath the epitheca is divided into a 
series of parallel longitudinal spaces by the parts of the 
zocecia above described. The linear figure formed by the 
insertion of the zocecinm into the basal epitheca is in some 
cases bifurcated proximally. On the frontal surface (figs. 
20, 21) each orifice is surrounded by a somewhat irregular 
ring of calcareous matter, from each side of which is given 
off a strong condyle (fig. 21, cond.) or “ denticle,” the two 
condyles forming the hinge of the operculum. The frontal 
surface is strengthened by a highly variable arrangement of 
calcareous bars, the general position of which is shown in 
fig. 20. The bars are in the main flattened, their flat surfaces 
being parallel with the surface of the branch, but in curving 
down into the lateral walls they usually give off vertical 
flanges from their free surface, and these form bridges across 
the depressed intervals between two zooecia, joining similar 
flanges in the adjacent zocecia. At the free end of the 
branch the proximal parts of the bars are first formed, and 
they grow in a distal direction beneath the epitheca. 

Along the lateral margin of the branch runs a tube (fig. 
20, m.c.)! which has usually been described as a “ chitinous 
fibre”? in other forms of Flustrine habit. This is merely a 
part of the branch which is not divided into zocecia, and 
calcareous bars (c. b.) extend from the marginal zocecia nearly 
to the outer edge of its free surface; it contains, moreover, 
strands of funicular tissue which pass across its lumen. This 
space runs as a continuous tube along the whole margin of 
the branch, and it communicates with the cavity which lies 
between the backs of the marginal zocecia and the basal 
epitheca. 

K. clathrata, like H. obtecta, is characterised by the 
dimorphism of its opercula. This is shown in figs. 20, 21, 
representing the ordinary type (“a’’) and the second form 

1 Cf, Waters, 1896, p. 291. 


270 SIDNEY F. HARMER. 


(“pn”). The zocecia to which the two kinds of opercula 
respectively belong also show some dimorphism. In the 
A-zocecia the condyles for the articulation of the operculum 
are long recurved teeth (fig. 21), while in the B-zocecia they 
are short tubercles. he distal calcareous wall of the 
A-zocecium is at the same time the proximal wall of the next 
zocecium in the same longitudinal line. In the B-zocecium 
this is not the case. The distal zocecium has a proximal 
wall (p. w.) of its own, from which some of its calcareous bars 
may spring, and this is much thinner than the distal wall 
(d. w.) of the proximal zocecium, from which it is separated 
by a narrow crescentic space, passing about half round the 
operculum of the proximal zocecium, This suggests that 
the B-zocecia possess a vestigial ovicell. The condition of 
my specimens unfortunately prevents me from ascertaining 
whether the production of ovaries is limited to B-zocecia. In 
two cases counted at random, about three A-zocecia occurred 
for every B-zocecium, no regular arrangement of the two 
kinds being apparent. The additional breadth of the 
B-operculum is correlated with a slightly increased trans- 
verse diameter of the zocecium itself, immediately on the 
proximal side of the operculam, This makes an appreciable 
difference in the capacity of the zocecium, a fact which is in 
favour of the view that the B-zocecia are female. Similar 
differences in the opercula are commonly met with in other 
Cheilostomes, in which the operculum of an ovicell-bearing 
zocecium may be wider than that of the ordinary zocecia. 
Two features in HK. clathrata demand especial attention, 
namely, the compensation-sac and the operculum. For the 
discovery of the compensation-sac Jullien (1888, 1) is 
entitled to full credit,'! although his results have been re- 


1 See my preliminary paper (1901) on this subject. Jullien’s accounts of 
the compensation-sac were very short, and his figures were not adequate. He 
nowhere brings out the importance of the sac in the discussion of the mor- 
phology of the Cheilostomata. So far as I know, he mentions its parietal 
muscles in only one place. This is in the explanation of a figure of Cribri- 
lina figularis (1888, ii, p. 272), in which he uses words which seem to 


THE MORPHOLOGY OF THE CHEILOSTOMATA. Da | 


ceived with scepticism, which proves to have been un- 
justified. 

The CompENsAtIon-sac of this species is a very large cavity 
which underlies practically the whole of the frontal surface 
of the zocecium. In longitudinal section (fig. 26, ¢. s.) it is 
seen to be perfectly distinct from the body-cavity, from 
which it is separated by a delicate membrane, constituting 
the floor of the sac. his membrane passes continuously 
into the base-line of the operculum, immediately proximal to 
which the sac opens to the exterior. When the operculum 
is closed the aperture of the sac is a virtual transverse slit. 
In the case of ordinary Escharine forms this aperture is so 
little apparent that it has commonly been supposed that the 
base-line of the operculum actually articulates with the 
adjacent part of the calcareous wall—a state of affairs which 
is often by no means the case. The roof of the compensation- 
sac is protected by the calcareous bars (f. b.), and the frontal 
wall of the zocecium is bevelled off on the proximal side of 
its aperture, in the characteristic way shown in fig. 26. The 
purpose of this arrangement is obvious. ‘The sharp edge of 
the frontal wall of the zocecium is in contact with the base- 
line of the operculum when the latter is closed. When the 
operculum is open (fig. 24) the plane of its free surface 
becomes parallel with that of the bevelled edge of the frontal 
wall, so that water can have unobstructed access to the sae. 
As supposed by Jullien, there can be little doubt that the 
entry of water into the sac renders the protrusion of the 
polypide possible, but it almost necessarily follows that the 
constant change of the water in the sac makes this structure 
an important organ of respiration. 

In this species I have been able to obtain only unsatis- 
factory evidence, in consequence of the state of its preserva- 


indicate that he had appreciated the morphological importance of his discovery. 
The words are as follows :—“ Muscles rétracteurs de l’abdomen (fibres muscu- 
laires pariétales des auteurs). Elles rétractent en réalité la paroi abdominale, 
mise en évidence par la découverte de la chambre a eau de compensa- 
tion ouchambre compensatrice.” 


Wiz SIDNEY F. HARMER. 


tion, with regard to the existence of the “ parietal muscles ” 
which elsewhere dilate the compensation-sac ; but the analogy 
of other Cheilostomes leaves little doubt as to their presence. 

The OpercuLum is a very remarkable piece of mechanism, 
which forms a most efficient means of protecting the entrance 
to the tentacle-sheath. Most of the existing descriptions of 
Cheilostomatous opercula take account of the appearance 
of the outer surface only of this structure—a very inadequate 
way of arriving at its real relations. Although K.clathrata 
has an operculum which, judged by the descriptions available 
for comparison, would appear to be unusually complex, it is 
in the highest degree probable that a renewed examination of 
other Cheilostomes will show that 1t 1s by no means unique 
in this particular. 

The first part of the introvert which leads to the mouth of 
the polypide is constituted by the ‘ diaphragm” or 
“ vestibule” (fig. 26, vest.), which is a muscular invagination 
projecting into the tentacle-sheath, and communicating with 
it by a central aperture.!. The way in which the vestibule 
opens and closes has been aptly compared with the action of 
a clasp-purse. The structures which protect the vestibule 
of HK. clathrata have a superficial resemblance to the skull 
of a turtle (Chelone), the skull with the upper jaw being 
represented by the operculum, and the lower jaw by a 
chitinous lower lip (figs. 23—27, /b.), which I propose to term 
the“labium.” ‘The labium was described by Hincks? in a form 
from the Queen Charlotte Islands, named by him Lepralia 
bilabiata, in allusion to the existence of this structure. It 
is probable that it will hereafter be found in numerous 
Cheilostomata. 

The labium can be clearly seen in those zocecia of the dry 
colony in which the operculum is open (fig. 19). Between it 
and the edge of the downwardly projecting flange (/l.) of the 


1 See, for a description of this structure, Nitsche (1871, p. 432), Jullien 
(1888, 4, p. 38), and Calvet (1900, pp. 180, 201). 

7 «Ann. Mag. Nat. Hist.’ (5), xiii, 1884, p. 49. The labium is the ‘‘ upper 
lip” of Hincks’s description. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. Dilee 


operculum is seen the entrance to the vestibule, while the 
aperture of the compensation-sac is situated between the base 
of the operculum and the calcareous frontal wall of the 
zocecium. 

The relations of the same parts are explained by the thick 
longitudinal section shown in fig. 26, in which the operculum 
is very nearly closed. By deep focussing can be seen the 
calcareous condyle (cond.) which constitutes the hinge. The 
divaricator muscles (div.) of the operculum are paired ; 
each originates from one of the lateral walls of the zocecium, 
and passes obliquely towards its frontal surface, crossing the 
condyle on its proximal side, to reach its insertion into the 
basal sclerite of the operculum. The contraction of these 
muscles will obviously have the double effect of opening the 
vestibule and of opening the compensation-sac. ‘The oper- 
culum is prolonged laterally into a very large triangular 
flange (fl.), whose plane is at right angles to that of its free 
surface (see also figs. 24, 25, 27). The occlusor muscles 
(ocel.) are similarly paired, each originating from the lateral 
wall of the zocecium at a deeper level than the divaricator, 
and passing obliquely across the base of that muscle to 
reach its insertion, on the distal side of the condyle, into 
the tip of the triangular flange. Fig. 25 shows that the 
insertion is by means of a broad tendon. Since the 
labium articulates with the operculum in this region, the 
effect of the contraction of the occlusor muscles will be, 
not only to close the operculum itself, but also to retract 
the labium from the position shown in fig. 24 to that shown 
in fig. 25. In the closed condition, the labium lies just 
inside the vertical flange of the operculum. 

The lateral flanges are not really independent structures, 
but they pass into one another in a continuous curve round 
the distal side of the operculum, their free border, continuing 
the comparison with the turtle’s skull, constituting the biting 
edge of the upper jaw. In longitudinal section (fig. 26) the 
upper jaw appears to be strengthened by a strong buttress 
(buttr.) which passes from the free surface of the operculum. 

voL, 46, PART 2,—NEW SERIES. s 


74 SIDNEY F. HARMER. 


This is a part of the Q-shaped figure which is seen in a 
surface view (fig. 21) of the operculum, an appearance which 
results from the fact that the thickness of the chitin is not 
everywhere the same. The whole of the region included 
within the two limbs of the Q is much thickened (figs. 
26, 27), while the rest of the free surface, as far as the edge, 
is composed merely of thin cuticle (fig. 29). From the 
median thickening a curved chitinous buttress diverges 
(figs. 283—31) on each side, passing down into that part of 
the body-cavity which immediately underlies the free surface 
of the operculum. In young zocecia, the space between the 
buttress and the outer lamella of the operculum contains a 
conspicuous epithelium, which secretes the cuticular sub- 
stance of which the operculum is composed. Remains of this 
cellular material may be seen, in the same position, in the 
adult operculum. 

The greater part of the thickness of the operculum is con- 
stituted by a superficial layer of cuticle which readily stains 
yellow with picric acid. ‘The whole of the deeper surface of 
this layer (and therefore both sides of the strengthening 
buttress which depends from the free surface), is lined by a 
much thinner layer of chitin (not indicated in the figures), 
which takes ared colour in sections stained with borax car- 
mine and picric acid. A two-layered chitinous cuticle has 
been described in certain Gymnolemata by Calvet (1900, 
p. 164). 

The buttress of the operculum is crescentic in shape, as 
seen from above (fig. 21, buttr.). Thus, if the thin lateral 
parts of the outer lamella of the operculum were removed, 
there would be exposed a crescentic cavity, the limbs of the 
crescent being directed towards the proximal side of the 
zocecium, the floor of the cavity being constituted by 
the slightly concave upper surface of the buttresses. Distally 
the buttress meets and fuses with the vertical flange of the 
operculum (fig. 26). Laterally the fusion is incomplete 
distally (figs. 27, 28), so that the cavity above the buttress is 
here continuous with the general body-cavity. In _ the 


THE MORPHOLOGY OF 'THE CHEILOSTOMATA. PAT) 


proximal half of the operculum the buttresses again meet the 
lateral flanges (figs. 29—31). 

The distal half of an operculum is shown in fig. 27, which 
represents a transverse slice of a specimen with its calca- 
reous parts, cut freehand! after embedding in paraffin. By 
focussing deeply are seen the labium (/b.) and the curved shit 
between it and the edge of the operculum. Into the ends of 
the lateral flanges of the operculum, and at the same time 
into the two sides of the base of the labium, are inserted the 
occlusor muscles (occl.). By focussing less deeply the free 
edges of the lateral buttresses are seen ; while the cut surface 
passes through the region of the condyles (cond.). Fig. 26 
shows that the operculum works on its condyles as a lever of 
the first order, the occlusors (occl.) and the divaricators (div.) 
passing on opposite sides of the fulcrum. 

Fig. 28 shows a transverse section of the distal part of the 
operculum where the buttresses have a free edge. Fig. 29 
is through the hinge-line. The vertical flange is still deep 
(cf. fig. 26), and is here strengthened by the buttresses, 
which have completely fused with it. On the proximal side 
of the condyles the depth of the flange rapidly diminishes. 
In the entire operculum a small circular mark (fig. 21, 7) 
appears on each side near the proximal end of the Q-shaped 
figure. In transverse section (fig. 30) this is seen to be due 
to the fact that at this point the buttresses leave the vertical 
flanges and are inserted into the free wall of the operculum, 
so that at this point the body-cavity is separated from the 
outside by a single layer of chitin. The small tube thus 
formed often stains deeply in carmine preparations. I am 
unable to explain the meaning of this arrangement. ‘The 
tube does not communicate with the space above the 
buttresses. On its proximal side the operculum becomes 
very shallow, as indicated in fig. 31, a section just on the 
distal side of the basal sclerite. This forms, with the median 
thickening, a horizontal L-shaped figure, the outline of which 

The hard nature of the chitin makes if difficult to prepare microtome 
sections of the opercula. 


276 SIDNEY F. HARMER. 


gives rise to the Q-shaped mark seen in an external view of 
the operculum (fig. 21). 

In decalcified sections the calcareous part of the condyle is 
seen to be surrounded by a strong layer of chitin. In longi- 
tudinal sections this has the appearance indicated in figs. 25, 
26 (cond.), the chitin being developed especially on the 
proximal side of the condyle. Fig. 24, which represents an 
open operculum, shows that this chitinous investment is pro- 
longed into the sharp edge (scl.) which surrounds the cavity 
in which the operculum lies (see also figs. 27,29). ‘This edge 
is continuous with a layer of chitin forming the outer wall of 
the cavity (figs. 28—81) ; and this, in its turn, becomes con- 
tinuous with the deep end of the vertical flange of the oper- 
culum. ‘The flange is strongly bent inwards soas to lie along 
the surface of the condyle. 

I think there is great probability that these arrangements 
are in the nature of a spring, and that if the partially opened 
operculum shown in fig, 29 be imagined to be closed by the 
occlusor muscles, the shape of the lateral flanges and of the 
buttresses will be somewhat distorted by the surfaces of the 
condyles. The elasticity of the operculum would thus tend 
to commence the opening of the orifice and of the compensa- 
tion-sac. This supposed elasticity probably gives a reason 
for the free ends of the buttresses in the distal part of the 
operculum, an arrangement which would facilitate the slight 
alterations in the vertical flanges during the movements of 
the operculum. : 

The whole apparatus, consisting of the operculum, the 
labium, and the chitinous cutting edge which surrounds the 
cavity containing the operculum, is obviously one of remark- 
able perfection, and is well adapted to defend the entrance 
to the tentacle-sheath. 

The most external part of the vestibule is lined by a 
distinct layer of cuticle (fig. 26), which is continuous with 
the cuticle of the operculum and with that of the labium. 
The actual “biting ” surface of the jaws is constituted by a 
thickened part of the cuticle, as shown in the same figure. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 277 


That of the operculum is so close to the insertion of the 
buttress into the vertical distal wall as to receive support 
from the buttress. That of the “lower jaw ” is continuous 
with the main thickening of the labium, which is the deeper 
wall, as shown in fig. 26. 

I have few observations on other points in the structure of 
this species. In the material at my disposal a certain 
amount of disintegration appears to have taken place before 
the specimens were preserved, ‘I'he number of tentacles is, 
however, about twenty-two. 


Kuthyris obtecta, Hincks,! Pl. 16, figs, 32—87, 

The B-zocecia are much larger than the A-zocecia, the 
ratio in volume being probably at least 2:1 (figs. 32, 35). 
The A-opercula are of much the same shape as in 
K. clathrata, while the B-opercula are relatively much 
broader. The condyles are weaker than in E. clathrata, 
and are situated nearer the proximal border of the oper- 
culum. Bat what most strikingly differentiates the two 
species is the character of the calcification. The calcareous 
wall is everywhere complete in EH. obtecta, the basal wall 
being uniformly rounded along its whole extent. Both kinds 
of zocecia are flask-shaped (figs. 34, 36), the opening of the 
neck of the flask being filled by the operculum. A con- 
tinuous flat epitheca overspreads the entire zoarium, being 
stretched out in the plane of the opercula on the frontal 
surface of the branch, and at a considerable distance from 
the zocecia on the basal surface (fig. 34, ep., b. ep.). The 
space beneath the epitheca (ep.c.) is thus extremely large, 
and that of the frontal surface is continuous with that of the 
opposite side by means of the undivided space which runs 
along each edge of the frond (figs. 32, 35, m. ¢.). 

On the frontal surface the epitheca is supported in the 
main by the edges of the orifices of the zoccia, but in 
addition by a few calcareous tubercles (fig. 34, cale. p.), 
which arise irregularly from various points of the surface of 

1 “Ann. Mag. Nat, Hist.’ (5), x, p. 1645; (6), xi, p. 177. 


278 SIDNEY F. HARMER. 


the calcareous wall (fig. 32), and pass vertically to the 
epitheca. Basally these tubercles are longer and are rather 
more regularly arranged. ‘Their arrangement is seen in 
figs. 34-36, from which it is apparent that they serve to 
keep the epitheca stretched out at a considerable distance 
from the calcareous part of the zocecium. In my preliminary 
note (1901, p. 16) I have pointed out that there is evidence 
that many Cretaceous Cheilostomes were provided with a 
basal epitheca, as is indicated by the presence of calcareous 
papillae resembling those of E.obtecta. It may perhaps 
be suggested that a function of the epitheca is to protect the 
calcareous walls from the attacks of boring organisms (e. g. 
the Infusorian Folliculina) which infest many calcareous 
Polyzoa. 

The calcified wall of E. obtecta shows no trace of the 
bar-like arrangement so characteristic of KH. clathrata. It 
is, on the contrary, a continuous calcareous film, traversed 
by pores. Asin the other species, these pores partly form 
communication-pores traversing the partition-walls between 
zocecia, and they are partly in relation with the space 
beneath the epitheca. It is hardly necessary to point out 
that in neither case are they open pores, although the cal- 
careous matter is deficient in these regions. ‘The portions of 
the vertical walls which are actual partition-walls are of 
limited extent (fig. 36), so that but few of the pores are com- 
munication-pores (c. p.) between zocecia. I have seen no 
evidence that the number of these is restricted to eight, as 
stated by Waters (1896, p. 282). The majority lead to the 
spaces beneath the epitheca, particularly to that on the basal 
side of the frond (figs. 35, 36). They are less numerous on 
the frontal side (fig. 36). 

The marginal part of the frond may be strengthened by a 
system of bars (fig. 87) which recalls those already described 
in the zocecia of HK. clathrata. ‘This system makes its 
appearance first in the angle between the two lobes of a 
bifurcation, and begins as a set of regular calcareous bars, 
running near the frontal surface, and originating from the 


THE MORPHOLOGY OF THE GHRILOSTOMATA. 279 


outer sides of the marginal zocecia. Each bar is at first 
narrow, but soon dilates into a thickened, tuberculated, 
rounded head, the heads being closely apposed to form a con- 
tinuous calcareous border. From the border a longitudinal 
vertical ridge stands up at right angles to the frontal 
epitheca, which it meets. A similar ridge (/) connects the 
border with the basal epitheca. In older branches each bar 
may be dilated at two points, and a second longitudinal ridge 
may be formed on each surface (fig. 37, J, 1). 

The marginal thickening,’ thus constituted of a series of 
calcareous thickenings, gives flexibility as well as strength to 
the margin of the frond. ‘There can be little doubt that the 
space beneath the epitheca is a kind of colonial body-cavity. 
The marginal bars of E. obtecta can thus be regarded as 
directly comparable with the zocecial bars of EK. clathrata. 
From the irregularity and variability of the bars in the latter 
I am inclined to regard this as a species in which the calcifica- 
tion has been reduced from a condition like that in E. obtecta. 

The general arrangement of the viscera in E. obtecta is 
shown in fig. 34. The operculum has conspicuous lateral 
flanges, and there is a delicate labium (lb.). The occlusor 
muscles (occl.) are inserted into the apices of the lateral 
flanges. Some of the fibres of the divaricator muscles (div.) 
appear to reach the base-line of the operculum, but some are 
probably inserted into the adjacent part of the floor of the 
compensation-sac (c. s.). This structure is very large, and in 
an adult zocecium it underlies the whole of the frontal 
surface. Fig. 33 shows its appearance in a B-zocecium in 
back view. The greater part of the basal wall (b. w.) of the 
zocecium and part of the polypide have been removed. The 
tentacle sheath depresses the sac medianly, but the sac bulges 
out on each side into a strongly convex lateral lobe. Each 
lobe thus formed is rounded off distally, but by deep focussing 
the two lobes can be seen to unite on the far side of the 
tentacle sheath into a single cavity, which can be traced to 


1 A similar thickening is well known in certain other flexible Cheilostomes. 
See, for example, Levinsen’s account (1891, p. 274) of Flustra carbasea. 


280 SIDNEY F. HARMER. 


the proximal border of the operculum. The divaricator 
muscles are seen at div.; occl. indicates the position of the 
origin of one of the occlusor muscles, while a pair of parieto- 
vaginal muscles are seen at p. v.m. Arising from the sides 
of the zocecium are a series of delicate parietal muscles (p. m.). 
There is a tendency for these to be arranged in a grouped 
manner. ‘They can be traced along the basal surface of the 
compensation-sac, into which each is inserted. 

A polypide and a compensation-sac occur in both forms of 
zocecia. Reproductive organs are, unfortunately, absent, so 
that it is not possible to ascertain whether the dimorphism 
has any relation to reproduction. 

The compensation-sac develops in what I shall term the 
Lepralioid manner ; that is to say, as an invagination formed 
at the base of the operculum after the calcification of the 
front wall has been completed. Some of the details of this 
process are described in the account of the next species. 

Kuthyroides, n. gen.—I suggest this term for Carbasea 
episcopalis, Busk,! a form placed by Hincks ? in the genus 
Euthyris. The diagnosis of the genus may be given as 
follows: 

Zoarium of Flustra-like habit, bordered along each edge 
by a tube, interrupted at intervals, which represents a part of 
the body-cavity not divided into zoccia, but without other 
spaces beneath the epitheca. Frontal wall more or less 
calcareous, covering a well-developed compensation-sac. Com- 
munication-pores large, typically four on each side. Ovicells 
large, external, with a wall composed of two calcareous layers. 

The genus differs from Huthyris in the absence of spaces 
beneath the epitheca and in the presence of large external 
ovicells, and of a very different type of communication-pore. 

The zocecia of EK. episcopalis (Pl. 15, figs. 13—17) are 
extremely elongated. Their vertical and basal walls are 
calcified, but there is so little caleareous matter in the frontal 

1 Busk (1852), p. 52; MacGillivray, ‘ Prod. Zool. Vict.,’ I, Dee. v, 1880, 
p. 28. 

2 «Ann. Mag. Nat. Hist.’ (5), x, p. 164. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 281 


wall that this part may not entirely retain its shape in 
drying. 

Large, oval communication-pores, or rosette-plates, occur 
in the lateral walls, at about the middle of the interval 
between the frontal and basal surfaces (fig. 14, ¢.p.). Hach 
zocecium is normally bounded by two zocecia on each side, 
and it usually communicates with each of its four lateral 
neighbours by means of two (rarely three) rosette-plates. 
In the terminal partition-walls, there is a horizontal row of 
small pores in place of definite rosette-plates. The vertical 
walls have no other pores. 

The transparency of this species makes it a favourable 
one for the study of the compensation-sac. At the growing 
ends of the branches the frontal surface of the zocecium is at 
first an uncalcified membrane, in which calcification begins 
at the proximal end and gradually extends distally. The 
outline of the operculum becomes apparent before calcification 
invades its immediate neighbourhood. Shortly before the 
edge of the calcified frontal wall (fig. 15, v) reaches the 
region of the future operculum, the part of the uncalcified 
membrane immediately proximal to the opercular base-line 
shows a special accumulation of nuclei (c. s.), towards which a 
number of muscle-fibres radiate through the body-cavity from 
both lateral walls. When the calcification has advanced so 
far as to mark out the future orifice, two lateral calcified 
processes and a median tongue-like structure begin to grow 
up just proximal to the operculum (fig. 14, 1. p., tg.). The 
nuclei are arranged in a more definite mass along the proximal 
margin of the orifice, to which the lumen of the tentacle- 
sheath (t¢.s.) now extends. Still later, the two calcareous 
processes meet, although the suture between them is per- 
sistent. There is thus left, between them and the calcareous 
tongue, a crescentic pore (fig. 15), the concavity of which is 
directed proximally. ‘This is an aperture of the compensa- 
tion-sac, which, however, opens to the exterior at the proximal 
edge of the operculum as well, an arrangement which is most 
obvious in the fertile zocecia. The occlusor muscles of the 


282 SIDNEY F. HARMER. 


operculum are large, and originate from the lateral parts of 
the basal wall, close to the distal end of the zocecium. The 
divaricator muscles originate from the lateral walls of the 
zocecium, in the same neighbourhood. 

The zoarium is bounded by a tubular cavity, which is, 
however, not continuous, as in the preceding species. The 
cavity is a direct prolongation of one of the longitudinal 
lines of zocecia, and may be compared with what Smitt calls 
a ‘‘Samknopp” (1865, p. 6). After extending a length of 
perhaps five ordinary zocecia it ceases abruptly, and a new 
marginal tube is formed as a prolongation of the series of 
zocecia next internal to it. The lateral tubes are connected 
with the adjoining zocecia by communication-pores. 

The ovicell (fig. 16) of this species is well known from 
Busk’s description. It is very prominent, and is provided 
with a median longitudinal keel, on each side of which is a 
large elongated fenestra (f.). This is simply a membranous 
deficiency in the outer wall of the ovicell, which is composed 
of two calcareous layers. The fertile zocecium is distin- 
guished from the others by a peculiarity in its calcareous 
frontal wall. Instead of having a single pair of prominences 
on the proximal side of the operculum, it has several such 
prominences (figs. 16, 17), the arrangement of which is some- 
what variable. There are either several tubular calcareous 
bars radiating towards a point in the middle of the zocecium 
(fig. 16), or a short series of bars disposed more definitely in 
pairs (fig. 17). The latter arrangement seems to occur 
typically in those fertile zocecia which have an ovicell on 
their proximal side, and the former in a fertile zocecium 
which is the first of a longitudinal series. 

The resemblance of these bars to the frontal bars! of a 


1 T suggest the term “ frontal membrane” for the membranous body-wall 
which is stretched over the “aperture” of a Membranipora (fig. 43). 
The “ frontal bars,” or ‘ costules”” of some authors, are the bars which grow 
over this membrane in a Cribrilina (figs. 8, 44); while I propose the term 
“frontal shield” for tle calcareous part of the frontal surface of a Cheilostome. 
The frontal shield is probably not homologous in all Cheilostomes, but the 
term may be used as a purely descriptive one, 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 283 


Cribrilina is very striking. They arch over a greatly 
reduced frontal membrane, into which parietal muscles are 
inserted in the immature condition of the zocecium (fig. 17). 
The bars are calcareous tubes, opening by a foramen into the 
body-cavity just external to the frontal membrane, and each 
bar has a minute membranous fenestra near its tip. After 
the complete formation of the bars there is found (in the 
fertile zocecium) a stage in the development of the com- 
pensation-sac precisely like that shown in fig. 15 in a 
zoceclum without an ovicell, and it can now be seen that the 
operculum is continuous with the floor of the compensation- 
sac, a wide opening into which is left between the operculum 
and the first pair of bars. In the mature fertile zocecium the 
compensation-sac extends under the greater part of the 
frontal shield, as in the ordinary zocecia. 

There is little difference between the relations of the frontal 
bars in the fertile zocecia of KH. episcopalis and that found 
in certain species of Cribrilina (e. g. C. figularis), in 
which the frontal membrane surrounded by the bars is of 
reduced extent. 

In the immature ovicell of EH. episcopalis (fig. 17) the 
inner calcareous wall is a concave plate (7. w.) lying on the 
surface of the zocecium next distal to that to which the ovicell 
belongs. ‘The outer calcareous layer (0. w.) rises up con- 
centrically outside it, and between the two is a mass of living 
tissue. It is impossible not to be struck by the resemblance 
between the development of the ovicell and that of the frontal 
bars. The ovicell may be compared with two greatly ex- 
panded bars, composed, like the others, of two layers of 
calcareous matter surrounding living tissue. 

The median keel of the mature ovicell represents the 
line along which these bars meet, and corresponds, I believe, 
with a complete septum between their cavities. It may thus 
be suggested that the ovicell is formed by the fusion of a 
pair of greatly expanded oral spines, the bases of which 
should communicate with the fertile zocecium on each side 
of the operculum. JI cannot claim to have proved this to 


284, SIDNEY F. HARMER. 


be the case, though I have obtained some evidence pointing 
in that direction. 

Dr. G. M. R. Levinsen has kindly given me permission to 
allude to his very interesting observation (which he proposes 
to publish hereafter), that in Alysidium parasiticum, 
Busk,! the ovicells develop as two arched, hollow valves, 
corresponding with the oral spines which occur on the 
ordinary zocecia. I do not at present know what conclusions 
Dr. Levinsen deduces from this observation, with which I was 
acquainted before making my own on KH. episcopalis. 

In Heterowcium amplectens, Hincks*? has described 
a Membranipora-like Cheilostome, in which the ovicell is 
constituted by a number of spines placed distal to the oper- 
culum, which meet in a Cribrilina-like manner, and form a 
structure which, in other respects, resembles a normal ovicell. 

Calvet (1900, pp. 57, 58, 152, pl. u1, fig. 14) states that in 
Bugula there is a communication-pore between the ovicell 
and the distal zocecium. 

It follows from the account given by Calvet and others, 
that the cavity of the ovicell, internal to its inner layer, is an 
external space which is overarched by the double wall of the 
ovicell. This is obviously true in a case like that of 
Mucronella coccinea, where the most distal oral spines 
of the fertile zocecium are actually inside the cavity of the 
ovicell. 

Further investigation is necessary to decide the morpho- 
logical nature of the Cheilostome ovicell. The existence of 
three possibilities is generally recognised :—(1) That the 
ovicell belongs to the fertile (proximal) zocecium ; (2) that it 
belongs to the distal zocecium ; (8) that it is a modified indi- 
vidual, as believed by Nitsche and others. The second possi- 
bility would seem to be indicated by Calvet’s observation 
above referred to. The relation of the operculum of the 
fertile zocecium to the ovicell, the occurrence of the “ internal 


1 1859, p. 14. 
2 «Ann. Mag. Nat. Hist.’ (5), viii, 1881, p. 129, pl. ili, fig. 7; ibid. (6), 
ix, p. 332, 


THE MORPHOLOGY OF THE CHEILOSTOMATA, 289 


ovicells” found in certain species of Flustra and elsewhere, 
and in particular Dr. Levinsen’s account of Alysidium 
parasiticum, seem to be in favour of the view that the 
ovicell is a part of the fertile zocecium. 

In his original account of Flustra militaris,! Waters 
suggests that this species is allied to HE. episcopalis, calling 
attention to the resemblance between the ovicells of the two 
forms, but noting the existence of differences between the 
opercula. ‘The two species agree in the great length of their 
zocecia and in their rosette-plates; but a striking difference 
is seen in the frontal surface, which is membranous in F. 
militaris, except for a small proximal calcified region. 
Parietal muscles appear to be inserted into this membrane, 
asin an ordinary Flustra. The two strong suboral spines 
of F. militaris may correspond with the two projections 
which in HK. episcopalis cut off the median pore. If the 
frontal surface has really the importance in classification 
attached to it in the present paper, I see no way of admitting 
the affinity of the two species (which undoubtedly resemble 
one another), except by assuming either that F. militaris 
is a species which has secondarily opened its compensation- 
sac, or that EH. episcopalis is a modified Flustrine form. 
The material at my disposal is not in a condition which 
allows me to make a further examination of these points. 


IV. THe MorpHoitocy oF THE COMPENSATION-SAC AND OF THE 
OPERCULUM. 


I have come to the conclusion that the evolution of the 
compensation-sac has not been identical in all Cheilostomes 
which possess that structure, but that in some cases it has 
been formed by the overarching towards the middle line of 
a series of marginal spines in such a way as to cover the 
primitive frontal membrane.? The evidence for the occur- 
rence of this process is as follows : 

(1) Many species of Membranipora exist in which the 

1 *Ann. Mag. Nat. Hist.’ (5), xx, p. 93. 
* See p. 282. 


286 SIDNEY F. HARMER. 


arrangement of the marginal spines foreshadows the condition 
above indicated. 

(2) The Cribrilinide are transitional from Membranipora 
to some at least of the Lepralioid genera. 

(3) The arrangement of the muscles connected with the 
compensation-sac is derivable from the condition found in 
Membranipora. 

(4) The study of the primary zocecium. 

(5) Paleontological evidence. 

The view that the Cribrilinide are intermediate between 
the Flustrina and the Escharina is not a new one. Smitt 
(1868, i, p. 401) states explicitly that the frontal bars of 
Membraniporella nitida are homologous with the free 
marginal spines of Membranipora lineata; and in his 
next paper (1868, 1, p. 48) he shows that Cribrilina marks 
a further transition to the Escharines. He leaves Membrani- 
porella in the Flustrina, while placing Cribrilina in the 
Escharina. Hincks, on the contrary (1880), places the two 
genera in the family Cribrilinide. 

But although taking this view of the intermediate position 
of the Cribrilinide, Smitt was not in a position to show in 
detail how the Flustrine zocecium could be modified into an 
Escharine zocecium. ‘The compensation-sac enters into the 
question as supplying the clue necessary for the solution of 
the problem. 

The foregoing instances have given some idea of the 
relations of the compensation-sac in its fully developed form. 
I next proceed to the proposition that the compensation-sac 
of some Lepralioid genera has been derived from a Mem- 
branipora-like condition through a stage similar to that of 
existing species of Cribrilina. 


(A) Flustrina. 
Flustra pisciformis, Busk.1—Fig. 4 shows the general 
anatomy of a young zocecium of this species. In the lateral 


1 1852, p. 50. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 287 


walls the calcareous matter is in two layers separated by a 
chitinous lamella,! as described by Nitsche (1871, pp. 421, 
455). Hach zocecium thus has its own calcareous lateral walls, 
distinct from those of its neighbours. In two regions, 
respectively proximal and distal to the broadest part of the 
zocecium, the lateral wall is thickened at its frontal edge. 
At the proximal narrow end the calcareous wall becomes 
deficient, and the terminal partition wall, which is thicker 
than any of the others, belongs to the proximal zoccium of 
the two which it divides. This accords with Levinsen’s 
statement (1891, p. 251) that in many Cheilostomes the 
terminal wall is single, while the lateral walls are double, so 
that the longitudinal rows of zocecia can be isolated from one 
another by boiling with caustic potash. ‘lhe frontal surface 
is entirely membranous; the operculum is merely a part of 
this membrane, and has no basal sclerite. When the oper- 
culum is open (fig. 3) it is seen to have a vertical flange, 
produced into lateral points, as in Huthyris. A similar 
arrangement is described by Nitsche ? in Membranipora 
membranacea. The occlusor muscles (occl.) are inserted 
into these points, and originate, as in Huthyroides episco- 
palis, from the basal wall. A pair of strong parieto-vaginal 
muscles (p. v. m.) pass from the tentacle sheath to the basal 
wall, and a smaller pair (p. v. m.’) connect the tentacle sheath 
with the frontal wall; but none pass to the vertical walls 
(cf. Calvet, 1900, p. 199). <A pair of strong parieto-dia- 
phragmatic muscles (p. d.) sprmg from the basal wall, just 
internal to the origin of the occlusor muscles, and are inserted 
into the diaphragm or vestibule. The retractor muscles (7. m.) 
of the polypide have their origin in one of the corners of the 
fish-tail-like proximal end of the zocecium. About four 


1 In a recent paper by Schulz (‘ Arch. f. Naturg.,’ Ixvii, Bd. i, Heft 2, 
1901, p. 118, pl. vi, fig. 4) this chitinous lamella is figured as a much thicker 
layer than anything I have ever seen. The species investigated by Schulz was 
M. “membranacea” (= M. monostachvys, Busk; cf. Levinsen [1891], 
pe 277): 

= 1871, p. 422. 


288 SIDNEY F. HARMER. 


groups of parietal muscles (p. m.) are seen, of which the distal 
group (p. m.’) originates from the basal wall, and the rest 
from the vertical walls, close to their junction with the basal 
wall. 

These arrangements are moderately constant in all the. 
zocecia, The distal group of parietal muscles is always strong, 
and is commonly stronger than the others. ‘he number of 
groups does not vary much, but it is important to notice that 
the distal group may (rarely) originate from the base of a 
vertical wall, and conversely that the other parietal muscles 
may originate from the basal wall. The distal parietal muscles 
thus clearly belong to the same series as the rest. 

The mode of action of the parietal muscles has been well 
described by Nitsche (1871, p. 426), who showed that the 
pressure on the fluid of the body-cavity due to their contrac- 
tion was the main cause of the protrusion of the polypide. <A 
more detailed account of the same process is given by Calvet 
(1900, p. 63). 

F., pisciformis has no distinct divaricator muscles. ‘The 
opening of the operculum is probably largely due to the fluid- 
pressure brought about by the contraction of the parietal 
muscles generally ; but it seems to me highly probable that 
the distal group of these muscles, by pulling on the membrane 
with which the base of the operculum is immediately 
continuous, may have a special effect in opening the operculum. 
I regard the distal parietal muscles as the starting point from 
which the divaricator muscles of Huthyris and its allies 
have been derived. 

Flustra papyrea, Pall.—This species is very similar to 
F. pisciformis in essential respects. The more noteworthy 
differences are (1) the parietal muscles are more numerous, 
usually numbering about six to seven groups on each side ;! 
(2) while each of the ordinary groups consists of not more 
than two or three fibres, the distal group is a much broader 
and more definite muscular band, composed of a considerably 
larger number of fibres; (3) the distal as well as the other 


1 They are shown in my paper (1892), pl. ii. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 289 


parietal muscles originate typically from the basal end of the 
lateral walls. The operculum resembles that of F. pisci- 
formis. 

I have observed the occurrence of typical parietal muscles 
in five other species of Flustra. 

Membranipora.—The operculum and the parietal muscles 
in M. membranacea and other forms which have been 
examined are essentially Flustrine in their arrangement. 
They have been described by Nitsche (1871, p. 426) and 
others. 

Farciminaria hexagona, Busk.'—The whole of the 
very long frontal surface is membranous. The operculum 
is like that of Flustra pisciformis, with no basal sclerite, 
and with a marginal vertical flange prolonged into two 
lateral points, into which the occlusor muscles are inserted. 
The orifice is markedly bilabiate, the operculum with the vesti- 
bule opening precisely hke a clasp-purse, of which the mar- 
ginal flange of the operculum forms the supporting rim of 
one half, while the distal edge of the zocecium supports the 
other half. There are about six groups of parietal muscles 
on each side, the distal group probably acting as in Flustra 
pisciformis. 

The next few species are ordinarily placed with the 
Cellularina, but their affinities appear to me to be with the 
Flustrina. 

Dimetopia spicata, Busk 2 (fig. 5).—This is an instance 
of a dendritic form, in which the lateral zocecial walls constitute 
the greater part of the exposed surface of the funnel-shaped 
zocecium. ‘lhe frontal membrane is small, and is situated at 
the broad end of the funnel, the edge of which is surrounded 
by seven or eight hollow calcareous spines. ‘These originate 
from a massive calcareous ring which is notched externally at 
the base of each spine to permit the cellular tissue of the 
spine to communicate with that of the general body-cavity. 

1 Busk (1884), p. 51. 

2 Busk (1852), p. 35; MacGillivray, ‘Prodr. Zool. Vict.,’ Dee. v, 1880, 
p. 33. ; 

vou. 46, PART 2.—NEW SERIES. . 


290 SIDNEY F. HARMER. 


The lumen of the spine has a characteristic globular dilatation 
at the point where it crosses the calcareous ring. The 
calcareous wall of this dilatation is completed externally by 
the outer wall of the zocecium. ‘The bases of the spines are 
connected with one another by a thin calcareous web, while 
part of the frontal membrane may project just beyond the web, 
internal to the bases of the spines. 

The zocecia are arranged back to back in pairs, new 
zocecia (z.’) originating from the axial side of their distal end. 
The orifice is close to this region, on the axial side of the 
frontal wall (fig. 5). The ocelusor muscles (occl.) originate 
from the lateral walls. In correlation with the reduction of 
the leneth of the frontal membrane, there is but a single pair 
of parietal muscles (p. m.). 

Bicellaria grandis, Busk, var. producta, MacGilli- 
vray! (fig. 1)—This is another example of an infundibuli- 
form zocecium. ‘The frontal membrane is obliquely placed, 
and is of restricted extent. The orifice is at the extreme 
outer end of the membrane, and the zocecium is prolonged 
beyond it into a lobe which bears three or four very long 
spines. On the axial side of the zocecium is a calcareous 
thickening? which surrounds the communication-pores (c. p.). 
The zocecia have the alternate arrangement characteristic 
of the genus, and in most cases the calcareous thicken- 
ing includes two sharply marked rings, one for the alternate 
zocecium of the other side of the branch, and the other for 
the next zocecium on the distal side. In fig. 1, however, 
there are three calcareous rings, the zocecium being the one 
which precedes a bifurcation, and having consequently to 
give rise to two distal zocecia. A peculiarity of this species, 
which I have not found in other species of Bicellaria, is 
that the proximal part of the “aperture” becomes filled in 
with a thin calcareous film (erypt.) which ends in a hook-like 
point. I am unable to state with certainty what is the 
function of this hook, but I think that its distal edge forms 


1 * Prodr. Zool. Vict.,’ Dec. vi, 1881, p. 38. 
2 Cf. Goldstein, ‘Trans. Proc. Roy. Soc. Vict.,’ xviii, p. 43. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 291 


a guide over which the tentacle sheath works in the move- 
ments of protrusion and retraction. I have not definitely 
settled whether the calcareous film is a superficial calcifica- 
tion or an internal cryptocyst, but the existence of a delicate, 
free, vertical edge all round the “aperture” probably in- 
dicates that there is living tissue adjacent to it (which I 
believe that I have been able to make out), and that the film 
is of the nature of a cryptocyst. 

The operculum is like that of Flustra, and there is a 
single pair of parietal muscles, as in Dimetopia. 

Bugula neritina,! L. The aperture is elongated, and 
occupies the greater part of the length of the zocecium. The 
parietal muscles are arranged in a series of Flustra-like 
groups, which lie very close to the lateral walls of the 


zooecium. 


(8) Cribrilinide. 


It has long been known that the characteristic frontal 
wall of certain Cribrilinidz develops as a series of marginal 
spines, originating from the periphery of the aperture and 
converging towards one another until they meet, and so 
form a calcareous roof in which the intervals between the 
bars remain either as simple slits or as a series of pores. An 
excellent account of this process is given, for instance, by 
Hincks (1880, p. 199). 

It is easy to see, by an examination of dry specimens of 
various Cribrilinide (e.g. C. annulata) that the operculum 
is not continuous with the calcareous frontal wall, but with 
the underlying frontal membrane. ‘This may be observed 
either in immature zocecia or in those from which the wall 
formed by the union of the bars has been broken away. 
Goldstein (1880, p. 48) has made the interesting observation 
that in C. monoceros(?) a Rotifer swam into the space 
between the bars and the frontal membrane. Although this 
is not expressly stated, it presumably entered at the distal 


1 Busk (1852), p. 44. 


292 SIDNEY FI. HARMER. 


end of the series, between the bars and the base of the 
operculum. 

The growth of these bars can easily be studied in the 
common Membraniporella nitida,' Johnst. The young 
zocecium is completely Membranipora-like, even in the 
character of its operculum. The bars are hollow calcareous 
spines, which originate as a series of crenulations from the 
edge of the vertical wall of the zocecium, each of the future 
bars being indicated by a round bay of the edge. By the 
closure of this bay the commencement of a tube is formed, 
which arches over the frontal membrane in such a way as to 
leave a considerable space between itself and the membrane. 
The shits between the bars are probably permanently open. 
Another point which is worth notice is that the calcareous 
bars are part of a series of which the distal ones persist 
as the oral spines. In Cribrilina annulata the bars 
originate in the same way, but the intervals between suc- 
cessive bars persist as a series of pores which probably 
remain open. 


Cribrilina philomela, Busk.? 


Dry specimens show at once that the operculum is con- 
tinuous with the original frontal membrane,® which becomes 
covered over in the same way as that above indicated 
(fig. 44). The fully developed zocecium (fig. 8) is covered in 
front by a series of alternate hollow bars, the lumen of each 
of which is for the most part narrow, but dilates into a 
rounded end near the middle line of the zocecium. The 
rounded end lies more superficially than the rest of the 
cavity. The outer wall of each bar is considerably thicker 
than the inner wall. 

The operculum is Flustra-like, with no basal sclerite, but 

1 Hineks (1880), p. 199. 

2 1884, p. 132. 

3 This is shown in a diagrammatic longitudinal section of a Cribrilina 
figured by Canu (1900, p. 441, fig. 53, iii), where the roof formed by the 
united caleareous bars or ‘‘ costules,’ is seen to be entirely independent of the 
operculum. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 293 


it is of firmer texture than in the preceding forms. It has 
the usual lateral projections of the vertical flange for the 
insertion of the occlusor muscles. Condyles are barely in- 
dicated, but they can just be made out as two slight knobs 
distal to the proximal broadest part of the operculum. 

A Flustrine series of groups of parietal muscles! occurs 
(fig. 9), among which a stronger group (p. m.’) on each side, 
immediately proximal to the operculum, probably acts as a 
divaricator, as in Flustra. 

A few of the zocecia are modified as vicarious avicularia. 
The zocecium is normal, but the operculam is longer than 
usual, and the occlusor muscles originate from as much as 
the distal two thirds of the basal wall. These zocecia are 
provided with typical parietal muscles. Polypides are 
absent in most parts of the preparation, but in several cases 
the vicarious avicularium possesses a polypide. These 
modified zocecia are not provided with ovicells, which occur, 
here and there, on the ordinary zoecia. 


Cribrilina radiata, Moll (fig. 7). 


This species, which differs in important respects from C. 
philomela, is considered in the following section (p. 326), 


(c) Lepralioid Genera, 

Under this heading I shall consider not only some of the 
genera which are usually referred to the Escharina, but a 
certain number of the branching forms at present included 
in the Cellularina. 


Umbonula verrucosa,? Esper (fig. 11; see also the dia- 
gram, fig. 12). 

The frontal membrane of this handsome species is perma- 

nently Flustrine; but it becomes overarched by a strong 

1 The parietal muscles of Membraniporella nitida are figured by Smitt 

(1865), pl. vi, fig. 1; those of Cribrilina figularis by Jullien (1888), 2, 


pl. xX 


2 Hincks (1880), p. 317. 


294, SIDNEY F. HARMER. 


caleareous roof, which extends to about the middle of the 
orifice, where it ends in a prominent shoulder on each side, 
rising in the middle line into a massive suboral umbo. ‘This 
bears a strong avicularium at its base, immediately over the 
entrance to the compensation-sac. The rounded mandible, 
when closed, is directed towards the end of the umbo, and 
then lies in a vertical plane, transverse to the long axis of 
the zocecium. The umbo, which probably belongs to the 
avicularium, is supported by a series of radiating buttresses 
of the frontal shield, and between them are deep pits, the 
marginal areole (ar.), closed by a layer of living tissue or 
epitheca. 

Fig. 11 shows the commencement of the calcification of the 
frontal shield. ‘The lateral partition-walls consist of a thin 
chitinous lamella, with a layer of calcareous matter on each 
side, belonging respectively to the two zocecia separated by 
the wall. At a level a little lower than the free edge, the 
lateral and proximal partition-walls give off a calcareous film, 
which in the young specimen figured reaches no further than 
the edge of the frontal membrane. his film is pierced by a 
series of round holes (p.), one of which corresponds to each 
interval or areola (ar.) between two of the future buttresses. 
These holes establish a continuity between the living tissue 
of the body-cavity and that which occurs on the outside of 
the frontal shield. Beyond the origin of the frontal shield 
the vertical wall splits into two membranous lamella, one of 
which passes over each of the contiguous zocecia to form its 
epitheca, and becomes continuous with the free edge of the 
frontal shield (fig, 12). The frontal membrane (/. m.) becomes 
covered by a continuation of this process—in other words, by 
the formation of a crescentic fold, of which the deeper 
lamella is calcareous, and the superficial layer is composed of 
a living membrane. In the incompletely calcified zocecium 
the edge of the crescentic fold is always membranous. Irom 
the deep lamella rise the radiating ridges which form the 
buttresses. 

‘he polypide is still young in fig. 11. With the distal end - 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 295 


of its tentacle sheath are connected the two “ opercular 
glands,”’? while somewhat more proximally are inserted the 
conspicuous parieto-vaginal bands (p.v.m.). The operculum 
is of delicate Flustrine structure, and has no special divari- 
cators. Of the strong occlusor muscles (occl.), that of the 
(apparent) left side is seen to originate from the distal vertical 
wall, and that of the right side from the lateral wall. A 
series of groups of parietal muscles (p. im.) originate from the 
lateral and proximal walls, and are inserted into the frontal 
membrane. 

There can, I think, be little doubt that the space below the 
calcareous roof is the homologue of the similar space in 
Cribrilina. The main difference between the two spaces is 
that that of Cribrilina is covered by a series of originally 
separate bars, while in Umbonula it is from the first a 
continuous crescentic film. The marginal pores,*? which are 
the only ones that occur in the film, have, however, precisely 
the same relations as the communications between the cavities 
of the hollow Cribrilina spines and the general body-cavity. 
I regard the covering of the compensation-sacin Umbonula 
(fig. 50) as having been derived from a Cribrilina-like con- 
dition (fig. 44) by the lateral fusion of the spines with one 
another, the edges of the spines being indicated by the 
buttresses, and their cavities by the marginal areole and by 
the pores leading from those spaces to the general body- 
cavity. This involves the necessity of assuming that the 
outer calcareous layer of the Cribrilina spines is now repre- 
sented by an uncalcified membrane, the epitheca. The very 
common occurrence of marginal areole in Escharine forms 
may be appreciated by turning over the plates of almost any 
paper dealing with a collection of recent or fossil forms. 


1 Cf. Waters (1892), p. 272 (‘gland-like bodies’), and Calvet (1900), 
p. 200. 

2 The relations of these pores are well shown in a figure given by Neviani 
(‘ Boll. Soe. Geol. Ital.,’ xv, 1896, p. 24). 


296 SIDNEY F. HARMER. 


Umbonula pavonella, Alder (= Mucronella 
pavonella, auctt.'). 


The structure of the adult zocecium (fig. 10) closely re- 
sembles that of U. verrucosa, and I venture to place the 
two species in the same genus. ‘There is no umbo nor median 
avicularium, but the frontal shield ends in a rounded median 
lobe which is not raised above the level of the rest of the 
shield. Hach of its lateral shoulders (f. sh. d.) bears a con- 
spicuous avicularium (avic.). Its free surface rises into a 
series of radial buttresses, and the areolew (ar.) between these 
communicate with the general body-cavity by pores arranged 
as in U. verrucosa. My preparations show with great 
distinctness that cords of living tissue (w) traverse these 
pores, and unite to forma continuous sheet of living substance 
(indicated by the brilliantly stained nuclei) which overspreads 
the entire frontal shield even in old zocecia. The basal wall 
is not provided externally with a living membrane. 

It is well known that the front surface of Cheilostome 
zocecia may alter greatly with age. ‘lhe original pattern 
may become lost, the thickness of the wall may increase very 
ereatly, and the orifices of the zoccia may finally be 
completely covered by secondary calcification. All these 
changes can be understood when it is realised that the frontal 
shield of an Mscharine Cheilostome is covered by living 
tissue. Milne-Edwards (1836, p. 27), in discussing the 
alterations which take place with age in Kscharine forms, 
came to the conclusion that the calcareous matter was a 
living tissue which grows like a bone. Goldstein (1880, 
p. 48) has stated that in Mucronella ellerii the brilliant 
scarlet colour ‘seems to be located in a fleshy epidermis, 
with which the stony polyzoary is coated.” This is the 
“epitheca ” of many authors, commonly seen in dry prepara- 
tions as a membrane overspreading the calcareous frontal 
shield, and consisting, in reality, of an external cuticle and 
subjacent living tissue. 


' Hineks (1880), p. 376. 


THE MORPHOLOGY OF 'THE CHEILOSTOMATA, 297 


The existence of the epitheca is responsible not only for 
the ordinary form of secondary calcification, but it explains 
the condition so commonly found in Cellepora, where new 
zocecia originate on the free surfaces of the old ones. This 
appears to be the result of the separation of the epitheca from 
the calcareous wall, the subjacent space (which is morpho- 
logically part of the body-cavity) increasing in size and 
forming the body-cavity of a new zocecium. 


Lepralia pallasiana,! Moll (Pl. 17, fig. 41). 


The frontal wall of the young zocecium is at first 
Membranipora-like. Calcification begins at the proximal 
end by the formation of a thin calcareous film, which is per- 
forated by large, uniformly spaced pores. This film is 
formed in situ beneath a delicate epithecal membrane. ‘The 
process proceeds with considerable rapidity until the whole 
of the frontal wall is calcified, with the exception of the 
operculum, between which and the distal end of the zocecium 
occurs a portion of the frontal wall, with a single row of large 
pores. In the earlier part of this process the polypide bud 
is small, and is at the proximal end of the zocecium, being 
connected with the membranous frontal wall by a cord of 
cells. This becomes hollowed out to form the tentacle sheath 
(fig. 41), at the distal end of which appear two thickenings, 
the opercular glands (op.gl.). When the calcification is com- 
pleted, the compensation-sac (fig. 41, c. s.) begins to be obvious 
as a well-marked cavity, extending from the base of the 
operculum beneath the calcareous frontal wall, and at first 
much shorter than the operculum. Into the floor of the sac 
are inserted the parietal muscles (p.m.), which, in the im- 
mature zocecium, radiate even from distant parts of the 
lateral walls of the zocecium. 

The sac rapidly grows to such an extent that it underlies 
the whole of the frontal wall, the parietal muscles thereby 
acquiring an arrangement similar to those of Cribrilina. 


1 Hincks (1880), p. 297. 


298 SIDNEY F. HARMUER. 


The process is identical with that which has been described 
above in Huthyroides episcopalis. The fact that the 
wall of the compensation-sac becomes tightly pressed against 
the lateral wall of the zocecium makes it difficult to see its 
outline in the old zocecia; and the same applies to many 
other Cheilostomata. But in a decalcified preparation the 
characteristic parietal muscles can be seen, even in old 
zocecia, with the utmost distinctness. 

The operculum articulates with two small condyles which 
are situated just on the distal side of its proximal broader 
part. ‘The appearance of two linear longitudinal submarginal 
sclerites (fig. 41) is due to the special development of the 
vertical flange into two lateral ridges, the distal connecting 
part of the flange being less developed. The occlusor 
muscles (occ/.) are inserted into these lateral ridges, and they 
originate low down from the vertical walls of the zocecium, 
either from the lateral walls or from the distal wall. The 
divaricator muscle is the distal group of parietal muscles on 
each side. I am not certain whether the insertion of these 
is directly into the base of the operculum or into the floor of 
the compensation-sac close to the operculum. 

The small, non-porous, suboral region of this species may 
be taken to indicate the former presence of a suboral avicu- 
larium! (as in Umbonula verrucosa). This view is con- 
firmed by the fact that a few of the zocecia actually possess 
a small avicularium with a semicircular mandible, which, 
when closed, lies horizontally, and points away from the 
operculum. 

L. pallasiana is one of the forms which has been specially 
studied by Calvet (1900), many of whose statements I can 
confirm, although he has not observed the compensation-sac. 
In the younger zocecium shown in his pl. xii, fig. 20 (L. 
foliacea), it appears to me that he indicates the develop- 
ment of the compensation-sac as an ectodermic invagination 
at the proximal end of the operculum. In his pl. vi, fig. 1, 

1 Waters (1853, p. 430) calls attention to the common occurrence of the 
suboral avicularium in Chicilostomes. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 299 


which in other respects gives a good idea of the genera 
structure of the zocecium, I think he has put the insertion of 
the occlusor muscles too near the proximal end of the oper- 
culum. 


Lepralia dorsiporosa, Busk'! (fig. 45). 


In one respect this species appears to me to show more 
primitive characters than L. pallasiana. The distal pro- 
longations (f. sh.d.) of the frontal shield do not completely 
surround the orifice, the point where they meet being 
commonly indicated by a slight emargination on the distal 
side of the zocecium. Each distal prolongation typically 
bears an avicularium (avic.). 

The compensation-sac is seen with great distinctness in a 
decalcified preparation. It underlies the whole of the frontal 
wall on the proximal side of the operculum, and it is provided 
with typical parietal muscles. The distal group on each side 
is specially strong, and it appears to me that the part of the 
compensation-sac immediately adjacent to the base of the 
operculum, into which these muscles are inserted, is some- 
what fascia-lke, an arrangement which confirms the view 
that the muscles function as divaricators. ‘The condyles 
(cond.) ave well-developed recurved hooks. 


Lepralia sincera, Smitt.” 


In this species, for specimens of which I am indebted to 
Dr. Levinsen, the compensation-sac and its muscles resemble 
those of the preceding species. The distal shoulders of the 
frontal shield extend no further than about the middle of the 
operculum. One of these shoulders is occasionally provided 
with an avicularium. 


Lepralia haddoni, n. sp. (figs. 38, 39). 
Zoarium encrustivg or bilaminate (Hscharine). Zocecia in 
regular longitudinal rows, each about 700—900 x long and 
about 370 x broad. Surface covered by a distinct epitheca. 


1 1884, p. 143. 
2 «Oty. k. Vet.-Ak. Forh.,’ xxiv (Bilang, 1867), 1868, pp. 28, 177. 


300 SIDNEY F. HARMER. 


Proximal part of the frontal shield with numerous large 
pores, which do not extend to the distal end of the zocecium. 
The oral end of the zocecium may be somewhat raised above 
the general surface of the flat frontal shield. A rounded 
suboral lobe occurs, and may have almost the appearance of 
a mucro. Ordinary orifices and opercula somewhat longer 
than broad, with a concave proximal margin, 210—250 
broad; fertile zocecia with trifoliate orifice and operculum, 
320—840 «x broad. External ovicells only represented by the 
somewhat hood-lke calcareous wall at the distal end of the 
zocecium. An avicularium, with long linear mandible, 
commonly occurs on one side of the orifice. Basal surface 
smooth, its intersection with the four vertical walls forming a 
regular oblong. Pore-chambers absent. 


Torres Straits, A. C. Haddon collection, 1888-9. 


This species somewhat resembles L. feegeensis, Busk, but 
differs from it in the absence of external ovicells. It is of 
interest in throwing light on the meaning of the occurrence 
of dimorphic opereula. The outline of the ordinary oper- 
culum (fig. 38) is not unlike that of the A-opercula in 
Kuthyris obtecta. ‘The B-opercula are much wider, and 
are trifoliate. In either kind of zocecium there may be an 
avicularium on one side of the orifice, its acute mandible 
sloping, when closed, towards the distal end of the zocecium. 
The compensation-sac is present in both kinds of zocecia, 
and has typical parietal muscles. In some zocecia it contains 
long wavy filaments, which may be parasitic Alge or Fungi. 
In the zocecia with trifolate operculum an ovary (fig. 39) is 
developed at the basal end of one of the lateral walls. A 
spherical ovisac, with delicate walls, makes its appearance at 
the distal end of the zocecium, on the basal side of the oper- 
culum and tentacle sheath. Into this sae are inserted 
muscle-fibres, which radiate to it from the adjacent parts of 
the basal and lateral walls. In one zocecium an egg has 
matured and has become filled with yolk, its diameter nearly 
equalling that of the zocecium. ‘This egg has passed into the 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 301 


interior of the spherical ovisac, while the ovary still remains 
in its original position. 

My preparations do not enable me to trace the origin of 
the egg-containing sac, which may, however, bea diverticulum 
of the vestibule, as described by Calvet (1900, fig. 42, on 
p. 266) in Lepralia pallasiana. 

The conclusion pointed to by these facts is that the dimor- 
phism of the opercula indicates the derivation of the species 
from forms provided with ovicells; in which, as is well 
known, the operculum of the fertile zocecium is commonly 
larger than that of the other zocecia. 


Flustra cribriformis, Busk.! 

This species is introduced at this point in further illustra- 
tion of the morphology of the ovicell. Its frontal wall is, of 
course, typically Flustrine. The ovicell is of the type usually 
described as “internal,”” and is merely represented by a 
more or less hemispherical bulging of the distal zocecial wall 
into the next zocecium. Into this space projects a spherical 
vesicle, which lies just beneath the vestibule, as in the fore- 
going species. The polypides are retracted to the extreme 
proximal end of the zocecium. In some cases I have been 
able to detect a very small ovary in one of the proximal 
corners of the zocecium, near the basal wall; but in conse- 
quence of the position of the polypides I cannot say whether 
the ovary is parietal or attached to the polypide. The ege 
enters the spherical sac, in some manner which has not been 
observed, while it is still very small. It receives its yolk while 
in the sac, and when it is full grown it is so large that with its 
investing sac it occupies as much as half of the body-cavity. 


Schizoporella linearis,® Hass. (figs. 48, 52). 
The characteristic “ sinus ” 


of the operculum of this genus is very similar to the part of 


or projecting proximal tongue 


the operculum which in certain species of Lepralia occurs 
1 1852, p. 51; 1884, p. 58. 

Cf. Levinsen (1891), p. 275. 

3 Hincks (1880), p. 247. 


bo 


302 SIDNEY F. HARMER. 


on the proximal side of the condyles. It appears to me that 
the function of the tongue is partly to close the aperture of 
the compensation-sac, and partly to give suitable leverage 
for the action of the divaricator muscles. 

S. linearis is characterised by the possession of a pair of 
suboral avicularia, with pointed mandibles. Hincks (1880, 
p. 251) has described in this species a form of “ ocecium,” 
associated with a rudimentary zocecium. My preparations 
show clearly that these supposed ovicells (fig. 48, avic.) are 
in reality gigantic suboral avicularia,! of the type well known 
in Retepora monilifera, var. munita, Hincks.? The 
greatly swollen region distal to the mandible is occupied by 
the enormous occlusor muscles, which differ only in size from 
those of the ordinary avicularia, their relation to which is 
further shown by the fact that a single gigantic avicularium 
may lie obliquely across the zocecium instead of longitudinally 
as in fig. 48. The compensation-sac (fig. 52, c.s.) of the 
zocecia is small, and even in zoccia with fully mature poly- 
pides is commonly no larger than in the specimen figured. 

The sinus of the operculum (fig. 52) fits into a correspond- 
ing emargination in the frontal shield. At a deeper level 
are seen the two condyles (cond.) which constitute the hinge. 
The frontal wall is of considerable thickness, and its distal 
margin, in the neighbourhood of the sinus, is bevelled off in 
precisely the way noticed in Huthyris clathrata (fig. 26). 
This results in the formation of a groove, the outer end of 
which is closed by the sinus of the operculum, while the inner 
end (fig. 52, a) is distinetly larger. When the operculum is 
open, its plane must be approximately parallel to the part of the 
bevelled surface which lies in the middle line. A short tube 
is thereby formed by which water can enter or leave the 
compensation-sac. 

The compensation-sac develops as in Lepralia palla- 
siana. The calcification takes place beneath the epitheca, 


1 Waters (1885, p. 6, and 1892, p. 274) has previously stated that this is 


the case. 


2 «Ann. Mag. Nat. Hist.’ (5), i, 1878, p. 361. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 505 


since a layer of nuclei can be seen superficial to it. The 
parietal muscles are very delicate. The occlusor muscles 
originate from the distal wall, but I have not certainly seen 
divaricator muscles. 


Schizoporella sanguinea, Norman (fig. 46). 


This is the form described by Waters,! in his paper on the 
Polyzoa of Naples, as Lepralia pertusa, var. sinuata. 

The operculum is rather stronger than in S. linearis, and 
it has a well-marked marginal sclerite which is distinctly 
thickened at the points where it articulates with the condyles. 
The compensation-sac (fig. 46, ¢.s.) is beautifully shown in 
specimens from which the basal wall has been removed. 
The parietal muscles radiate into its lateral walls, and the 
external parts of its basal wall, from the lateral and proximal 
walls of the zocecium. ‘he distal parietal muscles (p.m.’), 
which stretch transversely across the zocecium, probably act 
as divaricators. 

Dry specimens show the epitheca, stretched over the tuber- 
cles of the reticulately thickened frontal shield. 


Schizoporella australis, Haswell.” 

This species is characterised by having an elongated avicu- 
larium on the proximal side of the orifice, with an acute 
mandible directed obliquely towards the proximal end of the 
zocecium when closed. The condyles are strong, and have 
denticulated edges (fig. 47). The continuity of the floor 
of the compensation-sac with the sinus of the operculum is 
clearly seen in the preparations. The sac resembles that of 
S. linearis. 


Urceolipora nana,® MacGill., 1881. 


The operculum is Schizoporella-like. At its proximal 
border opens a large compensation-sac which extends nearly 

1 «Ann, Mag. Nat. Hist.’ (5), ii, 1879, p. 31, pl. viii, fig. 5. 
‘Proce. Linn. Soc. N. 8. Wales,’ v, 1881, p. 41. 

3 MacGillivray, ‘ Prodr. Zool. Vict.,’ Dec. xi, 1885, p. 19; Busk (1884), 
p. 82 (as Calymmophora lucida), 


to 


304 SIDNEY F. HARMER. 


to the proximal end of the zocecium, a great part of the cavity 
of which it occupies. ‘The two calcareous spines at the sides 
of the orifice have the effect of keeping the conspicuous 
- epitheca, which invests all parts of the zoarium, stretched out 
at some distance from the wall of the flask-shaped zocecium. 
The curved lines shown by Busk (1884, pl. xxxii, fig. 3 b) on 
the sides of the zocecia indicate the edges of the thicker 
lateral calcified walls, and are probably the actual junction of 
the lateral and frontal walls. Their level corresponds nearly 
with that of the floor of the compensation-sac, which has 
parietal muscles. The sac originates with great distinctness 
as an invagination from the proximal border of the operculum. 


Smittia trispinosa, Johnst., var. arborea, Lev.! (fig. 42). 


The specimens here referred to were sent to me, determined 
as above, by the Upsala Museum. Fig. 42 shows that there 
is a large compensation: sac (¢. s.), the opening of which is 
overhung by a median lobe of the frontal shield. he 
condyles (cond.) are at a much deeper level, and are long, 
recurved denticles, which are crossed by the submarginal 
lateral sclerites of the operculum. They clearly form the 
hinge on which that structure moves. ‘Typical parietal 
muscles are present. his form can hardly be regarded as 
a very typical Smittia, 


Smittia reticulata,’ J. MacGillivray (fig. 40). 


My evidence with regard to this species is not complete. 
In specimens from which the basal wall has been removed | 
have just been able to detect a large compensation-sac with 
walls of great tenuity, into which parietal muscles are in- 
serted. The median denticle (m.t.) or “lyrula”® appears to 
belong to the suboral avicularium (as in Rhynchozoon). 


1 « Bryozoer fra Kara-Havet,” ‘Dijmphna-Togtets Zool. bot. Udbytte,’ 1886, 
p. 16. 

2 Hineks (1880), p. 346. 

3 Cf. Waters, ‘Ann. Mag. Nat. Hist.’ (6), iv, p. 14. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 305 


The principal interest of this species, in the present connexion, 
is as an illustration of the group of Cheilostomes in which 
there is but a single series of conspicuous marginal pores or 
areolee (ar.). 


Catenaria lafontii,! Aud. (fig. 49). 


I consider this species in this place on account of its 
possession of a typical compensation-sac ; but I shall discuss 
its affinities later. The zocecia are very long, curved, narrow 
proximally, and dilating distally. From the back of the 
distal end usually originate a pair of new zocecia (z.’), each 
starting in a calcareous base containing a body-cavity, into 
which is fitted a chitinous joint (j.), which forms the connexion 
with the younger zoccium. ‘The oral extremity of the 
zocecium is surrounded by a circlet of short, hollow spines 
(sp.). The ovicell (ov.) is placed obliquely on the distal 
side of the orifice, and an extremely strong suboral avicu- 
larium (avic.) guards the entrance to the compensation-sac. 
There is some evidence that the base of the avicularium 
has been derived from a single pair of suboral frontal 
spines. ‘The porous region of the zocecial wall in the main 
corresponds with the limits of the compensation-sac (c. s.), but 
a few pores occur outside that region, even on the basal 
surface. 

The compensation-sac is extremely distinct, and is indicated 
in a figure by Calvet (1900, pl. viii, fig. 9), who does not, 
however, make any reference to the structure. It is provided 
with several groups of parietal muscles (p. m.). The oper- 
culum is placed obliquely, and its base is distinctly continuous 
with the floor of the compensation-sac, the orifice of which is 
rather large. Hach occlusor muscle (occl.) originates from 
the partition wall between the body-cavity of the zocecium and 
the base of the daughter-zocecium of its own side. I have 
not distinguished any divaricator muscles. 


1 Busk (1852), p. 14. 
voL. 46, PART 2.—NEW SERIES. U 


306 SIDNBY F. HARMBR. 


Vittaticella cornuta,! Busk (fig. 56). 


Maplestone* has recently suggested the generic name 
Vittaticella for the “ vittate” species of Catenicella, the 
subdivision of which appears to me desirable. 

Fig. 56 shows that Vittaticella is provided with a well- 
developed compensation-sac (c.s.). The younger zoccium 
further illustrates the fact that the development of this 
structure is typically Lepralioid. The sac is figured by 
Jullien (1888, 3, pl. xi), in two species of Catenicella, in 
both the young and the adult condition. A wide crescentic 
opening into the sac is seen in the younger zocecium (fig. 56) 
between the base-line of the operculum and the proximal 
border of the orifice. Amongst the parietal muscles which 
radiate from the wall of the sac can already be distinguished 
a distal group (p. m.’), the future divaricator muscles. In the 
next zocecium the compensation-sac is so large as to underlie 
nearly the whole of the frontal surface. 

Kach vitta (v.) is a tubular cavity, running longitudinally 
along the edge of the zocecium, and bounded externally by a 
delicate uncalcified membrane. ‘lhe cavity communicates by 
a single series of funnel-shaped tubes with the body-cavity.® 
In the young zocecia the whole of this arrangement is filled 
with a brilliantly staining cellular material ; but the structure 
is identical in the older zoccia, in which remains of cells 
may still be made out. 

I refer to the morphology of the vittee below when dealing 
with Catenicella hastata. ‘Their function is, perhaps, to 
assist im the deposition of the calcareous wall. The 
zocecium in this species is strengthened by a great thicken- 
ing of its edge, which is more pronounced in the region 
of the vitta than elsewhere. ‘lhe vitta extends along the 
whole length of the lateral thickening, and in the more 
elongated zocecia it may have as many as ten pores. 

1 Busk (1852), p. 11. 
* “Proc. Roy. Soc. Vict.’ (N. §.), xiii, 1901, p. 201. 
3 Cf. Waters (1881), p, 318. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 307 


The characteristic horns of this species are probably to be 
regarded as modified avicularian zocecia. In certain zocecia 
one of the horns is replaced by a large norma] avicularium. 


Catenicella alata, Wyv. Thomson! (fig. 53). 


The zocecial structure of Catenicella has been described 
by Busk (1852, p. 4), and by Waters (1883, p. 428). The 
zocecium develops on each side three structures, which to- 
gether form the great wings so characteristic of this species. 
Taking a case which is not complicated by the presence of 
the twin zocecium (cf. fig. 56) which occurs just before a 
bifurcation of the branch,? the zocecium is found to be some- 
what spindle-shaped, each end passing into a tubular sheath 
which surrounds the chitinous joint by which it is connected 
with its neighbours. The basal surface of the zocecium is 
extremely gibbous, and projects in this direction far beyond 
the wings. The frontal surface is nearly flat. The large 
orifice has, on its proximal side, a scutiform calcareous region 
bearing five fenestrae closed by membrane. ‘The zocecium is 
considerably wider than this scutiform plate, its convex. 
lateral surfaces being overlapped by parts of the wings. 

Kach wing consists of three parts: (1) a large proximal 
cavity (infra-avicularian + pedal compartments of Waters), 
which is almost two thirds of the length of the zocecium, and 
is provided on its frontal surface with two large membranous 
fenestre, one at each end (fig. 53, if. avic.) ; (2) a transverse 
cylindrical cavity, the free outer end of which has a mem- 
branous vacuity (avic.) ; (3) a large distal cavity with a single 
large fenestra, its distal point being uncalcified (swp. avic.). 
The second of these cavities is morphologically an avicularian 
zocecium, as is proved by comparision with other species of 
Catenicella. 

The five suboral fenestrae perforate the entire thickness of 
the calcareous frontal wall. At the level of their internal 


1 *Nat. Hist. Review,’ v, 1858, “ Proc. of Societies,” p. 137. 


2 Waters terms the solitary zomcium a “ globulus,” and the two associated 


zoecia a “ biglobulus.” 


308 SIDNEY F. HARMER. 


openings a sharply marked curved line (x) indicates the edge 
of a calcareous plate similar to that described below in C. 
plagiostoma. ‘he markings between the fenestrae and the 
operculum are somewhat variable. In a specimen which 
partly dried up during its preparation, fine slits containing 
air were noticed, radiating from the fenestra to the orifice. 

The compensation-sac of this species is described by 
Jullien (1888, 3). It is well developed, but its complete out- 
line is not easily seen in the mature zocecium in consequence 
of the fact that its breadth is greater than that of the exposed 
scutiform frontal wall. 


Catenicella plagiostoma, Busk, var. setigera, 
MacGillivray? (fig. 54). 

This species is remarkable for the obliquity of its orifice, 
and for the enormous development of its avicularia, which 
may have very different forms, even on opposite sides of the 
same zocecium (fig. 54). ‘The infra-avicularian compartment 
(inf. avic.) is almost entirely closed by membrane, the single 
fenestra so formed extending mainly over the side and basal 
surface of the zocecium. ‘The supra-avicularian compartment 
(sup. avic.) similarly has membranous walls, a very large 
fenestra being common to it and the avicularian zocecium. 

The structure of the frontal wall is more easily made out 
than in C, alata. The outer calcareous layer is reduced to a 
system of conspicuous bars which unite with a calcareous 
framework surrounding the operculum. ‘The internal 
calcareous layer is an obliquely oval plate? (pl.), which projects 
towards the proximal end of the zocecium. ‘This forms part 
of the roof of the compensation-sac, from which muscles (p. m.) 
can be seen radiating out to the adjacent parts of the wall of 
the zocecium. The distal groups of parietal muscles (p. m.’) 
appear to act as divaricators. 

In a back view of a zocecium it is seen that the avicularian 
zocecium has a very large oblique fenestra proximal to the 

1 * Prodr. Zool. Vict.,’ Dec. iii, 1879, p. 17. 
? This structure is described by Waters (1883, p. 429). 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 209 


mandible, and another distal to the same structure, and 
common to it and the supra-avicularian compartment. From 
various parts of the membrane covering the dorsal fenestre 
rise the spine-like processes to which the variety owes its 
name. 


Catenicella hastata, Busk! (fig. 55). 


The supra-avicularian compartment (swp. avic.) is calcified 
except at its extreme tip, the walls being perforated by small 
pores. It forms a broad, flattened spine at the upper lateral 
corner of the zocecium. ‘The avicularian mandible is minute, 
but the avicularium possesses, in addition to its occlusor and 
divaricator muscles, a polypide rudiment (pol.), as in various 
other Cheilostomes. Ihave noticed a similar structure in other 
species of Catenicella. The infra-avicularian compartment 
(inf. avic.) is completely divided into two. The outer wall of 
both these cavities is usually completely calcified, but the 
proximal one may have a slit-like membranous fenestra, 
which is usually lateral. Between the proximal portion, 
designated by Waters (1883, p. 428) the ‘ pedal compart- 
ment,’ and the zocecium, was noticed a row of three com- 
munication pores (c.p.). A comparison with Vittaticella 
suggests that the vittee shown in fig. 56 are the last remains 
of the pedal compartments with their communication pores. 
This view has already been maintained by Waters (p. 428), 
and, though not quite in the same manner, by Jullien (1888, 3). 
I have no suggestion to make with regard to the function of 
the lateral “ compartments ” in Catenicella. 

The frontal wall has from seven to nine small fenestra, from 
which tubular cavities pass transversely towards the middle 
line. The arrangement is strikingly Cribrilina-like, and 
the resemblance is intensified by the existence of irregular 
slit-like cavities alternating with the fenestra, which suggest 
an incomplete lateral fusion of frontal bars. 

‘he compensation-sac is usually clear. Its floor appears to 
be deeply pigmented, but this may be the result of the 

1 Busk (1852), p. 7; 1884, p. 10. 


310 SIDNEY F. HARMER. 


method employed, this preparation being the only one which 
IT have made with Mayer’s cochineal tincture. The species 
seems, however, to have a special development of pigment, 
as indicated by Maplestone.! ‘The compensation-sac develops 
as in Vittaticella. 


Catenicella lorica,” Busk. 


The remarkable fertile zocecia which characterise the 
genus Catenicella are well seen in this species to be 
provided with a large compensation-sac, with strong parietal 
muscles. ‘he three fenestrae of the ordinary zocecia appear 
to perforate only a single thin calcareous layer, the inner 
layer being completely absent. 


Catenicella wilsoni,’® MacGill. 


The great size of the fenestra of the infra-avicularian com- 
partment makes it an easy matter to see the outline of the 
large compensation-sac of this species, and the arrangement 
of its muscles. 


(p) Microporelloid Genera. 


Under this heading I consider a few of the forms with a 
“median pore,” though I am by no means certain that all 
such forms are related to one another. 


Calwellia gracilis, Maplestone* (figs. 61, 62). 


The zocecia are in pairs, back to back, the plane uniting 
the middle line of two zocecia being at right angles to the 


1 «Trans. Proc. Roy. Soc. Vict.,’ xvii, 1882, p. 49. 

2 Busk (1852), p. 6. 

3 MacGillivray, ‘Prodr. Zool. Vic.,’ Dec. ix, 1884, p. 30. 

The form which I describe in this paper was figured by Maplestone, with- 
out description, in a paper entitled “ Observations on Living Polyzoa,” in 
‘Trans. Proc. Roy. Soc. Vict.,’ xviii, 1882, p. 48, fig. 9 [the plate has no 
number], It has not the triangular shape mentioned by Wyville Thomson 
(1858) in his original account of C. bicornis. It is not the species described 
by MacGillivray as C. gracilis in ‘Trans. Proc. Roy. Soc. Vict.,’ xxii, 
p. 128; see also the same journal (N. 8.), i, p. 106. 


= 


THE MORPHOLOGY OF THE CHEILOSTOMATA. Sal 


similar plane of the next pair. The narrow proximal part of 
each zocecium extends the whole length of the subjacent pair, 
in such a way that each zocecium originates from the next 
lower one which looks in the same direction as itself. 

The two lateral horns are not calcified at their tips (fig. 
62, f). Below the oblique orifice is a triangular region 
sloping to the median pore (m.p.), which occupies the most 
projecting part of the zocecium. The middle line of this 
region is marked by a suture (s.), on each side of which is an 
‘oculiform ” pore. Besides the communication pore at each 
end, by which it is connected with its neighbours in the same 
longitudimal series, each zocecium has a pair of distal com- 
munication pores placed on the basal side, by which its soft 
tissues are continuous with those of the narrow proximal 
half of the next distal pair of zocecia, while its own proximal 
half is similarly provided with a pair of pores (c.p.) which 
lead to the two zocecia of the subjacent pair. 

The fully developed compensation-sac (fig. 62, c.s.) occu- 
pies a large proportion of the body-cavity. Fig. 61 repre- 
sents a compensation-sac which is about half grown, with its 
parietal muscles. The sac does not extend to the semi- 
circular operculum, the straight base-line of which appears 
to constitute the hinge. It develops in a Lepralioid manner, 
a small, apparently solid mass of cells appearing beneath the 
newly formed median pore; and to this mass the parietal 
muscles radiate from various parts of the zocecium. Some- 
what earlier the uncalcified, oblique, terminal wall of the 
zocecium extends as a triangular point as far as the median 
pore, so that the sac is probably derived from the proximal 
part of the orifice. The suggestion has often been made 
(cf. Hincks, 1880, p. 257) that the Microporellid median pore 
has been formed by the closure of the sinus of a Schizo- 
porella-like ancestor. The above-described immature stage 
of C. gracilis appears to point in this direction, 

The occlusor muscles (occl.) are easily seen, but I have 
found no divaricators. 


312 SIDNEY F. HARMER. 


Calwellia sinclairii,! Busk (fig. 60). 


Waters (1888, p. 17) has described the compensation-sac 
in this species, as well as in C. bicornis and Urceolipora 
dentata, and he correctly states that 1t opens to the exterior 
by means of the median pore. I agree with Waters that 
C. sinclairii should be placed in the genus Calwellia. It 
differs from C. gracilis principally in the fact that the 
proximal half of the zocecium is almost as broad as the 
distal half, and the zocecia do not therefore appear to 
be so definitely arranged in pairs. 

The operculum is a segment of a circle somewhat greater 
than a semicircle, and having a straight base-line. From the 
middle line of this a longitudinal suture (s.) passes to the 
transversely elongated median pore (m.p.), and on each side 
of the suture is a round fenestra. The suture can usually be 
traced down the entire length of the zocecium. ‘The position 
of the lateral horns of C. gracilis is indicated by a fenestra 
(f.), in the neighbourhood of which are one or two smaller 
pores. The median pore is somewhat crescentic, with an 
anterior concavity. The lateral edge of the large compensa- 
tion-sac does not pass to the base of the operculum, but turns 
sharply inwards at the level of the median pore, and runs 
into the broad tongue-like lobe of the wall of the zocecium 
which gives the pore its crescentic form. 

The operculum has a well-marked triangular lobe of its 
vertical marginal flange (/l.), into the apex of which the 
occlusor muscle is inserted, as in Flustra. 


Calwellia dentata,? MacGillivray. 


It appears to me that this species should be separated from 
U. nana,’ and placed in Calwellia, the generic character of 
which would then have to be amended so far as relates to the 
arrangement of the zocecia. In C. dentata these are placed 

1 Busk (1884), p. 103. 
2 *Prodr. Zool. Vict.,’ Dee. xi, 1885, p. 19. 
3 As suggested by Waters (1888, p. 10). 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 313 


back to back, alternately, all the zocecia looking towards one 
or the other edge of the branch. The zocecial characters 
agree with those of C. sinclairiil. There is a large compen- 
sation-sac, opening by a median pore, which is transversely 
elongated, but the concavity of which is on the proximal 
side. Between this and the operculum are two fenestra, 
placed as in C. sinclairii. In addition to two lateral horns, 
similar to those of C. gracilis, there are three other oral 
spines placed more distally. 


Ichthyaria oculata, Busk! (fig. 59). 


There can be no doubt that this species was incorrectly 
placed by Busk in the Bicellariide, a family with which it 
appears to have no affinity. Its more natural position would 
probably be in the Calwelliide,? with which it agrees in the 
characters of its -zocecium and of its ovicell. he zocecia are 
arranged in a double alternate series, all facing in one 
direction. The median pore (im. p.) is circular, and is situated 
on the axial side of the middle of the zocecium. It is con- 
nected with the operculum by a longitudinal suture (s.), 
on each side of which is a funnel-shaped fenestra, as in 
Calwellia. In correlation with the asymmetrical position 
of the median pore the abaxial fenestra is much further from 
the suture than is the other one. There are one or two 
distal fenestrae, probably indicating the former development 
of oral spines. 

The compensation-sac (c.s.) 1s arranged as in Calwellia. 
It opens by the median pore, and therefore ends at some 
distance from the operculum. The parietal muscles (p.m.) 
occur as a series of definite groups, asin Flustra. Calcifi- 
cation is deficient along a line extending the whole length 
of the basal side of the zocecium. ‘The line is narrow for the 
greater part of its course, but dilates at each end. The 
older parts of the colony are supported by a calcareous 

1 Busk (1884), p. 46. 

2 Waters (1888, p. 10) places it in Calwellia, a view which seems not 
unreasonable. 


314 SIDNEY F. HARMER. 


thickening, which runs along each margin of the branch. 
The retractor muscles of the polypide originate from the two 
lateral walls. On each side there is a strong fan-shaped 
group passing to the base of the tentacles, and another 
passing to the junction of the pharynx with the cesophagus. 


Onchoporella bombycina, Busk!? (not Ell. and Sol.). 


Although I have had only dry material of this species to 
work with, I can state that there is a well-developed com- 
pensation-sac, opening by the crescentic median pore, the 
tongue-like lobe to which the crescentic shape is due being 
distal, and seeming to be a triangular membrane which curves 
into the distal border of the compensation-sac (as in fig. 60). 
‘he operculum is of very delicate texture. On either side of 
the orifice, somewhat proximally, there is a short spine, on 
the proximal side of which is a pore, which is either a simple 
funnel-shaped deficiency of the calcareous wall, or it consists 
of two, or even three, funnel-shaped tubes contained in a 
single longitudinal groove. <A similar arrangement is shown 
in fig. 60 (f.). One or two pores of the same kind may 
continue the line of this groove down the frontal surface of 
the zocecium. 

‘he branches have a calcareous marginal thickening, which 
may have a zigzag course ; it is composed of numerous small 
pieces, an arrangement which clearly gives a certain amount 
of flexibility to the frond. 


Microporella malusii,”? Aud. (fig. 63). 

The median pore (m. p.) has a distal concavity, due to the 
projection into it of a tongue-like process which is partly 
membranous. Between this and the straight base-line of the 
semicircular operculum is a triangular group of pores, the 

1 Busk (1852), p. 52. The species described by Ellis and Solander (‘ Nat. 
Hist. Zooph.,’ 1786, p. 14), from the Bahama Islands, appears to be a 
Euthyris. 

2 Hincks (1880), p. 211. 


THE MORPHOLOGY OF THE CHEILOSTOMATA, 31,5) 


apex of the group pointing to the median pore. Round the 
edge of the zocecium is a single line of pores. This arrange- 
ment is constant in my Neapolitan specimens. The communi- 
cation pores open into “ pore-chambers” (p. ¢.), as pointed 
out by Levinsen (1891, pp. 250, 285). 

The compensation-sac (c. s.) is thin-walled, its outline being 
somewhat heart-shaped when seen from the basal surface. 
The median pore opens into it considerably on the proximal 
side of its distal margin, a triangular bay (corn.) of the 
body-cavity indenting the membrane of the compensation-sac 
in the immediate neighbourhood of the frontal surface. The 
parietal muscles (p. m.) are strong, and are arranged in 
distinct groups. 

The function of the median pore in M. malusii has fre- 
quently been discussed, and my results do not agree with 
those of any of my predecessors. Jullien (1888, 4, pp. 
36, 39) gives the name “fenestrula” to the median pore, 
which he describes as forming the narrow end of a funnel 
(“cornicula”’), the other end of which opens into the tentacle 
sheath. Jullien figures the cornicula in pl. xy, fig. 2, from 
which it appears to me clear that the funnel is the part of the 
body-cavity which I have described above (fig. 68, corn.) as 
indenting the distal outline of the compensation-sac. 

Pergens (1889, p. 507) does not believe in the existence of 
Jullien’s compensation-sac, and describes the diaphragm or 
vestibule as opening into the body-cavity and not into the 
tentacle sheath. 

The commonly received opinion that the vestibule is 
traversed by the tentacles during their protrusion is, 
however, based on too many observations to be lightly 
dismissed. 

Pergens further states (p. 506) that the median pore of 
M. malusii opens into the body-cavity. The “ longitudinal 
muscles’ mentioned by him were probably parietal muscles, 
while I suspect that the appearance of a “sack ” between 


1 For the pore-chambers of Cheilostomata see also Waters (1898), 
p. 658. 


316 SIDNEY F. HARMER, 


the operculum and the median pore can be explained in the 
same way as Jullien’s “cornicula ” mentioned above. 
Levinsen (1891, p. 285) states that the median pore, like 


the other pores, is closed by a membrane. 


Microporella ciliata,! Pall. 


The compensation-sac is like that of M. malusii, which 
this species further resembles in the grouped condition of the 
parietal muscles. I have not made out the details of the 
opening of the sac. There is a membranous portion in the 
tongue of the crescentic pore, as in M. malusii. 


(ze) Microporoid Genera. 


Micropora, sp.” (figs. 57, 58). 

The large lateral foramina serve for the transmission of 
the tendons of the depressor muscles (fig. 58, depr.) of the 
frontal membrane, as correctly stated by Jullien (1888, 4, 
pp. 77—81, pl. xiv, fig. 1), who terms them “ opesiules.” 
‘he opesiule is a complete calcareous tube (/.7.), the base of 
which joins the basal wall or the lateral wall of the zocecium, 
the two conditions often occurring on opposite sides of the 
same zocecium. ‘The tube usually gives off a curved lamella 
(fig. 57) on its distal side, near the basal surface. From the 
recess thus formed originates the occlusor muscle (ocel.), 
whose long tendon passes obliquely to be inserted into one of 
the lateral corners of the vertical flange of the operculum. 
The depressor muscle (depr.) is much stronger. Its fibres 
converge to a tendon which passes through the opesiule 
and is inserted into the frontal membrane. I regard these 
muscles as a special development of the parietal muscles, 
which are otherwise unrepresented. The calcareous frontal 
wall (erypt.) is complete, and the relations of the depressor 
muscles indicate that it is a cryptocyst. 


1 Hinceks (1880), p. 206. 
= Torres Straits, A. C. Haddon Coll. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 317 


Steganoporella alveolata, Harmer:.! 


I have previously alluded? to the muscular system of this 
species. That of the B-zocecia is more highly developed 
than in any other Polyzoon with which I am acquainted. 

The B-operculum has the form shown in figs. 64, 66. Its 
mode of articulation differs in a striking way from that of 
the Cheilostomes previously considered. Although its base 
is continuous with the unmodified frontal membrane which 
stretches across the whole of the proximal half of the 
zocecium, a strong basal sclerite (b. s.) is differentiated. Into 
the two ends of this are inserted, beneath the frontal mem- 
brane, two ligaments (lig.) which, continuing the direction 
of the sclerite, become attached to the strong condyles (cond.). 
The operculum is thus slung by two tight transverse cords 
between the two condyles, and this arrangement constitutes 
the hinge. The basal sclerite probably has the further 
object of playing over the projecting part of the structure 
which I have called the median process (fig. 66, m. pr.), and 
of preventing the operculum from being pulled too far by 
the contraction of the enormous occlusor muscles. Of these 
there are two pairs, which originate, with the divaricator 
and depressor muscles, deep down within the lateral recesses 
at the sides of the median process. ‘The insertion of the 
occlusors is shown in fig. 64. One pair (occl.), which pro- 
bably correspond with the normal Cheilostome occlusors, 
pass by strong tendons (tend.) into the occlusor tubercles 
(ocel. t.) of the operculum, but the tendon is also continuous 
with a fascia (7.), which is connected with the strongly pro- 
jecting proximal end of one of the main sclerites (m. s.) of 
the operculum. The second pair, which I distinguish as the 
distal occlusors (occl.’), are inserted into a strong transverse 
fascia (f.’) which lies between the two main sclerites. In 
fig. 64, which represents an operculum torn away from its 
zocecium, this fascia has been displaced from its natural 

1 1900, p. 287. 
2 Loc. cit., p. 230. 


318 SIDNEY F. HARMER. 


position, which is shown in figs. 65, 66. From these it is 
apparent that the median part of the fascia is exactly trans- 
verse and vertical when the operculum is closed, and that on 
each side it curves round in order to be attached to the 
proximal end of the main sclerite. It thus results that while 
not much more than the edge of the middle part of the fascia 
is seen in a view from the basal or from the opposite side, 
the surface of the lateral parts is seen under the same con- 
ditions. I have been unable to ascertain with certainty 
whether the median part of the fascia is inserted into the 
operculum or whether it has a free edge in this region. 

The origin of the muscles can best be seen in a basal view 
of a zocecium from which the basal wall has been removed in 
the way recommended on p. 265. The lateral recesses (J. 7.) 
are here seen to be almost entirely filled by four strong 
muscles. Of these the distal occlusors (occl.’) are far the 
strongest, and they originate from the distal half of the floor 
of the lateral recesses. The proximal or normal occlusor 
(occl.) originates from the proximal half of the lateral recess, 
in its outer part. The course of its tendons (fend.) is indi- 
cated in figs. 64—66. 

The distal occlusors, in helping to close the “ orifice,” 
probably tend to pull the whole operculum towards the distal 
end of the zocecium, since their fascia is inserted largely into 
the projecting proximal ends of the main sclerite (figs. 695, 
66). he proximal occlusors correct this tendency, as may 
be inferred from the direction taken by their principal 
tendons. When it is remembered that in this species each 
of the strong teeth of the operculum fits, when closed, into 
its own socket in the calcareous part of the zocecium, it is 
obvious that the operculum must shut with great accuracy, 
and a proper correlation between the contractions of the two 
sets of occlusors no doubt has to be maintained. ‘The 
accuracy of the closure is probably helped by the movement 
of the basal sclerite over the median process, and, of course, 
by the transverse ligaments uniting the operculum to the 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 319 


condyles, as well as by the tension of the frontal membrane ! 
continuous with the base of the operculum. 

Although the distal occlusors are larger than the proximal 
pair, their tendency to shift the operculum from its proper 
position is probably adequately corrected by the proximal 
occlusors, which, although weaker than the others, have a 
more effective pull, in consequence of the greater obliquity 
of their tendon. 

It appears to me that in certain positions assumed by the 
operculum, the distal occlusors would work to less advantage 
than the others. When the operculum is opened so far that 
its outer surface forms a right angle with the frontal mem- 
brane, the projecting ends of the main sclerites extend far to 
the distal side of the hinge. ‘The operculum seen from its 
edge would have the appearance of a right-angled triangle, 
of which the hypothenuse is formed by the main sclerite, and 
a short side by the piece connecting the proximal end of the 
same sclerite with the base of the operculum. Both pairs of 
occlusors now probably pull, by means of their fasciz inserted 
into the projecting ends of the main sclerite, on the short side 
of the triangle which is, in fact, an arm of a right-angled 
lever hinged at its angle. In so acting, the proximal occlusor 
probably pulls over the ligament (lig.) which connects the 
operculum with the condyle, the ligament thus forming a 
kind of pulley. The distal occlusors are probably specially 
important in the final stages of the closure. 

‘'wo muscles remain to be described. The depressor 
muscles (figs. 65, 66, depr.) originate from the innermost part 
of the lateral recess. ‘I'heir tendon passes obliquely towards 
the proximal end of the zocecium, ascending at the same time 
to reach the frontal membrane, into which it is inserted (fig. 
66). Between the depressor and the proximal occlusor is a 
group of muscles which constitute the divaricator (div.). 


1 In my former paper | described this as the epitheca, a more general term 
which it would be better not to use in cases where the homology of the mem- 
brane with the frontal membrane of a Membranipora seems to be demon- 
strated. 


320 SIDNEY F. HARMER. 


At its origin this muscle is often difficult to distinguish 
from the depressor, but its tendon takes a different direction 
and is inserted into one of the ends of the basal sclerite of 
the operculum, or perhaps into the frontal membrane imme- 
diately adjacent to the basal sclerite. 

I regard the divaricator and depressor muscles as a special 
differentiation of the parietal muscles of other Cheilostomes, 
and they are probably together the representatives of the 
depressor muscles of Micropora (figs. 57,58), which occupy 
special cavities that may be compared with the lateral 
recesses of Steganoporella.! 

In the A-zocecia the distal occlusors seem to be completely 
absent. The zocecium is more like that of an ordinary 
Cheilostome, with a single pair of occlusor muscles. 

In view of the possibility that the B-zocecia of Stegano- 
porella may represent avicularia,! it is worth mentioning that 
in the gigantic avicularia of Flustra abyssicola (Davis 
Straits) I have noticed that the occlusors are in two groups. 
Of these the distal group consists of an enormous mass of 
muscle which fills up more than the distal half of the 
avicularian zocecium, and its fibres are inserted into a broad 
median tendon. ‘The proximal occlusors are paired ; their 
tendons are at a deeper level than that of the distal occlusor, 
and they converge towards the base of the operculum. 
Jullien (1888, 4, pl. xii, fig. 3) indicates the occurrence of two 
sets of occlusors in the avicularium of Beania magel- 
lanica. 


V. ae Primary Zoacium on ANCESTRULA. 


The zocecium which results from the metamorphosis of the 
larva is frequently found to show characters which differ 


1 This conclusion is confirmed by the figure given by Waters (1885, 
pl. xiv, fig. 42) of the mandible of Onychocella angulosa (? =antiqua), 
in which the formation of the hinge appears to be like that foundin Stegano- 
porella, while a fascia shown in the same figure closely resembles that of the 
distal occlusors of Steganoporella. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 321 


from those of the ordinary individuals of the colony. This, 
the primary zocecium, has been termed the “ ancestrula”’ by 
Jullien (1888, 4, pp. 27, 29), a name which he appears to 
have used merely to convey the idea that it was the actual 
ancestor (“zocecie mére”’) of the other individuals of the 
colony. Smitt (1868, 1, p. 306) described the same thing as 
the “Tata form” of the Cheilostomes,! and clearly recognised 
their importance, stating that “Tata is for the Cheilostomes 
what Alecto showed itself to be for the Cyclostomes.”’ 

Tam fully of Smitt’s opinion with regard to this matter. 
The “ ancestrula”’ is in all probability ancestral in the sense 
of retaining characters of phylogenetic importance. Smitt 
returns to the same subject in a later publication (1896), in 
which he again maintains the importance of the Tata form. 
Neviani (1898, p. 165) supports the same conclusion. 

The frequency with which the Tata-like ancestrula (figs. 
2, 6, 7) occurs in Cheilostomata is indeed remarkable, and 
there is a curiously small range of variation in the number of 
its marginal spines. I subjoin a few references to cases 
where an ancestrula of this type has been described. 


Number of 
Species. Reference. marginal Remarks. 
spines. 


Kleetra pilosa, L. Barrois, 1877, p. 241, 5—7 
pl. xv, figs. 5—8 
Membranipora flem-|Hincks, 1880, p. 163 8—9 

ingil, Busk 
Membranipora crati-|Hincks, 1880, p. 147 | About 9 

cula, Alder 
Scrupocellaria scru-/Smitt, 1868, 1, p. 320, 8 |ERlongated. 


posa, L. pl. xvii, fig. 42, ete. 

Scrupocellaria sca-/Smitt, 1868, 1, p. 314, 9 3 
bra, Van Ben. pl. xvii, fig. 27 

Scrupocellaria rep-|Barrois, 1877, p. 183) — * 
tans, L. pl. x, fig. 16 


Acad. Roy. Belgique,’ xvi, pl. ii, fig. 14) are primary zoccia of a Mem- 
branipora, 


vo. 46, PART 2.—NEW SERIES. Ww 


OS 
ine) 


SIDNEY F. HARMER, 


Species. 


Menipea ternata, Hil. 
and Sol. 

Bicellaria ciliata, L. 

Bugula sabatieri, 
Calvet 

Membraniporella 
nitida, Johnst. 


Cribrilina punctata, 
Hass. 

Cribrilina 
ceros, Busk 

Schizoporella spini- 
fera, Johnst. 


mono- 


Schizoporella  uni- 
cornis, Jolnst. 
Schizoporella crus- 


tacea, Smitt 

Mucronella coccinea, 
Abildg. 

Mucronella coccinea, 
Abildg. 

Mucronella_ peachii, 
Johnst. 


Mucronella vario- 
losa, Johnst. 

Porella concinna, 
Busk 

Porella concinna, 
Busk 


Phylactella labrosa, 
Busk 
Microporella ciliata, 


Pall. 


Microporella 
lusii, Aud. 


ma- 


Microporella ma- 
lusii, Aud. 


| pl. xvi, fig. 15 
pl. xviii, fig. 1 
(fig. 13) 

Hincks, 1880, p. 


pl. xxvil, fig. 6 


Hincks, 1880, p. 191 


fig. 9 


pl. viii, fig. 37 
fig, 13 

Waters, 1900, p. 
pl. vill, fig. 40 
fig. 17 
fir. 16 


Hincks, 1880, p. 366 


Barrois, 1877, p. 155, 
pl. vil, fig. 16 
Hineks, 1880, p. 357 


‘Hineks, 1880, p. 207, 
pl. xxviii, fig. 3 


Hincks, 1880, pp. 211, 
212, figs. 10, 11 


Neviani, 1898, p. 6, 


fig. 3 


“A few” 


9, of 
which 7 
are oral 
About 7, 

of which 2 
are oral 


10 


Number of | 
Reference. marginal Remarks. 
spines. 
Smitt, 1868, 1, p. 306, 10—11 | 
Smitt, 1868, 1, p. 335, 13 \tafandibuliferane 
iCalvet, 1900, p. 112 5 
201,) 14, of 
which 4 
are oral 
12 
Jullien, 1888, 4, pl. iti, 9 
Hincks, 1880, p. 242, 9—10 
fig. 13 
Barrois, 1877, p. 152, 8 
Waters, 1900, pl. viii) ?9 
158, 13 
Hincks, 1880, p. 378, Tat 
Hincks, 1880, p. 361,) 11, of Modified Tata. 
which 6 
are oral | 
1O—11 |? Tata. 
Smitt, 1868, 2, p. 1384,, 7—9 |Variable. 
pl.xxvi,figs. 109—11] 
8 


? Tata. 


Variable. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. aoe 


Cases in which the ancestrula is not Tata-like: 


Species. Reference. Remarks. 


MembraniporellaHincks, 1880, p. 203 |Aperture membranous, but 
melolontha, Busk not surrounded by spines. 
Cribrilina annulata, Hincks, 1880, p.194 | Like the ordinary zoocia. 
Fabr. 
Lepralia pallasiana, Hincks, 1880, p. 298 | Like the ordinary zocecia. 
Moll 
Lepralia pallasiana, Barrois, 1877, p. 145, Like the ordinary zocecia. 
Moll pl. vu, fig. 1 
Schizoporella (Hip-Barrois, 1877, p. 168, Like the ordinary zowcia. 
pothoa) hyalina,L.) pl. ix, figs. 9—16 
Schizoporella (Hip- Jullien, 1888, 4, p. 29,Tata-like; resembling the 


pothoa)hyalina,L.| pl. iv, figs. 1—4, ordinary zocecia; or inter- 
mediate. 
Schizoporella cris-Hincks, 1880, p. 254, Intermediate. 
tata, Hincks pl. xl, fig. 6 @ 


Smittia reticulata, Hincks, 1880, p. 346 Like the ordinary zocecia. 
J. MacGill. 
Temachia opulenta,Jullien, 1883, p.13(sep.),,Cribrilina-like. 
Jullien pl. xiv, figs. 26, 29 


There is no other form of ancestrula which is known to 
occur in a number of different types of Cheilostomes, and it 
is impossible not to agree with Smitt and others that the 
view that the primary zocecium shows ancestral characters 
has much to be said for it. In the majority of described 
cases the ancestrula is like that shown in fig. 2, and is a 
typical Membranipora-like zocecium. In certain forms 
with an infundibuliform zocecium the ancestrula has already 
acquired this type. This is the case in Bicellaria ciliata, 
while in Scrupocellaria the ancestrula may show distinet 
tendencies in the same direction. It is hardly necessary to 
point out how thoroughly the common occurrence of a Tata- 
hke ancestrula confirms the general conclusions arrived at in 
this paper. It may be remarked that this type of ancestrula 
is by no means unlike that of a Cyclostome (which has, 
however, no spines); and these Polyzoa may probably be 
regarded as forms in which the frontal surface remains 


324 SIDNEY F. HARMER. 


circular and unealcified, while the greater part of the zocecium 
is constituted by the cylindrical lateral walls. 

I do not know of any Flustrine form in which the ancestrula 
is other than Flustrine in respect of its frontal wall. In 
Membraniporella nitida the ancestrula is Tata-like, and 
shows clearly that the four oral spines are of the same nature 
as the remaining ten marginal spines figured by Hincks. 
This species is particularly interesting in the fact that some 
of the older members of the colony produced by budding 
may retain the Membranipora-lke condition (Hincks, 
1880, pl. xxvii, fig. 7), and possess a series of irregular 
marginal spines which do not unite. The remaining zocecia 
have their frontal spines arranged with the regularity 
characteristic of the Cribrilinide in general. 

The consideration of the Cheilostomata in which the 
ancestrula is not Ta ta-like is almost as instructive as that of 
the first set of cases. 

In Hippothoa hyalina, Jullien (1888, 4, p. 29) states 
that the characters of the ancestrula are variable. In Kuro- 
pean specimens the frontal wall is fully calcified. ‘he same 
condition was found in specimens from ‘ierra del Fuego, 
although in most of the colonies from this locality the 
ancestrula was a circular 'ata-like zocecium, with a large 
membranous aperture bordered by six inwardly curving 
marginal spines (Jullien, pl. iv, figs. 1—8), while in a third 
case (ibid., fig. 4) the ancestrula is at first Tata-like, but the 
membranous aperture subsequently calcifies, so that the orifice 
normal for the species is contained in a calcified oral region 
surrounded by the circlet of marginal spines. ‘This leads to 
the form described by Julhen as Diazeuxia reticulata 
(ibid., fig. 5), in which the orifice of the ancestrula is 
completely Schizoporelliform. 

Neviani (1898, p. 165) describes three varieties of the ances- 
trulaof Microporella malusii. The first of these isan ordi- 
nary 'l'ata form with ten marginal spines. ‘The second more 
nearly resembles the normal zocecium of this species, but 
agrees with one of the ancestrule described by Jullien (see 


THE MORPHOLOGY OF THE CHEILOSTOMATA. Seo 


above) in having the orifice contained in a calcareous region 
which is surrounded by eight short spines. The third is parti- 
cularly interesting in having a Schizoporella-like orifice. 

Schizoporella cristata, according to the description 
given by Hincks, is another form with an intermediate type 
of ancestrula. The Schizoporelliform orifice is contained in a 
calcareous oval area, on the proximal side of which are three 
spines, while the orifice is provided with six oral spines. 

The ordinary zocecia of the form described by Jullien as 
Temachia opulenta are lageniform, and without spines, 
somewhat resembling those of Mucronella abyssicola. 
The ancestrula is obviously Cribriliniform. I have myself 
observed an analogous case. A Cribrilina (sp.?) from 
Funafuti, in the collection of the British Museum, has an 
irregular arrangement of pores on its frontal surface, some- 
thing lke that found in C. monoceros. The primary 
zocecium is Membraniporella-like, with a perfectly 
regular double series of transverse bars, separated by 
simple slits. 

Fig. 2 shows the ancestrula of a colony of Microporella 
malusii, from Naples. The frontal surface and the oper- 
culum are Flustrine. The marginal spines, some of which 
have been lost, were ten in number (as in one of Neviani’s 
cases). They originate, like the frontal spines of Cribrilina, 
from foramina in the calcareous wall, a little outside the 
frontal membrane. The spines resemble the three oral 
spines with which most of the other zocecia of the same 
colony are provided. 

An ancestrula of M. ciliata, also from Naples, is somewhat 
similar; but it differs in having a distinct “area,” or calcifi- 
cation of the proximal part of the frontal surface. The living 
tissues have disappeared, and I am, unfortunately, unable to 
state certainly whether the calcified area is a cryptocyst or 
not. This is, however, indicated by the way in which it 
slopes down steeply from its periphery, so that its inner edge 
lies at a deeper level than its outer margin. ‘There are ten 
long marginal spines. 


526 SIDNEY F. HARMER. 


One or two colonies of M. impressa (S. Devon) are par- 
ticularly interesting in having a Schizoporelliform ancestrula. 

In Schizoporella vulgaris (Naples) I have found an 
ancestrula (fig. 6), like that of Microporella ciliata, with 
ten marginal spines and a calcified plate (crypt.) which 1s 
probably a cryptocyst. This resembles the cases which I 
have described as “ intermediate” between the typical Tata 
form (fig. 2) and the adult condition with a completely 
calcified frontal wall. It appears to me that these ‘ inter- 
mediate” Tata forms deserve special attention. It remains 
to be seen whether the calcified plate internal to the circlet 
of spines is really a cryptocyst or not. Taking into considera- 
tion the frequency of the occurrence of a cryptocyst in the 
Amphiblestrum group of Membranipora (although 
the morphology of this part has yet to be determined in many 
of these cases) and in such a typical Membranipora as 
M. delicatula, Busk (e. g.)—and noticing, moreover, the 
presumed occurrence of a cryptocyst in Bicellaria grandis 
(fig. 1),—the presumption appears to be that the plate marked 
crypt. in fig. 6 is really a cryptocyst. ‘The bearing of these 
cases on the phylogeny of the Cheilostomata is considered in 
a later part of this paper (p. 304). 

Cribrilina radiata,! Moll.—The study of a young colony 
of this species has led me to altogether unexpected conclu- 
sions, Which probably indicate that this species is not really a 
Cribrilina at all. The ancestrula (fig. 7) is a Tata form 
with eleven marginal spines, of which the proximal (sc.) is broad 
and somewhat scutum-like, lying horizontally over the frontal 
membrane, while the others are erect or somewhat reclined 
outwards. Hach of the erect spines bears a minute calcareous, 
denticulated lobe which originates about halfway up the spine, 
and lies horizontally, directed inwards over the frontal mem- 
brane. ‘The outline of the ancestrula is somewhat concealed 
by the basal lobes of the younger zocecia. 

Two zocecia in contact with the ancestrula are smaller than 
the rest, and are doubtless those which were first formed by 

‘ Hincks, ‘ Brit. Mar. Pol.,’ 1880, p. 185. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. si 


budding. One of these (z.’) is on the distal side of the 
ancestrula, and the other (not drawn) is on its apparent left 
side. On each of the five oral spines of these zocecia is a 
minute calcareous lobe resembling those of the marginal 
spines of the ancestrula. The lobes may be indicated by 
still smaller projections on the oral spines of the other 
Zocecia, 

In all the ordinary individuals there occur, on the proximal 
side of the operculum, a pair of delicate, membranous, 
elongated structures (sp.), which are described by Hincks 
(1880, p. 186) as vibraculoid setz. Smitt! ascribes to them 
a sensory function. 

The structures in question have no claim to be regarded as 
vibracula. There is no trace of any muscles connected with 
them, and there can be no doubt that they continue the line 
of the oral spines. ‘his line is, in reality, prolonged all 
round the zocecium, the succeeding spines (sp.’) being delicate 
membranous papille which barely project abovethe calcareous 
surface, and are obviously in series with the oral spines and 
the “ vibraculoid setve.”” The base of each papilla is a pore in 
the calcareous frontal wall. Comparison with the ancestrula 
and with the calcareous lobes of the oral spines of the younger 
zocecia seems to indicate that the membranous papillee, together 
with the oral spines, correspond with the marginal spines of 
the ancestrula. ‘The calcareous lobes of that individual are, 
in consequence, the representatives of the bars which have 
united to form the frontal wall of the ordinary zocecia, while 
the scutum-like spine of the ancestrula is probably the 
equivalent of one of the denticulated lobes on the other 
spines. 

It has usually been assumed that the radiating series of 
pores correspond with the slits between the bars. ‘his is 
certainly the case in C. philomela (fig. 8); but im C. 
radiata (fig. 7) there can be no doubt that the pores are in 
the same radii as the membranous marginal spines. I can 

1 © Floridan Bryozoa,” I, ‘ K. Svenska Vet.-Ak. Handl.,’ xi, 1878, No. 4, 
p. 22. 


328 SIDNEY F. HARMER. 


find no indication of a double calcareous wall in this species, 
and it therefore appears to me that the frontal wall is in a 
Lepralioid condition, the inner calcareous wall alone being 
developed, and being perforated by a series of vacuities, 
while the outer wall remains membranous. The probability 
that a layer of living tissue (epitheca) overlies the calcareous 
wall is increased by the fact that the surface of the latter is 
ridged, a condition which is usually associated with a covering 
of living tissue. 

The lateral junctions of the frontal bars are indicated by 
prominent radial ridges, each of which rises to a small tubercle 
just inside the line of the membranous papille above described. 
The pores consequently le, as described by other observers, 
in radiating furrows. That the union between the bars has 
not been complete is indicated by the fact that a thin line of 
air in some cases underlies the ridge. 

The consideration of these facts suggests that the frontal 
wall has here not been formed by the overarching of the 
main spines, but by the development of the inner lobes of 
branched spines. It is impossible not to be reminded of the 
condition described by Jullien (1886, p. 609) in the 
Cretaceous Steginoporide, where a calcareous wall! is 
formed by the growth of branched peristomial spines. 

I have, unfortunately, been unable to make out the con- 
dition of the compensation-sac in C. radiata, but it is well 
known that this species may possess a well-marked median 
pore (m. p.). Inaddition to this Microporelloid feature, it may 
be noticed that the zocecium is surrounded by a flat basal lobe 
formed of pore-chambers,” as in Microporella ciliata and 
M. malusii. These are not present in C. philomela. 

There can, I think, be little doubt that C. radiata cannot 
be retained in the same genus with C. philoinela, and it is 
possible that its affinities are rather with the Microporellide. 

Cribrilina figularis, Johnst. (Plymouth), is another 
species in which, so far as can be judged from dry material, 

1 Which, however, would not appear to correspond with that of C. radiata. 

? The details of the pore-chambers are not indicated in fig. 7. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 529 


there is evidence that the frontal shield is formed by a 
series of horizontal lobes developed from the main spines. 
These latter are indicated with great distinctness as a curved 
series of conical calcareous papille, the truncated end of 
which is a membranous fenestra, and stands up at a higher 
level than the frontal shield, which is encircled by the 
papillae. ‘The radiating series of pores here correspond with 
the intervals between the spines, as in C. philomela. 

The existence of great differences between the opercula of 
different species at present referred to Cribrilina suggests 
that the genus is an unnatural one, representing a stage in 
the evolution of the Lepralioid zocecium, which has been 
arrived at independently in several cases, and is merely the 
result of incomplete lateral fusion of the bars which compose 
the frontal shield. 


VI. CLASSIFICATION OF THE CHEILOSTOMATA. 


While recognising the danger of drawing taxonomic con- 
clusions from the study of a single organ, I cannot but think 
(with Jullien) that the consideration of the frontal surface is 
capable of settling some questions of this kind in the Cheilo- 
stomata. It would, I think, be premature to propose a new 
classification, but the observations above described suggest 
certain lines along which the existing classifications may be 
amended. I confine myself to groups to which I have paid 
some attention, leaving entirely out of consideration a large 
proportion of recent genera. 

(1) The Membraniporidez and Flustride are indicated as the 
most primitive group of Cheilostomes by the relations of their 
frontal membrane and parietal muscles, and by the evidence 
afforded by the primary zocecium of many other Cheilostomes. 
For this group I think that Smitt’s name Flustrina should be 
retained. It remains to be seen whether many of the 
Amphiblestrum-like forms of Membranipora, in which 
a calcareous cryptocyst (7) is developed, are more nearly 
related to this group or to the Microporoid series. 


330 SIDNEY F. HARMER. 


In many of the Flustrina there is an elongated frontal 
membrane which is typically depressed by a series of parietal 
muscles. A frontal membrane of this description occurs in 
the Membraniporide (part), Flustride, and Farciminariide, 
and in such genera as Bugula and Beania. 

In other Flustrine forms (Dimetopia, Bicellaria) the 
zocecium becomes elongated and usually infundibuliform, 
with a concurrent reduction in the size of the frontal mem- 
brane—a tendency which may be shown even in the ances- 
trula. The reduction of the frontal surface may be correlated 
with a reduced number of parietal muscles. 

(2) Another series seems to be constituted by Scrupo- 
cellaria, Menipea, and Caberea, a group of “ Cellularine” 
forms in which certain species in each genus have the frontal 
membrane protected by the excessive development of a single 
lateral spine, the fornix or scutum, which may, however, be 
vestigial or absent. In some cases, as in Menipea 
jeffreysii,! Norman, the scutum is so largely developed 
that a complete calcareous frontal wall appears to exist. 
Round the edge of this is a slit which is the interval between 
the edge of the oval scutum and the calcareous margin of 
the frontal membrane. The scutum originates from a broad 
calcareous base just proximal to the operculum, and on its 
inner side. It has a nearly straight suboral edge, which 
meets the basal sclerite of the approximately semicircular 
closed operculum. On the outer side the scutum locks firmly 
into a strong tubercle developed from the edge of the 
zocecium. If the scutum were to fuse completely with the 
edge of the frontal surface a compensation-sac would be 
produced which would have had a different phylogenetic 
history from that of many other Cheilostomes. I see no 
reason why cases of this kind should not exist,” and it is 
possible that evidence to that effect might be forthcoming in 


1 Norman, ‘ Ann. Mag. Nat. Hist.’ (6), xii, 1893, p. 446. 

2 The figure given by Waters (1888, pl. i, fig. 1) suggests that this might 
be the case in Catenaria bicornis; and the structure of C. lafontii 
(fig. 49) is not irreconcilable with this view. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. Sal 


the existence of a single communication between the living 
tissue of the frontal wall and that of the body-cavity, corre- 
sponding with the base of the scutum. 

The group is usually characterised by the high develop- 
ment of avicularia and vibracula. 

(3) The Microporoid series includes Micropora, Thala- 
moporella, Diploporella,’ and Steganoporella, all 
characterised by the development of a cry ptocyst? or calcareous 
lamella which grows beneath the frontal membrane. In the 
forms which have been examined, the parietal muscles are 
reduced to a single pair of highly developed depressor 
muscles which commonly traverse opesiules® to reach the 
frontal membrane; or, in Steganoporella, to a pair of 
depressor muscles and a pair of divaricators. The series 
probably includes a large number of the Cretaceous forms 
allied to Onychocella,as well as the recent O. abyssicola, 
Smitt, etc. Some of the recent species sometimes placed in 
the Membraniporide may also belong to this group. 

(4) The Cribrilinide, including Membraniporella and 
Cribrilina (part), are forms in which a calcareous frontal 
wall is developed, ontogenetically, by the overarching of a 
series of marginal calcareous spines. This process has been 
observed by various authors in a certain number of species. 
I have shown above that C. radiata differs in important 
respects from the typical Cribrilinidee. 

(5) In Umbonulaverrucosa the frontal shield originates 
ontogenetically as a calcareous lamina, perforated by a con- 
spicuous series of marginal pores, which grows over the 
Flustrine frontal membrane, leaving a wide space between 
itself and that membrane. ‘The parietal muscles develop in 
the situations which they will occupy in the mature zocecium. 
I consider that the marginal pores correspond with the 
origins of as many Cribrilina-like spines which have 
united to form the frontal shield. The external calcareous 


1 MacGillivray (1887), p. 207. 
2 Cf. Harmer (1900), p. 228. 
3 Tbid., p. 230. 


Su SIDNEY F. HARMER. 


layer of the spines is now represented only by a membranous 
epitheca, composed of living tissue bounded by an external 
cuticle, while the frontal shield itself corresponds with the 
deeper layer of the original series of spines (fig. 12). The 
frontal shield extends to the sides of the operculum as a pair 
of calcareous shoulders, which commonly bear avicularia, 
while the suboral part of the shield may support a median 
avicularium which guards the entrance to the incipient com- 
pensation-sac. “ Mucronella”’ pavonella, Alder (p. 296), 
probably belongs to Umbonula, which may also include 
M. bicuspis, Hincks.! 

So far as I can judge from dry preparations, the develop- 
ment of the frontal shield takes place in the Umbonuloid 
manner in Porella compressa, P.saccata? ( = P. 
elegantula, auctt.), P. (Palmicellaria) skenei, Mucro- 
nella coceinea, and Escharoides sarsii. In specimens 
of U. verrucosa from Naples* the suboral avicularium has a 
mandible which closely resembles that described by Waters* 
in species of Porella, being of more or less semicircular 
form, with denticulated margin and conspicuous A-shaped 
main sclerite. This confirms the association of Porella with 
Umbonula. I am inclined to think that the presence of 
marginal areole (as shown in figs. 10, 11) is of great import- 
ance as indicating affinities in this direction. 

(6) In the Lepralioid type of calcification (figs. 13—15, 41) 
the young zocecium frequently has a Flustrine appearance ; 
but the frontal shield seems to result from a direct calcifica- 
tion of the frontal membrane. ‘The compensation-sac 1s 
developed, after calcification is complete, as an invagination 
formed at the base of the operculum. ‘This is the case in 
Lepralia, Schizoporella, Urceolipora nana, Huthy- 
ris, Huthyroides, Vittaticella, and Catenicella. The 

1 *Aun, Mag. Nat. Hist.’ (5), xi, p. 201. 

2 Waters, 1900, p. 81. 

3 It is, however, possible that this form is not identical with the littoral 
form common on the south coast of England, which I have alluded to as U. 


verrucosa, 
4 1900, pl. xi. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. B tat) 


parietal muscles at first radiate to the base of the operculum 
(figs. 13, 14), but their insertions become shifted by the 
growth of the compensation-sac, until in the fully formed 
zocecium (figs. 33, 46) they have much the same relations as 
those of Umbonula (fig. 11) or of Cribrilina philomela 
(fig. 9). It seems difficult to avoid the conclusion that the 
floor of the compensation-sac represents in all cases the 
original frontal membrane, as indicated in the diagrammatic 
figs. 43, 44, 50. From this it would appear to follow that in 
the Lepralioid mode of development the phylogenetic history 
of the compensation-sac implied by those figures has been 
modified ontogenetically, in such a way that the apparent 
frontal membrane of the young zocecium is in reality an 
epitheca (cf. fig. 12, ep.), that the calcareous frontal shield 
is next formed beneath the epitheca, and that the frontal 
membrane only makes its appearance when the compensa- 
tion-sac is invaginated. The fully developed frontal wall of 
a Lepralia would, on this view, be represented by a com- 
pletely fused series of Cribrilinid spines, of which the outer 
layer 1s membranous ( = epitheca) and the inner layer is 
calcareous (fig. 50) ; the marginal pores on this view corre- 
sponding with the original communications between the 
Cribrilinid frontal bars and the general body-cavity. 

Another view is, however, possible. The membranous 
covering of the young zocecium may be, as it appears to be, 
the frontal membrane, and the frontal shield may be a cryp- 
tocyst which grows beneath the membrane towards the distal 
end of the zocecium. 

Thus in fig. 34 the calcareous frontal wall from which the 
papilla (cale. p.) is seen to arise would be the cryptocyst, which 
has grown distally beneath the frontal membrane (ep.). The 
invagination of the compensation-sac would then represent 
an ingrowth of a fold of the frontal membrane round the 
free distal border of the cryptocyst. The actual ontogeny 
of the organ would then agree completely with its supposed 
phylogeny in such cases as Lepralia, Schizoporella, and 
Kuthyris. ‘he Microporoid series of genera would thus 


334 SIDNEY F. HARMER, 


acquire great importance in throwing light on the structure 
of some of the Escharine forms. If a given species has a 
Cribrilinid ancestry, it might be expected that the roof of the 
compensation-sac would be formed by the deeper surface of 
the frontal shield (figs. 12, 50). If, on the contrary, the 
frontal shield is a cryptocyst, it might be anticipated that the 
compensation-sac would have a membranous roof distinct 
from the frontal shield. Although this arrangement is indi- 
cated in fig.34 (Huthyris obtecta) I cannot assert positively 
that it is the case in that species. There is, however, no 
doubt of the existence of a complete membranous roof to the 
compensation-sac in Huthyris clathrata (figs, 26, 27). 
There are other considerations which indicate that the 
frontal shield may, in certain Escharina, be a cryptocyst. 
Numerous Membraniporide, both recent and fossil, show a 
tendency to develop a cryptocyst, although in many cases 
complete proof that the calcified “area ” is of this nature is 
not yet forthcoming. The case of Bicellaria grandis 
(fig. 1) is again uncertain. But the fact which is, perhaps, 
most significant is that in certain species of Schizoporella 
(fig. 6), Microporella, etc., the region encircled by the 
marginal spines of the primary zocecium is partially calcified 
proximally. It appears to me to be of great importance to 
ascertain whether this calcified portion is of the nature of a 
cryptocyst or not;—in other words, whether the frontal 
membrane extends as far as the sharp ridge on the inner 
side of the base of the spines. I can hardly doubt that this 
is the case; and if so, the Microporoid origin of Hscharine 
forms in which the compensation-sac develops as an invagi- 
nation at the base of the operculum would appear to be 
indicated. The Cribrilina-like character! of certain species 
of Catenicella, in which the sac develops in this way, is, 
however, a difficulty which requires further explanation. 


1 This character is indicated in fig. 55 of the present paper, and in many 
figures on pls. i and ii of MacGillivray’s “ Monograph of the Tertiary Polyzoa 
of Victoria” (‘'‘T'rans. Roy. Soe. Viet.,’ iv, 1895). The Escharine nature of 
the Catenicellide is noticed by Smitt (1868, 2, p. 46). 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 3385 


It may perhaps indicate that the cryptocyst may have 
co-existed with a Cribrilina-like frontal shield, and it is not 
indeed impossible that the plate pl. in fig. 54 may represent 
the cryptocyst. 

Calvet (1900, pp. 163, 436), who gives an excellent account 
of the structure of many marine Hetoprocta, has already 
expressed the opinion that in Lepralioid forms the calcified 
frontal wall is a cryptocyst. He states that while in most 
species the cryptocyst and the more external “ ectocyst”’ 
(=epitheca) occur only on the frontal wall, in Retepora 
the whole external surface is two-layered. This and the 
analogous case of Urceolipora nana require further study. 

Calvet does not recognise the compensation-sac, although 
it is indicated clearly in his pl. vi, fig. 9 (Catenaria 
lafontii), and its development is partly described on p. 399 
and figured in pl. xii, fig. 20 (Lepralia foliacea). The 
account which he gives (p. 168) of the median pore of 
Microporella does not agree with my results. I think 
that the discrepancies are easily to be explained bythe fact 
that in highly calcified species the compensation-sac is so 
delicate that it may easily be overlooked in sections which have 
been distorted by decalcification and embedding in paraffin. 

Calvet, after quoting in extenso (p. 276) Jullien’s very 
accurate account of the mode of articulation of the operculum 
and its relations to the compensation-sac, proceeds to give it 
a complete denial (p.278). He explains the existence of the 
parietal muscles in species with a calcified frontal wall by 
supposing that wall to have an amount of flexibility sufficient 
to allow itself to be depressed by the parietal muscles. The 
fact that these muscles are not inserted into the calcified 
wall is a sufficient answer to this suggestion. 

Kuthyris may be regarded as a highly modified Hscharine 
form, which has acquired a Flustra-like habit. The 
irregularity of the bars of HK, clathrata, and the fact 
that they are entirely beneath the epitheca, suggest 
that they do not represent the tubular frontal bars of 
Cribrilina, but are due to deficiencies in the calcification of 


336 SIDNEY F. HARMER. 


what was once a completely calcified wall. The epithecal 
investment is particularly obvious in Euthyris. It is 
possible that the basal epitheca is a reminiscence of a time 
when the frond was composed of two layers of zoacia, placed 
back to back. I have previously (1901, p. 17) pointed out 
that there is evidence that many Cretaceous Cheilostomes 
were provided with a similar basal epitheca. 

(7) Inthe Microporelloid series the compensation-sac opens 
by a “median pore,” which has probably resulted from the 
closure of a Schizoporelliform sinus. ‘This view, which has 
been adopted by Gregory! as the basis of his division 
Schizothyriata, receives support from the occasional occur- 
rence in the group of an ancestrula with a Schizoporella- 
like operculum (cf. p. 324). Gregory’s classification places 
Schizoporella with Microporella, and separates both 
from Lepralia. It appears to me, on the contrary, that 
Lepralia and Schizoporella cannot be separated from 
one another, whereas Microporella is a distinct step in 
advance of the other two. 

This section includes Microporella, and perhaps “Cri- 
brilina” radiataamong encrusting forms; while Calwellia 
(in which I place Urceolipora dentata), Ichthyaria, and 
Onchoporella also belong to it. I think Miss Jelly * goes 
too far in placing Siphonicytara in Calwellia. Oncho- 
porella may include O. bombycina, auctt., O. ligulata, 
Busk,® and O. selenoides, Ortmann,! but there seems to be 
little except its habit to distinguish it from Calwellia. The 
suture which in the latter genus connects the middle line of 
the operculum with the median pore probably indicates the 


1 1893, p. 224. 

2 1889, p. 33. 

3 “Quart. Journ. Mier. Sci.,’ viii, 1860, p. 282. This species has a close 
resemblance to O. bombycina, auctt. It was clearly by an oversight that 
Busk (1884, p. 104,n.) referred Scruparia diaphana instead of Carbasea 
ligulata to his genus Onchoporella; the original figures of the two species 
having appeared on the same plate of this Journal. 

* ‘Arch. f. Naturg.,” lvi (i), 1890, p. 28. 


THE MORPHOLOGY OF THE CHEILOSTOMATA, 33 


incomplete fusion of two suboral calcareous spines by which 
the median pore has been cut off from the orifice. ‘The 
characteristic infundibuliform pores on either side of the 
suture are perhaps parts of these spines. 

It may be pointed out in conclusion that the principal 
results of this paper appear to be completely in accordance 
with paleontological evidence. In Canu’s important memoir 
on Cretaceous Cheilostomata (1900), the rarity of the 
Cretaceous Monodermata, a group instituted by Jullien, 
which corresponds in the main with the Escharina of other 
authors, 1s pointed out (p. 458), and only three pages are 
devoted to their consideration. Nearly all the remaining 
species belong to the groups designated as Flustride, 
Onychocellide, Opesiulidee, and Costulide. The Flustride 
correspond with the Flustrina of other authors. The 
Onychocellide are an important series, still represented by 
the recent genus Onychocella, characterised by the great 
development of the cryptocyst, and usually by the presence 
of a peculiar form of vicarious avicularium known as an 
onychocellium. The Opesiulide include the recent Micro- 
poride and Steganoporellida, he Costulidee are equivalent 
to the Cribrilinide. These four groups are all represented 
by a large number of genera and species. 

Gregory (1896, preface) states that “it is among the 
Jurassic deposits that we have to seek the ancestors of 
existing types of Bryozoa.’ His Catalogue of Jurassic 
forms includes only two Cheilostomes, one of which is a 
Membranipora and the other an Onychocella. 

It thus follows that the Jurassic and Cretaceous Cheilo- 
stomes belonged, for the most part, to the Diplodermata 
of Jullien (= Athyriata, Gregory, 1893, p. 223). Both 
Jullien, in a memoir (1888, 4, p. 8) published subsequently 
to the original introduction of his term Diplodermata, and 
Gregory include the Cribrilinide inthe sub-orders respectively 
adopted by them. It appears to me that the Cribrilinidze 
would be better left out of the sub-order. ‘They do not agree 
with Jullien’s definition of the Diplodermata as forms with a 

vot. 46, PART 2.—NEW SERIES. x 


338 SIDNEY F. HARMER. 


double “ectocyst;” that is, as forms in which there is a 
calcareous cryptocyst internal to the frontal membrane. 
Although Gregory was justified in regarding the Cribri- 
linide as forms in which the frontal membrane persists, 
the view which I take of the compensation-sac involves 
the conclusion that the membrane persists even in the 
Escharine genera. The Cribrilinide may be regarded as a 
transition-group between the Flustrine and the Kscharine 
forms; but, as I have pointed out above, it appears to me 
that the family, as at present constituted, is not a natural 
one. 

Although it is very difficult to draw a line between the 
more simple species of Membranipora and those in which 
a definite cryptocyst occurs, it is hardly possible to speak of 
some of the former as diplodermatous. Although the term 
Athyriata is antedated by Diplodermata, it is more in 
accordance with the structure of recent genera than Jullien’s 
term, and in the present state of our knowledge it would 
perhaps be well to accept it as the name of a group in- 
cluding those forms in which the frontal membrane persists 
in more or less its primitive form, whether a cryptocyst is 
present or not below it. 

The Athyriata are thus, on paleontological evidence, the 
oldest group of Cheilostomes. In the Cretaceous period they 
had already differentiated themselves into the types of 
structure represented by Membranipora, Onychocella, 
and Micropora, in the latter two of which the cryptocyst is 
well developed. The Cretaceous period is also characterised 
by the occurrence of a number of species belonging to the 
transitional Cribrilinid type. So far as the evidence goes 
at present, the true Hscharine forms were then much inferior 
in number to the dominant Athyriate group. 

With regard to the other parts of Gregory’s classification 
(1893, p. 225), it appears to me that the sub-order Stolonata 
includes several families of very different affinities, the 
Kucrateide at any rate probably belonging to the Athyriata. 
Some of the Cellularina are also Athyriate, while I agree 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 339 


with Gregory (p. 222) that Catenicella, which he places 
nevertheless in the Cellularina, has but little to do with the 
other Cellularines. I cannot accept the divisions Schizothy- 
riata (forms with a sinus or a median pore) and Holothyriata 
(other Hscharine genera), because it seems to me that there 
is no essential difference between the Schizothyriate 
Schizoporella and the Holothyriate Lepralia. Both 
these genera probably require a good deal of rearrangement 
and subdivision; but among recent species there are 
certainly cases in which it is an open question to which 
genus they should be referred. It would perhaps have been 
better to institute a new sub-order for the Microporelloid 
genera. The question is complicated by the fact that all 
“median” pores are not homologous structures. It is a 
well-known fact, for instance, that in some cases (e. g. 
Adeonella) the median pore is simply a_peristomial 
opening. 


VII. Susmmary or THE MORE ImporTANT RESULTS. 


(1) A large number of Cheilostomata are provided with a 
“ compensation-sac,” a structure described by Jullien, which 
opens at the proximal border of the operculum or by a 
“median pore.” Muscles which usually originate from the 
lateral walls of the zocecium are inserted into the floor of the 
compensation-sac, which they dilate by their contraction, 
thereby helping to force out the polypide. The constant 
change of water in the compensation-sac probably has a 
respiratory importance. The sac is frequently found in 
zocecia whose polypides have undergone histolysis. 

(2) The floor of the compensation-sac corresponds in whole 
or in part with the membranous frontal surface (for which 
the name “frontal membrane” is suggested) of a Mem- 
branipora; and its muscles are homologous with the 
parietal muscles of the same genus. ‘I'he compensation-sac 
appears to have been evolved in more than one way. 

(3) In Cribrilinide it is well known that a calcareous wall 
(for which and for other calcareous frontal walls the name 


340 SIDNEY F. HARMER. 


“frontal shield”? is suggested) is developed as a series of 
marginal spines, which overarch the frontal membrane. This 
leads to the condition found in some KEscharine forms, in 
which, as in Umbonula, the calcareous frontal shield 
similarly overarches the frontal membrane. 

(4) The frontal shield of the Microporidz and Stegano- 
porellidze is of the nature of a eryptocyst,'—that is to say, of a 
calcareous lamella, which grows horizontally across the body- 
cavity beneath the frontal membrane. This has, perhaps, 
led to the condition found in Lepralia, Schizoporella, 
and other genera, in which the compensation-sac is developed 
as an invagination at the base of the operculum and passes 
to the deeper side of the frontal shield, which is on this view 
a cryptocyst. 

(5) The epitheca, or layer of living tissue, bounded by a 
cuticle, which covers the frontal shield, may represent the 
entire frontal membrane (Microporide, etc.), or only a part 
of that membrane (Lepralia, etc.), or it may have been 
derived from the outer calcareous layer of the frontal bars of 
a Cribrilina-like form (Umbonula). The epitheca is 
frequently responsible for the addition of secondary calcareous 
matter to the frontal shield, and in some cases (Cellepora, 
etc.) for the formation of new generations of zocecia which 
are superposed on the older ones. 

(6) The operculum, in its more primitive condition, is merely 
part of the frontal membrane strengthened by a semicircular 
marginal flange. It acquires a firmer texture and a more 
elaborate arrangement of its thickened parts as the result of 
its articulation with calcareous portions of the zocecium. 
While occlusor muscles may be regarded as an essential 
adjunct of the operculum, definite divaricators seem to have 
been evolved as a modification of the distal pair of parietal 
muscles. 

(7) ‘he consideration of the relations of the frontal surface 
involves considerable rearrangements in the Cheilostomata 
(see section on classification, p. 329). 

' Cf. Harmer (1900), p. 228. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 341 


(8) The results of the study of the compensation-sac and 


frontal surface generally are confirmed by an examination of 


the primary zocecium or “ancestrula”’ (Jullien). 


(9) Two new species (Huthyris clathrata and Lepralia 
haddomi) are described, and a new genus (Euthyroides) 
is suggested for Huthyris episcopalis, Busk. 


(10) The operculum of KH. clathrata is of a specially 


complicated type. The vestibule, or first part of the introvert, 


is provided with a chitinous distal lip, for which the name 


“Tabium ’ 


1877. 
1852. 
1884. 
1900. 
1900. 
1880. 
1893. 
1896. 
1892. 
1900. 


1901. 


1880. 
1889. 
1886. 


ees a San 
is suggested. 


VIII. Lirerarovre.! 


Barros, J.—‘‘ Rech. Embryol. Bryozoaires,” ‘Trav. Inst. Lille,’ i. 
Busx, G.—‘ Cat. Marine Polyzoa . . . Brit. Mus.,’ part 1. 
Busx, G—‘ Report on the Polyzoa,” I, ‘Challenger Reports,’ part 30. 


Cauvet, L.—* Cont. Hist. Nat. Bryozoaires Ectoproctes Marins,” 
‘Trav. Inst. Zool. Montpellier’ (N.5.), Mém. 8. 
Canu, F.—*‘ Rev. Bryozoaires du Crétacé figurés par D’Orbigny : 
II, Cheilostomata,” ‘ Bull. Soc. Géol. France’ (3), xxviii, p. 334. 
Goupstein, J. R. Y.—‘ Notes on Living Polyzoa,”’ ‘Journ. Micr. Soc. 
Victoria,’ i, p. 42. 

Grecory, J. W.— British Paleogene Bryozoa,” ‘ Trans. Zool. Soc.,’ 
xill, 1895, p. 219. 

Grecory, J. W.—“ Cat. Foss. Bryozoa in the British Museum,” 
‘Jurassic Bryozoa.’ 

Harmer, 8S. F.—*“ Excretory Processes Marine Poiyzoa,’’ ‘ Quart. 
Journ, Mier. Sci.,’ xxxili, p. 123. 

Harmer, 8. F.—“ Rev. Genus Steganoporella,” ‘Quart. Journ. 
Micr. Sci.,’ xliii, p. 225. 

Harmer, 8. f.—‘Struct. Class. Cheilostomatous Polyzoa,” ‘ Proce. 
Cambridge Phil, Soc.,’ xi, p. 11. 

Hincxs, T.—‘ Hist. Brit. Marine Polyzoa.’ 

JeviLy, EH. C.—‘ Syn. Cat. Recent Marine Bryozoa.’ 

JuLuiEN, J.—‘ Les Costulidées, nouvelle Famille de Bryozoaires,” 
‘ Bull. Soc. Zool. France,’ xi, p. 60. 


1 Many of the references are given as foot-notes. 


342 SIDNEY F. HARMER. 


1888 (1). Juttren, J.—“ Sur la sortie et la rentrée du polypide,” ‘ Bull. Soe. 
Zool. France,’ xii, p. 67. 

1888 (2). Jutmien, J—* Du testicule chez la Lepralia figularis,” ‘Mém. 
Soc. Zool. France,’ i, 1888, p. 270. 

1888 (3). JuLtren, J.—“ Obs. Anat. Caténicelles,” ‘Mém. Soc. Zool. France,’ 
i, 1888, p. 274. 

1888 (4). JuLLreN, J.—‘ Mission Scient. du Cap Horn,’ t. vi, Bryozoaires. 

1891. Levinsen, G. M. R.—* Polyzoa,” ‘ Vid. Udbytte ‘* Hauchs ” Togter’ 
(Copenhagen). 

1887. MacGiuuivray, P. H.—‘* Cat. Marine Polyzoa Victoria,” ‘ Trans. Proce. 
Roy. Soe. Victoria,’ xxii, p. 187. 

1836. Mitne-Epwarps, H—‘‘ Rech. anat., physiol. ’et zool. Eschares,” 
* Ann, Sci. Nat.’ (2), vi, p. 5. 

1898. Neviant, A.—“ Appunti sui Briozoit del Mediterraneo,” ‘ Boll. Soe. 
Romana Stud. Zool.,’ vil, p. 165. 

1871. Nivscun, H.—‘ Anat. Flustra membranacea,” ‘ Zeitschr. f. wiss. 
Zoo).,’ xxi, p. 416. 

1889. Percens, h.— Untersuchungen an Seebryozoen,” ‘ Zool. Anz.,’ xii, 
pp. 504, 526. 

1865. Smirr, F. AA—“Om Hafs-Bryozoernas utveckling och fettkroppar,’ 
‘Ofv. k. Vet.-Ak. Forb.,’ 1865, p. 5. 

1868 (1). Suir, F. A.—* Krit. forteckn. Skandinaviens Hafs-Bryozoer,” iii, 
* Ofv. k. Vet.-Ak. Forh.,’ xxiv, 1867, No. 5, p. 279. 

1868 (2). Smrrr.—lbid., iv, ‘ Olv. k. Vet.-Ak. Forh.,’ xxiv, 1867, Bihang, p. 3 

1896. Smirr, ., A.,—La Filiation des Espéces d’Animaux,” ‘Compte-rendu 
3e Congres Int. Zool.’ (Leyden, 1895), p. 235. 

1881. Warenrs, A. W.—* Fossil Chilostomatous Bryozoa 5.W. Victoria,” 
‘Quart. Journ. Geol. Soc.,’ xxxvii, p. 309. 

1883. Warrers, A. W.—* Foss. Chilostomatous Bryozoa Muddy Creek, Vic- 
toria,” ‘Quart. Journ. Geol. Soc.,’ xxxix, p. 423. 

1885. Warers, A. W.—‘ Use of the Avicularian Mandible in the Determina- 
tion of the Chilostomatous Bryozoa,” ‘Journ. Roy. Mier, Soe.’ (2), 
v, p. 774. 

1888. Waters, A. W.—* Suppl. Report on the Polyzoa,” ‘Challenger 
Reports,’ part 79. 

1892. Warens, A. W.—* Gland-like Bodies in the Bryozoa,” ‘Journ. Linn. 
Soc.,’ xxiv, p. 272. 

1896. Warners, A. W.—* Interzocecial Communication in Flustridae,” ‘Journ. 
Roy. Mier. Soe.,’ 1896, p. 279. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 343 

1898. Waters, A. W.—“ Observations on Membraniporide,” ‘Journ. Linn. 
Soc.,’ xxvi, p. 654. 

1900. Waters, A. W.—“ Bryozoa from Franz-Josef Land,” ‘Journ. Linn, 
Soc.,’ xxviii, p. 43. 


HXPLANATION OF PLATES 15—18, 


Illustrating Mr. Sidney F. Harmer’s paper “On the Mor- 
phology of the Cheilostomata.” 


REFERENCE LETTERS. 


ar. Margival areole. avic. Avieularium. 4. Polypide-bud. 4.4. Brown 
body. d.ep. Basal epitheea. dué/r. Buttress of operculum. c@. Caecum of 
stomach. ale. p. Calcareous papilla. coud. Condyie, or chitin surrounding 
the decalcified condyle. ¢. p. Communication pore. ¢. s. Compensation-sac. 
crypt. Cryptocyst. depr. Depressor muscle of frontal membrane. div. Divari- 
eator muscle of operculum. d. w. Distal wall of zocecium. ep. Epitheea. 
ep. ¢e. Cavity beneath epitheca. % Membranous fenestra. /. 4. Frontai bars or 
costules. jl. Vertical flange of operculum. 2 m. Frontal membrane. //. sé. 
Frontal shield. . sh. d. Distal lobes of frontal shield. izf. avie. Infra-avicu- 
larian compartment. 7. Chitinous joint. 76. Labium. ¢. 7. Lateral recess. 
m.¢. Marginal cavity of frond. m.p. Median pore. m.¢. Median tooth 
(lyrula) of Smittia. oee/. Ocelusor museles of operculum. op. Opereulum. 
op. gl. Opercular glands. ov. Ovicell. p. Pores. p. ¢. Pore-chambers. 
p. d. Parieto-ciaphragmatic muscles. p.m. Parietal muscles. p.m.’ Distal 
group of parietal muscles. p.v. m., p. v. m.! Parieto-vaginal muscles and 
bands. p. w. Proximal wall of zoecium. p. z. Primary zocecium. 7. m. 
Retractor muscles of polypide. s., s.’ Suture in caleareous wall. sc/. Chitinous 
sclerite. sp. Spine. sap. avie. Supra-avicularian compartment. ¢. Tentacles. 
é. s. Tentacle sheath. v. Vitta. vest. Vestibule (= diaphragm). z.’ Daughter- 
zocecium. 

Fig. 12 (PI. 15) and Figs. 43, 44, 50, and 51 (PI. 17) are diagrammatie 
representations of the points respectively illustrated by them, Fig. 18 (Pl. 16) 
is X 1. Figs. 19, 20, 22, 32, 35, and 36 (PI. 16) were drawn to the same 
scale (camera lucida, Zeiss, A obj.). Fig. 19 was not reduced, but the others 
were reduced 3. All the remaining figures are more highly magnified (Zeiss, 
C obj.; afterwards x 3). 


The specimens figured belong to the Museum of Zoology at Cambridge. 


344, SIDNEY F. HARMER. 


PLATE 15. 

Fie. 1.—Bicellaria grandis, Busk, var. producta, MacGill. Showing 
the frontal membrane (/. m.), into which a single pair of parietal muscles 
(p. m.) are inserted; the plate erypé. is probably a eryptoeyst. The zocecium 
has given rise by budding to two younger zocecia (z.’), and is connected with 
its lateral neighbour by the communication pore (c. p.). Hach of the three 
communication pores is surrounded by a strong calcareous ring; occl., 
occlusor muscles.— Victoria. 

Vie. 2.—Microporella malusii, Aud. Primary zocecium or ancestrula. 
The frontal membrane is typically Flustrine, its calcareous margin bearing ten 
spines, of which three are oral spines.—Naples. 

Fig. 3.—Flustra pisciformis, Busk. Distal view of operculum, showing 
the vertical flange #7.—Bass’s Strait (Challenger Collection). 

Fic. 4.—F. pisciformis. An entire zocecium from the same slide. The 
distal groups of parietal muscles (p. m.') probably act as divaricators of the 
operculum. 

Fig. 5.—Dimetopia spicata, Busk. The funnel-shaped zocecium is 
closed by a terminal frontal membrane which is depressed by a single pair of 
parietal muscles (p. m.).—Victoria. 

Fie. 6.—Schizoporella vulgaris, Moll. Primary zocwcium. The region 
inside the marginal spines is partly occupied by a calcareous plate (erypt.), 
which is probably a eryptocyst.—Naples. 

Fic. 7—Cribrilina radiata, Moll. Primary zocecium (p. z.) and three 
younger zoccia. The Flustrine frontal membrane of p. z. is overarched by 
a proximal scutum-like spine (se.), while the ten other spines bear minute 
horizontal calcareous lobes which project from their inner sides. The oral 
spines of the zocecium z.’ bear similar lobes. In all except the primary 
zocecium the series of oral spines is continued by the pair of elongated mem- 
branous spines (sp.) on either side of the median pore (m. p.), and further by 
the small papille (sp.’), which correspond with the marginal spines of the 
primary zoccium. The series of pores are in the same radii as these modified 
spines. ‘The pore-chambers are not indicated.—Naples. 

Fig. 8—Cribrilina philomela, Busk. The frontal shield is composed 
of hollow ealeareous bars (/. 4.), alternately arranged on the two sides of the 
zocecium. The rows of pores correspond with the intervals between the bars. 
The compensation-sae (c. s.) and some of the parietal muscles (p. m.) are indi- 
eated.—Marion Is. (Challenger Collection). 


Fie. 9.—Basal view of another zocecium from the same colony. The dista 
parietal muscles (p. m.’) are stronger than the other groups (p. m.). 


lic. 10.—Umbonula pavonella, Alder. A series of large areola! occurs 


1 Cf. Gregory (1893), p. 221. 


THE MORPHOLOGY OF THE CHELLOSTOMA'TA. 345 


round the proximal and lateral borders of the frontal shield, which extends as 
two wing-like processes (f. sh. d.) along the borders of the orifice. Hach of 
these processes bears a lateral oral avicularium (avic.). The calcareous floor of 
each areola is perforated by a pore through which a strand of living tissue («) 
passes to join a sheet of similar tissue which overlies the frontal shield, The 
frontal membrane lies at a much lower level than the frontal shield (cf. Fig. 12), 
and the orifice is not represented.—North Sea. 

Fic. 11—Umbonua verrucosa, Esper. Young zoecium. The frontal 
shield is represented by a crescentic calcareous film, already divided by radiat- 
ing buttresses on its upper surface into distinct areolee (ar.), in the floor of each 
of which is a pore (p.). The distal wings (/ sf. d.) of the shield are distinctly 
indicated. The parietal muscles ( p.m.) are seen through the calcareous film.— 
Plymouth. 


Fie. 12.—Diagrammatic longitudinal bisection of an Umbonuloid Cheilo- 
stome. 


Fies. 13—17.—Kuthyroides episcopalis, Busk.—Victoria. 
Fig. 13.—Young zocecium. The limit of calcification is the line 2, 
distal to which is an accumulation of nuclei (c. s.), to which the parietal 


muscles (p. m.) radiate. ‘The distal wall (d. w.) of the zocecium is not 
completely calcified ; e., growing edge of zoarium. 


Fig. 14.—An older but much shorter zocecium, in which calcification 
is nearly complete. ‘The median pore is becoming delimited by the 
simultaneous growth of the median tongue (¢g.) of calcareous matter and 
the lateral processes (/.p.). The parietal muscles (p. m.) radiate towards 
a mass of tissue at the base of the operculum ; c¢. p., communication pores. 

Fig. 15.—Older zocecium, with completely calcified walls. The com- 
pensation-sac (¢. s.) now possesses a distinct cavity, but is still of small 
extent. 

Fig. 16.—Mature ovicell, borne by a fertile zocecium which has an 
ordinary zocecium on its proximal side. 

Fig. 17.—A fertile zocecium with a young ovicell (ov.) represented by 
a concave plate, which will constitute the inner wall (¢. w.) of the ovicell, 
and by a second plate, which will form the outer wall (0. w.). The 
Cribrilina-like frontal bars (f 6.) differ from those shown in Fig. 16 
in correlation with the presence of an ovicell (0v.’) on the proximal side of 
this zocecium, On the distal side of the young ovicell (ov.) is a still 
younger fertile zocecium whose frontal bars are only half developed. 


PLATE 16. 


Kies. 1$—31.—Huthyris clathrata, n. sp.—Port Jackson. 
Fig. 18.—Ends of several branches, x 1. 


346 SIDNEY F. HARMER. 


Fig. 19.—Operculum and labium (2) from a dry specimen. The arrow 
indicates the entrance to the compensation-sac. 


Fig. 20.—Part of the frontal surface of a branch ; m., lateral margin of 
the frond, the epitheea of which covers a continuous marginal cavity 
(m. c.), strengthened by calcareous bars (c. 6.) in its frontal wall. Two of 
the opercula (B, B’) are different from the others (A). 

Fig, 21.—The zocwcium B and parts of its neighbours, of the preceding 
figure ; d. w., distal wall of B; p. w., proximal wall of its distal neighbour. 
(For 2, see Fig. 30.) 

Vig. 22.—Basal view of several zocecia; m, mesentery-like lamella 
of chitin connecting the proximal part of the basal wall of the zocecium 
with the basal epitheca. 

Vig. 23.—Labium, seen from the distal side. 

Vig. 24.—Open operculum (op.) and labium (/b.), in side view; coxd., 
chitinous lamella covering the condyle (decalcified) ; sc/., sclerite surround- 
ing the cavity in which the operculum lies; ¢.s., begmuing of floor of 
compensation-sac ; ves/., entrance to vestibule. 

Fig. 25.—Closed operculum and labium ; occé., occlusor muscles ; div., 
divaricator muscles. (lor a, see Fig. 50.) 

Fig. 26.—Longitudinal section (thick) of a zocwcium; & ep., basal 
epitheca; ep.c., cavity beneath epitheca; / 6.,a bar belonging to the 
frontal shield. 

Vig, 27.—Distal part of the operculum, with the labium (/d.), seenina 
thick transverse section of the zocecium (not decalcified). ‘The condyles 
(cond.) are in the foreground, and the part of the operculum immediately 
connected with them (cf. Fig. 29) is not indicated. ‘The labium and the 
biting edge of the operculum are seen at a much deeper focus; p., pores. 

Vics, 28—81.-—Transverse sections of opercula. 

Fig. 28.—On the distal side of the condyle, showing the great extent 
of the vertical lateral flanges,! and the free terminations of the buttresses 
of the operculum. 

Fig. 29.—Through the region of the condyles; the operculum is par- 
tially open. 

Vig. 80.—On the proximal side of the condyies ; showing the way in 
which the cireular marks (@) which appear in Fig. 21 are formed, 


Vig. 31.—Immediately distal to the basal sclerite. 


! Owing to the hardness of the chitin of the opercula, 1 have uot succeeded 
in obtaining completely satisfactory thin sections showing the relations of the 
labium to adjoining parts. 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 347 


Fies. 832—387.—Huthyris obtecta, Hincks.—Torres Straits. 

Fig. 32.—Part of the frontal surface of a branch, showing two of the 
large zocecia (B); m.c., marginal cavity of frond; cale. p., calcareous 
papillee supporting the frontal epitheca ; p., pores. 

Fig. 33.—Basal view of a B-zocecium from which most of the basal 
calcareous wall (d. w.) has been removed. The tentacles (¢.), contained in 
their tentacle sheath, lie in a groove of the compensation-sae (ce. s.) which 
bulges out on each side of the tentacles. ‘The operculum (9p.) is seen 
partly through the distal wall (@. w.). 

Fig. 34.—Thick longitudinal section, showing the frontal epitheca (ep.) 
and the basal epitheca (4. ep.) held at a distance from the calcareous walls 
of the zocecia by the ecaleareous papille (cae. p.). 

Fig. 35.—Basal view of a B-zocecium (8) and several others ; from the 
edge of a frond. 

Fig. 36.—B-zocecium and A-zocecium, calcined. The zocecia are in 
contact with their neighbours by small parts only of their walls, which 
are perforated by communication pores (c. p.). The remaining pores (p.) 
are in relation with the cavity beneath the epitheca. 

Fig. 37.—Basal view of part of the marginal thickening of an old 
branch; z., zocecia; m., free margin of branch; /.,2., longitudinal cal- 
careous ridges. 


PLATE 17. 

Fic. 38.—Lepralia haddoni, n. sp.; showing an avicularium (avic.) and 
the two kinds of opercula.—Torres Straits. 

Fig. 89.—L. haddoni. Basal view of azocecium with trifoliate operculum, 
from the same slide; m. ov., muscles of ovisac. 

Fre. 40.—Smittia reticulata, J. MaceGill.; showing the marginal areol 
(ar.).— Naples. 

Vig. 41.—Lepralia pallasiana, Moll. Young zoccium. ‘The compeusa- 
tion-sae (c.s.) is still small. The pores extend round the distal margin of the 
operculum.— Naples, 

Fie. 42.—Smittia trispinosa, Johnst., var. arborea, Lev.; showing 
the compensation-sac (ce. s.).—Greenland. 

Fig. 43.—Diagrammatic transverse section of a Flustrine form, showing the 
parietal muscles (p. #.) inserted into the frontal membrane (/. m.). 

Fic. 44.—Similar section of a Cribrilina. The frontal membrane is 
covered by the frontal bars (/ 4.) or costules. 

Fie. 45.—Lepralia dorsiporosa, Busk. The distal lobes of the frontal 
shield (f. sh. d.) have nearly united on the distal side of the operculuww.— 
Torres Straits. 


348 SIDNEY F. HARMER. 


Fic, 46.—Schizoporella sanguinea, Norm.—Basal view of a zocecium 
from which the basal wall has been removed.—Naples. 

Fig. 47.—Schizoporella australis, Haswell. Operculum with the den- 
ticulated condyles.—Torres Straits. 

Fic. 48.—Schizoporella linearis, Hassall. Showing avic., a gigautic 
avicularium (= ‘‘ ocecium,”’ Hincks) ; prow., proximal end of the zocecium, into 
which the polypide extends, beneath the avicularium.—Naples. 


Fic. 49.—Catenaria lafontii, Aud. Side view, showing the compen- 
sation-sac ; z.’, daughter-zocecia.—Naples. 

lie. 50.—Diagrammatic transverse section of an Umbonuloid Cheilostome 
The cavity of the marginal areole (a7.) communicates with the general body- 
cavity by the pores p.; 7. sd., frontal shield, with calcareous tubercles on its 
outer surface; ep., epitheca; /. m., frontal membrane, or floor of compensation- 
sac (c} 3:). 


Fig. 51.—Diagrammatic obliquely transverse section through a Microporoid 
Cheilostome, passing on the right side through the horizontal part of the 
cryptocyst (crypé.), and on the left side through a lateral recess (/. 7".) con- 
taining a depresser muscle (depr.) or modified parietal muscle; /. m., frontal 


membrane (cf, Pl. 18, fig. 58). 


Fie. 52.—Schizoporella linearis, Hassall. Zocecium showing the com- 
pensation-sac (¢. s.). The emargination of the frontal shield coutaining the sinus 
of the operculum enlarges as it passes through the thickness of the calcareous 
wall, so that its outline on the inner side of the frontal shield is represented 
by the line «—Naples. 


PLATE 18. 


Vie. 58.—Catenicella alata, Wyv. Thoms. The zocwcium, which forms 
the median part of the “globulus,” is overlapped by the lateral wings, con- 
stituted by sup. avic., the supra-avicularian compartment; avic., the vestigial 
avicularium; and ¢#/f. avic., the infra-avicularian compartment; v, edge of 
calcareous plate corresponding with p/. in Fig. 54.—Victoria. 

lia, 54.—C. plagiostoma, Busk, var. setigera, MacGill. The avicu- 
laria are here of enormous size, and are different on the two sides of the 
globulus. The roof of the compensation-sac (e¢.s.) is partly a caleareous 
plate (p/.) connected with the calcareous framework of the operculum; the 
free edge of this plate is seen at @. ‘The outer calcareous layer consists of a 
system of narrow bars separated by large membranous fenestrae (/.).—Vic- 
toria. 

Vie. 55.—C. hastata, Busk. The avicularium (eviec.), although small, 
possesses a rudimentary polypide (pod.). The infra-avicularian compartment 
(uf. avic.) 1s completely divided into two parts, of which the lower appears to 


THE MORPHOLOGY OF THE CHEILOSTOMATA. 349 


correspond with the vitta of the next species. The frontal shield (f/ s#.) is 
Cribrilina-like.—Victoria (Challenger Collection). 


Fic. 56.—Vittaticella cornuta, Busk.—Distal end of an old zocecium 
with a young “biglobulus ;” 7, fully formed chitinous joint ; 7.’, young joints 
h., the lateral horns; 4.’, vestigial horn of the proximal zocecium of the biglo- 
bulus; v., vittee, filled with deeply stained material in the young zocecia. The 
compensation-sae (c.s.) is fully developed in the proximal zocecium of the 
biglobulus, and is quite young in the distal zocecium.—Victoria. 

Fic. 57—Micropora, sp. Basal view, showing the lateral recess (/. 7.) 
of the ecryptocyst, containing the depressor muscle (depr.), and giving off 
distally a caleareous flange, from which the occlusor muscle (oce/.) originates. 
—Torres Straits. 

Fie. 58—Micropora, sp. (from the same slide). Frontal view, showing 
the eryptocyst (erypt.) with the two lateral recesses (/. 7.), and the depressor 
muscles (depr.). 


Fie. 59.—Ichthyaria oculata, Busk. Zocecium preceding a bifurcation 
of the branch, and therefore less curved than most of the zocecia. ‘The com- 
pensation-sac (¢.s.) opens by the “median” pore (m.p.), which is asym- 
metrical ; s.s.’, sutures in calcareous wall.—S.E. of Buenos Aires (Challenger 
Collection, Stat. 320). 

Fic. 60.—Calwellia sinelairti, Busk.—Seen somewhat obliquely. The 
compensation-sac (¢.s.) opens by the erescentic median pore (m.p.).—S. of 
Kerguelen Is. (Challenger Collection, Stat. 1538). 


Fie. 61.—C. gracilis. Maplestone. Young zocecium and parts of ifs 
neighbours.— Victoria. 
Fie. 62.—C. gracilis (from the same slide). Two mature zocecia, showing 


the large compensation-sacs (¢. s.); ¢c. p.. communication pores; prow., the 
narrow proximal part of a distal zocecium. 


Fie. 63.—Microporella malusii, Aud. Basal view of a zocecium which 
has lost its polypide ; 4., polypide bud; ¢. s., compensation-sac, opening by the 
median pore (m.p.); corz., Jullien’s ‘cornicula;”’ p.c., pore-chambers ; 2. 
neighbouring zocecia.—Naples. 


Fras. 64—66.—Steganoporella alveolata, Harmer.—Torres Straits. 


Fig. 64.—B-opereulum with its oeclusor muscles, seen from the inner 
side, and somewhat distally, so as to be considerably foreshortened ; cozd., 
condyles, united to the basal sclerite (4. s.) of the operculum by the strong 
ligaments, dig.; ocel., proximal oeclusor musele, the tendon (¢end.) of 
which is inserted into the occlusor tubercle (oce/. ¢.) of the operculum, 
and gives off a fascia (7) which connects it with the projecting proximal 
eud of the main selerite (m.s.); ocel.’, distal oeclusor, inserted into the 
fascia f.’; a, line along which the operculum passes into the roof of the 
vestibule (ef. Fig. 26). 


300 


SIDNEY F. HARMER. 


Fig. 65.—Basal view of a B-zocecium (basal wall removed). The erypto- 
eyst. (erypt.) is seen from its basal surface; part of the floor of the 
lateral recesses (2. 7.) is broken away. Between the two recesses is the 
passage (¢ube) by which the tentacle sheath passes to the orifice; 
occl.’, distal ocelusors, with their transverse fascia (f.’) ; occl., proximal 
occlusor, with its tendon (¢enxd.); depr., depressor muscles of frontal 
membrane ; div., divaricator muscles of operculum; m.s., main sclerite of 
operculum. The movements of the tentacle sheath are restrained by 
four delicate muscles, two of which originate from the wall of the tube. 

Fig. 66.—Frontal view of a B-zocecium ; oce/.¢., occlusor tubercle ; 
lig., hinge-ligaments, connecting the condyles (covd.) with the ends of 
the basal sclerite (4. s.); 2 m., frontal membrane, into which the 
depressor muscles (depr.) are inserted; m.pr., median process, arising 
from the roof of the tube (cf. Fig. 65); other letters as in Fig. 65. 


Posrscript.—Since this paper was sent to press I have 
received from Dr. G. M. R. Levinsen a copy of his recently 
published preliminary note entitled “Studies on Bryozoa”’ 
(‘ Vidensk. Medd. fra den Naturh. Foren. i Kjébenhavn,’ 
1902). The paper deals with the morphology to the Chei- 
lostomata, and the questions considered are largely identical 
with those to which | have here paid attention. Dr. Levinsen’s 
contribution is too important to be discussed within the limits 


of a postscript.—s. ielel 
July 27th, 1902. 


ON THE DEVELOPMENT OF SAGITTA. 301 


On the Development of Sagitta; with Notes on 
the Anatomy of the Adult. 


By 


L. Doneaster, 
King’s College, Cambridge. 


With Plates 19—21. 


I. HuistToricat. 


Tur development of the Cheetognatha was first investigated 
by Gegenbaur, but the earliest full account was given by 
Kowalevsky in 1871 (1). He described the gastrulation, the 
formation of the archenteric pouches, and the subsequent 
stages up to the hatching of the young Sagitta. Biitschh 
studied another species taken off the coast of Norway in 1873 
(2), and found considerable difference between this species 
and that which Kowalevsky had described. ‘he most 
important new facts which he added were, firstly, the forma- 
tion of separate anterior cavities of the ccelom, separated off in 
the earlier stages of the embryo from the archenteric pouches ; 
and secondly, the very early separation of the cells which 
remain unaltered during the whole embryonic development, 
and which ultimately give rise to the male and female genera- 
tive organs, 

In the species studied, which unfortunately was not iden- 
tified, the tissues of the embryo had a much more epithelial 


352 L. DONCASTER. 


character than those observed by other workers, and the 
structure of the embryo is therefore almost diagrammatic in 
its simplicity ; the development after hatching was, however 
not followed. 

The most complete and accurate account of the embryology 
of the group which has yet appeared was published by O. 
Hertwig in 1880 (8). Besides making a careful investigation 
of the structure of the adult in several species, he gives a full 
description of the development, from the first cleavage of the 
egg up to the tenth day after hatching. He described for 
the first time the remarkable cleavage, and pointed out that 
the cells which give rise to the sexual organs are at first two, 
and later four, and not two or four groups of cells, as was 
stated by Biitschi. He gives a fuller and more accurate 
account than his predecessors of the formation of the mouth 
and alimentary canal, and suggests, without asserting 
definitely, that the longitudinal septum in the tail region of 
the adult is formed by a continuation of the alimentary canal 
into the tail, and not by the prolongation of the splanchnic 
mesoderm, as Kowalevsky believed. 

Hertwig, however, failed to confirm Biitschli’s observation 
on the origin of the anterior pair of coelomic cavities. After 
the young Sagitta has hatched he describes the formation of 
the body-cavities, which have been obliterated in the embryo, 
the migration of the genital cells from the splanchnic meso- 
derm to the body-wall, and mentions that the transverse septa 
across the body appear also at the same time. He also men- 
tions the formation of a lumen in the intestine, the ciliation 
of its walls, and refers to the origin of the anus. The origin 
of the muscles and nervous system is also shortly described, 
but the account of the development after hatching is 
altogether much less complete than that of the embryonic 
stages. 

Grassi (4) was the next investigator to take up the study of 
the Cheetognatha, and his account published in 1885 gives a 
very full if not always satisfactory description of the anatomy 
of the adult, but no new observations of importance were 


ON THE DEVELOPMEN'T OF SAGITTA. 3038 


added on the development. Since his time the only impor- 
tant work that has been done on the embryology is a note by 
Jourdain. ‘This must be referred to, because he contradicts 
the statements of all previous workers as to the formation of 
the mesoblast, which he says is not formed by archenteric 
diverticula, but by a process of delamination between the 
ectoderm and endoderm. ‘This process is not fully described, 
and there are no figures, so that his exact meaning is hard to 
discover. 

It will be seen, therefore, that a number of important ques- 
tions concerning the Chetognatha have been left unsolved ; 
with regard to some points nothing was known, while with 
others observers have arrived at different results. Leaving 
aside the note by Jourdain, which reopened the whole question 
of the origin of the mesoderm and the ccelom, there was 
nothing certainly known about the mode of development of 
the transverse septa or of the oviducts and sperm-ducts, both 
of which are matters of great importance on account of their 
bearing on the theory that Sagitta is related to the Annelida. 
It was important also that Biitschli’s observation on the 
formation of head-cavities at an early stage should be, if pos- 
sible, confirmed, or at least that his account of the develop- 
ment should be reconciled with that given by Hertwig. 
There are also many minor points ; for example, Hertwig in his 
monograph describes the nervous system as being of two 
parts, one ectodermal and sensory in function, the other meso- 
dermal and motor. ‘This idea was founded on the anatomy of 
the adult, but no attempts have been made hitherto to confirm 
or deny it by study of the embryology. 

The present work, undertaken to solve if possible these 
questions, and to throw light on the systematic position of 
the Chetognatha, was carried on chiefly at Naples, between 
October, 1900, and June, 1901, while I occupied the Cambridge 
table at the Zoological Station. I wish to take this oppor- 
tunity of acknowledging my indebtedness to Professor Dohrn 
and all the officers of the station for their unvarying kindness 
and willingness to give help. 

voL. 46, PART 2,—NEW SERIES. ¥ 


304 L. DONCASTER. 


Il. Marerrat AnD MeErtruops. 


The material for my work was obtained partly at Naples 
and partly in the “ Pantano” at Faro, near Messina. The 
Pantano is a lagoon about half a mile in diameter, con- 
nected with the Straits of Messina by a shallow canal. In 
summer, when the water is exceedingly warm, one species of 
Sagitta is very abundant, and during the first two weeks of 
July, 1900, I obtained a number of eggs in all stages of 
development by fishing with the tow-net at a depth of fifteen 
to twenty feet. After allowing the “ Auftrieb” to settle for 
a short time, the bottom layer was drawn off with a syphon 
and searched with a lens, and in this way quantities of eggs 
could generally be obtained. I have not been able to deter- 
mine this species of Sagitta with absolute certainty; the 
adults were generally found at a considerably greater depth 
than the eggs, although the young occurred at all depths ; but 
as only one species was found, there can be little doubt that 
the eggs belonged to it. The adult Sagitta belonged in all 
probability to the species bipunctata, although it does not 
entirely agree with Grassi’s description; but I have found that 
this species is very variable, and am inclined to believe that 
possibly two species have been included under the name 
* bipunctata.” 

The eggs obtained at Faro developed very rapidly. At 7 
a.m. gastrule were found, and the young Sagitta generally 
hatched between 6 and 8 p.m. the same evening. The eggs 
were therefore preserved at intervals during the day in order 
to get a complete series of the embryonic development. ‘The 
methods of preservation adopted were (a) with sublimate con- 
taining 20 per cent. of acetic acid, and (b) with osmic acid for 
a few minutes followed by Miiller’s solution for several hours. 
Of these methods, the first gave good results ; the second was 
fairly good for the young after hatching, but the embryos in 
the shell treated in this way were rarely satisfactory, being 
very brittle and much contracted. Possibly the osmie acid 


ON THE DEVELOPMEN’ OF SAGITTA. 30D 


remained inside the shell, and therefore its action on the 
tissues was too much prolonged. 

At Naples eggs of several species were obtained, some in 
May and June, by searching the “ Auftrieb ” with a lens as at 
Faro; but since many species are common at Naples, these 
could not be identified with certainty. 

The most satisfactory method was to put a number of 
Sagitta in a jar with two or three litres of fresh sea water, and 
next day by drawing off the lower layer of water with a 
syphon the eggs could be found. In this way a large 
number of eggs of S. enflata were obtained in the autumn 
of 1900 (from November Ist to December 11th), and again in 
the spring of 1901, In the cold weather no eggs were laid ; 
the last were obtained on December 11th, and the first in the 
spring on March 8th. These eggs developed more slowly 
than those obtained in Sicily, possibly in consequence of the 
lower temperature, for it was noticed that the development 
took longer as the weather got cooler. They were also less 
regular than the Sicilian species. At 9 a.m. all stages from 
the unsegmented egg up to well-developed gastrule were 
found, and hatching took place usually but not always on the 
following day. In Sicily it was not found possible to keep 
alive the young Sagitta more than afew hours after hatching, 
but the S. enflata larve at Naples were kept easily for a 
week, and in one case for fifteen days. They were kept in 
jars of about three litres, covered with a glass plate, and 
without any especial care would live for several days. Those 
that lived for the longest time were kept at an even tem- 
perature by placing the jars in running water, and once or 
twice a day fresh sea water was run in, in such a way as to 
carry down bubbles of air to aérate the water. The young 
Sagitta always remained near the surface, so that the water 
could be drawn off with a syphon from below to make room 
for the fresh supply. ‘The water was in no case filtered, and 
the temperature ranged from 16°5° to 18°5°. 

I also obtained eggs from 8. bipunctata at Naples during 
April and May, and kept the young alive for several days. 


306 L. DONCASTER. 


Attempts to obtain eggs from Sagitta minima and Spa- 
della draco were unsuccessful. 

At Naples the eggs were generally preserved with con- 
centrated corrosive sublimate, to which a small portion (about 
5 per cent.) of acetic acid was added. This method also gave 
excellent results with the young after hatching; some of 
these were also preserved with osmic acid, followed by 
Miiller’s solution for a day or two. The latter method was 
more successful with the advanced young when the tissues 
had already become differentiated; for purely embryonic 
tissue ib gave poor results. Some of the young were also 
preserved in a solution of potassium bichromate, to which a 
little acetic acid was added, a method which gave results on 
the whole similar to those produced by sublimate and acetic 
acid. 

In most cases the material was embedded in celloidin and 
then in paraffin, and cut in sections ‘004 or ‘005 mm. thick ; 
this method facilitated the orientation and gave good results, 
but to get thinner sections paraffin alone was used. ‘The 
sublimate material was generally stained in bulk with borax 
carmine, the rest on the slide with hematoxylin. Sections 
were made of all stages from the young gastrula up to the 
ninth day after hatching, but the earlier embryonic stages 
can be most satisfactorily studied in the living egg, which is 
extraordinarily transparent. After about the sixth hour, 
however, the structure becomes somewhat complicated, and 
sections are therefore of great value in confirming and 
amplifying the observations made on the living embryos and 
young. 

For staining larve for mounting whole Mayer’s carmalum 
gave much the best results. 


III. Empryonic DEVELOPMENT’. 


- ' ae 
"he eggs of Sagitta are about 2 mm. in diameter, and are 
Jaid in the early morning, and develop floating not far from 


ON THE DEVELOPMEN'T OF SAGITTA. 357 


the surface. ‘They are enclosed in a firm transparent shell, 
which may almost fit the egg, as in 8. enflata, or may leave 
a large space containing fluid in which the egg is suspended 
(S. bipunctata). 

O. Hertwig’s account (8) of the early embryonic develop- 
ment, as seen in living embryos, is so complete and accurate 
that it is unnecessary for me to go over it in detail. I con- 
firmed in all points his account of the cleavage, gastrulation, 
and formation of the archenteric folds, and further the closing 
of the blastopore and origin of the mouth, and found that the 
genital cells appear in the archenteron as he describes. They 
lie slightly either dorsally or ventrally of the middle line, but 
I was not able to determine on which side. This eccentricity 
is shown in one of Hertwig’s figures, but not mentioned in 
his text. In one important point, however, my observations 
differ from his, and confirm the account of the development 
given by Biitschli (2). After the mouth has been formed 
the two lateral lobes of the archenteron become much 
restricted by the narrowing of the embryo, and their anterior 
ends are then separated off as distinct cavities by the meeting 
and fusion of their walls. he two anterior “ head-cavities ” 
so produced lie at the sides of the pharynx (PI. 19, figs. 3—5), 
and are from the first very small and soon become obliterated ; 
but the mesoblast enclosing them persists, separated off from 
the remainder, and gives rise to the mesodermal structures of 
the head. The size of the cavities varies in different species 
according to whether the egg-shell is large and leaves plenty 
of room, or the reverse. In 8. bipunctata they are easily 
recognisable, while in 8S. enflata they are from the first 
mere dots, and might easily be overlooked were not other 
species available for comparison (Pl. 19, fig. 4). 

The backward growth of the folds and displacement of the 
genital cells into the coelomic cavities takes place as Hertwig 
describes, and at this stage a faint line can sometimes be 
seen running back on the dorsal wall of the archenteron from 
the free end of the folds to the point where the blastopore 
closed. This looks in the living embryo rather as if the 


358 L. DONCASTER. 


dorsal wall of the archenteron took some share in the back- 
ward growth, but sections lend no support to this assumption. 
The folds grow back to the posterior end of the archenteron, 
separating off the coelomic cavities completely, and the folds 
become pressed together so as to form a thin longitudinal 
septum running through the animal. As the embryo elongates 
its cavities become reduced, and this is further increased by 
the structure of the somatic mesoderm, the inner boundaries 
of which become irregular, and cracks appear between the 
cells, so that the layer becomes very indistinct and appears 
like mesenchyme. ‘This is perhaps the origin of Jourdain’s 
statement that the mesoblast arises by delamination between 
the ectoderm and endoderm, and not by archenteric diver- 
ticula. 

The gradual elongation and narrowing of the embryo 
causes the obliteration of the cavities, and the whole embryo 
becomes solid, as described by Hertwig. As it grows in 
length it curls in the shell, and this ventral curvature 
becomes more pronounced as development proceeds, so that 
in S. bipunctata the tail meets the head (PI. 19, fig. 6), and 
in 8. enflata the embryo is curled through fully a turn and 
a half before hatching. In the latter species, in which the 
shell fits the egg, the curvature begins much earlier than in 
the large-shelled species. When the embryo begins to curl 
it is easy to get optical transverse sections showing the 
cephalic mesoderm at the sides of the stomodeum in the 
head, and in the trunk the two semicircular mesodermal 
masses separated by the thin endodermal septum, which 
expands somewhat dorsally and ventrally. My observations 
on the origin of the ganglionic rudiments and the remaining 
changes before hatching are so nearly in accord with 
Hertwig’s that it is not necessary to give them in detail. In 
the species that I studied hatching took place at a time 
varying from sixteen to fifty hours after the eggs were 
laid, differing according to the temperature and the species. 


ON THE DEVELOPMENT OF SAGITTA. 359 


IV. FurrHer Stupy or tHe Empryonic DEVELOPMENT BY 
MEANS OF SECTIONS. 


In the earliest stages of the embryonic development, 
sections show very little which cannot be seen in the living 
embryo. ‘The earliest sections made were of the egg shortly 
after gastrulation had begun, when the archenteric cavity 1s 
still small. They show comparatively few large cells, with 
moderately distinct cell boundaries, and the nuclei arranged 
close to the outer limits of the cells, in the epiblast, and at 
the free surface in the cells lining the archenteron. I have 
not been able to make out with certainty whether the genital 
cells are already differentiated at this stage, but in one 
section two nuclei larger than the rest le side by side close 
to the archenteric cavity, and it seems probable that these 
are the genital cells. In sections of a later gastrula there is 
no very important change ; the cells are much more numerous, 
so that there is an almost continuous band of nuclei round 
the outside of the embryo and round the archenteron. Two 
of the latter may be seen to be larger than the rest, and 
project a little into the cavity. At a stage when the 
archenteron is divided into three branches by the folds, 
already four genital nuclei are seen, although the cells still 
appear as two when seen alive; but as the four are packed 
so closely together, they would not be easily distinguishable 
in the living state. (Pl. 19, fig. 7, only one genital cell 
[gen. c.] appears in this section.) 

At a period slightly later than this the closing of the 
blastopore may be seen; it lies now not quite terminally, but 
slightly in front of the posterior end, but I do not know with 
certainty whether dorsally or ventrally, owing to the absence 
of anything to distinguish the dorsal from the ventral surface 
at this stage. Observations on the living embryo, however, 
lead me to believe that it is the ventral surface on which the 
blastopore comes to lie before it closes. 

As the embryo elongates, and the different layers come 


060 L. DONCASTER. 


into such close contact that observations in the living state 
are more difficult, the sections become more necessary. At 
a stage when the curvature of the embryo is still not much 
marked, when the mouth invagination and the head cavities 
have recently been formed, a transverse section through the 
head (PI. 19, fig. 9) shows the epiblast thickest dorsally and 
laterally, and quite thin on the ventral side. The mouth 
invagination is seen opening ventrally, and on each side of 
it are the two masses of cephalic mesoderm in which the 
cavity is already obliterated, but the nuclei can be seen 
arranged in a double row, showing how the cavity originally 
lay between them. These masses of cephalic mesoderm are 
dorsally in contact with one another, but ventrally they are 
separated by the mouth invagination, so that their shape in 
transverse section resembles that of a horseshoe. When 
traced back into the neck it is seen that they overlap the 
front ends of the trunk mesoderm, which lies nearer the 
middle line, so that in some sections parts of both can be 
seen at once, and hind ends of the head mesoderm masses 
lying at the outer sides of those of the trunk. 

The ectoderm of the head, besides its anterior dorsal 
thickening which gives rise to the cerebral ganglion, is also 
thickened at the sides in the mouth region, where it forms a 
layer, two or three cells deep laterally, while it is only one 
cell thick on the dorsal surface. This thickening is the 
rudiment of the hood (‘Kappe’’), the formation of which 
will be more fully described later. 

The alimentary canal in the head fills up a considerable 
space; it is roughly oval in transverse section just behind 
the mouth, and consists of a layer of well-defined cells which 
in the earlier embryos enclose a distinct space, but this 
disappears later. The alimentary canal in the head is 
entirely derived from the ectodermal invagination by which 
the mouth is formed, and the true endodermal part only 
begins in the neck region. As the embryonic curvature 
proceeds, the mouth lies more on the ventral (inner) side, 
and becomes elongated, so as to form a somewhat slit-like 


ON THE DEVELOPMENT OF SAGITTA. 361 


opening lying longitudinally on the ventral side of the head, 
and it therefore is visible in a number of consecutive trans- 
verse sections. When traced backward to the neck the 
alimentary canal suddenly becomes laterally compressed, and 
lies more toward the dorsal surface of the body, marking the 
point where the ectoderm of the stomodzum meets the true 
endodermic gut. ‘The latter is from the first very narrow, 
and appears simply as a sort of partition between the masses 
of mesoblast lying on each side (PI. 19, fig. 8). Later it 
becomes still narrower, and forms a thin lamina, slightly 
thicker dorsally and ventrally than in the middle, and this 
condition persists until a considerable time after hatching 
(Pl. 19, figs. 13, 14, end. sep.). Just at the posterior end, 
however, this extreme lateral compression does not take 
place, and there, up till near the end of embryonic life, the 
cellular nature of the septum remains visible (fig. 14, 
end. sep.). 

The mesoblast of the trunk and tail region is from the 
beginning sharply distinguished into splanchnic and somatic 
layers, which have different origins, the somatic being 
derived directly from the primary hypoblast, the splanchnic 
from the outer walls of the folds. In early stages a distinct 
coelomic cavity is seen enclosed by the mesoblast ; it is tri- 
angular in transverse section, and placed so that the 
splanchnic layer forms the base, and the somatic layer the 
other two sides of the triangle (fig. 8). In these ccelomic 
spaces lie the genital cells, two on each side, and in contact 
with the splanchnic layer ; and behind them, at a stage when 
the folds have not yet finished growing backwards, the two 
cavities open into one another, and a transverse section shows 
a single archenteric cavity. 

This condition with open ccelomic spaces does not, how- 
ever, persist very long. As the embryo increases in length it 
becomes correspondingly narrower, and since itis enclosed in 
the egg-shell its growth is restricted, and apparently in con- 
sequence of this its internal cavities become obliterated. In 
sections rather later than those described above (PI. 19, figs, 


362 L. DONCASTER. 


9—12) it is seen that no coelomic space is present, and that 
the mesoderm forms two strands running through the embryo 
and separated from one another by the endodermic septum. 
These mesodermic strands appear in cross-section roughly 
circular in outline, and near the circumference the cell 
boundaries are well marked ; but it is seen that the nuclei are 
no longer at the inner ends of the cells, and that beyond them 
the space which in early embryos was free and unoccupied is 
now filled with cell-substance, in which the boundaries 
between the separate cells are very indistinct. This condi- 
tion appears to have been brought about, firstly, by the com- 
pression of the whole embryo and consequent reduction of its 
cavities; and secondly, by the sinking of the mesodermic 
nuclei towards the bases of the cells, that is towards the 
plane separating the mesoderm from the ectoderm. The cell 
protoplasm of the centre of the mesodermal strands, that is 
within the ring of nuclei, is of a much looser and more watery 
character than that near the circumference, which probably 
accounts for the lack of distinction between the cells; in 
embryos in which the cytoplasm is less well preserved it tends 
to form a mass of strands between which are clear spaces. 
That this is not due entirely to faulty fixation and consequent 
maceration is certain from the study of the living embryo, in 
which an exactly similar appearance is seen at this stage as 
was described in Section III. The basal part of the somatic 
mesodermal cells is composed of much firmer protoplasm, 
and the lines separating the cells can be seen distinctly at 
this stage, and also when the embryo is more advanced. As 
development progresses the nuclei become aggregated in a 
dorsal and ventral mass on each side, while in the lateral 
areas they disappear or become very scarce. Between the 
nuclei and the base of the cells the protoplasm is becoming 
modified, so that it now takes a deep brown stain with osmic 
acid (figs. 13, 14), and a longitudinal section shows that the 
cells are becoming elongated in the direction of the animal’s 
length. The study of the larva after hatching shows that 
these dorsal and ventral groups of cells of the somatic meso- 


ON THE DEVELOPMENT OF SAGITTA. 363 


derm give rise to the longitudinal muscles of the body, and 
that while the greater part of the cell-substance is trans- 
formed into muscle, the nucleus and a little protoplasm 
remain at the inner end and ultimately form the lining of the 
ecelom. ‘his will be further described in dealing with the 
sections of larvee. 

It will be convenient at this point to describe more fully 
the four genital cells, which have now reached their per- 
manent condition. ‘hey lie one behind the other in the 
coelomic cavity on each side, and when this is obliterated 
they become embedded in the mesoderm and lie at the sides 
of the endodermic septum. ‘The male and female cells appear 
exactly alike, and are characterised by their extremely large 
nuclei, each of which is generally oval in shape, is enclosed 
by a definite membrane, and contains numerous nucleoli. 
The latter are arranged round the edge of the nucleus, close 
to the nuclear membrane, and they are generally of a rather 
elongated oval shape. In the remainder of the nucleus a 
network of fine threads can generally be made out, and since 
only the nucleoli take up the stain, in some sections the 
genital nuclei appear to be made up of a large number of 
very small cells, each with a stained nucleus. Closer exami- 
nation with a high power, however, shows that this is not the 
case; the genital nuclei are exactly like those of the rest of 
the body, except that they are more than twice as large, and 
the nucleoli are less crowded together. The cell protoplasm 
of the genital cells is small in amount, and its limits are hard 
to see, for when embedded in the mesoderm they become 
enclosed in a sort of envelope of mesodermal cells, which, 
however, do not at this stage form a definite epithelial 
sheath. In some sections this mesodermal envelope is not 
conspicuous, but in others (Pl. 19, fig. 15) small nuclei can be 
seen closely appressed to the genital cells, and these differ 
from the other nuclei of the body in staining evenly through- 
out, instead of consisting of a mass of nucleoli embedded in a 
colourless matrix. 

The ectoderm of the trunk is characterised chiefly by the 


364 L. DONCASTER. 


development of the ventral ganglion. In the early stages 
the ventral ectoderm does not differ from the dorsal, but 
when the archenteric folds have not yet reached the pos- 
terior end of the embryo, the ectoderm of the ventral 
surface from just behind the head along the greater part of 
the length of the body undergoes changes somewhat similar 
to those described above in the somatic mesoderm. ‘The 
nuclei, which at first lie at the outer edges of the cells, sink 
in till they reach their bases, and the cells at the same time 
become larger than on the dorsal surface (fig. 8). There is 
then a proliferation of nuclei along two bands lying in the 
ventro-lateral areas, and at the same time the whole ectoderm 
of the ventral half of the body increases in thickness (figs. 10, 
11). In this way two bands of closely packed nuclei appear 
just beneath the developing muscle-cells of the somatic 
mesoderm, and when the latter are deeply stained the two 
look very similar, but are always separated by a clear 
dividing line. In the dorsal half of the body, and also 
ventrally in the tail, this thickening and immigration of the 
nuclei does not take place, and the cells retain longer their 
original character. As development proceeds the ectodermal 
cells along the mid-dorsal region become very thin, and con- 
stitute a mere membrane separating the mesoderm from the 
exterior ; but traced downwards from the mid-dorsal line the 
ectoderm becomes steadily thicker, and is thickest along the 
ventral middle line. 


V. DEVELOPMENT OF THE YOUNG SAGITTA. 


The young Sagitta at the time of hatching differs very 
markedly in structure from the adult, so that the term larva 
might almost be used to describe it, especially as a rather 
sudden change of structure takes place after a few days, 
which may be compared with a metamorphosis. It is about 
1 mm. in length, and so transparent as to be almost invisible, 
but can be recognised by the naked eye as a minute shining 


ON THE DEVELOPMENT OF SAGITTA. 365 


body, especially when it swims. The larve, as a rule, le 
motionless near the surface, but swim in a jerky manner 
when disturbed, just as does the adult Sagitta. Inthe young 
at hatching (PI. 20, fig. 16) the tail-fin is already present to 
some extent in all cases, but in 8. enflata there is already a 
lateral fin beginning to be formed, continuous with the tail- 
fin and extending forward as far as the genital cells. The 
head is rounded and separated from the trunk by a slight 
neck. ‘The line separating the ectoderm from the underlying 
mesoderm is sharply defined, and the mesoderm forms a 
solid cord running through the length of the animal, sepa- 
rated into two halves by the endodermic septum (end. sep.), 
which is expanded in the head into a bulb, like a thermo- 
meter. The ectoderm of the anterior half of the body is 
much thickened ventrally, forming the rudiment of the 
abdominal ganglion (gang. v.); and scattered about the 
surface, especially in the head region, are tactile organs, like 
those of the adult, consisting of extremely fine bristles 
arranged in a fan-like manner transversely to the length of 
the animal (¢. 0.). At the base of each of these organs there 
are groups of sensory cells, which become more obvious in 
the later stages; but one of them, which later is very 
pronounced, is already noticeable just in front of the tail on 
each side. 

Not much of the internal structure can be seen in the 
living larve, but the four genital cells can generally be 
observed embedded in the mesoderm in close contact with 
the endodermal septum, just behind the middle of the 
animal. ‘They are seen better in specimens stained and 
mounted in balsam, and these show the rudiments of longi- 
tudinal bands of muscle just below the ectoderm. 

On the second day the larva has slightly increased in 
length and the fin has grown larger, but has still a very 
ragged and irregular appearance. Early on the third day 
no pronounced change has taken place, the fin is larger and 
more regular, long fin-rays having appeared at even intervals. 
The muscle-bands are more developed, and show a transverse 


366 L. DONCASTER. 


striation as in the muscles of the adult. During the third 
and fourth days very considerable changes take place, giving 
rise to a type of structure closely resembling that of the 
adult animal. Up to this time the body has been solid, 
containing no cavities whatever, and the first indication of 
the change is the formation of a cavity in the mesoderm of 
each side of the body and also in that of the head. At the 
same time the endodermic septum dividing the two meso- 
dermic cords from one another becomes thickened in its 
anterior half, namely, as far back as the genital cells, so 
marking out the rudiment of the alimentary canal (al.), 
though a lumen does not appear in it for several days. In 
the head, however, a cavity appears in the front end of the 
alimentary canal, which was formed in the embryo by 
epiblastic invagination, so that a buccal cavity is now present 
opening by the mouth to the exterior. The rudiment of the 
hood (“ Kappe ” or “ prepuce”’) of the adult has been form- 
ing as an ectodermal fold, and at this stage has reached 
considerable proportions, but since it can only suitably be 
studied in sections it need not be dealt with here. Under it, 
however, at this stage four or five small hooks make their 
appearance, the most posterior of which is the longest, while 
the front one is hardly visible; as they increase in size new 
ones begin to grow in front, until the normal number for the 
species is reached (PI. 20, figs. 18, 19, hks.). 

The ectodermic swelling in the front of the head, which 
was already visible in the embryo, has grown larger and 
forms the rudiment of the brain (gang. cb.), while at the sides 
of the mouth a rounded body appears on each side, which is 
seen in section to be the lateral ganglion of Hertwig (the 
vestibular of Grassi). The eyes (e.) also appear at this stage 
as a pair of minute black specks on the surface of the head, 
so small that their structure cannot be made out or compared 
with that of the adult, except that, as in the adult, the black 
pigment is surrounded by an oval transparent area. The 
muscles of the head also take their definitive form during the 
third and fourth days; before that time the cells have had 


ON THE DEVELOPMENT OF SAGITTA. 367 


an embryonic character, but now they develop into muscle- 
fibres arranged as in the adult, with special reference to the 
movements of the mouth and the action of the hooks. 

Behind the head striking changes are also taking place. 
The abdominal ganglion is becoming somewhat reduced in 
size, and at each end of it the ectodermal cells assume a 
remarkable structure, which is especially pronounced in the 
neck region, but occurs also at the posterior end of the 
ganglion, between it and the fin, which now extends con- 
siderably further forward than the genital cells. The ecto- 
dermal cells in the parts indicated lose their cell-contents 
almost entirely, and become so vacuolated as to resemble the 
parenchyma of a plant in appearance; this occurs most 
markedly at the sides of the body, but to a less extent 
dorsally and ventrally also (Pl. 20, figs. 18, 19, pt.). The 
cells so modified resemble exactly the curious vesicular tissue 
found in the trunk of Spadella draco, the only difference 
being that the cells in the Sagitta larve are much smaller 
and are found in a much less area. This parenchymatous 
tissue 1s most prominent about the fourth and fifth days, but 
it persists as long as I have been able to keep the young 
animals alive (fifteen days). 

In the tail region the fin has become divided into two 
parts, a tail-fin (f. ¢.), extending a considerable distance 
forward along the body, but separated by a gap from the 
lateral fin of each side. In the gap is the large tactile organ 
mentioned above, situated on a prominent ectodermal swell- 
ing (fig. 18, t. 0.). There is at this stage no trace whatever 
of the anterior pair of fins of the adult, so that the young 
animal closely resembles the genus Krohnia, in which they 
are permanently absent, and in which the tail-fin extends 
forward as in the young Sagitta. 

The most important changes which take place during the 
third and fourth days are those affecting the ceelom. The 
solid condition of the mesoderm described above persists 
until the third day after hatching, but on that day, unless 
development is retarded by cold weather, the mesoderm cells 


368 L. DONCASTER. 


begin to separate again into two layers, a somatic layer next 
to the body-wall, and a thin splanchnic layer surrounding the 
endoderm, and in this way the ccelomic cavities reappear. 
At first there are two small cavities in the head, and two 
large ones extending the whole length of the body, those in 
the head being separated from those behind by a transverse 
septum. These may be regarded as the same as those 
formed early in embryonic life, when the head cavities are 
separated off from the rest of the cclom. After the two 
posterior ccelomic cavities have reappeared, and extend 
through the whole length of the body, important changes 
take place in the region of the genital cells, which result in 
the division of the ccelom into an anterior and posterior part, 
and so give rise to the condition found in the adult. 

Up to this time the genital cells have lain embedded in 
the mesoderm a little behind the abdominal ganglion, and 
pressed close to the alimentary canal; they are oval in shape, 
and their longest direction coincides with the long axis of 
the body (fig. 17). At about this time, however, they begin 
to change their position, and gradually come to lie with their 
long axes placed transversely to the length of the animal. 
They then move slowly across the coelomic cavity until they 
reach the body-wall on each side, when they again come to 
lie end to end with their long axes in the same direction as 
the length of the body. This process is gradual and takes 
several hours, and while the large and conspicuous oval 
nuclei travel across, protoplasmic connections can be seen 
still attaching them to the wall of the alimentary canal, and, 
when they approach the body-wall, also to the splanchnic 
mesoderm (Pl. 20, figs. 20, 21). 

While they are traversing the ccelom the two of each pair 
lie side by side close together, but not in contact, and 
during their progress a transverse septum (sep. tr.) is formed 
between them, so that when they arrive at the outer sides, 
against the body-wall, a septum is left across the ccelom, 
dividing the body-cavity of each side into an anterior trunk 
and a posterior tail portion, as inthe adult. ‘The way in which 


ON THE DEVELOPMENT OF SAGITYA. 369 


this septum is formed is not absolutely clear, for the small size 
of the animal and the extreme minuteness of the cells made the 
actual process difficult to follow with certainty. There are 
two possible means by which such a septum might arise: 
first, by the splanchnic mesoderm rising up as a fold, and 
carrying the genital cells across with it till they reached the 
body-wall (or, what would amount to the same thing, the 
genital cells moving across and drawing the splanchnic 
mesoderm with them in the form of a fold); secondly, the 
septum might arise by the coalescence of cellular envelopes 
in which the genital cells are enclosed. Such envelopes were 
described and figured by Hertwig, but in the species which I 
have studied they have been exceedingly difficult to make 
out, and I have only rarely been able to see them. By 
comparing a number of larvee, both mounted whole and in 
sections, there seems to be no doubt that the genital cells 
are enclosed in a membrane which is separate from them, 
and which contains here and there a few nuclei (PI. 21, fig. 
32). The nuclei are much less numerous than those repre- 
sented in Hertwig’s figures, but they indicate that the 
envelope is a cellular structure, which is no doubt derived 
from the mesoderm in which the genital cells have been 
embedded since an early embryonic stage. When the 
genital cells move across the body-cavity their envelopes are 
elongated transversely to the body of the animal, and 
between the two cells their respective envelopes lie parallel 
with one another, almost, if not quite, in contact. If when 
the migration of the genital cells begins their envelopes 
remain attached to the splanchnic mesoderm at the point 
between the genital cells, while elsewhere they become free 
and move across with the cells which they enclose, when the 
latter have crossed the cavity and reached the body-wall a 
two-layered septum will have been produced across the body, 
with the two genital cells lying on opposite sides of it. 
After comparing a large number of larva, alive and stained, 
it appears to me that the septum is formed in this way, for I 
have never seen any indication of a fold of the splanchnic 
VOL, 46, PART 2,—NEW SERIES, Z 


370 L. DONCASTER. 


mesoderm, either in the early or later stages of the migration. 
The body-cavity is narrow dorso-ventrally, so that the cells 
appear to be in contact with both upper and lower walls 
during their passage. In many larve the septum appears 
from the first single, but it can be seen in some that it is two- 
layered, for the two layers are not in contact (fig. 21). This, 
however, would also be the case if it were formed by a fold. 
As the genital cells cross the cavity they retain connections 
with the splanchnic mesoderm for a time, apart from the 
septum; the anterior (female) cell generally has a strand of 
tissue crossing to the wall of the alimentary canal in front of 
the septum, and separated from it by a space, while the 
posterior (male) has a similar strand behind the septum (figs. 
20, 21). These connections are fainter than the true septum, 
and appear to contain no nuclei; when the genital cells 
reach the body-wall and take up their permanent position 
the connections with the splanchnic mesoderm disappear. 

At first there are only very few nuclei in the septum, and 
one of these generally appears at its outer end, next to the 
body-wall, and this makes that end of the septum very con- 
spicuous as soon as it is complete, almost suggesting at times 
that a fold of the somatic mesoderm is forming, although this, 
as a matter of fact, does not take place. ‘The nuclei of the 
envelopes seem to get collected between the genital cells, and 
so appear in the septum, while on the outer sides of the two 
cells no nuclei appear during the migration. 

Taking all the facts together, it seems almost certain that 
the transverse septum is formed in the way here described ; 
but if it should turn out that it is formed by a mesoblast 
fold, as suggested by Hertwig, the difference is really of less 
importance than appears at first. The envelopes of the 
genital cells are mesoblastic structures, derived from the 
mesoderm in which the cells have been embedded, so that 
in either case the septum is formed by a double sheet of 
splanchnic mesoderm in relation with and in consequence of 
the migration of the genital cells from one side of the body- 
cavity to the other. In any case the septum is formed 


ON THE DEVELOPMENT OF SAGITTA. 371 
\ 


directly in connection with this migration, and in a distinctly 
different way from that which gives rise to the anterior 
septum between the head-cavities and the rest of the ccelom. 
The theoretical bearing of these points will be discussed more 
fully later. 

The completion of the septum between the trunk and tail 
brings to an end the period of development which may be 
described as larval, for the animal has now essentially the 
form and structure of an adult. It still differs from the 
latter in many important respects,—for example, the ovaries 
and testes of each side are still each represented by a single 
cell, and there are no genital ducts. The alimentary canal is 
without a lumen, except in the mouth region, and the anus 
has not yet been formed. The abdominal ganglion is also 
enormously larger than in the adult, relatively to the size of 
the animal; there are no anterior fins, the parenchymatous 
tissue in the neck region and behind the ganglion is very 
conspicuous, and many minor differences still exist. he 
changes, however, which transform the young of this stage 
into the adult condition are very gradual, extending over 
several weeks, if not more, and comparatively few could be 
observed in the young Sagitta raised from the egg. 

During the first few days after hatching the alimentary 
canal increases steadily in thickness, and its nuclei become 
prominent, but the thicker portion ends abruptly opposite 
the genital cells, so distinguishing the true gut from the 
longitudinal tail septum, with which it is continuous. A 
lumen begins to appear in the alimentary canal shortly after 
the reappearance of the ccelom. As the cclomic cavities 
grow wider the alimentary canal becomes supported by a 
dorsal and ventral longitudinal mesentery, and about the 
seventh day its end can be seen bending down in the 
mesentery to touch the ectoderm. On the eighth or ninth 
day the anus is formed at the junction of ectoderm and 
endoderm ; it is not exactly opposite the transverse septum, 
but a short distance in front of it, and a space is left between 
the two layers of splanchnic mesoderm, extending from the 


372 L. DONCASTER. 


posterior end of the gut to the transverse septum. The 
space is in the ventral half of the body; dorsally the longi- 
tudinal mesentery extends continuously from the trunk into 
the tail. This arrangement is also found in the adult 
Sagitta, but is most pronounced in 8. lyra, in which the 
distance between the anus and the transverse septum is 
considerable. When the anus is formed a swelling can 
generally be seen in the rectal portion of the gut, just in 
front of it, and the endoderm cells develop cilia, which are 
especially active in this region. Small particles can be seen 
circulating in the distended rectum, so that the young 
animal probably begins to take food at this stage. 

On the eighth day a horseshoe-shaped group of nuclei is 
seen in stained specimens on the back of the head, rather 
behind the eyes; the two points of the horseshoe are 
directed backwards, and the nuclei are arranged in a double 
line. It is probable that this is the rudiment of the ciliated 
“ olfactory organ” (the “ corona ciliata ”’ of Grassi). 

The abdominal ganglion begins to be gradually reduced in 
size relatively to the rest of the body, and its elements 
become more completely separated, so that the nuclei are 
packed in a dense mass at each side of the ganglion, while 
the mid-ventral region appears clear, and is composed 
exclusively of fibres. This gives the ganglion a markedly 
bilateral appearance, especially in stained preparations. 

In the head the muscles have already assumed the arrange- 
ment of the adult by the sixth day, and the cerebral and 
vestibular (lateral) ganglia are now conspicuous. The hood 
is arranged just as in the adult, and a few days later the 
cuticular rods which support the teeth (“ Stutzplatten” of 
Hertwig) make their appearance, running forwards and 
inwards from the region of the hooks, and ending in front 
below the cerebral ganglion. No teeth, however, are as yet 
present. 

These changes are all practically complete on the tenth or 
twelfth day, and on a few occasions when I succeeded in 
keeping the young alive beyond that time no further altera- 


ON THE DEVELOPMENT OF SAGITTA, 373 


tions took place. The further development, leading to the 
adult condition, has therefore to be studied in specimens 
taken in the tow-net, and the consideration of it will be post- 
poned until after the sections of the early young have been 
discussed. 


VI. FurtHer Stopy or LarvaL DEVELOPMENT BY MEANS 
OF SECTIONS. 


The general course of the development after hatching up 
to the end of the second week has been described in the last 
section, and therefore only those points will be dealt with 
here which are better seen in section than in the living or 
mounted animal. It will be most convenient to consider first 
the sections of the head at different stages, and afterwards 
those of the trunk and tail. 

A transverse section through the mouth region of a larva 
killed a few hours after hatching differs little from that 
through the embryonic head; it is rather smaller, owing to 
the lengthening of the body at the expense of its width, and 
it is also becoming more obvious that the ectoderm at the 
sides of the head is becoming thickened and consists of more 
than one layer of cells (hd.). 

On the second day (fig. 22) little change has taken place ; 
the nuclei on the ventro-lateral parts of the ectoderm are 
more numerous, and at the sides of the head that layer shows 
signs of splitting, but otherwise the structure is closely 
similar to that of the first-day larva. By the third day, 
however, development has proceeded considerably; the 
tissues are beginning to take on their permanent form, and 
in consequence the definite structure of the head becomes 
marked out. 

Beginning with the ectoderm, it is seen in fig. 23 that the 
hood has now been formed by a splitting off of the two outer 
cell layers at the sides of the head, but on the dorsal surface 
they remain continuous with the general ectoderm, and so 


374 L. DONCASTER. 


appear to constitute a fold on each side (hd.). In sections 
taken more anteriorly these two folds are continuous with 
one another ventrally, and form a membrane covering the 
anterior part of the mouth; but further back they do not 
reach so far down, and appear in section as flaps at the sides 
of the head (figs. 23—26). There is a tendency for the two 
layers of the hood to split apart from one another in the 
region of its insertion, and so give rise to a cavity (hd. cav.) 
which is sometimes large in sections, but in life apparently 
always narrow if present at all. 

It is also seen that the epidermis is thickened under the 
hood, especially near the insertion of the latter; this con- 
dition persists in the adult in the anterior part of the head, 
and, since the thickening is in just the region from which the 
hooks ( Greifhaken”) grow out, it is possible that it is 
connected with their formation. Just at the corners of the 
mouth there is an aggregation of nuclei which will give rise 
to the lateral ganglia (gang. 1.) (called vestibular by Grassi). 
The nuclei plainly belong to the ectoderm, and some sections 
show the rudiments of the ganglia extending further forward 
than the mesoderm, and therefore Hertwig’s supposition that 
these ganglia belonged to the mesoderm must be regarded as 
incorrect. On the dorsal surface a lens-shaped thickening of 
the ectoderm is now visible on each side, forming the rudi- 
ment of the eyes which appear in this stage (e.). 

The mouth and alimentary canal of the head have not 
altered much on the third day, but the cell layers are 
becoming more definite, and a small cavity has already 
appeared. 

The mesoderm is beginning to be transformed chiefly into 
muscle, but on the third day this change is not yet complete ; 
the cells, however, have a looser and less regular appearance 
than before. 

During the fourth and fifth days the structures which were 
outlined on the third day are further developed, and by the 
sixth day they are approaching completion. Figs. 24—26 
represent three sections through the head of an individual of 


ON THE DEVELOPMENT OF SAGITTA. 379 


this age. The most prominent objects in the first two of 
these are the vestibular ganglia lying at the sides of the 
mouth, which are now almost completely developed, and 
consist of an inner mass of “ Punktsubstanz,” surrounded by 
a layer of deeply staining nuclei. 

The cerebral ganglion is also complete by this time; it lies 
at the anterior extremity of the head and to a large extent in 
the hood, in which it extends ventrally nearly as far as the 
mouth. In the adult it becomes more restricted, as does the 
abdominal ganglion. At this stage it consists of an outer 
layer of nuclei covering a deeper mass of ‘‘ Punktsubstanz,” 
but the latter is more restricted than the nuclear layer. Fig. 
25 shows on the dorsal surface of the mouth region two large 
masses of nuclei in the ectoderm, representing the eyes (e.), 
while in this figure and in fig. 24 the hooks are seen lying at 
the sides of the head, covered by the hood (hks.). 

The development of the mesodermal structures has now 
proceeded considerably, and the chief muscles of the head 
are already differentiated. Their general arrangement is 
indicated in fig. 26 (h. mus.), where it is seen that they do 
not differ much from those of the adult. Behind the mouth 
the pharynx lies near the dorsal surface, and when the 
muscles are formed a cavity appears just below it, which 
corresponds with the cavity called by Hertwig the head 
celom. A little further forward a cavity is present on each 
side more dorsally (figs. 24, 25, b. c.1.), and these are 
undoubtedly ccelomic. Already about the third day the 
pharynx is seen to be surrounded by a layer of cells which 
belong to the mesoderm, corresponding to the splanchnic 
layer (fig. 23), but, as in the trunk, these cells later be- 
come so closely connected with the alimentary canal as to 
be indistinguishable from it. About the fifth day, however, 
a cavity appears on each side between these cells and the 
outer layer of cephalic mesoderm, which is now being 
changed into muscles. This cavity is the head coelom, and 
that below the pharynx further back appears to be also 
ccelomic, and to be formed by the coalescence in this region 


376 L. DONCASTER. 


of the originally paired head cavities. At this stage the 
head-ccelom is clearly separated from the hood cavity, but 
later the latter seems to be obliterated, and the ccelom extends 
into the base of the hood. 

The alimentary canal of the head has now a well-defined 
cavity, which extends back into the neck; in its walls cell 
boundaries can no longer be distinguished, and they appear 
as continuous masses of tissue with nuclei at intervals. 

The latest sections made were of ninth-day larve (PI. 20, 
fig. 27); these show the same structures as are described 
above, but are further advanced, and approach more nearly 
the adult condition. 

Before turning to the structure of the trunk, that of the 
neck must be shortly described. When a series of sections 
of the sixth day is followed backward, a little behind that 
represented in fig. 26, a pair of oval cavities appear suddenly, 
one on each side, between the pharynx and the epidermis 
(fig. 28). These are the anterior ends of the trunk ccelom, 
which diverge a little in front (cf. the figures of the living 
head), and are overlapped both dorsally and ventrally by the 
mesodermal structures of the head. Followed back, how- 
ever, the latter disappear rapidly, and at the same time the 
pharynx bends towards the ventral surface, so that the two 
coelomic cavities, which take up the greater part of the 
section, are separated dorsally by a thin mesentery. The 
epidermis begins at this point to assume the vacuolated 
structure described above, but this is, as a rule, not well 
shown in sections owing to shrinkage; sometimes, however, 
as in fig. 29, p. ¢t., it is well seen. A paired mass of nuclei is 
also seen in the dorsal epidermis, which is the rudiment of 
the “olfactory organ” (“corona ciliata,” ol.). A little 
further back the ectodermal pharynx joins the true gut, the 
change being marked by the alimentary canal becoming very 
narrow and lying hke a septum from the dorsal to the 
ventral surface, supported at each end by a short mesentery, 
instead of being thick-walled and lying against the ventral 
body-wall, 


ON THE DEVELOPMENT OF SAGITTA,. 377 

The trunk and tail as seen in sections now remain to be 
described. On comparing a section through the trunk of a 
first-day larva with that of an embryo, the only differences 
are that the animal is more flattened dorso-ventrally, the 
cells which will give rise to the ganglion are more marked 
off from the epidermis, and the cells of the somatic mesoderm 
are more collected together. Traced backward, the ventral 
thickening dies out, and the animal becomes a narrower oval 
in section, with epidermis one cell thick covering the meso- 
derm dorsally and ventrally ; but laterally, where the fin is 
beginning to form, it remains slightly thickened. A little 
behind the ganglion the genital cells are seen pressed against 
the septum which divides the mesodermal strands. Behind 
them the animal gets gradually thinner, and more compressed 
dorso-ventrally, but otherwise there is no important differ- 
ence of structure. 

On the second day (Pl. 21, fig. 31) the most noticeable 
change is the reduction in size of the ventral ganglion. This 
has taken place by a shrinking of the cells; the nuclei remain 
as before, aggregated ventro-laterally, but the cell protoplasm 
which was so conspicuous between them and the epidermis 
has now largely disappeared. In the mid-ventral region, 
just below the alimentary canal, the rudiment of the fibrous 
part (“ Punktsubstanz”’) of the ganglion is appearing. Some 
of the nuclei (gang. nuc. 2) in the ganglion are noticeably 
larger than the others. In some specimens, both alive and 
in section, round cavities appear in the ganglion, one or two 
on each side, but as they are rare and seem to occur irregu- 
larly in the few cases where they are present, they are 
probably pathological. A change has also taken place in the 
mesoderm ; some of the nuclei of the somatic layer have again 
moved inwards owing to the growth, between them and the 
outer limit of the mesoderm, of a compact, rather faintly 
staining mass of tissue, most of which gives rise to the 
muscles (mus. c.). Behind the ganglion region the only point 
of importance is the further lateral outgrowth of the epidermis 
into the rudiment of the fin as a sort of lateral fold. 


378 L. DONCASTER. 


In the third-day larva it is seen at once that from the head 
as far as the genital cells the septum dividing the two halves 
of the mesoderm is much thicker in the middle, though 
dorsally and ventrally it narrows to a thin mesentery. It 
also contains numerous nuclei instead of the small number 
that were scattered at intervals up to this time. In the 
mesoderm the muscles are now well advanced, and appear as 
three pinnate groups (mus.) in each quadrant, especially in 
the tail region, where they are relatively larger (fig. 32). 
Over each group lies a nucleus, which is either that of the 
muscle-cell or belongs to avery slender peritoneal epithelium, 
but from sections of the adult I am inclined to doubt the 
existence of the latter. 

As the nature of the muscles and ccelomic wall is of great 
importance in fixing the systematic position of the Cheeto- 
enatha, it will be well at this point to review the whole 
evidence. It is seen in figs. 13, 14 (Pl. 19) that the nuclei of 
the somatic mesoderm, although collected into four groups at 
the outer sides of the mesodermal strands, do not lie actually 
in contact with the junction between the mesoderm and ecto- 
derm. ‘The nuclei in each group le on the inner face of a 
mass of tissue, which stains more deeply than ordinary cell 
protoplasm, and fig. 30 (Pl. 21) shows that this tissue is the 
rudiment of the muscles. It is found, however, that it never 
contains nuclei, while the nuclei which appear on the surface 
of each mass seem to belong to the cells which compose it, 
lying at their inner ends. As the somatic mesoderm is 
traced through the later stages the same condition is found ; 
in the first and second day after hatching the only difference 
is that the tissue beneath the nuclei is enlarging and 
becoming fibrous, while on the third and subsequent days the 
muscles, with their pinnate fibres, are fully formed. 

When a section of an adult Sagitta is examined (cf. fig. 36) 
the muscles have increased largely in number, but otherwise 
no alteration is found; the nuclei still appear at the inner 
ends of the pinnate groups of fibres. In some cases, e. @. in 
the series from which figs. 36, 37 were drawn, the muscles 


ON THE DEVELOPMENT OF SAGITTA. 379 


are often alternately large and small; the large ones run up 
into cells with distinct cell boundaries on each side, though 
continuous with the muscles below, while the smaller ones 
end abruptly. The conclusion from these facts is that in 
Sagitta the muscle-cells retain their protoplasmic character, 
with the nucleus at the inner surface next to the ccelom, and 
that the muscular part of the cell has a greater length than 
the protoplasmic portion, so that in transverse sections not 
all the muscles appear to be continuous with the protoplasmic 
part. Since the ccelom is bounded by muscle-cells there is no 
separate peritoneal epithelmm, and this conclusion is sup- 
ported by the way in which, in adult Sagitta, the whole 
mesoderm passes into an epithelium one cell thick at the 
lateral lines, and by the fact that in stained larve nuclei are 
seen in rows along the muscles, but not between them. 

Returning to the structure of the third-day larva, it is 
found that the rest of the mesodermal strands is much looser, 
and a ccelomic cavity is beimg formed, but is still crossed 
frequently by strands of protoplasm. In the ganglion no 
important alteration has occurred, but the nuclei seem to have 
increased in number, and the fibrous portion is becoming 
more conspicuous between the two groups of nuclei. | 

It is during the third day that the genital cells cross to the 
body-wall and that the transverse septum is formed, but 
although many series of sections were cut in the hope of 
throwing light on this question, yet none proved very success- 
ful. Fig. 32 represents a section through the male genital 
cells at the beginning of their movement, and shows how they 
are connected by protoplasmic strands with the mesoderm of 
the body-wall. In the lower half of the figure a nucleus 
(nue. g.e.) is seen close to the genital cell, but outside it; and 
this is almost certainly the nucleus of one of the investing 
cells which probably give rise to the septum. 

In fig. 33 it is seen that the alimentary canal is now quite 
thick, and the structure of the ventral ganglion is well shown ; 
the nuclear aggregations are sharply distinguished from the 
epidermis, and connected ventrally by a bridge of “ Punkt- 


380 L. DONCASTER. 


substanz.’’ It also shows two kinds of nuclei in the ganglion; 
a few large ones, faintly stained (gang. nuc. 2), and numerous 
smaller ones (gang. nuc.) which stain more strongly. 

After the third and fourth days the changes are gradual 
and less important. The ccelom becomes completely clear, 
and it also increases in size, while the body-wall becomes 
thinner. The alimentary canal gets larger and more rounded, 
so that the distinction between it and the dorsal and ventral 
mesenteries becomes more obvious ; while in the tail, on the 
other hand, the longitudinal septum becomes exceedingly 
thin. The genital cells become tightly pressed against the 
body-wall, and at the same time appear to be somewhat 
reduced in size. The fins enlarge, and can be easily seen to 
consist of a fold of the lateral epidermis enclosing a cuticular 
supporting plate. In the genital region a mass of nuclei is 
seen on the fin, which are those of the tactile sense-organ 
present in that region (¢.0.). The tactile organ mentioned 
above, lying between the lateral and tail fins on each side, 
appears as a mass of nuclei in the body-wall just at the 
posterior end of the lateral fin; the wall is so thin in this 
region that it is impossible to determine whether they belong 
to the ectoderm or mesoderm. ‘I'he nuclei in these sense- 
organs, like those of the ganglia, seem to stain more deeply 
than those of the rest of the body. Pl. 21, figs. 34, 35, 
illustrate the above description, and are taken from sections 
of a ninth-day larva; but after the sixth day hardly any 
change seems to take place in the trunk and tail except the 
further development of the alimentary canal and the forma- 
tion of the anus. 


VII. Post-LARVAL DEVELOPMENT. 


In the last two sections the development has been followed 
from the time of hatching up to the end of the second week, 
beyond which period no young were ever kept alive. ‘The 
young at that stage has the essential structure of the adult, 


ON THE DEVELOPMENT OF SAGITTA. 381 


but nevertheless differs from it in several important par- 
ticulars. With regard to many of these, however, the 
assumption of the adult form is very gradual, and takes 
place rather by a process of differential growth than by any 
real alteration of structure. Such changes as these are the 
reduction in relative size of the gangla, more especially of 
the ventral ganglion, which in a Sagitta a week old is nearly 
half the length of the animal, and also the formation of the 
anterior pair of fins, which are absent in the larva. These 
processes of development are comparatively unimportant, 
and require no further description, but there are a few points 
in which the young reared from the egg differ very materially 
from the adult, and these must be followed further. They 
concern chiefly the reproductive organs and their ducts, and 
since these have been used as characters for determining the 
systematic position of the group it is important that their 
mode of origin should be accurately known. ‘he oldest 
young which I was able to rear still retained the four 
primitive reproductive cells undivided, and none showed the 
least trace of any reproductive duct either during life or in 
section, and therefore to follow the development of these 
organs it was necessary to use young specimens taken in the 
tow-net, in which Sagitta of almost all stages were usually 
abundant. 

The youngest individuals which I obtained differed con- 
siderably from those reared from the egg, and were obviously 
considerably older ; they were fully twice as long, the anterior 
paired fins were present, the ganglion much reduced, and the 
parenchymatous tissue of the neck and behind the ganglion 
had disappeared, and in every case the genital cells had 
already divided to form groups of smaller cells. These 
groups of cells occupy the positions of the primitive genital 
cells, i.e. there is a pair of groups on each side, one in front 
of, and the other behind the transverse septum; in section 
they appear lens-shaped, and lhe closely pressed against the 
body-wall on the level of the lateral fin. In each group all 
the cells seem alike; they are moderately large, with pro- 


382 L. DONCASTER. 


minent nuclei, and at this stage also there is no vestige of 
genital ducts, either male or female. The period in which 
this condition lasts seems to vary in different species; in 
small species, such as 8S. minima, very small specimens are 
found further advanced; while in the larger ones, e.g. 
S. lyra, I have found individuals more than two centimetres 
long with the genital rudiments in this condition, and with 
no trace of ducts visible either alive or in sections. 

As the further development of the male organs differs 
greatly from that of the female, it will be convenient to deal 
with them separately, and since the male portion of the 
animal is the first to become mature, that will be described 
first. 

Grassi (4) has described the adult generative organs with 
great detail and accuracy, and made a study of the spermato- 
genesis ; he points out how when the testis reaches a certain 
size it gives off into the tail-ccelom groups of “ sperm- 
eumuli,” which there complete their development, and men- 
tions the fact that the ccelom of the tail becomes divided by 
secondary septa, the position of which in the large species 
differs from that of the smaller. He also describes and 
figures the structure of the vasa deferentia and vesiculz 
seminales of the adult, and it is therefore not necessary here 
to go over these points. As the animal develops, the group 
of cells mentioned above, which represents the testis, 
increases in size and grows backward along the body-wall, 
and at the same time projects somewhat into the cavity, 
becoming cylindrical in shape instead of lenticular, 

In 8. bipunctata at this stage the secondary longitudinal 
septa arise, one on each side, dividing each half of the tail 
cavity into two compartments; but since they occur only in 
the middle region, and die out both in front and behind, the 
cavity remains continuous at each end. ‘The septa are 
exceedingly slender, and contain here and there faintly 
staining nuclei; I have not been able to discover their mode 
of origin. When the testis has increased considerably in 
size it begins to give off masses of cells which fall into the 


ON THK DEVELOPMENT OF SAGITTA. 383 


body-cavity and there develop further, as described by 
Grassi, and at about the same time the male genital ducts 
begin to arise. In their first origin the latter appear as a 
thickening of the ectoderm in the lateral line in the space 
between the paired fin and the tail fin, just at the hind end 
of the former. The thickening increases and soon splits into 
two well-defined layers, and then the upper layer separates 
from the lower like a blister, leaving a space between them 
(Pl. 21, fig. 36, v. s.). Traced forward this space is found to 
narrow into a very fine canal, enclosed in a few well-defined 
cells and lying between the ectoderm and the mesoderm, but 
from its continuity with the larger space behind there can be 
no doubt that the whole is ectodermal in origin. This canal 
runs forward for a considerable distance, and its walls then 
join the lining of the celom into which the canal opens 
(fig. 37, v. d.). As the animal nears maturity the larger 
space behind increases in size, and at its front end a longi- 
tudinal groove appears in the outer wall, along which the 
opening to the exterior is formed. The whole space forms 
the seminal vesicle of the adult, and the chitinous “ calotte ” 
is formed only at maturity. 

In the female organs the course of development is different. 
After the stage is reached in which the rudiment of the ovary 
is a mass of similar cells, this mass grows forward and 
becomes more cylindrical in shape. A differentiation of the 
cells then begins to take place, those next to the body-wall 
becoming a sort of epithelium, while those towards the 
body-cavity become the primitive ova (fig. 38). The latter 
have two forms; those in the centre of the ovary become 
columnar, with elongated nuclei, and give rise to the genital 
epithelium, the cells of which (or some of them) later enlarge 
and become ova, as described by Grassi. Between this layer 
and the inner edge of the ovary, already in quite immature 
individuals larger rounded cells ave found, which will be the 
first ova to mature. Only parts of two or three of the latter 
appear in fig. 38 (0.) because the section is taken close to the 
base of the ovary, and they occur chiefly near its free end. 


384 L. DONCASTER. 


Between the epithelial layer which lies against the body-wall 
and the layer of columnar germinal cells there appears at 
this stage, at the base of the ovary, a mass of loose tissue 
with round nuclei (od. c.). A rather later stage shows these 
cells collecting themselves into a double layer, forming the 
lining epithelium of the oviduct, which, however, up to this 
point contains no lumen. The oviduct shortly before maturity 
has two walls,—an inner of cells with round nuclei which as 
yet enclose no cavity, and an outer which on the inner side 
forms the germinal epithelium (germ. ep.), while on the 
outside it consists of cubical cells constituting the limiting 
epithelium above described (fig. 39). The oviduct thus hes 
wholly within the ovary, except at the point at which it opens 
to the exterior, where it becomes continuous with the epi- 
dermis. There is, however, never any trace of invagination 
of the epidermis during the growth of the ovary, but as the 
latter grows forward the loose tissue above described (od. c.) 
grows with it near its outer border, and when the ova become 
nearly mature gives rise to the oviduct which runs along the 
whole length. ‘The development of the oviduct is very rapid 
and takes place just before maturity, and from the facts 
described it seems certain that it is developed from the ovary 
itself, and not by an invagination of the body-wall as was 
suggested by Hertwig, for no trace of such an invagination 
has ever been seen. 


VIII. Summary anp ConcLusIoNs. 


The development of the genus Sagitta may be summarised 
in its most important facts as follows: 

The cleavage is complete and equal, giving rise to a spherical 
blastula of apparently similar cells, the nuclei of which lie at 
their outer ends, and from this a gastrula arises by invagina- 
tion. When the gastrula is well formed two cells are separated 
off from its inner layer at the opposite pole from the blasto- 
pore; these soon divide into four, which constitute the prmitive 
genital cells, remaining unaltered until a late stage of de- 


ON THE DEVELOPMENT OF SAGITTA, 385 


velopment, and ultimately giving rise to the ovaries and 
testes. Soon after their appearance a pair of folds arise at 
the anterior pole, and, growing backward, divide the archen- 
teric cavity into three divisions, of which the middle one 
forms the alimentary canal, and the lateral ones the ccelomic 
cavities. 

The folds push before them the genital cells for some 
distance, and the latter ultimately come to rest in the 
coelomic cavities, two in each side. At about this stage the 
front ends of the ccelomic spaces become separated off as 
head-cavities, and at the same time the ectoderm of the 
anterior end is invaginated, and meeting and fusing with the 
endoderm produces the mouth. The blastopore comes to lie 
a little in front of the posterior end before it closes, and by 
the lengthening of the embryo in the shell and the general 
compression caused thereby, all cavities become obliterated. 

The ventral ectoderm of the body and that above the 
mouth become thickened, and many of the nuclei of the 
former sink in and become aggregated in two ventro-lateral 
masses against the mesoderm, forming the origin of the 
ventral and cerebral ganglia respectively, while at the same 
time most of the nuclei of the mesoderm form four aggrega- 
tions, a dorsal and a ventral in each section of the mesoderm, 
extending from the neck to the tail. The embryo assumes a 
pronounced ventral curvature in the shell, from which it 
escapes usually not more than two days after the eggs are 
laid. 

‘At hatching the larva is not more than 1 mm. in length 
and is rod-like in shape, tapering somewhat from head to 
tail.. There is already the rudiment of the lateral fins in the 
tail region, and some tactile organs are present on the 
epidermis. It is solid and contains no cavities; behind the 
head there is a single layer of epidermis, which ventrally along 
half the body is much thickened, and constitutes the rudiment 
of the ventral ganglion, and behind this it is laterally thick- 
ened to form the beginning of the fin-fold. The mesoderm 
consists of two solid strands from neck to tail, the nuclei of 

vou. 46, PART 2,—NEW SERIES, AA 


386 L. DONCASTER. 


which are aggregated into dorsal and ventral bands on each 
side, and in these regions some at least of the cells are 
already elongated and form the rudiments of the muscles. 
The genital cells lie in the mesoderm against its inner wall, 
at a point about halfway between head and tail. ‘The endo- 
derm forms a thin septum stretching from neck to tail, and 
separating the two halves of the mesoderm. In the head the 
ectoderm is thickened dorsally and anteriorly to form the 
brain, and laterally to give rise to the hood, while ventrally 
it is invaginated and forms the solid rudiment of the mouth 
and pharynx. The two masses of cephalic mesoderm lie at 
the sides of the latter, and meet one another dorsally. 

As development proceeds the fins increase in size, and the 
paired ones at the sides become distinct from that of the tail ; 
the ganglia become more sharply defined, the endoderm of the 
body increases in thickness as far as the genital cells, and the 
muscle-cells become definite muscles with transverse striation, 
and appearing pinnate in section. The epidermal cells of the 
neck and those just behind the ventral ganglion assume a 
vacuolated and parenchymatous nature. In the head the 
hood is formed by splitting, the mouth cavity opens, the 
vestibular ganglia are formed from ectodermal cells at its 
sides, and the eyes and hooks appear. The greater part of 
the cephalic mesoderm cells become muscles, and then a 
cavity, probably truly coelomic, appears between them and 
the pharynx. At the same time ccelomic cavities appear in 
the trunk, and the genital cells migrate across from the 
splanchnic to the somatic layer, forming as they go the 
posterior transverse septum, which is probably produced 
from the mesodermal envelopes of the genital cells. 

The coelomic cavities increase in size and their boundaries 
become definite, but it is not absolutely certain whether they 
are enclosed by a definite epithelium, or by the cells, the 
outer ends of which form the muscle-fibres; the latter view 
seems, however, more probable, since the nuclei bounding the 
coelom occur always in direct relation with the pinnate groups 
of muscle-fibres. The alimentary canal acquires a ciliated 


ON THE DEVELOPMENYT OF SAGITTA. 387 


lumen and becomes definitely supported by dorsal and ventral 
mesenteries, and the anus is formed just in front of the pos- 
terior transverse septum. 

The remaining changes concern chiefly the reproductive 
organs; the genital cells divide to form four groups of cells, 
and these grow into cylindrical masses, and form the young 
ovaries and testes. ‘I'he latter give off groups of cells into 
the tail cavities, and from them the sperm-cells arise, while 
the male genital ducts and seminal vesicles arise by thicken- 
ing and splitting of the lateral ectoderm of the tail. 

The ovaries grow forward as a cylindrical mass of cells, and 
as maturity approaches the oviducts appear along the outer 
edges, but completely enclosed in ovarian cells. They seem, 
therefore, to be formed actually as a cavity in the ovary, and 
not by ectodermal invagination as has been supposed. 

It now remains to consider what bearing these facts have 
upon the systematic position of the Chetognatha. There 
can be no doubt that the group is an exceedingly isolated one, 
and the fact that all the members of it are closely related to 
one another, and the absence of any skeleton which could be 
traced back in paleontology, make it peculiarly difficult to 
determine its relationships with certainty. ‘The Chetognatha 
have no close resemblances with any other group, and it is 
frequently found that when judged by different characters 
their nearest relationships appear in quite different directions, 
and in consequence they have been associated by various 
authorities with most of the phyla of the animal kingdom. 

In discussing their position, since their important charac- 
teristics offer so little help, it is necessary to descend to minor 
characters and details, and therefore it is impossible to arrive 
at an absolute certainty, but it is at least possible to discover 
in what directions the probability of relationship lies. The 
Cheetognatha have been associated, on various grounds, with 
many different groups, but it will not be necessary to discuss 
the merits of all. By the majority of authorities on the 
subject they have been placed either with the Annelida or 
with the Nematodes, but by some the threefold division of 


388 L. DONCASTER. 


their body-cavities has been considered as a reason for 
associating them with Balanoglossus, and possibly with the 
Echinoderms and Phoronis. I will discuss first their resem- 
blances with the Annelida. 

The resemblances between the Chetognatha and the 
Annelida were pointed out by Huxley half a century ago, and 
have since been emphasised by Hertwig and others. In 
transverse section there is a considerable likeness between 
Sagitta and an Annelid,—such, for example, as Polygordius. 
In each there is a large body-cavity divided by a longitudinal 
mesentery, which supports the alimentary canal; in each 
there are four groups of longitudinal muscles, which appear 
pinnate in section, and in which the fibres are striped. There 
are no circular muscles, and the epidermis is simple. If the 
section be taken through the ventral ganglion, the latter 
corresponds in position with the nerve-cord of Polygordius, 
and from its greater size in the larva may perhaps be 
regarded as a longitudinal cord which has become aggregated 
into a single ganglion. The section of Polygordius differs 
from that of Sagitta in the possession of oblique septa, 
nephridia, and blood-vessels. 

Further resemblances with the Annelids are found in the 
cerebral ganglion connected with the ventral by circum- 
cesophageal commissures, and in the transverse segmentation. 
It is supposed by the advocates of the Annelid theory that 
Sagitta is an animal of three somites, separated from one 
another by two transverse septa, one in the neck and the 
other near the anus. The number of segments has become 
greatly reduced, but those that remain are closely similar to 
the typical annelid somite, except that the alimentary canal 
is not continued into the last one; since, however, embryology 
shows that the endoderm is continued to the tail, this objec- 
tion has not much weight. Another important difference 
is the absence of nephridia in Sagitta, and it has been 
supposed that they are represented in a modified form by the 
genital ducts. It has been further argued that the hooks of 


ON THE DEVELOPMENT OF SAGITTA. 389 


Sagitta are similar in structure and formation to Chetopod 
sete, which, however, are absent in Polygordius. 

One weak point in the above argument is that it is based 
almost entirely on the anatomy of the adult, and that no 
account is taken of embryology. ‘There can be no doubt, 
ge wever, that the development of Sagitta is very much 
abbreviated, as is indicated by the fact that after the 
coelomic and enteric cavities have been formed they are 
closed by the compression of the whole animal, and only open 
again several days later. Similarly, the reduction of the 
endoderm to a simple septum, the formation of the ventral 
ganglion so greatly out of proportion to its future size, and 
other considerations all point to the abbreviated development. 
This, however, does not dispose of the objection that the 
development of the Chetognatha and Annelida is of a 
fundamentally different type. The production of the meso- 
derm in the one case by archenteric diverticula, in the other 
from pole-cells, shows a wide distinction, though it is pos- 
sible that the difference is less fundamental than at first 
appears. However, apart from the earliest stages, the 
development of Sagitta differs very widely from that of the 
typical Annelid. There is never a stage which resembles the 
Trochophore with its large segmentation cavity, its sense- 
organ, cilia, and head-kidneys, all of which are conspicuously 
absent in the young Sagitta. But in addition to the differ- 
ences in embryology, there is grave reason to believe that 
many of the apparent similarities between the Chetognatha 
and the Annelida are superficial, and do not indicate true 
affinity. The theory is based on the supposed homology of 
the three segments of Sagitta with Annelid somites, but this 
homology is by no means certain. Metameric segmentation 
has arisen in the animal kingdom in very various ways and 
independently in different groups, and it is very probable 
that segmentation apparently similar in character may arise 
in a variety of ways. In examining the embryological 
origin of the segments of Sagitta, the first thing that is 
noticed is that they arise at very different times; the head 


390 L. DONCASTER. 


cavities and the mesoderm associated with them are formed 
very early in embryonic life, while the division of the rest of 
the body does not take place until the animal has, in most 
important respects, the structure of the adult. The separa- 
tion between head and body is also much greater than 
between the body and tail, for while the latter are at first 
identical in essential structure, and in the adult differ only in 
the absence of the alimentary canal in the tail, yet the head 
is from the first widely different from the body, and becomes 
increasingly so as development proceeds. 

A further point of importance is the mode of origin of the 
two transverse septa, and here again there is no agreement. 
The anterior is produced by the meeting and fusion of the 
somatic and splanchnic mesoblast at a time when differentia- 
tion of tissues has not begun and the mesoblast is still 
continuous with the hypoblast ; while the posterior appears in 
close connection with the genital cells, and is not formed by 
the whole thickness of the mesoderm, but probably only by 
the cellular envelopes of the genital cells, and certainly by 
the splanchnic layer exclusively. 

The most probable origin of the posterior septum is hinted 
at by Grassi when he suggests that its object may be to 
separate the male genital organs from the female, and this 
view certainly gains support from its formation at the time 
of the passage of the genital cells across the body-cavity. 
From its mode and time of origin it seems reasonable to 
regard the posterior transverse septum as essentially part of 
the reproductive organs, and not closely connected with the 
general plan of the anatomy; but if this be the case it is 
necessary to account for the presence of the anterior septum. 
The most probable reason for its existence would be the 
early separation of the anterior part of the mesoderm for the 
production of the important head muscles, and since at the 
time of its formation in the embryo the body-cavity is widely 
open, this can only be done by folding off the front end as a 
separate space, enclosed in its own mesodermal walls. From 
the time of its separation the development of the head 


ON THE DEVELOPMENT OF SAGITYA. 391 


mesoderm differs greatly from that of the trunk, and in an 
animal which develops so rapidly as Sagitta it seems natural 
that the separation should be early and complete. 

If this theory be correct it destroys the chief ground for 
associating the Cheetognatha with the Annelida, for the three 
segments of the body would not be homologous with one 
another, and therefore cannot be compared with Annelid 
metameres. Such a separation seems justified by a closer 
examination of the other points of resemblance, the chief of 
which are the similarity of the body-cavity and muscles and 
the comparison of the reproductive ducts with nephridia. 
The longitudinal muscles of Sagitta certainly have a close 
similarity with those of Annelids both in structure and 
arrangement, but from the arrangement of the nuclei which 
border the ccelom in connection with the muscles it seems 
probable that there is in reality no ccelomic epithelium in the 
Cheetognatha, but that the body-cavity is bordered by a 
single layer of cells, the ends of which next to the cavity 
remain protoplasmic, while the deeper ends develop into 
muscle-fibres. Such an arrangement exists in the Nematoda, 
but it differs greatly from the Annelid plan of a definite 
coelomic epithelium overlying the longitudinal muscles. 
Further work on the histology of the group is required 
before this point can be finally settled. 

The relation of the genital ducts to nephridia can be dealt 
with more certainly, and this also gives evidence against the 
Annelid theory. In the first place, there is no trace of 
genital ducts in the early stages, and even in well-grown 
young they are quite absent until maturity approaches. In 
this point they differ greatly from nephridia, which appear 
at an early stage. In their mode of development they differ 
no less, for, as was shown above, the male ducts arise as 
epidermal thickenings, and are formed as splits in the 
ectoderm for the greater part, if not the whole, of their 
length, while true nephridia are always chiefly mesodermal 
in origin. 


The development of the oviducts is less certain, but lend 


392 L. DONCASTER. 


no support to the view that they are modified nephridia, for 
they are almost certainly formed chiefly if not wholly from 
the actual cells of the ovary, while the external aperture is, 
like the vas deferens, ectodermal. Although, therefore, the 
Annelid theory appears at first sight very probable, yet a 
further examination of the facts on which it is based shows 
that it is open to grave doubt, and that the resemblance may 
be very probably due to convergence of type. 

Of the remaining chief views of the affinities of the 
Cheetognatha, those, namely, which place them with the 
Nematoda or near the Enteropneusta, the latter is open to 
the objections which affect the Annelid theory; for if the 
segments of Sagitta are not homologous, but have arisen in 
different ways and at different times, no ground remains for 
a theory of their relationship which depends on the view that 
the three segments correspond with the three divisions of the 
body-cavity in Balanoglossus and its allies. The only view, 
therefore, that remains to be discussed is that of affinity with 
the Nematodes. 

The Chetognatha have been associated with the Nematoda 
chiefly in consequence of the very similar arrangement of the 
muscles in four longitudinal groups, and this comparison 
would be much strengthened if the muscles were shown to be 
certainly parts of the cells which line the body-cavity. 

The following comparisons are more doubtful, but it may 
be noticed that the vas deferens of Sagitta bears considerable 
similarity to the excretory duct of a Nematode, for each is 
ectodermal and lies in the lateral area between the dorsal and 
ventral muscles, and in rare cases in Nematodes (e. g. 
Lecanocephalus) the duct is said to open into the body- 
cavity. The oviduct of Sagitta may be compared with that 
of a Nematode in being continuous with the ovary and 
formed from the same rudiment. The development of 
Nematodes is unfortunately imperfectly known, but it is of 
interest that the genital organs arise, as in Sagitta, by the 
very early separation of two large cells, which later multiply 
and form the sexual glands and their ducts, and in the latter 


ON THE DEVELOPMENT OF SAGITTA. 393 


the innermost cells arrange themselves to form an epi- 
thelium. 

The differences between the two groups are, however, very 
wide ; in the first place, the body-cavity of the Nematoda is 
not certainly coelomic, and the intestine is not supported by 
mesenteries, nor are there transverse septa. The nervous 
systems of the two phyla are also very distinct, and when all 
the points of resemblance and difference are considered 
together it is evident that there can be no near relationship 
between them. The conclusion, therefore, must be that if 
the nearest connections of the Chetognatha are with the 
Nematodes, yet the two groups have diverged very widely 
owing to a difference of habit: the Nematoda were perhaps 
primitive Coelomata which have become degenerate through 
parasitism ; while the Chetognatha, if they branched off from 
the same original stock, have become fundamentally modified 
for pelagic existence. In the present state of our knowledge 
it seems safest to regard the Cheetognatha as descended from 
a primitive coelomate stock, from which the Annelida have 
arisen on the one hand; while, on the other, the Nematoda 
probably branched off, but lost many of their original 
characters owing to their parasitic habit. 


APPENDIX. 
On tHE ANATOMY oF SAGITTA MINIMA. 


Grassi, in his general account of the anatomy of the group, 
states that this species has irregular septa between the 
alimentary canal and the body-wall, but admits that he has 
not been able to throw any light on them by sections. 

These septa are very conspicuous in the adult S. minima, 
and form the readiest means of identifying it at a glance. 
‘hey occur especially in the anterior part of the body, but at 
intervals through the whole trunk region; they are quite 


394 L. DONCASTER. 


irregularly arranged, and not at all accurately transverse 
(P]. 21, fig. 40, sep. sp.). In the places where the alimentary 
canal is not in contact with the body-wall the septa never 
cross the space between them; they only touch the body- 
wall where the gut is in contact with it. In very young 
minima they are not present, and after studying sections of 
various ages I have come to the conclusion that they are not 
septa, properly speaking, at all. The alimentary canal in 
this species differs from that of others in being very wide, 
almost entirely obliterating the coelom; but this is less 
conspicuously the case in the young. If a series of sections 
of various stages is examined it is seen that the widening of 
the alimentary canal is due to an enlargement of its walls, 
while the lumen remains narrow as in other species. The 
cells composing the walls become exceedingly large, and 
ultimately lose most of their protoplasm; this change takes 
place soonest in the lateral region, while dorsally and ven- 
trally the cells remain less modified. At length, when the 
animal is sexually mature, and its life is probably nearly at 
an end, the alimentary canal consists of the inner wall 
surrounding the lumen, and an outer wall, which is largely in 
contact with the wall of the body ; and between these two is 
a cavity crossed at intervals by sheets of tissue, which are 
the remains of the cell walls of the endodermic cells. These 
sheets are irregular in their distribution, and seem to corre- 
spond in every way with the “septa” seen in the living 
animal (figs. 41, 42). At the posterior end of the ali- 
mentary canal, between the ovaries, this vacuolation and 
degeneration does not take place, so by following forward a 
series of sections of an adult 8S. minima the whole course of 
changes can be observed. ‘lo test this further, some living 
S.minima were put in a watch-elass with sea water in which 
carmine particles were suspended, and as the animals swallow 
water at intervals the carmine was taken into the gut and 
gradually travelled backwards to the rectum, which it 
reached in periods varying from five minutes upwards, and 
was finally expelled at the anus. During the whole of its 


ON THE DEVELOPMENT OF SAGITTA. 395 


passage it always kept exactly to the middle line, never 
passing laterally into the cavity at the sides, and the drop of 
water containing the carmine seemed to push apart the walls 
enclosing the lumen of the gut, which was otherwise almost 
or entirely obliterated by the walls coming into contact. 

The septa’? mentioned by Grassi seem, therefore, to be 
the remains of the endodermal cells, and are not any way to 
be regarded as mesodermal septa supporting the alimentary 
canal. 


PAPERS REFERRED TO. 


1. Kowatnvsky.— Embryologische Studien aun Wiirmern und Arthro- 
poden,” ‘ Mém. Acad. Pétersbourg,’ 7th series, tom. xvi, 1871. 

2. Birscuii.—* Zur Entwicklungsgeschichte der Sagitta,” ‘ Zeitschr. wiss. 
Zool.,’ Bd. xxii, 1873. 

3. Hertwic, O.— Die Chaetognathen,” ‘Jenaische Zeitschrift,’ Bd. xiv, 
1880. 

4, Grassi.—“ I Chetognati,” ‘Fauna und Flora des Golfes von Neapel, 
1883. 

5. Jourpain.—“ Sur l’embryogénie de Sagitta,” ‘Comptes rend. Acad. 
Sci.,’ Paris, 1892. 


EXPLANATION OF PLATES 19—21, 


Illustrating Mr. L. Doncaster’s paper “ On the Development 
of Sagitta.” 


The figures of sections were drawn with the help of a camera lucida, using 
a +.inch or a 34-inch objective. The remaining figures are mostly drawn 
from the living animal, and the camera lucida was not used. The magnifica- 
{ions given are approximate. 


396 L. DONCASTER. 


EXPLANATION OF THE LETTERING OF THE FIGUREs. 


a. Archenteron, al. Alimentary canal. al. div. Diverticula of alimentary 
canalinS.minima. b. ¢.1, 2, 3. Head, trunk, and tailccelom. b/. Blastopore. 
c. p. Caelom pouch. e. Eye. ect. Ectoderm. end. sep. Endodermic septum. 
epi. Hpidermis. f. p., f. €. Posterior and tail fin. jf. sk. Fin skeleton. Fs 7: 
Fin ray. f. a. Archenteric fold. gang. cb., gang. l., gang. v. Cerebral, lateral, 
and ventral ganglia. gang. nuc. Nuclei of ganglion. gang. nuc. 2. Large 
nuclei of ganglion. gang. fib. Fibrous layer (‘‘Punktsubstanz”). gen. c. 
Genital cells. gen. nuc. Nucleus of genital cell. germ. ep. Germinal 
epithelium. hk. Head. hd. Hood. hd. cav. Hood-cavity. h. mus. Head 
muscles. hks. Hooks. mes. 1, 2. Mesoderm of head, trunk. mes. so. 
Somatic mesoderm. mo. Mouth. mus. Muscles. mus. c. Muscle cell. mus, 
nuc. Nucleus of muscle cell. mnuc. g. e. Nucleus of genital envelope. o. 
Ovum. o.d.c. Lining cells of oviduct. ol. Olfactory organs. ov, Ovary. 
ph. Pharyux. p. t. Parenchymatous tissue. sep. (. Longitudinal septum. 
sep. tr. Transverse septum. — sep. sp. spurious septum (in 8S. minima). sh. 
Hgg-shell. . 0. Tactile organ. v.d. Vas deferens. v. s. Seminal vesicle. 


PLATE 19. 


Fic. 1.—S. bipunctata. Gastrula with genital cells. x 220. 

Fic. 2.—S. bipunctata. Formation of archenteric folds. x 220. 

Fic. 3.—S. bipunctata. Embryo with head-cavities recently formed. 
x 250. 

Fic. 4.—S. enflata. Formation of head-cavities. 

Fic. 5.—Sagitta, sp. Optical transverse section of embryo through 
recently formed head cavities. Xx 250. 

Fic. 6.—S. bipunctata. Embryo shortly before hatching. x 220. 

Fic. 7.—S. bipunetata. Longitudinal section of embryo during formation 
of the archenteric folds. x 875. 

Fic. 8.—S. bipunctata. Embryo at stage of Fig. 3. 
through the middle of the body. x 500. 


Transverse section 


Fics. 9—12.—Four transverse sections through an embryo at about the 
stage of Fig. 3. Fixed with osmic acid. x 600. 
Fig. 9.—Through the head. 
Fig, 10.—Through the ventral ganglion. 
Vig. 11.—Through the genital cells. 
Fig. 12.—Through the tail region. The archenteric folds have not 
yet grown back to this level. 


ON THE DEVELOPMENT OF SAGITTA. 397 


Fie. 13.—S, bipunctata. Longitudinal horizontal section through a late 
curled embryo, cutting the head and trunk. x 375. 

Fic. 14.—Section through the same embryo, showing the genital cells 
below and the tail above. The embryo is cut twice owing to curvature. 
x 375. 

Fie. 14 a,—Diagram to show the planes of section of Figs. 18 and 14. 

Fie. 15.—S. bipunetata. Section of embryo at about the same stage as 
Figs. 18, 14, showing the genital cells and their envelopes. x 900. 


PLATE 20. 


Fic, 16.—S. bipunctata. Larva just after hatching. x 200. 
Fie. 17.—S. enflata. First day. Genital cells. x 500. 
Fie, 18.—S. enflata. Fourth day. Ventral view. ‘Ihe eyes are seen 
through the head. x 200. 
Fie. 19.—S. enflata. Fourth day. Head and neck, lateral view showing 
parenchyma. x 600. 
Fie. 20.—S. enflata. Third day. Genital region in stained specimen, 
showing envelopes of genital cells. x 800. 
Fie. 21.—S. enflata. Fourth day (a late specimen). Showing migration 
of genital cells. x 600. 
Fie. 22.—S. enflata. Second day. Transverse section of head. x 
1000. 
Fic. 23.—S. bipunctata. Third day. ‘Transverse section of head. 
Only half of the section is represented. x 900. 
Fies. 24—26.—S. enflata. Three sections through the head of a sixth- 
day larva. x 375. 
Fig. 24.—Through the mouth. 
Fig. 25.—Through the eyes and beginning of the pharynx. 
Fig. 26.—Through the crescent-shaped transverse muscle in the 
posterior part of the head. 
Fires. 27—29.—Three sections through the head and neck of a ninth- 
day enflata larva. x 375. 
Fig. 27.—Through the posterior region of the head. 
Fig. 28.—Through the beginning of the neck. 
Fig. 29.—Through the neck, 


398 L. DONCASTER. 


PLATE 21. 

Fre. 30.—Part of longitudinal horizontal section through embryo shortly 
before hatching, to show development of muscles. x 1000. 

Fie. 51.—S. enflata. Second day. Transverse section through region 
of ventral ganglion. x 750. 

Fie. 32.—S. enflata. Third day. Transverse section through genital 
cells. x 900. 

Fic. 83.—8. enflata. Fourth day. Transverse section through ventral 
ganglion. x 750. 

Fic. 34.—S. enflata. Ninth day (same specimen as Figs. 27—29). 
Transverse section through ventral ganglion. x 375. 

Fie. 35.—Same larva; section through female genital cell. x 375. 

Fie. 36.—Transverse section through seminal vesicle of a young 8. 
enflata. x 550. 

Fie. 37.—Vas deferens of same specimen. X 550. 

Fig. 38.—Transverse section of young ovary in same specimen. X 2U0. 

Fig. 39.—Transverse section of ovary and developing oviduct in S. 
minima approaching maturity. x 550. 

Fie. 40.—Anterior part of adult S. minima to show spurious septa. 

Fic. 41.—Transverse section of adult S. minima near posterior end of 
trunk. x 200. 

Fie. 42.—Section of same near middle of trunk region. x 200. 


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. 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 Supprumenr 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 upa 
Volume. 


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


TO ‘CORRESPONDENTS. 


Authors of original papers s 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 eatra copies of their 
communications at their own expense can have them by applying 
to the Printers, 

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

For every four pages or less— 


25 copies : : ‘ : 5/- 
Bee : : : ; 6/- 
1OOuers: ' , : 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. 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., F.R.S. 
20; 


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, &e. 


Works published by the Association 


include the following :—‘ A ‘Treatise on the Common Sole,’ J.T. Cunningham, M.A., 
Ato, 25/-. ‘The Natural History of the Marketable Marine Fishes of the British 
Islands, J. IT. 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 1 Oper annum. 
LIFE MEMBERS . : : ; . 15 15 O Composition Fee. 
FOUNDERS . ; LOO SOO y > 


GOVERNORS (Life Members of Council) 500 0 O 


” 33 


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 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. 


2 4 
New Series, No. 183 (Vol. 46, Part 3), Price 10s. 


"\S'X) DECEMBER, 1902. | 
| THE 


QUARTERLY JOURNAL 


OF 


-MICROSCOPICAL SCIENCE. 


EDITED BY 


rl 7 sh 7 
KEK. 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 
SOCIRTY OF SCIENCES, AND OF THE ACADEMY OF THE LINCEI OF ROMF, 
AND OF THE 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 
PUYSICAL 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 5 


Weed WHEDON, MvA.| ERS, 


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


AND 
SYDNEY J. HICKSON, M.A., F.RS., 
BEYER PROFESSOR OF ZOOLOGY IN THE OWENS COLLEGE, MANCHESTER. 


WITH LITHOGRAPHIC PLATES AND ENGRAVINGS ON WOOD. 


© lO) ND ON: 
J. & A. CHURCHILE, 7, GREAT MARLBOROUGH STREET. 
1902. 


{London and Dorking 


Adlard and Son, | 


CONTENTS OF No. 183.—New Series. 


MEMOIRS: 
PAGE 
On a Cestode from Cestracion, By Wixitam A. HasweE.t, M.A., 
D.Sc., F.R.S. (With Plates 22— —24) . : : ; . 399 


The Development of Lepidosiren paradoxa.—Part III. t Develo 
ment of the Skin and its Derivatives. By J. Granam Kerr. 
(With: Plates 85—98) cs) of. ag). Pee ea ee 


The Metamorphosis of Corystes Cassivelaunus (Pennant). By 
Rogert Guryey, B.A.(Oxon.), F.Z.S. (With Plates “39="31)- ea 


Artificial Parthenogenesis and Fertilisation: a Review. By ‘'Homas 
H. Bryce . : : : : : : : Ars 


ON A CESTODE FROM CESTRACION. 399 


On a Cestode from Cestracion. 
By 


William A. Haswell, M.A., D.Sc., F.R.S., 
Professor of Biology, University of Sydney. 


With Plates 22—924. 


General Features. 


Tue Cestode, the results of a study of which are embodied 
in the present paper, occurs, usually in abundance, in the large 
intestine of the Port Jackson shark. It is one of these re- 
markable forms to which attention appears to have been first 
specially directed by P. J. van Beneden (1 and 2), in which 
the proglottides are set free from the posterior end of the 
strobila long before full maturity has been reached, and only 
attain a stage corresponding to that of the “ripe” proglot- 
tides of a Tenia after having pursued an independent 
existence for some considerable time. 

The strobila is actively locomotive, and appears to use the 
suckers more in connection with progression than as organs 
of permanent attachment. It is only 9 or 10 cm. long in the 
preserved condition. There is an elongated neck-region with 
a breadth, in the preserved specimens, of half a millimétre. 
The four sessile bothridia (fig. 1) are somewhat spoon-shaped, 
the anterior end being the narrower. The margin of the 
bothridium is very prominent, finely crenulate, and in the 
livmg condition extremely extensile, so that the shape is 
undergoing constant modification. In preserved specimens 

VoL. 46, pARY 3,—NEW SERIES. BB 


400 WILLIAM A. HASWELL. 


they are about 1 mm. in length. Each bothridium is so 
directed that a line running along the floor of its cavity in 
the direction of the long axis, and prolonged forwards, would 
meet the median axis of the neck at an angle of about 45°. 
The cavity is not divided or reticulated. At its anterior 
narrower end, where its margin is lowest, each bothridium 
bears a small circular accessory sucker. 

The last segment (fig. 2) is 5 mm. long and 2 mm. in 
breadth in the preserved specimens; relatively narrower in 
the extended living condition. 

Separated segments are to be found in abundance along 
with the entire strobile, moving actively through the intes- 
tinal contents. In the course of these movements the shape 
undergoes constant alteration, the phases through which it 
passes being comparable to those of a Ligula in its most active 
condition. ‘The anterior end becomes thrust sharply forwards 
until the “ head” becomes long and narrow and pointed, and 
the “neck” constriction becomes more or less completely 
obliterated. Then suddenly the anterior end becomes drawn 
together and thickened to form a distinct rounded knob, 
constricted off from the rest (fig. 3). The part behind this 
“head” now becomes drawn forwards, the region imme- 
diately following on the head gradually becoming thickened, 
while the head itself becomes gradually retracted until it 
nearly completely disappears, to become again thrust for- 
wards as before. The effect of these movements is clear 
enough. By the thrusting forwards of the narrowed head 
end, the thick matter contained in the intestine is readily 
penetrated, the subsequently formed knob at the anterior end 
then forming a point d’appui, towards which the rest of 
the proglottis becomes drawn forwards. 

These independent proglottides attain a relatively con- 
siderable size, the largest being about 11 mm. in length and 
1:75 mm. in greatest breadth. 

Attention has been recently directed by Liihe (18) to 
isolated proglottides from Acanthias,in which there is a 
distinct mobile “ head” similar to that above described, but 


ON A CESTODE FROM CESTRACTON. 401 


covered with spinules; and a similar case had previously 
been observed by Pintner. 

The early separation of the proglottides in this and other 
species is obviously correlated with the free locomotive habits 
of the strobila. With a much longer train of connected 
proglottides, the posterior loaded with eggs, such movements 
would be rendered difficult or impossible. The spiral valve 
in the intestine of the Elasmobranch renders it possible for 
the separated proglottides, without definite organs of adhesion, 
yet with an adaptation for creeping movement, to remain 
within their host until such time as the uterus has become 
fully charged with eges. 

This Cestode is to be referred to the genus Phylloboth- 
rium of P. J. van Beneden. In the definition given by that 
author! the bothridia are described as notched externally, 
but the notch is not present in one of van Beneden’s own 
species (P. auricula), and cannot be looked upon as of 
generic importance. I propose the name of P. vagans for 
the Cestracion parasite, which appears to be distinct from all 
the species described hitherto.” 

The only species of Phyllobothrium, of the structure of 
which a detailed account has been published, are P. thridax 
and P. Dohrnii. These have both been pretty fully described 
by Zschokke (20, p. 327 et seq.) ; but, as mature segments 
were not met with by that author, many features of import- 
ance, more particularly in the reproductive apparatus, were 
overlooked. 


Integument and Nervous System. 


The cuticle (fig. 4, ew.) is homogeneous and not divided into 
layers. Immediately beneath it are the usual external longi- 
tudinal (e.l.m.) and circular (e.c.on.) layers of muscular 
fibres. The subcuticular cellular layer is much _ better 
developed in the strobila than in the free proglottides, in 

1-1;,p.. 120; and 2, p:.123. 


2 | have not seen the original description of P. gracile, Wedl., from 
Torpedo marmorata, but only the brief definition given by Lonnberg (11). 


402 WILLIAM A. HASWELN. 


which it has undergone a reduction in thickness. A similar 
reduction is observable in the internal longitudinal layer of 
muscular fibres (7.1. m.), which are well developed in all parts 
of the strobila, and very conspicuous in transverse sections 
owing to their highly refracting character, whereas in the 
free proglottides they are barely discernible in transverse 
sections, and in longitudinal appear as a few inconspicuous, 
often degenerate, fibres. 

The nervous system (fig. 3, fig. 4, n.c.) 1s In no way 
remarkable, consisting of the usual head-ganglion in the 
scolex, and the pair of longitudinal nerve-cords with their 
branches and commissures. In the separate proglottides, 
owing to the reduction in the thickness of the subcuticular 
cellular and internal longitudinal muscular layers, the nerve- 
cords come to be situated more superficially than in the 
strobila. They meet anteriorly in the “head,” where there is 
a slight thickening of the nature of a rudimentary ganglion. 

As in many other forms, two of the four longitudinal 
excretory vessels of the anterior region—the dorsal pair— 
become reduced greatly in diameter in the posterior pro- 
elottides. In the last proglottis these open on the exterior 
at the posterior end. In the free proglottides (fig. 3) only 
the ventral pair remain. These are very narrow towards the 
anterior end, while posteriorly they are very wide and very 
sinuous; their external openings are situated near together 
at the posterior extremity. The excretory vessels in general 
have a wall consisting of a thin layer of fibrillated proto- 
plasmic material; but in the scolex and neck region the four 


main vessels have a fairly thick layer of iongitudinal muscular 
fibres. 


Reproductive Organs. 

The reproductive system will be best described first as it 
appears in its fully developed condition in the free pro- 
elottides. 

The testis (fig. 3, te.) consists of numerous rounded lobes 
extending from the neck to behind the genital aperture. 


ON A CESTODE FROM CESTRACION. 403 


They lie in the central or medullary region, and are thus 
situated on a deeper plane than the vitelline glands. They 
average about ‘(06 mm. in diameter. Hach lobe has a fine, 
thin-walled efferent duct; the ducts of neighbouring lobes 
anastomose to form a network. From this network are 
derived larger trunks, which towards the anterior end, and 
near the ventral surface of the proglottis, combine together 
to form a single median vas deferens (s.d.). The latter is 
a closely coiled, widish, thin-walled tube, situated in the 
middle of the region in front of the genital aperture. Its 
wall consists of a reticulated material with superficially placed 
nuclei. No muscular layer was definitely made out, but 
muscular fibres must be present, as in the living condition the 
tube is observed to undergo peristaltic contractions. The 
“prostate” cells described by various authors (see Braun, 5) 
as occurring in certain Cestodes, are not present. This main 
testicular duct is always packed full of sperms, and it plays 
the part of a vesicula seminalis as well as a vas deferens. 
It terminates by passing through the wall of the cirrus sac 
and becoming the ejaculatory duct. The cirrus sac has a 
wall composed of two layers of muscle. Within it, when the 
cirrus is not protruded, lies coiled up a long tube, contmuous 
internally with the vas deferens. This tube (fig. 5) has a 
muscular wall, consisting of an outer thicker layer of longi- 
tudinal fibres and an inner of circular fibres, Internal to this 
is a homogeneous cuticular layer, beset on its inner surface in 
the outer part of the tube with numerous excessively minute 
spinules. Outside the muscular layer is a layer of cells 
similar to the myoblasts of the oviduct and vagina. In the 
space between the wall of the cirrus sac and the enclosed tube 
are to be observed numerous muscular fibres which appear to 
run about in every direction. 

The outer end of the tube is continuous with the outer 
extremity of the cirrus sac, and might be described as 
invaginated within it were it not for the circumstance that its 
inner end is not free, but passes through the wall of the sac 
to become continuous with the vas deferens. 


4.04. WILLIAM A. HASWELL. 


The mode of protrusion of the cirrus is rendered evident on 
an examination of hving animals and of sections of speci- 
mens with the organs in various states. The strong muscular 
wall of the cirrus sac contracts, and the narrow outer end 
with which the invaginated tube is continuous becomes thrust 
out through the genital opening. Further pressure causes 
the tube to become evaginated as a narrow cylindrical process, 
the cirrus, with a double wall, the space between the two 
walls being continuous with the cavity of the cirrus sac. The 
retraction takes places through the agency of the muscular 
fibres that have been above referred to as situated in the 
cavity of the cirrus sac; when the cirrus is protruded these 
are put upon the stretch, and each of them is found to be 
connected internally with one of the myoblasts in the wall of 
the tube, and to run inwards towards the inner part of the 
wall of the cirrus sac. 

The ovary (figs. 3 and 6, ov.), as in many other Cestodes, 
consists of two large lateral portions and a small median 
isthmus connecting them together, the whole, on a dorsal or 
ventral view, resembling a letter H, with the limbs thick and 
near together and the transverse part very short. A trans- 
verse section shows that each lateral portion is itself double, 
consisting of a dorsal and a ventral lamina which coalesce 
internally towards the isthmus. The margins of the lamine 
are divided irregularly into a number of rounded lobes, but 
these divisions are quite superficial, the substance of the 
lamina consisting of a mass of ova with no trace of a tubular 
structure, except that irregular fenestrae occur here and 
there. The ova are somewhat smaller peripherally, largest in 
the neighbourhood of the isthmus. The mature ova are 
‘Ol mm. in diameter; their nuclei, ‘004 mm.; and their 
nucleoli, "(002 mm. Their cytoplasm appears homogeneous 
under the highest powers, binding them together in a small 
quantity of retiform connective tissue. Knclosing the whole 
ovary is a membrane having the appearance of a condensation 
of the parenchyma, but perhaps of muscular character. 

The isthmus, or connecting part, differs widely from the 


ON A CESTODE FROM CESTRACION. 405 


rest, and is to be looked upon rather as the beginning of the 
efferent duct than as part of the ovary proper. It is enclosed 
in a membrane continuous with that which encloses the 
lateral portions. The contained ova, instead of being closely 
ageregated together, are loosely distributed singly or in 
groups (figs. 11 and 12). 

The oviduct begins in a_ well-developed “ swallowing 
apparatus” (figs. 6, 7, 8, 9, 11, 12, 18, 14, 15, 16, sw.), such 
as has been described in various other Cestodes. This lies 
on the ventral side of the isthmus of the ovary and opens 
into its cavity. It is a bell-shaped structure, the wide mouth 
ot which, directed towards the dorsal surface, opens into the 
cavity of the isthmus of the ovary, while at the opposite 
extremity a very much smaller aperture leads into the ovi- 
duct proper. During life this swallowing apparatus was 
observed to perform rhythmical pulsating movements, the 
effect of which must manifestly be to seize the loose ova of 
the isthmus, one by one, and to pass them backwards along 
the oviduct. In sections it is found that the wall of the 
swallowing apparatus 1s continuous with the investment of 
the ovary and with the muscular layer of the wall of the 
oviduct. It has the character of a dense layer of fibres 
(figs. 15 to 16, sw. m.), which, though of extreme fineness, 
must be muscle-fibres. These are for the most part arranged 
circularly around the wall of the organ, but some are radial. 
Surrounding this fibrous layer is a single layer of cells 
(figs. 13 and 14, sw. my.) of irregular shape. Processes pass 
from these into the fibrous layer, and there can be little 
doubt that the majority of these cells are the myoblasts of 
the fibres of the swallowing apparatus. A small number 
(fig. 13) which give off processes both externally and in- 
ternally are probably nerve-cells. 

Through the oviducal opening of the swallowing apparatus 
projects for a short distance a sort of plug perforated by a 
circular aperture. The substance of this plug is continuous 
with the epithelium of the oviduct; but, though it contains 
several nuclei (fig. 16), it does not consist, so far as I have 


406 WILLIAM A. HASWELL. 


been able to ascertain, of definite cells. On its inner surface, 
i.e. the surface turned towards the ovary, it is fimbriated 
(fig. 14), and it is doubtless through the agency of these 
fimbriz that the ova are seized during the movements of the 
apparatus. 

The oviduct (figs. 6 and 7, od.’) runs, at first, nearly 
straight back from the swallowing apparatus, on the ventral 
side of the shell-gland and receptaculum seminis, and is 
joined by the narrow fertilise duct (f. d.) from the latter. 
In this part of its course (fig. 8) it has an epithelium com- 
posed of short prismatic cells. Internal to this is a thin 
cuticle beset on its inner face with numerous slender hairs, 
resembling cilia in appearance, but non-vibratile, which le 
with their apices directed backwards, i.e. away from the 
ovary, their arrangement thus being such as to prevent the 
ova received from the swallowing apparatus from passing 
forwards again towards the ovary. Hxternal to the epithe- 
lium is a muscular layer composed of external longitudinal 
and internal circular fibres. Surrounding this is a layer of 
cells of the same general character as those that surround 
the muscular layer of the swallowing apparatus. These 
appear to correspond to the cells which Zschokke (20) looks 
upon as glandular, and to those which Pintner! regards as 
the formative cells of the swallowing apparatus. In view of 
Blochmann’s* results on the subcuticular muscle, however, 
and Sabussow’s (17) extension of the same view to the re- 
productive ducts, | am more disposed to look upon these also 
as myoblasts. 

A little behind its point of junction with the fertilising 
duct the oviduct bends sharply round towards the dorsal 
side, and is joined by the main vitelline duct at the posterior 
limit of the shell-gland. From this point it runs forwards 
for some distance with a sinuous course on the dorsal side of 
the isthmus of the ovary and of the vagina, and then runs 

1 Sce Braun, 5. 

2 F. Blochmann, “ Ueber freie Nervenendungen und Sinneszellen bei Band- 
wiirmern,” ‘ Biol. Centralbl.,’ xv, 1895. 


ON A CESTODE FROM CESTRACION. 407 


straight forwards as a cylindrical tube with irregular dilata- 
tions. As this part of the oviduct contains fully formed 
eggs, and is something more than a mere passage, it will 
be convenient to designate it ootype, or primary uterus. 
Anteriorly it opens into the secondary uterus by a longi- 
tudinal slit, the extent and position of which vary in different 
specimens, situated on one side of the vagina. 

After it becomes joined by the maim vitelline duct, the 
oviduct changes its structure, the cuticular hairs are lost, 
and there is no epithelium, the wall of the duct now con- 
sisting of cuticle, muscular layer, and layer of myoblasts. 

The uterus (figs. 3, 6, and 18, s. w.) is a cylindrical un- 
divided chamber, extending from the level of the repro- 
ductive aperture to the interspace between the anterior 
portions of the lateral wings of the ovary. It has a lning 
membrane composed of a single layer of cells. It has no 
natural external aperture, but dehisces by the formation of a 
longitudinal slit along nearly the whole, or only a limited 
part of the length of its ventral surface. This dehiscence 
readily takes place when the specimen is manipulated, more 
especially when it is placed in sea-water, when the eggs are 
observed to be suddenly discharged with the appearance of 
a white cloud.’ 

The shell-gland is a compact oval body, ‘18 mm. in length, 
which surrounds the oviduct where the vitelline duct joims it. 


1 Shipley, in his description of the worms collected by Dr. Willey (19), in 
referring to a species of Phylobothrium, states that in the oldest pro- 
glottides the uterus had ruptured “about the centre of the dorsal surface.” 
But there cau be no doubt that the surface on which the dehiscence takes 
place is the ventral, and not the dorsal. This is made perfectly clear in the 
case of the Cestracion species by the relative positions of the various parts of 
the reproductive apparatus—as, for example, the vagina and vas deferens— 
and by the disposition of the longitudinal vessels of the excretory system. It 
may be remarked, however, that in the Australian land Planarian (Geoplana 
Mortoni) Steel has confirmed by observation on the living animal Dendy’s 
description of the rending of the dorsal body-wall on the discharge of the egg- 
apsules (‘ Proc. Linn. Soc. N.S.W.,’ 1900, p. 573, pl. 34, fig. 10, and pl. 41, 
fig. 6). 


408 WILLIAM A. HASWELL. 


Its cells, several hundred in number, are arranged in a radiat- 
ing manner round the oviduct, their narrow inner extremities 
evidently acting as ducts by which the secretion is dis- 
charged. Their nuclei are large, a little less than ‘005 mm. 
in diameter. Between the cells are a number of smaller 
nuclei indicating the presence of a certain amount “of inter- 
cellular tissue. 

The vitelline glands (fig. 3, v.) extend throughout a narrow 
belt of the lateral regions of the body from the neck to the 
posterior end. The lobes are spherical or subspherical in 
shape, and average about ‘03 mm. in diameter. Hach lobe 
has its slender duct, which joins those of neighbouring lobes 
to form larger ducts, and these again combine to form the 
main lateral ducts (fig. 7, v. d.). These converge from both 
sides towards the middle line, running on the ventral side of 
the ovary, and finally unite to give rise to an impaired main 
duct, situated slightly to the right of the middle lne. This 
runs backwards and joins the oviduct as already described. 
Near its termination it is usually distended with yolk, and this 
dilated part (figs. 7, 8, and 9, v. 7.) (‘05 mm. in diameter) 
might be looked upon as a yolk-receptacle. It is followed by 
a constricted part with thickened walls (fig. 8, v. 7. ©.) 
through which the yolk cells can only pass singly to enter 
the oviduct. The yolk matter leaves the lobes of the glands 
in the form of very regular spherical masses ‘012 mm. in 
diameter, each of which contains one, or sometimes two, 
rounded bodies which, as they are capable of being stained, 
though only slightly, are very lable to be mistaken for 
nuclei. These bodies will be further referred to in the 
description of the egg. Meanwhile it is of importance to 
emphasise the fact that they are not nuclei, and that the 
vitelline masses in which they are lodged are not cells.! 

The wall of the vitelline ducts consists of fibrillated proto- 
plasmic material with nuclei at intervals. In the main duct 

1 This is contrary to what is usually stated of Cestodes in general. Braun, 


for example, states: ‘* Die Ansicht Moniez’s dass die Dotterzellen keine 
echten sondern nur Scheinzellen seien entbelirt jeder Begriindung (5, p. 1468). 


ON A CESTODE FROM CESTRACION. 4.09 


the wall is thicker, and contains a large number of super- 
ficially situated nuclei. 

The vagina (figs. 6, 9, and 10, ra.) opens into the shallow 
genital cloaca by a narrow aperture immediately in front of 
the male aperture. The terminal part is somewhat dilated. 
From this point it bends round the sac of the penis as a narrow 
tube, which dilates again to a diameter of about ‘05 mm., 
as it runs straight backwards immediately above (1. e. on 
the dorsal side of) the secondary uterus. When it reaches 
the region of the ovary it again becomes narrower and more 
sinuous. Hventually passing backwards on the dorsal side 
of the isthmus, it becomes somewhat dilated again to form a 
vesicle, the receptaculum seminis (figs. 6, 7, 9, 10, 11, and 
12, r. s.). From the rounded posterior end of this a 
narrow duct, the fertilising duct (f. d.), runs to join the 
oviduct. 

In the posterior part of its extent the vagina has a thickish 
muscular wall consisting of external longitudinal and in- 
ternal circular layers. Internal to this is a cuticle beset 
with exceedingly minute spinules. External to the muscle is 
a layer of cells resembling those cells of the oviduct which I 
have supposed to be myoblasts. Anteriorly the muscular 
layers become reduced, and longitudinal fibres alone are 
present. The fertilising duct resembles the oviduct in 
structure, but the cuticular hairs are absent. In the 
posterior proglottides of the strobila (fig. 2) all parts of the 
reproductive apparatus are represented, though neither the 
male nor the female organs are mature, and there are no 
egos in the uterus. The latter has a comparatively narrow 
lumen surrounded by a thick layer of small cells; its aperture 
of communication with the primary uterus is already 
developed. In more anteriorly situated proglottides the 
uterus is represented by a solid cord of small cells running 
along on the ventral side of the vagina. 


410 WILLIAM A. HASWELL. 


Development. 


In the case of P. Dohrnii, Zschokke (20) states that the 
formation of eggs begins in the posterior proglottides of the 
strobila. In the form now under consideration this is not the 
case, eggs only occurring in well-developed free proglottides. 

The only recorded observations on the development of any 
member of the genus appear to be a few notes on P. thridax 
by Moniez (14, p. 28). I can trace no correspondence what- 
ever between the statements there made and what I have been 
able to observe in the species from Cestracion. 

The primary uterus contains only eggs with unsegmented 
ova. ‘The entire egg is in the form of a thick spindle about 
"045 mm. in length and ‘021 mm. in greatest breadth. The 
shell is at this stage not yet fully solidified, so that the 
shape is readily modified by pressure, and the eggs tend to 
adhere together in masses. he shell consists of two distinct 
layers 
fibrillee —which run in the direction of the long axis of the egg. 

The completed egg in the primary uterus contains (1) the un- 
segmented ovum ; (2) a large number of small, bright globules 
(3) one, or, more commonly, two, larger rounded masses. 
The last two are the substance of the vitelline spherule. 
When the eggs are acted upon by any weak acid the small 
globules tend to run together into larger (2 [14], p. 28) masses, 
and eventually these pass out through the shell at the ends of 
the egg, so that in preparations fixed and stained by any of 
the ordinary methods this constituent of the egg becomes 
completely lost, there being left behind merely some irregular 
granular matter, in which, presumably, the globules were 
enveloped. ‘These globules, from their appearance and 
behaviour, are most probably composed of oily matter. 

The larger bodies derived from the yolk (see fig. 20) are of 
an entirely different character. They are solid masses having 


an outer homogeneous and an inner made up of fine 


the central hilum and concentric lamination characteristic of 
the calcareous corpuscles. ‘They become coloured, though 
not strongly, by staining agents, the central mass colouring 


ON A CESTODE FROM CESTRACION. 411 


first. In fixed and stained preparations they become much 
altered, having apparently become partly dissolved, and the 
concentric lamination being no longer discernible, might very 
easily be taken for nuclei. Like the oil globules, these bodies 
consist, doubtless, of food materials ; but both these ingredients 
of the yolk persist, not greatly diminished in bulk, to the 
most advanced stage observed. Nothing was made out with 
certainty as to the processes of maturation and impregnation. 
The oosperm does not differ to any appreciable extent from 
the ovarian ovum. 

Very few, if any, unsegemented ova were found in the 
secondary uterus. No definite history of the process of seg- 
mentation could be traced, as there seemed to be great varia- 
tion in the details. The first two segments (figs. 20 and 21) 
are equal. One of these, or both, become divided into two 
equal parts (figs. 22 and 23), and from the three or four 
equal, or nearly equal cells thus formed, a number of smaller 
cells become segmented off (figs. 24, 25, and 26). Eventually 
the larger cells become reduced by division until a blastoderm 
is formed consisting of a disc of small cells (figs. 27, 28, and 
29), which are very irregular in size and shape, and present 
no definite arrangement. This disc becomes thickened to 
form a rounded mass, on the surface of which appears here 
and there a flattened cell. In this stage there appears to be 
no further cell-differentiation, except that there are present, 
in the most advanced embryos, one or two pairs of very small 
cells that become more intensely stained than the rest. It is 
conjectured, from their arrangement in pairs, that these are 
the cells destined to develop the hooks. 

No hooked embryos were found in the uterus of any of the 
numerous specimens examined. But of a number of eges 
which had been kept in pure sea-water for five days, a large 
proportion (figs. 30—52) were found to contain fully formed 
active hexacanth embryos. It would thus appear that passage 
to the exterior with the feces is, under normal circumstances, 
the necessary condition for the development of the hooked 
embryo. 


412 


ono uo 


12. 


13. 


14. 


15. 


16. 


ay: 


18. 


WILLIAM A. HASWELL. 


LIvERATURE. 


. BenepEN, P. J. Van.—‘‘ Recherches sur la faune littorale de Belgique : 


Les Vers cestoides,”? ‘ Nouv. Mém. de |’Acad. Roy. de Belg.,’ t. xxv, 
1850. 

3ENEDEN, P. J. Van.—“ Mémoire sur les Vers intestinaux,” Suppl. aux 
‘Comptes Rendus de |’Acad. des Sciences,’ 1861. 


. Benepen, KE. Van.—* Recherches sur le développement embryonnaire de 


quelques Ténias,” ‘ Archiv. de Biol.,’ vol. ii, 1881. 


. BLancuarp, —.— Recherches sur |’organisation des Vers,” ‘ Ann. Sci. 


Nat.,’ 3 sér., t. vii and viii, 1847. 


. Braun, M.—“ Vermes ”’ of Bronn’s ‘ Thierreich.’ 

. Dresine, C. M.—‘ Systema helminthum,’ 1850. 

. Leucxart, R.—‘ Die Parasiten des Menschen.’ 

. Lryton, E.—‘ Notes on Entozoa of Marine Fishes of New England,” 


‘U.S. Fisheries Reports,’ 1886 (publ. 1889). 


. Linton, E.—* Notes on Entozoa of Marine Fishes of New Wngland,” 


part 2, ‘U.S. Fisheries Reports,’ 1887. 


. Linton, E.—‘* Notes on Cestode Parasites of Fishes,” ‘ Proce. U.S. 


National Museum,’ vol. xx, 1897. 


. Lonnzerc, 0.—“ Bidrag till Kannedomen om i Sverige forekommande 


Cestoder,” ‘ Bih. till K. Svenska Vetensk.-Akad. Handlingar,’ Bd. xiv, 
1889. 

LonneBeErG, E.—‘*‘ Helminthologische Beobachtungen von der Westkiiste 
Norwegens: 1 Thi., Cestoden,” ‘ Bil. till K. Svenska Vetensk.-Akad. 
Handlingar,’ Bd. xvi, 1890. 

Linz, M.—* Ueber einen eigenthtimlichen Cestoden aus Acanthias,” 
‘Zool. Anz.,’ xxiv, 1901. 

Montez, R.— Mémoires sur les Cestodes,” ‘ Travaux de |’Institut Zool. 
Lille,’ t. iii, 1881. 

MonrIce.ui, F.—‘‘ Nota intorna a due forme de Cestodi,” ‘ Bollettino 
dei Musei di Zoologia ed Anatomia comparata della R. Universita di 
Torino,’ vol. vii, 1892. 

Otsson, P.— Bidrag till Scandinaviens Helminthfauna II,” ‘ Kgl. 
Svenska Vetensk.-Akad. Handl.,’ Bd. xxv, 1893. 

Sanussow, H.—‘ Zur Histologie der Geschlechtsorgane von Trisenophorus 
nodulosus, Rud.,” ‘ Biol. Centralbl.,’ Ba. xviii. 

ScHavinsitanv, H.—“ Die embryonale Entwickelung der Bothrio- 
cephalen,” ‘Jen. Zeitsehr. f. Naturw.,’ Bd. xix, Neue Folge, xii, 1885. 


ON A CESTODE FROM CESTRACION. 413 


19. Surrey, A. K.— Description of the Entozoa,”’ A. Willey’s ‘ Zool. 
Results,’ part 5. 


20. Zscnoxke, F'.—‘ Récherches sur la structure anatomique et. histologique 
des Cestodes ” ‘Mem. de l'Institut nation. Génévois,’ t. xvii, 1S86— 
1889. 

21. Zscuoxxke, F.— Studien itiber den anatomischen und _histologischen 


Bau der Cestoden,” ‘ Centralbl. f. Bakteriologie u. Parasitenkunde,’ i, 
1887. 


EXPLANATION OF PLATES 22—24, 
Illustrating Prof. William A. Haswell’s paper ‘‘ On a Cestode 


from Cestracion.”’ 


List oF Rererence Letters. 


ec. Cirrus. ¢.s. Cirrus sheath. ez. Cuticle. d. Depression at posterior 
end of free proglottis. d.v.m. Dorso-ventral muscular fibres. e.¢c. m. Exter- 
nal layer of circular muscular fibres. e. 2. m. External layer of longitudina 
muscular fibres. ea. Main excretory vessel. fd. Fertilising duct. 
h. “Head” of separate proglottis. 7./.m. Internal longitudinal layer of 
muscle. z.c. Nerve cord. o. d.' First part of oviduct. o.d.? Second part of 
oviduct. ov. Ovary. ov. m. Median part or istlimus of ovary. p.u. Ootype 
or primary uterus. 7.s. Receptaculumseminis. s.¢@. Sperm duct. s.g. Shell- 
gland. s.w. Uterus. sw. “Swallowing apparatus.” sw. m. Muscular layer 
of swallowing apparatus. sw. my. Myoblasts of swallowing apparatus. e. 
Lobes of testis. v. Lobes of vitelline glands. va. Vagina. v.d. Vitelline 
ducts. v.7. Vitelline reservoir. v.7.c. Constriction at posterior end of 
vitelline reservoir. @. Male reproductive aperture. 9. Female reproduc- 
tive aperture. 


PLATE 22. 
Fie. 1.—Scolex of Phyllobothrium vagans magnified. 
Fie. 2.—Last proglottis of strobila magnified. 
Fic. 8.—Free proglottis, dorsal aspect. Nervous system, blue; excretory 
vessels, green; testicular ducts, red. 
Fie. 4.—Portion of transverse section of strobila, showing integument and 
muscular layers. x 600. 


414 WILLIAM A. HASWELI.. 


Fic. 5.—Transverse section of cirrus. cz. Cuticle, with spinules. ¢.m. 
Layer of circularly arranged muscular fibres. 7. m. Layer of longitudinal 
muscular fibres. my. Layer of myoblasts. 


Fic. 6.—General view of the female reproductive apparatus as seen from 
the ventral side. 


Fie. 7.—Dorsal view of the median part of the ovary and of the neigh- 
bouring ducts. 


PLATE 23. 


Fic. 8.—From a series of longitudinal (horizontal) sections. Section 
passing through swallowing apparatus, first part of oviduct and main vitelline 
duct. x 450. 


Fic. 9.—From the same series. Section dorsal to that represented in 
Fig. 8, showing vagina, receptaculum seminis, and shell-gland. x 450. 


Fre. 10.—From the same series. Section dorsal to that represented in 
Fig. 9, showing receptaculum seminis and fertilising duct. x 450. 


Fic. 11.—From a series of transverse sections. Section passing through 
swallowing apparatus and median part of ovary. 


Fie. 12.—Section immediately behind that represented in Fig. 11. 
Fic. 13.—From an oblique series. Mouth of swallowing apparatus. x 600. 


Fic. 14.—From a transverse series. Showing swallowing apparatus and 
its relations to ovary. x 600. 


Fic. 15.—From a transverse series. Showing relations of swallowing 
apparatus to oviduct. x 600. 


Fic. 16.—From a transverse series. Swallowing apparatus and oviduct. 
x 600. 


Fic. 17.—From a horizontal series. Section of oviduct at the point where 
the ducts of the shell-gland open into it; an ovum in the act of union with a 
yolk-zell, 


Fic. 18.—Transverse section to show the relations of the primary uterus, 
the vagina, and the ruptured secondary uterus. 


PLATE 24. 
All the figures drawn under Zeiss’s apochromatic 2-0 mm, objective and 
compensation ocular 12, magnifying 1100 diameters. 
Fic. 19.—Kgg with unsegmented ovum. T'rom preserved specimen. 


Vic. 20.—Two-celled stage. Fresh specimen, showing the globules and 
coneentrically laminated bodies of the vitelline mass. 


Fic. 21.—Two-celled stage. Preserved specimen, 


ON A CKSTODE FROM CESTRACION. 415 


Fic. 22.—Three-celled stage. 

Fic. 23.—Four-celled stage. 

Fie. 24.—Stage of about eight, cells. 

Fie. 25.—Surface view of blastoderm of a somewhat later stage than that 
represented in Fig. 24. 

Fic, 26.—Stage of about fourteen cells. 

Fie. 27.— Surface view of dise-like blastoderm. 

Fies. 28 anp 29.-—Disc-like stages seen edgewise. 

Fig. 830.—Hexacanth embryo with the hooks retracted. 

Fie. 31.— Hexacantl embryo with the looks everted, 


Fic. 82.—EKgg containing hexacanth embryo. Fresh specimen. 


von. 46, PART 3,—NEW SERIES. C= 


6: be> iAP Seren te ° 


€ eee | ® 


; ot AAS ek 


% coll drg)) Tan ng a 
ae on 4! “45% ent ' hve Lert? \ a mtg 


12 yee 
‘ - = 

ol is ous tie 1 tometer bs oot oe fad 
: . it 

ine libtd 8 oF Mio.¢ <AG 
F eae VEL erie Gepale ugil-ted! a 2% 7o . 
a 7 sane had “ne & 5m ih, ue 7 a) boi he A) gage ‘e = 
ec dis itr ence i at) Ss i OES Pe ees 
vt an Fr a | » 4 hp) 0 TT) a -. e | 


+ 
- - 
. 
. * ’ 
ry 
? ¥ - > ape) 
7 - i) ty £ Wie 
_ be > a - ian 
J 
7 oa -\5 © Pom 


a ee fe 


7 hop Fo. ee Fa 
9 a Prana Jf 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 


417 


The Development of Lepidosiren paradoxa. 


By 


J. Graham Kerr, 


Regius Professor of Zoology in the University of Glasgow. 


Part III.—Development of the Skin and its Derivatives. 


With Plates 25—28. 


CONTENTS. 


The general epidermis 
The buceal cavity 
The teeth 
The hypophysis 
The central nervous system 
The brain of the adult : 
Development of main features of brain topography 
Thalamencephalon 
The hemispheres . 
Mesencephalon 
Rhombencephalon 
The sense organs: 
Olfactory organ . 
The eye 
Auditory organ 
General remarks 
Summary : 
Explanation of Plate 


INTRODUCTION. 


PAGE 


418 
423 
424 
427 
428 
428 
429 
434 
437 
458 
438 


438 
439 
445 
447 
452 
454. 


Iy the followmg pages I give an account of the chief 
results obtained from my investigation of the development 


418 J. GRAHAM ‘KERR. 


of Lepidosiren, in so far as they relate to the skin and 
certain organs associated with it. Some structures which 
ought logically to be described now have been deliberately 
omitted: such are the external gills, which I propose to 
consider at the same time as the branchial clefts. I also 
make no attempt to describe the various organs with an 
equal degree of fulness. In regard to some, which I have 
found specially interesting, I give a fairly detailed account ; 
in regard to others I am content. to outline the main 
features. 

I have endeavoured to avoid obscuring my description by 
going into masses of minute detail, feeling that by so domg I 
should often be trespassing in regions where individual varia- 
tion, and the “probable error” of observation, make results 
useless if not actually misleading. 


Tur GENERAL EPIDERMIS. 


It has already been shown that the first part of the 
epidermis to assume a fairly definite shape is that covering 
the ventral surface of the embryo, which is simply the 
persistent roof of the segmentation cavity. The epidermis 
covering the dorsal surface of the body, on the other hand, 
develops later. It has also been shown that during gastru- 
lation, in Lepidosiren, the dorsal hp of the blastopore is 
composed of a mass of undifferentiated cells, showing no 
distinction into layers. In this Lepidosiren differs from 
Ceratodus, where, as Semon pomts out, the epiblast is 
marked off by a distinct split right back to the lip itself. 
Elsewhere than at the blastopore lp epiblast is formed by 
delamination: from the large yolk-cells underneath. By 
Stage 14, when the process of gastrulation is finished, the 
embryo is already covered uniformly by a definite stratum of 
epiblast composed of two layers of closely apposed flattened 
cells, except just in the lp of the blastopore where the 
germinal layers are still, and will remain for some time 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 419 


undifferentiated. From the first the epiblast is thicker in 
the region overlying the archenteron, the cells of which it is 
composed being here somewhat columnar. The fate of this 
thickened region, which is destined to give rise to by far the 
most complicated product of the ectoderm, the central nervous 
system, may conveniently be left out of consideration until a 
little later. 

In regard to the general ectoderm there is little change 
to chronicle for a considerable period. In an embryo of 
Stage 25 it is still two-layered, the bounding surfaces of the 
two layers being smooth and parallel. Prominent yolk 
granules are still present in the cells, and the outer layer has 
formed on its surface a fine but distinct cuticle. When the 
tail begims to form the ectoderm at the tip of this thickens, 
its cells assuming a cuboidal form, but remains two-layered. 

The general ectoderm retains its two-layered condition 
for some time. Increase in thickness by division of the 
lower layer cells begins at a period varying from about 
Stage 52 to about Stage 55. 

In various young Lepidosirens, which have been fixed in 
strong Flemming’s solution, I have been able to make out that 
certain of the ectodermal cells are provided with tail-like 
processes of cell substance, closely resembling the tails so 
characteristic of the ccelenterate epithehal cells (PI. 25, fie. 1). 
The tails of the ectodermal cells in Lepidosiren are very 
difficult to observe, showing up distinctly only in well-fixed 
material in which the cells are shehtly separated from one 
another. They run along the mner surface of the epidermis, 
forming a kind of plexus-hke layer. Into this layer pass 
also processes from the underlying mesenchyme cells, so that 
it forms an organic connection between ectoderm and mesen- 
chyme. 

The glandular structures of the fully formed skin of 
Lepidosiren (PI. 25, fig. 2) are one of its most charac- 
teristic features. The tall unicellular mucus glands, which, 
in the adult, form a palisade-hke arrangement through the 
whole thickness of the skin, begin to appear about Stage 35 


420 J. GRAHAM KERR. 


as ordinary cells of the epidermis, whose cytoplasm assumes 
a clear vacuolated appearance, the whole cell remaining in 
form and size like its neighbours. By Stage 38 the gland- 
cells have become predominant by their size, and they are 
also elongating in shape. 

The multicellular glands appear about the same time as 
downgrowths of the deep layer of the epidermis, and here 
again we find that the rudiment is solid, and the cavity 
appears secondarily. This we can naturally not put down, 
as we do in the case of certain other organs arising similarly, 
to any such simple cause as the presence of yolk. By Stage 38 
a large cavity has appeared, but it is not yet open to the 
exterior. 


Cement Organ. 


A remarkable local development of epidermal gland-cells 
is afforded by the cement organ, which, as indicated before, 
retains through life the crescentic shape shown by Thiele to 
be characteristic of the organ in its early stages in Batra- 
chians. It is a curious point, however, to which my attention 
was first drawn by my friend Mr. Bles, that in the Amphibia . 
the organ is derived from the superficial layer of the epi- 
dermis, not the deep layer as in Lepidosiren. 

The first indication of the cement organ appears about 
Stage 23 (Pl. 25, fig. 3 4) as a sheht thickening of the deep 
layer of the epidermis, the superficial layer passing over it 
hardly affected. By Stage 25 the thickening has considerably 
increased, and the superficial layer now shows signs of break- 
ing down over the middle of the gland, so that here the deep 
cells are exposed to the external medium (fig. 5 3B). By 
Stage 31 the gland is fully functional. Its cells are tall and 
columnar with nucleus at the base, and protoplasm showing 
peripherally a clear transparent appearance. 

During the later stages of development the glandular 
surface becomes involuted slightly, and at the same time its 
lower edge becomes tilted up somewhat, so that the organ 
projects conspicuously above the adjoining skin surface. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 421 


Degeneration of the Cement Organ.—The process of 
atrophy of the cement organ is a comparatively rapid 
process, taking place about Stage 35. It is illustrated by 
fig. 5D. 

In the early stages of degeneration the glandular epi- 
thelium becomes penetrated by vascular loops, and leucocytes 
begin to concentrate in its neighbourhood. At a later stage 
(e.g. Stage 35, fig. 3p) there are crowds of leucocytes 
collected about the gland, and it is now seen that they are 
laden with fatty and other granules, the product of their 
active metabohism. The glandular part of the ectoderm 
becomes gradually consumed, and the adjoining epidermis 
becomes shrivelled and has its surface thrown into wrinkles 
as the gland cushion diminishes in size. 


Pigment Cells. 


About Stage 35 branched pigment cells begin to appear in 
the ectoderm. I believe that these are all mesodermic in 
origin. In sections from embryos of about this stage many 
examples of pigment-laden chromatophores may be seen in 
process of migration into the ectoderm. 

The only case, in fact, that I have found of pigment 
granules being formed to any conspicuous extent in epidermal 
cells is that of the pigment layer of the retina. 


Changes in Chromatophores caused by Alteration 
in the Amount of Incident Light. 


I have already referred (Part I, p. 320) to the remarkable 
difference in the appearance of a youne Lepidosiren 
during the day and mght. A Lepidosiren of Stage 38, 
which by day is of a deep rich brownish black, becomes at 
night-time quite colourless, the change being associated with 
the withdrawal of the dendritic pseudopodia of the chro- 
matophores. An inspection of Pl. 25, figs. 4a and 4.8, will 


4,22, J. GRAHAM KERR. 


illustrate the appearance of the skin of a Lepidosiren of 
the stage mentioned during the day and durimg the night. 
The night specimen had been exposed to faint lamp-lght for 
several minutes, and consequently the retraction of the 
pseudopodia is not quite complete. 

From fig. 4 it will be seen that the black chromatophores 
tend towards two distinct types, differing in the appearance 
of the contained pigment and in the degree of ramification of 
the pseudopodia. 

In type A, which is the less numerous, the pigment is very 
black, the cell body is compact, and the pseudopodia are 
long and comparatively slightly branched, and often present 
a varicose appearance. In type B the contained pigment is 
less opaque, of a brownish colour, and the cell body is often 
very irregular in shape, projecting in great trunks from 
which arise numerous short and very irregular pseudopodia. 

Of these type B appears to be the more highly sensitive to 
light, a much fainter amount of light sufficing to cause 
extrusion of its pseudopodia. 

When in a state of maximum expansion the pseudopodia 
frequently anastomose both with their neighbours and with 
those of other chromatophores. Anastomosis often takes 
place between pseudopodia belonging to chromatophores of 
the two different types. This, together with the presence of 
intermediate forms, indicates that the two types are not 
really distinct, but are merely the extremes of variation of a 
single type. 

It is instructive to compare vertical sections through the 
skin in different light conditions. Fig. 5 illustrates such 
sections from (1) a young Lepidosiren of Stage 38, killed 
at 9 p.amn., by faint lamp-light (fig. 5c), (2) one of the same 
brood taken from deep shade at 2 p.m. (fig. 5 B), and 
(3) a rather younger Lepidosiren taken from an open 
white enamelled dish with clear water and exposed to bright 
diffused dayheht (fig. 5 A). 

In (1) the chromatophores are in their state of maximum 
contraction, and | may mention that the scattered chromato- 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA, 4238 


phores deep down in the substance of the body are also 
contracted. 

In (2) the chromatophores have their pseudopodia fully 
extruded. In the case of pigment cells within the epidermis 
the pseudopodia pass between the cells up towards the 
surface. ‘The chromatophore tends to push its pseudopodia 
towards the hght; their movements are positively helio- 
tropic. In the case of chromatophores lying in the super- 
ficial layer of the dermis the cells flatten themselves out 
against the lower surface of the epidermis, forming with their 
pseudopodia a practically continuous heht-proof coat. 

In (5) the chromatophores are seen to have their pseudo- 
podia at the maximum of extension. 


THe “ Stomopzum.” 


In the young Lepidosiren up to Stage 30 there is no 
stomodzum present; the enteric rudiment, solid in this 
region and sharply marked off from surrounding tissues by 
its cells being packed with large yolk granules, extends right 
up to the external epiblast. 

About the stage mentioned a change is seen to be setting 
in in the anterior part of the enteric rudiment, corresponding 
to what will become the buccal cavity. The superficial layer 
of the still solid rudiment is seen to be approximating in 
character to the ectoderm. Its yolk granules become finely 
broken up, showing that active metabolism is taking place ; 
protoplasm and nuclei are becoming more abundant. In this 
way there arises a layer of definite epithelium continuous 
anteriorly with the external epiblast, sharply marked off 
from the embryonic connective tissue outside it, but im- 
ternally passing without any sharp boundary into the yolk- 
laden mass inside (cf. Pl. 25, fig. 64). It is, as it were, as if 
an influence were spreading inwards from the external epi- 
blast, gradually transforming the original “ endoderm” yolk- 
laden cells into ectoderm like itself. I find no evidence of 


424, J. GRAHAM KERR. 


an actual bodily involution of ectoderm such as is ordinarily 
associated with the term stomodzeum. On the contrary, the 


“ stomodeeal ”’ cells 


perfectly gradual transition between the 
and the typical yolk-laden endoderm cells shows quite con- 
clusively that the former are beg derived from the latter. 
The buccal rudiment retains its solid character till about 
Stage 31. About this time the cells in its interior begin to 
degenerate and break down, and so give rise to the cavity of 
the mouth. 

The tooth germs begin to appear long before there are any 
traces of lumen in the buccal cavity.’ Already in Stage 32 
they may be detected. 


Development of the Teeth. 


One of the most striking points brought out by Professor 
Semon’s researches on the development of Ceratodus has 
been the way in which the so characteristic tooth plates are 
formed by the joming together, by dermal bony trabecule, of 
originally separate denticles. On coming to consider the 
tooth development of Lepidosiren I not unnaturally ex- 
pected to find a similar state of affairs, and I was accordingly 
much astonished on using appropriate macerating media to 
fail completely to discover separate denticles. I then turned 
to young specimens of Ceratodus, and had no difficulty in 
completely confirming Semon’s description. In Lepido- 
siren the only possible reminiscence of such a stage in tooth 


1 In Urodele Amphibians the teeth similarly develop before a lumen is 
formed, and the lining of the buecal cavity appears to arise in them just as 
in Lepidosiren and Protopterus. Rése (Schwalbe’s ‘ Morphologische 
Arbeiten,’ Bd. iv, S. 182), in describing the development of the teeth in 
Urodeles, talks of the buceal cavity being “ mit Dotterplattchen und abgestos- 
senen schollenformigen Epithelzellen ausgefiillt.”” On the contrary, I should 
say, from a study of my own sections of Urodele embryos (Amblystoma 
and Triton). that the buecal cavity has not yet arisen at the stage of which 
Rose is speaking. In fig.6B I figure a section of the mouth region of an 
Amblystoma of the stage in question for purposes of comparison with the 
corresponding section from Lepidosiren. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 425 


arrangement is to be found in the fact that in the young 
individual the teeth are furnished with definite prominent 


pointed cusps—each probably representing the tip of an 
originally simple denticle,—although in ontogeny they de- 


velop as a perfectly continuous ridge from the beginning. 


Text-ric. 1.—Sagittal section through head region of a Protopterus 
larva (Stage 33). Cam. Zeiss a*, oc. 4. 
b.c. Buccal cavity. d.p. Dental papilla. ep, Pineal body. 
p. Paraphysis. ¢. ‘Thyroid rudiment. 


The first obvious rudiments of teeth occur about Stage 30 
in the form of a thickening of the oral epithelium,! under 
which the mesoblast becomes concentrated as it were, the 
nuclei being crowded much more closely together than 
elsewhere. 

By Stage 31 the thickening is growing downwards into 
the mesoblast so as to border on each side a ridge-like 
“ yapilla ” of mesoblast (Pl. 26, fig. 7 a). 

The first traces of hard structure in the tooth appear about 
Stage 52, when a conical calcareous cap appears beneath the 


1 For general topographical relations of this see Text-fig. 1. 


4.26 J. GRAHAM KERR. 


enamel organ. It is difficult to arrive at a certaim opimon on 
the morphological nature of this first formed cap. It adheres 
stronely to the enamel organ, as shown by the torn surfaces 
when the two structures have been pulled apart in process of 
preparation, and in many cases it 1s for a time sharply marked 
off from the underlying dentine. On the other hand, it 
differs from ordinary enamel in the much larger proportion 
of organic matter, which causes it to remain quite distinct 
even in decalcified specimens. 

On the whole, [ am inclined to look upon this structure as 
beg enamel, though of a somewhat modified kind. 

The structure of the palatopterygoid teeth about Stage 35 
may be gathered from the sections represented in figs. 7 B, 7 ¢, 
and 7 p. The enamel forms a distinct cap tapering off 
towards its edges, and sharply marked off from the under- 
lying dentine. It shows a faint striation perpendicular to its 
surface. In undistorted sections the flat mner ends of the 
enamel cells abut close against it (fig. 7D). 

At this stage there is a quite definite though still thin 
layer of dentine lying within the enamelcap. The broadened 
outer ends of the odontoblasts come into close contact with 
one another, and form, to the eye, a quite continuous mass 
(fig. 7p). As they pass into this their protoplasm shows a 
development of fine fibrillee crossmg one another in all 
directions. Traced still further out the fibrillar mass gradu- 
ally takes on more and more deeply the stain which, in 
Heidenhain’s hematoxylin preparations, indicates the pre- 
sence of calcareous matter. The thin outer layer is, im fact, 
fully calcified dentine, on its inner side passing by impercep- 
tible gradations into the ordinary protoplasm of the odonto- 
blasts, on its outer side marked off from the enamel by a 
sharp boundary. 

In Stage 36 (fig. 7 ») the formation both of dentine and of 
the bony trabecule which form the spongy basal support for 
the tooth is seen to have made considerable progress. ‘The 
ridges of the tooth now approach the surface of the oral 
epithelium, which is becoming thin over their apices prepara- 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 427 


tory to breaking through. 'The enamel layer is closely fused 
with the underlying dentine; the sharp line separating the 
two has completely disappeared, and it is only possible, by 
the use of very high powers, to distinguish the enamel by its 
clear appearance without any obvious structure from the 
dentine, which still shows a faint reticular or fibrillar struc- 
ture—the remnants of the more obvious structure of the 
same kind in the uncalcified odontoblast. 

Finally, in Stage 38 (fig. 7 F), when the young Lepidosiren 
has already begun to feed, the teeth have broken freely 
through the oral epithelium, the enamel organ having disap- 
peared entirely except for a vestigial flap (fig. 7 F, e.0.) stick- 
ing up all round the base of the tooth. The enamel is now 
no longer to be detected in my sections: it has probably been 
worn off, being doubtless, from its larger proportion of 
organic matter, much less hard than ordinary enamel. The 
mass of dentine has much increased in size. A little later its 
central portion assumes the hard glassy character of the 
vitrodentine of the adult (“Enamel,” ‘Tomes’ ‘ Dental 
Anatomy,’ fifth edition, p. 263). 


Hyporpnwysis. 


The hypophysis is somewhat obscure in Lepidosiren. 

In Stage 254 it is visible as a somewhat wedge-shaped in- 
erowth of the deep layer of the epiblast. 

In Stage 294+ the deep end of the structure has become 
slightly swollen, with indications of a longitudinal split in its 
middle; the portion connecting this with the ectoderm is 
thinned down to a narrow thread occupying the space between 
the closely approximated front end of the gut and the floor of 
the fore-brain. 

At a stage very slightly later (30) the connecting isthmus 
is nipped through, while the expanded extremity, whose 
split is now widening out into a definite cavity, les as a 
closed sac beneath the infundibulum. 


d 


428 J. GRAHAM KERR. 


Eventually the hypophysis becomes here as elsewhere 
closely united with the infundibulum, its dorsal portion 
becoming partly penetrated by tubular outgrowths of the 
latter (saccus vasculosus, cf. p. 432, Text-fig. 2, H). 


CrentrRAL Nervous System. 


As the brain of the adult Lepidosiren! has never before 
been investigated in the fresh condition, I give on Plates 
26 and 27 figures illustrating its conformation, and showing 
the roots of the cranial nerves, including those of the fourth 
and sixth, whose existence in Lepidosiren has hitherto been 
doubted. By a comparison with Burckhardt’s figures of 
Protopterus it will be seen that the two brains are very 
similar. In dorsal view the only difference is in the relative 
size of the different parts. In Lepidosiren the mid-brain 
region is relatively longer, the thalamencephalon relatively 
shorter than in Protopterus. In my figure I have not 
shown the extensive system of outgrowths from the saccus 
endolymphaticus which here, as in Protopterus, overlies and 
to a great extent hides the region of the hind brain. 

In the ventral view of the brain the cerebral hemispheres 
are not sharply marked off from the thalamencephalon. The 
swelling at the base of the olfactory nerves is much more 
conspicuous, owing to the smaller size of the post-olfactory 
lobe which underlies them. The lobi inferiores are much 
more strongly developed, the hypophysis is more rounded in 
form, and the hind brain is like the cerebral region less broad 
from side to side as compared with the thalamencephalon. 

In the side view of the brain the most striking difference 


1 In dissecting the brain of Lepidosiren one is struck by the extra- 
ordinary development of richly ramifying blood-vessels within the cranial 
cavity, forming a packing all round the brain. ‘This may possibly be an 
adaptation to the times at which it is impossible to make the blood rich in 
oxygen, during the final stages in drying up of the swamps, or during casual 
rainfalls in the dry season. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 429 


from what is found in Protopterus occurs in the cerebral 
hemisphere in the much smaller development of the post- 
olfactory lobe. The “lobus hippocampi” described by 
Burckhardt for Protopterus is less distinctly marked off in 
Lepidosiren, but is still distinctly visible. 


The Development of the Main Topographical 
Features of the Central Nervous System. 


I now proceed to describe in outline the main features in 
the brain and spinal cord of Lepidosiren. The minute 
structure and details of histogenesis I propose in this general 
account of the development to leave completely on one 
side. 

On Pl. 27, fig. 10, are given a series of drawings of the 
brain in side view, and on Pl. 26, fig. 8, are given selected 
stages as seen from the dorsal aspect, and an inspection of 
these figures will suffice to give a clear idea of the evolution 
of the external features of the brain without any elaborate 
verbal description. 

In my description of the early stages in development the 
brain and spinal cord were left (this Journal, vol. 45, p. 
23) when they were still in the condition of a partly solid 
rudiment. From the beginning the anterior or brain region 
is distinguished by its greater width. 

At about Stage 20 or 21 a slight constriction appears 
marking off the region of the hind brain from the region in 
front of it. 

At about Stage 25 a transverse wrinkle in the floor of the 
brain begins to appear to mark the commencement of cranial 
flexure.. By Stage 26 (fig. 10 4) this has become well marked. 
A slight bulging on each side of the thalamencephalon at this 
stage marks the rudiment of the cerebral hemisphere. 

By Stage 29 (cf. fig. 10 8) the cranial flexure has become 
more pronounced, and a depression of the brain-roof has 
begun to show itself in the region of the anterior limit of the 
hind brain. The anterior corner of the hind brain has grown 


430 J. GRAHAM KERR. 


out to form a prominent knob on each side, and the cerebral 
hemisphere has become more distinct. 

From Stage 30 to 32 (figs. 10 c, p, and ©) the chief changes 
consist in the close approximation of the infundibular region 
to the floor of the hind brain, in the appearance of the pineal 
outgrowth, and in the commencing forward growth of the 
two hemispheres. It is to be noted that up to nearly Stage 
32 there is no obvious separation of mid-brain from thalamen- 
cephalon. In a brain of about Stage 35 (fig. 10 F) the chiet 
advance consists in the considerable growth forwards of the 
hind brain on each side, so as in side view to completely hide 
the floor of the mid-brain. The roof of the thalamencephalon, 
forming for the most part the pmeal cushion,’ is now quite 
sharply marked off from the roof of the mid-brai. From 
now onwards to Stage 38 changes in external conformation 
are but slight, as will be seen from fig. L0G, the chief one 
being in the upward growth of the hemisphere-roof, so that 
it, with the pineal cushion, rises to about the same horizontal 
line as the summit of the mesencephalon. 

In the last stage (Stage 39), which I figure (fig. LOH), a 
marked advance towards the adult condition is seen, the 
cerebral hemispheres having undergone a™large increase 
in antero-posterior length. The olfactory lobe is already 
marked off. 

The chief change subsequent to this consists in the further 


ereat elongation of the brain axis. 


Dorsal Aspect of the Brain. 


The earliest stage which I have figured (fig. 8 a, Stage 31) 
illustrates (1) the relatively enormous size of the hind brain ; 
(2) the fact that mesencephalon and thalamencephalon form 
a single perfectly definite brain region, on the roof of which 
the pineal body has appeared; and (3) the paired independent 
rudiments of the hemispheres. 

In the view of Stage 35 (Pl. 26, fig..88) the forward 


' Zirbelpolster. 


THE DEVELOPMENT OF LNPIDOSIREN PARADOXA. 451 


growth of the hind brain on each side is seen, the elongation 
of the mesothalamencephalic region and its distinct division 
into a mesencephalic part behind, and a thalamencephalic 
portion in front, the latter thin-roofed, and forming a kind of 
pillow or cushion (pineal cushion), upon which the pineal 
body rests. Lastly, the elongation of the cerebral hemispheres 
has now begun. 

In the brain of Stage 38 (fig. 8c) this elongation of the 
cerebral hemispheres, of the mid-brain, and of the lateral 
angles of the hind brain is seen to have gone on still further. 

Finally, in the adult brain (fig. 8 p) the great elongation of 
the antero-posterior axis is very obvious, affecting all regions 
of the brain except the thalamencephalon. It will be noticed 
that the lateral angles of the hind brain have lagged behind 
in this lengthening, so that now they do not project forwards 
at all. <A curious shifting of the poit of origin of the 
olfactory nerves is also seen to have taken place in the later 
stages of development, the point of origin in the adult being 
terminal and anterior instead of external and some distance 
back from the front end of the hemisphere. 


Changes in the Sagittal Section of the Brain 
during Development. 


It is very instructive to compare sagittal sections of the 
brain taken at various stages of development, and in Text-fig. 
2 I give figures of a series of such sections. The outlines 
of the sections are from camera drawings. 

In Stage 25 (fig. A) a slight elevation of the brain floor is 
seen ; this marks the hinder boundary of the thalamencepha- 
lon, and the portion in front of it corresponds with what 
will later on be the floor and part of the anterior wall of the 
thalamencephalon. 

At Stage 28 (fig. B) the posterior boundary of the thala- 
mencephalon is more strongly marked: the floor of this part 
of the brain shows a shght thickening. 

Within the next two stages (cf. fig. C) considerable 

VOL. 46, PART 3.—NEW SERIES. DD 


Texr-riG,. 2.—Camera tracings of sagittal! sections through the brain of Lepidosiren 
at successive periods of development. Figs. A—G are drawn under the same magnifica- 
tion, Fig. H under a lower magnification. A stage 25, B stage 28, C stage 29x, D stage 
31, E stage 31x, F stage 35, G stage 38, H adult in second year. 

Reference Letters: a. Anterior end of floor of brain rudiment. a.c. Anterior commis- 
sure. c¢.h, Cerebral hemisphere. ch. Optic chiasma. ep. Pineal body. g.hab. Ganglion 
habenule. h.c. Habenular (superior) commissure. hyp. Hypophysis cerebri. par. Para- 
physis. p.c. Posterior commissure. pl. laf. Choroid plexus of lateral ventricle. s.v. 
Saccus vasculosus. vel. Plexus of third ventricle. 


_ Structures occurring not in the sagittal plane, but in sections parallel to but some 
distance from it, are shaded with oblique lines. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 4383 


developments take place. The fold in the brain floor behind 
the thalamencephalon has become still better marked, and 
over it the brain roof bulges upwards in what will later be the 
roof of the mid-brain. Posterior to this the brain-roof is 
distinctly thinner, foreshadowing the membranous condition 
of the roof of the fourth ventricle. 

In the floor of the thalamencephalon a depression has 
appeared, separating two elevations. Of these latter the 
anterior one is the rudiment of the anterior commissure, the 
posterior of the optic chiasma, while the depression between 
is the preoptic recess. 

Stage 31 (fig. D).—The fold of the brain floor behind the 
thalamencephalon has now increased very much in height; a 
downfolding of the roof has also appeared, bounding the 
mid-brain roof behind. 

The elevations in the floor of the thalamencephalon formed 
by the rudiments of anterior commissure and optic chiasma 
are now much more strongly marked though still without 
fibres. The pineal rudiment makes its appearance about 
this time as a small simple diverticulum of the brain roof, 
and in fact the thalamencephalon has assumed very much its 
definitive character. 

The anterior wall of the thalamencephalon at about the 
middle of its height is prolonged upwards and forwards as a 
cylindrical tube passing between the hemispheres, and ending 
blindly by a rounded end in front. This must be the para- 
physis, and it closely resembles the paraphysis of Urodeles.’ 

Stage 31 + (fig. H)—Fibres have now appeared in the 
chiasma and anterior commissure. ‘The front wall of the 
thalamencephalon is thinning out. The roof of mid-braim and 
thalamencephalon is of practically uniform thickness through- 
out, and there is no definite distinction between the two. 
Both the superior and the posterior commissures are well 


1 In regard to the region of the thalamencephalon, I propose to adopt the 
usage of the terms epiphysis, paraphysis, ete., favoured by Gaupp (Anat. 
Hefte, Abth. 2, Bd. vii, 8. 229). 

2 See Minot, ‘Amer. Journ. Anat.,’ vol. i, p. 93. 


434, J. GRAHAM KERR. 


developed, and fibres have appeared over most of the cerebral 
and mid-brain surface. 

The paraphysis has increased in size, forming a tube about 
‘05 mm. in diameter, and with walls composed of a single 
layer of cubical cells. This runs straight forwards and up- 
wards for a distance of about a quarter of a millimetre, and 
ends blindly in front. 

Stage 35 (fig. ).—The main new features are the appear- 
ance of a sharp depression marking off the floor of the mid- 
brain from that of the hind brain, and the fact that the region 
of the anterior commissure is undergoing a rotation forwards 
and upwards, so that it is becoming transferred from the 
floor of the thalamencephalon to its anterior wall. By this 
stage the process is still inconspicuous, but later on it becomes 
very obvious. 

The habenular commissure is very well developed. The 
paraphysis opens into the cavity of the thalamencephalon 
almost vertically under the root of the pineal body. 'The 
paraphysis has altered its position, having become rotated 
backwards so as to lie flat against the lamina terminalis 
almost vertically. 

Stage 38.—The changes seen in sagittal section are com- 
paratively trivial. The paraphysis is, however, relatively 
reduced, and the mid-brain is more sharply marked off from 
the hind brain both in roof and floor. 


Development of the Various Brain Regions. 


It will now be convenient to run over the main points 
in the development of each of the regions of the brain in 
turn. ‘ 


Thalamencephalon. 

We have already seen that the thalamencephalon begins 
to be marked off from the regions behind it as early as 
Stage 25, and it rapidly assumes its permanent features. 
The roof is for a long time perfectly continuous with that of 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 435 


the mid-brain. The first mdication of where the boundary 
between the two regions will come is given by the appear- 
ance of the pineal body. Appearing first about Stage 
31 as a simple flattened evagination of the brain roof, 
and never showing at any time any symptoms of division 
into a true pineal body and a parietal organ, the pineal 
body increases somewhat in length, and eventually becomes 
somewhat carrot-shaped, beimg attached by its hinder 
narrower end. Its walls remain simple, and the communi- 
cation of its lumen with that of the brain persists till 
Stage 38. In the brain of Stage 39 the lumen has become 
obliterated posteriorly. Anteriorly it still persists, but the 
cells bounding it show signs of breaking down, and the 
cavity is to a great extent filled with granular material 
apparently derived from them. At no time does the pineal 
body of Lepidosiren show any eye-like characters, and it 
always les well down below the skull roof. 

The roof of the thalamencephalon in front of the pineal 
body, while remaining thin in the median line, thickens out 
greatly on each side to form the habenular ganglion. These 
thickenings become conspicuous about Stage 31, and the 
commissure connecting them—the superior commissure—be- 
comes conspicuous about the same time. 

The pineal body lying on the roof of the thalamencephalon 
presses down the thin roof between the habenular ganglia, 
very much as a head does the pillow on which it rests. 
Consequently the name pineal cushion or pillow (Zirbel- 
polster) is a thoroughly suitable one for this region of the 
brain-roof. From the originally anterior wall of the thala- 
mencephalon is developed the paraphysis, the chief characters 
of which have been sufficiently brought out in the study of 
sagittal sections. Appearing first about Stage 31 as a small 
evagination of the anterior wall of the thalamencephalon at 
about its middle, it becomes a straight tubular structure, 
running upwards and forwards between the hemispheres. 
The paraphysis soon becomes relatively reduced in size. 
By Stage 39 it forms an irregularly twisted tube, somewhat 


436 J. GRAHAM KERR. 


dilated at its distal end, and lying between the choroid 
plexuses. The paraphysis does not give rise to plexus 
itself, although, from its very close relations with the choroid 
plexus on each side, it may on superficial examination appear 
to do so} 

On the floor of the thalamencephalon it has already been 
seen that the preoptic recess becomes distinct very early, 
marking off the region of the chiasma from that of the 
anterior commissure. The tip of the infundibulum remains 
quite simple as far as Stage 38, but in Stage 39 it is found 
to have sprouted out into narrow, tubular, and apparently 
glandular diverticula, which, winding hither and _ thither, 
fuse with and to a certain extent penetrate the dorsal part of 
the hypophysis, forming the so-called nervous part of the 
hypophysis. 

I must again draw attention to the curious differential 
growth in the anterior part of the floor of the thalamen- 
cephalon, which leads to a migration forwards and upwards 
of the anterior commissure, so that this structure, which at 
first lay on the floor of the thalamencephalon, comes to be 
situated well up on its anterior wall. The anterior com- 
missure by the early appearance of its rudiment performs 
an important service, for it acts as a landmark to indicate 
the position of the anterior end of the longitudinal axis of 
the base of the original brain rudiment, which hes just in 
front of it. From this follows the important consequence 
that the front termination of this original axis lies in the 
adult Lepidosiren, by no means at the tip of the infun- 
dibulum, but high up on the anterior wall of the thalamen- 
cephalon. 

It will have excited attention that I have, in treating 
of sagittal sections, said nothing about the velum. The 
velum is in fact a paired structure. In accurately mesial 
sections the anterior wall of the thalamencephalon is as 
shown in the figures, but upon either side of the mesial plane 


In this I entirely agree with Minot, op. ecit., p. 95. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 437 


and just posterior to the paraphysis it becomes invaginated 
into the third ventricle, forming a definite complicated velum, 
continuous on each side with the choroid ingrowth of the 


hemisphere. The velum develops late, but is present in 
Stage 39. 


The Hemispheres. 


The high development of the cerebral hemispheres, and in 
particular of the pallium, forms one of the most interesting 
features of the brain of Dipnoans. At their first appearance 
the hemispheres form paired separate bulgings of the lateral 
walls of the thalamencephalon. There is no forward growth 
of the anterior median portion of the wall. The hemispheres 
soon begin to grow up dorsalwards on either side of the 
thalamencephalon. Later on there sets in marked forward 
growth, and the immense antero-posterior length of the 
hemispheres becomes in the adult one of their most con- 
spicuous features. 

As the hemispheres grow forwards their outer and inner 
walls become equally thickened, except in their posterior 
portions. Here, where the mner wall of the hemisphere 
faces the thalamencephalon, it remains relatively thin. About 
Stage 55 a small rounded portion of this thin part of the wall 
bulges into the lateral ventricle. This rounded projection 
contains a vascular loop: it is the rudiment of the lateral 
plexus of the hemisphere. ‘The rudiments of the two lateral 
plexuses are thus seen to be perfectly separate bilaterally 
symmetrical structures. The plexus rudiment at once begins 
to grow forwards into the cavity of the lateral ventricle. By 
Stage 38 it has grown about half the length of the ventricle, 
and has become swollen out into an irregular crumpled lamina. 
Finally, in Stage 39 (cf. Text-fig. 2, H), the plexus has 
become greatly complicated, and fills almost the whole of 
the ventricular cavity. 


438 J. GRAHAM KERR, 


Mesencephalon. 


Of the mid-brain there is little to be said. The roof 
remains throughout development thin in the mesial plane, 
except anteriorly, where it thickens out as it approaches the 
posterior commissure. 

The floor of the mid-brain shows anteriorly very distinctly 
the two median pits which it has been suggested indicate the 
two anterior segments of the mid-brain. ‘These depressions 
are seen in the series of sagittal sections (‘Text-fig. 2, F—H) : 
they first appear about Stage 35 as shght crinklings of the 
posterior wall of the infundibular depression. 

Burckhardt points out, in regard to Protopterus, that 
the floor of the mid-brain is bounded posteriorly by mpush- 
ings of the two limiting membranes. The dorsal one of these 
becomes very sharply marked in Lepidosiren about Stage 
35. It increases in depth for a time, but then flattens out 
again, and in Stage 39 has become replaced by a slight 
elevation of the surface. There appears to be no trace of the 
lower groove. 


Rhombencephalon. 


The cerebellum becomes evident about Stage 51 as a shght 
thickening of the roof of the fourth ventricle. 'The thicken- 
ing is much more conspicuous laterally than in the middle 
line. 

Behind this the roof of the ventricle at an early stage 
becomes thin and membranous. 


Olfactory Organ. 


The olfactory organ arises from a solid ingrowth of the 
deep layer of the ectoderm. 

Within the few days after hatching (Stage 29 +) the 
cavity appears in the rudiment first in the form of a split. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 439 


This gradually increases in size (cf. Text-fig. 3), and about 
Stage 32 communication with the exterior is established, 
giving rise to the anterior nares. 


Text-FIG. 3.—Transverse section near the anterior end of the head of a 
Lepidosiren of Stage 31. 4.c. Rudiment of buccal cavity (still solid). 
c.h. Cerebral hemispheres. me. Mesencephalon. o/f. Cavity of olfac- 
tory organ, not yet open to exterior. 


The posterior nares arise later about Stage 34, 


The Hye. 


The rudiments of the eyes begin to develop between Stages 
20 and 21 as two solid “outgrowths” from the sides of the 
brain rudiment (fig. 12, A). 

By Stage 23 a cavity has appeared in each optic process— 
the fore-brain itself remaining’ still solid (fig. 12, B). The 
cavity rapidly enlarges and becomes continuous with the 
cavity of the fore-brain ; and at the same time, by differential 
growth, the attachment of the optic stalk becomes carried 


4.4.0 J. GRAHAM KERR. 


downwards to the ventral side of the brain (fig. 12, C). The 
brain being also meanwhile greatly increasing in depth, the 
optic rudiment is'caused to slope upwards to reach its point 
of contact with the skin. 

Later still, with the active growth in length of the brain 
rudiment, the point of attachment of the optic stalk becomes 
carried far forwards from the level at which its end portion 
is, so to speak, tacked on to the skin. The optic nerve now 
passes from its origin backwards to a great extent before 
reaching the skin. 

Already in Stage 25 the wall of the optic vesicle lying 
next the skin begins to show a thickening to form the retina 
(fig. 12, D), and at the same time, or very slightly later (fig. 
12, F), a shght thickening in the external epidermis fore- 
shadows the formation of the lens. The retina rapidly 
thickens (fig. 12, G), assumes a convex surface internally, 
and by Stage 30 fits closely to the surface of the inner wall 
of the vesicle (fig. 12, H). 

The lens, at first a simple thickening of the deep layer of 
the epiblast, begins to develop a cavity in its interior about 
Stage 30 (fig. 12, H). It develops somewhat ventrally to 
the optic vesicle, and there is at first a wide space between it 
and the ventral rim of the cup. 

As usual the invagination takes place partly from below, 
and in Stage 30 there is a wide choroid fissure, involving, 
however, only the retinal cup. The fissure closes remark- 
ably soon in Lepidosiren. Already by Stage 31 it has 
been obliterated. 

In embryos a little older than Stage 31 I find traces only 
persisting ; in one case a depression in the rim of the cup, 
in another mesoblast perforating its wall. The choroid fissure 
being of such a transitory nature is by no means so con- 
spicuous a feature as we are accustomed to in the higher 
vertebrates. 

The development of the lens from Stage 31 onwards seems 
to pursue a normal course, which is sufficiently indicated by 
fig. 12, I, J, etc. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA, 441 


Opie Nerve, 


Lepidosiren is not well suited for a study of the his- 
togeny of the optic nerve on account of its very slender 
character. 


Structure of the Wall of the Optic Cup. 


It will be convenient, before describing the histogenesis of 
the retina, to give a short account of its structure in a young 
Lepidosiren 78 mm. in length. 

Near its centre the retina, including the pigment layer, 

measures about 125 « in thickness. 
The retina consists of three definite layers of cell elements 
a layer of rod elements or percipient cells, a middle mass 
of nerve-cells, and a single layer of ganglion cells next the 
vitreous humour. These cellular layers are separated as 
usual by two “molecular” layers composed of the extremely 
fine ultimate ramifications of the cell bodies. On its inner 
surface the ganglion layer is covered by the thin layer of 
fibres going to the optic nerve. 


Layer of Visual Cells. 


Drawings of rod elements highly magnified are given in 
fie. 11, M and N. 

Each cell is cylindrical in shape; at its inner end it is 
continued by irregular processes into the outer molecular 
layer, while its outer end bears the rod. Hach rod is 
cylindrical, or rather very slightly conical. It measures from 
‘011 to ‘015 mm. in length, i.e. it is much shorter than the 
corresponding structures in Amphibia. At its outer end the 
rod is rounded. The substance of the rod shows a division 
into alternate dim and clear layers, the latter beimg the 
narrower. I have not observed any longitudinal mark- 
ings. 

The above measurements refer to a specimen killed during 
daylight. Rods from the corresponding part of the retina in a 


44,2 J. GRAHAM KERR. 


young Lepidosiren of the same brood, kept under identical 
conditions, but killed at 9 p.m., measured from ‘018 mm. to 
‘028 mm. in length, the thickness beimg correspondingly 
diminished. The difference is seen on comparing fig. 11, M 
and N. 

The rod rests on the outer end of the visual cell. The 
inner moiety of the cell is almost completely occupied by the 
ellipsoidal nucleus. This has a very definite appearance, 
which marks it off from the nuclei of the other layers, 
inasmuch as the chromatin is collected in large rounded 
masses, so as to give the nucleus a very coarsely granular 
appearance. The chromatin network, of which the rounded 
masses are the greatly thickened nodes, extends equally all 
through the interior of the nucleus. 

The end of the cell next the rod is occupied by a large 
spherical vacuole containing a clear fluid. The protoplasm 
bounding this vacuole is very dense, and stains deeply with 
carmine, forming what looks lke a distinct membrane with 
irregular thickenings. 

As is not very clearly shown by fig. 11, there is associated 
with the elongation of the rod in darkness a remarkable 
diminution in the size of the vacuole—a diminution so striking 
as to suggest the possibility of there being an intimate 
mechanical connection between the two phenomena, the 
rod being elongated by the passage into it of fluid from the 
vacuole. 

Sometimes the wall of the vacuole is in direct contact 
with the nucleus, at other times the two structures do not 
touch. 

Between the two is a meniscus-shaped second vacuole, 
differing from that previously described by its contents not 
staiming black with osmic acid. 

The protoplasm of the cell-body forms a thin cylindrical 
shell round vacuoles and nucleus, while beneath the nucleus 
it thickens out into a more considerable mass of protoplasm, 
irregular in shape and giving off prolongations into the outer 
molecular layer. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 445 


The membrana limitans externa lies immediately external 
to the visual cell-nuclei. Some of these nuclei do not reach 
the membrane, others are in contact with it and even bulge 
it outwards. Occasionally one is also able to make out that 
the membrane passes up round the principal vacuole (cf. 
fig. 11). 

There is no differentiation of the fully formed retinal 
elements into rods and cones. The structure figured by 
Schiefferdecker for Ceratodus (‘Arch. mk. Anat., Bd. 
xxvill, Taf. xxiv, fig. 80) is, I have no doubt, a young rod- 
cell which has not yet fully developed its rod. 


Pigment Layer.—The pigment layer of the retina con- 
sists of a single stratum of cells, polygonal, very usually 
pentagonal, in outline as seen in surface view. 

The inner portion of the cell-body is laden with melanin 
eranules, and during the exposure to light this is extended 
in a multitude of fine thread-like processes which pass up 
between the rods and rod vacuoles, many of them reaching 
to the membrana limitans externa. 

In a specimen of the same brood killed by faint lamplight 
during the night the processes were somewhat shorter, more 
clumped together, and more pigment was visible in the 
general protoplasm of the cell-body. 

I shall now proceed to describe the features which I have 
been able to make out in the histogenesis of the retina, more 
especially in regard to the development of the layer of visual 
cells.” 


Histological Evolution of the Optic Cup. 


Histological differentiation of the optic cup becomes evident 
about Stage 31. At this stage pigment granules are seen in 
the posterior wall of the cup, appearing in each cell along the 

' Cf. for frog, Bernard, this Journal, vol. 43, page 30, 

* For the most modern contribution to this subject see Levi, ‘ Osservazioni 


sullo svilluppo dei coni e bastoncii della Retine degli Urodeli,” ‘Lo Speri- 
mentale,’ anno liv, 1900. 


444. J. GRAHAM KERR. 


face turned towards the lens, as well as in the part of each 
cell which adjoins its neighbour. The nucleus occupies the 
greater part of the body of each cell, and the part of the cell 
behind it is free from pigment. 

In the retina itself the commencement of differentiation is 
seen in the fact that its central and inner portion is becoming 
free from nuclei, the cell-bodies in this region being composed 
of clear transparent protoplasm. 

The development of the rods is the last important thing to 
take place in the development of the retina. The precise 
period of their appearance is, however, extraordinarily vari- 
able. While as arulethey do not appear till about Stage 35, 
I have met with them at least partially developed as early as 
Stage 32. 

The appearance of the rod is heralded by the development 
here and there in the protoplasm of a visual cell of a small 
spherical droplet apparently of a fatty nature, staining black 
with osmic acid. Asa rule these appear about Stage 35. 

In the case of a visual cell which is about to develop a rod, 
the fatty droplet just mentioned increases in size, and comes 
into close contact with the external limiting membrane. 
As the droplet increases in size the membrana limitans be- 
comes pushed out over it to form a hemispherical bulging. 
This bulges out more and more, until eventually it forms a 
deep pocket in which hes the fatty vacuole and a small 
quantity of protoplasm round it. As the vacuole continues 
to increase in size after it has passed into the pocket the 
latter often assumes a pear shape, being narrower at its base 
of attachment to the hmiting membrane. This is shown in 
fig. 11, J to L. 

The nucleus may project more or less into the pocket, and 
its peripheral end squeezing against the vacuole often 
pushes in its imner wall, so as to convert the sphere into a 
concavo-convex lens. 

The rudiment of the rod appears as a conical projection of 
protoplasm beyond the vacuole. This condition is illustrated 
by fig. 1,J. The cuticular boundary of the rod rudi- 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 445 


ment (i.e. the limitmg membrane covering it) becomes less 
and less distinct, and the mass of protoplasm within becomes 
clear and transparent, and is converted bodily into the 
rod. Very soon traces of alternate dim and clear zones 
appear (fig. L). At first these are only two or three in 
number, but they increase in number as the rod increases in 
length. 

In the completely developed rod, as already shown, the 
alternating dim and clear discs are very numerous, and the 
limiting membrane is no longer to be detected ensheathing 
the rod. It is, I believe, quite correct to say that now the 
whole of the protoplasm at the peripheral end of the visual 
cell has become converted into a cuticular product, the rod. 

The rods develop first in the region of the axis of the eye, 
from which region their development gradually spreads 
outwards, 


Auditory Organ. 


The auditory organ appears as a solid ingrowth of the 
deep layer of epiblast. A split is already present almost 
from the first in its interior, e. g. Stage 21, and about Stage 
23 this dilates at its inner end into a definite cavity. 


The Peripheral Nerves. 


The development of the peripheral portions of the nervous 
system is of such interest and importance that I propose to 
devote a separate paper to its description. I may here simply 
summarise the chief facts which I have observed in regard 
to it in Lepidosiren. 

The first and the all-important fact which has to be men- 
tioned is that I find that the motor nerve connection between 
neural rudiment and myotome already exists at a time when 
these structures are in close apposition. Where the myotome 
and spinal cord remain in contact in the section it is of course 
practically impossible to demonstrate the existence of the 
nervous bridge between them. If, however, such an embryo 
be examined as that figured on PI. 10, fig. 24, of the first part 


446 J. GRAHAM KERR. 


of this work!—which has been excised from the egg and laid 
out in one plane while still alive, and then treated with the 
fixing agent, it is found that the myotomes have been pulled 
shghtly away from the neural cord, and now the nerve-root 
is quite visible (cf. fig. 12 a) as a pale strand passing from 
the neural cord near its ventro-lateral angle, outwards to 
the inner surface of the myotome. At this period of develop- 
ment (Stage 24) there are scarcely any mesenchyma cells 
between myotome and spinal cord. Where occasionally such 
a cell appears it is heavily laden with yolk granules, and is 
by that easily distinguishable from the pale nerve-root. 

As development progresses the myotome recedes from the 
neural canal with the development of mesenchymatous con- 
nective tissue between the two, and the nerve-trunk is corre- 
spondingly stretched out, or, to put it more correctly, it grows 
in length accordingly. 

By Stage 27 (fig. 12 B) the nerve-trunk shows a very dis- 
tinct fibrillation, and at its outer end its fibrils radiate out in 
a beautiful cone-like manner over the inner face of the myo- 
tome. In the middle of the cone the fibres are seen to pass 
directly into the protoplasm, and the whole arrangement of 
nervous tails to the muscle-cells, the latter having at this 
stage developed contractile fibrils in the protoplasm adjoining 
their three surfaces, dorsal, ventral, and external, irresistibly 
calls to mind the nerve-muscle arrangements of the nematode 
worms. Outside the conical arrangement of fibres mentioned 
I have not been able to demonstrate the connection of muscle- 
cell with nerve-root, though of course there can be little 
doubt that a similar relation holds. In all probability every 
muscle-cell in the myotome passes at its inner end into a 
nervous tail, but these are only conspicuous where they pass 
away from the inner surface of the myotome at a considerable 
angle; over the greater part of the surface they are lying 
closely apposed to it, and are hence difficult to see. 

As the mesenchymatous reticulum increases in quantity 
between the myotome and the spinal cord its protoplasm (and 


1¢ Phil. Trans.,’ vol. exci, }. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 447 


nuclei) tends to concentrate round the nerve-trunk, so that 
the latter now (fig. 12 c) hes embedded in a thick strand 
of the reticulum. The protoplasm forming this contains yolk 
granules, and doubtless serves to supply the nerve-trunk with 
nourishment. ‘The nerve-trunk is in need of this, for it is 
growing actively, both in length—as the muscle-cells are 
and 


pushed farther and farther away from the spinal cord 
also in thickness. Whereas at Stage 24 the nerve-trunk 
measured about 1 «in diameter, now it measures about 5 p. As 
the fibrils composing it do not appear to have undergone a 
corresponding thickening they must have increased in 
number. As to how this increase takes place, whether by 
splitting of original fibres or by formation of new fibres by 
the surrounding protoplasm, it is obviously impossible to 
make definite assertions, based merely on observation, which 
shall be at all rehable. Theory, of course, would favour the 
first view. 

As regards the sensory nerves, I do not propose to say 
anything at present, except that my evidence, so far as it 
goes, points to exactly similar processes taking part in their 
development as I have just outlined in regard to the motor 
nerves. 


GENERAL REMARKS. 


Effect of Light upon the Chromatophores. 


It will be possible from the present paper to form a clearer 
idea of the process of change in the chromatophores to which 
T alluded in my description of the external features. I still 
hold to the view that we have to deal with an actual 
retraction of pseudopodia, and not a mere movement of 
pigment granules into the body of the cell as held by 
Ballowitz. Stray pigment granules may be observed fre- 
quently remaining where the pseudopodia have been. I take 
these not as indicating that the cell processes are still pro- 
truded, but as having been left behind by the pseudopodia 

vot. 46, PARY 3,—NEW SERIES. u E 


448 J. GRAHAM KERR. 


during their retraction, after the manner familiar to any one 
who has watched living Foraminifera. 

Circumstances, unfortunately, did not allow of my making 
any experiments as to whether a nervous reflex plays any 
part in the contraction of the chromatophores. I am inclined 
to think, however, that we have to do with a direct action of 
heht upon the individual cell. 

I wish to draw attention to one point; that there is a 
fundamental similarity in the reaction to light stimulus of the 
general chromatophores of the body and the cells of the 
pigment layer of the retina. In both we have to do with 
relatively slightly differentiated cells, and in both their 
reaction to light (in addition probably to the formation of the 
pigment itself) consists in their pushing out positively helio- 
tropic pseudopodia. 


General Morphology of the Fore-brain Region. 


It is fashionable to accept the view that the “secondary 
fore-brain ” is fundamentally an unpaired structure, forming 
topographically the anterior end of the cerebro-spinal axis of 
the adult—that it is a “telencephalon.” 

Tt will be seen that the conditions holding in the ontogeny 
of Lepidosiren go completely against this view. On the 
contrary, they support the opposite and older view, the view 
held by von Baer, Reichert, Goethe, and in recent times 
including as one of its chief exponents Studni¢ka, that the 
hemispheres are fundamentally paired bulgings outwards of 
the lateral wall of the thalamencephalon, of a nature roughly 
comparable with the outward bulgings which give rise to the 
eyes. The portion of brain from which they project out- 
wards is simply the anterior portion of the original fore- 
brain,’ which, as has been shown, becomes marked off at a 
very early period of development by an upward bend in its 
floor,—that is to say, it is simply and solely thalamence- 
phalon. 

1 Vorbirn of Kupffer. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 449 


Studnicka has recently asserted that “Der Vorderhirn 
aller Cranioten ist paarig angelegt.”! Lepidosiren fur- 
nishes nothing to throw doubt on this generalisation, and the 
general tendency of modern research, especially upon forms 
in which the topographical relations are not lable to be 
distorted by the presence of a large mass of yolk with its 
accompanying exaggerated cerebral flexure, seems on the 
whole to support it. If true it ought to involve the rejection 
of His’s brain regions telencephalon and diencephalon, as not 
representing true morphological entities. The old expressions 
thalamencephalon and prosencephalon are much more accu- 
rate, but it seems questionable whether it would not be 
better to drop the latter name and merely use the descriptive 
term “hemispheres,” as this conveys no erroneous suggestion 
that we have to do with anything more than a local hyper- 
trophy of the thalamencephalon wall upon each side. 

The enormous importance of the primitive fore-brain—that 
brain region which assumes its definitive characters so extra- 
ordinarily early in development, and which appears to be 
already marked off in Amp hioxus,’ can hardly be exag- 
gerated. Compared with it the hemispheres, however im- 
portant they may be physiologically, are morphologically of 
relatively little account. 

It appears to me, biassed possibly by my observations of 
what happens in Lepidosiren, that too much morphological 
weight is attached to the velum. It is looked upon as being 
the landmark which indicates in the brain roof the boundary 
between two important brain regions, and structures lying 
anterior to it are regarded as belonging to the secondary 
fore-brain. In Lepidosiren we have seen that the velum is 
a paired structure, also that it arises simply as a prolongation 
backwards from each lateral plexus. ‘The evidence of Lepi- 
dosiren pointing towards the secondary fore-brain being 
prinitively paired and to the lateral plexus being similarly 


1 «S. B. Bohmisch. Ges.,’ 1896, xv. 
* Kupfler, ‘Studien zur vergl. Entwicklungsgeschichte des Kopfes der 
Kranioten,’ Heft 1, 8. 71. 


450 J. GRAHAM KERR. 


paired structures, it seems probable that in the paired nature 
of the velum we have also to do with a primitive character. 

It is also seen from the study of sagittal sections through 
the brain at succeeding stages of development that the 
unpaired portion of the brain lying anterior to the level at 
which the velum will later appear is nothing more than the 
anterior portion of the thalamencephalon. It follows from 
this that the paraphysis does not belong to the secondary 
fore-brain at all;! it is a projection of the anterior wall of 
the original thalamencephalon. 

Regarding speculations as to the original nature of the 
paraphysis, Lepidosiren seems to offer little definite evi- 
dence. On the whole, however, its topographical relations in 
Urodeles and Dipnoans suggest that it is a part of the brain 
wall which has been drawn out by the interposition of 
mesenchyme between skin and brain, through its outer end 
being attached to the skin; in other words, that it probably 
represents either an ancient sense-organ (Selenka) or a part 
of the primitive connection of the brain with the outer skin. 

There has been considerable discussion as to the nature of 
the smaller, more ventrally situated of the two vesicles 
found in the pineal region of Petromyzon. The relations 
of the paraphysis of Lepidosiren, about whose homology 
there can be, I imagine, no doubt, support the view ex- 
pressed by Kupffer? that the vesicle referred to is also 
morphologically the paraphysis. If this homology hold, the 
discovery by Retzius of nerve-fibres in the structure in 
question in Petromyzon would, as Gaupp points out, sup- 
port the idea of the at one time sensory nature of the para- 
physis. 


The Peripheral Nerves. 


It will, I think, be admitted that the facts which I have 
shortly sketched in regard to the development of the motor 


1 Gaupp, op. cit., p. 229. 
2 Op. cit., Heft 2,5. 10. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 451 


nerves have considerable bearing upon some not unimportant 
questions of general morphology. 

1. I have shown that those muscle-cells of the myotome 
whose investigation is easy are already in organic connection 
with the nervous system while they lie in close apposition to 
it. As these muscle-cells are carried further and further away 
from the nerve-centre by the interposition of mesenchyme, 
and by the action of differential growth, they trail behind 
them the ever-lengthening nervous strand,! and there is 
therefore no question about calling in—so far as regards 
these particular muscle-cells—either a growth outwards 
towards them of freely ending nerve-trunks or their pro- 
vision with nerves through the conversion into nerve of 
other tissue. They are connected while still close to the 
nerve-centre with a potentially nervous bridge, and as 
development goes on this merely increases in length and 
thickness. 

That this holds I have been able to definitely establish 
only in the case of certain muscle-cells in the region of the 
myotome nearest the nerve-root, but it is extremely unlikel y 
that motor nerves do not develop all according to the same 
general plan. They probably all develop just as do those 
which I have described. 

2. I have shown that in the young Lepidosiren there 
exists in the myotome a neuro-muscular apparatus of an 
extraordinarily primitive character—a simple though enlarged 
epithelial cell with contractile fibres developed in the peri- 
pheral regions of its protoplasm, and with its cell substance 
passing out at its inner end into a kind of tail which is 
continued along the nerve rudiment, doubtless to pass— 
though this would be very difficult to demonstrate—at its 
central end into a nerve-cell in the neural tube. 

The full consideration of muscle development will come 
in a later part of my work, but even in passing I think we 

1 This of course explains quite clearly the meaning of the rectilinear course 


of young nerves, upon which His and others lay stress. Cf. ‘Arch. Anat. 
Phiys;; 1887, 5.070. 


452 J. GRAHAM KERR. 


must recognise in the transient phase which I have described 
the indication of a far back phylogenetic stage in the history 
of vertebrate muscle. 

The facts which I have summarised above in regard to the 
development of nerve and muscle in Lepidosiren support, 
on the whole, the theoretical views of Hensen in regard to 
the development of nerve, and are totally opposed to the 
views of Bidder and Kupffer and of His that the nerves 
grow out towards their motor end structures, and unite with 
them secondarily. They also lend additional support to the 
view which Sedgwick has consistently taught, and which has 
been supported by much recent work, such as above all that 
of Apathy upon the structure of the nervous system, that we 
must regard the vertebrate individual in the first place as an 
organic and continuous whole, whose various regions are 
linked up in organic continuity, rather than as an aggregation 
of separate cells or organs. 


SuMMARY OF THE MORE ImporTANT FACTS DESCRIBED. 


1. Certain of the ectoderm cells are provided with tail-like 
processes which pass into a subepidermal layer, into which 
also pass processes of the underlying mesenchyme cells. 

2. The flask-shaped glands of the skin arise as solid thick- 
enings of the deep layer of the epidermis which develop a 
lumen later. 

3. The cement organ arises as a thickening of the deep layer 
of the ectoderm over which the superficial layer degenerates 
and disappears. Its atrophy is caused mainly by the action 
of phagocytes. 

4. The chromatophores of the skin are mesodermal in 
origin. 

5. There is no invagination of ectoderm to form a true 
stomodzum. The epithelium of the buccal cavity is developed 
in situ from the outer layer of the solid anterior portion of 
the yolk-laden enteric rudiment. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 403 


6. The tooth-plates of Lepidosiren are not at any stage 
represented by numerous separate denticles as is the case 
with Ceratodus. 

7. The tooth germs appear while the mouth region is still 
without a lumen. They develop according to the “ Placoid” 
type. 

8. There is a well-developed enamel organ, and a layer of 
what appears to be a peculiar type of enamel is the first of 
the hard parts of the tooth to be laid down. 

9. The dentine appears to arise by metamorphosis of the 
peripheral portions of the cell-substance of the odonto- 
blasts. 

10. The brain of the adult Lepidosiren closely resembles 
that of Protopterus, differing only in details. The fourth 
and sixth cranial nerves are present though extremely thin. 

11. The thalamencephalon and mesencephalon do not 
become marked off from one another until relatively late. 

12. The cerebral hemispheres arise as two separate lateral 
bulgings of the wall of the thalamencephalon. 

13. The choroid plexuses of the lateral ventricles are simi- 
larly paired originally; the plexus of the third ventricle 
arises from them and is also paired. There is no velum in the 
middle line. 

14, The pineal organ is simple, without any trace of a sepa- 
ration of a part of it to form a “ parietal organ.” 

15. There is a well-developed paraphysis closely resembling 
that of Urodele Amphibians. 

16. The paraphysis is a product of the thalamencephalon, 
not of the secondary fore-brain. 

17. The point in the brain of the adult which corresponds 
to the anterior end of the floor of the neural rudiment of the 
embryo, hes well up on the anterior wall of the thalamen- 
cephalon just below the root of the paraphysis. 

18. The rudiments of the eyes, like those of nose and ear, 
are at first solid, the cavity of the optic vesicle arising 
secondarily, rather before the ventricle of the brain at its 
level has appeared. 


ADA J. GRAHAM KERR. 


19. The lens also arises by a solid rudiment from the deep 
layer of the ectoderm, its cavity arising secondarily. 

20. In the histogenesis of the visual layer of the retina the 
development of the rods is preceded by the formation of the 
oil globules. 

21. Each oil globule passes with surrounding protoplasm 
into a pocket-lke projection of the external limiting mem- 
brane. 

22. The rod is formed directly from the protoplasm lying 
in the conical apical part of the pocket. 

23. The motor nerve-trunks are in Lepidosiren already 
laid down at a period when myotome and neural tube are still 
in close apposition. As development proceeds and the myo- 
tome recedes from the spinal cord the nerve-trunk lengthens 
out, increases in thickness, and becomes ensheathed in mesen- 
chymatous protoplasm. 

24. In the earliest stage observed the motor nerve-trunk is 
seen to be perfectly continuous with the protoplasm of the, at 
this stage, simple epithelial muscle-cell of the myotome. 


EXPLANATION OF PLATES 25—28, 


Illustrating Professor Graham Kerr’s paper on “ The Deve- 
lopment of Lepidosiren paradoxa,” Par6 ITI. 


The figures of sections have been drawn with the Abbe camera under Zeiss 
objectives a*, a, D, 8 mm. dry apochromatic, and 1 in. homogeneous immer- 
sion. The figures of the entire brain have been done either from actual 
dissections, or from reconstructions by the method described by me in an 
earlier number of this Journal.! 

By Stage n in the descriptions of the following figures I mean the stage re- 
presented by fig. n in Part 1.2 Stages slightly younger or older than the 
fizured are described as n- or n+. 

1 «Quart. Journ. Mier. Sci.,’ vol. xlv, p. 5. 
2 «Phil. Trans.,’ vol. excii, p. 299. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 4055 


Figures of a selected number of stages were copied in Part II (this 
specimen Journal, op. cit.). 


T have again to express my great indebtedness to Mr. Edwin Wilson for the 
eare which he has devoted to working up the figures. 


GENERAL List oF Rererence LEtTTErs. 


a.d. Archinephric duct. aud. Auditory vesicle or capsule. @.v. Annular 
vacuole. 4. Bony trabecula. d.c. Buccal cavity. 4.0. Blood-vessel. due. 
Solid yolk-laden cells occupying position of buccal cavity. ¢,. Chromatophore 
of Type A. cy. Chromatophore of Type B. ¢.4. Cerebral hemisphere.  d. 
Dentine. dg. Dorsal root ganglion. e¢. Enamel. e.' Superficial layer of 
ectoderm. ¢%. Deep layer of ectoderm. e.o. Enamel organ. ep. Pineal body. 
J. Fat globule. 7. Fore-limb. g.c. Unicellular glands. g.é. Flask-shaped 
gland. hyp. Hypophysis cerebri. 7/2. Lobus inferioris. m.c. Mesenchyme 
cells. mes. Mesencephalon. m./. Membrana limitans. m.a7. Motor netve- 
trunk. my. Myotome. 2. Notochord. x.a. Neural arch. x.c. Neural 
tube. xz. Nucleus. 0. Odontoblast. o.e. Hye rudiment. og. Oil globule. 
olf. Olfactory capsule. p. Pocket. ph. Phagocyte. 7. Rod. sed. Sclerotome. 
stom. Lining epithelium of buccal cavity arising in situ. ¢h. Thalamen- 
cephalon. v.IV. Fourth ventricle. 

The cranial nerves are indicated by the ordinary Roman numerals. In 
addition :—A.m. Hyomandibular branch. /a/. Lateral branch. at. gang. 
Lateral ganglion. 7.c. Communicating branch. s.p. Superior palatine branch. 
vise. gang. Visceral garglion. 


PLATE 25. 


Fre. 1.—Example of tailed cells from the ectoderm of a young Lepido- 
siren of Stage 835-—. Zeiss 3 mm. apochrom., oc. 2. 


Fic. 2—Section of skin from the head of an adult Lepidosiren to show 
the large unicellular glands (g.c.). Zeiss A, oc. 2. 


Fie. 3.—Illustrating the history of the cement organ. Zeiss A, oc. 2. 
A. Stage 23, showing simple thickening of the deep layer of the ecto- 
derm. 
B. Stage 25, showing the disappearance of the superficial layer over 
the gland rudiment. 


456 J. GRAHAM KERR. 


c. Stage 31, the gland at its full development. 
p. Stage 35, illustrating the atrophy of the gland. 4.v. Blood-vessel ; 
J. Fat globules; ph. Phagocytes. 


Fic. 4.—Portions of skin stripped off two Lepidosirens of Stage 38 
(belonging to the same brood). Zeiss A, oc. 4. 
A. Killed at 2 p.m. in daylight. 
B. Killed at 9 p.m. in faint Jamp-light. 
c., Chromatophores of Type A; ¢.. Chromatophores of Type B; 
g.c. Unicellular gland; g/. Multicellular gland. 


Fic. 5.—Vertical sections through the skin of young Lepidosirens 
killed under different conditions of light and darkness. Zeiss D, oe. 2. 
A. From white enamelled vessel in bright daylight. 
B. From deep shade under a verandah. 
c. Dropped into the fixing agent at 9 p.m. in darkness. 
[Norz.—In a the epidermis has become pushed upwards by the action 
of the acetic acid in the fixing solution. | 


Fie. 6.—Illustrating the formation of the lining of the buccal cavity. 


A. Part of a sagittal section of a Lepidosiren of Stage 30. Zeiss D, 
oc. 2. 
B. Part of a similar section of an embryo of Amblystoma, measuring 
75 mm. in length. 
buc. Still solid, yolk-laden cell-mass in position where the 
buccal cavity will develop later; e!. Superficial layer of 
ectoderm; e®. Deep layer of ectoderm; som. Lining epi- 
thelium of buccal cavity arising in situ. 


PLATE 26. 


Fie. 7.—Illustrating the development of the teeth. 


A. Portion of a section parallel to the sagittal plane at Stage 31, 
showing one of the palato-pterygoid tooth germs. The buccal 
cavity (4.c.) is beginning to form, but it is still closed anteriorly 
(to the right in the figure). Zeiss D, oe. 2. 

B. Portion of a similar section at Stage 35. Zeiss D, oc. 2. The figure 
shows the enamel cap (¢.) and the commencing formation of 
dentine (d.). ¢.o. Enamel organ; 0. Odontoblast. (‘The split 
round the enamel is, as shown by other sections, artificial.) 


TIE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 457 


c. Part of a transverse section of Stage 35, cutting the palato-pterygoid 
tooth germs longitudinally. Letters as before, and .v. Blood- 
vessel; 6. Bony trabecule. Zeiss A, oc. 4. 


p. Part of the preceding section under a higher power (Zeiss 3 mm. 
apochrom., oc. 2) to show the minute structure of the dentine 
which is being formed, and also the distinctness of the enamel 
layer. 

E. Part of section similar to A andB, but from Stage 36. Zeiss A, oc. 2. 

F. Part of similar section from Stage 38. Zeiss A, oc. 2. 


Fre. $8.—Illustrating the brain of Lepidosiren as seen from the dorsal 
side. 
a. Stage 31+. Reconstruction Zeiss a*,oc.2. The thalamencephalon 
and mesencephalon are not yet marked off from one another. 
B. Stage - 35. Reconstruction Zeiss a*, oc. 2. 
c. Stage 38. Dissection. 
p. Adult. 


aud. Auditory vesicle or capsule. ch. Cerebral hemispheres. 
d.g. Dorsal root ganglion. ep. Pineal body. 2. Fore-limb. 
mes. Mesencephalon. z.a. Neural arch. o/f. Olfactory capsule. 
th. Thalamencephalon (the line stops over the left habenular 
ganglion). v.LV. Roof of fourth ventricle. 


Craniat Nerves.—I. Olfactory. IL. Optic. ILI. Oculo-motor. 
IV. Pathetic. V. Trigeminal. V.! Ophthalmic division, 
V.? Maxillary division. V.3 Mandibular division. VII. Facial : 
VU, 4m. Hyomandibular; VII, /aé. Lateral; VII, s.p. Supe- 
rior palatine; VIL and X, 7.c. Commissure connecting lateral 
portions of VII and X. VIII. Auditory. 1X. Glosso-pharyn- 
geal. X. Vagus: X, /a¢. Lateral trunk ; X, /a¢. gang. Lateral 
ganglion; X, vise. gang. Visceral ganglion. 


Fie. 9.— View of brain of adult from below. 
hyp. Hypophysis. 7.7. Inferior lobe. VI. Abducent. 


PLATE 27. 


Fic. 10.—The brain of Lepidosiren in side view. A, B, C, D, B, represent 
reconstructions; F, G, H, I, dissections. 
a. Stage 26. 
B. Stage 29+. 
c. Stage 30. 


458 J. GRAHAM KERR. 


D. Stage 31. 
gE. Stage 31+. 
r. Stage —35. 
Gc. Stage 38. 
H. Stage 39. 
1. Adult. 
oc. Kye rudiment. Other letters as in Figs. 8 and 9. 


Fig. 11.—Illustrating the development of the eye. All the figures illus- 
trating the development of the eye are taken from sections transverse to the 
long axis of the head. 

A. Stage 21. Zeiss A, oc. 2. 
th. Still solid thalamencephalon ; oc. Hye rudiment. 
B. Stage 23. Zeiss A, oc. 2. A cavity is now arising in the optic 
rudiment ; the brain is still solid in this region. 


c. Stage 25. Zeiss A, oc. 4. Thalamencephalon and optic vesicle are 
now hollow. The retinal wall of the optic vesicle is beginning to 
thicken. 

p. Stage 26. Zeiss A, oc. 2. Showing advanced thickening of 
retina. 

E. Stage 30. Zeiss A, oc. 4. Optic cup—lens rudiment with com- 
mencing cavity (/.c.). 

F. Stage 31. . Zeiss D, oc. 2. 

G. Stage 35. Zeiss A, oc. 4. 

H. Stage 38. Zeiss A, oc. 4. 


PLATE 28. 


Fic. 1], 1—n. Illustrating the development of the percipient elements. 

1. Stage 35. 3; hom. imm., comp. oc. 6. Showing the formation 
of the fatty globules and their passage into pocket-like projec- 
tions of the internal limiting membrane. 

og. Globule; p. Pocket; m./. Membrana limitans; xz. 
Nucleus of visual cell. 

J. Stage 35. 3, hom. imm., comp. oc. 6. Showing first rudiment of 
rod as a protoplasmic projection (7.). 

K. Stage 35. =; hom. imm., comp. oc. 6. Slightly later stage of 
rod development. 

L. Stage 35. 3}; hom. imm., comp. oc. 6. Rod showing first sigus 
of stratification. Appearance of annular vacuole (a.v.). 

M. Stage 38. 1; hom. imm., comp. oc. 6. A fully formed rod 
element, fixed in the light. 

n. A similar rod element from a young Lepidosiren of the same age 
and same brood, but fixed in the dark. 


THE DEVELOPMENT OF LEPIDOSIREN PARADOXA. 459 


Fig. 12,—Illustrating the development of the motor nerve-trunks. All 
three figures represent portions of sections perpendicular to the longitudinal 
axis of the body. 

a.d. Archinephric duct ; me. Mesenchyme cells ; m.z.7. Motor nerve- 
trunk; my. Myotome; z.c. Neural tube; x. Notochord; sed. 
Sclerotome. 

A. Stage 24. Zeiss A, oc. 4. Section of an embryo excised from the 
ege and flattened out, so as to slightly separate myotome and 
spinal cord, and display the rudiment of the nerve-trunk which 
even at this stage unites the two. 

B. Stage 27. Zeiss D, oc. 2. The myotome has become separated 
from the neural tube by interposed mesenchyme. The nerve- 
trunk is lengthened out, and externally is continued into the 
muscle-cells of tlhe myotome. 

c. Stage 29+. Zeiss D, oc. 2. The nerve-trunk is now ensheathed 
in mesenchymatous protoplasm containing nuclei and yolk 
granules, 


METAMORPHOSIS OF CORYSTES CASSIVELAUNUS. 461 


The Metamorphosis of Corystes Cassivelaunus 
(Pennant). 


By 


Robert Gurney, B.A.(Oxon.), F.Z.S. 


With Plates 29—31. 


Tue material upon which the following account of the 
development of Corystes is based has been derived entirely 
from tow-net collections, taken generally in or just outside 
Plymouth Sound, between February and May of the present 
year. Occurring first on February 21st, the larve were 
obtained in small numbers fairly regularly from that time 
onwards. Advanced stages occurred first on March 15th, 
and in considerable numbers. The tow-net collections of 
April 14th, taken in the West Channel, were remarkable for 
the extraordinary abundance of the zowa stage of Corystes. 
They abounded on that date to the almost complete exclusion 
of all other zovas, nearly all the specimens being at an 
advanced stage of development. Since April 14th the 
zovas have continued to occur, but very few have been 
found in the early stages. 

Zoewas taken in the tow-nets have been kept successfully 
in plunger jars for considerable lengths of time, and several 
have moulted into the Megalopa stage, the latter being kept 
through the next moult, either in a jar immersed in running 
water or in a beaker, in which the water was kept in motion 
by a stream of air bubbles. 


4.62 ROBERT GURNEY. 


Habits of the Larve. 


The zoxa of Corystes is probably the same as that 
described by Weldon (1889), and figured as “a Portunid 
zoea,”’ though it differs from his figure in several respects. 
Weldon came to the conclusion that the long spines have the 
function of directing the movements, and enabling the animal 
to swim rapidly in a straight line. There can be no doubt 
that his conclusion is correct, for the larva always moves in 
the direction of the long axis of the spines. This is well 
shown when a zoea is caught in an eddy in the plunger 
jar. It then makes a violent effort to escape, darting 
upwards, sideways, or straight to the bottom, according to 
the position of the spines at the moment. Normally the 
larva rises and falls in the water, swimming upwards ver- 
tically for a short distance, and then resting. In the 
plunger jar larve have a tendency to collect, especially the 
younger stages, at the surface against the side of the jar. 
Here they often push their dorsal spines through the surface 
film, and are hereby suspended. Sometimes they rest against 
the rod of the plunger, suspended in this manner, and to this 
habit I have owed the death of several specimens which were 
caught and drawn up upon the plunger rod. If a light is 
brought to the jar at night the zowas become extremely 
active, swimming rapidly towards the hght, the dorsal spine 
directed forwards. The body may, however, be in any 
position with regard to the axis of the spines, the animal 
swimming upon its back, side, or ventral surface. 

It is of interest to note, in considering the function of the 
spines, that the period of their presence corresponds exactly 
with the period of the absence of an “auditory” organ. 
The latter, as Prentiss (1901) has shown, is not developed 
functionally till the Megalopa stage, when only traces of the 
spines remain. There is a rapid reduction in the length of 
the spines as compared to the total length of the body in the 
third and fourth larval stages, and it is in the last zowa 


METAMORPHOSIS OF CORYSTES CASSIVELAUNUS. 463 


that the first trace of the auditory invagination is to be 
found. It seems likely, then, that the spines may perform to 
some extent the balancing and orienting function of the 
auditory sac. 

Besides being of balancing and directive function, the 
spines probably also serve as a protection. It is, of course, 
hard to say what are the especial enemies of the zowas in 
nature. They are certainly preyed upon to some extent by 
Medusze and Ctenophores, and also by each other, and in 
these cases the spines can be of little value. They must, 
however, serve as a defence against the attacks of small fish. 
That this isso was shown by presenting the zoza to Gobius 
ruthensparri, a fish about 4°5 em. long, and with a width 
of mouth of about 3°5 mm. In the first experiment a goby 
seized and rejected the zova six times, each time failing to 
swallow it. Finally the fish gave up the attempt, and the 
zowa soon recovered and swam away, being, however, 
attacked and swallowed by a second goby after a number of 
failures. Other experiments showed the same thing, though 
the fish had no difficulty in swallowing small Brachyurous 
zovwas without great development of spines—for instance 
Carcinus. 

A peculiar habit of the zovza at all stages is that of 
frequently turning its abdomen backwards till the forked 
telson reaches and embraces the dorsal spine, scraping 
upwards asif to clean it. This action is so frequent that it 
seems not to be connected with the process of moulting, 
though possibly the stretching of the abdomen entailed may 
assist in preparing for the act. Moulting seems usually to 
take place at night, and must be a rapid process, for a 
successful moult was never observed. Zowas were fre- 
quently found half freed from the larval skin, but these 
specimens never succeeded in completely freeing themselves. 
In fact, the new skin seems to harden so rapidly that unless 
the process is completed at once failure results. 

The Megalopa stage is remarkably interesting from the 
point of view of its habits. It has most of the characters 

voL. 46, PART 3.—NEW SERIES. FF 


464 ROBERT GURNEY. 


and habits of the adult. Like the Megalopa of the majority 
of Brachyura it is very active, swimming rapidly by means 
of its pleopods, the antenne being carried stretched straight 
forwards and parallel to one another, the thoracic legs bent 
up under the body. Unlike the zowa it seems to be in- 
different to a strong light at night, being neither attracted 
nor apparently repelled by it. 

It does not seem to be a pelagic form properly speaking, 
and was only once obtained in the tow-net, and then within a 
fathom or so of the bottom, in deep water. Some of the 
specimens moulted in my plunger jars were provided with 
fine sand, and at once burrowed until covered completely 
except for the antenne. The act of burrowing is performed 
just as in the adult by means of the four posterior legs, the 
chelipeds taking practically no part. If the sand, which 
must be exceedingly fine, is not deep enough to completely 
cover the body, the Megalopa pushes itself backwards till the 
sand is heaped up above it, often moving backwards in this 
way for some distance. When covered the antenne are not 
necessarily held quite parallel, but the position seems to a 
considerable extent to depend upon the depth to which the 
animal has burrowed. In the buried position the respiratory 
current is reversed, and sets down the antennal tube, as Mr. 
Garstang has shown it to be in the adult. The efficiency of 
the antennal tube as a strainer was well shown by the sand 
grains resting on the interlocking hairs. One specimen of 
the Megalopa, and one of the succeeding post-larval stage, 
were obtained in some sand from Whitsand Bay. 

As to the food of the earlier stages of the larva I have 
no direct observations to record. Though provided with a 
constant supply of the smaller plankton organisms, and with 
other small zozas, I have never seen the zowa of Corystes 
taking any food. From the appearance of the contents of 
the gut it seems likely that the food consists entirely of 
floating alge or diatoms. I have several times seen the 
zowas of other forms, such as Eupagurus, devourimg other 
larvee, but it is quite possible that the zowas of Corystes, at 


METAMORPHOSIS OF CORYSTES CASSIVELAUNUS. 465 


least in the earlier stages, are exclusively vegetable feeders. 
The Metazoza and Megalopa, on the other hand, have more 
than once been found eating zowas, and even those of their 
own kind, and I have found small pieces of worm or shrimp 
muscle a very satisfactory food for the Megalopa. 


Development of the Larva. 


There appear to be four distinct stages in the development 
of the larva preceding the Megalopa, but I am unable to 
say how many moults are included in this period. The 
zowas referred to the second stage differ among themselves 
to a certain extent in the relative development of the parts 
of the body, and this stage, consequently, is not very sharply 
separated from these preceding and following it. 

First Stage (Pl. 29, fig. 1)—The measurements of the 
zovwa at this period are as follows, the figures given being 
the average of ten specimens. I have given here, as also 
in the stages succeeding, the minimum and maximum for 
each measurement. 


Average Minimum and maximum 
length. lengths. 
1. Length of rostrum . 14mm... 1:3to0 15 mm. 
2. i, dorsal spine 1:9- cs. 1D 5.320 
Se Liprtostip okepines «nA sx. cores Ow 
4. Length of body . bs ADR aie” “we nah On ane Tere 
5. 43 thorax as ee ee ee ee oe 
6. Ratio of 4 to 3, 1: 1°66. 


The zowa of Corystes is distinguished from all other 
Brachyurous zozas hitherto described by the following 
features :—The total length from tip to tip of the dorsal and 
rostral spines greatly exceeds the total length of the body. 
In the majority of zowas the two measurements are 
approximately equal. he posterior edge of the carapace 
bears a fringe of short sete. The forks of the telson have 


4.66 ROBERT GURNEY. 


the normal form, but bear only a single! lateral spine on 
each side. The resemblance between the zoza of Corystes 
and that of ‘hia polita as figured by Claus (76) and Cano 
(91) is very striking, but the latter differs in having a much 
longer lateral thoracic spine, and in having two lateral spines 
on each fork of the telson. 

The colour and its distribution in the body are also charac- 
teristic. The long dorsal spine is a rich orange colour, 
deepest towards the tip. The rostrum has the same colour, 
but the chromatophores appear to be less numerous. The 
labrum contains a dendritic black chromatophore, and similar 
chromatophores are found in the carapace, one above the 
mandible, and two near the postero-ventral and postero- 
dorsal edge of the carapace. A small orange chromatophore 
lies at the base of the dorsal spine. The alimentary canal is 
enveloped in black chromatophores, which run back along it 
as far as the end of the second abdominal segment. A large 
ramified black chromatophore hes at the jomt between the 
third and fourth, fourth and fifth, and fifth and sixth abdo- 
minal segments. ‘To the naked eye the liver and gut appear 
as a yellowish-black mass continued back as a black streak 
through the thorax, and the orange colour of the spines is 
conspicuous and distinctive. 

The appendages of the zowa are of the usual type, 
differing in no important respects from those of Portunus, for 
example, and do not need detailed description. 

The second maxilla (Pl. 30, fig. 9) is the only cephalic 
appendage which calls for any remark. In it the exopodite 
(scaphognathite) is characterised at this stage by the pos- 
session of only five sete, the fifth springing almost directly 
from the posterior edge, and not, as in Portunus, from the 
end of a narrowed prolongation of the edge. 


1 Since writing the above I have found a single specimen in the second 
stage of development, in which tle left fork of the telson bears two lateral 
spines as in Thia polita, while the right fork bears but one. The rarity and 
asymmetry of this structure seems to show that its presence is due merely to 
au individual variation (see fig. 4), 


METAMORPHOSIS OF CORYSTES CASSIVELAUNUS. 467 


The two pairs of maxillipedes have each a two-jointed 
exopodite bearing distally four long ciliated sete. 

Behind the first and second maxillipedes there are already 
developed rudiments of the six remaining pairs of thoracic 
legs. The first pair, or third maxillipedes, are longer than 
the rest, and bent forwards between the second maxillipedes. 
The fourth pair is covered by the third, and hence is not 
visible without dissection (see fig. 1). 

As breeding females of Corystes are not easy to obtain, 
and I have consequently not been able to hatch the zowa 
from the egg, it is possible that an earlier stage remains to 
be discovered. Still, the early development of the posterior 
thoracic limbs is not uncommon among Brachyurous 
zowas. In many forms the third maxillipede is already 
marked out in the first zovea, and in some all the thoracic 
appendages are visible on hatching. This is the case in 
Portunus puber,and more especially in Inachus dorset- 
tensis, where even the pleopods are distinctly traceable. 

The abdomen, in the first stage of Corystes, consists as 
usual of five distinct segments, the second bearing a for- 
wardly curved process on either side. The second, third, 
fourth, and fifth segments each bear a short hair on their 
posterior dorsal edge on either side of the middle line. 

The telson (fig. 2), with which the sixth segment is united, 
has the usual forked shape, and bears three strong sete on 
the inner surface of either fork. Hach seta is minutely 
ciliated, the first, however, bearing several much longer cilia 
about the middle of its length. There is only one external 
spine (the sixth of Mayer’s nomenclature). The spine 
formula is therefore 5 + 5 instead of the normal 7 4+ 7. 

Second Stage (fig. 3)—The following measurements 
are the average of the first ten specimens of a number 
measured, and the limits of variation in those specimens. 


468 ROBERT GURNEY. 


Average Minimum and maximum 

length. lengths. 
1. Length of rostrum . 200mm. ... 1°75 to 2°6 mm. 
2. a dorsal spine 2°7 ,, le QA hs Sat he 
3. Tip to tip of spines’ jeg eo tay, be 
4. Length of body . a GA gin Sas SO-P Saree 
5 “ thorax peas Garey reek. Gall ve eis ae 
6. Ratio of 4 to 3, 1: 1°67. 


It will be seen that though the absolute length of the 
spines is much greater at this stage than in the preceding 
one, the ratio between the total length and the total length 
of the body has increased only by ‘01. 

The more important differences between the zozxa at this 
period and that of the preceding one are the following :—In 
the second antenna the flagellum, which is barely indicated 
in the first stage, has grown out to nearly the length of the 
exopodite (spina mobilis). The exopodite and inner spinous 
prolongation of the stem are unchanged. 

The maxille have changed but little in form, but bear 
more sete. The scaphognathite of the second maxilla (fig. 
10) has now nine sete along its Inner margin, and there are 
three terminal sete instead of the single one of the preceding 
stage. 

The first two maxillipedes are unchanged, except that the 
exopodite bears distally six ciliated sete instead of four. 
The remaining thoracic legs are more distinctly developed, 
and there are traces of six gills on each side, i. e. those of 
the third maxillipedes and three succeeding limbs. 

In the abdomen the pleopods are marked out as knobs on 
each segment except the first, which remains lmbless 
throughout the larval development. ‘The third, fourth, and 
fifth segments are produced into a short spinous process on 
each side. ‘The sixth segment is separated from the telson. 
In the latter (fig. 4) the number of internal sete 1s increased 
by one or eyen two pairs, so that there are either six or seven 
on each side altogether. 


METAMORPHOSIS OF CORYSTES CASSIVELAUNUS. 469 


Third Stage.—Measurements (average and range of 
variation in eleven specimens) : 


Average Minimum and maximum 
length. lengths. 
1. Length of rostrum . 30mm. ... 2°7 to 3°5 mm. 
2. pe dorsal spine 36 ,, rey ater: (eee 
2. Lip totip Or spines: -. SO: 4. 4.0 O70) 470-0 Ns 
4, Length of body . Sy ety et ae 
5. i thorax BO Be ochre, WO ee Tore 
6. Ratio of 4 to 3, 1; 1:53. 
7. Length of third pleo- 
pod . : : Son ae Saas 1 as 
8. Length of fourth abdo- 
minal segment . a 2 Bs eae ae” | ee 


At this stage the reduction in length of the dorsal and 
rostral spines as compared with the total length of the body 
has become very noticeable, and the increased completeness 
of the development of the limbs shows an evident approach 
towards the Megalopa. 

The differences between the zova at this period and that 
of Stage 2 consists more in the increased development of 
parts already formed than in the acquisition of new ones. 
The internal branch of the first antenna is developing, and 
the base of the antenna shows a certain degree of swelling in 
preparation for the formation of the auditory organ. In the 
second antenna the increase in length of the flagellum is very 
striking, and it is now about twice the length of the exo- 
podite, showing traces, beneath the cuticle, of segmentation. 
The mandibular palp is present as a small two-joited process. 

In the first maxilla no change has taken place, but in the 
second maxilla (fig. 11) the scaphognathite has not only 
increased in size, but is provided with a very greatly in- 
creased number of sete, which fringe its edge, and are of 
more or less uniform size. ‘Those of the posterior border are 
not longer or stouter than the rest. These changes are 
probably associated with the further development of the 


470 ROBERT GURNEY. 


gills, which perhaps become functional at this stage. The 
first and second maxillipedes retain their original form, but 
the exopodite bears now twelve setz instead of six at its 
distal extremity. Each bears also at its base a small 
epipodite, but there is as yet no trace of gills. The succeed- 
ing six pairs of thoracic limbs are all well developed, and 
show distinct joints beneath the cuticle. The third maxilli- 
pede is the only one that bears an exopodite—a simple 
unjointed process,—the others developing directly to the 
adult form. 

The gills of the posterior thoracic region (fig. 8) are all 
distinctly formed except that the podobranch of the third 
maxillipede is not yet separated from the epipodite. As yet 
also they have not acquired the lamellate form of the adult 
oill. In the abdomen the pleopods are further developed on 
the last five segments, each being about half the length of 
the segment succeeding it. The lateral spines of the seg- 
ments are now more conspicuous than before. 

The telson at this stage (fig. 5) has generally developed a 
new pair of setze on its internal edge in front of the others, 
but in some specimens the number was found unequal on the 
two sides, the formula being normally 8 + 8, but occasionally 
ete 0h 

Fourth Stage (fig. 7)—Measurements (average and 
variation in fourteen specimens) : 


Average Minimum and 

length. maximum length. 
1. Length of rostrum , o4- mm... 3°) to SS" mm, 
2. 7. dorsal. spme,.. 41 ,,. “a. 3S ., tit 
5. Tip to tip of spine a OB gh he, SOD by. te ane 
4. Length of body . ss ch Oe olds d CBtiaune Die via ig 
». Length of thorax . Fa) a nr ee eee! LR hs 
6. Ratio of 4 to 3, 1: 1:31. 
7, Length of third pleopod, ~ °515, | 2.) voy | cag 
8 og fourth abdo- 


minal segment . ., 100: ss. cue 20 3 eee 


METAMORPHOSIS OF CORYSTES CASSIVELAUNUS. 471 


This stage is characterised by the greater development of 
the antenne, gills, and pleopods, but otherwise shows no 
essential difference from the preceding one. In the first 
antenna the inner branch is more developed, and the outer 
branch shows signs of segmentation beneath the cuticle and 
an increased number of sensory rods. The base is much 
swollen, and the auditory pit is forming. 

The second antenna has now a flagellum (endopodite) 
nearly two thirds the length of the rostrum, showing two 
distinct jomts at its base and a number of indistinct joints 
beneath the cuticle. The exopodite and spinous process are 
relatively unimportant structures, and at the approach of the 
moult their contents are absorbed and only the chitinous 
cuticle remains. 

The first maxilla shows no change, but in the second 
maxilla the sete are more numerous upon the scaphog- 
nathite, though they are comparatively shorter than before. 

The epipodites of the first and second maxillipedes are 
larger, but neither podobranchs nor arthrobranchs are 
formed. 

In the third maxillipede, however, the podobranch is being 
separated from the epipodite, and the arthrobranchs are both 
present. The anterior arthrobranch, however, shows no 
signs of lamellar structure, though the posterior one, like the 
succeeding pairs of gills, is distinctly lamellate. 

The gill formula at this stage is therefore as follows: 


| 
A | B. c ree 
Mi. Ep. — | — — ==) aie 
Vit. Ep. — | — — = Kp. 
VALUE Ep.4+1 at ii == =3+Ep. 
1D — il il — —} 
Ext; — — | — il =] 
XI. — —_ | _ il = || 
XII. -= — | = —_ 
>QUule — — - — 
| 7+38 Lp 


472 ROBERT GURNEY. 


The pleopods are now well developed, each exceeding the 
length of the next succeeding segment. The first four pairs 
consist of a broad basal part bearing a long exopodite and a 
short stump representing the endopodite, but there are no 
setee and no trace of segmentation. The fifth pair, on the 
sixth segment, are simple unbranched appendages. The 
telson is exactly the same as in the preceding stage. 

The Megalopa (fig. 13) —Measurements (average of ten 
specimens) : 


Length of carapace. ‘ . 36 mm. 
Breadth across third lateral spine eas 2 iy Mee 
Length of antenne . : ee 5 ae 


The last larval stage passes by a single moult to the 
Megalopa, which is distinctly recognisable as Corystes, 
though retaining certain features characteristic of the zowa. 

The rostrum and dorsal spine are still present, though very 
greatly reduced. 

The rostrum has now the form of a broad plate extending 
forwards between the eyes, its lateral margin arched upwards 
and crenulated. Its extremity is trifid, the median process 
representing the last trace of the original long rostral spine 
and retaining the orange chromatophores of the previous 
stage, the lateral processes by which it is flanked beg new 
formations. A few hairs are borne upon the upper anterior 
surface of the rostrum. 

The dorsal spine is now an inconspicuous orange-red 
process, situated not immediately over, but somewhat behind 
the heart. From it a ridge runs forwards for some distance 
along the middle line of the carapace. 

On either side of the middle line, in the region of the 
stomach, there is a single short spine on the dorsal surface. 
These spines appear first at this stage, and are lost again 
with the next moult. 

Laterally the carapace bears three strong teeth on either 
side, the first immediately behind the eye, and the third 
above the first ambulatory leg. The postero-lateral margin 


METAMORPHOSIS OF CORYSTES CASSIVELAUNUS. 473 


of the carapace is fringed with a number of setz. The 
appendages have now taken on essentially the form of those 
of the adult. The second antenne are considerably longer 
than the carapace, many-jointed, and provided with the 
characteristic dorsal and ventral row of sete. The antenne 
have, as already described, the same function of serving as a 
respiratory tube as they have in the adult. The mandible 
palp is now three-jointed, the distal jomt bearing a number 
of setee and overhanging the mouth opening in front. 

The first maxilla differs from that of the preceding stage 
in the form of the endopodite, which is now not jointed, and 
bears but a single well-developed seta. ‘This reduction in the 
number of sete is remarkable from the fact that in the adult 
there is a rich clothing of sete. 

There is but little change in the form of the inner lobes, 
and the only change from this stage to the condition in the 
adult consists in a relative reduction of the superior lobe and 
an increase in number of spines. 

In the second maxilla there is a great increase in size of 
the scaphognathite and simplification of the structure of the 
endopodite (fig. 12). 

The first and second maxillipedes show an intermediate 
condition between the swimming limb of the zoza and the 
masticatory limb of the adult. The two-jointed exopodite is 
practically unchanged, except that in the first pair it bears 
but five terminal sete, and in the second pair eight. ‘The 
endopodite of the first pair (fig. 15) is no longer jointed, but 
has not acquired the lamellate form characteristic of Corystes. 
The two basal joints are richly setiferous at their imner 
margin, and the epipodite is greatly developed. The endo- 
podite of the second maxillipede (fig. 17) has practically the 
adult form, while the podobranch and small arthrobranch 
are both developed. 

The third maxillipede develops directly to the adult form, 
the second joint of the endopodite having the characteristic 
anterior prolongation. ‘he remaining thoracic legs have in 
all essential respects the form of those of the adult. The 


474, ROBERT GURNEY. 


abdomen still retains some larval characters. The lateral 
spines of segments 2—5 are still retained, and the telson still 
shows traces of bifurcation, being deeply indented posteriorly. 
The five pairs of pleopods have the shape characteristic of 
the typical Brachyurous Megalopa. Those of the first 
four pairs each consist of a stem bearing a long exopodite 
armed with numerous long ciliated setae. The endopodite is 
very small, and interlocks with that of the opposite append- 
age as a retinaculum. ‘The last (fifth) pair of pleopods have 
no endopodites, and are shorter than the telson itself. 
First Post-larval Stage (fig. 14).—Measurement : 


Length of carapace . : . 40 mm. 
Breadth (across third lateral spines) POEs 
Length of antenne . : Pet) i fh 


The cast skin of the specimen from which these measure- 
ments were taken had the following dimensions : 


Length of carapace . ; . 34mm, 
Breadth . : : Sy gd eas 
Length of antenne . : «pared: 


The Megalopa stage lasts, according to my observations, 
from eighteen to twenty days, but possibly a more abundant 
food supply in natural conditions would somewhat shorten 
the period. 

The young Corystes has now attained the structure of the 
adult in almost all respects. ‘The rostral spine is reduced to 
an insignificant tubercle lying at the base of the indentation 
between the two anterior spines. The dorsal spine is com- 
pletely lost, though a small orange chromatophore still marks 
its position on the carapace. The dorsal surface of the cara- 
pace is smooth, the median ridge of the previous stage and 
the two anterior dorsal spines having disappeared. 

Besides the three lateral teeth of the Megalopa a fourth 
tooth is developed behind on each side close to the posterior 
edge of the carapace, so that the number characteristic of 
the adult is attained. The cephalo-thoracic appendages show 
no changes worth noting, except that the endopodite of the 


METAMORPHOSIS OF CORYSTES CASSIVELAUNUS. A475 


first maxillipede has attained its final lamellar form (fig. 16). 
The abdomen, however, has changed considerably. It is now 
kept normally bent up under the body, the young crab having 
taken definitely to a burrowing habit. The first two seements 
are broad and flattened at the sides, while the remaining 
segments narrow out posteriorly and bear no lateral spines. 
All the segments bear sete on their lateral margins. The 
telson has now an evenly rounded posterior margin. 

The pieopods are no longer swimming organs, haying lost 
all their sete. The first four pairs remain biramous, and of 
about the same size as before, but the fifth pair is reduced to 
a simple stump. There is still no appendage upon the first 
abdominal segment, so that apparently in the female this 
appendage never develops, while in the male it is retarded 
till at least the second post-larval stage. I have hitherto 
obtained no later stage than that now under consideration, 
so that I cannot say at what period the distinctive sexual 
characters appear. The specimens of the first post-larval 
stage in my possession show also no difference in the relative 
size of the chelipeds. 


CoNCLUSION. 


The Corystide, though placed by Milne Edwards (1834) 
and by Heller (1863) among the Oxystomata, have by more 
recent authors, such as Claus and Miers (1886), been assigned 
to the Cyclometopa. The resemblance between the Corystidee 
and the true Oxystomata has been shown by Mr. Garstang 
(1897, etc.) to be largely superficial, and due to adaptive 
modifications of an essentially different character, though 
directed to the same ends. He has, in fact, brought forward 
clear evidence that the Corystide and the Oxystomata have 
been independently derived from Cyclometopous ancestors. 

This view is to some extent supported by my observations 
on the development of Corystes, though the great uniformity 
in the structure of the zowa throughout the Brachyura pre- 
vents any conclusion being drawn from the earlier stages. 


4.76 ROBERT GURNEY. 


In fact, it must be confessed that the most striking feature 
of the zowa of Corystes, namely, the great length of the spine, 
recalls the zozas of such Oxystomata as Dorippe and Ethusa 
—forms from which it differs essentially in other respects— 
more than those of the Portunidee. 

Still the final stages of the metamorphosis show that the 
peculiar emarginate rostrum of the adult (which recalls that 
of the Oxystomata) is preceded by a three-toothed rostral 
prominence which exactly resembles that found in most 
Portunids. That the central tooth represents more than a 
mere entogenetic stage in the reduction of the long rostral 
spine of the larva is also confirmed by the retention of a 
trifid rostrum in the adult of Pseudocorystes and Trachycar- 
cinus (Faxon). 

The existence of this Portunid stage in the development of 
Corystes was, I understand, the subject of a verbal communi- 
cation made by Mr. Garstang to the Toronto meeting of the 
British Association in 1897 under the title ‘On Recapitula- 
tion in Development, as illustrated in the Life-history of the 
Masked Crab (Corystes).” As Mr. Garstang has been unable 
hitherto to write up his observations for publication, and as 
he informs me that the material at my disposal is more 
complete than in his own case, I am glad to be able to give a 
full account of the metamorphosis, and to confirm his obser- 
vations. I may here express my indebtedness to him for his 
kind advice and many suggestions during the carrying out of 
my work. 

PiymoutnH ; May, 1902. 


BIBLIOGRAPHY. 


1. Cano, G.— Sviluppo postembryonale dei Dorippidei, Leucosiadi, Corys- 
toidei, e Grapsoidi,’’ Napoli, ‘Atti della R. Accad. d. Sci. Fisiche e 
Matem.,’ 1891. 

2. Ciaus, C.—‘ Untersuchungen zur Erforschung der genealogischen Grund- 
lage des Crustaceensystems,’ Wien, 1876. 


METAMORPHOSIS OF CORYSTES CASSIVELAUNUS. 477 


3. Faxon, W.—‘ Reports on an Exploration off the West Coasts of Mexico, 
Central and South America, and off the Galapagos Islands: XV, The 
Stalk-eyed Crustacea,” ‘Mem. Mus. Comp. Zool. Harvard Coll.,’ 
xvill, 1895. 

4. Garstanc, W.—‘‘ The Habits and Respiratory Mechanism of Corystes 
Cassivelaunus,” ‘Journ. Mar. Biol. Ass.,’ iv, pp. 223—232, 1896. 

5. Garstane, W.—‘ The Functions of Antero-lateral Denticulation of the 
Carapace in Sand-burrowing Crabs,” ‘Journ. Mar, Biol. Ass.,’ iv, 
pp. 896—401, 1897. 

6. Garstanc, W.—‘On some Modifications of Structure Subservient to 
Respiration in Decapod Crustacea which burrow in Sand,” ‘ Quart. 
Journ. Mic. Sci.,’ xl, 1898, p. 211. 

7. Heiter, C.—‘ Die Crustaceen des Siidlichen Europa,’ Wien, 1863. 

8. MityE Epwarps, H.—‘ Histoire naturelle des Crustacés,’ 1834. 

9. Prentiss, C. W.—‘‘ The Otocyst of Decapod Crustacea: its Structure, 
Development, aud Functions,” ‘ Bull. Mus. Comp. Zool. Harvard Coll.,’ 
xxxvi, No: 7, 1901. 

10. Wexpon, W. I’. R.—‘ Note on the Functions of the Spines of the Crus- 
tacean Zowa,” ‘Journ. Mar, Biol. Ass.,’ i, n. s., 1889-90, p. 169. 


EXPLANATION OF PLATES 29—31, 


Illustrating Mr. Robert Gurney’s paper on ‘The Metamor- 
phosis of Corystes Cassivelaunus (Pennant).” 


All figures drawn with the aid of the camera jucida. 


Vie. 1.—( x 32.) Zowa of the first stage, showing distribution of chro- 
matophores. 
Fig. 2.—(x 100.) Telson of the first zoa. 
Fic. 3.—( X 47°5.) Zowa of the second stage. 
Fic. 4.—( x 100.) Telson of the second zowxa. 
Fie. 5.—( x 47°5.) Telson of the third zowa. 
Fig. 6.—(X 35.) Third maxillipede of the first post-larval stage. 
Vie. 7.—(x 26.) Zowa of the fourth stage. 
Fig. 8.—(x 65.) Second maxilla and thoracic appendages of the third 
zowa. 
Hp. 1—Kp. 3. Epipodites of maxillipedes 1—3. 
a l—a3. Arthrobranchs 1—8. 
p', p?. First and second pleurobranclis. 
mxp*, Third maxillipede. 


478 


Fic. 
Fic. 
Vie. 
Fic. 
Fie. 
Iie. 
Fie. 


Ge 


Fie. 


9.—(x 260.) 
100561703) 
11.—(x 105.) 
12.—(x 65.) 


13.—(X 20.) 
14.—( x 20.) 
15.—(X 45.) 
16.—( xX 40.) 


17.—(x 45.) 


ROBERT GURNEY. 


Second maxilla of the first zowa. 
Second maxilla of the second zowa. 
Second maxilla of the third zova. 
Second maxilla of the Megalopa. 
The Megalopa. 
The first post-larval stage. 
First maxillipede of the Megalopa. 
First maxillipede of the first post-larval stage. 
Second maxillipede of the Megalopa. 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 479 


Artificial Parthenogenesis and Fertilisation : 
A Review. 


By 


Thomas HH. Bryce. 


. Tats article is an effort to gather together, in so far as they 
relate. to the phenomena of fertilisation in the sea-urchin 
ego, the results obtained by experiments. It does not pre- 
tend to consider the problem of fertilisation as a whole, nor 
the phenomena save in Hchinoderms, and no attempt will be 
made to establish comparisons with other forms in which the 
details may to some extent differ. ‘The limitation to one form 
is in so far appropriate, that practically all the experiments 
have been made on Hchinoderm eggs. 

I have personally studied fertilisation in the egg of Echinus 
esculentus—specially in sections,—and though I have 
nothing fresh to add to the description of the facts, this 
article may in a measure be considered a sequel to a paper 
on maturation in the same form. In that paper my attention 
was chiefly directed to the chromosomes, and I did not follow 
out the results of observers in the experimental field, but 
as some of the phenomena described are of interest in con- 
nection with these results, I shall take the opportunity of 
returning to them. 

The two mitotic divisions characteristic of the maturation 
phases, differ markedly from those which take place in the 
segmentation phases. In many respects there is a close 
resemblance to phenomena observed in the eges of ‘Toxo- 
pneustes, which develop parthenogenetically under the 
influence of magnesium chloride solution (Wilson, 1901), 

VOL, 46, PART 3.—NEW SERIES. GG 


480 THOMAS H. BRYCE. 


On the dissolution of the nuclear membrane the site of the 
germinal vesicle is occupied by a “kinoplasmic” mass, 
derived either entirely from the nuclear network, or also 
partly from protoplasm differentiated on the distribution of 
the nuclear substance into it. In fixed material this area has 
a fibrillar appearance. This may be the result of the fixing 
reagents used, but in any case it indicates the accumulation 
at this part of protoplasm which has undergone some change 
in constitution physical or chemical. In the nuclear area, 
out of this material, asters are formed, and ultimately the 
first polar amphiaster. Besides the asters concerned in 
the formation of the bipolar figure, there are secondary 
asters, which seem to have only a temporary existence. 
In some few cases multipolar figures were observed. No 
structure recognisable as a centrosome or centiole was found 
before the germinal vesicle broke down, and therefore the 
centrosome was either derived from the nucleus, or formed 
de novo in the nuclear area. ‘The astral radiations are 
confined to a small and superficial part of the egg, and a very 
unequal division results in the formation of the polar bodies. 
When the two divisions are over all the radiations and the 
remains of the kinoplasmic area disappear, the cytoplasm 
assumes its alveolar structure throughout, the nucleus retires 
from the surface, and no centrosome can be recognised 
in relation to it. 

On the breaking down of the germinal vesicle the greater 
part of the nuclear material disappears as such, and not only 
is a change in the constitution and distribution of the proto- 
plasm to be recognised, but experiment proves that the egg 
has undergone a physiological change of state. Whereas a 
spermatozoon can neither fertilise an egg with the germinal 
vesicle intact, nor a fragment without the nucleus (Delage, 
1901), after the polar bodies are formed the egg becomes 
capable of fertilisation.’ This cytoplasmic maturation, de- 
pendent, probably (Delage, 1901), on the influence of the 


! For evidence and theories regarding the influence of the stage of matura- 
tion in Amphibian eggs see Bataillon (1901). 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 481 


nuclear sap set free from the germinal vesicle, ig accompanied 
by the conversion of the large vesicular nucleus related to the 
metabolic changes underlying the growth of the ovum, into a 
small morphologically equivalent nucleus, possessing the same 
number of chromosomes as the sperm nucleus. This has been 
proved by their enumeration when the nuclei undergo in- 
dependent transformation, and the number is one half that 
found in the segmentation divisions. 

The eggs thus matured remain, in the case of the sea-urchin, 
for a considerable time quiescent within the ovary before they 
are discharged—for the process of ripening in the ovary 
is a gradual one. When discharged into sea water it seems 
that, like the eggs of some other forms (O. Hertwig, 1893, 
p- 239), after lying for many hours unchanged, the sea- 
urchin eggs show spontaneously, karyokinetic transforma- 
tion; for instance, R. Hertwig (1896) observed in eggs which 
had been deposited prematurely during transport, analogous 
changes to those produced by treatment with strychnine. 
This phenomenon is one apparently of wide range. In an 
interesting review entitled “ Giebt es bei Wirbeltieren Par- 
thenogenesis’”’ (1900), Bounet, after examination of all the 
literature up to that date, comes to the conclusion that, 
according to our present knowledge, the phenomena in verte- 
brates are due to degenerative divisions, and in meroblastic 
eggs to fragmentations, and the alleged parthenogenetically 
divided tubal, uterine or laid eggs, are either over-ripe, and 
therefore badly fertilised, or are eges normally fertilised 
with defective spermatozoa. In the light of the facts of 
artificial parthenogenesis, it may be that this segmentation 
in unfertilised eggs, at least in certain invertebrates, is an 
effort in the direction of true parthenogenesis which is abor- 
tive, the egg dying before the tardy process is accomplished. 

In 1876 Greeff described parthenogenetic development in 
Asterocanthion. The eggs were obtained from animals 
early in the season, before the spermatozoa were mobile, and 
the blastulee formed differed from those preduced in normal 
fertilisation. O. Hertwig (1890) recorded some observations 


482 THOMAS H. BRYCE. 


on spontaneous parthenogenesis. In confirmation of Fol, he 
found that eggs from fully-matured animals did not seement 
spontaneously, but only after a considerable time underwent 
changes considered pathological. The nucleus enlarged 
more and more, and after ten to fifteen hours the eggs 
died and fragmented. Only among hundreds of eggs here 
and there one had divided into two. At Trieste, however, 
in a season when the animals were late in maturing, and at a 
time when males were rarely got, he observed in a limited 
number of cases (in Asterias glacialis, and Astero 
pecten) that after the polar mitosis had occurred, the nucleus 
did not come to rest, but continued to divide. There resulted 
an irregular division, but here and there a blastula was found 
which had no vitelline membrane. Into the interesting 
observations and suggestions regarding the failure of the 
second polar body extrusion, and the union of two vesicular 
nuclei in the egg, we cannot here enter. The main point 
established was, that fully-matured eggs did not develop 
parthenogenetically, but that in some few cases immature 
eges did divide irregularly, and in a small number of cases 
blastulee were formed. A number of observers have 
described the occurrence of natural parthenogenesis in 
Kehinoderms, and it is an open question; but apart from its 
possible relation to immaturity of the ovum, the sources of 
error in the matter of infection by spermatozoa are so many, 
and the causes which artificially start parthenogenetic 
development in certain cases are so slight, that all cases of 
so-called “ natural parthenogenesis” are open to suspicion, 
but even granting that it may occur, it is a matter of no great 
moment in the question of “artificial parthenogenesis.” It 
would be only additional evidence of the fact, that there is in 
these forms a tendency to parthenogenetic development, 
which, however, does not normally occur. 

Mitotic division may be excited in unfertilised eggs in a 
variety of ways. 

Kirst, by increasing the degree of concentration of the 
sea water (Morgan, Hunter), or by increasing the osmotic 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 483 


pressure in various other ways (Loeb). This may be 
done by adding various inorganic salts to the sea water, 
especially the salts of magnesium, potassium, sodium, and 
calcium, in definite proportions (Loeb, Morgan, and others), 
or by adding sugar or urea (Loeb). Other salts have also 
given results, e.g. chloride of manganese (Delage). The 
effect is produced when the eggs, after being left from 
half an hour to two hours in the solution, are transferred to 
pure sea water, and is due to the disturbance of the osmotic 
pressure leading to loss of water by the egg, followed by 
rehydration (Bataillon, Loeb, Giard, etc.), not to specific 
chemical stimulation. 

The nuclear activity may also be roused by other chemical 
bodies, as strychnine (Hertwigs, Morgan), chloroform, ether, 
alcohol, by lack of oxygen (Mathews), by very dilute hydro- 
chloric acid (Loeb, Delage). 

Further, purely physical agents may have the same effect— 
heat (Mathews, Bataillon, Delage, Viguier), cold (Morgan, 
Greeley), and, most important, agitation (Mathews). Mathews 
had previously proved for Asterias, and Morgan for Arbacia 
also, that shaking of unripe eggs caused them to form the 
polar bodies—the shaking presumably causing dissolution of 
the nuclear membrane. Ripe eggs of Asterias, but not of sea- 
urchin, act in the same way, but only after they have lain 
some time in water; after two hours larve begin to appear 
on shaking; after four hours, hard shaking produces a large 
proportion of larvae, and the mere transference of the eggs 
by a pipette from one vessel to another is sufficient to form a 
few larve. A few hours later the slight amount of shock 
experienced in the transference of the eggs, causes a large 
number to begin to develop, though they do not go beyond 
the late segmentation stages. At this time shaking causes 
all to develop, but none reach the blastula stage. Loeb and 
Fischer have extended this observation to the Annelids, 
Cheetopterus and Amphitrite. 

The mitotic phenomena produced artificially are apt to be 
irregular, and the division of the cell body is often unequal 


4.84 THOMAS H. BRYCE. 


when it occurs. Thus the nucleus may divide repeatedly 
without division of the cytoplasm, and then the egg 
may break into as many segments as there are nuclei 
(Wilson and others). It is only in a relatively small pro- 
portion of eggs that division is regular enough to permit of 
development to the larval stage. Further, the eggs of the 
same species behave capriciously to the same agents under 
different conditions (temperature, etc.), and the eggs of closely 
allied species seem to react differently to the same agent. 

There is no reasonable doubt, however, that true artificial 
parthenogenetic development has been demonstrated for the 
Kchinoderms—sea-urchins and star-fish—and for at least two 
Annelids, though the same amount of independent testimony 
is not available for the latter. 

Actual development to a larval stage has been obtained 
only by certain of the agents enumerated above. 

1. Increase of Osmotic Pressure-—The most successful 
results (Loeb, 1902) are obtained at a temperature about 
20° C., by the addition of the chlorides of potassium or sodium 
to sea water, the optimum degree of concentration being 
determined by experiment for each set of observations.! 
After the eggs have remained in this for half an hour to two 
hours, the optimum being again tested by experiment, they 
are restored to normal sea water. 

Sea-urchins (Loeb, Wilson, Giard, Prowazek, Delage, 
Viguier, Hunter). Annelids: Chatopterus, Amphitrite, 
Nereis (Loeb, Fischer). 

2, Agitation. — Asterias (Mathews), Cheetopterus and 
Amphitrite (Loeb, Fischer), but not sea-urchins, (Mathews, 
Viguier). 

3. Hlevation of Temperature.—Asterias during maturation 
(Delage) ; not for ripe eggs (Greeley, Viguier). 

4. Depression of Temperature.—Asterias (Greeley) ; not 
sea-urchins (Viguier). 

' Loeb (1902) uses a stock solution of 23 n. HCl, and adds this in different 


proportions, 8, 10, 12, 14, 16, 18 e.em., to 100 ¢.c. of sea water in six 
vessels to determine the best grade of concentration, 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 485 


5, Exposure to weak HCl in sea water, and subsequent 
restoration to pure sea water. Asterias (Loeb, Delage). 

6, Continuous exposure to a solution of a specific chemical 
substance at the same osmotic pressure as normal sea water. 

Potassium Chloride. Cheetopterus (Loeb). 

Calcium Chloride. Amphitrite (Fischer). 

The Ions of potassium and calcium are said to be specific 
for these forms respectively.! 

With regard to the influence of the state of maturation, 
Delage gives results to show that in Asterias glacialis, 
when the eggs are placed in sea water to which is added an 
equal quantity of a solution of HCl, raising the molecular 
concentrations of the mixture to 0°660, different results are 
got according to the stage of maturation. Among the eggs 
placed in the liquid before maturation, 20 per cent. of 
blastule were got, at the appearance of the first polar 
body 95 per cent., and after the appearance of the second 
polar body 5 per cent., while none of the controls showed 
any normal segmentation. 

From all this it seems that changes in the osmotic 
pressure between the egg and its surrounding medium, and 
mechanical agitation, are the chief agents so far as yet 


1 Delage (‘ Compt. Rendus de l’Acad. des Sciences,’ October 13th and 20th, 
1902) announces that he has found an agent which is as certain and effective 
as the spermatozoon, in producing development to advanced larval stages, in 
Asterias. It is sea water aérated by carbonic acid gas, and at the same osmotic 
pressure as ordinary sea water (or lower?). When the eggs, at what he 
calls the “ critical stage ’—i.e. when the nuclear membrane of the germinal 
vesicle is dissolved, up to the expulsion of the first polar body—are placed in 
this, and after one hour transferred to pure sea water, practically all the eggs 
develop. His view is, that the maturation is arrested temporarily, and on 
restoration to pure sea water, the carbonic acid gas is quickly eliminated and 
division proceeds ; but it is not partial, as in the polar mitoses, but complete, 
and goes on to the formation of the normal larval forms. The result is not 
obtained at a stage after the polar bodies are extruded and the ovum has again 
come to rest, nor is it applicable in sea-urchin, in which the maturation is 
over before the ova are shed. His theory as to the action of the gas is, that 
it is a temporary poison which arrests maturation completely, and is quickly 
removed afterwards without altering the characters of the protoplasm. 


486 THOMAS H. BRYCE. 


known, which tend to the production of artificial partheno- 
genesis, but that in the case of the Annelids there is evidence 
to show that certain Ions may have a specific effect. 

According to Loeb (1902) the solutions must act, first, by 
favouring the solution or dissolution of the nuclear mem- 
brane; and second, by changing, in some sense, the physical 
properties of the protoplasm (viscidity, etc.). 

Mathews (1900), as a conclusion from his experiments on 
Arbacia eggs, pointed out that the known methods of causing 
liquefaction in protoplasm will induce karyokinesis in these 
eggs, and also shows that loss of water has a liquefying 
action. 

Before considering further the bearing of the physiological 
and physico-chemical conceptions regarding fertilisation, I 
shall proceed to the morphological changes which have been 
described in unfertilised eggs which undergo parthenogenetic 
development. 

R, Hertwig (1896) studied the changes in the egg after 
treatment with strychnine. On the breaking down of the 
nucleus, half spindles, and in a few cases whole spindles, 
supposed to arise from the fan spindles, were formed. The 
fan spindle fibres he regarded as derived from the achro- 
matic network of the nucleus. The chromosomes derived 
from the nucleoli became attached to the primary rays. 
Later, protoplasmic rays also appeared, centering on the 
focal point of the half spindle. At this central poimt, and 
derived from the central parts of the rays, there appeared a 
rounded body resembling in every way a centrosome, though 
none such was to be found before the nucleus broke down. 
The body was an ovocentrum, formed from the achromatic 
portion of the nucleus, and, according to Hertwig, the 
‘individualised centrosome is ultimately a derivative of the 
nucleus—is, in fact, an achromatic nucleus. 

Doflein (1897), contrariwise, examined the phenomena 
of karyokinesis of the sperm nucleus in eggs which, after 
fertilisation, had been treated by chloral solution after the 
manner of the experiment of O. and R. Hertwig. The nuclei 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 487 


did not unite, but underwent independent transformation. 
Doflein, like R. Hertwig, considered: the middle piece of the 
spermatozoon as equivalent to the centrosome, and from the 
experiments concluded, that from the centrosome a complete 
spindle could form, and out of this, again, the achromatic 
nuclear network. Thus, compared with Hertwig’s results, the 
ripe sperm nucleus contains all the parts, even as the ripe 
ege nucleus, which are necessary for a further development. 

In Hertwig’s results we have evidence of a centrosome 
arising from the nucleus de novo. Morgan, in 1896, 
described the formation of artificial astrospheres in the 
cytoplasm of the eggs of Arbacia treated by salt solutions, 
and from his further observations published in 1899 and 
1900 he decided, that in spite of certain differences these 
artificial astrospheres corresponded to the normal spheres 
which occur at the apices of the spindles in the sezmentation 
stages; further (1900), that both artificialand normal spheres 
are due to accumulation of a specific substance, and that the yolk 
spheres are excluded from the substance of the astrospheres. 

His view of the astral radiations is that they serve to 
transport the chromosomes, but are not concerned in the 
division of the cytoplasm. 

Evidence of free formation of the centrosomes is found 
also in the appearance of asters in the cytoplasm in various 
forms, Hchinus among them, on the breaking down of the 
germinal vesicle in maturation. 

Boveri (1901) in essence accepted Hertwig’s definition of 
the structure described by him as an ovocentrum, and _ its 
origin apparently de novo. He argued that phylogenetically 
the centrosome is an individualised cytocentrum, derived 
from a centro-nucleus in which the centrosome or its 
equivalent is not differentiated from the chromatin nucleus. 
To the nucleus of the sea-urchin egg must necessarily 
be attributed the properties of a centro-nucleus, with the 
capacity of producing out of itself, under the action of 
certain stimuli, individualised centrosomes, when such fail to 
be supplied in the normal way in fertilisation. If even under 


4.88 THOMAS H. BRYCE. 


similar conditions, the sperm centrosome be present, the cyto- 
centrum remains latent. The centrosome looked at in this 
way, is not a specific cell organ in the sense that it must 
consist of a specific chemical substance, but that parts of a 
substance contained in the nucleus, undergoing certain 
changes, and grouping themselves together, are organised 
into a centrosome, 

Thus the ovocentrum of the sea-urchin egg is not to be 
considered an individualised centrosome, but an intra- 
nuclear latent cytocentrum, and the nucleus is a centro- 
nucleus. Thus the centrosome in such a case is not some- 
thing strictly new, but arises by the transformation in a 
definite manner of a cytocentrum already present. It isa 
case not of new formation, but of “ reparation.” ‘ Gervisse 
Centronuclei sind im stande unter bestimmten Bedingungen 
Centrosomen zu reparieren,” 

Morgan’s artificial astrospheres he did not admit to have 
true centrosomes—the essential character of capacity for 
division was not proved for them. 

This brings me to Wilson’s very interesting and important 
paper on the morphological phenomena in parthenogenetic 
eggs. 

The main results are that under the influence of the 
magnesium chloride solution, not only are asters produced 
de novo in connection with the nucleus, but also in the 
cytoplasm. “Not only the asters connected with chromo- 
somes (nuclear asters), but also the supernumerary asters 
unconnected with nuclear matter (cytasters), may multiply 
by division; the cytasters contain deeply staining central 
granules indistinguishable from centrosomes, that divide 
to form the centres of the daughter asters. ‘These asters 
operate with greater or less energy as centres of cyto- 
plasmic division. ‘Typical cytasters, often containing deeply 
staining central granules resembling centrosomes, are formed 
in the magnesium solution in enucleated egg fragments 
produced by shaking the unfertilised eggs to pieces, and 
these asters likewise may multiply by division, though 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 459 


no cytoplasmic cleavage takes place. The cleavage centro- 
somes first make their appearance outside the nucleus, but 
directly on the nuclear membrane, and the evidence renders 
it nearly certain that they arise by the division of a single 
primary egg centrosome that is formed de novo. All the 
evidence goes to show that the cleavage centrosomes are of 
the same general nature as the central bodies of the cytasters.” 

Among many interesting details I will refer here only to 
the changes described for eggs which underwent segmenta- 
tion, and were capable of developing into swimming embryos, 
because in certain particulars they are reminiscent of what 
takes place in the formation of the first polar amphiaster. 

I may summarise as follows :—(1) The first change that 
occurs is a coarsening in the appearance of the protoplasm, 
better marked in eggs treated by stronger solutions. (2) A 
primary radiation appears centering on the nucleus, better 
marked in eggs treated with weaker solutions. (3) A varying 
number of secondary radiations appear in eggs especially 
treated with stronger solutions. The extent of the primary 
radiations is inversely in proportion to the number of the 
secondary radiations. These latter appear as vague clear 
spots in the cytoplasm, which gradually become surrounded 
with radiations, and finally assume the form of asters. They 
always appear in situ, and do not change their position till a 
later period. (4) Coincident with the appearance of the 
radiations there is a gradual growth of the nucleus. (5) 
Round the nucleus appears a clear perinuclear zone of hyalo- 
plasm. (6) The nuclear membrane fades out, and a vague 
irregular clear space is left, to which the hyaline zone con- 
tributes. (7) The rays then diminish, and, indeed, almost 
disappear. 

The eggs at this point were restored to pure sea water, and 
after a pause the radiations reappear and advance centri- 
fugally towards the periphery. In eggs capable of develop- 
ment the principal rays are now focussed on two centres at 
opposite poles of the nuclear area, which now forms a spindle 
connecting the two asters. If the amphiaster is typical, 


4.90 THOMAS H. BRYCE. 


division proceeds as in normal fertilisation. If more than 
two asters are formed from the nuclear area, multipolar figures 
form, and irregular cleavage results. If there is only a single 
radiation which does not resolve itself into a bipolar figure, 
the egg never properly segments, but there are regularly 
alternating phases of nuclear transformation. 

Analysing the meaning of the phenomena, Wilson says, ““ We 
may therefore state that the first general effect of the stimulus, 
whether the magnesium solution or the spermatozoon, is to 
arouse an activity of the cytoplasm, one result of which is the 
establishment of a centripetal movement of the hyaloplasm 
towards one or more points at which the hyaloplasm accumu- 
lates.’ The rays in this view are the expression, in part at 
any rate, of centripetal currents, and the substance flowing 
in, is the hyaloplasm or interalveolar substance. The hyalo- 
plasm spheres at the centres of the asters are local accumula- 
tions of this hyaloplasm. In fixed material, studied in 
sections, the radiations are fibrillar in appearance, and as 
they stain much more deeply than the general network 
the hyaloplasm in the rays must probabiy have undergone 
some physical or chemical change. ‘he centrosome is a 
well-defined body of considerable size and of spongy con- 
sistence, composed of intensely staining granules, which often 
give the centrosome the appearance of a minute nucleus 
containing a chromatin reticulum. ‘The hyaloplasm spheres 
in the living egg correspond to the centrosome, the clear area 
round it, and the innermost darkly staining radiated zone of 
the aster taken together. 

hus Wilson has proved that structures which cannot be 
distinguished morphologically from “true centrosomes” 
appear in the cytoplasm de novo; and further, that they 
divide to form the apices of bipolar figures, even in enucleated 
fragments. 

In a recent paper Meves (1902), using Boveri’s nomen- 
clature, expresses the view that the centrosome is only the 
mantle of the centriole, and is only present in rapidly-dividing 
cells like the blastomeres. The ‘ Doppelkérchen” of the 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 491 


tissue-cells are to be considered as centrioles, and ‘‘ nur von 
den Centriolen nicht aber von den Centrosomen, kann daher 
gelten, dass sie allgemeine und dauernde Zellorgane sind.” 
The results of Morgan and Wilson can only then be held to 
prove that centrioles under certain conditions may, by the 
action of salt solutions, be excited to form centrosomes and 
radiations round them, for their results might be explained by 
a multiplication of the two centrioles which the egg has 
derived from the last division of the division period, and the 
distribution of these centrioles through the cell. Even in 
enucleated fragments there is no proof that the fragment did 
not contain the centriole of the cell. 

Such a supposition admits of neither proof nor disproof, 
and the presence of a free “centriole” in the unfertilised 
sea-urchin ege@ has not been demonstrated. I have seen in 
young oocytes minute bodies, stained black with iron 
hematoxylin—sometimes double bodies,—but I have not been 
able to convince myself that they are more than accidents of 
staining and fixing. 

Turning now to the phenomena of fertilisation in the sea- 
urchin, there is to be recognised (1) a local stimulation at 
the place of contact of the chosen spermatozoon,! manifested 
by the streaming out of the protoplasm to form the entrance 
cone. (2) A general stimulation, manifested by the throw- 
ing off of the vitelline membrane, and by a change in the 
constitution of the protoplasm. It becomes more viscid for a 
time (Morgan) ; a funnel-shaped area of darkly staining sub- 
stance follows the path of the sperm head (Wilson). (3) 
A protoplasmic movement focussed on the situation of the 
middle piece giving rise to the sperm aster. This appears 
soon after the entrance of the spermatozoon, when the head 
has begun a movement of rotation. The rotation goes on 


Buller (1902) has studied the question of the bearing of chemotaxis on 
fertilisation in Echinoderms. His conclusion is that chemotaxis plays no 
réle in bringing the sexual elements together. The meeting is a matter of 
chance. The passage through the gelatinous coat is radial in direction, and 
probably purely mechanical, though possibly due to stereotaxis. 


4,92 THOMAS H. BRYCE. 


through 180° till the base of the conical sperm head is 
directed inwards. The rays of the aster now extend widely, 
and at their centre is a clear area. Meantime the sperm head 
becomes converted into a small round nucleus. The move- 
ment of the sperm head is, at first, radial; then there is a 
change, and it assumes a new direction towards a point not 
quite in the centre of the egg; when this change of path is. 
taken up the egg nucleus begins to move towards the point 
where the nuclei ultimately meet (Wilson and Giardina). 
The aster now comes in contact with the egg nucleus, and 
as the nuclei approach, the clear area at its centre spreads 
out over its side. ‘The aster then divides and the nuclei 
conjugate. The radiations now die down during a pause in 
which the nucleus grows in size (Wilson), to redevelop again 
focussed at the poles of the nucleus. 

According to Hertwig, Doflein, Erlanger, and Wilson’s 
earlier account, the centrosome corresponds to the whole 
middle piece, but later Wilson described the middle piece as 
cast aside, and in the centre of the aster is a small darkly- 
staining granule. Boveri (1901) represents the sperm 
centrosome as a spherical body smaller than the middle piece, 
and containing within it two centrioles shortly after its 
entrance into the ege. 

Various other observers have represented a dark-staining 
eranule at the centre of the aster. My own observations 
are inconclusive, and do not warrant me in expressing an 
opinion.! 

1 The character of the fully-formed centrosome in the sea-urechin egg is 
still subject to difference of opinion. The form in which I see it in osmic 
acid material is that of a largish sphere of very finely alveolar structure. In 
Wilson’s papers on magnesium and etherised eggs, “it appears as a well- 
defined body of considerable size, consisting of intensely stained granules, 
which often give the centrosome exactly the appearance of a minute nucleus 
containing a chromatin network.” This becomes in the anaphases more homo- 
geneous, and flattens down into a plate-form, which in the telophases often 
lies directly on the membrane of the newly-formed nucleus precisely as Boveri 


(1901) has described for Echinus. Boveri (1901) represents it in several 
forms. In one set of preparations it is a largish sphere of very finely alveolar 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 4985 


The essential difference between the processes seen in 
magnesium eggs and normal fertilisation is that whereas in 
fertilisation there is only one, and that a definitely localised 
point of astral activity, in the magnesium eggs there are a 
number of foci, and development in large measure depends 
on the accident of their number in the nuclear area. 

There is the same want of unity of purpose that is seen in 
polyspermic eggs, in which the number of points of astral 
activity depends on the number of spermatozoa which gain 
an entrance. 

It has lone been recognised that the union of the nuclei 
and the initiation of division are co-ordinated, but in a 
measure independent factors in fertilisation. Partheno- 
genetic development under artificial agents is the iatest 
proof of this. The possibility of the development of 
enucleated ego fragments when entered by a spermatozoon, 
as described by Boveri, and afterwards named merogony by 
Delage, is another. Hither nucleus is sufficient in itself. 

With the problems underlying the nuclear conjugation 
this article is not concerned. It starts from the assumption 
that the union of equivalent nuclei is the end of fertilisation, 
but not the means (Boveri). 

The cause of the nuclear conjugation is not as yet under- 
stood. The first possibility is that the aster is concerned in 
bringing them together. Giardina (October, 1902) brings 
the latest suggestion on this line. Starting from the basis 
of the alveolar structure of protoplasm, he suggests that the 
aster is the expression of both centripetal and centrifugal 
currents. The centrosome is concerned in the diffusion of 
chemotrophic substances into the egg, while at the same time 


structure. In another set, in which the centrosome had reacted differently, 
there is a centriole within the centrosome, which divides before the centrosome, 
so that it is double in the metaphase. In Wilson’s earlier account there was 
no central body, but in later descriptions there was a mass of granules in a 
well-defined sphere, which succeeded a single granule of earlier stages. In 
my previous paper, I regret that I misrepresented Professor Wilson’s nomen- 
clature by referring to this as his centrosome. The sphere, as a whole, is 
named the centrosome. See note to page 314, “ The cell, ete.,” 1900, 


4.94, THOMAS H. BRYCE. 


the hyaloplasm flows in towards the centre. He points out 
that the germ nucleus does not move till the rays of the aster 
have reached it, and the aster has assumed a position of 
equilibrium towards the centre of the egg. The union is 
thus the result of the chemotactic forces of which the aster 
is the expression. 

Wilson (1901 8) shows, however, that the nuclei may 
unite in the entire absence of an aster. When eggs, im- 
mediately after fertilisation, are placed in a weak solution of 
chloral (O. and R. Hertwig), or ether (Wilson), no aster is 
developed, but when replaced in sea water the rays reappear 
and the nuclei unite. In a certain proportion of cases, 
which will be referred to later, the nuclei remain apart and 
undergo independent transformation ; but in some instances, 
also while the eggs are still in ether, the nuclei enlarge, and 
later conjugate in the entire absence of an aster, This 
happens, however, only when the spermatozoon has entered 
at a point not too far from the egg nucleus. Giardia holds 
that this fact, and the other—that the nuclei quickly unite 
whenever the eggs are put in pure sea water, and the aster 
develops,—makes Wilson’s observation insufficient to exclude 
his hypothesis. Other explanations, such as mass attraction 
and direct chemical attraction, both observers reject. 
Wilson thinks the latter improbable. Again, the idea of 
protoplasmic currents such as suggested by Butschli, 
Erlanger, and Conklin, is not proved by actual evidence in 
normal conditions in the sea-urchin egg (Wilson). The 
changes of shape of the germ nucleus might suggest 
amceboid movement on its part; but, again, this does not 
apply to the sperm nucleus, which travels through a longer 
path (Wilson). The changes in form might be due to the 
exercise of chemotactic forces on the nucleus (Giardina). 

The phenomenon described by Boveri (1888) under the 
name of “ Partial Fertilisation,” has recently been worked 
out in detail in Boveri’s fixed preparations by Teichmann 
(1902). The method by which the results were obtained 
was that eggs which had lain fourteen hours in unrenewed 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION, 495 


sea water were fertilised with spermatozoa, which were 
treated with a ‘05 per cent. solution of potassium hydrate 
until only a few were mobile. While polyspermy occurred in 
more than half the eggs, the remainder were fertilised by a 
single spermatozoon. In these cases, however, the sperm 
nucleus did not unite with the germ nucleus, but the aster 
became detached from it, and advanced alone to the germ 
nucleus, a bipolar figure was formed and division proceeded. 
The sperm nucleus took no share in the process, but passed un- 
altered into one of the blastomeres. later, however, either in 
the two- or the four-cell space, it broke up into its chromosomes, 
which entered into the equatorial plate of the cell in which it 
was included, and which now divided like its neighbours. 
Such eggs were capable of developing to the blastula stage. 

The question presented itself: Was this aster and the 
amphiaster the result of the activity of an ovocentrum, or 
were they the product of the sperm aster? 

_ In monospermic eggs Teichmann found the early stages 
very scarce, and, though very suggestive, too few for absolute 
proof, and the phenomena seen in dyspermic eggs are 
described to fill the gap. It may be admitted that in these 
eges, in spite of the apparent inactivity of the sperm nucleus, 
the sperm aster with its centrum is the operative factor in 
starting the developmental process. The appearances are 
very similar to those in the etherised eggs described by 
Wilson (1901). In that form, as in Hchinus, the nuclei con- 
jugate when they are very unequal in size, and before the 
division of the aster. In Asterias and other forms an amphi- 
aster is developed before conjugation, and the nuclei are 
nearly equal in size. In the experiments the union was 
delayed, as in “ partial fertilisation,” and the amphiaster was 
formed before the conjugation. 

Among 'l'eichmann’s observations I shall refer only to those 
of monospermic eggs. The main feature is the detachment 
of the sperm aster from its nucleus, its application to the egg 
nucleus, and its normal division, followed by normal segmen- 
tation. The fate of the sperm nucleus depends on the 

vou. 46, PART 3.—NEW SERIES. HH 


4.96 THOMAS H. BRYCE. 


position it assumes in the egg, relative to the cleavage 
plane. If it lies outside the equatorial plate of the spindle 
it passes unchanged into one of the blastomeres; if it 
lies within the field of the first spindle, it does not 
actually unite with the chromatin of the female nucleus, but 
its chromatin undergoes a marked relaxation. Though it 
shows a marked resistance to the tractive forces, it is drawn 
out and torn into several shreds. It thus passes undivided 
into one of the blastomeres, and no chromatin elements 
derived from it are found at the poles of the spindle. In 
several cases where the nucleus lay exactly at the equator, 
and the traction of the poles was nearly equal, it was 
observed that the chromatin mass was much broken up, and 
was torn into two parts. The cleavage of the cell body may 
have helped to complete the division. The loosening of the 
sperm chromatin mass in the first spindle seems to have 
broken its power of resistance, for when the next division is 
initiated, the two nuclei lying side by side in one of the 
blastomeres unite in the equatorial plate stage, and the 
chromatin of both is equally distributed in the next division. 
The number of chromosomes is now different in the blasto- 
meres, sometimes double, perhaps quadruple in some cases 
(though an accurate count was not possible), as if the chromo- 
somes of the sperm nucleus emerged in double number, 
though the first division was suppressed. 

The sperm nucleus in the case where it has not lain within 
the power of the spindle in the first division, may now, in 
analogous fashion, be caught in the second division, to unite 
later with the chromatin of one of the blastomeres of the 
four-cell stage, or it may even pass over into the eight-cell 
stage, as seen in living eggs by Boveri. 

Teichmann concludes that the radiations are derived from 
the sperm centrosome as their starting point, and on the 
supposition that the centrosome is introduced by the sperma- 
tozoon into the egg, that it has suffered less from the chemi- 
cal reagent than the nucleus. The centrosome behaves as 
in ordinary fertilisation, the sperm nucleus is passive, and 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 497 


seems to form no hindrance to the normal processes in the 
ego, and it seems to be of no significance, whether it enters 
earlier or later into union with one of the descendants of the 
ego nucleus. The difference between the phenomena of 
“partial fertilisation’ and the normal process is the non- 
union of the nuclei. In certain cases Boveri (1890) 
described an independent transformation of the nuclei 
under normal conditions, but the elements from both entered 
into the equatorial plate of the first cleavage spindle, and 
normal division took place. The mere want of union is not 
of moment, if the sperm nucleus hes near enough the germ 
nucleus to be influenced by the nuclear fluid of the egg 
nucleus. ‘lhe absence of this hastening factor may explain 
the fact that the karyokinesis of the sperm nucleus in 
enucleated fragments is much slower than in the cleavage 
spindle. But such an explanation alone will not hold for 
cases of dyspermy in these experiments, where the sperm 
nuclear descendants remain far behind the derivates of the 
ego nucleus, and it must be assumed that a change has taken 
place in the sperm nucleus itself, a kind of paresis, produced 
by the potassium hydrate. This holds for the monospermic 
egos also, and explains why, even in spite of its position, the 
nucleus does not enter into union. 

Another factor is a change in the egg, defined as an over- 
ripeness. In many cases the germ nucleus is a stage ahead, 
compared with the normal process, of the centrosome. When 
the centrosome met the egg nucleus, the latter must already 
have been in a way prepared for division, and this great 
readiness to enter into division may be part explanation of 
the lagging behind of the sperm nucleus. ‘here has not 
been time for the sperm nucleus to undergo transformation 
before the egg nucleus has submitted to division. Teichmann 
does not explain in what the over-ripeness consists. It may 
perhaps be that, since the eggs had lain fourteen hours in 
unrenewed sea water, the early preparatory stages of the 
natural transformation had supervened, which takes place in 
eggs after lying long in sea water. 


2) 


498 THOMAS H. BRYCE. 


Regarding the main point, it may be admitted that the 
aster and its centrosome here concerned is that belonging to 
the sperm nucleus, and that in its behaviour we have a 
beautiful demonstration of the independence of the two 
factors in fertilisation, or, in other words, of the two fune- 
tions of the spermatozoon, and that the two functions have 
been disturbed in unequal degree. At the same time the 
egg protoplasm, after the fourteen hours’ sojourn in 
unrenewed sea water, was approaching to that stage in 
which it acquires spontaneously the tendency to develop 
astral activities, and it might be held that the conditions are 
the same as In magnesium eggs in which, as a result of a 
general stimulation, asters and centresomes which are 
unconnected with the nucleus appear de novo in the 
cytoplasm. While the results are of interest in connection 
with the apparent independence of the factors in fertilisation, 
they also show how they are co-ordinated together. The 
sperm nucleus becomes dissociated from the aster, and fails of 
union, because it has not undergone the transformation which 
properly corresponds to the phase reached in the cycle of the 
centrosomal changes. Further, while the nuclei may be re- 
solved into chromosomes before union, and yet unite in the 
equatorial plate stage, a certain stage in the transformation 
of the dense mass of chromatin of the sperm head into 
a nucleus with distinct chromatin network, must be reached 
before union can take place. ‘This seems to show that several 
co-ordinated factors are at work in the nuclear conjugation. 

Further insight into the behaviour of the factors in 
fertilisation is given by an experiment described by Ziegler 
(1898). This consisted in carrying newly-fertilised eggs by 
a gentle current of water in his compressorium against 
threads of cotton wool. The egg was caught on a thread 
and nearly cut through, leaving only a slender bridge of 
protoplasm between the two portions of the egg. ‘The one 
contained the sperm nucleus, the other the germ nucleus. 
While the sperm nucleus regularly divided, followed by 
division of the cytoplasm, the egg nucleus merely underwent 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 499 


alternate changes of disintegration and reconstruction with- 
out division of the cytoplasm. Radiations appeared and dis- 
appeared, and after three cycles, owing to the segmentation 
of the portions containing the sperm nucleus, the ege-nuclear 
portion became detached and disintegrated. In another 
experiment the egg-nuclear portion underwent changes of 
form suggesting abortive attempts at cleavage. The mitotic 
transformation of the egg nucleus was not synchronous with 
that of the sperm nucleus, but always a little behind. 

These observations show that under the conditions of the 
experiments the egg nucleus is excited to division without 
direct contact with the sperm nucleus cr aster, but that the 
mitotic phenomena are ineffective to produce cytoplasmic 
cleavage. Ziegler refers this to the general stimulation of 
the egg by the spermatozoon, manifested also by the throw- 
ing off of the vitelline membrane. Boveri has shown that 
the same phenomena occur in egg fragments produced by 
shaking some minutes after fertilisation, and he (1902) refers 
to cases of this kind in which he has observed divisions of 
the nucleus followed by cell cleavage. The division was 
repeated a second time, and thus the four-cell stage was 
reached, but development then ceased. Another example of 
the effect of this general stimulation is to be seen (Boveri, 
1902) in the cases in which the egg is incited to throw off the 
polar bodies by the entrance of the spermatozoon. 

This brings me to a further reference to Wilson’s observa- 
tions on etherised eggs (1901 B). As has already been said, 
under this treatment the sperm and germ nuclei remain 
apart and undergo independently karyokinetic transforma- 
tion. ‘The most striking fact is that, while the sperm aster 
often gives rise to a perfect and symmetrical bipolar figure, the 
ego nucleus in a great number of cases produces a monaster, 
which seems at first incapable of resolving itself into a 
bipolar figure.” In typical cases the ege nucleus gives rise 
to a monaster such as described by Hertwig (’96), and such 
as occurs In magnesium eggs. While the egg monaster does 
not at first give rise to a dicentric figure, it does so later, as 


500 THOMAS H. BRYCE. 


may be gathered from the description of an egg continuously 
observed in the living state. At the height of its develop- 
ment the egg monaster lay at one side, the sperm amphiaster 
at the other, and no spindle was formed between them. The 
ege divided into three cells, two larger and somewhat 
irregular containing two daughter sperm nuclei, and a small 
one in which the single egg nucleus re-formed. At the second 
division each of the sperm nuclei gave rise to a perfect amphi- 
aster, and divided into two, the accompanying cytoplasmic 
division resulting in the formation of two complete cells and 
one binucleate cell. The single egg nucleus gave rise to a 
tetraster, and divided into three cells, one binucleate, the 
nuclei of the latter quickly fusing together. The embryo 
now consisted of six cells—three containing maternal, three 
paternal nuclei. At the ensuing division fifteen cells were 
formed, of which eight larger ones contained paternal nuclei, 
while seven much smaller ones containing maternal nuclei lay 
in a definite group at one side. The egg observed afterwards 
died. Wilson has not seen an egg monaster become dicen- 
tric at the first division, but the above observations prove 
that it may operate as an effective division centre, without 
establishing a spindle connection with either of the sperm 
asters, and that it may divide later. A centrosome was 
demonstrated in the monaster, in the same form as in the 
sperm aster, and as in magnesium eggs. The possible action 
of the chemical as the exciting agent of the karyokinetic 
transformation was excluded by control experiments, and it 
was therefore concluded, that it was due to a stimulus effected 
by the spermatozoon, as in Ziegler’s experiment. These 
observations, added to the results obtained in the magnesium 
eggs, “demonstrate that under appropriate stimulus the egg 
inay give rise to a centrosome capable of progressive division, 
but the etherised eggs show in the clearest manner that this 
centrosome is less effective than the sperm centro- 
some.” 

I shall not venture on the general problem of the asters and 
centrosomes. It will suffice for the present purpose if it be 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 501 


admitted that, without prejudice to the question either of the 
individuality, or the persistence of the centrosome, the body 
and its aster represent a kinetic phase of protoplasm, which 
reveals itself in cycles of activity, and that the centrosomes 
and asters constitute together, in some sense, a divisional 
apparatus, though that term is not used in any definite 
mechanical sense. 

The egg both before and after maturation lacks the power, 
for and by itself, to produce in normal circumstances such a 
divisional apparatus as will regularly and equally divide the 
cell. 

In Ziegler’s and Boveri’s experiments on separated portions 
of the egg containing only the ege nucleus, a divisional 
apparatus is called up under the general stimulation of the 
spermatozoon ; but it is ineffective, or only very partially 
effective. In the etherised eggs it is slow in appearing, and 
less effective than that associated with the sperm nucleus. 
In magnesium eggs the effect of the disturbance of equilibrium 
is to cause a change of state in the protoplasm which results 
in the differentiation at many foci of kinetic centres, and it is 
only in the cases where a single such centre, which divides 
into two, appears in the nuclear area, or at most two centres, 
that normal division proceeds. 

In fertilisation there is only one kinetic centre, and this is 
localised on the middle piece of the spermatozoon. Its 
activities are rapidly unfolded, and dominate all the other 
latent astral activities of the egg. ‘The latent capacity of 
both nucleus and cytoplasm to give rise to centrosomes is in 
this case wholly inhibited.”’! By union of the nuclei its 
activity is transferred to the cleavage nucleus, and “ becomes 
a part of an activity on the part of the ege nucleus that 
would have ensued even had the germ nuclei not united.”! 

Thus it may be said that the spermatozoon supplies the 
lack in the egg, by providing a powerful and effective 
“divisional apparatus.” How is this effected? Does the 
spermatozoon act by giving a general or diffused stimulus 


1 Wilson, 1901 a, pp. 581, 582. 


502 THOMAS H. BRYCKH. 


to the ege, or by disturbing the general equilibriam in some 
such way as the loss of water does, when the normal osmotic 
relations are disturbed ? Or does the spermatozoon carry into 
the egg some specific chemical substance which produces a 
local differentiation, of which the centrosome and the aster 
are the expression? Or does it import “ a highly active centro- 
some or centroplasm about which the cytoplasmic energy is 
brought to a focus?” 

Boveri (1902) holds that a general stimulation of the egg, 
with the sperm head as the point of predilection for the 
formation of the aster, as in magnesium eges the ege nucleus 
is the point of predilection, is insufficient as an explanation. 
There is much rather something special present in the sperma- 
tozoon, which determines that the aster shall appear at that 
point, and that poimt only ; and thus he thinks that still the 
appearances may best be described as being due to the intro- 
duction of a centrosome. Hven admitting—which, as has just 
been indicated, he does not—that the spermatozoon acts like 
Loeb’s agents, and in view of the demonstration by Morgan 
and Wilson that their effect is to cause the egg to produce 
centrosomes de novo, only a modification of secondary 
importance would be required in his theory of fertilisation, 
viz., that instead of saying that the spermatozoon brings a 
centrosome into the egg, it would be necessary to say that it 
causes the formation of a centrosome in the egg, from the 
division of which the rest follows. 

Taking the sperm aster as the manifestation of activities 
produced by the spermatozoon, and looking to its sharp 
localisation on the site of the middle piece, it seems reason- 
able to suppose that the localised excitement is the effect of an 
agent operative in fertilisation, and that it is probably related 
to the middle piece; but the actual continuity between the 
centrosome of the spermatozoon and that in the aster has not 
been absolutely demonstrated, and the new facts in regard to 
the centrosome put the matter in another light. Thus it 
remains for the future to decide which of the two latter 
alternatives stated above shall be adopted, and perhaps after 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 505 


all there is only a formal difference, for the fundamental 
problem is the same when the question is raised how the 
centrosome exercises its activities. 

Loeb (1901), on the physico-chemical side, suggests that a 
catalytic substance is carried by the spermatozoon into the 
egg, that is one which accelerates physical or chemical 
processes which would occur without it. The K ions act as 
catalysers, and the loss of water acts also, though less 
directly, in the same way, and it may be that it gives rise to 
substances which act catalytically. Inasmuch as in Cheetop- 
terus the normal development does not show the character- 
istics of a treatment of the eggs by K, it is probable that 
normal fertilisation is not brought about by K ions. 

Delage (1901) considers that the egg is in an unstable 
state of equilibrium, which is readily upset by various 
agencies—loss of water, heat, etc., and he lays some weight 
on the specific action of the salts. He finds that the 
chloride of manganese has, for Asterias, a specific action 
superior to that of the alkaline salts. ‘Together with his son, 
he showed that in the case of the sea-urchin there was less 
magnesium chloride in the sperm than in the eggs, by about 
1 per cent., so that this salt could not have a specific action. 

Among other possible factors in the action of the sperma- 
tozoon, he gives prominence to its abstraction of water from 
the cytoplasm, During maturation the nuclear sap is shed 
into the cytoplasm; until this is effected by the solution of 
the nuclear membrane, fertilisation is not possible ; it is just 
at this “critical stage” in Asterias that he finds artificial 
parthenogenesis most hable to occur. In the specialisation 
of the sexual elements, the egg thus becomes rich, while the 
spermatozoon has become poor, in water. After the sperm 
head has entered the ovum it increases in size by abstraction 
of fluid from the egg protoplasm, and this abstraction of 
water by the sperm nucleus has to be reckoned with as a 
possible factor in fertilisation. 

Apart from the large assumptions involved in such an 
hypothesis, the facts of ‘‘ partial fertilisation, and the local- 

VOL. 46, PART 3.—NEW SERIES, I] 


504. THOMAS H. BRYCE. 


isation of the aster on the middle piece, are in opposition 
to it.” 

It has been suggested that the centrosome is the seat of 
formation of a ferment. Mathews (1901), from the results of 
his experiments on the eggs of Arbacia, believes that “ what- 
ever the details of the process may prove to be, the essential 
basis of karyokinetic cell division is the production of localised 
areas of liquefaction in the protoplasm.” “The centrosome 
might be a liquefying enzyme.” 

Experiments on this line have been tried, but without 
definite result. Pieri’s results (1899), from which he sup- 
posed he had obtained a ferment ‘ ovulase,” have not been 
confirmed. Dubois (1900) showed that there was no 
question of a ferment being obtained by Pieri’s methods. 
He made various experiments on sperm and eggs, from 
which he concluded that there was evidence of the existence 
of a “ zymase,” which he provisionally named ‘ Spermase,” 
in the spermatozoa, and in the egg a substance, at least 
modifiable by ‘‘spermase,” provisionally named “ Ovulase.” 
Spermase cannot enter the egg by diffusion or osmosis, but 
only by a mechanical means, which is the raison d’étre of 
the spermatozoon. Winkler’s experiments (1900) are also 
inconclusive. He used sperm shaken for half an hour in 
distilled water and filtered five or six times through three- 
fold filter-paper. ‘The filtrate was added to sea water, the 
precaution being taken of keeping the mixture at the same 
degree of concentration as the sea water. While the sperm 
in heated sea water produced no results, the liquid caused in 
the case of Spherechinus and Arbacia eggs, though in a rela- 
tively small number, the beginnings of segmentation. ‘These 
results may have been due to osmotic influences, 

Loeb (1900) states that up to that date he had found no 
enzyme save papain which had an effect in causing the egg 
to segment, and it was uncertain whether this was not due 
to some accidental constituent of the enzyme preparation used. 
Gies (1901) made a complete study of the effects of extracts 
of sperm made by the ordinary methods for the preparation of 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 505 


enzyme solutions. His results were wholly negative, and he 
concluded that, used in certain proportions and under certain 
conditions at any rate, such extracts did not possess any 
power of causing proliferation of the ripe ovum. No 
evidence could be furnished of the existence of a zymogen 
in spermatozoa. Extracts of fertilised eggs in the earlier 
stages of development seemed likewise devoid of any seg- 
mental activity. ‘he results, Gies adds, do not, however, 
certainly show that enzyme action is impossible, after, or at 
the time of union of the spermatozoon with the ovum, within 
the latter. 

The same negative result was got this spring by R. T. 
Lieper at Millport Marine Biological Station, using an 
extract of sperm prepared by spreading fresh sperm on 
sheets of glass, then drying in air and sun, and afterwards 
triturating the dried extract in sterilised sea water. The 
filtrate from this fluid produced no segmentations, though 
control experiments with eggs from the same ovaries 
normally fertilised, nearly all developed. 


LITERATURE. 


Artota.— Atti Soc. Ligrest. Sc. Nat. e Geogn.,’ ann. xii, fase. 3. 
Bataruton, 1900.—‘ Compt. Rendus de |’Acad. des Sciences,’ t. exxxi, p. 115 
BararLton, 1901.—‘ Compt. Rendus de l’Acad. des Sciences,’ t. exxxii, pp 
852, 1134. 

BatarLion, 1902.—‘ Compt. Rendus de l’Acad. des Sciences,’ t. exxxiv, p.918 
Barattton.— Archiv fiir Entwickelungsmechanik,’ t. xi, Heft 1. 
Batattton.— Archiv fiir Entwickelungsmechanik,’ t. xii, Heft. 4. 

Bonnet, 1900.—Merkel und Bonnet- Ergebnisse u. s. w.,’ 1900, 


Bovert, 1888.—‘ Sitzungsber. des Ges. f. Morph. und Physiol. in Miinchen, 
Bd. iv, Heft 2. ; 


Bovert, 1901.—‘ Zellen Studien,’ Heft 4, Jena, 1901. 
Bovert, 1902.—* Das Problem der Befruchitung,’ Jena, 1902. 
Bryce, 1902.—‘ Quart. Journ. Mier. Sci.,’ vol. xlvi. 
Buuier, A. R., 1902.—‘ Quart. Journ. Mier. Sci.,’ vol. xlvi. 


506 THOMAS H. BRYCE. 


DetaGe, 1899.—* Archiv. de Zool. Expér.,’ t. vii. 

Decace, 1901.—‘ Rev. Gen. des Sciences,’ 12 année, No. 19. 

Dexace, 1901.—‘ Archiv. de Zool. Expér.,’ t. ix, Nos. 2, 3. 

DorteEin, 1897.—‘ Arch. f. mikroskopische Anat.,’ Bd. 1. 

Dusors, 1900.—‘ C. R. Soe. Biol.,’ Paris, vol. lii, p. 197. 

Fiscner, 1902.—-‘ Amer. Journ. of Physiology,’ June 2nd, 1902. 

Granrp, 1900.—‘C. R. Soe. Biol.,’ Paris, vols. lii, liii. 

Giarpina, 1902, a.—‘ Anat. Anzeiger,’ Bd. xxi, No. 20. 

Grarpina, 1902, d.—‘ Anat. Anzeiger,’ Bd. xxii, Nos. 22, 33. 

Giss, W. J., 1902.—* Amer. Journ. of Physiol.,’ vol. vi. 

Greerr, It., 1876.—* Sitzungsber. der Ges. zur Beford. der gesammten Natur- 
wiss. zu Marburg,’ No. 5, quoted from O. Hertwig, 1890. 

Greevey, A. W., 1901.—‘ Amer. Journ. of Physiology,’ vol. vi, p. 296. 

Hertwic, O. and R., 1887.—*‘ Jenaische Zeitschrift fiir Naturwissenschalt,’ 
BileexXe 

Hertwie, O., 1890.—‘ Jenaische Zeitschrift fiir Naturwissenschaft,’ Bd. xxiv 

Hentwie, O., 1893 —*‘ Die Zelle und die Gewebe.’ 

Hertrwie, R., 1896.—* Uber die Entwick. des unbefurchteten Seeigeleies 
Leipzig, 1896. 

Hunter, 8. J., 1901.—‘ Amer. Journ. of Physiology,’ vol. vi, p. 177. 

Lorp, J., 1900.—‘ Amer. Journ. of Physiology,’ vol. ili, No. 9. 

Lorn, J., 1900.—‘ Amer. Journ. of Physiology,’ vol. iv, p. 178. 

Lors, J., 1901.—‘ Amer. Journ. of Physiology,’ vol. iv, p. 423. 

Logs, J., 1902.—* Archiv f. Entwickelungsmechanik,’ vol. xiii, Heft 4. 

Lorx, J., 1902.—‘ Archiv f. Entwickelungsmechanik,’ vol. xiv, Heft 4. 

Maas, O., 1901.—‘ Sitzungsber. d. Ges. f. Morph. und Physiol. in Miinehen,’ 
Bd. xvii. 

Matuews, A. P., 1900.—‘ Amer. Journ. of Physiology,’ vol. iv, p. 3438. 

Matuews, A. P., 1901.—‘ Amer. Journ. of Physiology,’ vol. vi, p. 142. 

Marturws, A. P., 1901.—‘ Amer. Journ. of Physiology,’ vol. vi, p. 216. 

Meves, I’., 1902.—‘ Verhand. des Anat. Ges. Halle,’ 1902, p. 132. 

Morean, I’. H., 1896.—*‘ Arch. f. Mntwickelungsmechanik,’ Bd. iii Heft 3. 

Moreay, I’. H., 1899.—* Arch. f. Entwickelungsmechanik,’ Bd. viii, Heft 3. 

Moreay, I’. H., 1900.—‘ Arch. f. Eutwickelungsmechanik,’ Bd. xii, Heft 2. 

Norman, W. W., 1896.—‘ Arch. f. Entwickelungsmechanik,’ Bd. iii, Heft 1. 

Prent, J. B., 1899.—* Archiv. de Zool. Expér., Notes et Revue,’ vii, 3. 

Prowazek, S., 1900.—*‘ Zool. Anzeiger,’ No. 618, p. 358. 


ARTIFICIAL PARTHENOGENESIS AND FERTILISATION. 507 


TrIcuManN, Ernst, 1902.—‘ Jen. Zeitschrift fiir Naturwiss.,’ n. F., Bd. xxx, 
Heft 1. 

VicurER, 1900.—‘ Compt. Rendus Acad. Sci.,’ Paris, t. exxxi, p. 118. 

Vicuier, 1901.—‘ Compt. Rendus Acad. Sci.,’ Paris, t. exxxii, p. 1436. 

VicuIER, 1902.—‘ Compt. Rendus Acad. Sci.,’ Paris, t. exxxv, p. 60. 

VieurEr, 1902.—‘ Compt. Rendus Acad. Sci.,’ Paris, t. exxxv, p. 197. 

Wepexinp, 1901.—‘ Ber. tib. der Verhand. d. 5 internat. Zool. Congress,’ 
Berlin, 1901. 

Wutson, Ep. B., 1901.—‘ Arch. f. Entwickelungsmechanik,’ Bd. xii, Heft 4. 

Witson, Ep. B., 1901.—-‘ Arch. f. Entwickelungsmechanik,’ Bd. xiii, Heft 1. 

Wink Ler, Hans, 1900.—‘ Nachr. K. Ges. Wiss. Gottingen, Math. Phys. K1.,’ 
1900, Heft 2, p. 187. 

ZiEGLER, H. E., 1898.—‘ Arch. f. Entwickelungsmechanik,’ Bd. vi, Heft 2. 


voL. 46, PARY 3.—NEW SERIES. KK 


4 ret fe T’ 143. Sno pecall qn *— O00L ft 
it iierey ale’ ieee Sg ‘ier a 


Ps wt 
AY 2xx9 Tid teal nee Ib to, 200F 


ke th ata )\aiieeeataa fiw J.otk Sn fOCL ah eas 
ltalé jie inoe ae ase A nlvald 000! Metal L ohs 


area hha arehelctiere SE APM ETSA 


Wad awed Fat oh Gael Maat a EAR 
; t folk 


8 aye otto fl rad OR em 


Stew e 3b Seed ah co oll =F 


oer i 


st enataodsy atid? of A POOR EE ot, 


NOL i 8 thot COURS 
TL", daadoomapuistciwiad 1 Jord SCRE ot ae 


- 


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 SupptemMent 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. 

Each Number is sold at 10/-, and Four Numbers ria up a 
Volume. 


London: J: & A. CHURCHILL, 7 Great Alarberougs = 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 eatra 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/- 
dU, ty, é : ; : 6/- 
100s 7/- 


Plates, 2/- per 25 if uncoloured ; it coloured, at the same rate for 
every colour. 
Prepayment by P.O. Order is requested. 
ALL COMMUNICATIONS FOR THE EDITORS TO BE ADDRESSED 10 THE CARE 
or Messrs. J. & A. CuurcHintt, 7 GREAT MARLBOROUGH SrREBT, 
Lonpon, W. 


~ THE MARINE BIOLOGICAL ASSOCIATION 


| UNITED KINGDOM. 


Patron—HIS MAJESTY THE KING. 


President—Prof. E, RAY LANKESTER, LL.D., F.R.S. 
03 


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, &e. 


Works published by the Association 


include the following :—‘ A Treatise on the Common Sole,’ J. ‘I’. 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. 


ANNUAL MEMBERS . F 3 . £1 1 Oper annum. 

Lire MEMBERS . : : : . 15 15 O Composition Fee. 
FOUNDERS . : : : i - + 100%28:"0 ee a 
Governors (Life Members of Council) 500 0 0 + Ms 


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. 


MAR La | 


New Series, No. 184 (Vol. 46, Part 4), Price 10s. 


“\S 20) FEBRUARY, 1903. 
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 ROMER, 

AND OF THE AMERICAN ACADEMY OF ARTS AND SCIENCES OF BOSTON 4; 

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 3 


W... E.R WELDON... MAG FERS. 


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


AND 


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


BEYER PROFUSSOR OF ZOOLOGY IN THE OWENS COLLEGE, MANCHESTER. 


WITH LITHOGRAPHIC PLATES AND ENGRAVINGS ON WOOD. 


Ji 


LOND ON : 
J. & A. CHURCHILL, 7, GREAT MARLBOROUGH STREET. 
1905. 


Adlard and Son,] [London and Dorking. 


CONTENTS OF No. 184.—New Series. 


MEMOIRS: 


The Movements and Reactions of Fresh-water Planarians: a Study 
in Animal Behaviour. By Raymonp Peart, Ph.D., Instructor in 
Spee in the University of eae Ann Avnae Michigan, 
U.S.A. : : ; : , : ‘ 


On the Dividers IV. On the Central Complex of Cephalo- 
discus dodecalophus, Mcl. By A. T. Mastrerman, M.A,, 
DSc; Lectur er on Zoology, School of Medicine, Edinburgh. (With 
Plates’ 32, 33) : : : _ : : : 


On Hypurgon Skeati, a New Genus and Species of Compound 
_Ascidians. By Icerna B. J. Sorzas, B.Se.Lond. (With Plates 
34,35) Sy bgt Bees ee 


The Anatomy of Arenicola assimilis, Ehlers, and of a New Variety 
of the Species, with some Observatious on the Post-larval Stages. 
By J. H. Asuwortn, D.Sc. (With Plates 36,°37) 


TitLe, Contents, aNnD INDEX. 


wt ———KLK 


PAGE 


509 


715 


729 


737 


MAR 12 1903 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 509 


The Movements and Reactions of Fresh-water 
Planarians: a Study in Animal Behaviour.' 


By 
Raymond Pearl, Ph.D. 


(Instructor in Zoology in the University of Michigan, Ann Arbor, 
Michigan, U.S.A.) 


ConrTENTS. 
PAGI 
A. INTRODUCTION . : : i s Bill 
B. RESUME OF LITERATURE . : : : 5 le 
I. Morphological and Systematic — . ‘ +, BLS 
II. Physiological ; : : ; . 520 
c. MarTEeriaL : , 523 
p. Hapits anp Natura. History : t ; 25 
I. Occurrence and Distribution E . 526 
II. Activities 3 ‘ E 3 = Hey 
a. Sensitivity : ; : : » b27 
4. Secretion of Mucus : 2 : . 529 


c. Periods of Activity and Rest 
d. Formation of Collections 


5 

5 

e. Movement on Surface Film : m5 
III. Food ; ' : : 4 i 
5 

5 

5 


“ot Ee Cs 


LV. Defecation ‘ : 
V. Summary of Factors in Behaviour 
rE. Normat Motor Acrtivitirs 


Mm CO Oo to OO Ww OH WH OH CS DW WO WD DW ®W 2 
S wo 


I. Locomotor Movements 539 
a. Gliding : F 539 

1. Rate of Gliding Movement 545 

2. Direction 548 


' Contributions from the Zoological Laboratory, University of Michigan, 
Ann Arbor, Michigan, No. 58. 


von. 46, PART 4.—NEW SERIES. LL 


510 RAYMOND PEARL. 


PAGE 
b. Crawling Movement : ; ; . 548 
1. Direction . ; : ~ boo 
Y, Stimuli which induce Graminge 5 Syl! 
c. Movement on the Surface Film : 552 

d. Relation of the Movements of Triclads to tigse of ones 
Forms : - = 553 
Il. Non-locomotor Mawaments ' : DDD 
. Contraction of the Body. : ; . 555 
i Extension of the Body —. ; : . 556 
@ Rest, ; ; . sbby 
1. Formation of Walleetions ; ; . 566 
d. The Effect, of Operations on Movement : . 570 
F. REACTIONS TO STIMULI. F ¢ - 576 
T. Reactions to Mechanical Stimnuly : : /. 696 
a. Methods : 3 : ; 5 56 
). Description of Reactions 577 
1. Reactions to Stimuli applied to tie Head Region 577 
a. Reactions to Strong Stimuli 578 
B. Reactions to Weak Stimuli 582 

2. Reactions to Stimuli applied to the Middle Region 5 
the Body 4 . 588 
a. Reactions to Strong Stimuli ‘ . 588 
B. Reactions to Weak Stimuli ‘ 589 

8. Reactions to Stimuli applied to the Posterior Bevin 
of the Body. A 1og2 
4. Reactions to Stimulation of the Ventral Statice . 594 

5. Reactions of Resting Specimens to Mechanical 
Stimuli : . 595 

6. Reactions to Stimuli given by Open Proce- 
dure . 5 DUo 
7. The Effect of Meehaniedt Emaeaue to Maveineah 5 BS) 

ce. The General Features of the Reactions to Mechanical 
Stimuli. ; , . 600 
d. The Mechanism of the mason : . 602 
1. The Relation of the Brain to the Baactions . 602 
2. The Neuro-muscular Mechanism 606 

e. Features in the General Behaviour of the rgatiem mined 
the Reactions to Mechanical Stimuli explain . 619 
J. Summary : ‘ . 693 
Il. Reactions to Food and Ghemtcate Sismalt ; . 623 
. Food Reactions . . 694 
1. Food Reactions of Spaiment aiden Saratinid . 637 


2. Summary of Food Reactions. 5 . 640 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 511 


PAGE 
6. Reactions to Chemical Stimuli—Chemotaxis . > 643 

1. Reactions to Localised Chemical Stimuli . . 643 

a. Methods : : . . 643 

B. Results A : : . 646 

2. General Summary : Se Oni! 

3. Unlocalised Action of @henen's : . 669 

IT. Thigmotaxis and the Righting Reaction . . 670 
a. ‘Thigmotaxis ‘ ‘ : 2 600 

b. The Righting Reaction —. d : 2) 83 

The Mechanism of the Reaction — . : . OG 

ec. Summary : : 3 : . 684 

IV. Electrotaxis : ; : : . 685 

. Methods 4 : 2 ; . 685 

; Results ; : f = 685 

ec. Mechanism of the Reachions F : . 690 

d, Summary : ‘ : : mE OOD 

V. Reaction to Desiccation . : : ; (695 
VI. Rheotaxis : 697 
G. GENERAL SUMMARY AND Drecoetee OF esters : . 698 
1. List oF LITERATURE : . ? : YG 


A. INTRODUCTION. 


THe present study has for its purpose the analysis of the 
behaviour of the common fresh-water planarian into its com- 
ponent factors. It is well known that, aside from the 
researches of a few investigators on a small number of forms, 
we have little detailed knowledge of the behaviour of lower 
organisms. It is coming to be realised, too, that knowledge 
of what an animal does is just as important in the general 
study of life phenomena as a knowledge of how it is con- 
structed, or how it develops. But it must be admitted that 
until quite recent times the study of the activities of living 
things was a much neglected field in biology. ‘lhe 
publication of the ‘Origin of Species’ gave the biological 
pendulum a swing towards the study of phylogeny, from 
which it is only just beginning to return. 

As a consequence of this concentration of interest on other 
subjects, we possess an accurate and full knowledge of the 


512 RAYMOND PEARL. 


activities of very few lower organisms. ‘The behaviour of the 
Protozoa has been quite fully described and analysed by the 
work of Verworn (’89) and Jennings (’97, 799, 99a, ’99b, 
’99e, : 00, : 00a, :00b, :00c,: 01, Jennings and Moore : 02). 
In the earlier work of Verworn the general features of most 
of the reactions of the Protozoa are described, special atten- 
tion being paid to the rhizopods. The reactions of the 
Infusoria have been very thoroughly worked out by Jennings. 
In the case of the Infusoria we now know exactly the 
mechanism of the reaction to a large number of stimul. The 
reactions and general behaviour in the case of two groups of 
echinoderms are quite thoroughly known from the early work 
of Preyer (86, ’87) on the starfish and the recent brilliant 
work of von Uexkiill (96, ’96a, 799, :00, : 00a) on the sea- 
urchin. These few instances are the only ones in the 
literature where the movements and reactions of an organism, 
or group of organisms, have been investigated in any com- 
prehensive ‘‘ monographic ” way, There is a great body of 
literature dealing with isolated reactions in a variety of forms, 
but the thorough investigation of the activities of animals in 
a way comparable to that in which their morphology has 
been investigated remains in large degree yet to be done. 

It, appeared highly desirable that this sort of knowledge be 
extended, and it was with this idea in mind that this work 
was undertaken. The form used, Planaria, was chosen for 
several reasons. In the first place, it has come to be a sort of 
paradigm for work on regeneration, and its biology from that 
standpoint is already well known. Furthermore, in some one 
or more of its species it is an almost universally distributed 
form and can always be ebtained in quantities. Finally, and 
particularly, it is a representative of an animal type about 
whose activities we know only the most general facts. It is 
a symmetrical aquatic organism of low organisation, and its 
behaviour is rather complicated. The importance of possess- 
ing a detailed knowledge of the activities of a bilaterally 
symmetrical, free-moving, low organism will be apparent 
when it is considered that such an organism has never been 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 015 


made the subject of such a study. The behaviour of typically 
unsymmetrical organisms, the Infusoria, has been analysed, 
as has also that of some radially symmetrical animals, and in 
both cases there is found to be a very close interrelation- 
ship between the general form of the body and the reactions. 

To investigate, then, in a comprehensive way the activities 
of a bilaterally symmetrical organism standing low in the 
animal series was the purpose of this work. ‘The most 
general problem which presents itself is the establishment 
of the animal’s position in the objective psychogenetic series. 
Are its activities relatively simple or are they complex ? Do 
they fall under the same general type as those of the 
Infusoria or those of the higher organisms, or do they occupy 
an intermediate position? Another general problem of im- 
portance is whether there is any marked correlation between 
the behaviour and the form of the body, such as has been 
found to obtain in so marked a degree in the case of the 
Infusoria and the rotifers (vide Jennings, loc. cit.). We 
have in the flat-worm a symmetrical animal; are its reactions 
of a symmetrical type? Besides these broad fundamental 
problems there are, of course, a large number of subsidiary 
questions which readily suggest themselves in connection 
with a work of this sort. These need not be specifically 
mentioned here, but will be brought out in the course of 
the paper. 

As to the scope of the work as actually done, the following 


” of the animal 


may be said:—The general “ natural history 
was studied as completely as possible. All the normal move- 
ments were studied in detail. ‘The reactions to mechanical 
stimuli; the food reactions and reactions to chemicals in 
general ; electrotaxis ; thigmotaxis; rheotaxis; the righting 
reaction; the reaction of cut and regenerating pieces ; and 
hydrotaxis and the reactions during desiccation were investi- 
gated. No work was done on the phototaxis or thermotaxis, 
A study of the phototaxis was omitted for two reasons ; first 
on account of the fact that during the progress of this in- 
vestigation Parker and Burnett (: 00) reported their work on 


o14 RAYMOND PEARL. 


the same subject, and furthermore on account of lack of 
opportunity. As a result of some incidental observations 
made during the course of this work, it has appeared that it 
would be profitable to extend the work of Parker and 
Burnett, and this, together with a study of the thermotaxis, I 
hope to be able to do in the future. Another field for further 
work is afforded in the study of the reactions of regenerating 
individuals. As this subject did not fall immediately into 
the general plan of this work, but comparatively little atten- 
tion has been given to it, yet the work done gives much 
promise of important results to be gained by more extended 
investigations. 

So far as possible the details of the movements and 
reactions will be described fully. It is not easy to see why 
there is not as much need for a complete knowledge of 
details in physiological work as in morphological, yet in 
much of the recent work in comparative physiology only the 
most general results are reported. To gain a knowledge of 
the details one must do the work over again. While such 
more or less general papers are easy to read, and put the main 
results in such a form as to be easily accessible, yet it is 
believed by the writer that the solid foundations of com- 
parative physiology and psychology must consist of detailed 
“fine” work, just as has been the case in morphology. It 
seems to the writer that the tendency to abandon the detailed 
descriptive method in favour of the extreme experimental 
method in biological work is unfortunate. Both ways of 
working are methods of getting at the truth, and, as proven 
by their results, both are good methods. The current notion 
of the sufficiency of the experimental method to the exclusion 
of others is not only an evident exaggeration of the facts in 
the case, but, in the opinion of the writer, the exclusive use 
of the “ crucial-experiment ”? method in work upon the move- 
ments and reactions of organisms has in some cases hindered 
rather than helped us to gain a clear understanding of the 
phenomena. ‘The importance of close observational work in 
the study of animal behaviour has been strongly emphasised 


MOVEMENTS, BTC., OF FRESH-WATER PLANARIANS. 515 


recently by Whitman (99). The aim in the present work 
has been to get as extensive and detailed a knowledge as 
possible of the behaviour of the organism by direct observa- 
tion before resorting to experiments. 

At this point I wish to acknowledge my indebtedness to 
the officials of the laboratory in which this work has been 
done. To Professor H. 8S. Jennings, under whose general 
oversight this investigation has been prosecuted, I wish to 
extend my heartfelt thanks for his uniform kindness in 
freely giving advice, suggestion, and kindly criticism of im- 
measurable value. Any adequate expression of my indebted- 
ness to him is impossible. I further wish to express my 
thanks to Professor Jacob Reighard for the numerous 
facilities which I have enjoyed during my stay in his 
laboratory, and for his kindly interest, which has made work 
there a pleasure. Finally, I desire to acknowledge my in- 
debtedness to Professor F. C. Newcombe, of the Botanical 
Department of the University of Michigan, for many 
valuable suggestions and advice. 


B. Resume or LITERATURE. 


But little has been done on the physiology of the move- 
ments or on the psychology of the ''urbellaria, and, as in 
the case of most of the literature dealing with these subjects, 
what has been done has been in comparatively recent years. 
Investigators of the old ‘ natural-history ” school which 
flourished before the time when Darwin’s work changed the 
course of zoology seem not to have given much attention to 
planarians, while the later systematists and morphologists 
for the most part carefully avoided any reference to the 
activities of the forms which they studied. 


I. Morphological and Systematic. 


Among the papers devoted primarily to the systematic or 
morphological treatment of the group, there are occasional 
references to points in the behaviour of the organisms which 


516 RAYMOND PEARL. 


are of importance from the present standpoint. Among 
such references the following may be noted : 

Moseley (’74), in a paper concerned principally with the 
anatomy and histology of the land planarians, devotes a sec- 
tion to a discussion of the habits of these forms. He com- 
ments on the “avoidance of light” (negative phototaxis) of 
land and aquatic planarians, and discusses the habitat and 
food of the animals. He reaches the conclusion that all 
planarians are carnivorous, but gives no account of the 
method of feeding. He quotes Rolleston as having found 
that Planaria torva and Dendrocelum lacteum in a 
dish in which had been placed a freshly killed earthworm 
“crowded on to the worm’s body and soon sucked all the 
hemoglobin out of it, leaving it white and pulpy.” Brief 
mention is made of the habit of the land planarians of 
secreting a mucous thread and hanging from it as a mollusc 
does. Finally, the method of movement of Bipalium with 
the head raised and waved from side to side as the animal 
proceeds is described. <A bibliography of previous literature 
is given. 

In another paper Moseley (77, pp. 273, 274) gives an 
account of the movements and general habits of Geoplana 
flava, a Brazilian species. This species was found to keep 
in shaded and moderately lighted places. The direction of 
the ciliary currents was tested by placing small bits of paper 
on the surface of the body, and it was found that when the 
animal was in active movement the effective beat of the 
cilia on the anterior part of the dorsal surface was forward 
and outward, while on the posterior portion of the dorsal 
surface the beating was backward and outward. ‘The 
currents on the ventral surface were always straight back- 
ward. ‘The author concludes that the function of the cilia 
on the dorsal surface is to quickly remove foreign bodies. 
When the organism was at rest there was no movement of 
the dorsal cilia; “the animal moves to a large extent by 
muscular action, the body alternately contracting and ex- 
panding during motion. When moving it lifted its anterior 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 9517 


extremity often, ... . .and moved it to and fro as if to fee| 
or see its way.” “ When the anterior extremity of the body 
was cut off the remainder of the animal seemed still to move 
with definite purpose, avoiding obstacles and retreating from 
the hght, while the cut end was raised and thrust in various 
directions as if to search for an object on which to climb.” 

In a brief note Zacharias (?88) mentions the occurrence of 
Geodesmus terrestris between the lamelle of Agaricus 
deliciosus. Particular points mentioned are: the slow 
movement, characterised by the raised anterior end, and the 
hanging by a mucous thread after passing over the edge of 
a glass plate. Light stimulation of the anterior end with a 
needle induces a very strong contraction of the whole body. 

Gamble (93), in a systematic paper on marine Turbel- 
laria, describes briefly the movements of a number of 
species of rhabdocceles and triclads. 

Lang (’84), in his monograph on the polyclads, devotes a 
chapter to the habits, movements, and natural history of this 
group of planarians (loc. cit., pp. 631—641). While not 
done particularly from the physiological standpoint and not 
treating the subject experimentally, this work contains 
numerous valuable observations. Points especially treated 
are the habitat, colouration, food and method of feeding, 
defecation, movements, including swimming, copulation, 
respiration, regeneration, growth, and duration of life. ‘The 
details in the behaviour of the polyclads recorded in this 
monograph will be discussed later in connection with the 
points on which they have direct bearing. 

The most important paper dealing with the movements 
and general behaviour of planarians which I have been able 
to find in the literature is that of Lehnert (91). This work 
is principally devoted to an account of the biology of three 
forms of land planarians, viz. Bipalium kewense, B. 
kewense var. viridis, and Geodesmus bilineatus. 
Besides the work on these land forms, Lehnert also made 
some comparative studies on several fresh-water dendrocceles 
and rhabdoceeles. He gives an admirably full and detailed 


) 


518 RAYMOND PEARL. 


account of the movements of land planarians ; in fact, by far 
the best description of these phenomena in the literature. 
In this account the relation of. the movement to the mucous 
secretion from the ventral surface of the body is brought out 
in great detail. The principal factors in producing the 
movement in the case of the land planarians he gives as 
(a) ciliary movement on the ventral surface, (b) rhythmical 
contraction waves passing longitudinally over the ventral 
surface, (c) secretion of slime, and (d) snake-like movements 
ot the whole body. A comparison with the movements of 
other planarians (fresh-water) 1s made. In this connection 
it may be mentioned that Lehnert considered rhythmical con- 
traction waves passing along the ventral surface of the 
animal to be a factor in the movement of fresh-water 
planarians (Dendrocelum lacteum, Planaria_ poly- 
chroa, and Polycelis tenuis). This I am unable to con- 
firm from observations on the planarians which I have 
studied. This point will be discussed more fully later. The 
food and the method of taking food in case of the land 
planarians, Lehnert worked out very thoroughly. They were 
found to be carnivorous, and in the case of Bipalium the 
pharynx was capable of being stretched over a large piece of 
earthworm, so that it resembled a very thin transparent skin 
covering it. The relations to other phases of the environ- 
ment, e.g. air, water, temperature, light, solid bodies, etc., 
are described very briefly. 

Raspail (93), in a brief note, mentions the feeding of a 
planarian. 

Van Duyne (96) mentions briefly the movements of 
heteromorphic forms of Planaria torva (?). He found 
that the parts of two-headed individuals moved inde- 
pendently of each other, and that each piece would move 
away from the other until they had completely torn apart. 

Willey (97), in a brief note, describes the structure of a 
remarkable asymmetrical planarian, for which he proposes 
the generic name Heteroplana, having the left side of the 
body almost completely atrophied. Regarding the loco- 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 919 


motion of this remarkable form he says, “In Heteroplana 

. the locomotion is usually conducted in a somewhat 
one-sided fashion,’ and he furthermore figures the animal 
as constantly moving towards the left. It is to be regretted 
that no reference is made to how this form reacts to stimuli, 
as it would be of great interest to know whether the reactions 
are asymmetrical, and in general how they compare with 
the normal planarian type. 

A series of papers by Morgan (98, :00, : 01) contains 
numerous references of importance on the movements of cut 
and regenerated specimens of various fresh-water planarians. 
He finds (98), in confirmation of van Duyne, that in two- 
headed individuals each head tends to move in its own proper 
direction. In the case of a heteromorphic form with two 
heads pointed in opposite directions, this likewise held true ; 
but one component being stronger this determined the move- 
ment of the whole. The lack of movement in certain forms 
of cut pieces is also noted. In his :00 paper Morgan notes 
the readiness with which “ Planaria, sp.”! and Planaria 
maculata take food, although no account is given of the 
method of the feeding reaction. An interesting observation, 
and one of considerable theoretical importance, is also re- 
ported in this paper. In an individual split longitudinally in 
the median line from the posterior end forward, in which the 
two parts were united only by a small connecting band of 
tissue at the anterior end, it often appeared “ as though these 
pieces would pull apart, but as soon as the tension on the 
connecting band becomes too strong, the rest of the piece, by 
a sort of adaptive response, ceases pulling in its former 
direction.” In the most recent paper cited (:01) Morgan 
corrects a statement of Bardeen? regarding the feeding of 
Planaria. It is maintained (and I may mention at this 
point that my own observations agree entirely with those of 
Morgan) that Planaria ‘responds freely ” to food sub- 


1 Later identified by Woodworth as Planaria lugubris. 
= 'To be reviewed later. 


520 RAYMOND PEARL. 


stances not actually in contact with it. This point will be 
discussed in detail later. 

Lillie (: 01) brings out the fact that cut posterior parts of 
the body of Dendroccelum lacteum show very little 
movement, and in general fail to give the typical reaction to 
light after removal of the brain. 

Finally, there are accounts of the natural history and 
habits of various planarians in numerous “natural histories” 
and text-books. As such accounts are for the most part 
brief and of no great significance from our standpoint, they 
will not be referred to in detail. 


II. Physiological. 

The literature dealing with the planarians from a purely 
physiological standpoint is very meagre. Furthermore, for 
the most part it deals only with special phases of the 
physiology of these organisms, there being very little work 
attempting to bring the behaviour of planarians into relation 
with that of other forms. 

The most important work dealing experimentally with the 
physiology of the movements of flatworms which I have 
found is that of Loeb (?94).. The purpose of his work was to 
determine in how far the reactions of such low organisms as 
worms were dependent upon the brain. The planarians used 
were ''hysanozoon brocchil, and Planaria torva. In 
Thysanozoon he found that if the animal were quickly cut 
into two pieces transversely with a sharp scalpel or scissors 
the anterior piece crawled on undisturbed, while the 
posterior piece showed no further movement. The conclu- 
sion is then drawn that ‘‘ Die Spontaneitét der Progressiv- 
bewegungen ist also bei Thysanozoon eine Funktion des 
Gehirns.” This form shows no definite ‘ geotropic ” reac- 
tion, but crawls about with the axes of the body forming any 
angles with the direction of gravitation. The very strong 
“ stereotropism ” (thigmotaxis) of the ventral side of Thy- 
sanozoon, which always tends to keep in contact with a 
solid body, is noted. ‘This reaction is found to be inde- 


MOVEMENTS, ETC., OF FRESM-WATER PLANARIANS. 521 


pendent of the brain. There was found to be co-ordination 
between the anterior and posterior pieces of a worm in 
which the lateral longitudinal nerves had been cut, but in 
which a narrow connecting strip of tissue had been left 
between the pieces. In P. torva Loeb states that posterior 
parts of the body which have been separated by a transverse 
cut from all connection with the brain crawl “ ebenso munter 
weiter, wie die orale Halfte.’” The reaction of this form to 
changes of light intensity is discussed in considerable detail, 
it being shown that in strong leht the organism is stimu- 
lated to active movement, while in the shade it remains quiet 
or moves very slowly. ‘This was found to occur as well in 
decapitated as in normal worms. The “ stereotropic” reac- 
tion in this form is also mentioned. In concluding, the 
author holds that in worms there is no “ associative Gedicht- 
niss,’” and hence no consciousness. ‘These results have been 
recently incorporated without essential change into a larger 
work (Loeb : 00). 

In an earlier paper Loeb (’95) first described the reactions 
to hight of Planaria torva. These results were incor- 
porated without essential change into the “94 paper men- 
tioned above. 

Hesse (97), in his morphological studies on the eyes of 
flat-worms, devotes a section to the subject of the reactions 
to light of Euplanaria gonocephala and Dendroccelum 
lacteum. His results are confirmatory of Loeb’s, nothing 
of particular significance being added. 

Steiner (798) found that posterior pieces of Planaria 
Neapolitana (=Stylochus pilidium, Lang) separated 
from the brain by a transverse cut would move about freely 
after recovery from the operation. He believes this ability 
to move is conditioned by the presence of ganglion-cells in 
other parts of the body than the brain (along the lateral 
nerve-cords). 

Parker and Burnett (:00) have recently made a thorough 
study, using very careful experimental methods, and treating 
the results statistically, of the reactions of Planaria 


522 RAYMOND PEARI. 


gonocephala to light. This form moves away from the 
source of the light. The amount of directive influence was 
measured. It was found that specimens without eyes, i.e. 
in which the anterior end had been cut off, react in much the 
same way to light as do normal individuals, “in that they 
have a tendency to turn away from the course when directed 
towards the source of light, and to keep in it when directed 
away from the source, though with less precision, and often 
to less extent, than planarians with eyes.’? Furthermore, 
figures are given showing that planarians from which the 
anterior end has been cut off move more slowly than normal 
animals. This is thought to be due to the absence of the 
eyes. 

The most extensive paper dealing with the physiology of 
planarians is that of Bardeen (:01). This paper is mainly 
devoted to a study of regeneration in Planaria maculata, 
but before entermg upon the discussion of this subject the 
author devotes considerable space to an account of the 
anatomy and physiology of the organism. In the section 
devoted to physiology, the author discusses, under the 
caption ‘ Environmental Activities,” sensation, movement, 
and the central nervous system. The author makes the 
remarkable, and obviously incorrect, statement that the 
planarian is sensitive only to light and contact. A very few 
inconclusive experiments having reference to thigmotaxis, 
geotaxis (?) and hydrotaxis, are reported. ‘I'he statement is 
made that specimens “ would remain unmoved by the presence 
close by their side of a piece of fresh snail, a food much 
prized by them.” ‘Two forms of movement are described— 
“swimming” and crawling. The author’s description of 
what he calls the “swimming” movement will be discussed 
later in this paper. Brief and very general statements 
regarding the reactions to mechanical stimuli are presented. 
Under the heading “Internal activities” are discussed 
deglutition, food dispersion, defecation, and respiration, in a 
rather loose and hypothetical way. The author makes the 
following contribution regarding excretion in Planaria :— 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 528 


“ Hixeretion is carried in part through the intestines by the 
act of defecation; in part it is doubtless carried on by an 
excretory system opening on the surface.” A more detailed 
discussion of various points raised by Bardeen will be entered 
into in connection with the parts of this work on which they 
have direct bearing. 

A second paper by the same author (Bardeen, : 01a) 
describes briefly the normal food reactions of Planaria, and 
shows that a decapitated specimen will not find food material 
in a dish, although one such a specimen could “ be made to 
eat if it were placed on its back on a slide in a small drop of 
water. Under the conditions mentioned the pharynx is 
usually protruded, and will engulf bits of food placed in the 
mouth.” An experiment was performed in which the part 
of the head in front of the eyes was cut off. Such specimens, 
from which merely the tip of the head had been removed, re- 
acted normally to food. It is also shown that specimens 
from which the part of the body posterior to the pharynx has 
been removed feed like normal worms. Regarding the 
method by which planarians find food in their immediate 
vicinity, Bardeen says (p. 176), “It is difficult to determine 
the source of the impulse which gives rise to this purposeful 
activity. It is possible that the auricular appendages here 
act as delicate organs capable of stimulation by slight 
currents in the water set up by the minute organisms that 
prey at once upon the flesh of the dead snail.” Experiments 
to be reported in the course of the present paper show, I 
think, that the mechanical and chemical stimuli given by food 
are the ones which affect planarians. 


c. MATERIAL. 


The following species have been principally used in this 
study :—Planaria maculata, Leidy; Planaria gono- 
cephala, Dugés; Planaria dorotocephala, Woodworth.! 
Of these P. dorotocephala and P. maculata have been 


' Excellent figures and descriptions of these three species have been pub- 
lished by Woodworth, 97. 


524 RAYMOND PEARL. 


most used, both on account of their abundance and, further- 
more, because P. dorotocephala is a form particularly 
favourable for the study of reactions. It is very active, and 
after being disturbed continues in movement longer than 
either P. maculata or P. gonocephala, as has already 
been noted by Woodworth (loc. cit., p. 7). I have found 
also that it moves faster than either of the other two species. 
There is a general precision and positiveness of response in 
its behaviour which make it especially favourable for experi- 
mental work. A large number of experiments have been 
made with a view to determining whether there was any 
difference in the reactions of these three species, but no 
essential difference has been found. The form of the reactions 
is the same in all cases. Whatever differences there are are 
differences of degree, such as would be conditioned by the 
relative sluggishness and activity. 

Certain forms of reaction to mechanical stimuli, and to 
chemical stimuli, are rather more easily induced in P. 
dorotocephala than in either of the others, yet, as will be 
shown later, these reactions will be given, under the proper 
conditions, by the other species. This being the case, and 
since P. dorotocephala was, for reasons noted above, most 
used in this work, it will be employed throughout the paper 
as the type form, and it will be understood, when there is no 
statement of the species, that P. dorotocephala is the form 
meant. 

No account of the anatomy of these forms will be given 
here, because it has been very fully treated in other readily 
accessible papers. The most important papers dealing with 
the morphology of the fresh-water triclads are those of 
Jijima (84), Lang (81, ’81a), Kennel (’88), Chichkoff (’92), 
and Woodworth (’91 and ’97). 

Besides the species mentioned above, on which the most of 
the work was done, a number of observations and experi- 
ments have been made on several other species of triclads 
and rhabdoceeles. he other triclad most frequently met, 
and whose reactions have been found to agree closely with 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 525 


those of the species of Planaria, is a form which agrees 
with the description of Dendroccelum lacteum, except in 
respect to the colour. This form is usually coloured from a 
light grey to nearly black. The colouring is uniform. In 
only one specimen have I found any deviation from this 
typical coloration, and in that case there was a band of 
black pigment extending the whole length of the body 
along the mid-dorsalline. In width this band occupied about 
one third the whole width of the body. The margins were a 
pure white, without the faintest trace of pigmentation. ‘This 
specimen was kept under observation for some time, and 
there was no doubt that it belonged to the same species 
as the grey form. ‘The specimen struck one at once as being 
transitional between the ordinary white to cream-coloured 
Dendroccelum lacteum, and the grey form found about 
Ann Arbor. Being in some doubt as to the true taxonomic 
position of this grey species, I shall refer to it throughout 
this paper as Dendroccelum, sp. 

Besides the forms mentioned, several undetermined triclads 
have been collected and worked with to some extent, but as 
no new factors presented themselves in their reactions they 
will not be considered in this paper. 

A large rhabdoceele, which I have identified as Mesostoma 
personatum, O. Schm., is found rather commonly in certain 
localities about Ann Arbor in the spring. I have done some 
work on this form. Another rhabdoccele whose actions I 
have studied to some extent is Stenostoma leucops, O. 
Schm. No detailed investigation of the behaviour of the 
rhabdocceles was made, but as opportunity offered they were 
used for comparison. 

The methods used in experimentation will be given under 
the separate headings dealing with the reactions. 


p. Hasrrs! ano Narurat History. 


In the course of more than two years’ study of planarians 


1 In this section the word “habit” will be used to signify merely those 
activities of the organisms which are frequently observed to occur under 


von. 46, PART 4,—NEW SERIES. MM 


526 RAYMOND PEARL. 


numerous observations have been made on their general 
natural history. It is thought desirable to present a general 
account of this here for two reasons: first, because there is no 
adequate discussion of the natural history of the fresh-water 
triclads in the literature; and furthermore, because it will 
bring out prominently the phenomena for which we are 
seeking an explanation. In other words, it will present the 
problems with which this study has had to do. 


I. Occurrence and Distribution. 


The species of Planaria (maculata, dorotocephala, and 
gonocephala) used in this study have been collected mainly 
from the Huron River near Ann Arbor. They are found, for 
the most part, on the under surfaces of stones in places where 
the current is of moderate swiftness, and the substrate on 
which the stones rest is rather soft. They are also found 
among the fronds of such water plants as Ceratophyllum 
and HKlodea, although less abundantly than on stones. I 
have obtained these species only very rarely in collections 
from ponds and small pools of stagnant water. ‘They appear 
to be, in general, much more abundant in shallow water than 
in deep. 

Rhabdocceles I have found in great abundance in small 
ponds and pools of stagnant water, and, with the exception 
of Stenostoma leucops, almost never in running water. 
Dendroccelum, sp., is also much more abundant in stagnant 
water than in streams. 

‘There is no marked seasonal distribution of the species of 
Planaria studied. ‘They appear to be slightly more abun- 
dant in the fall than in the spring. [Ihave found no evidence 
of any migration into deep water during the winter in the 
case of these forms, as has been described by Child (: 01, pp. 
978—981) as occurring in Stichostemma. ‘The seasonal 
distribution of Dendrocclum, sp., appears to be well marked, 


natural conditions, without necessarily implying the same idea as that embraced 
in the term ‘habit ” as used by the psychologists. 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 527 
individuals being found in considerably greater numbers in 
the spring than at any other time in the year so far as my 
observations go. ‘This seems to be true also of the rhab- 
docceles. 

Relatively the most abundant species of planarians in this 
region are Planaria dorotocephalaand maculata. The 
numbers of these two are about equal, with a slight advantage 
in favour of P. dorotocephala. Next in abundance I have 
found to be P. gonocephala, but this is considerably below 
the other two. Finally comes Dendrocclum, sp., which I 
have never found in an abundance to be compared with the 
species of Planaria. 


Il. Activities. 


The movements of planarians will be discussed in detail in 
a later section of the paper,! but it is desired to take up here 
certain general activities and relations to the environment 
which properly fall under the general subject of natural 
history. 

The first of these subjects is— 

a. Sensitivity.—The flat-worm is extremely sensitive to 
a variety of stimuli. Among the different stimuli which pro- 
duce specific reactions, and to which we must therefore 
conclude it is sensitive, are the following :—Mechanical 
disturbances of the general environment (shaking, jarring, 
movement of water, etc.), contact (localised mechanical 
stimulation), chemical changes in the environment (in the 
widest sense, including food substances), ight, the electric 
current, desiccation, a current of water, and heat. 

Its extreme sensitivity, which makes it responsive to very 
slight changes in the environment, may be shown by a very 


1 It may be stated here, for the convenience of the reader before reaching 
the full discussion of the movements, that the progressive movements of 
triclads are of two sorts. These are (a) gliding movements, in which there is 
little or no muscular action ; and (4) crawling movements, in which the motion 
is effected by muscular contractions involving the whole body. ‘The crawling 
has some general resemblances to the method of progression observed in a 
leech of the genus Clepsine. 


528 RAYMOND PEARL. 


simple experiment. If a dish containing specimens not in 
any way stirred up by rough handling, but gliding along the 
bottom, be jarred ever so slightly, every individual will 
instantly stop, contract, and remain immoyable. If only a 
single jar is given, the worms will start after only a momentary 
pause. A further experiment shows more strikingly the same 
thing. If in a dish containing water to a depth of not more 
than 1 to 1°5 cm. a single specimen gliding quietly 1s 
selected, and a needle is touched to the surface of the water 
above or to one side of it, it will be seen, if closely watched, 
to give the same momentary pause and partial contraction. 
If the needle is pushed down through the water towards the 
worm in any but the quietest and gentlest way the contracted 
state will continue. Only at the moment when quietness in 
the surroundings intervenes again will movement be resumed. 
I have frequently tried to introduce a needle close beside the 
animal without causing this momentary pause. With a layer 
of water not over a centimetre in depth covering the worm I 
have not been able to do this, except in rare instances. 
After the point is once through the surface film it may be 
brought nearer the worm without causing a persistence of the 
contraction, provided it is advanced in line with itself, 1. e. 
not slid up laterally. In order to observe this extreme sensi- 
tiveness to disturbance of the water one must take care that 
the animals have not been violently disturbed just previously. 
Any marked disturbance or persistent, more or less violent 
stimulation puts the animals in a condition which may be 
called, for lack of a better term, “ excited.” Such a condition 
is characterised by increased rapidity of movement and in- 
creased general activity, and in this condition the animals do 
not give the “finer” responses,—that is, responses to weak 
stimuli. I shall have occasion to discuss this matter in more 
detail later. 

This marked sensitivity and its associated behaviour are 
remarkably similar to what has been found by Whitman (799) 
to obtain in the case of the leech Clepsine. He has further 
pointed out that lack of attention to this extreme sensitivity, 


MOVEMEN'S, ETC., OF FRESH-WATER PLANARIANS, 529 


which is apparently quite generally distributed among lower 
organisms, ay be an important source of error in work on 
the behaviour. 

Regarding the statement of Bardeen (:01, p. 14) that he 
does not find that Planaria “ is sensitive to anything but light 
and contact,” nothing need be said here. The detailed 
accounts of the reactions of the organism to a variety of 
stimul: which follow in this paper are in themselves a 
sufficient criticism. 

b. Secretion of Mucus.—There is secreted at all times 
over the surface of planarians a sticky slime, apparently of 
the nature of mucus. ‘This secretion is increased when the 
animal is irritated, and is under normal conditions more 
abundant on the ventral than on the dorsal side. If a 
needle or fine glass rod is touched several times on the 
surface of the body its end becomes covered with this 
secretion. For this reason it 1s necessary in applying 
localised mechanical stimuli to wipe the mucus off the end 
of the needle frequently, in order to obtain good results. 
Similarly, if one is using a sharp scalpel to cut the animals, if 
the edge is left in contact with the surface of the body any 
leneth of time before the decisive cut is made, the edge will 
become so coated with mucus that a clean cut is impossible ; 
instead, the animal will slip from under the knife. 

When the animal moves about it leaves behind a more 
or less heavy string of this mucus, so that if several speci- 
mens are placed in a clean glass dish the bottom will, in 
a short time, become covered with a network of interlacing 
mucus threads. The same phenomenon occurs in other 
Turbellaria and among the Nemerteans (cf. Child, : 01, 
and Wilson, :00). he threads when first secreted are so 
transparent as to be invisible, but in larger quantities they 
appear opalescent, and may be picked out of the dish with 
forceps. 

The function of this secretion in locomotion is evidently to 
attach the body to the substrate. Secretions for such a 
purpose occur widely among aquatic organisms. 


530 RAYMOND PEARL. 


The mucus also undoubtedly plays an important part in 
the attachment of the animal to the under side of the surface 
film. When the worm leaves the surface film in open water, 
i.e. when it cannot reach any solid body, it hangs by the 
mucus thread in much the same way that a terrestrial 
molluse, like the common slug, does when it passes through 
the air from a higher to a lower point. This observation I 
have made many times, though generally in an indirect way. 
As has been said before, the mucus thread is invisible when 
first secreted, so that when a worm leaves the surface film it 
seems to e@lide freely through the water. If, however, one 
passes a needle horizontally through the water immediately 
above the posterior end of a worm which has just left the 
surface film, it will be seen that at a certain point (where the 
needle strikes the thread) the end of the worm will be jerked 
to one side. Furthermore, one may with care pick up the 
invisible mucus thread with forceps and raise the whole 
worm, provided the attempt is made before the anterior end 
reaches the bottom. I have seen specimens of P. maculata 
crawl back upon the thread after going a part of the way 
down to the bottom, and again regain a position on the 
surface film. The same thing is frequently done by slugs. 
When the animal has fully reached the bottom, connection 
with the thread which has served to suspend it in the water 
is usually broken by several sharp jerks of the posterior end 
of the body from side to side. 

The relation of the organism to this slimy secretion is 
much the same in the land planarians, according to the 
observations of Lehnert (’91). He distinguishes in case of 
these forms ‘ Kriechfaden,”’ ‘* Briickenfaden,” and “ Gleiten- 
faden”’ formed from the slime, the distinction being based 
on the relation of the thread to the surroundings. The 
“ Kriechfaden ” are the threads left behind as the organism 
moves over a continuous solid body, and the “ Gleitenfaden ” 
ave the threads on which the animal hangs in passing 
through the air from a higher to a lower level. Both these 
forms of threads I have found in case of the fresh-water 


MOVEMENTS, BTC., OF FRESH-WATER PLANARIANS. 551 


planarians. The “ Briickenfaden,” which are formed by the 
land planarians when they pass from one solid body to 
another at about the same level, | have never observed in 
case of fresh-water planarians, though I see no reason why 
under proper conditions they would not be formed. Lehnert 
(loc. cit., p. 17) says, “ Die Wasserplanarien bilden wie die 
Landformen ihren Kriech-, Briicken-, und Gleitfaden.” He 
also noted that Polycelis tenuis was able to crawl back 
upon a mucus thread after passing for some distance down 
over it. 

Nothing like the formation of “cysts” from this mucus, 
such as Child (:01, pp. 989 to 993), found in the case of 
Stichostemma, has been observed in the case of planarians. 
Its only biological significance in these forms is in relation 
to movement, as pointed out above. 

In connection with the subject of mucus secretion it may 
be well to point out the tenacity of the attachment of the 
flat-worm to the bottom. It will be found in attempting to 
dislodge the animal that the extreme anterior end and the 
extreme posterior end stick very firmly to the substrate. 
Whether this holding is the result of a sucker-like action of 
the ends of the body, or is due merely to the stickiness of the 
mucus, I have been unable to decide. It is easily possible 
that the muscles could be so contracted as to form out of 
either end of the body a practical sucker, but whether this is 
done or not it is impossible to say. Woodworth (’97) has 
described a permanent anterior adhesive disc in Dendro- 
celum lacteum, but considers that this “is not a true 
sucker, nor does the animal employ its anterior end for the 
purpose of attachment to any greater degree than the 
posterior or lateral margins of its body, along the ventral 
surface of which numerous mucus glands have their 
openings. In truth, it is the margins and posterior end that 
adhere more firmly to a support; often when the animal is 
forcibly removed from the sides of the aquarium the parts of 
the margin or the posterior end will adhere so firmly to the 
glass that the points of attachment are drawn out into 


532 RAYMOND PEARL. 


digitate processes.” I incline to the view that in Planaria 
it is the mucus which attaches the organism to the support, 
although it must be said that the appearance is at times 
strikingly as if the anterior and posterior ends acted as 
suckers. 

c. Periods of Activity and Rest.—There is in the 
case of freshly collected planarians a certain periodicity in 
the activities. First, there is the rather marked difference 
in the amount of activity in the night and day. It has been 
stated by a number of investigators that planarians were 
probably nocturnal in their habits, i. e. more active at night 
than during the day. ‘This can easily be seen to be the case 
in the following way :—In a dish containing a large number 
of planarians, together with some plant material like Cera- 
tophyllum, usually comparatively few specimens will be 
seen during the day. Nearly all will be in among the fronds 
of the plant material in a quiet condition. If, however, one 
comes into the laboratory at 8 p.m. or later at night, so that 
(in case of winter days) there has been two and a half or 
more hours of darkness, a large number of the specimens 
will be found on the sides and bottom of the dish in active 
movement. Again, one will frequently find in the morning 
that the specimens are scattered about all over the sides 
and bottom of the aquarium dish at rest. By noon many 
of these will have disappeared, or, in other words, gone 
in among the plants, where they are protected from the 
light. 

Besides this day and night periodicity there is another 
fact that may be mentioned; this is that during the day, 
at any rate, they seem to be incapable of continuing 
movement more than a certain, not very great, length of 
time. ‘Then a period of rest must intervene. Thus -one may 
see a specimen which has been moving about come to rest, 
and after a length of time, varying from a comparatively few 
ininutes to several hours, it will start into spontaneous move- 
ment again, and repeat the whole cycle over and over. It 
seems that the periods of quiet are really for the purpose of 


MOVEMEN'S, ETC., OF FRESH-WATER PLANARIANS. 933 


resting, i. e. the animal seems to be quickly fatigued by its 
own movements. ‘This is indicated by the fact that if one 
stirs up a specimen, and sets it into activity again just as 
soon as it comes to rest, the periods of spontaneous activity 
will become progressively shorter, until finally the worm will 
only move a very short distance before coming to rest again. 
The periods of activity are longer and more frequent in 
P. dorotocephala than in any of the other species I have 
studied. 

d. Formation of Collections.—There is a well-marked 
tendency for specimens of planarians to form well-defined 
groups or collections when they come to rest on an open 
surface like the bottom or sides of a glass dish, or on the 
under side of rocks, under natural conditions. Of course, 


Fig. 1.—Diagram showing the appearance of a collection of resting 
planarians. 


this is in part a result of their reaction to light, as has been 
noted by Loeb (94). Besides this there seems to be some 
other factor at work, for in the same dish one frequently 
finds several localised collections from one to two inches in 
diameter in different parts of the dish. In these collections 
the specimens may be closely packed together, and with 
some specimens overlapping and lying partly over others, yet 
in the species I have studied a looser arrangement of the 
character shown in fig. 1 is the more usual one. On the 
under surface of stones such groups are frequently seen ; 
two or three may be found on the same medium-sized stone. 
In this case light as a factor cannot be present, since the 
conditions of all with reference to this stimulus are equal. 


534 RAYMOND PEARL. 


We have, then, here a case of what appears superficially to 
be “ social instinct.” 

e. Movement on Surface Film.—As is well known, 
flat-worms and a number of other animals frequently move 
about on the under side of the surface film at the top of the 
water. On account of the flexibility of the support, motion 
under these conditions is very slow, and usually, after having 
been on the surface film for a short time, the worm will 
loosen its hold and pass down to the bottom in the way which 
has been described above. The worms do not remain 
customarily in the angles formed by the surface film with the 
side of the dish, as does Stichostemma (Child, :01), but 
instead pass out at once on to the free surface. Further, the 
flat-worms never push through the film at the side of the 
dish and pass up out of the water as the nemertean does. 
The occurrence of planarians on the surface film is not the 
result of any thigmotactic reaction (using thigmotactic in the 
sense ordinarily understood), but is brought about by a 
simple reflex act, and is the result of the configuration of the 
surface of the water and the side of the dish. This will be 
brought out in more detail later. It is probable that fresh- 
water planarians, in their normal habitat, very rarely take up 
a position on the surface filu. Among other organisms 
(Mntomostraca, Hydra, ete.) this habit probably has a much 
ereater biological significance than in planarians (ef. 
Scourfield, 794, :00, :01). When on the surface film the 
worm behaves in nearly every respect as it does when on the 
bottom. The head is frequently raised (with reference to 
the worm) and waved about in the water just as occurs in 
the normal movement. ‘hat the situation is a more or less 
abnormal one, however, is shown by the fact that, so far as I 
have observed, the worm never comes to rest on the surface 
film, but instead, always keeps in active movement till it 
leaves it. 

The means by which the animals maintain their position on 
the under side of the film is undoubtedly the mucous secre- 
tion from the ventral surface. ‘lhis is very sticky, and holds 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 539 


the animal to the film, the surface tension being sufficiently 
great to support a considerably greater weight than that of 
a flat-worm. 

It is interesting to note in this connection that the land 
planarians are able to move about on the top of the surface 
film of water to a limited extent (cf. Lehnert, loc. cit., 
p- 16). The immediate means of support here, as in the case 
of the fresh-water planarians, is the mucous secretion. 

The leaving of the surface film by means of the mucus 
thread described above apparently does not take place if it is 
possible for the same result to be accomplished in any other 
way. Before it occurs the worm usually stretches the 
anterior end down into the water, and turns it in all direc- 
tions. If it comes in contact with something solid the 
anterior end becomes attached to this, and pulls the posterior 
end of the body away from the film. If nothing solid is 
within reach the worm will usually, after a time, drop down 
on a mucus thread as described. 


Ii Hood: 


Planarians will take almost any sort of animal food very 
readily. I have used mainly, in the feeding experiments, 
crushed pieces of fresh-water molluses, such as Physa, 
Planorbis, etc. One of these molluses, removed from the 
shell and placed in a dish containing a large number of 
planarians, will, in a short time, be literally covered with the 
worms feeding. Ifa worm is gently lifted off the pile the 
greatly stretched pharynx will be brought into view. The 
worms will eat any other kind of animal tissue (fresh meat, 
parts of insects, pieces of fresh-water worms, etc.), so far 
as I have observed, the only condition being that the meat 
must be fresh. As will be shown later, the juices from the 
food act as chemical stimuli, so that it is necessary that the 
tissue be crushed or bruised so that its juices can escape into 
the water. A partially crushed specimen of Planaria, even 
though still able to move about, will be seized upon and eaten 


536 RAYMOND PEARL. 


as quickly as any other food. I have several times seen 
specimens thus eaten. It is, in fact, possible, with a little 
patience, to make a specimen eat a small piece cut off the 
posterior end of its own body! This eating of each other 
does not occur, so far as I have observed, unless an individual 
is bruised so that some of the tissue underlying the epidermis 
is exposed. Under these conditions juices escape from the 
body and act as stimuli on the other worms. Under normal 
conditions contact of one individual with another does not 
start the feeding reaction, which is a purely reflex pheno- 
menon, capable of being started only by a certain set of 
stimuli. Promiscuous cannibalism, such as Child (:01) 
suspects to occur among individuals of Stichostemma, I 
have seen no evidence for among the Turbellaria.! 

In the feeding the worm hes fully distended, with the 
posterior two thirds of the body on the meat, or whatever 
else is being used for food; the pharynx is extruded, 
frequently to nearly half the length of the body, andits end is 
attached to the meat. During the feeding the very anterior 
end of the worm is attached to the bottom of the dish, provided 
the piece of food is not so large as to make this impossible. 

Besides the animal food which the worms will take so 
readily, they also normally, probably to some extent, feed on 
vegetable matter, although I have not been able to induce 
the typical food reactious (to be discussed later) with vege- 
table material. The evidence for the statement that 
vegetable food is used by planarians is of two sorts: (a) 
specimens are frequently found extended on the stalks of 
water plants, with the pharynx extruded and attached to the 
stalks; and (b) the feces which have been observed immedi- 
ately after defecation have been found to consist largely of 
finely divided plant tissue. It would appear, however, that 


' Bardeen (: 01, A, p. 176) says, “Strong planarians often prey upon weak 
ones. In such instances the strong individual attaches its pharynx somewhere 
upon the body of the weak one, usually near the head.” I have never seen 
even the largest specimens cat smaller ones unless these latter were bruised 
in some way. 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 537 


vegetable food is not alone sufficient to keep the animals in 
good condition, for specimens kept in an aquarium dish with 
plenty of living plant material, on which they stay the 
greater part of the time during the day, will steadily grow 
smaller unless animal food is given them. 

The food is in part digested, or at any rate softened, and 
physically changed outside the body. <A piece of mollusc, 
on which a number of worms have been feeding for some 
time, has a white, fluffy appearance, similar to that of meat 
after partial gastric digestion. ‘I'his is apparently brought 
about by a secretion poured out of the end of the pharynx. 
The necessity for some such action is apparent, because the 
flat-worm has no teeth or other means of separating a portion 
of ordinary tough fibrous tissue off from a mass so that it can 
be swallowed. This can be done, however, if the mass is first 
softened and partially dissolved. There are certain other 
evidences that a secretion is poured out from the pharynx 
during the feeding process. ‘These will be taken up in 
another connection. 

After the worms have fed undisturbed for a certain length 
of time they will leave the meat, and, after a short period of 
activity, come to rest. 

The worms are able to live for a considerable length of 
time (at least two months) without food, although they 
continually grow smaller during this time. This marked 
decrease in size while starving has been noted by several 
observers, and especially studied by Lillie (:00). his 
author finds that the decrease in size is accompanied by a 
simplification of structure—a sort of ‘development back- 
ward,” such as has been described by Patten (’96) for ab- 
normal embryos of Limulus. 


IV. Defecation. 


The process of defecation has been observed by Bardeen 
(: 01). The process consists of three or four general con- 
tractions involving the whole body, during which the 


538 RAYMOND PEARL. 


contents of the intestine can be seen to be in rapid motion. 
Soon after the beginning of the contractions, which are in 
character different from any other of the movements of the 
body which I have observed, and which cannot be adequately 
described, the intestinal débris is shot out of the pharynx. 
‘The force of the expulsion is so great that the feces spread 
out in the water some considerable distance from the opening 
of the pharynx. I have observed the process only a few 
times ; apparently it occurs only at infrequent intervals. 


V. Summary of Factors in Behaviour. 


From the above sketch the behaviour of the flat-worm can 
be seen to be of considerable complexity. ‘lhe movements 
show many variations in character, rate, and direction. The 
animal shows apparent preferences for certain situations 
while avoiding others. It reacts to a variety of stimuli in 
ways which, on the whole, further its preservation and well- 
being perhaps as well as if guided by careful thought. It 
chooses its food, taking certain things and passing by others. 
It forms gatherings of a sort which apparently indicate that 
the flat-worm prefers to be in the company of his fellows ; 
in other words, it seems to have something of “ 
instinct.’ On the whole, as the further analysis will show, 
it fits itself to its environment by its activities in a way 
which would not be discreditable to a being possessed of 
considerable powers of reasoning. 

Our problem now is to analyse, as far as possible, this 
complex behaviour into its component factors. Hach activity 
will be taken up in detail and subjected to thorough scrutiny, 
to determine its essential nature, and whether it may not be 


social 


resolved into simple components. With this analysis com- 
pleted, it will be possible to assign the organism a definite 
position in the objective psychological scale. With the 
internal psychological factors—those of which there is no 
objective criterion—we shall not attempt to deal. The 
purpose of the paper is to furnish the data which may be 


MOVEMENTS, HTC., OF FRESH-WATER PLANARIANS. 539 


obtained by an objective study of the phenomena: precisely 
what these imply as to internal factors would doubtless be a 
subject of dispute among psychologists of different schools. 


E. Normat Moror ActIvitiEs. 


Under this heading will be included all the purely motor 
phenomena of the organism. This will include the 
movements (without reference to special reaction to stimuli), 
the coming to rest, and the general resting condition of the 
organism. 

The movements naturally fall into two categories; (a) 
locomotor movements, and (b) non-locomotor, including such 
movements as contractions and expansions and the like. 


I. Locomotor Movements. 


As has been mentioned above (p. 19), there are two sorts 
of locomotor movements, the gliding and the crawling. The 
gliding is the smooth, even motion by which the flat-worm 
shps about over surfaces without showing a perceptible 
ripple of muscular movement. This is the characteristic 
movement when the organism is not particularly stirred up. 
The crawling is the characteristic movement when the animal 
is, or has been recently, strongly stimulated. It is a purely 
muscular movement, 

a. Gliding.—The movement which I have called gliding 
is apparently the same as that which has been called “ swim- 
ming ”’ by Bardeen (loc. cit., p. 15), yet it must be stated that 
in all of my observations on a very large number of plana- 
rians I have never seen anything corresponding to some of 
the details which this author mentions in this movement. 
In the first place, he speaks of the worms moving progres- 
sively when not in contact with a solid body, i.e. of a 
movement freely through the water. This I am unable to 
understand, as I have never seen the slightest indication of 
the organism moving freely in the water without contact 
with a solid body or something which served the purpose of 


54.0 RAYMOND PRARIL. 


a solid (viz. the surface film). The only possible exception 
to this is the passage of the animals from the surface film 
to the bottom on a string of mucus, as described above. 
Furthermore, so far as I can find in the literature, no one 
else has ever seen a fresh-water triclad swim freely through 
the water. 

The movement takes place with the body in contact with 
a surface either of a solid or of the surface film. There is, 
of course, between the ventral surface of the body and the 
surface on which it is moving, the thin layer of mucus which 
is constantly being secreted. It is in this mucus layer 
rather than the free water that the cilia beat. 

This gliding movement is, so far as I have been able to 
ascertain, brought about by the action of the cilia on the 
ventral surface. There may also be some very slight 
muscular movement of the ventral body-wall comparable to 
that in the foot of a molluse like Physa, which assists in the 
locomotion ; but in the case of the flat-worm this factor, if it 
exists at all, is very insignificant. Only in a few instances 
have I been able to satisfy myself that any such movement 
was taking place, and then it did not have the characteristic 
rhythm seen in a molluse.’ If this factor has any effect at all 
on the gliding movement it must be an extremely slight one. 

The cilia beat strongly backward, i.e. towards the poste- 
rior end of the body. I have not been able to induce any 
reversal of the ciliary beats in these ventral cilia. Bardeen 
(loc. cit., p. 15) states that when the head is suddenly drawn 
back from some object the movement of the cilia on the 
antero-lateral margin of the head is reversed, and further 
sugeests that “this reversed action may possibly be set up 
by the mechanical friction of the water.” It would appear 
that the suggestion is the correct explanation, and that this 
is not a true reversal of effective beat. 

The cilia which are mainly effective in producing the 
eliding movement are distributed on the ventral surface of 
the body, as shown in Fig, 2. There is a band down the 
centre of the body, which widens ont at the anterior end so 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 54 


as to cover nearly the whole of the ventral surface of the 
head. The beat is the strongest down the median line of 
this band, and diminishes in intensity towards either edge 
until at the margins there is no ciliary movement at all. At 
the anterior end the cilia near the side of the head beat 
backwards and at the same time inward towards the median 
line, so that the currents take the course indicated by the 
arrows in that region in Fig. 2. The distribution and action 
of these cilia were made out by stirring finely powdered indigo 
in the water, and then either directly observing the ciliary 
action on these suspended particles as the animal glided on 
the surface film, or by indirectly observing it in a mirror 
placed below the bottom of the glass dish in which the 
worms were. Both of these methods gave the same results, 


Fic. 2.—Diagram of the ventral surface of Planaria, to show the 
distribution of cilia. The stippled area is that which bears cilia. 
The arrows indicate the direction of the ciliary currents. (The 
pharynx is omitted for the sake of clearness.) 


and showed very clearly the distribution of the effective 
cilia. 

I have found no evidence of ciliary action on the dorsal 
surface of the body. Around the margins of the head there 
are cilia, but in other parts of the body, either on the dorsal 
surface or the edges, I have found no evidence of their 
presence. Particles of indigo dropped on the dorsal surface 
of a worm will remain in the same place for hours at a time. 
This is in striking contrast to the conditions in the land 
planarians as described by Moseley (’77), where the dorsal 
surface is thickly covered with cilia, which serve the purpose 
of keeping the body freed of foreign matter. 

In the gliding movement the head is raised slightly from 
the bottom so as to form an angle with the rest of the body. 

voL. 46, PART 4,—NEW SERIES. NN 


542 RAYMOND PEARL. 


This position is shown in Fig. 3. As will be brought out 
later, the head is a particularly sensitive portion of the body, 
and apparently its elevation is related to its sensory function, 
in that it practically brings the head into close relation with 
a large environmental field. The head is not held in a fixed 
raised position, but is in constant though shght movement 
whenever the animal as a whole is moving. These “ feeling” 
movements (“tastende Bewegungen”’) of the head are very 
characteristic. The head as a whole is raised and lowered, 
and turned from side to side, while at the same time the 
antero-lateral margins are moved up and down and extended 
and retracted. These ‘‘ feeling” movements of the head 
region are usually very slight, and escape notice except 
under close observation. When the organism is much 
stirred up, however, they may become qumte apparent. 
Their purpose is evidently to increase the chances of re- 
ceiving stimuli, so that any stimulus in the neighbourhood 


————————— ae 


Fie. 3.—Diagrammatic side view of a gliding planarian. 


may be quickly received. Constantly different sensory 
surfaces are presented to the environment. ‘The head region 
acts in movement as a single great tentacle-like organ which 
is constantly testing the environment as the animal pro- 
ceeds. At the same time the auricles are fully extended and . 
raised. I do not think that this marked sensory activity 
functions so much for the protection of the organism against 
harmful environmental influences as it does to give prompt 
notice of useful stimuli,—for example, stimuli due to the 
presence of food material. In general it would not appear 
that such an organism as the flat-worm runs as great risk of 
elimination from enemies as it does from not finding food 
material for its own support. In the ctenophore Mnemiop- 
sis Leidyi, whose reactions I have studied, no trace was 
observed of a reaction adapted to the purpose of getting the 
organism out of dangerous surroundings, but its only specific 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 543 


reaction is one which would bring it towards any food 
material which might be encountered.! While, as will be 
shown later, there is in the case of Planaria a reaction 
which is adapted to getting the organism out of danger, yet 
it is not called forth by so weak a stimulus as is the food 
reaction, and it is evidently for the purpose of receiving 
stimuli of the lowest intensity that the “feeling” move- 
ments are adapted. 

In addition to the slight “ feeling movements” of the 
head, described in the preceding paragraph, the organism 
frequently in the course of its gliding raises the whole 
anterior part of the body off the bottom and waves it abont 
in the water. The portion of the body so raised may in- 
clude the whole anterior half. he gliding is usually 
entirely stopped or very much decreased in rate while these 
waving movements are taking place. The head is swept 
from one side to the other and raised high in the water, 
covering a considerable area. This movement is also un- 
doubtedly for sensory purposes. 

In the gliding movement the body back of the head is kept 
in an approximately straight line ; that is, there is no sinuous 
bending of the body such as is observed, for example, in 
Stichostemma (Child, loc. cit., p. 981), or at times in the 
movement of the earthworm. Furthermore, I have never 
observed any regular undulation of the margins of the body 
durmg movements such as take place in case of many poly- 
clads, e.g., Leptoplana tremellaris (cf. Lang, 784), 
Bardeen (loc. cit., p. 15) seems to imply that such motions 
occur, and are an aid in the locomotion, but I am unable to 
confirm this statement. ‘here are, of course, slight move- 
ments and changes of contour of the margins of the body, 
but they are not of a prominence or character to warrant 
thinking that they in any way contribute to the propulsion 
of the animal. In fact, it seems more probable that they are 
in part passive results of the motion of the whole body, and 


' A brief preliminary account of the reactions of Mnemiopsis has been 
published in ‘ Science,’ N.8., vol. xii, No. 311, pp. 927, 928, : 00. 


544 RAYMOND PEARL. 


in part the expression of local changes in the tonic contrac- 
tion of the muscles. 

In the gliding movement the body is in close contact with 
the surface alone which the animal is moving. When an 
animal passes from the resting condition into movement one 
can see the body lengthen and flatten so as to hug the 
surface. By observing with a compound microscope an 
animal gliding alone the vertical side of a dish so that the 
edge is brought sharply into view, the closeness of the con- 
tact of the margin of the body with the surface can be well 
seen. Furthermore, in specimens in which the posterior part 
of the body has been split longitudinally in the middle line 
to a point just behind the head, it is found that the half of 
the body which is determining the direction of the move- 
ment is always in close contact with the surface, while the 
other half only lightly touches it. 

It would appear from all the observations which have been 
given that the glidmg movement is brought about in the 
following way :—The ventral surface of the body constantly 
secretes mucus in greater or less quantity. This mucus can 
be shown experimentally to be very sticky immediately after 
it is secreted into the water. As it is secreted under normal 
conditions it immediately sticks to the surface on which the 
animal is reposing. Thus there will be constantly between 
the animal and the surface on which it is moving a layer of 
mucus which is adherent to the substrate. We can think of 
the lowest part of this mucus layer where it is stuck to the 
surface as of denser consistency than its upper layers which 
are in contact with the animal. In this upper layer of the 
mucus the cilia are beating and constantly pushing the 
animal forward. Of course, what really takes place is that 
the cilia are pushing the secreted mucus backward, but as 
this layer of mucus becomes fixed to the substrate as soon as 
it is secreted, the practical result is that the animal is pushed 
ahead. ‘his relation is shown in Fig. 4. A represents a 
side view of a gliding worm; D is the substrate ; C the cilia 
ou the ventral side of the organism; and B the mucus secre- 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 549 


tion, represented disproportionately exaggerated in thick- 
ness. ‘I'his sticks to the surface of the substrate, and the 
backward beating of the cilia drives the worm ahead. 

1]. Rate of Gliding Movement.—There is no very 
marked difference in the rate of the gliding movement in 
case of the species of Planaria studied. On the whole, 
specimens of P. dorotocephala move more rapidly than do 
those of the other two species, but there are large individual 
differences in this matter. Active specimens of Dendro- 
ccelum, sp., move much faster than any other planarians I 
have observed. Large specimens of this form will sometimes 
glide along with simply amazing rapidity, not showing the 
slightest tremor of the surface of the body. 

As to the absolute rate of the crawling, some statistics 


B AC 


Fig. 4.—Diagram to show the mechanism of the gliding movement. A 
represents a specimen of Planaria seen from the side; B, the layer 
of mucus secreted by the animal. (This layer is represented as 
greatly exaggerated in thickness in proportion to the animal.)  C, 
cilia. D, the substrate. The arrow indicates the direction in which 
the organism is moving. For further explanation see text. 

have been collected and will be presented. The statistics 
were obtained in the following way :—A paper was ruled into 
centimetre squares; over this was placed a flat Petri dish 
containing the worm to be tested. Normal active specimens 
of P. maculata were used, and nothing was put into the 
dish but fresh clean water. ‘The experiments were performed 
at night, and the source of illumination was a 16-candle 
power electric light enclosed within a ground glass globe. 
This lamp was 35 em. above and 35 cm. distant in a hori- 
zontal direction from the centre of the dish, so that the heht 
struck the animal at an angle of approximately 45° on its 
dorsal surface. ‘The worm was allowed to get into an even, 


546 RAYMOND PEARL. 


normal glide, and then to come around, as it usually would in 
a short time, so that it was headed in the direction of the 
light. Then the time which it took the worm to glide three 
centimetres was taken by means of a stop-watch. If the 
animal started crawling, or abruptly changed its direction, 
the trial was ruled out. The average rate in millimetres per 
second determined in this way from twenty trials on two 
individuals is 1°34. This rate is considerably higher than 
those obtained by Parker and Burnett (:00) for P. gono- 
cephala. In that form they found a rate of 1:04 mm. per 
second in the case of individuals moving toward a hori- 
zontal light; 1:12 mm. per second when movement was away 
from a horizontal light, and 1:08 mm. per second when the 
animals were moving under a vertical light. There seems to 
be a well-marked correlation between the size of individual 
and the rate of gliding, as would be expected on general 
grounds, and is apparent from merely qualitative observations 
on the movement. One of the specimens from which obser- 
vations were taken was 11 mm. long when extended, and its 
rate of gliding was 1°48 mm. per second; while the other 
specimen, which was only 6 mm. long when extended, showed 
a correspondingly slower rate of 1:23 mm. per second. The 
statistics are, of course, very meagre, and are not offered for 
any other purpose than to give a concrete idea of the approxi- 
mate rate of the gliding movement. <A thorough quantitative 
study of this matter of the rate of movement in planarians 
and other related organisms, and of the effect of different 
agents on the rate, would, I believe, be very interesting, and 
might lead to valuable results. I hope to be able to make 
such a study at some future time. 

Lehnert (loc. cit., p. 17) gives a table of the rate of move- 
ment (presumably in the case of the fresh-water forms the 
rate of the gliding movement) of several species of flat-worms. 
He gives no account of how the data were collected, but his 
values may be inserted here for the sake of comparison. His 
‘ates for Bipalium kewense and B. kewense viridis are— 
Usual rate, 1 to 1:33 mm. per second; occasional rate, 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. O47 


1°83 mm. per second. This agrees very closely with the rate 
for P. maculata given above (1°34). His rates for 
Geodesmus bilineatus and Dendrocelum lacteum are 
considerably slower (0°5—0°66 mm. per second and 0°75— 
1:33 mm. per second respectively). Polycelis tenuis 
(1°66—1-83 mm. per second), Planaria polychroa (2°:16— 
2°5 mm. per second, exceptionally 5°35 mm. per second), and 
Mesostomum tetragonum (2°66 mm. per second) show a 
markedly faster rate than the forms I have studied. 
Regarding the effect of different agencies on the rate of the 
gliding movement no special study has been made, and I can 
only report a few incidental observations. Such a study 
should be made by exact quantitative methods, and this I 
have not had the opportunity to do. What the effect of light 
on the rate is, it seems to me, impossible to say with entire 
certainty. Cole and myself! have found that light of great 
intensity (that obtained from a projection lantern with an 
electric arc as its source of illumination) causes a definite 
increase in the rate of gliding, but this increase has not been 
measured. The results of Parker and Burnett do not help us 
to answer this question of the effect of the intensity of light 
on gliding, as they are concerned only with the direction of 
its rays. The well-known phenomenon of ‘ Unterschieds- 
empfindlichkeit””? for light which Planaria shows (Loeb, 
°93, et al.) would indicate that increased light causes 
increased rapidity of movement. The electric current causes 
a very marked diminution in the rate of gliding in the weakest 
intensities which affect the organism at all. The effects of 
chemicals on the rate of gliding are not altogether uniform. 
Solutions of all chemicals tried with this point in view, when 
above a certain strength, caused a marked diminution in the 
rate of the gliding, or else an entire inhibition of it, and the 
substitution of some other form of movement. The action of 
weak solutions varied with the different substances. Very 
weak acids slightly increased the rate. Weak sugar solutions 
had no observable effect so far as rate of movement was con- 


1 Unpublished observations. 


548 RAYMOND PEARL. 


cerned. Copper sulphate causes an entire inhibition of the 
gliding movement in moderately weak solutions, even when 
these are not immediately fatal. Weak mechanical stimuli 
applied at the posterior end of the body cause a slight 
increase in the rate of gliding, but this is not marked, as any 
decided stimulus in this region of the body causes the crawl- 
ing motion to supervene. 

It is probable that the various agents affect the rate merely 
by causing changes in the general tonus of the animal. There 
is much evidence to support the view that the rate of gliding 
of an individual is a direct function of its tonic condition. 
Thus in the resting condition, which is characterised by a 
general lowering of the tonus, as will be brought out later, 
there is little, if any, ciliary movement. Again, after opera- 
tions which result in lowering the tonus, the gliding is very 
slow or entirely absent. 

2. Direction.—The direction of the gliding is, so far as 
my observations go, always forward. I have never been able 
to make the animal glide backward. This is in agreement 
with the finding of Child (:01) in the case of Sticho- 
stemma. It, of course, indicates that the effective beat of 
the cilia cannot be reversed. In the case of the planarians 
on which this work was principally done, a lateral change 
of direction of movement is not brought about by the 
stronger beating of the cilia on one side. In other words, 
when the animal turns to one side it does so by a muscular 
bending of the body in that direction, and not by ciliary 
action. In an undetermined species of triclad, however, I 
found that the most usual method of turning towards one side 
was by the stronger beating of the cilia on the opposite side 
of the body. As an individual was gliding along the bottom 
of the aquarium dish it would swerve off at an angle to its 
former course without bending its body in the slightest 
observable degree. 

b. Crawling Movement.—The second form of loco- 
inotor activity, the crawling movement, is distinctly a 
muscular movement. It takes place only when the animal 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 549 


has been stimulated in certain ways, and is of much less 
frequent occurrence than the gliding. 

The crawling is always induced when the posterior end of 
the body is strongly stimulated. The characteristics of the 
movement are as follows :—If the posterior end of a worm 
which is gliding smoothly along is touched with a needle the 
posterior half of the body immediately contracts longi- 
tudinally ; an instant later the anterior end stretches out far 
in front and fastens to the substrate. ‘Then there is a longi- 
tudinal contraction which begins just back of the head and 
runs posteriorly. This, of course, at once draws forward the 
posterior part of the body, and as this comes forward and 
gets a hold on the surface on which the animal is crawling, 
the anterior end is again extended far in front and attached 
to the substrate. This process is repeated until the animal 
settles down into the regular glide again. It consists essen- 
tially in a stretching out of the head followed by a pulling 
of the body forward by an active muscular contraction. 
When the animal is very strongly stimulated the por- 
tion of the body posterior to the pharynx usually takes no 
part in the crawling after the first general contraction. In 
fact, the posterior half of the body may even be held 
slightly raised off the bottom, while the region between the 
head and the origin of the pharynx is actively expanding and 
contracting and sending the body ahead. These strong ex- 
pansions and contractions of the anterior end which make up 
the crawling movement may follow each other in rapid 
succession as described above, or there may be a considerable 
interval between one contraction and the next. In this in- 
terval the body as a whole keeps moving ahead as a result of 
the ciliary action; that is, the gliding movement continues 
during the crawling, so that the latter may be regarded as 
an additional movement for the purpose of advancing the 
animal faster than the gliding alone can do it. ‘The crawling 
may take place, however, with the ciliary beat entirely 
stopped. 

The duration of the crawling movement after it is induced 


5a0 RAYMOND PEARL. 


is usually rather short. A single strong stimulus at the 
posterior end of the body—such, for example, as is given by 
running a needle through the body—will not usually cause 
more than three of the strong contractions of the crawling 
movement, and then the animal will relapse into the usual 
elide. ‘The limits in this matter I found to be from a single 
contraction and expansion as a minimum up to six or seven 
asa maximum. ‘his is, of course, in response to a single 
stimulus only. By repeating the stimuli the animal may be 
made to continue the crawling indefinitely. 

The effective rate of this form of progression is faster than 
that of the gliding movement. The crawling rate of one of 
the worms used for the measurement of the rate of gliding 
(the specimen 11 mm. long) was measured in the same way 
as was the gliding rate. The worm was stimulated with a 
needle at its posterior end just enough to keep it crawling, 
i.e. prevent it from settlmg into the regular glide. The 
average rate of the crawling was found to be 1°66 mm. per 
second. Merely qualitative observation shows that the worm 
ets along somewhat faster in the crawling than in the 
eliding. 

1. Direction.—The crawling may take place so as to 
carry the animal either forward or backward. The back- 
ward crawling is induced by very strong stimulation of the 
anterior region of the body. It does not always occur even 
after such stimulation, there being apparently some in- 
dividual differences among the specimens in this respect. One 
factor which will call forth persistent backward crawling is 
partial desiccation. If the dorsal surface of the organism is 
allowed to dry, it will attempt to crawl backward violently. 
The mechanism of the backward crawling is just the reverse 
of that which obtains when the animal moves forward. ‘The 
posterior end is extended and fastened to the bottom; then 
a wave of contraction, starting in this case from the posterior 
end, draws the remainder of the animal backwards, and then 
the posterior end is again extended. 'The backward crawl- 
ing is usually induced when the worm is excessively stimu- 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 551 


lated or injured at the anterior end. This movement almost 
always occurs when the head is cut off, and may usually be 
induced in such decapitated specimens for a considerable 
period after the operation by stimulating the anterior end. 
The backward crawling is not so rapid as the same move- 
ment forward. The reason for this appears to be that the 
posterior end is unable to take so firm a hold upon the 
bottom as does the anterior end. The backward crawling is 
usually not very long continued, the animal soon coming to 
rest. The inability of the animal to glide in a backward 
direction should, of course, be noted in this connection. 
Strong chemical stimulation of the anterior end will cause 
the backward movement to appear in some cases. Light, so 
far as I have observed, will not, nor will the electric current. 
There is considerable variation as to the appearance of this 
backward crawling. Some individuals cannot be induced to 
do it at all, or only in a very slight degree, while others will 
crawl backward for considerable distances after injury to the 
anterior end. It appears to be a complex of reflexes which 
under normal circumstances is inhibited, and only appears 
in any pronounced way under comparatively rare condi- 
tions. It is not, as might be expected, a method ordinarily 
used by the organism to get out of danger. This is one of 
the cases quite frequently met where an organism has among 
its available assets, so to speak, a reaction which is well 
adapted to a certain end, but of which use is nop made at all, 
or but very little. 

2. Stimuli which induce Crawling.—It may be said 
in general that almost any strong stimulus applied to the 
posterior portion of the organism causes the forward crawl- 
ing movement to appear. Mechanical and strong chemical 
stimuli applied in this region will do this. Light, either of 
ordinary intensity, or of such high intensity as that from an 
arc light, so far as I have observed, will not cause the crawl- 
ing movement. ‘The electric current does cause it, but 


greatly diminishes the rate. Any operative treatment—as, for 


example, cutting the body in two in the middle—almost in- 


552 RAYMOND PEARL. 


variably causes the portion in front of the cut to advance by 
the crawling movement, and, as has been mentioned in the 
preceding section, at the same time frequently causes the 
posterior piece to crawl backward. There is no reason to 
suppose that the operative procedure acts in this respect in 
any other way than merely as a strong mechanical stimulus 
applied at the point ef the cut. Other stimuli which induce 
the backward crawling have been taken up in the preceding 
section. 

c. Movement on the Surface Film.—Motion on the 
surface film is practically confined to the gliding movement. 
This gliding is slower in rate than that on the bottom, largely 
on account of the greater flexibility of the surface on which 
the animal is moving. While the mechanism of the move- 
ment is the same in the two cases, the surface film is elastic, 
and does not give so firm a basis as does a solid body. The 
effect of this elasticity of the film is very well seen when the 
animal attempts to change its course and turn to one side. 
The film offers little resistance to the posterior end, so that 
this cannot easily serve as a fixed point for the anterior part 
to turn about. Furthermore, in case the anterior end is left 
in contact with the film when the turn is attempted, as is 
usually the case, there 1s almost as much resistance against this 
turning of the anterior part as there is resistance to hold the 
posterior end fixed as a pivot support. The consequence is 
that the worm is unable to change its direction of movement 
quickly when on the film, and it has to go through a succes- 
sion of muscular twists and jerks towards one side before the 
result is attained. I have not been able to induce well 
co-ordinated crawling movements in a worm while on the 
surface film. ‘lhe preliminary contraction of the posterior 
part of the body occurs when that region is stimulated, but 
the subsequent stretching out of the head and drawing up of 
the body does not usually follow. I have tried stimulating 
both the exposed ventral surface of the animal and the dorsal 
surface from below, but neither method is effective. ‘lhe 
reason for this is probably to be found again in the elasticity 


MOVEMENTS, E'TC., OF FRESH-WATER PLANARIANS. 553 


of the film. The anterior end is unable to get any firm 
attachment so that the rest of the body may be drawn 
forward. Furthermore, similar resistance is offered to the 
stretching of the head forward as to the turning of it 
towards one side. 

When the animal is gliding on the surface film the same 
raising of the head (with reference to the worm, of course: 
in this case, a lowering with reference to the centre of the 
earth) and waving it about in the water occurs as under 
normai circumstances. In some cases two thirds or three 
fourths of the whole body will be thus raised and waved 
about, extending itself to its utmost capacity, and apparently 
seeking some solid body on which to attach itself. In these 
cases only a small portion of the very posterior end of the 
body will be left in contact with the film to support the 
whole. 

On coming to the side of the dish when eliding on the 
surface film, the worm almost invariably leaves the film and 
turns down the side of the dish. The reaction which is the 
cause of this and of the organisms passing from the side of 
the dish on to the film will be brought out later. 

d. Relation of Movements of T'riclads to those of 
other Forms.—In respect to their movements, the triclads 
studied occupy a somewhat intermediate position between 
certain other groups. The rhabdocceles are in general 
characterised by free movements in the water, brought 
about by cilia covering the whole body. ‘Their movement 
in general features resembles that of the holotrichous 
Infusoria. A type showing well this class of movement 
among the Turbellaria is Stenostoma leucops. The 
movement is not at all or very httle dependent upon muscular 
activity. On the other hand, the movement of many of the 
polyclads is characteristically muscular. An example of this 
is found in the case of Leptoplana tremellaris, where 
the movement is largely effected by the rhythmical beating 
of the margins (cf. Lang, 84). In fact, this form of 
movement has become so well developed in these animals 


554. RAYMOND PEARL. 


that we may think of the margins of the body as special 
locomotor organs. Ciliary action plays little if any part in 
the movement of such a form. It is to be noted, however, 
that both the rhabdocceles and the polyclads are capable of 
performing true swimming movements, 1.e. movements free 
in the water without contact with any solid body, In the 
fresh-water triclads, especially of the genera Planaria 
and Dendroccelum, the cilia have become much diminished 
in comparison with the rhabdocceles, and are restricted to a 
portion of the ventral surface only. Consequently they are 
not numerous and strone enough to support and move the 
disproportionately heavier body freely through the water. 
The movement of the cilia merely serves in these forms to 
propel the body while insufficient to support its weight. 
Consequently we find the principal form of movement to be 
a gliding over the surfaces of solid bodies.t| On the other 
hand, the fresh-water triclads have not attained the high 
development of muscular locomotion which the polyclads 
have. There is a purely muscular movement in their case, 
but it is not by far the most important form of locomotion, 
and is not so highly developed as is that of .the polyclads, 
Evidently, then, the fresh-water triclads seem to form a 
transitional stage in respect to locomotor phenomena between 
the rhabdocceles on the one hand, where purely ciliary 
locomotion obtains, and the polyclads on the other hand, 
where we find the locomotion largely if not entirely 
muscular. Whether this has any phylogenetic significance 
is not certain. 

‘he land planarians occupy a position very similar to that 
of the fresh-water forms so far as their movements are 


1 [ do not wish to imply, in this discussion of the different forms of move- 
ment as related to the number and distribution of the cilia, any belief that 
structure gave rise to function or function to structure. I wish merely to 
point out the evident correlation which exists in the matter. It seems to me 
most probable that structure and function changed together ; but in this, as 
in many other similar cases, positive evidence is lacking, and consequently 
attempts to settle the phylogenetic development of the phenomena would 
appear to be fruitless, 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 595 


concerned. There is, however, a noteworthy difference 
between the two groups. In the movements of the land 
planarians the muscular factors (rhythmical wave motion of 
the ventral surface, and snake-like movements of the whole 
body) are more important relatively to the ciliary component 
than in the fresh-water forms. In these land planarians 
there is evidently the beginning of the characteristic 
rhythmical wave motion of the part of the body in contact 
with the substrate, which reaches its highest development in 
the case of the Mollusca. 


II. Non-Locomotor Movements.! 


Under non-locomotor movements will be included the 
phenomena of contraction, expansion, “ feeling movements,” 
movements of the pharynx, etc. The purpose of discussing 
these phenomena, which are not immediately included in the 
general standpoint, is to give an account of their mechanism 
which may be referred to in succeeding portions of the paper. 
These movements are the physiological foundations on which 
the locomotor movements and the reactions are based, and it 
is necessary to determine their mechanism in order to bring 
the analysis of the behaviour to completion. 

a. Contraction of the Body.—bBy the term “ contrac- 
tion of the body,” when applied to forms like the flat-worm, 


is usually meant the shortening of the body lengthwise. In 
the flat-worm this movement is brought about by the con- 
traction of the longitudinal muscle-fibres. It may involve 
the whole body or only a portion of it. Most frequently only 
a part of the body contracts longitudinally after stimulation ; 
thus, if the anterior end is rather strongly stimulated in the 
middle line, the resulting contraction will usually involve only 
the anterior third of the body. In this longitudinal contrac- 


1 Tn discussing the musculature I have used throughout the nomenclature 
of Jijima (loc. cit.), in whose paper a very full description of this system will 
be found. I have identified in sections of P. maculata the following groups 
of muscle-fibres :—(a) outer longitudinal, (4) circular, (¢) oblique (?), (¢) inner 
longitudinal, (¢) dorso-ventral, (/°) transverse. 


556 RAYMOND PEARL. 


tion all the sets of muscle-fibres other than the longitudinal 
must be relaxed completely, because as the animal shortens 
it grows broader and thicker, which would be impossible if 
the ring, or transverse or dorso-ventral musculature, also con- 
tracted. The longitudinal musculature is. apparently better 
developed and more effective on the ventral side of the body 
than on the dorsal, because after very strong stimulation of 
the anterior end there is a well-marked tendency for the 
middle portion of the body to be raised and the head some- 
what curled in under it. ‘his relation is shown diagram- 
matically in Fig, 5. This curling under of the head does not 
appear to be a specific reaction, but, on the contrary, merely 
an expression of the fact that the ventral musculature is 
capable of shortening its side of the body more in maximal 
contraction than the dorsal side. Jijima (loc. cit., p. 378) 


Fig. 5.—Diagram showing the appearance in side view of a 
maximally contracted planarian. 


finds, from a histological study of the musculature, that the 
bundles of fibres in the main longitudinal muscle layer are 
thicker on the ventral than on the dorsal side. 

b. Extension of the Body.—The mechanisms by which 
extension of a soft-bodied animal is brought about are 
probably very different in different groups. In the case of 
the flat-worm extension is produced by the contraction of the 
circular muscular layers surrounding the body, and of the 
transverse and dorso-ventral systemsof musculature. Probably 
also the oblique musculature, when present, assists, by its 
contraction, in the extension of the body. ‘The mechanical 
necessity for extension of the body, after contraction of these 
muscles, is readily apparent. If the body, for simplicity’s 
sake, be considered a cylinder, contraction of circular 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 557 


muscles must cause it to lengthen, while with the form of 
body which really exists in the flat-worm, the contraction of 
the well-developed dorso-ventral fibres must bring about the 
flattening seen in the fully extended gliding animal. 

Probably by far the most important sets of muscle-fibres 
for producing the general extension are the dorso-ventral and 
circular. It is to be noted that contractions of any of the 
sets of fibres may take place in localised regions, producing 
extensions or contractions of that region, according to the set 
affected. The sensory, or “ feeling” movements of the head 
are brought about in this way. 

The extension, and probably also in large part the extrusion 
of the pharynx is brought about by contraction of its well- 
developed circular musculature. 

In the case of the marine mollusc, Aplysia limacina, 
Jordan (: 01, pp. 11—15) has recently shown that extension is 
brought about in an entirely different manner. It results 
from passage of fluid from vesicles in the skin into the spaces 
in the body parenchyma when the body muscles are relaxed. 
When the animal contracts this fluid (blood) is pressed out 
into the vesicles, which become very much extended and 
swollen; then, when the muscles are relaxed, the elasticity of 
the walls of the vesicles forces the fluid back into the body, 
and thus causes its extension. As a result of this method of 
expansion it is possible to kill the animal fully extended by 
the use of such poisons as cocaine. 

It is not unusual to consider the fully extended condition 
of such an organism as a flat-worm as one of approximate 
relaxation. Instead of this, it is, in fact, a condition in which 
a certain part of the musculature is in a state of well-marked 
tonic contraction. This furnishes a reason for the fact that — 
it is impossible to kill these animals in a completely extended 
condition by the use of poisons which tend to produce a 
relaxation of the muscles. Under these circumstances the 
animals take on the typical relaxed form, which is quite 
different from that of extension. 

c. Rest.—Inasmuch as a very large, if not the larger 

voL. 46, PART 4.—NEW SERIES. 00 


558 RAYMOND PEARL. 

portion of the life of a planarian is spent in a condition of 
rest, it will be well to discuss this matter; and it may, per- 
haps, best be taken up under the general heading of 
“ activities,’ although really the opposite of activity. 

The appearance of the worm when resting is, as has already 
been mentioned, quite different from its appearance in the 
active condition. The body is shorter, wider, and thicker. 
The ordinary contour of the head is almost entirely lost, and 
in place of the sharply pointed anterior end of a form like P. 
dorotocephala, the end is evenly rounded. The auricles 
disappear almost entirely, and their position is indicated only 
by the difference in the pigmentation at that part of the 
dorsal surface. The lateral edges of the body frequently have 


es ns 


Fic. 6.—Diagram showing the typical appearance of a resting planarian. 
The dotted line bounds approximately the area covered in the 
“testing movements ”’ which precede the coming to rest (ef. text). 


a wavy line instead of the straight one of the active condi- 
tion. he anterior end of the body is in contact with the 
bottom, and not raised as in movement. The general appear- 
ance of a resting planarian is shown in Fig. 6. 

The coming to rest of a gliding animal is usually done ina 
very characteristic way. Tirst, the animal-glides more and 
more slowly for some distance before reaching the point at 
which it will finally stop. The distance before reaching the 
stopping place in which the worm: glides appreciably slower 
is not, however, in most cases very considerable—usually not 
more than two or three times its own length. It is to be 
noted that this slower gliding which precedes the coming to 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 509 


rest is not in form or rate distinguishable from the other 
slow gliding motion of the worm which is not followed by 
rest. In other words, a specimen may glide slowly for a long 
time without stopping, so that one cannot prophesy with 
certainty from the rate of movement whether the specimen is 
soon coming to rest or not. ‘The coming to rest is practically 
always preceded by a period of slower gliding, but all slow 
gliding is not immediately followed by rest. After a brief 
period of this slower glidmg the worm suddenly stops, and 
the posterior half of the body remains fixed in precisely the 
same position. he anterior half of the body is slowly moved 
about over the bottom from side to side, the head being 
touched frequently to the bottom or any other solid object in 
the neighbourhood. ‘The anterior part in this “feeling” 
movement moves about the posterior part as a fixed point, 
the latter very rarely changing its position after it has once 
stopped. The thoroughness of this “testing” of the sur- 
roundings by the sensory anterior end varies much in different 
cases, but in practically all cases one can see some indication 
of it. I have in some instances seen it done very thoroughly, 
so that the whole surroundings within a radius of 5 mm. 
were gone over. Finally, when thisis done the animal comes 
to complete rest, and assumes the typical relaxed condition 
shown in Fig. 6. The apparent significance of the “ testing ” 
movements at the time of stopping is that it is a piece of pro- 
tective behaviour. The worm examines the surroundings 
before coming to rest, to see if there is anything dangerous 
(either of a solid nature or a harmful chemical) in the 
immediate neighbourhood. Whether or not this explanation 
is the true one, and further, whether natural selection 
developed this reaction for protective purposes, seems to me 
to be very doubtful, for reasons brought out in another place 
(cf. pp. 542 and 543). In some cases I have seen the worms 
come to rest by simply stopping without any appreciable trace 
of the “ feeling ” 
cedure. In coming to rest in one of the collections already men- 


movements, but this is not the usual pro- 


tioned, the “feeling” movements are usually very well marked. 


560 RAYMOND PEARL. 


There is a well-marked tendency for the planarians studied 
to come to rest in such a way that the long axis of the body 
forms a right angle, or nearly a right angle, with the lines 
of the force of gravitation. The cases in which the organisms 
come to rest with the long axis forming an angle of less than 
thirty degrees with the line of gravitation are rather few. 
Of course, when they come to rest on the bottom, the angle 
formed is approximately ninety degrees. A large number of 
observations on individuals which came to rest on the sides 
of dishes of various shapes have given the general result 
stated above. ‘here are, of course, exceptions, but there is, 


Fre. 7.—Diagram showing the positions taken by planarians coming to 
rest in a dish (see text). 


after making due allowance for these, a tendency to a hori- 
zontal position as the position of rest. ‘This is the only 
behaviour of the organism which bears any resemblance to 
a geotactic reaction. Another tendency, less marked than 
the former, is for the animals to come to rest in the angle 
formed by the sides and bottom of the dish. Not only do 
specimens come to rest lying directly in the angle, as shown 
at A in fie. 7, but also, and more frequently, they lie in such 
a position that a part of the body is on the side of the dish 
and a part on the bottom, as shown at B, fig. 7. In this 
position the animal usually lies obliquely rather than at right 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS, 561 


angles to the line of the junction of the side and bottom of 
the dish. The animals usually come to rest in this position 
atter they have been gliding on the side of the dish. When 
they come to rest from movement on the bottom of the dish, 
a position frequently taken is that shown in GC, fig. 7, where 
only the very anterior end is in contact with the side. This 
coming to rest in the angle of a dish is apparently a reaction 
which agrees with those usually called thigmotactic reactions. 
But it is not, as has been stated by several writers, due to a 
tendency to get more of the body in contact with something 
solid, than is in such contact under usual conditions; for in 
the case of an organism like a flatworm, it is impossible for 
any more of the surface of the body to be in contact with a 
solid when it is bent, as shown in Fig. 8, A, than when it is 
flat, as in B. There is the same amount of surface in contact 
in either case. The ventral surface of the flat-worm is 


Sa 


A B 


Fic. 8.—Diagrammatic cross-section of a planarian at rest—A, in 
an angle; and B, on a plane surface. 


strongly positively thigmotactic under all circumstances, and 
the dorsal surface negatively thigmotactic, but this does not 
help us understand why the animal comes to rest frequently 
in angles. This behaviour of the flat-worm in dishes is due to 
the same sort of reaction as that which causes them to come 
to rest on unevennesses on rocks, and also causes the same 
phenomenon ina more marked degree in the case of Lit- 
torina, as recently described by Mitsukuri (:01). The 
common factor in the reaction in all cases is that different 
parts of the body are brought into such positions that they 
form unusual angles with each other. Since this phenomenon 
is distinctly different from any embraced by the term thig- 
motaxis as used in its true sense, it seems desirable that it be 


562 RAYMOND PEARL. 


given a specific name. I would propose for this reaction the 
term goniotaxis.! 

When a flat-worm starts from a resting condition the 
nature of the movement, i.e. whether gliding or crawling, 
depends in large measure on the intensity of the stimulus 
which starts it. If the resting animal is rather strongly 
stimulated it will start at once into a crawling movement, 
_which changes to gliding after three or four, or fewer, con- 
tractions, provided the stimulus is not again renewed. It is 
possible by the use of a very weak stimulus to start the 
resting animal off at once into the gliding movement, or with 
only the faintest indication of a single crawling contraction. 
When the animal starts spontaneously into movement it 
usually begins at once with the glide. When starting spon- 
taneously the glide is usually preceded by some of the 
“ feeling’? movements of the head end, such as precede the 
coming to rest. The purpose of these is evidently the same 
as in the former case. Any sort of strong stimulus wiil start 
the resting animal into movement. 

The physiological condition of the resting animal is, as has 
already been mentioned, one of relaxation. All of the 
muscular systems of the body are in an apparently com- 
pletely relaxed condition. ‘This is evidenced by the form of 
the resting animal, which differs from that of one in move- 
ment, in being shorter, wider, and thicker, and in not showing 
such features as the auricles, or the pointed tip of the head. 
This relaxed condition is evidently one of lowered tonus, as 
may be determined by simple observation. Stimuli of an 
intensity which would cause a marked reaction in an in- 
dividual in an active condition, will produce no effect on a 
resting animal. ‘This point has been tested with a variety of 
stimuli, including mechanical, chemical, food, etc., and the 
markedly lower tonus of the resting animal is very evident. 
The reactions which are produced, provided the stimulus is 
made strong enough to be just effective, are weak. Of 
course, if the stimulus is above a certain strength, or is con- 


' From ywria = angle. 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 563 


tinued for some time, the animal will become generally 
stirred up and glide or crawl away. ‘This condition of 
lowered tonus in the resting animal reminds one of the con- 
ditions found in sleep in the higher animals. There, as in 
this case, the general sensory and muscular tonus is greatly 
reduced, and there seems to be no good reason why the 
resting condition of these lower organisms may not be con- 
sidered and called “sleep.” The two things appear to be 
fundamentally the same physiologically, and would appear 
to serve the same purpose. Furthermore, there is no 
apparent reason why the lower organisms should not have as 
great a need as the higher for periods of rest or sleep, 
during which the anabolic processes are in considerable 
excess over the katabolic. The fact that some lower 
organisms are so balanced physiologically that they 
apparently do not require such periods of rest is not con- 
clusive evidence that other low organisms must be similarly 
balanced. So far as is known to the writer, there has been 
comparatively little attention paid to the physiological con- 
dition of lower organisms during different phases of their 
activities. An animal which is not moving is loosely said 
to be in a “resting condition,’ when in many instances, 
as in Clepsine, the quiet animal is m a -condition of 
heightened rather than lowered tonus (cf. Whitman, loc. 
cit.). . 

As was noted in the section on “ Natural History,” the 
periods of activity of Planaria are separated by periods of 
rest of greater or less length. ‘lhe time spent in the rest- 
ing condition, at least during the daytime, is considerably 
greater, on the average, than that spent in movement. 
Probably, however, this is reversed during the night, when 
the activity is greater than during the day. ‘This periodicity 
in the activity is just what would be expected if there is a 
necessity for rest at intervals as in the higher animals. 

The causes which immediately induce the coming to rest 
may now be considered. The principal cause, as has been 
indicated above, is that the animal becomes fatigued by 


564 RAYMOND PEARL. 


movement, and its general tonus becomes lower and lower. 
Asa result of this it must remain relaxed for a certain time 
in order that recovery may take place. When in the course 
of the activity of the animal its general tonus gets below a 
certain point it stops, the actual process of coming to rest 
being a more or less gradual one. <A strong piece of 
evidence in favour of this view is the fact already given in 
the section on ‘‘ Natural History,’ namely, that if the 
animal is stirred up and made to start moving again im- 
mediately after coming to rest each time, it will be found 
that the periods of activity become progressively shorter. 
Furthermore, when the general physiological condition of 
the organisms is weakened by keeping them for a time in 
the laboratory, it is found that the periods of rest become 
progressively longer in proportion to the periods of activity. 
The general “ predisposing condition ” to the coming to rest 
is then probably a lower tonus due to fatigue. The im- 
mediate causes determining the exact place chosen are of 
three sorts. First, and probably most important of these, is 
the intensity of the light. It is well known that planarians 
tend to come to rest in regions of comparatively low hght 
intensity, the reaction having been first noted by Loeb (793), 
and called by him “ Unterschiedsempfindlichkeit.” This 
factor seems to be the most important of any in determining 
the region in which the animals come to rest, both under 
experimental conditions and in the natural habitat. In 
aquarium dishes placed close to a window, and containing 
considerable plant material, the worms will be found resting 
practically always in the half of the dish away from the 
window. ‘The largest number of individuals will be 
entangled in the plant material, and usually for the most 
part invisible ; while of those specimens resting on the sides 
and bottom of the dish the greatest number will be found in 
such places that there are heavy masses of plant material 
between them and the window. <A few will come to rest far 
around on the sides of the dish where the glass itself cuts off 
some of the light. ‘This last position has been mentioned by 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 965 


Loeb as the one most frequently taken by planarians in a 
dish containing only water. This behaviour towards hght 
is not, however, an absolutely precise reaction. Many times 
during experiments I have seen specimens come to rest in 
the very lightest parts of the dish and remain there; but in 
general this reaction will cause most of the animals to gather 
in shaded areas. It is probably the principal factor in 
causing the animals to take positions beneath stones in their 
natural habitat. 

The second immediate cause in determining where the 
animals shall come to rest is the goniotaxis mentioned above. 
If an animal in the proper physiological condition of reduced 
tonus comes to an unevenness in the surface on which it is 
moving, it will in most cases come to rest there. ‘This, again, 
is not a very precise reaction; not sufficiently so as to make 
it possible to predict beforehand where any given individual 
will stop. In this case, just as in the case of light, much 
depends on the animal’s physiological condition, and when in 
the proper condition they may come to rest on a perfectly 
smooth surface. ‘Thus in a dish individuals will always be 
found at rest on the smooth sides and bottom, yet there is a 
distinctly marked tendency, when the animals are put under 
experimental conditions and closely observed, for them to 
come to rest in the angle of the dish. This reaction probably 
also plays a considerable part in the habit of coming to rest 
among the branches and leaves of the plant material. In 
the natural habitat it is undoubtedly the factor which causes 
them to take positions on the uneven parts of stones. It 
may be that the immediate cause of the stopping in this case 
is the increased resistance to movement afforded by the 
unevenness of the surface. This, acting on an animal ina 
fatigued condition, might give the necessary stimulus for 
the stopping. 

The third factor in determining where the animals shall 
come to rest is one about which I am doubtful. There seems 
to be some evidence, from the behaviour of the animals 
themselves, that in the formation of the groups or collections 


566 RAYMOND PEARL. 


previously mentioned (pp. 533, 534) there is a sort of chemo- 
kinesis. That is to say, the presence of some chemical 
substance in the water causes the animals to stop. The 
evidence for this factor will be taken up with the discussion 
of the formation of collections. It probably does not play 
any part in determining where a single individual shall come 
to rest outside of a collection. 

It must be emphasised that all of these three factors are 
secondary in importance as compared with the physiological 
condition of the animal, which may be said to prepare it for 
the resting state. An active animal, in which the tonus is at 
or near the maximum, will pass through regions of low 
illumination, uneven surface, or collections of other imdi- 
viduals without stopping. Only when the animal is in the 
right general condition do these factors come in to determine 
the precise point where the stop shall be made. 

1. Formation of Collections.—Since the formation of 
collections is dependent on the animals coming to rest in a 
certain area, it may properly be taken up in this section. 
‘he collections are fairly well-defined groups of from six or 
eight up to twenty or more individuals. The general appear- 
ance of such a group is shown in Fig. 1. The individuals 
composing it have no definite orientation, but are scattered 
about with the anterior ends directed in whatever way they 
happeued to be pointed when the individuals stopped. The 
distance separating the individuals varies much in different 
cases. In some cases it may be as much as a half-centimetre, 
or again may be the width of an individual worm or less. 
This formation of collections of this sort might be considered 
the result of a “social instinct ” by animal psychologists of 
the Binet school. Actually, it appears to be due to two 
simple reactions taken in conjunction with the general 
physiological condition of the individuals composing it. ‘The 
first of these reactions is that to light. ‘That is to say, when 
individuals come to a comparatively restricted area of a 
certain degree of illumination, if they are in a certain 
condition of reduced tonus, they stop. ‘Those which are 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 067 


very active pass on through the region, but necessarily in 
course of some time several individuals will have stopped, 
and a group will have been begun. When once started 
another reaction apparently enters to assist in enlarging it. 
This reaction appears to be due to some chemical substance, 
and belongs to the class of reactions which Engelmann has 
suggested should be called “kinetic,” in this case chemo- 
kinesis. It would appear that planarians excrete or secrete 
some chemical substance towards which they are themselves 
positively chemotactic, and which also causes them to come 
to rest. When several individuals remain quiet in a small 
area this substance, of course, accumulates and affects other 
individuals passing. That some such a substance is sepa- 
rated from the bodies of the animals is evidenced by two 
phenomena. First, in the case of the food reaction, which 
will be taken up in detail later, it is found that after one or 
two individuals have attached themselves to a piece of food 
material and begun feeding the mass of food and planarians 
is a much more effective stimulus to positive chemotaxis than 
is the same food substance alone, even though it may have 
remained in the water a greater length of time. The “zone 
* (vide infra, p. 626) of the food and feeding 
individuals together is much wider than that of the food 


of influence’ 


alone. Specimens are affected at a greater distance from 
the food and react more sharply. Asa result of this, dense 
ageregations of planarians will be formed in a comparatively 
short time after the first two or three individuals have found 
a bit of food. As there is no reason to suppose that the 
action of the food itself is different in the two cases, we 
must conclude that the greater effectiveness of the food and 
feeding individuals is due to some chemical substance coming 
from the organisms themselves. 

The second line of evidence for the existence of a reaction 
to a chemical in the formation of collections is found in the 
behaviour of specimens coming near a group of individuals 
resting on the bottom of a dish. When some distance away 
from the outer boundary of such a group a gliding animal 


568 RAYMOND PEARL. 


will frequently be seen to give a well-defined positive 
reaction, and turn towards the group. The reaction is of 
precisely the same character as that given by the organism 
to weak chemicals (to be described later), and the behaviour 
convinces one observing it that the specimen is stimulated 
by some chemical diffusing out from the group. After 
turning towards the group the specimen will glide into it 
and usually come to rest, in the manner which has been 
described above. 

What the nature of the chemical substance present in the 
region about the groups is, I have not been able to discover. 
Neither rosolic acid nor methyl orange is discoloured by it. 
Whatever its nature, it must be in an extremely diluted 
state. ‘his seems evident for two reasons: first, because it 
does not affect delicate indicators; and second, because it 
does not have any effect on active specimens of Planaria. 
A large number of experiments have been performed to test 
this latter point, but always with the same result. Unless 
the individuals were in the proper predisposing condition of 
lowered tonus, they would pass by or through groups of 
other individuals without giving any reaction. 

Attempts to produce, artificially, collections of planarians 
in chemicals have been unsuccessful. I have tried various 
solutions (such as sugar, weak alkalies, etc.) to which the 
organisms showed a well marked positive chemotaxis when 
tested by other methods, but have not been able to get any 
format of collections in them. ‘The animals would give 
the positive reaction on coming to the edge of the diffusing 
chemical and pass into it, but would not come to rest. This 
failure to produce collections artificially is not surprising 
when one considers the number of conditions necessary for 
the production of the desired result. The organism must be 
in just the right physiological condition, the chemical must 
be of a certain concentration, and finally, it must be located 
in an area of a certain light intensity. It is practically 
almost or quite impossible to fulfil all these conditions at 
the same time in an experiment. 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 569 


The coming to rest in the collection seems to be due 
simply to the direct effect of the chemical on the organism. 
There is no evidence that the animals are held in the group 
as a result of a negative reaction to the surrounding water, 
as is the case in the collections formed by the infusoria 
(cf. Jennings, 799, b). The method of formation of collections 
in chemicals in the case of the infusoria is as follows :— 
Specimens swimming about at random come to the edges of 
drops of chemicals purely by chance. If, for example, the 
chemical happens to be a weak acid, the specimens will pass 
into the drop without giving any reaction. When, however, 
they reach the opposite edge of the drop and attempt to pass 
from the chemical back into the water they are stimulated, 
and give their usual motor reaction. This turns them back 
into the drop, in which they are, as it were, “caught in a 
trap.” As a consequence of this method of reaction a very 
dense collection will be formed in a short time. With the 
flat-worm the case is very different in that an active in- 
dividual frequently passes into and out of one of these 
collections without showing the faintest trace of a reaction 
on either side. The only way in which any stopping in the 
region is brought about by a chemical is by a chemokinetic 
reaction. The fundamental difference in the reactions of the 
two groups of organisms on which this difference in the matter 
of forming collections is based will be brought out in the 
section on the reactions to chemicals. 

To sum up, the formation of collections of individuals 
seems to be due, in the first instance, to the tendency of the 
organisms to come to rest in areas of a certain degree of 
intensity of light, and in a lesser degree to a tendency to 
turn towards and come to rest in areas containing some 
substance secreted or excreted by the worms themselves. A 
prerequisite in the formation of collection, as in the coming 
to rest under any circumstances, is a proper physiological 
condition of reduced tonus. 

There does not appear to be any special biological sig- 
nificance to this tendency of the animals to collect in groups. 


570 RAYMOND PEARN. 


The behaviour is not of any evident benefit to the organisms, 
as it is in the case of infusoria, where it is apparently closely 
connected with the obtaining of food. On the contrary, it 
seems to be, at least potentially, a harmful thing, because 
any accident or enemy would affect a number of individuals 
rather than a single one when they were so collected. 

d. The Effect of Operations on Movement. — It 
may be well to put together in one place the results which 
were obtained with reference to the movements from animals 
which had been cut in various ways. From these we can 
form some idea of the relation of the nervous system of the 
planarian to its movements.! 

The immediate effect of any operation is that of a very 
strong mechanical stimulus applied to the same part of the 


Fic. 9.—Operation diagram. The heavy straight line indicates the cut 
made. For results see text. 


body, and the sort of movement resulting in each piece 
depends on the position of the cut. The details of this im- 
mediate effect will be described in connection with other 
mechanical stimuli. What concerns us here is the permanent 
after-effect of operations on the movements. We can best 
oet at this matter by taking np some specific cases. 


1 All the operations were performed with a sharp scalpel, in most cases 
with the specimen in a dish of water, In some cases the worm was transferred 
toa drop of water on a soft board for the cutting, but in all cases where 
immediate observations were wanted, the operations were performed in the 
dishes used for the experiments. The only difficulty in performing operations 
on planarians arises from the fact that if the edge of the knife is allowed to’ 
rest on the surface of the body for even a very short time before the cut is 
made, it will become covered with the stieky slime from the animal, and then 
any clean cut is impossible. ‘The edge will slip off the back of the worm 
without penetrating. 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 571 


If a planarian is cut squarely across the body in the 
region a short distance behind the head, as indicated in 
Fig. 9, the anterior piece will continue to move after the 
operation at approximately the same rate as the whole 
animal did before. After the immediate effect of the opera- 
tion is past the glide is its ordinary movement, and it will go 
about the dish and behave in general like a whole individual. 
At the outstart its periods of activity and rest are distributed 
about as in a normal individual, or, in other words, its power 
of spontaneous movement is not impaired, at least for a time. 
On the other hand, the posterior piece comparatively soon 
comes to rest after the operation. Its glidig movement is 
slower, and the periods of rest become longer and longer in 
comparison with the periods of activity. Its power of spon- 
taneous movement becomes very greatly diminished within a 
comparatively short time after the operation, and it remains 


ee aes 


Fic. 10.—Diagrammatic side view of a decapitated specimen performing 
the gliding movement. 


in the relaxed resting condition during the greater part of 
the time spent in the process of regeneration. When this 
posterior piece does glide about soon after the operation its 
anterior end is usually raised off the bottom considerably 
higher than is the head of a normal flat-worm under similar 
circumstances. This is shown in Fig. 10. There are no 
“feeling” movements of the anterior end of such a piece, 
but instead this end is held very stiffly in the raised position. 

If, instead of making the cut so close behind the head, it is 
made back in the middle region of the body, the anterior 
piece behaves as before, i.e. like the normal animal. The 
posterior piece, however, moves slower than did the corre- 
sponding piece in the previous experiment, and it comes to 
rest sooner after the operation, and remains quiet longer. 

In the same way cuts may be made nearer and nearer the 
posterior end ; the posterior piece will move more and more 


572 RAYMOND PEARL. 


slowly, and come to rest sooner. At the same time the 
anterior pieces will appear more and more like the normal. 
In both sets of pieces the crawling movement may be in- 
duced by proper stimulation of the posterior ends. 

Oblique transverse cuts produce the same results as do 


Fic. 11.—Operation diagram. Heavy lines indicate the cuts made. 
For results see text. 
direct ones. The same laws hold as to the movements of the 
pieces. In case a strip is cut from the side of the body, as 
shown in Fig. 11, the smaller piece A curls up, and does not 
make any further progressive movements, although it 
remains alive, and will eventually regenerate in most cases. 
The main part B contracts on the cut side, and hence becomes 
curved in that direction after the operation. It is able to 
move about, but at a somewhat slower gliding rate than 
normal, and in a path curved towards the cut side. In case 
a worm is slit down the middle line at the anterior end, as in 
Fig. 12, it is able to glide, but at a slower rate than 
normal. It performs the crawling movement in response 
to stimulation at the posterior end, and each half of the head 
performs feeling: movements independently of the other half. 


Fig. 12.—Operation diagram. The heavy line indicates the cut 
made. For results see text. 
An individual slit up in the middle line from the posterior 
end, as in Fig. 13, glides at approximately the normal rate, 
provided the cut is not carried:too far forward. If the cut 
extends into the head region the gliding becomes im- 
mediately slower. Such a specimen performs the crawling 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 573 


movement upon stimulation of the posterior end of either 
piece, but in a peculiar way, which will be described later. 
Putting all these results together, we see that there is a 
general tendency for animals on which operations have been 
performed to glide at a slower rate than normal. In some 
of the pieces this tendency is very slight, and frequently 
hardly noticeable. In others the movement is very much 
slower than normal. In all cases the periods of rest are 
longer during the time of regeneration than normally. This 
tendency for the animals to remain quiet during regenera- 
tion increases up to a certain point as regeneration proceeds. 
A piece of a planarian may be quite active for three or four 
hours after the operation, while during the following three 
or four days it will scarcely move atall. After the regenera- 
tion is practically complete the worm will begin to move about 
again approximately as it normally does. During the re- 


Fic. 13.—Operation diagram. The heavy line indicates the cut made. 
For results see text. 
generating process the anterior pieces, bearing an uninjured 
head, are much more inclined to move about than are the 
posterior parts. These latter usually remain entirely quiet 
during regeneration. 

This behaviour of the posterior parts during regeneration 
appears to be distinctly purposive, and to belong to the class 
of phenomena called regulatory. The general tonus of these 
pieces is immediately lowered by the operation, and conse- 
quently they keep quiet. Yet at the same time the processes 
of morphallaxis and, in many cases, growth begin at once, 
and proceed very vigorously till the missing parts are 
restored. If we consider that the worm or part of a worm 
has at the beginning a certain sum-total of energy available 
for all activities, including movement, growth, morphallaxis, 
and all its other vital processes, then it would appear that 

voL. 46, PART 4.—NEW SERIES. PP 


574 RAYMOND PEARL. 


the performance of any single set of activities in excess must 
sause a corresponding diminution in other activities. This 
is exactly what we find to be the case with the regenerating 
planarian. While the processes concerned in regeneration 
are at their maximum activity, we get a decided reduction in 
the amount of movement. It would seem, then, that a large 
part of the energy which is ordinarily expended in movement 
is used after operation or injury in the processes of regene- 
ration. As the regeneration nears completion more and 
more energy is available for, and used in movement. ‘This 
would seem to be a sort of “energy regulation.” The 
behaviour is evidently further beneficial in the case of the 
posterior pieces, because their anterior ends are very insensiI- 
tive as compared with the head of the normal animal, and 
if they moved about they would certainly be more apt 
in the long run to get into difficulties than if they remained 
quiet. 

It may be well in closing the section to point out the 
relation of the nervous system to the movements. Loeb 
(:00) has maintained that “if we divide a fresh-water 
planarian, for instance Planaria torva, transversely, the 
posterior half, that has no brain, crawls just as well as the 
oral half. Spontaneity in Planaria torva is, therefore, by 
no means a function of the brain.” If by “crawl” in the 
first sentence we understand “ elide” to be meant, the state- 
ment is not strictly accurate. The posterior pieces do not 
“move just as well as the oral” (anterior), but, as has 
already been brought out, more slowly. For a very short 
time after the operation the statement would in some cases 
be correct, but it certainly would not be twenty-four hours 
later, according to all the observations I have been able to 
make on the subject. As for the spontaneity of the move- 
ment, that also becomes very much lowered with the loss of 
the brain, as I have attempted to show above. ‘The very 
much lessened activity of posterior pieces of planarians has 
been mentioned by Lillie (: 01, pp. 152, 133). 

Irom my own observations it seems clear that the principal 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 575 


function of the brain of Planaria with reference to move- 
ments is to maintain the tonus of the ciliary system. That 
neither the crawling nor the gliding movements are specific 
functions of the central nervous system is evident, because 
both sorts of movement may take place after its removal. 
Yet all my observations tend to show that after injury to or 
loss of the brain the gliding movement becomes, almost 
immediately, markedly slower. This relation is especially 
well indicated by the experiments noted above on splitting 
the animal longitudinally from the anterior and the posterior 
ends. In the one case the gliding movement becomes at 
once distinctly slower, while in the other case there is only a 
slight difference in the rate, evidently conditioned by the 
fact that only comparatively few of the cilia can get a hold, 
so to speak, so that they can function. The force of the 
argument will be impressed if one glances at the relative size 
of the cuts in Figs. 12 and 13, and then remembers that the 
rate of gliding of the specimen figured in Fig. 13 is faster 
than that of the one in Fig. 12. With the co-ordination of 
movements, including the crawling, the central nervous 
system has very little to do in the case of Planaria. With 
regard to the spontaneity of movement it is difficult to decide 
in how far the brain functions. It is certain that regenerat- 
ing anterior pieces show more spontaneous movement than 
do posterior pieces, yet the anterior pieces are behind the 
normal worm in this respect. The brain probably plays 
some part in the performance of normal spontaneous move- 
ments, but, as has been pointed out, in these operation 
experiments the whole matter is very definitely related to 
the regenerative process, and loss of substance plays nearly, 
if not quite as great a part as loss of nervous system. 
Summarising, we may say that 


1. For the performance 
of the crawling or gliding movements the brain is not specifi- 
cally necessary. ‘hese movements are normally co-ordinated 
in the absence of the brain. 

2. The maintenance of the tonus of the ciliary system 
(which produces the gliding movement) is a specific function 


576 RAYMOND PEARN. 


of the brain, and is, further, its most important function so 
far as movement is concerned. 

3. The brain plays a certain part in the production of 
spontaneous movements. 


F. REACTIONS TO STIMULI. 


I. Reactions to Mechanical Stimuli. 


Since the reactions which are given by Planaria to 
mechanical stimuli are in a sense the foundation on which 
the reactions to other stimuli are based, it may be well to 
consider them first. After thoroughly working out the 
reactions to mechanical stimuli we have a very definite clue 
to practically all the animal’s behaviour. 

a. Methods.—For rough, general work with mechanical 
stimuli a needle or a sharp-pointed scalpel may be used as 
the stimulating agent. For the finer work in sharply local- 
ising the stimulus, I at first made use of pieces of glass 
tubing drawn out to capillary fineness. This method was 
not, however, satisfactory, as the glass was too stiff to admit 
of reaching all points of the body under some circumstances. 
Furthermore, this stiffness, together with the sharpness of 
the end, made it almost impossible to give the animal a 
moderately strong stimulus without wounding it. A far 
better plan was found to be to fasten with sealing-wax a 
moderately stiff piece of human hair to a piece of glass 
tubing, the latter to serve as a handle. With such an 
arrangement the stiffness of the stimulating point can be 
varied by varying the length of the hair. Danger of 
wounding the animal is avoided, yet repeated strong stimuli 
may be given, while, further, the flexibility of the hair makes 
it possible to stimulate the animal at any point and from any 
desired direction. 

An annoying difficulty in connection with this work was 
the clinging of the slimy secretion of the body to the point 
used for stimulating. Once coated with this slime the 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 577 


sharpest point will slide off the body without giving any 
effective stimulation. 

b. Description of Reactions.—The reactions can. best 
be described by taking up in order the typical results 
following stimulation of the different parts of the body. 

1. Stimulation of Head Region.—If a_planarian 
gliding along on the bottom of a dish be touched with a 
needle on one side of the head, it will, under normal circum- 
stances, in the majority of cases, turn the head and anterior 
one fourth of the body away from the side stimulated, and 
continue gliding along in the new path determined by the 
turning of the anterior end. This “ turning away” reaction, 
or, as we may call it for economy of words, negative reaction, 
will always be given if the stimulus is made sufficiently 
strong. ‘There is a certain intensity of stimulation below 
which the negative reaction may or may not be produced, 
depending on the physiological condition of the individual, 
but above which it always occurs. If, again, a normally 
eliding planarian be selected for stimulation, and this time 
the stimulating point (preferably something finer and more 
flexible than a needle) be touched very lightly to the 
edges of the sides of the head or the auricles, we get, 
provided the specimen is in the proper physiological con- 
dition, a very graceful and striking reaction, quite different 
from that obtained in the former case. his time the flat- 
worm will stop for the briefest instant, turn the head and a little 
of the anterior end of the body towards the side stimulated, 
and at the same time raise the head from the bottom, until 
finally the tip of the head points exactly towards the point 
from which the stimulus came, and then glide forward in that 
direction. This “turning towards” or positive reaction is 
given only in response to very weak mechanical stimuli, and 
then only when other conditions are favourable. It is a very 
precise and characteristic reaction when it does appear. 

Having outlined the two main reactions following mechan- 
ical stimulation in the head region, we may proceed to con- 
sider each of them in more detail. 


578 RAYMOND PEARL. 


a. Reactions to Strong Stimuli.—The negative reac- 
tion is the characteristic reaction given to all strong stimuh, 
whether mechanical or of some other sort. It is, further, the 
same type of reaction which most organisms with fairly well- 
differentiated reactions give in response to strong stimulation. 
It takes the animal away from what might be a dangerous 
object. 

In Planaria the portion of the body which takes part 
in the turning away varies with the strength of the stimulus 
to a certain degree. Stimuli just strong enough to call forth 
the negative reaction will cause only the head to be turned 
away. The first turn away of the definite reaction never in- 


B 


Fic. 14. —Diagram showing the form of the negative reaction to mechan- 
ical stimuli. A shows the position just before the stimulus is 
applied, and B the position after the reaction. 


cludes any of the body back of the pharynx, so far as I have 
observed, except in the case of very strong and repeated 
stimuli. In the typical and most often observed form of the 
negative reaction the portion of the body which turns away 
is that anterior to a point about halfway between the level 
of the eyes and the point of origin of the pharynx. This is 
shown in Fig. 14. With stronger stimuli the point of turning 
is farther back on the body. 

The number of degrees through which the head is turned 
in the negative reaction depends on the intensity of the 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 579 


stimulus. With stimul just effective in calling forth the 
reaction the turn is only slight, and since it affects only the 
head end the direction of movement of the whole animal may 
be scarcely changed at all. The amount of turning of the 
anterior end is typically from 30° to 40°, 

There is in the negative reaction a pause at the instant of 
stimulation, preceding the turning away. The first effect of 
the stimulus is to cause the animal to stop its relatively rapid 
movement. ‘his pause may be so slight as to be almost 
imperceptible in the case of comparatively weak stimuli, or, 
on the other hand, may lengthen to a quite noticeable 
interval when the stimulus is very strong. It is a character- 
istic feature of both the positive and negative reactions of 
planarians, and is evidently due merely to the fact that before 
a reaction (1. e. something involving a change of motion) the 
former movement must stop. 

The effect of localisation of mechanical stimuli in the head 
region may next be considered. As has already been men- 
tioned, stimulation of the sides of the head produces the 
positive or negative reaction according to the intensity of the 
stimulus. There are no special regions of specific sense- 
organs connected with either of these reactions. The nega- 
tive response is given after strong stimulation of any part of 
one side or the other of the head and, so far as it is possible 
to observe, just as decidedly after stimulation of one part as 
of another. It is of interest to know what happens after 
stimulation of the head in the median line. It is very 
difficult to get a stimulus exactly in the median line, but one 
may come very near it by stimulating the dorsal surface of 
the head in the region between the eyes. The reaction pro- 
duced is a longitudinal contraction of the anterior part of the 
body, drawing the head back away from the stimulus. The 
head is then turned to one side or the other as in the usual 
negative reaction, and the animal starts ahead again in the 
new direction. ‘The side towards which the turn is made 
after median stimulation is indeterminate—that is, there is no 
tendency to turn in more cases towards one side than towards 


580 RAYMOND PEARL. 


the other, as has been found by Frandsen (:01) to be the 
ease with Limax. ‘This is what would be expected in the 
case of the flat-worm, because it 1s a perfectly bilaterally sym- 
metrical organism. Probably what actually determines which 
way the organism shall turn after attempted median stimula- 
tion is the fact that the stimulus really acts a little to one 
side or the other, and the turning is really the negative 
reaction. 

By repeatedly stimulating the anterior end of a worm with 
moderately strong mechanical stimuli its reactions may be 
modified. In the beginning of such a series of stimulation 
the worm turns away farther and farther from each suc- 
ceeding stimulus, at the same time remaining at the same 
place in the dish, i. e. not making any progressive move- 
ments. ‘This process tends to make the animal describe a 
circle away from the stimulus, about its posterior end as a 
fixed point. It never completely describes a circle, how- 
ever, but after several stimuli have been given, to which it 
has responded progressively more vigorously, it finally 
jerks back with a strong longitudinal contraction, and turns 
the anterior end through a considerable arc, so that it points 
in an entirely different direction. ‘his final strong reaction 
in the majority of cases turns the anterior end towards the 
side from which the stimulation is coming, or, in other words, 
in an exactly opposite direction to that of the previous 
reactions. This reaction appears as if, after the animal has 
tried in vain to get away from an uncomfortable stimulus by 
its ordinary reaction, it finally tries a wild jump in the oppo- 
site direction. This curious change in the reactions induced 
by a repetition of strong stimuh I have observed many times. 
It indicates the effect of the organism as a whole on its 
reflexes. As von Uexkiill (:00, p. 73) has well brought 
out, we must consider that in the case of a higher organism, 


like a dog, the animal moves its legs, while with a lower 


5? 
organism whose activities are reflex—for example, the sea- 
urchin—it is really the “legs” (1. e. locomotor organs) which 


move the animal. In the flat-worm the movements of the 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 581 


whole organism are determined by definite stereotyped 
reflexes, yet in such exceptional cases as the one just 
described the organism as a whole takes control, and does 
something quite different from what the normal reflex fitted 
to the case would accomplish. 

Very strong mechanical stimulation of the anterior end, 
such as to wound the animal, causes a very much more 
vigorous reaction than the ordinary negative one, and of a 
slightly different form. The animal contracts strongly longi- 
tudinally, and, as a result of the heavier musculature on the 
ventral surface, curls the head in under the body. Then the 
anterior end is turned to one side through a larger angle 
than is usually the case, and the worm straightens out in this 
new direction. The point of importance to be noted in this 
reaction to maximal stimuli is the curling under of the head. 
The turn away from the side stimulated frequently is so great 
as to turn the animal squarely about, so that it heads in the 
direction opposite to that before stimulation. Besides this 
effect of maximal stimuli just described, they may also pro- 
duce a change in the movement from gliding to crawling. 
The crawling does not usually follow stimulation of the 
head end of the body, but it is possible in some cases to pro- 
duce it by very strong stimulation here. I have also been 
able in a normal animal to induce crawling backward by 
very strong and continued stimulation of the anterior end of 
the body. ‘This backward crawling, when it occurs, is of the 
same character as the same movement in a forward direction, 
except that all the factors are reversed. It has been 
described above (cf. p. 551). It is much more easily pro- 
duced after certain operative procedures, and in connection 
with them further details regarding it will be brought out. 

The negative reaction, i.e. that to strong stimuli, is given 
more frequently than any other in the course of the activity 
of the individual, and apparently does not depend on the 
presence of any special physiological condition. It is given 
in response to stimuli covering a wide range of intensity. 
The lower liminal value of the stimulus producing it (there 


582 RAYMOND PEAR. 


is apparently no upper limit) varies to some extent with the 
physiological condition of the individual. ‘Thus in some 
specimens at certain times stimuli which would ordinarily 
produce a rather strong negative reaction will call forth 
nothing but the positive reaction. ‘his condition is only a 
transitory one, and the reason for it seems to be a heightened 
tonic condition of the animal. Specimens exhibiting this 
relation to rather strong stimuli are always very active, and 
move about with great rapidity, frequently raising the 
anterior end of the body and waving it about through the 
water as they glide along. Persistent strong stimulation 
of the organism rapidly changes the general physiological 
condition. ‘This is not more true of stimulation applied to 
the head region than of strong mechanical stimulation of any 
part of the body. The animal becomes “ stirred up” gene- 
rally, moves about with increased rapidity, its sensitiveness 
to stimuli becomes diminished, and it will give only the 
negative response to stimulation of the anterior end. ‘This 
change in the physiological condition of the animal as a 
result of continued stimulation of any sort, as in a series of 
experiments, is a matter of great practical importance in 
connection with reaction work. One may get totally differ- 
ent appearances from an individual which has been “ stirred 
up” from what are seen in the case of one which is in the 
normal condition. This is only one of a number of factors 
which must be taken into account in work on the reactions 
and behaviour of an organism if one is to obtain trustworthy 
results. It is almost an absolute necessity that one should 
become familiar, or perhaps better intimate, with an organism, 
so that he knows it in something the same way that he 
knows a person, before he can hope to get at even an approxi- 
mation of the truth regarding its behaviour. 

B. Reactions to Weak Stimuli.—The positive reac- 
tion is the characteristic reaction given to all weak stimuli. 
It is an orienting reaction in the sense that it brings the 
anterior end of the animal in a position such that it points 
approximately towards the source of the stimulus. On 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 583 


account of its fineness of adjustment with reference to the 
strength of the stimulus and the general physiological 
condition of the animal, it is a response which mht be 
very easily overlooked in a superficial examination of the 
behaviour. As the worm gives this positive reaction in 
response to a gentle stimulus, turning the head towards the 
source of stimulation, and at the same time raising it, it 
gives one the impression that it is seeking something, and 
such the behaviour would doubtless be called by some 
animal psychologists. This impression is enhanced by the 
fact that if the head does not come in contact with the 
stimulating object at the first reaction, the animal advances 
in the direction from which the stimulus came, with the 
anterior part of the body raised and waving from side to 
side in the water. 

As has been mentioned, the reaction is very delicately 
adjusted physiologically. In the majority of cases the 
animal must be in a comparatively quiet condition,—that is, 
not “stirred up” or excited, and gliding smoothly at the 
ordinary rate, in order that the reaction may appear at all. 
The stimulus must ordinarily be very weak, and given so as 
not to disturb the animal by abruptly changing the sur- 
rounding conditions. It is possible to produce the reaction 
by the use of a needle or scalpel point if sufficient care is 
taken, but better results are obtained by the use of a hair 
as the stimulating point. The point should be lightly touched 
to the edge or dorsal surface of the head, and then quickly 
drawn a short distance away. Even when all these precau- 
tions are taken one may fail to produce the characteristic 
response. I have frequently found that the same specimen 
which at one time would give the positive reaction in a very 
definite and characteristic way to every light touch on the 
head could not be made to show it a few hours later. This 
shows how closely it depends on general physiological 
conditions. On the other hand, specimens will frequently 
be found that for short periods of time (two or three hours) 
can hardly be induced to give any other response to mechan- 


084 RAYMOND PEARL. 


ical stimulation of the head. Stimuli strong enough to be far 
above the usual upper liminal value for this reaction will call 
it forth. Such specimens show the reaction in a much more 
pronounced type than is usually the case. After a stimulus 
has been given they will turn towards it, and if the source is 
not touched immediately they will remain in the same spot 
waving the head about the region from which the stimulus 
came, at the same time stretching the anterior end of the 
body far out in all directions, precisely as if in search of the 
stimulating body. Usually this hypersensitive condition 
passes off in a short time, and the animals behave again in a 
more normal fashion. It was thought that possibly this 
condition was due to hunger, but experiments! devised to 
test this question indicated that this was not the case. 
We can only say that it is due to some intimate physio- 
logical condition, the exact nature of which we do not know. 
Another fact which may be mentioned in this connection is 
that sometimes a specimen in normal condition will give the 
positive reaction in response to a certain strength of stimulus 
only a part of the time. Other trials result in entire indiffer- 
ence on the part of the organism. Of course, it is not 
possible to give mechanical stimuli always of the same 
strength, yet with the closest possible approximation to this 
by an experienced operator, some of the trials will not affect 
the animal in any way except to cause a slight local con- 
traction at the point on the head stimulated. The worm 
glides along without any change in rate or direction. 
Altogether we must conclude that the reaction is one which 
is very closely dependent on the existence of certain definite 
internal conditions as well as the external ones. 

‘The typical course of the reaction is, as has been described, 
first a momentary pause, followed by a turning of the head 
towards the stimulus, accompanied by a raising of the ante- 
rior part of the body. From this typical form of the reaction 
there are many variations. The raising of the anterior end 
from the bottom just before and during the time it is being 


1 See section on “ Reactions to Food and Chemicals.” 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 585 


turned towards the source of the stimulus may be entirely 
omitted. In this case the head is swept around towards the 
stimulus without being any further raised from the bottom 
than in the ordinary glide. The duration of the pause 
immediately following stimulation is likewise subject to 
great variation. It may be so diminished as to be imper- 
ceptible, the worm sweeping the anterior end around through 
the water without any change in the rate of the glide. The 
amount of the turn varies with the point of application of the 
stimulus, the head being turned just far enough to point in 
the direction from which the stimulus comes. _ ‘'l'his orienta- 
tion, if we may so call it, is generally quite exact. If the 
stimulus is near the middle line on the edge of the head the 
turn will be only through a few degrees, while if the auricles 
are touched it will amount to nearly 90°. This fact indicates 
the remarkably well-developed co-ordination of the reaction, 
There is a great deal of variation with regard to what takes 
place after the turn has been made, and the anterior end is 
directed towards the stimulus. If the stimulating point is 
removed immediately after stimulation, so that the animal 
does not touch it by means of the first reaction, a normal 
specimen will usually lower the head and continue gliding in 
the new direction. As has been meutioned, however, in 
some cases a specimen will continue ‘ feeling ”’ about in the 
locality for some time. If the stimulating point is held in 
about its original position after the stimulus has been given, 
the first reaction will in most cases bring the head into 
contact with it. In this event the animal usually moves the 
tip of the head about over the hair (or other point) for a 
short time, and then drops back to the bottom and continues 
gliding. In other cases it will clasp the anterior end about 
the hair (as in the feeding reaction to be described later), 
and then ina moment start gliding up over it. When this 
happens the hair or needle may be moved about in the water 
or even lifted out of it, and the animal will not let go its 
hold and drop off. If the needle is held quiet, however, the 
animal will in a short time glide down off it and proceed on 


586 RAYMOND PHARL. 


its way along the bottom. ‘This behaviour when the animal 
is able to reach the stimulating object is evidently the action 
which most frequently occurs in natural environmental con- 
ditions. 

With regard to the localisation of the stimulus producing 
this reaction, I may say that I have been able to produce it 
by proper stimulation of any part of the edge or dorsal 
surface of the head region under favourable circumstances. 
It seems to be more certainly produced—that is, in a larger 


number of cases—by stimulation of the auricles than of any 
other part of the head, and it may be that in this is to be 
found an indication of the chief function of these sense-organ 
bearing structures. At any rate, this is the only indication 
of a special function for them which I have been able to 
discover. he positive reaction given in response to light 
stimulation of the dorsal surface of the head is necessarily 
somewhat different from the typical reaction which has been 
described. In this case there can be no turning towards one 
side, because if this were done the head would not be directed 
towards the source of the stimulus. Instead, what takes 
place is this: the head is sharply raised and twisted, so as 
to form a part of a spiral in the region posterior to the head. 
‘This brings the anterior end into a position pointing towards 
the source of the stimulus, and at the same time the ventral 
surface is brought around so as to be, in most instances, the 
first portion of the body to touch the stimulating point. 
This reaction, following stimulation of the dorsal surface of 
the head, is not an easy one to obtain. I have succeeded 
best in producing it in the case of individuals in the hyper- 
sensitive condition mentioned above. 

With regard to the strength of the stimulus necessary to 
call forth the positive reaction, only very relative statements 
may be made. Unfortunately we have no method of measur- 
ing the intensity of such weak mechanical stimuli as are 
used in work on lower organisms. Our only idea of the 
strength of the stimulus must come from the reaction of the 
organism itself. It must suffice to say, regarding the reaction 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 587 


under discussion, that in an animal in a condition of hyper- 
sensitivity I have been able to produce the reaction by the 
weakest stimult which I was practically able to give. Under 
normal conditions of sensitivenes it takes a slightly stronger 
stimulus. No absolute value can be given for the upper 
limen of the reaction, beyond which it does not appear, but 
gives place to the negative reaction. ‘his value varies 
greatly with different individuals. The general statement 
may be made that the positive reaction is the characteristic 
response to stimuli of very low intensity, and its production 
is very closely dependent on the proper gradation in the 
intensity. ‘This dependence is so close that it is possible to 
obtain a part of both the negative and positive responses 
combined in the same reaction by the use of a stimulus of 
the proper intensity. I have been able in a few very favour- 
able cases to produce by a single stimulus a pronounced 
raising of the head, such as is characteristic of the positive 
reaction, followed by a turning away from the source of the 
stimulus. Now the raising of the head is no part of the typical 
negative reaction, and, furthermore, was done in the very 
characteristic way in which it occurs in the positive reaction. 
The stimulus which produced it was evidently about in- 
termediate in intensity between what, in the case of the par- 
ticular animal used, would have called forth either the positive 
or negative reaction, as the case might be. This experiment 
shows in a very striking and conclusive way that in both 
the positive and negative reactions we are dealing with a 
complex of reflexes, since here a part of one of the reactions 
is associated with a part of the other. This point will be 
alluded to again in another connection, and its significance 
more fully pointed out. 

The evident purposeful character of the positive reaction is 
plainly apparent. It is a reaction admirably suited, on the 
whole, to bring the organism into contact with beneficial 
things, such as food, ete. It seems to me that it must be by 
far the most important reaction of the animal in the struggle 
for existence. In the conditions under which planarians 


588 RAYMOND PEARL. 


live a reaction which gets it food, or helps to, is of far 
greater importance for the survival of the individual than a 
reaction which takes it out of danger; for, so far as observa- 
tion can show, the dangers it encounters are relatively few. 
It does not move over large areas of territory, and, so far 
as is known, it does not furnish a considerable part of the 
food supply of any other organism. Altogether its chief 
struggle for existence would seem to consist in obtaining 
subsistence for itself, and for this the positive reaction to 
mechanical stimuli would appear to be an important aid. 
As will be shown later, the food reaction proper consists 
largely of this same response. 

We may now pass to a consideration of the— 

2. Reactions to Stimuli applied to the Middle 
Region of the Body.—I use the term “ middle region of 
the body” to distinguish that portion extending from the 
posterior border of the head to about the middle of the 
pharynx. The separation of the body behind the head into 
a “middle” and a “ posterior” region is based entirely on 
physiological considerations, and is not defined morpho- 
logically. 

a. Reactions to Strong Stimuli.—Strong mechanical 
stimulation of the middle region of the body along the lateral 
edges causes, in the first instance, a local contraction of the 
body in the immediate region of the stimulus. This local 
contraction 1s well marked; much more distinct than that in 
the head region. If the stimulus is sufficiently strong, and 
especially if the stimulating point is applied to the edge 
from above rather than from the side, the previous gliding 
movement will be changed to crawling. ‘This will continue 
for a brief interval, usually from two to four crawling con- 
tractions being given; then the animal will relapse again 
into the glide, provided the stimulus is not repeated. In 
the case of a strong stimulus applied to the side of the 
middle region of the body, especially if the stimulus is 
several times repeated, we get the negative reaction—a 
turning away from the side stimulated—just as in the similar 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS, 589 


case in the head region. The nearer to the anterior end of 
the middle region the stimulus is applied, the more easily 
will the negative reaction be produced, while back in the 
pharyngeal region it follows even strong stimulation in fewer 
eases. In all cases where this reaction is not produced the 
direction of movement is either unchanged by mechanical 
stimulation, or the anterior end may be brought around 
very slightly towards the side stimulated as a mechanical 
result, of the local contraction on that side. By repeating 
the strong stimuli on one side of the middle region, summa- 
tion effects similar to those described above as taking place 
when the head is similarly treated are not produced. The 
animal crawls faster and faster away from the stimulus. Its 
direction of movement is changed, but usually not more than 
thirty to forty degrees. We see here evidence of precise 
response to localisation of stimulus. Stimulation of the head 
causes the animal to turn to one side, and, in case the 
stimulus is very strong, to contract longitudinally strongly 
before doing so. As we go back from the head in the 
middle region of the body, the tendency to crawl rapidly 
ahead away from the stimulus increases. At the same time 
the turning away from the stimulus becomes less and less 
marked the farther back it is applied. In no case do we get 
any strong retraction of the anterior end, which, in case of 
stimulation of the middle region of the body, would tend to 
bring the animal, or at any rate a part of it, into further 
contact with the stimulus. 

Strong stimulation applied to the dorsal surface of the 
middle region of the body causes the animal to change from 
the glide to the crawl. ‘This change of the form of motion 
may be regarded as a specific reaction in response to strong 
stimulation of the middle or posterior regions of the body. 
Stimulation of the dorsal surface of the middle region does 
not change the direction of the movement unless the stimulus 
is applied near the lateral margin, in which case it may 
cause the negative reaction, as mentioned above. 

B. Reactions to Weak Stimuli.—Weak mechanical 

vol. 46, pARY 4.—NEW SERIES. QQ 


590 RAYMOND PEARL. 


stimulation of the sides of the middle region of the body 
causes, in the first instance, a small local contraction at the 
place stimulated, without any effect on the general direction 
of movement of the whole organism. Under favourable 
circumstances, however, it is frequently. possible to get a 
quite different result by the use of a weak stimulus on the 
lateral margins of this region. Specimens of Planaria may 
be induced to give the characteristic positive reaction 
described above by stimulation of a point as far back as the 
middle of the pharyngeal region. The stimulus inducing it 
is of the same intensity and character as that which will pro- 
duce the same result in the head region. 

Some experiments bearing on this point will be reported 
in full, on account of their important bearing on theoretical 
questions to be taken up later. ‘These experiments were to 


Fic. 15.—Diagram to show the portion of the side of the body (a 4) 
within which weak stimulation produces the positive reaction. 


test the effects of weak stimuli on the sides of the middle 
region of the body, especially with reference to the relation 
of the physiological condition of the individual to its 
reactions. The experiments were performed on _ large 
specimens of P. maculata, and the region of the body 
stimulated was that from a to b in Fig. 15. Most of the 
stimuli were applied in the region between the auricles and 
the origin of the pharynx. ‘The stimuli were given by 
moving the point of a fine scalpel along the bottom of the 
dish till it came into contact with the margin of the body. 
In this way no general disturbance was produced. The 
attempt was made to make the stimuli of as nearly as 
possible the same intensity each time. The results were 
classed as positive, negative, or indifferent, according as the 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 591 


specimens gave the usual positive or negative reaction, or 
kept their course without change, showing only the local con- 
traction. The first series which I will report was on an 
individual which was in a condition of great excitation, 
moving about with more than normal rapidity, and generally 
“stirred up.’ The results of twenty-three stimulations on 
this specimen were— 


Expt. I. Positive responses . ; Lg eee 
Negative responses ; 0 pecimen in state o 


excitation. 
Indifferent responses. 29 ; 


A similar experiment with another individual in an 
entirely normal unexcited state, gliding at a moderate rate, 
gave the following results : 


Expt. II. Positive responses 20 
Negative responses : 
Indifferent responses. 


Specimen in normal 
condition. 
13 


The striking preponderance of the positive reaction in the 
case of the unexcited individual is notable. The same 
individual used in Experiment II was now “ stirred up” by 
poking it violently about the dish with a needle for about 
five minutes. It was then allowed to settle into a glide 
which was at a more rapid rate than normal, and another 
series of stimulations was made, with the following results : 


Expt. III. Positive responses 
Negative responses. 
Indifferent responses . 


1 
[Specimen in condi- 
3] tion of excitation. 


Here, again, the indifferent responses are in excess, and 
there are practically no positive reactions. The specimen 
was again “stirred up” in the same way as before, and 
another series taken. 


Expt. 1V. Indifferent responses (trials 
1 to 11 inclusive) 
Positive responses (trials 12; 
and 13) 


Negative responses (0). 


Specimen in state of 
excitation. 


592 RAYMOND PEARL. 


The specimen was again stirred up in the same way and 
another series taken, with the following results: 


Expt. V. Indifferent responses (trials 


1 to 9 inclusive ? ; 
) Specimen in state of 


Positive responses (trials 10>; eee 
E (tri: excitation. 


and 11) . 
Negative responses (0) ; 


The positive, responses in all these experiments were very 
definite and characteristic. I have obtained the same results 
in many other series of experiments, which need not be 
recorded in detail. ‘The experiments show very clearly that 
in order for the animal to give positive responses to weak 
stimuli it is necessary that it be in an unexcited condition, 
These results have also an important bearing on the question 
of the mechanism of the positive response, in that they show 
conclusively that the reaction does not depend on the stimu- 
lation of special sense organs located in the head regions 
alone. 

Weak mechanical stimulation of the dorsal surface in the 
middle region of the body is usually without any effect other 
than the causing of a slight local contraction at the point 
stimulated. If any specific effect on the whole animal is 
produced, it is merely a change from the gliding to the 
crawling movement, such as results from strong stimulation 
in the same region. 

3. Reactions to Stimuli applied to the Posterior 
Region of the Body.—By “ posterior region of the body ” 
I mean that part of the body from the pharyngeal region to 
the posterior end. This region is not sharply marked off 
physiologically from the middle region, and it is impossible 
to say in any given individual at just what level the demar- 
cation will be found. The physiological distinction between 
the two regions is founded on the fact that it is possible by 
unilateral stimulation of the middle region of the body to 
produce a change in the direction of the movement of 
the animal as a whole, while in case of the posterior region, 


MOVEMENTS, BTC., OF FRESH-WATER PLANARIANS. 093 


as will be shown, this cannot be done. On this account it 
will not be necessary in the description of the reactions to 
sharply distinguish between the effects of stimulation of the 
margins and of the dorsal surface, as has been done in the 
previous cases. 

Strong mechanical stimulation of the posterior region of 
the flat-worm produces as a specific reaction an immediate 
change from the gliding to the crawlmg movement. ‘The 
direction of the crawling is the same as that of the gliding ; 
that is to say, the worm keeps on in a straight line, taking 
itself directly and in the quickest possible way away from 
the stimulus. The duration of the crawling movement 
following stimulation of the posterior region varies with the 
relative intensity of the stimulus and the physiological con- 
dition of the specimen. ‘The most usual number of the 
strong, crawling contraction waves following strong stimula- 
tion is three or four. We may get a smaller number than 
this, and very frequently do, but in the species studied I 
have very rarely seen more than four of the general con- 
tractions following a single stimulus. This is evidently all 
that would be necessary under normal circumstances, since 
four of these strong contractions will carry the animal a 
considerable distance ahead, and probably out of reach of 
the stimulating agent. ‘he weaker the stimulus is, the fewer 
are the contractions and the shorter the distance crawled. 
In some individuals it is at times almost impossible to induce 
the crawling movement except by repeated stimulation. 
Such specimens will merely draw up the posterior end in a 
single crawling contraction after stimulation, and then im- 
mediately relapse into the glide. If a strong stimulus is 
repeatedly given at the posterior end the crawling is con- 
tinued, becoming more and more rapid. This is the only 
effect of continued stimulation in this region, there being no 
summation effect corresponding to that produced by stimu- 
lating the anterior end. No different effect is produced by 
stimulating the margins of the posterior region of the body 
from what takes place when the point stimulated hes near 


594. RAYMOND PEARL. 


the middle line. There is no turning towards or away of 
any part of the body. The lack of any special effect of 
unilateral stimulation is not surprising, for the reason that 
rapid movement in a forward direction will get the animal 
away from harmful stimuli affecting this region, in the long 
run, more quickly than any other. Further, there would be 
no advantage in the production of a positive reaction by 
stimuli at the posterior end. If we think of these reactions 
as having been developed by natural selection there would 
be no possibility of such a reaction having arisen, for the 
reason that practically any favourable stimulus would be 
encountered by the anterior end before it possibly could be 
by the posterior. Very weak mechanical stimulation of the 
posterior end of the body causes only a local contraction at 
the point stimulated. 

4, Reactions to Stimulation of the Ventral Sur- 
face.—In the descriptions of the reactions to mechanical 
stimuli up to this point we have been considering stimuli 
applied to the dorsal surface and to the margins of the body. 
It may be well to describe briefly what the reactions in 
response to localised stimulation of the ventral surface are. 
This matter can best be tested when the animal is moving on 
the under side of the surface film, with its ventral side 
uppermost. It might be supposed before the trial was made 
that this habit of the animal would afford ideal conditions for 
testing its reactions to ventral stimulation, but, as a matter 
of fact, the conditions are anything but ideal. ‘The flexibility 
and elasticity of the surface film makes it almost impossible 
to touch it with a stimulating point anywhere within a radius 
of a centimetre about a planarian without causing the animal 
to be jerked bodily to one side or the other, quite sharply 
and for some little distance. This is, of course, a mere 
mechanical effect, which takes place with lifeless bodies also. 
Furthermore, as has been mentioned in an earlier section, it 
appears to be very difficult for planarians to quickly change 
the direction of their movement when on the surface film (as 
is necessary in reacting to stimuli). On account of these 


MOVEMEN'IS, ETC., OF FRESH-WATER PLANARIANS. 595 


conditions it is very difficult to get any certain and trust- 
worthy results from the stimulation of the ventral surface. 
My results have been as follows :—strong stimulation of the 
anterior end on one side of the middle line causes the 
negative reaction just as when the stimulus is applied at a 
corresponding point on the dorsal surface. For mechanical 
reasons the response is not as extensive as when the animal 
is on a solid, but there seems no doubt of its character. The 
positive reaction to weak stimuli I have not been able to 
produce in any certainly recognisable form in response to 
stimulation of the ventral surface, but I think this negative 
result is due probably to the external conditions, and not to a 
real failure of the organism to react. Strong stimulation of 
the posterior end of the body causes the gliding to change 
to the crawling just as under other conditions. Very 
strong mechanical stimulation of the ventral surface of the 
body causes the animal to let go its hold and pass down 
to the bottom. 

5. Reactions of Resting Specimens to Mechanical 
Stimuli—A resting specimen gives no response whatever 
to weak stimuli which are still strong enough to produce a 
definite reaction when the worm is in the active condition. 
The stimulus is simply below the threshold of the resting 
animal’s sensitiveness. To stronger stimuli the reactions 
correspond in form with those given by the active animal, 
but are less pronounced. For example, rather strong stimu- 
lation at the anterior end induces a weak negative reaction ; 
similar stimulation of the posterior end sets the animal off 
into the crawling motion. Strong stimulation of any part of 
the body besides producing the characteristic reaction for 
that region (that is the negative reaction) will also in most 
cases start the animal into movement. ‘This will always be 
the case if the stimulus is of sufficient strength, or is several 
times repeated. As would be expected from the low sensi- 
tiveness of the resting flat-worm, it is impossible to call 
forth from it any positive reaction. 

.6. Reactions to Stimuli given by Operative Pro- 


596 RAYMOND PEARL. 


cedure.—Hvidently when a planarian is cut the cutting 
induces a strong stimulation, which is of the same kind as 
that induced by ordinary mechanical stimuli, only much 
more intense. ‘The immediate effects of operations may then 
be taken up in this section. 

If we take first the typical case given by cutting the 
animal transversely in two in the region between the pos- 
terior border of the head and the origin of the pharynx, and 
make the cut by a single stroke of a sharp scalpel, we find 
that the effect on the anterior piece is precisely the same as 
that of an ordinary strong mechanical stimulation of the 
same place. That is, this piece merely changes from the 
oliding to the crawling movement, and after giving three or 
four crawling contractions settles down again into the 
elide. This is the same result essentially as that obtained 
by Norman (: 00) and earlier by Loeb (94 and :00). In the 
behaviour of the posterior piece in this experiment under 
discussion there is a great deal of variation. In about 70 
per cent. of all cases in which I have observed the results of 
such an operation, the posterior piece crawled backwards 
as a result of the cut. In the remainder of the cases the 
piece either stayed in the same place and contracted 
violently, or else glided ahead. The amount of the back- 
ward crawling when this occurs varies greatly, from a 
short distance involving only one longitudinal crawling con- 
traction to several times the length of the worm, the move- 
ment lasting in this latter case for over a minute. In order 
that this backward crawling may appear in a well-marked 
and distinct form it is necessary that the posterior piece be 
above a certain size. Very small posterior pieces after 
operation usually remain quiet. 

A cut so made as to split the anterior end of the body in 
the middle line in most cases causes the worm to crawl back- 
wards just as does a transverse cut. In some cases this, as 
well as other operations, merely causes the animal to contract 
violently and squirm about at the same place. Splitting the 
posterior end of the body in the middle line causes the parts 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 597 


on either side of the cut to give violent longitudinal con- 
tractions, while the worm as a whole starts crawling ahead ; 
that is, it changes from the gliding to the crawling 
movement. 

Oblique cuts produce essentially the same effects as would 
transverse cuts in the same part of the body, i. e. forward 
crawling of the anterior piece, and usually backward crawling 
of the posterior piece. This is true unless the cuts are very 
oblique, so as to form very acute angles with the sagittal 
plane of the body. In such cases the effects produced more 
nearly resemble those obtained in complete longitudinal 
splitting of the body. If the body is split completely into 
two parts longitudinally, there is usually very little pro- 
gressive movement of either piece afterwards. ‘The pieces 
contract strongly on the cut sides very soon after the opera- 
tion is performed, so that they take on the form of a bow, 
which in many instances becomes a nearly complete circle. 
This being the case, any progressive movement, either by 
gliding or crawling, is nearly or quite impossible. Cuts 
involving only a small portion of one side of the body 
produce, if in the anterior region, the characteristic negative 
reaction given to other strong mechanical stimuli, while if in 
the posterior region they cause the crawling ahead. 

Cuts made on the resting animal produce essentially the 
same effects as on the gliding specimen. Unilateral cuts 
have the same effect in producing the negative reaction. 

7. The Effect of Mechanical Hindrance to Move- 
ment.—A series of experiments was performed on Dendro- 
coelum, sp., with reference to the behaviour of the animal 
when progressive movement was made impossible, and yet 
the animal was stimulated strongly at the same time. ‘These 
conditions can be realised by thrusting a needle through the 
centre of the body from above, and then holding it fixed in 
position. The results of this procedure varied somewhat, 
according to the portion of the body through which the 
needle was thrust. In case the hindrance is in the posterior 
region of the body, e. g. at a point just behind the posterior 


598 RAYMOND PEARL. 


end of the pharynx, the effect immediately following the 
thrusting in of the needle is a strong longitudinal contraction 
of the whole body. After this first strong contraction the 
animal remains perfectly quiet in the contracted form for a 
varying length of time (in some cases as long as five minutes, 
but usually less). After this period of quiet a series of 
rhythmical waves of contraction pass longitudinally over the 
still contracted body. The purpose of these waves is 
evidently to loosen the restraining object by making the 
hole in the body through which it passes larger. This is the 
same behaviour that I have observed in the deposition of the 
large egg. ‘This process of rhythmical longitudinal contrac- 
tion 1s continued for a time; then the animal stretches to its 
extreme length, attaches the anterior end to the substrate, 
and attempts to crawl away. The movement of the anterior 
end is precisely the same as in crawling. The animal turns 
and twists and struggles violently in this attempt to crawl 
away, and the cilia beat strongly. If the needle occupies a 
position near the edge of the body this first struggle will 
usually be sufficient to tear the body loose from the needle, so 
that the animal may then move ahead freely. Such specimens 
will, of course, have a large jagged wound in one side of the 
body, which, however, closes in and heals in a short time. In 
case the first struggle of the extended animal to crawl ahead 
is not effective, that is if the needle is too far in towards the 
centre of the body to make the tearing out possible, the 
animal, after continuing the struggle for a time, contracts 
strongly longitudinally and goes through the whole series of 
stages of quiet, rhythmical, longitudinal contraction and 
attempted crawling again. The only difference between the 
first and succeeding series of trials is that the stages in which 
the animal is strongly contracted longitudinally tend to 
become shorter with each repetition. 

In case the needle is thrust through the body in front of 
the pharynx, the strong longitudinal contraction appears as 
before, and is followed after some time by an extension of the 
part in front of the needle, while the rest of the body re- 


MOVEMENTS, ETO., OF FRESH-WATER PLANARIANS. 599 


mains quiet and contracted. This short anterior region, 
including hardly more than the head, goes through the 
crawling movements, but on account of its small size is very 
ineffective so far as pulling the body away from the needle 
is concerned. In my experiments I have never seen any 
worm succeed in getting free from a needle put through the 
body in this position. 

This general behaviour of the animal in response to restraint 
of movement is very interesting, especially in the cases where 
the restraint is at the posterior end, as showing the relation 
between the behaviour and the capability of regenerating. 
The organism tears itself loose from a restraining body with 
entire nonchalance, as it were, and its confidence is well 
founded because no permanent harm comes from the action. 
The lost and wounded parts are regenerated and healed in a 
short time. The behaviour takes advantage of the ability to 
regenerate. Whether the form of behaviour (pulling away 
from restraining objects) or the power of regeneration and 
reparation appear in the organism first we cannot say, for 
either might very well follow, in a more or less remote causal 
connection, the other. What we do know is that at present 
there is a very nice condition of mutual adaptation between 
the two things. 

The effect of the hindrance of a rather light weight at the 
posterior end of a worm is to induce the crawling movement. 
This can be seen in case the animal is feeding on a small 
piece of food material, and, as frequently happens, starts into 
movement before the pharynx is withdrawn. The piece of 
food attached to the end of the pharynx is dragged along 
behind, and the movement is the crawling. Frequently, also, 
in feeding experiments pieces of food will get stuck to the 
posterior end of the worm by means of the mucous secretion 
of the body, and these have the same effect in inducing the 
crawling movement. 

Having now obtained a descriptive basis we may pass to a 
discussion of some general features of these reactions. We 
may first take up— 


600 RAYMOND PEARL. 


c. The General Features of the Reactions to 
Mechanical Stimulii—From the above description it 
appears that the nature of the reactions to mechanical stimuli 


depends upon several factors. These are 

1. The intensity of the stimulus. 

2. The localisation of the stimulus. 

3. The physiological condition of the organism. 

The reactions given may be of several different kinds, de- 
pending on the factors mentioned above. These are chiefly 
as follows: 

1. The resting individual may begin locomotion. 

2. The gliding movement may be changed to the crawling 
movement. 

3. The forward movement may be transformed to move- 
ment backward. 

4, The animal may turn away from the source of the 
stimulus (the “ negative ”’ reaction). 

5. The animal may turn towards the source of the 
stimulus (the ‘ positive ” reaction). 

It is evident that the reactions last named—the negative 
and positive reactions—are the most important and most 
interesting from the theoretical standpoint. It is of the 
greatest interest to note that these two qualitatively opposite 
reactions’ are induced merely by differing intensities of 
stimuli, the stimuli being otherwise identical throughout. 

lt is to be noted further that the positive and negative 
reactions have the characteristics of purely reflex acts. Hach 
reaction has a perfectly definite and characteristic form. 
While, in some cases, which of the two reactions will be 
given in response to a particular stimulus depends on the 
physiological condition of the organism, yet it 1s practically 
always either one or the other of the typical reactions. Only 
very rarely do we get any deviation from the type forms, 
and in such cases the reaction is evidently a combination of 
easily recognisable components of the two typical complexes 
of reflexes. 

These two reactions are evidently not single simple 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 601 


reflexes, but are complexes of several simple reflex acts. It 
may be well to present in tabular form the different com- 
ponents in each of these reactions, indicating by the position 
in the table the relations of the parts. 


Component Phases of the Reactions to Mechanical 
Stimuli, with special reference to the Head 
Region. 


PosITIve. NEGATIVE. 
A. Momentary stopping of pre- A. Same as in positive. 
vious movement. Referred 
to as ‘‘ pause” or ‘ hesita- 
tion’ in description. 
B. Longitudinal extension of the 6. Longitudinal contraction of ante- 


anterior end to greater or rior end of greater or less 
less extent. Amount de- intensity. ‘Tends to make A. 
pends on previous extension. appear more pronounced and 
Usually distinctly noticeable. longer in duration. 

/C. Turning towards one side, viz. C. Turning towards one side, viz. 
that stimulated. This side that not stimulated. Defined 
is defined by the position of as in positive. No sharp 
the source of the stimu- “ orientation.” 


lus, not structurally, Sharp 
“orientation.” 

C’. Raising of anterior end. This 
takes place at the same time 


as C. 

D. Movement towards stimulus. D. Movement away from stimulus. 
Direction determined by Direction determined as in 
position taken by anterior positive. 


end at termination of C. 


Time relations are indicated by vertical position in the table. Components 
occurring at the same time are included in braces. 


Kach of the components before D may be considered as a 
single reflex, and thus there are in one case four and in the 
other case three simple reflexes which go to make up the 
whole reaction. ‘That these reactions are composites of the 
distinct parts is evidenced, first, by direct observation of the 
reactions themselves; and second, by the fact that it is 


602 RAYMOND PEARL. 


possible by varying the strength of the stimulus to produce 
only certain parts of the whole reaction without the 
remainder, and, furthermore, that a part of one reaction may 
in rare instance be combined with a part of the other 
(v. sup., p. 587). 

d. Mechanism of the Reactions.—A question which 
is of the greatest importance in all work on the reactions of 
organisms is, what is the mechanism of the reaction? In 
the case of the flat-worm this becomes, what is the neuro- 
muscular mechanism of the reactions? Very little direct 
evidence bearing on this question can be obtained from the 
reactions themselves. Taking the positive and negative 
reactions as they occur, there are several different sets of 
muscles and of nerve connections by means of which they 
might conceivably be brought about. ‘The best evidence on 
the question is the indirect evidence from operation experi- 
ments, in which parts of the mechanism are injured or 
removed. 

1. Relation of the Brain to the Reactions.—The 
first specific problem which may be taken up may be stated 
thus: is the brain necessary for the performance of the 
normal reactions to mechanical stimuli? Or, in other words, 
will a planarian from which the brain has been removed 
react normally to stimuli? This question can be answered 
from the study of specimens which have been cut in two 
transversely, and consequently we may proceed at once toa 
description of the reactions of the pieces resulting from such 
an operation. A typical specimen is cut in two transversely 
at the level of a point about halfway between the head and 
the origin of the pharynx, as shown in Fig. 16. As has been 
mentioned above, the cut itself acts as a strong mechanical 
stimulus, and the immediate effect of the operation is to set 
both pieces crawling, the anterior one ahead and the 
posterior one usually backward. 

If now the pieces are allowed some hours to recover from 
the immediate effect of the operation, and then stimulation is 
tried, the following results are obtained :—With the anterior 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 6083 


piece A, containing the brain, the results are entirely similar 
to those obtained in case of the normal animal. Strong 
unilateral stimulation of the head causes the negative 
reaction, weak stimulation of the same sort the positive 
reaction. Stimulation at the posterior end causes the crawl- 
ing movement to appear, and altogether the appearances are 
essentially the same as in the normal complete specimen. 
The posterior piece B (lacking the brain) behaves in a 
somewhat different manner. If the anterior end of this piece 
is given a stimulus of moderate intensity anywhere on the 
cut surface the piece will usually start crawling straight 
backwards. This is almost always true for a short time 
after the operation, and is especially well shown in such 
specimens as started crawling backwards as a result of the 
cut. When from twenty-four to forty-eight hours have 
elapsed after the operation this tendency of posterior pieces 


Fie. 16.—Operation diagram. Heavy line indicates cut. 


to crawl backward on stimulation of the anterior end begins 
to grow less marked, and, as regeneration proceeds, finally 
disappears. In many such posterior pieces I have been able 
to produce this backward crawling in a very pronounced 
form, and of comparatively long duration (three or four 
minutes at a time). The character of the movement has 
been described above. If the stimulus is applied to one side 
or the other of the anterior end of such a posterior piece, 
instead of squarely against the cut surface, a well-marked 
negative reaction is produced; that is, the anterior end 
turns away from the stimulus just as a whole animal would. 
The reaction is very definite, and of precisely the same 
character as the normal negative reaction. The only 
difference to be observed is that in proportion to the strength 
of the stimulus the reaction is not so pronounced as in the 


604 RAYMOND PEARL. 


normal animal, this being due to the generally lowered tonus 
in such a piece. I have not been able to obtain any positive 
reaction (i.e. turning towards the stimulus) in such a 
posterior piece after operation. Stimuli which are at all 
effective produce the negative response. This experiment 
has been tried many times, but always with the same result ; 
the positive reaction never appears. If the posterior end of 
such a posterior cut piece is stimulated the crawling move- 
ment is produced just as in case of the normal complete 
animal. As has been noted in connection with the move- 
ment, there is a general reduction of tonus in the posterior 
pieces resulting from transverse cuts. This low tonus in- 
volves not only the motor functions, resulting in slower 
movement, but also to a less extent the sensory functions. 
Such a piece is somewhat less sensitive to mechanical stimuli 
than normally. The cut surface is more sensitive to 
mechanical stimuli than any other part. 

Now it will be seen from the above description of the 
reactions of a piece from which the brain has been removed, 
that the most striking difference in the behaviour of such a 
piece from that of a normal animal is to be found in the 
absence of the positive reaction. 

There are three conceivable possibilities as to the cause of 
the absence of the positive reaction in pieces from which the 
head has been removed. First, the positive reaction might 
be due to the stimulation of certain sense organs which are 
removed by the operation. But this is decisively negatived 
by the fact that in an entire worm stimulation of points 
posterior to the level of the cut removing the anterior end 
will cause the positive reaction. 

Second, it might be conceived that the reaction is brought 
about by a special localised muscular mechanism, which is 
removed or destroyed by the cut. But there is no evidence 
of the existence of such a mechanism ; and further, it will be 
shown later that the ordinary musculature of the body, which 
is of course uninjured in the posterior part, is sufficient to 
bring about the reaction. 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 605 


Finally, the positive reaction might im some way be a 
specific function of the brain, which is removed by the 
operation. As the evidence seems to be decisive against the 
first two possibilities this seems probably true. Is this 
because the brain contains a special “ centre”? whose function 
it is to produce the reaction ? 

There is no reason to think of the reaction as a function of 
the brain in the sense that that organ forms a centre which 
originates the impulses which cause the reaction. On the 
contrary, it seems much more probable that the loss of the 
brain causes the loss of reaction for the following reason. 
It has been shown that removal of the brain causes a general 
lowering of the tonus of the organism, and further that the 
appearance of the reaction in a normal animal is closely 
dependent on the tonic condition of the organism. Probably, 
then, the chief reason for the non-appearance of the positive 
reaction in posterior pieces is that in these the conditions of 
general tonus are so changed by the loss of the brain that the 
reaction is no longer possible. Expressing it in another way, 
the animal is too sluggish to give the positive response. 
This being the case, it would be expected that it might be 
possible to induce the positive reaction in a decapitated 
specimen provided the tonus were raised in some way. Asa 
matter of fact, as will be shown later, positive reactions to 
certain chemical stimuli have been observed in a few cases 
(cf. p. 649). In its form and mechanism the positive reaction 
is not directly dependent upon the brain. 

Summing up the evidence on the relation of the brain to 
the reactions of the flat-worm, it may be said that all the 
reactions to mechanical stimuli shown by the normal animal, 
with the single exception of the positive reaction, are given 
by specimens from which the brain has been removed. 
The relation of the brain to the positive reaction is, in large 
part, so far as evidence can be obtained, an indirect one, viz. 
it is necessary for the maintenance of the proper tonic 
conditions of the organism. ‘Thus far there is no evidence of 
any special “ centre ” functions of the brain, similar to those 

vol. 46, PART 4,—NEW SERIES, RR 


606 RAYMOND PEARL. 


supposed to exist in the cortical centres, for example, of a 
mammal. 

2. The Neuro-muscular Mechanism.—In the negative 
reaction to mechanical stimuli the anterior end of the body 
is turned sharply away from the source of the stimulation, 
while in the positive reaction it is equally sharply turned 
towards the source. These relations immediately suggest the 
following questions :—Is the negative reaction the result of a 
crossed impulse, which, originating at the point stimulated, 
crosses over to the other side of the body and causes the 
contraction of the longitudinal muscles on that side, thus 
producing the turning away from the stimulus? What is the 
course of the nerve impulse which produces the positive 
reaction ? What sets of muscles are concerned in the pro- 
duction of each reaction ? 

The discussion of the negative reaction may be taken up 
first. If the nervous impulse producing this reaction 
crosses the body to produce a contraction on the side 
opposite from the stimulus, the experiment cited in the 
section above shows that this crossing cannot occur entirely 
in the brain, but must also occur in some part of the body 
posterior to the brain; or at any rate, be capable of so doing 
in a quite normal fashion immediately after removal of the 
brain. In this experiment where the body has been cut 
in two behind the brain, the posterior piece performs the 
negative reaction in a quite normal way immediately after 
the operation. ‘This experiment may be carried farther, 
and the animal cut in two transversely in places nearer 
and nearer to the posterior end of the body. In all of these 
cases, until the piece becomes too small to show definite 
movements of any sort, the negative reaction may be obtained 
by strong unilateral stimulation. ‘This shows conclusively, 
then, that if the negative reaction is to be considered a 
crossed reflex, there must be all along the body a series of 
cross-commissures which are at all times ready to bring about 
in co-ordinated perfection a result with which they have never 
previously had anything to do. This conclusion seems in- 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 607 


evitable because, as has been shown above, unilateral 
stimulation of the posterior region of the body in a normal 
individual does not cause the negative reaction, but instead 
merely causes the animal to move ahead faster by crawling. 
If these paths for the crossing of impulses which are so 
immediately effective after the operation are present in the 
uninjured specimen, one would expect the reaction to be of 
quite a different character from what actually occurs. A 
stimulus applied near the posterior end would naturally cross 
over at once and produce a bending on the opposite side at 
the same level. Or the stimulus might diffuse, so that the 
entire opposite side would be affected and the worm would 
become uniformly curved on that side. But as a matter of 
fact we find that the turning affects only the anterior portion 
of the body. If it is urged that after operation the crossing 
of impulses takes place through the general protoplasm the 
difficulties encountered are no less, for it must be shown how 
passage of an impulse through the protoplasm to cause a 
perfectly well co-ordinated reaction can appear so quickly and 
produce such perfect results at once. If tested immediately 
after the operation, before the general lowering of tonus is 
felt, the reaction time for the negative response of a posterior 
piece of the body will not differ appreciably from that of a 
normal worm. Now, according to the views of the advocates 
of the theory that after operations involving loss of nervous 
tissue, impulses may be conducted through the general 
protoplasm, it is held that such conduction is always at first 
appreciably slower than in nervous tissue. It would also 
seem on purely a priori grounds that this must be true. 
Thus it is seen that there are serious objections to the view 
that the negative reaction is the result of a contraction on the 
side of the body opposite to that stimulated—that is, that it is 
a crossed reflex. The question now arises, if the reaction is 
not produced in this way, in what other way can it be 
produced? Kvidently it is quite possible that the anterior 
part of the body can be turned away from the stimulus by a 
lengthening of the side stimulated, quite as well as by a 


608 RAYMOND PEARL. 


shortening or contraction of the opposite side. We may now 
consider the evidence as to whether or not the turning away 
is actually due to a lengthening of the side stimulated. 

Very little evidence can be obtained regarding this from 
observation of the normal moving animal, because the 
general appearance in the turning would be the same 
whether it were due to a shortening of one side or a 
lengthening of the other. The results from certain sorts of 
operation, however, give definite evidence on the question. 

A specimen split longitudinally in the posterior end, as 
shown in Fig. 17, a, and the cut was extended forward to the 
posterior border of the head region. Several days were 


A 
B = 


b x 
Fie. 17.—a. Operation diagram. 0b. Showing side A supported on 

B. For further explanation see text. (The pharynx is omitted 

for the sake of clearness.) 
allowed for recovery from the shock of the operation, care 
being taken to prevent the two parts from growing together 
again. By this time the cut edges had healed well, and the 
specimen was in good condition for experimentation. The 
results of mechanical stimulation were as follows: strong 
stimulation of the head or anterior part of the body on either 
side caused the negative reaction; the anterior end turned 
away from the stimulus. But it was possible to tell in this 
case which of the two pieces or halves of the body were 
effective in producing the turning. It could be seen clearly 
that the half stimulated, immediately on stimulation, flattened 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 609 


out slightly ventrally, thus bringing the ventral cilia in close 
contact with the bottom, as is necessary for their effective 
working. At the same time it lengthened along its outer 
side, thus forcing the anterior end around towards the side 
opposite from the stimulus. That the ‘side opposite ” had 
nothing to do with the turning could be observed in many 
cases directly, for this side (B) would remain in an almost 
entirely relaxed condition after the stimulus was given, and 
not get any effective hold on the bottom so that it could 
affect the movement. It was further possible by a little 
manipulation to get the piece B laid over on A so as to be 
practically entirely supported by it, as shownin Fig. 17,6. If 
with such conditions the worm was stimulated rather strongly 
on the A side of the head, it gave a strong negative reaction, 
the point about which the turn was made being as far back 
asa. Evidently with part B up on the dorsal surface of A, 
and consequently having no hold on the bottom, it could 
have no effect in the reaction. The reaction must have been 
due to the side A alone. The same thing could be shown by 
very gently lifting on a needle the side B so that it was not 
in contact with the bottom, and then stimulating A, when 
again the negative reaction occurred. This experiment I 
have repeated with variations many times, but always with 
the same result, showing that the side stimulated is the 
effective one in producing the turning. 

It may be mentioned here that the effect of strongly 
stimulating the posterior end of either of the two pieces of a 
specimen slit in this way was to cause a local contraction of 
the piece stimulated, and a crawling movement of the short 
portion of the body in front of the slit. ‘This crawling was 
not very effective, since so small a portion took part in it, 
but it is of interest to note that what crawling appeared 
involved only the uncut part of the body. 

It being established that the side stimulated produces the 
turning, the question may be raised, how, supposing in these 
longitudinally split individuals that this side does produce 
the reaction, is it known that it does this by lengthening 


610 RAYMOND PEARL. 


along its outer margin rather than by actively contracting on 
its inner cut margin? This question may be answered by 
operative experiments of a different character. If the side 
stimulated, acting independently, produces the reaction by 
lengthening on its own outer side, then an isolated longitu- 
dinal half of the body ought to be able to give only one 
reaction wherever stimulated, or, in other words, it ought 
always to turn towards the same side. Furthermore, such a 
piece onght always to turn towards the cut edge, since only 
on the side opposite to this has it a margin possessing 
the necessary circular muscles for extension (vide sup., 
pp. 556, 557). On the other hand, if the contrary view is 
correct, that the turning away is due to contraction of the 
longitudinal muscles on the side opposite that stimulated, 


Fie. 18.—Showing the appearance of a longitudinal half of a 
planarian when at rest. 


then such an isolated longitudinal half of the body ought to 
be able to turn either way, according to the localisation of 
the stimulus, since there are longitudinal muscle-fibres along 
the cut edge as well as along the other. We may determine 
from experiments which of these two views is correct. 
Unfortunately, it is impossible to get any clear evidence 
on this point from entirely separated longitudinal halves of 
the worm. When a planarian is split in two lengthwise 
each of the pieces immediately becomes strongly contracted 
longitudinally on the cut side, the apparent purpose of this 
reaction being to reduce the exposed surface at once to a 
minimum, After this strong contraction has taken place, 
giving the piece the form shown in Fig. 18, no further 
progressive movement can take place, and the general tonus 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 611 


becomes immediately very much lowered. In view of these 
facts it is impossible to get any very trustworthy results 
from the stimulation of such a piece. 

There is another operation, however, which, while it does 
not isolate completely two longitudinal halves of the body, 
yet does separate into longitudinal halves the essential 
reacting parts, namely, the head regions. This is the 
splitting of the worm in the middle line for a short distance 
back from the anterior end, as shown in Fig. 12. After this 
operation the two anterior pieces move about violently and 
independently for a time, taking all the various positions 
shown in Fig. 19. The animal soon recoyers from the imme- 


Sy ee - 
ma 


Pi) e 
Fie. 19.—Diagram showing the different positions taken by the two 
components resulting from longitudinal splitting of the head. 


diate effects of the operation, glides about in a normal way, 
only at a rather slow rate, and responds well tostimul. The 
anterior piece keeps comparatively straight, there bemg much 
less tendency to contraction on the cut side than when the 
split extends the whole length of the body. The reactions of 
such a specimen to mechanical stimuli are as follows. ‘lo 
stimuli applied at the posterior end along the sides of the 
body the reactions are precisely the same as those already 
described for the normal individual. Stimulation in the 
regions aa (Fig. 20) of moderate or strong intensity produces 
the negative reaction, The organism turns away from the 


612 RAYMOND PEARIL. 


side stimulated quite as promptly and in the same way as 
does a normal specimen. If now the cut edges A and B (Fig. 
211) are stimulated in the same way (a needle may best be 
used for this) the specimen will always turn towards the 
stimulus. This can best be brought out by describing a 
typical case in which a series of fifty stimulations in the 
regions A and B were made ona favourable individual cut in 
this way. In thirty-nine of the reactions the animal turned 
towards the stimulated side. That is, if the stimulus was 
applied at A the animal turned in the direction of the arrow 
a; while if B was the stimulated edge the reaction was in the 
direction of the arrow b. In eight of the remaining eleven 
trials the reaction was indifferent. The animal stopped at 


Fic. 20.—Operation diagram. See text. 


stimulation and then started moving straight ahead again, 
the stimulus evidently having been ineffective so far as 
special reaction is concerned. In only three cases out of 
fifty did the specimen turn away from the stimulus. Since 
it required the greatest care in manipulation to give the 
stimulus to one eut edge without touching the other side, 
especially in view of the fact that the animal was moving all 
the time, it seems very probable that in these three cases a 
stimulus was accidentally given to the side which it was not 
intended to stimulate. The same general result of turning 

' After this operation the two parts of the head usually take the position 


shown in this figure after the first’ spasmodic movements following the opera- 
tion have ceased. 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 613 


towards the stimulus when applied to the cut edge was 
obtained in several other series with this same specimen, and 
many times with other specimens similarly mutilated. — It 
will be seen that this is the result which would be expected 
if the turning away is due to lengthening of the side stimu- 
lated. Stimulation of either side of the cut portions, inner 
or outer, causes turning in the same direction, and that 


Fic. 21.—Diagram to show the reactions to mechanical stimuli and 

their mechanisms in the case of a specimen in which the head 

has been split longitudinally. For further explanation see text. 
direction is the one in which turning would be caused pro- 
vided each piece did actively lengthen on its outer side. 
There seems to be no reason whatever, if the turning away 
were due to contraction of the side opposite that stimulated, 
why the specimen should not turn away from stimuli applied 
to the cut inner edges. This it does not do. ‘There seems 
to be no escape, then, from the conclusion that the turning 


614 RAYMOND PEARL. 


away from the stimulus (negative reaction) is due to a 
lengthening of the side stimulated. 

It may possibly be objected to the last experiment that 
the impulse from a stimulation at, for example, B (Fig. 21) 
took the path indicated by the dotted line in that figure, and 
caused a contraction on the left side of the body, so that 
really the observed turning was the result of a contraction on 
the side opposite that stimulated. To this objection it may 
be answered that by stimulating different points alone the 
edge B it is possible to cause the point about which the 
turn occurs asa pivot to be located anywhere along the linea y. 
It is very evident that contraction of muscles in the region N 
can have nothing whatever to do with turning of the right 
piece about the point x So this objection is without force. 

As the process of regeneration of a cut longitudinal half of 
the body goes on, the piece will straighten out from the 
curved form it takes after the cut is made, and it is conse- 
quently possible to obtain specimens in which the regenera- 
tion of the missing half of the body has produced only a very 
small amount of new tissue, and which are at the same time 
nearly straight in outline and able to make progressive move- 
ments. The reactions of such partially regenerated speci- 
mens are of importance as throwing light on the normal 
mechanism of the reactions. The reactions of a typical 
specimen of this sort may be described in detail. On October 
10th, 1901, a small piece of the anterior end of a specimen of 
P. maculata was isolated. The piece was cut as nearly as 
possible in the form shown in Fig, 22, a. On October 16th 
the piece had the form shown in Fig. 22, >. A narrow strip 
of new tissue had formed down the right side, and the forma- 
tion of the outline of the head and of the right eye was just 
beginning. At this time the reactions of the specimen were 
as follows. Stimuli applied at y caused the head to turn 
sharply away from the stimulus (typical negative reaction). 
This reaction was quite like that given by a normal individual 
stimulated in the same way. Stimulation at x, however, 
produced no trace whatever of a negative reaction, On 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 615 


stimulation at this point the specimen contracted longitudi- 
nally, and then started moving ahead again in exactly the 
same direction in which it was going before stimulation. It 
was impossible to induce any turning away following stimula- 
tion of the side 2, although this was tried many times. 

Now it is evident that this specimen comes very near 
to being an isolated longitudinal half-planarian. All the 
structures of the original one half are present, and there 
is only a very little of the other side of the body produced in 
the line of new tissue, down the originally cut edge. In this 
new tissue there is probably very little differentiation, and the 
muscle layers are not well formed. It was brought out above 
(p. 610) that an isolated half of the body ought to be able to 


Fie. 22.—a. Operation diagram. b. Piece which regenerated from A 
in Diagram a. The new tissue is indicated by stippling. 

give only one reaction, or, in other words, ought to be able to 
turn the body in only one direction in response to stimulation, 
provided this turning is due to an extension of the stimulated 
side. We find precisely this result in the regenerating speci- 
men just discussed. It turns away from stimuli applied at y 
because on that side are present all the muscles necessary for 
extension just as in a normal animal. It does not turn away 
from stimulation of the side v because it has not the necessary 
muscles for extension on that side. On the view that the 
turning away is due to contraction on the side opposite that 
stimulated, there is no reason why stimulation at « should not 
cause the animal to turn away from the stimulus, because the 
opposite side (y) has all its muscular mechanisms intact. 


616 RAYMOND PEARL. 


The reason why the specimen in this last experiment does 
not turn towards the stimulus when stimulated on the side a, 
is apparently because the regeneration has proceeded only far 
enough to produce just enough new tissue to form the 
beginning of a new side to the body. 'l'his new side receives 
the stimulus and is sufficiently potent to determine the re- 
action of the whole (the straight longitudinal contraction), 
but is lacking in the mechanism necessary to produce its own 
proper reaction, the negative reaction. On the other hand, 
in the case of the individual with the split anterior end, each 
piece turns towards the stimulus after stimulation of the cut 
edge because here only one half the organism is present 
either to be stimulated or to react; there is not even the 
beginning of the formation of a new side along the cut edge. 

Putting all the evidence together, I think it must be re- 
garded as demonstrated that the turning away from the 
stimulus in the negative reaction to mechanical stimuli is due 
to an extension of the side of the body stimulated. ‘This 
extension is brought about by the contraction of the circular 
and dorso-ventral muscle-fibres—probably also assisted by 
the transverse and oblique systems of fibres—in the region 
stimulated. This reaction is a simple reflex act involving 
only the side stimulated. The normal organism, so far as this 
response is concerned, is to be considered as composed of two 
identical, but in a certain sense independent longitudinal 
halves. Thus, representing these halves diagrammatically, 
asin Fig. 23, a, the evidence presented indicates that stimula- 
tion of one side of the worm, as A, causes a reaction in that 
side, and, so far as essential features of the directive reactions 
go, only in that side. The movements of half A after its 
stimulation determine and, in fact, cause the reaction of the 
wholeanimal. Furthermore, these longitudinal halves retain 
their individuality as halves if they are isolated from each 
other. A separated half-worm (longitudinal) reacts as a 
half-worm, just as it did when in connection with the other 
half in the body, and not, as might perhaps be expected on 
a priori grounds, as a whole worm. It reacts as a whole 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 617 


worm only after a new half has been regenerated along its 
cut edge. The various stages in the change from the 
reactions as a half-worm to those as a whole worm can be 
followed step by step as regeneration proceeds. he new tissue 
formed along the cut edge very quickly takes on some of the 
functions of a side. When only a narrow strip has been 
formed it serves for the reception of the stimulus, and hence 
stops the reaction of the opposite side, as in the experiment 
last discussed. ‘To make the meaning more clear, reference 
may be made to diagrams b and ¢ of Fig. 23. In b is repre- 
sented, in a straightened position, the half B of anormal worm 


a aads sa 


sae iye seen erere re 1 
ee ae er ee 


b. 
c 


A 

Fig. 23,—Diagrams to show the relations of the halves of the body 

of Planaria to the reception of stimuli, and the reactions 

thereto. See account in text. (The pharynx is omitted for 

the sake of clearness.) 

immediately after being separated from the other half, while 
c represents the same half after regeneration has begun and 
a strip of new tissue has been formed down the cut edge. 
Now stimulation of the cut edge of b causes the anterior end 
of the piece to turn towards the stimulus, 1. e. to give its 
own proper negative reaction (cf. experiment given above on 
shtting anterior end). This is because in this case it is 
side B that is stimulated, although along its inner edge. 
Stimulation along the right-hand edge of ¢ does not cause 
the turning towards the stimulus, because in order that this 


618 RAYMOND PEARL. 


may take place it would be necessary for the side B to give 
its proper negative reaction. It cannot do this because it is 
not directly stimulated, but the new very small side A is 
stimulated. ‘This side may not have the necessary muscles to 
give a negative reaction itself—as in the experiment 
described above,—yet may receive the stimulus and so 
indirectly prevent B from reacting. Another way of ex- 
pressing this same fact is by saying that in regenerating 
longitudinal halves of planarians the physiological middle 
line remains at the line of the former cut edge for some time 
after regeneration has begun.' In connection with this 
discussion of the reactions of half-animals it is greatly to be 
regretted that Willey (97) did not get any data on the 
reactions of the remarkable form Heteroplana. In this 
form we have a natural “ half-planarian,” or very nearly that. 
One side is so greatly atrophied as to be practically absent. 
It seems to me very probable that this organism would react 
to stimuli in much the same way that a longitudinally split 
specimen of Planaria, which had begun to regenerate, 
does. 

I do not wish it to be understood from the analysis of the 
negative reaction which has been given that I intend to 
maintain that in this reaction the side opposite that stimu- 
lated never contracts longitudinally. It probably often does 
this, especially in cases of very strong stimulation which cause 
a general excitation and reaction of the whole body. I have 
merely wished to show that the fundamental basis of the 
negative reaction is the extension of the side stimulated. It 
seems to me quite possible that it may be shown by close 
analysis in other cases that supposedly crossed reflexes are 
not fundamentally such at all. 

We may now pass to a brief consideration of the mechanism 
of the positive reaction of the planarian to mechanical 

1 I have records in my notes of experiments which show that in the case of 
oblique cuts the physiological middle line remains at the cut edge until after 


the new head is well formed in the new tissue on the oblique edge. Lack of 
space forbids detailed description of these experiments here. 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 619 


stimuli. As has been shown above, removal of the anterior 
end of the body containing the brain causes the disappearance 
of this positive reaction, and this result is probably due rather 
to the lowering of tonus than to the removal of any special 
centre having the causation of this reaction as its function. 
Additional evidence on this view that lowering of the tonus 
is the chief cause of the disappearance of the reaction is found 
in the fact that other injuries to the head, such as longitudinal 
splitting, which produce a lowering of the general tonus, also 
cause the disappearance of the positive reaction. 

This very close dependence of the reaction on the general 
tonic conditions of the organism makes its analysis difficult, 
but it seems most probable that its mechanism is as follows :— 
a light stimulus, when the organism is in a certain definite 
tonic condition, sets off a reaction involving (1) an equal 
bilateral contraction of the circular musculature, producing 
the extension of the body; (2) a contraction of the longi- 
tudinal musculature of the side stimulated, producing the 
turning towards the stimulus (this the definitive part of the 
reaction); and (3) contraction of the dorsal longitudinal 
musculature, producing the raising of the anterior end. In 
this reaction the sides do not act independently, but there is 
a delicately balanced and finely co-ordinated reaction of the 
organism as a whole, depending for its existence on an 
entirely normal physiological condition. It is to be noted, 
however, that the definitive part of the reaction, namely, the 
turning 


for 
This point is one of fundamental importance for the general 


is a response of the side of the body stimulated. 


theory of the reactions. 

The mechanism of the other reactions to mechanical 
stimuli are evidently very simple. The crawling movement, 
which must be considered as the specific reaction to mechanical 
stimulation of the posterior region of the body, is due to 
rhythmical contraction of the longitudinal musculature. The 
only other reactions to mechanical stimulation are local con- 
tractions, whose mechanism is evident. 

e. Features in the General Behaviour of the 


620 RAYMOND PEARL. 


Organism which the Reactions to Mechanical 
Stimuli explain.—That much of the behaviour of pla- 
narians in their natural surroundings is the result of the re- 
actions above described is very evident to any one watching 
them. Among specific features of this sort in which these 
reactions play a part may be mentioned the escape from 
enemies or harmful surroundings, the getting of food (to be 
discussed in detail later), the localities chosen for coming to 
rest, the behaviour on meeting solid obstacles in the path of 
movement, the passing on to the surface film, ete. All 
of these need not be discussed specifically, as their relations 
will be evident enough on a moment’s thought, but the last 
two deserve special mention. 

The behaviour of planarians on meeting solid bodies in 
their path in the course of movement is entirely made up of 
reactions to mechanical stimuli. The behaviour in detail is 
as follows :—If a gliding specimen meets squarely head-on an 
obstruction of considerable size, so that it cannot glide over 
it without changing to some extent the position of its long 
axis, it will stop an instant, raise the head, let it drop down 
till it touches the obstruction again, and then glide directly 
up on to and over the solid body. ‘This behaviour is invari- 
able, so far as my observations go, if the worm meets the 
obstruction squarely. It is at once seen to be merely a 
special case of the usual reaction to a weak mechanical 
stimulus, characterised by the raising of the head. ‘The 
behaviour is evidently purposeful in the long run, because it 
will take the organism up on to food material just as well as 
indifferent bodies, If the gliding worm meets the obstruc- 
tion obliquely the behaviour depends in large part on the 
physical nature of the cbject. If it is food material, or some- 
thing else of a rather soft and yielding texture—as, for 
example, another planarian,—the worm will immediately raise 
the head, turn it towards the object, and crawl up over it. 
This behaviour is evidently the typical positive reaction to a 
weak mechanical stimulus. A special and rather curious 
case of this positive reaction, which I have twice observed, 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 621 


appeared when two specimens eliding along, with the anterior 
ends shghtly raised in the normal manner, met head-on, 
Both were simultaneously stimulated to the positive reaction 
and raised the anterior ends, and then let them drop again. 
As they came down the two ventral surfaces were brought 
squarely together in the way shown in Fig. 24; then each 
started gliding up the ventral surface of the other. In a 
movement as a result of the constantly changing form of the 
body, the ventral surfaces slipped off from one another and 
the two worms went on their way. When the obstruction is 
a hard body, asa piece of glass, the specimen meeting it 
obliquely usually turns the head away slightly at the first 
contact (negative reaction), and then glides along parallel to 
the edge of the body fora distance. If it happens to touch 
if again with the side of the head, it frequently gives the 
negative reaction and turns away again. After the solid body 


Fie, 24.—Side view of two planarians starting to glide up on the 
ventral surfaces of each other. 


has been touched several times, however, the positive reaction 
is usually given, and the worm passes at once up on to the 
solid body. This behaviour is shown in Fig. 25. The precise 
form of the behaviour on meeting obliquely a solid body in 
the path varies considerably with the general physiological 
condition of the individual. In case it is much excited, the 
first touch will induce a strong negative reaction, and the 
individual will turn away and pass out of the neighbourhood. 
In the cases where the final positive reaction is preceded by 
two or three negative ones, it would seem as if repetition of 
what must be an almost identical stimuius causes it to be- 
come in effect weaker. Leaving aside all variations in the 
exact character of the behaviour on meeting a solid, the 
important point to be brought out is that all this behaviour is 
based on the simple reactions to mechanical stimuli. The 
VOL. 46, PART 4,—NEW SERIES. Ss 


622 RAYMOND PEARL. 


exact behaviour in any given case depends on _ several 
different factors. These are the position of the animal with 
reference to the obstruction, the physical nature of the 
obstruction, and the physiological condition, whether of 
ereater or less excitation. 

So, again, with reference to the habit of the animal of 
moving about on the surface film, a problem is presented. 
When a specimen, gliding up the side of a dish, touches its 
anterior end to the surface film at the point where the latter 
joins the glass, it immediately gives a characteristic positive 
reaction, precisely like that in response to any other weak 
mechanical stimulus. The head is raised and turned towards 
the side from which the stimulus came, and then dropped 


5 6 


Fic. 25.—1, 2, 3, 4, 5, and 6 are successive stages in the reactions of 
Planaria on meeting obliquely an obstacle in its path. The heavy 
straight line represents the obstacle. 


again. As a consequence of this reaction, the head end 
comes to rest on the under side of the surface film at a point 
some little distance out from the side of the dish. The 
ventral surface of the anterior end of the body flattens out on 
the surface film, and the animal glides out on to the film, 
following the direction determined by the reaction of the ante- 
riorend. Thus itis seen that the going on to the surface film is 
only aspecial case of a response to a weak mechanical stimulus, 
i.e. the positive reaction, the film itself acting as the stimulant. 
The leaving of the surface film and passing down the side of 
the dish is evidently also due to the same positive reaction. 


ra 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 623 


There are a number of other points in the general behaviour 
which are directly related to the reactions to mechanical 
stimuli, which will be taken up later in connection with the 
other reactions. 

jf. Summary.—Before passing on to a discussion of the 
next subject, it may be well to summarise briefly the chief 
findings with reference to the effect of mechanical stimuli on 
planarians. 

It has been shown that the planarian responds in a well- 
nigh perfect manner to the localisation and intensity of 
mechanical stimuli. It turns away from strong stimuli (in 
the long run harmful) applied to the side of the body ; turns 
towards weak stimuli (in the long run beneficial, almost never 
harmful) ; it crawls rapidly away from strong stimuli applied 
to the posterior end; backs and turns away from similar 
strong stimuli applied at the anterior end. 

It has been shown, further, that these reactions have all 
the characteristics of reflex actions, complex, it is true, but 
still reflexes. 

The mechanisms of the reactions to unilateral stimulation 
are unilateral, and lie in the side stimulated. 

Discussion of the implications of these results on 
mechanical stimulation, with reference to the psychology of 
the organism and the general theories regarding the reactions 
of organisms to stimuli, is deferred till the results from other 
sorts of stimuli are in hand. 


Il. Reactions to Food and Chemical Stimuli. 


Evidently one of the most important factors in the sum 
total of the activities of any aquatic organism is its reactions 
to chemical substances. Its ability to receive chemical 
stimuli and react to them must be of prime importance in its 
struggle for existence, for in its natural habitat such an 
aquatic organism must be almost constantly encountering 
different chemical substances. Some of these may be harm- 
ful and some beneficial, and it would seem that if a species is 


624 RAYMOND PEARL. 


to survive, its individuals must have some sort of reaction 
whereby they may avoid the harmful and take advantage of 
the beneficial. In the case of planarians, the reactions to 
chemicals seem to be of about equal importance with the re- 
actions to contact stimuli in the general activities. Since the 
reactions to food substances are a special case of the reactions 
to chemicals in general, they may be discussed first. 

a. Food Reactions.—The nature of the things used as 
food by fresh-water planarians has been discussed already in 
the section on ‘‘ Natural History,’ 
us here, 

A typical case of the food reactions to a bit of crushed 


> and hence need not detain 


Fie. 26.—Diagram showing the successive stages in the normal food 
reaction of Planaria. A represents a small bit of meat. 
mollusc may first be described, to serve as a basis for the 
account.!. Ifa piece of the body of Physa which has just 
been extracted from the shell and crushed between the points 
of a pair of forceps is placed in a small dish containing a 
number of active planarians, it will result from chance alone 
that some of the flat-worms must in course of time pass near 
the food material. For a very short time after the food has 


1 The food reactions of Planaria have been briefly described by Bardeen 
(:01, a). 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 625 


been placed in the dish specimens may pass very near it— 
within two or three millimetres—without being affected in 
any way. They simply glide straight by as if there were no 
food there. After a few minutes have passed, however, it 
will be found that a worm coming near the food is affected 
in a very characteristic manner. Its behaviour is as 
follows: —When within about three or four millimetres of 
the piece of meat (Fig. 26, a) it stops abruptly, raises the 
head, and turns it towards the food (Fig. 26, b). As the head 
is raised and turned the gliding is resumed, and the head 
being almost immediately lowered, the movement is directly 
towards the food. ‘Thus far the reaction is evidently 
precisely like the positive reaction to weak mechanical 
stimuli, and so we may speak of it as the positive reaction to 
food, the reaction being the same in the two cases, though 
the stimulus differs. When the anterior end of the head 


eee eA 


Fie. 27.—Diagrammatic side view of Planaria to show the 
*‘oripping” of a bit of food, A. 


touches the food it flattens down upon it, and, if the con- 
figuration is such as to make it possible, “grips” ib 
(Fig. 26,c). The details of this “ gripping ” (shown in side 
view in Fig. 27) are as follows:—The anterior end closes 
down over the very edge of the piece of food, or over the 
whole piece provided it is small enough, and then 
apparently squeezes it by contraction of the longitudinal 
muscles on the ventral surface of the head. The action is 
very characteristic, and evidently forms an integral part of 
the normal food reaction. Its probable function will be 
brought out later. While it is taking place the worm as a 
whole stops its progressive movement and remains quiet. 
After the “ gripping” has continued for some time the worm 
starts gliding ahead up on to the food. It passes forward 
till the point where the opening for the extrusion of the 
pharynx is located is approximately over the place pre- 


626 RAYMOND PEARL. 


viously “gripped” (Fig. 26,d). Then the pharynx is ex- 
truded and feeding begins (Fig. 26,¢). After a time the 
worm voluntarily leaves the food and glides off over the 
bottom. 

Having described the typical case of a food reaction, we 
may take up some of the more important variations from the 
type, and describe the various phases in the reaction in 
vreater detail. 

Starting with the very beginning of the reaction, it may 
be said that the distance from the food at which any effect 
on the planarian is produced varies greatly, as is to be 
expected. This distance, of course, depends on the extent 
which the juices or chemicals of the food have diffused from 
it. When a piece of meat is first put into the water 
specimens will pass very close to it without being stimulated. 
In fact, if a specimen finds a piece of food within three or 
four minutes after it is put into the dish, it wil usually have 
done so as a result of accidentally coming in contact with it. 
As has been brought out above, when a gliding worm 
touches anything of a rather yielding texture, like food, it 
immediately gives the positive reaction and passes up over 
it. This plays an important part in the getting of food, 
because, as I have found in experiments, unless the food is 
crushed and pressed with forceps the juices diffuse rather 
slowly, and for some time specimens will not give the 
positive reaction unless they actually touch the food. On 
the other hand, after the food has been in the water for 
some time, so that diffusion has taken place, the distance at 
which specimens may be affected becomes quite considerable. 
I have seen specimens gliding by a small piece of meat at 
a distance of 1} cm. from it give the positive reaction and 
turn towards it. At greater distances than this food is not 
effective, according to my observations. ‘The distance from 
food at which a given specimen will give the positive 
reaction and go towards it depends also on the physiological 
condition of the individual. Specimens in a state of general 
excitation will, as I have frequently observed, go closely by 


MOVEMEN'S, ETC., OF FRESH-WATER PLANARIANS. 627 


a piece of food without turning towards it, while other 
Specimens in a more normal condition will give the positive 
reaction some distance from it. 

After the first specimen has begun feeding on a piece of 
material the zone of influence of that piece becomes almost 
immediately widened appreciably. As the number of feed- 
ing specimens increases the area in the surrounding water 
which affects others becomes correspondingly greater. ‘This 
phenomenon is very striking in many cases, as an illustration 
will indicate. Several pieces of crushed snail were put in a 
dish with a number of planarians. In a short time a 
specimen in gliding about the dish had come near to one 
of these pieces, had given the positive reaction and begun 
feeding. At almost the same time another of the pieces of 
food had “attracted ”’ another specimen. The other bits of 
food were quite similar in every way to these two, and lay in 
the dish not far from them. Yet at the end of fifteen 
minutes the two pieces by which the first two worms had 
been affected were completely covered with feeding 
specimens, while the remaining pieces of food, with a single 
exception,! did not have a specimen on them. ‘This increase 
in the effectiveness of the food as a stimulus must be due 
to the diffusion of more chemical substance from it. 
Apparently the increase is due either to some secretion of 
the feeding animals or to some change which they induce in 
the food. It is probably due to a combination of these two 
factors. That a digestive secretion is poured out through 
the pharynx of the feeding worm is well known, and clearly 
shown by the appearance of a piece of food on which a 
specimen has been feeding. ‘The surface of the meat is 
turned white, and rendered very soft and almost flocculent. 
It is probable that this digestive secretion acts as a positive 
chemotactic stimulus to other worms, and that coupled with 
this there is an increased diffusion of juices from the food 
itself caused by the changes which it is undergoing. 

The reaction which is caused by this chemical stimulus 


’ One piece farthest removed from the others hada single specimen on it. 


628 RAYMOND PEARL 


from the food is evidently essentially the same thing as the 
positive reaction given to weak mechanical stimuli. It con- 
sists in a turning of the anterior end of the body towards 
the source of the stimulus. There is no reason for supposing 
that its mechanism is in any way different from that of the 
same reaction to mechanical stimuli, and hence this need not 
be further discussed here. A question of prime importance 
with regard to this positive reaction in response to chemical 
stimuli, which was not taken up before, is— how well localised, 
with reference to the stimulus, is the reaction ? or, in other 
words, how precisely does the anterior end point towards the 
source of the stimulus,—in this case food? Have we here a 
clear-cut orienting response? In answer to this problem it 
may be said that when the worm is only a short distance 
from the food the response is very precise. The anterior 
end is brought by the first positive reaction so as to point 
exactly towards the meat, and as the worm glides ahead it 
never misses it. This is true where the specimen is near 
enough (usually within three quarters of its own length), so 
that the stimulus which reaches it is a fairly strong one. 
In case the worm is stimulated near the edge of a large 
diffusion area when the stimulus is very weak, the first 
reaction may not suffice to direct the animal straight towards 
the food. In this case the behaviour is usually like that 
shown in Fig. 28, in which the line B, B, B, represents the 
effective margin of the diffusion area of the piece of food A. 
(By “ effective margin” is meant the line outside of which no 
effect is produced by the food on passing specimens.) The 
first reaction which the worm gives on reaching this diffu- 
sion area (Fig. 28, 1 and 2) is a weak positive one. It then 
proceeds on the new path into this area, but not directly 
towards the food. After a short time, however (Fig. 28, 3), 
it is again stimulated to a positive reaction (4). ‘This time 
both the stimulus and the reaction are stronger than before, 
and the worm is directed more nearly towards the centre of 
diffusion, but still not exactly. When it gets opposite the 
food again (5) another positive reaction (6) is given, and this 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 629 


time, since the stimulus is a rather strong one, the reaction 
is a very precise one, and the subsequent movement carries 
the animal directly to the food (Fig. 28, 7). This behaviour 
is typical for this sort of stimulation, but may vary in its 
component phases, depending on the relative strength of the 
stimulus—the distance from the food when first stimulated. 


et ee 
~_ 
~ 


_- 
- ~ 


~ 
3S 


Fig. 28.—Diagram showing the reactions of Planaria to food (A) 
from which juices have been diffusing into the water for some time. 
B, B, B, represent the effective margin of the diffusion area of the 
food A. 1, 2, 3, 4,5, 6, and 7 are successive positions taken by 
the organism. 


Thus either two or as many as four positive reactions may be 
necessary to bring the animal to the food. This shows 
clearly that with reference to chemical stimuli, the precision 
of localisation of the positive reaction decreases as the in- 
tensity of the stimulus diminishes. Indeed, I have observed 


what is evidently an unlocalised positive reaction, although 


RAYMOND PEARL. 


this seems paradoxical. The behaviour was as follows :—A 
large diffusion avea had been formed, and a specimen was 
stimulated to a weak positive reaction at a distance of about 
twice its own length from the food (Fig. 29, 1). It passed 
into the diffusion area, but did not give another positive reac- 


~ 
~ 
> 
BS 
“ 
NS 
an 
\ 
N 
\ 
ww 
‘\ 
. 
s 
N 
.y 
‘ 
\ 
\ 
\ 
\ 
\ 
\ 
\ 
‘ 
‘ 
\ 
i) 
' 
\ 
' 
‘ 
' 
{ 
U 
' 
‘ 
} 
‘ 
/ 
‘ 
/ 
‘ 
U 
/ 
/ 
4 
/ 
4 
/ 
4 
Pid 
7 B 
a 
(4 
oe 
- 
aon 


- =- 
wWeme eae seo” 


Fic. 29.—Showing the reaction of a planarian to a very weak food 


stimulus. Letters as in Fig, 28. 
tion when opposite the food, but instead glided by and away 
from it. When it had gone some distance in this direction 
it stopped and gave a very clear and characteristic positive 
reaction, so far as the form of the reaction indicated, but 
with the turn away from instead of towards the centre of 


MOVEMENTS, E'TC., OF FRESH-WATER PLANARIANS. 631 


diffusion. ‘There was no doubt of the character of the 
reaction; the head was raised and the body turned in the 
usual manner of the positive reaction, which one can never 
mistake after once having become familiar with it. The 
specimen kept on in the path determined by this last 
reaction (Fig. 29, 4), and passed entirely out of the region of 
the food. Evidently in this the worm was stimulated very 
weakly by a chemical, and the stimulus was nearly as strong 
on one side of the body as on the other, and when the reflex 
was set off it was on the wrong side of the body. This is 
not the usual result of weak stimulation, and has been 
observed in only two cases, but it serves very well to show 
the decrease of the power of localisation when the stimulus 
is very weak. 

When, as frequently happens, the worm approaches the 
food exactly head-on, the reaction usually consists merely 
of that portion of the reflex expressed in the raising of the | 
head, while the worm keeps on in its straight path till it reaches 
the food. The head may be waved from side to side slightly, 
but the general direction of motion is not changed. The 
action evidently corresponds to the positive reaction following 
weak mechanical stimulation of the dorsal surface of the head 
in the middle line, as described above. In some cases, how- 
ever, I have observed very active and hungry specimens of 
Dendrocelum, sp., which were going straight towards the 
food, give a complete positive reaction and turn to one side 
and start off in a new direction away from the food. This, 
however, of course brought the specimen at once into a 
position where the stimulus was acting unilaterally, and it 
again gave a positive reaction, this time heading it again 
for the food, which it usually reached without further 
reaction. But in some cases I have observed the specimen 
give so strong a reaction as to be taken almost directly 
away from the food by the subsequent movement, and, 
passing out of the area of diffusion, fail to reach it at all. 
Specimens behaving in this way were “‘wild” in their 


general reactions. ‘I'he responses were very vigorous, but 


632 RAYMOND PEARL. 


not localised with reference to the stimulus with the usual 
precision. 

The “ gripping” of the food substance by the anterior part 
of the worm is a very characteristic feature of the normal 
food reaction. Its exact form depends on the configuration 
of the food or other body “ gripped.” In its most typical 
form, where the food material 1s in the form of a cylinder, or 
approximately such, the action reminds one of the action of 
the human hand in grasping a stick. The tip of the head 
closes over the material in the same way that the fingers do, 
while the region just behind the auricles bears the same 
relation as does the proximal part of the palm, just in front 
of the wrist, in grasping. After the head has been placed 
over the material in this way it can be seen to contract 
rather strongly, and thus literally squeeze the food. In case 
the surface contour of the food does not admit of this reflex 
being carried out in its typical form, as close an approxima- 
tion to this is made as possible. ‘To compare again with the 
human hand, when the surface is flat, or forms the surface of 
a cylinder of large radius, the ventral surface of the head is 
pressed closely to it, the tip attempting to dip in, as it were, 
below the surface, in just the same way that a man “ claws” 
with his finger tips in attempting to obtaim a hold on a 
similarly configured body, too large for complete grasping. 

While the “ gripping” is in general a very characteristic 
feature of the food reaction, it may be omitted in rather 
exceptional cases. he cause for the omission where it 
occurs, or any laws governing the matter, I have not been 
able to discover. A necessary accompaniment of the 
> of the food is the cessation of the forward 
movement of the animal as a whole. This pause when the 


“ eripping’ 


food is first touched by the anterior end and before the worm 
passes up on to it, occurs in practically every case, whether 
the gripping accompanies it or not. The length of the pause 
is, of course, considerably greater when the “ gripping ” 
occurs than when it is absent. 

The function of the “ gripping” of the food material before 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 6383 


feeding begins is not immediately apparent, but I am inclined 
to think its purpose is to intimately test the substance with 
regard to its availability as food. Some evidence on this 
point and further discussion regarding it will be introduced 
later. 

After the preliminary pause and “ gripping” of the food 
the worm glides up on to it to begin active feeding. The 
position taken by the worm brings out a very nice correlation 
in reflexes. In a very large number of cases (certainly over 
75 per cent., so far as my observations haye gone) the worm 
advances over the food until the pharyngeal opening is 
exactly over the place where the first “ gripping” occurred, 
and there the pharynx is extruded and feeding begins. 


Fic. 30.—Diagram showing great extension of the pharynx. The 
stippled area represents food substance on which the planarian 
1s resting. 


When the worm reaches this position the posterior part of 
the body relaxes and takes on the appearance character- 
istic of the resting specimen. The pharynx is thrust out, 
and becomes attached very quickly. As it passes out through 
the opening in the body-wall it becomes usually considerably 
extended, and its diameter becomes correspondingly smaller 
than when it is in the pharyngeal sac. It may or may not 
attach to the food directly beneath the body. When con- 
ditions are favourable it usually does, and consequently 
canuot be seen on looking down on the animal from above. 
On the other hand, I have frequently seen it stretched out 
and attached some little distance to one side of the body, as 
shown in Fig. 30. The stimulus, causing the extrusion of the 


634. RAYMOND PEARL. 


pharynx, is the contact of food or other solid body with the 
pharyngeal region of the ventral surface, together with an 
appropriate chemical stimulus. The pharynx is not extruded 
until the animal gets up on to the food so that the opening of 
the pharyngeal sac is in direct contact with it. This can be 
demonstrated by direct observation by the use of a very 
small piece of food material and a plane mirror placed 
beneath the glass dish in which the specimen is moving. By 
lifting gently the posterior end of the body on a needle it can 
also be seen that the pharynx is not extruded before it is over 
the food. The most striking illustration of the correlation in 
the reaction which brings about the extrusion of the pharynx 
when it is just over the food, is to be seen when a specimen 
of the nemertean Stichostemma asensoriatum is used as 
food, and the long axis of the planarian and of the nemertean 
are at right angles to each other. After first “ gripping” 


—@ ee, 
a x b 
Fig. 31.—Diagrammatic longitudinal section of a planarian feeding 
on a nemertean (shown in cross-section at x ). 
the nemertean the planarian glides along over it until the 
pharyngeal opening is just above it, and then pauses, and the 
pharynx is extruded and attached (a and b, Fig. 31). ‘hese 
facts strongly indicate that the effective stimulus for pharyn- 
geal extrusion is received, at least in part, in the pharyn- 
geal region itself. That it is necessary for both contact 
and chemical stimuli to act to produce the extrusion of 
the pharynx may be shown by experiments on specimens 
gliding on the surface film ventral side uppermost. If, with 
such a specimen, a chemical known to produce under other 
conditions extrusion of the pharynx, is allowed to come in 
contact with the pharyngeal region, there is no result. Of 
course in performing this experiment proper precautions 
were taken not to disturb the animal by allowing the solution 
todrop uponit. Another demonstration of the same fact that 
a chemical stimulus alone does not suffice to cause extrusion 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 635 


of the pharynx is that specimens immersed in favourable 
solutions, such as sugar solutions, do not show this phe- 
nomenon. ‘That mechanical stimulation alone does not suffice 
is demonstrated by the fact that planarians pass over and rest 
on other planarians without extruding the pharynx, although 
the consistency of their bodies is evidently much the same as 
that of the animals used as food. In fact, they will be 
used as food frequently if they are wounded so as to afford 
the proper chemical stimulus. ‘The stimulation of the anterior 
end of the body by the food seems also to be necessary before 
pharyngeal extrusion takes place. The data on this point will 
be presented later in connection with operation experiments. 

The appearance of the body on the food is quite charac- 
teristic. As mentioned above, when the pharynx is extruded 
forward, movement stops, and the posterior part of the body 
becomes more or less relaxed. The anterior third of the 
body, however, keeps in movement during a considerable 
part of the time the specimen is feeding. The head is waved 
about from side to side, and touched to the food or the 
bottom of the dish here and there. It keeps its character- 
istic extended form to a greater or less degree. A favourite 
position is for the anterior third or half of the body to lie on 
the bottom and move about, while the posterior part lies up 
on the food. This is the position most frequently seen in 
specimens feeding on a rather small piece of meat. When 
the anterior end gets on the bottom it gives every appear- 
ance, in many cases, of attempting to glide away, and being 
only restrained by the attachment of the pharynx to the 
food. In other cases, however, the anterior end remains 
quiet. The importance of the attempted movement will be 
brought out later. As has been mentioned above, the flat- 
worm is able to move off and drag the food still attached to 
the extruded pharynx along behind it. In the fastening of 
the food to the body in this case the sticky slime undoubtedly 
assists the pharynx. 

After the food has been softened by the digestive juices, 
it is taken into the body through the pharynx. 


636 RAYMOND PEARL. 


After the worm has been feeding for a certain length of 
time it will detach the pharynx and spontaneously move off 
from the food, the pharynx being withdrawn again into its 
sac. The length of time after the beginning of the feeding 
at which this takes place varies very greatly in different 
cases. I have observed a specimen which fed on a piece of 
molluse for as long as an hour and thirty minutes, while in 
other cases the worm may stay on the food only ten minutes, 
or even less. Judging from the rate at which food is taken 
up while the animal is feeding during the day, and from the 
fact that pieces of meat left in the dish overnight are almost 
entirely consumed by morning, it would appear that much of 
the time during the mght is spent in feeding when any 
material available for the purpose is at hand. While the 
anterior end of the feeding worm retains its normal sensi- 
tiveness to stimuh, it nevertheless requires considerable 
stimulation to induce a feeding worm to leave the food. 
Shaking of the dish, which would ordinarily set all resting 
specimens into rapid movement, has little or no effect on 
feeding specimens. If a worm is suddenly pulled off a piece 
of meat on which it is feeding a very good view of the 
extruded pharynx may usually be had, as this organ is 
retracted somewhat slowly when torn from food in this way. 

So far as I have been able to discover, the presence of 
food in the immediate neighbourhood of a resting planarian 
has no effect upon it. Apparently the stimulus afforded by 
crushed meat is not sufficiently strong to produce a response 
from such an individual. ‘The following experiment copied 
from my notes will show this. 

May 14th, 1901, 3.10 p.m.—A piece of freshly crushed 
snail was placed 1 mm. distant from the anterior end of a 
resting specimen. No reaction or other effect produced. 

3.30 p.m.—Worm in same position as before. 

4.5 p.m.—No change. (At this time the worm was acci- 
dentally started into movement and the experiment conse- 
quently ended.) 

This lack of effect of food on resting specimens may be 


- 


MOVEMENTS, E'C., OF FRESH-WATER PLANARIANS. 637 


the reason for the statement of Bardeen (loc. cit., p 522) “ that 
worms which had been kept in pure rain water for a week or 
two, and were thus in a hungry condition, would remain 
unmoved by the presence close by their side of a piece of 
fresh snail, a food much prized by them.” 

1. Food Reactions of Specimens after Opera- 
tions.—For the purpose of throwing light on the general 
mechanism of the food reaction, experiments were tried on 
specimens cut in different ways. It is unfortunately very 
different from practical reasons to get many certain results 
from these experiments. Many of the results are negative, 
and hence not entirely conclusive. Since, however, some 
important facts have been brought out by these experiments, 
they will be described. 

The first operation which will be discussed is that of 
cutting the animal in two transversely. If such a cut is 
made in the region in front of the pharynx, the anterior 
resulting piece, after it has recovered somewhat from the 
shock effect of the operation, will show the following reac- 
tion. On coming into the zone of diffusion about a piece of 
meat it gives the positive reaction just as a normal worm 
does, and turns towards the food. On reaching the edge of 
the meat its behaviour is again like that of the normal 
animal; it stops, usually “grips” the food, and then passes 
on over it. At this point appears the striking difference 
between the behaviour of this anterior piece, which, it must 
be remembered, has no pharynx, and the behaviour of the 
entire worm. The anterior piece after gripping the food 
glides up over it, and without the slightest change, even in 
the rate of gliding, passes down off of it on the other side. 
There is not the slightest indication of any stopping for the 
pharynx to be extruded. 

If the transverse cut is made farther back, so that the 
pharynx is included in the anterior piece, this will then 
behave with reference to food quite as a normal animal does. 
It will stop on the food and extrude the pharynx. 

The posterior pieces resulting from transverse cuts do not 

VOL. 46, PART 4,—NEW SERIES, gy 


388 RAYMOND PEARL. 


give any definite food reaction, so far as I have been able to 
ascertain, until they have been regenerated to some con- 
siderable extent. Posterior pieces from which only the head 
has been cut will glide by pieces of snail on which other 
worms are feeding, without giving the slightest reaction.! 
In experiments so arranged that the gliding posterior piece 
would just touch with its anterior end the edge of a piece of 
food, it gave no reaction. This same arrangement with a 
normal worm practically never fails to call forth the positive 
reaction and bring the worm up on to the food. Posterior 
pieces placed gently on pieces of food material do not extrude 
the pharynx and start feeding, but immediately e@lide down 
froin it and over the bottom of the dish. These experiments 
with posterior pieces have been tried many times and under 
varied conditions, in the hope that some sort of positive 
results might be obtained, but never with success. This is 
true for three days after the operation. After a new head 
has been fairly well formed the animal will react to food 
again. ‘The behaviour of one of these posterior pieces on 
touching with the anterior end a piece of food is very 
strikingly different from that of a normal animal. The cut 
piece, if it touches with the front or sides of the anterior end 
the smallest shred of food material, or any other substance, 
gives a well-marked negative reaction, and goes in a new 
direction away from the obstruction. It does not, as a rule, 
crawl up over anything which it meets squarely “ head-on,” 
but instead turns away. 

Thinking that possibly the pharynx might play a more or 
less independent part in the normal food reaction, i. e., that 
it might have a set of reflexes of its own, not determined by 
the rest of the body, I tried experiments with the isolated 
pharynx removed entire from the body. Such an isolated 
pharynx will remain alive for a considerable period, and 
respond to stimulation. When first removed from the body 

' Bardeen (:01, a) has shown that if the transverse cut is in the region in 


front of the eyes the posterior piece (comprising in this case nearly the whole 
worm) will react normally to food, 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 639 


it contracts rhythmically in a longitudinal direction for a 
time, and then comes to rest at about its normal leneth when 
in the body. Mechanical stimulation causes merely longi- 
tudinal contraction, while the presence of food near it has no 
effect whatever. Freshly crushed snail meat placed within a 
millimetre of such an isolated pharynx had no effect upon it 
in the course of an hour. I have tried laying the isolated 
pharynx directly on pieces of meat to see if there would be 
any tendency for the end of the organ to attach itself as it 
normally does. This was not done, nor was any other 
definite reaction produced. 

These operation experiments show, so far as they go, that— 

(1) The presence of the pharynx in the body (i.e., the 
functional] ability to take food) has nothing to do with deter- 
mining the reaction of the anterior end of the body to food 
stimuli. The anterior part of the body gives the same re- 
action to food in every case, without regard to whether so 
doing actually puts the animal ina position to get food or 
not. The reaction is only purposive under certain circum- 
stances; when changed conditions make it no longer purpo- 
sive, no adaptive change in the behaviour of the anterior end 
occurs. This shows clearly how little basis there is for con- 
sidering the behaviour towards food as anything of the 
nature of intelligent behaviour, 

(2) The stopping of the worm on the food under normal 
circumstances is due to the posterior half of the body, not the 
anterior. The behaviour of the anterior cut piece in gliding 
directly over the food is what one might be led to expect from 
the behaviour of the same part of the body under normal 
circumstances. As described aboye, it was seen that the 
anterior end of the normal individual gives every appearance 
of attempting to continue moving forward while the posterior 
part is feeding, and is only prevented from doing this by the 
mechanical hindrance of the attached pharynx. In a sense, 
we may consider that in a large degree the work of the 
anterior end of the body with reference to feeding is over 
when it gets the animal up on to the food. 


640 RAYMOND PFARI. 


(3) The reception of the food stimulus is a function of the 
head. In other words, the head is the only part of the body 
capable of receiving very weak chemical stimuli. 

(4) Decapitated specimens do not extrude the pharynx, so 
far as my observations go, even though the proper normal 
stimuli are given the pharyngeal region. Presumably the 
brain is the necessary organ in this connection, as we have 
already seen that the sense organs concerned with the act of ex- 
trusion are not those of the head, but of the pharyngeal region. 

Bardeen (: 01, a, p. 178) states that ‘the simple reflexes of 
extending the pharynx and of swallowing are preserved after 
removal of the head. I found, by repeated trials, that one of 
the headless pieces could usually be made to eat if it was 
placed on its back on a shide in a small drop of water. Under 
the conditions mentioned the pharynx is usually protruded, 
and will engulf bits of food placed in the mouth.” Regard- 
ing this conclusion, I can only say that in a large number of 
experiments with decapitated specimens I have never been 
able to induce extrusion of the pharynx, under conditions 
approximating as closely as possible to the normal. I do not 
wish to affirm that the decapitated planarian cannot extrude 
the pharynx, but merely that it does not when placed in 
situations which normally produce pharynx extrusion. 

(5) The pharynx is not an independent organ in its reactions, 
since, when separated from the body, it does not react with 
reference to the localisation of the stimulus, as it does when 
normally connected with the remainder of the body.! 

2, Summary of Food Reactions.—It is shown above 
that planarians have a very definite and characteristic set of 
reactions to food substances which enable them to become 
aware of the presence of food, and find it. The importance . 
of these reactions in the life of the individual can hardly be 
over-estimated. While planarians, like many other lower 
organisms, can live for a considerable time without food, yet 
in the long run they must, of course, have it. The question 


1 Evidence on this latter point will be brought forward in connection with 
the reaction to chemicals, 


MOVEMEN''S, ETC., OF FRESH-WATER PLANARIANS. 641 


of how a lower organism gets its food, taking advantage of 
the good and rejecting the bad, and thus apparently choosing 
one thing from several, is one of the most interesting and 
important in comparative psychology. 

The food reaction of planarians consists of an extremely 
well co-ordinated set of reflexes, which may be set into action 
by stimuli of two sorts,—first, chemical; and second, 
mechanical. Both sorts of stimuli are, of course, given by 
the food. he first and most important of all the reflexes in 
the food reaction is the turning of the head towards the 
source of stimulation, followed by movement in that direction. 
This is the reaction which enables the animal to find food. 
Evidently it is the same thing exactly as what has been 
described as the positive reaction to mechanical stimuli; or, 
in other words, the positive reaction to mechanical stimuli is 
only a special case of the general food reaction. Its primary 
function is evidently the getting of food, whatever the stimulus 
which calls it forth. The reason for a food response following 
mechanical stimulation is to be found in the fact that it most 
frequently happens that many things (e. g., whole animals) 
which are available for food are not emitting chemical sub- 
stances into the water in sufficient quantity to cause an 
effective stimulus. If the planarian did not give a positive 
reaction after contact with such bodies they would be missed, 
and no advantage taken of them as food. By reacting 
positively to weak mechanical stimuli the animal is in a 
position to take advantage of the presence of food of all sorts, 
whether it is in condition such as to diffuse chemical sub- 
stances through the water or not. This fact that the animals 
react to food substances as a result of mechanical stimulation 
affords a possible explanation of the “ gripping” phase of the 
general response. ‘The purpose of this ‘ gripping” may be 
to bring the sense organs of the head, which are sensitive to 
chemical stimuli, into very close contact with the substance 
in order to determine whether it possesses the chemical 
characteristics of food. In other words, this reaction is a 
2 


“tasting ” reaction, which is made necessary by the fact that 


642 RAYMOND PEARL. 


the organism turns toward all bodies of a certain physical 
texture under most circumstances. The active squeezing of 
the material in the “gripping” undoubtedly helps to press 
out to the surface any juices which may be in the material. 

In closing the section on food reactions it may be well to 
give a sort of general picture of the whole behaviour of 
fresh-water planarians towards food. The method by which 
the planarian finds material suitable for food is as follows : 

1. Chemical substances diffusing from food come in con- 
tact with the sensitive head region of the planarian ; or— 

‘The moving animal touches with the head some soft sub- 
stance, and as a result of either of these two sorts of 
stimulation— 

2. The organism gives a positive reaction, i.e. turns 
towards the source of the stimulus. This reaction is very 
precisely localised in most cases, and is the most essential 
part of the whole food behaviour. Its mechanism has been 
previously described (v. sup., p. 619). 

3. When the anterior end squarely touches the food as a 
result of this reaction it typically closes tightly over it, 
giving what I have called the “gripping” reaction. This 
reaction is evidently a very much specialised feeling move- 
ment for the purpose of closely testing the chemical nature 
of material. It is produced by a contraction of the ventral 
longitudinal muscles of the head region, While it is taking 
place progressive motion ceases. 

4. After this pause the worm glides over the piece of food 
till the opening of the pharyngeal sac lies over or nearly 
over the place “ gripped,” and there the posterior part stops 
and the pharynx is extruded and attached to the food. 'The 
factors determining the place where the pharynx shall be 
extruded are (a) the stimulation of the ventral surface of 
the body in the pharyngeal region of the food (pure reflex 
factor), and (b) the presence of the brain, which probably 
acts as a co-ordinating centre for this reaction, 

5. A digestive fluid is poured out through the pharynx, 
and the food is partly digested before being taken up. 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 645 


6. The softened food is taken into the body through the 
pharynx. 

7. The animal spontaneously stops feeding after a certain 
time. 

The question now arises, if the normal process of getting 
food is at bottom in the majority of cases a reaction to a 
chemical stimulus, what is the nature of the chemical sub- 
stance causing it? Can the same response be induced by 
the use of different inorganic and organic chemicals? Is 
there any relation between chemical composition and the 
intensity or form of the reaction? ‘To answer these and a 
number of other questions arising out of them recourse must 
be had to experiments in which the nature and concentra- 
tion of the chemicals affecting the organisms may be con- 
trolled. All the experiments of this kind I will group 
together under the heading— 


b. Reactions to Chemical Stimuli—Chemotaxis. 


1. Reactions to Localised Chemical Stimuli.— 
This particular phase of the general subject of the effects of 
chemicals may be considered first, since it is most closely 
related to what has preceded on the food reactions. The 
plan of the experiments was to try the effect of a series of 
substances when applied to restricted areas of the body. A 
sufficiently large number of chemicals were used to include 
representatives from each of the main groups of substances 
which have been found to have marked effects on organisms. 

a. Methods.—The method which was found to give the 
most satisfactory results in the application of localised 
chemical stimuli was the use of a capillary tube filled with 
the solution whose effects it was desired to test. The form 
of the tube used is shown in Fig. 32. The tubes were 10 to 
15 cm. long, and were made from glass tubing of about 
2°5 mm. internal diameter. Hach end was drawn to capillary 
fineness, and then broken off so as to give an opening of the 
desired size. ‘The opening at the upper end was made 


644. RAYMOND PEARL. 


slightly larger than that at the lower, which was used in 
giving the stimulus. The tube was filled with solution by 
suction. ‘lhe rate of diffusion can be regulated by changing 
the sizes of the openings, and can be determined for each 
tube from the rate at which the fluid sinks at the upper end 
of the tube. Considerable experimenting is necessary in 
order to get the best rate of diffusion for work on planarians. 
Since the animal is moving rather rapidly while the stimulus 
is being applied it is necessary to have reasonably rapid 
diffusion or the worm will not react at all, or not for so long 
a time after the stimulation has begun that one cannot be 
certain of the results. It is easily possible to get the 
capillary so fine that no results can be obtained. On the 
other hand, when it is too large the solution affects too large 
a portion of the body at one time, and furthermore, as will 
be shown later, may cause a rheotactic reaction of the 
organism. ‘lhis, of course, introduces a possible source of 


Vic. 32.—Glass tube used in giving localised chemical stimuli. 


serious error. It can be avoided by frequent and proper 
control experiments. 

It will be well to describe in advance the conduct of a 
typical experiment and the precautions taken, so that it may 
not be necessary to repeat these details in the account of 
each experiment. Six to ten normal active planarians were 
taken from the aquarium dish and put in a Petri dish of 
about LO cm. diameter, in freshly drawn, filtered tap water. 
Kuough water was put in the dish to give a depth of about 
lcm. ‘lwo or three of the capillary tubes with different 
sized openings were filled with the test solution. ‘These 
tubes were all tested before a final experimental series was 
begun, and usually only one which had been found to allow 
diffusion at the satisfactory rate was used. In some cases, 
however, varying degrees of sensitiveness among the 
different specimens made it necessary to use for some in- 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 645 


dividuals capillaries of faster or slower rates than what may 
be called the standard. After preliminary experiments to 
determine the relative sensitiveness of the different parts of 
the body to chemicals, attention was devoted almost entirely 
to stimulation of the head region, and consequently in the 
experiments which will be reported first the stimulus was 
applied only to the head, unless otherwise stated. The 
method of applying the stimulus was to place the point of 
the capillary tube a short distance (about 2 mm.) from the 
place on the body to be stimulated. ‘The animal was stimu- 
lated as it was gliding along in the normal way, and hence 
it was necessary to move the capillary tube at the same rate 
the animal moved in order to keep it opposite the same point 
in case the reaction was not given at the instant the capillary 
was put into place, which, of course, almost never happens. 
With a little practice one can move the tube along as the 
worm glides so as to keep the relative position of the two 
almost identically the same. Just as soon as a reaction had 
been obtained with a given specimen the capillary tube was 
removed from the water, so as to permit as little as possible 
of the chemical to get into the water surrounding the 
organism. After a series with any substance, the worms 
were transferred at once to a dish of fresh water before 
beginning another series. Further, in any long series, when 
for any reason it might be supposed that the water was 
becoming contaminated with the chemical to an extent 
sufficient to affect the results, the worms were transferred to 
another dish of fresh water. All through the course of an 
experiment frequent control tests were made by trying the 
effect on the worms of the water surrounding them when 
diffusing out from the same tube used previously for the 
chemical. After each experiment the tubes were thoroughly 
rinsed by drawing distilled water back and forth through 
them many times. ‘lhe tubes were also frequently discarded 
and new ones substituted. 


646 RAYMOND PEARL. 


The following substances were used in the experiments : 
(esas 
Mineral acids . +» Hydrochloric 
| gakphat: 
(eae 
Citric 
Formic 


Organic acids 


{Sodium hydrate 


Alkahes : ; ; 
| Sodium carbonate 


Salts of heavy metals | COPPer SEAS 
: | Zine sulphate 

Bane chloride 

Sodium bromide 


Other salts . rer : ‘ 
Potassium chloride 


Magnesium chloride 
Cane-sugar. 
Distilled water. 


Since distilled water was found to have a decided effect in 
producing a reaction, the solutions were prepared in both 
distilled water and in filtered tap water. In case of any 
doubt, as with very dilute solutions, the effects of the solu- 
tions prepared in each sort of water were tested and 
compared. 

Since only qualitative results were desired, and for the 
practical reason of greater convenience, percentage rather 
than molecular solutions were used. 

(3. Results.—The results are, in a way, so remarkable that 
they will be presented in some detail. 


Mineral Acids. 


Nitric (sp. gr. 1:42), + per cent.—This solution causes 
strong negative reaction, If applied to the head region the 
animal turns away from the side stimulated immediately, and 
strongly. If the stimulus is long continued the animal 
writhes and twists about violently. 

Stimulation of the posterior region causes the part where 


MOVEMENTS, E'TC., OF FRESH-WATER PLANARIANS. 647 


the solution strikes to contract very violently, and the whole 
animal to start crawling ahead rapidly. This concentration 
is very injurious, and if its action is continued, quickly kills 
the individual. It will be noted that its effects are the same 
essentially as those of strong mechanical stimuli applied to 
the same parts of the body. 

ify per cent. and 54, per cent.—Results the same as in + 
per cent. The animal is not as quickly and extensively 
injured by these solutions as by the former. It is to be 
noted that with these comparatively strong solutions the 
reaction time after stimulation of the posterior end of the 
body is so slow that this part of the body is permanently 
injured or destroyed before the animal gets away. 

zy per cent.—In some cases a well-marked positive re- 
action was caused by stimulation of the head region with 
this solution. The head would turn towards the mouth 
of the pipette in the characteristic fashion of the food reaction, 
or the reaction to weak mechanical stimuli. In other in- 
dividuals the reaction given was weakly negative, while still 
other specimens were indifferent. In cases where there was 
an indifferent reaction there was a local contraction of the 
side of the head stimulated. 

sy per cent.—Clearly marked positive reaction in large 
majority of cases after the stimulus has acted for some 
time. ‘This solution never caused the negative reaction. Some 
individuals were, in a few cases, indifferent to this solution. 
This solution is too weak to start a resting specimen into 
movement. 

chy per cent. and weaker.—Indifferent reactions or weak 
positive. 

This acid appears to be a very strong stimulus for the 
negative reaction in conceutrations down to 34, per cent., 
while below that it is a rather ineffective stimulus, and the 
reaction when induced is positive. 

Hydrochloric, ;4, per cent.—Strong negative reaction. 
There is noticeable in some cases a teudency for some 
individuals to turn very slightly towards the source of 


648 RAYMOND PEARL. 


stimulation before giving the strong negative reaction. 
Stimulation of the anterior end of a decapitated specimen 
caused a slow negative reaction with long reaction time. 
This solution causes the change from the glide to the crawl 
when applied to the posterior end of a normal worm. 

zy per cent.—Negative reaction; rather weaker than with 
preceding solution. With this solution one specimen would 
turn towards the source of the stimulus until the head came 
into the strong acid near the mouth of the pipette, and then 
give the sharp negative reaction. 

qo per cent.—Specimen A gave positive reaction in every 
case ; specimen B in about 50 per cent. of all cases, while 
in the remainder of trials gave weak negative. Other 
specimens negative reaction. 

sy per cent.—Specimen A as in preceding case. Specimen 
B gave positive reaction in about 90 per cent. of all trials. 
Other specimens weakly negative reactions. 

tuo per cent.—All specimens give well-marked positive 
reaction. They glide up to the end of the capillary and 
“orip”’? if with the anterior end as in the food reaction. 
After holding on for a moment they let go and give a sharp 
negative reaction, indicating that the stimulus is still too 
strong when continued. ‘his behaviour will indicate the 
machine-like character of the positive reaction. 

sxo per cent.—In the majority of cases indifferent re- 
action. Remainder positive. 

to give an idea of the dependence of the reactions to 
chemicals on the physiological condition of the organism, the 
following series of experiments with HCl in solutions of >45 
per cent. and weaker concentrations may be described. It is 
to be understood that these experiments were carried out 
on different animals from those just given. 

teo per cent.—No sharp positive reaction. Specimens 
will give a weak negative reaction if the opening of the 
capillary is held very near the head. In most cases reactions 
are indifferent. 


gives positive reaction and 


gzo per cent.—One specimen g 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 649 


goes through whole food reaction on the end of tube. 
The remainder still give weak negative reactions. 

six per cent.—Reactions essentially the same as in zt 
per cent. 

At this point this series was discontinued. It shows that 


0 


any absolute concentration for a chemical solution which will 
cause al] planarians to give the positive reaction cannot be 
assigned. How a given individual will react to a given 
concentration of chemical depends almost, if not quite as 
much, on the individual as it does upon the solution. 

Sulphuric, =, per cent. and 1, per cent.—Caused imme- 
diate and violent reaction. Decapitated worm reacts like 
normal. This is evidently a very strong stimulus, 

qi per cent.— Caused strong negative reaction in majority 
of cases. One specimen reacted as follows :—the capillary 
tube being held some distance away from the head, it 
first gave a well-marked positive reaction. On coming 
into the stronger solution near the mouth of the tube it 
began strong convulsive contractions (evidently on account 
of too strong stimulation). It remained, however, at the 
same spot, and after a few minutes extruded the pharynx 
and swept it about over the bottom. ‘The specimen re- 
mained this way for some time. The tube was, of course, 
removed immediately after the first positive reaction was 
given. A decapitated specimen in one case gave a very 
distinct positive reaction to this solution, the tube being 
held some distance away from the specimen. 

zy per cent.—Negative reaction. Decapitated specimen 
gave positive reaction once. ‘This solution, applied to 
the posterior end of the body, induces the crawling move- 
ment. 

xia per cent.—Negative reaction. Isolated pharynx con- 
tracts into a ball when stimulated with this solution. 

xia per cent.—Positive reaction in one case. Remainder 
negative. Same result with pharynx as in ;4, per cent. 

gin per cent., zs, per cent., and 5,4,5 per cent.—With 
these solutions the reactions were for the most part negative. 


650 RAYMOND PEARL. 


In a few cases positive responses were produced, but not 
recularly. 
_1__ yer cent. — Positive reaction in all cases. The 
5120 
whole food response was produced in case the end of the 


ce a) 


tube was left in position. The worms “ gripped” it, glided 
up on toit,and extruded the pharynx, in many cases running 
the latter up into the lumen of the tube. Anterior piece, 
resulting from cutting animal in two transversely, acts like 
whole worm (positive reaction), but less strongly. Decapitated 
worm gave no response. In order to make sure that in this 
ease it was the extremely diluted acid which was producing 
the result, numerous controls with distilled water and culture 
water and fresh tap water were tried on the same speci- 
mens, in alternation with trials with the acid. With tap 
water and culture water the specimens were indifferent ; 
but with the acid solution (34,5 per cent.) mixed in either 
tap water or distilled water they gave a well-marked positive 
reaction. This showed clearly that the results were due to 
the acid. 

Summary.—With the three mineral acids tested it was 
found that to concentrations above a certain point the speci- 
mens always gaye the negative reaction, while to concentra- 
tions below this point the positive reaction was given. The 
absolute value of this “critical point” varies widely with 
different individuals. The behaviour is essentially the 
same as that in response to mechanical stimulation, viz. to 
strong stimuli the negative reaction is given, to weak the 


positive. 
Organic Acids. 
Oxalic, + per cent. and ;4, per cent.—Sharp negative 


reaction. his solution affords a very strong stimulus 
and quickly kills the specimen. ‘The negative reaction 
is very violent when once induced, but several specimens 
were killed before they turned away. ‘There was notice- 
able a shght tendency to turn towards the stimulus the 
instant it was perceived, and before this could be replaced by 


MOVEMENTS, ETC,, OF FRESH-WATER PLANARIANS. 651 


the negative reaction the specimens were nearly or quite 
killed. 

zy per cent.—Convulsive negative reaction in the great 
majority of cases. In one case stimulation was followed 
by sharp positive reaction, succeeded by extrusion of the 
pharynx. 

> per cent. and 4, per cent.—A few specimens on some 
trials give positive reaction, and then go into convulsive 
twisting movements as they get into stronger solution. 
Remainder negative. 

sia per cent. and =4, per cent.—Positive and weak nega- 
tive reactions about equally divided. 

giv per cent. and 5,4, per cent.—Positive reactions 
becoming proportionately more numerous. Negative re- 
actions are very weak when given in response to these 
solutions. In the cases where there is a positive reaction 
the full response is not given; the specimens go up to 
the mouth of the tube, but do not grip it nor extrude the 
pharynx. 

seu per cent.—With this solution all but one specimen 
give the positive reaction. Specimens will follow the end of 
the pipette about the dish if itis moved slowly. his is done 
by a series of positive reactions. Specimens will give the 
complete food reaction on the end of the tube. 

Citric, 2 per cent.—Strong negative reactions. 

1 per cent.—Less marked negative reactions. ‘l'endency 
to positive in some cases. 

;'; per cent.—Positive reactions in nearly all cases. Re- 
mainder indifferent. 

=, per cent.—Indifferent. 

Citric acid in weak solutions seems to be a very ineffective 
sort of stimulus, not causing pronounced reactions of any kind. 

Formic, + per cent. and ;', per cent.—Prompt and de- 
cided negative reaction. Causes a resting worm to give 
a weak negative reaction of the anterior end, but does not 
start the whole animal into movement, provided the tube is 
withdrawn after the first reaction is obtained. 


652 RAYMOND PEARL. 


si; per cent.—Negative reaction, but decidedly less pro- 
nounced than with preceding concentrations. Does not 
cause any movement whatever in resting specimen. 

4, per cent.—Negative reaction, less strong than in pre- 
vious cases. In some cases positive reaction. Noticeable 
tendency to give slight positive reaction just before the 
definite negative response. 

=i, per cent.—Well-marked positive response. 

Summary.—The same conclusions are to be drawn from 
the experiments on organic acids as from those on mineral 
acids, viz. that to strong concentrations of a given substance 
the negative reaction is given, while weak concentrations 
cause a positive response. Oxalic acid is rather peculiar in 
that it appears to furnish in all concentrations a stimulus of 
the proper quality to induce the positive response, but is at 
the same time excessively harmful in any above the weakest 
solutions. 


Alkalies. 


Sodium Hydrate, + per cent., ;4, per cent.,and 54 per 
cent. — Immediate strong negative response. Specimens 
turn away very sharply. In ; per cent. the reaction is 
slightly weaker than in the other two. 

zi, per cent.—Negative reaction. Stimulus applied to pos- 
terior end of body is sufficiently strong to cause crawling 
movement. 

, per cent.— Weaker negative reaction. Sufficiently 
strong to start resting animal into movement. 

xiv per cent.—Weak negative reaction. Ineffective on 
resting worm and on posterior end of body of moving 
specimen. 

zi; per cent.—Very weak negative response. In one 
specimen sharp positive reaction; performs whole food re- 
action on the end of the tube. 

aiy per cent.—Positive reactions from all specimens. The 
complete food reaction is given. 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 653 


To solutions below this concentration the organisms are 
either indifferent or, in a few cases, weakly positive. 

Sodium Carbonate, + per cent.—Rather weak negative 
reaction. 

zg per cent.—Majority of all reactions positive. Remainder 
weakly negative. 

zg per cent.—Well-marked positive reaction in all cases. 
The specimens can be led around the dish by moving the tube 
slowly. 

Below this concentration the reactions were either in- 
different or weakly positive. 

Summary.—A caustic alkali (NaOH) and a salt of strong 
alkaline reaction (Na,CO;) produce essentially the same 
results as the acids. In strong solutions they cause negative 
reactions ; in weak, positive. 


Salts of Heavy Metals. 


Copper Sulphate, ~, per cent.—At the very first 
trials the animals all turned sharply and immediately towards 
the stimulus (positive reaction), but the solution was so 
strong as to throw the animal into convulsions, when the 
head came very near the mouth of the tube. Subsequent 
trials produced the negative response. 

zy and =, per cent.—All specimens give positive reaction. 
The head is brought up to the tube, and the worm glides up 
over the latter. 

With all concentrations of CuSO, there is a very well- 
marked local contraction of that part of the body which is 
stimulated. 

In another series of experiments with this same substance, 
the strongest concentration which would produce the positive 
reaction in all cases was 74,5 per cent. 

Zine Sulphate, + and 54, per cent.—Sharp and imme- 
diate negative reaction. 

zs per cent.—Negative reaction, but less pronounced than 
in former cases. 

zy and ,!, per cent.—Specimens stop when stimulated, 
vou. 46, part 4, 


NEW SERIES, UU 


654 RAYMOND PEARL. 


and wave the anterior end about in the water, first 
away from and then towards the source of stimulation. 
As the head comes nearer to the end of the tube, where the 
solution is strongest, it is more strongly stimulated, and 
gives a definite negative reaction. As it gets out into the 
weaker zone again it is stimulated to a positive reaction 
once more. If the tube is now removed the specimen will, 
in some cases, after a short time turn sharply towards the 
place where it was, and move in that direction. In other 
cases the negative reaction finally predominates. It not in- 
frequently happens that in the earlier part of this reaction the 
anterior end only moves very slightly towards, or very 
slightly away from the stimulus, so that the body seems, at 
first sight, to be fixed in one position. The planarian, in this 
strenuous reaction, probably comes as near to the hypothe- 
cated behaviour of the famous “ Buridan’s ass” as anything 
is ever likely to in actual practice. 

=i; per cent.—One specimen gave clearly marked positive 
reaction in every case. Others as in the preceding solutions 
(1; per cent. and 4, per cent.). 

xi; per cent.—Well-marked positive reaction. Specimens 
give complete typical food reaction. 

In one case, with a small worm, I was able to produce 
crawling ina backward direction by continuous stimulation 
of the anterior end in the middle line of the body with 
wy per cent. ZnSO,. 

Summary.—The results from solutions of salts of two 
heavy metals are in accord with those obtained with other 
chemicals. 


Other Salts. 


Sodium Chloride, + per cent. and ;4, per cent.—Nega- 
tive reaction; distinct, but not as strongly marked as the 
negative reaction to strong acids. 

si; per cent.—Weak negative reactions and weak positive 
reactions in about equal numbers. Many of the trials produce 
no response whatever. 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 655 


ga per cent.—Weak positive reactions in nearly every 
case. No negative reactions. The typical, complete food 
reflex I have not been able to induce with sodium chloride. 

Concentrations below this do not produce any definite re- 
action. 

In general, NaCl is a very ineffective stimulus to pla- 
narians, either to the positive or the negative reaction. Dis- 
tilled water is a considerably stronger stimulus to the positive 
reaction. 

Sodium Bromide, 2 per cent.—Weak but distinct 
negative reaction in all cases. 

2 per cent.—Well-marked positive reaction in all cases. 
Complete normal food reaction is produced. 

Potassium Chloride, 2 per cent.—The animals usually 
react in a peculiar way to this and stronger solutions of KCl. 
When stimulated they stop, turn the anterior end either 
shightly towards or slightly away from the source of stimula- 
tion, and then stay in the same place and squirm and twist 
the body. In some cases there is a well-marked negative 
reaction. 

4+ per cent.—Some specimens give negative reactions 
in the first few trials; afterwards give definite positive 
responses, as do other specimens in all cases. In one case 
the specimen gave marked positive reaction, and after the 
head was turned towards the stimulus, remained quiet in the 
same position as long as the chemical acted. 

zg per cent.—All specimens give positive reaction or 
are indifferent. ‘The whole food reaction took place on the 
end of the tube. In this experiment it could be clearly 
demonstrated that the pharynx is positively chemotactic to 
this substance. It is probably positively chemotactic to all 
substances which induce the preceding portions of the feed- 
ing reaction. If, after the pharynx had been extruded, the 
tube was turned about so that the ventral surface of the 
animal could be seen, and the posterior part of the body was 
moved with a needle, so as to change the position of the 
pharynx with reference to the mouth of the tube, it could be 


656 RAYMOND PEARL. 


seen that this organ bent directly towards the mouth of the 
capillary. ‘The pharynx oriented itself with reference to the 
issuing chemical. 

The cases in which specimens were “indifferent” to this 
solution (i.e. did not give either the positive or negative re- 
action) were evidently not due to the fact that the animal 
was not stimulated, but, on the contrary, that it was stimu- 
lated about equally to negative and positive responses. This 
was indicated by their restless behaviour when “ indifferent.” 
While the animal as a whole moves in a straight line, the 
head constantly moves slightly towards and away from the 
stimulus. Evidently the solution is not quite strong enough 
to induce a definite negative reaction, nor quite weak enough 
to cause a clear positive response. 

zy per cent., z45 per cent., and 4, per cent.—Distinct posi- 
tive reaction in all cases. 

ziz per cent.—Positive reactions in some cases, mainly 
indifferent. The ‘ indifference”? is now due to lack of 
stimulation. 

Below ;4, per cent. I have been unable to get definite 
responses of any sort with KCl. 

Magnesium Chloride, 4 per cent.—Usually sharp 
negative reaction. In some cases a slight turn towards the 
stimulus preceded the negative response, and in some few 
other trials the animal was indifferent. 

|; per cent.—Weaker negative reaction. In one case 
clear positive reaction. No local contraction of the region 
stimulated is caused by this chemical. 

sly per cent.—Positive reaction in all cases. Complete 
food reaction could be induced. 

zi; per cent.— Weak positive reaction or indifferent. 

Summary.—To the salts NaCl, NaBr, KCl, and Me(l, 
the planarians react as to other chemicals, by giving the 
negative response to strong concentrations and the positive 
to weak. 

Cane-sugar.—Sugar solutions, in all concentrations 
above ;'; per cent., so far as I have been able to discover, 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 657 


sause well-marked positive reactions in all cases. his is 
the only chemical which I have found that causes only one of 
the reactions. 

Distilled Water.—To distilled water applied by the 
capillary method the organisms give a well-marked positive 
reaction in all cases. That the reactions to very dilute solu- 
tions of chemicals were not due to the distilled water in cases 
where this was used as the solvent, rather than to the chemical 
itself, was proven in the following way :—Parallel experiments 
were performed, using tap water as a solvent, and in every 
case the same reaction was given to the tap-water solution as 
to that in distilled water. At the same time the specimens 
would not react to clear tap water applied in the same way by 
the same tube. 

2. General Summary.—Putting all the results on the 
effects of localised chemical stimuli together, we are forced to 
the somewhat remarkable conclusion that practically all sub- 
stances are both “ attractive ” and ‘ repellent ” to planarians. 
HKvidently, then, the chemical composition of a substance is 
not of the first importance in determining how the individuals 
shall react to it; but, on the contrary, its concentration is the 
important matter. To weak solutions of any chemical the 
animals give positive responses, while to strong solutions they 
give negative. 

Between the behaviour towards chemical stimuli and 
towards mechanical stimuli there is a very close parallelism, 
or, perhaps better, identity, which is evidently something of 
fundamental importance. In order to bring this out more 
clearly it may be well to arrange in tabular form the results 
of the study of the reactions to these two stimuli. 


658 RAYMOND PEARL. 


Mechanical Stimuli. Chemical Stimuli. 


Strong. Weak. Strong. | Weak. 


| Unilateral stimula- | Negative re- | Positive | Negative re- | Positive 
tion of head region | action reaction | action reaction. 


| 


ene zie Py 
Stimulation of head Hithera very | Positive | Strong nega- | Positive | 


region on median strong ne- | reaction tive reac- | reaction. 
line | gative reac- tion, or 
tion, or crawling 
crawling backwards | 
backwards 


Stimulation of middle | Essentially the same as | The same as for stimula- 
region of body for stimulation of the] — tion of the head, except 
head that the sensitivity is 
much less, and dimin- 
ishes more rapidly pos- 
teriorly than in case of 

mechanical stimuli. 


Stimulation of pos- | Crawling) Local con-]Crawling | No effect, 


terior region of| ahead traction ahead or slight 
body local con- 
traction. 


From this close parallelism we must conclude, I think, that 
in the behaviour of planarians the qualitative character of a 
stimulus is of little importance in comparison with its quanti- 
tative relations. Or, to express it differently, to all stimuh 
which are of low intensity the flat-worm gives the positive 
reaction, while to stimuli which are of high intensity it gives 
a negative response. ‘his sort of behaviour will at once be 
seen to be, in the long run, purposive, and is, further, of a 
kind which might very well have been developed by the 
action of natural selection. In the long run the planarian’s 
reactions will take it away from injurious substances and 
into favourable surroundings. 

These results on chemicals are interesting in connection 
with the work so much done in recent times on the specific 


MOVEMENTS, E'TC., OF FRESH-WATER PLANARIANS. 6059 


effects of ions and the conclusions based on very fine quanti- 
tative results with chemicals. Two such series of experi- 
ments as those quoted above from HCl and CuSO, indicate 
what would be the worth of the assignment of an absolute 
value for the concentration of either of these two substances 
which would produce the positive reaction in planarians. 
Such instances might be multiplied, and they serve to bring 
out the fact, apparently so frequently lost sight of, that what 
an organism will do when stimulated is quite as much a 
function of the physiological condition of the organism itself 
at the time as it is of the stimulus. 

A. comparison of these results with those of Yerkes (: 02) 
on the reactions of Gonionemus is of much interest. ‘This 
author finds that though there is a well-marked and 
characteristic food reaction, which is given in response to 
food substances, whether in solid or liquid form, yet this 
reaction cannot be induced by other chemicals. It is stated 
that a number of chemicals were tried in all concentrations 
for the special purpose of determining whether the food 
reaction might not depend upon intensity rather than quality 
of stimulus. ‘his was not found to be the case. We must, 
then, conclude that Gonionemus 1s a stage farther along in 
its psychic development than is the flat-worm, for the medusa 
reacts with reference to the quality as well as to the in- 
tensity and location of the stimulus, while with the flat-worm 
the intensity and location of the stimulus are by far the most 
important factors. It is necessary in the case of the flat- 
worm, to be sure, that there be mechanical and chemical 
stimuli acting together in order to produce the complex of 
reflexes forming the complete food reaction, thus indicating 
some relation to quality of stimulus. But for the production 
of what is, in one sense, the most important phase of the re- 
action, the turning towards the source of stimulation, the 
quality of the stimulus is not significant. 

With an understanding of the method of reaction to 
localised chemical stimuli, a number of interesting special 
problems present themselves. While it will not be possible 


660 RAYMOND ‘PEARL. 


to take up all of them in this paper, a few of the specially 
important ones may be considered. 

One such important general question which arises is the 
problem of orientation to diffusing chemicals. Do planarians 
orient themselves along radial lines of diffusion and proceed 
towards the centre of diffusion? It would seem that in the 
case of such a perfectly bilaterally symmetrical organism as 
Planaria, if anywhere, Loeb’s theory of orientation ought to 
hold good. ‘This theory accounts for orientation by sup- 
posing that when an organism is stimulated unilaterally its 
motor organs are caused to act either more strongly or more 
weakly, as the case may be, on that side than on the other. 
This results in bringing the long axis of the body parallel 
with the lines of action of the stimulus; and then, since 
symmetrical points on either side of the body must be equally 
stimulated, the organism moves in a straight line towards or 
away from the stimulus. Jennings has shown (: 01) that for 
most stimuli this theory of orientation does not hold in the 
case of the Infusoria. 

From the account of the reactions of planarians to 
chemical stimuli which has been given, it will be at once 
seen that there is in this case, to some degree at least, 
an orienting reaction. With weak chemical stimuli the 
head turns towards the stimulus in such a way as to point 
the anterior end very directly towards the source of stimula- 
tion. It mght be thought that this marked a pure orienta- 
tion, but it must be remembered that the organisms turn the 
head just as precisely towards the point from which a weak 
mechanical stimulus comes. ‘The two reactions are evidently 
exactly the same thing. However, a single mechanical 
stimulus can hardly be considered a directive stimulus of 
the sort which induces an orientation, such as, for example, 
the electric current. The orientation of unicellular organisms 
to the constant current is the purest type of an orienting 
response, however, and the most characteristic thing about it 
is that the organism, after having the anterior end turned 
towards one of the poles, keeps the long axis of the 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 661 


body parallel to the lines of action of the stimu- 
lus. This movement of the animal in a constant relation to 
a constantly acting stimulus is, as I understand it, the funda- 
mental criterion of an orientation according to the theory 
above mentioned. Now it we find, as has been shown above 
to be the case, that the organism gives precisely the same 
reaction to a chemical unilaterally applied as it does to a 
single weak mechanical stimulus similarly applied, it seems 
doubtful whether we can consider that there is such an 
orientation in the case of the chemical, even though the head 
is directed very precisely towards it. On the contrary, it 
seems apparent that we are dealing here with a well co- 
ordinated motor reflex only—such as, for example, the reflex 
of a frog’s hind leg, which brings its foot very exactly to the 
point stimulated on the side of the body. 

A crucial test of this point may be obtained by submitting 
the animals to the action of some chemical to which they are 
known to give the positive reaction when it is applied locally, 
only arranging the experiment so that it is diffusing over a 
large area. Under these conditions, if the organism shows 
positive orientation, it ought to move along the lines of diffu- 
sion straight up to the source of diffusion. To test this mattcr 
I constructed a trough of the form shown in Fig. 33, I. 
On a plate of glass A was fastened the trough B, which was 
cut froma block of paraffin. The internal dimensions of this 
trough were 50mm. x 50mm.-x 5mm. Only the sides were 
of paraffin, the glass plate serving as the bottom. A hollow 
was cut in one end of the trough, and a glass tube D, about 
4 cm. long, was fastened into it in an upright position. Then 
from the point a on the inside of the trough a fine needle was 
thrust through the paraffin till it came out into the hollow 
previously cut in the wall. A sectional view of this part of 
the device is shown in Fig. 33, II. When it was desired to 
use the apparatus the trough was filled with filtered tap 
water and a number of planarians placed in it. Then into 
the tube D was introduced a certain amount of the solution 
whose effects were to be tested. By varying the amount of 


662 RAYMOND PEARL. 


the solution introduced, the rate of its diffusion through z 
into the water could be very nicely controlled. This matter 
was thoroughly tested, and the apparatus in a_ sense 
calibrated by the use of coloured solutions before the actual 
experiments were begun. 

A considerable number of experiments were tried with this 
diffusion trough, with the following results :—In no case was 
there any observable orientation of the organisms. A typical 
experiment will illustrate what actually took place. A 
‘per cent. solution of Na,CO;, which by the capillary 


20 


method always produces a sharp positive reaction, was put 


Fie. 33.—I. Diffusion trough used for testing the reactions of planarians 
to diffusing chemicals. A, A. Glass base plate. 5B, B, B. Paraffin 
trough. «. Point of opening of diffusion tube. C. Cavity of 
trough in which the specimens are placed. D. Tube in which the 
solution to be tested is placed. If. Enlarged sectional view of the 
end of the trough bearing the diffusion tube. Lettering as in I. 


into the tube D in sufficient quantity to give a diffusion of 
moderate rate. After it had been diffusing for some time (by 
test with coloured solutions long enough to reach the middle 
of the trough) specimens were introduced at the end C. They 
started gliding about in random directions at once. Some 
passed diagonally up to the end D; others remained nearer 
the end C; while still others went up on the paraffin sides to 
the end D. None went straight towards « after they had 
come into the region where the chemical had diffused. No 
reaction of any sort was given in the course of the passage 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 663 


towards the end Din the majority of cases. In some few 
instances an individual would give a weak positive reaction 
(i.e. turn slightly towards x) at some point in its course, but 
this was so small in amount that it did not in most cases 
turn the animal directly towards a. Further, the direction 
of movement was frequently changed considerably, and turned 
away from «# after this» weak positive response. In other 
words, the animals moved about in the trough practically at 
random, giving only slight reactions in a few cases while in 
the area of diffusion. Many of the individuals, after reaching 
end D of the trough, turned around and went back to the 
other end again, just as they would have done provided no 
chemical had been present. Other specimens would glide 
across the trough on the paraffin of the end D. Only these 
specimens showed any definite response to the chemical. 
When they came within the length of their own bodies from 
the opening w they gave a well-marked positive reaction and 
went to a Having arrived there, they explored and 
“oripped” the edge of the hole with the head, and then 
extruded the pharynx. The pharynx was usually stretched 
up into the diffusion opening, and the worm proceeded to 
feed for a time on Na,COs,. 

These experiments were repeated many times with a 
variety of chemicals of various concentrations, and diffusing at 
various rates. It was very certain in all cases that there was 
no definite orientation along lines of diffusing ions. When 
the organism by chance came near the diffusion opening ~, it 
would give a positive reaction if the solution was of the proper 
concentration, and then proceed to give the complete food 
reaction over the hole, but there was no continued orientation. 

There was a similar absence of a negative orienting 
response when strong solutions of acids were used. In this 
case the animals stayed at end C of the trough, but this was 
because when, in the course of their random movements, they 
struck the diffusing chemical where it was of sufficient con- 
centration, they gave the usual negative reaction, turning the 
anterior ends about 30° away, and starting off on the courses 


664 RAYMOND PEARL. 


so defined. If they came in contact with the strong solution 
again they repeated the reaction. In no case did they turn 
squarely arcund with their heads directly away from w and 
the long axis parallel to the lines of diffusion. 

It would be unprofitable to further multiply accounts of 
these experiments, since all led to the same result. No 
definite orientation occurred, but only the positive and nega- 
tive motor reflexes coupled with random movements. 
Whether, as some maintain, we have in these positive and 
negative reactions the “ Dinge an sich” of orientations is a 
question for the metaphysician rather than the physiologist 
to decide. The objective reality of the matter is that in the 
behaviour of planarians towards chemicals there is no orien- 
tation in the lines of diffusing ions, i.e. no phenomenon lke 
the orientation of Paramecium to the electric current. 

Another problem of importance in connection with the re- 
actions of the organisms to chemicals has to do with the 
formation of collections of individuals. Are collections 
formed in certain chemicals, as is the case with certain of the 
Infusoria as described by Jennings? As this author has set 
forth, Paramecia will form dense aggregations in drops of 
various chemicals, particularly weak acids, introduced into 
the culture water. The method by which this is done is as 
follows :—Individuals swimming about at random strike the 
drop of acid by chance and pass into it without giving 
any reaction; when, however, they come to the opposite 
side of the drop, and start to pass from it to the water again, 
they are stimulated and give their characteristic motor re- 
action (jerk back and turn towards the aboral side). This 
reaction turus them back into the drop, which forms, as it 
were, a trap for all that enter it. In a short time a dense 
ageregation is formed. ‘This is almost the only method of 
active reaction, known aside from orientation, which will 
produce collections of organisms in chemicals. Its essential 
feature is not the getting of the organisms into the chemical, 
this being purely a matter of chance, but the holding of 
them in the chemical after they have entered it, by what 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 665 


amounts to a negative reaction to the surrounding water. 
The question, then, is, can we get any such formation of 
collections by the retention of those specimens which have 
entered an area by chance in the case of. Planaria ? 

This problem was attacked in a number of different ways, 
but the clearest results could be obtained by the “ two-drop ” 
method of Massart. ‘Two drops of fluid of equal size are 
placed near each other on a slide, and a narrow connecting 
band is made between the two by drawing some of the fluid 
across with a needle. One of them was usually of culture 
water, while the other was of the solution to be tested. Now 
evidently, if the animals form collections by the ‘ motor 
reflex ”’ 
without any reaction, but when they attempt to pass back 
into the water drop they should be stimulated to a negative 
reaction and thus turned back. 

An experiment with a solution to which the animal gives a 
sharp positive reaction may first be reported. One of the 
drops was tap water, and the other was 1 per cent. sugar 
solution, to which the specimens gave a strong positive re- 
action. Several small planarians were put into the water 


method, they ought to pass into the drop of solution 


drop. They glided rapidly about this drop, and soon one 
came up to the bridge connecting the water with the sugar. 
It was headed straight for the sugar drop, and passed over 
into 1t without any reaction whatever. Up to this point the 
behaviour is like that of the Infusoria towards the acid drop. 
This specimen circled about in the sugar drop, and after a 
time became directed towards the connection between the 
sugar and water, and passed back into the water drop with- 
out giving the faintest trace of a reaction of any sort. All 
the specimens passed back and forth between the two drops 
without giving any reaction, except in some cases a weak 
positive one. ‘lhe conditions under which a positive reaction 
is given are that a specimen should come more or less trans- 
versely across one end of the connecting bridge, as shown in 
Fig. 34. It then usually gives a weak positive reaction and 
turns slightly towards the other drop. It may do this on 


666 RAYMOND PEARL. 


passing either from the water to the sugar or vice versa. 
When in sugar solution it gives a positive reaction to tap 
water, whether applied by the capillary tube method or as just 
described. It is evident, from this experiment, that collec- 
tions are not formed by planarians in the same way that they 
are by Infusoria. The animals are not negative to the 
surrounding water after they have been in the solution, ‘To 
test and verify this conclusion the experiment was repeated 
with solutions of different substances. It was found that in 
case of all substances in concentrations to which the animals 
eave a positive response when stimulated by the capillary 
method, the specimens would pass back and forth from water 
to solution and vice versa, indifferently. If solutions were 
used in concentrations to which a negative reaction was given 


Fic. 34.—Diagram showing the arrangement of “ two-drop ”’ experi- 
ment with chemicals. 


when stimulation was by the capillary method, the specimens 
merely stayed in the water drop. When they came to the 
boundary line of the strong solution they gave the negative 
reaction, and hence stayed in the water. This immediately 
raises the question, why would there not be a permanent 
collection of the planarians formed in a drop of a substance 
to which they give the positive reaction, provided they were 
first put in a drop of some substance to which they were 
strongly negative ? There is evidently no theoretical reason 
why this should not take place, but there is an important 
practical one. ‘This is that any solution which would cause a 
negative reaction, under these circumstances, will, so far as I 
have found, also seriously modify the animals’ movements, if 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 667 
they are immersed in it. They will simply squirm about and 
make no progressive movements, and hence not get into 
the drop of substance to which they are positive. But it is 
quite possible that by making a long enough series of experi- 
ments on this point, one unght get a solution just strong 
enough to cause a negative reaction, and in which the 
organisms would still move well. We would then get a 
collection in the positive drop. The important thing, how- 
ever, is that to the water in which they live the animals do 
not, under any circumstances, give a negative reaction, and 
hence under normal conditions no collections can be found by 
a “motor reflex”? method. 

It may be well, before leaving this subject, to point out the 
fundamental physiological difference between the Infusoria 
and the planarians, on which the difference in the behaviour 
towards chemicals is based. It is that in the case of the 
Infusoria there is but one form of reaction (the ‘‘ motor reflex”’ 
turn towards a structurally defined side) regardless of whether 
the stimulus is strong or weak, while in the case of the 
planarian there is a qualitatively different reaction to strong 
stimuli from that which is given to weak. When the in- 
fusorian passes into the drop of acid it is apparently not 
stimulated at all (for what reason we do not know). When 
it attempts to pass from acid to water it is given a stimulus 
which must be in the nature of things a rather weak one, yet 
it responds with the only reaction it has, and is, as a 
consequence, kept in the acid. With the planarian any slight 
change in environmental conditions gives a weak stimulus, 
and the specimen turns towards the source of stimulation. 
This serves, together with random movements, to get it into 
the drop of solution; but when it strikes again the water, 
which again must furnish a weak stimulus, it gives the same 
positive reaction and passes out into the water. The ability 
to differentiate in the reactions between the strong and weak 
stimuli gives the organism a far greater range in_ its 
activities. 

Another problem which is of interest in connection with 


668 RAYMOND PBRARL. 


food and chemical reactions is the relation of the condition 
of the organism as regards hunger to its reactions to stimuli. 
It might be supposed thatan individual which had not had food 
for some time would be more apt to give the positive reaction 
to a given stimulus than one which had just fed. 

To test this point parallel experiments were instituted with 
specimens allowed to feed till they left the food spontaneously 
about three hours before the experiments, and specimens 
which had been kept for three weeks in a dish of clear water. 
NaBr was used as the stimulating solution, and was applied 
by the capillary method. ‘The specimens chosen were of the 
same species, P. dorotocephala, and as nearly as possible 
of the same size. The only difference which could be detected 
between the fed and the unfed animals in their behaviour 
towards a 2 per cent. solution of NaBr was that the unfed 
animals gave the whole food reaction on the end of the 
capillary tube, while the recently fed specimens only went so 
far as to give the positive reaction, and touched the end of 
the tube with the anterior end of the head. ‘They did not 
“orip’’ it and pass up on to it, as did the others. In the 
main point at which I was working, namely, the giving of 
the definite positive reaction, there was no discoverable 
difference between the fed and unfed specimens. One set 
cave the reaction just as promptly and decidedly as did the 
other. Next a weaker solution, ;4, per cent., was tried. 
With this solution about 50 per cent. of the specimens in 
ordinary condition give a weak positive reaction, and 50 per 
cent. are indifferent. This concentration, being about on the 
border line between that which affords no stimulus at all and 
that which is a definite stimulus for the positive reaction, 
ought to bring out any differences which may exist between 
fed and unfed individuals in the sensitivity to stimuli for the 
positive reaction. As a matter of fact, no difference in the 
behaviour of the two sets was to be observed. One gave a 
well-marked positive reaction in as many cases as did the 
other. In some instances the reaction time of the fed 
specimens seemed to be slightly greater than that of the 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 669 


unfed, but this was neither marked nor of general occur- 
rence. This experiment was afterwards repeated with other 
specimens, and with sugar as the stimulus, with essentially 
the same results. I have also repeatedly tried stimulating 
with various solutions specimens which had just ceased 
feeding, and in these cases found no certain difference 
between their behaviour and that of specimens which had 
not been fed for some time, with regard to the giving of the 
positive reaction. It would appear, then, that so far as the 
giving of the positive response to weak stimuli is concerned, the 
amount of food the animal has previously had is of very little 
consequence. The failure of fully fed specimens to give the 
full feeding reaction on the end of the capillary tube indicates 
that the physiological changes induced by recent feeding 
affect the performance of the food-taking rather than the 
food-seeking reflexes. 

3. Unlocalised Action of Chemicals.—An extensive 


Fic. 35.—Diagram showing the form of crawling movement exhibited 
by Planaria when placed in 10 per cent. NaCl. 
series of experiments on the effects of immersing planarians 
in various solutions was performed, but as the results threw 
but comparatively little light on the general nature of the 
behaviour, they will be reported only briefly. Immersion in 
any strong solution causes marked changes in the move- 
ments. The gliding is made very much slower or entirely 
disappears. In 10 per cent. NaCl a peculiar form of crawling 
appears. Very pronounced contraction waves pass over the 
body longitudinally, giving it the appearance shown in 
Fig. 35. In 2 per, cent. CuSO, the animals make no pro- 
gressive movements, but wave the head violently from side to 
side. In strong solutions of acids the worms squirm violently 
without making any effective progressive movements. In all 
these strong solutions the sensitiveness to all stimuli is 
vot. 46, PART 4,—NEW SERIES. sas 


670 RAYMOND PERARL. 


greatly diminished. This can best be shown with mechanical 
stimulation. In strong solutions of NaCl (10 per cent.) the 
animals make no attempt to right themselves if placed with 
their dorsal surfaces down. Another peculiar effect of strong 
solutions of NaCl is to cause the extrusion of the pharynx. 
This organ is thrust out of the body and extended to a much 
greater length than is usual. Immersion of the animal in 
weak solutions that cause the positive reaction—as, for 
example, 1 per cent. sugar—has no definite effect on the 
movements, but when in these solutions the animals will give 
the positive reaction to tap water when the latter is applied 
by the capillary tube method. Under such circumstances 
contact with water is a slight environmental change, and acts 


as a weak stimulus. 


Ill. Thigmotaxis and the Righting Reaction. 


a. Thigmotaxis.—lf a specimen of Planaria is turned 
over and placed dorsal side down on the bottom, it will 
immediately right itself. This is done by a very characteristic 
reaction, and is one of the first things to attract the attention 
of one studying the behaviour of the organism. Loeb (94, pp. 
251—252) held that the righting reaction in the polyclad 
Thysanozoon was due to the negative and positive thig- 
motaxis (stereotropism”’) of the dorsal and ventral surfaces 
respectively. The evidence offered for this view was that 
when the thigmotactic relations of these two surfaces were 
reversed, the animal reacted strongly, and that this result 
could not be due to any effect of gravitation, since the animal 
assumed all possible relations to gravity, and kept them for 
considerable periods. of time. It seemed to me desirable to 
get, if possible, some further evidence on this subject, and to 
work out the mechanism of the righting reaction. 

That the dorsal surface of the animal is negatively thigmo- 
tactic 1s certain, and can be shown in other ways than by 
laying the animal on its dorsal surface. For example, if a 
piece of cover-glass be gently laid on the dorsal surface of 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 671 


either a resting or a moving specimen, it will very promptly 
move out from under it. Further, if crevices are arranged of 
this form (—) by supporting cover-glasses at two corners, 
and letting the two opposite corners rest on the bottom of the 
dish, specimens will not go into them. ‘The moment the 
dorsal surface touches the cover-glass above, the worm 
begins to react violently, changing its direction of movement, 
and goes out from under the cover. 

With the existence of an apparent negative thigmotaxis of 
the dorsal surface established, however, there still arises the 
question as to whether this is the sole cause which induces the 
inverted animal to right itself. The following experiment 
throws light on this point :—A specimen is placed ventral 
side up on a dry spatula in the air, and then the spatula is 
placed just beneath the surface of the water in a tall jar or 
large test-tube and quickly pulled out from under the worm, 
so that the latter starts falling through the water in an in- 
verted position. Another way in which the worm may be 
started falling ventral side up is by holding it on a scalpel 
point above the water, and then dropping it beneath the 
surface in the desired position. Before the worm has 
dropped any great distance it will give the characteristic 
righting reaction, and turn itself over so as to bring the 
ventral side down again. ‘This is done in precisely the same 
way as when the animal is inverted on the bottom (to be 
described later). After the falling animal has thus righted 
itself it may again give the same reaction, and thus turn 
itself over so that the dorsal side is down again. In a few 
cases I have seen a worm after righting itself the first time 
keep right side up during the remainder of the fall. The 
most usual behaviour is for the animal to keep giving the 
righting reaction all the time that it is falling, although this 
does not, of course, keep it all the time with the same side 
uppermost. JI have performed a large number of these 
dropping experiments in which the animals were started in 
both upright and inverted positions, and in all cases they 


gave the righting reacting one or more (usually more) times 


672 RAYMOND PEARL. 


before reaching the bottom, provided the distance through 
which the drop was made was greater than 7—10cm. ‘This 
result seems to indicate that there is something more con- 
cerned in the righting reaction than the negative thigmo- 
taxis of the dorsal surface for the following reasons :—(1) the 
dorsal surface is not in contact with any solid of this experi- 
ment; (2) it is in contact with water only, just as is normally 
the case when the animal is right side up. It may be 
objected that the experiment is not conclusive, because, as a 
result of the falling, there is an increased water-pressure on 
the dorsal surface, and this may act as a _ thigmotactic 
stimulus. This objection is met by two different facts. 
First, the animal gives the righting response in some cases 
while falling ventral side down, under which circumstances 
there can be no increased pressure on the dorsal surface. 
Second, if a stream of water from a pipette is directly 
squarely against the dorsal surface of a worm normally 
gliding about on the bottom the righting reaction is not 
induced, regardless of the force of the stream. LHvidently 
this stream of water against the dorsal surface produces a 
pressure on the dorsal surface similar to that when the 
animal is falling, and if the righting reaction in the falling 
is due to increase of pressure on the dorsal surface, we might 
suppose that some indication of it would be produced in this 
case. As a matter of fact it is not. We must conclude, 
then, that the righting reaction is due, at least in very large 
part, to some other cause than the negative thigmotaxis of 
the dorsal surface. This is indicated also by the fact that 
when solid bodies are laid on the back of a specimen in its 
normal position, the reaction which is caused is not the 
righting action, as would be expected if the latter were 
due solely to the negative thigmotaxis of the dorsal surface. 
The righting reaction is clearly not due to gravitation, since 
the flat-worms move on the surface film with the dorsal 
surface downward. This leaves, as the only factor te which 
the reaction can be due, the positive thigmotaxis of the 
ventral surface. Iam convinced that it is to this factor that 


MOVEMENTS, EBTC., OF FRESH-WATER PLANARIANS. 673 


the reaction is chiefly due. While the negative thigmotaxis 
of the dorsal surface plays some part in the reaction, it is, as 
the experiments described above show, a comparatively un- 
important factor. ‘The specific relation of these two factors 
to the definite righting reaction will be brought out in the 
next section, in which the form and mechanism of this 
reaction will be set forth. 

b. The Righting Reaction.—The righting reaction is 
a very characteristic piece of behaviour, and can best be 
described in a single phrase by saying that when the animal 
is placed on its back it throws itself into a spiral in such a 
way that the ventral surface of the head comes into contact 
with the bottom. This ventral surface then attaches itself to 
the bottom by means of the mucous secretion, and starts 
eliding ahead. As it goes forward it unwinds the remainder 
of the spiral, as each successive posterior part of the ventral 


Fre. 36.—Showing the form taken by Planaria in the righting reaction. 


surface comes into full contact with the bottom. ‘he form 
of this spiral just after the ventral surface of the head has 
come into contact with the bottom is shown in Fig. 36. The 
spiral is thrown very quickly after the dorsal surface touches 
the bottom, and usually includes the whole length of the 
body at once. However, by observing a specimen in which 
it takes place a little more slowly than usual, it can be seen 
that the movement is started at the anterior end. Beginning 
with, for example, the right side of the head, this is turned 
under, while at the same time the left side is raised. ‘This, 
of course, brings the ventral surface of the head region down, 
and at the same time makes a twist in the body, just back of 
the head. In some cases this is the only twist that is made, 
while in otliers another similar twist is thrown in the body 
farther back. As the anterior end after it is righted elides 


674 RAYMOND PEARL. 


ahead, the spiral is unwound by the raised edge of each 
twist dropping down and attaching to the bottom as soon as 
it is in a position where this is possible. Thus, of course, 
when the animal has traversed a distance equal to its own 
length it will have come entirely into the normal position 
again. The reaction is really a rotation of the body on its 
long axis through 180°. The mechanism of the turning is 
such that only a part of the body rotates at a time,—first 
the anterior end, then the portion next behind that, and so 
on, till the whole animal has turned over. This rotation by 
sections, as it were, causes the spiral form which the animal 
takes on in the reaction, 

The number of turns into which the body is thrown in 
forming the spiral varies with the length of the individual, 
and apparently to some extent with its physiological con- 
dition. ‘There may be only a half-turn in the whole body, or 
there may be one complete turn; or, again, one and a half 
turns; or, finally, as many as two complete turns in the body. 
One complete or nearly complete turn, as shown in Fig. 36, 
is the usual form of the reaction. In large individuals 
more twisting is frequently seen. Hvidently all the twisting 
that is absolutely essential for the righting of the specimen 
is the half-turn given by the turning of the anterior end 
ventral side down. 

The determination of the direction in which the spiral is 
thrown, or, in other words, the side of the body towards 
which the anterior end turns in order to get right side up, 
was for some time a very puzzling problem. A collection of 
statistics on the matter showed that the anterior end twisted 
towards the right and towards the left! in an approximately 
number of cases. This is precisely the result which would 
be expected if the matter were due to chance only, but the 
reaction did not give the appearance of being a chance 
matter. Finally, the determining factor was found to be the 
relation of the dorsal surface to the bottom. A cross-section 


' In the figure (Fig. 36) the worm is represented with the spiral thrown 
towards the left, 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 675 


of the body of Planaria has the form shown in Fig. 37. It 
is convex in outline on the dorsal side, and nearly straight on 
the ventral, As a consequence of this shape of the dorsal 
surface the animal when placed in an inverted position very 
seldom lies exactly on the mid-dorsal line, and if it does at 
first it almost immediately tips over to one side or the other, 
so that its cross-section has the relation to the bottom shown 
in Fig. 37, Band C. It is then found that the side of the 
body which is in contact with the bottom determines in 
which direction the spiral shall be thrown. If the right 
side of the dorsal surface is down the right side of the head 
will turn under towards the left and the left side will be 
raised up over towards the right, or, in other words, the 
head as a whole will rotate from right to left, i.e. in a 


Dorsat 


LIES 


VENTRAL 
A 


ath BE BEA ESB 0 


B C 


Fig. 87.—Diagrammatic cross-section of Planaria to show the contact 
relations of the dorsal surface of the body to the substrate in the 
case of a specimen in an inverted position. 

counter-clockwise direction. If the left side of the dorsal 
surface of the body is down at the beginning, the head will 
rotate from left to right. This relation may be made out 
easily by direct observation in all cases where the reaction is 
not too rapid. 

The righting reaction is a fairly rapid one. The head is 
turned over and the spiral thrown in the case of a normal in- 
dividual almost immediately when the dorsal surface touches 
the solid. The length of time which it takes a specimen to 
eet completely righted evidently depends on the length of 
the body, because the longer spiral which must be unwound, 
the more the time which must be taken. The following 
figures will bring out this relation between the size of the in- 
dividual and the time taken in righting. In ten trials with 


676 RAYMOND PEARL. 


an active but large specimen (about 12 mm. long) of 
P. dorotocephala the average time taken to regain com- 
pletely the normal position after being inverted was 8°68 
seconds. With a small specimen (5°5 mm. long) the average 
time taken in righting in ten trials was 5°22 seconds. The 
time taken in the reaction also depends, of course, on 
the general physiological condition of the animal. Thus in 
ten trials with a sluggish specimen, approximately 9 mm. 
long (thus shorter than the first specimen mentioned), the 
average time taken in regainine the normal position was 
10:90 seconds. 

The thigmotactic irritability may be modified or reduced 
in several ways, and, as a consequence, the righting reaction 
will disappear entirely or in part. One of these cases has 
been mentioned above (p. 670) where it was shown that a 
specimen placed on its back in a 10 per cent. solution of 
NaCl makes no attempt to right itself. Sinmularly a specimen 
put in an inverted position on a dry surface, care being taken 
that no water surrounds the animal, will not give the righting 
reaction. In both of these cases the specimens are able to 
move. 

The Mechanism of the Reaction.—It is a very 
difficult matter to determine exactly the muscular mechanism 
of this righting reaction, since it is such a complicated move- 
ment, and is ordinarily done in its most essential feature— 
the formation of the spiral—so very quickly. Furthermore, 
as will appear from the operation experiments to be described, 
it is almost impossible to devise crucial experiments of a 
character which will demonstrate what the mechanism is. 
What I shall do, then, will be to present a tentative explana- 
tion of the mechanism of the reaction, together with the 
evidence for it which I have been able to obtain. I may say 
that the view to be presented is the result of a long and 
careful study of the phenomena both in normal and operated 
worms, and I believe that it is a correct explanation. 

The mechanism of the righting reaction is probably as 
follows :—The half of the body of an inverted specimen which 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 677 


is in contact with the bottom extends (by the mechanism 
previously described, pp. 556,557) in response to the stimulus 
given by the contact of the dorsal surface of that side of the 
body with the bottom. At the same time the opposite half of 
the body, by active muscular contraction, keeps its length 
the same. Thus any bending of the body away from the 
side stimulated as in the ordinary negative reaction is pre- 
vented, or, in other. words, the long axis is kept straight by 
the opposite side maintaining actively its normal length. 
Now the necessary mechanical result of keeping one side of 
a flexible system at a constant length while the other side 
lengthens must be that the lengthening side will be thrown 
into a series of waves. In other words, it is mechanically 
impossible for the lengthening side to keep its whole edge in 
the same plane. Furthermore, if in such a system it is 
possible for rotation about a longitudinal axis to occur, the 
system will be thrown into a spiral of the form which the 
planarian takes in the righting reaction. Again, as soon as 
one side of such a system under elongating stress changes its 
level with reference to the remainder of the system, and thus 
starts the formation of the spiral, the lone axis of the 
system (i.e. the centre of the spiral) will keep itself straight. 
Any further force elongating one side will merely throw the 
spiral into tighter coils without having any tendency to bend 
its long axis. This fact is of importance in the case of the 
planarian where the maintenance of the initial straightness 
of the long axis is done by the opposite side of the body. Of 
course, a symmetrical spiral cannot be formed unless the two 
edges are of equal length, but the moment the spiral of the 
planarian is started all necessity for one side keeping a con- 
stant length ceases. It must be kept in mind, however, as 
has been indicated above, that the force which produces the 
spiral must act on one side only, and hence the side of the 
planarian opposite that initiating the movement must be 
moved passively by the other in the spiral formation after 
this has once begun. The direction in which the spiral shall 
turn will evidently not be determined by the mere lengthen- 


678 RAYMOND PEARL. 


ing of one side of the body. The determinant of this is 
evidently a difference of tension on the upper and lower sides, 
the spiral turning towards.the side of greatest tension.’ This 
ereatest tension is evidently, then, in the normal reaction on 
the dorsal surface, as we should expect on a priori grounds, 
since that is the part directly stimulated. 

To sum up, the spiral righting reaction of the planarian, 
as I have worked it out, is due to an elongation of that side 
of the body whose dorsal surface is in contact with the solid, 
while the opposite side of the body actively maintains its 
original length. As the elongation occurs the various parts 
of the body rotate freely about its long axis, and hence the 
whole worm takes on the spiral form. The spiral turns 
towards the dorsal surface in every case (i.e. so as to bring 
the ventral surface of the head down), as a result of the 
ereater tension of the dorsal musculature on the elongating 
side. 

The reaction is thus seen to be of almost the same cha- 
racter as the ordinary negative reaction to strong mechanical 
stimuli, in that the primary reaction is an extension of the 
side stimulated. The difference between the two is that in 
one case there is a bending of the longitudinal axis of the 
body, while in the other there is a rotation about this axis. 
On the view just given of the mechanism of the righting 
reaction the specific parts played by the positive and negative 
thigmotaxis of the ventral and dorsal surfaces are evident. 
The positive thigmotaxis of the ventral surface is the primary 
cause of the whole reaction, and is evidently the stronger 
factor of the two, as shown by the experiments of laying 
solid bodies on the dorsal surface of the animal when in a 
normal position. It will be recalled that such treatment does 
not call forth the specific righting reaction. Further 
evidence of this same thing is found in the fact that speci- 


' The statements as to the mechanical principles of a spiral have been veri- 
fied with different sorts of models, including plastic clay, rubber bands, ete. 
Lack of space will not permit the enumeration of these experiments in detail, but 
anyone can verify for himself the various statements with very little trouble, 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 679 


mens will remain in the normal position on the bottom of a 
dish when there isa layer of plant débris a half-centimetre 
in thickness above them, and necessarily in contact with the 
dorsal surface. The negative thigmotaxis of the dorsal 
surface plays its part in the righting reaction in determining 
in which direction the turning shall take place. 

It has so far been shown that the view of the mechanism 
of the righting reaction presented is in accord with all the 
mechanical principles necessary to produce the observed 
results. The attention may now be turned to an examination 
of the evidence that this mechanism is the one which actually 
brings about the reaction. ‘This evidence is obtained from 
experiments with worms on which operations have been per- 
formed. Obviously, if the mechanism described is the one 
by which the reaction is produced, any operation which 
destroys or throws out of working order any essential part 
of the mechanism will cause the typical reaction to disappear, 
or be greatly modified. 

We may first consider the reactions of the pieces resulting 
from cutting the animal in two transversely in the middle of 
the body. It is found that each of the pieces resulting from 
such a cut will perform the righting reaction in the typical 
manner. ‘The spiral is formed, but there is usually only one 
half-turn of the body, i. e. just enough to bring the anterior 
end ventral side down. This then attaches itself to the 
bottom and starts gliding, unwinding the spiral just as under 
normal circumstances. There is observable the same rela- 
tion between the side of the body, which is in contact with 
the bottom and the direction of the turn as in the normal in- 
dividual. The only striking difference in the behaviour of 
the anterior and posterior pieces is that the reaction time of 
the former is much shorter than that of the latter. The 
anterior piece rights itself practically as quickly as does the 
normal animal, while the posterior piece took in one series of 
experiments | minute and 38°1 seconds (average of ten trials) 
for complete righting. This slower righting reaction is 
another expression of the generally lowered tonus of. such 


680 RAYMOND PEARL. 


posterior pieces. By varying the position of the cuts, seg- 
ments of the body of various lengths may be obtained. All 
of these, which are about 14 mm. in length, will usually right 
themselves by as close an approximation to the typical spiral 
reaction as is possible under the circumstances. ‘The side of 
the body which is lowest can be seen to elongate in these 
very short pieces, and just enough of a twist is found to 
bring the ventral surface of one corner of the anterior end 
into contact with the bottom. Of course, no complete spiral 
‘an be found in such short pieces. Their reaction time is 
very slow. 

Next, experiments were tried with the pieces resulting 
from splitting longitudinally anterior halves of worms in the 
middle line. These pieces had the form shown in Fig. 38. 
Evidently such pieces have only a half of the mechanism 
necessary for the performance of the spiral righting reaction, 


Fie. 38.—Operation diagram (see text). 
according to the view given above, and therefore should not 
be able to give the typical response. They have one com- 
plete side which may elongate, but they have no other side 
to keep the middle line straight, and so make the elongation 
effective in forming a spiral. Such pieces, when placed with the 
dorsal surface down, reacted immediately by bending strongly 
towards the cut side, i. e., so that the concavity was on the 
cut side, This was kept up for a time, the animal squirming 
about violently, but it was finally replaced by another reaction. 
The ventral longitudinal muscles contracted strongly, and 
‘aised the anterior end of the piece well up from the bottom 
(shown in side view in Fig. 39,a). After a strong raising con- 
traction the piece would extend and settle back again. Then 
after a time the raising was repeated, and it soon became 
noticeable that the piece was rising higher each time and 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 681 


settling back less after each trial. Successive stages of this 
rising are shown in Fig. 39, b, c,d. Finally, it worked up till 
it stood directly on the posterior end (e), and then the next 
contraction caused it to fall over of its own weight and come 
down right side up (f). The sticky mucous secretion at the 
posterior end was undoubtedly what held the piece up after 
each successive trial. This behaviour, as described, was 
uniform in all the trials. 

The behaviour of these pieces brings out several points of 
importance. First, it is to be noticed that no trace of the 
typical spiral righting reaction is to be seen; yet, on the other 
hand, we find the pieces bending strongly towards the cut 
side when first inverted, which is just the effect which would 


2 f 
Pie. 89.—Diagram showing the method of righting adopted by one 
of the pieces shown in Fig. 38. 


be produced by the lengthening of the stimulated side in the 
normal righting reaction, provided, as actually obtains in this 
case, there was no opposite side to keep the long axis of 
the piece straight. Thus we get precisely the result 
which would be expected if the view given of the mechanism 
of the reaction is the correct one. Another fact that is 
brought out by this experiment is the apparent adaptation 
shown. Whenthe animal is unable to give the usual reaction 
for righting itself it very quickly reacts in an entirely 
different way, but attains the same end result. 

A worm was cut so as to give a piece of the form shown at 
A in Fig. 40. This piece was placed in an inverted position 


682 RAYMOND PEARL. 


and its reactions observed. Evidently, so far as injury of 
the mechanism by the operation is concerned, such a piece 
is in essentially the same condition as the pieces described 
in the previous experiment. It has only one complete side 
of the body. The piece when inverted squirmed about consider- 
ably at first, but gave no indication whatever of the normal 
spiral reaction. In a short time the violent movements ceased, 
and a notch was noticed in about the middle of the uncut 
edge (cf. Fig. 40, b). This soon grew larger, and extended 
more and more towards the ends of the piece, as shown in 


e oe f 


Vic. 40.—a. Operation diagram. Heavy lines indicate the cuts. 4, ¢, 
and d. Successive stages in the righting reaction of the piece A of 
diagram a. e and /. Cross-sections through A at two successive 
stages in the righting process. See text for further explanation. 


cand d. By close observation the cause of this appearance 
was found to be that the thin mobile edge was folding 
under and attaching its ventral surface along the bottom, A 
cross-section through the worm at this stage had the outline 
shown ine. As soon as a considerable portion of the edge 
had so folded wnder and become attached, the piece gave 
a series of strong contractions and literally “ flopped” over 
the attached edge and came down right side up. A stage in 
this process is shown in cross-section in f. ‘This behaviour 
was so peculiar, and at the same time precise, that the 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 683 


experiment was repeated many times on this piece and on 
others cut in the same way. The same method of righting 
was always observed. After the first few times the turn 
is made in this way; it is done more quickly at each succes- 
sive trial. 

This experiment leads to the same conclusion regarding 
the mechanism of the righting reaction as did the previous 
one. It affords another and more striking example of regu- 
lation in reactions. The piece attains the end (normal 
position) by a reaction which it undoubtedly never had 
occasion to practise before. 

Tsolated longitudinal halves of the body react in the same 
way as did the piece described in the preceding experiment. 
They right themselves by folding under the edge, and then, 
by violent contraction, drawing the rest of the body up over 
it. There is no trace of the spiral righting reaction. 

A specimen cut in the manner shown in Fig. 41 shows a 


Fie. 41.—Operation diagram (see text). 


very peculiar righting reaction. When placed dorsal side 
down the portion posterior to the median longitudinal 
slit immediately gives the spiral righting reaction, and drags 
the two passive anterior pieces over. ‘I'he process is slow but 
very characteristic, so that there is no doubt of the nature of 
the reaction. This shows that in that part of a single piece 
of a worm where the necessary mechanism is present we get 
the spiral righting reaction, while in other parts it does not 
appear. 

The same point can be brought out by sphtting a worm 
longitudinally from the posterior end up to a poimt near the 
head. The complete anterior part of such specimens gives 
the normal spiral reaction, while the posterior parts remain 
passive so far as this reaction is concerned. 

A considerable number of different experiments were per- 


684 RAYMOND PEARL. 


formed for the purpose of testing the righting reactions after 
operations, but since none of them bring out anything 
different in principle from the results already given, they will 
not be reported here. But it may be said in general, that all 
the experiments gave the same results with reference to the 
mechanism of the reaction, namely, that so long as the 
mechanism described above was intact the typical spiral re- 
action was given; when this mechanism was destroyed or 
injured the reaction was not given, but the animal, if it 
righted itself at all, did it by a different method. 

When the animal falls freely in the water the righting re- 
action is induced because the ventral surface is no longer in 
contact with a solid. There is no reason for thinking that 
the mechanism of the reaction in this case is any different 
from what it is when the animal is placed in an inverted 
position on the bottom. The direction in which the spiral is 
thrown in the case of the falling animal is probably deter- 
mined by slight differences of pressure on the two sides of the 
body. 

c. Summary.—The flat-worm is positively thigmotactic on 
its ventral surface, and negatively thigmotactic on its dorsal 
surface. As a result of this it gives a characteristic righting 
reaction whenever the normal relations of either surface are 
changed. This righting reaction consists in throwing the 
body into a spiral in such a way as to bring the ventral sur- 
face of the anterior end down into contact with a solid (in all 
vases except when the animal is dropped into free water). 
The anterior end starts glding and unwinds the spiral, 
thus righting the whole body. The thigmotactic reaction 
may be modified by chemical and other stimuli. <All the 
evidence shows that the spiral righting reaction is due to a 
lengthening of the side whose dorsal surface is in contact 
with a solid, while the other side of the body keeps the long 
axis straight. ‘The direction of the turn in the spiral is 
determined by the side of the body which is in contact with 
the solid. This reaction is thus seen to be closely related to 
the negative reaction to mechanical and chemical stimuli, so 


MOVEMENTS, E'TC., OF FRESH-WATER PLANARIANS. 685 


far as mechanism is concerned. Cut pieces, in which the 
normal mechanism for the righting reaction has been 
destroyed, right themselves in various ways, thus showing a 
sort of regulation in reactions. 


IV. Electrotaxis. 


In view of the sharp and precise reactions of planarians to 
other stimuh, it was thought that they would furnish excellent 
objects for the study of electrotaxis, but unfortunately this is 
not the case. Their reactions to the constant current are not 
clear-cut, since the specimens become wholly or partially 
paralysed in a very short time after the current begins to 
act, and as a consequence the reactions become feeble and in- 
distinct. For the sake of completeness, however, and since 
some facts of importance are brought out, the experiments on 
this subject will be briefly reported. 

a. Methods.—The following methods were used :—The 
constant current used was obtained from the general lighting 
circuit of the University, and reduced to the proper intensity 
by interposed resistance. This apparatus for getting the 
current I have described fully elsewhere (: 00, : 01), so that 
it need not detain us here. In the circuit a rheostat was 
inserted for regulating the strength of the current. Ordinary 
unpolarisable brush electrodes were used. The specimens 
were placed either ina trough with clay ends, to serve as 
poles, and with paraffin sides of 5mm. depth, or else ona 
slide under a cover supported by several layers of moistened 
filter-paper. These filter-paper ends then serve as the poles 
of the preparation, the brushes of the electrodes being Jaid 
upon them. The layer of water m which the specimens were 
in this sort of a preparation was approximately 2°5 mm. in 
thickness. Identical results were obtained by both the trough 
and the filter-paper method, but since the latter is the neater 
and generally more satisfactory method, it was used almost 
entirely in preference to the trough. 

b. Results.—The typical result of the action of the 
current on specimens in such a position that the long axis of 

VOL. 46, PART 4.—NEW SERIES. a 


686 RAYMOND PEARL. 


the body is approximately at right angles to the direction of 
the current, may be described first. If a number of speci- 
mens are gliding about at the normal rate, and a current of 
from weak to medium intensity is made through the prepara- 
tion, the first reaction of all the specimens is to stop their 
forward motion, turn towards the kathode, and start crawling 
very slowly towards this pole. The orientation towards the 
kathode is at the first trial usually rather precise. The whole 
animal gets squarely into line with the current and moves 
slowly towards this pole. While the current is acting the 
anode end of the body, in this casé the posterior end, remains 


BeBe : 


i + 
a 


Fie. 42.—Diagram showing the typical electrotactic reaction of Planaria. 
a. Position at the moment of making the current, 4, ¢c, and d. Sue- 
cessive phases of the reaction, 


rather strongly contracted, presenting the same appearance 
as when mechanically stimulated. Movement occurs only for 
a short time after the current begins acting. ‘The worm soon 
comes to rest, and further stimulation serves only to cause 
contraction of various parts of the body without producing 
any progressive movement. ‘The successive stages of the first 
typical reaction to the constant current are shown in Fig. 42. 

In succeeding experiments on a given individual, and in 
many cases with the very first experiment, the reaction is 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 687 


much less pronounced. ‘The animal in the transverse position, 
at the moment of making the current, will simply turn the 
anterior part of the body somewhat toward the kathode and 
then stop. Reversal of the current causes the head to swing 
a short distance towards the new kathode, and then stop 
again. The orientation becomes less and less precise the 
longer the current acts. The position most frequently taken 
by a specimen after it has been submitted to the action of the 
current for a short time is shown in Fig. 45, where it is seen 
that the orientation of even the anterior end is not very 
precise. In all such cases the specimen remains perfectly 
quiet after the first turn towards the kathode until the 
current is reversed or broken. 

The behaviour described is that which is typical for currents 
of medium to fairly weak intensities. With very weak 
currents no striking effect is produced. With a current 


s Ye 


Fig. 43.—Diagram showing partial orientation of Planaria to the 
constant current. 


which is just stong enough to cause a general movement of 
Paramcecium towards the kathode, the only effect on a 
planarian gliding at right angles to the current is to cause in 
some cases a very slight turn of the head towards the kathode 
at the moment of making. The specimen does not stop the 
eliding movement, and is not forced into any orientation, but 
may give a she¢ht turning reaction, which changes its course 
from one squarely at right angles to the current to one 
turned a little diagonally towards the kathode. In many 
cases such a current produces no effect whatever. With very 
strong currents the planarian stops at the moment of making, 
jerks the anterior end around towards the kathode more or 
less, and then curls up into the form shown in Fig. 44, as a 
result of very strong contraction of the ventral longitudinal 


688 RAYMOND PEARI.. 


musculature, and dies. I have never been able to produce 
disintegration on the anode side with any current strength at 
my disposal except in a single case, where disintegration 
began in the region just behind the pharynx in a specimen 
strongly curled up in the way described. 

In case the long axis of the planarian is parallel with the 


Fic. 44.—Diagrammatic side view of a planarian subjected to the action 
of a very strong constant current. 

direction of the current, and the head is towards the kathode 
at the moment of making, with a perfectly fresh specimen the 
effect is to cause a cessation of the gliding movement and a 
change to a very slow crawling. The direction of the move- 
ment is not changed. There is a well-marked contraction of 
the anode (posterior) end of the body. The reaction of the 
animal in this position is shown in Fig. 45, b. Very weak 
currents have either no effect on a specimen in this position 
or else may cause a very slight contraction of the ventral 
longitudinal fibres mentioned above. 

When the long axis of the body is parallel to the direction 
of the current, and the head is towards the anode at the 


“KC ea 


Fic. 45.—Dmgram on the electrotactic reaction of Planaria when 
the long axis of the body i is in line with the current direction, and 
the head is towards the kathode. Contracted portions are indicated 
by heavy lines. 

moment of making, the effect of a current of medium intensity 
is to cause the gliding movement to stop. At the same time 
there is a very definite contraction of the anode (head) end 
of the body. As the current continues to act the specimen 
begins to squirm about, and very soon gets out of line with 
the current. Then the anterior end is turned towards the 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 689 


kathode slowly, and this process may be continued until com- 
plete reversal is brought about and the animal comes to lie 
again in line with the current, but with the anterior end now 
directed towards the kathode. This reversal into the usual 
orientation is the typical reaction for fresh specimens at the 
first trials of the current ; it is shown in Fig. 46. Jn case the 
specimens have been under the action of the current for some 
time, there is no reversal of the position. The specimen 


Se pace) no a ae 


‘ £ 
Vic. 46.—Diagram showing the electrotactic reaction of Planaria when 
the long axis of the body is in line with the current direction, and 
the head is towards the anode. Contracted portions are indicated 
by heavy lines. 
simply remains in the same position and contracts strongly at 
the anode (head) end of the body. 
’ u . . 
Strong currents have the same effect as described in the 
preceding experiment. Very weak currents either have no 
definite effort, or else cause a shght jerking back of the head, 
and turning a little to one side at the moment of making. 
After the animals have become partially paralysed by the 


690 RAYMOND PEARL. 


action of the current, the nature of the contractions and 
relaxations of different parts of the body can be very clearly 
seen, and since these are the most significant features of the 
animal’s reactions to the electric current, they may be de- 
scribed a little more fully. These reactions for the three 
chief positions are shown in Fig. 47. The essential features 
are contraction of the anode end of the body when in lne 
with the current, and convexity on the anode side when at 
right angles. Besides this there seems to be some shght 
expansion at the kathode end of specimens in line with the 
current, but this appearance is not constant. Reversal of the 
current in these paralysed specimens causes contraction at 


= Gf Ree) lia 


4 


Fic, 47.—Diagram showing the contractions caused by the current with 
the body in the three principal positions. 
the new anode end or bending towards the new kathode. On 
breaking the current the contracted portions relax. 

c. Mechanism of the Reactions.—It will be seen from 
the figures, and the account which has been given of the 
responses to the electric current, that there is an apparent 
anomaly in the behaviour. The specimen contracts always 
at the anode end of the body, but apparently not on the 
anode side of the body. I believe that the explanation for 
this apparent difference in behaviour is to be found in the 
structure of the animal, and in a peculiarity in the action of 
the constant current which has been noted in another case. 
When the animal is in line with the current the contraction 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 691 


observed at the anode end is, as shown by the form taken by 
the part reacting, a contraction of the longitudinal muscle- 
fibres, while the fibres of the circular and transverse system 
are relaxed. In other words, the current only affects those 
fibres which bear a definite orientation with relation to direc- 
tion of its flow, viz. those which are parallel with it. Now it 
has been shown in an earlier part of this paper that in the 
ordinary negative reaction the turning away from the stimulus 
is produced by a contraction of the circular, transverse, and 
dorso-ventral fibres (principally the circular) on the side 
stimulated. Kvidently when the animal is at right angles to 
the direction of the flow of the current the only muscle-fibres 
in the body whose longitudinal axes are in line with the 
current are the fibres of the circular and transverse systems. 
Unless it is assumed that the current acts differently in one 
case from in another there is no apparent reason why, when 
the animal is in the transverse position, the fibres which are 
in line with the direction of the current on the anode side of 
the body should not contract. If the fibres fulfilling these 
conditions as to location and orientation (the circular system) 
do contract, they will cause the anterior end to be turned 
towards the kathode and the anode side to become convex,— 
in other words, produce the actually observed result. The 
fibres of the longitudinal system should not be affected, and 
there is no evidence that they are. This explanation assumes 
that the current produces its effect by directly causing the 
contraction of properly oriented muscle-fibres, possibly, or 
even probably, without relation to the stimulation of any 
sense-organs of the animal. Or, to put it in another way, the 
responses according to this view might not necessarily be 
reactions of the organism at all, im the sense of being some- 
thing that the animal does after receiving and transforming 
a stimulus, but are direct effects of the stimulus acting on the 
motor organs. It has doubtless occurred to the reader that 
another explanation is possible for these reactions, namely, 
that they are in no way essentially different from what would 
be produced if the animal were given strong mechanical 


692 RAYMOND PEARD. 


stimuli on those parts of the body which are nearest the anode 
in the several positions. In other words, the constant current, 
from the standpomt of the planarian, produces the same 
effect on the anode side or end of the body that a strong 
mechanical stimulus applied in the same place would. 

Which of these two views is the correct one the planarian 
does not show clearly. Yet there is some inferential evidence 
which makes it seem probable that the first view as to the 
sause of the reaction is the correct one, viz. that the current 
produces direct contractions of muscle-fibres oriented in line 
with its direction. The evidence for this view is as follows :— 
(1) In the case of specimens which have been for some time 
under the action of the current, and are, as has been 
mentioned, almost completely paralysed, the essential features 
of contraction on the anode side or end can still be pro- 
duced by a fairly weak current. At the same time it takes a 
very strong mechanical stimulus to get any reaction from 
these pieces, indicating that their sense-organs are almost 
completely paralysed, and their general sensitivity gone. If 
the current acts merely as a stimulus qualitatively lke 
others which produce the same reactions, it 1s not apparent 
why it should be effective in weak intensities when another 
stimulus fails in strong intensity. If it acts directly on the 
muscles we should expect that it would be capable of pro- 
ducing an effect after the general sensory functions had been 
lost. (2) The contractions produced by the current are 
sharply localised, 1. e. they involve only a certain definite 
part of the body whether the current is strong or weak 
(within certain limits) ; whereas mechanical stimuli appled 
to the same places with an intensity sufficient to cause the 
sane definitive reaction will also cause a marked general 
response of the whole organism. ‘This is just what would be 
expected if the current affects only the muscles oriented in 
line with it and lying at the anode pole of the worm. (3) 
By analogy with other forms—for example, the Protozoa—it 
would be expected that the current would produce some 
ether effect than that of an ordinary stimulus applied at the 


MOVEMENTS, ELC., OF FRESH-WATER PLANARIANS. 693 


same point. In the case of the Infusoria the current causes 
an entirely different reaction from that produced by any 
other known stimulus. 

For these reasons, then, I am inclined to think that in the 
vase of the flat-worm the current affects certain definitely 
oriented muscle-fibres directly, and by this means produces in 
the main the characteristic reactions. That the current does 
not also stimulate the sense-organs, and so act hke other 
stimuli apphed to the same places, I am not prepared to say, 
but it seems probable that the phenomena observed are not 
primarily caused by such action. 

It has been brought out by inference that the cila play 
no part in the electrotactic reaction of planarians. This is 
the true state of the case. The current in any intensity 
sufficient to cause the definite reactions stops immediately, so 
far as I have been able to observe, all ciliary movement. The 
evidence for this is twofold. First, all gliding movement 
stops in effective currents; and second, by direct observation 
of specimens crawling ventral side up on the surface film no 
ciliary currents can be observed while the electric current 
acts. This result is of interest in connection with the 
reactions of the rhabdocele Stenostoma leucops, O. 
Schm. This form, which normally moves freely through the 
water by the activity of its cilia, reacts to the electric current 
in essentially the same way as do the Infusoria (cf. Pearl, : 00). 
That is to say, the cilia on the kathode half of the body take 
a reverse position when the current is made, and their effective 
stroke is towards the anterior end. The different relations of 
the cilia in different positions of the body are shown in 
Fig. 48. This relation of the cilary beat, coupled with the 
form of the body, causes, as a mechanical necessity (cf. 
Ludloff, 795), the animal to orient with the anterior end 
towards the kathode. This method of reaction of Stenostoma 
I worked out by precisely the same methods as I used in a 
previous study of the electrotaxis of the Infusoria (: 00). 
This reversal of the position of the cia as a result of the 
action of the current has hitherto been observed only in 


694. RAYMOND PEARL. 


the Infusoria, and to find the same thing in a multicellular 
organism is a matter of considerable interest. It is out- 
side the scope of the present paper to discuss the relation 
of this result to current theories of electrotaxis, as I hope 
to be able to do in a later paper, but it may be said that 
this furnishes another strong piece of evidence that in the 
case of these lower organisms the current does not cause the 
observed reactions in any way comparable to that in which a 
mechanical stimulus causes a reaction, 1. e. by furnishing a 
certain “sensation.” On the contrary, the current acts as 
a physical force on a structure organised in a certain way. 
Experiments on the electrotactic reaction of cut pieces of 


+ eee hee 
a b 


Fie. 48.—Diagram showing the electrotactic reaction of the rhabdocele, 
Stenostoma leucops, O. Schm. 


Cc 


planarians have been tried in considerable numbers, but with, 
on the whole, unsatisfactory results. Anterior pieces result- 
ing from transverse cuts are the only ones from which I have 
been able to obtain any constant results. Such pieces react 
like the normal animal in every way. Posterior pieces from 
transverse cuts show the contractions on the anode side and 
ends in a slight degree, but there is no constant production 
of orientation. Specimens slit longitudinally in the middle 
line from the posterior end nearly to the head react 
essentially like a normal specimen, although much more 
weakly. I have observed in one case fairly precise orienta- 
tions of such a specimen. From specimens slit longitudinally 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 695 


in the middle line from in front backwards I have never been 
able to obtain any definite results. They simply squirm 
about in an aimless way for a moment when the current is 
made and then become quiet, and remain so while the current 
passes. The direction in which the current is flowing makes 
no difference in their behaviour. All operated specimens 
become very quickly paralysed by the current. 

d. Summary.—tThe constant current very quickly para- 
lyses planarians. Its specific effect is to cause a contraction 
of the anode side or end of the body. This produces in the 
case of fresh specimens a well-defined orientation, with the 
anterior end towards the kathode. All progressive move- 
ment after the making of an effective current is by the crawl- 
ing method, the cilia being stopped or very greatly slowed 
in their beat. The electrotactic reaction, so far as the attain- 
ment of orientation is concerned, is essentially the same as 
the negative reaction to mechanical stimuli. In the rhab- 
doceele Stenostoma leucops there is found to occur a 
reversal of the cilia on the kathode half of the body, such as 
occurs in the case of the Infusoria. 


V. Reaction to Desiccation. 


A series of experiments was performed to determine the 
reactions of the animal on drying. This is an environmental 
condition which planarians probably have had to meet with 
relative frequency in the course of their history as a species, 
and it is a matter of interest to determine whether they have 
any method of reacting which protects them from it. 

Experiments were first performed in the following 
manner :—Specimens were taken from the aquarium dish on 
the point of a scalpel or a spatula, and lightly touched to a 
filter-paper for a moment to remove any adherent water, and 
then laid upon a dry surface—either glass or paper. The 
behaviour was usually as follows :—The worm would curl up 
closely and thrust the head under the body, as shown in 
Fig. 49. The purpose of their behaviour seems to be to get 
the body into as small space as possible, and especially to keep 


696 RAYMOND PEARL. 


the head from drying. At fairly frequent intervals the 
animal straightens out and extends the head in front as far 


> movements. It is then 


as possible, and makes ‘feeling’ 
withdrawn, and the animal curls up again. After the drying 
has proceeded for some time the most characteristic feature 
of the whole reaction appears. This is a lengthening of the 
posterior part of the body to its fullest extent. The posterior 
end then attaches itself to the surface, and strong waves of 
contraction, like those in the crawling movement, pass over 
the body from the posterior end forward. No progressive 
movement is made, but backward crawling is evidently 
attempted, and is only prevented by the dry surface which 
the animalis on. There may be considerable variation in the 
first part of the reaction with regard to the curling up; this 
may appear or may not, but the attempted backward crawl- 


Fic. 49.—Diagram showing the reaction of Planar ia to desiccation. 


ine movement of the posterior part of the body I have found 
to be a constant feature in the experiments which I have 
performed. When the dorsal surface of the worm becomes 
dry all movement ceases. If quickly put back into the water 
the worm will usually recover completely, even though all 
movement has ceased in the air. 

If the worm is put on a slide in the centre of a small area 
which has been wet, but on which there is no standing water, 
it will squirm about and extend the head frequently, as in 
the last experiment. If the head goes outside the wet area 
it is very quickly jerked back, and the specimen gives the 
negative reaction, i.e. turns away from the side stimulated. 
The attempted backward crawling occurs in this case just as 
in the others, a short time before the dorsal surface dries off. 

It is to be noted that there is never any actual progressive 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 697 


movement of a specimen in the air. If a specimen is placed 
on very wet filter-paper it is not able to progress unless 
water 1s kept constantly dropping on it from above, so that it 
is at any time surrounded by a layer of water. On account 
of this lack of ability to move when out of water, there is no 
true hydrotaxis in the sense of movement towards water. 

As has been mentioned before, specimens placed on a dry 
surface dorsal side down do not show the righting reaction. 

To sum up, it is found that planarians, when removed 
from the water and subjected to a process of drying, are 
unable to make progressive movements. At a certain stage 
in the drying process they attempt to crawl backwards—a 
form of movement which, under certain circumstances, might 
get the animal back into water. On meeting a dry surface 
with the anterior end the animals give a well-marked negative 
reaction, The animal does not give the righting reaction on 
being inverted on a dry surface. 

On the whole, the general behaviour when subjected to 
drying is purposeful; that is, it would tend to prevent the 
animal ever becoming dried up under natural conditions. 
There is nothing im the behaviour of planarians to indicate 
how the change from aquatic to terrestrial life could be 
brought about. The fresh-water Triclads, so far as I have 
observed them, never leave the water and crawl up into the 
air above the surface film as some other forms do. 


VI. Rheotaxis. 


A large number of experiments were performed early in 
the course of the work with various sorts of devices to deter- 
mine whether the animal showed any distinct reaction to 
currents in the water, but without success. Streams of water 
from a pipette, currents made by filling the tube of the 
diffusion apparatus described above (pp. 661, 662) with water 
and blowing into it, and other methods gave no results. If the 
currents were made with sufficient force to threaten dislode- 
ment of the animal from its hold on the bottom it would stop 
moving and contract longitudinally, and thus attach itself 


698 RAYMOND PEARL. 


more firmly to the substrate. Weaker currents caused no 
effect whatever. I was inclined to believe that the longitu- 
dinal contraction and the gripping of the bottom were the 
only rheotactic reactions which the organism exhibited. It 
was found later, however, that there was a very precise 
rheotactic reaction of a different character. In the course 
of the experiments on reactions to localised chemical stimuli 
by the capillary tube method, it was discovered that by using 
a tube with a relatively large opening (from 4 to } mm, in— 
diameter) and lettmg the ordinary tap-water in which the 
animals were flow out of it, by its own weight, a current of 
just the right intensity to cause a positive reaction could be 
produced. The animals would turn very sharply towards the 
source of such a current, the reaction being evidently the 
same as that given to other weak stimuli (chemical and 
mechanical). This reaction is localised m the same way as 
the usual positive reaction. It is given only when the current 
is directed against the head or anterior part of the body. 

It is thus seen that the planarian is positively rheotactic to 
very weak currents, the form of the reaction being precisely 
the same as that given to other weak stimul. It seems very 
doubtful if this reaction is of any importance in the normal 
activity of the animal. 


G. GENERAL SUMMARY AND Discussion oF ReEsvUts. 


As was stated earlier in the paper, the problem with which 
this study deals is the analysis of the behaviour of the 
common fresh-water planarian. The movements and reac- 
tions to all the more important stimuli, with the exception of 
light and heat, have been described and analysed into their 
component factors in the body of the paper. It is believed 
that it is of the greatest importance to have as complete and 
detailed an account of the various activities as possible, and 
as a consequence full details have been given in the case of 
each subject treated. Since this method of treatment 
necessarily makes the account of considerable length, it has 
a tendency to obscure the general and significant results in a 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 699 


mass of detail. It is desirable, then, to state clearly at the 
end the important general facts which have been brought out 
by this study, and to discuss to some extent their significance. 
In this place I shall state the results in a categorical manner, 
making no attempt to indicate the evidence on which the 
conclusions are based. This will avoid needless repetition. 

1. The locomotor movements of Planaria are of two sorts, 
eliding and crawling. The gliding movement is produced 
by the beating of the cia on the ventral surface of the 
organism. It is by far the most usual method of locomotion. 
For its production it is necessary that there be a layer of 
sticky, mucous slime between the ventral surface of the body 
and the substrate. In this slimy secretion the cilia beat and 
so propel the animal (cf. pp. 544 and 545). The organism 
never moves freely through the water without some sort of 
mechanical support. The rate of the gliding is changed by 
the action of various agents, such as light, chemicals, elec- 
tricity, etc. Its direction is always forward. 

The crawling movement is produced by strong longitudinal 
waves of muscular contraction passing over the body from 
the anterior to the posterior end. It is more rapid in rate 
than the gliding. It appears only after strong stimulation 
of the organism, and its purpose is evidently to get the 
animal quickly away from harmful stimuli. Its direction 
may be either forward or backward. 

Periods of movement alternate with periods of rest in the 
course of the animal’s daily activity. When at rest the flat- 
worm is in a condition of relaxation and generally lowered 
tonus, corresponding to the condition of a higher organism 
in sleep. The causes which induce the coming to rest are— 
(a) a more or less fatigued condition of the organism. This 
is the primary cause; without it the other causes are ineffec- 
tive. (b) A relatively low intensity of hght. (¢) Roughness of 
the substrate. This brings the body into a position such that 
its different parts form angles with one another, and causes 
the animal to come to rest as the result of a reaction which 
I have called goniotaxis (p. 562). (d) Certain chemical con- 


700 RAYMOND PEARL. 


ditions. Asa result of the action of some one or all of these 
above-mentioned factors, collections or groups of planarians 
are frequently formed. 

Planarians which have been injured by operative procedure 
move comparatively little during the course of regeneration, 
thus showing a sort of regulation or correlation between 
behaviour and morphogenetic processes (pp. 573, 574). 

2. There are two principal qualitatively different reactions 
to stimuli, the positive and negative reactions. 

The negative reaction is given in response to strong 
unilateral stimulation of the anterior portion of the body. 
It consists essentially in a turning of the head away from 
the side stimulated. It is brought about by the extension of 
the body on the side stimulated. This extension is produced 
by a contraetion of the circular, dorso-ventral, and transverse 
systems of muscle-fibres. The purpose of the negative 
reaction is evidently to get the organism away from harmful 
stimull. 

The positive reaction is given only in response to weak uni- 
lateral stimulation of the anterior portion of the body. It is 
essentially a turning of the head towards the source of the 
stimulus. This reaction is one of considerable precision, 
bringing the anterior end into such a position that it points in 
most cases exactly towards the source of the stimulus. The 
turning is brought about by the contraction of the longi- 
tudinal muscle-fibres of the side stimulated. The evident 
purpose of the positive reaction is to get the animal into 
regions of beneficial stimuli. 

3. Whether the negative or the positive reaction shall be 
oiven in response to a particular stimulus depends primarily 
on the intensity of the stimulus, and secondarily on its loca- 
tion. Neither reaction is given unless some part of the body 
in front of the pharyngeal region is stimulated. The negative 
reaction is given only im response to stimuli above a 
certain intensity (strong stimuli). This relation between 
intensity of stimulus and form of reaction holds for both 
mechanical and chemical stimuh. 


4 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 701 


4, The reactions of Planaria to a variety of chemicals, in- 
cluding representatives of several of the most important 
chemical groups, were studied. It was found that to a weak 
solution of any substance, regardless of its chemical composi- 
tion, the organism gave a positive reaction identical with the 
positive reaction to mechanical stimul. To strong solutions 
of the same substances (with a single exception, see p. 657) 
the organisms responded by a negative reaction identical with 
that caused by strong mechanical stimuh. 

Planaria does not orient itself to a diffusing chemical in 
such a way that the longitudinal axis of the body is parallel 
to the lines of diffusing ions. Its reactions to chemicals are 
motor reflexes identical with those to mechanical stimuli. The 
positive reaction is an orienting reaction in the sense that it 
directs the anterior end of the body towards the source of the 
stimulus with considerable precision, but it does not bring 
about an orientation of the sort defined above. 

5. Several important features in the normal behaviour of 
the flat-worm are found upon analysis to have their explana- 
tion in the positive and negative reactions to mechanical and 
chemical stimuli. 

The method by which the organism gets its food is simply 
a special case of the positive reaction. From substances 
which serve as food for the planarians, various juices diffuse 
into the surrounding water. When the planarian meets any 
of these diffusing substances it gives the positive reaction,— 
that is, turns in the direction from which the stimulus comes. 
The food substance acts as a weak chemical stimulus, to 
which the animal reacts in the same way as to all other weak 
chemicals. 

The direction of the planarian’s movement, and its behaviour 
with reference to obstacles in its path, are usually deter- 
mined by its reactions to mechanical stimuli. 

The behaviour of the organism with reference to the 
surface film is determined by its reactions to mechanical 
stimuli. 

6. Strong stimulation—either mechanical or chemical 


of 


vou. 46, parr 4.—NEW SERIES. ZZ 


702 RAYMOND PEARL. 


the posterior portions of the body induces the crawling move- 
ment. This is to be regarded as the specific reaction of this 
portion of the body. Weak stimulation of the same region 
causes local contraction at the point stimulated in the case of 
mechanical stimuli, while weak chemical stimuli apphed to 
this region are ineffective. 

7. The ventral surface of the body of Planaria is strongly 
positively thigmotactic, and the dorsal surface is negatively 
thigmotactic. 

8. When the organism is placed in an inverted position it 
performs the righting reaction. This reaction consists im a 
turning of successive parts of the body about the longitudinal 
axis through 180°. During the process the animal takes the 
form of a spiral. The anterior end is brought into the up- 
right position first. On analysis the righting reaction is 
found to be a special case of the reaction to strong stimuli 
(the negative reaction). It is brought about by an extension 
of one side of the body, while the other side maintains its 
original length (pp. 676—679). The reaction is given when- 
ever the ventral surface is removed froma solid or the surface 
film of the water. 

9. To the constant electric current Planaria reacts by 
turning the anterior end towards the kathode. Complete 
orientation and movement towards the kathode may occur. 
The turning towards the kathode is brought about by an 
extension of the anode side of the body. The current causes 
a contraction of muscular elements whose long axes are 
parallel to the direction of the current (pp. 690—693). The 
current very quickly paralyses planarians on which it acts. 

The rhabdocele Stenostoma leucops orients to the 
current with the anterior end towards the kathode, and 
moves towards this pole. This orientation is brought about 
by changes in the positions and consequent effective beat of 
the cilia, exactly like those which occur in the case of the 
ciliate Infusoria. Cilia, on the portions of the body directed 
towards the kathode pole, take on reversed positions. 

10. All the normal reactions to stimuli are of the nature of 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS, 705 


reflexes, more or less complex. What the animal will do 
after a given stimulus, or in a given situation, can be predicted 
with reasonable certainty. There is, however, some variation 
in the behaviour, depending on the physiological or tonic 
condition of the individual at the time of stimulation. Thus 
a stimulus sufficiently weak to induce the positive reaction in 
one specimen may cause the negative reaction in another; or 
at different times the same individual may show different re- 
actions—either the positive or negative—to the same stimulus. 

11. Psychological Position of Planaria. 
tive psychological position of any organism is evidently deter- 


The objec- 


mined by the relative simplicity or complexity of what it 
does. With a view of determining what the position of 
Planaria in the psychological scale is, it may be well to 
make a catalogue of the things which it does in the course 
of its ordinary existence. 

The animal performs the following acts : 

a. It moves progressively by two methods, a ciliary motion 
and a muscular motion. 

b. It turns, by a complex of simple reflex acts, towards all 
weak stimuli investigated. 

c. It turns, by another set of simple reflex acts, away from 
_all strong stimuli investigated. 

‘d. It comes to rest in certain definite environmental 
situations. 

e. When stimulated in a certain way it extends the pharynx 
and feeds. 

f. When its ventral surface is removed from contact with 
a solid body (or the surface film), a reflex of essentially the 
same character as that of ¢ brings this surface again into 
contact with the solid. 

From these essential factors is composed a behaviour 
whose complexity one has only to study to realise. 

The behaviour is thus seen to be, in the main, what may 
be characterised as reflex. It is very simple to say that an 
animal’s activity is composed of a series of invariable reflex 
acts in response to stimuli, but I doubt whether the full 


704 RAYMOND PEARI. 


significance of such a condition is always realised. It implies 
that the animal as an individual “ does” nothing in the sense 
that a man “does” things. It is moved about from place to 
place by its locomotor organs; it is put into certain definite 
and invariable relations to its surroundings by its reflex 
mechanisms. Considered as a whole, such an organism is a 
sort of shell to hold a series of mechanisms, each of which is 
independently capable of doing a certain thing, and in the 
doing produces some effect on the shell as a whole. We may 
perhaps get a clearer picture of what such a reflex existence 
means by considering for a moment what would be the effect 
if all a man’s activities were composed of invariable reflexes, 
to be set off by the appropriate stimuli. Under such circum- 
stances, whenever a man saw or smelled food he would have 
to go to it and eat it. Whenever anything touched him he 
would have to move in a new direction very closely related 
to the position of the object which touched him. Whenever 
he touched water he would have to take a bath, or perhaps 
drink till he could hold no more. During the day he would 
have to move always in a definite direction with reference to 
the sun, and so on ad infinitum. All he did would be 
definitely fixed and, in a sense, predetermined by the things 
about him. 

It is apparent that the behaviour of Planaria is not thus 
entirely and purely reflex, because there is a certain amount 
of variation in it. As has been brought out im several places 
in the body of the paper, and in paragraph 10 of these 
conclusions, this variation in the behaviour is the result of 
the physiological condition of the individual. ‘To put this in 
a more concrete form, we may say that a fatigued animal 
or an animal in a state of great excitation does not always 
react to a certain stimulus by the same set of reflexes as that 
by which a normal animal would react. Furthermore, there 
is a variation in the intensity of the negative reaction 
dependent upon the intensity of the stimulus producing it. 

Another point in which the reactions of Planaria differ 
from what would obtain in the case of an organism whose 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 705 


behaviour was composed of invariable reflexes is found in the 
behaviour following repeated strong stimuh applied to the 
anterior end (vide pp. 580, 581). In this case the organism 
shows an evident modifiability in reaction, for after giving 
for some time the ordinary negative reaction, and not thereby 
getting away from the stimulus, it finally turns directly 
towards the source of the stimulus. Again, in the righting 
reactions of pieces of the body we see entirely new forms of 
reaction appearing (pp. 680—683). 

In order to give a concrete idea of the psychological 
position of Planaria it may be well to present in parallel 
columns the principal factors which make for simplicity in 
the behaviour on the one hand, and for complexity on the 
other hand. 


Factors which tend to make Factors which tend to make 


the Behaviour Simple. the Behaviour Complex. 
A. Mssential reflex character at the A’. Comparatively large number of 
basis of all the reactions. qualitatively different general 
reactions. 
B. General lack of modifiability of B’. Marked qualitatively different re- 
reactions. actions to differing intensities of 
stimulus, 
C. Comparatively small number of C’. Definite relations of reactions to 
qualitatively different reflexes location of stimulus. 
composing the general reac- 
tions. 


D’. Rather close dependence of reac- 
tions on the physiological condi- 
tion of the individual. This brings 
about variation in the reactions. 


The behaviour of Planaria is evidently much more com- 
plex than that of the Infusoria, as described by Jennings 
(loc. cit.). In the case of the Infusoria, all the factors A’, B’, 
C’, D’, which make the behaviour of Planaria so complicated, 
are nearly or quite absent; and in respect to C these organ- 
isms are at a much lower stage than Planaria. The 
Infusoria have practically but one purely reflex reaction to 
nearly all stimuli, and this reaction is not localised with 


706 RAYMOND PEARL. 


reference to the location of the stimulus. Again, the Infu- 
soria do not show qualitatively different reactions to differing 
intensities of stimuli, as does Planaria to a marked degree. 
We thus see that Planaria stands considerably higher in 
the psychological scale than the Infusoria, and that the 
development is taking place along two main lines: (a) the 
higher organism reacts differentially with reference to the 
location and intensity of the stimulus; and (b) the physio- 
logical balance in the higher organism is much more deli- 
cately adjusted than in the lower, and as a consequence we 
see much more variation in the physiological condition. 
These variations in the physiological condition bring about 
variability im the reactions. 

In the case of the ctenophore Mnemiopsis Leidyi we 
have an intermediate stage between the Infusoria and 
Planaria. Here the animal reacts with reference to the 
position, but not the intensity of the stimulus. This condi- 
tion, in which an organism reacts with relation to the position 
of a stimulus, and not to its imtensity, must be for the indi- 
vidual a precarious one, because the animal must either go 
towards or away from all stimuli alike, whether good or 
harmful. Chances are theoretically equal that after each 
stimulus it may get a toothsome morsel of food, or, on the 
contrary, serve in that capacity itself. Further development 
beyond the point in the behaviour series where Planaria 
stands must be in the line of further differential reactions 
with reference to quality of stimulus. A beginning along 
this line is made by the planarian, and the process is carried 
a step farther in the case of Gonionemus, as recently 
described by Yerkes (loc. cit.). 

12. Relation of Behaviour and Structure.—tThe reac- 
tions of organisms are evidently, in any case, very closely de- 
pendent on the structural relations of the given organism, and 
on the conditions under which it lives, i.e. its environment 
in the broadest sense. Thus we find the asymmetrical Infu- 
soria, which live freely in the water and move about by means 
of cilia, all reacting in the same way, and the determinative 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 707 


factor in the reaction is the asymmetry of the body (cf. 
Jennings, :00). Now Jennings has further found! that 
certain rotifers, which live freely in the water and move 
about by the activity of cilia in a similar way, and further- 
more are asymmetrical in fundamentally the same way that 
the Infusoria are, react in essentially the same manner as do 
the Infusoria. Similarly, I believe that the general reactions 
method of the planarians may be found to be in the main the 
method by which all organisms presenting the same general 
structural relations and mode of life react. Only one 
example on which this conviction is based may be given 
here. In the case of such fresh-water molluscs as Physa it 
is apparent that the actual locomotor and sensory organisa- 
tion is symmetrical in form, and furthermore these forms live 
in fresh water on the surface of solid bodies just as do 
planarians. Now I have found, in a series of observations 
not yet published, that in the case of several of these molluscs 
the fundamental scheme of reaction is like that in the 
planarian. They react in the same way with reference to 
the location and intensity of the stimulus, and these are the 
fundamental things. In fact, the general behaviour is 
strikingly alike in the two widely separated groups. 

13. Purposive Character of Reflexes.—A fact which 
is strongly impressed on one working on the behaviour of an 
organism whose activities are largely reflex is the purposive 
character of these reflexes. They are so adjusted that in the 
long run they keep the animal out of danger, and get it into 
favourable conditions. Inthe flat-worm these two things are 
very well done in general by the negative and_ positive 
reactions. Of these two reactions it 1s easy to see that the 
positive is the more highly developed, in particular in the 
fact that it is much more precisely localised with reference to 
the position of the stimulus. We can see a reason for this in 
the fact that under the conditions of the planarian’s life the 

1 Complete observations not yet published. or preliminary account see 


‘Science,’ N.8., vol. xv, pp. 524 and 525 ; and Jennings, : 01, in bibliography 
at the end of this paper. 


708 RAYMOND PEARL. 


eetting of food is of far more importance in the struggle for 
existence than the avoidance of danger. This point has, 
however, been discussed earlier in the paper, and need not 
detain us here. The real problem is presented in the attempt 
to discover how any of the purposive reflex acts in the 
organisms arose. I see no reason for denying that many of 
them—such as, for example, the positive reaction which gets 
the animal its food—were developed by natural selection. 
There are other evidently purposeful reactions, however, with 
whose development it hardly seems as if natural selection 
could have had anything to do, since they cannot themselves 
be of selective value. This pomt has been well brought out 
in a recent paper by Morgan (:02, p. 281). I think a pos- 
sible explanation of some of these may be found in their 
analysis into component factors, when it may appear that 
only a very few simple reflexes had to be formed by natural 
selection, and then all the reactions are built up from these. 
An example will make my meaning clearer. In the righting 
reaction of the planarian we have a fairly complex reaction 
which is evidently immediately purposeful. Yet we find on 
analysis that this reaction is at bottom nothing but a slight 
modification of the ordinary negative reaction, which might 
very well have been developed by natural selection. And 
thus it is with other reactions and pieces of behaviour. They 
are for the most part built up from a very few simple 
purposive reflexes. If we can get them subdivided and 
spread out, as it were, so that we can see what goes to 
compose them, we may find that our problem has diminished 
very much, and we shall have to deal with only a few factors 
where before there appeared to be many. 

A difficult problem in purposeful behaviour presents itself 
when we find that new methods of reaction appear at once if 
the usual reaction is prevented. The best examples of this 
are found in the righting reaction of cut pieces of planarians. 
Here we find pieces of the body, in which the normal 
mechanism of the reaction has been destroyed, immediately 
reaching a certain end (the righting) by a method differing 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 709 


entirely from any that planarians ever used before to attain 
the same end, so far as we have evidence. These phenomena 
have a considerable resemblance to such phenomena as the 
well-known regeneration of the lens from the iris in some 
Amphibia. It is not easy to see how such behaviour comes 
about, and natural selection helps us very little. The matter 
belongs apparently to the same class of phenomena as 
morphological regulations, and probably has ultimately the 
same explanation. What this explanation is we do not know. 

14, Functions of the Nervous System.—The most im- 
portant function of the brain is the preservation of the tonus 
of the organism. After its removal the general tonus rapidly 
diminishes, and on this account the positive reaction—which 


depends rather closely on the physiological condition—can 
be obtained only with great difficulty im such decapitated 
specimens. There is no evidence of the presence of special 
centres in the brain. ‘he nervous system, as a whole, has 
its main function in the rapid conduction of impulses. 

15. Subjective Psychic Attributes.—One of the 
principal questions which forever recurs with regard to work 
on animal behaviour is, does the animal possess conscious- 
ness? Now although it has been shown what the component 
parts of the activities of the planarian are, yet it cannot be 
said, as it seems to me, that the planarian does not, or, on the 
other hand, that it does, possess consciousness. Al] that any 
such an organism ever has done in the past, or ever will do 
in the future, cannot tell us whether it was conscious in the 
doing or not. Any “ objective criterion” of consciousness 
does not exist. Furthermore, whether consciousness is or is 
not present in any given case is not, in any event, the greatest 
concern of the physiologist, who rests content with the objec- 
tive explanation of how results are brought about, regardless 
of what the animal is thinking about the matter. On this 
subject Claparéde (: 01, p. 24), in concluding an interesting 
and valuable discussion, has said, “A la question ; les 
animaux sont-ils conscients ? la physiologie—et méme la 
psychologie en tant que cette science est explicative—doivent 


710 RAYMOND PEARL: 


3 


done répondre non seulement, ‘Je Vignore, mais encore, 


‘“Peu m’importe’!” With this standpoint I am in thorough 


accord. 


H. List or Literature. 


Barpeen, C. KR. :01.—“On the Physiology of the Planaria maculata, 
with Kspecial Reference to the Phenomena of Regeneration,” ‘ Amer. 
Journ. Physiol.,’ vol. v, pp. 1—55. 

BarpeEN, ©. R. : 0), a.—* The Funetion of the Brain in Planaria 
maculata,” ibid., vol. v, pp. 175—179. 

Cricukorr, G. D. ’92.—‘‘ Recherches sur les Dendrocceles d’eau douce 
(‘Triclades),” ‘ Arch. de Biol.,’ t. xii, pp. 435—-568, pls. xv—xx. 

Cuitp, C. M. :01.—“'Tlie Habits and Natural History of Stichostemma,” 
‘Amer. Nat.,’ vol. xxxv, pp. 975—1006. 

CraparepE, BE. :01.—‘* Les Animaux sont-ils Conscients ? ”’ ‘ Rev. Philos.,’ 
t. li, pp. 481—498. 

Duyner, J. van. °96.—‘* Ueber die Heteromorphose bei Planarien,” ‘ Arch. 
f. d. ges. Physiol.,’ Bd. Ixiv, pp. 569—574, pl. x. 

Franpsen, P. :91.—‘‘ Studies on the Keactions of Limax maximus to 
Directive Stimuli,” ‘Proc. Amer. Acad. Arts and Sci.,’ vol. xxxvii, pp. 
185 —227. 

Gameie, I’, W. °93.—* Contributions to a Knowledge of British Marine 
Turbellaria,’ ‘Quart. Journ. Mier. Sei.,’ vol. xxxiv, pp. 433—528, 
pls. xxxix—xli. 

Hesse, R. ’97.—“ Untersuchungen iiber die Organe der Lichtempfindung 
bei niederen Thieren: 11, Die Augen der Plathelminthen, insonderheit 
der tricladen Turbellarien,” ‘Zeitsehr. f. wiss. Zool.,’ Bd. Ixii, pp. 527— 
582, Taf. xxvii and xxviii. 

Jennines, H.S. °97.—* Studies on Reaction’ to Stimuli in Unieellular 
Organisms: I, Reactions to Chemical, Osmotic, and Mechanical Stimuli 
in the Ciliate Infusoria,” ‘Journ. Physiol.,’ vol. xxi, pp. 259—3822. 

Jennines, H. 8. ’99.—* Studies on Reactions to Stimuli in Unicellular 
Organisms : II, The Mechanism of the Motor Reactions of Paramecium,” 
‘Amer. Journ, Physiol.,’ vol. ii, pp. 311—841. 

Jennines, H,8. 799, a.—* Studies on Reactions to Stimuli in Unicellular 
Organisms: III, Reactions to Localised Stimuli in Spirostomum and 
Stentor,” ‘Amer. Nat.,’ vol. xxxiii, pp. 873—3889. 

Jennines, H.8. 799, d.—*Studies on Reactions to Stimuli in Unicellular 
Organisms: LV, Laws of Chemotaxis in Paramecium,” ‘ Amer. Journ. 
Physiol.,’ vol. ii, pp. 855—379. 


MOVEMENTS, EIC., OF FRESH-WATER PLANARIANS. 711 


Jennines, H.S. °99, e.—The Psychology of a Protozoan,’ ‘Amer. Journ, 
Psychol.,’ vol. x, pp. 503—515. 


Jennines, H. 8. :00.—*‘Studies on Reactions to Stimuli in Unicellular 
Organisms: V, On the Movements and Motor Reflexes of the Flagellata 
and Ciliata,” ‘ Amer. Journ. Physiol.,’ vol. iii, pp. 229—260. 


Jennines, H. 8. :00, ¢a.—‘* The Behaviour of Unicellular Organisms,” 
‘ Woods Holl Lectures for 1899,’ pp. 93—112. 


Jennincs, H. 8. :00, 6.— Studies on Reactions to Stimuli in Unicellular 
Organisms: VI, On the Reactions of Chilomonas to Organic Acids,” 
‘Amer. Journ. Physiol.,’ vol. iil, pp. 897—403. 


Jenninos, H.S. :00, ¢e.—‘* Reactions of Infusoria to Chemicals: a Criti- 
cism,” ‘ Amer. Nat.,’ vol. xxxiv, pp. 259 —265. 


Jennincs, H.8. :01.—*On the Significance of the Spiral Swimming of 
Organisms,” ibid., vol. xxxv, pp. 8369—378. 


Jennines, H.S., anp Moors, i. M. : 02.—* Studies on Reactions to Stimuli 
in Unicellular Organisms: VIII, On the Reactions of Infusoria to 
Carbonic and other Acids, with Especial Reference to the Cause of the 
Gatherings spontaneously formed,” ‘Amer. Journ. Physiol.,’ vol. vi, 
pp. 233—250. 


Jima, J. ’*84,—“ Untersuchungen iiber den Bau und die Entwicklungs- 
geschichte der Siisswasser Dendrocoelen (Tricladen),” ‘ Zeitschr. f. wiss. 
Zool.,’ Bd. xl, pp. 8359—464, Taf. xx—xxiil. 


Jorpan, H. :01.— Die Physiologie der Locomotion bei Aplysia lima- 
cina,” ‘Inaug. Diss. Bonn,’ 1901 (Druck von Oldenbourg, Miinchen), 
p. 51, Taf. i. Also in ‘ Zeitscir. f. Biol,’ Bd. xli, pp. 196—238, Taf. ii. 


KENNEL, J. von. ’88.—‘‘ Untersuchungen an neuen Turbellarien,”’ ‘ Zool. 
Jahrb.,’ Abth. Anat. und Ont., Bd. iii, pp. 447—486, Taf. xviii, xix. 


Lane, A. ’81.—“ Untersuchungen zur vergleichenden Anatomie und His- 
tologie des Nervensystems der Plathelminthen: IV, Das Nervensystem 
der Tricladen,” ‘Mitth. a. d. Zool. Stat. Neapel,’ Bd. ii, pp. 53—76, 
Taf. v, Vi. 

Lane, A. 81, a.—‘ Untersuchungen zur vergleichenden Anatomie und 
Histologie des Nervensystems der Plathelminthen: V, Vergleichende 


Anatomie des Nervensystems der Plathelminthen,’ ‘ Mitth. a. d. Zool. 
Stat. Neapel,’ Bd. iii, pp. 76—95. 


Lane, A. *84.—* Die Polycladen (Seeplanarien) des Golfes von Neapel und 
der angrenzende Meeresabschnitte,” ‘Fauna u. Flora d. Golfes von 
Neapel,’ XI Monographie (Leipzig, Engelmann), pp. x and 688, 
Daf. XxI%. 


712 RAYMOND PEARL. 


Leunert, G. H. *91.—* Beobachtungen an Landplanarien,” ‘ Inaug.-Diss.,’ 
Leipzig (Berlin), pp. 1—47. Also in ‘Arch. f. Naturgesch., 57th 
Jahrg., pp. 806—350. 

Linu, Fr. R. :00.—*Some Notes on Regeneration and Regulation in 
Planarians: I, The Source of Material of New Parts and Limits of 
Size,” ‘ Amer. Nat.,’ vol. xxxiv, pp. 173—177. 

Linum, F. R. :01.—*‘* Notes on Regeneration and Regulation in Planarians,” 
‘Amer. Journ. Physiol.,’ vol. vi, pp. 129—141. 

Logs, J. ’93.—‘ Ueber kiinstliche Unwandlung positiv heliotropischer 
Thiere in negativ heliotropische und umgekehrt,” ‘Arch. f. d. ges. 
Physiol.,’ Bd. liv, pp. $1—107. 

Lorn, J. °94.— Beitrage zur Gehirnphysiologie der Wiirmer,” ibid., 
Bd. Ixi, pp. 247—269. 

Lors, J. :00.—‘ Comparative Physiology of the Brain and Comparative 
Psychology,’ New York (G. P. Putnam’s Sons), 1900, pp. x and 309. 


Logs, J., anD Maxwett, 8.8. °96.—“ Zur Theorie des Galvanotropismus,”’ 
‘Arch. f. d. ges. Physiol.,’ Bd. Ixiiil, pp. 121—144, Taf. 1. 

Luptorr, K. °95.—‘* Untersuchungen tiber den Galvanotropismus,” ‘ Arch. 
f. d. ges. Physiol.,’ Bd. lix, pp. 525—554. 


Mitsuxurt, K. :01.—‘‘ Negative Phototaxis and other Properties of Litto- 
rina as Factors in determining its Habitat,’ ‘Annot. Zool. Japon.,’ 
vol. iv, pp. 1—18. 

Morcan, T. H. °98—‘‘ Experimental Studies of the Regeneration of 
Planaria maculata,” ‘Arch. f. Entwickelungsmechanik,’? Bd. vii, 
pp. 364—397. 


Morean, T. H. +:00.—* Regeneration in Planarians,” ibid., Bd. x, pp 
58—119. 


Morean, T. H. :01.— Growth and Regeneration in Planaria lugubris,” 
ibid., Bd. xiii, pp. 179—212. 


Moreay, I. H. :02.—“* The Reflexes connected with Autotomy in the 
Hermit Crab,” ‘Amer. Journ. Physiol.,’ vol. vi, pp. 278—282. 


Mosexey, H. N. °74.—* On the Anatomy and Histology of the Land 
Planarians of Ceylon; with some Account of their Habits, and a Deserip- 
tion of Two New Species, and with Notes on the Anatomy of some 
Huropean Aquatic Species,” ‘ Phil. Trans. Roy. Soc. London,’ 1874, 
pp. 105—17], pls. x—xv. 


Mosetgy, H., N. ’77.—‘ Notes on the Structure of Several Forms of Land 
Planarians, with a Description of ‘Two New Genera and Several New 
Species, and a List of all Species at Present known,” ‘ Quart. Journ, 
Micr. Sci.,’ N.S., vol. xvii, pp. 273 —292. 


MOVEMENTS, ETC., OF FRESH-WATER PLANARIANS. 718 


Norman, W. W. :00.—* Do the Reactions of the Lower Animals against 
Injury indicate Pain Sensations?” ‘Amer. Journ. Phiysiol.,’ vol. iii, 
pp. 271—284., 

Parker, G. H., anp Burnert, F. L. :00.—‘ The Reactions of Planarians, 
with and without Kyes, to Light,” ibid., vol. iv, pp. 873—385. 


Parren, W. ’96.—* Variations in the Development of Limulus poly- 
phemus,” ‘Journ. Morph.,’ vol. xii, pp. 17—148, pls. ii—x]. 


Pear, R. :00.—** Studies on Electrotaxis: I, On the Reactions of certain 
Infusoria to the Electric Current,” ‘Amer. Journ. Physiol., vol. iv, pp. 
96—123. 

Peart, R. :01—*Studies on the Effects of Electricity on Organisms : 
II, The Reactions of Hydra to the Constant Current,” ibid., vol. v, 
pp. 801—320. 

Preyer, W. ’86.—“ Uber die Bewegungen der Seesterne: ein vergleichend 
physiologisch-psychologische Untersuchung,” Erste Halfte, ‘Mitth. a. 
d. Zool. Stat. zu Neapel,’ Bd. vii, pp. 27—127. 

Preyer, W. ’°87.—‘‘ Uber die Bewegungen der Seesterne: ein vergleichend 
physiologisch-psychologische Untersuchung,” Zweite Hialfte, ibid., 
Bd. vil, pp. 191—233, ‘Taf. vii. 

RaspaiL, X. °93.—“ Note preliminaire sur une Planaire, sp.? ” ‘ Bull. Soc. 
Zool. France,’ t. xvili, pp. 49, 50. 

Scourrietp, D. J. *94.—** Entomostraca and the Surface Film of Water,” 
‘Journ. Linn. Soe. Zool.,’ vol. xxv, pp. 1—19. 


ScourFIELD, D. J. :00.—‘ Note on Scapholeberis mucronata and the 
Surface Film of Water,” ‘Journ. Quekett Micr. Club,’ Series 2, vol. vii, 
pp. 809—312. 

ScourFriELD, D. J. :01.—*‘ Hydra and the Surface Film of Water,” ibid., 
Series 2, vol. vill, pp. 137—142, pl. viii. 

Sterner, J. ?98.—‘ Die Functionen des Centralnervensystems und ihre 
Phylogenese: Dritte Abth.,’ ‘Die wirbellosen Thiere,” Braunschweig 
(Vieweg und Solin), 1898, pp. x and 154, Taf. i. 

Urexkt, J. von. °96.— Ueber Reflexe bei den Seeigeln,” ‘ Zeitschr. f. 
Biol.,’ Bd. xxxiv, pp. 298—3818. 

Urxktt, J. von. 796, a—‘* Vergleichende sinnesphysiologische Unter- 
suchungen: II, Die Schatten als Reiz fiir Centrostephanus longi- 
spinus,” ‘ Zeitsclir. f. Biol.,” Bd. xxxiv, pp. 319—339, Taf. i—iii. 

Urxktt, J. von. *99.—“ Die Physiologie der Pedicellarien,” ibid., Bd. 
xxxvil, pp. 384—403, Taf. iv, v. 

Uexkiitt, J. von. :00.—* Die Physiologic des Seeigelstachels,” ibid., 
Bd. xxxix, pp. 73—112. 


714. RAYMOND PARI. 


Uexkt., J. von. :00, a—* Die Wirkung von Licht und Schatten auf die 
Seeigel,” ibid., Bd. xl, pp. 446—476, Taf. 1. 

Verworn, M. ’89.—‘ Psychophysiologische Protistenstudien,’ Jena (lisher), 
1889, pp. vill and 218, Taf. i—vi. 

Wurman, C, O. ’99.—‘ Animal Behaviour,” ‘ Woods Holl Lectures for 
1898,’ pp. 285—338. 

Wittey, A., 97.—‘‘ On Heteroplana, a New Genus of Planarians,” ‘ Quart. 
Journ. Mier. Sci.’ N.8., vol. x], pp. 203—205. 

Wiuson, C. B. :00.—“The Habits and Karly Development of Cere- 
bratulus lacteus (Verrill): a Contribution to Physiological Mor- 
phology,” ‘Quart. Journ. Micr. Sci.,’ N. 8., vol. xlili, pp. 97—198, 
pls. ix—x1. 

Woonvwortn, W. McM. °9]1.—‘‘ Contributions to the Morphology of the 
Turbellaria: I, On the Structure of Phagocata gracilis, Leidy,” 
‘ Bull. Mus. Comp. Zool.,’ Harvard College, vol. xxi, No. 1, pp. 1—42, 
pls. i—iv. 

Woopwortn, W. McM. °97.—‘ Contributions to the Morphology of the 
Turbellaria: IJ, On some Turbellaria from Llinois,” ibid., vol. xxxi, 
pp. 1—16, pl. i. 

Yerkes, R. M. :02.— A Contribution to the Physiology of the Nervous 
System of the Medusa Gonionemus Murbachii: Part I, The 
Sensory Reactions of Gonionemus,” ‘Amer. Journ. Physiol.,’ vol. vi, 
pp. 434—449. 

ZaAcHARIAs, O. ’88.—‘* Landplanarien auf Pilzen,” ‘ Biol. Centralb.” Bd. 
Vill, pp. 542, 543, 


ON THE DIPLOCHORDA. 715 


On the Diplochorda.' 


IV. On the Central Complex of Cephalodiscus 
dodecalophus, MclI. 


By 
A. T. Masterman, M.A., D.Se., 


Lecturer on Zoology, School of Medicine, Kdinburgh. 


With Plates 32 and 33. 


INTRODUCTION. 


Iv the following description a number of the organs involved 
already possess a plurality of names, arising from the fact 
that various observers have recognised differing homologies, 
and emphasised themin the nomenclature. ‘Thus the “ buecal 


shield” of Cephalodiscus is also commonly termed the 


d d) 


In the case of 
the two canals opening from the cavity of the buccal shield to 


“ epistume,” “oral disc,” and “ pre-oral lobe. 
the exterior there is the same difficulty. From homology with 
Balanoglossus they are termed tle ‘ proboscis pores,” 
although the buccal shield has never been termed the proboscis; 
and they are more than pores, being of the nature of definite 
canals. As this work indicates an even closer structural 
resem blance to Balanoglossus than heretofore recognised, 
it would be well to retain as far as possible the nomenclature 
indicating this relationship; hence the terms collar cavities 
and trunk cavities are retained, whilst the term pre-oral canal 
is used as a synonym of proboscis pore with its attendant 
canal. The term subneural gland is retained for the ‘ noto- 


' Read before the Royal Society of Edinburgh, May, 1901. 


716 A. T. MASTERMAN. 


chord” of Harmer, with its homology to the similarly named 
structure in Balanoglossus. Lastly, the terms pericardial 
sac, glomerulus, and ectodermal pit are adopted for organs 
in Cephalodiscus which seem to be homologous with simi- 
larly named organs in Balanoglossus. 

In the region lying between the buccal shield and the 
collar of Cephalodiscus are several important organs, the 
exact relationships of which have not previously been fully 
determined. This area may be described as the central com- 
plex, a convenient term already applied to the same region in 
Balanoglossus. In this region there are externaily the 
ectodermal pit, the pre-oral pores, and the central nerve-mass ; 
internally are the subneural gland, certain important blood- 
vessels, the pericardial sac, and the mesenteric walls of the 
pre-oral and collar cavities, together with the glomerulus and 
muscular strands. ‘lhe general outlines of the subneural 
gland and the pre-oral and collar cavities have been indicated 
by previous workers (McIntoshs, Harmers) and by myself, 
but a detailed examination by carefully orientated and serial 
sections has brought to hght some interesting new facts. 

We may describe the organs under the following headings : 

1. Kcetoderm and nervous system, ectodermal pit, pre-oral 
pores. 

2. Subneural gland and pharynx. 

3. Pericardial sac, pre-oral cavity, pre-oral canals, glome- 
rulus, collar cavities, and blood-vascular system. 

1. Hetoderm.—Figs. 1—8 all illustrate the condition of 
the ectoderm in this region. Ventrally the ectoderm on the 
buccal shield consists of long narrow epithelial cells with 
numerous unicellular glands, which form a buccal gland as 
described by McIntosh (6). This epithelium is not shown 
here. Dorsally it consists of columnar epithelial cells with 
a very definite cuticle. In the region of the central nerve- 
mass the inner ends of the cells are seen to pass downwards 
as delicate fibres, terminating in peculiar conical ganglion- 
cells (figs. 1 and 2). At their base these ganglion-cells give 
off other delicate fibres running forwards and backwards. 


ON THE DIPLOCHORDA. rally) 


These fibres make up the main nerve-mass lying over the sub- 
neural gland; they are seen in transverse section in figs. 6— 
8, and in longitudinal section in figs. 1—5. Forwards they 
run along the dorsal surface of the buccal shield, and back- 
wards they branch outwards to form the two lateral cords; 
hence, in longitudinal sections such as figs. 1—5, the nerve- 
fibres appear to terminate abruptly backwards against the 
wall of the pharynx, which here is in contact with the dorsal 
ectoderm. The same remark applies to the dorsal blood- 
sinus. 

Immediately in front of the central nerve-mass is seen a 
pit or depression, the ectodermal pit (fig. 1). This pit lies 
exactly over the apex of the subneural gland, and extends 
transversely in a slightly crescentic form. At the outer ends 
of the crescent open the pre-oral pores, which are situated 
at the posterior termination of the pre-oral cavity, at the 
level of the central complex. The ectodermal pit, therefore, 
represents the line of division dorsally between the head or 
buccal shield and the collar, as do the “‘epidermistache” 
of Balanoglossus and the epiblastic pit of Actinotrocha 
(formerly termed the neuropore). 

It is well to notice in dealing with this part that the 
nerve-mass is co-extensive with the collar, as this part is 
considerably narrower in the dorsal region; and further, that 
the subneural gland les entirely in the collar area. At first 
sight one is inclined to suppose, reasoning from prior know- 
ledge of Balanoglossus, that the two collar cavities are 
produced forwards into the pre-oral cavity, but such is not 
the case. The buccal shield is produced backwards ventrally, 
but the subneural gland lies in its primitive position in the 
collar, and is in no way produced into the pre-oral cavity. 
It is a backward ventral extension of the buccal shield 
which makes the subneural gland lie in front of the mouth,— 
not, as in Balanoglossus, a forward median extension of 
the subneural gland into the pre-oral cavity. 

2. The Subneural Gland and Pharynx.—The sub- 
neural gland is an elongated cecal tube or prolongation of 

vot. 46, PART 4,—NEW SERIES. Aas 


718 A. T. MASTERMAN. 


the anterior wall of the pharynx. Its total length is usually 
about ‘14 mm., and its breadth about ‘02 mm. It usually 
has for about four fifths of its length a central lumen, which 
opens posteriorly into the pharynx, and terminates anteriorly 
in a variety of ways. Fig. 1 shows the subneural gland cut 
throughout nearly its whole length. ‘The lumen usually, as in 
this case, contains a rod of glandular secretion of the nature 
of mucus. At its apex the gland is bent dorsally. Throughout 
the greater part of its extent its wall is composed of a simple 
glandular epithelium, but at its distal extremity the cells 
show a chordoid modification. The cells become vacuolated 
and reticular, producing the well-known chordoid structure of 
the “notochord” of Balanoglossus, Actinotrocha, and the 
Vertebrata. ‘his is well seen in figs. 8 and 9. The extent 
of this chordoid modification varies immensely, and it is only 
the largest (oldest) individuals which show such a complete 
chordoid apex as in figs. 8 and 9. This specimen also shows 
a not uncommon feature in the complication of the central 
lumen. In the apex it forks out into two lateral canals as 
well as the median central canal (fig. 7)—a character also 
found in some Enteropneusta. 

The relationships of the subneural gland to the pharynx 
have been already described elsewhere (10), and its connection 
with dorsal pharyngeal and peripharyngeal grooves has been 
demonstrated. In fig. 5 the commencement of the dorsal 
and peripharyngeal grooves is shown with their numerous 
nnicellular glands. The commencement of the pleurochord 
is seen in fig. 5. It is important to notice the relationships 
of the subneural gland to the pre-oral and collar cavities. 
It is bounded laterally throughout its extent, except at the 
apex, by the walls of the two collar cavities, and ventrally by 
the wall of the pre-oral cavity. Above it the two collar 
walls form a median dorsal mesentery (fig. 2), and then 
diverge under the ectoderm to form the dorsal blood-sinus. 
At its distal end or apex the subneural gland reaches just 
beyond the collar walls, and plugs up the mouth of the heart, 
as described below (figs. 1 and 2), 


ON THE DIPLOCHORDA. 719 


3. The Pericardial Sac and Heart, Pre-oral Cavity, 
Pre-oral Canals, Glomerulus, Collar Cavities, and 
Blood-vascular System.—The pericardial sac lies ante- 
riorly to the distal extremity of the subneural gland. 
In most specimens it is nearly square in cross-section, 
but may be compressed at its base as in fig. 6. Roughly 
its cross-section is about ‘05 mm., and its length about 
‘(08 mm. It appears to be a closed sac formed of very 
delicate endothelium; its posterior wall is invaginated to 
form the heart. This inner wall is thickened, and has 
numerous muscular fibres stretching across the cavity of the 
sac to its outer wall (figs. 6 and 2). It is doubtless con- 
tractile, and the shape of the pericardial sac varies greatly 
according to its state of contraction. On its ventral wall 
there is a fairly constant transverse groove (fig. 1). The sac 
lies in the blood-space or cavity between the walls of the 
pre-oral and collar cavities, and its walls do not differ except 
in their extreme delicacy from those of these cavities. In 
transverse sections it is seen that the pericardial sac is bent 
over the apex of the subneural gland dorsally and ventrally 
(figs. 7 and 8). laterally it is bounded by the wall of the 
pre-oral cavity, which is thickened into an epithelial lining of 
the pre-oral canal. In fig. 6 both pre-oral canals are clearly 
seen, and the right pre-oral canal is cut throughout its length 
from the pore at the base of the ectodermal pit to the mner 
opening on the wall of the pericardial sac. The canal is 
lined by a delicate columnar epithelium, apparently ciliated. 
In this connection we may note the statement of Ehlers (2) 
that the “ proboscis canals”? of Cephalodiscus end in blind 
sacs. ‘here can be no doubt whatever that McIntosh and 
Harmer were perfectly correct in stating that they open 
freely into the pre-oral cavity, though in a specimen examined 
as a transparent object the pre-oral canal might appear to 
terminate in the pericardial sac. 

I have elsewhere (7) described the blood-vascular system 
of Cephalodiscus, and we have here to notice that, as 
indicated by Harmer (4), the organ I first took to be the 


720 A. T. MASTERMAN. 


heart now proves to be a pericardial sac, containing the true 
heart! within it. The dorsal sinus can be seen running along 
immediately under the ectoderm and above the dorsal collar 
mesentery (figs. 1—4). Anteriorly it terminates against the 
posterior wall of the pericardial sac (which in a large number 
of specimens is ruptured). Here it is also joined on each 
side by a branchial vessel coming from the branchial plumes 
(fig. 7). Further, the anterior end of the dorsal sinus is 
continued into the cavity of the heart by paired lateral 
canals, the relationships of which are not easy to find nor to 
describe. If we could pull the apex of the subneural gland 
backwards from the mouth of the heart it is clear that the 
dorsal sinus would communicate directly with the heart. In 
the normal condition, however, this wide aperture of the 
heart is almost completely plugged up by the apex of the 
subneural gland. Dorsally and ventrally (figs. 7 and 8) this 
organ rests closely up against the pericardial wall, but 
laterally a small canal remains running downwards from 
dorsal sinus to heart (fig. 4). This canal is bounded pos- 
teriorly and laterally by the wall of the collar cavity, and 
anteriorly by the wall of the pre-oral canal (pre-oral cavity). 
It is doubtless through this paired canal that the blood finds 
its way from the dorsal sinus to the heart. 

Below the subneural gland is a well-defined ventral sinus, 
which passes backwards to the level of the mouth and round 
it on either side. It is wide and large posteriorly, but passes 
forwards, getting narrower and narrower till it is lost in the 
olomerulus (fies. 7 and 8). Ventrally it is bounded by the wall 
of the pre-oral cavity, which also extends ventrally, laterally, 
and anteriorly to the pericardial sac. Various parts of this 
wall (pre-oral cavity) are thrown out into cecal prolongations 
into the cavity, with thickened protoplasmic walls. The 
cavities of these czeca are in direct communication with the 
blood-sinuses. They produce an appearance closely similar 
to that of the glomerulus or pericardial gland of Balano- 
glossus, with the exception that the walls are simple and 


'The “ pre-oral sac’ of my previous work. 


ON THE DIPLOCHORDA. Fea | 


not of a definite cellular structure. There is usually a paired 
patch of this glomerular tissue on the antero-lateral surfaces 
of the pericardial sac (figs. 1 and 5) in close proximity to the 
internal apertures of the pre-oral canals. Further, the wall 
of the ventral sinus shows a similar structure (figs. 3, 4, 
and 7). In many cases the glomerular tissue of the ventral 
sinus is also paired, and the ventral sinus is then almost 
constricted into two paired sinuses. 

There is little doubt that this glomerular tissue is homolo- 
gous with the pericardial gland or glomerulus of Balano- 
glossus. Antero-dorsally to the pericardial sac we may 
notice a pre-oral sinus bringing blood back from the buccal 
shield (fig. 2) to the glomerulus. 

We may now briefly run over the figures given here, noting 
the special points of each. Figs. 1 to 5 are selected from a 
series of very nearly sagittal orientation. In fig. 1 the sub- 
neural gland is cut almost throughout its length, its opening 
into the pharynx being more to the right. The right collar 
cavity is cut just at the apex of the gland, so that the sinus 
is rather more to the right anteriorly than posteriorly. The 
right glomerulus is also seen, whilst the cavity of the heart is 
spacious, although not at its largest (in the median line). 
The dorsal sinus is cut throughout its length, and two oral 
grooves may be recognised. In fig. 2 the collar mesentery is 
cut for some portion of its extent, and the glomerulus of the 
ventral sinus is coming into view; the left dorsal groove is 
also just appearing. In fig. 3 the pericardial sac is cut in 
the median line; the peculiar shape of the heart is noticeable. 
Further back the left collar cavity is alone seen, the dorsal 
sinus is restricted, and the subneural gland is interrupted. 
The left peripharyngeal and dorsal grooves are differentiated. 
In fig. 4 the heart is no longer visible, but the left canal from 
dorsal sinus into heart is seen. The posterior portion is still 
more to the left, showing the grooves as before. Fig. 5 is 
eight sections further to the left. In following the sinus one 
notices the left pre-oral canal becoming gradually more pro- 
minent, first laterally and then dorsally; the left glomerulus 


1422 A. T. MASTERMAN. 


appears, and the left collar cavity increases greatly in size. 
Posteriorly the first trace of the left pleurochord is seen, 
lying laterally to the dorsal groove. In this section the 
ganglion-cells are no longer seen, and the nerve is inter- 
rupted at the pre-oral canal. 

In figs. 6 to 8 we have selected sections from a transverse 
series. The right side of the figures is slightly posterior to 
the left. Thus in fig. 6 the right pre-oral canal (on the left) 
is cut throughout its length, but the left only in part. In 
this section the pericardial sac is cut transversely, and the 
heart is seen in its greatest size. Fig. 7 is a few sections 
further back. Here the apex of the subneural gland is cut 
through, and shows three internal canals and a chordoid 
structure. Dorsally the pericardial sac is still cut, and below 
the ventral part is the glomerulus of the ventral vessel. On 
the right is seen the left branchial sinus leading out from the 
dorsal sinus, and a wider right branchial sinus opposite. The 
two horns of the ectodermal pit are also seen. Fig. 8 is still 
further back. The pericardial sac is no longer seen dorsally, 
but is still cut ventrally. The walls of the two collar cavities 
are approximately in the middle line, and behind the sub- 
neural gland will form a dorsal mesentery. The lateral nerves 
of the post-oral ring are seen in this section. 

In fig. 9 the chief features here described are reproduced 
in a semi-diagrammatic median section of the entire animal. 
I have also shown the pharyngeal structure formerly described, 
i. e. the pleurochords, the dorsal and ventral grooves, and the 
oral grooves. 

The facts above described must inevitably tend to bring 
Cephalodiscus into even closer union with Balanoglossus 
than heretofore. Not only is every organ in the central 
complex of the former to be directly compared to its homo- 
logue in the latter, but the latter has no organ in this region 
which does not occur in the former. The only essential 
difference is one which several years ago appeared to me to 
be of fundamental importance, but which must now be regarded 
as of secondary value. In Balanoglossus the pericardial 


ON THE DIPLOGHORDA. i235 


sac, glomerulus, and subneural gland protrude forwards 
into the pre-oral cavity, and hence are covered dorsally as 
well as ventrally; but in Cephalodiscus they protrude 
upwards between pre-oral cavity and collar cavities, 
and they are therefore dorsal and posterior to the former. 
In this way the pericardial sac lies in contact with the dorsal 
ectoderm, and the subneural gland is only separated there- 
from by the dorsal sinus. ‘This difference cannot be regarded 
as fundamental in view of the anatomical resemblance, and 
we have seen above that it is due to the forward protrusion 
into the pre-oral cavity of the subneural gland in Balano- 
glossus, whereas in Cephalodiscus it remains in the 
collar. 

Of other points the homology of the subneural gland is a 
most important question. It appears desirable to adhere to 
this term, firstly, because it is unquestionably glandular in 
function; secondly, because it has precisely the same re- 
lationship to a system of dorsal and ventral grooves in 
the pharynx as is the case with the similarly-named organ 
in Tunicata (10); and thirdly, because its anatomical position 
is exactly under the main nerve-mass. These and other 
facts led me to doubt its homology with the “ Hicheldarm”’ 
of Balanoglossus, but its relationships to perivardial sac 
and glomerulus and the chordoid structure of its apex 
appear to me to be conclusive in favour of accepting Har- 
mer’s original comparison. I would extend the appellation 
of subneural gland to the organ in Balanoglossus, for, as 
in so many other features, Cephalodiscus would appear to 
show us amore primitive condition of the organ than Balano- 
glossus. In making this comparison it appears to me to be 
questionable how far the subneural gland is at all comparable 
to the notochord of the Vertebrata. As indicated elsewhere 
(11, p. 412), a chordoid histological structure by itself 
cannot be regarded as an absolute criterion of homology, and 
the occurrence of chordoid organs of the same nature as, but 
not homologous with, the Vertebrate notochord is to be 
expected in these low chordates. The view of Willey that 


724 A. T. MASTERMAN. 


Cephalodiscus is to be regarded as a degenerate ally of 
Balanoglossus has not much to commend itself; the conse- 
quent assumption that the former has lost numerous gill- 
slits perforating its anatomy in all directions, not to mention 
numerous other organs, has no justification in fact. We may 
with Lang (5) suppose that Cephalodiscus has undergone 
certain important modifications due to a semi-sedentary habit, 
but the assumption that its proximate ancestors had many 
pharyngeal clefts and gonads has nothing to recommend it 
but its necessity for Willey’s theory. I would prefer to 
regard Cephalodiscus as the more primitive form, as its 
want of metameric segmentation and its primitive method 
of feeding would imply (9). On this basis the “ Hichel- 
darm” of the Enteropneusta must be regarded as a glan- 
dular specialisation of the anterior end of the pharynx, to 
be termed the subneural gland, owing to its functions and 
structural relationships. 

In Cephalodiscus its distal end often exhibits a com- 
mencing degeneration into chordoid tissue (which, by its 
development in Actinotrocha, is clearly an arrested form of 
glandular epithelium), whilst it is still functionally active as 
wu gland. In Balanoglossus, with a specialised burrow- 
ing habit, the original function has been largely lost (though 
the “ Hicheldarm” of Balanoglossus is unquestionably 
glandular), and the chordoid tissue with supporting function 
becomes still more in evidence. The organ to which the 
name of subneural gland was given in Actinotrocha 
occupies exactly the same position as in Cephalodiscus, 
but as it is only embryonic its walls would hardly be expected 
to be of a definitely glandular nature. 

The pericardial sac of Cephalodiscus and its contained 
heart are so similar to the pericardium (Herzblase) and 
heart respectively of Balanoglossus, and so different from 
any structures found elsewhere, that the homology need not 
be insisted upon. In a similar manner the mutual relations 
of the ectodermal pit, the pre-oral canals, the pericardial sac, 
and the surrounding blood-sinuses speak for themselves. 


ON THE DIPLOCHORDA. pa) 


Lastly, there can be little question that we have in the 
glomerulus a homologue of the proboscis gland of Balano- 
glossus. Each is a proliferation of the pre-oral ccelomic 
endothelium in the neighbourhood of the pericardial sac and 
pore canals, consisting of czcal vascular processes. 

It is evident that in the study of the budding processes 
(10) the origin of the pericardial sac must have been over- 
looked ; but as we do not yet know how this organ arises in 
the demersal larva of Balanoglossus, nor even with 
certainty in ‘l'omaria, this is not surprising. From certain 
indications it appears that in Cephalodiscus it is a portion 
of the pre-oral cavity constricted off from its posterior end, 
and therefore ccelomic in origin. 

During the progress of my work on Cephalodiscus Cole 
(1) has published a short paper upon the bulbous termina- 
tions of the twelve branchial plumes. His results appear to 
indicate that the migration of oval lens-like bodies out of the 
epithelial cells to the exterior is to be regarded as a normal 
process, and that McIntosh’s view (6) that these organs are 
masses of unicellular glands is correct. Assuming that the 
migration might be an abnormality, I had suggested that “it 
seems most reasonable to regard them tentatively as primitive 
eyes,” a view I had already abandoned before the unexpected 
appearance of Cole’s work. Cole further finds that the 
glandular bodies break up to form rhabdites, which I think 
quite probable, especially as I had already found and described 
indications of ‘‘ one or more areas iu the centre (of the bodies) 
staining more deeply than the rest (7).” I cannot agree with 
Cole’s description of the epithelium in these terminal knobs 
as normally correct, as such a vacuolated swollen mass with 
little or no cuticle occurs commonly in other parts of the 
body, and seems to be an abnormality ; the vacuolated con- 
dition of the bulbs is undoubtedly present, especially in 
older specimens. Cole denies the existence of a cuticle, of 
pigment, and of nerve-endings in the cells. In respect to the 
cuticle [am hardly prepared to inaugurate a discussion upon 
the line of distinction between a “ peripheral deeply-staining 


726 A. 'l. MASTERMAN. 


membrane and a cuticle. ‘There is lttle doubt that the in- 
tracellular bodies under discussion arise in the young form 
in close contact with this limiting membrane, but it is 
possible that they do not actually arise from it. I am still of 
the opinion that fine brown pigment granules are scattered 
throughout the cells (McIntosh [6] previously remarked upon 
the “deep yellowish tint” of this region) ; and I still believe 
that the inner end of each cell “ tapers to a fibre-like thread, 
which I believe to have in some cases traced into the main 
nerve of the plume” (7, p. 344). Indeed, it is rather difficult 
to understand otherwise in what region the very evident nerve 
down each plume terminates. None of these features are 
opposed to the “battery” function as suggested by Cole, 
though I have not as yet seen the rhabdites, which appear to 
require special staining. If their presence is corroborated 
it would form by no means the least interesting feature of 
Cephalodiscus. Cole, as a critic of the work of his prede- 
cessors, might perhaps make a somewhat more sharp distinc- 
tion between a tentative suggestion and a definite statement 
of fact; but leaving this apart we may regard his work as 
confirming McIntosh’s previous interpretation of the bulbous 
endings as masses of unicellular glands, the glandular se- 
cretion being extended to the exterior through the surface 
of the cells. Further, there is every reason to believe that, 
according to Cole, some at least of the glandular masses 
break up into rods. 


LITERATURE. 


Cotz, F. J. ‘Journ. Linnean Soc.,’ vol. xxvii, No. 175. 
Turers. ‘ Abb. k. Gesellsch. Wissensch. Gottingen,’ Bd. xxxvi (1890). 
Harmer, 8. F.—Appendix to “ Challenger” Report, vol. xx. 
d ‘Zool. Anzeiger,’ No. 545. 
“Take, A.—‘ Jenaische Zeitschrift,’ xxv (1890). 
McIntosu, W. C.—*On Cephalodiscus dodecalophus,” ‘ “ Chal- 
lenger” Report,’ vol. xx. 


DAP OP 


ON THE DIPLOCHORDA. TPA | 


7. Mastermay, A. T.— On the Diplochorda,” part ii, ‘Q. J. M.8.,’ Aug., 


1897. 
8. 7. “On the Notochord of Cephalodiscus,” ‘Zool. 
Auz.,’ No. 545, 1897. 
9. - *©On the Origin of Vertebrate Notochord and Pha- 
ryngeal Clefts,” ‘ Rep. Brit. Assoc.,’ Sept., 1898. 
10. = On Further Anat. and Budding Processes of 


Cephalodiscus,” ‘Trans. Roy. Soc. Edinb.,’ 
vol. xxxix, pt. ill, No. 17. 

15 s “Qn the Diplochorda,” part ii, ‘Q. J. M. S., 
vol. xlili, pt. il. 


DESCRIPTION OF PLATES 32 & 33, 


Illustrating Mr. A. T. Masterman’s paper “On the 
Diplochorda.” 


Fres. 1—5.—Selected sections (1, 3, 5, 7, 15) from a longitudinal sagittal 
series through Cephalodiscus dodecalophus (Zeiss, obj. 7, eyep. 1). 


Figs. 6—8.—Selected sections (1, 4, 6) from a transverse series through 
Cephalodiscus dodecalophus (Zeiss, obj. 7, eyep. 1). 


Fic. 9.—A semi-diagrammatic right half of a polyp. 


‘= 
: : 
7 7 (a PVG y 


‘4 Mid y Ae) 
” 
t aly 4, 
tal ( 1 aq pte 
be May 4 r 
*! iw ® 7 b F 
4 
a 
hi 
. 
2 = 
. 


a 


Pe) 
“a 


“I 
ro 
eo) 


HYPURGON SKEATI. 


On Hypurgon Skeati, a New Genus and Species 
of Compound Ascidians.' 


By 


Igerna B. J. Sollas, B.Sec.Lond., 
Bathurst Student of Newnham College. 


With Plates 34 and 35. 


Amone the marine sponges from the Malay Peninsula 
collected by Mr. R. Evans, of Oxford, at present Curator 
of the Government Museum in Georgetown, Demerara, and 
very kindly handed over to me by Dr. Harmer for description, 
there were included two specimens of the new genus of 
Synascida Didemnida, which I have endeavoured to describe 
below. 

The locality named on the collector’s label in the case of 
each of the two specimens is Pulau Bidang. 

The association of the Tunicate with a sponge was merely 
fortuitous, and due solely to participation in the same surface 
of support. 

The colony forms a thin sheet, little over 1 mm. in thick- 
ness, adherent to the substratum. The colour in spirit is a 
dirty yellowish brown. 

The appearance of the colony when examined by reflected 
light under a low power of a binocular microscope is repre- 


11 take the Greek troupydc, and by lengthening the 6 get irovpydr, 
meaning a place where things are made serviceable. 


730 IGERNA-B. J. SOLLAS. 


sented in Pl. 34, fig. 2. This external view shows at once 
the character to which the generic name alludes, namely, the 
presence in the test of numerous ovoid fecal pellets. These 
are seen through the transparent substance of the test, and 
now appear of an opaque cream-white colour. Clusters of 
calcareous asters (fig. 3) mark out the oral siphons, since 
they make a conspicuous snow-white patch around each 
siphonal aperture. These white spots are visible also with 
the naked eye. 

The arrangement of the ascidiozoids is irregular. A large 
number of them share the same atrium, the atria being 
shallow but extensive cavities with but few and small 
siphons. The siphons are not visible in surface view, but in 
section it is seen that their lps are formed of transparent 
test-tissue destitute of spicules. 

The bulk of the common test, which consists of actual 
tunicin, is small, its substance being excavated by numerous 
oval spaces, in which the fecal pellets he. To reach this 
position the pellets must, after being ejected into the atrium, 
sink through the excessively thin epithehal wall of that 
cavity. The cellular elements in the test are of the usual 
types; bladder-cells are specially abundant near both upper 
and under surfaces, and round the oral siphons. Spicules 
occur in small numbers, chiefly aggregated round the oral 
siphons and in the neighbourhood of the branchial sac. 
They may be isolated or packed in dense clusters (Pl. 34, 
fio. 3). Finally the renal vesicles, described presently, are to 
be reckoned among the structures included in the test. The 
ascidiozoids, as is common among Didemnida, have a sharp 
constriction between the branchial region of the body and 
the abdomen. 

The number of lobes round the oral siphons varies from 
four to six. The tentacles are twenty-four in number ; 
twelve long ones alternate with twelve short. The branchial 
sac has four rows of five stigmata on each side. Connectives 
(Hancock; trabecule, Yves Delage) are absent. The 
dorsal languets are long and median in position. The sub- 


HYPURGON SKEATI. Tar 


neural gland has a simple opening with a swollen lower lip 
(fig. 4, d. ¢.). 

Through the narrow aperture of communication between 
the two regions of the body the cesophagus descends to open 
into the stomach, while the intestine passes upwards into the 
rectum, which lies above the constriction, so that the anal 
opening is close to the base of the branchial chamber. 

The walls of the stomach are raised up round the termina- 
tion of the cesophagus ; or, in other words, the cesophagus 
has its opening deep in the cavity of the stomach; the ter- 
minal part of the cesophagus is richly ciliated. 'The intestine 
of a young bud is frequently found attached at both ends to 
the cesophagus, to which it owes its origin. When this is 
the case the thoracic portion of the same bud is to be seen 
lying in the test at the opposite side of the cesophagus. The 
budding is thus of the type known as pyloric (Giard), 
and found among Didemnidee in the tribe Didemnine (Y. 
Delage). 

The walls of the stomach are smooth; seen en face from 
the outside they show a beautiful reticulum formed of the 
more deeply staining protoplasm which surrounds and con- 
nects the nuclei of the cells of the gastric epithelium. 

The intestine as it leaves the stomach is richly ciliated; in 
passing thence to the anus its walls become continually 
thinner, the walls of the rectum being almost membranous. 
The anus has thickened lips. The alimentary canal is bathed 
by blood-sinuses along its whole course. 

The heart in its pericardium runs more or less vertically 
between the upper and lower walls of the abdominal 
cavity. Its lower end abuts against and sends a large vessel 
into a prominence of the test, the sides of which are covered 
by a patch of specially large cells of the mantle which form 
the glandular part of the renal organ (r. gl., figs. 5 and 7). 
The excreta of these glandular cells appear to be picked up 
by wandering cells—presumably corpuscles of the blood con- 
tained in neighbouring vessels or sinuses. These cells would 
then migrate into the test, carrying their burden with them. 


732 IGERNA B. J. SOLLAS. 


Large numbers of vesicular cells contaming concretions are 
to be found embedded on each of the above-mentioned 
prominences of the test, while in older kidneys there may be 
a relatively enormous rounded mass of such vesicles more 
deeply situated in the test substance (fig. 7, k.). Some such 
masses may be found in the basal layers of the test at a 
distance from the abdominal cavity of any zooid; these have 
evidently been left behind, the zooid to which they belonged 
having shifted upwards as the floor of the cloacal cavity was 
raised by the continual addition of fresh pellets. 

Thus the excretory organs of Hypurgon agree with the 
simple type of excretory organ found in Botryllus, in that 
the urinary concretions are stored in the cavities of single 
vesicular cells; but apart from this particular they are of a 
type unlike any yet described (Dahlgriin, ‘ Archiv fiir mikr. 
Anat.,’ vol. lviii, 1901) among Tunicates, and are far less 
simple than any known in other Synascida. 

The reproductive organs lie in shallow depressions of the 
wall of the abdominal cavity (fig. 9). The testis is oval, 
and the vas deferens makes four or five turns of a spiral 
around it. The ovary has membranous walls, and contains a 
string of eggs of successive ages. I have not seen an oviduct. 

Any mature ova that I have seen have sunk deep into the 
test, and so have come to lie in a great recess of the abdomi- 
nal cavity (fig. 10), communicating with it by a narrow 
aperture. The material contains but one larva, which was 
developing in a completely closed cavity in the test (fig. 11). 
This may or may not be the normal course taken by the 
developing eggs. Hggs are not to be seen being sheltered 
by any other part of the organism than the test, though eggs 
of all ages were found in the ovaries. 

The feecal pellets, which contribute so largely to the 
formation of the test, show a very remarkable degree of 
coherence. If a piece of the colony be boiled in sulphuric 
acid, the residue consists of faecal pellets which retain their 
form perfectly, and continue to do so even if the boiling be 
much prolonged. Even thin sections of pellets, isolated by 


HYPURGON SKEATI. 733 


boiling microtome sections of the colony in sulphuric acid, 
may still be mounted whole after this treatment. Boiling in 
aqua regia and boiling in fuming nitric acid are equally 
ineffectual in disintegrating the pellets; when these latter 
reagents are used the test naturally forms part of the residue, 
since they are not capable of dissolving tunicin. 

When isolated by means of sulphuric acid the pellets have 
a black colour, due to the action of the acid on the organic 
matter contained in them. These blackened pellets may next 
be washed and calcined, and though raised repeatedly to 
cherry heat they still remain intact, and are now opaque 
white when examined by reflected light. Mounted in oil, or 
passed through oil into balsam, they become transparent. 
Calcined pellets dissolve completely in hydrofluoric acid. 
Prolonged boiling in a strong solution (nearly saturated) of 
caustic soda resulted in the dissolution of calcined pellets. 

It seems, then, that the strong coherence of the pellets must 
be due either (1) solely to cohesion and adhesion between the 
foreign particles contained in them, or (2) to a deposition of 
silica between these particles. The siliceous nature of the 
greater part of this foreign matter makes it impossible to 
determine between these two alternatives. It naturally 
suggests itself that this property of coherence of the pellets 
is an adaptation to enable the aninal to utilise waste organic 
matter with impunity. But it must be mentioned that the 
pellets are porous, taking stains readily both before the 
treatment described above, and also at every stage during it. 

It is curious that the pellets are also highly fragile; they 
crumble at once under pressure of the cover-slip. 

Melicerta tubes were boiled in avid for comparison: the 
form of the component pellets was lost almost immediately— 
as soon as the cementing substance between neighbouring 
pellets disappeared. 

A parasitic crustacean was found in one ascidiozoid, 
occupying a large part of its branchial chamber. ‘The body 
of the parasite is a mere sac filled with ova in an advanced 
state of segmentation. There appear to be six pairs of 


von. 46, PART 4.—NEW SERIES. BBB 


734 IGERNA B. J. SOLLAS. 


appendages belonging to the anterior region of the body, 
besides one foremost pair which serves as an organ of 
attachment, and is inserted into the tissues of the host. 

The systematic position and diagnosis of the genus may 
be stated as follows:—Synascida Didemnida Didemnina, 
(Y. Delage). Colony thin; ascidiozoids with four rows of 
branchial shts and twenty-four tentacles; vas deferens 
spirally coiled round the testis ; feecal pellets included in the 
test, in which organ the renal vesicles are likewise contained. 

In conclusion, it gives me much pleasure to take this 
opportunity of expressing my thanks to Mr. Graham, Kerr 
for kind help and advice. 


LITERATURE CONSULTED. 


Herpman.— Voyage of H.M.S. Challenger,’ Tunicata 11. 
Yves Detacre.—‘ Zoologie concrete,’ vill. 
Dantertn, W.—‘ Archiv fiir nat. Mikr.,’ lviii, 1901. 


Granp.—‘ Arch. d. Z. expér.,’ i, 1872. 


EXPLANATION OF PLATES 34 & 35, 


[llustratine Teerna B. J. Sollas’s paper “On Hypurgon 
Skeati, a New Genus and Species of Compound 
Ascidians.” 


as. Caleareous spicule. aé.s. Atrial siphon. éd.c. Blood-corpuscle. d/. e. 
Vesicular cell of test. 02.8. Blood-sinus. d./. Dorsal languet. d. ¢. Dorsal 
tubercle. exd. Mndostyle. ££ Fusiform cell. g. Nerve ganglion. 4. Heart. 
int. Intestine. ¢. Larva. x.éc. Notochord. @. Cisophagus. ov. Ova. p. 
Fecal pellet. p.e. Pericardium. rect. Rectum. 7. Renal organ. 7. gl. 
Glandular cells of renal organ. 7.¢c. Renal concretion. sf. Stomach. ¢. 
Testis. v.app. Vase. appendage. v.d. Vas deferens. 


HYPURGON SKEATI. 735 


PLATE 34. 


Vig. 1.—A piece of a colony of Hypurgon Skeati, slightly larger than 
natural size. 

Fic. 2.—A portion of the surface of a colony seen under a binocular micro- 
scope. X 75. 

Iie. 3.—Caleareous spicules from the test of Hypurgon Skeati. a@and d 
from one colony ; ¢, e, and ffrom a second. /, acluster of spicules. 

Vie. 4.—A vertical section through a part of a colony of Hypurgon 
Skeati, showing the branchial sac and parts of the abdominal cavity of one 
zooid (slightly reconstructed from neighbouring sections). x SO. 

Fig. 5.—Vertical section of an abdominal cavity. 

Fic. 6.—Diagrammatic reconstruction of a slice of acolony of Hypurgon 
Skeati, showing one zooid from the left side and one from the dorsal surface. 
Drawn as though it were transparent. 


PLATE 35. 


Fie. 7.—Section of a renal organ of Hypurgon Skeati which has been 
functioning long enough to form the considerable accumulation of concre- 
tions K. 

Fic. 8.—Portion of the test of H. Skeaticontaining renal vesicles, more 
highly magnified. 

Fic. 9.—Section of an abdominal cavity of zooid of H. Skeati, to show 
reproductive organs. 

Fic. 10.—Section of alarge ovum of H. Skeati ina recess of the abdominal 
cavity. 

Fig. 11.—Section of a tailed larva of H. Skeati developing in a closed 
cavity in the test. 


me .OMAA Ce 


a 8 pitzeg 


gl i eit « oe abs Oe oe Chee OF iaaly 


i 
4 ete Ber Cyien® & iy SG Aes CFO 14 (ad ae 
CF 
ito torn el lh heten pe: xu! coven ede Lg 
- * iritee 4 mi ae ee ii Nisag , ok oR 
i} ’ i : a Lh epee? prea a iy 
a j j + er Od ee 
Ae 7 ; i H i a“ rege _ 


Obl sh i reer iat 
H : : i i vi A yeas ALL! —, i 
“ro ’ a wl 1) da ih @ Udreee Hivay Qn ~All ide 


forts imine 2 loeeS 
-_ 


© 
S @iaire 
t . 4.07 wv" ? ij ' i bi Le re 
7 Pe ee ey pets we iipihe felis ete @ th) baa its 
% ak Tea Bie ine The’ oie’ a Ee ea | Bi 
4 
il iy ieee ti es & mal mad te nls 1 it jew le 


er 
meee ere ae! Ee 


fn nigeend’ lb ged? J ine Var ss 


-_ _— ad n= , a ‘ pe 7 
. © So see oh cn 
=" g > i> iy Se 


: uo .) Pe Gua 2 

Pa oe ee. =. 

— ! — sees ated _-oehl 
‘_ ° aaa 2.0) haa 


~JI 


(Sy) 


ba | 


ANATOMY OF ARENICOLA ASSIMILIS. 


The Anatomy of Arenicola assimilis, Ehlers, 
and of a New Variety of the Species, with 
some Observations on the Post-larval Stages. 


By 
J. H. Ashworth, D.Se., 
Lecturer on Invertebrate Zoology in the University of Ndinburgh. 


With Plates 36 and 37. 


ConTENTS. 

PAGE 

I. Introduction . ; 
II. Arenicola eect Ehlers. : Se 4 
III. Specimens of Arenicola from New Zenon 75 
IV. Systematic Position of A. assimilis and of the Specimens Frey 


New Zealand . 760 

V. Post-larval Stages of Arenicola on the Falkland Telomtte . 164 
VI. Adult Specimens of Arenicola from the Falkland Islands . 768 
VII. Distribution of Arenicola assimilis , : mee ike: 
VIII. Specific Characters of the Caudate Arenicolide 775 
IX. Summary of Results. : : ' Bae AT 
itt) 


X. Literature 


I. Introduction. 

In response to my inquiry regarding the occurrence of 
Arenicola on the shores of New Zealand, Professor Benham 
kindly sent to me three specimens of this worm from Otago 
Harbour, and one from the Macquarie Islands. 

‘The specimens were caudate Arenicolide resembling A. 
marina, Linn., and A. claparedii, Levinsen, in external 
form. A rapid examination of the grosser anatomical 
features of one of the Otago specimens seemed to point to its 


738 J. H. ASHWORTH. 


close affinity with the latter species, for it was at once seen 
that the New Zealand specimen possessed multiple ceso- 
phageal glands and that there were no pouches on the first 
diaphragm—two features known only in, and considered to 
be almost diagnostic of, A. claparedii. At first, also, only 
five pairs (the number occurring in A. claparedii) of 
nephridia were seen in the Otago specimen, but finally a 
much reduced pair was found in the segment anterior to the 
one bearing the first fully developed nephridia. In the 
other three specimens sent by Professor Benham there are 
six pairs of fully developed nephridia, which is evidently the 
normal condition. The lateral lobes of the prostomium of 
these specimens were found to be more feebly developed than 
those of A. claparedii. There were therefore two points 
in which the southern specimens agreed with Levinsen’s 
species, viz. the presence of multiple cesophageal glands and 
the absence of diaphragmatic pouches; and two features in 
which they differed, viz. the form of the lateral lobes of the 
prostomium and the number of nephridia. 

On sectioning the anterior end of one of the Otago speci- 
mens a pair of large otocysts was found, each opening to the 
exterior by a narrow tube. The presence of these well- 
developed organs, in conjunction with the important differ- 
ences above mentioned, finally settled that the New Zealand 
specimens do not belong to the species A. claparedii, in 
which the absence of otocysts is so characteristic and 
remarkable a feature. 

These specimens agree with A. marina in the number, 
position, and character of their gills, in the number of their 
nephridia, and in the general anatomy of their otocysts ; but 
the southern specimens are clearly distinguished from 
A. marina by their multiple cesophageal glands, by the 
absence in the former of diaphragmatic pouches, and by 
other less obvious features. 

The only other known species to which the New Zea- 
land specimens show any close similarity is A. assimilis, 
Khiers (1897, pp. 103, 104), of which only the external 


ANATOMY OF ARENICOLA ASSIMILIS. 739 


characters are briefly described. Ehlers states that this 
species closely resembles A. marina in general external 
characters, but that in A. assimilis there are twenty 
cheetigerous segments and asa rule there are thirteen pairs 
of gills, the first being situated on the eighth chetigerous 
annulus, but occasionally only twelve pairs of gills are 
present. Hhlers also finds that, compared with A. marina, 
the median lobe of the prostomium of A. assimilis is 
proportionately smaller than the lateral ones, and the noto- 
podial setee somewhat more feebly feathered. 

The New Zealand specimens differ from Ehlers’ species in 
the number of cheetigerous segments and in the position of 
the first gill, but unfortunately Ehlers does not mention the 
nephridia or nephridiopores, the cesophageal glands or the 
otocysts,—important diagnostic characters concerning which 
information was essential before the affinity of the New 
Zealand specimens with A. assimilis could be either 
accepted or rejected. My thanks are due to Dr. Michaelsen, 
of the Hamburg Museum, who collected the specimens 
examined by Ehlers, for kindly sending to me two complete 
examples of A. assimilis from Uschuaia, in Tierra del Fuego, 
and an incomplete specimen from Punta Arenas, in the Straits 
of Magellan. 

As Ehlers has given only a brief account even of the 
external characters of his species,’ I propose to describe the 
principal features of the three specimens given to me by Dr. 
Michaelsen before proceeding to consider the New Zealand 
specimens in detail. 


! Since writing the above I have received, through the kindness of Pro- 
fessor Ehlers (1901), a copy of his recently published monograph, ‘ Die Poly- 
cheeten des magellanischen und chilenischen Strandes,’ in which he describes 
(pp. 177, 178) the structure of the otocysts of A. assimilis, and also states 
that the alimentary canal, the vascular system, and the nephridia of this 
species agree, so far as he can ascertain, with those of A. marina. ‘his 
agreement is, however, not quite so close as Mhilers’ statement would lead 
one to expect, since, for example, the cesophageal glands in the latter species 
are a single pair, while in A. assimilis there are several pairs. These points 
are further discussed below, 


740 J. H. ASHWORTH. 


II. Arenicola assimilis, Ehlers. 


The two complete specimens found at low-tide mark at 
Uschuaia are 105mm. and 120 mm. long respectively. The 
specimen from Punta Arenas would have been about 120 mm. 
long if complete. Their colour in spirit is a fairly uniform 
yellowish brown, but the tail of one specimen is of a some- 
what darker colour. The body is slightly swollen in the 
anterior region. The animal closely agrees in form with 
specimens of A. marina of the same size. 

External Characters.—The prostomium (see fig. 20) is 
moderately developed; the two lateral lobes are in the form of 
a V, the arms of which embrace the median lobe.’ ‘The 
prostomium resembles that of A. marina and A. cristata, 
except that the median lobe is proportionately smaller in 
Ehlers’ species. ‘lhe nuchal organ is similar in its structure 
and relations to that of A. marina. The metastomial 
grooves indicating the track of the cesophageal connectives 
are well marked. 

There are twenty cheetigerous segments, the last thirteen 
of which are branchiferous; the first gill is thus on the 
eighth chetigerous annulus. In this character A. 
assimilis differs from all other caudate Arenicolidz, in 
which the first gill (except in those abnormal cases in which 
the first true gill is missing) is on the seventh cheetigerous 
annulus, Of the three specimens in my possession only one 
has thirteen pairs of fully developed gills; in another the first 
right gillis very minute; while in the third specimen the first 
left gill is small, and the last gill on each side is considerably 
smaller than the preceding one. Ehlers (1897, p. 104) records 
specimens with only twelve pairs of gills. 

The gills of the specimens from Uschuaia are dense bushy 
structures resembling the dendritic type of gill found in 

’ There is a rough sketch (‘‘d’aprés Ehlers”) of the prostomium of A. 
assimilis ina memoir by Fauvel (1899, p. 178), which, however, does not give 
an accurate impression of its form, as the two lateral lobes are represented as 


separate, whereas they are actually united posteriorly to form the V-shaped 
structure described above. 


ANATOMY OF ARENICOLA ASSIMILIS. 741 


littoral specimens of A. marina. There are in most of them 
about eight main stems, each 2°5 mm. to 3 mm. long, bearing 
five or six dichotomously subdivided branches on each side. 
The gills of the specimen from Punta Arenas are larger and 
of a somewhat more regular form. ‘The eight or nine stout 
main stems are about 4 mm. long, and are regularly arranged 
in radiating fashion; each bears six or seven pairs of 
branches which divide dichotomously. Although at first 
sight this gill seems to approach to the pinnate type, the 
lateral branches are neither so numerous nor so regularly 
arranged as in pinnate gills, and the gill may be regarded 
as merely a well-developed example of the dendritic type. 

The skin is subdivided into annuli. Between the prosto- 
mium and the first chretigerous annulus there are five rings 
(see fig. 20). The first four of these represent the region 
found in other species of Arenicola which has been shown 
to be composed of the peristomium (here represented by the 
first two of these rings) fused with the first body-segment of 
the post-larva, the sete in which disappear very early (see 
figs. 19, 20). The fifth ring is the first annulus of the first 
chetigerous segment, this segment being composed of three 
annuli, viz. a chetigerous one and the annulus preceding 
and following it. ‘The second and third chetigerous seg- 
ments also consist of three annuli, the middle one bearing the 
sete. The fourth and succeeding segments up to the end of 
the branchial region are composed of five annuli, the fourth 
of which is chetigerous. The region between any two 
cheetigerous annuli behind the third is therefore subdivided 
into four rings. 

The epidermis of the tail is raised into numerous papille. 
The segmentation of this region is only feebly marked, but it 
is indicated, especially in the anterior portion of the tail, 
by the presence of somewhat larger annuli placed at regular 
intervals, upon which the epidermal papille are distinctly 
larger than those on the intervening annuli. Each of these 
larger rings is followed by a slight constriction, denoting the 
presence internally of a septum, best seen in those parts of 


742 J. H. ASHWORTH. 


the tail which were stretched at the moment of death. The 
space between two of the larger annuli is subdivided at the 
anterior end of the tail into two or three rings, but further 
back into from four to ten. These smaller annuli also bear 
epidermal papilla, but in the anterior tail region they are 
distinctly smaller than the papille found on the larger annuli. 
Proceeding backwards along the tail, the difference in the 
size of the annul and of the papille they bear may be 
clearly recognised until the middle of the tail has been passed ; 
then the papille become almost equal in size, and near the 
anus if is impossible to distinguish any difference between 
those of the various annuli. ‘There are about twenty-eight 
segments in the tail of each of the complete specimens. 

Setz.—The capillary sete (figs. 1, 1a) of the notopodium 
are very similar to those of A. marina. They attain a 
length of 4:5 mm., and on their distal fourth bear small 
pointed processes, which, as Khlers (1897, p. 104) remarked, 
are not so well developed as those of A. marina. The 
processes are usually present on both sides of the seta; they - 
are moderately obvious on one side, but on the other they are 
very minute, and are borne on the edge of a thin border or 
lamina. This lamina, which seldom exceeds 6 u in width, 
extends along the seta for a distance of about one third its 
length. In some of the setee the lamina is not denticulate at 
its margin, and in others is only very faintly so; but it is 
crossed by fine oblique lines, the intervals between which 
correspond roughly to the size of the teeth on the dentigerous 
lamin. From an examination of the sete of A. assimilis 
and A. marina, it seems probable that the lamina at first 
possesses an entire margin, but later this tends to break up 
from the edge inwards, thus giving rise to the minute teeth 
which are usually seen on full-grown sete. This explanation 
would account for the fact that in some setz the margin of 
the lamina is entire, while in others it bears either exceed- 
ingly minute denticulations or the more obvious teeth shown 
in fig. La. These three conditions are occasionally seen at 
different points along the border of a single seta. 


ANATOMY OF ARKENICOLA ASSIMILIS. 745 


Setz similar to those above described are present in A. 
marina, and in some examples the lamina is very well 
marked, e. g. in a specimen of the Laminarian variety the thin 
border extends for nearly a millimetre along the seta, and 
attains a width of 20 uw. Similar sete are present in A. 
eristata, but in A. claparedii! the lamina is not so well 
developed, being short and narrow. In A. ecaudata and A. 
grubii the lamina is also very narrow, seldom exceeding 
about 3 in width. 

The neuropodia of A. assimilis are easily seen, even in 
the first segment. ‘They are especially well developed in 
the branchial region, where each resembles a pair of closely 
applied tumid lips, between which is the row of crotchets. 
These (fig. 9) are often curved, and are 0°6 mm. to 0°7 mm. in 
length, being considerably longer than cheetz from specimens 
of A. marina of the same size. (The longest crotchets found 
in a specimen of A. marina 125 mm. long are only 0°47 mm. 
in length.) The rostrum is short and blunt, even in unworn 
cheetee. There is a small subrostral enlargement, and about 
six to nine teeth are present just behind the rostrum. 

Musculature.—The musculature calls for little comment ; 
it is similar to that of A. marina, except that the oblique 
muscles are present along the whole animal from the first 
diaphragm to the end of the tail. They are exceedingly thin 
bands, somewhat broader in the posterior part of the gill 
region, but even here seldom exceeding 0°5 mm. in width, and 
as a rule they are only 0:2 mm to 0°3 mm. wide. There is a 
dorsal mesentery in the first and second cheetigerous seg- 
ments supporting the dorsal blood-vessel. The three dia- 
phragms are, as usual, situated at the anterior ends of the first, 
third, and fourth cheetigerous segments. ‘There are no 
pouches on the first diaphragm. This condition was con- 
sidered to be so marked a feature of A. claparedii that it 
was given as one of the diagnostic characters of this species 
(Gamble and Ashworth, 1900, pp. 533, 541), since all other 
Arenicolide whose anatomy is fully investigated possess 

' See Gamble and Ashworth, 1900, pl. xxiii, fig. 23, 


744 J. H. ASHWORTH. 


diaphragmatic pouches, and in some species they attain a 
large size, e.g. in A. cristata they may reach a length of 
12mm. Many of the blood-vesseis which cross the ccelom 
obliquely to the nephridia and gills are provided with a very 
obvious connective-tissue strand or band, which gradually 
increases in size in the posterior segments of the gill region, 
forming in the last two or three segments of this part of the 
body an almost complete septum supporting the afferent and 
efferent branchial vessels. There are well-developed caudal 
septa. 

Alimentary Cana].—The most striking feature of the 
internal anatomy of A. assimilis is the presence of multiple 
cesophageal pouches. ‘These are placed on the sides of the 
cesophagus, just behind the third diaphragm. There are in 
each of two specimens six, and in another eight, pouches on 
each side. The anterior pair is long—-12 mm. in one speci- 
men, 17 mm. in another,—and each of them is usually swollen 
at or near its anterior free end, having a club-shaped appear- 
ance. Their abundant blood-supply is evidenced by the 
network seen in their walls. In the contracted condition 
these anterior pouches are digitiform structures with a some- 
what moniliform appearance. ‘The smaller posterior glands 
are from 1 mm. to 4 mm. in length, and are pear-shaped or 
oval sacs with rather thicker walls. As in other Areni- 
colide, the cavity of each pouch is partially subdivided by 
numerous septa produced by infolding of the wall; each 
septum, therefore, is composed of two lamellw of glandular 
epithelium, between which is a cavity filled with blood. The 
partitions are very obvious in the smaller pouches, and in 
the larger pouches when in a contracted condition; but when 
these are fully distended the septa become mere ridges on 
the inner wall of the pouch. In the presence of multiple 
cesophageal pouches A. assimilis conforms to another of 
the features hitherto considered to be peculiar to A. 
claparedii, as in all other species of Arenicola in which 
the pouches are known there is only a single pair.! In other 


' Ehlers (1901, p. 177) states that the gut of A.assimilis agrees with 


ANATOMY OF ARENICOLA ASSIMILIS. TA5 


respects the alimentary canal resembles that of A. marina. 
The ventral groove, which is well seen in the intestine, may 
be traced forwards into the stomach to about the level of the 
eighth or seventh seta. 

Vascular System.—The vascular system closely agrees 
with that of A. marina (see Gamble and Ashworth, 1898, 
pl. u), except in the fifth and seventh chetigerous seg- 
ments. In each of these there is only one pair of vessels, 
afferent branches of the ventral vessel, passing to the 
nephridia. The first pair of efferent branchial vessels is 
situated in the eighth segment; this and the four succeeding 
pans open into the subintestinal vessels, while the last eight 
gills, 1. e. those of the thirteenth to twentieth segments, 
return blood to the dorsal vessel. The body-wall is well 
supphed with blood-vessels, especially in the anterior region ; 
in sections of the peristomium and first cheetigerous seg- 
ment there are numerous vessels lying either in the con- 
nective tissue or in small ccelomic canals (see below) just 
beneath the epidermis (fig. 22); in sections of some of the 
posterior segments the vessels are not so abundant. 

The heart is of moderate size, and has the usual relations. 
There is a cardiac body formed by ingrowths, chiefly of the 
posterior wall of the heart, and this is well developed in one 
of the specimens 120 mm. long. 

Celom.—The ccelom is spacious, as in A, marina. A 
remarkable feature noticed at once in sections (fig. 22) of the 
anterior end of A. assimilis is the large number of coelomic 
spaces in the body-wall and between the muscles. In this 
portion of the animal there are exceedingly numerous coelomic 
canals lying in the subepidermal tissue of the body-wall, 
penetrating into the muscle-bands, especially of the buccal 
musculature, insinuating themselves between the brain-lobes 
and between the brain and the prostomial epithelium, and 
often accompanying the blood-vessels which supply the body- 


that, of A. marina, but the presence of multiple cesophageal glands in the 
former while there is only a single pair in the latter species is a point of 
difference of considerable systematic importance. 


746 J. H. ASHWORTH. 


wall. In each of these canals the thin lining of ccelomic 
epithelium may be easily recognised, and ccelomic corpuscles 
may also be found in many of them. Similar canals are 
present in A. marina (Gamble and Ashworth, 1898, p. 28), 
in A. grubii, and to a less extent in the other species ; 
but the development of these outgrowths of the ccelom 
reaches its maximum in A. assimilis. They probably act 
as nutritive, and possibly also as excretory and respiratory 
channels. ‘here are very few coelomic canals in the pos- 
terior part of the animal. The coelomic fluid and corpuscles 
resemble, as far as can be ascertained in preserved specimens, 
those of A. marina. 

There are six pairs of nephridia, the ex- 


Nephridia. 
ternal openings of which are slightly posterior to the dorsal 
ends of the fourth to the ninth neuropodia. The funnels of 
the first pair of nephridia lie on the anterior face of the 
third diaphragm. This condition is found again only in 
A. marina; the first pair of nephridia of other species 
corresponds in position to the second pair of A. assimilis 
and A, marina. The dorsal lip of each nephrostome 
(fig. 17) bears about twelve to fourteen spatulate or triangular 
ciliated processes attached by their narrower ends. These 
are subdivided distally, some of the larger ones into five or 
six. The edge of the ventral lip of the nephrostome is 
thrown into folds or frills, so that although it agrees in 
general shape with the ventral lip of the nephrostome 
of the marina section of the genus, the ventral nephro- 
stomial lip of A. assimilis is quite distinguishable by this 
peculiar character. his frilling is probably not due to con- 
traction on killing, as it is not seen in specimens of 
A. marina, A. claparedii, and A. cristata which have 
been killed in a similar manner, 

The nephrostome of the first nephridium lies on the 
anterior face of the third diaphragm, and is directed an- 
teriorly. It is smaller than any of the other nephrostomes, 
its dorsal lip bears only eight to eleven processes, and the 
frilling of the edge of the ventral lip is not so well marked. 


ANATOMY OF ARENICOLA ASSIMILIS. 747 


The first nephridium is usually distinctly smaller than any 
of the others, a condition frequently noticed in A. marina. 
Although this nephridium possesses a gonidial vessel, no 
gonad is developed upon it. 

Gonads.—The gonads are, as in other species, associated 
with the nephridia, and are present on all except the first 
pair. Hach gonad is a club-shaped mass of cells about 1°5 
mm. long (fig. 17), formed by proliferation of the cells 
covering the gonidial vessel (Gamble and Ashworth, 1900, 
p. 921) immediately behind the nephrostome. The forma- 
tion of ova and spermatozoa follows the same course as in 
A.marina. ‘The ova present in the ccelomic fluid of the 
specimen from Punta Arenas are apparently mature,! and 
have a distinct but thin vitelline membrane (3 , thick). 
They are not spherical, but somewhat discoidal. The face of 
the disc is usually oval, and measures 0°19 to 0°20 mm. by 0°15 
to 0°16 mm. in diameter. The thickness of the egg is about 
0-075 mm. Measurements of a considerable number of well- 
preserved unshrunk eggs from the ccelomic fluid show that 
the three axes above named are fairly constant in propor- 
tion. It will be convenient to correct here a statement in 
the memoir by Dr. Gamble and myself (1900, p. 527) in 
which- the ova of A. marina, A. claparedii, and A. 
cristata are described as spherical. This isa mistake, as the 
ova of all these species are flattened in one plane, like those of 
A. assimilis described above. In A. marina? the face of 
the egg is either circular (usually) and about 0°14 to 0°15 mm. 
in diameter, or it is oval, with diameters of 0°16 and 0:12 to 
0-14 mm., and the third axis of the egg is from 0:08 to 0:09 
mm. in length. ‘The ova of A. claparedii are usually only 
slightly oval, the two diameters of the face of the egg being 
about 0°16 mm. and 0-14 to 0°16 mm. respectively, and the 


1 This specimen Was taken in September, 1892. 

2 The following description and measurements may replace those given on 
p. 527 in the memoir cited above; they have been drawn either from living or 
well-preserved ova, most of which have come into my hands since the com- 
pletion of that memoir. 


748 J. H. ASHWORTH. 


thickness of the egg 0°07 mm. In A. cristata the three 
axes of ova removed from a ripe female in Naples measure 
0°155mm., 0°145 mm., and 0°07 mm. respectively.! The ova of 
A. grubiiand A. ecaudata are not compressed in this way, 
or only very slightly so. They are usually ovoid, and ripe ova 
of the former species are 0°17 mm. long and 0°15 mm. broad 
and thick. The largest ova of A. ecaudata which I have seen 
have slightly smaller dimensions, but they are probably not 
quite mature. The ova of A. grubii and A. ecaudata are 
distinguished by their stout vitelline membrane, which is 5 
to 6 4 thick; while in'A. claparedii, A. cristata, and A. 
assimilis itis 2 to 3 m,and in A. marina only slightly over 
1 «x in thickness. 

Brain.—The brain of A. assimilis conforms to the 
general plan seen in the marina section of the genus. It 
consists of a pair of anterior lobes placed well forward in 
the prostomium, a pair of posterior lobes which lie below the 
nuchal organ, and an intermediate region which connects the 
anterior and posterior lobes. The anterior lobes are short 
but very broad ; in fact, this is by far the broadest part of 
the brain; behind these lobes the brain gradually tapers. 
The brain may be roughly compared in shape to two slightly 
flattened pears lying side by side with their narrower faces 
adjacent and fused along the middle third of their length. 
The broad forwardly directed ends of the pears represent 
the anterior cerebral lobes, while the tapering ends represent 
the posterior lobes, which are continuous with two nerve- 
tracts lying below the epithelium of the nuchal organ. The 
anterior brain-lobes are separated in front by a ccelomic space. 
Hach gives off anteriorly and dorsally a series of nerves 
to the epithelium of the prostomium. ‘The anterior part 
of these lobes consists almost entirely of small cells situated 
in clusters and separated from one another by fibrous tracts 
and by neuroglial tissue. Further back the delicate neuro- 
pile which forms the core of these anterior lobes is well seen, 

' See also Child (1900), p. 592, for further observations on the ova of A. 
cristata. 


ANATOMY OF ARENICOLA ASSIMILIS. 749 


surrounded by clusters of nerve-cells and neuroglial tissue. 
Bundles of nerve-fibrils may be traced from the bases of the 
prostomial epithelial cells into the neuropile (see fig. 23). 
Larger unipolar ganglion-cells are found immediately out- 
side the neuropile, and particularly on the side nearest the 
middle line. The cesophageal connectives arise from the 
anterior lobes at the point where the neuropile reaches its 
greatest development (fig. 22). The eyes are found on the 
dorsal side of this part of the brain. There are four or five 
on each of the anterior lobes. A little further back large 
pyriform ganglion-cells become more numerous, and 
especially on the inner side of the two anterior lobes just 
before they unite in the middle line and for some distance 
after their union (see fig. 23). Passing backwards along 
this region, it is seen that the ganglion-cells become more 
restricted to the dorsal and lateral faces of the brain, the 
middle and ventral parts being composed largely of neuro- 
pile, in which also neuroglial cells and fibrilla may be 
recognised. The ganglion-cells of this region are more 
intimately associated with the median part of the prostomium. 
The posterior brain-lobes are small and tapering, and 
eradually merge into two nerve-tracts which lie on the inner 
side of the nuchal organ just below the sensory epithelium. 
The brain of the specimen 120 mm. long is 0°65 mm. in 
length, 1 mm. broad across the anterior lobes, and about 0-4 
mm. deep in this region. It is most nearly like the brain of 
A. marina, but is wider anteriorly. (The brain of a speci- 
men of A. marina 120 mm. long is about 0°7 mm. wide.) 
Gisophageal Connectives.—The csophageal connec- 
tives arise from the lateral region of the broadest part of the 
anterior cerebral lobes, i.e. just in front of their point of 
union. Each is a stout fibrous cord with numerous cells on 
its outer face, which, in the first part of its course, lies about 
a millimetre below the epidermis, and is slung up in a 
muscle sheet, which is attached to the subepidermal muscu- 
lature by numerous muscle-strands (fig. 22). It is only in 
the ventro-lateral region that the connectives approach the 
voL. 46, PART 4.—NEW SERIES. Cee 


750 J. H. ASHWORTH. 


body-wall and finally come to lie upon the layer of circular 
muscles. The connectives, the course of which is indicated 
externally by the metastomial grooves, unite in the hinder 
portion of the third annulus (cf. fig. 19). Hach gives off 
numerous branches to the epidermis of the region through 
which it passes ; in fact, the nerve-supply to the skin and 
following segment is enormous—nerves pass into the raised 
areas upon the skin and repeatedly branch, their terminations 
lying in close contact with the bases of the epidermal cells 
(fig.22). In addition to one or two nerves derived from each 
connective, the skin of the region immediately below the 
prostomium receives two moderately stout nerves which 
arise from the ventral portion of each anterior cerebral lobe 
close to the point of origin of the connective. These nerves 
run on to the roof and sides of the mouth, and their branches 
may be traced a considerable distance along the pharynx. 
They are apparently more numerous on the dorsal than on 
the ventral region of the pharynx. In many sections the 
nerves may be seen sending branches along the axes of the 
buccal papille (fig. 22). The connective gives off a very 
short but stout nerve to the otocyst. Ganglion-cells are 
present in moderate number around the point of origin of 
this nerve. The nerve comes into contact with the otocyst 
at the point where the tube leads off to the exterior, and is 
intimately related to both structures. It provides the otocyst 
with a sheath of nervous elements, which lies just below the 
sensory epithelium, and also sends a small nervous sheath 
along the tube. 

Nerve-cord.—The nerve-cord is situated within the 
layer of circular muscles. The right and left fibrous portions 
are separated by a median vertical sheet of neuroglia. ‘The 
ganglion-cells are distributed along the whole length of the 
cord, and are not aggregated into ganglia. ‘They are 
numerous, unipolar, pyriform, usually quite small cells with 
deeply staining nuclei, aud most of them are situated in the 
ventro-lateral regions of the cord; but some few larger ones 
measuring from 15 to 30 «in length, and having vesicular 


ANATOMY OF ARENICOLA ASSIMILIS. 751 


nuclei, are found rather nearer the middle line in many of 
the sections examined. The process of each cell is directed 
dorsally into the lateral portion of the fibrous mass. The 
spinal nerves arise in the same manner and position as in 
other Arenicolide (Gamble and Ashworth, 1900, pp. 482, 
483). Giant-fibres to the number of two or three are seen 
in sections of the branchial region and tail. The giant-cells 
are regularly arranged, being situated close to the posterior 
border of each segment. In eight of the nine segments 
examined there is only one giant-cell per segment, but in the 
other segment two cells are present near together. The 
giant-cells are placed in the extreme lateral regions of the 
cord, and in the first piece of cord which was sectioned they 
are situated alternately on the right and left sides, i.e. in 
the seventeenth to twentieth chetigerous segments, and in 
the first tail segment of this specimen they are situated 
respectively R, L, R, L, R (see fig. 12). In sections of the 
same specimen taken further forwards (tenth to thirteenth 
segments) this curiously regular arrangement was not found, 
the cells present in these segments being situated respectively 
L, L, RR (two cells are present in this somite), R. The 
average size of the cellsis 0:065 mm. long and 0:05 mm. broad 
and deep. Each cell is pyriform, surrounded by a fibrillated 
sheath, and sends out usually ouly one process, which passes 
at once into the fibrous portion of the cord towards the 
lateral giant-fibre. ‘lhe protoplasm of one cell, however, is 
drawn out dorsally into five processes, one of which is much 
thicker than the others, and may be easily traced into the 
mid-dorsal region of the fibrous part of the cord. The 
slender processes are traceable only a very short distance, 
being lost either between the small ganglion-cells or imme- 
diately on entering the fibrous part of the cord. There is in 
most of the giant-cells a more deeply staining area in the 
protoplasm close to the nucleus, due to the presence of 
chromophilous granules. This probably corresponds to the 
similar but better marked centrosphere seen in A. grubii 


(Gamble and Ashworth, 1900, pp. 487, 488, and fig. 76). 


752 J. H. ASHWORTH. 


Sense Organs.—The eyes, of which there are four or five 
on each side of the prostomium, are similar to those of other 
Arenicolide (Gamble and Ashworth, 1900, pp. 506, 507). 
They are situated either among the small nerve-cells on the 
dorsal surface of the anterior cerebral lobes just in front of 
their point of union, or in the epidermis immediately dorsal to 
this region. Hach eye is composed of a cup-shaped mass 
(6 to 12 « in diameter) of reddish-brown pigment spherules 
grasping the base of a spherical or ovoid lens. 

The otocysts! are remarkable for their large size. They 
are oval sacs whose three internal diameters are 0 37, 0°36, and 
0°25 mm. respectively. Their size may be better appreciated 
after comparison with that of the otocysts of other species 
(see figs. 13, 16). The largest otocysts seen while examining 
four large specimens of A. marina for this purpose were 
found in one about 180 mm. long, where they measure 0:07, 
0-16, and 0:17 mm. along each of their three internal dia- 
meters ; wlile in a specimen 250 min. long the corresponding 
measurements are 0°15, 0:12, and 0°15 mm. (fig. 16). The 
nearly spherical otocysts of full-grown specimens of A. 
cristata (800 mm. long), A. grubii (180 mm.), and A. 
ecaudata (180 mm.) have a mean internal diameter of 
0-17 to 0:18 mm., 0°16 to 0°17 mm., and 012 to 0°13 mm. 
respectively. From these figures it will be seen that the 
otocysts of A. assimilis are much larger than those of 
any other species. Hach opens to the exterior by a narrow 
curved tube. The external opening is very minute, and at the 
bottom of a groove situated immediately in front and to the 
outer side of the lateral portion of the nuchal organ (fig. 22). 

1 Ehlers (1901, pp. 177, 178) has recently described these organs and their 
numerous spherical otoliths, consisting of a concentrically layered, evidently 
secreted, material. These vary in size from 0°003 to 0°03 mm. The small ones 
are compared to those of A. claparedii (the author really means A. grubii). 
The larger otoliths are often irregular, and in a few a central foreign body may 
be observed. ‘The opening of the tube into the cyst is very small, and Ehlers 
thinks this is correlated with the character of the otoliths. He also states 
that the otoeyst is apparently larger and its external opening nearer the 
brain than in A. marina, 


ANATOMY OF ARENICOLA ASSIMILIS. 7538 


The opening is close to the point of origin of the cesophageal 
connective, i.e. at the dorsal end of the metastomial groove, 
so that it is more dorsally situated than the corresponding 
opening in A, marina. ‘The lumen of the tube is small, 
and in two of the four examined is almost obliterated alone 
part of its length by approximation of the walls. In three of 
the tubes there are fine particles of foreign matter at one or 
more points. The otocyst and tube are lined with a cuticle 
about 3 in thickness. The epithelial wall of the otocyst is 
comparatively thin (80 to 40 ,). The sense cells are not 
easily distinguishable, at first sight, from the supporting cells, 
but in one series of sections they may be distinguished by 
the presence of neuro-fibrille: in the former. Hach sense cell 
is seen to be traversed by asingle fine fibril, which terminates 
immediately below the cuticle. These cells and fibrille are 
especially abundant in the wall of the otocyst near the 
entrance of the tube, and they are also present in the 
adjacent part of the tube. Below the epithelium is the 
nervous sheath, among the fibres of which occur scattered 
fusiform or stellate cells. The nerve-supply to the otocyst is 
derived from the cesophageal connectives (see above, p. 750). 
The otocyst contains the coagulated remains of the fluid with 
which it was filled in life. Among this coagulum are 
numerous minute spherical deeply staining granules, which 
are probably secreted by some of the cells in the wall of the 
otocyst (fig. 13). There are about forty or fifty otoliths in 
each otocyst; they are usually spherical, but a few are oval, 
and some are irregular, but have a rounded outline. They 
nearly all show concentric markings indicating the method 
of their formation by deposition of layer upon layer of a 
secretion produced by cells in the wall of the otocyst. ‘he 
largest otoliths are 35 w to 45 « in diameter. In the centre 
of a few of them there is a minute refringent body, evidently 
of foreign origin, forming the nucleus around which the 
secreted matter has been deposited. Besides the contents 
already named, there are in the otocyst several deeply 
staining bodies varying in size from the minute granules 


7a4 J. H. ASHWORTH. 


present in the coagulum to spherical, oval, or elongate masses 
10 w in diameter, which are either free or adhering to the 
surface of one of the otoliths. Their appearance suggests 
that they are composed of a substance similar to that of 
which the otoliths are formed, although the latter are usually 
much more lightly stained (fig. 15). 

The nuchal organ of A. assimilis resembles that of A. 
marina in its main features. The epithelium of the organ is 
composed of exceedingly slender columnar cells; some of 
these—the sense cells—are 70 p to 80 pw long, and only about 
2 » wide, and have deeply staining nuclei. The intervening 
supporting cells are a little stouter, and their nuclei stain less 
deeply. Many of these cells are ciliated, and some are 
elandular. Beneath the epithelium there is a layer of nerve 
elements in connection with the posterior brain-lobes. From 
this layer neuro-fibrillae may be traced into and through the 
entire length of many of the sensory cells. 

Similar fibrils may be seen in some of the epidermal cells 
of the general body surface, and of the papille of the 
proboscis. 


III. Specimens of Arenicoia from New Zealand. 


Three of these were collected in Otago Harbour, and are 
respectively 136, 126, and 90 mm. long. Another specimen 
from the Macquarie Islands is 217 mm. long. The Otago 
specimens are of a light brown colour, the two larger ones 
being darker in the anterior gill region, and the Macquarie 
specimen is dark brown throughout its length. 

External Characters.—The prostomium (fig. 20), the 
nuchal organ, and the metastomial grooves agree in form and 
relations with those of A. assimilis. ‘There are nineteen 
cheetigerous segments, of which the last thirteen usually bear 
gills. The first gill is thus situated on the seventh segment, 
as in A. marina, A.claparedii, and A. cristata. The 
gills of the two larger Otago specimens are all fully developed, 
but in the smallest specimen the last right gill is smaller 


ANATOMY OF ARENICOLA ASSIMILIS. 75D 


than any of the others, and its fellow on the left side is 
represented by a single filament about half a millimetre in 
length and bifid at the tip. In the Macquarie specimen 
there are only twelve pairs of gills, the first being well 
developed and situated on the eighth chetigerous annulus. 
The true first gill! is totally absent, a condition frequently 
met with in A. marina. The gills of the Otago specimens 
are of the pinnate type, and are beautifully regular in 
arrangement. Hach consists of fourteen to sixteen main 
stems, usually 3°5—4 mm. long (but in several cases reaching 
6 mm.), which radiate from the base of the notopodium, and 
are connected near their bases by a web-like membrane. 
Each stem bears eleven to eighteen pairs of pinne, which are 
either opposite or almost alternate in arrangement and 
usually divide dichotomously. These gills are remarkable 
for the enormous size of the afferent and efferent blood- 
vessels, best seen in the main stems and in the larger 
pinne. They closely resemble the gills of the Laminarian 
variety of A. marina figured by Gamble and Ashworth 
(1898, pl. i, fig. 2), except that the webbing at the base is 
more pronounced in the southern specimens. The gills of 
the Macquarie specimen are of a different type. They have 
only seven or eight main trunks 3°5—5:°5 mm. long, each 
bearing six or eight pinne on each side, and these are less 
regularly arranged than in the Otago specimens. There is 
no connecting membrane at the bases of the main trunks. 
These gills resemble in form, but are larger than, those of the 
Uschuaia specimens of A. assimilis (see p. 740). 

The annulation of the skin is exactly like that of A. 
assimilis (see p. 741). 

Setz.—The notopodial sets (figs. 2, 24) taper more 
abruptly at the tip than those of A. assimilis. Those of 
the Otago specimen are 4 mm. long. The usual pointed 
barbs or processes are present on one side of the shaft of the 
seta, while on the other is a well-marked lamina, which 
in most sete is 12 uw, and in some is 15 « broad. The 


The pair of efferent vessels of this segment is also completely suppressed, 


756 J. H. ASHWORTH. 


margin of the lamina is in many cases entire, but im some is 
very finely dentate. The sete of the Macquarie specimen 
are about 5 mm. long, and the lamina is much narrower, 
being only about 6 w in width. 

The neuropodia are well developed ; the crotchets show an 
interesting feature. On examining the post-rostral region 
(fig. 7), there are seen to be about five to seven teeth—that is, 
five, six, or seven teeth are in focus at the same time as the 
rostrum, and lie approximately in the same plane. On focus- 
sing slightly above or below this level, there comes into view 
a number of teeth situated on the sides of the rostrum, so 
that the latter projects from the centre of a series of teeth 
arranged around its base (fig. 8). The small subrostral pro- 
cess marks the position of the base of the lowest tooth of the 
series. Only those cheetee which have not been much worn 
show these lateral teeth.! The rostrum is slightly longer and 
more pointed, and the enlargement near the middle of the 
shaft better marked than in Ehlers’ species. Many of the 
crotchets are strongly curved. ‘They vary in length in the 
Otago specimens from 0°66 mm. to 0°8 mm., and in the Mac- 
quarie specimen they reach a length of 0°86 mm. The last- 
named cheetz are stouter, the rostrum more rounded at the 
tip, and more nearly in line with the shaft, and the teeth are 
more feebly marked ; these characters are due to the greater 
age of the specimen from which the cheeta were taken. 

Musculature.—The musculature is very similar to that 
of A. assimilis, except that the muscles in the region of the 
first diaphragm are more slender in the Otago specimens. 
Oblique muscles are present alone the body from the first 
diaphragm to the end of the tail. There are no pouches on 
the first diaphragm. 

Alimentary Canal.—The alimentary canal agrees most 
minutely with that of A.assimilis. Multiple cesophageal 


1 They are very well seen in the crotchets of post-larval stages (see fig. 10). 
A similar series of teeth is present around the base of the rostrum of the 
cheete of other Arenicolidee, but they are not so easily distinguished as in the 
specimens above described, 


ANATOMY OF ARENICOIA ASSIMILIS. LO 


pouches, to the number of seven on each side, are present in 
the three specimens examined. The anterior pair is long 
and filiform or club-shaped, measuring in one case 15 mm., 
and in the other two specimens 24mm. in length. The other 
pouches are pyriform or ovoid, and 5 to 5 mm. long. 

Vascular System.—The vascular system agrees closely 
with that of A. marina, and only differs from that of A. 
assimilis in the position of the first efferent branchial 
vessel. The first six efferent branchial vessels open into the 
subintéstinal vessels, and the last seven into the dorsal 
vessel. The heart of the specimen 126 mm. long contains a 
moderately developed heart-body. 

Nephridia.—There are six pairs of nephridia, opening, as 
in A.marinaand A. assimilis, on the fourth to the ninth 
segments. In the three specimens examined the first pair 
is smaller than any of the others, and one of the nephridia is 
considerably reduced, no funnel being visible. The funnels 
of the first nephridia le on the anterior face of the first dia- 
phragm, ‘heir dorsal lips bear about six broad, but usually 
undivided, ciliated processes, and their ventral lips, though 
small, are thrown into several of the peculiar folds or frills 
as described above (p. 746) in the nephrostomes of A. 
assimilis. The funnels of the other nephridia are larger ; 
their dorsal lips bear about sixteen spatulate processes, most 
of which are subdivided terminally into five or six, and their 
ventral lips are thrown into some twenty or more folds. The 
vesicles of the nephridia had been recently greatly distended 
but are now almost empty. 

Gonads.—The gonads are small and occur in the usual 
position. None are present on the first pair of nephridia. 
It is probable that the breeding season of these specimens 
was practically over at the time of their capture (September, 
1899). It is evident that the nephridial vesicles had been 
recently subjected to great distension, and this was probably 
due to the accumulation therein of genital products. Similar 


' Except that in the Otago specimens there is in connection with the third 
nephridium only one blood-vessel, viz. an afferent branch of the ventral vessel. 


758 J. H. ASHWORTH, 


distension of the vesicles occurs during the breeding season 
in A. marina, A. grubii, and A. ecaudata (Gamble and 
Ashworth, 1898, pl. ii, fig. 15, and 1900, pl. xxvi, fig. 47). On 
staining and clearing the nephridia the vesicles are found to 
still contain either a few large ova or masses of spermatids. 
‘he ova are of the same somewhat flattened shape as those of 
A. assimilis. ‘Their three axes measure 0°195 to 0°20 m.m, 
0°16 to 0°175 mm., and 0:075 mm. respectively. (For the 
measurements of the ova of other species see p. 747.) 

Central Nervous System.—The brain resembles that 
of the Uschuaia specimens, except that the anterior lobes are 
much broader. In the specimen 126 mm. long the brain is 
about 0°7 mm. long, and is broadest across the anterior lobes 
at the point of origin of the cesophageal connectives. The 
breadth of the brain at this point is 15 mm. and its depth 
04mm. After the fusion of the two anterior lobes the brain 
rapidly narrows, so that its middle region is only about 
0-7 mm. broad. The structure of the anterior lobes is 
exactly as described for A. assimilis on pp. 748, 749. Near 
their point of union larger ganglion-cells occur near the middle 
line, gradually increasing in number posteriorly and being found 
over the whole dorsal face of the neuropile of the mid-brain. 
In the posterior part of this region there are a few groups of 
pyriform, fusiform, or pyramidal ganglon-cells, the stout 
processes (usually only one to each cell) of which are united 
into a number of bundles. These processes pass downwards 
into the ventral portion of the neuropile, where they 
branch freely (see fig. 23). Similar ganglion-cells extend 
some distance into the posterior brain lobes. In other 
respects the brain of these specimens conforms to the 
description given on pp. 748, 749. 

The cesophageal connectives arise, as usual, from the 
posterior part of the anterior cerebral lobes. They lie 
immediately below the epidermis of the metastomial groove, 
and give off numerous nerves to the skin and buccal muscula- 
ture. There is a slight swelling on the connective at the 
origin of the nerve to the otocyst. 


ANATOMY OF ARENICOLA ASSIMILIS. 709 


The situation and structure of the nerve-cord agree with the 
description given on p. 750. Sections taken in the mid- 
branchial region show one, two, or sometimes three giant- 
libres. Serial sections of three segments (fig. 11) of this part 
show that in the first and last seements there are two giant- 
cells, and in the middle one only one cell. When two cells 
are present they lie, asin A. grubii, one behind the other, 
the anterior one being only a little distance posterior to the 
parapodium. The cells are laterally situated, pyriform in 
shape, and their single process is directed into the adjacent 
fibrous portion of the cord. 

Sense Organs.—The nuchal organ and the reddish-brown 
eyes have the usual structure and position, 

The otocysts are somewhat smaller and le more laterally 
than in A. assimilis. They are almost spherical sacs 
(fig. 15) about 0°21 mm. in diameter, which communicate with 
the exterior by a tube, whose external aperture occupies a 
similar position to the corresponding opening in A. marina. 
It is situated near the metastomial groove, but further from 
the brain than in A. assimilis. The otoliths are of purely 
external origin. They consist of numerous irregular bodies 
(quartz-grains, fragments of spicules, etc.), without any of 
the chitinoid covering which is usually associated with the 
otoliths of A. marina, and which forms the major portion 
of each otolith of A. assimilis. There are in each otocyst 
from twenty to fifty moderately large bodies, the largest 
being 55 w long and 27 w broad, and also a quantity of finer 
débris of similar origin and character. ‘he lumen of the 
tube is slit-lke, and about halfway down one of the tubes 
there are at two points large foreign bodies. In the structure 
of its wall the otocyst agrees with that of A. assimilis 


(p. 753). 


760 J. H. ASHWORTH. 


IV. Systematic Position of Arenicola assimilis and of 
the Specimens from New Zealand. 


Arenicola assimilis is clearly distinguished from all 
other species by the following characters :—(a) externally, by 
its twenty chetigerous segments and the presence of the 
first gill on the eighth segment; and (b) internally, by the 
possession of six pairs of nephridia opening on the fourth to 
the ninth segments, by the presence of numerous cesophageal 
elands and of large otocysts opening to the exterior, and by 
the absence of the pouches on the first diaphragm. 

This species obviously falls within the caudate section of 
the genus Arenicola. It has practically no points in com- 
mon with A. eristata except those of generic value ; the two 
species differ in every one of the characters named above, 
Khlers’ species has some points of resemblance to A, 
claparedii; in fact, the two most characteristic features of 
the latter species are found in A. assimilis, viz. the 
multiple cesophageal glands and the absence of diaphrag- 
matic pouches. But these two species are clearly distin- 
guished by the differences in the number of segments, the 
position of the first gill, the number of nephridia, and the 
presence in A. assimilis of large otocysts, such organs being 
absent in A. claparedii. 

The structures hitherto beheved to be diagnostic of 
A. marina are also found in Khlers’ species, viz. six pairs 
of nephridia opening on segments 4 to 9, and a pair of 
open otocysts. These two species may be easily differentiated 
by an inspection of the number of segments, the position of 
the first gill, the cesophageal glands, and the first diaphragm 
(to ascertain the presence or absence of pouches), 

So that, although related in some degree to A. marina 
and A. claparedii, A. assimilis is qumte distinct from 
either, and may be easily determined by reference to the six 
characters given above. 

The determination of the systematic position of the New 


ANATOMY OF ARENICOLA ASSIMILIS. 76! 


Zealand specimens is a matter of considerable difficulty. In 
the number of chetigerous segments and position of first gill 
they resemble A. marina and A. claparedii, but the pro- 
stomium is more nearly like that of the former. Internally 
there are four characters, two of which are in agreement with 
those of A. marina and in contrast to those of A. claparedii, 
and two are vice versa 


(1) the number and position of the 
nephridia and the presence of open otocysts, and (2) the 
presence of multiple cesophageal glands and the absence of 
pouches on the first diaphragm. The absence of otocysts in 
A. claparedii is so remarkable and characteristic a feature 
that their presence in the Otago specimens, taken in con- 
junction with the important differences in the number of 
nephridia and the form of the prostomium, is sufficient to 
exclude the southern specimens from Levinsen’s species. 
While their relationship with A. marina rests on a stronger 
basis, the internal differences are too important to be passed 
over, and one must look elsewhere for a nearer ally. 

Throughout the description of the anatomy of the New 
Zealand specimeus it is striking how frequently a perfect 
agreement occurs between them and A. assimilis. Their 
prostomia are practically identical, and they further agree 
in almost every internal character—the number and position 
of their nephridia, their cesophageal glands, the absence of 
pouches on the first diaphragm, the form and structure of the 
brain, the large size of their open otocysts and of their ova. 
The only differences are externally in the number of seg- 
ments and the position of the first gill, and internally in the 
vessels of the seventh segment and in the nature of the 
otoliths. It is a question whether these differences are of 
sufficient importance to justify the separation of the New 
Zealand specimens as a distinct species. 

The form of the otoliths is certainly very different. In 
A. assimilis they are rounded, and consist almost entirely 
of a substance secreted by the cells in the wall of the 
otocyst, while in the New Zealand specimens they are 
irregular foreign bodies (figs. 749, 751).  Hhlers has re- 


762 J. H. ASHWORTH. 


marked (1901, p. 178) on the small size of the opening 
from the tube into the otocyst in A. assimilis, and con- 
sidered that this was connected with the form of the 
otoliths. I had previously arrived at the conclusion that 
their shape was due to the closure of the lumen of the tube, 
and had examined a number of otocysts of A. marina to 
obtain further evidence on this question. ‘The anterior ends 
of nine specimens of the latter species have been sectioned, 
und show considerable differences in the character of their 
otoliths. Six of the specimens are comparatively young (from 
about 17 to 65 mm. in length), and their otoliths are irregular 
foreign bodies such as quartz-grains, portions of spicules, 
frustules of diatoms, etc., which are almost naked, i. e. they 
have either no secreted covering, or else it is a mere film, the 
presence of which is indicated by its staining with haema- 
toxylin. Of the remaining three older specimens, one, which 
is about 170 mm. long, has irregular otoliths like those 
described above, but in the other two, which are about 130 
and 250 mm. long respectively, the otoliths have quite a 
different appearance. They were at first irregular, but the 
original particles have been covered by layer upon layer of 
secreted substance, and the resultant otoliths have rounded 
outlines (see fig. 16). The tubes of these two pairs of 
otocysts are found to be practically closed along almost their 
whole length, either by apposition of the walls or by the 
blocking of the lumen by a granular substance secreted by 
the gland-cells in the wall of the tube. In each case the 
walls of the tube are so closely apposed that the lumen along 
the greater part of its length is reduced to a slit not more 
than 3 or 4 across, and even this space is occupied by a 
thin band of the secreted substance mentioned above, thus 
effectually closing the passage. The variation in the nature 
of the otoliths is probably dependent on the condition of the 
tube. At any rate, it is interesting to note that in the large 
specimen (170 mm, long) with irregular otoliths mentioned 
above, the lumen of the tube, as seen in section, is a fair-sized 
slit, and is not encroached upon to any extent by secretion 


ANATOMY OF ARENICOLA ASSIMILIS. 763 


such as blocks the tubes in the specimens with rounded 
otoliths. It is also worthy of note that in the other species 
of Arenicola (cristata, grubii, ecaudata) which have 
rounded or spherical otholiths, formed largely of secreted 
inatter, the otocyst is a closed vesicle. It seems probable, 
therefore, that the presence or absence of an open passage 
connecting the otocyst to the exterior has considerable 
influence upon the character of the otoliths, which varies 
even in different specimens of the same species. The fact 
that the otoliths of A. assimilis are rounded, while those of 
the New Zealand specimens are irregular, is not of funda- 
mental importance; it probably indicates that in the former 
the tube leading from the otocyst to the exterior very soon 
became blocked, and the otoliths are therefore largely com- 
posed of material deposited around the small particles which 
had gained access to the otocyst before the closure of its 
tube. The otocysts of A. assimilis and of the New Zealand 
specimens agree in the most important character, namely, 
that each possesses a tube leading to the exterior; and the 
modification which takes place in the former, causing a differ- 
ence in the nature and shape of the otoliths, may be regarded 
as of secondary importance, since a similar, though not so 
marked a difference, may be observed within the limits of a 
single species (A. marina). Tor further remarks on this 
subject see p. 771. 

While the number of chetigerous segments in the ecaudate 
Arenicolidee varies greatly (from about twenty-four to forty in 
A. grubii, and thirty-five to fifty-six in A. ecaudata), it is 
peculiarly constant in three of the caudate species, there being 
invariably nineteen in A. marina and A. claparedii, and 
seventeen in A. cristata. In most American specimens of the 
last-named species there is, however, an extension into the 
tail of structures which are usually associated only with 
parapodia, Small gills and cirriform processes occur upon 
the first two or three tail segments of one specimen examined 
(Gamble and Ashworth, 1900, p. 442, figs. 31, 32), and similar 


processes are commonly present on American specimens, but 


764 J. H. ASHWORTH. 


have never been recorded in any Neapolitan specimen of this 
species. 

With this example in mind it is not difficult to believe that, 
in a species probably widely ranging over the enormous 
coast-line of the South Atlantic and Pacific Oceans, some 
specimens may have become modified in the direction above 
indicated, so that finally a condition was reached in which 
some members of the species possess nineteen and others 
twenty segments. If we suppose that an additional para- 
podium and gill have been produced, the only alteration 
necessary to bring such a form into line with A. assimilis 
would be the loss of a gill at the anterior end of the series. 
The reduction and absence of the first gill are so frequently 
observed in A. marina (and to a less extent in almost all 
other species) tlat the reduction and eventual loss of the first 
eill of the hypothetical form are quite conceivable. 

In my opinion the possession of an extra cheetigerous seg- 
ment, though striking, is scarcely a sufficiently important 
character to form by itself a test of specific value, and to be 
used as the sole means of distinguishing two otherwise 
identical forms. It seems preferable to regard the New 
Zealand specimens as forming a variety of the species A. 
assimilis, to which the name affinis may be given indicat- 
ing its close connection with and resemblance to the type. 


V. Post-larval Stages of Arenicola from the Falk- 
land Islands. 


After finding multiple cesophageal pouches in adult speci- 
mens of Arenicola assimilis, it occurred to me that I 
might be in error in the determination of the species of 
certain post-larval Arenicolide from the Falkland Islands, 
aud a re-examination of them became necessary. The 
specimens were preserved in, and handed to me in, formalin, 
and | examined them in that fluid two years ago. On finding 
multiple cesophageal glands I had little hesitation in refer- 
ring them to the species A. claparedii, because at that 


ANATOMY OF ARENICOLA ASSIMILIS. 765 


time the occurrence of several pairs of cesophageal ceca was 
known only in this species, and, indeed, was considered to be 
one of its most characteristic features. At the same time I 
looked for the otocysts, but did not succeed in finding them." 
They would have been moderately easy to see in post-larval 
stages of A. marina of the same size, and finding no similar 
structures in my post-larvee I therefore concluded (wrongly, 
as it now appears) that otocysts were absent. With these 
two features in mind, but relying especially on the very 
obvious presence of several cesophageal pouches, I identified 
the specimens as post-larval stages of A. claparedii, and as 
such they were recorded by Miss Pratt (1901, p. 12). The 
fact that these specimens had been obtained in the region in 
which A.assimilis is found had not escaped my notice, 
but as Ehlers’ account (1897, pp. 105, 104) stated that his 
species closely resembled A. marina, it was naturally 
concluded that the presence of multiple cesophageal glands 
might still be regarded as a diagnostic character of A. 
claparedii. 

These post-larvee have now been carefully re-examined, 
both entire and in sections, with the result that my previous 
determination is found to be wrong; they are the youug 
stages of A. assimilis, var. affinis. 

The three specimens were found on the surface of the sea 
near the Falkland Islands, by Mr. R. Vallentin, of New Quay, 
and were handed over to me by Miss Pratt, of Owens 
College, Manchester, who was working over Mr. Vallentin’s 
collection of Polycheetes. 

The specimens are 7°6, 8°7, and 11:1 mm. long respectively. 
The largest specimen is provided with a transparent gelatin- 
ous envelope about 1 mm. in diameter, which covers the 
animal, except for a distance of a little over a millimetre at 
each end. In general aspect these post-larvee resemble 
those of A. marina, 


! The nephridiopores were also examined, but on account of their minute size 
it was impossible to make certain of their presence or absence on the critical seg- 
ment (the fourth), and therefore their number could not be definitely ascertained. 


vou. 46, PART 4,—NEW SERIES. DDD 


766 I, H. ASHWORTH. 


There are sparsely scattered greenish-brown pigment cells 
in the epidermis. 

The conical prostomium bears from two to four small 
brownish-red eyes on each side (fig. 18). It is followed by a 
region divided into two by a faint groove (figs. 18, 19). 
The anterior portion of this region is the true peristomial 
segment, and in the largest specimen is itself encircled by a 
groove, which subdivides it into two annuli. ‘The posterior 
part of the region above named corresponds to the segment 
bearing the minute vestigial seta in the’ post-larve of 
A. marina (Benham, 1893, p. 49). There is no trace of sete 
in this segment in the post-larvee now under consideration. 
In adults (see fig. 20) the region between the prostomium 
and first chetigerous segment is divided into four rings 
(see p. 741), in the third of which the cesophageal connec- 
tives unite. By comparison with these post-larvee, it is seen 
that the first two rings of the adult belong to the peristo- 
mium and the other two to the first true body-somite, which, 
in Arenicola, has lost its setae and has become fused with 
the peristomium. 

There are nineteen cheetigerous segments, in each of which 
crotchets and sete may be clearly distinguished. ‘There are 
two kinds of sete present in the notopodia. The more 
numerous and longer ones are very similar to those of the 
adult (fig. 3). They are about 0°3 mm. long, and bear a 
lamina along about half their length. The shorter sete, about 
0°25 mm. long, are obviously laminate for a short distance on 
both sides (figs. 4, 5). They are almost lanceolate in shape, 
and drawn out into long, slender tips. Only one of these 
sete is usually present in each notopodium, in which there 
are two to four sete altogether. There is a tendency, more 
marked in the lanceolate sete, for the lamina to break up, 
from the edge inwards, into fine, pointed teeth (figs. 4, 6). 

The crotchets are 0:07 to 0°08 mm. long, andare distinguished 
by the presence of a thickening, forming an encircling ridge 
upon the shaft of the cheeta (fig. 10). This ridge lies just 
below the level of the epidermis. As described on p. 756, 


ANATOMY OF ARENICOLA ASSIMILIS. 767 


the teeth are not confined to the region immediately behind 
the rostrum, as on careful focussing they may be found also on 
the sides of the rostrum. There is really, therefore, a circular 
series of teeth from the centre of which the rostrum projects, 
and the subrostral process is the lowest of this series. 
Fie. 104 shows the appearance of the crotchet when the 
rostrum is in sharp focus; in fig. 108 the other teeth seen on 
focussing slightly upwards are added. 

Each of the posterior tail segments is divided by shght 
constrictions so as to present a tri-annulate appearance. The 
anal seement is somewhat swollen, and the lips of the anus 
are crenate. 

There are no gills present in any of the specimens. ‘The 
blood is light red in colour (in formalin). 

After staining and clearing the specimens the alimentary 
canal could be well seen (fig. 18). The muscular pharynx 
leads into the thick-walled cesophagus, which bears on the 
dorsal surface of its posterior portion the cesophageal glands, 
of which there are from six to eight visible on each side ; the 
anterior one is the largest. Just behind this point the 
cesophagus is slightly constricted, and the two hearts lie on 
its lateral walls. The stomach is a wider tube, and upon its 
walls may be clearly seen the vessels of the gastric plexus 
bounding the chlorogogenous areas. ‘The intestine, like that 
of the adult, is thrown into concertina-like folds. 

Sections show that the anterior part of the cesophagus is 
ciliated, and that the stomach is lined by columnar cells, 
many of which contain a vacuole near the end which adjoins 
the digestive cavity. 

In sections of the anterior ends of the two smaller speci- 
mens the otocysts are not easily found ; they are much less 
clearly differentiated at this stage than those of corresponding 
post-larval stages of A. marina and A. ecaudata. They 
are seen to be two small pits in the epidermis, the lips of 
which are approximated so as to form a very short tube 
(fiz. 21). Hach otocyst is somewhat triangular in section; 
its apex is directed laterally and leads to the external open- 


768 J. H. ASHWORTH. 


ing. There are in each otocyst from four to six foreign 
bodies (otoliths), all of which are apparently quartz-grains 
except two; these are obviously fragments of spicules. The 
otocysts of the specimen 11:1 mm. long are faintly visible in 
a stained preparation of the whole animal cleared in thick 
cedar-wood oil. They are about 404 by 25, in internal 
diameter (fig. 18). 

The nuchal organ is easily recognisable ; its cells are richly 
ciliated (figs. 18, 21). 

Neither giant-cells nor giant-fibres can be identified in the 
nerve-cord at this stage. 

Six pairs of nephridia may be traced in sections. The first 
nephridium is small, and its anterior end runs forward and 
pierces the third diaphragm. On the sixth nephridium the 
gonidial vessel has a covering of cells which have large 
spherical nuclei. This is the gonad, and it may also be dis- 
tinguished, though not so clearly, on the fourth and fifth 
nephridia. 

The above-described post-larval stages are evidently not 
young specimens of A. claparedii,as is shown by the presence 
of otocysts and six pairs of nephridia. They are the young 
stages of the variety of A. assimilis. 

Ehlers (1897, p. 104) has recorded from Uschuaia a similar 
aill-less specimen about 6°5 mm. long, which bears nineteen 
cheetigerous segments. This post-larval stage was found 
amone the “roots” of seaweeds (Tangwiirzeln), and had 
probably recently settled down to its littoral habitat. 


VI. Adult Specimens of Arenicola from the Falkland 
Islands. 


When the foregoing account was ready for press I received, 
through the kindness of Mr. R. Vallentin, of New Quay, 
Cornwall, five adult specimens of Arenicola from the Falk- 
land Islands, and have thus been able to confirm some of the 
observations described in the former part of this paper. 


ANATOMY OF ARENICOLA ASSIMILIS. 769 


The specimens were dug from the sand in Whale Sound, 
Stanley Harbour, during the early months of this year (1902). 
‘They are respectively 187, 185, 135, 128, and 121 mm. long. 
Each has nineteen chetigerous segments, the seventh of 
which bears the first and invariably small pair of gills. The 
other external characters, e. g. the prostomium, annulation, 
etc., agree exactly with those of the Otago specimens, while 
internally the agreement is scarcely less perfect. In the two 
specimens dissected there are six pairs of nephridia opening 
on the fourth to the ninth segments. The first nephridium 
is small, and its nephrostome is on the anterior face of the 
third diaphragm. ‘The edge of the ventral lip of the larger 
nephrostomes is thrown into numerous folds or frills, as 
figured (see fig. 17). The vascular system agrees exactly with 
that of the Otago specimens. 

There are multiple cesophageal glands to the number of 
twelve or thirteen on each side, the anterior ones digitiform 
or club-shaped, the others pyriform or ovoid. There are no 
pouches on the first diaphragm. 

The only feature of an unusual character in the body of the 
animal is the presence of a partial septum one segment 
behind the third diaphragm. ‘This structure is homologous 
to the septa met with in the posterior branchial region of 
this and other Arenicolide. It is amembrane supporting both 
the afferent and efferent vessels to the second pair of 
nephridia, and is nearly 3 mm. across in its widest part. It 
is not so extensive as the third diaphragm (which in the 
same specimen is over 6 mm. across), as it does not reach 
either the dorsal or the ventral body-wall. It may be 
regarded as merely an exaggeration of the supporting 
strands which are usually present in other species alongside 
either one or both of the vessels to the nephridia (see, for 
example, A. grubii, Gamble and Ashworth, 1900, pl. xxvi, 
fies. 53, 54). 

The two specimens examined are females which have 
probably spawned, as only a very few ova are present in the 


1 In one specimen the true first gill is absent on the right side. 


770 J. H. ASHWORTH. 


body-cavity. ‘These are large, and measure across their flat 
faces 0°2 x 0°17 mm. (see pp. 747 and 758). 

The anterior end of one of the specimens was cut into sections. 
A pair of large otocysts is present (fig. 14). ‘hey are much 
larger than those of the Otago specimens and a little larger 
than those of the worms from Uschuaia (cf. figs. 13, 14, 15). 
Their three diameters are respectively about 0°36, 0°38, and 
0:28 mm. (compare the measurements on pp. 752 and 759). 

The otoliths are all spherical or nearly so, and are com- 
posed of a yellowish or brownish secreted substance. There 
are in each otocyst two otoliths (fig. 14) considerably larger 
than the rest. They are about 0°055 mm. in diameter, and in 
the centre of each is a small irregular foreign body, probably 
a quartz-grain. The smaller otoliths are usually from 0°02 
to 0°03 mm. in diameter, and only rarely is a foreign particle 
visible in them, though doubtless each has a very minute 
central nucleus of this description. The two large otoliths 
described above are probably the first otoliths of the post- 
larval stage, which always remain distinguished by their 
ereater size from those which are formed later. A similar 
condition exists in A. ecaudata, in the post-larval stage of 
which there is for some time only one otolith, which always 
remains conspicuous, owing to its larger size (Gamble and 
Ashworth, 1900, p. 504 and fig. 64). 

The otocyst opens to the exterior by a tube, the external 
opening of which corresponds in position to that of the Otago 
specimens and of A. marina. ‘The lumen of the tube is of 
moderate size along the greater part of its length, but is 
reduced near its entrance to the otocyst in one case to a very 
narrow passage, and in the other is practically obliterated. 
The wall of the tube is remarkable for the presence of 
large gland-cells, which are practically confined to the dorsal 
wall. ‘They are almost ovoid in shape, and their cell- 
contents are in the form of a reticulum. In the ventral 
wall of the tube there are numerous elongate fusiform sense 
cells. 

The remaining structures shown in sections of the anterior 


ANATOMY OF ARENICOLA ASSIMILIS. val 


end are so exactly similar to those of the Otago specimens 
that no further description of them is necessary. 

The specimens above described are interesting from their 
bearing on the discussion regarding the taxonomic value of 
the shape of otoliths (see p. 761). The only difference 
between the Falklands specimen and those from New Zealand 
is that in the former the otoliths are spherical and composed 
almost entirely of a secreted substance, while in the latter 
they consist of irregular foreign bodies, such as sand-grains 
and fragments of spicules. There can be no doubt that the 
two sets of specimens belong to the same species, or rather 
to the same variety, so that (as was also proved for A. 
marina, see p. 762) the shape of the otoliths varies in 
different specimens of the same species or variety. The 
closure of the tube of the otocyst along part of its length and 
the presence of the numerous large gland-cells in its wall are 
probably the principal factors in determining the shape and 
nature of the otoliths of the Falklands specimens. Having 
proved the presence of spherical otoliths in some examples of 
A. assimilis, var. affinis, it will be noticed that one of the 
differences (discussed on pp. 761—763) between this new 
variety and the type of the species disappears ; so that now the 
only features by which they may be distinguished are (1) the 
presence of twenty chetigerous segments in the type of the 
species, whereas the new variety possesses only nineteen, and 
(2) the slightly different position of the external opening of 
the otocyst. As the latter is too fine a character for ready 
application in systematic work, it may be said that the deter- 
mination rests upon the number of chetigerous segments. 

Another striking feature about the otocysts is the great 
difference in their size in specimens of the variety from the 
two localities. Whereas in the Otago specimens their 
average diameter is 0°21 mm. (in a specimen 136 mm. long), 
in one (128 mm. long) from the Falklands their average 
diameter is 0°34 mm. (cf. figs. 14, 15), so that the internal 
volume of the latter is about four times that of the former. 

Adult specimens of the new variety are now recorded from 


772 J. H. ASHWORTH. 


the same locality as the post-larval stages described on pp. 764 
—768. There can be no doubt that the latter are stages in 
development of the former. Judging from Ehlers’ record 
(1897, p. 104) of the capture near Uschuaia of a aill-less 
specimen 6°5 mm. long with nineteen chetigerous segments, it 
seems probable that the variety occurs at this place along 
with typical specimens of the species. 


VII. Distribution of Arenicola assimilis. 


Ehlers (1901, p. 178) records the occurrence of A. 
assimilis in collections from the Straits of Magellan (Punta 
Arenas and Susanna Cove), the Beagle Channel (Uschuaia 
and Lapataia Nueva), South Georgia, Chile (Schmarda), 
Kerguelen (Grube), and California. 

Schmarda’s (1861, pp. 51, 52) A. piscatorum from Chile 
and Grube’s (1878, pp. 511, 554) A. piscatorum, Cuv., var., 
from Kerguelen, are both included by Ehlers under the 
species A.assimilis. Although Schmarda gives a brief 
description of some points in the anatomy of his specimens 
he unfortunately does not mention any characters which 
enable their identity to be definitely settled. With respect 
to Grube’s specimen from Kerguelen the only information 
given is that most of the branchiferous segments are divided 
into only four annul, and owing to this feature Grube dis- 
tinguished his specimens as a variety of A. piscatorum, 
‘There is no evidence to show that any of these specimens 
belong to the species A. assimilis, 

Khlers (1897, p. 104) states that in the Gottingen collec- 
tion there is a species of Arenicola! from California in 
which there are twelve pairs of gills which agree in position 
with those of A. assimilis, and these specimens are dis- 

1 | thank Professor Ehlers for sending to me by letter the further informa- 
tion that this is a duplicate from Professor Agassiz’s collection, which was 
sent to Godttingen to be worked over, ‘The rest of the specimens were 
returned to Professor Agassiz, and are doubtless those referred to on the next 
page. 


ANATOMY OF ARENICOLA ASSIMILIS. 773 


tinguishable from A. marina only by this character. No 
mention is made of other features which would have been 
much more valuable as diagnostic characters, but the 
difference in the number of gills is accepted as a sufficient 
ground for separating the specimens from A. marina, not- 
withstanding the well-known liability to reduction (from 
thirteen to twelve pairs) in the number of branchie in this 
species. As will be seen from the discussion below, it is very 
probable that Ehlers’ specimen does not belong to either of 
these species, and that this is an example of the confusion 
due to placing an implicit reliance on the value of external 
features in discriminating species of Arenicola. On such 
a variable and insufficient character as the number of gills 
Ehlers bases his diagnosis of the Californian specimen, and 
refers it to the species A.assimilis. ‘This is the only 
evidence in support of his record of this species from Cali- 
fornia. 

I have recently re-examined specimens of Arenicola 
from a collection made by Professor Agassiz, near Crescent 
City, California, sent to Dr. Gamble and myself from the 
Harvard Museum, and identified by us (1900, p. 423) as 
A. claparedii. These specimens are the more interesting 
because they are accompanied by a label? indicating that they 
have passed through the hands of Professor Ehlers, and that he 
considered them to belong to a new species nearly related to 
A. marina (= piscatorum). Itisalmost certain that these 
are the same specimens which Ehlers has recorded as A. ass1- 
milis. ‘There are five specimens, in three of which there 
are twelve pairs of gills, the first situated on the eighth 
chetigerous segment. In each of the other two specimens 
there are twelve gills on the left side (the first being on the 
eighth segment), accompanied in one case by thirteen gills 
on the right, the first being very small and borne on the 
seventh segment, while on the right side of the other 

‘The writing upon the label, which is now faint, is as follows :— 


*Arenicola, n. sp. nahe piscator. 7 vor Segm. 12(13) Kiementrag. 
Californien (H. Ehlers).” 


774 J. H. ASHWORTH. 


specimen there are only eleven gills, the first of which is on 
the ninth segment. It may therefore be said that it is usual 
to find the first gill in these specimens on the eighth segment 
as in A. assimilis. Dissections of two of the specimens 
show that there are five pairs of nephridia, multiple ceso- 
ageal glands, and no pouches on the first diaphragm; and 
sections of the anterior end prove conclusively that there are 
no otocysts. All these points are so characteristic of 
A. claparedii that there can be no doubt that the speci- 
mens belong to this species. 

I am indebted to Dr. H. P. Johnson for two specimens of 
Arenicola from Puget Sound, Washington. In one of 
these! there are thirteen pairs of gills, but in the other the 
seventh segment bears a gill only on the right side, the first 
left gill being on the following segment. Dissections of the 
specimens and sections of the anterior end of one of them 
fully confirm the determination of their species made by Dr. 
Johnson (1901, p. 421) ; they are undoubted A. claparedii. 

It is therefore highly probable that Ehlers is in error in 
recording A. assimglis from California. In the first place, 
his determination of the species of the Californian specimens 
rests solely upon a character which is very variable and 
almost useless for distinguishing species ; secondly, a re- 
examination of what are probably the very same specimens 
proves them to be A. claparedii, and this species has been 
recorded from another point on the west coast of the 
United States. 

A revision of Ehlers’ record of the distribution of A. assi- 
milis therefore becomes necessary, and may be given as 
follows :—Adult typical specimens of A. assimilis have been 
recorded from several places in the extreme south of the 


' It is remarkable that of the seven specimens examined from the west 
coast of the United States this is the only one which possesses the full 
number of gills, On the contrary, it is unusual to find any departure from the 
normal number in Neapolitan specimens of A. claparedii; out of thirty-nine 
examined only two show a reduction in the number of gills; in each case 
there are thirteen on the left side, but only twelve on the right. 


ANATOMY OF ARENICOLA ASSIMILIS. tO 


American continent and from South Georgia. Others 
forming a new variety but agreeing with the type, except in 
the number of chetigerous segments, are now recorded from 
Otago Harbour, the Macquarie Islands, and the Falkland 
Islands. Post-larval stages of the variety have been obtained 
off Stanley Harbour (Hast Falkland) and near Uschuaia. 


VIII. Specific Characters of the Caudate Areni- 
colide. 


Appended isa revised summary of the characters of the 
caudate Arenicolide, which clearly shows by what features, 
both external and internal, A. assimilis may be readily 
recognised and distinguished from other species with which 
it is hiable to be confused. It cannot be too strongly urged 
that attention should be directed by systematists chiefly to 
internal characters in the discrimination of the species of 
Arenicola. No determination of A. marina, A. clapa- 
redii, or A. assimilis can be considered complete or 
entirely trustworthy which relies solely on external characters. 
It is impossible to distinguish these three species with 
certainty unless reference be made to the nephridia, 
cesophageal glands, and otocysts, the two former being of 
especial use in this connection. 

The characters! of the caudate Arenicolide may be briefly 
stated thus: 

A distinct tail present; the parapodia and gills do not 
extend to the posterior end of the animal. The body is 
often swollen anteriorly. Gills, pinnate or derivable from 
the pinnate type, eleven to thirteen pairs, the first (which 
may be small or even absent) on the seventh or eighth 
chetigerous segment. Prostomium consisting of a median 
and two lateral lobes. Nephrostomes with dorsal lip well 
provided with flattened, spatulate, ciliated, vascular processes ; 

" The following is a revision of a part of the summary published by Dr. 


Gamble and myself (1900, p. 540), to which reference may be made for the 
characters of the genus and of the ecaudate species. 


776 J. H. ASHWORTH. 


ventral lip ciliated, entire (i. e. not deeply notched as in the 
ecaudate Arenicolidz). Gonads small, ova discoidal. 

(a) A. marina, Linn.—Nineteen chetigerous segments. 
Thirteen pairs of gills; the first, which is on the seventh 
segment, may be reduced (or suppressed). Otocysts opening 
to the exterior. Otoliths, numerous foreign bodies (quartz- 
grains, etc.), which may, however, be covered with a layer 
of secreted chitinoid substance, giving them a rounded out- 
line. Six pairs of nephridia opening on segments 4 to 9. 
One pair of cesophageal pouches, cylindrical, club-shaped, 
or conical. Diaphragmatic pouches (on the first diaphragm) 
small, globular, or flask-shaped. 

Found on both sides of the North Atlantic. 

(b) A. assimilis, Ehlers.—Twenty chetigerous segments. 
Thirteen pairs of gills, the first of which is situated on the 
eighth segment (the first gill is liable to be reduced or 
suppressed). Otocysts large, opening to the exterior. Otoliths 
numerous ; spherical or rounded chitinoid bodies. Six pairs 
of nephridia opening on segments 4 to 9. Several pairs 
of cesophageal pouches ; the anterior pair long, club-shaped, 
or filiform; the others much smaller and pear-shaped. No 
pouches on the first diaphragm. 

Recorded from the extreme south of the American 
continent. 

(c) A. assimilis, var. affinis, Ashworth.—Nineteen 
chetigerous segments. Thirteen pairs of gills, the first 
(liable to reduction or suppression) on the seventh segment. 
Otocysts large, opening to the exterior. Otoliths numerous, 
aud composed either of foreign bodies (quartz-grains, etc.) 
or of spherical chitinoid bodies. Other characters as in the 
type of the species (see above). 

Recorded from Otago Harbour, New Zealand, the Mac- 
quarie Islands, the Falkland Islands. 

(d) A. claparedii, Levinsen—Nineteen chetigerous seg- 
ments. ‘Thirteen pairs of gills, the first on the seventh 
segment (this pair of gills is liable to reduction or suppres- 
sion, especially in specimens from the west coast of North 


ANATOMY OF ARENICOLA ASSIMILIS. What 


America). Lateral lobes of prostomium well developed. No 
otocysts. Five pairs of nephridia opening on segments 5 
to 9. Two or more pairs of oesophageal pouches, the 
anterior pair long and slender or club-shaped, the others 
shorter, usually pyriform. No pouches on the first diaphragm. 

Recorded from the Mediterranean and from the west coast 
of the United States. 

(e) A. cristata, Stimpson.—Seventeen chetigerous seg- 
ments. Eleven pairs of gills, the first on the seventh 
segment. Otocysts, closed spherical sacs each containing a 
single large, spherical, chitinoid otolith. Six pairs of 
nephridia opening on segments 5 to 10. One pair of 
cesophageal pouches cylindrical or club-shaped. Diaphrag- 
matic pouches (on the first diaphragm) large and finger- 
shaped. 

Found in the Mediterranean, in the West Indies, and on the 
eastern shores of North America south of latitude 40° N. 


IX. Summary of Results. 


1. The anatomy of Arenicola assimilis is fully de- 
scribed for the first time. Although Ehlers states that the 
species differs from A. marina only in the number of 
chetigerous segments (nineteen in the latter, twenty in the 
former), in the position of the first gill and in the relative 
size of the middle lobe of the prostomium, further examina- 
tion shows that there are other important points of difference, 
e. g. A. assimilis possesses multiple cesophageal glands 
and has no pouches on the first diaphragm. 

2. Specimens of Arenicola are described from Otago 
Harbour (New Zealand) and from the Macquarie Islands 
which differ from the type in the number of chetigerous 
segments (nineteen) and situation of the first gill. There is 
also a difference in the shape of the otoliths ; in the type they 
are spherical or rounded, while in the New Zealand specimens 
they are irregular. These specimens belong to a new variety 
(var. affinis) of the species. 


778 J. H. ASHWORTA. 


3. In the discussion of the systematic position of the 
Otago specimens it is concluded that the form of the 
otoliths is not sufficiently reliable to form a character by 
which species may be discriminated. In A. marina the 
otoliths are usually irregular, but two out of nine specimens 
examined possess rounded otoliths. In these cases the 
otoliths were at first irregular foreign bodies, but they 
have been covered with layer upon layer of secreted sub- 
stance, and now have a rounded outline. In each of these 
two cases the tube which placed the otocyst in communica- 
tion with the exterior has become either wholly or partially 
blocked, either by apposition of its walls or by the secretion 
into the lumen of a glandular substance which forms an 
effectual plug. In the seven specimens of A. marina with 
irregular otoliths the tubes are not closed in this way. It is 
concluded that the presence of spherical or rounded otoliths 
is associated with the closure of the tube, and it is pointed 
out in support of this conclusion that the other species 
(cristata, grubii, and ecaudata) in which spherical 
otoliths are found have closed otocysts (pp. 761—763). 

4, The brain is well developed. The ganglion-cells of its 
middle region are large (especially in the Otago specimens) 
and send processes into the neuropile, where they branch 
freely. 

5. Giant-fibres and segmentally arranged giant-cells are 
present in the nerve-cord. They have the same structure as 
in A. erubii. 

6. The otocysts of A. assimilis are distinguished by their 
size. ‘They are considerably larger than those of any other 
species. Neuro-fibrille may be traced from the nervous 
sheath of the otocyst into and along the whole length of the 
sense cells of the otocystic epithelium. These cells and 
fibrils are especially abundant near the point of entrance of 
the tube to the otocyst. 

7. There is a large nerve-supply to the skin and proboscis. 
Neuro-fibrilla may be seen in some of the cells of the general 
body-surface and of the papilla of the proboscis, 


ANATOMY OF ARENICOLA ASSIMILIS. 779 


8. Post-larval specimens of A. assimilis, var. affinis, are 
described from the Falkland Islands. They possess an 
acheetous segment between the peristomium and_ first 
chetigerous segment (as in similar stages of A. marina and 
A. ecaudata). By comparison with the adult the hmits in 
the latter of the peristomium and acheetous body-segment 
may be determined (figs. 19, 20). 

9. Adult specimens of A. assimilis, var. affinis are also 
described from Stanley Harbour, East Falkland. They are 
remarkable for the large size of their otocysts, the internal 
volume of which is about four times that of the otocysts of 
the Otago specimens (figs. 14, 15). It is evident that a con- 
siderable variation in the size of these organs may occur in 
specimens of the same species or variety from different 
localities. The otoliths, several of which contain an irregular 
foreign body, are spherical, and in one specimen two of them 
are much larger than any of the others. ‘They are the first 
two otoliths of the post-larval stage which have continually 
received fresh depositions of secreted substance, and always 
remain distinguished from those formed later by their larger 
size. These specimens from the Falklands differ from the 
Otago specimens in the nature of their otoliths. Here is 
additional evidence that the character of the otoliths con- 
tained in otocysts provided with a tube leading to the 
exterior is not a feature upon which much value should be 
placed in systematic work. Blocking of the tube (as occurs 
in the Falklands specimen) converts the otocyst into a closed 
sac, in which spherical otoliths are formed, while in other 
specimens (e.g. those from Otago Harbour) in which the 
tube remains open the otoliths are irregular foreign bodies, 
such as sand-grains, which are able to gain access to the 
otocyst. 

10. Ehlers records A. assimilis from the Straits of 
Magellan, the Beagle Channel, South Georgia, Chile 
(Schmarda), Kerguelen (Grube), and California. It is shown 
that there is no evidence in support of the last three 
records. Schmarda’s and Grube’s specimens are insufficiently 


780 J. H. ASHWORTH. 


described, and no character is mentioned by which their 
species may be determined. Ehlers’ diagnosis of the 
Californian worms rests solely upon a character which is 
very variable and almost useless for distinguishing species. 
A re-examination of what are probably the same specimens 
proves them to be A. claparedii, and this species has also 
been recorded from another point on the west coast of the 
United States (pp. 772—774). Incidentally it may be men- 
tioned that specimens of A. claparedii from the west coast 
of North America almost invariably bear only twelve gills 
either on one or both sides, while Neapolitan specimens have 
usually the full number (thirteen pairs). 

A. assimilis may be regarded as the characteristically 
southern species of the genus. Adult typical specimens are 
recorded from several points in the extreme south of America 
and from South Georgia. A new variety (var. affinis), 
differing from the type only in the number of cheetigerous 
segments, is now recorded from Otago Harbour (New Zea- 
land), the Macquarie Islands, and the Falkland Islands. 
Post-larval stages of the variety have been taken off the 
Falklands and near Uschuaia, in the Beagle Channel. 


X. LITERATURE. 


1861. Scnmanrpa, L. K.—‘ Neue Wirbellose Thiere,’ Band i, p. 51, Leipzig, 
1861. 

1878. Grune, K.—“ Anneliden-Ausbeute S.M.S. Gazelle,” ‘ Monatsber, d. K. 
Akad. d. Wissensch. zu Berlin, aus dem Jahre 1877,’ pp. 511, 554, 
Berlin, 1878. 

1893. Bentnam, W. B.—‘ Journal Marine Biol, Association,’ New Series, 
vol. ili, p. 48, 1898. 

1897. Enters, E.—* Polycheten,’ ‘Hamburger Magalhaenische Sammel- 
reise,’ pp. 108, 104, Hamburg, 1897. 

1898. Game, F. W., anv Asnwortn, J. H.—‘ Quart. Journ. Mier. Sci.,’ 
vol. xli, p. 1, 1898. 

1899. Fauver, P.—* Mémoires de la Société nationale des Sciences naturelles 


et mathématiques de Cherbourg,’ tome xxxi, p. 178, Cherbourg, 
1899. 


ANATOMY OF ARENICOLA ASSIMILIS. 781 


1900. Cutnp, C. M.—‘ Archiv f. Entwickelungsmechanik,’ Band ix, p. 587, 
1900. 

1900. Gameie, F. W., anp AsHwortH, J. H.—‘ Quart. Journ. Mier. Sci.,’ 
xiii, p. 419, 1900. 

1901. Prarr, E. Mi—‘ Memoirs and Proceedings of the Manchester Lit. and 
Phil. Society,’ vol. xlv, p. 12, 1901. 

1901. Jounson, H. P.—‘ Proceedings of the Boston Society of Natural His- 
tory,’ vol. xxix, p. 421, 1901. 


1901. Enters, £.—‘ Die Polychaten des magellanischen und chilenischen 
Strandes,’ pp. 177, 178, Beriin, 1901. 


EXPLANATION OF PLATES 36 & 37, 


Hlustrating Dr. J. H. Ashworth’s memoir on ‘The Anatomy 
of Arenicola assimilis, Ehlers, and of a New Variety 
of the Species, with some Observations on the Post- 
larval Stages. 


List oF REFERENCE LETTERS. 


A. B.S. Achztous segment of body between peristomium and _ first 
chetigerous segment. Axt. Cer. Z. Anterior lobe of brain. B/. Bladder of 
nephridium. B.V. Blood-vessel. C. Cuticle. Ch. Seg.’ First cheetigerous 
segment. Cal. Ceelom. Conx. Tiss. Connective tissue. pid. Epidermis. 
£.T. Entrance to tube leading from otocyst to exterior. Ht. Op. Of. External 
opening of otocyst. Gang.C. Ganglion-cell. Gast. Lat. Lateral gastric 
vessel. G@/.C. Gland cell. Goxz. Gonad. Goz. V. Gonidial vessel. Hé. 
Heart. Met. Gr. Metastomial groove. Middle Comm. Middle commissure 
of brain. M/. Cire. Circular muscles. MM. Long. Longitudinal muscles. AZo. 
Mouth. J/. Sh. Buce. Muscular sheath of buccal mass. NV. Band of nerve- 
fibres from prostomial epithelium to brain, WV. 4f. Afferent vessel of 
nephridium. MW. Buee. Nerve to buccal mass and papillae of ‘ proboscis.’ 
N.C. Nerve-cord. WN. Epid. Nerve to epidermis. Nm. Ch, Neuropodial 
chet. Not. S$. Notopodial sete. Nphm.D. Dorsal lip of nephrostome. 
Nphn. V. Ventral lip of nephrostome. pile. Neuropile. Nue. Gr. Nuchal 
groove. Oc. Kye. i. (Hsophagus. WM. Conn. (Msophageal connective. 
.G/, (Esophageal gland. O¢. Otocyst. O¢h. Otolith. O¢. 7. Tube of 
otocyst. Pap. Papilla of “proboscis.” Per. Peristomium, PA. Pharynx. 
Post. Cer. L. Posterior lobe of brain. Prost. Prostomium. Prost. Epith. 
Epithelium of prostomium. Prost. Lat. Lateral lobe of prostomium. 


vou. 46, pARkT 4,—NEW SERIES. EEE 


782 J. H. ASHWORTH. 


Prost. Mid. Middle lobe of prostomium. Sp. Fragment of spicule.  S¢. 
Stomach. V.V. Ventral vessel. 


PLATE 36. 


Fre. 1.—Distal half of a seta from the fifth notopodium of a specimen of 
Arenicola assimilis from Uschuaia. x 100. 


Fic. 1 a.—A portion of the seta more highly magnified. Note the lamina 
on the left bearing fine teeth. x 750. 


Fie. 2.—Distal portion (two fifths) of a seta from the sixth notopodium of 
a specimen of A.assimilis, var. affinis, from Otago Harbour. xX 100. 


Fig. 2a.—A portion of the same seta more highly magnified. Nete the 
broad lamina crossed by fine oblique lines. x 500. 


Fics. 3, 4, 5.—Notopodial sete from a post-larval specimen (7°6 mm. long) 
of A. assimilis, var. affinis. Most of the sete are of the kind shown in 
Fig. 3, but in each notopodium there is one seta of the type seen in Fig. 5. 
Sete of the kind shown in Fig. 4 are much less common than the preceding ; 
only two specimens were seen in ten notopodia. The lamina is breaking up 
on one side near its tip into fine teeth. x 320. 


l'1c. 6.—Tip of a seta of the same kind as shown in Fig. 5. Note the fine 
teeth on the margin of the lamina on one side. x 600. 


Fic. 7.—A crotchet from the fifteenth neuropodium of an Otago specimen 
(var. affinis). x 100. 


Fic. 8.—The head of a crotchet from another Otago specimen, to show the 
teeth situated on the sides of the rostrum. x 280. 


Fic. 9.—A crotchet from the fifteenth neuropodium of A. assimilis from 
Uschuaia. xX 100. 


Fic, 10.—Two erotchets from the sixteenth neuropodium of a post-larval 
specimen (7°6 mm. long) of A. assimilis, var. affinis, 4 shows the appear- 
ance of the cheta when the rostrum and post-rostral teeth are in focus; in B 
the teeth on the sides of the rostrum are also shown, ‘The subrostral process 
is now seen to be one of the series of teeth, ‘The dotted line indicates the 
level of the epidermis, x 600. 


Fie. 11.—Diagram of a portion of the nerve-cord of A. assimilis, var. 
affinis, from Otago Harbour, to show the distribution of the giant nerve- 
cells. The transverse lines indicate the position of the neuropodia, the 


numbers of which they bear. The nerve-cord is magnified about 10 times 
and the cells 40 times. 


Fre. 12.—Diagram of a portion of the nerve-cord of A. assimilis from 
Uschuaia, to show the distribution of the giant nerve-cells. The trans- 
verse lines indicate the position of the neuropodia, ‘The last giant-cell shown 


ANATOMY OF ARENICOLA ASSIMILIS. 783 


situated (probably) in the first tail segment. The nerve-cord is magnified 
about 10 times and the cells 40 times. 


Fic. 138.—Shows the size of the otocyst of a specimen (120 mm. long) of 
A. assimilis from Uschuaia. The oval outline is a camera drawing of the 
cuticle which lines the otocyst. The tube which leads from the exterior 
enters the otocyst near the point marked #7. 7. The otoliths are rounded, and 
many of them show concentric markings, indicating their deposition in layers. 
In the centre of several of them small foreign bodies may be distinguished. 
Attached to some of the otoliths are other small rounded bodies of a similar 
nature, but which stain more deeply. The minute deeply staining spherules 
(indicated by the dots) are also composed of a similar substance. x 210. 
(Cf. Figs. 14, 15.) 


Fie. 14.—Camera drawing of the cuticle lining the otocyst of a specimen 
(128 mm. long) of A. assimilis, var. affinis, from the Falkland Islands. The 
otocyst is somewhat larger than the one shown in Fig. 13. The otoliths are 
nearly spherical. (They are not all present in one section; some are added 
from another section.) The two larger ones are probably the first otoliths of 
the post-larva, which are easily distinguished by their size from those which 
are formed later. In the centre of each of the large otoliths an irregular 
foreign body may be seen. x 210. (Cf. Figs. 13, 15.) 


Fie. 15.—Camera drawing of the cuticle lining the otocyst of a specimen 
(136 mm. long) of A. assimilis, var. affinis, from Otago Harbour. The 
otocyst is much smaller than either of the two preceding. The otoliths are 
irregular bodies, chiefly quartz-grains, but two small fragments of spicules 
(Sp.) are seen lying close together. x 210. (Cf. Figs. 18, 14.) 

Fic. 16.—Camera drawing of the cuticle lining the otocyst of A. marina 
(about 10 inches long). The tube connecting this otocyst to the exterior is 
almost blocked, and in consequence the otoliths, which were originally small 
irregular foreign particles, are now assuming a rounded outline, due to the 
deposition upon them of layer upon layer of secreted substance; see, for 
example, the otolith containing the spicule fragment (Sp.). Note the small 
size of the otocyst compared to those shown in Figs. 18, 14, and 15. x 210. 


Fie. 17.—Fifth nephridium of A. assimilis (specimen from Punta 
Arenas). The dorsal lip of the nephrostome (Nphm. D.) bears the usual 
ciliated vascular processes, while the edge of the ventral lip (Nplm. V.) is 
thrown into numerous folds or frills. Note the gonad, a somewhat club- 
shaped mass of cells around the gonidial vessel (seen by transparency through 
the gonad). x 10. 


Fic. 18.—Left aspect of the anterior portion of a post-larval specimen of 
A. assimilis, var. affinis, from the Falkland Islands. ‘The total length of 
this specimen is 11‘l1 mm. Note the four cup-shaped eyes on the prostomium. 
Between the prostomium and the first chetigerous segment is a region 


784 J. H. ASHWORTH. 


imperfectly divided into two by a groove ; the anterior part—the peristomium 
(Per.)—is again partially subdivided into two annuli, in the anterior of which 
the otocyst (O/.) may be faintly seen ; the posterior part (4. B. S.) is the first 
body-segment, which, however, does not bear any traces of sete. The suc- 
ceeding segments hear notopodial and neuropodial sete. The former usually 
consist of two or three capilliform bristles and one lanceolate seta (see also 
Figs. 3—6). The alimentary canal is seen by transparency through the body- 
wall. The buccal mass leads into the cesophagus, which dilates in the fourth 
chetigerous segment, and in the sixth bears the glands (@. G/.), seven of 
which may be recognised. Immediately behind this the cesophagus is con- 
tracted, and leads into the stomach, on whose surface the almost rectangular 
chlorogogenous areas are already differentiated. The nephridiopores are not 
shown, as they are too minute to be detected with certainty. x 50. 


Fic. 19.—The anterior end of the same specimen. Ventral aspect. The 
slit-like ccelomic cavity between the two anterior brain-lobes is seen by trans- 
parency in the prostomium. The metastomial field—the area included 
between the oesophageal connectives—is slightly raised above the level of the 
ceneral epidermis. The specimen shows the peristomium (Per.), the achatous 
hody-segment (4..B.8.), and the first chetigerous segment (Ch. Seg.'). See 
description of previous figure. Compare these parts with those of the adult 
shown in Fig. 20. x 50. 


Fic. 20.—The anterior end of an. adult specimen of A. assimilis, var. 
affinis, from Otago Harbour. Dorsal aspect. Note the prostomium with 
its V-shaped lateral lobes embracing the median one. ‘The nuchal groove 
(Nuc. Gr.) and the origin of the metastomial groove (Met. Gr.) are shown. 
The first chetigerous segment (Ch. Seg.’) consists of three annuli. The region 
between this and the prostomium is divisible into two almost equal parts, an 
anterior part—the peristomium (Per.)—in which the two annuli are not very 
regular, and a posterior part, consisting of two annuli, which form the achi- 
tous body-segment (4. B.S.). x8. (Cf. Fig. 19.) The form of the pro- 
stomium and the annulation of the skin of type specimens of A, assimilis 
from Uschuaia are exactly the same as shown in this figure. 


Fre. 21.—Transverse section of a post-larval specimen of A. assimilis 
var, affinis, 8°7 mm. long. The section passes somewhat obliquely through 
the posterior brain-lobes, the nuchal organ, and the otocyst of the right side 
The two posterior cerebral lobes (Post. Cer. Z.) are seen in the prostomium 
separated by a portion of the ccelomic cavity. They are closely applied to the 
nuchal organ, the ciliated epithelium of which is well seen on the left, but is 
cut obliquely on the right. The right otocyst is seen as an invagination of 
the epidermis, the mouth of the pit being narrowed to form the tube of the 
otocyst. Already a few foreign bodies (otoliths) have gained admittance. 
Note the large gland-cells scattered in the epithelium, tie sections of the 


ANATOMY OF ARENICOLA ASSIMILIS. 785 


cesophageal connectives, the pharynx, the circular aud longitudinal muscles. 
The nuclei of the muscle-fibres have been omitted. x 210. 


PLATE 37. 


Fic. 22.—Transverse section of the anterior end of a specimen of A. 
assimilis, from Uschuaia. The section passes through the anterior part of 
the brain at the point of origin of the cesophageal connectives. The two 
anterior cerebral lobes are seen in the prostomium in close relation to its 
epithelium ; bands of nerve-fibres (V.) may be seen passing from the latter 
into the brain. The cells in the brain are represented by the dots shown in 
the figure. The spaces seen in the brain and in the subepidermal tissue are 
canaliform prolongations of the coelomic cavity. On the left the section has 
passed somewhat obliquely through the skin, and shows the numerous 
branches given off from the cesophageal connective, and ending in the basal 
part of the epidermis. The dots in these nerves represent the nerve-cells 
which are present. Immediately to the left of the prostomium is the external 
opening of the otocyst (Hv. Op. Ot.), the tube being cut through along a con- 
siderable part of its length; a transverse section of the inner part of the tube 
is seen just ventral to this. In the lower part of the figure the buccal mass 
is seen cut across; the elevations of the epithelium (Pap.) shown are the 
papille of the “proboscis.” The nerves (WV. Buce.) which supply these are 
shown. Note the strong musculature of the buccal mass. x.30. 

Fic. 23.—Transverse section of the middle portion of the brain of a speci- 
men of A. assimilis var. affinis, from Otago Harbour. Only the left half 
of the section is drawn; the median plane is indicated by the two vertical 
lines. In the upper part of the section the intimate relation of the prostomial 
epithelium and the brain may be observed; the ganglion-cells extend up to, 
and lie among the bases of, the epithelial cells. The dots in the brain repre- 
sent the nuclei of small nerve-cells, which are usually arranged in groups or 
clusters. The large cells in the middle of the figure show the form and posi- 
tion of the larger ganglion-cells of the brain. Their processes extend down- 
wards into the neuropile, where they branch. Note the numerous fibres 
passing from the left half of the brain across the middle line to the right, form- 
ing the middle cerebral commissure. The isolated cell on the right side is 
drawn from another section. x 210. 


VoL. 46, PART 4.—NEW SERIES. FYE 


TRE AR ie fh Ateg & 9 ; 


- * 
: - aan ‘Sige bd =d4 
: Ads oe et O Jt 00h 04 gis i. ate ott) 


f bj io Die si Z i. ii ngs sreliges ie 
lymd¢ nes ra Wr @ rel | bag wit a Subadil ie | 
ark. oi MP wat oor 2a : AE A op he SS 
t syelwers 4 Rie i a ae ia -cice) 

f Ons Ny 4 i ay Cae od: Gn ais os Tsu pli teu 


- i Brin ~ : d, ee aT 
‘i hrs: ile @ hind an! lin Boewiiren tit 
jt vie Cette my agtres 
i i : Jn © | Tiel sie 
evel f v a -_< a. : : ~ 
cu rr t y a 7 iis eb} y Lita 
rw Hi 7 | 4 Lf pos 
’ aT een i f ih 
UDooes SW ay ! ou a 
seit? Awe 
1} a eliiea-<ht Th no ky as i 
: s bh ’ jal ey tee 
ai iM hwes 2) wat os ae ah ie 
Th gig Wl g. he epeteionae 
’ A i Vey AL Cage é avr, 6) Coe aca 
J i bylemuay ot omaha. eel wy So 1s £10 The ah 
\ a Cu (an Ten $4 1 een Ae 


; ae rly cc... Jee Sarre Gee 
AG hid = tue i ry 7 hee 

, wl 4. 324 itis Roath Sarah fe 

- 4 eae. wel Ms ait 7— & Cee ge 

se ’ “ery carl eu 28 fe ung “Te; : 

! Vn af) wie ater wih ee Bree le ihe 
F y agit 2 ai) Mens ett. ud) aot 

- iso cA) ge Homa tam glh 9 apt ls ekil 

ae ar 


INDEX. TOs Vou: 


46, 


NEW SERIES. 


Arachnid entosternite, by Pocock, 
225 

Arenicola assimilis, anatomy of, 
by J. H. Ashworth, D.Sc., 737 

Ascidians, new genus and species of, 
by Igerna B. J. Sollas, 729 

Ashworth, on the anatomy of Areni- 
cola assimilis, 737 


Bernard, studies in the 
Parts 111, 1V, and V, 25 

Bryce, Dr. Thomas H., on artificial 
parthenogenesis and fertilisation; 
a review, 479 

— on the maturation of the ovum in 
Kehinus esculentus, 177 

Buller, on chemotaxis aud fertilisa- 

tion, 145 


retina, 


Cephalodiscus, central complex of, by 
Masterman, 715 

Cestode from Cestracion, by Haswell, 
399 

Cestracion, cestode from, by Haswell, 
399 

Cheilostomata, the morphology of, 
by Harmer, 263 

Chemotaxis in the fertilisation of 
eggs, by A. H. Reginald Buller, 
B.Se., Ph.D., 145 


Corystes, metamorphosis of, by 
Gurney, 461 


Dendy, Prof., 
mirabilis, 1 

Diplochorda, by Masterman, 715 

Doncaster, on the development and 
anatomy of Sagitta, 351 

Drummond, Isabella, on the deve- 
lopment of Paludina vivipara, 
with special reference to the urino- 
genital organs and theories of 
Gasteropod torsion, 97 


on Pelagohydra 


Kchinus, maturation of ovum in, by 
Dr. Bryce, 177 

Entosternite of Arachnids, by Pocock, 
225 


Fertilisation and artificial partheno- 
genesis, 479 

Fertilisation and chemotaxis, 145 

Fleure, on the relations of the kidneys 
in Haliotis tuberculata, 77 


Gasteropods, torsion in, 97 

Gurney, Robert, on the metamor- 
phosis of Corystes Cassive- 
launus, 461 


voL. 46, pART 4,—NEW SERIES. 


788 


Haliotis, kidneys of, by H. J. Fleure, 
ag 

Harmer, on the morphology of the 
Cheilostomata, 263 

Haswell, Prof., on a cestode from 
Cestracion, 399 

Hypurgon Skeati, by Igerna B. 
J. Sollas, 729 


Kerr, Prof. Graham, on the develop- 
ment of Lepidosiren paradoxa, 
Part III, 417 

Kidneys of Haliotis, by H. J. Fleure, 
77 


Lepidosiren paradoxa, develop- 
ment of, by Prof. Graham Kerr, 
Part IIT, 417 


Masterman, on central complex of 
Cephalodiscus, 715 


Ovum, maturation of, in Kchinus, 177 


INDEX. 


Paludina vivipara, development 
of, by Isabella M. Drummond, 97 

Parthenogenesis, artificial, by Bryce, 
479 

Pearl, Raymond, on the movements 
and reactions of Planarians, 509 

Pelagohydra mirabilis, a free- 
swimming Hydroid, by Professor 
Arthur Dendy, 1 

Planarians, movements and reactions 
of, by Raymond Pearl, Ph.D., 509 

Pocock, on the Arachnid entosternite, 
225 


Retina, studies in, by Henry M. 
Bernard, Parts III, LV, and V, 25 


Sagitta, development and anatomy of, 
by L. Doncaster, 351 

Sollas, Igerna B. J., on Hypurgon 
Skeati, a new genus and species 
of compound Ascidians, 729 


Torsion in Gasteropods, by Isabella 
M. Drummond, 97 


PRINTED BY ADLARD AND SON, 
BARTHOLOMEW CLOSE, E.C.; 20, HANOVER SQUARE, W. ; 
AND DORKING. 


Arthur Dendy del 


PELAGOHYOR® MIRABI IS. 


f 


Wz, 


Fig, 


Er. Buay 


4 


- End Bud. 


¥ 


Huth, Litht London. 


Azthur Dendy del 


wo 

5 

a 

xo 

ire 

= 
aly 
= q 
z= 
Oo y 
1a] 


RBA 


—— 
Xo} FAOOS vig 


OR 
75) 


MB. del AS Huth, Litht Lendon : 
4 


~ 


LENSHLA, 


¢ 


7 


: 
- 
: 


AS Huth Lith® London 


1 


HMB. de 


AS Huth Lith® London. 


Duarte Isrn Mir Sei be 


Flaure del AS Huth Lith, 


= 


West,Newman lith. 


ic 
3 
as 
is 
3 
ce] 
5 
ic} 
E 
Pp 
z 
A 
= 
a 


S & af 
Cae Wake 


i 7 y 


) MDremmond ad nat del 


5 West.Newman lith. 
) 


| MDrummond & PJ Bayzand adnat del. 


. y 
: 
. 
: 
: 
A 
; 
: 


: 
. 
‘ 
HT 


. : F ‘ 
Beg i 
7 7 et ie es) 


y, 


YL 7 
[Z 


Quart. Zour Mur Sc. Uo. LONE. 


Huth, Lith? London. 


Huth, Lith® London. 


RLPocock del 


a a ca 


= 
S 
s 
& 
5 
= 
é 
= 
a 
EI 
ss 


RPocock del 


i 


50, 100. 200, 300. ee : Huth, Lith’ London. 
S.F Harmer gel_ 


Quart. Foupn. Mor Sei, Vet 4 SN SPLIE 


S.FH ex del 
SFHarmer d Huth, Lith? London 


SEHarmer del, 


| 


we 


ik 


S.R Harmer del. 


Huth, Lith? London. 


” a 
> 


Huth, Lith London. 


it 
i 


b.2.8. 


gep.U. == ==-f—=- = 


es 


6 
§ 
is 
s 
re 
te 
= 
at 
3 
s 
= 


"gas el 


Huth, Lith® London. 


2uart. Sourn. Mim Se Vl, LONSPL £2. 


4 
Ag 
oo 
“aD 
# 
ey 
G) 
he 
Al 
wa 
3 ad ,, 
GF Op eV 
: Med Ne Anes : 
8 i We : ‘ | 
iS : ; | 
PSK oy £ & ; ; 
as : 
i) 8S 
res 
tel 3 sgait ss euiamao PRUESET LRN 
on Sze 5D00 HOPS LE Eve ens ay EN 
Bo! z E 
. SSeS ScoF OSS / 
Coeca.gerged e500. 90 oe 
a) 239000, $0 20 9969S © 2 Se OX 
' ES 03,°.0 POP 6° OOO 
1 oo 6929 9,8 0G 0 Oo) f 
: S Ooo 2OS>OA OOS F BS 
. : COPRode™ OS OE ods 
' Sri eee fears, 9.8 
‘ 


W.A.H. del 


~~ 


1 
\ 

n 

1 
Ae 
8 

9 


—— 


(Si) “ 


Se eeararicy 


( 


WAH. del. 


Huth, Lith? London. 


> 
a ‘ _ > cone Da ae 


Figs.5a, 5), 5c) (es 


Zdwin Wilson, Cambridge. 


Figs.4a,40 joes 


J 
ey 


: 


» "hg SS. ce 
ie ae 
aed a 


Figs 7ev,. 


Fig 7c 9 oT ie ak 


Fig 7d COTE DEES SORE OS i re 


" 2 3 


Figs Te, 7f 2 a mm. 
Figs 8a.8b o 


/ 


= 
Tes a 
4. \ 
*. 
. 
ey 
x = ‘ - 
° 
. 
re 
tyre 
Fe 
. = 
~ 
i 
+ 
Fi 


Fig. (Oh C1 


fo) 1 2 
FIGS 0a =O oss SS tir rl 


; Bees eich ah 5 KK 
Figgas 6. eseauoeeee ca paetee aay pote 


Fig. llc. ee ee tae 


Figssllid;Hegll gil) hussein ce ea eat oe 


He] q 4 | 
BiguiliP> cus ee enne Nena on 


< 7 7 "ae Ven 
; , ieee Se 
: — 
, © 5 ets 
a 
: ¢ 


| ~. nai ig ees pe ae . 
Quart Spurn Mire Le: VMN AIC 2 


sel. 12a. 
Fags.did, Aid, (rade, Wj ce eee re re 
7 “0 aI . c: 
oe ‘ “ Fig.12q | 1 n (ee i r iy 
pi e 
y >. Figs.124,12¢ 1% ea 


Edwin Wilson, Cambridge. 


Fig. 6. 


R Gurney del. 
Huth,Lith® London. 


he ae | 
4 Ps oa. 
ee ot 


Fig. 8. 


Fig.10. 


R. Gurney del Huth, Lith’ London. 


. RB Gurney dvi 
7 


: St eS 
PCCLE CEP SIECEEC CEC CCCN COMIC RNC 
SS a 


Huth Lith? London. 


Left collar cavity. 


ectodermal pie. 


” wg ht g domerulas. 


7 1 Wy z 7 
pericardiad sae ' Le enor , right cotlar cavity. ‘\ 
J ' subneural gland ~ c \ 
heart WY TAULLS Nt 
ESD tes 
} aorsat 000 SU 
ectodermal pit LEN KEEN MES RI Ik 


pre-oradl sis. 


Pericardial 
sac 7 


~ nerve mass. 


left eanal.- 
from dorsal sims 


median dorsal 
ES ERLE 
‘ = 1 
/ left collar cavity, | 
= ri Tay 
sub-ne y na n 
xe TEEN ECT ventral blood sinus 


peripharyngeal witty glomerwuius 
gTroaove 


venirad blo 
oral grooves 


nus pericardial sae 
pnd glomerulns. | 


“aii 
post-orat ane 


lamella 
Gg tomernus. 
oesophagus a 
. 


stomach Ee) f 


y 
\ py 


ep is tomial conrtty. 


. ectodermal pt. 
sub-neurar gland 


ize \ dorsad blood sinus 
i \ 
\ 


dorsal groove of pharynx 


\ Plewrochord oa alan i 
Masterman del intestine. 9. | Huth, Liti ondon 


oO 
Yh 47 
ALLL 


pigat pre-oral canar 


4 , CV 5 Hie 
oe Libs say hObaye 3) AL AT S 
VL FOU Fa 72 HtHlTeT ey, O72 2 vo. 


CRUP aL TLELTVE IR 


perveardial 


SAC. ~~$§ 


pes 
——_ 
f XJ Nasi 4 : é = z : Pon ap Reg. 
ee a2 ea : —- TRO-REUPAL GARD, 
es. 1 3 ; 


right b ranchipl — 


SUES za 


\ 
me \ pericardial sae. 
} Ae ventral blood a] 


f Vy, e . . ; - 
/ /} epis tomial: cavity 


4 / ‘ t ‘ 
f he ‘ 
laterat pons aes ae Se : 
+ oe z : 
. Perucaraal sae. as ; 


Masterman dei. 8. Huth, Lith’ London. 


~ 
2 
. 
° 
/ 
‘ 
- 
e 
* 
a - 
pies . 
' 
i] 


London. 


Huth, Lith 


Sollas del. 


2 
- od | 
, 
. 
. 
. 
. 
‘ 
; LJ 
od = he - a . 
° 
‘ 
‘ a « 
es ~ 
‘o > ‘ 
. 
o 
7 
7 « a 
| . : - 
a 
. 
> 
; 
¢ 
: 
a * 
‘ 
. 
> 
, 
. 
+ 
. 
. 
. 
= 
. 
= — 


A/ YY &Y 


fo) os 
¥-A4.€ 


oe 


) f ee tf - VY ep f “pf 
Luan 7 OWN. Mier ICL. Vol. 4s 


ty, 


0125 TUN. 


Huth, Lith® London. 


Sollas del. 


15.x ste 


JH Ashworth del. 10.x 600 3.x 320. 4. x320. 


Quart. Fourn. Mer Sc VALENS PU37, 


wera NE pid. 
BVin@e. -~_ erp _ | Foot. Op Ot. 
ae WG 2 x Prost.Epith. 


Muscle and 


2. 3.% 210 


8 
JH. Ashworth del. 3 Huth, Lith® London. 


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 Jourwat, which remain in print, are 
now sold at an uniform price of 10/-. 

The issue of Suppnument Numpers being found inconvenient, 
and there being often in the Editor’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. 

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 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 eatra copies of their 
communications at their own expense can have them by applying 
to the Printers, 

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

For every four pages or less— 


25 copies : ; i 5/- 
BON sae : : : 6/- 
HOO, 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 
oF Messrs. J. & A. Courcuitt, 7 Great Mariporouca SrrEp?, 
Lonpon, W. 


OF THE 


UNITED KINGDOM. 
Patron—HIS MAJESTY THE KING. 


President—Prof. E. RAY LANKESTER, LL.D., F.R.S. 
20: 


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 BRIVISH 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 'l'reatise on the Common Sole,’ J. I’. 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. 


ANNUAL MEMBERS , : : . £1 1 Oper annum: 

LIFE MemBERs . , j 4 . 15 15 O Composition Fee. 
FOUNDERS . = ; ; ; LOOk OO ; £ 
Governors (Life Members of Council) 500 0 0 °F o 


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 tothe 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. 


ye he, 
vie Yo bene 4 ie 


ae Aine 
bs Paar e's 


~ = a — = ae 
? noel . -- SS SII Senn GE IIE LTO LY ISN OCA LS At Pee eee pean nt oe ai aaa Gast ond a semen aa SER ee 
Sn et ne en es A ae i “ * now. hin Cm ot. ent ert | ee ns < an 
— — = . - _ + osu ee tt rn eee A me Sem ae AO le mk ee ee ee ew ee 
SRL a TS Rae aye —— - — y-~ 


© 


Se Ee 


a 
—