pore.

Spey

jp Nee
2 ae

=
‘SS

Bes

ey
+ eae
“RS
Petr ay
BY a
fie
0 See
cope
bye XA

iat
Py a

i “ee / iy SR ea wy " si neatenn i uN
‘ ‘ ‘ Aye \ ' a “ fe
Fs eal ba ne ; SR GnCL
RAG i Ab hentai ‘i an t
ae dow 4Ash
y 4 : ' shea
4 Was ‘ . we \:
a hat) ae sh {3 ‘abit wees ae
Ape et ey Sean ‘ :
e PK aT nial RIC } ; Sa Ae x
ah ' Y, : oe rian ERNE acts LR oh
: o Sats! fas ite
1 1 ’ at Vibe ens
. . ~*~
ay FAT ; Seon Sy oe ie actistista The
, Ms) eae S2heatacin i hime ha Ta
J . § Cry ig tal ‘ eye Maha na" au
f ‘ ’ ‘ : 3 1 Hot
“ ; .
iri Sa ‘ ! 2, Malte a
Ne ai teat ik acer
. : Sipps 44 Siw Nhe ys
j : es i ‘ My
rel Ni J ei . ae ode Rast SA
P j RR eS ye ‘ . St
y ' mn runt
’ vor Hew cH
Te 4 re ‘
si i, SAP coe 84

iat

Poy
Jirche dead pew em No een
obiseheay Var

Hea abaay
oaey hea Pha eit
wae

as day

és ies Ee eee ee
sage ae fee Bibb iA

ida ee Lf LEN ge a
pail oe Vein pas pee

eis
ap sbasen ke fete
hen Bes ah

fai ete

eZ

Se
a
re Re

Beant a
ety W ie ee

Heit aad i? be seis
OM Bef F
iy: ate

POR THE PEOPLE
FOR EDVCATION
FOR SCIENCE

LIBRARY
OF

THE AMERICAN MUSEUM

OF

NATURAL HISTORY

= rel -
* - =
‘, :
2
Pa
; en es
)
"
; \
t
= a
rk. '
=

‘i

oo ; ‘ |
e [
a

--

QUARTERLY JOURNAL

MICROSCOPICAL SCIENCE

HONORARY EDITOR :

Sm RAY LANKESTER, K.C.B., M.A., D.Sc., LL.D., E.RB.S.

EDITOR:

EDWIN 8S. GOODRICH, M.A., F.RBS..

LINACRE PROFESSOR OF ZOOLOGY AND COMPARATIVE ANATOMY
IN THE UNIVERSITY OF OXFORD 5

WITH THE CO-OPERATION OF

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

BEYER PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF MANCHESTER ;

GILBERT C. BOURNE, M.A., D.So., F.R.S. ;
J. GRAHAM KERR, M.A., F.R.S.,

REGIUS PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF GLASGOW }

E. W. MACBRIDE, M.A., D.Sc., LL.D., F.B.S.

PROFESSOR OF ZOOLOGY AT THE IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY;

G. P. BIDDER, M.A., Sc.D.

VOLUME 67. New Series.
WITH LITHOGRAPHIC PLATES AND TEXT-FIGURES

OXFORD UNIVERSITY PRESS,
HUMPHREY MILFORD, LONDON, H.C. 4.
1923

(4a)

or 7

1) HAY he ay .

POET ayy 1.

FAATUOU. FaTS he

ee wAA4 An tcooeem

21}-Q4279 &
ae Ky
PRINTED IN ENGLAND

ss AT "THE OXFORD UNIVERSITY PI

ee
mp »
f |
‘et
4
a
are eee ii is 2 STAG AN DOHTTN
Pa
“4
edi,
> t if y 7 y i ets (VT Lara
4 ei
1% i AW bape al.
As ,
“4

CONTENTS

CONTENTS OF No. 265, N.S., April, 1928.

MEMOIRS :
Some Observations upon the Development of the Teeth of Physeter
macrocephalus. By Frank E. eae D.Sc., M.A., F.R.S.
(With 13 Text-figures) :

The ‘ Segmentation Cavity’ of the Egg of ‘ire. oe By jAcaeccaann
Meek. (With Plate 1 and 1 Text-figure) ‘

Nuclear Divisions in Amoeba proteus. By Monica Taye OR, S. N. D.
(With Plate 2) . : 5 ; : ; :
~ On Amphilina paragonopora, sp.n., gudlenithertalundce aed
Phase in the Life-history of the Genus. By W. N. F. Woopranp,
D.Sc. (London). (With Plates 3-5) ;
On the Strobilization of Aurelia. By E. PERctvAL, B. Se.s ne
of Zoology, The University, Leeds. (With Plate 6 and 3 Text- figures)

Histology of the Soft Parts of Astraeid Corals. Pa G. Marruat, M.A.
(With Plates 7 and 8)

The Yolk-Sac and Allantoic Placenta in Peace By 7. womens
Friynn, D.Sc., Ralston Professor of Biology, University of Tasmania.
(With Plates 9-11 and 4 Text-figures) . : : ; :

CONTENTS OF No. 266, N.S., July, 1928.
MEMOIRS :

The Male Meiotic Phase in two Genera of Marsupials (Macropus and
Petauroides). By W. E. Acar, F.R.S., Professor of ey in the
University of Melbaunie. (With Plates 12- —14)

Marsupial Spermatogenesis. By A. W. GREENWOOD, B.Sc., University
of Melbourne. (With Plates 15 and 16)

On Sexual Differentiation in the Infusoria. By Tako V. i.
DocreL, Zootomical Laboratory, eS of eer a
Plate 17 and 1 Text-figure) ;

‘Sanguinicola from the Sudan. By W. N. F. Woobwinn: ole
Bureau of Scientific Research. (with Plate 18) . F ;

On Centropygus joseensis, a Leech from Brazil. By CHARLES

Bapuam, B.Sc., M.B., Ch.M. (With 10 Text-figures) :
¥ Some Observations on the Hypophysis of Petromyzon and of Amia.
By G. R. DE BErEr, B.A., Be.C. (With 34 Text-figures) . :

The Relation of the form of a Sponge to its Currents. a G. P. BippER

Se.D. (With 12 Text-figures) ‘ : ‘

PAGE

101

183

203

219

233

243

257

293

CONTENTS OF No. 267, N.S., October, 1928.
MEMOIRS :

y On the Development of the Hypobranchial, Branchial, and Laryngeal
Muscles of Ceratodus. With a Note on the Development of the
Quadrate and Epihyal. By F. H. Epcewortu, M.D. (With
39 Text-figures) 3 A 2 ; ; ; : : -

On Golgi’s Internal Apparatus in spontaneously absorbing Tumour
Cells. By C. Da Fano, Reader in Histology, King’s Colleges Uni-
versity of London. (With Plates 19 and 20) ,

The Golgi Bodies of a Coccidian. By SHana D. Kine and J. BRonTE
GaTENBY. School of Zoology, Dublin University. (With Plate 21) .

Some Observations upon Spirostomum ambiguum (Ehrenberg).
By Ann BisHop, M.Sc., Victoria University, Manchester. ses
Plates 22 and 23 and 9 Text-figures) : . ; :

” On some remarkable new Forms of Caryophyllaeidae from the Anglo-
Egyptian Sudan, and a Revision of the Families of the Cestodaria.
By W. F. N. Wooptanp, Wellcome Bureau of Scientific Research,
London. (With Plates 24 and 25 and | Text-figure) é :

Studies in Dedifferentiation. IV. Resorption and Differential Inhibi-
tion in Obelia and Campanularia. By J. 8S. Huxiny, M.A., and
G. R. pE Breer, B.A., B.Sc. (With Plate 26 and 7 Text-figures)

CONTENTS OF No. 268, N.S., December, 1923.

MEMOIRS :

) The Cleavage of the Egg of Lepidosiren paradoxa. By AGNES
FE. Miruer, M.A., Department of ockeey Vanes of Cl
(With 12 Text- figures) :

The Morphology of the Nudibranchiate Mollusc Melibe (syn.
Chioraera Leonina Gould). By H. P. KyERscHow AGERs-
sporc, B.S., M.S., M.A., Ph.D., Williams College, Wiliams
Massachusetts. (With Plates 27 to 37)

Observations upon the Behaviour and Structure of Hydra. By SHerna
MarsHALL, B.Sc., Assistant Naturalist, Scottish Marine ° Bole
Station, Millport. (With 4 Text-figures)

Head Length Dimorphism of Mammalian Spermatozoa. By A. 8.
Parkes, B.A. (Cantab.), Department of Boole Univeua of
Manchester. (With 3 Text-figures)

On Brinkmann’s System of the Nemertea Enopla and Siboganemertes
weberi, n.g. n.sp. By Dr. Grerarpa Sriasny-WisnuHorF, Leiden.
(With 26 Text-figures)

PAGE

325

369

381

391

435

473

497

507

593. -

Some Observations upon the Development of the
Teeth of Physeter macrocephalus.

By

Frank E. Beddard, D.Se., M.A., F.R.S.
With 13 Text-figures,

On the following pages I describe the nature of the tooth
rudiments in the very young embryo of Physeter macro-
cephalus, whose general external characters I have already
commented upon and figured (Beddard, 1). Since the material
upon which I have worked is, so far as I am aware, a unique
specimen only 41 inches long, I am particularly grateful to
the Curator of the Durban Museum (Mr. E. C. Chubb) for
placing it in my hands and to the authorities of the Museum of
the Royal College of Surgeons of England for allowing their
highly skilled assistant, Mr. Steward, to prepare a series of
sections for study.

GENERAL CHARACTERS OF TEETH IN ForTAL SPERM
WHALE.

Although the teeth of the adult sperm whale, of both lower
and upper jaws, are well known (Ritchie and Edwards, 5),-
the development of the teeth in this cetacean is described in
only one memoir (Pouchet and Beauregard, 4), so far as I am
aware. This memoir contains a description of the teeth in
an embryo of 80 cm.; those of both jaws are described and
figured. This embryo, it will be noted, is about twice the size
of that dealt with in the present communication. There is not,
however, as it would appear, a great deal of difference in the
condition of development of the teeth; hence I have not
very much to add to the description of Messrs. Pouchet and
Beauregard. Apart, however, from such new facts as I am

NO. 265 B

rd FRANK E. BEDDARD

able to set forth here, certain peculiarities in the growth of the
teeth of Physeter macrocephalus are worth confirma-
tion, though I doubt whether I am able to settle definitely the
question of the homologies of the teeth of the adult, whether—
that is to say—they belong to the milk or permanent dentition.
The former view of the odontocyte dentition seems to be the
one generally held; but by others the question is considered
to be still open. It should be added, however, that these views
are not at all based upon a consideration of the facts described
by Pouchet and Beauregard, whose memoir has been largely
overlooked. This is a further reason for again directing the
attention of zoologists to this subject.

I may commence by directing attention to a matter not
illustrated in the figures of the teeth published by Messrs.
Pouchet and Beauregard ; this is the position of the teeth of
the upper and lower jaw with reference to each other. It will
be seen by an inspection of Text-fig. 1 that the teeth of the
upper jaw are divided from each other by a space that is less
than the space dividing the two teeth rows of the lower jaw—
that is to say, the upper teeth are distinctly within the lower
teeth. Furthermore, the upper teeth are quite vertical in
position, and at right angles with the longitudinal axis of the
head, while the two rows of teeth in the lower jaw are at an
angle to each other, and to the same axis of the head. Thus the
teeth of the lower jaw look outwards as well as upwards,
while those of the upper jaw are directed downwards only.
This state of affairs is more marked in the anterior region of the
lower jaw. It is due to the varying contour of the lower jaw,
which anteriorly is more rounded while posteriorly the upper
surface is straighter. Thus the teeth are, so to speak, carried on
to what is largely the lateral surface.

This figure also shows a character which is to be seen more
in detail in subsequent illustrations of the teeth of this foetus.
The cavity in which the tooth rudiments lie is not at all invaded
by the upgrowth from below forming the dental papilla, which
is only represented in this stage by a thickening of the meso-
dermic tissue shown by a closer approximation of the nuclei

TExt-Fi@.. 1,

‘4i »

In this and the ensuing figures the following general statement
holds good of all, and need not be repeated. The figures are of
sections cut deliberately thick, the diameter being 25 u (i.e.
go mm.). The direction of all the sections is transverse to the long
axis of the head. Where necessary for explanation, the sign * is
on the lingual side. The following lettering is employed :

a, rudiment. of milk tooth. 6, rudiment of permanent tooth.
c, residual lamina. d.p., rudiment of dentine papilla. £.,
Epithelium of mouth cavity which gives origin to dental lamina.
S, side of head. U.J., upper jaw. L.J., lower jaw.

Fig. 1 represents a portion of the upper and lower jaw to show
the position of the tooth germs of these two jaws in relation
to each other. The tooth of the lower jaw is seen to lie outside of
that of the upper jaw.

4 FRANK E. BEDDARD

of this tissue. The peculiarly large extent of this cavity,
which will be dealt with immediately, is possibly to be looked
upon as a preparation for the subsequent growth of the dental
papilla.

The ingrowths of the epidermis to form the enamel organ,
instead of lying within—and firmly embedded therem—the
mesoderm tissue underlying the epidermis, depend freely
into a spacious cavity just referred to, which forms the tooth
follicle. Pouchet and Beauregard figure extensive spaces
surrounding these ingrowths, but not to so large an extent as
I have found in the young embryos examined by myself.
In the lower jaw of my foetus the space invariably commenced
immediately below the epidermis ; but in the upper jaw there
was frequently a layer of mesoderm immediately underlying
the epidermis and perforated by the ingrowth. In other
mammals spaces are apt to occur in the same way. The cavities
are so large that they are only moulded in the roughest way
to the enamel ingrowth ; it 1s to be remarked, however, that
—in the upper jaw at any rate—the labial outgrowths (see
Text-figs. 9-12) of the dental lamina, which will be dealt with
later, lie in one or two diverticula of this cavity, and in the
same way free within it ; they are not closely adpressed to its
walls. I imagine that this state of affairs is not altogether
natural, but is due to reagents and consequent shrinkage.
I have no means, however, of ascertaining whether any part
of this cavity is normal. In any case the practical result is that
both in upper and in lower jaw a canal is formed which is quite
continuous from one end to the other of both jaws. This
cavity gradually narrows at the extreme end of the series of
teeth and finally ceases to exist close beneath the epidermis.
It is possible that it is associated with the groove which in this
and other cetaceans lodges the teeth in the adult animal.

It is, furthermore, possible that something of the same kind
led to the erroneous views upon the development of the teeth
expressed by Goodsir, whose figures persisted until quite
lately in text-books such as ‘ Quain’s Anatomy ’.t Here the

! 8th ed., vol. ii, 1876, p. 315, fig. 214, 3 and 4,

a

TEETH OF PHYSETER o

srowing tooth is represented as the dental papilla only, growing
upwards into a cavity.

A rough survey of the series of sections, one of which is
shown in Text-fig. 1, shows also that there is apt to be con-
siderable difference in size between some of the tooth cavities
of the upper jaw and those of the lower. ‘This is not always the
case ; but posteriorly (i.e. nearer to the condyle of the jaw)
the follicles of the upper jaw teeth are deeper than those of the
lower jaw, even twice as deep, in that and other regions. This
is correlated with a greater length of the dental lamina, which
will now be deseribed.

It seems to be a general rule—so general that Sir Charles
Tomes (6) makes his diagrams of the developing teeth conform
to it—that the dental lamina is oblique in direction, running
indeed at times almost parallel with the oesophageal epithelium,
of which it is a downgrowth. In Physeter, as the figures
of Pouchet and Beauregard indeed show, this lamina is
absolutely at right angles to the oesophageal epithelium.
This will be apparent from an examination of Text-figs. 3 and 4,
&e., annexed hereto. The origin of the lamina shows no points
of particular interest. It arises from the malpighian stratum,
at both sides of an ingrowth of the superficial layer of cells
which thus forms its core.

The histological condition of the material was good enough
to show the difference between the central core of the lamina
and the often cubical layer of epidermis which surrounds it
externally. I do not, however, attempt any special description
of the various cells, as they do not seem to present any features
of disagreement with those of the developing teeth in other
mammals. :

TEETH OF THE LOWER Jaw.

My examination of the teeth of the lower jaw was hindered by
the condition of the sections of the anterior region of these
jaw rami. The tooth follicles upon which I have already com-
mented were often entirely empty of contents, the jaws being
here obviously more exposed to external injury. I have, how-

6 FRANK E. BEDDARD

ever, no other reason to doubt that the teeth rudiments were
like + those situated more posteriorly, which presented very few
such lacunae. Still, in attempting general statements concern-

TExt-Fia. 2.

Tooth of lower jaw. This figure and the two following are a nearly
continuous series, one section only lying between each section
figured. The mouth epithelium is seen to be ruptured owing
to the swelling of the tooth follicle, and the dental lamina with its
tooth germs to be dislocated towards the lingual side. The
dental papilla is no more than a closer agglomeration of the
mesoderm cells at the base of the tooth follicle.

1 T could only find one series of teeth in the anterior part of the jaw
and these were quite early. The only difference from those which are
described more in detail below is that the actual tooth germ is longer
and more parallel with the residual lamina, and that the latter tends to
disappear between successive tooth germs ; there is thus an approximation
in structure to the teeth of the upper jaw. But the dental lamina remains
very short as in the posterior teeth of this jaw. These anterior teeth are
evidently more advanced in development.

TEETH OF PHYSETER 7

ing the mandibular teeth the defective condition here referred
to must be borne in mind. I shall have, for example, to indicate
actual structural differences between the teeth of the mandibles
and those of the maxillae and premaxillae.

These teeth are readily comparable at first sight with the

TEXT-FIG. 3.

As Text-fig. 2, but the rudiment of permanent tooth is larger.

teeth of other mammals in a corresponding stage of develop-
ment. The tooth rudiment shown (Text-fig. 4) as a_bell-
shaped swelling seems to be clear; and beyond this, i.e. distally
from the place of origin of the dental lamina, is a prolongation
which would seem to correspond to the residual lamina of
other teeth. They all presented more or less the appearance
shown in Text-figs. 2,3, 4. The entire organ developed from
the enamel germ often lay closely adpressed to one side of the

8 FRANK E. BEDDARD

copious tooth follicle, but in other cases it lay more in the middle
of that follicle removed from its walls. ‘The growing tooth
was small compared with those of the upper jaw, which will be
dealt with immediately. Hach was distinctly marked into three

Text-Fic. 4.

Te

As in two previous sections ; the rudiment of the permanent tooth
has acquired its full size.

regions, which are very plain—as is shown in the figures
already referred to. The tooth itself les to the lmgual side of the
ingrowth, and at its middle or thereabouts is rather bell-shaped,
or at least divided into two lobes ; these look inwards and not
downwards. ‘The stalk of which this is an outgrowth, i.e. the

TEETH OF PHYSETER 9

dental lamina is straight and quite at right angles to the
oesophageal epithelium. Beyond the origin of the tooth rudi-
ments, as will be seen in the figures, it 1s continued onwards
in the same straight line as the dental lamina, and it is this
region which I have termed above the residual lamina. In
neighbouring sections to that represented in Text-fig. 4 (see
Text-figs. 2, 3) the dental and residual laminae show no
differences, but the actual tooth germ is more slender. But

TEXT-FIG. 5.

y

ef

fe

SY
aN
v

SS

KS

LE

2 2.

= St
5S

PZ
S

945
ae
‘a

ay
cyt!
ia)
.)

Three sections, near to each other, from the condylar end of the
lower jaw, representing three stages in the growth of the tooth
germ. The left-hand figure represents the initial stage, in which
the entire tooth germ is a mere swelling of the dental lamina.
Later (the middle figure) the residual lamina becomes differen-
tiated ; and the third section shows the complete differentiation
of the rudiment of the permanent tooth.

it does not appear ever to vanish between successive teeth,
but to be continued as a lamina, the actual tooth germ being
local thickenings of this. The residual lamina undergoes no
change in the intervals between the teeth, but it may at times
terminate in a more club-shaped or at least slightly swollen
extremity than in other places. At the very beginning of the
series of tooth rudiments of the lower jaw—at the end nearest
to the condyles—the rudiment consists (see Text-fig. 5) of
a swollen extremity supported by a short stalk. This resembles

10 FRANK E. BEDDARD

the first of the series of the upper jaw, which will now be
described. It will be noted, therefore, that here, as in the
upper jaw, the tooth series develops from behind forwards.

TEETH OF THE Upper JAw.

In the case of the upper jaw I was also able to trace the
dental lamina to the very end, close to the condyle. It begins
here as a much shorter fold than it becomes farther forward.
The fold at the very first is more or less oval in transverse
section, and later appears club-shaped. It becomes separated
in fact into the oval free extremity and a much more slender
stalk. The expanded free edge of the dental lamina may be
traced forwards into the series of rudimentary teeth. The only
change at first in this region of the dental lamina is the increase
in size, at intervals and for the distance of a few sections, of
the expanded free extremity. Nevertheless, it is particularly’
to be noted that where there are no signs of tooth formation
the ending of the lamina is still swollen. It has been asserted
and denied* that a swollen extremity of the dental lamina
argues the actual presence of a rudimentary tooth germ. It
would seem likely in the present case that the region of the jaw
which we are now considering will ultimately be furnished with
teeth. But I have no positive facts to fix the validity of this
decision. And, moreover, in view of the apparent agreement
in age of all the undoubted tooth germs in both jaws, it might
be argued with equal force that the terminal region is not to
be invaded by teeth. In this event the swelling of the edge
of the dental lamina will be an argument in support of those
who see in a terminal swelling no actual prophecy of teeth
in the same situation, however rudimentary those teeth may be.

The accompanying figure (Text-fig. 6) shows an early tooth
follicle with the dental lamina therein and that it ends in a slight
swelling. It will be noticed that the terminal swelling is con-
tinued in the same straight line as the rest of the dental lamina.
The next figure (Text-fig. 7) shows a section some way farther
on towards the symphysis of the jaws; and in this a slight

1 See, for a brief summary of these views, Tomes (6, p. 357).

TEETH OF PHYSETER HH

alteration is to be noted. The terminal swellimg—not particu-
larly strongly marked and so far like the first of this group of
sections which has just been described—is not in the same

TEXT-FIG. 6.

An early stage in the development of a tooth of the upper jaw.
This, with Text-figs. 7, 8, is to be compared with Text-fig. 5,
which represents three more or less equivalent stages in the
development of the lower jaw teeth. The much greater length
of the dental lamina will be noted,

straight le with the remainder of the dental lamina. It is
distinctly turned lingually at almost, or in some sections quite
at, right angles to the rest of the lamina. <A further stage of

12 FRANK E. BEDDARD

development of the tooth series of the upper jaw is to be seen
in Text-fig. 8, which is the third of the present series. Here
we have a dental lamina with terminal tooth swelling shaped—
as in the last—much as a tobacco pipe, the ‘ bowl’ bemg

TEXT-FIG. 7.

x
J

¢
Asties4

ede Ee

sit

A stage farther than that represented in Text-fig. 6. The tooth
rudiment, instead of being a straight continuation of the dental
lamina, is slightly bent to lingual side,

turned inwards (lingually) at right angles to the ‘stem’.
But there is here a further addition. There are two slight
processes upon the labial side which are very short but evidently
correspond to the more highly developed processes whose

TEETH OF PHYSETER 138

form and nature will be dealt with presently. We see in these
tooth rudiments, early in the series, a commencing of the various
characteristics of the fully developed embryonic teeth of this

TEXT-FIG, 8.

A more advanced tooth rudiment. In addition to the more marked
bending of the rudiment of the permanent tooth, the commence-
ment of the milk rudiment and the residual lamina is to be seen.
In all these three figures the dental papilla is a mere thickening of
the mesoderm tissue.

young foetus. It must not, however, be assumed that the
labial bud-like outgrowths of the dental lamina are subsequent

14 FRANK E. BEDDARD

in time of origin to the germ of the persisting tooth of the adult
which arises from the end of the dental lamina. If this were
proved to be the case the homologies of both would need
another view than that put forward here. The actual time
at which the end of the dental lamina becomes the rudiment
of a tooth is hard to decide.

In the memoir of Messrs. Pouchet and Beauregard the authors
figure! six sections of teeth, of which one only (Text-fig. 6)
represents a tooth of the lower jaw. All of the figures are
approximately alike, and an inspection of these drawings does
not lead to any possibility of distinguishing between the teeth
of the two jaws in this cetacean, or at any rate in an embryo
of 30 cm. in length. I have already mentioned the fact that
I was only able to discover anything that looked definitely
like a residual lamina in the teeth of the lower jaw in my younger
embryo. Nor do the figures of Pouchet and Beauregard (4)
show anything similar beyond all doubt to the projection of
enamel tissue which I have figured (see Text-fig. 4) to the
labial side of the undoubted tooth rudiment. There is, however,
in the former figures a thick process, forming really part of the
terminal expansion of the dental lamina, to be seen; but it
will be noted that this is on the lingual side. The process does
not seem to me to be distinctive enough to set it down as
a residual lamina, even without going into the question of its
position with regard to the main ingrowth of the dental lamina.
And in any case those authors complain of the condition of their
specimen, which would render it unwise to do more than call
attention to such characters as are obviously not masked by
the inferior histological state of this specimen.

In my younger example there was nothing that could be
compared to this residual lamina (if it be so) which is figured
by Pouchet and Beauregard.

This, however, may easily be due to the fact that the state

1 On Pl. vii of the memoir already quoted. I may call attention to the
slight confusion in the description of Pls. vii and viii: the former is stated
to be of an embryo of 1:30 m., the latter one of 030m. The reverse is
obviously the case,

TEETH OF PHYSETER 15

of the tooth development was younger than that of the foetus
of Pouchet and Beauregard, and that therefore a definite
residual lamina could hardly be expected until a little later.

TEXT-FIG. 9.

Tooth germ from upper jaw at a more advanced stage of develop-
ment. The milk rudiment is long and conspicuous, expanded at
the extremity.

The teeth of the upper jaw, however, though they show no
prolongation of the dental lamina precisely comparable to that

16 FRANK FE. BEDDARD

to which attention has just been called, are provided with
an outgrowth or outgrowths of which the nature is also diffi-
cult to decide. These are figured in Text-figs. 9, 10. They lie

Text-Fic. 10.

The next section but one to that figured in Text-fig. 9. The residual
lamina is not to be seen,

in the sections which I have in my possession invariably on the
labial side of the dental lamina. There are quite frequently,
perhaps always, two of them—or even three, of which two
may even arise by a common origin (Text-fig. 11) from the

TEETH OF PHYSETER 17

dental lamina well behind its termination in a tooth germ.
These outgrowths end in a swollen termination egg-shaped in

TExt-Fic, 11.

Another tooth rudiment in the neighbourhood of those repre-
sented in Text-figs. 9, 10. It shows the peculiarity of a double
milk outgrowth.

outline. The appearances which I have seen, and which are
represented in the annexed figures (Text-figs. 9-11), are
No. 265 (a)

18 FRANK E. BEDDARD

obviously like those of the developing teeth of the older
foetus described by Pouchet and Beauregard as regards these
outgrowths. But there is one important difference which
is particularly to be noted. In my foetus the outgrowths are
always on the labial side; but the figures of the two writers
just referred to show that these outgrowths may be also on
the lingual side, a fact which will have to be borne in mind
in considering the nature of these parts of the tooth germs of
the cachalot. They figure two sections in which the outgrowths
are on the lingual and not on the labial side (4, Pl. vu, figs. 3
and 5). Furthermore, in the older foetus these outgrowths may
become, or be accompanied by, a luxuriant crop of similar out-
erowths (Pouchet and Beauregard, 4, Pl. vu, fig. 1) on the
same side as the single outgrowth or from the oesophageal
epithelium (4, Pl. vu, fig. 5). I have seen nothing of the kind
in my specimen. The nearest approach to it is the double
outgrowth (Text-fig. 11) im my specimen. Finally, I call
attention to the fact that in one section represented by Pouchet
and Beauregard there are two outgrowths precisely as in several
sections of the series from my younger foetus (4, Pl. vil, fig. 4).
I should mention also that in my specimens one of the two
outgrowths only, or it may be two where there are two or three,
certainly corresponds to the diverticulum or diverticula of the
tooth follicle to which I have referred on a previous page.
T am not sure whether this is always the first of the series where
there are two or three ; for the great development of the cavity
prevents an accurate fitting to the enamel organ. It should
be noted also that the long ‘ stalk ’ of these lateral outgrowths
ending in a swollen extremity reproduces more exactly the form
of the stalked tooth germ of which they are an outgrowth than
in the case of the older foetus. This will be apparent from
a comparison of Text-figs. 9-11 with those of the French
authors. What are these outgrowths of the dental lamina ?
Messrs. Wilson and Hill (7, Pl. ii, figs. 52-4, woodeut figs. 2,
3, pp. 568, 569) figure and describe certain outgrowths from the
dental lamina of marsupials (in Perameles) on the labial
side, and call attention to the reference by the late Mr. Martin

TEETH OF PHYSETER 19

Woodward (9, Pl. xxlvii, figs. 25a and b) to similar out-
growths, which (in Petrogale) the latter regarded simply
as tooth germs. Wilson and Hill, however, find that these
outgrowths (in Perameles it must be remembered, not
Petrogale) are really sheets arising from the dental lamina,
and not to be confused with dental rudiments. They refer
them to the ‘ labio-alveolar lamina’ and trace them back in
their origin to the oesophageal epithelium, finding sometimes
no connexion at all with the dental lamina. I do not attempt
to criticize, and content myself with the briefest account
of this view of the labial outgrowths in question. I am con-
vinced, however, that the structures which I describe here in
Physeter macrocephalus really originate from the
dental lamina, and are limited to narrow outgrowths like
tooth germs. They are not to be seen throughout a long series
of continuous sections ; but only in three or four consecutive
sections. I have dwelt upon their resemblance to the undoubted
tooth germs in this cetacean ; indeed, the only difficulty in
the way of regarding them as tooth germs might be considered
to be their origin from the labial instead of from the lingual
side of the dental lamina. But in the first place these rudiments
do occasionally originate from the lingual side, as I presume
from the figures of Pouchet and Beauregard ; and in the second
place in a vertically developed dental lamina the actual side
of origin seems less important than in an obliquely. disposed
dental lamina.

Woodward also found no difficulty in referring such out-
growths to a milk dentition, while in the case of the incisors
he referred a lingual outgrowth to the permanent series.
This position as to the nature of a particular rudimentary
tooth is accepted and asserted by Tomes, who writes (6, p. 356) :
‘we are justified in saying that any additional specialization of
the dental lamina which is situated on the lingual side of
a formed germ belongs to a later generation of teeth, and con-
versely that any similar outgrowth of the lamina which lies
on the labial side of a formed tooth germ belongs to an ante-
cedent generation.’ But it would, as I think, be pushing this

C2

20 FRANK E. BEDDARD

generalization too far to regard the (? same) outgrowths of this
embryo of Physeter as a rudiment of a milk dentition when
they appear on the labial side and of a postpermanent genera-
tion when they are processes of the lingual surface of the dental
lamina. There are, however, as it appears to me from the facts
represented in my sections, and from the literature briefly
referred to above, considerable grounds for believing these
outgrowths of the dental lamina in Physeter to represent

Trxt-Fic. 12.

Three sections at an interval from each other of one section only,
nearer to the anterior end of the upper jaw than those sections
represented in preceding figures, and therefore at the most
complete stage of development shown in the foetus examined.
They show certain differences from the more posteriorly situated
sections of the upper jaw series. This chiefly affects the relative
positions of the milk rudiment (a) and the residual lamina (c) to
each other and to the germ of the permanent tooth (6) (assuming
that these several outgrowths are correctly identified).

vestiges of a milk dentition which never comes to maturity,
and that the permanent teeth of this cetacean are therefore
to be looked upon as the equivalent of the permanent dentition
of other mammals. This conclusion is not that of Kikenthal
(2), who, however, did not (probably was not able to) refer to

the memoir of Bouchet and Beauregard owing to nearly simul-
taneous publication.

TEETH OF PHYSETER 91

I have in a few cases (Text-fig. 12) found an apparent absence
of a second outgrowth of the dental lamina. ‘There is here
a specially long single outgrowth which is less hke a residual
lamina than in other teeth, as may be noted by a comparison
with Text-fig. 9. Nevertheless, there are some reasons for
regarding this outgrowth as the representative of the residual
lamina though the usually present additional outgrowth seems
to be absent. I believe, however, that this structure is not
absent, but present in the form of a process of epithelium of
a somewhat different appearance and origin. Close to the
stalk of the dental lamina, and close to its origin from the
oesophageal epithelium, is a pyramidal heap of cells which is
continuous with the stalk but also arises separately from the
oesophageal epithelium. This pyramid has its apex directed
upwards, its wider base being continuous with the oesophageal
epithelium. It is not large, as I could only find it in one to
three continuous sections. I think that this outgrowth may be
compared with that which I have described above as existing
in most of the teeth of the upper jaw. If this supposition is
wrong, I cannot at the moment compare it with anything else,
except perhaps as a dwarfed equivalent of the mass of indepen-
dent outgrowths which Pouchet and Beauregard have figured
as growing from the oesophageal epithelium in the immediate
neighbourhood of a tooth germ.

It should be noted (as is shown in Text-fig. 13) that this
pyramidal outgrowth is received into an excavation of the
mesoderm tissue surrounding the tooth follicle, as are other
parts of the developing tooth series. It is not, therefore, to
be looked upon as merely an outgrowth of the oesophageal
epithelium having no relation to a particular tooth germ.

It is possibly the case that the absence of the basal pyramidal
outgrowth of the oesophageal epithelium to which reference
has just been made is not always a reality, for im the three
consecutive sections in one case (Text-fig. 12) I could find this
outgrowth in only one or two sections, and there but small
and rather of a rounded than conical form. It may be, there-
fore, that this structure has been missed or—if present—not

99, FRANK E. BEDDARD

regarded as I regard it (i.e. as a part of a definite tooth out-
growth), in sections such as that which Kiikenthal represents
of the developing tooth of the Beluga. ‘This latter figure

Trxt-Fic. 13.

Tooth germ belonging, like Text-fig. 12, to those of the anterior
part of the upper jaw, but situated farther back than that
represented in Text-fig. 12.

(2, p. 391, fig. 60) consists of a tooth rudiment which is older
than those figured in the present paper, Imasmuch as one germ
is already bell-shaped, while the accompanying germ is racket-
shaped like those which I figure on the preceding pages. On

TEETH OF PHYSETER 93

the other hand the small anterior outgrowth, characteristic
of so many of the teeth of the upper jaw in Physeter, may
perhaps have been missed owing to its non-occurrence in the
sections actually figured by Kiikenthal. Or the pyramidal
outgrowth may have been missed for similar reasons, as
suggested above on general grounds. If this be so, then the
teeth of Physeter will come more into line with those of
other Cetacea, and be less abnormal than the facts described
and illustrated here would imply. But again it is as likely—
perhaps more likely—that the figure of Kiikenthal is to be
compared rather with the lower jaw teeth of Physeter.

Finally, the fact is to be emphasized that these more
basally situated outgrowths of a more or less pyramidal form
are only to be found among the teeth which are in the anterior
part of the upper jaw; and it is only farther back that the
long filiform outgrowth is to be seen. There is thus a differentia-
tion of the upper jaw teeth into an anterior and a posterior
series, which is remarkable.

CoMPARISON OF TEETH OF LOWER AND UppER JAw.

Having now dealt with the structure of the teeth of the
upper jaw, we are in a position to compare them more accurately
with those of the lower jaw. There is, I think, on the whole,
reason for believing that there are differences between these
two series. It is remarkable, however, that Pouchet and
Beauregard figure no differences between teeth of the upper and
lower jaw in their memoir. Such differences as I shall point
out, or recall from the above references, from the teeth of the
lower jaw, may be therefore merely a matter of age. Whether
this be the case or not, the younger foetus shows the following
apparent differences between the two series of teeth—those of
the upper and those of the lower jaw. Apart from size, and short-
ness of the cental lamina in the teeth of the lower jaw, the chief
—and indeed perhaps the only—difference between the two
series lies in the fact that whereas the dental lamina has only
one outgrowth in addition to that which forms the persistent
tooth in the case of the teeth of the lower jaw, there are at

24 FRANK E. BEDDARD

least and generally two such outgrowths in the teeth of the
upper jaw. The most anterior of these (i.e. that closest to the
persisting tooth rudiment) is to be regarded as the residual
lamina which is alone (?) met with in the teeth of the lower jaw.
Whether I am right or not in regarding the second outgrowth
as a milk rudiment, it is at least a point of difference between
these teeth and those of the lower jaw, where it appears to be
non-existent—at any rate in the embryo which I have
examined.

This again may be a matter of age. As to the residual lamina
its exact likeness to that of the lower teeth is not absolute.
There is this important difference. While in the case of the
lower teeth the lamina is a lamina continuous from section
to section, it is not so with the upper teeth ; here in fact the
small process (see Text-fig. 10) which may be its equivalent
disappears and reappears every two or three sections, thus
indicating a series of processes and not a continuous lamina
(cf. Text-figs. 9, 10). There is next to be seen a difference—
perhaps more apparent than real—between the mode of origin
of what I regard as the permanent tooth in the upper and
lower jaws.

As has been already mentioned, the tooth germ in both arises
as a thickening of the end of the dental lamina, which.is con-
tinuous as a thickened edge to that lamina. In examining the
whole series of teeth rudiments in the lower jaw, from their
commencement at the condylar end of the jaw, the followmg
stages may be detected. The oval thickening, shown in
Text-fig. 5, persists in section after section, but gradually
alters its shape to a more triangular outline, and at the pointed
end away from the origin of the dental lamina the residual
lamina gets gradually to be free, the rest of the thickening
remaining behind, so to speak, as the actual tooth germ.

In the upper jaw the series of events is rather different. The
same thickened edge appears at first, but its stalk grows
longer until a racket-like structure is produced, as shown in
Text-fig. 6. Instead of remaining as it is—as is the case with
the lower jaw teeth—it is bent over lingually, and the residual

TEETH OF PHYSETER 35

lamina appears as a new structure. There is, so to speak, no
freeing of the residual lamina from the compound mass.
Strictly speaking, therefore, there is not an exact homology
between the concave surfaces of the future cup-shaped enamel
cap in the two series of teeth. It is terminal in one and lateral
in the other.

But it must be remembered that, as I have pointed out above,
the more mature teeth, i.e. those at the symphysis end of the
jaw, apparently approach the teeth of the upper jaw in this
last-mentioned characteristic. The tooth germ, that is to say,
is more elongated and oval. But what we are dealing with
here is not the form of the growing tooth germ but its mode of
origin. This is undoubtedly different in the teeth of the two
jaws, as has been emphasized. But this latter consideration
may be regarded perhaps as suggesting comparisons between
the teeth of the two jaws which have not yet been closely
examined. On the views just advanced the one outgrowth of
the dental lamina beside the outgrowth which results in the
tooth of the adult is a residual lamina. Its form, moreover,
is highly suggestive of such an interpretation as is to be seen
in Text-figs. 1 and 4; and I have put forward other facts.
On the other hand, in the more mature teeth of the lower jaw
the shape of the whole tooth germ is not unlike that of the
Beluga as figured by Kiikenthal, as pointed out on another
page, and is of course also like that of the upper jaw of the
present species and specimen represented in Text-fig. 10 of the
present paper. Are there, in fact, after all, reasons for regarding
the process which I have lettered ‘c’ in the teeth of the lower
jaw (Text-fig. 4) as really the equivalent of the process lettered
‘a’ in the teeth of the upper jaw (e.g. Text-figs. 9, 10) ?

If this is so, it is clear that a different view may have to be
taken of the homologies of the two teeth rudiments than that
advanced in the present paper.

For if the labial outgrowth immediately following the lingual
outgrowth is a tooth germ, and not a residual lamina, it would
appear to follow that it is this which is the rudiment of the
tooth of the permanent dentition ; and therefore that the tooth

26 FRANK E. BEDDARD

which actually arrives at maturity is in reality of the milk
dentition—a view which is held of the cetacean teeth. But
to support this view requires some ‘ manipulation’ of the facts
set forth in the above pages, and in the memoir of Messrs.
Pouchet and Beauregard. It is true that the labial process
in the lower jaw teeth at the symphysis extremity of the
jaws is very like the tooth rudiments of Beluga, and in faet
many mammals, a likeness which is increased by the fact that
this outgrowth does not form a continuous lamina as does its
supposed homologue in earlier sections (i.e. at the condylar
end of the jaw), but decreases in successive sections and seems
to disappear. This, however, need only remind us of the
residual lamina (as I regard it) in the upper jaw (see Text-figs.
9, 10), which is not a continuous lamina but a series of out-
srowths.

A stronger argument in favour of the view advanced here is
that on the hypothesis now being considered we should have
to pay no attention at all to the conspicuous outgrowths of
the upper jaw, which are difficult to explain away as of no
importance ane without meaning. But even then, it will be
noted, we are left with an undoubted difference between the
teeth of the two jaws, lower and upper, which is evident in other
characteristics of these organs, and which is set forth in the
present section of this paper.

PECULIARITIES OF TEETH OF PHYSETER AND COM-
PARISON WITH THOSE OF OTHER MAMMALS.

It is possible to deduce from the foregomg pages such
a comparison, which does not, however, shed a great deal of
light upon the zoological relationships of the Cetacea, except
perhaps in one of the points raised.

It is clear, at any rate, that Physeter agrees with other
mammals in having the usual two dentitions and—as in many
cases—a residual lamina containing the promise or possibility
of a third dentition, sometimes abnormally developed farther
(e.g.inman). Furthermore, I have shown reasons for believing
that the permanent dentition, in this cetacean at least, is

TEETH OF PHYSETER 27

preceded by a milk dentition, thus conforming to the generally
accepted view that (as far at any rate as the facts contamed
in the present paper allow of a statement) the Cetacea are the
offspring of a stock already provided with the typical Kutherian
dentition. .

Messrs. Pouchet and Beauregard, as has been duly pointed
out in the above pages, register an apparent peculiarity of the
developing teeth of Physeter in the form of tufts of out-
growths from either the dental lamina or as a direct series of
buds arising—not from, but beside, the dental lamina. These
I have not been able to discover in the younger foetus described
by myself. But in any case they are not, as I believe, to be
regarded as a peculiarity of Physeter or of the Cetacea.
For others have dealt with structures which are, I think,
essentially similar.

Thus Leche (8, Pl. ix, fig. 70; Pl. xi, figs. 64, 90; Pl. xvi,
figs. 140-2) figures and refers to a number of small ‘ tags ’
attached to the dental lamina in Phoca groenlandica,
in the bat Phyllostoma, and in the marsupial Phasco-
larctus. The latter figures are copied by Wilson and Hill
(7, Pl. xxxi, figs. 76, 77). How far such outgrowths have any-
thing to do with tooth formation — phylogenetically for
instance—the facts at my disposal do not allow of a guess.
They suggest themselves as a mere state of perhaps abnormal
activity.

There is a final matter, however, in which a possibly ifnportant
difference from that generally observed in mammals is to be
seen in the developing teeth of Physeter macrocephalus.
This concerns the continuation along the jaw not only of the
dental lamina but of the actual tooth germs of the permanent
series only. The more usual state of affairs in mammals must
be referred to first. Thus in the earliest stage (Stage II) of
the embryos of Perameles studied by them Messrs. Wilson
and Hill (7, Pl. xxv, figs. 1, 2, and woodcut fig. 1 on p. 475)
represent the origin of a third deciduous incisor which grows
out of the dental lamina. In the first of these sections the
dental lamina is shown alone without a trace of the tooth which

28 FRANK E. BEDDARD

appears suddenly in the next sections as an outgrowth of the
dental lamina. There is no trace of any direct connexion—
additional to the dental lamima—hbetween the germ of this
tooth and preceding teeth; its enamel organ is a separate
outgrowth of the dental lamina. In the same way these
authors represent the growth of a premolar in the same animal
in a later stage (Stage III).

There are seven sections figured (a—c), each three sections
apart. It is not quite clear what is the exact connexion between
the second premolar (represented in figs. a and B) and the
dental lamina; but in any case the latter is shown as such
(i.e. without any tooth outgrowths) in figs. c-r. Then in ¢
suddenly appears—as an outgrowth of the dental lamina—the
rudiment of deciduous premolar three ; the whole tooth germ
—that is, the actual tooth, the dental lamina, and the residual
dental lamina, extending beyond the tooth—possessing a close
resemblance to one of the teeth of the lower jaw in Physeter,
such as is represented in Text-fig. 1 of the present paper.
There is no trace here either of any continuous lamina connect-
ing the individual tooth germs. A final instance is shown by
Woodward in a reconstruction of the teeth, deciduous as well
as permanent, of Sorex (10, Pl. xxv, fig. 19). In this figure
the teeth are seen to depend from the dental lamina. only,
and to be completely separate from each succeeding and
preceding tooth. Quite different is the state of affairs shown
in my séctions of Physeter. The processes of the dental
lamina which I have identified above with the milk dentition
are in fact a series of processes only arising at intervals from the
dental lamina. But the permanent teeth are produced at the
free end of the dental lamina (in the case of the upper jaw)
or from its lingual surface, leaving a continuous residual lamina
(in the case of the lower jaw). In the upper jaw the position
of the future teeth is shown by a bending inwards of the entire
dental lamina (see Text-figs. 6-8), and a thickening of the
same at intervals ; but there is no projection of the rudiments
of teeth beyond the edge of the dental lamina, which is con-
tinuous between the successive teeth and is only different in

TEETH OF PHYSETER 29

the interdental regions by its less swollen character. Pre-
cisely the same is to be seen in the lower jaw, where (Text-figs.
2-4) the dental rudiments are definite outgrowths of the dental
lamina, which is, however, a continuous outgrowth, being
merely thinner in the interdental intervals (Text-figs. 2, 3).
This will, I think, be made plain by an inspection of the figures
referred to. There is to be seen, as I interpret the facts ascer-
tained and figured by Messrs. Wilson and Hill (8, p. 141,
Text-fig. 1), a perhaps comparable state of affairs in the
developing teeth of Ornithorhynchus.

In the younger of two foetuses examined by those two
anatomists the entire dental lamina of both upper and lower
jaw (of one side) is figured and described. From those figures
and the descriptions it is to be inferred that the enamel organs
of two teeth are differentiated in the substance of the lamina
of each half-jaw as a thickening of it, and not as an outgrowth
therefrom—the connecting part of the dental lamina remaining
unaltered between those rudiments. This is, as I think, to be
compared exactly with any two succeeding teeth of the upper
jaw of Physeter, where the tooth thickening is merely the
lamina itself, and the unaltered lamina remains in the same way
between successive tooth germs. This—as it will not be for-
gotten—is different from the lower jaw teeth of Physeter ;
for these are outgrowths of the dental lamina in the form of
a continued laminal outgrowth thickened at intervals to form
the actual teeth rudiments, which remain connected by the
unaltered laminal outgrowth.

The fusion between successive teeth in this the youngest
stage of Physeter as yet known may have some bearing upon
the theory of tooth origin, i.e. as to whether separate teeth,
like those of Physeter, are primitive, or show signs of the
breaking up of a complex tooth series. Are the unions between
the individual teeth a promise of a later concrescence, or the
remains of a separation of the cusps of a complex tooth? Ihave
not, however, been able to ascertain any further facts which
bear upon this most interesting topic. I can, for example,
see no gaps which might mark the boundaries of pre-existent

30 FRANK E. BEDDARD

multicuspidate teeth, or, on the other hand, show by this
arrangement the specialization of sets of separate cusps—
a promise of separate multicuspidate teeth in the future. I
may remark, however, that these connexions between successive
teeth may possibly be related to the fact that in the adult the
individual teeth are connected by a tough gum which comes
away with them when they are forcibly removed from the bony
trough in which they lie.
ResuME.

As to the facts contained in the above pages they are really
summed up in the illustrations which accompany the letterpress.

There are tooth germs from end to end of both upper and
lower jaws, except at the posterior end of the series, where the
dental lamina is not specialized in the upper jaw for some
little distance. The teeth, in fact, are developed from behind
forwards.

The dental lamma extends into the subjacent mesoderm at
absolutely right angles to the horizontal plane of the head.
The dentine papilla is represented only by a condensation of
nuclei in the mesoderm ; it does not yet extend into the cavity
surrounding the enamel organ.

The tooth rudiments of the lower jaw are borne upon a
shorter dental lamina than those of the upper jaw. They
consist of the dental lamina which is prolonged beyond the tooth
germ as a ‘ residual lamina ’ and of the tooth germ arising from
the lingual surface of the lamina.

The tooth rudiments of the upper jaw are borne upon a
longer dental lamina than those of the lower jaw. They
consist of the dental lamina which is prolonged beyond the
tooth germ as a ‘residual lamina’; but this consists, not of
a continuous lamina as in the lower jaw, but of a series of
processes one to each of the successive tooth germs. In
addition to these there is always a second outgrowth of the
dental lamina on the labial side (sometimes doubled) lying nearer
to the origin of the dental lamina, which do not form a con-
tinuous lamina but are separate outgrowths corresponding

TEETH OF PHYSETER 31

each (or each two) to a tooth germ. In the anterior half of the
jaw the tooth rudiments also possess two labial outgrowths,
of which the first, i.e. that nearest to the tooth germ, is longer
than in the posterior series of tooth germs, while the second,
i.e. that nearest to the origin of the dental lamina, arises partly
from the dental lamina and partly from the oesophageal
epithelium, and forms a short pyramidal process. I have not
seen intermediate conditions.

There is also a differentiation into two series of the teeth
of the lower jaw, but the anterior teeth seem to differ merely
through greater age.

In both the teeth of the upper and of the lower jaws the
permanent tooth rudiments (i.e. those outgrowths from, and
on the lingual side of, the dental lamina) are not isolated
outgrowths of the dental lamina but are connected successively
by a continuous outgrowth of the dental lamina, as follows :

In both lower and upper teeth the individual tooth germs
arise from a marginal thickening of the dental lamina, but the
subsequent course of the development differs in the two series.
In the teeth of the lower jaw the thickening is shifted to the
lingual side of the dental lamina by the freeing from it of
a residual lamina on the labial side. In the upper jaw the
corresponding thickening at the distal edge of the dental lamina
is bent over to the lingual side, while a later formed residual
lamina continues at intervals the straight line of the dental
lamina; thus the tooth germ has grown to he laterally
instead of being formed in situ.

As to the homologies of the various regions of the embryonic
teeth, it has been attempted to show that there are reasons
for believing that the teeth of the adult correspond to the
permanent dentition of other mammals, that there are also
rudiments of precedent milk dentition, and that a residual
lamina succeeds the rudiments of the permanent dentition.

32

if

2.

3.

4.

5.

6.

a.

8.

9.

10.

FRANK E. BEDDARD

LITERATURE.

Beddard.—** Further Contributions to the Anatomy of the Sperm
Whale (Physeter macrocephalus) based upon an Examination of
two additional Foetuses ’’, ‘ Ann. Durban Mus.’, ii (part iv), 1919.

Kiikenthal.—‘‘ Walthiere ”’ in ‘ Denksch. nat. Ges. Jena’, Th. ii, 1893.

Leche.—** Entwicklungsgeschichte des Zahnsystems der Saéugethiere ”’,
* Bibl. Zool.’, vi, 1895, Heft 17.

Pouchet et Beauregard‘ Recherches sur le Cachalot’’, ‘ Nouv.
Arch. Mus.’ (3) i, 1889.

Ritchie and Edwards.—‘‘ On the Occurrence of Functional Teeth
in the Upper Jaw of the Sperm Whale ’”’, ‘ Proc. Roy. Soc. Edin.’,
xxxili, 1913, p. 166.

Tomes.—‘ Dental Anatomy ’, 6th ed., 1904.

Wilson and Hill—*‘ Observations upon the Development of the
Teeth in Perameles, &c.”’, ‘ Quart. Journ, Micr. Sci.’ (n.s,), xxxix.

‘* Observations on Tooth Development in Ornithorhynchus ”’,
ibid., li.

Woodward, Martin.—‘‘ Contributions to the Study of Mammalian
Dentition, Part I”’, ‘ Proc. Zool. Soc.’, 1893, p. 450.

‘* Contributions to the Study of Mammalian Dentition, Part I1”’.

ibid., 1896, p. 557.

The ‘Segmentation Cavity’ of the Egg of
the Frog.
By

Alexander Meek.

With Plate 1 and 6 Text-figures.

Tue ege of the frog, Rana temporaria, starts its career
of segmentation as does that of Amphioxus, but after the one-
celled blastula stage a delamination of cells takes place, with
the result that the wall becomes multicellular. The cavity is
excentrally situated, the anterior wall being thin and the
posterior wall thick. The latter consists of endoderm, and
while internally it is impossible to point to a distinct demarca-
tion between ectoderm and endoderm, externally an equatorial
zone is defined which marks the limits of ectoderm and endo-
derm, and this is emphasized when the zone becomes a ring of
proliferation indicating the beginning of the activities of the
blastopore. The proliferation is followed by the appearance
of a dorsal groove, the dorsal lip of the blastopore. An in-
vagination takes place, confined, however, to the dorsal aspect
of the embryo, and a horizontal slit-like cavity is produced
which penetrates anteriorly above the segmentation cavity and
is extended posteriorly as the dorsal lip of the blastopore
advances over the large cells of the yolk-plug. This cavity is
the enteron, and while it is forming the original cavity, which is
still called the segmentation cavity, acquires a cup shape,
the anterior wall being definitely resolved into ectoderm, and
the endodermal margin of the cup gradually narrows and
ultimately fuses, so that the cavity becomes enclosed in endo-
derm. It is almost universally believed that the segmentation
cavity is compressed by the expanding enteron and_ that
it finally disappears antero-ventrally in the endoderm. But

NO. 265 D

34 ALEXANDER MEEK

A. M. Marshall, O. Schulze, and O. Hertwig have all drawn
attention to cases of fusion between the two cavities, cases
which indicate that sometimes the segmentation cavity may
become converted into the forward part of the enteron. Such
have been regarded as abnormal, and upholders of the orthodox
view of the procedure which takes place within the egg of the
frog explain them as being due to faulty preservation or prepara-
tion. On the other hand, the accepted view is not supported
by a convincing series of figures. A stage is shown which
presents the segmentation cavity and the enteron. The next
stage is one which displays the enteron occupying the place of
the segmentation cavity and the latter as non-existent or
as a small cavity of the endoderm ventral to the enteron.
Even if such a change actually takes place it is plain that the
subsequent excavation and thinning of the ventral and lateral
walls of the endoderm restores the cavity to the enteron, and
the process could be explained as a variation of that produced
by an earlier fusion of the two cavities. With this in mind
it may be said, to begin with, that, if the segmentation cavity
sometimes produces the anterior part of the enteron and
usually disappears as a cavity enclosed by endoderm, the
nomenclature of the cavity is wrong and misleading.

With a view to obtaining evidence of what actually occurs,
I requested the laboratory steward, D. C. Geddes, to prepare
serial sections of a large number of embryos. The eggs were
gathered in the neighbourhood of Neweastle. They were in
the process of segmentation, and those preserved six to seven
days after collection furnished all stages from the end of the
period of delamination to that in which the yolk-plug is
reduced to a small spot at the posterior end of the egg. These
I have now examined.

I was quite prepared to find that the first-formed enteron,
the so-called segmentation cavity, was actually replaced by the
secondary or neurenteric enteron, but no such example pre-
sented itself. In all cases I found that the segmentation cavity
is converted by delamination into a primitive enteron distended
by dissolved yolk-products, and that it is put by fusion into
communication with the enteron formed underneath the

TEXT-FIGS. 1-6.

stages of development

Diagrams of sagittal sections to illustrate the

concerned,

nN

36 ALEXANDER MEEK

dorsal lip of the blastopore. The changes which take place are
illustrated by photographs of sections.

Passing over the stage of completed delamination, figs. 1 and 2
(Pl. 1) are taken from slides 4 and 12 ofa series of horizontal
sections of an embryo with a blastopore and invaginated dorsal
lip (Text-fig. 2). The primitive enteron is, as will be observed,
abstracting food material from the yolk-cells, and the yolk-
cells are being changed and broken down in the process. As
a result, the anterior wall is thinner than in the previous stage.
The blastopore is ushered in as an area of proliferation uniting
ectoderm and endoderm. It is later that the groove appears
dorsally, and still later that it extends ventrally to form
a circular lip. During the earlier period the activities of the
blastopore are mainly confined to the dorsal lip, and in that
region the blastopore is concerned not only in producing the
secondary enteron but in yielding mesoderm which presses
outwards on each side separating ectoderm from endoderm.

Fig. 3 (Pl. 1) is a typical sagittal section of a somewhat
later stage (Text-fig. 3), but it is scarcely necessary to point
out that in this series the anterior end has unfortunately
become flattened. The anterior wall may now be said to
consist of ectoderm. In association with the invagination of
the dorsal lip of the blastopore a margin of endoderm projects
anteriorly, providing for the forward extension of the enteron.
The margin, however, also advances all round the primitive
enteron, thus producing a cup-shaped cavity of the endoderm.
The condition defines more clearly the layers and may be
described as a reversion to a segmentation cavity under the
influence of the blastopore. But, as has already been stated,
this condition of the primitive enteron is a temporary one.

Figs. 4 to 8 (Pl. 1) are from successive slides of a series
of sagittal sections bearing the numbers 4 to 7 and 9 (Text-
fig. 4). These show that the invagination which produces the
secondary enteron is confined to the dorsal aspect of the
embryo, and that the margins of the cup of endoderm are
rapidly approaching and fusing at the anterior pole of the egg.
But, more importantly, they indicate that the invagiation
of the secondary endoderm is accompanied by a breaking down

“SEGMENTATION CAVITY’ OF RANA Bil

of the endoderm intervening between the primitive enteron and
the secondary enteron. The imtervening cells are thereby
reduced to a single layer. I have also a series of transverse
sections of this stage. This process creates an extension of
the primitive enteron and provides for the expansion of the
secondary enteron. The expansion takes place by the thin .
floor being depressed into the primitive enteron, and during the
process the intervening cells are absorbed.

Figs. 9 to 14 (PI. 1) are from a series of transverse sections
of the stage (Text-fig. 6) after the fusion of the two cavities
has taken place, and are photographs of sections from slides 3,
4, 5, 6, 8,9. It will be seen that they bear no evidence of the
loss of the primitive enteron ; the cavity is practically as large
as ever. Incidentally they show that the notochord is con-
tinuous with the rest of the mesoderm, and that it is formed
by splitting off from the lateral wings of mesoderm. ‘These
latter completely encircle the wide anterior region of the
enteron and already begin to exhibit the splanchnocoel.

But the process of fusion is better illustrated by the trans-
verse sections of the younger stage, figs. 15 to 18 (PI. 1),
from slides 11, 12, 13, 15—a stage intervening between Text-
figs. 5 and 6. In the first of the sections photographed, fig. 15
(Pl. 1), the secondary enteron is seen above, not far from the
dorsal lip, and the primitive enteron below. The succeeding
figures demonstrate that the two cavities are converted by
fusion into one cavity. It will be observed that even at this
stage the mesoderm has made rapid progress in enveloping the
endoderm, and not by a process of delamination.

T have preparations besides which show in sagittal section
the fusion depicted especially by Schulze and Hertwig (Text-
fig. 5). These, with the transverse sections submitted, prove
that the thin wall of cells which intervenes between the two
cavities is depressed and absorbed during the process. Some
degree of variation is possible as to the completion of a process
which I am led to regard as the normal one. If it is not the
normal process then a demonstration to the contrary is called for.

The morphological aspects of the results I do not intend to
discuss. But I may be allowed to point out that my observations

38 ALEXANDER MEEK

- demonstrate that the frog and its allies come into line with
the meroblastic Amphibia and with the rest of the terrestrial
Vertebrata. It is evident that the segmentation cavity of
Petromyzon is not, for the reasons given above, a true segmenta-
tion cavity, and evidence is required as to its fate; and this
raises the question of the interpretation of the segmentation
cavity in other aquatic vertebrates. Even amongst the
Urochordates there are examples of the formation of the
enteron from an excavation of endoderm.

SUMMARY.

The examination of series of sections of frog embryos has
shown that in all cases the ‘segmentation cavity ’ is converted,
by fusion with the secondary or neurenteric enteron, into the
forward part of the enteron.

LITERATURE.

1888. O. Schulze.—‘ Zeit. f. wiss. Zool.’, Bd. 47, p. 325.

1891. Robinson and Assheton——‘ Quart. Journ. Micr. Sci.’, vol. xxxii,
p. 451.

1893. A. M. Marshall‘ Vertebrate Embryology ’.

1900. H. V. Wilson.— Anatomischer Anz.’, Bd. 18, p. 209.

1902. A. Brachet.—‘ Arch. d. Biol.’, t. 19.

1906. O. Hertwig— Handbuch der Entwicklungslehre ’, i.

EXPLANATION OF PLATE 1.

Figs. 1 and 2.—Horizontal sections of an early stage (Text-fig. 2).

Fig. 3—-Sagittal section of typical stage indicating an early phase of the
invagination of the dorsal lip (Text-fig. 3).

Figs. 4-8.—Sagittal sections of a later stage, wherein the excavation
of the cells intervening between the primary and secondary cavities has
reached an advanced stage (Text-fig. 4).

Figs. 9-14.—Transverse sections of a stage after fusion of the cavities
(Text-fig. 6).

Figs. 15-18.—Transverse sections of a younger stage than the preceding
to illustrate the process of fusion (Text-fig. 5).

Quart. Journ. Micr. Sci. Vol. 67, N.S., PED

notochord

notochord

neural plate
mesoclerm

endoderm

enteron

ll

neural fold
mesoderm

ectoderm

_mesoderm

Nuclear Divisions in Amoeba proteus.
By
Monica Taylor, S.N.D.

With Plate 2.

No definite answer, so far as I have been able to ascertain,
has yet been given to the question that is evitably provoked
by the spectacular rapidity with which amoebae increase in
number in laboratory cultures, viz. “How does the nucleus
divide ?’ In attempting to answer this question it is important
to remember, a fact which five years’ experience in cultivating
amoebae on a large scale has taught, that the most luxuriant
cultures are subject to periods of depression. Unless the
necessary precautions be taken (5) they die down and remain
dormant for a varying period of time, then gradually recover.

The amoebae which first appear in such ‘ recovering ’ cultures
are minute. When they have attained full size they multiply
at a great rate, luxuriance being eventually again established
in the culture. Divisions of the nucleus by fission must there-
fore be distinguished from nuclear divisions connected with the
‘ sporulation ’ process.

It will be useful to recall at the outset the ‘ resting’ nucleus
and its behaviour in a living amoeba. In the following account
the descriptions given by Doflei (8), Schaeffer (4), Carter
(1, 2), having been verified, are embodied for the sake of viewing
the subject as a whole.

The nucleus is a large discoid body which is rolled over
passively by the streaming endoplasm. Its membrane is
strongly marked. Immersed in the nuclear sap is (i) a centrally
placed, plate-like, conspicuous karyosome which hes in a highly
vacuolated achromatic substance, and (ii) the chromatin.
The chromatin, according to Doflein (8) and Schaeffer (4), is

40 MONICA TAYLOR

confined to the blocks situated normally just under the nuclear
membrane (Pl. 2, fig. 1). The karyosome being plate-shaped
may appear circular (Pl. 2, fig. 1) or band-shaped (PI. 2, figs. 2,
3). For convenience of reference these two views of the
karyosome will be referred to briefly as the karyosome in
‘plan’ and in ‘elevation’ respectively. It consists of two
substances: (1) a ground substance which does not stain so
deeply as (2) another substance in the form of small blocks or
rods which stains like chromatin, but which, according to
Doflem, takes no part in the formation of the chromosomes’
which appear on the spindle of the nucleus when it divides by
karyokinesis. These karyosome blocks have a more or less
round outline when viewed in the plan of the karyosome.
They appear more rod-like in the elevation view of that
structure (Pl. 2, fig. 2).

As the nucleus is rolled about by the endoplasm the karyo-
some presents not only alternate views of its plan and elevation
positions, but a series of positions intermediate between these
two extremes. Another complication now to be described gives
the nucleus a successive variety of appearances when it is
examined in the living condition. The rolling over brings
about a redistribution of the nuclear sap. Now the more solid
nuclear contents, i.e. the karyosome, the chromatin, and the
achromatic network, are very flexible. They easily become
temporarily folded. Most frequently two such folds are
formed from the poles towards the equator, one fold being
nearest the proximal surface, the other nearest the distal surface,
of the nucleus under examination. This has the effect of making
the nuclear contents appear lenticular or dumb-bell-shaped
(Pl. 2, fig. 4). This appearance, which is exceedingly common,
is succeeded by the normal (PI. 2, fig. 1) as the nucleus gradually
steadies itself, and becomes stationary, only however to begin
a new series of convolutions as it is played upon by the streaming
endoplasm. In a well stretched, fairly young and aetive
amoeba which has got a good grip of the substratum these
successive changes in appearance (ef. Pl. 2, figs. 6 and 7) of
the nuclear contents can easily be observed. In order to

NUCLEAR DIVISIONS IN AMOEBA 41

study the phenomenon in ‘fixed’ specimens a fairly large
number of young amoebae should be placed on a slide in
a small drop of water and left in a damp chamber until the
animals begin to creep about on the glass. The fixative should
then be dropped quickly on to them, when, a cover-slip having
been provided, they can be examined microscopically (aceto-
carmine is very useful for this purpose) and the whole range of
positions studied. (The behaviour of a nucleus in a living
amoeba is highly reminiscent of the tossing of a pancake on
Shrove Tuesday !)

For the purpose of this investigation, in addition to an
unlimited supply of living amoebae for examination and of
specimens treated with aceto-carmine, whole cultures were
fixed in corrosive alcohol, in warmed corrosive acetic, and
stained in Delafield’s haematoxylin, carmine,' light green, ete.,
and the individual nuclei studied whole or sectioned. I am
sreatly indebted to Sister Carmela Hayes for much assistance
in making preparations.

In stained preparations the ‘ resting ’ nucleus is a much more
striking object when seen with the karyosome in elevation
than it is when the latter is viewed in plan. A newly formed
daughter nucleus can easily be picked out from a number of
older nuclei by the more brilliant colouring exhibited by the
karyosome, especially in aceto-carmine preparations.

Doflein states that the division of the nucleus is effected
when the amoebae have temporarily drawn in their pseudo-
podia and have assumed a spherical shape, having freed them-
selves from the substratum. They look extremely opaque in
this condition. It is a common sight to see several of these
opaque-looking spheres in any good culture of amoeba. In
one instance, Sister Carmela, who was tending the cultures
during my absence, reported to me that practically every
amoeba in a particular culture (and there must have been
many thousands) had rolled itself into a ball.?

1 A modified formula of Picro-magnesia-Carmine suggested to me by
Dr. J. 8. Dunkerly was also applied.
* The subsequent history of this culture, which shortly afterwards

42, MONICA TAYLOR

I have repeatedly verified the observation that such ball-
shaped amoebae are multinucleate. The commonest number
of nuclei is four, but binucleate, six- or eight-nucleate condi-
tions occur. The extreme rarity of mitotic figures in the many
preparations he examined caused Doflem (8) to leave it an
open question as to whether or no there was another method
of division. The results of my investigation lead me to con-
clude that the mitotic figures given by.Lucy A. Carter (2) for
A. proteus (var. X, Carter; A. dubia, Schaeffer), and
by Doflein (8) for A. proteus (var. Y, Carter) are connected
with sporulation, and ‘ encystment ’. The ordinary vegetative
multiplication, i.e. fission divisions, are effected by a sort
of *‘ budding’ of the nucleus. A search through innumerable
specimens at all times of the year, and at all hours of the day
and night, in many cultures of amoeba, some of them so
luxuriant that the bottom of the glass trough (8 in. in
diameter) appeared whitish because of the enormous numbers
of amoebae lying on it, has failed to give one single example of
a mitotic figure. It is the common experience of cytologists
that wherever an organism is developing rapidly mitotic figures
are sure to be forthcoming if such tissues be examined by
a suitable technique. Hence it would seem that the published
figures of mitosis in A. proteus belong to the sporulation
cycle of the life-history.

In most of the spherical amoebae I have examined the
separate nuclei have been close together. If, however, the
spherical specimens be examined as soon as they have assumed
this form (and this is quite a simple matter where large numbers
of cultures are accessible) appearances similar to those drawn
in figs. 8, 9, and 14 (Pl. 2) are to be found. Here the daughter
nuclei are clearly not completely formed, and the mother nucleus
is clearly not dividing by ordinary mitosis. In short, a series
of preparations may be made showing the gradual conversion
of a ‘lobed’ large nucleus into four daughter products. The
underwent a period of depression and is now (July) full of small amoebae,

would seem to indicate that many of the amoebae must have been pre-
paring to encyst.

NUCLEAR DIVISIONS IN AMOEBA 43

spherical opaque amoebae vary greatly in size, the mother
nucleus giving rise to only two daughter nuclei in the smaller
spheres. To facilitate description of this non-mitotic method
of division, I have used these smaller amoebae, and I have
drawn a series of stages with the karyosome (1) in elevation
(Pl. 2, figs. 15-19) and (2) in plan (PI. 2, figs. 25-31 and 34),
the product of the division bemg two daughter nuclei in
each case.

The process is not a simple halving of the nucleus into two
hemispheres which then roll themselves up into discoid figures,
but seems to be as follows : a patch of chromatin blocks divide,
each block into two (PI. 2, figs. 10-13). This process gradually
extends to the whole of the chromatin. Simultaneously the
karyosome also divides into two, each daughter karyosome
assigning itself to one daughter set of chromatin blocks. The
outer set of daughter products sloughs off gradually from the
inner by the increase of nuclear sap and by the formation of
the necessary additional areas of nuclear membrane and two
daughter nuclei thus result. The distance between these
gradually increases. Very frequently the daughter nuclei in
one preparation show different views of the karyosome (Pl. 2,
figs. 32, 33). Very complicated figures necessarily result during
such a process of division, a binocular eye-piece being essential
for the complete elucidation of the preparations in some Cases ;
especially is this true when the mother nucleus is undergoing
rapid successive divisions into four or more nuclear products
(Pl. 2, fig. 14).

If a spherical individual in which the daughter nuclei have
not been separated be placed on a slide and made to assume the
expanded position, the nucleus is seen to resemble the ‘ folded ’
nucleus described by Schaeffer (4) as bemg characteristic of old
or large specimens of A. proteus. In the light of the above
interpretations of fission divisions these folded ‘lobed’ nuclei
are due to the fact that some external physical condition has
interfered with the amoeba when division was in process, the
so-called ‘lobed’ nuclei being incipient multinucleate struc-
tures. Now the amoebae are easily disturbed. They are

44 MONICA TAYLOR

extremely sensitive to change of position, temperature, or light.
When, therefore, a spherical amoeba is removed from its culture
for purpose of examination it quickly accommodates itself to
its new surroundings, and crawls away without completing the
division of its nucleus. Very complicated appearances are
due to such nuclei rolling about in an active amoeba. With
care to preserve their cultural conditions, however, spherical
amoebae can be removed and isolated, when they are seen
to give rise to daughter products.

These ‘lobed’ nuclei, however, are formed before the
amoebae become spherical. They are very early stages in the
division process, for although the actual separation into
daughter nuclei takes place when the amoebae are spherical,
the division is initiated at a much earlier stage. Most of
the large individuals in a watch-glass culture will be found to
have a ‘lobed’ nucleus. If they become spherical this is
converted into a varying number of daughter nuclei, unless,
as described above, external conditions disturb the process.

BIBLIOGRAPHY.

1. Carter, L. A—‘‘Some Observations on Amoeba proteus”, ‘ Proc.
Royal Phys. Soc. Edin,’, vol. xx, part 4, 1919.

2. ——‘‘ Case of Mitotic Division in Amoeba proteus”’, ibid., vol. xix,
no. 4, 1913.

3. Doflein, F.—“‘ Die vegetative Fortpflanzung von Amoeba proteus, Pall.’’,
* Zool, Anzeiger ’, Bd. xlix, no. 10, 1918.

4, Schaeffer, A. A——‘‘ Note on Specific and other Characters of Amoeba
proteus, Pall. (Leidy), A. discoides, spec. nov., A. dubia, spec. nov.’’,
* Arch, fiir Protistenkunde ’, Bd. xxxvii, 1916.

5. Taylor, M., and Hayes, C—‘‘ The Technique of Culturing Amoeba
proteus ’’, “ Journal of the Royal Microscopical Society ’, pp. 241-4,
1921.

Quart. Journ. Micr, Sci. Vol. 67, N.S., Pl. 2

Scale of Figs. 04 mm.

‘05 m.m.
Scale of Figs. 1.2.6.20.21.25.26.32.33.

NUCLEAR DIVISIONS IN AMOEBA 45

EXPLANATION OF PLATE 2.

Illustrating Dr. Monica Taylor’s paper on ‘ Nuclear Divisions
in Amoeba proteus’,

All figures were drawn with the Abbé camera.

With the exception of figs. 4 and 8 the material for the
drawings was fixed in warm corrosive acetic, and stained in
Delafield’s haematoxylin.

Fig. 1—Resting nucleus. Karyosome in plan.

Fig. 2—Resting nucleus. Karyosome in elevation.

Fig. 3—Resting nucleus. Karyosome in elevation (daughter product
of newly divided nucleus).

Fig. 4—Resting nucleus. ‘ Biconcave’ form of contents (aceto-
carmine).

Fig. 5—Resting nucleus. Karyosome contracted.

Fig. 6—Resting nucleus. Karyosome rolling over.

Fig. 7—Resting nucleus. Nuclear contents assuming ‘ biconcave’
form.

Fig. 8.—Aceto-carmine preparation of nucleus from a spherical amoeba.

Fig. 9——Permanent preparation of nucleus from a spherical amoeba.
Two pairs of daughter nuclei not completely separated.

Fig. 10—Early stage in division of nucleus. Increase in number of
chromatin blocks.

Figs. 11 and 12.—Early stage in division of nucleus. Increase in number
of chromatin blocks. Division of karyosome.

Fig. 13.‘ Lobed’ nucleus—assuming ‘ biconcave’ condition. Chroma-
tin and karyosome dividing. Nuclear membrane beginning to divide.

Fig. 14.—Complicated nucleus consisting of four incipient daughter
nuclei.

Fig. 15.—Later stages in division. One daughter nucleus with its com-
ponent chromatin blocks and karyosome is freeing itself from the other.
Karyosome in elevation,

Figs. 16—-19.—Later stages in division. Karyosome in elevation.

Figs. 20 and 21.—Later stage in division. Karyosome in elevation.

Fig. 22—Daughter nuclei almost separate.

Fig. 23.—Daughter nuclei completely separate, but lying close together
in the amoeba.

Fig. 24.—Daughter nuclei rolling away from each other. Karyosome
partly in ‘ plan’, partly in ‘ elevation’.

Figs. 25 and 26.—Stages in division of karyosome (plan).

46 ' MONICA TAYLOR

Figs. 27 and 28.—Later stages in division of nucleus, ‘ Plan’ view
corresponding to ‘ elevation ’ view of karyosome shown in fig. 17.

Fig. 29.—Stage in ‘ plan’ corresponding to fig. 19. ‘ Elevation’ view
of karyosome.

Figs. 30 and 31—Daughter nuclei separating. Plan view of karyosomes,

Fig. 32.—Daughter nuclei showing plan and elevation view of karyo-
some.

Fig. 33—Ditto, but slightly older.

Fig. 34.—Slightly later stage than in fig. 29.

Fig. 35.—Nuclear contents greatly contracted—probably due to fixation.

On Amphilina paragonopora, gp. n.,
and a hitherto undescribed Phase in the
Life-History of the Genus.

By

W. N. F. Woodland, D.Se. (London),

Wellcome Bureau of Scientific Research,! 25-7 Endsleigh Gardens,
Gordon Square, Euston Road, London, N.W. 1.

With Plates 3-5

CONTENTS.

PAGE

Part I, THE STRUCTURE AND BIONOMICS OF AMPHILINA PARA-
GONOPORA, Sp.n.. - : : - : ees
(a) Habitat and External Characters. ‘ : : sl) kD
(b) The Reproductive System ; : ; wide
(c) The Proboscis:; its Musculature and Gaadecion: : ee ts)
(d) The Excretory System . ; ‘ : ‘ : gs Op
(e) The Central Nervous System. patel 67
({) The Histology of the Body-wall, and the Body AMoseulabure 7 OS
(g) Some Stages in the Development of the Larva . 3 68

(h) Re-definition of the Genus Amphilina, and the chief agin!
tions between the five Species. 6 ‘ ; oT aw

Part il. On THE [RREGULAR-FORM STAGE OF DEVELOPMENT IN THE

LIFE-CYCLE OF A. PARAGONOPORA ; ; : 74
Notes ON TECHNIQUE , ‘ ; ; : : : : 78
SUMMARY OF PRINCIPAL CONCLUSIONS IN Parts I anp II. Ba its:
LITERATURE REFERENCES . ; : ; : s ‘ 5 tet

1 A large portion of the work involved in this inquiry was done in
India, during my tenure of the Chair of Zoology at the Muir Central
College, Allahabad, U.P. (Indian Educational Service).

48 W. N. F. WOODLAND

Part I.

Tur StructuRE AND Bronomics or AMPHILINA
PARAGONOPORA, §p. N.

THE new species of Amphilina which forms the subject-
matter of this paper constitutes the fifth+ known to exist
and the third to be described in detail. I have supplied
a detailed description of the anatomy of this new species
because it appears to me that the descriptions of Amphilina
supplied by previous authors leave much to be desired in certain
respects, because the present species differs in several particulars
from A. foliacea and A. liguloidea, and because it
is necessary to elucidate the exact nature and mode of function
of the proboscis apparatus—the proboscis, the enormous
attached proboscis muscle (or bundle of gland-ducts according -
to one interpretation), and the conspicuous giant cells which
Salensky termed ‘ problematische Zellen ’—an apparatus which
has been much misdescribed and misunderstood. I am also
able to describe an important phase in the life-history of this
form, which has not been mentioned in connexion with any
other species of Amphilina or, indeed, Cestodaria, so far as
IT am aware.

I may add that A. paragonopora is the second species of
Amphilina discovered in or near India. The first (A. magna)
was found by Southwell (8) in the coelom of a marine fish
(Diagramma crassispinum) from the coast of Ceylon,
and in general shape and structure is apparently similar to
A. paragonopora, and it is to be regretted that Southwell’s
sole description of its genitalia is so vague. A. paragono-

1 The other four are A. foliacea, Rudolphi, 1819 (vide Salensky,
7: Cohn, 1; Hein, 2; Pintner, 6; Wagener, 9) and A. neritina,
Salensky, 1874 (7), from Europe; A. liguloidea, Diesing, 1850 (vide
Monticelli, 5; Janicki, 3), from Brazil, and A. magna, Southwell,
1915 (8), from Ceylon. The anatomy of A. neritina (assuming it to be
a distinct species) is apparently very similar on the whole to that of
A. foliacea, only differing in small details, and Southwell has only
roughly indicated the general disposition of the genitalia in A. magna,

ON AMPHILINA PARAGONOPORA 49

pora, as will be stated, is found in a fresh-water fish in the
Ganges and Jumna.

(a) Habitat and External Characters.

Amphilina paragonopora is parasitic in the body-
cavities of the two closely related specics of Siluroid,
Macrones aor and M. seenghala!—the common
*'Tingra ’—found in the Ganges and the Jumna at Allahabad
and elsewhere in the United Provinces and in the Panjab,
and supplied to the bazaars as food for the lower caste natives.
In size the parasite varies greatly according to its stage of
erowth and the degree of contraction of the body musculature.
The largest specimens (two) Ihave found measured when lving
(and uncontracted) 280mm. im length (when preserved the
specimens contracted to 170mm.) and 5mm. in maximum
breadth ; on the other hand, the smallest specimens only
measured 10 mm. or 11 mm. in length and 1 mm. in breadth.
Specimens measuring 60-70 mm. in length are fairly common ;
occasionally specimens are found which are slightly broader
in proportion to their length than the above measurements
indicate. The parasites are thus distinctly rnbbon- or strap-
shaped (PI. 3, fig. 1), the thickness bemg one-fifth to one-tenth
of the breadth. There is no scolex, but the ‘ anterior ’? end
tapers slightly, is either rounded or pointed (according to the
state of contraction), and carries a short muscular globular
or ovoid evaginable organ (PI. 3, fig. 1, k)—the * proboscis ’—
which is not a sucker (acetabulum) and is never used for

1 The chief distinction between Macrones aor and M. seenghala
is that in the latter the maxillary barbels are much shorter than in the
former. These fishes are said to grow to a length of six feet, but my
largest specimens did not exceed three feet and the majority were between
one foot and two feet in length. I may add that the food of these fishes
consisted of insects (chiefly Orthoptera) and small fishes, and an examina-
tion of the tissues of this food fauna may reveal later the intermediate
host or hosts of Amphilina paragonopora.

2 T apply the term ‘ anterior’ to the mobile end of the body possessing
the proboscis in a descriptive and not in a comparative morphological
sense.

NO, 265 i

50 W. N. F. WOODLAND

attachment. The ‘ posterior ’ extremity of the body is similar
in form to the anterior, save that at the extreme end a well-
marked bay or semicircular inlet is situated, in the centre of
which is a contractile papilla which bears the three separate
openings of the ductus ejaculatorius, terminal excretory duct,
and vagina (PI. 8, figs. 17,11,12). In colour the parasites vary
from a creamy-white to orange-yellow (the usual colour being
a distinct yellow), and are almost always identical in tint with
the masses of fat attached to the mesentery of the fish they
infest.

The above statements concerning the dimensions and shape
of the parasites are based on the appearance of the parasites
when the body-cavity of the fish is first opened. So long as
the parasites are not disturbed they exhibit no active move-
ments, save perhaps a to-and-fro motion of the protruded
‘ proboscis ’ which is only visible under the microscope (PI. 4,
fig. 15), but if a parasite be removed from the body-cavity
and placed on a glass slide the body almost immediately shrinks
to about one-half its former length and also becomes pinched
by two, three, four, or more deep transverse constrictions 4
which usually travel down the body antero-posteriorly, new
constrictions appearing anteriorly as the old disappear
posteriorly (Pl. 8, fig. 1, f, g, h); the ‘ proboscis ’ also either
becomes extruded (Pl. 4, fig. 15, a) and active in penetration
movements, or (more usually, especially when the animal is
placed in water) becomes tightly retracted and immobile
(Pl. 4, fig. 15, b).

Amphilina paragonopora is, as I have already stated,
usually found in the body-cavity of the fish. I have most
frequently found it on the mesentery, or on the peritoneum,
or on the surface of the liver. On some occasions I have
found specimens lymg underneath the peritoneum, and once
a specimen, 40 mm. in length, was discovered lying imme-
diately dorsal to the gas bladder, in the dense connective

1 Hein has described similar contractions as occurring in A. foliacea,
but states that they are feeble (the worm is very broad) and move postero-
anteriorly, i.e. in the reverse direction to that described in the text (!).

ON AMPHILINA PARAGONOPORA 51

tissue attaching the bladder to the dorsal muscles and vertebral
column. Another Amphilina, 200 mm. long, was found to
have bored through the wall of the gas bladder into its cavity,
half the worm being in the cavity and half still outside, and
another large Amphilina was found to be buried in the muscles
of the body-wall anteriorly. A. paragonopora is also
occasionally to be found on the external surface of the fish,
and my attention was first drawn to the parasite by discovering
part (56 mm. long, and a portion anterior to this, including the
proboscis, of unknown length, had become detached) of
a large specimen emerging from a perforation at the base of
the left pectoral fin and lying on the outside of the skin (PI. 3,
fig. 3). On cutting open the perforation I found that a further
portion of the worm (34 mm. long) lay in a cavity in the muscles
of the body-wall, and that the remainder of the worm (the
hind portion, 40 mm. long) lay inside the body-cavity. This
Amphilina, therefore, had evidently bored its way through
the body-wall in the neighbourhood of the base of the left
pectoral fin (where the body-wall is very thin), and, had it not
been caught in the act, would doubtless have altogether
escaped from the fish. The uterus of this Amphilina (180 mm.
in length plus an unknown length of anterior portion detached)
was full of nearly or fully mature larvae (Pl. 4, fig. 35). In
another fish also I found a small worm (only 11 mm. long)
lying on (not attached to) the surface of the external skin just
behind the anus. I cannot say whether or not this worm had
escaped from the body-cavity of this fish, but it is possible
(although the worm was of course immature sexually) since
in this fish a large perforation (some 3mm. in diameter) was
found at the base of the right pectoral fin, leading into a large
inflamed cavity in the body-wall and thence into the body
cavity. Out of fifty-one? Macrones aor and M. seen-
ghala, most but not all of which I specially examined for
perforations under the pectoral fins, seven were seen to
possess these perforations (situated either under the left

1 Tn all I examined during this inquiry about one hundred specimens
of Macrones sp., in twenty-two of which I found active Amphilina.

E 2

52 W. N. F. WOODLAND

or the right pectoral fin, or, as on one occasion, under both).
From sixteen only of these fifty-one fishes did I obtain active
A. paragonopora.

I may also mention that no relationship exists between the
size of the fish and the degree of infestation by or size of the
active parasite, since some of my largest Amphilina have been
obtaimed from very small fish: thus e.g. in one Macrones
aor 450mm. long [ found in the body-cavity my two largest
specimens of Amphilina (each 280 mm. long), also one about
40 mm. long, two 20mm. long, one 15mm. long, and two
10mm. long, and in a Macrones seenghala only just
over 220mm. long | obtained three Amphilma (70 mm.,
64mm., and 59mm. long), while in many other fishes, over
900 mm. in length, I obtamed only one or two Amphilina
measuring not more than 20-380 mm.

In addition to what I have called the active parasite, there
also exist in a still larger percentage of the fishes masses of
tissue of various sizes and usually of irregular form which
closely resemble the parasite in colour and to some extent
consistency (Pl. 3, fig. 2). These masses, which range from
minute spheres and ovoids (PI. 38, fig. 2, a) up to large shapeless,
often very thick, bodies (Pl. 3, fig. 2, c—7), are usually attached
to the mesentery, but are also to be found free in the body-
cavity, and, although of the same colour, vet can easily be
distinguished from masses of fat by their more solid texture.
Though evidently consisting of the same kind of tissue as the
parasites, yet they can usually be easily distinguished from
these, not only on account of their lack of definite form, but
by the fact that they cannot be flattened out between glass
slides. In some cases the connexion between these masses
and the parasites is betrayed by parts of these masses assuming
the characters of parts of the body of the active Amphilina
paragonopora (usually the anterior or posterior end—
PL. 3, fig. 2, m,n), and in other cases these masses are almost
exactly similar in form to the parasite (Pl. 3, fig. 2, g, h, 7, k)
and can only be distinguished by their total unmobility (lack
of contractions) when removed from the fish and the fact,

ON AMPHILINA PARAGONOPORA 53

already mentioned, that they cannot be flattened out between
glass slides. We must, therefore, assume provisionally that
these masses represent as a rule bodies which would, in the
ordinary course of events, become transformed into active
Amphilina paragonopora. They undoubtedly represent
a phase in the life-history of Amphilina paragonopora,
and they will be described and discussed in detail in Part II.

(b) The Reproductive System.

The general plan of construction of the reproductive system
of Amphilina paragonopora (Pl. 3, fig. 8) is similar
to those of A. foliacea (vide Pl. 3, fig. 9, adapted from the
figures by Salensky and Heim) and A. liguloidea (as
described and figured by Janicki), but, as we shall see later,
there are a few noteworthy differences. Previous to describing
the genitalia of A. paragonopora in detail, it is necessary
to decide as to which surface of the animal is to be labelled
‘dorsal’, since previous authors have been by no means
unanimous concerning even this essential preliminary: thus
e.g. while Salensky and Wagener consider that the side of the
body on which the uterus arises from the ovary is the left,
Monticelli assumes that it is the right (the view taken by
the present writer), while other authors apparently avoid the
subject. In typical mesozoan Cestoda it is the rule that the
surface of the body to which the ovary is the more adjacent
is to be regarded as the ‘ ventral’ surface ; and that the ootype
is situated on the ‘ ventral’ side of the ovary. In Trematodes
also,it is a rule that the vitelline ducts enter the ovary on its —
‘ventral’ aspect, whatever may be their subsequent course.
If we apply these rules to Amphilina paragonopora,
then Pl. 3, figs. 8, 10, 11, and 12 represent the organs as seen
from the dorsal! aspect, as proved by the study of serial
transverse sections. It will be observed from these figures
that when viewed from the dorsal aspect, (a) the uterus arises
from the right side of the ovary, runs forward nearly to the

1 Wagener’s ventral aspect.

54 W. N. F. WOODLAND

anterior end of the body and returns on itself forming a loop,
then crosses posteriorly to the left side of the ovary, and again
returning on itself (making a short posterior loop to the left
of the ovary) runs forward and opens on the left + side of the
proboscis at the anterior extremity, and (b) the sperm ducts
lie dorsal to the uterus where they cross. These two features
are apparently common to all three species of Amphilina, and
we may therefore assume for all three species that when the
animal is so placed that the proboscis is directed away from
the observer and the uterus arises from the right side of the
ovary, we are then viewing the animal from the dorsal
aspect.

I will first briefly describe the genitalia of A. paragono-
pora. Pl. 8, fig. 8 shows the general disposition of the
genitalia, and the three diagrammatic sections the relative
positions (dorsal and ventral) of the various organs relative
to each other. Pl. 8, fig. 10 shows the arrangement of the
various ducts in the neighbourhood of the ovary. The oviduct
(ovp) arises from the ovary towards the right posterior corner
and, as shown, soon opens into a small spherical chamber, the
fertilization chamber (Fc), into which also opens the slight
terminal dilatation of the vagina, the receptaculum seminis
(rs). The fertilized eggs then escape from the fertilization
chamber by an opening situated close to that of the oviduct,
which leads into what may be called the zygote or fertilization
duct (zp). The zygote duct shortly receives the common
vitelline duct (vp) and then dilates slightly to form the recepta-
culum vitelli (rv), after which its walls become thickened and
glandular to form the shell gland (suex), and this portion of the

1 Wagener figures the opening in A. foliacea as lying on the same
side of the proboscis as the first limb of the uterus, but if this was the
case his specimen must have been abnormal, since Salensky, Cohn, and
Hein all figure it on the opposite side in this species, and the opening
lies on what I designate the left side in A. liguloidea, A. magna,
and A. paragonopora. It is a pity that Braun (Bronn’s ‘ Thier-
reich’, Bd. iv, Abt. 1 b, 1894-1900, p. 1155) and Benham (Platyhelmia

in Lankester’s ‘ Treatise on Zoology’, part iv, 1901, p. 100) should have
popularized Wagener’s figure of A. foliacea,

ON AMPHILINA PARAGONOPORA 55

duct is known as the ootype (orp). The uterus (v) is continuous
with the ootype and is a thin-walled much convoluted duct of
wide diameter (capable of great expansion when filled with
grown larvae), considerably more than three times the length
of the entire body (taking the convolutions into account)
and having the characteristic conformation shown in the
figures—a conformation which is found in all five species of
Amphilina. The ducts lying in the proximity of the ovary—
from the receptaculum seminis to the beginning of the uterus—
are embedded in a dense cushion of connective tissue (c'r),
probably protective in function. The common vitelline duct
is formed by the junction of the vitelline ducts (vp) of the
two sides of the body and all run dorsal to the ovary and
the uterus. The vitellaria (vir) extend, as shown, in a row
along each edge of the body for nearly its entire length (Pl. 3,
figs. 8, 14).

The uterus does not contain eggs until the worm is at least
30-85 mm. in length, and then unsegmented eggs are only
present at the very commencement of the uterus, the rest of
the uterus being quite empty (observed in three specimens).
In a worm about 41 mm. long about a quarter of the first
(proximal) limb of the uterus is filled with embryos in the
blastomere stage of development—groups of four, eight, and
twelve blastomeres having been observed. In a worm circa
75 mm. long, the uterus is full of eggs in various stages of
development, but mature larvae are not present. In the worm
which I described as in the act of escaping from the fish through
the body-wall and which was more than 130 mm. in total length,
mature larvae in large numbers were found in the terminal
(third) limb of the uterus, the other two limbs contaiming larvae
in earlier stages of development. I may mention that I have
observed the larvae escaping from the uterus to the exterior
in an Amphilina not more than 100 mm. in length.

The testes (TES), as shown in Pl. 8, figs. 8 and 13, consist,
on each side of the body, of a row of saes (PI. 8, fig. 13) opening
at frequent intervals into the convoluted vas deferens (sp).
They lie just internal to the row of vitellaria, and, like these,

56 W. N. F. WOODLAND

extend over the greater part of the body length. The two
vasa deferentia unite to the left of the middle line at about
the level of the hind end of the ovary, the right vas deferens
crossing the ovary on its dorsal side. The common sperm
duct then runs posteriorly to the left of the vagina, and just
anterior to the terminal opening becomes invested by a thick
coat of muscular tissue 1 (mrp) and then forms the thick-walled
ductus ejaculatorius (DEJ). The arrangement of the terminal
openings of the ductus ejaculatorius, vagina, and terminal
excretory duct in A. paragonopora differs considerably
from those of other species of Amphilina and merits careful
description. Pl. 8, fig. 11 shows the general rather indeter-
minate appearance of these ducts as observed in flattened and
therefore distorted specimens of A. paragonopora.
This figure shows at least that all three ducts possess
openings to the exterior which lie very close together (hence
the name of the species ‘ paragonopora ’, suggested to me by
Dr. Baylis) and which are situated on the small median papilla
enclosed in the bay (Bay) at the posterior extremity of the
body, but to ascertain the exact inter-relationships of these
ducts and apertures it 1s necessary to study series of horizontal
and sagittal sections through this region (Pl. 3, fig. 12). Such
series show (1) that the vagina opens on the dorsal side of the
base of the papilla when this is extended (see Pl. 3, fig. 1, 1) ;
(2) that the opening of the ductus ejaculatorius to the exterior
is terminal on the papilla (pz); and (3) that the opening of
the terminal excretory duct (rep) is also terminal on the
papilla and lies, so far as I can ascertain from sagittal sections
and flattened whole-mount specimens, just to the left of the
opening of the ductus. In some specimens the opening of the
excretory duct appears to be confluent with that of the ductus.
Thus all three openings lie very close together—so close that
for a long time I thought they were all confluent. I may add
that the excretory duct is situated ventrad to the ductus
ejaculatorius where the former crosses the latter, and that

+ This is commonly called a ‘ prostate gland ’, but there is certainly no
evidence of its glandular nature in my preparations.

ON AMPHILINA PARAGONOPORA 57

the tip of the ductus is markedly thick and muscular, forming
a penis. There is no conspicuous cirrus sac and no penial setae.

I will now briefly note some differences which exist
between the genital system of A. paragonopora, as just
described, and the genital systems of A. foliacea, A. ligu-
loidea, and A. magna, as described by authors already
named. All these three latter species are apparently distin-
cuished from A. paragonopora by the fact that’ the
vaginal aperture lies at a considerable distance apart from the
penial aperture: in A. foliacea it is separated by about
one-third the distance between the base of the ovary and the
end of the body, and is situated on the edge of the left side
(the ‘ right * side of Salensky, who, according to my determina-
tion, viewed his specimens from the ventral surface) of the
body; in A. liguloidea it is separated by about three-
quarters of this same distance, and, according to Janick, is
paired 1—-one opening being median and ventral and the other
median and dorsal; in A. magna Southwell figures the
vaginal aperture as being separated by about one-fifth of this
same distance, and lies in the middle line aud it presumably
opens dorsally.

In A. foliacea the posterior half of the vagina hes to the
left of the ductus, but in the other three species the whole of
the vagina lies to the right of the ductus. In A. foliacea
the cirrus sac is not terminal but lies midway between the
ovary and the opening ; in A. liguloidea the sae is figured
as terminal; in A. paragonopora there is no distinct
cirrus sac. Only in A. foliacea are there penial setae in
connexion with the very long penis (setae and length correlated
with the lateral situation of the vagina and the distance
separating the two openings ?). In A. liguloidea alone
the vagina carries an anterior blind diverticulum extending
anterior to the ovary—the so-called ‘anterior vagina’. ‘The
testes are stated to be scattered in A. foliacea but in the
other three species ? they are linear in arrangement, lying just

1 Monticelli only describes a single aperture.
* Southwell states that the testes are ‘scattered about through the

58 W. N. F. WOODLAND

internal to the row of vitellaria on each side of the body.
In all four species the uterus opens at the extreme anterior
end on the left side at the base of the protruded proboscis
(Pl. 4, fig. 16)... Some small differences may exist in the
arrangement of the ducts adjacent to the ovary in the different
species, but, judging from the (in some cases rather doubtful)
figures, the general arrangement found in A. paragonopora
is found in all. I may add that the ventral position of the
vitelline ducts relative to the uterus shown in Hein’s fig. 14
(correctly indicated as dorsal in his fig. 18 however), and
relative to the sperm ducts and the ovary shown in Janicki’s
figs. 5 and 6, is probably an error in both cases—in both cases
the vitelline ducts should be shown as dorsal to these organs.

(c) The Proboscis: its Musculature and
Connexions.

When A. paragonopora is removed from the body-
cavity of a freshly opened fish and placed on a slide in body-
cavity fluid, the extreme anterior end of the worm is sometimes
seen to be protruded into a narrow process which moves
vigorously from side to side in groping movements (Pls. 3, 4,
figs. 1, k, and 15, a). More usually, however, this protrusible
portion of the anterior end is not visible, it having been tightly
retracted inside the body-contour (Pl. 4, fig. 15, b). This pro-
trusible portion of the anterior extremity I shall term the
‘ proboscis ’, and, as will be seen, it is essentially an introvert
in structure. In three or four of my specimens the proboscis
chanced to be preserved in a protruded condition (PI. 4, fig. 16),
while in all the others the proboscis was retracted (PI. 4, fig. 17).
From these figs. 16 and 17 it will be seen that the proboscis
essentially consists of a thickening of the wall at the extreme

parenchyma’ in A. magna, but his figure shows that the arrangement
of the testes is linear, asin A. liguloidea and A. paragonopora.

1 T am unable to understand why Southwell (8, p. 327) supposes that
A. foliacea has no uterine opening ‘at the base of the small anterjor
end of the worm ’,

ON AMPHILINA PARAGONOPORA 59

anterior end of the worm to form a thick-walled invaginable
conical bulb (Pls. 3, 4, figs. 1, k, and 16). In transverse section
(Pl. 4, fig. 18) it can be seen that the cavity of the retracted
bulb is star-shaped in outline, and in longitudinal sections of
the protruded proboscis it is seen that the portion of thickened
wall at the very extremity of the worm forms a kind of terminal
cushion (Pl. 4, fig. 16, rr). The lateral walls of this bulb
resemble the ordinary body-wall in general histological structure,
but there are three very striking and significant differences,
and had one of these been observed by previous writers, the
true function of the proboscis and the huge so-called * retractor *
muscle attached to it would have been obvious. The first
difference is the total absence of circular and longitudinal
muscle-layers, the second is the presence of very distinct large
radial muscle-fibres (PI. 4, fig. 16, rtm) which extend up to
the base of the cuticle (Pl. 4, fig. 19), and the third the minute
but well-marked serration of the cuticle (srr) which covers
the greater part of the outer surface of the protruded proboscis
(Pl. 4, figs. 16, 16a) but not the terminal thickening (t7),
which is altogether devoid of a cuticle. This serration of the
proboscis cuticle has not been previously observed. With
regard to the second difference mentioned, the radial fibres
(rtm) which originate immediately under the serrated cuticle
at first run longitudinally, i.e. parallel with the cuticle (Pl. 4,
fig. 16, a), but soon bend nearly at right angles t and run direct
to the surface of the big muscle (BM) shown in Pl. 4, fig. 16
as occupying the axis of the proboscis and then bend again and
run parallel to the surface of the big muscle posteriorly, and
at the hind end of the proboscis these muscle-fibres diverge,
extend posteriorly in the general parenchyma of the central
core of the body for a considerable distance, and finally
apparently become continuous with the parenchyma. It 1s
difficult to determine how far these fibres extend posteriorly
because they are not easily distinguishable from the ordinary

1 Some anteriorly attached fibres do not bend but run longitudinally

in the proboscis wall, crossing the greater number of fibres in so doing.
They do not form a layer of longitudinal muscles.

60 W. N. F. WOODLAND

longitudinal musculature of the body. These muscle-fibres
constitute the true retractor muscle of the proboscis
(rTM), and are very similar in form and disposition to the
retractor muscle-fibres of the proboscis of a simple Turbel-
larian, e.g. Pseudorhynchus bifidus, v. Gr. ‘The
terminal thickening (rr) of the proboscis consists mainly of
short columnar cells and forms a thin cellular pad covering
the anterior extremity ot the huge axial muscle just referred
to (Pl. 4, fig. 16, Bm). It is to be remarked that this introvert
proboscis does not possess extraneous radial protractor muscles
attached to the body-wall.}

Occupying the central axis of the proboscis is the huge muscle
(PL. 4, fig. 16, Bm) which is so conspicuous in Amphilina and
which previous authors have either labelled ‘ retractor’ or,
in view of its disproportionate size to the minute proboscis
which it was supposed to retract, have regarded as a bundle
of gland ducts! Before discussing, however, the views of
previous authors, I will describe the entire apparatus of the
proboscis. The huge muscle can be seen to extend posteriorly
as far back as the anterior end of the ovary (Pl. 4, fig. 20),
its shiny fibres showing plainly in whole mounted specimens
of the worm. Anteriorly, i.e. in the anterior fifth of the body-
length, it is of considerable thickness and occupies at least the
middle third of the body seen in transverse section (Pl. 3,
fig. 5), but more posteriorly it becomes attenuated (PI. 3, fig. 4)
and immediately in front of the ovary only consists of a few
centrally situated fibres. Hach of these fibres (some nearly
as long as the worm itself) can be seen, if traced through serial
sections, to run parallel with the central longitudinal axis
of the body for the greater part of its length, but just before
its termination it always bends at right angles (i.e. becomes
more or less radially disposed in a transverse section) and
becomes connected with one of the remark-

1 Cohn figured extraneous radiating protractor muscles in his drawing
of the proboscis of A. foliacea, but they certainly do not exist in
A. paragonopora. Salensky provides what is probably a much
more accurate figure and shows no extraneous muscles,

ON AMPHILINA PARAGONOPORA 61

able giant cells which Salensky labelled ‘pro-
blematic’. The ‘ problematic cells’ of Salensky are then
simply the attachment or ‘anchor ’-cells of the large axial
muscle. The mere distribution (as seen under a low-power
objective) of these anchor-cells, as I shall call them, is evidence
that they have something to do with the fibres of the muscle,
since in the anterior fifth of the body im which the muscle-fibres
are most numerous, the anchor-cells are most plentiful and
extend laterally to the region of the testes (Pl. 4, fig. 21) ;
whereas, more posteriorly, where the muscle-fibres are much
fewer in number, the number of anchor-cells is also much smaller
and is obviously roughly proportional to the number of fibres
in any given zone (PI. 4, fig. 21). The continuity of the giant
anchor-cells (each as big as the egg-filled uterus im transverse
section) with the fibres of the muscle can be easily seen both in
longitudinal and transverse serial sections, whereas in whole
preparations it is not easy to see the connexion, and this accounts
for the erroneous supposition of Salensky that the processes
from these © problematic ’ cells become sooner or later attenuated
and indistinguishable from the ordinary parenchyma with
which they are connected, since he doubtless did not trace
these processes through serial sections. All these anchor-cells
are elongated and their long axes im all cases lie in a plane at
right angles to the long axis of the body (Pl. 3, figs. 4, 5)—
a significant fact. These long axes are, in the case of cells
occupying the axial region of the body, disposed vertically
(the muscle processes arising from either their dorsal or ventral
ends—Pl. 3, fig. 4), but anteriorly, in those cells which are
situated nearer the sides of the body, the long axes are often
obliquely inclined towards the median axis (Pl. 3, fig. 5).
Cytologically the anchor-cells are, as Salensky remarked, not
unlike nerve-cells, and, indeed, they may be neuro-muscular
in function, though they apparently have no connexion what-
ever with the main nervous system of the worm. Each cell
(Pl. 4, figs. 22, 23) contains a nucleus (with a conspicuous
nucleolus) which hes in a small island of chromatophil cyto-
plasm (very evident in indigo-picro-carmine preparations, in

62 W. N. F. WOODLAND

which the island is stamed red and the rest of the cell-plasm
green), and the remainder of the cell is packed with con-
spicuous granules. The cell process grows out from one pole
of the cell and at first is packed with granules identical with
those in the cell, but at some distance from the cell the substance
of the fibres becomes longitudinally striated (the striations being
sinuous When the muscle is not fully extended) and much less
sranular ; in other words, comes to resemble muscle-substance
(Pl. 4, fig. 24). Another and very significant fact to be men-
tioned concerning the fibres of the big muscle is that, in the
proboscis and for some distance below it, these fibres, at least
in many cases, run obliquely, as is proved by the fact that the
fibres are seen to be cut transversely or obliquely in longitu-
dinal sections cut parallel with the long axis. Pl. 4, fig. 25
gives some indication in very diagrammatic form of the
general arrangement of the muscle-fibres and their connexion
with the anchor-cells—an arrangement also found in A.
foliacea and A. liguloidea and probably in A. magna.

What can be the function of this huge axial proboscis muscle ?
Its extraordinary size enables us at once to dismiss the idea
that it is simply the retractor muscle of the minute proboscis,
as supposed by Salensky? and other authors, especially in
view of the fact that a well-defined retractor muscle (indicated
in A. liguloidea by Janicki in his Text-fig. 9) already
exists. In no other worm do we find a proboscis, of the size
found in Amphilina, associated with a muscle (known to be
a retractor) of the dimensions just described. In Turbellaria
the small proboscis is always retracted by a small retractor
muscle consisting of short diverging fibres ; on the other hand,
when, as in the Nemertines, the retractor muscles are long,

1 Salensky described the fibres of the proboscis muscle as originating
posteriorly in two halves from the lateral subcutaneous muscle-layer,
and naturally could offer no explanation of the presence of his “ problematic
cells >. Hein and Cohn both failed even to find the anchor-cells (the former
confusing them with the myoblasts and the latter assuming that they were
in part identical with his flame-cells and in part ‘ Kunstprodukte ’ !)
though these cells can be easily seen under a magnification of 30 diameters
and less.

ON AMPHILINA PARAGONOPORA 63

the proboscis is also long, and even in the Tetrarhynchus
scolex in which the four retractor muscles of the four pro-
boseides reach, as in Amphilina, to the hind end of the
‘segment ’, each proboscis is at least one-fifth the length of
its retractor muscle.

Wagener (1858) and Lang (1881), on the other hand (neither
of whom can be supposed to have recognized the fact that the
fibres of the large proboscis ‘muscle’ are individually con-
nected with the giant cells which Salensky called * problematic ’),
adopted the view that the whole of the proboscis * muscle ’
is a bundle of elongated ducts connected with ‘Speichel-
driisen ’ (!) lying in the parenchyma, the ducts being supposed
to open on the surface of the proboscis. Needless to say,
apart from the superficial resemblance of the anchor-cells to
gland-cells (and ganglion-cells !), and the necessity of adopting
an alternative to the ‘ retractor’ theory, there is no justifica-
tion whatever for this view, though it has been adopted both
by Pintner and Janicki and referred to im such well-known
general works as Bronn’s ‘Thierreich’ and Lankester’s
‘Treatise on Zoology’. The three obvious facts (evident in
any series of well-stained longitudinal sections through the
proboscis), viz. (1) that there is no trace of a lumen in the
individual fibres, (2) that the fibres are distinctly composed of
muscle substance, and (3) that the supposed ducts do not reach
the surface of the proboscis,! are sufficient by themselves to
dismiss this gland-complex theory, quite apart from the further
facts, already described, of the greater part of the proboscis
being covered with serrated cuticle (the terminal cushion
provides far too small an area for the openings of such a large
number of supposed ducts) and the suggestive twisting of the
muscle-fibres anteriorly, and the consideration that it is
difficult to conceive the necessity for the existence of such an
enormous gland—a gland which, with its ducts, extends
throughout nearly three-quarters of the substance of the body.

The proboscis muscle then, being neither a retractor nor

1 Janicki, in his fig. 9 of the ‘ vordere KGrperspitze’, expressly refrains
from figuring the supposed ducts of his ‘ Frontaldriisenzelle ’ !

64 W. N. F. WOODLAND

a bundle of gland ducts, can only be associated with some
special function of the proboscis. The proboscis, as we have
seen, is not an organ of attachment—a sucker—but it is, on
the other hand, a very efficient organ of penetration. Amphi-
lina, as already related, normally bores its way through the
tough body-wall of the fish in order to liberate its larvae to
the outside world, and we know that it also occasionally bores
through muscles, the wall of the gas bladder, the kidney,
and other organs and tissues. But im order to overcome
resistance a penetrating organ must possess (1) some degree
of rigidity, (2) some instrument with which to tear tissue,
and (3) a powerful muscle to work the apparatus. The large
proboscis muscle—which [ shall in future term the boring
muscle (sm)—fulfils requirements (1) and (3), and the
serrated cuticle covering the outside of the protruded proboscis
fulfils requirement (2). The anterior thickness of the boring
muscle not only supplies the proboscis with a dense more or
less rigid core with which the animal can push its way into
resistant tissue, but the posterior extension and firm attach-
ment of the muscle to the large anchor-cells firmly embedded
in the axial parenchyma (and placed at right angles to the
length of the fibres) enables the worm to * put its whole weight ’
into the boring process and to draw the hind portion of the
body through the perforation or path made by the proboscis.
The slight twisting of the fibres of the boring muscle in and
below the proboscis also doubtless serves for a semi-rotary
movement of the proboscis, enabling the serrated cuticle to
saw its way through the tissue. The proboscis muscle thus
serves as a boring muscle, as a notochord for the anterior end
of the body and as a means of dragging the hind half of the
body along the path excavated by the proboscis, and these
important functions amply account for its huge size. Retrac-
tion of the proboscis is effected by the retractor muscle, and
protrusion of the proboscis probably results merely from the
slackening of the retractor (the stiff bormg muscle naturally
projecting forwards in a position of rest) but possibly also
from the contraction of the longitudinal muscles attached to

ON AMPHILINA PARAGONOPORA 65

the body-wall at the base of the proboscis—the proboscis
being exposed by this means.

(d) The Exeretory System.

Salensky has described the presence in Amphilina
foliacea of two lateral excretory channels (lying one on
each side of the body internal to the nerves) which receive
branches from the parenchyma. This description agrees
essentially with the plan of excretory system which I have
found in A. paragonopora, and I therefore venture to
doubt the accuracy of Hein’s account (with figures) of an
irregular close network of excretory channels occurring in
A.foliacea.! In A. paragonopora, asin A. foliacea,
a large excretory channel extends along the whole length of each
side of the body, lying immediately internal to the testes (Pl. 8,
fig. 6). Anteriorly (Pl. 4, fig. 17) each channel apparently
originates as a narrow channel or loop (equal in calibre to one
of the branches or loops given off more posteriorly) in the
parenchyma situated at the sides of the base of the proboscis ;
posteriorly the two lateral channels converge and meet in the
middle line in a slight dilatation (situated at about a third
of a millimetre from the terminal aperture—PI. 3, fig. 12, pxB)
to form the terminal single excretory duct (TED), which in its
turn runs directly to open externally on the papilla (Pap)
at the extreme posterior end of the body, the opening lying
adjacent to and to the left of that of the penis. In addition
to these two lateral main excretory channels there are four
series of subsidiary (as regards size) excretory channels, which
arise from the two lateral main channels along their entire
length. The first that may be mentioned comprises the
dorsal transverse channels (Pl. 3, figs. 4, 6, prc),
which put the two lateral main channels into direct communicea-
tion across the dorsal side of the body, lying between the

1 T am aware that Cohn speaks of definite lateral channels and a net-
work in A. foliacea, and that Janicki figures a fine network in
A. liguloidea.

No. 265 5

66 W. N. F. WOODLAND

internal longitudinal muscle-layer and the dorsal anchor-cells
and fibres of the boring muscle ; the second series comprises
the similar transverse channels situated on the ventral side of
the body—the ventral transverse channels (Pl. 3,
figs. 4, 6, vrc); the third comprises those channels which
arise from each lateral main channel on its dorsal side and turn
outwards towards the outer edge of the body—the dorsal
external channels (Pl. 8, fig. 4, pc); and the fourth
comprises the similar outwardly directed channels which
arise from the ventral side of the lateral main channel—the
ventral external channels (PI. 3, fig. 4, vec). In many
and perhaps in most cases, the dorsal external channels and the
ventral external channels join together to form a lateral
excretory loop (PI. 8, fig. 6, teL), but in other cases the
two channels do not appear to communicate. Thus, in its main
plan, the excretory system consists (1) of two lateral main
channels which unite posteriorly to form a single exit channel
which opens to the exterior at the posterior extremity, and
(2) of subsidiary smaller channels which take the form, roughly
speaking, of three series of ‘ rings "—the axially situated series
of flattened ‘ rings ’ formed by the dorsal and ventral transverse
channels, and the two lateral series of ‘rings’ arising from
and lying external to the lateral main channels. It must be
mentioned, however, that the upper and lower halves of these
‘rings ’ very rarely lie in the same transverse plane—they only
form a ‘ring’ when a considerable thickness of the body is
viewed end-on. In one worm measuring about 40mm. in
length I observed that in 5 mm. of this length approximately
21 external channels were given off from one of the lateral
main channels—about 150 in the entire length of the worm.

Apart from this system of channels I have been unable to
discover any other portion of the excretory system, though
I have searched most carefully, in both transverse and longitu-
dinal series of well-fixed sections, for flame-cells. According
to Hein and Cohn flame-cells exist in A. foliacea in large
numbers, and the former author figures them; but these
statements need confirmation, since Hein is possibly mistaken

ON AMPHILINA PARAGONOPORA 16H

in his deseription of the general plan of the excretory system
and admits that his preparations were ill adapted to show even
the central nervous system; and Cohn, as we have seen,
suggested that Salensky’s ‘ problematic ’ cells were themselves
the flame-cells! I may add, finally, that I was also unable
to detect cilia in any of the excretory canals.

(e) The Central Nervous System.

In 1874 Salensky remarked upon the presence in Amphi-
lina foliacea of two longitudinal nerve-trunks, and Lang (4)
in 1881 confirmed this, and also described a ‘ brain com-
missure ’ (a band of fibres piercmg the boring muscle) and
branches given off from the two lateral longitudinal trunks.
Cohn in 1904 added the information that the branches given
off dorsally and ventrally from each of the two lateral longitu-
dinal trunks meet dorsally and ventrally across the body so
as to form a series of nerve-rings throughout the length of the
body, but since these rings were not observed by Lang, and
are certainly not present in the elongated and more nearly
cylindrical A. paragonopora, I doubt their existence.
Pl. 4, fig. 29 illustrates the anterior end of the central nervous
system in A. paragonopora, and it will be seen that it
confirms Lang’s description in all respects. Anterior to the
‘brain commissure ’ the two lateral longitudinal trunks extend
forwards and end in the margin of the anterior end of the
body. Posteriorly, in A. paragonopora, the two lateral
longitudinal trunks converge slightly (in accordance with the
narrowing of the body) but do not appear to join: they
terminate separately at the sides of the posterior inlet or
semicircular bay at the posterior end, in much the same way
as the trunks terminate anteriorly. In transverse and longitu-
dinal sections each of the lateral longitudinal trunks is seen to
give off dorsal, ventral, and internal branches (Pls. 8, 4, figs. 4, 5,
29), and these are distributed to the body-wall muscles and
other organs. I have not ascertained the exact numbers of
these branches. The trunks are umform in diameter and
there are no special aggregations of ganglion cells.

F 2

68 W. N. F. WOODLAND

(f) The Histology of the Body-wall and the
Body Musculature.

Pl. 4, fig. 26 shows the appearance of the body-wall in trans-
verse section. The cuticle (cur) is not very thick, and imme-
diately underlying it is the thick ‘ subcuticula’ (suse). The
outer zone of the subeuticula apparently consists solely of
numerous radially disposed thin fibres similar in nature to’
those which compose the general parenchyma. ‘Three muscle-
layers lie in the outer zone of the subcuticula—the thin outer
circular muscle layer (oc), a thin layer of longitudinal muscle-
fibres (ouM), and a second thin layer of circular fibres (10M).
Internal to these three muscle-layers lies the inner zone of the
subcuticula (‘ epidermal’ layer), consisting of spindle-shaped
cells, in between which lie radially-disposed fibres and numerous
caleareous bodies (canc). Underlying the subcuticula is the
parenchyma (PAR), in the outer zone of which hes a powerful
longitudinal muscle-layer (11m), and internally to which there
is to be seen an attenuated scattered layer of oblique longitu-
dinal musele-fibres (optm). Gland-cells (Gc), excretory canals,
and nerve-fibres (Nr) are also of course contaimed in the
parenchyma. Included in this Pl. 4, fig. 26 is a drawing of
a giant anchor-cell (ancc) drawn to the same scale, which will
give some idea as to the enormous size of this class of cell.
Pl. 4, fig. 27 illustrates a longitudinal section through the body-
wall, and this shows the indentations of the cuticle and sub-
cuticula due to the outer substance ot the body-wall being
ridged transversely when contracted longitudinally.

(7) Some Stages in the Development of the
Larva.

The mature eggs in the ovary which are about to enter the
oviduct in an Amphilina paragonopora about 30 mm.
long are of the ordinary alecithal type. After having been
fertilized and passed through the ootype, the eggs normally
become encased, with a quantity of yolk material, in a relatively
thin irregularly oval shell, to one end of which is attached

ON AMPHILINA PARAGONOPORA 69

a short filament or ‘tag’ (PI. 4, fig. 30, /). Occasionally eggs
are shed into the uterus without a shell. Only unsegmented
eggs are to be found in a worm about 30 mm. long, and they are
situated in the portion of the uterus immediately adjacent to
the ovary, the rest of the uterus being empty. In an older
A. paragonopora (a little over 40 mm. in length and cut
into serial horizontal sections) I have observed early segmenta-
tion stages-—groups of four, eight, twelve and more blastomeres
—and in some of the early morulae it is possible to detect one
or two blastomeres which differ from the rest and which are
doubtless destined to form the investing membrane of the
embryo. Ina worm a little over 70 mm. in length the uterus
is full of embryos, the oldest stage being a solid morula which
fills the shell, the morula being surrounded by an investing
membrane (Pl. 4, fig. 31 depicts a very young morula). In
another worm (87mm. long when uncontracted, though it
shrank to 32mm. when placed on a slide) which I cut into
serial transverse sections, the third limb of the uterus was full
of embryos of the stage represented by Pl. 4, fig. 34, while
earlier stages (e.g. that represented by Pl. 4, fig. 82, which I
drew on account of the single large internally situated blasto-
mere shown—GBL) were present in the other two limbs. The
embryos in limbs 2 and 38 of the uterus all possess a definite
ectoderm, while in hmb 3 the embryos are further distinguished
(1) by the possession of a small group of large glandiform cells
(ac) which are drawn out towards one end of the embryo
and may be unicellular glands similar to those found in many
cerariae, and (2) by the elongation of the terminal cells at the
other end of the embryo (Pl. 4, fig. 33) and the secretion by
these cells of the ten (I think) calcareous hooklets so charac-
teristic of the Amphilina larva (Pl. 4, fig. 34). Pl. 4, figs. 38
and 34 are drawn from sections. In three worms each about
100 mm. long when uncontracted, the third limb of the uterus
was full of larvae still contained in their shells, which they
filled and were somewhat longer than the stage represented
by Pl. 4, fig. 34. At this stage the larvae are often liberated
and are about 100 microns in length. PI. 4, fig. 35 represents

70 ® W. N. F. WOODLAND

in optical section the type of larva found in large numbers
in the third limb of the uterus in a worm more than 130 mm.
long (the worm which was in course of boring its way through
the fish body : vide Pl. 3, fig. 3). In these larvae (which were
about 200 microns in length—twice the length of other mature
larvae observed by me), whether mounted whole or in section,
I found it difficult to detect for certain the central group of
gland-cells which is so distinctly figured by Salensky (vide his
fig. 84) and Janicki (vide his fig. 16) and which I have seen
clearly in sections of younger larvae, the probable reason
being that the gland-cells are much distended and full of
unstained secretion: it is difficult to suppose that the gland-
cells have disappeared at this stage. The anterior end of the
larva (the end opposite the hooklets) is usually drawn out
somewhat and probably carries the fine ducts of the gland-
cells. The larva at this stage has usually escaped from its shell
and has secreted a thin but definite cuticle, outside which lies
the remains of the investing membrane. I could detect no
trace of ciliation on any part of the surface.

The foregoing facts were, as I have stated, observed by me
in specimens of A. paragonopora ranging from 30 mm.
to 130+? mm. in length and require no particular comment.
In one of my two largest specimens of this worm (both 280 mm.
long when living and uncontracted), however, 1 found, both in
horizontal and vertical longitudinal sections, in portions of the
body mounted whole and in macerated preparations of the
body, two types of products in the uterus (all three limbs of
which apparently contained the same or similar products) :
(1) oval flattened larvae (Pl. 4, figs. 38, 39) with typical hooklets
and possibly with gland-cells (though I could not observe them
even in sections ; however, large spheres, not shown in PI. 4, figs.
3, 8, could occasionally be distinguished deep in the substance)
and only as long as the larvae liberated from worms 100 mm.
long, i.e. half the length of the larvae represented by Pl. 4,
fig. 351 (though the hooklets were the same size in both)

1 | am unable to say whether the larvae of Pl. 4, fig. 35 are abnormally
large (since they were only found in a single worm) or whether the flat

ON AMPHILINA PARAGONOPORA 71

and of much denser consistency, and (2) oval egg-shells but
little inferior in size to the larvae and contaiming only a few
large dissociated blastomeres of different sizes (Pl. 4, figs. 37,
39). These practically empty egg-shells were extremely
numerous—from five to ten times more numerous than the
oval larvae. They are best displayed by macerating portions
of the Amphilina, either fresh or preserved, in Marcacci’s fluid
(equal parts of nitric acid, glycerine, and water) over-night,
followed by teasing or, better still, by grinding up on metal
gauze suspended in water with a piece of flat wood—the
products of the grinding sink through the gauze and can be
collected by centrifuging. Lest it be thought that the half-
empty condition of the egg-shells was due to the Mareacci’s
fluid, I may mention that I obtained the same result with
simple maceration in water (Pl. 4, fig. 39) and that Mareacci’s
fluid does not damage even such objects as spermatozoa ; the
half-empty egg-shells can also be seen, as already mentioned,
in whole-mounted specimens and in serial sections mounted in
balsam. I may also mention that it was only by employing
Mareacei’s fluid that I first detected the filament or tag on the
egg-shells, it usually being difficult to see this structure in
sections and whole mounts. I suspect that similar maceration
would display a tag on the egg-shells of Amphilina magna,
Southwell.

The half-empty egg-shells undoubtedly represent larvae that
have degenerated. It is conceivable that if for any reason
a worm cannot escape from its fish host, the larvae degenerate
in consequence of not being liberated into water. Facts cited
in Part I afford additional evidence in favour of this view.

(h) Re-definition of the Genus Amphilina, and the
chief Distinctions between the five Species.

Wagener’s definition of the genus Amphilina (9) must be
amended in order to comprise the facts that in A. para-
gonopora the body is ribbon-shaped, the posterior end is

larvae (which are the same length as the largest larvae | have seen in
utero) are abnormally small,

72 W. N. F. WOODLAND

not pointed, the two surfaces are alike in curvature, and that
the proboscis does not bear the openings of a gland complex,
the latter bemg a large bormg muscle, the fibres of which
extend a considerable distance posteriorly (probably about
eight-ninths of the length of the body in the three best-known
species) and become connected individually with giant
‘anchor ’-cells. Thus amended the definition of the genus
reads as follows: Body flat and varying in outline from an oval
to a narrow ribbon. Anterior end pointed or slightly truncated
according to the state of contraction ; posterior end pointed,
rounded, or emarginate. A small evaginable proboscis is present
at the anterior end, and connected with this is a large boring
muscle, the fibres of which end posteriorly in giant ‘ anchor ’-
cells situated in the parenchyma. ‘The exeretory system
consists usually of two main lateral channels connected with
a system of smaller channels and opening posteriorly by an
approximately median single pore. ‘Testes numerous. Ovary
and openings of vas deferens and vagina posterior. Uterus
a long convoluted duct consisting of three limbs (N-shaped
when viewed from the ventral surface), each extending nearly
the entire length of the body, and opening anteriorly at the base
of the proboscis and on the left side (i.e. on the side of the body
opposite to that on which the uterus arises from the ovary).

T may add that in no species is there an ‘ acetabulate sucker’.

Some of the more conspicuous distinctions between the five
known species of Amphilina are stated in the Table on the next
page.

The specific distinctness of A. neritina from A. folia-
cea needs confirmation.

Three type-specimens (including one of maximum size) of
Amphilina paragonopora have been deposited in the
British Museum (Natural History) at South Kensington,
London.

JUSTOSURIIY TROUT
‘avjes Terued ON

(|
Jb

core TP
-prut ur Ssutuedo sny4
-onp 0} Tesiop ysni
AjIwe13x9 puly 4%
Ssuruedo pur snjonp
JO JYSsIL 0} eUIseA
“BIpuy YON JO
SIOATL “VP VY SUIS
yo IOV souordVT
‘97 8UlS
-1eula §=pua O14
-sod ‘papunoir 0
payutod pus rol1ejyuy

+ €-83-8-02/T

*(SupAq]) “WAU gz

(paArosoid) “uur O/T
‘asueIO

0} MoTjeA-AuTvAI_D

‘viodouoseard "vy

qUOTIASURIIY IBOUVT
avyes [erued ON

Ay

-WdI}xX9 PULY ULOTT

‘wut = Z oul, ap

-prut ut (4) ATTestop

suedo puv snjonp
JO JYSII 07 vUTSeA

uojAay) jo

qsvoo ‘urnutdstis
-SV190 VPULUWRAISVICG,

pepunoi pur
AOIIVU spud Tog

8-Z1-9-01/T
(SurAt]) “wut [gg
(poArosead) "wu (GZ

ayy -ATT
‘euseU ‘VW

"GG/[ SVM OLYBI YY SUdUITOEds SutAT] ysosrey Ava uy L
‘UYyO?) pur ‘uLePT ‘AYSuTeY WoT, SSUIMBAP DAY FO SJUBUTOINSvITT 4

qUsMIeSURIIe IROUTT
avyas [erued ON

wy]
-NOTJLIATP LOLIO}UB
sey vuiseA “AQT

-WldI}X9 PUTY WLOIF
WU & YOR sUTT
a]pprut ut ssuruedo
[eqjZUSA PUR [Tes.top
sey pure sngonp
JO JYSII 03 BUIseA

[izerg
‘se5rs vpuredviy

peqyutod pur
MOIIVU spue [Og

0-F-¢ 8/T
(paArasard) ‘wut 9g

ayy M-Aars)
“BAIPLO[RSIT ‘VW

pe194
1808 OG OP Pres
ayes [eIUag

oo

sutuedo sngonp uriIpeut
oy} wor Ave “wu Z ynoqe
Apoq jo ulsivut yjoy uo suedo
pue snjonp fo yey 0} Aya1or104
-sod sory vursea sotoeds yjoq uy

‘adommay
‘ds cosuadioy

pepunor

pue Mor
-Ivu spud YO

« 0&-1-Z-T/T
“UU (99-97%

aqtyM-ATUIRIIO
"BQ0VT[OF ‘VW

adoingy
‘ds aasuodtoy

pepunoar

pue Mor
-1eu spue Yo

0:2/T
“TUL QT

ugeis-Aar4)
"BUIQIIOU VW

$9489],
aBqas [VUIg

aan}
-1ad¥8 sJI puv BUTSvA

180H]

sorzruie1yxe Apo
yysue] 03

([ sev uayey) Apoqg jo

YjpRerq WMWIXeU

jo o1j7ea oyeturxoiddy

Yjsue] WNUIXPyTy

anojog

74 WwW. N. F. WOODLAND

Part II.

On THE [RREGULAR-FORM STAGE OF DEVELOPMENT IN THE
LIFE-CYCLE or AMPHILINA PARAGONOPORA.

[ have already mentioned in Part I that in addition to the
active Amphilina paragonopora there are to be found
attached to the mesentery and, in the case of the larger bodies,
lying free in the body-cavity of Macrones aor and
M. seenghala, large numbers of masses of tissue, varying
greatly in size and form and, except in the case of the smaller
bodies, rarely showmg any approach to a definite shape
(Pl. 38, fig. 2), but which, nevertheless, are similar in colour to
the active worm, though they differ in consistency, it bemg
impossible to flatten them between glass slides. They are quite
distinct from the similarly coloured masses of fat, being denser
in appearance and texture. These amorphous? masses are
stages in the development of the active Amphilina, a fact
which is proved by the discovery of stages transitional between
the two (vide Pl. 3, fig. 2, m, n), and they represent a distinet
part of the life-cycle of the species. The portion of life-cycle
from the formation and escape of the larva from the fish to
- the first appearance of the amorphous masses in the mesentery
of the fish is unknown.

The smallest amorphous masses which can be detected in
the mesentery by the naked eye are spherical or ovoid in
shape (Pl. 3, fig. 2, a),? and there exist all transitions from
these small masses up to elongated masses 80 mm. or more
in length (Pls. 3, 5, figs.2 and 46). I have cut serial sections of
some dozens of these masses and also through portions of
infected mesentery, and by close examination of these I have

1 TY use this term, which strictly speaking can only be applied to a gas,
for want of a better. So far as lam aware there is no term in use in English
to express the meaning of indefinite shape as apart from definite. Professor
Platt suggested ‘ ataxomorphic ’,

2 It is impossible to distinguish these with the naked eye from the
numerous Nematode cysts which also abound.

ON AMPHILINA PARAGONOPORA 75

found what are undoubtedly the earliest stages of growth of
these masses, and several other facts of interest.

The earliest certain developmental stage of the amorphous
mass which I have observed was approximately spherical and
measured 29 microns in diameter (PI. 5, fig. 40, x500). The
mass simply consists of a globule of plasm with a dozen or so
nuclei embedded in it, and is surrounded by a relatively thick
capsule of loose concentrically arranged fibres; usually
capillaries are to be seen near by. I have seen about a dozen
of these early stages + (one measuring 40-2 microns in diameter
is shown in Pl. 5, fig. 41, x500, and another larger one in
Pl. 5, fig. 42, x500). It will be noticed that these minute
masses are smaller than the larvae liberated from the uterus
of the worm, and it is therefore evident that the larva must
undergo some process of subdivision, probably in an inter-
mediate host, before infection of another fish can occur. I have
not been able to trace these masses back to a unicellular stage.
Eneapsulated masses larger than these early stages are very
numerous, and of course are easily distinguishable in section
from small Nematode cysts of the same size both by the
absence of the worm, the nature of the contents, and by the
structure of the capsule wall. As I have said before, all stages
of the growth of these masses are to be found, and since it
would be profitless to describe the gradual process of histo-
genesis which commence in masses of about the size shown in
Pl. 5, fig. 45, I have only figured sections of three of the later
stages, the magnifications stated giving some idea of their
relative sizes. Pl. 5, fig. 48 (112) illustrates an encapsuled
mass with tissue still quite undifferentiated, and Pl. 5, fig. 44
(x112) a larger similar mass, and Pl. 5, fig. 45 ( x35) shows
a later stage of the worm after it has become too long for the
capsule and is consequently becoming coiled. ‘The commencing
histogenesis is not indicated in the figure. In still later stages

1 These early stages must be distinguished from sections through nerve-
fibres in the mesentery, which are very numerous and are often twisted
into lumps of about the same size as the masses, and in some cases are sO
small as to resemble unicellular bodies.

76 W. N. F. WOODLAND

the worm becomes much more coiled inside the capsule,! and
at some period escapes from the capsule. Judging from the
widely different sizes, both of masses lying free in the body-
cavity and of the active Amphilina, the differentiated masses
must escape from their capsules at very different stages of
growth. It must be understood that although I have only
figured a few of the stages of development of these masses,
yet all stages transitional between the youngest and the oldest
can be observed in sections through infected mesentery. It
is evidently unnecessary to give figures of the entire series.

‘l'wo other kinds of encapsulated masses must be described. In
two of the largest amorphous masses (PI. 3, fig. 2, e, represents
one of these masses) I found enclosed in each a histologically
well-developed Amphilina of large size and _ considerable
length, the body being tightly coiled on account of the restricted
space. Many of the tissues of these encapsulated Amphilina
were well differentiated and especially the reproductive system,
the uterus being of large size and, to my great surprise, full
of large oval flat larvae with their charac-
teristic hooklets and equal in size to the similarly shaped
larvae in the uterus of the 280 mm. active Amphilina described
in Part I. The larvae, however, were very degenerate (in
most cases largely disintegrated) and were only recognizable
as larvae on account of their location, general shape, and the
characteristic hooklets. The only hypothesis which I can sug-
gest to account for these facts is that for some reason the
amorphous mass in this instance had been unable to escape
from its capsule, and beg compelled to undergo its develop-
ment inside the capsule, this development was both incomplete
(e.g. the proboscis was not developed) and one-sided, the
reproductive organs developing at the expense of other organs.
The larvae being formed and unable to escape, naturally
degenerated. Certain appearances suggest to me that in
places the substance of the worm was being invaded by histo-
lytic tissue derived from the walls of the capsule.

1 Salensky states that he once found a young A. foliacea contained
in a capsule on the peritoneum covering the liver of the Sterlet.

ON AMPHILINA PARAGONOPORA iff

In one other case—an isolated instance—I found a small
capsule (roughly 500 microns long and 384 microns in mean
transverse diameter) also to contain a dozen or so degenerate
flat larvae and some disintegrated matter, but in this case
no worm was enclosed in the capsule nor could ever have been
ina capsule of this small size (PI. 5, fig. 47). The only explana-
tion of this remarkable situation of the degenerate larvae
is that the larvae had been ejected from an Amphilina into
the body-cavity of the fish (and we have seen that larvae
are sometimes extruded from the uterine pore long before the
worm escapes from the body-cavity) and that the larvae having
come into contact with the mesentery, a histolytic capsule had
been formed by the mesentery tissue to isolate and destroy
them (the amorphous masses are themselves surrounded by
capsules formed from mesentery tissue, in much the same way
that Linguatulid larvae become encapsuled). It will be
noticed from Pl. 5, fig. 47 that the walls of the histolytic
capsule are very different in construction from the walls of the
capsules enclosing normal amorphous bodies.

It is thus of interest to note that larvae, if not liberated soon
enough, can become degenerate (1) in an active Amphilina
(as e.g. in the 280 mm. Amphilina described above), (2) in an
Amphilina permanently imprisoned in its capsule, and (3) when
liberated into but unable to escape from the body-cavity of
the host (fish). The only fact which it is difficult to understand
is why, in the 280 mm. Amphilina, only a small proportion of
the eggs had developed as far as the full-grown larva stage,
the rest becoming degenerate while still inside their shells.
But perhaps this was an anomalous, as it certainly was, in
my observations, an isolated, occurrence : in no other instance
have I observed the uterus to contain anything but normally
developing or developed larvae. I may emphasize that the
flat oval type of larva seen in the 280mm. Amphilina also
occurred in the capsule-imprisoned Amphilina and in the
histolytic capsule, and possibly this is the fully grown stage
of the larva, the type of larva figured by Selensky, e. g., being
not fully grown.

78 W. N. F. WOODLAND

Notes oN TECHNIQUE.

I fixed specimens of Amphilina paragonopora either
in Zenker’s fluid, Mann’s fluid, aceto-bichromate, or simply
in 6 per cent. formalin. Some of the specimens were first
flattened between glass slides, and these, when properly
stained, showed the reproductive organs well. Unflattened
specimens were either embedded for section-cutting (the
specimens being kept straight by being placed between small
squares of wire gauze while in the embedding bath) or simply
preserved in formalin and glycerime. Whole specimens and
sections were usually staimed either with Delafield’s haema-
toxylin (diluted with ten times its bulk of water, and immer-
sion over-night) and in some cases followed by eosin or other
plasma stain, or with borax carmine. Both staims gave good
results.

In conclusion I wish to express my indebtedness to Dr. H. A.
Baylis, who very kindly consulted for me, while I was in India,
several original papers and checked my references to previous
work, and to Dr. §. W. Kemp, who kindly sent to me some
volumes from the Indian Museum library at Calcutta while
I was in Allahabad. I am also indebted to Colonel G. E. F.
Stammers for some assistance in checking references,:and to
Messrs. B. K. Das, M.Se., and §. K. Datta, M.Se., for the
considerable assistance they have kindly given to me in obtain-
ing large numbers of Macrones and finding Amphilina.

SuMMARY OF PRINcIPAL ConcLUsIONS IN Parts I anp II.

1. Amphilina paragonopora is parasitic in the body-
cavity of the Siluroid fishes, Macrones aor and M. seen-
ghala, from the Ganges and Jumna, India. It is linear in
shape and varies in length from about 10mm. to 280 mm.,
but full-grown larvae are not formed in the worm until it is
at least 100 mm. in length. The ‘ anterior’ end of the parasite
is rounded and carries a small proboscis, which is a boring
organ and not a sucker. The uterus opens anteriorly to the
left and at the base of the proboscis. The ‘ posterior ’ end of
the body is marked by a semicircular inlet or bay, and on

ON AMPHILINA PARAGONOPORA 79

a median papilla in the bay open the ductus ejaculatorius
(at the extremity of the papilla in the middle line), vagina
(at the base of the papilla on its dorsal side, the opening thus
being practically terminal) and terminal excretory duct (imme-
diately to the left of the opening of the ductus and sometimes
confluent with it). The parasites when mature (the uterus
being filled with larvae in various stages of development)
apparently escape from the fish by boring through the body-
wall at the base of one of the pectoral fins. In addition to
the ‘ active’ parasite, an inactive stage in its life-cycle is to be
found in the form of irregularly shaped masses usually attached
to the mesentery.

2. The general plan of the reproductive system is similar
to that in Amphilina foliacea, but there are some note-
worthy differences, already summarized at the end of Part I.

3. The small boring proboscis (covered with a serrated
cuticle) is connected with and manipulated by a huge boring
muscle (formerly miscalled ‘retractor’ by some authors,
and by others interpreted as a bundle of gland ducts—the
* problematic ’ cells of Salensky being the glands) which is very
thick anteriorly and extends posteriorly, though in an
attenuated form, to the region of the ovary. The fibres of this
boring muscle end in the giant cells which Salensky called

‘problematic’ and which I have renamed ‘ anchor ’-cells.
The function of the boring muscle is (1) to give a semi-rotary
movement to the proboscis (its fibres being twisted anteriorly),
(2) to act as a stout support for the anterior end of the body
when engaged in boring, and (3) to drag the hinder portion of the
body through the perforation made by the proboscis—and these
three functions account for the enormous size of the muscle.
A true and distinct retractor muscle lies externally to the boring
muscle.

4. The excretory system consists of two lateral main channels
which unite posteriorly and form a short straight terminal
excretory duct which opens in the median line posteriorly,
and a series of smaller loops and branches given off from these
two lateral main channels, which appear to form, typically,
three series of ‘ rings * when the body is viewed end-on (PI. 8,
fig. 6). Flame-cells are absent.

5. The central nervous system is, in the main, that described
by Lang for Amphilina foliacea.

6. Some stages in the development of the larva are described,
and the larvae, when liberated from the uterus, appear to be
similar to those of Amphilina foliacea. Fully grown

80 W. N. F. WOODLAND

larvae found in a 280 mm. specimen of A. paragonopora
are oval in shape and flattened, and possibly do not possess
the large gland-cells, but these may be degenerate forms.

7. The genus Amphilina is re-defined and the more con-
spicuous specific differences between the five species of Amphi-
lina are stated.

8. A brief account of the irregular-form stage in the life-
history of A. paragonopora is given, from which it
appears that the amorphous masses found on the mesentery
of the fish and which give rise to the active Amphilina, arise
from small spherical multicellular masses (about 30 microns
in diameter) enclosed in capsules formed by the mesentery
tissue. All transitions can be observed from these smallest
masses up to the stages in Pls. 3, 5, figs. 2 and 45, and thence to
the active Amphilina.

Occasionally the masses develop into sexual Amphilina inside
the capsules which enclose the masses, smce Amphilina con-
taining full-grown larvae are occasionally met with inside large
capsules.

Oceasionally also larvae liberated into the body-cavity
become secondarily encapsulated by the mesentery tissue and
are apparently disintegrated.

LITERATURE REFERENCES.

1. Cohn, Ludwig.—‘‘ Zur Anatomie der Amphilina foliacea (Rud.)”’,
‘ Zeit. f. wiss. Zool.’, Bd. Ixxvi, 1904, p. 367.

2. Hein, W.—“‘ Beitrige zur Kenntnis von Amphilina foliacea”’, ibid.,
p. 400. ‘

3. Janicki, C. v—‘‘ Uber den Bau von Amphilina liguloidea, Diesing”’,
ibid., Bd. Ixxxix, 1908, p. 568.

4, Lang, A.—‘‘ Untersuchungen zur vergleichenden Anatomie und
Histologie des Nervensystems der Platyhelminthes ”’, ‘ Mittheil.
a. d. zool. Stat. Neapel’, Bd. 11, 1881, p. 394.

5. Monticelli, Fr. Sav—‘‘ Appunti sui Cestodaria”’, ‘ Atti d. R. Acead.
d. Sci. Fis. e Mat.’, vol. v, ser. 2 (6), Napoli, 1893, p. 1.

6. Pintner, Th-—‘‘ Uber Amphilina”’, ‘ Verh. Ges. deutsch. Naturf. u.
Aerzte, Leipzig ’, Bd. Ixxvi, 2 (1), 1906, p. 196.

7. Salensky, W.—‘‘ Ueber den Bau und die Entwickelungsgeschichte der
Amphilina (Monostomum foliaceum, Rud.) ”’, * Zeit. £. wiss. Zool.’,
Bd. xxiv, 1874, p. 291.

8. Southwell, T.—‘‘ Notes from the Bengal Fisheries Laboratory, Indian
Museum. No. 2. On some Indian Parasites of Fish, &c.’’, ‘ Records
Indian Museum ’, vol. xi (iv), 1915, p. 326.

Wagener, G. R.—‘*‘ Enthelminthica, No. V. Ueber Amphilina foliacea,
mihi (M, foliaceum, Rud.), Gyrocotyle, Diesing, und Amphiptyches,
Gr. W.’’, ‘ Archiv f. Naturg.’, Bd. xxiv, 1858, p. 244,

go

ON AMPHILINA PARAGONOPORA 81

EXPLANATION OF PLATES 3, 4, AND 5.

N.B.—The magnifications given for all figures are those at which the
figures were drawn. The extent to which these magnifications have been
reduced in printing can be estimated by comparison of the printed 5 cm.
scale with an actual 5cm. The majority of figures were drawn under the
camera lucida.

PLATE 3.

Reference Letters in Figures 1-7.

ANcC, giant anchor-cells ; BM, boring muscle ; CALC, calcareous bodies ;
DEC, dorsal external excretory channel; DEJ, opening of ductus ejacula-
torius; DNB, dorsal nerve branch; prtc, dorsal transverse excretory
channel; LEC, lateral main excretory channel; LEL, lateral excretory
loop; Lut, lateral longitudinal nerve trunk; Pp, perforation at base of
left pectoral fin ; RTM, retractor muscle-fibres of proboscis ; TED, terminal
excretory duct; TES, testes; U1, U 2, U 3, first, second, and third limbs
of uterus; VEC, ventral external excretory channel; vit, vitellaria ;
vnB, ventral nerve branch ; vop, opening of vagina; vic, ventral trans-
verse excretory channel.

Fig. 1.—a, b, c, d, e, outlines of young specimens of Amphilina
paragonopora (drawn natural size); f, g, , three specimens of A.
paragonopora showing transverse constrictions of body when removed
from the host; j, an A. paragonopora which measured 280 mm.
when alive and uncontracted (contracted to 170mm. when preserved).
The proboscis is evaginated ; k, the anterior end of the 280 mm, specimen
showing the evaginated proboscis as seen under the binocular; J, the
posterior end of the same specimen showing the opening of the vagina
(dorsal) at the base of the papilla and the opening of the ductus at the
extremity.

Fig. 2 (drawn natural size) illustrates some of the immobile, mostly
amorphous, masses which represent a phase in the life-history of the
species ; 2, a represents the young spherical and ovoid masses; in 2, m,
one end of a large mass has assumed the form of the anterior extremity
of the active worm, and in 2, », the posterior extremity of the active
worm is apparent.

Fig. 3 (drawn natural size).—Anterior portion of a large A. para-
gonopora which has escaped from the body-cavity of the fish through
the perforation at the base of the left pectoral fin.

Fig. 4 (xcir. 63).—Transverse section through A. paragonopora
about midway in the length of the body.

Fig. 5 (x cir. 63).—Similar transverse section at about the level of the
hind end of the anterior fifth of the body.

NO. 265 G

82 W. N. F. WOODLAND

Fig. 6 (x cir. 7).—Diagrams representing the general arrangement of
the excretory channels in the anterior third of the worm and the two
main channels throughout.

Fig. 7 (x cir. 950).—Portion of a very fine excretory canal.

Reference Letters in Figures 8-14.

BAY, terminal bay or inlet at hind end of body; BM, boring muscle ;
cmMo, median opening of the ductus ejaculatorius ; cr, connective tissue
investment of ducts; DEJ, ductus ejaculatorius; ExB, junction of two
lateral main excretory channels; Fc, fertilization chamber; LEC, lateral
main excretory channel; MTD, muscular tissue round ductus; 0, ovary ;
OTP, oOotype ; OVD, oviduct ; PAP, papilla with terminal opening of ductus ;
PE, penis; PRO, proboscis ; RS, receptaculum seminis ; RV, receptaculum
vitelli; sp, sperm duct; sHeL, shell gland; tTrep, terminal excretory
duct with opening to the left of ductus ; TEs, testes; U, uterus; U1, U2,
u 3, first, second, and third limbs of the uterus ; Uo, opening of uterus ;
VAG, vagina; VD, vitelline duct ; vir, vitellaria; vop, opening of vagina ;
ZD, zygote duct.

Fig. 8 (x cir. 7).—Dorsal aspect of the general reproductive apparatus
of Amphilina paragonopora, with three diagrammatic transverse
sections in the region of the ovary to show the relative dorsal and ventral
positions of the various ducts.

Fig. 9.—Diagram of the general reproductive apparatus of Amphilina
foliacea, for comparison with fig. 8.

Fig. 10 (x cir. 660).—The base of the ovary and associated ducts,

Fig. 11 ( cir. 660).—Surface view of the posterior ending of the ductus
ejaculatorius, vagina, and lateral main excretory channels.

Fig. 12 (x cir, 63).—Reconstruction from serial longitudinal sections
of the structures shown in fig. 11.

Fig. 13 (x cir. 330).—Testes opening into the vas deferens.

Fig. 14 (x cir. 330).—Vitellaria opening into the vitelline duct.

PLATE 4,

Reference Letters in Figures 15-25.

ANCC, giant anchor-cells ; BM, boring muscle ; CALC, calcareous bodies ;
LEC, lateral mainexcretory channel; 0, ovary; PRO, proboscis; RTM
retractor muscle of proboscis ; SER, serrated cuticle of proboscis ; TES,
testes ; TT, terminal thickening of proboscis; u 1, U 2, U 3, first, second,
and third limbs of uterus ; Uo, opening of uterus ; vit, vitellaria.

Fig. 15,—Sketch of proboscis everted (a) and retracted (6), drawn from
the living animal.

Fig. 16 (x cir. 78).—Everted proboscis in longitudinal section.

Fig. 16, a (x 1060).—Serrated cuticle covering proboscis.

Fig. 17 (x cir. 78).—Introverted proboscis in longitudinal section
(drawn from a whole-mount specimen),

>

= ie UO Gee Cnt
Qe) An ! GBR ins
G '

wy

N

\iS

OES Ge
2a)

ry, ota ,
A y “

—S= * @ ® a9
ap in|

HU ee t (g i]
Lip Z wh

yu

PAS

LeEL Vie

W. N. F. Woodland, del.

Vol, 67, N.S., Pl. 8

Quart. Journ. Micr. Sci.

bu

D RV.

Z

D

2.
in
rd

ay

rs

vy

a!

S

ard
Col

5.
rT
ce

‘it
cy

A}
O98
a8
en

ON AMPHILINA PARAGONOPORA 85

Fig. 18 (x cis. 150).—Retracted proboscis in transverse section.

Fig. 19 ( x cir. 310).—Portion of wall of proboscis in longitudinal section.

Fig. 20 (x cir. 2).—Diagram to illustrate the longitudinal extent of the
boring muscle of the proboscis.

Fig. 21 (x cir. 7).—Diagram to illustrate the distribution of the boring
muscle-fibres and the giant anchor-cells (Salensky’s ‘ problematic cells ’),
and the same structures shown in two transverse sections (a and 6) at the
levels indicated. :

Figs. 22, 23 (x cir. 300).—Giant anchor-cells with boring muscle
processes.

Fig. 24 (x cir. 300).—Fibres of boring muscle.

Fig. 25.—Diagram illustrating the connexion of the boring muscle-fibres
with the anchor-cells,

Reference Letters in Figures 26-39.

ANCC, giant ‘anchor’-cell; Bc, ‘ brain’ commissure; BM, boring
muscle ; CALC, calcareous body: cut, cuticle; EGSH, egg-shell; F, fila-
ment or tag on egg-shell; GBL, giant blastomere; Gc, gland-cell; HK,
hooklet ; 1cM, inner circular muscle-layer ; ILM, inner longitudinal muscle
layer ; INB, internal nerve branch; IvM, investing membrane of larva;
LLT, lateral longitudinal nerve-trunk; N¥FI, nerve-fibre; OBLM, oblique
muscle-layer ; OcM, outer circular muscle-layer ; oLM, outer longitudinal
muscle-layer; PAR, parenchyma; RTM, retractor muscle-fibres; SUBC,
subcuticula ; Uo, opening of uterus ; yo, yolk material.

Fig. 26 (x cir. 330).—Transverse section through body-wall of A.
paragonopora.

Fig. 27 (x cir. 330).—Longitudinal section through body-wall.

Fig. 28 (x cir. 980).—Early stage of growth of calcareous body.

Fig. 29 (x cir. 63).—Semi-diagrammatic figure of the anterior end of
the central nervous system.

Fig. 30 (x cir. 580).—Eggs in commencement of uterus with shells and
yolk.

Fig. 31 (x cir. 580).—Morula stage of embryo surrounded by investing
membrane.

Fig. 32 (x cir. 580).—Early larva in longitudinal section showing one
large blastomere (shell not shown).

Fig. 33 (x cir. 580).—Section through posterior end of larva showing
formation of hooklet-cells.

Fig. 34 (x cir. 580).—Early larva in section showing gland-cells and
hooklet-cells.

Fig. 35 ( x cir. 580).—Late larva seen in optical section (the gland-cells,
if present, were invisible).

Fig. 36 (x cir. 980).—Hooklet from posterior end of larva.

Fig. 37 (x cir. 580).—Three of the degenerate larvae in shells, which

G 2

84 W. N. F. WOODLAND

occupied most of the uterus of the 280 mm. specimen of A. para-
gonopora (macerated in Marcacci’s fluid).

Fig. 38 (x cir. 580).—An oval flat free larva from the same uterus
(i.e. of the 280 mm. Amphilina) seen in section.

Fig. 39 (x cir. 489).—Similar degenerate larvae in shells and free larvae
from the same large Amphilina, macerated in water only.

PLATE 5.

Figs. 40, 41, 42 (x 500).—Sections of early stages of development of
amorphous masses, from mesentery of Macrones sp.

Figs. 43, 44 (x 112).—Sections of later young stages of development
of amorphous masses.

Fig. 45 (x 35).—Section of amorphous mass elongated and slightly
coiled in capsule.

Fig. 46 (natural size).—An elongated large fully-grown form of amorphous
body lying free in hody-cavity of fish, which would soon become
transformed into an active Amphilina.

Fig. 47 (x 112).—-Young cyst developed from the mesentery and enclosing
degenerate larvae and disintegration products.

oS ISS"
IN ea
x NWS if

as

3 = 2 Z Z

<

|

|

a fell i
el i }

HNN
i) i H NW H
/ WAL ff
4 US Jo OF
Zz

Se

Le

a)

gussea
eaeeres
TREE EERE

rae)

W. N. F. Woodland, del

Quart. Journ. Micr. Sci. Vol. 67, N.S., Pl. 4

Hpi tua
(Hil

16M

|

| AU AAT
MTL ih Ht Wi a iy
hy aad
vee Nh
NCAR IAI

38
O@ UA ‘GO
pico & = 22GOG,

@ feyes
6 © 20,50 2/
a) &)
ow

“

he

Quart. Journ. Micr. Sci. Vol. 67, N.S., Pl. 5

oe a

3 be 5 om.

AA Ah

2

!
Figs.

W.N. F. Woodland, del.

ih a. ‘s +t) “e % se,
, at ry! ae
bs ’ ie” ae Sn lth P . «ee
> < ~ ALU a =
re is ] a ¥ ot a
= a 7
. .
‘
.
.
+
.
.

On the Strobilization of Aurelia.
By

E. Percival, B.Se.,

Department of Zoology, The University, Leeds.
With Plate 6 and 3 Text-figures.

1. InrRoDUCTORY AND HISTORICAL.

Iv is well known that in the production of a polydise strobila
the scyphistoma undergoes a series of constrictions as far
as the septal longitudinal muscle-strands, commencing just below
the wreath of tentacles and continuing downwards towards
the foot. There is always a portion at the pedal end which
does not undergo constriction, remaining ultimately as a polyp
after regeneration of oral disc and tentacles. The animal at
this stage is a pile of segments, the oldest and most defined
being at the oral end, while the youngest and least defined is
near the pedal end. Later, there usually arise from each
segment, or ephyra rudiment, eight radiating lobes enclosing
diverticula from the enteron, four in the interradii and four in
the perradu. At the end of each lobe develops a median out-
srowth containing an extension of the enteron, which becomes
the tentaculocyst. On each side of the tentaculocyst is
a wing-like lappet of ectoderm.

The seyphistoma, just before undergoing the external
changes, has four radiating septa perforated each by an ostium
beneath the oral disc, forming together a ‘ gastral ring-sinus ’.
As the body increases in length new ostia arise below the old
ones (Text-fig. 1), and the constriction takes place between the
sets of perforations which later take part in the formation of the
gastral cavity of the ephyra.

There appears to be unanimity of opinion between previous
workers on this portion of the change, but with regard to later
developments there is some diversity of view. Perhaps they

S6 E. PERCIVAL

are least unanimous on the subject of the origin of the manu-
brium of the non-terminal ephyra. ‘ According to Goette it
proceeds from a completely new formation after the casting
off of the preceding ephyra’ (Heric). Claus (1) figures Aurelia
as having, between two ephyra rudiments, a circular shelf
or horizontal fold of the two-layered wall of the connecting
tube. The ectodermal portion of the fold is apparently con-
tinuous all round, but the endodermal portion is interrupted in

TExtT-FiG. 1.

Interradius Ferradius

Early stage in strobilization.

each interradius. Its four radial components appear from within
as grooves, the adjacent ends of which grow towards each other
and meet between the longitudinal muscle-strand and the
ectoderm, thus forming the continuous ring-canal which Claus
calls the ‘ring-sinus of the proboscis (i.e. manubrial) dise ’.
As a result of this the septal or longitudinal muscle-strand in
the region connecting adjacent ephyrae is surrounded by
endoderm, and when the edge of the incipient manubrium
becomes split off from the exumbrella of the preceding ephyra,
it bends outwards, leaving the septal muscles to act as con-
necting strands between the two ephyrae. In fact the original

STROBILIZATION OF AURELIA 87

neck connecting one ephyra with its neighbour is now reduced
to the four longitudinal muscle-strands, each surrounded by a
strip of endoderm, while the wall of the neck has become con-
verted into the manubrium and spread out horizontally.

Heric (4), working on Chrysaora, shows the intersegmental
folds, but states that they occur only between the septa as
four bladder-like outgrowths of both layers. These, on breaking
away distally from the exumbrella of the upper ephyra,
spread outwards as four semicircular flaps the adjacent edges
of which meet and fuse in the interradil, so producing the
flat manubrium. ‘Thus, he says, the connecting strands have
endoderm on the inner side and ectoderm on the outside
(whereas in Aurelia, according to Claus as we have just seen,
ectoderm is entirely absent).

Claus and Heric show, then, that the manubrium is undere
going development at the same time as the rest of the ephyra,
and that the liming of the manubrium is of endoderm.

Claus also states that the manubrium of the polyp, which
remains after strobilization, is lined by endoderm, so that at
no time is there a stomodaeum such as 1s described by Goette.
The work of Friedemann (8) and Hein (2) on the embryological
aspect confirms the above as regards the endodermal lining
of the manubrium and the absence of stomodaeum.

Claus and Heric state that the gastral filament is derived
by the transformation of the columella, i.e. the axial portion
(internal to the ostium) of the taeniole containing the longi-
tudinal muscle-strand, which for a time connects the sub-
umbrella and exumbrella of a developing ephyra.

With regard to the septal muscle-strands Goette has described
them as hollow structures with the cavity continuous with
that of the peristomial pit. Claus (1), Friedemann (8), and
Heric (4) state that the muscles are solid structures, and
Friedemann figures the peristomial pit of a well-developed
scyphistoma as being distinct from the septal muscle.

There does not appear to be any record of an observation
on the development of the peristomial pits in a non-terminal
ephyra. Claus (loc. cit.), in one figure, indicates a peristomial

88 E. PERCIVAL

pit on one non-terminal ephyra but shows nothing further.
Heric concludes that they are new structures.

In the same way there is little information on the regenera-
tion of the oral dise of the polyp which remains after strobiliza-
tion. Heric indicates that the development of the proboscis
is similar to that of the manubrium of the ephyra, but says that
the four strands connecting the polyp with the ephyra above
pass directly to the wall of the enteron after having become
dissociated from the proboscis. According to this the polyp
would have no septal ostia, and the oral end of the longitudinal
muscle would be absent for a time. Whether this is so or not
in Chrysaora I cannot say, but, as shown below, it is not the
case in Aurelia. He goes onto say: ‘ How the oral dise (of the
polyp) is connected with the septa can be answered just as
little as the question whether the septal muscle, in its course,
is transformed by the remainder of the polyp to be replaced
by a new muscle originating from the oral disc, or whether
it is preserved and enters (secondarily) into connexion with the
oral disc ’.

A complete ephyra becomes free by the gradual extension of
the connecting strands and by their final rupture. Prior to
this event the exumbral opening has closed. Heric believes
that this closure takes place simultaneously with the breaking
of the strands. He did not find any trace of longitudinal muscle
in an ephyra immediately after detachment.

9. New INVESTIGATION.

The material used in this work was obtained from the Dove
Marine Laboratory, Cullercoats, Northumberland, and was
killed and fixed in saturated aqueous corrosive sublimate
solution with 2 per cent. glacial acetic acid. Delafield’s haema-
toxylin, and iron haematoxylin and eosin were used to stain
the serial sections. Delafield’s haematoxylin is very useful
for general observations, but the latter stains were better for
determining the limits of ectoderm and endoderm, especially
in the early stages of the manubrium.

Tt should be understood that my description of the forma-

STROBILIZATION OF AURELIA 89

tion of an ephyra does not refer to the transformation of the
original oral disc of a scyphistoma, but to the fate of the
segments.formed below the original oral disc.

Formation of Manubrium.—tThe constriction of
a polyp proceeds as far as the longitudinal muscle, the ecto-
dermal groove interradially cutting the septum horizontally.
Between the interradii the endoderm is pushed inwards, but the
edge so formed does not reach the imaginary line taken between
two adjacent longitudinal muscles. This edge is concave
internally, so that a transverse section between two ephyra
rudiments would show a roughly cross-shaped gastral cavity
with the longer diameters lying in the perradii. About this
time a groove appears in the subumbral endoderm near the inner
edge, beginning in the perradius as a very shallow groove
(Pl. 6, fig. 12a) and becoming deeper passes towards the
septum or taeniole. The fold of endoderm thus formed pushes
into the mesogloea between the subumbral endoderm and
ectoderm (PI. 6, fig. 12 b, Gr.). The groove, as it approaches the
taeniole, divides into two portions at its septal end. One of
these portions is a gutter passing round on the inside of the
septal muscle to meet its fellow from the other side, and the
other portion is a solid plug of cells (Pl. 6, fig. 12, P.) which
grows through the septum towards the interradius, where it
meets and fuses with a fellow from the other side, lying between
the septal muscle and the ectoderm. ‘The muscle is thus
surrounded, in this manubrial region, by endoderm. The
portion of the plug of endoderm cells lying in the interradius
causes a bulging of the ectoderm at that point (Pl. 6, fig. 9),
but the outer layer does not project all the way round. Four
horizontal perradial slits now appear in the connecting tube
in the same plane as the ectodermal projections (Text-fig. 2),
so that the gastral cavity comes into communication with the
outside. Finally, the slits unite across the interradial ectodermal
projections and the outer layer of an upper ephyra is cut
off from that of the next below. While this takes place, a split
occurs in the plug of endoderm cells connecting the two endo-
derm grooves (PI. 6, fig. 9, S.), which now become continuous

9g: - E. PERCIVAL

abaxially to the longitudinal muscle. The lip of what is now
the rudimentary manubrium, curls outwards involving the
subumbral wall of the endodermal groove. Perradially, where

TEXT-FIG. 2.

Ferradius

Determination of number of ephyrae and Jaying down of
form of polyp.

there is practically no groove (PI. 6, fig. 12 a) the curling causes
a considerable increase in the diameter of the manubrial
opening. This, no doubt, has some influence on the formation
of the cruciate cavity of the manubrium. As Claus and Herie
have shown, a manubrium formed this way has a lining of

STROBILIZATION OF AURELIA 91

endoderm. It is, for a time, a flattened plate having a central
opening into the enteron. This change does not involve the
appearance of a ‘ proboscis ring-sinus ’ as described and figured
by Claus, but it is possible that the solid rod of endoderm cells
and the ring-sinus may be developed under different conditions.

Peristomial Pit.—Associated with the development of
the manubrium is the formation of the peristomial pits. As
has been stated, the constriction in the interradius brings the
ectoderm very close to the longitudinal strand (they are
separated by a thin layer of mesogloea), and when the endo-
dermal plug pushes out the ectoderm a small interradial funnel
is formed in the angle made by the fold and subumbrella
(Pl. 6, fig. 9, P.P.); this is the rudiment of the peristomial
pit. The subsequent curling outwards and growth of the
manubrium and the increase in depth of the ephyra bring about
the deepening of the funnel (PI. 6, fig. 10, P.P.).

Longitudinal Muscle.—Contrary to what Claus,
Friedemann, and Heric have maintained, transverse sections
of a seyphistoma show that the septal muscles may be hollow
structures (Pl. 6, fig. 8). The cavity may not be continuous
along the whole length, but it generally appears a short distance
below the peristomial pit, and, as Friedemann has shown,
the end of the muscle and the apex of the pit are not con-
tinuous with each other. The muscle-cells are arranged with
their tails to the mesogloea and the protoplasmic parts towards
and projecting into the lumen (PI. 6, fig. 8). There are mesogloeal
ridges projecting into the cavity and on these are set muscle-
cells, and this probably serves to increase the amount of
muscular surface. Sometimes the cavity is quite wide, and
none would describe such a structure as a solid muscle. Towards
the foot the muscle-cells are usually closely packed, and this
portion of the strand can be described as solid.

In a strobila, mainly in the lower segments, sections show
that some parts of the muscles may be hollow, though the
cavities in such cases are short, extending along the depth
of an ephyra rudiment. As an ephyra becomes more developed
and the circular and radial muscles begin to function, the

92 E. PERCIVAL

longitudinal muscles begin to atrophy and are seen in longitu-
dinal sections as mesogloeal bands striated with degenerate
muscle-fibres (Pl. 6, fig. 11, R.O.M.). The bands stain more
deeply than the rest of the mesogloea when treated with
Delafield’s haematoxylin.

Gastral Filaments.—About the time when the manu-
brium of the ephyra has become curled outwards the gastral
filaments first appear. They occur in pairs, one pair per inter-
radius (PI. 6, figs. 1 and 5, G.F.). On each side of the columella,
and close to the exumbral endoderm, there grows laterally and
towards the central stomach an endodermal process (Pl. 6,
fig. 1, G.F.). Very soon the tip turns towards the oral opening
and remains pointing in that direction until the ephyra becomes
free (Pl. 6, fig. 5). Occasionally one or both filaments may be
suppressed. Thus, contrary to what has been maintained by
all writers hitherto, the longitudinal muscle takes no part in
the formation of the filaments which may be seen in a strobila
in various stages of development.

Separation of an Ephyra from the Strobila.—
In passing upwards towards the oldest ephyra the connecting
strands are seen to become slightly stretched and the covering
epithelium does not stain so deeply as that lower down. They
converge to the apex of the exumbrella when the apical opening
has become obliterated. The closure of this opening takes
place comparatively early in the history of the attached
ephyra (Text-fig. 2), and a free disc having such an opening
may be considered to have been prematurely detached. Such
may happen in the laboratory when a strobila is roughly
handled.

Shortly before an ephyra becomes separated it is seen that the
peristomial pits are very deep (PI. 6, fig. 10, P.P.), and the depth
from the subumbrella to the exumbrella is much greater than
that in a younger disc. Also the columella is stretched so that
the gastral filaments are brought some distance away from the
exumbrella (Pl. 6, fig. 5). The covering of this stretched portion
resembles in staining properties and cell-form that of the
connecting strand in the same condition. The longitudinal

STROBILIZATION OF AURELIA 93

muscle here is degenerate, and the apical portion of the ephyra
is thickened with mesogloea in which the muscle remnants
can be seen (fig. 5, .M.). As Heric suggests, the separation is
probably brought about by the violent action of the circular
and radial muscle-bands. The connecting strands break close
to the exumbrella. About the same time the stretched part of
the columella breaks, resulting in the carriage of the gastral
filaments to the subumbral side. The remnant of the stretched
portion is seen as a small papilla of cells between the bases of
the gastral filaments (PI. 6, fig. 3, R.C.).

The rupture of the columella brings about a rapid shortening
of the peristomial pit, which is now seen to be a small but
distinct funnel-shaped depression (fig. 3, P.P.) in the sub-
umbrella close to the bases of the gastral filaments. The greater
part of the wall of the pit has gone to form a portion of the
subumbral surface and the base of the manubrium. The pit,
which was close to the base of the manubrium, is now some
distance away. The gastral filaments no longer point to the
oral opening but come to be roughly in a line at right angles
to the interradius.

When an ephyra becomes free the connecting strands are left
projecting from the oral opening of the next ephyra (PI. 6,
fig. 10, C.S.) and are soon reduced in length. The flat manu-
brium of this next disc now begins to assume the tubular form,
which is not completed until separation-and the rupture of the
columella (Pl. 6, fig. 5).

Muscle remnants are to be found in a newly separated ephyra
in the thickened apex (PI. 6, fig. 4) and at the base of the manu-
brium on the inner side of the peristomial pit. A vestige of
the connecting strand and the part of the muscle between it
and the apex of the pit can be found at the base of the manu-
brium (PI. 6, fig. 3, R.O.M.).

Formation of the Proboscis of the Polyp.—
A strobila can be divided into two regions, viz. segmented,
giving rise to ephyrae, and unsegmented, producing the
polyp. Between every two adjacent constrictions le four
septal ostia arranged in a ‘ring-sinus’. Below the lowest

94 E. PERCIVAL

constriction are four ostia which usually remain to form the
gastral ring-sinus of the scyphistoma.

The proboscis of the polyp develops in a manner similar to
that of an ephyra, but there is evidence of the formation of a
proboscis ring-sinus such as Claus has described for the ephyra.
The edge of the proboscis curls outwards but the opening
is much wider than that in an ephyra. The passage of the

TEXT-FIG. 3.

Interradius Rrradius f
A.
\

Before liberation of last ephyra. Polyp complete.

connecting strands direct to the gastral wall of the enteron
and unconnected with the wall of the oral opening was seen
in only one case. Such a condition (which it will be remembered
Heric regards as normal in Chrysaora) is probably accidental,
and may be described in Aurelia as abnormal. In a ease of
this kind there are no septal ostia and so no gastral ring-sinus.
The upper portion of the enteron is as entire as in the free
ephyra. Usually the columellae persist and the resulting polyp
possesses all the features of the original seyphistoma from which
it has been formed (Text-fig. 3).

As far as has been ascertained the peristomial pit of the

STROBILIZATION OF AURELIA 95

polyp does not arise in exactly the same manner as in the
ephyra. It occurs in the same relative position but apparently
not as a pit. A specimen showing the earliest trace had a solid
strand of cells reaching from the oral dise obliquely inwards
and downwards to the longitudinal muscle (PI. 6, fig. 2, R.). At
the outer end was a slight depression which was probably the
commencement of the formation of a cavity in the strand.
Another specimen showed a complete pit the apex of which
bore the same relation to the longitudinal muscle as does that
of an ephyra (see Text-fig. 3). Probably, then, the strand of
cells becomes hollow from outside inwards. This cavity should
not be confused with that in the longitudinal muscle, nor do
the two ever communicate.

Ciliation of the Ectoderm.—Gemmill (5) has shown
that, in the ephyra, there are definite currents passing over
the ectodermal surface apparently for the purpose of carrying
small animals to the lappets, where they are pierced by stinging
threads and afterwards carried to the mouth by flexure of the
arm. He also states that the scyphistoma captures infusoria
in much the same way as the ephyra, the tentacles taking the
place of the arm lappets.

Powdered carmine suspended in sea-water will also serve
admirably to demonstrate these currents. In the seyphistoma
the carmine particles are carried upwards along the surface of the
body and become entangled in slime secreted by the ectoderm.
Between the bases of the tentacles ropes of particles and slime
may be seen carried along the disc and up to the edge of the
proboscis. The tentacles, along which the current passes to
the tip, sometimes curl over into the proboscis and the material
travels into the gastric cavity. Often a slowly revolving ball
is formed above the mouth and finally passes slowly down into
the enteron. The same applies to the ephyra, in which the
passage of the carmine ball into the gastric cavity can be easily
seen. Sometimes the ball will be slowly ejected from the
enteron immediately after being taken. Gemmill believes that
particles are taken in ‘ by a central inhalant current which is
compensatory to exhalant currents produced by ciliary action

96 E. PERCIVAL

in the floor of the mouth angles ’, adding * but this may not be
the whole explanation ’.

In both ephyra and scyphistoma the use of powdered carmine
seems to show that the lining of the proboscis or the manubrium
has cilia which may produce exhalant or inhalant currents
when necessary. The stream of particles passes down the angles
in the perradii as well as along the ridges of the interradii ;
further, expulsion of material may follow the same channels.
In only one case did I observe both currents acting simultane-
ously, and then the in-currents moved along the interradial
ridge and the ex-currents along the perradial angles. A
strobila shows a strong current from the foot upwards to the
uppermost ephyra. This is due to the fact that the surface
is composed of a considerable portion of the aboral surface of
each ephyra, on which surface the current is centrifugal to the
ends of the lappets. There did not appear to be any passage
of carmine in between any two ephyra rudiments. This
suggests that the intermediate ephyrae or ephyra rudiments
do not feed by means of these currents.

High power examination shows that the currents are caused
by flagellated ectoderm cells (PI. 6, figs. 6 and 7). The ectoderm
of the scyphistoma, except on the pedal disc, and of the
ephyra appears to consist entirely of flagellated cells. Hven
the surface of the tentaculocyst is provided with flagella.
When death takes place, either naturally or by poison, the
flagella usually disappear. Occasionally im a well-fixed specimen
they still persist and can be studied by means of sections.
In the case of the lining of the manubrium or of the proboscis
the flagella are almost always visible in sections, an interesting
point which corroborates the histological evidence as to the
endodermal nature of this lming.

Remarkable cases of Polyp Formation.—While
keeping under observation a young strobila for the purpose of
watching the transformation of the oral disc into an ephyra,
I was able to observe a number of interesting but unexpected
changes. There were five constrictions in the strobila, and
the two ephyra rudiments next below the oral dise proceeded

STROBILIZATION OF AURELIA 97

normally to give rise to ephyrae. The two rudiments below
these, however, produced eight tentacles each and no lobes.
The oral dise did not undergo any change except that it became
separated from the rest and lay on the bottom of the dish.
There was a large apical opening the presence of which suggested
that the disc had been prematurely separated. After some
days the hole closed and the tentacles were absorbed. No
lobes developed and later a pedal stalk grew out from the
apical end, the body ultimately becoming attached to the floor
of the dish. Later four perradial tentacles appeared and the
whole had the form of a normal scyphistoma. The next two
rudiments proceeded normally and produced ephyrae which
were subsequently liberated.

The rudiments possessing tentacles then became free and
moved over the surface of the dish for about a week by means
of the flagellated surface, one of them revolving, the other
progressing in a more or less straight line. Later each grew
a foot stalk from the apex and became attached to the vessel.
In the meantime one had increased the number of tentacles to
twelve.

A second young strobila also behaved in an unusual manner.
Here the two segments next to the oral disc gave rise each to
four tentacles and no lobes. They, along with the oral disc, be-
came separated in a body and remained attached to each other
while on the floor of the dish. The oral dise did not appear
to undergo any change, but the other two segments gradually
absorbed their tentacles and became converted into a long
narrow stalk at the end of which was a pedal disc. Attach-
ment with the dish was effected, the stalk became stouter, and
the whole body assumed the form of a normal scyphistoma.

The occurrence of segments which give rise to polyps leads
one to believe that there is a distinct difference between these
and the ephyra rudiments. It may be that they have been
derived from ephyra rudiments, the sequence of changes
necessary to produce an ephyra having been interrupted or
reversed, or they,may be different from the start. Again, it
may be that up to a certain point the segment may be regarded

NO. 265 A

98 E. PERCIVAL

as undifferentiated and, in the cases where polyps are produced,
they have remained so. It is interesting to note how they
agree in general behaviour with the upper part of the base of
the strobila, which produces oral dise and tentacles, while the
segment above gives rise to an ephyra.

The several ways in which a polyp may be formed are thus :

By direct development from the egg.

By the outgrowth of stolons from a polyp.

By the elaboration of the basal portion of a strobila.

By the separation of an unchanged oral disc.

By the elaboration of an intermediate segment.

I wish here to acknowledge my indebtedness to Professor
W. Garstang for the kindly criticism and help which, from time
to time, he has afforded.

SUMMARY.

The manubrium of a primarily non-terminal ephyra is formed
from the connecting tube between two ephyra rudiments.

The proboscis of the polyp remaining arises in a manner
similar to that of the manubrium of the non-terminal ephyra.

The apical opening of an ephyra normally closes before
liberation.

The connecting strands are covered with endoderm.

The longitudinal muscles of a polyp may be hollow and the
cavities are not in communication with the peristomial pits.

The formation of the peristomial pit is associated with the
development of the manubrium.

The gastral filaments are formed as paired outgrowths of the
endoderm of the columella. They do not involve the longitu-
dinal muscle in their production. They originate early in the
history of an ephyra.

There are definite currents over the ectoderm caused by
flagellated ectoderm cells.

The segments of a strobila may give rise to ephyrae or polyps
and the oral disc may separate unchanged to continue its
existence as a polyp.

——

End. Seepacesp REE

= tds

a
Q
a.

‘ingi*

= bE: iy Puta

|

mn
S
sate

F Percival de/ j

Quart. Journ. Micr. Sci., Vol.67, N.S., Pi. 6.

STROBILIZATION OF AURELIA 99

List oF REFERENCES.

1, Claus——‘ Entw. d. Seyphistoma v. Cotylorhiza, Aurelia u. Chrysaora”’,
* Arb. Zool. Inst. Wien ’, vol. 10, 1892.

2. Hein—“‘ Unters. iiber die Entwicklung v. Aurelia aurita’’, ‘ Zeit.
f. wiss. Zool.’, vol. 67, 1900.

3. Friedemann.—‘‘ Unters. iiber die postembryonale Entwicklung v.
Aurelia aurita ’’, ibid., vol. 71, 1902.

4, Heric_—* Zur Kenntnis der polydisken Strobilation vy. Chrysaora’’,
* Arb. Zool. Inst. Wien ’, vol. 17, 1909.

5. Gemmill—* Notes on Food-capture and Ciliation in the Ephyrae of
Aurelia ”’, “ Proc. Roy. Phys. Soc. Edin.’, vol. 20, 1921.

EXPLANATION OF PLATE 6.

Fig. 1—\ Longitudinal section of polydisc strobila showing relation
between gastral filaments and columella.

Fig. 2——Tangential section of base of strobila showing portion of solid
rod of ectoderm cells passing from oral surface to longitudinal muscle.

Fig. 3.—Vertical section through interradius of newly liberated ephyra
showing base of manubrium, peristomial pit, common base of gastral
filaments, and remnant of longitudinal muscle.

Fig. 4.—Vertical section through apex of newly separated ephyra showing
mesogloeal thickening and muscle remnants.

Fig. 5—Median vertical section through uppermost ephyra of polydisc
strobila.

Fig. 6—Flagellated cell from ectoderm of scyphistoma.

Fig. 7— Portion of living tentacle (a) extended, (b) contracted, showing
flagella and cells.

Fig. 8—Transverse section through connecting strand about half-way
along a polydisc strobila, showing cavity of muscle, and muscle fibres.

Fig. 9—Longitudinal section (somewhat oblique) cutting interradius
of two lowermost ephyra rudiments showing origin of manubrium.

Fig. 10— Longitudinal section (oblique) through interradius of strobila
showing relation between peristomial pit and longitudinal muscle of second
ephyra down.

Fig. 11. Longitudinal section through centre of third ephyra down
showing closure of apical opening.

Fig. 12—Three longitudinal section of lower ephyra rudiment in fig. 9
showing form of endodermal groove which provides lining of manubrium
(a) almost perradial, (b) about adradial, (c) near interradius and showing
commencement of abaxial rod of cells.

H 2

100 E. PERCIVAL

ABBREVIATIONS.

C.S., connecting strand. Cav., cavity of longitudinal muscle. Col.,
columella. Hct.,ectoderm. Hnd.,endoderm. Hz.Ect., exumbral ectoderm.
Ex.End.,exumbralendoderm. G., gastralcavity. G.F., gastral filaments.
G.P., gastral pouch. Gr., groove of endoderm to form manubrium.
L.M., longitudinal muscle. Mn., manubrium. P., plug of endoderm
cells to form interradial portion of manubrium. P.P., peristomial pit.
R., rod of ectoderm cells to form peristomial pit of polyp. #&.C., remnant
of exumbral end of columella after rupture. #.L.M., degenerate longitu-

dinal muscle. S., split in ectoderm and in endodermal plug. S.0O., Septal
ostium.

Histology of the Soft Parts of Astraeid Corals.
By

G. Matthai, M.A.,
_ Mackinnon Student of the Royal Society during the years 1914-17.

With Plates 7 and 8.

Tue following account of the histology of the soft parts of
the Astraeidae is supplementary to the description given
in the Introduction to my paper on ‘A Revision of the Recent
Colonial Astraeidae possessing Distinct Corallites ’ (25, pp. 1-32).
It is based on the study of a large number of polyps belonging
to various Astraeid species of the Indo-Pacific and Atlantic
regions, particularly of Favia favus (Forsk.), Favia
hululensis (Gard.), Coeloria daedalea (Ell. and Sol.),
Leptoria gracilis (Dana), Eusmilia aspera (Dana).
During a short stay at the Carnegie Marine Biological Station
at Tortugas (July 16-August 2, 1915), living colonies of all
coral species of that locality were kept under observation,
but at that time larvae were extruded only from colonies of
Favia fragum (Hsp.). These were fixed at different intervals
during the free-swimming stage—from eight hours to about
ten days—in Flemming’s fluid, corrosive acetic solution, and
Bouin’s fluid, and were subsequently sectioned serially to
thicknesses of 4¢, 64, 8, and 10 in order to compare their
histology with that of adult colonies. No larvae of any species
were obtained during a subsequent visit to the Bermudas
(Aug. 20—Sept. 14). Solid embryos lying in the coelen-
teri¢ cavities of polyps from a colony of Favia fragum,
which Dr. Vaughan forwarded to Professor Gardiner from
South Bight, Bahamas, and which have been sectioned, were
also studied.

The colonies from the Indo-Pacific region were fixed im
saturated solution of corrosive sublimate and in formic aldehyde
poured into sea-water; those from the Atlantic region were

102 GEORGE MATTHAI

narcotized in a partly expanded condition in weak solutions
of magnesium sulphate before fixation in formalin. They were
then brought up to 75 per cent. alcohol for preservation. The
decalcification was done in 2-3 per cent. solutions of nitric
acid in 75 per cent. alcohol, some of the colonies with hard
coralla taking as long as three months to decalcify, but, as
a rule, the histological condition of the soft parts has not been
affected to any extent by the process.

Various stainng methods were employed, chiefly Haiden-
hain’s iron-haematoxylin followed by eosin, aniline blue, and
orange G (Mallory), safranin O, borax-carmine followed by
picro-nigrosin and picric aniline blue. Well-preserved polyps
were subjected to teasing before and after maceration in the
Hertwig’s osmic-acetic solution; while at the American
Biological stations a similar investigation of fresh coral tissue
could not be made to my satisfaction owing to pressure of
other work, though it was found that the method of teasing
was not so suited as that of serial sectioning to reveal the true
histological relationships of lowly differentiated tissues like
those of the Madreporaria.

I have to thank the Governing Body of Emmanuel College
for financial aid in connexion with this research.

The soft parts of the Madreporaria consist of an outer
and inner protoplasmic sheet and an intermediate supporting
lamina. The two former are described in this paper under
the widely accepted terms Ectoderm and Endoderm,
which had originally been employed by Allman in 1853 to
denote the outer and inner layers of the Tubulariadae
(1, p. 868). But I have refrained from applying any of the
suggested names to the middle lamina since (as will be shown
in the course of this study), from its nature and formation,
this lamina does not appear to be essentially different from
the Mesoderm of the Triploblastica. It may also be
stated at once that these laminae in Astraeid corals are not
discrete layers, as has been the prevalent view, but are of the
nature of three strata in a continuous multinucleated sheet.

HISTOLOGY OF ASTRAEIDS 103

ECTODERM.

The ectoderm forms the entire outer lining of the soft parts,
i.e. of the oral-dises, tentacles, column-walls of polyps (where
it is termed the calicoblastic layer), and edge-zones and
coenosare (the ectoderm of their outer wall is a continuation
of the oral-dise ectoderm, while that of their inner wall is
a continuation of the calicoblastic layer).

In the oral-dise and outer wall of the edge-zone (figs. 1 and 2)
the ectoderm has an even free surface which, in sections,
is often seen to be covered with mucous secretion and is of
more or less uniform thickness. It has a thin free border con-
taining fine vertical striae and is provided with short. cilia.
Elongated nuclei are aggregated somewhat along the middle
of the ectoderm: smaller round ones which are less numerous
lie more or less scattered in its inner half. Both mucous and
granular vacuoles are present, the former being more abundant
than the latter, and a varying number of nematocysts also
occur in it. Outwardly diverging tracts are sometimes visible
in the protoplasm between the vacuoles. Fibrils are continuous
between the ectoderm and middle lamina. At the base of the
ectoderm is a finely granular stratum in which a faint network
can be discerned which is probably the result of intercrossing
of the basal processes of the nuclei of the ectoderm and the
fibrils which pass into it from the middle lamina. This stratum
has been usually regarded as nervous, but, on renewed examina-
tion, no such nerve elements as those described by the Hertwigs
in Actinians could be found in it. In the tentacles (fig. 3 and
25, fig. 44) the ectoderm is greatly thickened at intervals to
form the batteries in which nuclei are considerably increased
in number, the elongated ones lying in the upper half of the
batteries, smaller oval and round nuclei in the lower half.
The granular stratum at the base of the tentacular ectoderm
is thicker than at the base of the oral-dise ectoderm.

The calicoblastic ectoderm (figs. 10, 13-15) is very thin except
where the column-wall processes are being formed, and has
a somewhat irregular outline and granular protoplasm, the

104 GEORGE MATTHAI

granular appearance being more pronounced than elsewhere
in the ectoderm. Nuclei are few and are arranged in a single
row; they are large, oval or round, rarely elongated, finely
granular, and he tangentially in a single row at comparatively
wide intervals, except near the attachments of mesenteries
to the corallum, where the calicoblastic layer is usually thickened
and nuclei tend to become irregularly distributed without any
crowding. Most of the nuclei contain a brilliantly stamed spot
(perhaps the nucleolus) which is conspicuous under an oil-
immersion lens. The calicoblastic layer persists to the base
of polyps, though considerably attenuated.

From this account it will be seen that the ectoderm
consists of two histologically different regions, viz. the part
covering the exposed surface of colonies (i.e. the oral-dise,
edge-zone, and coenosare ectoderm) and the calicoblastic
layer, the differentiation being doubtless in accordance with
differences in their functions. In the free-swimming larva of
Favia fragum, the ectoderm of the column-wall (fig. 4) is
histologically similar to that of the oral-dise of the polyp
(the calicoblastic layer bemg non-existent at this stage),
and is, in this respect, comparable to the column-wall of
Actinians ; at the aboral pole the ectoderm is thickened,
presumably for future attachment. Tentacles had not appeared
in any of the larvae examined.

ENDODERM.

The endoderm forms the lining of the coelenteri¢e cavities,
i.e. forms the inner lining of the column-wall, oral-dise, edge-
zone, and tentacles, the outer lining ot the stomodaeum and the
double sheet of the mesenteries. It 1s usually vacuolated and is
apparently without cilia, while in Actinians the Hertwigs
found a single long cilium or flagellum on each endoderm cell ;
the vacuoles are often elongated, their longer axes are more or
less perpendicular to the width of the mesenteries, the vacuoles
bemg somewhat broader distally. Nuclei are oval or round,
smaller and less numerous than those of the oral-dise ectoderm.
The endoderm varies in thickness in different parts of the same

HISTOLOGY OF ASTRAEIDS 105

polyp and in different species. In the tentacles it is consider-
ably swollen in some species, almost filling their lumina ; in the
column-wall it is thin in the stomodaeal region, but becomes
highly vacuolated and reticular below this region, where it con-
taims comparatively few nuclei which are arranged in a row near
the free surface ; in the stomodaeal wall the endoderm remains
uniformly thin. In the mesenteries the endoderm is usually
swollen along the pleatal region and behind the filaments.
A narrow constriction is present behind the filament, which is
deeper in principal than in subsidiary mesenteries (figs. 7 and 8).
Numerous round bodies, usually regarded as symbiotie algae,
are present in the oral-disc, edge-zone, and tentacular endo-
derm, i.e. in the exposed regions of colonies ; in some polyps
they are so massed as to fill parts of the endoderm. In
the mesenteries these algal bodies occur in varying numbers,
chiefly in the exocoelic¢ side, but are scarce in the column-wall.
Fibrils are’ continuous between the endoderm and middle
lamina, somewhat as between the latter and ectoderm. In all
the larval stages examined, the endoderm has attained histo-
logical similarity with that of the polyp, but organic débris
containing scattered nuclei are still seen in the coelenteric
cavity (fig. 4) and are perhaps remnants of the contents of
the earlier solid planula stage (87, Pl. u, fig. 4).

The distmguishing characters of the endoderm are its vacuo-
lated condition, presence of algal bodies, the numerical
inferiority and somewhat scattered condition of its nuclei.
The endoderm, unlike the ectoderm, has a homogeneous
appearance, except for its relative swelling in different parts
and the varying number of algal bodies present.

INNER LINING OF STOMODAEUM (figs. 5 and 17).

In the larva, the stomodaeum is said to be formed by
invagination of one of its extremities, while it is found that
in colony-formation new stomodaea may be formed by invagina-
tion or by the union of the broader mesenteries in diverti-
cula (80). In the two former cases, the inner lining of the
stomodaeum is a continuation of the surface ectoderm, while

106 GEORGE MATTHAI

in the latter case it is an endodermal formation ; but, whether
ectodermal or endodermal in origin, the inner lining possesses
histological identity. It is raised into ridges over the attach-
ments of mesenteries (fig. 5); these ridges vary in thickness
and breadth in different species and frequently possess median
grooves as they approach the enterostome. The free border
of the mner lining shows the vertical striation better than in
the surface ectoderm. It is conspicuously ciliated, the cia
being longer in the median grooves of the ridges and in the
sulci between the ridges; these cilia would function in the
ingress and egress of currents of water. Below the striated
border is a somewhat finely granular non-nucleated stratum,
beneath which is a much thicker region containing massed
nuclei of varying length and size (mostly tapering at both ends).
Between the nucleated region and the middle lamina is a
fibrillar region which is thicker than all other regions and
comprises the lower half or two-thirds of the ridge ; it contains
a few small, round or oval nuclei, and the fibrils are continuous
with the middle lamina. At the base of the imner lining is
the fine granular stratum which shades off into the middle
lamina. The striated border and the non-nucleated stratum
underlying it are of uniform thickness over the entire stomo-
daeum, while the nucleated and fibrillar regions are consider-
ably thinner in the intermesenterial parts. Vacuoles are
usually present in the stomodaeal ridges and are fewer in the
intermesenterial region ; many of them contain granules which
- vary in their density, beg quite fine in some vacuoles. The
vacuoles extend to the surface of the ridges and, in sections of
some polyps, the granules are seen to have actually passed
into the lumina of the stomodaea. Nematocysts are sometimes
present in the inner lining of the stomodaeum. At the entero-
stome the inner lining of the stomodaeum becomes continuous
with its outer endodermal lining.
Since numerous stomodaea are present in a Madreporarian
1 Finely powdered carmine, when put into sea-water containing a live

,colony of Manicina areolata was passed into the stomedaea and
subsequently ejected.

HISTOLOGY OF ASTRAEIDS 107

colony, the ectoderm of the free surface is continuous with the
stomodaeal lining at frequent intervals. The inner lining of the
stomodaeum of the larva is not raised into ridges at the mesen-
terial attachments, hence both grooves and sulci are absent ;
its nuclei are not so crowded together nor so slender as in the
polyp, and are arranged along the middle of the layer whose
protoplasm is conspicuously granular and opaque above the
nuclei, while below them it is vacuolated and translucent ;
the condition of the stomodaeal lining of the larva is on the
whole intermediate between that of the oral-dise ectoderm
and the stomodaeal ridges of the polyp.

MESENTERIAL FILAMENTS (figs. 7, 8, and 18).

The epithelium of mesenterial filaments has essentially
the same structure as the inner lining of the stomodaeum,
the median lobe being similar to the stomodaeal ridge and the
ventro-lateral tracts to the parts between the ridges. The
median grooves of most of the stomodaeal ridges are continued
to some distance along the middle of the straight regions of
their corresponding mesenterial filaments, the cilia in the
grooves being longer than those over the rest of the filaments.
A transverse section through a principal filament just below
the stomodaeum, as shown in fig. 6, bears striking resemblance
to Ashworth’s figure of a transverse section through a dorsal
mesenterial filament of Xenia Hicksoni (8, fig. 19).
Granular vacuoles are frequently present in the straight region
of the filaments. Nematocysts are few in the straight region,
numerous in the convoluted parts, where they often become
massed together. Each of the ventro-lateral tracts of a fila-
ment is organically continuous with the mesenterial endoderm
on its side. Filament-epithelium is present on subsidiary
mesenteries (except on the very narrow ones) along the greater
part of their length, but is rudimentary in their upper half
or one-third, where it contains a few aggregated nuclei or
is sometimes entirely absent. Subsidiary filaments are smaller
in transverse section than principal filaments. Mesenterial
filaments of the free-swimming larva (whether of principal or

108 GEORGE MATTHAI

subsidiary mesenteries) are similar to the inner lining of its
stomodaeum, i.e. nuclei in them are not so closely aggregated
nor so slender as in filaments of polyps.

It is obvious that in a subsidiary mesentery of a polyp
the filament is formed by modification of the endoderm of
the mesentery along its free margin, attaining histological
similarity with the inner lining of the stomodaeum and with
filaments of principal mesenteries. Stages in this modification
are seen in subsidiary mesenteries of varying width. In larvae
of Favia fragum, also, filament-epithelium is present along
the margins of some subsidiary mesenteries which is undoubtedly
formed by modification of the marginal endoderm of those
mesenteries. I have previously described the presence of
filament-epithelium on the mesenteries of an extra-tentacular
bud of Favia hululensis (Gard.) (27).

In some species, Favia fragum, Mercilinaampliata,
Hyderophora maldivensis, Isophyllia dipsacea,
there are regions in the convolutions of mesenteries in which
the filament-epithelium is considerably-vacuolated and swollen,
in which nematocysts 1 or 1 are closely arranged. In dumb-
bell-shaped transverse sections of these, the filament-epithelium
at each end resembles the intervening endoderm. In other
words, histologically identical epithelia are found in stomodaea
and mesenteries, whether the inner linings of the former and the
filaments of the latter are ectodermal or endodermal in origin.
But it is to be noted that algae are absent from the imner
lining of the stomodaeum and mesenterial filaments.

Tue SupPorRTING MIDDLE LAMINA.

The middle lamina is found everywhere between the ectoderm
and endoderm and forms the median core of mesenteries.
Though Bourne could find no trace of structure in the middle
lamina of Fungia, he remarked that the use of proper
reagents might possibly have disclosed a fibrillar structure
(5, p. 310). Asa result of making careful microscopical prepara-
tions this lamina is now found to consist of (1) a homogeneous
matrix or clear cementing substance containing (2) fine fibres

HISTOLOGY OF ASTRAEIDS 109

and (3) nuclei (figs. 1, 11, and 12). The fibres are of two kinds :
those which have a wavy appearance and-run in various
directions but have chiefly a longitudinal and transverse
disposition—such fibres appear to be unbranched and are
closely cemented together to form the substance of the lamina ;
branching fibres which form a loose plexus in the lamina—these
are brought to view by carefully staining sections of not more
than 6» thickness. The apparently homogeneous appearance
of the middle lamina is due to the thinness and close cementing
of the fibres. Nuclei are comparatively few and lie scattered in
the lamina; they become evident in tangential or oblique
sections through the thicker regions. Each nucleus is oval in
shape, containing a conspicuous spot (the nucleolus), and lies
in thin finely granular protoplasm from which irregular pro-
cesses usually radiate into the substance of the lamina; not
infrequently a narrow clear space can be detected around the
protoplasm. In several West Indian species of coral Duerden
noted the presence of ‘ migrant connective-tissue cells, such as
occur in the larger Actinians ’ (9, p. 22).

The middle lamina is thickened in the mesenteries and is
raised on one side into longitudinal pleats whose breadth
and thickness vary in the different species (fig. 9). In the
stomodaeal region the pleats extend over part of the width of
mesenteries to a varying distance from their column-wall
attachments, while below the stomodaeum they cover almost
the entire width of mesenteries. The lamina is usually con-
siderably thickened where mesenteries join stomodaea and
column-walls. While at the stomodaeal attachment the
thickening is restricted to the ridge, at the column-wall inser-
tion the thickening usually spreads a short distance into the
adjacent middle lamina, these lateral thickenings appearing,
in transverse section, like two arms. The middle lamina is
thickened to a less extent in the tentacles and oral-dise ; in
the former, outer longitudinal pleats are present which are less
conspicuous than those of mesenteries. Processes arise from
the middle lamina over the entire extent of the column-wall
to attach the soft parts to the corallum, and are more numerous

110 GEORGE MATTHAI

and larger at the insertions of mesenteries. These processes are
composed of fibres and cementing substance, and are the
homologues, in the column-wall, of the pleats in mesenteries
and tentacles.

The superficial longitudinal fibres in the pleats of mesenteries
and tentacles are specially thickened. These specialized fibres,
which vary in thickness, appear to be composed of fibrils,
but had usually been supposed to be similar to the muscular
elements described by the Hertwigs in Actinians, Faurot in
1895 being the first to doubt their muscular nature. In teased
preparations and in sections, nuclei are not found in these
fibres nor is there any morphological or physiological evidence
for regarding them as muscular. Specialized fibres are present
on the exocoelic side of mesenteries (but not so thick nor so
close together as on the entocoelic side), although pleats are
absent from that side or only a few feebly developed ones
are present near the stomodaeal attachment. The striae in
the processes of attachment appear to be specialized fibres which
have a radial disposition.

In preparations of mesenteries with the endodermal lining
scraped off, the specialized longitudinal fibres of the middle
lamina could be seen running along its entire length. When
parts of the living tissue of Isophyllia were isolated from
expanded colonies, teased in sea-water, and stained in methy-
lene blue, the middle lamina took a purple or violet tinge
and its fibrous texture became quite apparent. The fibrous
condition could also be unmistakably seen in properly pre-
served tissue which had been teased after maceration and
removal of the protoplasmic sheets, as well as in such tissue
cut to 4v and 6, thicknesses. The branching fibres were
best seen by staining in safranin O and picro-nigrosin. The
specialized fibres of the middle lamina, whether in the pleats
or in the processes of attachment, were similarly coloured with
different stains, e.g. dark in iron haematoxylin, purple in
aniline blue and orange G, slaty blue in borax-carmine followed
by picro-nigrosin, such results suggesting identity of texture
of both sets of fibres. The absence of muscular fibres in the

HISTOLOGY OF ASTRAEIDS ile i

soft parts of the Astraeidae would also explain the absence
of a nervous system—central or peripheral.

In the larval stage the middle lamina is present everywhere
and is fibrous, though thinner than in the polyp. Pleats are
hardly recognizable in mesenteries, but specialized fibres are
present. H. V. Wilson found that in the development of
Manicina areolata the middle lamina appears in the
solid planula stage (87, fig. 5). This observation is corroborated
by my study of the solid embryonic stages occurring in the
coelenteric cavities of polyps of Favia fragum.

The middle lamina of Actinians as seen from a study of serial
sections of young polyps of Sagartia bellis, Metridium
senilis, and Corynactis viridis (which in alcohol
measured2mm. x 1 mm.,12mm. x 3mm., and 4mm. x 3-5mm.),
obtained from Plymouth, is in essential points similar to that
of Astraeid corals. In the former two species, the middle
lamina has a swollen somewhat loosely spongy core containing
many nuclei which is bounded by closely arranged unbranched
fibres; the plexus of the spongy core consists of branching
fibres which are more abundant than in coral polyps (fig. 16). In
Corynactis viridis the meshwork is closer, approaching
the condition in the Astraeidae. The principal difference
is that, in the column-wall of Actinians, the middle lamina is
considerably thickened. The fibrous condition of the middle
lamina of various Zoantharia had been previously observed
by Kolhker, Schneider, and Rétteken, von Heider, Jourdan,
the Hertwigs, and Faurot.t As early as 1875, Allman remarked
that the ‘ hyaline lamella ’ (= middle lamina) of Myriothela
consisted of “two layers—internally a perfectly transparent,
thin, structureless membrane, and externally a layer of fibrillae,
which adheres closely to the structureless membrane’ (2,
p. 554, fig. 6).

The histology of the middle lamina of the Madreporaria
resembles that of mammalian connective tissue, the massed

1 Hickson in 1883 described the middle lamina of Tubipora as

consisting of a “ homogeneous matrix’ in which might be found ‘ cells
and fibres’ (17, p. 11).

112 GEORGE MATTHAI

wavy unbranched fibres, the network of branching fibres, and
the nuclei with the granular protoplasm in which they lie, are
comparable respectively to the white fibres, branching fibres,
and the so-called connective-tissue ‘ corpuscles’; the various
elements lie in a clear matrix in both cases. It is probable that
the branching fibres in the middle lamina of coral colonies are
elastic like those of connective tissue.

The possibility has not been excluded that the fibrous
strands of the ‘ epithelio-muscular ’ or ‘ myo-epithelial’ cells,
so commonly figured im connexion with the histology of
Coelenterates, might only be fibres of the middle lamina torn
apart with the adjacent parts of the ectoderm and endoderm
in the process of teasing. Such results are to be expected ; the
ectoderm, middle lamina, and endoderm are organically con-
tinuous. (In teasing, protoplasmic parts are sometimes
dissociated from the fibres of the middle lamina, as in Hickson’s
figures 28b-e of Alcyonium digitatum.) It is not
improbable that, as in the Astraeidae, these strands may
be of the nature of connective-tissue fibres, for, in figured
examples of epithelo-muscular cells, the nuclei, unlike their
position in plain muscular fibres of mammals, lie in the proto-
plasm extrinsic to the strands (16, Pl. vi; 18, Pl. xxxix;
3; Pl. xvi):

Bourne, H. V. Wilson, Duerden, and others regarded the
middle lamina of the Madreporaria as a secretion of
one or both of the protoplasmic layers and ‘ not formed by the
direct metamorphosis of the ends of ectoderm or endoderm
cells’ (87, p. 198). The appearances in my various prepara-
tions, however, suggest that the middle lamina is formed by
modification of part of the protoplasm of the ecto-endoderm
into cementing substance and fibres, all or some of the nuclei
in the modified part of the protoplasm becoming the nuclei of
the lamina. The formation of the middle lamina is well seen
in the case of the processes of attachment which are formed in
the calicoblastic layer, stages in their development being
abundantly present in my preparations (figs. 13-15). Where
such a process is to be formed, the calicoblastic layer is raised

HISTOLOGY OF ASTRAEIDS 113

into a short, somewhat irregular eminence which may or, may
not contain a nucleus. Subsequently, the protoplasm of this
projection becomes modified from its periphery inwards to the
middle lamina of the column-wall and specialized fibres appear
init. At the attachments of mesenteries to the corallum,; the
calicoblastic projections are usually larger and each of them
often contains more than one nucleus ; when their protoplasm
has been modified, the attaching structures become connected
with the middle lamina usually by means of narrow necks,
while elsewhere they are smaller and arise directly from the
middle lamina. In my preparations there is no indication
that these processes are at first formed in cellular elements or
“desmocytes * which become subsequently connected with
the middle lamina by the modification of neighbouring ‘ cells ’
of the calicoblastic layer, as Bourne described (p. 329), but
they are the result of a continuous change in the multinucleated
calicoblastic layer, the transformation of the protoplasm
beginning from its periphery and gradually extending inwards
to join the middle lamina. The processes are the parts that
project beyond the outer margin of the ealicoblastic layer and
in which the specialized fibres lie. These fibres are pronounced
towards the periphery of the processes, gradually becoming
fainter as they reach the middle lamina, and probably they
merge into the fibrils of the latter.1

Part of the middle lamina is formed entirely in the ecto-
derm, viz. the processes of attachment in the ealicoblastic

1 According to Bourne in Caryophyllia Smithii, ‘where a
desmocyte is about to be formed, one, two, or three nuclei become sur-
rounded with a mass of darker, finely granular protoplasm. The next
phase is the appearance of a band-shaped or ovoid body in the centre
of the granular protoplasm which already shows faint signs of striation . . .
usually one nucleus remains in close association with this body ; the others
(if more than one combine to form the granular protoplasmic mass)
appear to be concerned in the formation of the mesogladal process which
will join the desmocyte to the mesogladal lamina. The striations next
become more defined, and the desmocyte, which was at first separate from
the mesoglada, becomes attached to it by a process developed, as it seems,
at the expense of neighbouring cells ’ (pp. 528-9). (A desmocyte containing
more than one nucleus he regarded as a cell.)

NO. 265 I

titA GEORGE MATTHAT

layer, part of it arises in the endoderm, viz. the median core of
mesenteries, and the remainder is contributed to by both the
ectoderm and endoderm, viz. the middle lamina of the column-
wall, oral-disc, edge-zone, and stomodaeum. While Bourne
in 1899 held that the processes of attachment, which were
essentially similar to and became part of the middle lamina,
were formed by modification of elements in the calicoblastie
layer or ‘ desmocytes ’, i.e. were intra-protoplasmic formations,
he had in a previous paper (5) inferred that the middle lamina
itself was a secretion of ectoderm and endoderm, i.e. was an
extra-protoplasmic product.

The middle lamina appears to be essentially a supporting
stratum, i.e. has the function of connective tissue of Vertebrates
and, like the latter, has a fibrous texture. It 1s best developed
in mesenteries, since they support the oral-dise with the ten-
tacles and keep the stomodaeum in position, the longitudinal
pleats giving additional strength to the mesenteries. Owimg
to the presence of a calcareous skeleton to support the column-
wall, the middle lamina in the Madreporaria is very thin,
whilst in Actinians theabsence of such a skeleton has necessitated
a considerable thickening of the middle lamina in the column-
wall (being best developed in this region), which, when the
column-wall is folded longitudinally as in Metridium
senilis, is swollen into longitudinal ridges within the folds
or rugae (fig. 16). The column-wall processes are analogous to
tendinous structures in Vertebrates, since, doubtless, they
attach the soft parts to the corallum. This function would
account for their sucker-shape, usually concave attaching
surface, and comparatively small size, combined with their
numerical abundance ; the specialized fibres in them presum-
ably impart additional toughness to these processes.

Since the middle lamina has a spongy texture, the infilling
of its meshes with fluid would help in the distention of polyps,
which is, however, mainly effected by the ingress of sea-water
into the polyp cavities, while the general contractility of the
middle lamina would help in the retraction of polyps.

HISTOLOGY OF ASTRAEIDS 115

GENERAL CONSIDERATIONS.

In the tissues of the Astraeidae, cell-limits cannot be
discerned, the nuclei lying immersed in the general protoplasm.
This is particularly the case in the surface ectoderm, inner
lining of stomodaeum, and in mesenterial filament. While
in the endoderm, nuclei tend to lie between vacuoles, in the
middle lamina nuclei are few. and the protoplasmic areas con-
taining them are not definitely circumscribed but appear to
be organically connected together by means of their radiating
strands. Sections of 4 and 6 p» thicknesses treated with silver
nitrate failed to show any cell-limits, nor is a cellular structure
seen in sections cut in gelatin with Aschoff’s CO, freezing
microtome, nor again in celloidin sections of polyps. Duerden
observed that m Siderentsea radians the endoderm
of the wall of the polyp lining the uppermost parts of the
skeleton is “a syncitium showing no signs of cellular divisions ’,
and that the calicoblastic layer “in the growing areas of the
skeleton shows no evidence of cell limitations ’ (9, pp. 30, 31).
Gardiner, too, could not find definite cell outlines in Coeno-
psammia and Flabellum (18 and 14). Such outlines, so
frequently represented in figures of the ectoderm and endoderm
of the Anthozoa, are doubtless conventional and arbitrary.

The products of teasing of the tissues, whether before or
after maceration, cannot be regarded as separated units of
structure or * cells ’, but are really bits of protoplasm inevitably
torn apart with the nuclei in the mechanical process of teasing.
Hence, it is generally found that such pieces of protoplasm
possess neither regular nor uniform contour and are sometimes
torn apart with fibres of the middle lamina. If an appearance
of cellular strands is noticeable in some preserved tissues, it is
due to the shrinkage of protoplasm around the nuclei, which
probably act as centres of force. H. V. Wilson regarded the
endodermal mass of the solid planula stage of Manicina
areolata as a © plasmodium which was subsequently broken
up into cells’ (87, p. 200) ; he regarded the earlier blastosphore
as composed of cells, although their inner ends were ‘ not

I2

116 GEORGE MATTHAI

distinctly marked off from the solid endoderm’ (p. 197). In
many of Bourne’s figures of the soft parts of the Anthozoa
definite cell boundaries are not visible, although such limits
have been presupposed in the descriptions. Indeed, in
Caryophyllia, Euphyllia, Madrepora, and “several
other’ corals Bourne could not find any cell outlines in the
calicoblastic layer (7, p. 532), the latter bemg an irregular
multinucleated sheet of protoplasm. When more than one
nucleus was present in a mass of protoplasm, it was assumed to
be a coenocyte formed by the fusion of uninucleated cells ;
for example, in referring to scleroblasts or spicule-forming cells
of Aleyonaria, Bourne remarked that they were * often
coenocytes contaiming two, three, or more nuclei’ (p. 509).
It would appear to be more likely that the scleroblastic tissue
was of the nature of a syncytium in which spicular bodies
formed.

The mucous and granular vacuoles in the outer lamina of the
Madreporaria have also been regarded as cells, but nuclei
are not definitely related to them, some of them having more
than one nucleus while others show none at all. The only
cellular elements in the soft parts of the Astraeidae are
nematocysts, algal bodies, and the reproductive elements ;
these are all characterized by their definite and uniform outline.
Nematocysts are secondary formations in the ectoderm for
special purposes. Algal bodies are restricted to the endo-
derm, but little is known of their life-history ; it is doubtful if
they are symbiotic organisms, as is generally supposed, since
they are found in newly hatched larvae and even in earlier
embryonic stages (87, Pl. 11, fig. 4). Ova and spermatozoa lie
in spaces in the middle lamina (25, figs. 9, 10, and 49).

The laminae of Astraeid corals are therefore to be regarded
as syncytial, and since, as has been seen, there is organic
continuity between them, i.e. the ectoderm is everywhere
directly continuous with the middle lamina and the latter with
the endoderm, and, further, the ectoderm passes into the
endoderm by way of the inner linings of the stomodaeum, the
tissues form one nucleated continuum which has undergone

HISTOLOGY OF ASTRAEIDS LA 7)

partial differentiation into three strata. Towards the base
of the column-wall, the middle lamina is absent in places, the
calicoblastic layer and the endoderm merging into each
other. In such places the appearance is that of one sheet
of nucleated protoplasm with an outer granular stratum con-
taining large oval nuclei tangentially placed at intervals, which
represents the calicoblastic layer, and an inner stratum whose
nuclei are smaller but more numerous and placed vertically,
which represents the endoderm (fig. 10).

Since the middle lamina of the Astraeidae is nucleated,
formed early in development, and is of the nature of connective
tissue, it is comparable to the mesoblast and mesoderm of other
animals.t Bourne in 1887 restricted these terms to denote the
intermediate layer of the triploblastica (which he apparently
identified with the coelomata), on the view that the middle
lamina of Coelenterates was neither embryonic nor ‘ cellular ’,
to which he gave a new name, mesoglaea (5, p. 311). This
nomenclature was subsequently accepted by most authors—
Haddon, van Beneden, Hickson, Ashworth, MeMurrich,
Duerden—who regarded the nuclei occurring in the middle
lamina of Coelenterates as belonging to cells which secondarily
migrated into the gelatinous secretum from one or both of the
protoplasmic laminae—a view to which my studies on the
Madreporaria lend no support. Moseley, von Heider,
O. and R. Hertwig, and other earlier zoologists had deseribed
the intermediate layer of the Anthozoa under the term
mesoderm. This prior usage was resumed in 1895 by Faurot,
who, from his comparative study of many Actinian species,
disagreed with Bourne in regard to its supposed extra-proto-
plasmic formation and structureless consistency and the need
for a new terminology. It is also clear from the embryological

1 Bourne states that “ by mesoblast is meant a layer of undifferentiated
cells, developed in the embryo before the differentiation of other organs
or tissues from either one or the other or both of the primary germ-layers,
the epiblast and hypoblast. By mesoderm and its adjective mesodermic
are meant all such tissues in the adult as are clearly derived from the
mesoblast ’ (5, p. 314).

118 GEORGE MATTHAI

studies of Jourdan on Actinia equina and Balano-
phyllia regia, of E. B. Wilson on Renilla, and of
H. V. Wilson on Manicina areolata, that the middle
lamina appears early in development—in the solid planula
stage. I have found this to be the case in the solid embryos
of Favia fragum. Jourdan distinguishes between a
‘membrana propria’ and a granular mass ; while the origin of
the former was uncertain, the latter was said to be formed
by the severance and fusion of the mner ends of the ectoderm
cells of the body-wall, which subsequently become fibrous."
E. B. Wilson, who made a more or less similar distinction,
also found that the middle lamima of the body-wall was
formed by the separation and fusion of the swollen inner ends
of ectoderm cells, though he somewhat arbitranly termed the
process ‘a peculiar form of cuticular secretion’ (86, p. 759).
Bourne’s account of the formation of the middle lamina in
Heliopora (6) is not different from those of Jourdan and
E. B. Wilson.

Although a discussion of the highly controversial subject
of the history and homology of the germ-layers of the Metazoa
does not lie within the scope of this paper, it will be seen
from the foregomg account that there was not adequate
reason for withholding the application of the term mesoderm
to the middle lamina of the Anthozoa. j

BIBLIOGRAPHY.

1. Allman, G. J—‘‘ On the Anatomy and Physiology of Cordylophora ”’,
* Phil. Trans.’, exliii, 1853.

2. ——-‘‘ On the Structure and Development of Myriothela”’, ibid.,
elxv, p. 549, 1875.

3. Ashworth, J. H—‘‘ The Structure of Xenia Hicksoni, nov. sp., with
some Observations on Heteroxenia Elizabethae, Kollicker”’,
‘ Quart. Journ. Micr. Sci.’, xlii, p. 245, 1899.

1 The mode of formation of the middle lamina of certain Alcyonarians
described by Kowalevsky and Marion (28) is essentially similar to that
of Jourdan, but in their subsequent discussion they, however, came to the
conclusion that the middle lamina of Coelenterates was not homologous
with the mesoderm of Coelomates.

HISTOLOGY OF ASTRAEIDS 119

4, Beneden, Edouard van.—‘ Les Anthozoaires de la Plankton-Expedi-

tion’, * Résultats de la Plankton-Expedition der Humboldt-
Stiftung ’, ii, K.e., 1897.

5. Bourne, G. C.—“‘ The Anatomy of the Madreporarian Coral Fungia ”’,

6.

10.

at

12.

13.

14.

15.

16.

17.

18.

19.

20.
21.

22.

23.

* Quart. Journ. Micr. Sci.’, xxvii, p. 293, 1887.

—— ‘On the Structure and Affinities of Heliopora coerulea, Pallas.
With some Observations on the Structure of Xenia and Helero-
xenia ’’, ‘ Phil. Trans.’, clxxxvi, p. 455, 1895.

‘** Studies on the Structure and Formation of the Calcareous

Skeleton of the Anthozoa’’, ‘ Quart. Journ. Micr. Sci.’, xli, p. 499,

1899.

. Duerden, J. E——‘‘ West Indian Madreporarian Polyps ”’, ‘ Nat. Acad.

Sci. U.S.A.’, viii, p. 403, 1902.

“The Coral Siderastrea radians and its Post-larval Develop-
ment ’, Carnegie Institution, Washington, 1904.

Faurot, L.—‘ Etudes sur l’Anatomie, I’Histologie et le Développe-
ment des Actinies ’’, ‘ Archiv. de Zool. Exp. et Gén.’, 3° sér., iii,
p. 43, 1895.

—— ‘‘ Développement du Pharynx des Couples et des Paires de
Cloisons chez les Hexactinies ”’, ibid., 4° sér., i, p. 359, 1903.

*“ Nouvelles Recherches sur le Développement du Pharynx et
des Cloisons chez les Hexactinies ”’, ibid., vi, p. 333, 1907.

Gardiner, J. Stanley—‘*‘ On the Anatomy of a supposed New Species
of Coenopsammia from Lifu”’, ‘ Willey Zool. Results’, part iv,
p. 357, 1899.

“South African Corals of the Genus Flabellum rubrum, with
an Account of their Anatomy and Development ”’, “ Marine Invest.
South Africa’, ii, p. 117, 1902.

Heider, A. R. von.—** Sagartia troglodytes—Ein Beitrag zur Anatomie
der Actinien ’’, ‘ Sitzb. kais. Acad. Wien’, Ixxv, Abth. I, p. 367,
1877.

Hertwig, Oscar and Richard.— Die Actinien ’, Jena, 1879.

Hickson, Sydney J.—** The Structure and Relationships of Tubi-
pora’’, * Quart. Journ. Micr. Sci.’, xxiii, p. 556, 1883.

“The Anatomy of Aleyonium digitatum ”’, ibid., xxxvii, p. 343,
1895.

Jourdan, Et.—‘‘ Les Zoanthaires du Golfe de Marseille”, ‘ Ann. Sci.
Nat., Zool.’, 6° sér., x, no. 1, 1879-80.

KGllicker, A. yon—‘ Icones Histologicae ’, Leipzig, 1863.

Korotneff, A~—‘* Histologie de Hydre et de la Lucernaire ”’, ‘ Archiv.
de Zool. Exp. et Gén.’, v, p. 369, 1876.

Kowalevsky, A—‘** Zur Entwicklungsgeschichte der Alcyoniden ”’,
* Zool. Anzeig.’, 1879.

Kowalevsky, A., and Marion, A. F.—‘‘ Documents pour l’Histoire

120 GEORGE MATTHAI

embryogénique des Alcyonaires”’, “Ann. Mus. Hist. Nat. Mar-
seille ’, Zool. I, no. 4, 1883.

24. Lacaze Duthiers, Henri de.—‘‘ Faune du Golfe du Lion ”’, ‘ Archiv. de
Zool, Exp. et Gén.’, 3° sér., v, p. 1, 1897.

25. Matthai, G—‘ A Revision of the Recent Colonial Astraeidae possessing
Distinct Corallites ’’, ‘ Trans. Linn. Soc.’, London, 2nd ser., Zool.,
xvii, p. 1, 1914.

26. ** Reactions to Stimuli in Corals ’’, ‘ Camb. Phil. Soc.’, 1918,

27. * Colony-formation in Astraeid Corals,’

28. McMurrich, J. Playfair— Contributions on the Morphology of the
Actinozoa.’

29. —— (i) ‘‘ The Structure of Cerianthus Americanus *’, * Journ. Morph.’,

hy [Os Laila Urs

30. (ii) “‘ On the Development of the Hexactiniae”’, ibid., p. 303,
1891.
31. ——(v) ‘‘ The Mesenterial Filaments in Zoanthus sociatus (Ellis) ”,

* Zool, Bull.’, ii, no. 6, p. 251, 1899.

32. Moseley, H. N.—‘‘ On the Structure and Relations of the Aleyonarian
Heliopora coerulea, &c.”, ‘ Phil. Trans.’, clxvi, p. 91, 1875.

‘* Report on certain Hydroid, Aleyonarian. and Madreporarian
Corals ’’, ‘ Challenger Reports ’, Zool. II, part vii, 1881.

34. Schneider and Rétteken.—‘‘ On the Structure of the Actiniae and
Corals ’’, ‘ Ann. Mag. Nat. Hist.’, vii, p. 437, 1871.

35. Wilson, E. B.—‘‘ The Mesenterial Filaments of the Alcyonaria”’,
‘Mitt. Zool. Stat. Neapel’, v, p. 1, 1884.

36. ‘* The Development of Renilla’, ‘ Phil. Trans.’, p. 723, 1882.

37. Wilson, Henry V.—‘‘ On the Development of Manicina areolata’”’,
* Journ. Morph.’, ii, p. 191, 1889. ;

33.

EXPLANATION OF PLATES 7 AND 8.

LETTERING EMPLOYED.

alg., algal bodies. c.l., calicoblastic layer. ect., ectoderm. end., endo-
derm. gr.st., granular stratum at base of ectoderm. g7.v., granular
vacuole. lg.f., longitudinal fibres of middle lamina. m.f., mesenterial
filament. m.l., middle lamina. muc.v., mucous vacuole. %,, type L
nematocyst. ,, type IL nematocyst.

Fig. 1—Coeloria daedalea (Ell. and Sol.). Part of somewhat
oblique section through oral-disc and mesentery.

Fig. 2—Eusmilia aspera. Part of vertical section through edge-
zone. The endoderm is crowded with algal bodies.

Fig. 3—Coeloria daedalea (Ell. and Sol.). Part of transverse

G. Matthai, del.

Quart. Journ. Mier. Sei. Vol. 67,.N.S., Pl?

Tega i OEP Pace

ts

af

G. Matthai, det.

®
HISTOLOGY OF ASTRAEIDS 21

seetion through a tentacle showing a sub-terminal battery. Since the
section was cut somewhat obliquely, the middle lamina and its specialized
longitudinal fibres (lg.f.) appear thicker.

Fig. 4—Larva of Favia fragum (Esp.) ten hours after extrusion.
Part of transverse section through column-wall. Note fibrous condition
of middle lamina and protoplasmic remains in coelenteric cavity.

Fig. 5—Coeloria daedalea (Ell. and Sol.). Part of transverse
section (slightly oblique) through stomodaeum, showing a ridge and
adjacent intermesenterial areas.

Fig. 6—lIbid. Part of transverse section through a principal mesentery
just below stomodaeal region, showing continuation of median groove on
filament, -

Fig. 7—Ibid. Part of transverse section through a principal mesentery
below stomodaeal region, showing straight region of filament and the two
endodermal lobes.

Fig. 8—Ibid. Part of transverse section through a subsidiary mesentery,
showing straight region of filament and the two endodermal lobes.

Fig. 9.—Ibid. Part of transverse section through pleatal region of
a principal mesentery in stomodaeal region.

Fig. 10—Favia hululensis (Gard.). Part of transverse section
through column-wall at base of polyp. The middle lamina is thin and
somewhat discontinuous at this level.

Fig. 11—Coeloria daedalea (Ell. and Sol.). Longitudinal fibres
of the middle lamina of a mesentery after maceration in osmic-acetic
solution and staining in borax-carmine and picro-nigrosin.,

Fig. 12—Ibid. Part of tangential section (6 » thick) through a mesen-
tery, showing the network of branching fibres of the middle lamina. Note
the mass of unbranched fibres and nuclei in the middle lamina.

Figs. 13-15. Showing stages in the formation of column-wall processes
in the calicoblastic layer.

Fig. 13—Leptoria gracilis. Part of transverse section through
intermesenterial region of column-wall at level of stomodaeum.

Fig. 14—Ibid. Part of transverse section through column-wall at
attachment of a mesentery in stomodaeal region.

Fig. 15—Coeloria daedalea (Ell. and Sol.). Part of transverse
section through column-wall at attachment of a mesentery in stomodaeal
region.

Fig. 16—Metridium senilis. Part of transverse section through
column-wall, showing a ridge. In the middle lamina note (1) the swollen
loosely spongy core which is less open towards the ectoderm and endoderm,
the network itself consisting of branching fibres; (2) circularly arranged
unbranched fibres bounding the spongy core, the fibres being massed
against the endoderm; (3) nuclei in the spongy core; a thin granular
protoplasmic area can be seen around most of them. The points marked

72) GEORGE MATTHAI

ol al

in the meshwork are probably transverse sections of longitudinal fibres,
being more numerous where the meshwork is less open.

Fig. 17—Larva of Favia fragum (Esp.) ten hours after extrusion.
Transverse section through stomodaeum. Four mesenteries have joined
stomodaeum ; their stomodaeal attachments are shown in figure.

Fig. 18—Larva of Favia fragum (Esp.) ten hours after extrusion,
Part of transverse section through a primary mesentery, showing filament,

The Yolk-Sac and Allantoic Placenta in
Perameles.!

By

T. Thomson Flynn, D.Se.,
Ralston Professor of Biology, University of Tasmania.

With Plates 9-11 and 4 Text-figures.

CONTENTS.
PAG

1. INTRODUCTION : ; ; : « bd
2. REVIEW AND CRITICISM OF Soon Ware : ; . . 126
3. MATERIAL. ‘ ; : E ; ; : . 134
4, TERMINOLOGY : : . : : . 135
5. DESCRIPTIVE ACCOUNT OF Tol ATERIAL : : : : . 36
Stage 1. Perameles obesula, 61mm. : : . 136
(a) Significance of the Uterine Syncytium in Perameles 143
(b) General Remarks on the Fixation of the Embryo. . 145
Stage 2. Perameles gunni, 66mm. . : : . Tat
Stage 3. Perameles obesula, 7mm. . 5 ‘ . 158
Stage 4. Perameles obesula, 8-8:75mm. . ; 262
Stage 5. Perameles obesula, 12-5mm. : ps AEGZ
Stage 6. Perameles nasuta, post-partum . d . 165
6. SuMMARY OF CONCLUSIONS . . ; ‘ ; j . 165
7. CONCLUDING REMARKS . ; . : ; : ; . 168
(a) The Placental Conception . : ‘ : . 168
(b) Placental Phases in Perameles : : » WTA
(c) Placental Phenomena in Marsupials venbially ; ZZ

(d) The Relation of the Allantoic Placentation of Pera-
meles to that of the Eutheria . : : 5 ne
8. PAPERS REFERRED TO IN THE TEXT : s : ; lle:
9. DESCRIPTION OF FIGURES : ; : ; ; : . 180

1 The present communication was awarded the University Medal when
presented to the University of Sydney as a thesis for the degree of Doctor
of Science. The absolute necessity, in the present time, of reducing
publication costs to the minimum, has resulted in the omission of some
of the original figures, these being reduced from fifty-two to the present
limit.

124 1. THOMSON FLYNN

1. INTRODUCTION.

Ir is proposed in the present communication to initiate the
publication of a series’ of papers dealing with marsupial
embryology. The facts and interpretations embodied will be
based on result of an investigation into an amount of mar-
supial material now collected in the Biological Laboratory of
the University of Tasmania and fairly representative of the
Tasmanian fauna.

It is evident that any attempt to solve the phylogenetic
problems presented by the mammalian group without having
at our disposal a large number of the facts relating to the
morphology and embryology of the Metatheria will be
extremely unsatisfactory.

Yet in the not very distant future, with the spread of settie-
ment and the almost incredible devastation caused among
these animals by fur-hunters and trappers—a depletion but
little compensated for by restrictive legislation-it may be
considered certain that the connected stages necessary for the
elucidation of their intra-uterme development will be practi-
cally unobtainable.

In this connexion it may be instructive to quote the words
of Hubrecht (1909), who says, m discussing placental arrange-
ments in the Marsupialia, that ‘the early ontogenetic events
and the different phases in the mutual relation of the blasto-
cyst and mucosa ought to be fully known in order to furnish
us with all the data that can be brought to bear on this impor-
tant question. And it is to be fervently hoped that those
genera that are very rapidly diminishing in their native land,
some of them even on the verge of disappearance, may yet
be fully investigated before they have been exterminated,
and have thereby become as mute on this important point
as are their fossil predecessors.”

Of such disappearing genera two may be mentioned, Thyla-
cinus and Sarcophilus, both certainly primitive in
the morphology of their genital organs, and for that reason
alone possessing Important possibilities with regard to their

PLACENTA IN PERAMELES 195

intra-uterine nutritional arrangements. Both genera are
absolutely on the verge of extinction. The intra-uterime
development of Thylacinus will probably never be known,
and it is extremely unlikely that detailed investigation will
be possible in the case of Sarcophilus.

Under these circumstances it is a pleasure to acknowledge
gratefully any help extended towards an investigation of the
ontogenetic and phylogenetic relationships of this fast dis-
appearing fauna.

First, in this way, I must express my most grateful thanks
to the Trustees of the John Ralston Bequest and particularly
to their chairman, Dr. L. G. ‘Thompson, these gentlemen having
placed at the disposal of the University of Tasmania a sum of
money to be used chiefly in marsupial investigation. It is
mainly by their financial aid and unswerving support that
the important collection of embryological material in the
University of Tasmania has been brought together.

Secondly, I am indebted to the Committee of the British
Association for the Advancement of Science, who, at their
Australasian meeting in 1914, placed at my disposal a grant
with the object of procuring material for the study of the brain
and embryology of marsupials.

The preliminary portion of the exanunation of the material
of the present paper was carried on 12 my own laboratory, but
lack of facilities for consulting almost all the literature, and
other disadvantages of isolation in T'asmania, led me to consult
my old friend Professor 5. J. Johnstone as to the possibility
of accommodation in one of the laboratories of the University
of Sydney.

This was arranged and the work was carried on under these
altered conditions.

Still more recently, by the courtesy of Professors Hill and
Watson, I have been accommodated at University College,
London. I am indebted to the Senate of the University of
Sydney and to the Research Committee of that University
for a grant from the McCaughey Bequest to defray the cost
of the plates illustrating this paper.

126 T. THOMSON FLYNN

Finally, I must not omit to pay a tribute to the friendly
interest which has been shown in my work by my former
teacher, Professor W. A. Haswell, who has been always ready
to help with kindly criticism and advice. To him, also, I am
indebted for the loan of literature otherwise inaccessible to me.

9. Review AND CRITICISM OF PREVIOUS WorRK.

The discovery by Hill of a true allantoic placenta in the
two forms, Perameles obesula and nasuta, was of
the greatest importance in regard to the phylogeny of the
marsupial group and its relationship to the remainder of the
mammalia. The actual placental material, upon which Hill
was able to base his conclusions, consisted of three stages,
The first of these, Stage C, concerned an embryo 7 mm. long.
This was followed by a Stage D describing the placental con-
nexion in embryos 8 to 8-75mm. long. In a subsequent
paper Hill described the similar phenomenon in a 12-5 mm.
embryo which was also designated D. For the sake of clearness,
when speaking of these stages, I will designate them by their
length in each ease.

In his earliest (7 mm.) stage the fixation of the embryo was
already completed and the allantoic placenta was well on the
way towards being fully established.

From an examination of these three stages Professor Hill
showed that the epithelium of the uterus becomes conyerted
into a vascular syncytium, the nuclei of which arrange them-
selves in groups in the lower portion of this layer. ‘At the same
time maternal capillaries pass up between the syncytial lobules,
penetrate the syncytial protoplasm, and form a network on
and just beneath the surface ° (1897, p. 387).

The fixation of the embryo is brought about im the usual
way by means of the chorion. The ectodermal portion of this
membrane, by which actual approximation of the foetal
membranes to the uterine wall is first achieved, consists of
a single layer of large cells by means of which the attachment
is brought about.

With the somatic mesothelial layer of the chorion, the

PLACENTA IN PERAMELES LOT

splanchnic mesoderm of the allantois fuses, and the vessels
ramifying on the surface of the allantois thus come to lie
immediately below the layer of true chorion. Up to this point
the development of the placenta may be regarded as being fairly
typical, but, from now on, the development of this organ,
according to Hill’s account, results in the appearance of such
structural peculiarities and modifications as to give rise to the
general impression that the placentation of Perameles,
at any rate as concerns its more intimate development, is
without parallel in the whole mammahan group.

The allantoic placenta is completed ‘ by the gradual degenera-
tron and resorption of the enlarged chorionic ectoderm cells
over the placental area proper. ‘These cells thus take no further
share in placental formation.’ The result of this is the close
apposition of maternal and foetal blood-vessels, the two blood-
streams being ‘now only separated by their thin endothelial
walls and perhaps a thin layer of syncytial protoplasm ’ (p. 388).

A yolk-placenta is present, formed by the close apposition
of the vascular area of the yvolk-sac to the highly vascular
uterine syncytium outside the allantoic placental area.

It is not necessary at this stage to enter into any discussion
as to the significance of the presence of an allantoic placenta
in Perameles other than to indicate that it has been
definitely accepted by most embryologists, that there is now
no reason to doubt the common origin of Metatheria and
Kutheria from a primitive placental stock. Hill says it is
‘exceedingly improbable that an allantoic placenta should
have been twice independently acquired and in such a funda-
mentally similar manner within the limits of the mammalian
class ’ (p. 433).

Now, outside the intrinsic mterest of the presence of an
allantoic placenta in a member of a group formerly regarded
as aplacental, it would be expected, on a priori grounds,
that the occurrence would be of importance in elucidating the
phylogeny of the mammalian placenta or, at least, in giving
us some means of arriving at a definite idea of the method of
placental formation in the original protoplacental group.

5 T. THOMSON FLYNN

—
to

It need hardly be said that in these respects the placentation
of this animal, as interpreted by Hill, is extremely disappointing.

The combination of a persistent uterine syncytium with
a degenerating chorionic ectoderm is without parallel in the
Eutheria, and any attempt at a comparison of the placenta-
tion of the two groups results in a deadlock.

Although Hill’s results have been accepted by many, never-
theless Hubrecht (1909) ventures to question whether, morpho-
logically, the placenta of Perameles will not prove on further
investigation to be more comparable with some one or other
of the various placental styles found in the Kutheria. And
that there is some foundation for Hubrecht’s opinion is un-
doubtedly apparent from an examination of Hill’s figures
alone.

Before passing to a consideration of this question it will
perhaps be necessary to have before us a short account of some
special details of the placentation of this marsupial as described
by Professor Hill.

His most important and peculiar point concerns the history
of the foetal ectoderm. The portion of this which is concerned
in the fixation of the embryo is said to disappear almost
completely, being represented in the 12-5 mm. stage by a few
scattered cells in the original ectodermal position. The dis-
appearance is stated to be of the nature of a degeneration, this,
apparently, not being inaugurated until the allantoic attach-
ment has taken place and the allanto-chorionic fusion com-
pleted. In the earliest placental stage (Stage C, 7mm.) the
allantoic placenta is already on the way towards full establish-
ment, the allantoic vesicle being already attached to the chorion
by its placental face. Excellent figures are given of this stage
(1897, Pls. xxix and xxx, figs. 5 to 12).

At a later stage (D, embryos 8-8-75 mm.) the ectoderm has,
apparently, almost completely disappeared. In the description
of Stage C, it is stated that the cells towards the central portion
of the foetal ectoderm are ‘ of a very varying size and shape,
and, in places, through the disappearance of the outlines
between adjacent cells, large multinucleate cells have been

PLACENTA IN PERAMELES 129

formed. ... In many of the ectoderm cells shown in fig. 9
the nuclei are also seen to be in various states of disintegration.
Many of them stain only slightly ; the nuclear membrane is
becoming indistinct, while the chromatin is found broken up
and diffused in the form of small granules throughout the
delicate nuclear reticulum. Eventually the position of the
nucleus is only marked by a few straggling irregularly thickened
remnants, which finally become diffused through the proto-
plasm and lost to view’ (p. 404). [Italics mine.]

In this stage (embryos 8-8-75 mm.) the ectoderm is repre-
sented centrally, ‘only by more or less isolated degenerating
cells ’ (p. 413).

In the 12-5 mm. embryo Hill states that ‘ over the placental
area usually single, much degenerated, and deeply-staining
chorionic ectoderm cells are still to be found * (1899, p. 9).

From the above it appears that one of the causes of the
disappearance of the ectoderm is the loss of the chromatic
constituents of the nucleus, these being absorbed into the
surrounding cytoplasm.

As to the other possible reasons for the presumed degeneration
and disappearance of this ectodermal layer, Hill states further
that, mm some cases, ‘the inner ends of the cells are greatly
vacuolated, a fact which suggests that a process of vacuolation
may also play a role in the retrogression of the chorionic
ectoderm ’ (1897, p. 404).

Further, referring again to this layer, he is ‘inelined to
believe that the allantoic capillaries, so closely related to its
inner surface, are by no means the least active agents in effecting
its removal. Of direct fusion of the degenerate ectoderm with
the underlying syncytium there can be no question. All the facts
negative such a view’ (1897, p. 404). [Italics mine.]

From the above it appears that Hill is of the opinion that
the layer of chorionic ectoderm disappears, and that the
following processes mutually assist in causing this removal :

(a) Degeneration in situ with or without vacuolation.

(b) Removal by allantoic capillaries with or without previous

degeneration.

NO. 265 K

130 T. THOMSON FLYNN

Thus it is definite enough, according to Hill’s views, that
the chorionic ectoderm takes no share in the formation of the
placenta proper.

The completion of the placenta is brought about by the allan-
toic capillaries comimg into intimate relationship with the
maternal capillaries ramifying on the surface of the maternal
syncytium.

The above is, | am inclined to think, an accurate précis
of Hill’s results. His interpretation of the facts and his
strongly-expressed opinion that there is nothing of the nature
of a fusion between the foetal ectoderm and maternal tissue
has given rise to the idea—almost generally accepted—that
the placentation of Perameles is of ‘a peculiar type not
met with anywhere else ’ (Jenkinson, 1913, p. 216).

The correctness of Hill’s conclusions has been questioned by
A, A. W. Hubrecht in his famous essay on ‘ The Early Onto-
genetic Phenomena in Mammals’ (Hubrecht, 1909).

Hubrecht’s criticism:—

The opinion of this gifted investigator was arrived at evidently
on somewhat theoretical grounds, but he was confirmed in his
ideas by an examination of material placed at his disposal by
Professor Hill. It is important to understand Hubrecht’s
standpoint thoroughly, because, in my opinion, it is to some
extent justified. He believes that the foetal ectoderm, so
far from disappearing, penetrates into the maternal syncytium
to form a ‘ mixed syncytium ’, corresponding to what Schoen-
feld has described (1903) for the dog. “The Perameles
placenta may be said to be a somewhat simpler—because
thinner—form of placenta than that of the Carnivora,
but at the same time to approach more closely to that type ;
whereas amongst the Insectivora, Sorex provides us
with an example of a yet more extensive proliferation of the
material uterine epithelium before the allantoic attachment
of the blastocyst comes about than even Perameles.
At all events, the placentation of Perameles, characterized
by so intimate a fusion between foetal and maternal elements,
should never be classified amongst those forms of placenta

PLACENTA IN PERAMELES 131

which are either primarily primitive (as yet unknown to us)
or secondarily simplified (Ungulates, Lemurs, Cetacea,
&e.) ’ (1909, p. 117).

With similar material before each of them hardly could two
authors come to more different conclusions. Whereas Hill is
of the opinion that there is no interfusion of maternal and
foetal tissues other than that caused by the intergrowth
of allantoic tissue—which is comparatively insignificant—
Hubrecht is just as emphatic that an invasion of the maternal
syncytium by the foetal chorion does occur and that therefore
the type of placentation in Perameles corresponds closely
with that of some of the Eutheria, notably the Car-
nivora.

Interpretation of Professor Hill’s Figures:—

Without as yet adducing any evidence from my own material,
I may be permitted to say that from the evidence of Hill’s
figures alone there appears to be a considerable defence for
Hubrecht’s standpoint in this matter.

According to Hill’s account (1897, 1899) the histology of the
placental area is extremely simple. With the exception of the
syncytial nuclei, endothelial nuclei, and leucocytes, the only
elements of this region are the cells of the foetal chorion,
which in no case enter into the constitution of the placental
thickening but degenerate and disappear.

With this interpretation before us it will be possible to
proceed to examine his figures, particularly those of Pls. xxix
and xxx, figs. 7, 8, and 9 (1897), representing portions of the
placental area with the chorionic ectoderm attached. These
figures are all representative of Stage C (7 mm.).

As Hill has shown, there is a marked difference in the condi-
tion of the chorionic layer dependent more or less onits distance
from the centre of the fixed area. In general terms it may
be stated that the farther from the centre of the attached area
the less alteration is evident.

Examining PI. xxix, fig. 7, representing the edges of the
placental area of Stage C, it will be seen that the ectoderm is

K 2

132 T. THOMSON FLYNN

almost intact throughout. At certain localized points, however,
the ectodermal cells are deeper than the average. Some cells
have become multinucleate and the nuclei are in many cases
situated basally, such cells being deepened considerably com-
pared with their original condition. Below the ectoderm the
syncytial nuclei are arranged in their lobules. These nuclei
are rounded, and although membranate the chromatic contents
are spare. Hach nucleus contains usually a well-defined
nucleolus. Examination of the syncytial nuclei of the extra-
placental area shows that the suppression of the chromatin
contents and the presence of a single rounded nucleolus is
a general characteristic. Hyven with iron-haematoxylin staining
only a very faint network can be made out. It is quite other-
wise with the nuclei of the cells of the ectoderm layer. A refer-
ence to Hill’s Pl. xxx, fig. 9, will serve to show that the nuclei
of these cells contain quite a well-developed chromatin network
with a number of karyosomatic aggregations, even as many as
half a dozen in some eases. Further, these nuclei are seldom
spherical, but irregularly elliptical, ovalish, or lenticular, and
generally of large size.

The difference in shape and_ histological characteristics
between these and the syncytial nuclei is most marked, and
in my preparations they can immediately be distinguished from
one another.

Bearing in mind, then, the difference between the rounded,
bead-like, maternal syncytial nuclei— typical resting nuclei “—
now congregated in groups in the syncytial lobules whether
within or without the placental area—and the more robust
layer of irregularly-shaped nuclei of the foetal ectoderm, each
with its network of easily-stained chromatin, it will be possible
to follow out their migration and rearrangement ina general way.

In Hill’s Pl. xxix, fig. 7, which represents a marginal
portion of the placental area with the chorion attached (embryo
7mm.), the foetal ectoderm (ch.ect.) is apparently complete
and unbroken, while on the lower side of the figure are
the syncytial lobules containing groups of syncytial nuclei.
Between these two sets of elements are to be seen a number of

PLACENTA IN PERAMELES 133

nuclei of the origin of which nothing is said by Hill. Apparently
he leaves it to be inferred that they are of the nature of syncytial
nuclei which have not yet reached the lobules.

In my opinion it is quite definitely indicated that these nuclei
originate from the foetal ectoderm through its active prolifera-
tion, and such centres of proliferation are to be seen in this
figure. It needs, I think, no other evidence than that of Hill’s
fig. 7 to show conclusively that such a process of proliferation
is IM progress.

Any possible doubt, however, must be dispelled by an
examination of figs. 8 and 9 of Pl. xxx of the central portion
of the placental area. These are drawn at a greater magnifica-
tion than fig. 7, and show the features I have indicated above
with more certainty. I cannot see that any other conception
than the one | have suggested can be possible. Particularly
is this evident in the case of fig. 9. Here a most active prolifera-
tion and migration of the chromatically rich trophoblastic
nuclei is quite apparent. They have so far advanced as to
invade the syncytial lobules. In the latter position the original
maternal epithelial nuclei are being overwhelmed by the ad-
vancing ranks of foetal nuclei. In some cases but one or
two trophoblastic nuclei have entered the syncytial nests ; in
extreme cases maternal syncytial nuclei appear to be entirely
absent, their place being taken by the newly-arrived, evidently
phagocytic, foetal nuclei. One result of this is that the original
chorionic ectoderm is now no longer a perfectly discrete layer.
Another is the inclusion of the maternal capillaries by the eyto-
plasm of the advancing trophoblast and their consequent
approach towards the stratum of allantoic capillaries.

The above statements being granted, it will be easy enough
to apply the new interpretation to the remainder of Huill’s
stages and figures. ‘To the further consideration of this I will
return in the descriptive portion of the present paper.

Gland Alteration.—There is still, however, one point to
which I would like to draw attention at this stage. It concerns
a drawing of a gland in PI. xxxi, fig. 18, one of Hill’s many
figures in which a gland is depicted. In the whole of the paper,

134 T, THOMSON FLYNN

no mention is made of any gland alteration, nor would any such
be likely to occur under Hill’s conception of a degenerating
chorion with a passive syncytium—yet distinct traces of such
alteration is evident in this figure. It will be seen that up to the
point where the gland enters the presumable syncytium its
epithelium consists of the somewhat low cubical cells so charac-
teristic of many of the glands of this stage. From this point to
the opening of the gland, however, there is abundant evidence
of degeneration. Apparently this consists of a syncytialization
similar to—though not as marked as—that occurring in the
Carnivora. It was the evidence of this phenomenon in my
own material which first drew my attention to the possibility
of the occurrence of a more complex process in Perameles
than was described by Hill, and in itself lends sufficient colour
to the view I have expressed above that there is something
more to be reckoned with in the placentation of this animal
than a simple degeneration of the foetal ectoderm.

Accepting the fact that the trophoblast of the placental area
in Perameles proliferates and that the uterine epitheliam
after a preliminary preplacental extension remains afterwards
passive, then we can come to the conclusion that placental
phenomena in Perameles can now be brought more or less
into line with similar phenomena in the Eutherian mammals.
Here, further, | may be allowed to state that I will be able to
show by the aid of my own material that the chorionic ectoderm
after attachment proceeds to form by proliferation two
structures :

(a) a plasmodiblast, plasmoditrophoblast, or

plasmodium,

(b) a cytoblast or cytrophoblast,

_and that Hill failed to recognize the presence of the plasmodi-
blast nuclei, the structure which he ealls the chorionic ectoderm
being really only the cytoblastic portion of that layer.

3. MATERIAL.

At my disposal for the examination of the foetal membranes
of Perameles, I have two intra-uterine stages both of which

PLACENTA IN PERAMELES 185

are important. ‘The younger one, a specimen of Perameles
obesula, shows the first fixation of the chorion; the older one,
which belongs to the Tasmanian form, P. gunni, is a stage
in which the first attachment of the allantois is in progress.

Both specimens were preserved in Hill’s fluid (picro-nitro-
aceto-osmic). Sections were stained, sometimes with Ehrlich’s
haematoxylin, sometimes by the iron-haematoxylin method,
counter-stained in each case by means of eosin.

In addition, | have been able, by the courtesy of Messrs.
L. Harrison and K. A. Briggs of the Zoological Department
of the University of Sydney, to refer to several excellently-
preserved sections in the collection of that department and
representing some of the material mounted by Dr. Hill of the
stages described in his paper. This very important collection
is as follows :

Perameles obesula, 7mm. stage, one microscope slide
containing one representative section of the uterine wall,
showing the placental and extraplacental areas and the attach-
ment of the allantois.

Perameles obesula, 12-5 mm. stage, one slide with five
sections similar to the above.

Perameles nasuta, post-partum stage, one slide with
one section.

All the above are stained with haematoxylin and eosin.

Hill’s Stage D, representing embryos 8-8-75 mm., is, unfor-
tunately, not represented in the Sydney University collection.

4+. TERMINOLOGY.

The expressions “ omphalopleure ’,* vascular omphalopleure ’,
and * bilaminar omphalopleure ’ first used by Hill are so con-
venient and expressive as to need no apology for their continued
employment.

The term ‘chorion’ or ‘ true chorion’ will be used by me
in the same sense as by Minot and Hill to indicate that part of
the extra-embryonic somatopleure which remains after separa-
tion of the amnion.

I will follow the example of most embryologists in using

136 T. THOMSON FLYNN

Hubrecht’s term * trophoblast ° for the outer ectodermal layer
of the mammalian blastocyst without, however, associating
it with the more recent theoretical and hypothetical meaning
with which Hubrecht has invested it.

I shall also use Minot’s expression ‘ trophoderm ” for that
portion of the trophoblast which proliferates and enters into
relationship with the maternal epithelium. As will be seen
later, this, in Perameles, undergoes changes comparable
to those occurring in the EKutheria. Similarly we find
a complementary structure, the maternal * trophospongia ’, used
to indicate (Hubrecht, 1909) ‘maternal cell proliferation,
specially intended for the fixation of the blastocyst ’.

5. Descriptive Account oF MATERIAL.
Stage 1. Perameles obesula, 6-1 mm.

This is perhaps the most important stage which has yet
been examined in the placentation of Perameles. Its
investigation definitely shows the fundamental connexion
between the placentation of this animal and that of the
Eutheria.

The specimen of Perameles obesula, on the examina-
tion of which the followmg account is based, was trapped by
myself some miles from a small town in the Tasmanian midlands.
It had apparently been dead in the trap for about an hour, but
no sign of post-mortem change was to be detected. It was
dissected on the spot and both uteri were found to be swollen.
Conditions at the time made any dissection of the uteri madvis-
able. They were opened slightly and placed in fixing solution
(Hill’s fluid). Later detailed examination showed that the
uterus of the right side was pregnant, but. unfortunately,
the delicate foetal membranes had been somewhat damaged.
The general condition was that the chorion was already
attached to the uterime syncytium over a small area, but that
the allantois had not yet come into relation with the conjoint
layer so formed.

Pregnant Uterus.—tThis was found to contam an

PLACENTA IN PERAMELES 13y

embryo of 6:-1mm. direct length, attached to the uterme
epithelium by a portion of the true chorion.

In the uterus the allantoic placental area at this stage is
distinguishable by the fact that its surface is marked by folds
noticeable at once by their depth and distinctness.

To one of these folds the foetal trophoblast is attached.

Sections show that the uterine epithelium has become con-
verted by loss of cell outlmes and by proliferation and migra-
tion of the nuclei into a syncytium as described by Hill.

I will proceed first to give a description of the maternal
structures afterwards passing to those more concerned with the
embryo.

Morphology of the Synevtium.—this varies greatly
in character according to the locality in the uterus. Over the
main wall of the uterus the syncytium is thin, 0-035 mm., while
in the region of attachment of the trophoblast it measures as
near as can be judged about 0-07 mm.

In the allantoic placental region it is that the complexity
of the syncytium has reached its maximum. Here, as Hill
has already shown and as happens over the remainder of the
uterus to a less degree, the nuclei of the original epithelium
have proliferated and migrated to the deeper portion of the
layer, which has now markedly thickened. The result is the
formation of a syncytium in which the deeply-situated nuclei
assume a particular form and arrangement. ‘These nuclei
become aggregated mainly in rounded masses or nests situated
in lobular projections of the syncytial protoplasm. The lower
surface of the syncytium has a wavy appearance due to the
presence of these lobules.

The syncytial nuclei at this stage are rounded with a well-
defined membrane, a distinct nucleolus, and indefinite chromatin
network which, however, is slightly more evident than it is in
later stages. Their lack of staining qualities makes them easily
distinguishable from the newly-formed trophoblastic nuclei.
Careful investigation of the arrangement of the epithelial nuclei
in each lobule shows that, when finally at rest, they are more
or less definitely arranged round a central cavity. This arrange-

138 T. THOMSON FLYNN

ment is, In many Cases, somewhat irregular, as is shown im
figs. 1 and 2 (cav.), but a glance at the further fig. 3, which
depicts quite a common arrangement, will show that in many
cases the syncytial nuclei come to form a more or less definite
layer round the central space.

The latter becomes filled by infiltrated lymphatic material
(fig. 3, onf.), which in fact is copiously distributed throughout
the syncytium in and between the lobules. At this stage also
the syncytium is well vascularized, each capillary bemg enclosed
in its delicate endothelial layer. Endothelial cells and leuco-
cytes are a well-marked feature of the syncytium. They will
be referred to in more detail later on.

The syncytium outside the allantoic placental area gradually
decreases in thickness. In the region opposite the bilaminar
omphalopleure the arrangement of the nuclei is essentially in
groups similar to those of the placenta area, yet these aggrega-
tions are not so distinctive nor so individual as in that area.

This portion of the uterine epithelium is also extremely well
supplied with capillaries, many of which reach the surface.
Lymph also finds its way into the uterine lumen through the thin
portions of the uterime epithelium between the ill-defined
syncytial nests.

Leucocytes are also present in this region but not nearly sO
plentifully as in the allantoic placental area.

Remainder of the Mucosa.—The main portion of this
consists of the much-branched connective-tissue cells, the
branches of which are extremely delicate and contain in their
meshes abundant lymph material. In the stroma are contained
elands and blood-vessels, and around these the connective tissue
is condensed to form a thin investing layer.

One feature of this stage is that the glands in the allantoic
placental region are narrower and more closely packed than i
the remainder of the mucosa. No doubt proliferation of the
glands has occurred, their length has increased, and their
courses become more tortuous without there bemg as yet
a sufficiently accommodating increase in the thickness of the
mucosa. ‘They measure in this region, on the average, -05 mm.

PLACENTA IN PERAMELES 139

in diameter while outside this region the average width is
0-065 mm. A result of this is that the mucosa of the former
region has a more compact appearance than in the latter.
Another feature to which attention should be drawn is the
presence of at least one branched gland, a photograph of which
is shown in fig. 10 and an outline, obtained by superimposing
a number of sections, in Text-fig. 1. This figure by no means
represents all the branches of the gland. This, I believe, is the
first record of any but simple glands in the marsupial uterus.

TExt-Fic. 1.

Diagram of a branched gland from the uterine mucosa of
Perameles.

In all cases the glands in their lower portions are narrower
and more coiled than in the upper, where in general they widen
out and retain their epithelium unchanged up to the point of
opening into the uterus. This latter characteristic, however,
is not shown by glands opening into that portion of the syns
cytium to which the chorionic ectoderm is attached.

The gland epithelium is of the usual character, consisting of
a single layer of cells with peripherally situated dark-staining
nuclei. The secretory activity of the glands is most marked,
particularly in the allantoic placental area. Migrating through
the glandular epithelium, to be added to the secretion, are
numbers of leucocytes.

Abundant in the connective tissue of the placental regions are

140 T. THOMSON FLYNN

cells containing pigment in the form of black streaky and
granular deposits. ‘These cells occur throughout the whole
wall of the uterus. They are very abundant in the serosa
and are found distributed through the muscularis, the connec-
tive tissue, and the glandular epithelium. Such pigmented
cells have often been noted in the virginal and pregnant uteri
of Eutherian mammals.

Foetal Structures.—In general the arrangement and
histology of the foetal membranes are in agreement with the
description given by Hill, and I find I can add nothing of
importance to his description of these structures.

Allantois.—the vesicular portion of this is a somewhat
flattened body, taking, however, a curved shape corresponding
to the dorsal curvature of the trunk of the embryo. In surface
view it is somewhat elliptical, measuring 5mm. by 3-1 mm.
The point of attachment of the stalk is placed a little nearer
the posterior than to the anterior end of the vesicle. The
difference in thickness and texture between the placental and
coelomic surfaces of the allantois is easily seen with the naked
eye. The coelomic surface is an extremely tenuous sheet
bearing the larger blood-vessels, while the outer or allantoic
surface is more opaque and abundantly supplied with a network
of capillaries derived from or supplying vessels which pass round
the margin in the manner described by Hill. For a full deserip-
tion of these allantoic vessels I would refer the reader to Hill’s
account. The allantoic stalk has the usual relations and
structure.

Fixation of the Embryo.

The importance of this stage rests on the fact that, over
a very small area, the trophoblast is now attached to the
thickened maternal syncytium (trophospongia). ‘This
portion of the foetal ectoderm is, of course, the outer layer
of the chorion, which consists, in addition, of somatic meso-
derm. The latter is a thin mesothelial layer consisting of
flattened cells with oval, somewhat deeply-staining, nuclei.

The chorionic ectoderm typically consists also of a single
cell-layer as it undoubtedly does in the margial free portions.

PLACENTA IN PERAMELES 141

Over the area of fixation to the uterme syncytium, however,
an important and highly significant alteration has been im-
pressed upon the ectoderm by which it becomes converted into
a typical trophoderm (Minot) or ectoplacenta (Duval)
so characteristic of this layer in the Hutheria.

At certain points in this portion of the trophoblast, cell
proliferation takes place. The cells of the original layer divide
to give rise to nucleated groups in which the cell outlines have
disappeared. These cytoplasmic aggregations possess an
irregular contour due to the presence of pseudopodial pro-
cesses, so there is distinctly present here a layer definitely
homologous with the plasmodial structure (plasmodium,
plasmodiblast, or plasmoditrophoblast) so charac-
teristic of the placentation of the Hutheria. The appear-
ance of the plasmodiblast at this stage is shown in figs. 4, 5,
and 6. At various points the plasmodial nuclei invade the
uterie syncytium. The soldering of the foetal trophoderm to
the maternal syncytium is brought about by the above-
mentioned pseudopodial processes, in the meshes of which
numerous spaces are enclosed. ‘The remaining basal cells of the
trophoblast layer form the cytoblast or cytotropho-
blast. This is by no means at first a definite cell-layer. It
is apparently not till a little later that the basally-situated
nuclei divide in a regular way to form the more definite cellular
layer known as the cytoblast.

A point of the greatest significance, and one to which I shall
later refer, is the fact that the localities of proliferation are
determined by the presence of the syncytial nests, and it is
into these that the plasmodial masses pass. Fig. 4 shows this
phenomenon, while it is also indicated in fig. 8, in which,
however, the nests are not quite cut centrally. The effect of
the growth of the attached trophoblastic cells on the maternal
structures is shown in figs. 4 and 7. In fig. 4 the chorionic
ectoderm cells show but little departure from their original
linear arrangement, but have already begun to give off pseudo-
podial processes which immediately phagocytically attack the
syncytial nuclei of a neighbouring syncytial lobule, only part of

142 T. THOMSON FLYNN

which latter is shown in the drawing. Only the nuclei of the
nest outside the range of attack are seen to preserve their
original shape and structure, the others bemg in the. state of
degeneration. Apparently this consists in the loss of contour
through the breaking down of the nuclear membrane followed
by loss of chromatin and virtual disappearance. Figs. 5 and 6
show stages in the proliferation of the chorionic ectoderm
cells. In all cases they give the impression of thrusting forward
wedge-shaped plugs which penetrate into the aggregations of
maternal nuclei.

There happens to be but one gland present in the small
area to which the trophoblast is attached at this stage. This
is shown in fig. 7, which depicts a section through the actual
opening of the gland. Its uterme mouth is seen to be blocked
by the overlying chorion, the ectodermal portion of which may
now be regarded as forming two distinct portions, a basal
portion, the cytoblast (cyt.), and a plasmodial portion, the
plasmodiblast. The growth of the plasmodium has extended
a considerable distance, particularly on one side, where the
protoplasmic processes have caused degeneration in the gland
epithelium cells similar to that which occurs in the maternal
nuclei. Here, as before, the nuclei have lost their contour,
their position in the cell-lme, and their chromatin. They
decrease in size and disappear, bemg evidently ingested.
Further degeneration of the gland-cells is foreshadowed by
the presence of protoplasmic processes involving them. Also
in this figure will be seen remnants of syncytial nuclei and other
remains, haematids and leucocytes.

That the cells of the plasmodiblast are phagocytic cannot be
doubted. Their effect on the syncytial and gland nuclei is
some evidence of: this, but the presence of numerous rounded
granules such as those shown to the left of fig. 7 (ing.) and other
cellular débris makes the matter certain.

Pigmented cells are present in the trophoblast. This pig-
ment is black and is arranged either as minute granules or as
an aggregation of streaky lines, usually in the neighbourhood
of the nucleus.

PLACENTA IN PERAMELES 143

From the above account it will be evident that the method of
fixation of the embryo to the uterine wall in this marsupial
is fundamently similar to the general type occurring in the
Eutheria. If, for example, the figures of this stage which |
have given (figs. 4, 5, 6, 7, 29, and Text-fig. 4) be compared with
the essentially similar drawings of the comparable phenomena
recorded by Schoenfeld for the dog (1903), (PI. xxi, figs. 14
to 17) (see Text-fig. 3), it will be seen that the difference in
the two forms rests mainly, in the early stages, on the behaviour
of the uterine epithelium. This, in the dog, remains as a distinct
layer until it undergoes degeneration as a result of the inroads
of the plasmodiblast.

The Non-pregnant Uterus.—This was examined in
sections and found to have undergone changes corresponding
to those which had occurred in the right. The ovary of this
side, however, had a well-developed corpus luteum. I will,
therefore, not enter into a detailed description of the structure
of this uterus further than to mention the following :

The epithelium has developed into a syneytium abundantly
supphed with capillaries, many ramifying on the surface, some
of which discharge their blood by extravasation into the
lumen. A point of interest is the fact that, on the dorso-
mesial side of the uterus, the syncytium is somewhat thickened
and the mucosa is here developed into deep folds. At its edges
this thickening passes off gradually into the rest of the syney-
tium. This thickened portion possibly represents the maternal
trophospongia. Throughout the whole uterus the glands
show no degeneration. They are in an active state of secretion
and the inner ends of the gland-cells are torn and frayed out
through abundant breaking off of cellular secretion. Cilia,
therefore, are not to be found.

(a) Significance of the Uterine Syncytium
in Perameles.
Proliferation and syncytialization of the uterine epithelium
is a well-marked feature of maternal preparation for allan-
toic placentation in many Kutheria. In many cases the

144 T. THOMSON FLYNN

preparatory proliferation is soon interfered with by the destruc-
tive action of the blastocyst on the uterine epithelium.

Hill (1897, p. 393) institutes a comparison between Pera-
meles and Sorex in the matter of the proliferation of the
uterine epithelium. As a consequence, however, of his belief
in the degeneration of the chorionic ectoderm and the
persistence of the uterine epitheltum in Perameles_ he
could not carry the comparison far enough.

In Sorex, Hubrecht states that over the future allan-
toidean and omphaloidean placental areas the epithelium
undergoes a tremendous proliferation and development into
a cell aggregate of relatively great thickness.

The history of this maternal formation in the allantoic
placental area of Sorex is, I think, worthy of particular
consideration and of comparison with what happens in
Perameles. MHubrecht shows (1894, p. 492 seq.) that, m
the shrew, the nuclei of the epithelial proliferation become
arranged in fan-shaped groups at comparatively regular
distances, the centre of each group being without nuclei
(1894, fig. 69).

‘In the following stage ’, says Hubrecht, ‘ this arrangement
becomes converted into a functionally more important one.
The centre of the fan-shaped structure becomes an open crypt,
the protoplasm breaking up and the peripheral nuclei forming
the epithelial liming of the crypt. The uterine epithelium
breaks away from under the crypt and the inner lining of the
erypt solders with the surrounding epithelial surface at the
lower border ° (p. 493).

It is quite easy to see the resemblance between the shrew
and Perameles in respect of the phenomenon here deseribed.
In both there is intense epithelial proliferation, particularly
in the placental area. The resulting nuclei or cells are in both
arranged in nest-like groups. While in the bandicoot these
groups remain practically unaltered, in Sorex they are
transformed into epithelial crypts. Nevertheless, as I have
shown above, the proliferated nuclei in Perameles
take on a more or less definite arrangement as

PLACENTA IN PERAMELES 145

a layer bounding a central space. This is indicated
in figs. 1, 2, and 3, whilst in other figures, 4, 7, and 9, I have
drawn attention to the fact that the trophoblastic proliferations
bear a definite relation to the groups. Particularly in the next
stage, it will be seen that the foetal nuclei invade and fill the
nest with consequent more or less complete disappearance of
the maternal nuclei. Under these circumstances it does not
require any extraordinary stretch of imagination to recognize
in this highly characteristic and important phenomenon the
remains of a much more elaborate system of placental forma-
tion. The conclusion is certainly obvious to me that here in
Perameles, in the formation of the peculiar syncytial
groups, there is to be recognized an abortive attempt at the
formation of crypts such as occur in the placental area of
Sorex, and further. while in the latter crypt-building is con-
fined to the placental area, in Perameles the comparable
phenomenon occurs at all points of the uterine epithelium,
although in a lesser degree opposite to the omphalopleure
than in the placental area.

(b) General Remarks on the Fixation of
the Embryo.

Here it will be convenient to interpolate a few remarks on
the method of fixation of the blastocyst and on the general terms
used to express the nature of the structures taking part in it.

Fixation is brought about in Perameles as in others by
the junction of a circumscribed portion of the trophoblast,
the chorion, with a corresponding area of preplacentally
proliferated maternal tissue, the trophospongia.

There is a fundamental difference in the character of the
two uniting layers—the foetal beg an active, the maternal
a quite passive layer. To a mobile, virile formation of the
former type—of foetal origm—the general term plasmo-
dium is applied, while the corresponding muitinucleate
structure, of maternal origin, usually acting as a pabulum
for the foetal plasmodium, is known as a syncytium (see
Schoenfeld, 1903).

NO. 265 L

146 T. THOMSON FLYNN

It is evident enough that in all cases of placentation where
the uterine epithelium is not immediately destroyed, the area
of fixation consists for a longer or shorter time of a conjoint
layer of foetal and maternal epithelia, each having the charae-
teristics outlined above. Development of the former and
degeneration of the latter proceed in Perameles side by
side, and it would be convenient if a single expression could
be coined to denote the composite layer consisting of the
two. No convenient term seems to exist, and I propose to
use the name ‘ diploplasma’ to indicate the conjoint layer
consisting of foetal chorionic ectoderm and maternal tropho-
spongia.

The diploplasma consists in Perameles of three zones.
Along the line of junction of foetal and maternal tissues, the
syncytium is undergoing degeneration and resorption by the
plasmodium. Such a degenerating syncytium is called a
symplasma (Schoenfeld, 1903), a term which can be correctly
appled only to maternal structures of a degenerate nature
contaimed in the plasmodium. Recently, however, Willey
has suggested its use to indicate the junctional portion where
there is an intimate mixture of active foetal elements and
degenerating maternal material.

Accepting this suggestion (Willey, 1914), the three zones of
the diploplasma in Perameles consist of the following :
a middle junctional layer composed of mixed foeto-maternal
tissue (symplasma) with, on one side, a pure layer of foetal,
and on the other side, pure maternal material.

The foetal portion differentiates early into a basal layer,
the cytoblast, and a plasmodial layer, the plasmodiblast.
Contrary to what has beenstated by Willey for other mammals,
the cytoblastic layer in Perameles is well in evidence before
the time of attachment of the allantois.

The plasmodiblast has a twofold duty concerned with
(a) attachment, (b) nutrition. Both functions are performed
with the aid of root-like pseudopodial processes which attack
the maternal elements converting them into symplasmatic
débris which is ingested. Such nutritional material is passed

PLACENTA IN PERAMELES 147

on to the cytoblast, the cells of which exercise a secretory
and selective role.

The cytoblast appears to have as one of its duties the ampli-
fication of the area of attachment. Cell outlines are always
distinct between the cytoblastic cells due to radial divisions.
These result in the formation of new cells and consequent
increase of the area of attachment. As a result, new marginal
zones of plasmodial formation are brought into being, and these
attack fresh sources of nutrient material in the maternal
trophospongia. ‘This areal increase is subject, of course, to
certain limitations, and in Perameles, as will be seen
later, the growth in thickness of the plasmodial formation is
determined also by the thickness of the proliferated maternal
epithelium. In many mammals, however, as is well known, the
plasmodiblast extends much further, even in some cases as far
as the muscularis, causing the degeneration and disappearance
of practically all the structures in its track.

A second function of the cytoblast cells is, as indicated
above, the elaboration of a secretion which is passed into the
extra-embryonal coelome. ‘They therefore exercise a certain
selective capacity on the material passed to them by the
plasmodium.

Abundant evidence of this secretory activity is to be seen in
sections, the actual secretion being easily observed. From the
above it will be seen that structurally and physiologically the
proliferating trophoblast of Perameles over the allantoic
placental area is quite comparable with the corresponding layer
in the Kutheria.

Stage 2. Perameles gunni, 6-6 mm.

The specimen on which the following description is based
was brought to me by a trapper on September 17, 1919, about
two hours after being trapped. Both uteri were swollen and
found to be pregnant. The left uterus after being opened was
fixed in Hill’s fluid, and the whole uterus with embryo in
situ was later sectioned.

The right uterus was examined in salt solution. It was

L2

148 T. THOMSON FLYNN

carefully slit open along the ventral side without any injury
to the foetal membranes. It was noted that although there
was a close apposition of the yolk-sac wall to that of the
uterus, the folds of the former fitting into the hollows of the
latter in a very intimate way, yet there was absolutely no sign
of fusion or organic connexion. It will be remembered that
Hill, at first of the opinion that there was a protoplasmic
connexion between the yolk-sac wall and the uterine syncytium,
found later that this was not the case.

In the case of P. gunni it was possible by means of careful
manipulation to make out the details of the yolk-sac circulation.

This, at this stage, is in a state of considerable activity,
a condition no doubt to be correlated with the fact that the
allantoic placental connexion and circulation is now just on
the way towards completion and definite establishment. By
removing the lower portion of the bilaminar omphalopleure
and reflecting the remainder over the embryo a full view of
the vascular area was obtained, and the whole course of the
vessels could be made out in detail. After these had been
sketched and photographed the lower portion of the yolk-sae
was removed, whereupon it was found that the connexion
between allantoic vesicle and uterus was so slight that the
whole embryo with its allantois could be removed intact. The
point of attachment was visible in the case of this embryo as
a somewhat central raised area on the placental surface of the
allantoic vesicle. The elevation of this area was due to the
adherence of a slight amount of maternal tissue brought from
the uterine wall.

This being the only pregnant uterus of P. gunni so far
known, I may be excused from entering with some detail mto
the description of the relation of the embryonic and maternal
structures.

General Relations of Embryo and Uterus.
Foetal Membranes.—These agree essentially in their
arrangement with what Hill has described for other species of
Perameles. For that reason I will concern myself in the

PLACENTA IN PERAMELES 149

following account only with those features which seem to be
peculiar to the Tasmanian form.

The Allantois.—This agrees in all respects with that of
other species.

In the vesicle the distinction between the outer or placental
wall of the allantoic vesicle and the inner or coelomic wall is
obvious. A feature worthy of note is the somewhat dense look
of the allantoic vesicle in surface view. ‘This is due to the thick-
ness of its walls, the mesodermal layer of which is much denser
and thicker than in the species examined by Hill.

The Vascular Omphalopleure.—The vascular om-
phalopleure (fig. 23, v. omph.) is essentially similar in structure to
that of other species. It is, however, as well to point out that
the tenuity of the ectodermal layer of this area, on which Hill
laid so much stress, is not so pronounced a feature in P. gunni
as in the forms examined by him. It is, however, still much
thinner thanin Didelphys (Selenka, 1886-7). In P. gunni
the ectoderm cells are somewhat columnar with ovalish nuclei,
basally situated. The free end of the cell-body is frayed out into
pseudopodia-like processes. In the vascular omphalopleure
is contained a large proportion of the extra-embryonal vascular
system which in the stage under discussion forms a vascular
absorptive organ of some complexity and size. No marsupial
so far described appears to have a yolk-sac circulation of
such an elaborate type as is characteristic of this stage of
Poecunni,

Bilaminar Omphalopleure (fig. 23, bal. omph.).—This
of course consists of two layers, trophoblastic ectoderm and
yolk-sac entoderm. The histological details of a portion of
this wall are shown in fig. 22. It will be seen that the ecto-
dermal layer consists of cells which are large and are remarkable
for the immense amount of vacuolation which occurs in their
cytoplasm.

These vacuoles are of all sizes occupying in the aggregate
almost the whole of the interior of each cell, separated from one
another by thin bridges of protoplasm in a special condensa-
tion of which the nucleus is contained. The outer ends of the

150 1, THOMSON FLYNN

cells are often rounded, but in other cases possess an irregular
profile suggesting the presence of pseudopodial processes.

No doubt the vacuolation of the ectodermal cells of this
region of the blastocyst is to be associated with an active
absorption on the part of these cells of carbohydrate (? glycogen)
in process of being transferred to the embryo per medium of
the yolk-sac. The cells closely resemble the glycogenic cells
figured by authors for certain other mammals (e. g. Jenkinson,
1902, 1. vi).

The entoderm of this region consists of a layer of cells
somewhat darkly staining with haematoxylin. ‘These are
sometimes rounded, sometimes flattened. I have found the
entoderm cells occasionally shghtly vacuolated in the manner
described by Hill for his species.

It is a well-known fact that the vacuolation of cells con-
cerned in embryotrophic processes varies greatly at different
stages. Hill found that the entoderm cells of the yolk-sac, but
slightly vacuolated in early stages, became greatly so im his
12-5 mm. specimen. This he regarded as a forerunner of
degeneration, but I rather think it is to be associated with the
active absorption and internal transmission by the yolk-sae
wall of substances (probably carbohydrate) secreted by the
mother and destined for the nutrition of the foetus. ‘

The Yolk-sac Placenta.—As stated above, this is
brought about by the intimate apposition of the vascular
omphalopleure to the extra-placental portion of the uterme
wall. Fig. 28 shows the details of this circulation and fig. 24
is a somewhat diagrammatic representation of the area with
regard to the blastocyst wall and the embryo.

In fig. 23 most of the bilamimar omphalopleure has been
removed and the remainder of the blastocyst wall has been
reflected over the embryo. The vascular area is therefore seen
from its inferior aspect. ‘The anterior end of the embryo is
contained in the proamnion (proa.), here large and persistent.

The vitelline artery (vit.a.) is a very large and thick trunk
which, after leaving the yolk-stalk, keeps at first a little to the
left side of the body and passes over the surface of the tail to

PLACENTA IN PERAMELES 151

reach the right side. Here it passes into the outer leaf, 1. e.
from the yolk-sac splanchnopleure to the vascular omphalo-
pleure. In the course of its passage along the body it gives
rise to the numerous fine and extremely characteristic vessels
which extend in their peculiarly parallel manner into the
yolk-sac splanchnopleure which they supply. ‘These branches
are extremely long and pass for some little distance over into
the vascular omphalopleure where they alternate with corre-
sponding venous factors to the vitelline vein.

Immediately on entering the vascular omphalopleure the
vitelline artery divides to form the sinus terminalis (s.t.).
Each portion of the simus passes forward and at first has the
usual course. Instead, however, of passing directly forward
in the usual way, each branch, at about the level of the fore-
limb, takes a sudden turn ventrally so that the anterior portion
of each lateral branch forms another curve with a ventral
convexity. The artery on each side unites with its fellow in
front of the head by anastomosis.

The smus gives off to the area vasculosa a wonderfully rich
plexus of branches quite as complex or even better developed
than in any marsupial so far described.

The pecuhar arrangement of the sinus terminalis results
in the division of the vascular area on each side into two regions,
anterior and posterior, the former being somewhat smaller in
area than the latter. Hach of these areas is drained by a
separate factor of the vitellime vei on each side. The posterior
factor from the posterior area is the larger, and after receiving
numerous fine capillaries and branches travels along the
dorsolateral aspect of the vascular omphalopleure, parallel to
and some little distance from its dorsal edge till the vein
reaches the height of the neck flexure of the embryo, where it
passes over into the yolk splanchnopleure. Here it receives the
anterior factor which drains the anterior area. The branches
and capillaries which go to make up the latter are extremely
rich. Each lateral vitelline vem (vit.v.) formed of the factors
just mentioned passes down and unites with its fellow just
before entering the body. Between them in this region is the

152 T. THOMSON FLYNN

permanently non-vascular area spoken of by Semon and by
Hill.

It should be mentioned that the posterior factor of each
lateral vitelline vein receives on its inner side branches from
the yolk-sac splanchnopleure corresponding to the fine branches
from the vitelline artery supplying that membrane. These pass
upwards in the splanchnopleure, and then continue outwards
into the omphalopleure to jom the posterior factor as stated.

The measurements of the vascular area are as follows :

Across anterior portion. , : . ema
Across posterior portion . : : . 9-9mm.
Greatest length, anterior portion : . 4mm.
Greatest length, posterior portion. . 5mm.

Formation of the Allantoic Placenta in
Perameles gunni.

Various stages of this are shown in figs. 13-16, 19.

Maternal Structures.—The wall of the uterus is
divisible, as stated above, into two general portions, placental
and extra-placental regions.

Placental Region.—The mucosa varies greatly in thick-
ness, from 0-60 to 1-1 mm., due to folds in the uterine wall.
The glands are numerous, long, and tortuous, measuring in
diameter from 0-037 to 0-051 mm. They are of the usual type,
being unbranched and lined by a single layer of columnar cells,
with deeply-staining nuclei, peripherally situated. I am not
able to detect any trace of the cilia which Hill and O’ Donoghue
have observed in Perameles. ‘The glands are in a highly
active state of secretion, their basal portions being filled with
cellular and other material.

The inter-glandular tissue is extremely thin and tenuous,
but condensed where it immediately surrounds glands and
blood-vessels. Distributed through-the connective tissue is an
abundance of lymph material. This is particularly evident
just below the uterine epithelium, in which position there is
a space up to 2mm. wide filled with lymph which bathes the
lower surface of the uterine syncytium. Here and there this

~

PLACENTA IN PERAMELES 153

space is crossed by glands and blood-vessels. I have called this
space the sub-epithelial lymph-space. Owing to the presence
of this coagulum and the sparsity of cellular elements the
whole stroma presents a very homogeneous appearance.
Scattered through it, however, are numbers of leucocytes,
and here and there are erythrocytes which have escaped from
the capillaries by extravasation.

Attached to and intimately united with the epithelium of
this area is the embryonic chorion.

Epithelium.—the condition of the epithelial elements of
the allantoic placental mucosa shows, unquestionably, that in
P. gunni the same process of syncytialization has occurred
as in other species, and that here is a similar aggregation of
the syncytial nuclei into nests arranged in lobular masses of
the syncytial protoplasm. The distribution, as I have sug-
gested above, I believe to represent an abortive attempt at
the formation of simple uterine crypts. Between these nests
the capillaries and leucocytes of the remainder of the mucosa
gain access to the syneytimn.

The Extra-Placental Region.

Mucosa.—In all essential respects this resembles the
mucosa of the placental area, bemg highly vascular and con-
taining a large quantity of lymph. Here are to be found
leucocytes and red corpuscles, but neither are so numerous as
in the placental portion. The glands are in an active secreting
phase and their contents—cellular detritus and other nutritive
substances—are poured into the uterine cavity to be received
into the trophoblastic layer of either region of the omphalo-
pleure.

The sub-epithelial lymph-space, as such, becomes less
evident the farther we proceed from the placental area and
the glands are somewhat smaller in diameter.

Syneytium.—this has been formed by a similar process
to that which has resulted in the modification of the epithelium
of the placental area. The difference is mainly one of degree.
Thus the extra-placental syncytium is thinner, being about

154 T. THOMSON FLYNN

0-043 mm. opposite the vascular omphalopleure and 0-023 mm.
opposite the bilamimar. Similar proliferation and migration
of the nuclei have taken place but the lobular nests are much
smaller and less individual, and, further, more and more
nuclei remain outside the nests until in some places lobules
and nests as such are scarcely distinguishable. The nuclei are
quite similar to the syncytial nuclei of the placental area bemg
rounded, bead-like, and vesicular with little chromatin.

Chorion.—this shows, both in the amount of area of
attachment and of proliferation, a considerable advancement
on the condition found in the last stage. The fusion with
maternal tissue is so complete that the resulting layer is one
in which occasionally there is some difficulty im distinguishing
maternal and foetal cytoplasm. Here, as before however,
we have the infallible test of the difference im structure, shape,
and fate of the maternal and foetal nuclei.

The Diploplasma.—The thickness of this layer measures
now from 0-135 to 0-190 mm.

This increase is due mostly to the great growth of the
chorionic ectoderm, which consists of a well-marked basal
cellular layer, the cytoblast, and a much thicker portion im
which cell outlines are not visible, the plasmodiblast. The
maternal portion of the diploplasma consists of the remains
of both the maternal syncytial protoplasm and of the syncytial
nuclei. When unaltered the latter have the characteristics
which have been before noted, viz. small rounded vesicles in
which the chromatin is suppressed but containing a more or
less prominent nucleolus (see figs. 13, 14, and 19, syn.mn.).
They have an exactly similar appearance to the nuclei of the
extra-placental area.

At this stage, however, nests are seldom found intact. In
almost every case they have been invaded by plasmodiblast
nuclei (pb.n.), so that by this time a fair proportion of maternal
nuclei have degenerated and disappeared.

The Trophoderm.—this is divisible as stated above into
cytoblast and plasmodiblast.

The Cytoblast (cyt.) is a remarkably definite layer of

PLACENTA IN PERAMELES 155

cells with distinct cell outlines. The cytoplasm is granular
and the nuclei large and plump. Their chromatin contents
are in the form of an extremely rich network with a number
of karyosomes. A nucleolus may be present, and if so is of large
size. At the edge of the placental area the cytoblast (cyt.)
passes over into the ectoderm of the marginal chorion, whose
cells gradually decrease in size till they attain, at the junction
with the vascular omphalopleure, the normal dimensions and
appearance of trophoblastic ectoderm cells. Some idea of the
difference in size of the cells of this layer may be obtained from
the fact that, at the margin of attachment, the cells are twice,
while in the centre of the area of attachment they are eight
times the height of the ordinary trophoblastic ectoderm cell.

The cytoblast cells, therefore, have already the appearance,
particularly in the centre of the placental area, of typical
placental megalokaryocytes.

Intense proliferation is to be seen in the cells of the cytoblast
of the whole of this area. I have not been able to observe
mitotic figures, but it must be remembered that in the case
of my material as well as that of Hull, fixation was not possible
until a couple of hours after death.

Proliferation of the cytoblast (cyt.) and its results are shown in
figs. 18416, 19. The plasmodiblast nuclei (pb.n.) resulting from
this division are usually more darkly stained than the original
cytoblastic nuclei and for this reason are easily distinguishable,
particularly from the original syncytial nuclei (syn.n.). The
plasmodiblast nuclei are occasionally isolated, but more often
are arranged in clumps to form multinucleated masses (g.c.)
of sometimes large size. These resemble the giant or monster
cells which have been found in similar positions in connexion
with the placentation phenomena of higher HEutheria.
The main result is that the plasmodiblast makes its way mto
the nests of syncytial nuclei, in which one or two or, in most
cases, a large number of these nuclei can be distinguished.
In such cases the syncytial nuclei have to a great extent dis-
appeared, the tendency being to replace them in their nests by
the newly-arrived foetal nuclei.

156 T. THOMSON FLYNN

The pseudopodial processes characteristic of the plasmodium
of the previous stage are greatly in evidence in this stage also.
The result is that spaces occur in the diploplasma giving it a very
uneven appearance. ‘The arrangement of these plasmodial
spaces suggests that new nuclei are formed in a somewhat
spasmodic manner in waves one after the other.

One result of the advance of the plasmodium is the inelu-
sion of the maternal capillaries which pass up
between the remains of the syncytial lobules
and ramify below the layer of cytoblast. Hach
capillary is contained in its delicate endothelial layer. Further
than this in the symplasmatic zone of the diploplasma are to
be found remains of maternal nuclei, maternal blood corpuscles
and various granules. These are all obtained phagocytically.
An evidence of this action in the plasmodiblast is the strong
affinity which some of its nuclei have for pigment. In this
stage much of the pigment found in other cells of the previous
stage has disappeared. It is now almost entirely confined to
the cytoblast and plasmodiblast. Here it appears usually as
black granules deposited round the nucleus, but in other cases as
narrow irregular lines in the cytoplasm (figs. 25-27). A similar
pigment (related to haematoporphyrin ?) is found in Ungulates,
and is regarded by Jenkinson and Duval as being the remains
of ingested haematids whose iron has been passed on to the
embryo. The result is that the pigment remains in the tropho-
blast increasing in amount up to the time of parturition.

Leucocytes.—These are extremely characteristic of the
uterine mucosa, particularly in the placental regions. They are
found in the connective tissue, from which they migrate through
the epithelium of the glands and mingle with their secretion.
They also pass into the diploplasma, from which they reach the
extra-embryonal coelome. ‘They can be seen abundantly
present in all sections of this region at this stage. In the
connective tissue they are small, taking the form of small
mononuclear leucocytes. In the diploplasma, however, and
particularly when they reach the cytoblast, they have increased
greatly in size, forming large mononuclear leucocytes or macro-

PLACENTA IN PERAMELES 157

cytes. In this position they occasionally displace cytoblast
cells. All the leucocytes carry well-defined pigment. Isolated
patches of pigment occur occasionally in the plasmodium.

The coagulable material (proteid, lymph) passing through
the epithelium of the uterus is particularly abundant in the
placental area. The secretion of the cytoblast cells is visible
on their coelomic faces as small rounded swellings which break
off to form spherical bodies floatmmg in the coelomic fluid.

I can find no evidence of fat secretion either in the cytoblast
cells or in the gland cells, but it must be remembered that no
special means have been employed for the detection of fat,
glycogen, or iron.

The chorionic mesoderm, where it can be made out, is a thin
mesothelial layer consisting of flattened cells with ovalish
darkly-staining nuclei. I must confess that I find it impossible
to discern this layer over most of the placental area.

The free margin of chorion forms the connecting link between
the fixed chorion on the one hand and the vascular omphalo-
pleure on the other. The mesoderm here is similar to that of
the central portion of the chorion, but the trophoblastic layer
consists of somewhat flattened cells near the vascular omphalo-
pleure, these increasing greatly in height as they approach the
fixed portion.

Gland Alteration.—A word about the condition of the
glands in this stage. The body of the gland is lmed by an
epithelium which is somewhat lower than in the preceding
stage. In its upper portion the gland swells out and becomes
somewhat barrel-shaped. The whole of this portion enclosed
in the diploplasma suffers a degeneration of its epithelium.
The result is that here the gland appears to be a mere space
lined only by plasmodiblast. The gland-mouth is closed by
a layer consisting of the cytoblast plus a certain amount of
plasmodiblast. The latter by no means takes the form of
a plug as in the last stage, but passes down on either side
apparently causing the disorganization of the gland epithelium
on its way. There appears to be a struggle here between the
downward force of the growing plasmodiblast and the upward

158 T. THOMSON FLYNN

pressure of the gland secretion, evidenced by the swelling of
the gland and the formation of lateral fissures in the plas-
modium. Such appearances are characteristic of all the glands
in this stage.

When such glands are sectioned obliquely or transversely
they appear in sections of the diploplasma as irregular spaces
without any definite epithelial bounding layer.

Attachment of Allantois.—At this stage this has
occurred over an approximate area of 0-5 by 0-21mm. The
attachment is due to an intimate fusion of the splanchnic
mesoderm of the allantois with the somatic mesoderm of the
chorion. Apparently the allantois spreads its area of attach-
ment very rapidly, and one of its almost immediate results is
a quickening of the proliferating activity of the cytoblast cells.

In the centre of the area of attachment the distinction
between the plasmodial layer and the cytoblast to a great
extent breaks down, the cell outlines of the latter disappearing
and its nuclei being converted into plasmodiblast nuclei. It is
in these positions that apparently the union of the allantois
with the chorion is best effected. This dissolution of the
cytoblast allows maternal capillaries to come closer to the
surface of the diploplasma, and this can be seen in some cases
even before the allanto-chorionic fusion is in beimg (fig. 14).

Stage 3, Perameles obesula, 7 mm.

This is Hill’s Stage C, being his earliest placental stage. It
is represented in the Department of Zoology of the University
of Sydney by a single slide containing one section. The
specimen, however, is well stained and mounted, and shows
all the details of the relations of the foetal membranes to the
uterie wall.

The excellent general account given by Hill of the foetal
membranes of this stage renders any further description un-
necessary. I will therefore confine myself to a consideration
of the maternal and foetal structures associated with the
allantoic placentation.

The attachment of the allantois may now be said to be com-

PLACENTA IN PERAMELES 159

plete, the allanto-chorion being united with the maternal
epithelium over an area whose diameter, following the well-
marked uterine folds, is some 15-14 mm.

The area of attachment is not co-extensive with the whole of
the chorion since there is a marginal zone of the latter quite
free, as in the 6-6 mm. stage.

Structural Details of the Allantoic Placenta.

As would have been expected from the description of the last
stage there is by this time a most intimate fusion of the maternal
and foetal tissues in the diploplasma, the proliferation of the
trophoblastic ectoderm and its invasion of the maternal
syneytium having advanced considerably beyond the condi-
tion found in the 6-6 mm. embryo of P. gunni.

Here again, therefore, in the trophoblastic proliferation there
can be recognized two portions, a basal cellular layer corre-
sponding to the cytoblast or cytotrophoblast of higher mam-
mals, and externally to this a proliferating plasmodial portion,
the plasmodiblast or plasmoditrophoblast.

The cytoblast is throughout its extent fairly distinct. Over
a fair area towards the centre, however, it has already begun
to lose its integrity—having disappeared in many places as
a cellular layer and become converted here ito plasmodiblast
(fig. 18). It was upon this condition of the eytoblast in its
central portion that Hill based his suggestion of the degenera-
tion of the chorionic ectoderm.

In the marginal portion (fig. 21) the cytoblast still consists of
high columnar cells separated from one another by distinct
cell walls. Externally, however, the boundaries of these on
their plasmodial aspect are indistinct. Where intact, they
have somewhat elongated nuclei arranged at right angles to
the surface of the uterus. These nuclei stain well and are rich
in chromatin. They do not, however, stain so darkly as the
more granular plasmodial nuclei to which they give origin.

At the edge of the placental area the development of plas-
modiblast from cytoblast is in its minimum condition of activity,
as is shown in the figure (fig. 21).

160 T. THOMSON FLYNN

Here it can be seen that the arrangement of the syncytial
nuclei of the extra-placental area is similar to that of the
preceding stages. ‘I'he presence of an internal space, free of
nuclei, in the syncytial nest should be noted.

Passing from this region into the placental area the gradually
increasing activity of the proliferating chorionic ectoderm
becomes evident. More and more of the foetal nuclei occupy
the nests. The plasmodial nuclei in a large number of cases
form multinucleate groups. In many cases there are but two
nuclei in each group, these being comparable to the binucleate
cells ‘ diplokaryocytes ’ described for some EHutheria (for
example in Ungulates, Assheton, 1906). More often
three or more nuclei are contained in one plasmodial mass.

An examination of a more central section shows that in this
portion of the placental area many nests are now quite filled
with the plasmodiblast nuclei (fig. 18). So closely are these
packed in the nests that instead of being oval or irregularly
shaped they take a polygonal form due to mutual pressure.
The condition of the original epithelial nuclear nests can be
gauged from the fact that of twenty-seven observed in one
central portion of the allantoic area four were untouched ;
eleven were partly, twelve completely filled by invading
nuclei, the nests having the appearance of solid multinucleated
masses. j

In such positions also it is that the original cytoblast layer
has practically disappeared. It is evident, therefore, that
a large proportion of the maternal epithelial tissue has been
replaced by intruding foetal material. One effect of this is that
the maternal capillaries of the allantoic area become enclosed
by the advancing plasmodium, and in many cases are now to be
found at the surface of the diploplasma, where they directly
underlie the allantois and even come into contact with the
allantoic vessels.

Pigment is not so common as it is in the previous stage but
it is still to be found, particularly in leucocytes and in the
cytoblast cells. Here and there in the plasmodium are to be
seen isolated patches of pigment pointing to an active ingestion

PLACENTA IN PERAMELES 161

of maternal cellular material. Haematids are often to be found
contained in the plasmodium in process of being absorbed.

Glands.—Most of the glands of the allantoic placental
area have undergone certain characteristic changes. Instead
of being narrow and tortuous, they have now become generally
straighter and wider.

The necks of the glands are particularly spacious, and often
the gland on its entrance into the diploplasma shows a barrel-
like enlargement. The aperture of the gland in the placental
area is closed by the cytoblast, while the plasmodiblast has
disappeared from its mouth but bounds this portion of the gland
on either side where it has been responsible for the disorganiza-
tion of the gland epithelial cells (fig. 28). The cells of the body and
neck of the glands have now been transformed into a low cubical
epithelium, and their nuclei instead of being oval are rounded.
The gland-cells throughout are ciliated. Occasionally some of
the glands swell to a relatively enormous size. A further
feature of note is that a number of glands have penetrated
into the muscularis.

The Allantois.—For the general description of this
I would refer the reader to Hill’s account. At this stage the
splanchnic mesoderm of the allantois is found fused with that
of the chorion over the full extent of the placental area.
There is but a small amount of penetration of allantoic tissue
into the plasmodiblast, and this only becomes possible where
gaps have occurred in the disorganized cytoblast. The complex
‘interlocking system ’—by which apt phrase Hill describes
the mutual apposition of foetal and maternal capillaries—is
brought about by short finger-like downgrowths of the allan-
toic mesoderm combined with the opposite (inward) tendency
of the maternal capillaries. This mutual process becomes more
and more easy the more the basal cytoblasti¢ layer loses its
integrity and becomes converted into plasmodiblast. Where the
cytoblast remains practically intact as at the margins (fig. 28),
the approximation of foetal and maternal vessels does not
occur. ‘The maximum penetration of the allantois appears
to be a little more than the thickness of the cytoblast. One

NO. 265 M

162 T. THOMSON FLYNN

result of this penetration is the tendency to form occasional —
cytoblastic islands, containing usually one or two nuclei,
isolated by the growth round and behind them of the allantoic
capillaries (fig. 24).

Stage 4. Perameles obesula, 8-8-75 mm.

Of this stage I have no material available, so will content
myself by stating that Hill’s figures (1897, figs. 15 to 21) show
that the various processes of allantoic placental formation
which have been initiated in the case of the younger embryos
can be recognized as being continued in this.

The growth of the plasmodiblast has gone on apace, with
a corresponding diminution of the amount of maternal tissue
contained in the diploplasma. This results in a very homo-
geneous appearance of the tissue of the placental area. In
Hill's figures very few syncytial nuclei are recognizable with
certainty. On the other hand, it can be seen unmistakably
that, by this time, most of the cytoblast has been converted
into plasmodiblast, and that in this formation giant multi-
nucleate cells are a very prominent and characteristic feature
(see especially his fig. 17).

Here again the greatest activity is being shown towards
the centre of the placental area. Of the still remaining basal
cytoblast cells Hill says (p. 414), “im some cases they are multi-
nucleated . . . or the single nucleus is also hypertrophied and
vesicular,’ a statement which well accords with the facts to
which I have already drawn attention in the preceding pages.

Many of these remaining cytoblast cells have the appearance
of diplokaryocytes. Some of them plainly show a tendency
towards plasmodial formation, as can be seen in Hill's fig. 21,
where the cell marked ch.ect. has very much the appearance
of a giant cell with plasmodial processes.

Stage 5. Perameles obesula, 12-5 mm.
Of this stage I possess five consecutive sections stained in
haematoxylin and eosin.
Here, as before, I will confine myself to a description of the

PLACENTA IN PERAMELES 1638

structural modifications associated with the formation of the
allantoic placenta.

A section through the very folded allantoic placental area
reveals the presence of an extremely homogeneous granular
layer with lobed lower margins, indicating the positions of the
original syncytial nests.

Unfortunately, the gap between this and the preceding stage
is too wide to make it possible to closely follow and be certain
of the histolytic changes which have taken place in the placental
region.

Within the lobulated portions of the diploplasma are very
irregular groups of nuclei corresponding in position to the
groups of the original syncytium.

An examination of such an aggregation at this stage shows
that it consists mainly of plasmodiblast nuclei which are in
a marked stage of degeneration. One simple group is shown in
fig. 26 of this paper, while many others are depicted by Hill
(1899, Pl. xlin, figs. 6 and 7).

One particular type of degeneration change in the case of
a plasmodial nucleus is shown in fig. 17. This corresponds
closely with what has been recorded by Jenkinson (1902,
figs. 24-5) for the megalokaryocytes of the mouse. The
origial robust nucleus loses its contour through shrivelling
of the nuclear membrane and escape of the nuclear sap. The
chromatin becomes irregularly arranged. Thereupon the whole
nucleus flattens and becomes of the nature of an extremely thin
rod, which by further absorption is seen as a few darkly staining
particles in the general ground-mass, finally disappearing.

There are to be found also degenerating vesicular elements
which no doubt are the remnants of the formerly numerous
maternal epithelial nuclei. On this important point, in view
of the paucity of the material available, I regret I am not able
to make any certain statement.

If, however, this interpretation be correct, then Hubrecht’s
suggestion (1909) that the placenta of Perameles consists
in its final stage of a ‘ mixed syncytium’ is not very far from
the truth.

M 2

164 T. THOMSON FLYNN

It is difficult to suggest a cause for the degeneration and
resorption which has taken place in the case of these nuclei,
but it is a significant fact that associated with the degenerating
groups are to be found abundant leucocytes of the small and
large mononucleate types.

Situated just internally to the allantois are to be found
occasional large trophoblastic cells, bemg remains of the
cytoblast. These also are undergoing degeneration.

In places the homogeneous nature of the ground-tissue is
less evident, and here multinucleated masses of protoplasm are
still in evidence. Apparently the foetal chorionic cells have
performed their function, viz. the fixation and a portion of
the preliminary nutrition of the embryo, and are now in
a process of degeneration and disappearance.

The maternal vessels are extremely numerous, and their
finer branches now ramify at the surface of the diploplasma,
where they come into intimate apposition with the vessels of
the allantois in the way Hill describes.

The allantoic mesoderm has penetrated but little mto the
trophodermic layer. No spaces are formed in which maternal
blood flows, all maternal vessels being contained in definite
endothelial walls. The two blood-systems, however, are
elaborately interlocked, and very often foetal and maternal
blood-streams are separated merely by the thickness of two
endothelial walls.

As regards the glands of the placental area, they are now
characterized by the possession of a low cubical epithelium.
Their cavities are wide, particularly near the opening of the
gland. The cells are ciliated. The mouth of each gland is
sometimes closed by the allantois with its numerous vessels,
at other times by degenerating remains of the trophoblast.
The condition of the portion of the gland contained within the
diploplasma is similar to that characteristic of the preceding
stage.

Outside the placental area the maternal nuclei of the epithe-
lium still take the form of rounded vesicles in which it is difficult
to distinguish even a nucleolus. The glands of this portion of

PLACENTA IN PERAMELES ' 165

the uterine wall possess the usual columnar epithelium and
are also ciliated. They are narrower than the glands opening
into the placental area.

Seeing that the newly-born young of Perameles measures
but 14 mm. in length, the condition of the allantoic placenta
in the 12-5 mm. stage may be accepted as being practically
that of the full-term placenta.

Stage 6. Perameles nasuta, post-partum.
With regard to this stage the only way in which I can
supplement Hill’s description is by pointing out that the
amount of foetal tissue left behind in the uterus is considerably
more than would be the case if Hill’s conception of the placenta-
tion of this animal were correct. Instead of consisting merely
of the allantois—with the addition of a few remaining foetal
cells left behind after the degeneration of the chorionic ecto-
derm—there is really comprised in the contra-deciduate
portion, in addition to the allantois, the whole thickness of the

diploplasma, of which undoubtedly the greater part is foetal.

6. SUMMARY OF CONCLUSIONS.

The conclusions arrived at in the preceding pages may be
summarized as follows :

1. Allantoic placenta.

(a) The fixation of the embryo is brought about by means
of the chorionic ectoderm at a time when the embryo measures
about 6 mm. direct length.

(b) The chorionic ectoderm develops directly into two
portions, (1) a basal cellular cytoblast which by proliferation
gives rise to (2) a plasmodiblast.

(c) The plasmodiblast phagocytically attacks the maternal
tissues, particularly the maternal nuclear aggregations or nests.

(d) The proliferating foetal ectoderm and maternal syncytium
are thus intimately fused to form a structure, the diplo-
plasma.

(e) Sooner or later most of the cytoblast layer disappears,
beimg converted into plasmodiblast.

be

TEXT-FIG.

mat. cap.

RUAN XX UX x x= MXN
1% Fa) x

oe. aa e

troph.

Diagrams showing the development of the allantoic placenta in Perameles.

(a2) Shows the character of the preplacental maternal trophospongia with
its lobules, between which vessels and glands penetrate. The capillaries
ramify at the surface.

(b) Chorionic attachment has now taken place and the chorion has proli-

PLACENTA IN PERAMELES 167

(f) The outgrowth of foetal plasmodium does not extend
any farther than to involve the proliferated maternal epithelium.

(g) The attachment of the allantois is effected when the
embryo has attained a length of approximately 6-5 mm.

(h) The outward migration of the basal cytoblast cells (when
converted into plasmodiblast) gives opportunity for the
maternal and foetal vessels to come into intimate apposition.

(i) In the final stage the foetal nuclei in the placenta are
found to be in a state of degeneration.

(j) Remains of the maternal epithelium still probably exist
in the full-term placenta.

(k) The uterine glands persist throughout gestation, but the
portion of their epithelium within the diploplasma disappears.

(2) All maternal vessels have definite endothelial walls :
hypertrophy of the endothelial cells does not occur and lacunae
are not formed.

(m) An allantoic placenta is recorded for P. gunnt.

2. Yolk-sae Placenta.—A virtual yolk-sac placenta is
present in P. gunni as in other species of Perameles,
brought about by the intimate apposition of the complex
system of vessels in the vascular omphalopleure with the highly
vascular portion of the uterine syncytium just beyond the
placental area.

ferated to form a basal cytoblast and an external plasmodiblast. The
proliferations of the latter are related to glands and nuclear nests. The
capillaries ramify just below the cytoblast layer.

(c) Allantoic attachment is now in its first stages. Further proliferation of
the foetal trophoblast has taken place and the plasmodiblast now occupies
the greater part of each nest. Within the diploplasma the gland
epithelium has disappeared and the gland is closed in above by cytoblast
plus plasmodiblast. Capillaries still ramify below the cytoblast.

(d) The foetal plasmodium now occupies the whole of each nest with the
exception of (possibly) a few remains of maternal tissue. The cytoblast
layer has broken down, having mostly been transformed into plasmo-
diblast. A few islands of trophoblastic tissue are, however, still left
in the cytoblast position. The very much branched and interlocked
foetal and maternal! vessels have now come into intimate apposition.
The gland epithelium in the diploplasma has entirely disappeared, and
the glands are closed in above by all antoic vessels alone.

all., allantois; all. cap., allantoic capillary ; cyt., cytoblast; gl., gland ;
mat.cap., maternal capillary ; pb., plasmodiblast ; troph., trophospongia.

168 '. THOMSON FLYNN

Further, considerable vacuolation of the cells of the bilaminar
omphalopleure at certain stages mark it as being an absorptive
organ of some importance in embryotrophic processes.

7. ConctupING REMARKS.

Now that we have some clear idea of placentation in Pera-
meles, the question arises what relation exists between this
and the placentation of other marsupials and of Eutheria
generally ? Further, can any light be thrown on the phylogeny
of the mammalian allanto-placenta by a consideration of this
question ?

Before, however, entering on a discussion of these, it 1s,
I think, necessary to get a clear idea of the morpho-physio-
logical conception of the term ‘Placenta’.

(a) The Placental Conception.

It is inevitable that the conception of placentation which
has arisen in the minds of investigators should, until fairly
recently, have been associated with the very complex and
highly advanced structure developed in the intimate fusion or
apposition of the foetal membranes of the commoner and best-
known Eutheria with the uterine mucosa. ’

It might, however, be taken for granted that—in the case
of an organ so prominent in mammalian developmental pro-
cesses and rightly regarded ontogenetically and phylogeneti-
cally as of the highest importance—there would be no two
opinions as to its definitive structure or its physiological
significance. But such a supposition would be wrong, and
a very superficial examination of the writings of the more
recent investigators soon shows that they hold radically different
views as to what is understood by the term ‘ placenta’.

All the important recent works treating of the comparative
anatomy of the placenta to which I am able to refer (Strahl,
1905; Grosser, 1910; Jenkinson, 1913) are insistent that
the fundamental idea of placental formation lies in the apposing
of two blood-streams—one foetal, one maternal—to form

PLACENTA IN PERAMELES 169

a structure by which the physiological processes intended for
the well-being of the embryo can be carried out.

Such a conception makes no allowance whatever for the
work of the bilaminar omphalopleure of marsupials—itself
non-vascular but physiologically of considerable importance.

Even less acceptable is the suggestion of Professor Hubrecht
(1909), who insists that ‘fusion of embryonic with maternal
tissue is a conditio sine qua non, and so we must admit
a placenta in the case of Didelphia (Perameles) and deny it to
certain Monodelphia (Equus, Sus, Nyticebus, Galago, and
others) *. As Assheton has pointed out, under this scheme
‘the sheep is a placental, a cow a non-placental mammal ’.

Reaction between mother and embryo, or rather dependence
of the latter on the mother for food, oxygen, and the removal of
its waste products, may be said to commence from the time of
the first appearance of the ovum in the uterus. * The mam-
malian ovum, says Hill (1910, p. 113), ‘ already in the mono-
tremes greatly reduced in size as compared with that of reptiles,
and quite minute in the Metatheria and Eutheria, con-
tains within itself neither the cubic capacity nor the food
material necessary for the production of an embryo on the
ancestral reptilian limes. We accordingly find that the primary
object of the first developmental processes in the mammals
has come to be the formation of a vesicle with a complete
cellular wall capable of absorbing nutrient fluid from the
maternal uterus.’

Our knowledge of the physiology of the early stages of
mammalian intra-uterine development is admittedly as yet
very incomplete. Nevertheless, it is quite certain that the
amount of nutrient material present in the ovum is absolutely
insufficient for even the most elementary developmental pro-
cesses, and has to be supplemented, from the very beginning,
from outside sources.

Even a cursory consideration of the above will serve to
indicate that in the uterine development of the viviparous
mammals there occur two distinct phases, differing entirely
in the means by which the necessary physiological processes

170 T. THOMSON FLYNN

of the embryo are arranged for. In the first of these, extending
over the time of cleavage and of blastocyst formation, absorp-
tion and exchange are performed solely by means of the tropho-
blast.

When blood-vessels appear and are functional, quite a new
phase is inaugurated, lasting to the end of pregnancy, during
which these vessels come to the aid of the trophoblastic layer
in more quickly and efficiently performing the necessary
embryonic services.

Grosser’s (1910, p. 94) terms ‘ embryotrophic ’ and * haemo-
trophic ’ could have been conveniently employed to indicate
these periods, but, unfortunately, his use of the word ‘ exelu-
sively ’ in the definition of the latter term has made it applicable
alone to haemochorial placentae, and it is even doubtful if the
definition would be strictly correct in their case.

Nor does Resink’s (1902) arrangement suit the case any
better. This author regarded the intra-uterine life of the hedge-
hog as falling into two well-defined periods as follows :

(a) Preplacental period, during which maternal and
foetal preparation for the allantoic placentation takes place ;
the trophospongia and ectoplacenta are formed and the
embryo fixed. Broadly, this period may be said to occupy the
earlier portion of intra-uterme existence up to the time of
attachment of the allantois.

(b) Kuplacental period, in which the allantoic attach-
ment is made and the placenta completed.

The weakness of this arrangement as apphed to mammals
generally is to be found, in my opinion, in the extreme impor-
tance given to the allantoic placentation and the inclusion of
the yolk-sac (vascular) placentation in the first of these periods.
Such a scheme becomes difficult of application, particularly to
the Metatheria, in which an allantoic placenta occurs, so
far as is known, in but one genus, various methods of tropho-
blastic attachment in others, in many no attachment what-
ever, a yolk-sac placenta in all.

It is therefore apparent that allantoic placentation is only
of the greatest importance in one group of mammals, the

PLACENTA IN PERAMELES 171

Eutheria, and it is due to the fact that the most detailed
investigation has been expended on this group and to the
prominence of the allantoic placenta in it that other features
of embryonal intra-uterine life have been for so long over-
looked.

Viewed in the light of what we already know of the morpho-
logy and physiology of the foetal membranes in the two groups
of viviparous mammals, it is evident that placentation, as
generally understood, is but part of a much larger conception
which has to do with the whole physiological intimacy, durmg
intra-uterine life, between foetus and mother.

Thus I am fully in accord with Assheton’s suggestion (1909)
that the term ‘placenta’ should be applied to all organs
consisting of an intimate apposition or fusion of the foetal
membranes with the uterine wall for the purpose of carrying
out physiological processes destined for the well-being cf the
embryo.

Such a conception would include the following types of
placenta :

(a) That im which the trophoblast is vascularized from the
allantois—allantoplacenta.

(b) That in which the trophoblast is vascularized from the
yolk-sac—omphaloplacenta.

(ce) That im which no foetal blood-vessels are concerned.
This is the case of the bilaminar omphalopleure of marsupials
whether there is a fusion of part of this with the uterine mucosa
(Dasyurus, Phascolarctos), or merely, as is more usual, intimate
apposition. For this type of placenta 1 propose the term
‘metrioplacenta’. These may be illustrated by referring
to Perameles, the genus which is the subject of investiga-
tion in the present paper.

(b) Placental Phases in Perameles.

Preliminary Phase.—During this period the blasto-
eyst is formed and the physiological processes are carried on
by means of the trophoblastic cells. There is no union in

172 {. THOMSON FLYNN

Perameles between the trophoblast and the uterine
wall, and there is no absorption by means of foetal blood-
vessels.

The work of this phase is carried on in later intra-uterine
life by the bilaminar omphalopleure.

Intermediate Phase.—tThis is the stage of the vascular
yolk-sac placenta. It comes into being with the functional
formation of the vascular area. ‘here is a close apposition
between the foetal and maternal blood-vessels. This phase
reaches its most active condition before the attachment of the
allantois, and although, maybe, less efficient, endures, with the
existence of the vascular area, until the end of pregnancy.

Final Phase.—the allantoic attachment takes place and
the allantoic placenta is completed.

It is evident that in marsupials, with the exception of
Perameles, the preliminary and intermediate stages are the
more important, in fact the only ones present, while in general,
in Eutheria, the preliminary and the final phases are of
the greater value.

Under these circumstances we can denote the placental
periods in Perameles as metricoplacental, omphaloplacental,
or allantoplacental, according to the type of placenta which is
the dominant one for the period concerned.

x

(c) Placental Phenomena in Marsupials
generally.

In reviewing these I will commence with the most specialized
groups.

Macropodidae.—tThe works of Owen (1834-7, Macro-
pus major), Semon (1894, Aeprymnus rufescens),
and Hill(1895,M. parma, M.ruficollis, M.robustus,
and M. major) emphatically show that in these forms the
allantois throughout life remains small, buried in the splanchno-
coele. From my own observations I am able to state that
this is also the case for Potorous tridactylus and
Bettongia cuniculus. It is possible that in some
Macropods the allantois reaches the chorion, although Caid-

PLACENTA IN PERAMELES 173

well’s statement (1884) that there is such a union in the case
of Halmaturus ruficollis, as well as his testimony of
a fusion between the bilaminar omphalopleure and the uterine
wall, have not yet been confirmed. An omphaloplacenta is
well developed in Macropods.

Phalangeridae.—In Trichosurus vulpecula (Hill,
1889) and Petaurus sciureus (Semon, 1894) the allantois
ig similar to that of Macropods. I am also able to state that
this is the case for Pseudochirus cooki. Here again the
embryo depends on the work of the trophoblast both of the
vascular omphalopleure and of the bilaminar omphalopleure.

Phascolarctus.—this genus is particularly interesting in
possessing, according to Caldwell (1883) and Semon (1894),
a respiratory allantois. There is a well-developed omphalo-
placenta and also a union in the metrioplacenta between an
annular zone of the bilaminar omphalopleure (just outside the
simus terminalis) and the uterme mucosa.

Didelphys.—tThe allantois does not meet the chorion.
The omphaloplacenta is well developed. Certain portions of
the ectoderm of the bilammar omphalopleure are stated by
Selenka to form absorptive proliferations similar to those found
in certain Eutheria, for example Manis (Weber), and
Equus (Ewart).

Dasyurus.—tThe allantois shows interesting stages in
degeneration. At a particular stage it becomes applied to the
chorion which is itself in intimate association with the uterine
mucosa. Later the allantois withdraws from the chorion and
degenerates considerably, its vascular system practically
disappearing. An omphaloplacenta is present as well as a
similar annular fusion of the bilaminar omphalopleure with
the uterine wall as occurs in Phascolarctus.

Perameles.—tThis is a most primitive form possessing
a well-developed allantoic placenta and an omphaloplacenta,
and there is considerable evidence of absorption in the bilaminar
omphalopleure.

From the above abstract it will be seen that we can, as yet,
hardly be said to have a detailed knowledge of the structure,

174 tT. THOMSON FLYNN

physiology, and ontogeny of the foetal membranes of most
marsupials. Particularly in such primitive genera as Thyla-
cinusand Sarcophilus, it may be expected that investiga-
tion will help to shed a clear light on the phylogeny of the
placenta in this group.

Another point of importance (of greater value, I think, than
Assheton would have had us believe) lies in the behaviour of the
uterine mucosa. Of this our knowledge in the marsupials is
particularly meagre. Yet it is extremely important, since
there is naturally a mutual reaction of embryo and uterus.
An investigation of the modifications of the uterme mucosa
during pregnancy would, there is not the slightest doubt, be
of great value in shedding a light on ancestral placental arrange-
ments in marsupials.

Pseudochirus cooki is instructive in this regard.
Preliminary investigations which I have already made in the
case of this diprotodent marsupial have shown that the uterine
epithelium in a very early stage of pregnancy consists of a single
layer of very high columnar cells with correspondingly elongated
deeply-staining nuclei. Below the epithelium the connective
tissue is condensed to form a layer in which run the capillaries.
This stage can be recognized as being very similar to one occur-
ring in many Eutheria.

At a later stage of gestation, cell outlines have disappeared
and a vascular syneytium is formed similar to that of Pera-
meles, except that it is composed apparently not only of the
epithelial cells but of those of the sub-epithelial capillary layer.
These capillaries now ramify at the surface as is the case in
Perameles.

Here without doubt can be recognized the remains of an
ancestral trophospongial proliferation.

From the consideration of the above facts, particularly as
regards the condition of the foetal membranes in Perameles,
Dasyurus, and Phascolarctus, bearing in mind the
complementary modifications of the uterme wall where they
are known, it must be evident that these conditions in marsu-
pials represent a degeneration from a more complex system

PLACENTA IN PERAMELES 175

of placentation which undovbtedly obtained in the original
protoplacental group.

Into a full treatment of this there is no need for me to enter.
It has been ably discussed by Hill and the facts and conclu-
sions embodied in the preceding pages can only be regarded as
confirming and strengthening his expressed opinions.

(dq) The Relation of the Allantoic Placenta-
tion of Perameles to that of the Hutheria.
This question I will discuss but briefly, reserving its full

treatment for some future occasion when adequately fixed
and preserved late gestation stages of Perameles may
perhaps be available.

It is with some pleasure that I have been able to bring the
method of allantoplacental formation of Perameles into
line with that occurring in the simpler Eutherian forms. In
fact it may be said in general that the only difference between
the two is one of degree. There are the same characteristics
of passivity of the uterine epithelium and activity of the
trophoblast with a division of the latter into a cytoblastic and
plasmodial layers. After preliminary diploplasmatic prepara-
tion the allantois becomes fixed and an apposition of the two
blood-streams becomes effected. I might here briefly refer
to the resemblances between the earlier stages of allanto-
placentation in Perameles and the dog and rabbit. In
Text-fig. 3 I have indicated the main points of Schoenfeld’s
fig. 14 (1903) representing an early stage of chorionic invasion
in the dog. A somewhat comparable stage in Perameles is
represented by Text-fig. 4. The agreement in the method of
foetal invasion is evident. In the dog, however, according to
Schoenfeld, the uterine epithelium does not form a syncytium.

In the rabbit, on the other hand, as in many other Kutheria,
such a maternal syncytium is formed, and here the early stages
show an even more significant resemblance to those occurring
in Perameles. Particularly I may refer to Schoenfeld’s
(1903) figs. 4, 5, and 6, Pl. xxi, and those of Maximow (1900,
figs. 1 and 2, Pl. xxx).

176 T. THOMSON FLYNN

The phylogenetic importance of the presence of large multi-
nucleate masses of foetal origin in the allantoplacenta of
Perameles, the dog, the rabbit, and others cannot be over-
estimated.

Bearing in mind the accepted origin of the Metatheria

TEXT-FIG. 3.

Fig. 3—An early stage of development of the dog showing chorionic
attachment (after Schoenfeld). cyt., cytoblast; pb., plasmodi-
blast ; ep., uterine epithelium.

TExtT-FIG. 4.

Fig. 4—A stage in Perameles comparable with that of the dog
in Text-fig. 3. Lettering as in Text-fig. 3.

and Eutheria from a primitive diphyodont protoplacental
stock (Hill, 1897, p. 432), it is possible to state with certainty
that, in that early group, the same conditions of passivity of
the uterine epithelium and active phagocytic quality of the
trophoblast were already in existence; with the further
differentiation of the latter layer in its placental portion into
two distinct layers, respectively cellular and plasmodial.

The foregoing facts make this conclusion inevitable. This
being so, those Eutheria, particularly the Ungulata,

PLACENTA IN PERAMELES 177

in which there is no union between the trophoblast and mucosa
in the allantoic placental region, must have reached this condi-
tion, in the course of their phylogeny, by a process of secondary
simplification.

It is evident enough then that the attempt made by Strahl
(1906) to group Perameles with the Ungulates and others
in the Semiplacenta breaks down. At the same time it is
not easy to suggest any arrangement by which Perameles
will take its proper place in placental classification.

The only possible course at this stage appears to be to
examine briefly what is the relation of the allantoic placentation
in Perameles to some one or other of the groupings at
present in use.

Assheton’s suggestion to divide the Placentalia into
Placentalia cumulata and plicata seems to be the
most promising, since in addition to having a structural basis
these divisions are to some extent physiological. Assheton has
given a table of the characteristics of the two groups as he
conceived them (1909), and even a cursory glance at these will
indicate that the allantoplacenta of Perameles structurally
and physiologically occupies a place somewhere between the
two but more primitive than either.

Thus in the * heaping up ’ of the trophoblast—a very funda-
mental point—it agrees with the cumulate type, while in many
other features, absence of lacunae and the mildness of attack
on maternal tissues, it approaches the pleate type. The
secretion of the placental glands appears to be at first of minor
importance in Perameles, but, later, absorption by the
allantoic vessels is direct, increasing the value to the embryo
of glandular secretion tremendously.

It is apparent that if we regard the above grouping as a
rational one, the allantoic placenta can be regarded as being
of a central primitive type from which development in either
direction might easily have proceeded.

And the very real relation of the allantoplacentation in
Perameles to that of the Carnivora—the relation of the
simple to the slightly more complex—shows that the Carni-

NO. 265 N

/

178 T. THOMSON FLYNN

vora, as suggested by Hubrecht and upheld by Assheton,
exhibit, of the Hutheria, the most undifferentiated arrange-
ment. ‘The trophoblastic “heaping up’ is common to the
allantoplacenta of both Perameles and the Carnivora.
This is fundamental, and is sufficient evidence of the fact that
the cumulative type of placenta is the more primitive.

8. PAPERS REFERRED TO IN THE TEXT.

Assheton, R. (1906).—‘‘The Morphology of the Ungulate Placenta,
particularly the Development of that Organ in the Sheep, with Notes
on the Placenta of Hyrax and the Elephant’, ‘ Phil. Trans.’, Ser. B,
vol. cxcviii.

—— (1909)—* Professor Hubrecht’s Paper on the Early Ontogenetic
Phenomena in Mammals: An Appreciation and a Criticism”, ‘ Quart.
Journ. Micr, Sci.’, vol. 54, part ii.

Caldwell, W. H. (1884).—‘** On the Arrangement of the Embryonic Mem-
branes in Marsupial Animals”, ibid., vol. 24.

(1887)—** The Embryology of Monotremata and Marsupialia’’, part i,

‘ Phil. Trans.’, Ser. B, vol. clxxviii.

~ Chapman, H. C, (1881)—‘* On a Foetal Kangaroo and its Membranes ”’,
‘Proc. Acad. Nat. Sci.’, Phil., part iii.

Chipman, W. (1903).—*‘ Observations on the Placenta of the Rabbit, with
Especial Reference to the Presence of Glycogen, Fat, and Iron”, * Lab.
Repts., Roy. Coll. Phys.’, Edinburgh.

Ewart, J. C. (1917).—** Studies on the Development of the Horse. 1, The
Development during the Third Week’, ‘ Trans. Roy. Soc.’, Edinburgh,
51,

Grosser, O. (1910)—‘** Development of the Egg-membranes and Placenta ”’,
in Keibel and Mall, * Manual of Human Embryology ’, Phil. and Lond.

, Hartmann, C. G. (1916)—** Studies in the Development of the Opossum,

Didelphys virginiana, (i) History of the Early Cleavage, (ii) Formation of

the Blastocyst ”, ‘ Journ. of Morph.’, vol. 27.

1919).—** Studies in the Development of the Opossum, Didelphys

virginiana, (iii) Description of new Material on Maturation, Cleavage, and

Entoderm Formation, (iv) The Bilaminar Blastocyst ’’, ibid., vol. 32.

© Hill, J. P. (1895)—** Preliminary Note on the Occurrence of a Placental

Connexion in Perameles obesula and on the Foetal Membranes of

Certain Macropods ”’, * Proc. Linn. Soe.’, N.S.W., vol. 10, part iv.

/ —— (1897).—‘*‘ The Placentation of Perameles”’ (Contributions to the

Embryology of Marsupialia, i), ‘ Quart. Journ. Mier. Sci.’, vol. 40.

(1899)—** On a Further Stage of the Placentation of Perameles and

on the Foetal Membranes of Macropus parma ”’, ibid., vol. 43.

—

PLACENTA IN PERAMELES 179

Hill, J. P. (1900).—*‘ On the Foetal Membranes, Placentation, and Parturi-
tion of the Native Cat (Dasyurus viverrinus) ”’, ‘ Anat. Anz.’, vol. 18.
(1910)—‘** The Early Development of the Marsupialia, with Special
Reference to the Native Cat’, ‘ Quart. Journ. Micr. Sci.’, vol. 56.
and C, H. O'Donoghue (1913)—‘* The Reproductive Cycle in the

Marsupial Dasyurus viverrinus ”’, ibid., vol. 59.

Hubrecht, A. A. W. (1889)—\‘‘ The Placentation of Erinaceus europaeus”
(Studies in Mammalian Embryology, i), ibid., vol. 30."

——— (1894).—‘** The Placentation of the Shrew’, ibid., vol. 35.

(1899).—‘** Ueber die Entwickelung der Placenta von Tarsius und

Tupaja nebst Bemerkung iiber deren Bedeutung als hamatopoietische

Organe ”’, ‘ Proc. Int. Cong. Zool.’, Cambridge, 1898.

(1909)—*‘ Karly Ontogenetic Phenomena in Mammals and _ their
Bearing on our Interpretation of the Phylogeny of the Vertebrates ”’,
* Quart. Journ. Micr. Sci.’, vol. 53.

Jenkinson, J. W. (1902).—*‘ Observations on the Histology and Physiology
of the Placenta of the Mouse ’”’, ‘ Tijd. Nederl. Dierk.-Ver.’, (2) vii.

—— (1906)—“‘ Notes on the Histology and Physiology of the Placenta
in Ungulata ’’, ‘ Proc. Zool. Soc.’, 1906.

— (1913)—‘ Vertebrate Embryology ’, Oxford.

Matschie, P. (1915)—“‘ Kinige Beitrage zur Kenntniss der Gattung
Pseudochirus, Ogilb.”’, ‘ Sitzb. Ges. natf. Freunde’, Berlin, 1915.

Maximow, Alex. (1900)—** Die ersten Entwickelungsstadien der Kanin-
chen-Placenta”’, ‘ Arch. f. mikr. Anat.’, lvi.

Minot, C. 8. (1889)—‘* Uterus and Embryo, (i) Rabbit, (ii) Man ”’, * Journ.
of Morph.’, vol. 2.

(1890)—‘*‘ Die Placenta des Kaninchens”’, ‘ Biol. Centralblatt ’,

1890.

(1891)—** A Theory of the Structure of the Placenta ”’, ‘ Anat. Anz.’,

vol. 6.

(1911)—* A Laboratory Text-book of Embryology ’, London, 1911.

Nolf, P. (1896).—*‘ Etude des modifications de la muqueuse utérine pendant
la gestation chez le Murin (Vespertilio murinus)”, ‘Archiv. de
Biol.’, 14.

~ Osborn, H. F. (1883).—‘‘ Observations upon the Foetal Membranes of the

Opossum and other Marsupials’, “ Quart. Journ. Micr. Sci.’, vol. xxiii,

——. (1888).—** The Foetal Membranes of the Marsupials: the Yolk-sac
Placenta in Didelphys”’, ‘ Journ. of Morph.’, vol. 1.

Resink, A. J. (1902)—\‘‘ Bijdrage tot de kennis der placentatie van
Erinaceus europaeus ” (with abstract in German), * Tijd. Nederl. Dierk.-
Vereen.’, (2) vii.

Schoenfeld, H. (1903)—‘‘ Contribution 4 Vétude de la fixation de Put
des Mammiféres dans la cavité utérine et des premiers stades de la
placentation ’’, ‘ Archiv. de Biol.’, 19.

180 T. THOMSON FLYNN

\ Selenka, E. (1886)—‘ Studien iiber Entwickelungsgeschichte der Thiere :
Das Opossum (Didelphys virginiana).’

Semon, R. (1894)—\‘‘ Die Embryonalhiillen der Monotremen und Mar-
supialer’’, “ Zool. Forsch. in Australien’, &c., Bd. 2.

Strahl, H. (1890)—\‘‘ Ueber den Bau der Placenta von Talpa europaea
und tiber Placentardriisen ’’, ‘ Anat. Anz.’, vol. 5.

(1906).—‘‘ Die Embryonalhiillen der Siaiuger und die Placenta”’,
Hertwig’s ‘ Handb. der vergl. Entwicklungsgesch.’, Bd. 1, Th. 2.

Van Cauwenberghe, N. (1910)—‘‘ Etude sur les cellules géantes du
placenta de la Taupe’, * Arch. de Biol.’, 25.

Willey, A. (1914)—‘‘ The Blastocyst and Placenta of the Beaver”’,
‘Quart. Journ. Micr. Sci.’, vol. 60.

9. DESCRIPTION OF FIGURES.

The outlines of all figures have been drawn with the aid of
Zeiss’s camera lucida, then enlarged by means of the panto-
graph and details filled in by freehand.

List oF REFERENCE LETTERS.

all., allantois. all.cap., allantoic capillary. all.ent., allantoic entoderm,
all.mes., allantoic mesoderm. all.pen., penetration of the allantoic meso-
derm. all.st., allantoic stalk. all.ves., allantoic vesicle. amm., amnion.
ana., anastomosis of the two portions of the sinus terminalis. bil.omph.,
bilaminar omphalopleure. cav., cavity contained in syncytial nests.
ch.mes., chorionic mesoderm. ch.ect., chorionic ectoderm. coel.w., coelomic
wall of the allantoic vesicle. cyt., cytoblast (cytotrophoblast). . emb.,
embryo. ea.coel., extra-embryonal coelome. ex.syn., maternal syncytium
of the extra placental area. ea.syn.n., Syncytial nuclei of the extra-placental
area. g.¢., multinucleate giant cell. gl., gland. gl.ep., gland epithelium.
inf., infiltrated material (? lymphatic) contained in the cavity of a syncytial
nest. ing., material ingested by the plasmodiblast. leuc., leucocytes.
mat.cap., maternal capillary. m.ch., marginal chorion. muc., uterine
mucosa. musc., muscularis. pb., plasmodiblast. pb.n., plasmodiblast
nuclei. pgm., aggregations of pigment in the plasmodiblast. plac., allan-
toic placenta. pl.syn., maternal syncytium of the placental area. proa.,
proamnion. syn.n., nuclei of the maternal trophospongia. troph., maternal
trophospongia. vasc.omph., vascular omphalopleure. v.omph., vascular
omphalopleure. vit.a., vitelline artery. vit.v., vitelline vein. yk.cav.,
yolk-sac cavity. y.s., yolk-sac. yl.spl., yolk-sac splanchnopleure.

PLACENTA IN PERAMELES 181

EXPLANATION OF PLATES 9-11.

PLATE 9.
Figs. 1-8, 10-12, PERAMELES OBESULA, 6:-1mm. Fic. 9, PERAMELES
OBESULA, 12°5 mm.

Figs. 1, 2, and 3.—Nections showing the arrangement of syncytial
nuclei as a more or less irregular layer round a central cavity. Note the
vesicular shape and chromatic characteristics of these nuclei. Note also
intruding leucocytes.

Figs. 4, 5, and 6—Phases in the earliest growth of the plasmodiblast.
In fig. 4 the syncytial nuclei are already undergoing degeneration under
the effect of the plasmodial advance processes,

Fig. 7—Plasmodial attack on a gland. Note the breaking down of the
gland epithelium on one side. Here also ingested material is evident in the
plasmodiblast.

Fig. 8—Another stage in plasmodiblast formation. The cytoblast is
a well-defined layer. Note presence of leucocytes.

Fig. 9—Shows the vesicular and degenerate appearance of foetal
nuclei at this stage with close apposition of maternal and foetal vessels.

Figs. 10, 11, and 12.—Photomicrographs of sections through a branched
gland just outside the area of the first fixation of the chorion (see Text-
fig. 1).

Prate 10.
Fics. 13-16, 19, PeraMELES GUNNI, 66mm. Fics. 17, 18, PERAMELES
OBESULA, 125mm. Fics. 20, 21, PERAMELES OBESULA, 7 mm.

Fig. 13.—Placental area towards the centre, showing the commence-
ment of the disorganization of the cytoblast allowing maternal capillaries
to approach the surface. The very superficial position of one of these
capillaries is, however, very exceptional for this stage. Note the penetration
of the plasmodiblast nuclei into the syncytial nests.

Fig. 14.—Section of somewhat more peripheral portion of the same area.
A distinct cytoblast is present and a plasmodiblast in which giant cells
and ingested material are outstanding features. This figure should be
compared with Maximov’s fig. 1 of the rabbit.

Figs. 15, 16, and 17.—Show attachment of allantois and disorganization
of cytoblast to allow of the apposition of foetal and maternal blood-vessels.
Figs. 17 a, b, c, d, e-——Stages in the degeneration of a foetal nucleus,

Fig. 18—Section of the central portion of the placenta showing the
following : breaking down of cytoblast, almost complete filling of syncytial
nests by foetal nuclei, penetration of allantoic capillaries and almost final
apposition of maternal and foetal blood-streams.

182 T. THOMSON FLYNN

Fig. 19—Placental area showing section through a gland. Note dis-
appearance of gland epithelium, the enlargement of the gland lumen at the
apex and the tendency to form lateral slits in the plasmodium. This
section shows well the intruding leucocytes of the large mononucleate
type. The plasmodiblast shows a particularly noteworthy giant cell,
pigment patches, and ingested material.

Fig. 20.—Section of placental area towards the margin. Cytoblast is
intact. The nests are well filled with plasmodiblast nuclei.

Fig. 21.—NSection of placental area at the margin. For description see
text.

Pate 11.
Fies. 22-27, PERAMELES GUNNI, 6-6 mm. Fic. 28, PERAMELES OBESULA,
Tmm. Fre. 29, PERAMELES OBESULA, 6:1 mm.

Vig. 22.—Section showing structure of the bilaminar omphalopleure.

Note the extreme vacuolation of the ectoderm cells.

Fig. 23.—The yolk-sac circulation. For description see text.

Fig. 24——Diagram showing the relation of the embryo to its membranes.

Figs. 25, 26, and 27——Pigment-bearing cells: fig. 25 of the serosa,
fig. 26 from the connective tissue, and fig. 27 from the trophoblast.

Fig. 28.—ection through the mouth of a gland.

Fig. 29.—Shows the relation of a plasmodial proliferation to a group of
syncytial nuclei,

\\

~—_

~
Y
rR
~~
=
=
S
~

i" > Quart. Journ. Mier, Sci. Vol. 67, N.S., Pl. 9.

= oe
: +

ww
—~—_
So te
$2 re
a: i
‘

55S

TO

Tiny
e<
=e:

ch.mes
Z ch. act

pgm .
mat Cap
Ung /

pb

4 A.
°
f I

L.
19 qe

T. T. Flynn del,

oe. ’
Geen
REC

seu

oa et

e = oi

Sy UL. MES

Quart. Journ. Micr. Sci. Vol. 67, N.S., Pl. 10.

mh. pil gS
ey M2? IS

ch. ye di

bed, hee

oy a

— re

ie ey
“¥
7 j
' =
M
= —
‘ :
x
<n i =
: a
3 + 7 5
. ~
= a ; =
t
J
i
M |

'
oo
a -
= \
(i *
7
9
7 . i
'
|
: a

See Pao he ita

°
ys
~ ‘ 2s

~ - : ; re. :

| + » A r

; =
A a Pay
: pe 7
a? «
- ’ .
4 ’
f rs
= . ‘; =
*
: _
‘
J *
’
a o M4 =
=: ,
v 4 ‘

- i

ae

“
P »>
‘ =
.
’
i
7
5
i
!

The Male Meiotic Phase in two Genera of
Marsupials (Macropus and Petauroides).

By
W. E. Agar, F.R.S.,

Professor of Zoology in the University of Melbourne.

With Plates 12-14.

A vary slight experience of cytological research is sufficient
to impress the worker in this field with the different facilities
for accurate research afforded by different organisms, and also
with the importance of discovering the most favourable objects
for such research. Consequently I have been making a cyto-
logical survey of such groups of Australian animals as seemed
most likely to afford useful cytological material, paying at
first particular attention to the Marsupials, for two reasons.
The first is the well-known technical difficulties presented by the
Kutherian mammals on account of the usually rather large
number and tendency to clump of their chromosomes, and the
second is that Jordan’s work on the American opossum (1911)
showed that in this marsupial the number of chromosomes is
comparatively small. Jordan determined the number as seven-
teen (male), but Painter (1922) has raised it to twenty-two.
Up to the present we have made in this laboratory a pre-
liminary survey of some fourteen species of Marsupials, and
this paper presents an account of the more important features
of two of these which have been worked out in more detail.

At present the most interesting feature of this work is
undoubtedly the determination of the conditions of the sex
chromosomes, a problein which has always presented in
mammals considerable elements of dubiety. Of primary
importance is probably the confirmation of the occurrence of

NO. 266 O

184 W. E. AGAR

Y-chromosomes in the Mammalia—first clearly established by
Painter (1922) in Didelphys. In Macropus this element
is very minute, and, though I had shown it in several figures,
I failed to interpret it properly until we received Painter’s
paper, which arrived at a moment when Mr. Greenwood (whose
results are to be published in this journal) was paying attention
to a very minute chromatic body occurring in several other
species of Marsupials at which he was working. It imme-
diately became apparent that this body was the Y-chromosome.
Comparison with his work, and with Painter’s description of
Didelphys, quickly established that the small body already
observed in Macropus is also a Y-chromosome.

It is a pleasure to acknowledge my indebtedness to several
people for assistance in obtaining this material. Dr. T. L.
Bancroft, of Eidsvold, Queensland, has sent me in the last two
years a great number of specially preserved testes of Marsupials,
Monotremes, Ceratodus, and many other Australian animals.
For living material utilized in the present paper I have to
thank Dr. Colin MacKenzie, Director of the Australian Institute
of Anatomical Research, and Mr. W. H. D. Le Souef, of the
Zoological Gardens, Melbourne.

MATERIAL AND METHODS.

The two species dealt with in the present communication
are Macropus ualabatus and Petauroides volans.
These genera belong respectively to the Diprotodont families
Macropodidae and Phalangeridae.

The material was mostly preserved in Flemming, Bouin, and
Allen’s modification of Bouin (1913, 1915). I have found the
latter method excellent, especially when followed, as Allen
recommends, by anilin and bergamot oils in place of the higher
alcohols and xylol. For general purposes I have found this
method the best I have yet tried for mammals.

The standard stain used was Heidenhain’s iron haematoxylin,
though safranin, methyl green, and acid fuchsin, and others,
were used as controls.

As in former work, I have found thick sections much more

MEIOTIC PHASE IN MARSUPIALS 185

valuable than the thin ones usually used by cytologists. Most
of the work has been done with sections of 15-20 p» in thickness,
mounted between two coverslips instead of in the usual way
between a slide and a coverslip. This allows any nucleus to
be examined from both sides. For convenience of examination
the lower coverslip is temporarily attached to a microscope
slide by a drop of immersion oil.

A. Macropus.

Number of Chromosomes.—tThis animal is remarkable
among mammals for its small number of chromosomes. The
chromosome formula is of the type first clearly established for
mammals in the case of Didelphys (Painter, 1922), the
males being of the XY type. The diploid chromosome number
in the male is 10+ XY, or twelve, and in the meiotic division
there are five autosome bivalents and the XY bivalent. This
very small number allows of great certainty i counting the
chromosomes, Macropus being obviously far more favourable
for this purpose than any other known mammalian genus.

Although counts of the meiotic chromosomes leave no doubt
that the number is as just stated, it is only rarely that there
are twelve separate chromosomes in the spermatogonial mitoses.
As a rule there are quite indubitably only eleven (PI. 12, fig. 2),
of which one, usually occupying the centre of the ring, is very
minute. This is the Y-chromosome. In a small percentage
of cases, however, there are equally plainly twelve (Pl. 12,
fig. 3), the extra one being smaller than any of the others
except Y. ‘This is obviously the X-chromosome, for the
meiotic phase shows that X is much smaller than the auto-
somes. In the 11-chromosome spermatogonial mitoses X is
presumably attached to one of the autosomes, though I have
not been able to identify with certainty the chromosome to
which it is joined. Owing to its small size it would not add
much to the length of one of the longer chromosomes. I have
frequently found a constriction near the end of one of the
longer chromosomes, but in view of the widespread tendency of
chromosomes to develop such constrictions it would be unjustifi-

02

186 W. E. AGAR

able to assume that this represents the point of attachment of
the X-chromosome.

As will be described in more detail below, similar conditions
are found in the meiotic division. The XY bivalent is possibly
sometimes independent, but more often it is attached to one
of the autosomes. In this phase, however, the XY is easily
identifiable even when attached to an autosome.

In the fact of its usual attachment to an autosome but
occasional independence, both in spermatogonial and meiotic
mitoses, the X-chromosome in Macropus_ resembles that
of Ascaris megalocephala (Kdwards, 1910). In many
male Orthoptera also the X-chromosome is temporarily or
permanently united to an autosome (McClung, 1905 ; Wilson,
HOU).

In the female (Graafian follicle cells) the small Y-chromosome,
so characteristic of the spermatogonia, is not present. I have
never been able to find more than ten separate chromosomes,
and here, as in the male, the small number of chromosomes
makes it easy to find a large number of dividing nuclei in
which every chromosome is distinct (Pl. 12, figs. 4, 5). Sinee
there is no Y, and since in the male, X is generally attached to
an autosome, it is quite safe to interpret the ten chromosomes
of the female as 10+XX, and the two X’s attached to auto-
somes. The condition here is again comparable to that found
in A. megalocephala, where Frolowa (1912) found that
the two X-chromosomes are generally attached to autosomes
in the female.

The Meiotie Phase.—tThe spermatogonial nuclei (PI. 12,
fig. 1) contain a very scanty chromatic reticulum and a large
central nucleolus. This is apparently a plasmosome impregnated
with chromatin, for it stains densely with iron haematoxylin,
but in well-balanced methyl green and acid fuchsin preparations
it takes up the fuchsin. In the early prophase of the spermato-
gonial mitoses this nucleolus loses its chromatic staming
reaction even with iron haematoxylin, and becomes a typical
plasmosome. This nucleolus is the only compact body m the
resting spermatogonial nucleus, so unless they are somehow

MEIOTIC PHASE IN MARSUPIALS 187

incorporated with the plasmosome, it is clear that the sex
chromosomes at this stage are in a diffused condition like the
autosomes.

The earliest stages of the primary spermatocytes which are
distinguishable from the spermatogonia are early leptotene
stages—a rather later leptotene nucleus being shown in fig. 6.
The large nucleolus is shown by its staining reaction to
be a plasmosome, and is still the only compact body in the
nucleus. The X-chromosome is therefore at this stage in the
leptotene condition like the autosomes. The same presumably
applies to the Y-chromosome, though this is too small to permit
of definite statement.

The synizetic contraction, though unmistakable, is not very
pronounced (PI. 12, fig. 7).

The process of syndesis is difficult to follow in this animal.
It begins about the stage shown in fig. 6, and is completed
by the stage illustrated in fig. 7, which is a pachytene nucleus.
About the stage of fig. 6 frequent duplicity and parallelism of
threads can be observed, from which parasyndesis may be
inferred. The direct evidence for this mode of syndesis is,
however, certainly not strong in this species, but the indirect
evidence is very convincing. Firstly, this mode of syndesis
can be observed in Petauroides, and the general course
of meiosis is so similar in the two genera that it is incredible
that the mode of syndesis should not be the same. Secondly,
as we shall see below, there is no doubt that the components
of the definitive bivalents in Macropus are derived from the
pachytene threads by the longitudinal splitting of these, and
not by their doubling over. That being so, it follows that the
mode of syndesis must have been by longitudinal fusion, unless
one of the most fundamental hypotheses of modern cytology—
that is to say the individuality of the chromosomes—is false.

In the early pachytene nuclei (Pl. 12, fig. 7), as in still earlier
stages, the sex chromosomes are not visibly different from the
other chromosomes. As the pachytene threads begin to
contract, however, X soon becomes visible by reason of its
much more rapid condensation, so that it soon comes to form

188 W. E. AGAR

a compact rounded mass in sharp contradistinction to the still
thread-like autosomes (PI, 12, figs. 8-10). From its first appear-
ance onwards it is attached to the end of one of the autosomes.
At first it is not possible to identify the minute Y, but later, as
the autosomes lose in staining capacity, Y becomes conspicuous
by reason of its denser stain. At first it is distinct from X,
but they soon fuse to form a bivalent (Pl. 12, fig. 9).

A large pale plasmosome makes its appearance at the time
that the sex chromosomes are uniting and in close contact with
them. The nature of this plasmosome, and its relation to the
plasmosome of the earlier stages, is discussed below.

In the late pachytene stage (PI. 12, figs. 9-10) the staining
capacity of the autosomes becomes greatly diminished, and
their outlines become somewhat blurred by the development
of outgrowths and anastomoses between the different chromo-
somes. The compact XY bivalent is now very conspicuous,
owing to the fact that the general decrease in staining capacity
does not affect 1t nearly so much as the autosomes.

Fig. 11 represents the diplotene stage. This is perhaps
chiefly interesting on account of the confidence with which
its mode of derivation from the pachytene stage can be deter-
mined. In many late pachytene nuclei, such as about the stage
shown in fig. 10, the five pachytene bands can be counted
with ease and certainty, and one can follow step by step in
great detail the conversion of each of these bands into one of
the diplotene bivalents by the appearance and gradual widening
of a longitudinal split down its middle. The three stages
figured (Pl. 12, figs. 10, 11, 12) will, however, probably be
enough to carry conviction that the gemini of the diplotene
nucleus are derived from the pachytene bands in this way, and
not by their doubling over as is required by the theory of
telosyndesis.

Fig. 13 shows a stage in the contraction of the diplotene
loops into the definitive bivalents. The nuclear membrane
has by now disappeared. The great increase in bulk of the
chromosomes which has taken place between the stages shown
in figs. 11 or 12 and that shown in fig. 13 is remarkable. A

MEIOTIC PHASE IN MARSUPIALS 189

rough estimate of the relative volumes of the total chromatin
content at these two stages made by means of plasticene
models showed that the volume of the chromatin is more
than twice as great in the later as in the earlier stage.

The XY bivalent is visible as before, attached to an auto-
some. It is now, however, seen to be attached to one limb
only of the bivalent.

Figs. 14-16 (PI. 13) represent metaphases of the first meiotic
division, to show the relations of the sex chromosome. In
fig. 14 they form a compact body attached to one end of one
of the autosome bivalents. In fig. 15 they are similarly attached,
but somewhat drawn out towards the equator of the spindle.
At this stage no distinction between X and Y is visible, but
as the metaphase progresses the two components begin to
separate, as shown in figs. 164 and B. The minute Y is very
characteristically pulled out along the spindle-fibre at this
stage.

Fifty metaphase I’s were examined especially in respect to
the mode of attachment of XY.

In fourteen cases it was attached as in fig. 14.

In twenty-five cases it was attached as im fig. 15.

In eleven cases it was apparently free, forming an independent
bivalent. In many of these cases, however, it was probably
attached as in the manner of fig. 15, but by a longer and finer
thread. There is also little room for doubt that an attachment
as in fig. 14 indicates an early metaphase, and that later the
relations shown in fig. 15 are always assumed.

I have not been able to determine whether the autosome
to which XY is attached is always the same one, but it appears
probable that it is (note the distinctive shape of this chromo-
some in the two groups figured in fig. 18, and in figs. 14 and 15).
The bivalent in question is, however, certainly always one of
the larger ones.

The first meiotic division is the differential division for X
and Y (Pl. 13, figs. 16, 17). During anaphase Y becomes still
further pulled out along the line of the spindle-fibre, and
presents in the late anaphase the characteristic appearance

190 W. E. AGAR

illustrated in fig. 17. In Heidenhain preparations both X and
Y are at this stage slightly paler than the autosomes.

Two kinds of secondary spermatocytes are therefore pro-
dueed in the well-known manner, one with the X-chromosome
and the other with the Y. There is a complete, and apparently
prolonged, resting stage between the two divisions. ‘The
difference between the two kinds of secondary spermatocytes
is conspicuous in the young nuclei, one member of each pair
containing a dense chromatic body (presumably X) which is
lacking in the sister nucleus or represented by a very much
smaller speck (PI. 13, fig. 18). A group of fully resting secondary
spermatocytes, presenting the same dimorphism, is shown in
Hie Wo.

Fig. 20 shows a prophase for the second division in a
secondary spermatocyte containing the larger chromatic body
(X) which is seen attached to one of the chromosomes.

The expected two types of second division are easily found.
An early, and rather irregular anaphase with the Y-chromosome
is shown in fig. 21. Here Y has just divided. <A later anaphase
of the other type of second division is illustrated im fig. 22.
Here we have apparently only five chromosomes present in
each group; this must clearly be interpreted as a division in
which X is present and fused, as usual, with an autosome.

It is noticeable that there is no trace in Macropus (nor
in Petauroides) of the second pairmg of chromosomes to
give a quarter of the diploid number which has so often been
described for the second division in birds and mammals.

Jordan (1911) described such a second pairing for Didel-
phys, but Painter (1922) found that it did not occur in his
material.

B. PErTavuRoIDES VOLANS.

Number of Chromosomes.—The determination of the
number of chromosomes in this species presents more difficulty
than in the case of Macropus, owing to their greater
number. The number countable in the spermatogonial
mitoses is generally twenty-two, forming typically a ring of
twenty, with two smaller, shghtly unequal ones, in the centre.

MEIOTIC PHASE IN MARSUPIALS 191

These are presumably X and Y, the latter beg much larger
than the corresponding element m Macropus, and being,
indeed, but slightly smaller than X. I have, however, found
some spermatogonial mitoses with apparently only twenty-one
chromosomes, and yet containing this pair in the centre.
I am therefore in doubt, from the spermatogonial mitoses,
whether the number is 20+ XY or 20+X. J have some quite
unequivocal counts of polar views of the first meiotic metaphase,
and some of these show eleven and some twelve separate
elements. When twelve are present, one is always distinctly
smaller than any of the others. Presumably, when the number
is eleven, there are ten autosome bivalents and the XY bivalent.
When the number is twelve, X and Y have dissociated. Side
views of the meiotic metaphase (of which I have never found
one that could be counted) show that one of the bivalents
(2? XY) commonly dissociates much in advance of the others.
The conditions in the early pachytene nuclei also poimt to the
presence of two sex chromosomes. It appears, therefore, to
be fairly well established that the formula for the male
Petauroides is 20+ XY, or twenty-two in all. This corre-
sponds with Painter’s enumeration for Didelphys. It
will also be noticed that the number of autosomes is double
that of Macropus. None of my ovarian material proved
suitable for chromosome counting.

The Meiotic Phase.—The spermatogonial nuclei of this
animal differ from those of Macropus, in that the place of
the fine reticulum of the latter genus is taken by a number of
irregular blocks of chromatin, united by anastomoses. In the
young spermatogonia these chromatic bodies are in approxi-
mately the diploid number, and from a study of the spermato-
gonial pro- and telophases it appears probable that these
blocks are of the nature of ‘ prochromosomes’, being the
undiffused remains of the telophase chromosomes, and passing
directly into the chromosomes of the following prophase.
When the spermatogonia pass into a more profoundly resting
stage the number of these bodies becomes more difficult to
determine, owing to their becoming broken up and _ their

192 W. E. AGAR

fragments scattered, till at last a stage is reached where the
chromatin is finely distributed throughout the nucleus except
for three or four masses representing the remains of the larger
blocks which have escaped complete fragmentation. Very
often, however, the number of the chromatic masses remains
at approximately the diploid number throughout the whole
interphase from the telophase to prophase.

The interstitial nuclei of the two genera present a similar
difference, those of Macropus being finely reticular, and
those of Petauroides containing about the diploid number
of irregular chromatic masses.

The leptotene stage develops from a primary spermatocyte
having the same structure as a spermatogonium, 1. e. containing
a number of massive chromatic bodies (PI. 13, fig. 23). Hach
of these becomes the centre of a process of thread formation
(Pl. 18, figs. 24, 25), to produce in the aggregate the leptotene
nucleus (Pl. 13, fig. 26). This process resembles that by which
the leptotene stage in certain insects is developed from a
nucleus containing the diploid number of chromatic bodies
(Wilson, 1912). In Petauroides, however, the evidence
that these blocks represent each a single chromosome, though
strong, is not complete.

The massive centres from which the thread-spinning starts
persist for a long time, and indeed appear to form the
basis of the synizetic knot. Synizesis is more pronounced in
Petauroides than in Macropus—at least in my material
(Pl. 13, fig. 27).

Parasyndesis occurs during the synizetic contraction (Pl. 13,
figs. 27, 28), but not with the regularity observable in those
animals in which the leptotene nucleus is orientated mto a
bouquet. In the nucleus shown in fig. 27 it is proceeding over
one length of the thread, but appears to be already completed
over the rest of the nucleus. That syndesis concerns, not the
chromosomes as a whole but their constituent chromomeres, is
beautifully shown in Petauroides (PI. 13, figs. 27, 28). In
this animal the chromomeres are unusually distinct and large
at this, and some other, stages.

MEIOTIC PHASE IN MARSUPIALS 193

Figs. 27 and 28 show that, (1) the chromomeres in a single
chromosome differ greatly in size, (2) the series of chromomeres
in a pair of conjugating chromosomes closely corresponds,
(3) conjugation takes place between the corresponding (homo-
logous) chromomeres. It will be noticed also that the final
union of chromomeres appears to be very intimate, all external
trace of duplicity having disappeared in the most completely
fused pairs.

Syndesis is followed by the usual pachytene stage (Pl. 14,
figs. 29-33), during the early part of which two compact bodies,
presumably X and Y, are conspicuous (PI. 14, fig. 31). These
soon unite into a single body, one or both of them often being
pulled out into irregular shapes during the process (Pl. 14,
fig. 32). The autosomes remain filamentar and suffer a tem-
porary diminution of staining capacity. An important feature
which is very conspicuous in Heidenhain preparations subjected
to the right amount of extraction is that the stain is retained
much more tenaciously by certain of the chromomeres than
by others (Pl. 14, fig. 32). The general significance of chromo-
meres is discussed below.

As the chromosomes regain their staining powers towards
the end of the pachytene stage, the chromomeres again become
very conspicuous (Pl. 14, fig. 33), as they are also in the diplotene
nucleus (Pl. 14, figs. 34, 35). It is interesting to compare the
chromosomes shown in detail in fig. 85 with those in fig. 28,
the latter representing the conjugation of the chromosomes,
the former their separation. The correspondence between the
chromomeres of homologous chromosomes is still evident in
the diplotene stage, but now they are beginning to run together
on the contracting chromosome, ultimately to give rise to the
smooth chromosomes shown in fig. 36.

I have not been able to trace with certainty the movements
of the sex chromosomes in the first meiotic division. In
regard to the second division, all that can be said is that the
number is clearly about ten or eleven, showing that there is
no second numerical reduction.

The dimorphism of the secondary spermatocytes is not so

194 W. E. AGAR

conspicuous as in Macropus, probably because the X and Y
chromosomes are more nearly of the same size. Indeed, for
a long time I thought there was no visible difference between
the two types, but closer examination has shown that such
a distinction exists. All secondary spermatocytes have a com-
pact chromatic nucleolus, but in the case of young sister nuclei,
whose relation to each other can still be seen by the persistent
spindle remains, one of them constantly has a distinctly larger
nucleolus than the other. Sometimes, as in the pair figured
(Pl. 14, fig. 37), this is expressed by one of them being bilobed
and the other single. In other cases it is merely a difference in
size. The difference between the two classes of secondary
spermatocytes is thus very small, but once recognized it is seen
to be constant.

DIscUSSION OF SOME SPECIAL PROBLEMS RAISED BY THE
FOREGOING DESCRIPTIONS.

(1) Chromomeres.—Many cytologists maintain that
chromomeres are purely artefacts, due to unequal contraction
of the chromatic thread under the influence of the fixative,
or else, in the case of smaller chromomeres, are mere optical
effects of angles, &c., in the thread. At any rate, according to
this view, they do not represent any real local differentiations
of the substance of the chromosome.

The strength of the criticism that the chromomeres are
artefacts depends much upon the exact meaning attached to
that word. It is of little importance whether or not chromo-
somes which are beaded when fixed appear smooth in life (a very
difficult observation in any case!) For the sake of argument
it may be granted that the beading is produced by the action
of the fixative. The important question is: Is the beading of
such a nature that it could be produced mechanically by
precipitation in and contraction of a homogeneous thread, or
is it the expression of a pre-existing though perhaps invisible
differentiation in the living chromosome ? Doubtless so-called
chromomeres have been deseribed which might have been

MEIOTIC PHASE IN MARSUPIALS 195

produced by the action of the fixative on a homogeneous
thread, but the view that the chromosomes are composed of
differentiated chromomeres cannot be disposed of by demon-
strating mistaken interpretation in individual cases. On the
contrary, we have now a large accumulation of observations
where the answer to the above question seems undoubtedly.to
be in the negative; observations, that is to say, which lead
to the conclusion that whether the beads exist as such in life,
or whether they are produced by the fixative at the moment of
death, they must be expressions of local differentiations of the
substance of the chromosome—and that is all, of course, that
is required by the theory that connects the chromomeres with
the linear arrangement of hereditary factors in a chromosome.

The reasons which seem to exclude the view that such chromo-
meres are produced mechanically, and so to speak, accidentally,
on a homogeneous thread which is contracting unequally under
the influence of unequal stresses in different parts are :

(1) The chromomeres in a single chromosome may differ
very greatly as regard size (PI. 13, figs. 27, 28).

(2) Had they been produced by unequal contraction of parts
of a homogeneous thread, larger chromomeres would be
separated from each other by longer intervals of connecting
thread than those which separate the smaller chromomeres.
A glance at fig. 28 shows that this rule does not hold.

(3) There is a close correspondence between the chromomeres
of homologous chromosomes, both during syndesis and the
diplotene stage.

(4) Wenrich (1916) has described the constant arrangement of
the principal chromomeres on a given chromosome—a con-
stancy which is maintained not only in all the nuclei (of the
same stage) in a single animal, but even in different animals.
Unless this observation be doubted, it supplies conclusive
evidence that the chromomeres as seen in fixed nuclei corre-
spond to definite local differentiations of the substance of the
chromosome.

(5) It is now well known that the shape assumed by the long
type of chromosome common in many forms of mitosis 1s

196 W. E. AGAR

characteristic and constant for any given chromosome. ‘This
shape is commonly some form of V, with equal or unequal
limbs. The point at which the chromosome bends to form
the V—whether in the middle or towards one end—(which
is also the spot to which the spindle-fibre is attached) varies
from chromosome to chromosome, but is constant for any
given chromosome. Similarly, the transverse constrictions
which develop across the chromosomes of so many organisms
are constant in position for a given chromosome. This con-
stancy in position is proof of the existence of a constant
differentiation, in a lengthwise direction, of the substance of
the chromosome.

The tendency of certain chromosomes to develop transverse
constrictions at spots constant for each particular chromosome
is specially significant in estimating the value of the criticism
that chromomeres are ‘artefacts’. In Lepidosiren
(Agar, 1913) the long somatic chromosomes usually show no
trace of a transverse constriction. Hach chromosome is a
smooth curved rod or V of approximately uniform thickness.
When the chromosomes become shorter and thicker, as happens
regularly in the meiotic phase, and occasionally, from unknown
reasons, In somatic tissues also, the transverse constrictions
develop in a spot characteristic for each particular chromosome
(which is also the point at which the apex of the V is situated,
when the chromosome is in this shape).

This shortening and thickening of the chromosomes was
produced in several plants by Sakamura (1920) by the action of
various reagents, such as chloral hydrate and chloroform.
These artificially shortened chromosomes showed well-marked
and characteristically placed constrictions, though the position
of these is barely indicated in normally fixed tissue. The con-
stancy in position of these constrictions shows that they can
only be called ‘ artefacts ’ in the sense that they make visible
a pre-existing heterogeneity of the chromosome substance,
which is concealed from view in the ‘ typical’, well-fixed, and
apparently uniform chromosome. Indeed, to deny that the
chromomeres correspond to pre-existing local differentiations

MEIOTIC PHASE IN MARSUPIALS 197

of the chromosome substance on the ground that they only
appear in tissues treated in a certain way, would be as illogical
(granting that it is true) as to deny the distinction between
a plasmosome and a chromatin nucleolus because the differ-
ence only becomes visible under the action of appropriate
stains.

We conclude, therefore, that the chromomeres which appear
in certain stages of mitosis in fixed tissues correspond to real
local differentiations of the substance of the chromosome,
though the actual shape which they assume (namely, bead-like
swellings on a fine thread) may be assumed, or at least exag-
gerated, under the stress of the fixative.

(2) Crossing over.—Whether or not the phenomenon of
crossing Over occurs in mammals is still in doubt. Castle
(1921) has described such a case in rabbits for the linked genes,
English and non-English, and short-haired and Angora.
Three individuals were tested, two males and a female, and
crossing over was found in all of them. If this is established,
it will show that the phenomenon in mammals is not quite
comparable to that in Drosophila and Bombyx, where
it occurs only in the sex which is homozygous for the sex
chromosomes. By analogy with these, crossing over is to be
expected in the female mammal alone. Considering the
cytological evidence only, it would certainly seem that the
conditions supposed to be necessary for crossing over are
provided in the male diplotene nuclei of both these genera.

This is specially clear in Petauroides, because of the
chromomeres. As fig. 27 shows, fusion of chromomeres in
syndesis is intimate. Indeed, no sign of duplicity may remain.
In fig. 35 (diplotene stage) the intertwined chromosomes are
still held together at certain of the crossing places by unsplit
chromomeres, and in view of their intimate union it is not
difficult to imagine that when they finally do separate they may
do so in such a way that the portions of the two chromosomes
on either side of the point of union have been interchanged.

(3) The Relation between the Sex Chromosomes
and the Plasmosomes in the Meiotic Phase.—The

198 AGAR E. W.

condition of the plasmosome in Macropus has already been
described. In Petauroides a plasmosome can occasionally
be seen in the primary spermatocytes, but usually none can be
identified with certainty—probably because, asin Macropus,
this plasmosome is partly chromatic in the resting nucleus,
and therefore does not stand out clearly from the other
chromatic bodies. In the leptotene, synizetic, and early
pachytene nuclei of Petauroides, a plasmosome can
sometimes be identified, but is usually concealed among the
dense tangle of chromatic threads. In the later pachytene
stages, two plasmosomes are plainly visible (Pl. 14, fig. 32).
One is considerably darker than the other, and has no relation
to the sex chromosomes. This one I take to be the plasmosome
of the earlier stages. The other plasmosome is in close relation
to the sex chromosomes, and indeed appears to be formed out
of their substance. It makes its appearance as a large pale
body at the time that the sex chromosomes are uniting into
a bivalent, and at first is an elongated structure closely attached
to the bivalent. In some cases its shape and relations suggest
that it is the persistent part of the bivalent from which the
chromatin has flowed away into the rounded mass which forms
the condensed sex bivalents: this appearance is strengthened
by the fact that sometimes rounded granules or drops of
chromatin are left embedded in the plasmosome. In other
cases it is pear-shaped and is attached by its neck to the
bivalent, irresistibly suggesting that it has been squeezed out
of the contracting chromosomes like a viscid fluid from a
narrow aperture. These appearances are illustrated in figs. 32
and 88. In later stages this plasmosome becomes approxi-
mately spherical and often becomes detached from the sex
bivalent, though always lying close to it.

For a time the two plasmogomes—one presumably the
remains of the pre-leptotene nucleolus, and the other apparently
formed out of material (plastin or linin ?) derived from the sex
chromosomes—coexist, the former being the first to disappear.

In Macropus the larger pale plasmosome which appears in
close connexion with the uniting sex chromosomes has also the

MEIOTIC PHASE IN MARSUPIALS 199

appearance of being formed out of their substance, though no
figures were discovered quite so striking as those illustrated for
Petauroides.

(4) Chromatoid Bodies.—In Petauroides one or
two bodies staining densely with iron haematoxylin appear
suddenly in the cytoplasm in the pachytene stage. Their
origin and fate have not been determined, but they seem to be
distributed capriciously at cell-division. In fig. 87 they have
all passed to one daughter cell (that containing the sex chromo-
some), but this mode of allocation is not invariable. In
Macropus chromatoid bodies are either absent or incon-
spicuous.

SUMMARY.

Macropus ualabatus has twelve chromosomes, namely
10+ XY in the male and 10+ XX in the female.

In Petauroides the number is almost certainly twenty-
two, the male being of the formula 20+ XY. No female counts
were obtained for this animal.

In the male Macropus X is generally attached to one of
the autosomes in spermatogonial mitoses. Y, which is exceed-
ingly minute, is free. During the pachytene stage, while the
autosomes are still elongated, X and Y condense into a bivalent.
In the first meiotic division this bivalent is attached to an
autosome.

As a result of the first meiotic division the usual two classes
of secondary spermatocytes are formed one with X and the
other with Y. In the second meiotic division, those with
X show only five separate chromosomes, showing that X, as
usual, is fused with an autosome. The other class of second
divisions shows five autosomes and the minute Y.

In the female Macropus the sex chromosomes were never
found free from the autosomes in the ovarian follicle cells,
which therefore show only ten separate chromosomes.

In Petauroides the sex chromosomes cannot be distin-
guished with certainty from the autosomes. An unequal pair
of small chromosomes usually situated in the centre of the

NO. 266 P

900 W. E. AGAR

spermatogonial metaphase plates probably, however, are
X and Y. Early pachytene nuclei show two compact bodies
which unite into one, presumably the sex bivalent.

The second reduction of the chromosome number to one-
quarter of the diploid total in the second meiotic division,
which has been described for several species of birds and
mammals, does not take place either in Macropus or
Petauroides.

Chromomeres are very prominent in Petauroides in
the zygotene and diplotene stages.

Probably in Macropus, and more convincingly in
Petauroides, the cytological conditions to permit of
‘ crossing over ’ are present in the male.

The plasmosome which appears in the pachytene stage is
probably formed from the plastin or linin basis of the con-
tracting sex chromosomes.

REFERENCES TO LITERATURE.

Agar, W. E. (1913).—‘ Quart. Journ. Mier. Sci.’, 58.
Allen, E. (1913).—* Journ. Morph.’, 31.

(1915).—‘ Anat. Record’, 10.

Castle, W. E. (1921).—‘ Science’, 54.

Edwards, C. E. (1910).—‘ Arch. f. Zellforschung’, 5.
Frolowa, S. (1912).—Ibid., 9.

Jordan, H. E. (1911).—Ibid., 7.

McClung, C. E. (1905).—* Biological Bulletin’, 9.
Painter, T. S. (1922).—‘ Journ. Exp. Zool.’, 35.
Sakamura, T. (1920).—‘ Journ. Coll. Science Tokyo’, 39.
Wenrich, D. H. (1916).—‘ Bull. Mus. Comp. Zool. Harvard’, 60.
Wilson, E. B. (1911).—‘ Journ. Morph.’, 22.

(1912).—‘ Journ. Exp. Zool.’, 13.

MEIOTIC PHASE IN MARSUPIALS 201

EXPLANATION OF PLATES 12, 18, AND 14.

The scale given on Pl. 12 applies to all the figures except
19 and 38, which are on a smaller scale.

LETTERING.

P, pre-pachytene plasmosome. Px, plasmosome which appears at the
time that the sex chromosomes are uniting. xX, Y, the sex chromosomes,

Figs. 1-22, Macropus ualabatus; Figs, 23-38, Petauroides
volans.

PLatTE 12.

Fig. 1.—Resting spermatogonium.

Fig. 2.—Metaphase, spermatogonial mitosis, eleven chromosomes,

Fig. 3.—Metaphase, spermatogonial mitosis, twelve chromosomes.

Figs. 4, 5.—Metaphase, ovarian follicle cells, ten chromosomes,

Fig. 6.—Late leptotene nucleus.

Fig. 7.—Synizesis.

Fig. 8.—Early pachytene nucleus showing condensation of X.

Fig. 9.—Later stage, showing pairing of X and Y.

Fig. 10.—Late pachytene nucleus.

Fig. 11.—Early diplotene nucleus.

Fig. 12.—The chromosomes of an early diplotene nucleus shown
separately.

Fig. 13.—Contracting bivalents in two adjacent cells.

PuaTeE 13.

Figs. 14, 15.—Two metaphases of the first meiotic division, to show
modes of attachment of XY to an autosome.

Figs. 164, B.—Autosomes, with attached XY bivalent, from more
advanced metaphases, to show separation of X and Y.

Fig. 17.—Anaphase of first division.

Fig. 18.—Pair of young secondary spermatocytes, still connected by the
spindle remains, one with large compact chromatic body, the other without.

Fig. 19.—A group of secondary spermatocyte nuclei to show the
dimorphism. Half with large and half with small chromatic body (pre-
sumably X and Y).

Fig. 20.—Early prophase of a second division with the X-chromosome.

Fig. 21.—Anaphase of a second division with the Y-chromosome. A
small cyphlasmic inclusion is shown at the bottom right-hand corner.

Fig. 22.—Anaphase of a second division with the X-chromosome
(indistinguishably fused with an autosome). To avoid overlapping the
two groups, havebeen slightly shifted laterally in drawing.

Pp 2

202 W. E. AGAR

Fig. 23.—Resting primary spermatocyte.

Figs. 24, 25.—Fragments of developing leptotene nuclei to show con-
version of the massive blocks of the resting spermatocyte into the leptotene
threads.

Fig. 26.—Leptotene nucleus.

Fig. 27.—Synizesis and syndesis.

Fig. 28.—Three short lengths of conjugating chromosomes from three
different zygotene nuclei.

PLATE 14.

Fig. 29.—Syndesis complete.

Fig. 30.—Early pachytene nucleus, synizesis loosening out.

Fig. 31.—Pachytene stage, with two compact bodies, presumably X
and Y.

Fig. 32.—Later stage. X and Y, one of them greatly pulled out, uniting
into a bivalent.

Fig. 33.—Late pachytene nucleus showing evidence of commencement of
diplotene stage. Chromatoid body in the cytoplasm.

Fig. 34.—Early diplotene nucleus. Chromatoid body in cytoplasm. —

Fig. 35.—Three bivalents from an early diplotene nucleus, .

Fig. 36.—Late diplotene nucleus, chromatoids in cytoplasm.

Fig. 37.—A pair of young secondary spermatocytes, still connected by
the spindle remains. Note larger and bilobed chromatic body in the upper
nucleus. Chromatoids in the cytoplasm of one cell.

Fig. 38.—Outline drawings of four nuclei, about the stage of fig, 32,
to show relations between the sex chromosomes and the plasmosome.

Quart. Journ Micr. Sct. Vol.67, N.S.,PU.12.

W E.Agar del

Quart. Journ. Micr. Sa. Vol.67, N.S. PL.13.

Quart. Journ. Micr: Sci. Vol.67, N.S. PL.14.

WE.Agar del.

Marsupial Spermatogenesis.

By
A. W. Greenwood, B.Se.,

University of Melbourne.

With Plates 15 and 16,

INTRODUCTION.

Tuis work is a contribution to the cytology of the Marsupials,
a group which, owing to the small number of their chromosomes,
affords peculiar advantages over other mammals for this
kind of study. The sex chromosomes are particularly clear in
this group, and Painter’s discovery of a Y-chromosome in the
American opossum (Didelphys) has been fully confirmed
in several Species of Australian Marsupials examined in this
laboratory. In the present paper I give the results of the
study of three species of Marsupials belonging to two different
families, two species belonging to the family Dasyuridae of the
sub-order Polyprotodontia, and one to the family Phalan-
geridae of the sub-order Diprotodontia.

The form examined in most detail is Phascolaretus
cinereus. In the other forms I have done little more than
determine the number and the behaviour of the sex chromo-
somes.

The following work was undertaken under the guidance,
and with the assistance, of W. E. Agar, F.R.S., Professor of
Zoology in the University of Melbourne.

SPERMATOGENESIS OF PHASCOLARCTUS CINEREUS.

Material.

The animals were obtained through the courtesy of Mr. Ker-
shaw of the National Museum, from the National Park at

204 A. W. GREENWOOD

Wilson’s Promontory. The first set of testicular and ovarian
material was obtained by Professor Agar, who killed and fixed
the material. All the other animals obtained were killed and the
gonads fixed at the University laboratory. The animals were
received at different times during the year, namely in the
months of January, April, May, July, and August. In all
testis, except those obtamed in April, the tubules were filled
with various stages in spermatogenesis; in the latter the
tubules contained a relatively large number of Sertoli cells,
but other stages in spermatogenesis were few in number.

Hixine!

Various methods of fixing the material were employed—
Bouin, Allen’s modification of Bouin, cold Flemming, and
corrosive acetic. The only satisfactory fixatives were the
Bouin fluids and the Flemming. For early prophases the
Flemming-fixed material gave the best results, but the Bouin
fixatives gave very clear figures of the division stages.

Staining.

Sections of 10 and 20, were cut. Some were mounted
in the ordinary way on glass slides, and others were mounted
between coverslips so that the nuclei could be examined from
both sides.

Heidenhain’s iron haematoxylin with iron alum was used for
staining the sections. Staining with safranin and gentian
violet was also tried, but the results were not very satisfactory.

Number of Chromosomes.

The diploid number of chromosomes in Phascolaretus
is sixteen. This number was obtained from numerous counts
in both male and female material. Female counts were obtained
from the prophases of large cells of the corpus luteum. One
such cell is figured (PI. 15, fig. 1). Altogether sixteen chromo-
somes can be seen. In the female it should be noted that no
chromatic dot is present, such as is shown so clearly in the

MARSUPIAL SPERMATOGENESIS 205

spermatogonial metaphase plates (Pl. 15, figs. 2 and 3). In
the female there are the comparatively large autosomes (14)
and two much smaller chromosomes. ‘These two smaller
chromosomes, similar in size and shape, are the sex chromo-
somes. The chromosome complex of the female is therefore
14+XX.

In the male counts were obtained from equatorial plates of
the dividing spermatogonia, and from first and second meiotic
division figures. The number of chromosomes obtained from
these stages was sixteen. In the spermatogonial plates the
chromosomes are arranged in a circle around a central clear
space. Inside the circle of chromosomes a chromosome much
smaller than the others is seen; also near this small chromo-
some a small chromatic dot is to be seen (Pl. 15, figs. 2 and 3).
From their subsequent behaviour these are identified as the
X- and Y-chromosome respectively. The chromosome formula
of the male is therefore 14+ XY.

Structure of Testes.

The testes have the typical mammalian structure of numerous
convoluted tubules. Close to the wall of the tubules are
situated the Sertoli cells, spermatogonia, and the early meiotic
prophases. Passing in towards the lumen of the tubule, the
later stages of the maturation divisions occur, leading up to
the formation of the spermatozoa. These are found nearest
the lumen of the tubule. There appears to be a definite layering
of the cells of the different stages, although some overlapping
of the inner layers occurs.

Spermatogonia.

The nuclei of the early spermatogonia are oval in shape, and
are frequently lobed. The nuclei of the later generations of
spermatogonia are much smaller and are approximately circular
in shape. In the resting condition the spermatogonial nucleus
contains a well-defined nucleolus which stains very densely with
the iron haematoxylin. This is evidently of the nature of
a plasmosome impregnated with chromatin. The chromatin

206 A. W. GREENWOOD

of the resting cell occurs in the form of rather faintly stained
blocks connected by fine strands. The blocks of chromatin
appear as loose masses, frayed out at the edges, and the number
present is approximately the same as the diploid number of
chromosomes.

The onset of the prophase is marked by an increase in the
staining capacity of the chromatin blocks which now give rise
to short irregular threads. These threads increase in length,
probably at the expense of the nucleolus, which now shows up
as a large pale body with a few deeply staining granules
embedded in it. The long irregular threads begin to contract,
ultimately giving rise to the thick chromosomes marking the
end of the prophase. The nucleolus has given rise to a large,
oval, faintly stained body, a typical plasmosome.

Meiotic Phase.

The origin of the leptotene stage has not been determined
with absolute certainty, but it appears that the telophase of
the last spermatogonial division does not pass into a complete
resting stage, but the chromatin remains in the form of blocks
situated close underneath the nuclear membrane. These blocks
are present in approximately the diploid number. They are
more compact and stain more deeply than the chromatin
blocks seen in the resting spermatogonial nucleus. The
leptotene stage appears to be derived from this by the formation
of long threads from these chromatin blocks. The early lepto-
tene nucleus consists of a tangle of fine threads. On the
threads chromomeres can be seen. These are spaced rather far
apart and vary considerably in size. In the centre of the
nucleus the pale plasmosome can be seen. Following this stage
the threads begin to contract away from one side of the nucleus,
and, at the same time, begin to contract in length. This is
the earliest indication of synizesis (Pl. 15, fig. 4). Although the
exact time of syndesis could not be determined, it is probable
that it begins to take place now.

It is significant that the leptotene nucleus entering synizesis
shows that the threads contract away from that side of the

MARSUPIAL SPERMATOGENESIS 207

nucleus which is opposite to that on which the archoplasmic
mass is found, and it appears as if this body exerts some
influence, if it is not wholly the cause of the synizetic con-
traction.

In the early leptotene stage no sign of a compact X- or Y-
chromosome could be seen, so that they are evidently threaded
out like the autosomes at this stage.

Fig. 5 shows a much later stage in syndesis and synizesis.
The nucleus is not complete, but shows very clearly the pairing
of the chromomeres in homologous threads. The chromomeres
exhibit great variability in size. In one of the threads syndesis
appears to be nearly complete. In the centre of the nucleus
can be seen the compact mass of the synizetic contraction.

The leptotene stage is followed by the pachytene stage.
Fig. 6 shows an early pachytene stage consisting of thick,
looped chromosomes which have emerged from the synizetic
contraction. These chromosomes are seen to be distinctly
double in composition, the presence of the chromomeres showing
up as darkly stained bodies in the more lightly stained thread.
The threads are now very thick. The ends of the threads
at this stage are directed towards the archoplasmic mass. Later,
these threads lie scattered through the nucleus and all trace
of duplicity is eventually lost. The X-chromosome makes its
appearance in the early pachytene stage. It appears first as
a thin, deeply stained thread (Pl. 15, fig. 7), but contracts down
to form a round mass, which is typical of the X-chromosome
in later prophase. I have been unable to identify the minute
Y-chromosome at this stage. Whether the Y-chromosome fuses
with the X-chromosome to form a bivalent could not be
determined with certainty owing to the minuteness of the
Y-chromosome and the presence in the nucleus at this stage
of several deeply staiming granules.

Always in contact with the X-chromosome there is a large
pale plasmosome (PI. 15, figs. 7, 8, 9). This varies considerably
in shape, and usually contains a number of deeply staining
granules. In later pachytene stages two plasmosomes are
visible, each containing one or more deeply staining granules

208 A. W. GREENWOOD

(Pl. 15, fig. 9). One of the plasmosomes (Px) remains in contact
with the sex chromosome, the other (Px’) appears to be formed
from a division of this plasmosome.

The early pachytene stage is followed by a late pachytene
stage in which the chromosomes show a marked diminution in
staining capacity. They become diffuse and furry in appear-
ance, the X-chromosome or possible XY bivalent alone remain-
ing as a deeply stained body. .The onset of the diplotene stage
is marked by the recovery of the staining power of the chromo-
somes. The chromosomes begin to split so that two long,
thin threads are formed. The chromomeres can be distinctly
seen on these threads occurring in pairs (PI. 15, fig. 10).

The thin threads now begin to contract, but never entirely
separate from one another, remaining in contact at two or more
points (PI. 15, fig. 11). At this stage the nucleus, which has been
increasing in size from the onset of the meiotic prophase, has
now attained its maximum volume. Following this stage the
nuclear membrane breaks down and the chromosomes lie free
in the cytoplasm.

In a few cases during the diplotene stage I have seen the
X-chromosome apparently attached to the end of one of the
autosomes, but this condition appears to be exceptional. In
most cases it lies free. It is in the metaphase of the first meiotic
division that the Y-chromosome can first be identified with
certainty. Figs. 12 and 13 are of metaphase plates. The first
is a Gell just after the nuclear membrane has disappeared. The
seven large bivalents and the separate X- and Y-chromosomes
can be seen attached to one another by threads. Fig. 13 is
a rather later view. In Phascolarctus the X- and the
Y-chromosome do not usually form a bivalent. In division
figures they are always found separate. In the meiotic prophase
they may possibly be in the form of a bivalent, but, as men-
tioned before, this point could not be determined.

Fig. 14 shows a metaphase side-view. In this it will be seen
that the X- and Y-chromosomes are on opposite sides of the
mass of chromosomes at the equator of the cell, and are travel-
ling to opposite poles of the cell ahead of the other chromosomes.

MARSUPIAL SPERMATOGENESIS 209

In many cases, on the other hand, the sex chromosomes lag
behind on the spindle. This is shown in fig. 15 of an anaphase.
In this figure it will be noted that the X- and the Y-chromosome
are attached to the same spindle-fibre.

The first division is therefore the reductional division, and
gives rise to two daughter secondary spermatocytes which are
dimorphic, one containing the X-chromosome and the other
containing the Y-chromosome. ‘This dimorphism of the
secondary spermatocytes is shown in fig. 16 (P1.16). That these
are two daughter spermatocytes is shown by the remains of the
spindle-fibres connecting the two cells. In one of the cells can
be seen a rather large, deeply stained body in the nucleus which
I take to be the X-chromosome. In the other cell nucleus
there is a much smaller body not so deeply stained, which
probably represents the Y-chromosome.

Second Meiotiec Division.

Before the onset of the prophase of the second meiotic
division, the secondary spermatocyte undergoes a prolonged
resting stage and increases greatly in size. From the resting
stage with its faintly staining network, the onset of the prophase
is shown by the recovery of the staiming power of the chromatin
in patches. From this the deeply stained, irregular threads of
the prophase are formed (PI. 16, fig. 17). Right through the
meiotic stages the X-chromosome has retained its staining
capacity and does not thread out, except in the early prophase
of the first meiotic division. During the prophase of the second
division the X-chromosome remains compact and does not
thread out. The division follows on as before, but this time the
sex chromosomes divide, one half going to each pole of the cell.
The second division is therefore equational. No further reduc-
tion in the number of chromosomes takes place during this
division.

Fig. 18 shows an anaphase of the second division with the
X-chromosome divided and lagging behind on the spindle.
Fig. 19 shows a late anaphase of the same division showing the
presence of the Y-chromosome at both poles of the cell.

910 A. W. GREENWOOD

Cytoplasmic Inclusions.

So far in my description of the spermatogenesis of Phasco-
laretus I have made no mention of cytoplasmic structures.
Beyond noting their occurrence in the germ cells I have done
little to determine their nature.

In the cytoplasm of the spermatogonia and meiotic stages
a large round body is to be seen (Pl. 15, fig. 6). This varies
greatly in size and staining capacity at different stages. In
some stages it is quite deeply stained, notably in the spermato-
gonia, leptonema, and early pachynema. Later it stains
capriciously, and eventually, during the first meiotic division,
becomes quite pale (Pl. 15, figs. 12 and 13). It does not divide
during the division but passes indiscriminately to one or other
of the secondary spermatocytes (Pl. 16, fig. 16). It occasionally
is found in the early spermatid nuclei, but I have been unable
to find it at any later stage.

This body, I believe, is probably the same as that figured
by Gatenby in his work on the * Cytoplasmic Inclusions of
Germ Cells ’, as an excretory granule.

Another cytoplasmic inclusion seen in all stages of spermato-
genesis 18 a very pale, somewhat sausage-shaped body lymg
close up against the nuclear membrane (PI. 15, figs. 6 and 10).
It is towards this body that the synizetic contraction takes
place. It is found in all the secondary spermatocytes and
spermatids, although I have never found any sign of its division
during any of the stages in spermatogenesis. From its behaviour
I believe this to be the archoplasmic mass.

Other cytoplasmic inclusions are the chromatoid bodies.
These are conspicuous in the sections fixed with Flemming,
but are pale and inconspicuous in those sections fixed in Bouin.

The leptotene nucleus always contains one or more deeply
staining granules. In the pachytene stage also the nucleus often
contains a deeply stained granule usually lying close beneath
the nuclear membrane (PI. 15, fig. 7). These granules appear
to give rise to the chromatoid bodies seen in the cytoplasm of
the germ cells at different stages (Pl. 16, fig. 28).

MARSUPIAL SPERMATOGENESIS 11

These chromatoid bodies vary in size and in number. During
the first meiotic division they are distributed indiscriminately
between the two daughter cells. Further than this they have
not been traced.

The Sertoli Cell in Phascolarctus.

In Phascolarctus the Sertoli cell nucleus is very large
(Pl. 16, fig. 20). It is about three times as large as the nucleus
of the primary spermatocyte when it has attained its maximum
volume. The nucleus lies at the foot of the Sertoli cell close
to the wall of the tubule. The outlines of the cell could not
be distinguished. The chromatin of the nucleus is in the form
of a coagulum distributed through the nucleus. In the centre
of the nucleus there is a clear space surrounding a granular
mass. This granular mass is probably formed by the degenera-
tion of the nucleolus.

The cytoplasm of the Sertoli cell contains, usually lying close
up against the nucleus, a varying number of refractive, rod-
like bodies. These are a constant feature of the Sertoli cell
in Phascolarctus, and I found them present in all

_the material examined by me whatever time of the year the

material was obtained. Usually in close association with the
rods, a pale yellow, fluffy mass can be seen. This probably
consists of deutoplasmic material. This shows up well in
sections which have been fixed in cold Flemming before
staining with the iron haematoxylin.

In some of the outside tubules of the sections, especially in
the Flemming-fixed material, instead of the bundle of rods a
mass of over-lapping, very pale plates can be seen (PI. 16, fig. 21).

The rods appear pale yellow in those sections fixed in the
Bouin, but often in the Flemming-fixed material are quite
black. The rods are found together, lying approximately
parallel. They vary in length, some of them consisting of small
pieces lying end to end and probably formed by the fracture
of one of the longer rods. Although the rods are usually com-
paratively straight, in many cases they are seen to possess
a very wavy outline.

312, A. W. GREENWOOD

Much has already been written regarding these rod-like bodies
present in the cytoplasm of the Sertoli cells in Phasco-
larctus. Ido not propose to discuss at much length the
theories already advanced, but will give my conclusions
arrived at by a study of these bodies, and the results of experi-
ments undertaken by me to determine if possible their nature.
The experiments were undertaken with the view to ascertaining
whether the rods served a nutritive function. The fact that
these rods were found only in the Sertoli cells led me to believe
that possibly these rods were a source of nutriment, or con-
nected in some way with the supply of nutriment to the develop-
ing spermatozoa.

Below I give the results of some digestion experiments under-
taken to prove this point if possible.

Fresh material was cut by means of a freezing microtome.
Owing to the difficulty of picking up the rods in unstaimed
material, the tissue was stained. The stains used were neutral
red and methyl green.

The sections after staining were submitted to the action of
a weakly acidic mixture of pepsin and glycerine. Cells with
rods were picked out and their position marked by means of
a micrometer eyepiece. Then the microscope was placed in
an electric oven and kept at a constant temperature of 30° Centi-
grade. The progress of the action was watched from time to
time. The accumulation of the products of digestion after a time
tended to stop the action and only partial digestion took place.
In all cases of partial digestion the rods were still visible. With
the use of more of the digestive fluid complete digestion of the
cytoplasm occurred. In this case, with complete digestion, the
cells moved about in the fluid and were difficult to find again.
In the majority of the experiments the rods were identified
even after complete digestion, but in some cases they could
not be found.

The same experiment was carried out, using an alkaline
mixture of zymine and glycerine. Again, here the rods were
still visible after partial digestion ; but in the case of complete
digestion, owing to the difficulty of picking up the rods in the

MARSUPIAL SPERMATOGENESIS 213

resulting fluid, I could not be absolutely certain whether the
rods were dissolved or not, although in the majority of cases
the rods were still visible.

Bardeleben refers to these bodies as crystals of haematoidin.
He also describes the presence of similar bodies in the great
blood lacunae in the material of the testis. With regard to the
composition of these bodies I have not been able to determine
whether they are composed of haematoidin or not.

Fresh tissue was boiled in chloroform. The tissue was then
embedded and sections cut, stained, and mounted. Upon
examination it was found that in no case did the chloroform
have any action on the rods in the Sertoli cells.

This does not of course confirm Bardeleben’s view that they
are crystals of haematoidin, but goes to show that they are,
at any rate, not a blood derivative which is an acid.

With regard to the occurrence of similar bodies in the blood
lacunae of the testis, in all the material I have examined I have
not found any bodies comparable with the rods found in the
Sertoli cells. The presence of somewhat similar rods has been
described by several authors. Montgomery has shown that the
Sertoli cells in man are derived from the spermatogonia, and
that the nature of the resultant cell is determined by the
presence of a rod-like body in the cytoplasm, i.e. all spermato-
gonial cells containing the rod-like body give rise to the Sertoli
cells. ‘This, however, does not appear to be the case in
Phascolarctus. Although I am convinced that the Sertoli
cellsin Phascolarctus arise from a division of the spermato-
gonia, I have never been able to find any trace of the rods until
the nucleus of the cell, by its peculiar structure, is definitely
defined as a Sertoli cell. No Sertoli cell divisions have been
found in Phascolarctus, but in Perameles (in this
animal the Sertoli cells are comparable with Phascolarctus
in point of size and nuclear structure) I have found an apparent
diplotene nucleus which from its size and position in the
tubule appears to have originated from a Sertoli nucleus.
From this it seems safe to assume that the Sertoli cell has been
derived from the same cells as give rise to the germ cells.

214 A. W. GREENWOOD

SPERMATOGENESIS OF SARCOPHILUS URSINUS.

Both male and female animals of this species were obtained.
The technique followed was the same as in Phascolarctus,
viz. fixatives used were Allen’s modification of Bouin, and cold
Flemming ; this was followed by staining in Heidenhain’s iron
alum haematoxylin. In the testis the same more or less
definite layering of the germ cells in the tubules is noted as in
Phascolarctus.

Number of Chromosomes.

In the female the number of chromosomes was obtained
from metaphase plates of dividing follicle cells surrounding
the ovum. The number found was fourteen. Of this number
twelve are large and the other two are much smaller. These
two smaller ones, similar in size and shape, are the sex chromo-
somes. The chromosome formula of the female Sarcophilus
is therefore 12+ XX (PI. 16, fig. 22).

In the male chromosome counts were obtained from spermato-
gonial plates and first meiotic division stages.

From the metaphase plates of the spermatogonial divisions
the number of chromosomes was found to be fourteen. Here,
asin Phascolarctus, the presence of a small X-chromosome
and a minute Y-chromosome, usually in the centre of a circle
formed by the twelve larger autosomes, was again noticed
(Pl. 16, fig. 28). In fig. 24 the spermatogonial chromosomes are
dividing or have already divided. Two of the chromosomes
as yet show no sign of splitting (Pl. 16, fig. 24, a and b). The
division of the Y-chromosome is very clearly shown in this
figure.

The chromosome formula of the male Sarcophilus is
therefore 12+XY.

Meiotic Phase.

T have made no attempt to follow out in detail the phenomena
of the meiotic phase, but stages similar in appearance to those
of Phascolarctus are found in Sarcophilus.

MARSUPIAL SPERMATOGENESIS 215

The side-view of an early anaphase of the first meiotic division
is shown in fig. 25. The X- and the minute Y-chromosome are
on opposite sides of the central mass of chromosomes, and are
travelling towards the poles of the cell ahead of the other
chromosomes. ‘The first meiotic division therefore acts as
the reductional division and the two daughter secondary
spermatocytes produced are dimorphic, one containing the
X-chromosome and the other containing the Y-chromosome.
Satisfactory second meiotic division figures have up to the
present not been obtained.

SPERMATOGENESIS OF DASYURUS MACULATUS.

Only the male of this species was obtained, and so a check
count of the number of chromosomes of the female could not
be obtained. However, very good counts from spermatogonial
metaphase plates were obtained leaving no doubt as to the
number of chromosomes present. The number of chromosomes
in this animal is the same as in Sarcophilus, a member
of the same family. Of the fourteen chromosomes, twelve are
the autosomes, one is the small X-chromosome, and the other
the minute Y-chromosome (PI. 16, figs. 26 and 27).

From figures of the first meiotic division, the separation of the
X- and the Y-chromosome is seen to take place as in the other
animals (Pl. 16, fig. 28).

SUMMARY.

In the three animals studied the total number of chromo-
somes in the male is as follows:
Phascolarctus 16 (14 autosomes + XY).
Sarcophilus 14 (12 autosomes + XY).
Dasyurus 14 (12 autosomes + XY).

In the female the number of chromosomes is as follows :
Phascolarctus 16 (14 autosomes + XX).
Sarcophilus 14 (12 autosomes + XX).

Tn all animals dealt with in this paper the Y-chromosome is

very minute in size compared with the other chromosomes ;
NO. 266 Q

916 A. W. GREENWOOD

also the X-chromosome is much smaller than any of the auto-
somes.

Chromomeres are conspicuous during syndesis, early pachy-
tene, and early diplotene stages.

The early pachytene stage is followed by a late pachytene
stage in which the threads become diffuse and lose their capacity
for taking up the stain.

Except in the early merotic prophase the sex chromosome
remains compact and deeply stained and does not thread out
hke the autosomes.

In all the above animals the first meiotic division is reduc-
tional, separating the X- and the Y-chromosomes, and the
second division is equational, in each cell the sex chromosome
dividing. The spermatozoa are therefore of two kinds, one
containing an X-chromosome and the other containing a
Y-chromosome.

No further reduction in the number of chromosomes takes
place during the second meiotie division.

The Y-chromosome could not be identified during the meiotic
phase until the metaphase of the first meiotic division. At
this stage in Phascolarctus the sex chromosomes are
separate and do not form a bivalent.

The archoplasm seems to exert some influence on the
chromatin threads at synizesis and during the early pachytene
stage. In the former case the contraction takes place to that
side of the nucleus at which the archoplasmic mass is situated ;
in the latter the chromosomes are in the form of thick loops
with the ends of the chromosomes pointing towards the archo-
plasmic mass.

In Phascolarctus the Sertoli cells are very large and
possess peculiar rod-like bodies, the origin and function of
which was not arrived at. The result of experiments seem to
show that the rods are not affected by the action of digestive
fluids.

Quart. Journ. Micr. Sci. Vol. 67, N.S., Pl15.

AW Greenwood del.

ol

MARSUPIAL SPERMATOGENESIS 917

EXPLANATION OF PLATES 15 AND 16.

REFERENCE LETTERS.

A, archoplasmic mass. £, excretory body. D, deutoplasm. P, plasmo-
some, Pz, Plasmosome associated with X-chromosome. Px’, plasmosome
derived from Px. Chr, chromatoid body.

PuaTE 15.

Phascolarctus cinereus.

Fig. 1.—Female. Nucleus of a cell from the corpus luteum. The two
sex chromosomes (XX) are much smaller than any of the fourteen auto-
somes. (Bouin.)

Fig. 2.—Spermatogonial plate with fourteen autosomes surrounding
the X- and Y-chromosomes. (Bouin.)

Fig. 3.—Spermatogonial plate. The constrictions seen at the end of
some of the chromosomes are not usually present. (Bouin.)

Fig. 4.—Late leptonema, and beginning of synizesis. Only some of the
threads are shown. Chromomeres are distinct on some of the threads.
(Flemming.)

Fig. 5.—Late synizesis and syndesis. Incomplete nucleus. Showing the
pairing of the chromomeres on homologous chromosomes. Central mass
representing the synizetic contraction. (Bouin.)

Fig. 6.—Early pachynema. Thick-looped chromosomes still showing
some duplicity. (Flemming.)

Fig. 7.—Showing the condensation of the sex chromosome, also the
associated plasmosome containing deeply staining granules. C, chromatic
granule in nucleus which probably gives rise to chromatoid body. (Bouin.)

Figs. 8, 9.—Showing the presence of the second plasmosome (Pz2’).
( Bouin.)

Fig. 10.—Early diplonema, Compact X-chromosome or XY bivalent.
Paired chromomeres distinct.

Fig. 11.—Later diplonema. (Bouin.)

Fig. 12.—Metaphase plate of first meiotic division just after nuclear
membrane has disappeared. (Bouin.)

Fig. 13.—Later metaphase plate. (Flemming.)

Fig. 14.—Metaphase (side-view). X- and Y-chromosomes travelling to
opposite poles of the cell ahead of the other chromosomes. (Bouin.)

Fig. 15.—Anaphase of first meiotic division. X and Y lagging behind
on the spindle. (Bouin.)

Q 2

Be

218 A. W. GREENWOOD

Puate 16.
Phascolarctus cinereus.

Fig. 16.—Daughter secondary spermatocytes (dimorphic) with remnants
of spindle-fibres between the two cells. One cell nucleus contains a deeply
stained body—the X-chromosome. (Bouin.)

Fig. 17.—Prophase of second meiotic division, Irregular threads,
X-chromosome compact. (Flemming.)

Fig. 18.—Anaphase of second meiotic division showing division of the
X-chromosome. (Bouin.)

Fig. 19.—Anaphase of the second meiotic division showing division of
the Y-chromosome. (Bouin.)

Fig. 20.—Sertoli cell nucleus with associated rods and deutoplasm.
(Flemming.)

Fig. 21.—Sertoli cell nucleus with accompanying pale plates. The
chromatin of the nucleus is represented semi-diagrammatically. (Flem-
ming.)

Sarcophilus ursinus.

Fig. 22.—Female. Metaphase plate of follicle cell. 12+ XX chromo-
somes. (Flemming.)

Fig. 23.—Spermatogonial plate. 12+ XY chromosomes. (Bouin.)

Fig. 24.—Splitting of chromosomes during spermatogonial mitosis.
All chromosomes except two (a and b) have divided. (Bouin.)

Fig. 25.—Metaphase (side-view) of first meiotic division showing separa-
tion of the X- and the Y-chromosomes. (Bouin.)

Dasyurus maculatus.
Fig. 26.—Spermatogonial plate. 12+XY chromosomes. (Bouin.)
Fig. 27.—Spermatogonial plate. 12+ XY chromosomes, (Bouin.)
Fig. 28.—Metaphase (side-view) of first meiotic division showing the
separation of the X- and the Y-chromosomes. (Flemming.)

LITERATURE.

Bardeleben, K. von.—‘* Die Zwischen-Zellen des Siiugetier-Hoden Ps
* Anat. Anz.’, Band 13.

- Benda, C.—‘‘ Die Spermiogenese der Marsupialier”, “ Zoologische For-

schungsreisen in Australien’, Semon, Band 3.

Bronté-Gatenby, J.—‘‘ Cytoplasmic Inclusions of Germ Cells”, * Quart.
Journ. Micr. Sci.’, vol. Lxii.

Jordan, H. E.—‘‘Spermatogenesis of the Opossum”, ‘Archiv f. Zellfor-
schung’, Band 7.

Montgomery, T. H.—“ The Human Cells of Sertoli”, * Biological Bulletin ’,
xx, LOU

Moore, J. E. S., and Walker, C. E.—‘“* The Meiotic Process in Mammalia”,
‘ First Report on the Cytological Investigation of Cancer ’, 1906.

Pinter, T. S.—* Studies in Mammalian Spermatogenesis ”, ‘ Journal of
Experimental Zoology ’, vol. 35, no. 1.

Quart. Journ. Micr: Sci. Vol. 67,N.S.,PU.16.

\

\

}

a \--?

4 c

fap

rs
: ay A
a

16

AW Greenwood. del.

-

On Sexual Differentiation in the Infusoria.
By
Prof. V. A. Dogiel,

Zootomical Laboratory, University of Petrograd.
With Plate 17 and 1 Text-figure.

THE object of this preliminary note is to describe a very
curious type of conjugation which may throw light on the
process of sexual differentiation in the Infusoria—a type leading
from conjugation between two hermaphrodite individuals to
copulation between male and female specimens. In his very
lucid account of the genetics of the Ciliata, C. Dobell (1914),?
with good reason, points out that in cases of typical conjuga-
tion both the conjugants are to be looked on as herma-
phrodites which are performing cross-fertilization. Only in
Vorticellids there are males and females, and the sexual process
changes to copulation, while the dwarf-hke male perishes after
fertilizing the larger female. In other Ciliata there are no well-
proved cases of sexual differentiation of the conjugants, although
some hints at it are to be found in the works of Cull, Doflein,
and Enriques. Doflem states that the conjugants of the same
pair in Paramecium putrinum differ very much in size,
a feature which he is ready to attribute to sexual differentiation.
Enriques relates of Chilodon that the conjugants, which
are very much alike at the beginning of conjugation, become
differentiated into a longer and a shorter one during the process
of conjugation. The meaning of such conjugants—termed by
Enriques ‘male and female hemisexes ’—remains obscure, as
also the extensive discussion of the author on the subject of
sexual differentiation. The observations of Cull on ‘ incipient

1 * Journ. of Genetics,’ vol. iv, p. 131, 1914.

22.0 Vv. A. DOGIEL

sexuality in Paramecium proved to be erroneous (Jennings
and Lashley), and need no further mention. But in this paper
I propose to deseribe a case of conjugation in the Cilata where
the members of a pair show very marked differences, which may
have a relation to sex.

The species of ciliate which I have studied belongs to the
genus Ophryoscolex (order, Oligotricha ; family, Ophryo-
scolecidae). The representatives of this family lead a parasitic
mode of life in the stomach or the intestine of different
Ungulata. Ophryoscolex janus, the form described, is
a new species, found in the stomach of African antelopes,
Bubalis cokei and Madoqua sp. The antelopes were
shot by me during my expedition to British Hast Africa in
1914, on the shores of Lake Naivasha.

In order to make my description more comprehensible
I am obliged to describe in some detail the morphology of
O. janus. The ordinary or, as we shall call them, neuter
individuals of O. janus have the following structure.
The body (PI.17, fig. 1) is oblong, and approximately cylindrical,
its posterior half being a little wider than the anterior one.
The anterior end of the body bears the mouth, lying on the
top of the oral cone: the posterior end is provided with a long
and slender terminal spine. The ciliary apparatus consists of
an adoral zone (Adz) and a dorsal crescent (Dz) of mem-
branellae. This crescent in O. janus is removed very far
backwards, lying a little behind the middle of the body. The
mouth leads into a long pharynx, with a very complicated
structure. It begins with a rather narrow oral cavity, which
farther backwards widens into the pharynx proper; the
pharynx follows the right side of the body to its hinder end,
growing gradually narrower backwards. Close to the posterior
end of the body the pharynx communicates with the endoplasm.
This latter is so well defined by a thin continuous membrane that
we can term it (in agreement with other authors) the mid-gut.
The mid-gut sends out anteriorly a long conjugation outgrowth,
which lies parallel to the pharynx and during conjugation
serves as a bridge for the migration of the male pronucleus.

SEX IN INFUSORIA D1

A short tube-like rectum ends with a circular anal aperture
at the base of the terminal spme. As a characteristic feature
of O. janus we can mention the very strong development of
the inner skeleton. It is represented by a thin plate (PI. 17,
fig. 1, skp) of alveolar structure, lying under the tough super-
ficial cuticle of the body. Anteriorly the skeletal plate surrounds
the body, forming under the cuticle a sort of stiff collar, whose
opposite ends meet on the dorsal side of the animal to form
a suture-line. In the posterior half of O. janus the right
side of the plate retains its superficial subcuticular position,
while the left one separates from the cuticle and dips into the
interior, forming a wing-like outgrowth, which surrounds and
supports the hinder part of the pharynx. ‘There are special
myonemes, described in other Ophryoscolecidae by Sharp
and Braune, closely connected with the pharynx. They
form a continuous thin layer on the sides of the pharynx
supported by the skeletal plate, to whose imner surface the
myonemes are attached. The myonemes begin at the posterior
end of the pharynx, pass forwards, and, after reaching the
anterior third of the body, detach themselves from the skeletal
plate and converge to the centre of the body, following the
wall of the pharynx. In so domg this muscular layer forms
a sort of oblique diaphragm (fig. 1, D) which is concerned with
the ingestion of food particles.

The strong development of the skeletal collar removes the
nucleus and the contractile vacuoles far backwards. The
macronucleus is usually lemon-shaped and lies on the right
side of the animal, somewhat nearer to the ventral surface of
it than to the dorsal one ; in a cup-like depression of the macro-
nucleus the small ball-shaped micronucleus is situated (PI. 17,
fig. 1, Mz). A little dorsally from the macronucleus lie both
the contractile vacuoles (V,, V,), the anterior being con-
siderably larger than the posterior one.

Besides the neuter individuals I met (in both the antelopes
investigated) with conjugating pairs of O. janus. The great
majority of pairs proved, to my great astonishment, to consist
of individuals widely different in dimensions and several other

993, Vv. A. DOGIEL

morphological features—differing in several points not only from
one another but also from the neuters. Let us call them micro-
and macroconjugants, without giving to these names
the significance of sexual differentiation. The macroconjugant
(Pl. 17, fig. 2, left-hand individual) differs less than the micro-
conjugant from the neuter individuals. The macroconjugants
are individuals 82-112 » long and 42-55 » broad (at the level
of the macronucleus), the great bulk of them measuring from
90 » to 100. The posterior part of the body is somewhat
shortened and inflated in comparison with the neuters. The
most prominent differences from. the neuter individuals,
however, consist in the lack of the hind contractile vacuole
and in the character of the terminal spine. Long and slender
in neuters, the spine is short and thick in the macroconjugants,
being about one-third as long as in the neuter specimens.

The microconjugants are even more strongly modified
(Pl. 17, fig. 2, right-hand individual). They are 60-75 p» long
and 20-5 » broad. It is easy to see (Pl. 17, figs. 2 and 12) that
the general aspect of the body becomes quite different from
that of the macroconjugants, the body being long, slender,
and somewhat vermiform. ‘The difference will appear still
greater if we compare the respective volumes of both types of
individual. Thus for a microconjugant of 75 » x25 » we shall
have an approximate volume of 34,000 cubic microns, while
a macroconjugant of 90x45 will give a total of 137,000
cubic microns—that is, about four times the volume of a micro-
conjugant.

Notwithstanding its small size the microconjugant possesses
a long and slender terminal spine, of just the same length as
that of the neuter specimens. The skeletal plate, so charaec-
teristic of O. janus, is wholly lacking. This absence of the
endoskeletal plate does not remain without influence on several
other internal characters of the microconjugant. The outline
of the anterior half of the body becomes folded and wrinkled,
while in the macroconjugant it appears smooth and even.
The pharyngeal myonemes, having lost their line of insertion,
hang loosely back and form a sort of long muscular cone

SEX IN INFUSORIA 43}

(Pl. 17, fig. 2). The muscular diaphragm which is usually seen
in neuters and macroconjugants becomes obliterated. Finally,
there is a difference in the shape of the macronucleus. In the
macroconjugants and neuters the macronucleus resembles
a lemon. In the microconjugants (PI. 17, fig. 12) it is more
elongated, being at the same time rounded at both ends. Both
the micro- and macroconjugants differ from the neuters in
having only one contractile vacuole instead of two.

Taken altogether, the differences between the conjugants are
so great that the individual members of a pair could be con-
sidered as belonging to different species, were they not found
in a state of conjugation.

The conjugants adhere one to another with their anterior
ends, diverging at an acute angle from the point of conjunction.
I could follow on my slides all phases of the nuclear changes
characteristic of a typical conjugation. A cross-fertilization
takes place—both individuals interchanging their migratory
nuclei. Furthermore, I have found conjugating pairs with
a synearyon in both partners, and a large number of excon-
jugants of both kinds. The latter show different stages in the
reconstruction of a normal nuclear apparatus.

At first we find in the exconjugants two pronuclei sur-
rounded by a plasmatic halo (described by Prandtl in
Didinium, and present, I believe, in most of the Infusoria).
The old macronucleus persists in the microconjugant after
separation from its partner, while in the macroconjugant it
is already dissolved at the moment of disjunction. The double
synearyon (PI. 17, figs. 3 and 4) forms the first division spindle,
and the reconstruction of the nuclear apparatus is very simple,
following the type represented by Chilodon. ‘The first
division spindle by its fission gives rise to a pair of
nuclei which become respectively a new macro- and micro-
nucleus. An interesting feature of this division is its hetero-
polarity—tfrom the diaster stage onwards. One of the polar
swellings of the dividing synearyon is smaller and stains more
intensely with nuclear stains than the other one. It ultimately
becomes still more condensed, and develops into a micro-

294 Vv. A. DOGIEL

nucleus ; while the larger continues to swell and becomes the
new macronucleus (Pl. 17, fig. 5, Ma).

Several particularly successful preparations help us to
elucidate the further fate of the micro-exconjugants. They
complete their reorganization and return to the type of the
neuter individuals. Specimens with the nuclear apparatus
not fully reconstituted show the rudiments of a new and very
thin skeletal plate, which is finely alveolar—its alveoli corre-
sponding to those ultimately constituting the fully formed
plate. At the same time the fibres of the pharyngeal muscular
mantle apply themselves to the inner surface of the plate, and
in so doing form the pharyngeal diaphragm. Further modifica-
tion produces tiny neuter specimens, of the same length as the
microconjugants but with fully developed skeletal plates. It
might be objected that the stages described can be interpreted
in an inverse sense, i.e. that the tiny neuters by losing their
skeletal plates may become microconjugants. But there can
be no such alternative, for the formation of microconjugants
takes place in quite a different way. They arise as a result
of an unequal fission of neuter individuals. O. janus
possesses two different kinds of fission. One of them, the
ordinary one, leads to the formation of neuters. The other
one, which may be called progamic, results in the formation
of two preconjugants which differ in size and other morpho-
logical characters. The beginning of the fission is in both
cases manifested by the elongation of the posterior half of the
body and of the macronucleus as well, while the micronucleus
assumes the shape of a short spindle. At the same time the
first rudiments of the new adoral and dorsal zones of mem-
branellae (Pl. 17, fig. 6, Dz.) appear under the cuticle.

In cases of ordinary fission (Pl. 17, fig. 6) the phase just
described clearly indicates the initial stages of the building up
of a new skeletal plate in the posterior individual (PI. 17, fig. 6,
skp.,). Further phases lead to the complete formation of the
posterior individual, with fully developed skeleton and all the
characters of the neuters. As a result of fission we have two
neuters with long terminal spines. During the whole process

SEX IN INFUSORIA 995

of fission the micronucleus retains its small dimensions ; and
freshly separated daughter micronuclei are always connected
by a long fibrous strand, whose terminal swellings (i.e. groups
of chromosomes) stain deeply with all nuclear stains. In
cases of progamic fission (PI. 17, fig. 7) the posterior individual
does not show any signs of a skeletal plate from the beginning
to the end of the fission—it is the microconjugant in statu
nascendi, and gets its long and slender terminal spine from
its neuter parent. ‘The anterior individual, which gets the
skeletal plate of its parent, differs from the latter by the
shortness and thickness of its newly developed terminal spine.

The progamic fission can be easily recognized, moreover,
by the behaviour of the micronucleus, which swells to enormous
size, becoming at the same time very feebly stainable with
nuclear stains (Pl. 17, fig. 7, Mz). The micronuclei of both
the daughter individuals remain in this condition from the
moment of their disjunction to the beginning of conjugation.
Such individuals may be termed preconjugants (Pl. 17, figs. 8
and 9). From what has just been said it follows that in a
population of O. janus one can easily distinguish the
preconjugants from other individuals (neuters) which are
incapable of conjugating.

The conjugation of O. janus becomes still more interesting
from the fact that this species exhibits a certain percentage
of isogamous pairs, exclusively of the macroconjugant type.
The number of such isogamous pairs amounted, in both the
antelopes examined, to about 20 per cent. of the whole number
of pairs, the remaining 80 per cent. being of the anisogamous
type. On examining such isogamous pairs more closely, it is
seen that about 60 per cent. of them present different stages
of nuclear changes characteristic of conjugation (PI. 17, fig. 10)
from the first maturation spindle to the stage of migrating
pronuclei. In the remaining 40 per cent. of isogamous pairs
the micronuclei remain very small and lie embedded in a flat
depression of the macronucleus (Text-fig. 1) without showing
any preparation to fission. As a further peculiarity, the con-
jugants of such pairs appear to assume a position somewhat

226 Vv. A. DOGIEL

different from the normal one. The anterior end of one con-
jugant is sometimes engulfed by the introverted pharynx of
its partner (Text-fig. 1), while both the conjugants lie on the
same level without forming the characteristic angle of about
40 degrees. The long slender terminal spines of both the
conjugating individuals indicate that they are neuters. The
formation of such pairs might, however, be explained in a
different way. It might possibly be the result of a fortuitous
snapping of one individual at another in its endeavours to
engulf food particles; but such an interpretation does not
appear to me very plausible. It might also represent an
abortive attempt to conjugate on the part of neuter individuals

TExtT-FIc. 1.

A conjoined pair of neuters.

which are unable to conjugate successfully. In all the 60 per
cent. of macroconjugants conjugating inter se with success,
the terminal spine is short and thick and the posterior con-
tractile vacuole is wanting ; and it is thus evident that these
specimens have gone through the preparatory process of
progamic fission.

Such are the most important results of my investigation of
the conjugation of O. janus. A detailed general discussion
of many interesting questions arising therefrom will be given
in my full account of conjugation in the Ophryoscolecidae,
but I may now venture to draw some conclusions regarding the
more striking features of the processes described.

First of all, the question arises as to which type of conjuga-
tion is primary in O. janus—the isogamous or the aniso-
gamous ? Comparison with the other Ophryoscolecidae and
with the great majority of free-living Infusoria seems to

SEX IN INFUSORIA 997

indicate isogamous conjugation as the more primitive mode of
sexual process.

In other species of Ophryoscolex the conjugants always
possess a fully developed skeleton. It is therefore natural to
suppose that in O. janus also the primary type of con-
jugation was that between two macroconjugants formed by
equal fission of a neuter individual. At the present time this
old mode of conjugation is retained by about 20 per cent. of
pairs only, while in the rest it has been superseded by a manifest
anisogamy.

Secondly, what causes have evoked this change in the sexual
process ? This problem is extremely difficult to solve and
permits of many different explanations. One that seems to
me to be a probable one depends upon the consideration that
the change mentioned procures for the preconjugants the
advantage of being able to conjugate as rapidly as possible.
A fission accompanied by the building up of the whole complex
skeleton in the posterior individual and by the growing of the
latter to the size of a neuter, would perforce require much more
time to perform than a fission which is confined to a simple
cutting off of the posterior third of the body. The presence of
only one contractile vacuole in the preconjugants also speaks
in favour of this supposition. In ordinary fission the anterior
individual gets the first vacuole, the posterior individual the
second ; while those which are wanting are soon afterwards
formed anew. There, as we have seen, each of the conjugants
possesses only one vacuole, which it obtained at the progamic
fission : no reconstruction of the missing vacuole takes place.
Admitting that conjugation occurs in critical circumstances
menacing the existence of the population, we could . thus
understand the tendency to abridge the preparations for this
process.

Has the observed differentiation of two kinds of conjugants
the significance of sexual differentiation, and can we thus
regard the micro- and macroconjugants as males and females ?
It appears that notwithstanding all the differences in size and
structure both kinds of conjugants act as hermaphrodites.

228 Vv. A. DOGIEL

Ample evidence for this is afforded by the cross-migration of
‘male’ pronuclei, and by the reconstruction of the nuclear
apparatus and skeleton in microconjugants. Of course, we
could suppose that during the progamic fission the material of
the micronucleus is distributed unevenly, so that the male
elements are taken by the microconjugant, the micronucleus
of the macroconjugant getting only the female ones. If so, then
before and during the conjugation one specimen acts as a male,
another as a female, returning only after conjugation to the
hermaphrodite state. But there is no real evidence in fayour
of this view, and the possibility of conjugation between two
macroconjugants (in 20 per cent.) speaks against it.

But even so, we can still see in the complex processes of
conjugation in O. janus the first hint of approaching sexual
differentiation. Indeed, conjugation seems to have become
impossible between two microconjugants, and this circum-
stance raises them physiologically to quite another category
in comparison with the macroconjugants. This loss of the
faculty to conjugate inter se speaks clearly in favour of
a male tendency in the microconjugants. Then, again, certain
abnormal cases of conjugation also confirm my view of the
possible future transformation of microconjugants into males.
Among about a hundred micro-exconjugants which came under
my inspection there happened to be the followmg abnormal
specimens. One exconjugant, instead of old macronucleus +
synkaryon (or its derivatives), possessed only the old macro-
nucleus. Another one, instead of having two pronuclei or
a syncaryon, showed only one small nucleus of micronuclear
type and the remains of the old macronucleus. A few abnormal
pairs of conjugants permit us to guess how such exconjugants
arise. In one of the pairs (Pl. 17, fig. 11), for mstance, the
microconjugant contains only a macronucleus, while the
macroconjugant has got the macronucleus and four micro-
nuclei (Pl. 17, fig. 11, Mi,-Mi,). One of the latter (M2,) is
lying in the hind half of the body, another one (Mz,) just
at the line of junction of the conjugants, on the way to the
partner, while the remaining micronuclei (Mi, and Mi.) he

SEX IN INFUSORIA 229

in the endoplasmatic conjugation outgrowth of the macro-
conjugant. The only possible interpretation of such a pair is
the following one. The conjugants have reached the stage
of pronuclei. Both the pronuclei of the microconjugant have
migrated into the larger partner, while the corresponding
process in the latter was retarded so that the ‘ male ’ pronucleus
of the macroconjugant has not yet succeeded in penetrating
into the partner. And we must say that, in the pair under
discussion, a further migration of the male pronucleus (7,)
into the smaller partner would become absurd, as the corre-
sponding female pronucleus of the microconjugant (J/1,) has
penetrated into the larger individual. Let us imagine that the
separation of such a pair is accomplished, and we should then
have before us a micro-exconjugant provided with only one
nucleus, the old macronucleus (as in the above-mentioned
abnormal exconjugant), and a macro-exconjugant with a
double set of nuclei. A strong confirmation of such an inter-
pretation is supplied by a macro-exconjugant in the stage of
reconstruction of the nuclear apparatus. In this specimen
there are two micronuclei and two macronuclei instead of one
micro- and one macronucleus. There, as I believe, the pro-
nuclei of the microconjugant penetrated into its partner (as
in the case mentioned above) and copulated with the corre-
sponding pronuclei—thus producing a pair of syncarya. The
latter then divided, giving rise to a double set of micro- and
macronuclei. The micro-exconjugant with an old macro-
nucleus and a small micronucleus (see above) is to be thought
of as a specimen whose ‘ male’ pronucleus has migrated into
the larger partner, while the corresponding migration of the
macroconjugant’s ‘male’ pronucleus did not take place ; and
the exconjugant represents, therefore, an individual with its
old macronucleus and a ‘ female’ pronucleus.

These and some other abnormal cases give us grounds
enough for framing the following hypothesis. If analogous
anomalies become more common, a time may come when in
conjugating pairs only the ‘male’ pronucleus of the micro-
conjugant will migrate, the reciprocal process being suspended.

230 V. A. DOGIEL

The macroconjugant, transformed into a fertilized female,
will thus go on living and multiplying, while the microconjugant,
now to be regarded as a dwarf male, is predestined to die after
conjugation.

It is very interesting to note that the sexual phases of
O. janus remind us of the reproduction in a group of Metazoa,
namely, in some of the Cirripedia. It is well known that
several representatives of the Cirripedia possess, besides the
large hermaphrodite individuals, so-called complementary
dwarf males. I cannot help comparing the macroconjugants
of O. janus to hermaphrodite Cirripedia, while the micro-
conjugants are on the way to become complementary males.

Another poit worthy of special mention is the progamic
fission. Several species of Infusoria (Dileptus, Didinium,
Paramecium, &c.) are known to undergo, before conjuga-
tion, several ‘ hunger-divisions’. KR. Hertwig, postulating
a causal connexion between these divisions and conjugation,
says that hunger-divisions may correspond with the maturation
divisions of multicellular organisms. Still, it is uncertain how
far these divisions are indispensable for the beginning of con-
jugation, and how far the preconjugants differ from the
ordinary (neuter) individuals. All the Ophryoscolecidae
examined, and O. janus more than the rest, prove that the
fissions preceding conjugation have a peculiar character. They
have a close connexion with the commencement of sexual
reproduction and must bear the special name of progamic
fissions. The individuals -resulting from these fissions are the
preconjugants, which differ in several points from the
neuters. Only the preconjugants are able to conjugate ; and
this compels us to consider the progamic fission as a process
which may be compared with the attaining of puberty by
multicellular organisms. The individuals which are formed by
the progamic fissions are sexually mature. On the other hand,
the reduction divisions of the micronucleus during conjugation
are the real homologues of the maturation processes of sexual
cells in the Metazoa.

The same rule—that conjugation is possible only between

SEX IN INFUSORIA 981

the preconjugants—is applicable, I believe, to all the rest of the
Infusoria. We have already seen the very marked differences
which characterize the preconjugants of O. janus in com-
parison with the neuter specimens. The same differences,
though less evident, exist in other Ophryoscolecidae (in all the
three species studied by me): and some further indications
of the existence of preconjugants in other forms are scattered
here and there in different papers on Infusoria ; but I do not
intena to discuss them in this preliminary note.

The conception of sexual puberty preceding conjugation
casts a new light on the question of experimental induction
of conjugation by means of different external stimuli (hunger,
&c¢.). In all such cases the specimens treated in the experiment
evidently remain still unable to conjugate, whatever be done
to them ; but the stimuli applied to the Infusoria make them
begin the progamic fission, which produces sexually mature
individuals ready for conjugation. It is noteworthy that the
cases of ‘ reconjugation ’ prove this sexual maturity to persist,
in some exceptional cases, even after conjugation has occurred.

There is a further point to be discussed, although I do it
with some reserve. I refer to the heightened viability of the
conjugants as compared with the rest of the population. The
high mortality amongst exconjugants is already known. In
my material also dying exconjugants are often to be found ;
in Ophryoscolecidae they are easily recognized by many striking
characters, as I shall show in another note. The same symptoms
of death are observed in many (17 per cent.) of the neuters.
The conjugating pairs, on the contrary, were never found in
a dying condition. This circumstance makes me_ believe
that the stages of sexual puberty and conjugation are the most
viable. If this is so, we can easily understand why ciliates
acquire a tendency to conjugate in bad conditions of life ; for
the processes of progamic fission and conjugation would enable
the animals to acquire, for a couple of days, a heightened
resistance to harmful external conditions. It remains to test
this experimentally on some free-living Infusoria. If we accept
this hypothesis the cases of reconjugation are easily explained :

NO. 266 R

932 Vv. A. DOGIEL

since under bad life-conditions, continuing for a longer space
of time, the exconjugants of the first conjugation would hasten
to reconjugate, so as to become again more resistant to their
surrounding medium.

EXPLANATION OF PLATE 17.

All figures depict Ophryoscolex janus n.sp. and were
drawn with the help of Abbe’s apparatus, under a Zeiss 2 mm.
homogeneous immersion objective with compensating ocular
no. 4. The figures were reduced to three-quarters in re-
production.

Fig. 1.—A neuter individual.

Fig. 2.—Conjugation between a macro- and microconjugant. Each
individual has a single macronucleus and two micronuclei.

Fig. 3.—A micro-exconjugant with a synkaryon and the remains of the
old macronucleus.

Fig. 4.—A macro-exconjugant with synkaryon ; the old macronucleus
is dissolved.

Fig. 5.—A micro-exconjugant with old macronucleus (Oma), new
macronucleus (Ma), and new micronucleus (Mi).

Fig. 6.—An ordinary fission of O. janus giving rise to a pair of neuters ;
note the skeletal plate (Skp,) in the posterior individual.

Fig. 7.—A progamic fission of O. janus; note the absence of a skeleton
from the posterior individual and the large size of the micronuclei (7).

Fig. 8.—A micro-preconjugant.

Fig. 9.—A macro-preconjugant.

Fig. 10.—Conjugation between two macroconjugants.

Fig. 11.—An abnormal case of conjugation ; all the micronuclei (M/7,-
Mi,) lie in the macroconjugant, the microconjugant retaining only the
old macronucleus.

LETTERING.

Adz, adoral zone of cilia. Cpr, conjugation process of endoplasm.
D, oblique diaphragm. Dz, dorsal crescent of membranellae. Ma, macro-
nucleus. J/i, micronucleus. Oma, old macronucleus. Phm, pharynx.
f, rectum. Skp, skeletal plate. Sy, synkaryon. Vj, anterior vacuole.
V.,, posterior vacuole.

é Uf ly

ne

Oveurs
BY i ; Roe
PS SAP

PL
S

SS ————

WD
Za
nwo
aco | [Sees ay ai

V. A. Dogiel, del.

Quart. Journ. Micr. Sct. Vol. 67, N.S, Pl. 17

ig
te
(2

aes

ES
> a
fo

ee Se

ea

Sanguinicola from the Sudan.

By

W. N. F. Woodland,

Wellcome Bureau of Scientific Research, 25-7 Endsleigh Gardens,
Euston Road, London, N.W. 1.

With Plate 18.

In 1908? the description by Dr. Marianne Plehn of Sanguini-
cola as the only known example of a Cestode (Cestodarian)
infesting the blood aroused considerable interest, and though
the subsequent conclusion by Odhner in 1911 2 that Sanguinicola
was rather to be regarded as an extreme form of suckerless
haematobic Trematode somewhat detracted from the signifi-
cance of the discovery, yet on account of its plesio-Turbellarian
structure and possible Turbellarian affinities, Sanguinicola still
remains a zoological type of importance.

So far as I am aware, Sanguinicola has, up to the present,
solely been found in Germany in the blood of Cyprinidae,?*
and it was therefore a source of gratification to me to discover
in the large collection of slides of Helminth material made by
the late Dr. A. J. Chalmers when Director of the Wellcome
Tropical Research Laboratories at Khartum and kindly
presented to the Wellcome Bureau by his successor, Major
R. G. Archibald, some thirty specimens of Sanguinicola
obtamed from the blood in the heart of the Nile Siluroids,

v + “ Kin monozoischer Cestode als Blutparasit (Sanguinicola armata
u. inermis, Plehn)”, ‘ Zoologischer Anzeiger’, Bd. xxxiii, p. 427,
1908-9.

| ® T. Odhner, “ Sanguinicola M. Plehn—ein digenetischer Trematode ! ”
ibid., Bd. xxxviii, p. 33, 1911.
J 3 Max Liihe, “ Parasitische Plattwiirmer II: Cestodes” in Brauer’s
‘Die Siisswasserfauna Deutschlands ’, Heft 18, 1910.

R 2

234 — W. N. F. WOODLAND

Synodontis schall, Bloch-Schneider 1801, and Aucheno-
glanis occidentalis, Cuv. and Val. 1840. Since I have
not examined personally the Sanguinicola armata and
S. inermis of Plehn, I am unable to say, especially in view
of the somewhat diagrammatic figures and, in the case of
certain organs, discordant accounts of the anatomy of these
two species, whether the Sanguinicola from the Sudan is a
distinct third species or not, but its structure is evidently
sufficiently similar to enable me to venture to correct and
amplify in some respects the descriptions supplied by Plehn
and to furnish some additional evidence in connexion with the
possible affinities of this organism.

My material consisted in all cases of specimens already
stained and mounted in balsam. Twenty were from Aucheno-
glanis occidentalis and fourteen (including one very
young form, three cut into horizontal and transverse sections,
and four which I lost during restaining but which I had pre-
viously examined) from Synodontis schall. In length
the adult or nearly adult specimens varied between 647-4 and
1,228-4 microns in length, and all were of the‘ armed’ (i.e.
bearing spinelets on the edge of the body) type.

One external feature which Dr. Plehn has not described for
S. armata and §. inermis is the presence on the surface
of the body which does not carry the genital openings of
a furrow or groove, shallow over the greater length of the
body but deep posteriorly and terminating just anteriorly to
the hind end of the animal in a distinct pocket (Pl. 18, figs. 3, 7,
Z, Y¥,;X, W, V). This furrow or groove is formed by the inturning
and posterior fusion of the edges of the body, is apparently
a permanent formation judging by the posterior pocket, and
is similar in form to the similar body grooves to be found in such
Trematodes as Hemistomum clathratum and the male
Schistosoma, but differs in that it does not contain the sexual
apertures and therefore cannot be of use in copulation. The
edges of the body which border this furrow contain the spinelets +

1 If mounted specimens of Sanguinicola be not well flattened out,
spinelets will appear to be absent on the ‘ edges’ of the body and can only

SANGUINICOLA 935

or ‘ Hiikchen’ described by Plehn, except posteriorly where the
edges turn mediad in order to unite (PI. 18, fig. 1, sp). The spine-
lets are distinct rods of the form shown in fig. 9, and are almost.
entirely embedded in the subcuticula and parenchyma save
for their outer extremities, which project slightly. These
spinelets attain a maximum length of about 27 microns, and
they are apparently identical in nature with the chitinous spines
found in the skin of many Malacocotylea.

For the reason to be supplied later when describing the genital
apparatus, I shall consider the side of the body bearing the
furrow as the dorsal surface, from which it follows that the
genital openings are situated, as in most Turbellaria and
Trematoda, on the ventral surface.

In general shape (Pl. 18, fig. 1) the Sudan Sanguinicola
resembles the Sanguinicola described by Plehn. As Plehn
describes, and as is evident even in preserved material, the
body is highly contractile and in much contracted specimens
may be a broad oval in outline.

As regards the internal organization, the two chief systems
of organs to be described are the gut and the genital apparatus ;
T have nothing to add to Plehn’s description of the nervous and
excretory systems and the general histology.

The gut, as shown in figs. 1, 4, 5, 6, is in a much reduced
condition as compared with the gut in most Turbellaria and
Trematoda; but that it is a gut and not a ‘frontal gland’ is
shown not only by the presence of a well-marked muscular
sucking pharynx (pH) and by the non-glandular character of
the wall of the gut sac (Gs), but also by the occasional presence
of distinct blood-corpuscles in its lumen (seen in two of my
specimens). The anterior mouth-opening (m) is, as in all

be detected by careful focusing; also in young specimens the spinelets
are smaller than in full-grown forms and in very young forms are absent.
I hardly like to suggest that a conjunction of these two conditions is
responsible for the description of the 8. inermis of Dr. Plehn, but it
appears to me that the suggestion might be applicable. The difference
of size between the two species S. armata and S. inermis quoted
by Lihe is of no specific value.

236 W. N. F. WOODLAND

blood-sucking animals, extremely minute, and this leads into
a short, very narrow, though distensible channel opening into
the thick-walled (when empty) elongated pharynx (pH) which
extends posteriorly as far as the transverse nerve-commissure
(nc). This pharynx is capable of great distension (PI. 18, figs. 4,
c,d). Its wall, when not distended, is of considerable breadth
and is transversely striated, and I believe is covered externally
by a layer of cytoplasm containing relatively few nuclei (judging
from the appearance seen in one of my specimens, on which
fig. 5 is based) ; but of this Iam not quite certain, and Dr. Plehn
may be correct in stating that nuclei are absent in this region
of the gut. In most preparations it is difficult to distinguish
such nuclei from the nuclei of the layer of muscle-cells (PI. 18,
fig. 1, pmus) which is connected with the outer wall of the
pharynx. From the posterior end of the pharynx to the gut
sac is a narrow oesophagus (o£), with a wall much thinner than
that of the pharynx and covered externally with a layer of
nucleated cytoplasm. The oesphagus apparently hes ventral
to the nerve commissure. The gut sac (Gs), situated at the
hind end of the anterior third of the body, is somewhat irregular
in shape, but is more or less compact and possesses none of the
distinct four or five lobes figured by Plehn. Its wall consists
of a thin, occasionally nucleated, layer of granular cytoplasm.
Dr. Plehn, in her first description * of Sanguinicola as a Turbel-
larian, naturally regarded the anterior canal and sac as a gut ;
but she states that it only contains a ‘ feinen Brei’ and that
blood-corpuscles are never found in the lumen, because the
mouth and anterior canal are too narrow to admit of their
entrance. But the mouth and all parts of the canal are capable
of considerable distension, as is shown in my specimens, and in
two cases I have found distinct rows and small masses of
evident fish blood-corpuscles in the anterior end and hind part
of the pharynx and in the sac. There is thus no question con-
cerning the gut-nature of the anterior canal and sac, and this
fact definitely settles the non-Cestode nature of Sanguinicola.

2 “Sanguinicola armata und inermis (n. gen. n. sp.), n. fam.

Rhynchostomida. Ein ento-parasitischer Turbellar im Blute von
Cypriniden ”, ‘ Zoologischer Anzeiger’, Bd. xxix, p. 244, 1905-6.

SANGUINICOLA 987

Sanguinicola is hermaphrodite and its reproductive organs
have been well described by Dr. Plehn in her first paper,?
but, judging by my specimens of the Sudan Sanguinicola, she
has erred in several respects in her second revised description
of Sanguinicola as a monozoan Cestode, as I shall indicate
later.

The ovary (ov) in the Sudan Sanguinicola is bipartite in form
and is extensive, occupying the margins of nearly the whole
of the anterior two-thirds of the body, external to the testes,
eut sac, and pharynx (PI. 18, fig. 1). The oviduct (ovp) is
median and, according to the convention I have already
adopted, dorsal in position, 1.e. on the side of the body bearing
the longitudinal furrow (PI. 18, fig. 2). Ovarian follicles are to
be found anteriorly at the sides of the pharynx (interspersed
with the muscle-cells surrounding this organ), and they extend
posteriorly, on each side, opening into the median oviduct
by some eight or more transverse ducts (hidden under the testes
in fig. 1). Posteriorly to the last pair of transverse ducts, the
oviduct, in a dorsal view (PI. 18, fig. 1, shows the ventral aspect),
turns to the right above the median vas deferens; then,
becoming much dilated, bends again to the left and somewhat
posteriorly, passmg ventral to the vas deferens, and finally,
becoming narrow, runs posteriorly on the left-hand side, to
open by a narrow aperture into the distinct spherical fertiliza-
tion chamber (orp) on its posterior aspect. From the anterior
side of the fertilization chamber and by a similar small opening,
arises the vagina (vAGN), which is short, narrow at first, and wider
after (vaaB), and opens on the ventral surface by a circular
vaginal pore (vacp), which lies at the end and at the bottom
of an elongated ridged muscular oval groove (vaGrR), directed
to the left and posteriorly, the vaginal groove (Pl. 18, figs. 1, 7,
Ty, 7x, 8). The course of the oviduct I have above described
is constant in all my specimens. I may also mention that the
oviduct, fertilization chamber, and vagina are all distensible
structures, but that I have never observed the openings
into and from the fertilization chamber to be otherwise
than narrow.

1 Tbid.

938 W. N. F. WOODLAND

The fertilization chamber is a quite distinct permanent
spherical dilatation, sharply demarcated from both the oviduct
and the vagina by the narrow openings just referred to, and
with distinctive thick walls. It usually contains a large cluster
of eggs, and eggs of course are usually to be seen in the oviduct,
both anteriorly and posteriorly, but I have not been able to
detect either eggs or spermatozoa in the vagina, though the
eggs must make their exit by this duct and spermatozoa must
enter. The eggs (Pl. 18, fig. 10) in the fertilization chamber
(which measure in diameter from 5-2 to 5-6 microns) and oviduct
appear to be fully mature, and are certainly not the mere
accidentally injected yolk-cells postulated by Dr. Plehn. One
trifling and yet in a way important point to mention is that
the muscular walls of the dilated portions of the oviduct,
situated just below the ovaries, as well as the walls of the vas
deferens, show a marked longitudinal striation, and this stria-
tion, seen in optical or actual section in slides stained with
haematoxylin, bears a superficial resemblance to a mass of
spermatozoa, and it was this appearance I imagine which led
Dr. Plehn to assume that the portion of the oviduct behind
the ovaries represented a vagina. Starting from this assump-
tion, Dr. Plehn was logically led into the further assumptions :
(1) that a second female duct was present—-the duct she
labelled * Uterus’ and supposed by her to be continuous in
the young animal with the duct she labelled ‘ Dottergang ’
(Pl. 18, fig.11);* (2) that an ootype must be situated between
the bases of the two ovaries, but, it not being visible, it was
necessary to assume (3) that the animal was not mature, and
in view of this immaturity, (4) that the eggs plainly to be seen
in the oviduct and in my fertilization chamber were only yolk-

1 Dr. Plehn says in her second paper that she had previously failed to
observe the ‘ganz typische Cestodenvagina’ full of spermatozoa, but
a comparison of the figures in her first and second papers proves that the
* Dottergang-uterus ’ is the new second female duct figured and not the
‘vagina’. In her first paper the ‘ vagina’ (i.e. the oviduct) is shown
clearly (and correctly) in all its winding course and full of eggs, not sperma-

tozoa ; whereas, in her second paper, it is represented as of the same
form but devoid of eggs.

SANGUINICOLA 939

cells which had become prematurely or accidentally shed into
the ‘ uterus ’.

Concerning the existence of vitellaria as distinct from ovarian
follicles, IT am unable to speak with certainty from actual
observation, since, as Dr. Plehn admits, they are not in any way
distinguishable, and I am willing to allow that, if vitellaria exist,
the upper part of the oviduct may function as a vitelline duct,
but as regards the existence of a separate female duct, i.e.
combined ‘ Dottergang’ and ‘ Uterus’ (Pl. 18, fig. 11) in addition
to the oviduct figured by Dr. Plehn in her original communica-
tion (and by me in fig. 1) and relabelled * vagina ’ in her second
paper, I am absolutely certain that, in the Sudan Sanguinicola,
such a second female duct does not exist (as I can demonstrate
in both whole-mounted specimens and in horizontal and trans-
verse sections) and, since there is no shell-gland and Dr. Plehn
did not observe this second female duct when first describing
the genitalia, 1 feel tolerably certain also that it does not exist
in the German Sanguinicola.t Further, the non-existence of
a separate vitelline duct almost implies the non-existence of
vitellaria, as also does the presence of globules of presumably
some sort of food material in the periplasm of the eggs (PI. 18,
fig. 10). As regards the existence of an ootype, a true ootype,
i.e. a chamber into which the ducts of the vitellaria and shell-
gland open, obviously cannot exist, since both vitellaria and
a shell-gland are absent ; but if it did, it would presumably be
found next to the fertilization chamber which I have described
and not between the ovaries. Jinally, I can see no reason to
suppose that the eggs found inside the fertilization chamber
are not mature: their position alone certifies their maturity.

In short, the female reproductive system of Sanguinicola is
in essentials constructed upon the plan found in many Rhabdo-
coelida and Polycladida, in which vitellaria are also absent and
in which the oviduct is also long and opens directly to the
exterior, its end part being differentiated into a fertilization

1 Tf this second female duct does exist in the German Sanguinicola, then

this is radically different from the Sudan form, but this I am unable
to credit without further evidence.

240 W. N. F. WOODLAND

chamber or passage and ootype when present and the ‘ antrum
femininum ’ serving as vagina, and Dr. Plehn’s first idea of
regarding Sanguinicola as an aberrant and much modified
Turbellarian has still much to be said in its favour, both from
the points of view of the genitalia and the gut. Odhner’s com-
parison of Sanguinicola with certain Malacocotylea was of
course based on Dr. Plehn’s second but, as I believe, erroneous
description of the genitalia of Sanguinicola.

The testes (TEs) are bounded anteriorly by the gut sac, and
laterally and posteriorly by the ovaries (PI. 18, figs. 1, 2). They
consist of large ovoid or spherical capsules connected with the
main vas deferens in the median line by transverse ducts.
The main vas deferens (vp) is conspicuous in stained prepara-
tions by reason of its longitudinally striated muscular walls,
and anteriorly it les ventral to the oviduct. In occasional
specimens there appear to be two or three main longitudinal -
trunks of the vas deferens anteriorly instead of one. Imme-
diately posterior to the ovaries the vas deferens lies below the
oviduct, but turning to the right (viewed dorsally) it passes
above the dilated oviduct, then, becoming considerably
dilated, bends sharply to the left (lying parallel with and close
behind the dilated limb of the oviduct), and then, as sharply
bends again to the right, where it enters the penis which lies
posteriad and to the left, and opens to the right of and slightly
behind the opening of the vagina. The directions in which the
penis and the vagina lie are approximately at right angles to
each other, and if we imagine two Sanguinicola to apply their
ventral surfaces together, then the direction of the penis of
one Sanguinicola will coincide more or less with the direction of
the vagina of the other, one (or both) of these Sanguinicola
perhaps holding on to the wall of the blood-vessel by the spiny-
edged furrow of the opposite surface.

During copulation—a process which I assume to result in
a mutual exchange of spermatozoa between two Sanguinicola—
the spermatozoa must, in each animal, traverse the vagina and
reach the fertilization chamber, where fertilization occurs.
After fertilization I assume that the eggs are extruded through
the vaginal pore into the blood. Since these eggs are devoid

SANGUINICOLA 241

of shells, it seems to be evident that the eggs must be removed
from the blood by the agency of some such external blood-
sucking parasite as (in the case of the German Sanguinicola
of the carp) the carp-louse, Argulus foliaceus, or (in
the case of the Sudan Sanguinicola) other species of Argulidae,
or perhaps a leech. In Odhner’s sketch of the possible life-
history of the German Sanguinicola he omitted to offer any
suggestion as to how the fertilized eggs reached the Molluscan
intermediate host postulated by him from the fish, other than
stating that he found a ‘ mature ’ egg, containing a miracidium,
in the kidney ofa carp, but to me it is difficult to believe, without
further evidence, that a miracidium can develop from a shell-
less egg in blood and thence take to water. All animals (except
perhaps sponges and other very low forms of aquatic life)
which extrude eggs into water, even when the animals them-
selves are living in water, and even when the eggs are not
fertilized until after contact with the water, protect the eggs
with shells or envelopes of one kind or another, and much more
should this be the case in parasitic forms (like Schistosoma)
which have to extrude the eggs first into vertebrate blood and
then into open water.

As regards the external vaginal groove, this is possibly
connected with the insertion of the penis, though it is difficult
to understand on that hypothesis why its direction should be
at right angles to that of the vagina and not in line with it.

Finally, I must state that I have adopted the convention of
regarding the surface of the body bearing the longitudinal
furrow as the dorsal side, because it is usual in both Turbellaria
and Trematoda for the sexual openings to be situated ventrally.
The fact that in certain Trematodes a similar furrow is situated
on the ventral side is but of little importance in this connexion
because it is probable that the furrow in Sanguinicola is a special
structure adapted to life in blood-vessels, and Iam by no means
convinced, in view of the Turbellarian conformation of the
genitalia in Sanguinicola, that Odhner is right in regarding
Sanguinicola as a Malacocotylean, despite the analogies (of the
gut and habitat) with Aporocotyle and Deontacylix, though it
is true that the genitalia of Sanguinicola may have become

249, W. N. F. WOODLAND

secondarily simplified. The evidence of Looss? to the effect
that a cercaria stage occurs in the life-history of Sanguinicola
requires confirmation before a decision can be arrived at,
and if it be confirmed, it will be of interest to know how the
naked eggs liberated into the blood of the fish are transported
to the intermediate Molluscan host.

DESCRIPTION OF PLATE 18.

o7, male aperture. 9, female aperture (vaginal pore), ‘DG’, ‘ dotter-
gang’. EE, narrow opening from fertilization chamber into vagina.
EGG, ripe eggs. GS, gut sac. M, mouth. NC, nerve commissure. OE, Oeso-
phagus. 000, narrow opening of oviduct into fertilization chamber.
otp, fertilization chamber. ‘orp’, hypothetical ootype. Ov, ovary.
OVD, oviduct. oOvpB, dilated oviduct. OVDN, narrow portion of oviduct.
PEN, penis. PH, pharynx. PMUS, muscle-cells surrounding pharynx.
SOL, solid posterior extremity of body behind dorsal furrow. sp, spinelets.
TEP, terminal excretory pore. TES, testes. ‘UT’, ‘uterus’. VAG, vagina.
“vaG’, ‘vagina’, vaGs, dilated part of vagina. VAGN, narrow part of
vagina. VAGP, vaginal pore (?). VAGR, vaginal groove on surface. VD, vas
deferens. vr, dorsal body furrow. v-v, W-w, X-X, Y-Y, Z-Z, planes of
sections across fig. 7 corresponding approximately to the figures of actual
sections in figs. 7, v, W, X, Y, Z.

Fig. 1 (x cir. 87).—Sanguinicola viewed from the ventral aspect.

Fig. 2 (x 180).—Transverse section across Sanguinicola behind the gut sac
(the convex surface is ventral).

Fig. 3 (x 39).—Sanguinicola in dorsal surface view.

Fig. 4.4 (x cir. 120).—Unusually small contracted pharynx.

Fig. 4, b, c, d (x cir. 120).—The undistended (6) and distended (c, d@)
pharynx in three Sanguinicola.

Fig. 5 (x 530).—The undistended pharynx.

Fig. 6 (x 530).—The gut sac and opening of oesophagus.

Fig. 7 (x 260).—The posterior genital ducts and openings from the
ventral aspect. V-Vv, W-W, X-X, Y-Y, Z-Z represent approximately the
planes of the sections shown in fig. 7, v-z (x 180).

Fig. 8 (x 530).—The posterior genital ducts and apertures magnified
to show the character of the walls of the former.

Fig. 9 (x 840).—The spinelets embedded in the edge of the dorsal body
furrow.

Fig. 10 (x 840).—Eggs contained in the fertilization chamber.

Fig. 11.—Diagram to show the genital ducts as figured and described
by Dr. Plehn in her second paper (1908), for comparison with fig. 1.

? Quoted by Odhner,

Quart. Journ. Micr. Sci. Vol. 67, N.S., Pl. 18

WN. F. Woodland, del.

wa > = :
a YY -_
: * me wes)
- . Fax >
é ad
.
J s
¥ : '

On Centropygus joseensis, a Leech
from Brazil.

By

Charles Badham, B.Se., M.B., Ch.M.,
Assistant Microbiologist, Board of Health, New South Wales.

With 10 Text-figures.

Durine the Percy Sladen Expedition to Brazil in 1918
Professor J. P. Hill collected six specimens of land leeches,
and while I was on leave from the Australian Imperial Force
in France, and acting as Demonstrator in his department, he
requested me to describe them. I was glad of this opportunity,
for these leeches, themselves earthworm-like in appearance,
were easy to determine as close relatives of Lumbrico-
bdella, that interesting leech from Brazil, which has so closely
copied the form and habit of an earthworm.

I have determined these specimens as

Centropygus joseensis (Grube et Oerstedt, 1859).
Syn. Centropygos joseensis (Grube et Oerstedt, 1859).
Centropygos jocensis (Grube et Oerstedt, 1859).
Cylicobdella lumbricoides (Grube, 1871).
Nephelis tergestina (R. Blanchard, 1892).
Liostomum joseense (Grube et Oerstedt, 1859; R.
Blanchard, 1896).
In view of the want of any note (save Weber, 1914) on this
leech, made in its natural state, the following extract from
Professor Hill’s diary is of value :

State of Rio, Brazil, at Government Orchard, Macieiras,
altitude 1,500 metres. Found two species (?) of leeches, first

944 CHARLES BADHAM

one under clod of earth coiled knot-like, bright red in colour
darkening to the posterior end; second specimen under a
stone. A third specimen found in earth on dislodging a buried
log, darker in colour, and probably belongs to the same species
as a fourth specimen, a much larger leech and slaty black in
colour. On following day found two more land leeches similar
to the first.

T have laid stress on this description, for the next year Weber
(1914), describing certain leeches from Columbia as “Centro-~
py gus joseensis ’, for the first time mentions their blood-red
colour.

The specimens collected by Professor Hill consist of five small
and one larger leech. The former measure from 40 to 80 mm.
long by 3 to 3-5 mm. in diameter—the latter single specimen
is 130 mm. long by 7 mm. diameter.

HISTORICAL.

The genus Centropygus was first established by Grube
in 1859 to contain a leech which Blanchard thinks came from
San José near Panama, the generic name being given because
the anus was erroneously supposed to open in the centre of the
posterior sucker. In 1871 Grube described the genus Cyclico-
bdella in which he placed C. lumbricoides, a leech from
Desterro.

Blanchard, who has examined these two types, found them
to belong to the same species, and in this species placed also
a leech described by him as Nephelis tergestina, and
gave the generic name Liostomum priority.

Later, he restored the name Centropygus to the genus
which contains ©. joseensis from ‘'l'rinidad, and another
characterized by its blood-red colour described by Kennel
in 1886 as C. coccinea, the colouring of C. joseensis
being still unrecorded.

Kennel (1886) gave an excellent account of this species, and
separated it from C. lumbricoides on certain anatomical
details which I will mention later.

Weber (1914) was the first to record the colour of living
specimens of C. joseensis, and for want of better informa-

CENTROPYGUS JOSEENSIS 945

tion he described the smaller of these specimens as C. coc-
cinea. He laid stress on the fact that apart from size he could
distinguish no difference in the specimens: the form, number
of annuli, and the position of the genital pores were the same,

small variations occurring
in the number of annuli, as
was already mentioned by
Kennel (1886) and Blan-
chard (1896).

Kennel (loc. cit.) separated
the species C. coccinea
from C. joseensis for the
following reasons: in OC.
coccinea the size is
smaller, the ovaries lie ven-
tral to the gut, between if
and the nerve-chain, and
the anterior part of the mid-
gut has no blind sac at its
transition into the mid-gut,
whereas in C. joseensis
as well as the greater size
of the leech, the ovaries lie
under the lateral blood-
vessel and a blind sac is
present. He admits that all
the specimens of C. coc-
cinea he examined were
sexually immature, but does
not think that this would
account for the ditferences

Trxt-Fic. 1.

Lateral view of C. joseensis from
a preserved specimen 42 mm. long.

enumerated. ‘lo determine the position of the specimens
I dissected one, and cut serial sections of another, and I will
be able largely to confirm Kennel’s work and bring it up to
date, especially as regards the relations of the ganglia and
annuli, and to add certain new details.

246 CHARLES BADHAM

DISTRIBUTION.

Blanchard (1896), describing members of this species collected
by Borelli in Paraguay and Uruguay, says: ‘ This species is
widely distributed in Central America, and has been collected
in Rio de Janeiro, Sao Paulo, Para, the basin of the Xinqu,

TEXT-FIG. 2.

Anterior end of C. joseensis from the ventral surface to show
the annulation and genital openings.

from San Bernardo, Paraguay, from Chiriqui (Central America),
from Caracas, Puerto Cabello, Venezuela, Santa Catharina
(Brazil), and Rio Grande do Sul.

‘This leech also extends very far towards the Hast. It is
recorded from Antisana, Ecuador, that is to say to the very
middle of the Eastern foothills of the Andes. Does this species
cross the mountains, and is it found on the Western slope ?
Nothing is known to this effect yet, but it is not impossible
since other species can live on either slope of the Cordillera.’

CENTROPYGUS JOSEENSIS 247

ANNULATION AND MEASUREMENTS.

In dealing with the annulation of these leeches it is necessary
to define precisely the method adopted in the enumeration of
the annuli. The earlier workers counted as first that annulus
which completely surrounded the body, so that the male pore
was given as opening between the 26th and 27th annuli.

To make their enumerations tally with those of recent workers,
there must be added the four annuli on the dorsal surface of
the anterior sucker and the annulus which is immediately below

TEXT-FIG. 3.

OH
Lateral view of the posterior end of C. joseensis. A., position
of anus. P.s., posterior sucker.

these divisions but is incomplete ventrally, where it forms the
lateral boundaries of the mouth cavity.

Again, in regard to the position of the anus, which both
Blanchard and Kennel show as between the annuli which are
second and third from the posterior sucker, I find that in only
two of my specimens is there any indication of the last annulus
which marks the dorsal surface of the posterior sucker (Text-
fig. 3).

T have shown in Text-fig. 4 the relations of the nerve ganglia
to the annuli, and defined the somites on a neuromeric basis.
This figure is drawn from the dissection of a specimen 80 mm.
long, and certain details added from the serial sections of
a specimen 60mm. in length. The six specimens which I
examined ranged in length from 40 to 180 mm., and in breadth
from 3 to 7 mm.

NO. 266 S

CHARLES BADHAM

248

+
iS)
=
i
H
ca
&

a

CENTROPYGUS JOSEENSIS

TEXT-FIG. 4.

Diagram of Centropygus joseensis showing the annuli
and their relation to the nerve ganglia, the somite limits, the
alimentary reproductive and nephridial systems as seen from the
dorsal surface of a dissected specimen. WNe.gn., nervia ganglia
and somite limits. Nph.p. 3, nephridiopore 3, &c. Nph.p. 19,
nephridiopore 19. Oes., oesophagus. A.g., anterior gut. A.m.g.,
anterior portion of mid-gut. JV/.g., mid-gut. H.g., hind-gut.
B.g., blind pouch of anterior portion of mid-gut. Sph.,, &c.,
sphincters of anterior part of mid-gut. T7'.,, testes, first pair.
31, 32, male opening. 33, 34, female opening.

TExtT-Fic. 5.

Diagram of the reproductive system of Centropygus
joseensis as seen from the dorsal surface of a dissected speci-
men. 7’.,, first pair of testes, &c. V.d., vas deferens. V.sem.,
vesicula seminalis. D.ej., ejaculatory canal. D.ej.t., terminal
portion of ejaculatory canal, Sp.s., spermatophore sac. Ov.,
ovary. 11, 12, &c., nerve ganglia.

$2

250 CHARLES BADHAM

The total number of annuli varies from 102 to 104. In all the
male pore is placed between the 31st and 32nd annulus, and
the female between the 83rd and. 34th annulus (Text-fig. 2).

ALIMENTARY SYSTEM.

The month placed at the base of the spoon-shaped anterior
sucker is bounded in preserved specimens by two well-marked
lobes, which abut on the ventral ends of the 5th annulus
(Text-fig. 2). These lobes are a very characteristic feature in all
the preserved specimens I have examined ; but an examination
of sections leads me to believe that they are partly of an
oedematous nature—several longitudinal folds furrow the
concavity of the anterior sucker.

The pharynx has a well-developed musculature of longi-
tudinal circular and radial fibres, and from it the oesophagus
extends and passes into the anterior gut which in its turn gives
place to the anterior portion of the mid-gut in the 14th somite.
There is a narrowing of the lumen at this point caused by
convolutions of this part of the gut. The anterior part of the
mid-gut extends to the 21st somite; its relations to other
structures, the sphincters surrounding it, and the blind-gut
which arises from it are shown in Text-fig. 4 (Sph.,,, B.g.)-

These last two features—the sphincters and the blind-gut—
are of considerable importance in distinguishing (. joseensis
from C. coccinea. There are five sphincters, one placed
at the end of each somite from the 15th to the 18th ; the 5th
sphincter is in the 21st somite immediately behind the blind-
gut (Text-fig. 4, Sph.,.;). Hach consists of a layer of circular
muscle-fibres surrounding the gut ; in preserved specimens, being
relaxed, they produce little constriction. The sphincter in the
21st somite is particularly well developed. .

Another feature of systematic importance is the development
of a single blind-gut which occurs in C. joseensis but not
in C. coccinea. This is given off at the termination of the
anterior portion of the mid-gut in the 21st somite; it extends
to the 24th somite, lying to the right of the mid-gut.
Kennel (1886) figures the blind-gut as coming off on the right

CENTROPYGUS JOSEENSIS 951

of the anterior part of the mid-gut, but I have seen it coming
off on the left in a specimen of 60 mm.

In structure the blind-gut resembles the anterior part of the
mid-gut, of which it appears to be a backward prolongation,
but no lymph spaces surround it and its circular musculature
is better developed (Text-fig. 9, B.g.).

The mid-gut extends from the 20th somite to the beginning
of the 24th somite; it is not surrounded by a lymph space

TEXT-FIG. 6.

Transverse section of Centropygus joseensis at the level
of the male opening, showing the junction of the two horns of the
terminal portions of the ejaculatory duct to form the spermato-
phore sac and parts in relation thereto. (x 30.) <A.g., anterior
gut. Ne.c., nerve-cord in ventral lacuna. £j.d., ejaculatory duct.
£j.d.t., terminal portion of ejaculatory duct. Sp.s., spermato-
phore sac. &., male bursa.

like its anterior part. At its termination on the 24th somite
there is a sphincter which separates it from the hind-gut.
The anus is placed above the 27th nerve ganglion.

REPRODUCTIVE SYSTEM.

There are ten pairs of testes, a pair being placed within the
1st and 2nd annulus in each somite from the 16th to the 25th
(Text-figs. 4, 5, T.,, &c.). Kennel (loc. cit.) states that he has
found twelve pairs of testes, but in my serial sections of a

252 CHARLES BADHAM

60 mm. specimen and in a dissection of an 80 mm. specimen
there are only ten pairs. Still, it is not an uncommon thing in
leeches for a pair or more testes to be wanting, and Kennel
shows in his figure that the 1st and 3rd testis of the right side
are wanting. Hach vas deferens lies ventral to the testes and
runs anteriorly from somite 25 to somite 12; here it turns
medially and then runs posteriorly to the 18th somite, where

TEXT-FIG. 7.

Big aaa
Le Vd Nec.

Transverse section of C. joseensis at the level of the 16th somite,
showing the relation of the various systems. (x 30.) J.1., intestinal
lacuna of the anterior portion of the mid-gut. Ov., ovary. L.t1.,
lateral lacuna. 7'., testes, first pair. V.d., vas deferens ascending
part under testis, descending part ventral to ejaculatory duct.
Ej.d., ejaculatory duct. Ne.c., nerve-cord.

it turns and passes forward as the vesicula seminalis, a con-
voluted tube packed with sperms, which lies with the vas
deferens in connective tissue dorsal to the ventral lacuna.
The vesicula seminalis develops muscular walls and becomes
the ejaculatory duct in somite 15. The ejaculatory duct of
either side pursues a convoluted course until it reaches its
glandular terminal portion ; this part joins with its fellow and
their distal ends form the spermatophore sac which opens into
the male bursa (Text-fig. 6, Sp.s.).

CENTROPYGUS JOSEENSIS 253

The ovaries (Text-figs. 5, 7, Ov.) are placed in a connective
tissue which lies close to the testes. They are much convoluted
and extend on either side from somite 12 to somite 16;
anteriorly they unite at the female atrium.

This reproductive system shows certain affinities with that of
other Herpobdellids, but there is not seen the conducting
tissue as described by Brumpt (1899) for H. atomaria.

NEPHRIDIA.

There are nineteen pairs of nephridia placed in somites 7 to
25. As usual the nephridiopores open on the latero-ventral
aspect of the annulus preceding that in which the nerve
ganglion lies.

The general structure of the nephridium follows that seen in
other members of the Herpobdellidae.

Ciliated funnels have not previously been found in Centro-
pygus joseensis. Kennel was unable to find them; he
says that ‘in spite of careful examination of serial sections
I have never seen a picture which would allow one to conclude
the presence of a widely open funnel, and if any opening exists
it should be fairly narrow’. These ciliated organs are found
in each somite, and in the testicular region one is placed in
a dilatation of a branch of the lateral lacuna (lateral blood-
vessel of early authors) which lies immediately ventral to each
testis (Text-fig. 8, N-f.).

The structure of the ciliated organ in this leech has some
resemblance to that of Nephelis as figured and described
by Graf (1899). A comparison of a transverse section of a
ciliated funnel of C. joseensis, as shown in Text-fig. 10,
with Graf’s figures of the ciliated organ of Nephelis quadro-
striata, shows the similarity of these organs. The terminal
vesicle of each nephridium is particularly well developed and
lies on either side at the level of the nerve ganglion in relation
to it; portion of the lumen of the vesicle is taken up by the
bulging into it of the testis.

254 CHARLES BADHAM

TEXT-FIQG. 8,

Transverse section of C. joseensis, in the 20th somite, showing
the position and relative size of the nephridial funnels. (x 30.)
C.m., circular muscle layer. JL.m., longitudinal muscle layer,
Ll, lateral lacuna. T’., testes, fifth pair. V.n., terminal vesicle
of nephridium. N./., nephridial funnel. V.g., anterior portion
of mid-gut. Ne.c., nerve-cord in ventral lacuna.

TEXT-FIG, 9.

Transverse section of C. joseensis in the 22nd somite, showing
the relation of the blind-gut to the middle-gut, (x 30.) C.m.,
circular muscle layer. Lpi., epidermis. L.m., longitudinal
muscle layer. D.l/., dorsal branch of lateral lacuna. L.J.,
lateral lacuna. V.n., terminal vesicle of nephridium. J.g.,
blind-gut. M.g., middle-gut. Ne.c., nerve-cord in ventral sinus.

CENTROPYGUS JOSEENSIS 955

CobLomic SystEM AND 1Ts MODIFICATIONS.

The lateral lacunae (lateral blood-vessels of authors before
Oka (1912)) extend throughout the body, being much reduced
at either end; in each somite there are regularly arranged
communications dorsally between the lateral lacunae and
ventrally with the ventral lacuna (Text-fig. 9, Z.l.). There
is no dorsal lacuna and this is absent also in other Herpo-
bdellids, and the ventral lacuna is small and contains only the

TeExt-Fic. 10.

Transverse section of a ciliated funnel of C. joseensis, showing
the ciliated crown cells and the granular cells. C.c., ciliated crown
cell. G.c., granular cells. A., ampulla,

nerve-cord and ganglia and forms no expansion anteriorly to
contain the reproductive organs.

SENSE ORGANS.

These consist of a series of flask-shaped organs chiefly
arranged along the edge of the oral sucker. They have been
described and figured by Kennel (1886), and I have nothing
to add to his description.

I wish to thank Professor J. P. Hill, of University College,
London, for his kindly assistance whilst engaged in this work
in his department.

256 CHARLES BADHAM

LITERATURE.

Kennel, J. (1886).—“‘ Ueber einige Landblutegel des tropischen America ”’,
‘Zool. Jahrbiicher ’, ii, 1886-7.

Blanchard, R. (1896).—“ Hirudinées. Viaggio del dott. A. Borelli nella
Repubblica Argentina e nel Paraguay ”’, ‘ Boll. Mus. Zool. Torino’,
xi, no. 263.

—— (1899).—‘“‘ Courtes notices sur les Hirudinées’”’, ‘ Bull. Soc. Zool.
de France’, xxiv.

Brumpt, E. (1899).—‘‘ Reproduction des Hirudinées”’, ‘Mém. Soc. Zool,
de France ’, xii.

Graf, A. (1899).—‘‘ Hirudineenstudien ”, ‘ Nova Acta C.L.C.G. Nat. Cur.
Halle ’, Ixxii.

Oka, A. (1912).—‘‘ Ueber das Blutgefasssystem der Hirudineen ”’, ‘ Annot.
Zool. Japon.’, iv, p. 2.

Weber, M. (1914).—‘“‘ Hirudinées colombiennes ’’, ‘Mém. de la Soc. des
Sciences Nat. de Neuchatel ’, v.

Some Observations on the Hypophysis of
Petromyzon and of Amia.

By

G. R. de Beer, B.A., B.Sc.,

Demonstrator in the Department of Zoology and Comparative Anatomy,
University Museum, Oxford.

With 34 Text-figures.

CoNTENTS.
I, PETROMYZON: PAGE
Introduction : J : : ; : 5a PT
The Hypophysis epression! : : ‘ 3 . 259
The Formation of the Hypophysial tae : : : . 265
Discussion ; ‘ : oe:
The Primitive Radian of the Hy po physis ; : : « ane
The Pars Tuberalis . : : : : : : 5 Casill
If. Amra:
Introduction 3 : : ; ; , . 283
The Origin of the Hypophysi : : ; : . . 284
SUMMARY F ‘ ‘ : : : : . 289
List oF LITERATURE Gino: : ‘ ; ; : : . 290

I. PETROMYZON.

Introduction.

Tur hypophysis of Petromyzon has been the object of
research of many observers, more indeed than will be referred
to in this paper since complete bibliographies are to be found
elsewhere, and it might almost seem as though any further
work on the same subject could not but be repetition. It
seemed to me, however, when considering the problems
presented by the hypophysis in the light of more recent

958 G. R. DE BEER

research that there are a few pomts which deserve further
attention.

The main outlines of the structure and development of the
hypophysis of Petromyzon have been known since the work of
Dohrn, Gétte, and Scott, and the problems of the hypophysis
then presented themselves, but they have been answered in
many contradictory ways by them and the many other
observers who have devoted their attention to them. Shortly,
the main problems are the following :

(1) What is the primitive position of the hypophysis, and has
it any primitive relations with the mouth or the nose
or any other organs ?

(2) What is the primitive method of its development given
the great diversity which exists in the vertebrates, and
especially the great difference presented by Cyclo-
stomes as compared with almost all other vertebrates ?

In this paper I am concerned with these two questions and
am led to discuss a few more, such as the relations of the hypo-
physial cavity to the glandular portions of the pituitary body
and the possible presence of one of these portions in Petromyzon
corresponding to the pars tuberalis of higher vertebrates.

I wish to express my thanks to Mr. Huxley and Dr. Hogben,
not only for stimulating my curiosity in some of these questions
but also for useful criticism. To Professor Goodrich also my
oratitude is due for helpful advice.

Methods eall for no special comment ; as complete a series
of embryos as possible was studied, cut into serial sections in
the transverse, sagittal, and horizontal planes. Staims such
as picronigrosin and Lichtgrim were useful in detecting
boundaries and limiting membranes where such were obscure.
The material of Petromyzon planeri was collected durmg
a stay at Naples, and some stages were supplemented by the
kindness of Professor Goodrich. The work was done in the
Department of Zoology and Comparative Anatomy at Oxford.

The origin of the hypophysis in Petromyzon from the anterior
surface of the head and close to the rudiment of the olfactory
organ is familar in every text-book, after the researches of

HYPOPHYSIS OF PETROMYZON AND AMIA 259

Dohrn (1883), Gétte (1888), Haller (1896), von Kupffer (1893),
Scott (1888), Stendell (1914), Sterzi (1904), and Woerdemann
(1914). (To Sterzi’s work I have unfortunately been unable to
gain access.) It is described as an invagination giving rise to
a cavity into which in the adult the olfactory organs open and
which extends posteriorly for a considerable distance between
the brain and the gut.

In those cases where the hypophysis arises within the stomo-
daeum (Selachians and Amniotes) it may develop either as
a hollow invagination (Rathke’s pocket) or as a solid ingrowth
of a mass of cells within which the hypophysial cavity may make
its appearance later. Where it arises outside the stomodaeum
as in Amia (Reighard and Mast 1908; the writer, see part i
of this paper) or in Amphibia (Gétte 1875, Atwell 1919)
no invagination occurs, but the rudiment is a solid plate pushing
in from the ectoderm. Now the hypophysis of Petromyzon
certainly arises outside the stomodaeum, but it is situated
next to the region which enlarges enormously during develop-
ment, viz. the upper lip. It was my first object to decide
whether the so-called invagination is due to the growth of the
parts composing its walls or to actual invagination, or to both
causes. Pending decision on this point I shall call it the
hypophysis depression.

The Hypophysis Depression.

In the earliest stage at which any rudiment of the future
organ can be discerned (Text-fig. 1) a thickening of ectoderm
on the antero-ventral surface of the head marks the region
where the hypophysis and olfactory organs will develop. On
the supposed primitive nature of the connexion between these
organs I have little to say except that I see no reason for
regarding it as anything more than a topographical one,
two organs both developing from the ectoderm at the same time
and in the same restricted region, viz. the antero-ventral
surface of the head of necessity arising together. There is at
this stage (just before hatching) no trace of invagination. The
stomodaeum is indicated and is separated from the hypophysis

260 G. R. DE BEER

rudiment by a certain distance, occupied by flat ectoderm ;
there is as yet no upper lip.

In an embryo twelve days old from the time of fertilization
(Text-fig. 2), which has consequently hatched, the anterior
region of the brain has extended slightly and the antero-ventral
surface of the head lies practically horizontal. A depression
U-shaped in section near the anterior extremity is the invagina-

TErxtT-Fig. 1.

Petromyzon planeri. Embryo just prior to hatching,
rudiment of hypophysis and stomodaeum.

AdO, adhesive organ. A.O., accessory organ. D.W., dorsal wall of
fore-gut. F.b., fore-brain. H.c., hypophysial cavity. 4H.d.,
hypophysial depression. H.r., hypophysial rudiment. 4H.s,
hypophysial strand. J., infundibulum. J.l., lateral lobes of
‘ Uebergangsteil’. N., notochord. N.c., cartilage of nasal
capsule. O.ch., optic chiasma. O.i., olfactory invagination.
O.n., olfactory nerve. P.A., pars anterior. P.E., pineal eye.
P.I., pars intermedia. P.N., pars nervosa. S.i.h., solid ingrowth
of the hypophysis. St., stomodaeum. T7\.c., trabecula cranili.
U., ‘ Uebergangsteil’. U.1., upper lip.

tion of the olfactory organ. Farther back is the larger similarly
U-shaped invagination of the stomodaeum ; but between them
is a slight dent on the surface corresponding to a region where
ectoderm cells are pushing in beneath the brain. This is the

HYPOPHYSIS OF PETROMYZON AND AMIA 261

rudiment of the hypophysis. The region of ectoderm between
it and the stomodaeum is no longer flat but convex, being the
beginning of the expansion of the upper lip. Is the dent the
beginning of a true invagination or is it due to the expansion
of the upper lip ? The latter I believe to be the true interpreta-
tion. It bears no resemblance to the normal appearance of
Rathke’s pocket or any invagination, as may be seen by com-

Text-Fig. 2.

-_domm Ul

Petromyzon planeri. Embryo just hatched, beginning
of hypophysis depression.

paring it with the stomodaeum or the olfactory organ. An
invagination presents one concave curve; the hypophysis
depression is made up of two convex curves meeting at an apex.
There is a solid inpushing of cells at this stage although the
depression is a mere dimple, whereas in those cases where
the hypophysis arises by invagination it is sunk beneath the
surface by means of that invagination which must be almost
as deep (see Text-fig. 23a).

The upper lip continues to increase in size, and the hypo-
physis depression comes to assume the form of a cleft which
deepens and the upper and lower sides of which become more
and more parallel as the upper lip enlarges. At the same time
the sides of the upper lip grow up with the sides of the anterior

262 G. R. DE BEER

surface of the head so as to leave the hypophysis depression
as a tube which becomes deeper by the lengthening of its walls.
The beginning of this process is shown in Text-fig. 8, and it
is more accentuated in Text-figs. 4 and 5.

The apex of the V, i.e. the limit of the depression, does not
move farther back relatively to the rest of the head. It never

TEXT-FIG. 3.

Petromyzon planeri. Embryo fourteen days old,
beginning of expansion of the upper lip.

quite reaches the optic chiasma. Meanwhile the solid ingrowth
has extended until it almost reaches the tip of the notochord.
To verify this I have resorted to measurements, though I do not
set much store by them owing to the difficulty of measuring
with accuracy and of selecting standard measuring lines since
no points can be assumed to be fixed.

A. Distance from apex of depression to notochord.

B. Distance from apex of depression to extremity of upper lip.

HYPOPHYSIS OF PETROMYZON AND AMIA 263

Embryo represented

In Figures : A, B.
3 30 15
4 30 25
5 30 60
6 50 120
iT 80 140
8 280 440

TEXT-FIG. 4.

Wk r>D :
x ; Lop fe >}!
ee GPUTTINSY |
oi Hd | SSS) |
( z St.
a Sih.

Petromyzon planeri. Embryo twenty-two days old.

Since the measurements are all to the same scale the units
are unimportant.

The fact that the distance between the apex of the depression
and the notochord does not decrease, while that between the
apex and the extremity of the upper lip increases rapidly,
supports my contention that the hypophysis depression is not

an invagination but is brought about by the relative growth
NO. 266 £

264 G. R. DE BEER

of the upper lip. The relations of the apex to the optic chiasma
confirm this.

The solid ingrowth of the hypophysis extending by itself
from the region of the optic chiasma to the notochord does so
not by the deepening of the hypophysis depression but by its
own lengthening.

It is usually stated that the monorhinal condition of the

TEXT-FIG. 5.

SS
SQ
QVVNIOS

9
Q

A)
dbo

ys
©

to mm.

Petromyzon planeri. Embryo thirty-two days old.

lamprey is due to the connexion of the hypophysis with the
olfactory organs which thus communicate with the exterior
by means of a single (hypophysial) aperture. It appears to me
to be easier to regard the agent as the upper lip in conjunction
with the sides of the anterior surface of the head, which by their
expansion cause both organs to be situated in the same
secondary depression.

The hypophysis of Petromyzon then belongs to the class of

HYPOPHYSIS OF PETROMYZON AND AMIA 265

those which develop by a solid inpushing. A diagrammatic
representation of the growth of the upper lip and formation
of the hypophysis depression is given in Text-fig. 6.

The Formation of the Hypophysial Cavity.

The solid ingrowth lies between the dorsal wall of the gut
and the floor of the brain, separated from them by well-marked
membranes. In the stage represented in Text-fig. 7 no glandular

TEXT-FIG. 6,

Petromyzon planeri. Diagram showing the expansion of
the upper lip and formation of the hypophysis depression.

differentiation has as yet occurred. The first indication of the
histological differentiation which will produce the future
pituitary body is to be found in the floor of the brain. Just
dorsal to the posterior portion of the hypophysis a thickening
of the brain-floor occurs composed of neuroglia (Text-fig. 8),
thus foreshadowing the pars nervosa. There is no trace of
hypophysial cavity.

In a later stage (85 mm., Text-fig. 9, and under higher magni-
fication in Text-fig. 10) the hypophysis is beginning to undergo
its histological differentiation (changes which will result in the
formation of the glandular elements which it contributes to the
pituitary body). Beneath the optic chiasma the strand of tissue

T 2

DE BEER

G. R.

266

cal
4.

TEXT-FIG

Embryo 10 mm. long.

Petromyzon planeri.

TEXT-FIG. 8.

Appearance of
all of the infundibulum.

go,
=)

o

Embryo 15 mm. lon

neuroglia thickening in the w

Petromyzon planeri.

HYPOPHYSIS OF PETROMYZON AND AMIA 267

TExtT-FIG. 9.

Petromyzon planeri. Embryo 35 mm. long.

Trext-Fic. 10.

Petromyzon planeri. Thesame embryo highly magnified.
Differentiation of the glandular elements. No trace of a hypo-
physia! cavity.

268 G. R. DE BEER

thickens and its cells become ovoid and glandular and less
closely packed. Dorsally this thickening is for the greater part
of its length closely adpressed to the floor of the infundibulum.
The layer of neuroglia which separates the cells lining the
infundibular cavity from the outer membrane of the brain-

Text-Fic. 11.

Petromyzon fluviatilis. Ammocoete. Transverse
section just posterior to the apex of the external depression
showing absence of hypophysial cavity.

wall has also thickened. All this time the apex of the hypo-
physis depression has not moved, and the differentiation of the
elandular hypophysis takes place from a solid strand of cells,
there being still no true hypophysial cavity.

In the Ammocoete a cavity appears in the region just anterior
to the gland. It has no connexion with the external depression

HYPOPHYSIS OF PETROMYZON AND AMIA 269

TrxtT-ric. 12.

Petromyzon fluviatilis. Ammocoete. Transverse
section slightly farther back and anterior to the gland showing
the hypophysial cavity in the middle of the hypophysis.

Text-ric. 13,

Petromyzon fluviatilis. Ammocoete. Transverse
section still farther back showing the relation of the gland
to the hypophysial cavity.

270 G. R. DE BEER

(Text-fig. 11) and is hollowed out in the middle of the strand
(Text-fig. 12), its walls being two or more cells thick. More
posteriorly the cavity extends for some distance ventrally to
the glandular body, the latter forming its dorsal wall (‘Text-
fig. 18). It extends farther back laterally than it does in the
median plane. Transverse sections posterior to this region
show no cavity at all (Text-figs. 14and 15). The relations of the

Trxt-Fic. 14,

Imm.

Petromyzon fluviatilis. Ammocoete. Transverse
section behind the limit of extension of the hypophysial
cavity showing the lateral lobes of the ‘ Uebergangsteil ’.

hypophysial cavity are shown in Text-fig. 16, which is a sagittal
section of an Ammocoete. It arises in the middle of the
hypophysis tissue without communi¢ation with the hypo-
physis depression. The glandular tissue forms its dorsal wall,
and at this time the pars anterior and pars intermedia, though
distinguishable, are in contact, being continuations the one of
the other in a straight line. The hypophysial cavity does not
separate them. Text-fig. 17 is a high-power view of this
stage. The distinction between the parts of the pituitary body

HYPOPHYSIS OF PETROMYZON AND: AMIA 271

Text-Fia. 15.

fi na
oe

limm.

Petromyzon fluviatilis. Ammocoete. Transverse
section showing the lapping of the lateral lobes back round

the pars intermedia.

TEexr-rIG. 16:

Spy ee

MNT Mitty
)

)

\

eck
aut I

Dae
F284,
rc

>
o

Ammocoete, Longitudinal section.

Petromyzon fluviatilis.

D2, G. R. DE BEER

is plain. Text-figs. 11 to 18 are of Petromyzon fluvia-
tilis, the conditions in planeri are similar.

At a later stage the hypophysial cavity acquires more definite
walls, and dorsally the wall separates itself from the glands
(Text-fig. 18). The pars intermedia is the first to lose connexion,
a band of connective tissue passing between it and the cavity

TEXT-FIG. 17.

0
OG POC e lore
PISGao. Oe a>

'
' %
9S Ve ' GFZ Z Gy
P Le ‘ ' “a S ee ‘8 Yj

He DW

‘

Petromyzon fluviatilis. Ammocoete larva more highly
magnified, showing the relations of the hypophysial cavity, the
pars anterior, pars intermedia, pars nervosa, and * Uebergangsteil’.

and also between it and the pars anterior separating it from
that also. The pars intermedia, practically devoid of blood-
vessels, is now in intimate connexion with the pars nervosa of
the infundibulum, and its condition is now similar to that
which it presents in the adult. The pars anterior is highly
vascular. The cavity becomes more spacious and extends
ventrally and posteriorly. It acquires a connexion with the
exterior through the hypophysis depression by a split which
occurs along the hypophysis strand. In the adult (Text-figs. 19,
20, and 21) the separation of the glands from the hypophysial
cavity has gone further and they are now firmly encapsuled

HYPOPHYSIS OF PETROMYZON AND AMIA 2738

TExt-Fig. 18.

Och _
\ “ae 5

= = —_ 4 SEE =
—— s
Selina
i —————

Petromyzon fluviatilis. Ammocoete. Further exten-
sion of the hypophysial cavity and separation from the glands.

TExtT-FIG. 19.

Petromyzon planeri. Adult. Longitudinal section show-
ing the separation of the glands from the hypophysial cavity and
the great extension of the latter.

274 G. R. DE BEER

in connective tissue. The hypophysial cavity has reached
huge dimensions, and is now the well-known sac extending
back between brain and gut and communicating with the
exterior by the single ‘ nasal’ aperture.

Trext-Fic. 20. TExtT-Fic. 21].

Teva LN

retetaligg
es)
at

%
"eer

Petromyzon planeri. Petromyzon planeri. ~ Adult.
Adult. Transverse section Transverse section through the pars
through the pars anterior. intermedia.

Discussion.

The hypophysial cavity of Petromyzon is remarkable for
several reasons. Briefly its characteristics may be considered.
It develops in the middle of the hypophysis tissue which was
a solid inpushing. This is paralleled by many other verte-
brates. It remains in connexion with the exterior throughout
adult life. This condition is only found elsewhere in Polypterus
and Calamoichthys. In these a persisting connexion between
the hypophysial cavity and the stomodaeum is referred to by
Sedgwick and Wiedersheim, and it would be of great interest
to know its development.

HYPOPHYSIS OF PETROMYZON AND AMIA D5

The connexion with the exterior may be regarded as a delayed
appearance of Rathke’s pocket. ‘The adaptive nature of this
connexion in Myxine where the hypophysial cavity also com-
municates with the gut suggests that the hypophysial cavity is
secondarily modified, as does also the size of the hypophysial
cavity in these forms. The fact that it has lost all connexions
with the glandular portions also lends support to this view.
For whatever the primitive function of the hypophysis may have
been, it was an ectodermal organ sunk beneath the skin and the
hypophysial cavity must be supposed to have served to keep
the organ in communication with the exterior. In vertebrates
above Cyclostomes where a hypophysial cavity exists it is in
association with the glandular portions of the organ, though
functionless since the latter have adopted the method of internal
secretion. In Cyclostomes the cavity is separated from the
glands in the adult by a good thickness of connective tissue,
although in earlier stages it was in contact with them and they
even formed its dorsal wall. It is this fact which enables one to
believe that the hypophysial cavity of Petromyzon is homo-
logous with that of other forms, a homology which would be
difficult to establish from the adult. But even here Petro-
myzon is peculiar. In vertebrates typically the hypophysial
cavity separates the pars anterior from the pars intermedia.
In Petromyzon, as we have seen, the glands differentiate before
the appearance of a cavity and the pars anterior and pars
intermedia he in a straight le with the cavity (when it does
form) beneath them. Starting from this condition, i.e. the
hypophysis arising from the front and growing backwards, and
the hypophysial cavity horizontal with the glands on the brain
side (dorsal), the pars nervosa of the pituitary very slightly
developed and not projecting ventrally ; the conditions obtain-
ing in higher vertebrates can be derived when the following
changes are taken into account.

(i) The hypophysis grows up from beneath and meets the
brain at right angles. The consequence of this is that that
portion of the roof of the primitive hypophysial cavity in
connexion with the pars anterior which was dorsal and

276 G. R. DE BEER

horizontal in Petromyzon is rotated through 90° and lies
anterior and vertical.

(1) The down-growth of the infundibulum in connexion with
the specialization of the pars nervosa forces the pars intermedia
down also so that it hes posterior to the hypophysial cavity.
In this manner the hypophysial cavity comes to lie between

TExtT-FiG. 22.

no ez loorere-

, Co

Diagram illustrating the difference in relations between the glands
and the hypophysial cavity in (a) Petromyzon and (6) higher
vertebrates. Pars anterior dotted, pars intermedia lined, pars
nervosa black.

the pars anterior and the pars intermedia. This is shown
diagrammatically in Text-fig. 22; and it is on such lines that
I would explain the difference between the relations of these
organs in Cyclostomes and higher forms.

Swale Vincent (1922) gives a figure of the pars intermedia
on both sides of the cleft, i.e. his pars anterior is separated
from the cleft by a strip of tissue labelled pars imtermedia.
If the cleft should represent the true hypophysial cavity, this

HYPOPHYSIS OF PETROMYZON AND AMIA OTT

is incompatible with the theory which I am putting forward
here. The strip of tissue in question is probably a portion of
the pars anterior the cells of which are slightly modified since

TEXxT-FIG. 23.

Diagrammatic comparison between (a) Squalus, (b) Amia, and
(c) Petromyzon showing the position of the hypophysis and
infundibulum.

here they act as an epithelium and line a cavity. Swale
Vincent does not describe this structure in the text, so that
not much importance must be placed upon his figure. The

278 G. R. DE BEER

evidence is in favour of the cleft representing the true hypo-
physial cavity, and to quote Biedl (1913) ‘the anterior and
posterior lobes are separated by a cleft more or less broad.
This represents the vestige of the embryonic hypophysial
cavity. The posterior wall of this cavity is directly opposed
to the posterior lobe and forms its anterior limit in the shape
of a strip .. . known as the pars intermedia.’

The identification of the hypophysis with Kélhker’s pit (and
therefore the neuropore) of Amphioxus (Willey 1894) is open
to the objection that in higher forms hypophysis and neuropore
are found in one and the same animal without any connexion.

With regard to the Tunicates the researches of Julin and van
Beneden led to these observers believing that the hypophysis
is represented by the subneural gland. But the later observa-
tions of Willey, Seeliger, and others show that the neural gland
is derived not from external ectoderm but from the nervous
system. The gland may have some connexion with the neuro-
pore, but later it is in communication with the buccal cavity
and a shallow ectodermal invagination (or several) meets the
tube growing out of the gland. It is possible that the shallow
ectodermal invaginations alone may represent the hypophysis
(Stendell 1914b). Dr. Hogben kindly permits me to make
use of the fact that he found that extracts from the subneural
gland of ascidians had no properties such as are present in
extracts of the posterior lobe of the pituitary of all classes of
Gnathostomes.

Dohrn’s (1888) view that the hypophysial cavity of Petro-
myzon represents a pair of gill clefts is open to the objection
that whereas gill clefts are formed by outgrowth of endoderm,
the hypophysis is ectodermal.

Haller (1896) believes that the hypophysis of Cyclostomes
is secondarily modified, but he describes a lumen in the pars
intermedia the existence of which has never been confirmed.

Woerdeman (1914) agrees that the glandular elements are
differentiated from a solid strand of tissue in the absence of
any cavity, and also that the greatest part (‘ Grésstenteils ’)
of the hypophysial depression is due to the overgrowth of the

HYPOPHYSIS OF PETROMYZON AND AMIA 279

upper lip. In the hypophysis of Gnathostomes he divides the
hypophysial cavity into three parts; the most posterior he
terms Rathke’s pocket and separated from it by a constriction
he distinguishes the remainder of the cavity as ‘ Mittelraum ’
and ‘ Vorraum’. Applying these homologies to Petromyzon
he gets the external hypophysis depression to correspond with
the ‘ Vorraum’, the glandular portion of the hypophysis with
Rathke’s pocket, and the upper lip with that region situated
just behind Rathke’s pocket in Gnathostomes. The consequence
of this is that the olfactory organs and mouths of Cyclostomes
and Gnathostomes are not homologous. Such conclusions are
unacceptable. The recognition of the secondary nature of
monorhiny, the relations of the olfactory organs to the olfactory
nerves and brain and those of the trigeminal and facial nerves
to the mouth are against his view. Besides, it does not appear
that the acceptance of his divisions of the hypophysial cavity
facilitates their interpretation.

We are left then with the view that the hypophysial cavity
of Petromyzon is homologous with that of other forms but
differs from them by secondary modifications.

The Primitive Position of the Hypophysis.

The most striking feature about the hypophysis of Petro-
myzon at first sight is the fact that it has nothing to do with
the mouth, but communicates with the exterior on the dorsal
side of the head. The latter feature is, as previously stated,
due to the great expansion of the upper lip with the front of
the head, so that when a stage before the upper lip has developed
(Text-fig. 2) is considered, the hypophysis faces ventrally.
But even then it has no connexion with the stomodaeum. The
limits of the stomodaeum are not easy to define since they are
not marked by any structural peculiarity ; but they may be
taken as being the points where the concave curve of the
invagination changes and becomes convex. The fact that in
the majority of vertebrates, viz. Selachians and Amniotes,
the hypophysis develops from within the stomodaeum has

NO. 266 U

280 G. R. DE BEER

led to the view that this is its primitive position (Stendell
19144).

Recently, however, attention has been drawn to those cases
where the hypophysis arises outside the stomodaeum and just
anterior to it. In this connexion the works of Gétte (1915)
and Atwell (1919) for Amphibia, of Reighard and Mast (1908)
and the writer for Amia, of Wells (unpublished) for Clupeus
may be mentioned. In these cases the hypophysis has no
connexion with either mouth or nose, and this position Scott
(1883) believes to be primitive. Where the hypophysis arises
within the stomodaeum, i.e. in Selachians and Amniotes,
there is a great development of the fore-braim and cranial
flexure ; this rotation of the anterior part of the head causes
the ventral elements of the head to he relatively farther back
and accentuates the stomodaeal invagination. This I believe
to be the cause of the hypophysis being situated in the stomo-
daeum in these forms, but the difference between the two types
is more apparent than real. In addition there is the fact that
the hypophysis is ectodermal tissue which must get into contact
with the infundibulum. Where the fore-brain is large and the
cranial flexure pronounced this can most conveniently be done
through the stomodaeum. But primitively the fore-brain
cannot have been large and there was less cranial flexure. _ This
condition is preserved in the Teleosts and Amphibia, and here
the typical position for the hypophysis to arise is outside and
dorsal to the stomodaeum. Text-fig. 23 is a diagrammatic
comparison between Squalus, Amia, and Petromyzon showing
the modifications brought about by the fore-brain and the
cranial flexure.

Not only are the ventral elements of the head of Squalus
pushed backwards, but the dorsal elements are pulled forward
owing to the outer side of the curvature of the cranial flexure
being dorsal. So the dorsal nerve-roots lead backwards to their
respective gill arches whereas in Amia they incline forwards.
The anterior position of the heart in Amia is partly due to the
embryo being flattened out on the yolk.

Support is lent to the view that this is the primitive position

HYPOPHYSIS OF PETROMYZON AND AMIA 981

of the hypophysis by Goodrich’s (1917) suggestion that the
hypophysis is represented in Amphioxus by the deep groove
and depression known as the pre-oral pit in the larva and wheel
organ in the adult. This, as its name implies, is anterior to the
mouth.

There remains the question as to whether an invagination
(Rathke’s pocket) or a solid ingrowth is the more primitive
method of formation of the hypophysis. It is dangerous in
a point of this kind to attempt to mduce phylogeny from
ontogeny, for the mode of development obeys embryonic
conditions. I should suggest that the hypophysis of the
priunitive vertebrate was an invagination as is the pre-oral pit
of Amphioxus, but that when the combination with the
infundibulum forming the pituitary body was evolyed, the
mode of development became influenced by the distance which
the ectodermal tissue has to travel.

So m Petromyzon or Amia or Amphibia the ingrowth is
solid, in Selachians and Amniotes it tends to be hollow.

Mre Pars Tuberalis.

To the description of the main glandular elements of the
pituitary body of Petromyzon I have little to add. The pars
anterior is made up of two portions, an anterior lobe composed
of chromophil cells, and a posterior of chromophobe. This
latter portion is termed by Stendell (1914) the * Uebergangs-
teil’, by Gentes ‘the middle lobe’. Stendell (1913) regards it
as morphologically part of the pars anterior (‘ Hauptlappen ’) ;
Sterzi as part of the pars intermedia. It is seen in the ammo-
coete of Petromyzon fluviatilis in longitudinal section
in Text-fig. 17, where it occupies the region between the pars
anterior and the pars intermedia. The outer sides of this
*“ Uebergangsteil ’’ lap back round the pars intermedia on each
side, as seen in transverse section in Text-fig. 15 and in hoyi-
zontal in Text-fig. 24. Text-fig. 25 is a reconstructed view
of the whole organ from the ventral surface. This fact was
also observed by Woerdeman (1914).

Recently attention has been paid to the pars tuberalis as

U 2

282, G. R. DE BEER

a distinct element of the pituitary body, notably, by Tilney
(1913), Woerdeman (1914), Baumgartner (1916), and Atwell

TEXT-FIG. 24.

Sear D>
PX)
a
D 7,9:
Cy

Petromyzon fluviatilis. Ammocoete. Horizontal section
showing the lapping of the lateral lobes round the pars intermedia.

Petromyzon fluviatilis. Reconstruction of the pituitary
body of an ammocoete seen from the ventral surface after
removal of the ventral wall of the hypophysial cavity.

(1919). Woerdeman considers the possibility that its homologue
is to be found in the accessory organ (Text-figs. 8 and 9)

HYPOPHYSIS OF PETROMYZON AND AMIA 983

situated in the nasal capsule described by Scott (1888). ‘This
organ has been studied and is being described by me else-
where. It develops in close connexion with the olfactory organ
and has no histological or topographical similarities with the
pars tuberalis of other forms.

In the description of its origin in other forms the pars tuberalis
arises as lateral processes (lobuli laterales) which develop
acini composed of chromophobe cells, and which are closely
adpressed to the brain. According to Tilney it arises in con
nexion with the pars intermedia (his pars infundibularis).
If the pars tuberalis is represented at all in Petromyzon,
I suggest that it is the ‘ Uebergangsteil’. Its cells are chromo-
phobe (Herring, 1910), it has lateral extensions which are closely
adpressed to the brain-wall, and is the only part of the organ
not otherwise accounted for. From its position between the
pars anterior and the pars intermedia it can hardly be said to
arise IN connexion with the one rather than with the other,
though Sterzi considers it to be more closely related to the
pars intermdia.

IJ. AmrIA CALVA.

Introduction.

The hypophysis of Amia calva is described by Dean
(1896) as developing late. He regarded it as ectodermal,
though he admitted that its ventral limit could not be distin-
guished from the endodermal roof of the fore-gut.

Subsequently Prather (1900) investigated the question and
concluded that the hypophysis of Amia was of endodermal
origin and derived from the roof of the fore-gut.

Later still, Reighard and Mast (1908), as a result of their
researches, believe that Prather was misled in his conclusions
by artifacts in his preparations due to shrinking, and they
claim that the hypophysis of Amia arises from the ectoderm
close to the neuropore, dorsal to and separate from the
stomodaeum.

Smith (1914) regards the hypophysis as of ectodermal origin,
but thinks that endoderm cells are contributed to it.

284 G. R. DE BEER

Stendell (1914a) makes no mention of Amia, and remarks
that the Ganoids are in need of much further study in this
respect. In view of the diversity of opinion on this matter and
the fact that the confirmation of either of these two theories
must lead to important conclusions with regard to the hypo-
physis, I determined to examine my preparations of embryonic
and larval stages of Amia.

The material consisted of sets of serial sections, sagittal,
horizontal, and transverse ; and I may say at once that the
remarks made by Reighard and Mast with regard to the care
necessary in making the preparations are well founded. By
shrinking of the egg-membrane the contaimed embryo is often
compressed and the limits of its organs are sometimes difficult
to make out with certainty. Thionin was found to be an
efficient stain although it unfortunately fades. For bringing
out the basement membranes Lichtgriin or acid fuchsin were

found to be useful as counter-stains after safranin and methylene
blue.

The Origin of the Hypophysis.

The earliest stage in which the hypophysis was visible is
shown in sagittal section in Text-fig. 26. The brain is in contact
with the antero-dorsal ectoderm in the region where the neuro-
pore has closed. The antero-ventral region of the head is
occupied by a mass of large endoderm cells containing abundant
yolk and destined to form the adhesive organ. Between the
latter and the brain is a tract of smaller cells almost devoid of
yolk-granules, in contact with the ectoderm antero-dorsally,
and postero-ventrally tapering into a point where the brain
comes into contact with the endodermal roof of the fore-gut.
There can be no doubt that these cells originate from the
ectoderm.

The next stage is shown in small scale in Text-fig. 27 and under
higher magnification in Text-fig. 28. The tract of ectodermal
cells described in the previous figure is more compressed and
denser. Anteriorly it shows traces of previous connexion with
the inner layer of superficial ectoderm of the front of the head.

HYPOPHYSIS OF PETROMYZON AND AMIA 285

The line of demarcation between it and the underlying endoderm
is not easy to see, but with an oil-immersion objective traces
of the original basement membrane of the endoderm can be
observed. More striking is the difference in yolk content, for
whereas the hypophysis (since such I believe these ectoderm
cells to be) is practically devoid of it, the endoderm cells both
of the roof of the fore-gut and of the adhesive organ contain

TEXxtT-FIG. 26.

OQ
?
2
y
On ie
30
wn
NAN
NINE

BW a BA
SR8 SS A
< > ©

wows ZO,
G)

(
es
) ve
LY
©) bi
f J g
‘ 0 0
KT
QW
SA
iS
ES
p=
eS

SS
Q
oC 3 SS
ES SS se DonwD D2
C0)5,0/0) GO AP :
BRE Ch ~DW

Amia calva. Young embryo (coiled round the yolk) showing
the rudiments of the adhesive organ and of the hypophysis.

numerous large granules. At the same time the brain appears
to be pressing down on the hypophysis and squeezing it against
the endodermic roof of the fore-gut.

In the next stage (Text-figs. 29 and 30) this pressure has
increased, for not only is the hypophysis less thick but it has
impressed its convex dorsal and ventral surfaces into the
corresponding concavities of the brain and gut-roof. The
floor of the brain which in previous stages was bent dorsally
in the region of the recessus opticus here continues flat for
a further distance forwards. It looks as if the brain and the

286 G. R. DE BEER

front of the head generally were growing forward and the
hypophysis caught tight between the brain and the gut-roof
were prevented from joining in this forward movement and

TEXT-FIG. 27.

Amia calva. Slightly later stage than the preceding, the hypo-
physis is beginning to lose connexion with its point of origin.

TEXxtT-FIG. 28.

Amia calva. The same as Text-fig. 27, higher magnification.

so comes to he relatively farther back. The demarcation
between hypophysis and gut-roof is again hard to pick out,
but the restriction of large yolk-granules to the latter is plain.

As development proceeds the brain and the front of the head
generally seems to have grown and extended anteriorly. But

HYPOPHYSIS OF PETROMYZON AND AMIA 287

the distance between the hypophysis and the anterior extremity
of the notochord is much reduced. This may mean that the
tip of the notochord has grown forward, or that the hypophysis
itself has moved back. The tip of the notochord appears to
be stationary and to bear fairly constant relations to the limits
of the hind-brain, so that the movement must be sought for

TExtT-FIc. 29.
Fb

oe

(Zz hi : ATTIRE
Ad. 0. <- | emesis = Sn
= yarn eda ou 4
tr ctl “he,
“gn ' >

’
' !

1H N

Amia calva, 6 mm. long. Hypophysis completely separated
from its point of origin and wedged between the brain and
the dorsal wall of the gut.

TExtT-FIG. 30.
Fb

chica

WG
(a:

= {oe Lo raise ?
55 Slr iM)
i gees
\ ——
'
L. ; :
Oo mm H DW

Amia calva. The same as Text-fig. 29, higher magnification.

in the hypophysis itself. Such movement could hardly be due
to active migration since the hypophysis is firmly held between
brain-floor and gut-roof, but is rather to be explained as the
passive result of the movement of the latter. This movement
may be due to the ‘ recoil’ resulting from the forward growth
of the anterior of the fore-brain as suggested by Reighard and
Mast, or to shrinkage of the gut-roof in the region posterior to
the hypophysis, which would have the effect of drawing back

288 G. R. DE BEER

the structures anterior to it. In transverse section the hypo-
physis about this stage appears as in Text-fig. 31. At first
sight it looks as if the hypophysis must have been derived
in situ from the underlying endoderm cells. But careful
observation with high powers reveals the differences that have
been met with before between the cells of the hypophysis and
the endoderm cells, viz. absence of yolk and different orienta-
tion. In many cases the limiting membrane of the gut-roof

TEXT-FIG. 31. TEXT-FIc. 32.

DW H to mm_,
Amia calva. Thesameas Text- Amia calva, 8mm. long. Trans-
fig. 29. ‘Transverse section. verse section. Beginning of the
hypophysial cavity.

appears to be continuous with that of the hypophysis, and this
fact puzzled me for some time, but I believe it to be without
significance and due to the close apposition of the hypophysis
to the gut-roof. In some sections the real discontinuity of the
membrane can be seen. The hypophysis is still in contact
with the gut-roof, and at this stage the begining of the hypo-
physial cavity can be observed (Text-fig. 32). Mesenchyme
begins to make its way between the floor of the brain and the
sut-roof, and in the next stage (Text-figs. 33 and 34), where
the hypophysis has separated off the endoderm, it is enclosed
by a layer of mesenchyme.

HYPOPHYSIS OF PETROMYZON AND AMIA 989

TEXT-FIG. 33.

Amia calva. 11 mm.long. The hypophysis with a hypophysial
cavity separated from the dorsal wall of the gut by connective
tissue,

TEXT-FIG. 34.

|
\
l ATS
LEO
' AX O@E2)
' AES LES i
A . CYT
Seeds Ae re
Ava 4 5 3 E C
SERS | cag Eee “es
= So ea A Np log (4) Zs ’ s
>, O- SN Ke 0. ODS <5 0 cay

SK ASIN OSCE EOFS %
oi} Whee 33 Bobs, i Nee

poe i)
SU (OPP, QS PSS AP ‘ee
= a Be 2 LAD <
ea 0s CABO ie = —_,
SS ORE OE
DES is Li yee
a a = a

Amia calva.

The same as Text-fig, 33, higher magnification.

290 G. R. DE BEER

The hypophysial cavity is now distinct with the cells arranged
radially around it. There is as yet no distinction between the
cells on opposite sides of the cavity, the differentiation into
pars anterior and pars intermedia has not yet appeared. On
the side of the brain there is no pars nervosa, though the
infundibular cavity and the recessus saccularis are already
very distinct.

SUMMARY.

1. The hypophysis of Petromyzon arises as a solid ingrowth.

2. The depression in which the hypophysis and olfactory
organs come to lie is formed by the great expansion of the upper
lip in conjunction with the sides of the anterior surface of
the head.

3. The beginning of the histological differentiation of the
glandular elements takes place from a solid strand before the
appearance of any cavity.

4, The hypophysial cavity arises late as a split in the thickness
of the hypophysis, and afterwards extends in both directions
communicating forwards with the external depression.

5. The homologue of the pars tuberalis is probably to be
found in the ‘ Uebergangsteil’ of Stendell, a chromophobe
portion situated between the pars anterior and pars intermedia.

6. The hypophysis of Amia is derived from the ectoderm,
thus agreeing with all other known forms.

7. It arises outside the stomodaeum on the anterior surface
of the head.

8. It is a solid ingrowth which loses all connexion with its
point of origin, and within which the hypophysial cavity makes
its appearance at a later stage.

9. Outside and in front of the stomodaeum is probably the
primitive position of origin of the hypophysis, without con-
nexion with either mouth or nose.

10. In Selachians and Amniotes where the fore-brain is early
very large and the cranial flexure pronounced, the hypophysis
arises further posteriorly and so is included in the hollow of the
stomodaeum.

HYPOPHYSIS OF PETROMYZON AND AMIA 291

11. Although probably primitively hollow, the rudiment
of the hypophysis is often solid. Such diversity is brought
about by embryonic developmental conditions, perhaps the
distance separating the rudiment from the infundibulum.

List oF LITERATURE CITED IN THIS PAPER.

Atwell, W. J. (1919).—‘‘ The Development of the Hypophysis in the
Anura’’, ‘ Anat. Record’, vol. 15, 1919. —

Baumgartner, E. A. (1916).—‘‘ The Development of the Hypophysis in
Reptiles ’’, ‘ Journ. Morph.’, vol. 28, 1916.

Biedl, A. (1913).—~“ Innere Sekretion’. Berlin.

‘ Dean, B. (1896).—‘* On the Larval Development of Amia calva’’, ‘ Zool.
Jahrb.’, 9, 1896.

YDohrn, A. (1883).—‘‘ Die Entstehung der Hypophyse bei Petromyzon
planeri’’, ‘ Mitt. Zool. Stat. Neapel’, vol. 4, 1883.

“Goodrich, E. 8. (1917).—‘* Proboscis Pores in Vertebrates”’, ‘ Quart.
Journ. -Micr. Sci.’, vol. 62, 1917.

Gétte, A. (1875).—* Die Entwicklungsgeschichte der Unke’. Leipzig,
1875.

—— (1883).—*‘ Ueber die Entstehung und Homologien des Hirnanhangs ”’,
* Zool. Anz.’, 1883.

Haller, B. (1896).—** Untersuchungen iiber die Hypophyse’’, ‘ Morph.
Jahrb.’, vol. 25, 1896.

Herring, P. T. (1913).—“*‘ The Pituitary in Vertebrates ”’, ‘ Quart. Jour.
Exp. Physiol.’, 1913.

Kupfier, K. von (1893).—‘ Studien zur vergleichenden Entwicklungs-
geschichte des Kopfes der Kranioten’. Lehmann, Miinchen and Leipzig,
1893.

’ Prather, J. M. (1900).—** The Early Stages in the Development of the
Hypophysis of Amia calva’”’, ‘ Biol. Bull.’, vol. 1, 1900.

’ Reighard, J., and Mast, L. A. (1908).—‘‘ The Hypophysis of Amia
calva’’, ‘ Journ. Morph.’, vol. 19, 1908.

/ Scott, W. B. (1888).—‘‘ Embryology of Petromyzon”’, ibid., vol. 1, 1888.

(1883).—* Development of the Pituitary Body in Petromyzon”’,
“Science ’, vol. 2.

vy Smith, P. E. (1914).—** Development of the Hypophysis in Amia calva”’,
“ Anat. Rec.’, 8.

y Stendell, W. (1913).—‘‘ Zur vergleichenden Anatomie und Histologie der
Hypophysis cerebri’’, ‘ Arch. f. mikr. Anat.’, vol. 82, 1913.

¥ —— (1914a).—* Die Hypophysis cerebri’’, ‘ Lehrbuch der vergleichenden
mikroskopischen Anatomie der Wirbeltiere ’, Oppel., part 8, Jena, 1914.

v

299, G. R. DE BEER

Stendell, W. (1914 6).—‘‘ Betrachtungen iiber die Phylogenesis der Hypo-
physis cerebri nebst Bemerkungen iiber den Neuroporus der Chordonier”’,
* Anat. Anz.’, 45, 1914.
* Sterzi, G. (1904).—‘‘ Morfologia e sviluppo della ipofisi nei Petromyzonti”’,
‘ Arch. Ital. Anat. e Embr.’, vol. 3, 1904.
Swale Vincent (1922).—‘ Internal Secretion and the Ductless Glands’.
London.
Tilney, F. (1913).—‘‘ Analysis of the Juxtaneural Elements of the Pitui-
tary”, ‘ Internat. Monatschr.’, vol. 30.
YWilley, A. (1894).—‘ Amphioxus and the Ancestry of the Vertebrates’.
Macmillan.
Woerdeman, M. W. (1914).—‘‘ Die vergleichende Anatomie der Hypo-
physis ”, ‘ Arch, f. mikr, Anat.’, vol. 86, 1914.

The Relation of the form of a Sponge
to its Currents.

By
G. P. Bidder, Se.D.1

With 12 Text-figures.

Aut zoologists know that from the large holes, which we call
oscula, on a sponge, an outgoing current may be detected
in life. During several months in Naples I investigated this
current, using litmus and carmine solutions, and carmine and
indigo in suspension. I worked with two calcareous species of
sponges, having oscula at the end of tubular prolongations,
which reach the size and shape of a child’s thumb in the case
of Leucaltis, and of a child’s finger in the case of
Leucandra aspera (Text-fig. 1). The solutions were either
placed on the surface of the sponge, to be sucked in by its
currents (Text-fig. 2), or dropped by a pipette through an
incision into the cloaca—the cavity of the tubular prolongation.
In the latter case the time taken for the colour to be thrown
out at the osculum, though liable to many corrections, afforded
on the whole the most trustworthy determinations of oscular
velocity : the cloaca bemg wider than the osculum, the observed

1 This paper was read before the British Association at Hull, September
1922. A preliminary note was published in ‘ Proc. Camb. Phil. Soc.’,
1888, vol. vi, p. 5. (See also ‘ Quart. Journ. Micr. Sci.’, vol. 38, p. 28;
* Proc. Roy. Soc.’, vol. 64, p. 61; and I. B. J. Sollas, ‘ Camb. Nat. Hist.’,
vol. i, p. 235.) The experiments were made in the Naples Zoological
Station in 1887, 1888, and 1889, where I occupied the Cambridge University
table, and in 1890, 1891, and 1892 at a table allowed me by the great
kindness of the late Professor Anton Dohrn. For a long time I proposed
to myself to make a further series of experiments to clear up doubtful
points, but recognizing that I shall not now do so, I have reconsidered
all the experiments this year (1922) and recalculated allresults and formulae.

294 G. P. BIDDER

cloacal velocity would be multiplied by 4, 5, or 6, as the case
might be, to obtain the oscular velocity. A pretty method was
arrived at accidentally (Text-fig. 3), when I found the coloured
jet marked by dark beads or nodes, caused by my pulse shaking

TEXT-FIc. l.

Leucandra aspera var. gigantea (A. 11).!

the pipette; the length between any two nodes, divided by
three-quarters of a second, gives the core-velocity at that part
of the jet.

The fastest oscular velocity recorded directly for the

1 This sponge came from the Porto Militare, and I regret that I have
drawn it erect instead of pendent. Vosmaer states that this metamp
of L. aspera is found only on the keels of boats (cf. p. 314).

FORM OF A SPONGE 995

L. aspera of Text-fig. 1 was 7 cm. a second, and this was the
basis adopted in the calculations of this paper. I have now
changed them, in consequence of the conclusion that the actual
mean oscular velocity when the sponge was in the sea was
8-5 em. a second (Appendix, Note 1).| From Parker’s experi-

TrEext-Fic. 2.

é ee
yen

y

ments on the pressure in Stylotella, a siliceous sponge,
its velocity is considerably higher than that of Leucandra;
we shall see later that this could be conjectured from its
structure.”

In quite still water such a current from Leucandra goes

* This was the conclusion from observations of length of jet and velocity ;
I cannot put it forward as an exact physical measurement, but as a final
judgement after considering upwards of 1,000 unsatisfactory and imper-
fect observations, of the nature described in the text. Note 6 (Appendix)
indicates that the figure 8-5 was a lucky judgement, and is probably close
to the true velocity.

* The velocity in Stylotella will be less than Parker calculates,
as he does not allow for the friction in the canals of the sponge.

NO. 266 x

296 G. P. BIDDER

10 or even 20 inches before coming to rest.1 Speaking exactly,
it does not really come to rest at this distance, but reaches a
velocity not higher than that of the slow return-current, slightly
indicated in Text-figs. 2 and 3, which is necessarily established
to fill up the region from which this water has been removed.
Text-fig. 4 is a diagram of the currents which must exist
around a bath-sponge in still, deep water. The swift vertical
jets from the oscula on the top surface carry the used and

TEXT-FIG. 3.

o> & > + soe) ee

beets jn cercep avon eae eer ee a 1 ena

Raa ote Rae Bi 6. te BS

fouled water to a stratum some feet above the sponge; slow
currents, in the plane at right angles to the jets, creep in from
all directions along the sea-bottom to feed the intaking pores,
which cover the general surface of the sponge. If the water
be absolutely still, there is established between these afferent
and efferent currents a re-entrant vortex, whose section is
a circle in any radial plane through the osculum.

The diameter of that circle I call the Diameter of Supply ;
and the angle between the directions of the intake and outflow
currents, which in sessile sponges (Text-fig. 4) is a right-angle,
and in pedunculated sponges (Text-figs. 9 and 10) is 110° to
120° or more, I call the Angle of Supply.

1 See Note 5, (12).

FORM OF A SPONGE 297

On these two factors depends the life of the sponge, or of
any other fixed or stationary organism in still water. The
outgoing current carries with it water which has been filtered
of food, in which carbonic acid has been substituted for oxygen,
into which the poisonous products of metabolism have been
excreted. In still, or nearly still water, the angle of supply
and the diameter of the circle of supply measure the chance
that some convection current or drift will carry away that water,

TEXT-FIG. 4.

useless for life, before the slow eddy of return brings it down to
the plane of the ingoing current. According to the distance
to which it is so carried is the percentage of clean, unused water
which enters the organism, and according to this percentage
is the chance of life of the organism ; and in a slow tidal channel
it is clear that the distance to which the foul water will be
carried by the tide before it is drawn back to the plane of the
ingoing currents, depends directly upon the length of the
oscular jet.

The length of this jet was shown by the experiments to vary
as the initial velocity—a result to be expected by elementary
theory, though a full theory would be difficult. With jets of
the same initial velocity, but from oscula of different size,

x2

298 G. P. BIDDER

the distance carried appeared to be proportional to the diameter
of the osculum, although, in consequence of the oscula used
having small range in size, this result was not so certain. The
rough formula indicated by the experiments is that, using
centimetres and seconds, the length of the jet approximates
numerically to twelve times the product: of its initial velocity
and the diameter of the osculum.

That among jets of equal velocity the distance carried should

TEXT-FIc. 5.

Rhagon.

be proportional to the oscular diameter, might also be expected
by elementary theory; simce per centimetre of length the
weight of water increases as the square of the diameter, while
the surface exposed to friction increases only as the diameter.
Consequently the ratio of the moving weight to the resisting
' surface increases as the diameter, and with equal velocity
a jet 4mm. wide may be expected to go twice as far as a jet |
2mm. wide.

This consideration enables us to understand the advantage
gained by the union of many unicellular flagellates to make
one thimble-shaped Olynthus, or by the union of many Olynthi
opening into a central cloacal cavity to form a Rhagon (Text-
fig. 5) or a Sycon. Suppose, in Text-fig. 6, we unite a hundred

FORM OF A SPONGE 299

Olynthi, each ejecting an efferent stream 1 mm. wide, and raise
up their colonial wall to enclose a common efferent aperture
1 cm. wide, thus forming a hypothetical Rhagon with an efferent
aperture having its area equal to the aggregate oscular area
of the 100 Olynthi. We can visualize the stream from each
Olynthus as a thread, and the stream from the Rhagon as a cord
of a hundred threads issuing side by side. We see at once

TExT-rigG. 6.

Efferent streams. Hypothetical Rhagon with wide aperture.

that most of the threads lie altogether inside the cord, with
their surfaces entirely protected by other moving threads from
friction against still water. The mass and initial velocity of
the Rhagon’s jet is that of its 100 constituent jets; but for
each centimetre of length the external surface of the Rhagon’s
jet 1s only one-tenth of the total surface of the constituent jets,
consequently the friction to which it is exposed is one-tenth
of their friction, and the combined jet will travel about ten
times as far as each of the constituent jets would go separately.

Coalescence is therefore advantageous in increasing the
diameter of supply ; and, as in all living things and in the

300 G. P. BIDDER

works of engineers, absolute magnitudes are determined by the
relations of the consequences of increase in length to the con-
sequences respectively of increases in area and in volume.

But sponges are not mere coalesced Protista: they have
varied cellular differentiation and at least two organs. The
first is the perforated membrane of tissue which is formed by
the flagellate cells, or on which they stand (Text-fig. 8), thereby
ensuring to them absolute separation of the water which leaves

TExT-FIc. 7.

Wall of Leucandra aspera.

them from the water which supplies them; this organ is
possessed by no choanoflagellate, and characterizes the sponges.
The second is the canal system, which we may better call the
hydraulic organ, in which the form of every part is wonderfully
modified for the advantage of the whole aggregate. In the
canal system all agree that progressive changes have occurred
again and again in similar order along many lines of descent
in sponges. We can show these changes to have as one of their
necessary consequences large increase in oscular velocity, with
consequent increase in the diameter of supply, and that there-
fore each successive change has left the sponge a more self-

FORM OF A SPONGE 301

sufficient food-catching machine than its ancestors, so as
eventually to make a fixed organism which is independent of
chance currents from waves and tides.

Watching under the microscope a flagellum with such a
rapid period of vibration that the eye only sees a mist

TEXT-FIG. 8.

Wall of flagellate chamber, with two afferent pores.

terminated by the two extreme positions, we are apt to think,
and to say, that the flagellum is moving rapidly ; and much
theory has been written about the mechanics of the collar-
cells and choanoflagellates, on the assumption that food-
particles are thrown through the water at express-train speed
by this rapidly moving flagellum, as a savage hurls a projectile
with a throwing-stick, or as civilized mandrives a ball with a club.

302 G. P. BIDDER

Really, far from express-train speed, no part of the most
rapidly moving flagellum ever attains the rapidity of motion of
a snail. We forget, as we look through a microscope magnifying
1,000 diameters, that though distance is magnified, time is not
magnified, and therefore any velocity is zoo What it appears
to be. I have estimated the vibrations of the flagella in healthy
and lively collar-cells at twenty to the second.’ The flagellum
is 80 long; therefore if it vibrates through an angle of 60°,
its tip travels 30 » each half-vibration and 60 » each complete
vibration, making twenty vibrations 1:2 mm. per second ;
which is 14 ft. an hour. This is the speed of only the extreme
tip of the flagellum, the base being motionless ; so that some
7 ft. an hour is the mean speed of the middle poimt of this
invisibly rapid flagellum.

I confess that I have often adduced this calculation to
show why there are no flagella in the cloaca of Leucandra,
where the mean velocity is 14 centimetres a second, or twenty
times the speed of the flagellum. But, really, viscous flow at
a mean velocity of 1 cm. means an axial velocity of 2 cm.
diminishing gradually to zero on the walls. The bareness of
the cloaca we must attribute to the fact that the sponge
flagella do not, and cannot, act like oars, an action which
requires direction of movement and nervous co-ordination.
All observers agree that the movements of sponge-flagella are
neither co-ordinated, synchronous, nor in parallel planes.”
A collar-cell flagellate surface is comparable mechanically to
a seine-net with a number of fishes fixed by their gills in the
meshes. Their movements cannot establish a current along the
face of the net: that would involve their tails all striking
strongly in the same direction and weakly in its opposite.
But if the net be fixed, the uncoordinated movement of their
tails will draw water through the net from the side on which
are their noses. Sponge collar-cells are similarly capable of

1 * Quart. Journ. Micr. Sci.’, vol. 38, p. 17.

2 Permanent sections of Oscarella show the pinacocytal flagella of
the afferent canals looking as if they work as oars, and far enough apart to
do so without collision; say, one to the area of 30 collar-cells.

FORM OF A SPONGE 303

drawing water through the perforated membrane on which
they stand ; from its position the walls of the cloaca cannot be
so perforated, so it bears no collar-cells. It is true that sponge-
flagella can accelerate a current down which they lie, but for
this their position is least useful when they stand on a wall
parallel to the current, and most useful on a wall at night
angles to it. ‘To this latter position they become more and
more limited in the progressive development of the canal
system, in which flagellate tubes become progressively shortened
until only the perforated hemispherical end of the chamber is
left.

The remarkable achievement of the perfected hydraulic
organ in sponges is that from this waving of hairs zoooo0
of an inch in thickness at a mean speed of 7 feet an hour,
there is produced an oscular jet with an axial velocity
of over half a foot a second (280 times the speed of the flagel-
lum), which in Leucandra throws to the distance of 9 inches
five gallons a day or a ton in six weeks (Appendix, Note 3).

It is, of course, clear that when we combine many slow
streams into one narrow channel the velocity is increased.
I computed in several sponges the aggregate cross-section of the
stream through the body in various parts of its course. In the
Leucandra of Text-figs. 1 and 7, 10 cm. long, there were
about 24 million flagellate chambers, with a total transverse
area of 524 sq. cm., or 1,700 times the area of the osculum,
which is 0-031 sq.cm.; so that the mean velocity at the osculum
is 1,700 times as high as in the chambers.

At 8} em. a second, a quarter of a cubic centimetre (0-26 ¢.c.)
will issue from the osculum each second, and to replace it
a quarter of a cubic centimetre must have passed through the
24 million chambers, that is ggso of a cubic millimetre through
each, or 116,000 cubic»; which through a chamber 54 » diameter
(transverse area 2,300 sq. ») implies a rate of flow of 50 p
a second, or z7oo Of that at the osculum, as above.

The flagellate chamber (Text-fig. 7) is a blind, thimble-
shaped tube, the water entering through perforations in the
walls. The total area of the wall-surface is eight times the cross-

304 G. P. BIDDER

section of the tube, so that the water enterimg has this trans-
verse area of channel as it passes the flagella (Text-fig. 8), and
therefore the velocity of only 6a second. Below the flagella
half the channel is occupied by the necks of the collars, so that
between them the water moves at the rate of 12 a second,
and a particle of food takes a second to travel the length of a cell.

Slow as we thought the movement of the flagella, at } mm.
a second, the water on which they act is stationary by
comparison, and they can get on their full work. And the
remarkable anomaly in sponges, that through considerable
evolution their motor-cells remain their ingestive cells, ceases
to be surprisimg when we realize that both functions are alike
localized at the position where the current is slowest.

The one second during which a food-particle is passimg the
collar-cell is not such a short time for its capture as would at
first appear. It allows of a good many events in the cell’s life:
we know of twenty double vibrations of the flagellum, with the
metabolic cycles which they imply. The biological magnitude
of an interval of time is measured by the number of events -
which can happen in it; and since every event requires the
motion of something from one position to another, therefore
where the distances between positions are smaller, events can
happen more rapidly. Every motion is produced by an
acceleration—such as gravity, or the stress of contracting
protoplasm—and with a given acceleration the time required
to move over a certain distance from rest is as the square root
of that distance: a stone takes a second to fall 16 ft., but to
fall a quarter that distance takes half a second. Therefore in
a biological world whose linear dimensions are zooo those
of our own, there will be some thirty times as many events
in a second as in our own, since thirty-two times thirty-two
is a thousand; and I suggest that in the biological time of the
flagellate cell the one second during which a particle of food is
passing would compare with half a minute in our external life.
(We must put the adjective ‘ external’, because our psychical
events which happen ‘ with the speed of thought ’, are events

1 Note 6 gives reason for supposing double these velocities in full health.

FORM OF A SPONGE 305

in a cellular or intracellular world where distances may be even
smaller than those about a flagellate cell. Hence, always, their
rapidity has been noted as of a different order to that of common
external events.)

There is also a purely physical point deserving attention
in the conditions of the world under an immersion lens. When
we watch flagella working under a high power, the water seems
to have lost its fluidity: a particle moved with apparent
swiftness by a flagellum loses its motion at once. The general
appearance is as if the flagella were labourmg in thick gum,
or treacle; and to understand microscopic physics it is a
serviceable short-cut to think of the water as treacle. The
energy of a projectile to overcome the resistance of the medium
through which it is thrown is as its mass multiplied by the
square of its velocity ; loss of energy from the resistance is as
its surface multiplied by the velocity and by the distance
traversed. We magnify its apparent mass as the cube of the
magnification, and the square of its apparent velocity as the
square of the magnification; so that the apparent energy of
the projectile is magnified as the fifth power ; but the energy
lost, measured by surface multiplied by velocity and distance
traversed, is only magnified as the fourth power. Consequently,
with 1,000 diameters, the water offers a thousand times the
retarding effect which we expect, on the projectile which we
think we see; and the ratio is even higher with the small
projectiles which concern us. With velocities among which
25 ft. an hour is the swiftest, at distances among which z5'o5
of an inch is very great, the viscosity of water is the pre-
dominant phenomenon ; and this world at which we are looking
is a world of pushing, not of throwing.

When the flagellum pushes in with its stroke a minute drop-
let of water into the flagellate chamber (Text-fig. 8), it creates

1 Sir J. J. Thomson kindly informed me that, according to Stokes’s law,
the resistance of water to the movement of a minute sphere is proportional
to the diameter, not to the square of the diameter. This would make the
apparent retardation under the microscope a million instead of a thousand
times the expected retardation.

306 G. P. BIDDER

a pressure there which forces an equal amount of water out
into the efferent canal (Text-fig. 7), and so the pressure created
in the chamber is transmitted to the efferent canal, and thence,
with a loss by friction, to the cloaca. The chamber is distended
by this flagellar pressure, as the elastic bag of a squeeze-pump
is distended, and the stretching of the chamber-walls is resisted
by the surface tension and elasticity of the tissue, as the
stretching of a soap bubble is resisted by the surface tension of
soapy water in air. The text-books have long pointed out that,
as the canal system is specialized, the diameter of the chambers
become smaller, and they change from cylinders to hemispheres
and spheres. This change, therefore, directly increases the
possible pressure in the chambers, and therefore the diameter
of supply. ‘There is twice the pressure in a soap bubble 1 in.
in diameter that there is in a soap bubble 2 in. in diameter ;
and similarly, reduction in the size of flagellate chambers pro-
portionally increases the pressure which their tension can
balance. In Leucandra I calculate from computation of the
oscular current ? and the friction in the canals that the pressure
is between 2mm. and 1¢mm. of water in the cylindrical
flagellate chambers 54, wide. ‘The same tension would
support double the pressure in spherical chambers of the
same diameter, and four times the pressure in spherical chambers
of half the diameter ; so that from my results the pressure in
a sponge with spherical chambers 27 » in diameter would be
2} to 5mm. of water, and in 85 p chambers would be 2 to
4mm. In spherical chambers of 35», by direct experiment,
Parker found the pressure of 34 to 4mm. of water in Stylo-
tella. We may therefore conclude with some safety that
the pressure in Leucandra is close to 1 mm. of water,
and that the healthy tension of the chamber-wall tissue
is nearly alike in this and in Stylotella (for the latter
0-00034 gm. weight per centimetre, or less than go of the
surface tension of petroleum in water). In the smaller spherical
chambers of Stylotella this tension can support three times
the pressure of the large cylinders of Leucandra, and so
1 See Appendix, Note 6.

FORM OF A SPONGE 307

we may expect nearly twice the oscular velocity, and conse-
quently nearly twice the diameter of supply, with this
admittedly more specialized type of canal system.

This pressure of 1 mm. water is transmitted, with a loss
from friction, through the efferent canals and cloaca to
the open osculum, where the potential energy of the com-
pressed water is converted into the kinetic energy of the swift
oscular jet.

We are familiar with such conversion of the potential energy
of compressed air in an air-gun into the kinetic energy of the
escaping bullet, and of the potential energy of the compressed
water at the bottom of a cistern, converted into the kinetic
energy of the jet from a garden-hose. We know well, with the
garden-hose, that when the water will only go 3 ft. from the
open hose, it will throw a jet 30 ft. from a fine tube-nozzle.
This is because with the open pipe, delivering perhaps 10 gallons
a minute, there is a flow of 5 ft. a second through the 1 in.
hose-pipe, with great loss of energy in every foot of the pipe
from friction. Putting on a nozzle which will only deliver
1 gallon a minute, the velocity within the hose is lowered to
6 mn. a second, 33 of the loss by friction is avoided, and
the potential energy of the cistern is transmitted almost
undiminished through the hose, in any part of which there is
nearly the full pressure of the cistern. If we stop the nozzle
with the finger, we have the full pressure of the cistern through-
out the hose, and can feel it on the finger. This is not per-
ceptibly diminished by allowing a fine thread of water to
escape, and its velocity of issue approximates to the full
theoretic velocity due to the cistern head, but it does not travel
far owing to the smallness of its mass compared with the surface
of friction it exposes to the air. As the jet is allowed to increase
in volume, so does the velocity increase in the hose-pipe, and
the consequent successive loss of energy in each yard of its
length ; and we can feel the pressure on the finger noticeably
diminish and can see that the velocity of the jet consequently
decreases, though with its greater volume it travels farther.

So with the sponge. The narrow osculum, 2 mm. diameter, in

308 G. P. BIDDER

Leucandra is comparable to the tube-nozzle on the hose. It
means that pressure is transmitted from the flagella through
the water of the canals, the water moving so slowly that loss
in friction still leaves enough energy at the osculum to make
a strong stream. As with the hose-pipe, were we to close the
osculum more, there would be less energy lost, and the velocity
at the osculum would rise, with a pin-point hole, from 8} em.
to 13 cm. a second, but the tiny jet could barely travel 15 em.
instead of 24 cm., while the quantity would not be a hundredth
of that necessary for nourishment. With the osculum half its
existing diameter, the velocity would be increased to 11 em.
instead of 8$em., but the length of the jet reduced to three-
quarters and the quantity of water to only one-third what we
now findthem. On the other hand, were the same Leucandra
shaped like a cornucopia, with the widest part of its cloaca as
osculum, the quantity of water passing would be increased by
one-sixth of the existing quantity, but the velocity would be
only 14 cm. a second, and the jet consequently less than half the
present length. For a given pressure, acting through a given
length of channel of fixed width, there is an optimum value
for the size of the osculum (as we have all found with the
garden-hose) above and below which it will not carry so far ;
with measurements of the flagellar pressure, and of the number
and dimensions of the canals, an equation can be made to
determine this optimum value. I have calculated it for the
Leucandra of Text-fig. 1 (Note 5), but my computations of.
the canals and currents are not close enough to say more than
that the theoretically best diameter for the osculum of this
specimen is 2-6mm.-++-}mm.; the preserved diameter being
2mm. Rough computations for other specimens, and for
a Sycon, confirmed the conclusion that the osculum is always
at any rate near to the optimum size for producing the greatest
diameter of supply ; and that this is the explanation of the
small and definite oscula which we have all noted as charac-
teristic of the majority of sponges with high canal system.

The secret of the repeated development of this common
type of sponge is the reduction of internal velocities, so that

FORM OF A SPONGE 309

a larger proportion of the energy produced by the flagella is
transmitted to the osculum in the untaxed form of pressure.
The cloaca must be much wider than the osculum for the
cloacal current to be slow; hence the bottle-shaped cloaca
with a small orifice, which led our forefathers to compare to
the stomach and mouth of animals the pressure-chamber

TEXT-FIG. 9.

Clathrina blanca. Angle of supply 110°.

and vent of the hydraulic organ of sponges, evolved with the
advantage, not of the retention and digestion of food, but of
the forcible removal of excreta.

I shall not here discuss the sluggisheurrents of Clathrina
and Leucosolenia; nor the evolution of subdermal spaces,
and their possible development as muscular pumps with the
collars acting as valves (as in the strange reversed canal

310 G. P. BIDDER

system described by Vosmaer in Spirastrella).! But
modifications concerned with the angle of supply demand
brief notice in regard to our main thesis.
Stalks of greater or less length (Text-fig. 9) increase the
angle of supply in many sponges, and when this is the case
in sponges of any size, the osculum opens out, as admixture of
incoming and outgoing streams becomes less probable, and
oscular velocity therefore less important. This gives rise to the

TEXT-FIG. 10.

15

Pie

Calyx lieberkuhnii (modified from O, Schmidt).

well-known Neptune’s-cup form (Text-fig. 10), found in many
sroups, and the expanded lip mtervenes between the two
streams.

If this cup becomes set on one side (Text-fig. 11), the efferent
stream passes away forwards, while the intake is at the back
of the cup, and the angle of supply approaches to 180°. In
still water, with the oscular jet horizontal, the length of the
jet becomes infinite, whatever its velocity ; because with the
angle of supply 180° there is no back eddy, and the friction
on the surrounding water serves by degrees to set 1t in motion

1 Siboga-Expeditie, “The Genus Spirastrella’, p. 49.

FORM OF A SPONGE 31]

TEXT-FIa. 11.

pene

Phakellia ventilabrum (combined from Johnston and Bowerbank).

TEXT-FIG. 12.

Phakellia conulosa (reconstructed from Dendy).
NO. 266 Y

312 G. P. BIDDER

in the direction of the jet. So, since maintenance of oscular
velocity is now of no advantage, the osculum opens out com-
pletely to allow the maximum quantity of flow regardless of
oscular energy. . The cloaca becomes a flat surface (Text-fig. 12),
and the whole sponge a disk or fan with one intake face and
one outflow face. If such a sponge be in really still water the
surrounding water will by degrees be all set into slow motion,
in the direction from intake to outflow, and the condition of
life approaches that of the deep-sea sponges and polypes.

It is a puzzling fact at first that in most of the Hexactinellida
we can detect no hydraulic evolution nor hydraulic efficiency ;
puzzling until we remember that in the great depths where
they live, an unchanging current sweeps slowly from the
poles to the equator. They have but to spread a net across it,
and, whatever their mechanical inefficiency, they have incoming
and outgoing streams 180° apart; the flagella have only to
work the water through the many meshes formed by the feet
of the collar-cells. The cavity of the Hexactinellida is no
pressure-chamber: it is even perforated to let the onflowing
water sweep out that which is befouled. Food is brought to
them, waste is taken away. For them in their eternal abyss,
with its time-like stream, there is no hurry, there is no return.
Such an organism becomes a mere living screen between the
used half of the universe and the unused half—a moment of
active metabolism between the unknown future and the
exhausted past.

APPENDIX.

Note 1. Tue VeEuocity or A JET OF MEASURED LENGTH.

The empiric formula, in centimetres and seconds,
LE =(12 + 2) V.B. x {1 —0-028 (20 —1)}
where L is the length of the oscular jet,
V its mean velocity at the osculum,
B the diameter of the osculum,
t the temperature in degrees Centigrade,

FORM OF A SPONGE 313

may be of use to naturalists who are able to observe in a large
tank the length of the jet from a molluse or other fixed organism,
and the width of the aperture. It is thus possible, for jets
not exceeding 180 cm., to obtain a value for the approximate
issuing velocity, and therefore of the quantity of water filtered
by the animal and of the effective work performed by its
cilia. The correction for temperature is theoretical.

Thus the figure of 8-5+1-5 cm. per second, which I have
finally adopted as the mean oscular velocity, at 18° C., of
Leucandra aspera gigantea (Text-fig. 1), healthy and
in natural conditions, is deduced from the oscular streams of
30 cm. and even 45 cm. long, measured from recently gathered
sponges. These indicate a mean oscular velocity of 8 cm. per
second, and from the reduction which this undergoes under the
best aquarium conditions, it may be concluded that the velocity
of any gathered sponge is less than in the sea.*

The formula was deduced from measurements of velocity
on sponges some time in the tanks whose streams had shortened
to 15 em., 10 cm., 5 em., or even 1 cm. ; such languid velocities
being more easy to measure. The constant (12+2) has been
adopted as a compromise between a group of experiments by
different methods which agree on the value 1 /VB =10-5+1 and
several experiments which group themselves about L/VB = 14.°
The distance to which the cloud drifts will obviously be affected
by the position of the jet with reference to the walls of the tank :
with accurate physical experiment and definition of the condi-
tions a closer result could be obtained; but for aquarium
observations it is probable that the conditions cannot be
sufficiently identical to reduce greatly the probable error of the
ratio L/VB.

1 Tn cold water, flagella have a lower metabolism with which to drive
a more viscous fluid, and the energy of the oscular current is less. When
increase of temperature first becomes injurious, diminishing vitality is
compensated by diminishing viscosity ; with still higher temperatures the
change in viscosity is less and the injury greater (cf. ‘ Linn, Soe. J.’, 34,
p. 317).

2 Experiments in June and July, temperature unfortunately not observed.

WZ

314 G. P. BIDDER

Norse 2. REGULATION OF THE SIZE OF THE OSCULUM.

In two examples of Leucandra aspera (one with an
iris-like oscular sphincter) I cut off the oscular end for preserva-
tion, and observed that the aperture contracted to one-third
of its normal diameter. Mechanical stimulation failed to induce
further contraction.

There was more than once evidence that the oscular aperture
contracted when the current through it grew feebler, though
the contraction was not sufficient to keep from diminution
either the oscular velocity or the length of the jet. In a
Leucaltis, observed over 24 days, the osculum contracted
to half its original diameter, while the length of the jet dimin-
ished from 18 cm. to 1 em. I give this series of measurements,
which shows also that when the current recovered (probably
with a lower temperature) the osculum widened again.

July 7, 7, 8 9, 9 11, 11, .11,. 20, 21,. 22, 23, 29, 29, SOlae eee
Diameter of -38 -35 -30 -25 -18 -32 -28 -25 -20 -205 -20 -20 -17 -16 -18 -18 -18 —
osculum
Length of

jet

16+ 9 765 213 8 1465 6-7 6244 9 -5 l-60ies

Parker found in Stylotella (‘ Journ. Exp. Zool.’, 8, p. 784)
that the oscula close when the external water is still; this may
be called ‘ Parker’s reaction’. In Leucandra and Leu-
caltis it is shown by the sphincter of the osculum being
inhibited from contraction by the movement of water over its
internal surface. If the reaction is such that there is neither
contraction nor relaxation in response to the velocity character-
istie of the species (Note 5 (12)) the optimum osculum will be
maintained at all stages of growth.

Note 3. DELIVERY OF WATER AND SECRETION OF
LIME.

Inthe Leucandra aspera of Text-fig. 1, taking the mean
oscular velocity at 8-5 cm. per second (see Note 1), the area of
the osculum being 0-031 sq. cm., the delivery per second was
0-26 c.cm., or 16 ¢.cm. per minute; that is -9 litre an hour,
90 litres in 4 days, and a ten in 45 days (see Note 6).

“7

FORM OF A SPONGE ae ea 25

This sponge was gathered on May 20, and from Vosmaer
(‘ Mitth. Z. S. Neapel’, v, pp. 486, 487) was probably only a
month old, during which time its volume and the volume of
its delivery must be supposed to have increased by some
40 per cent. every day in geometrical progression. Then the
total amount of water passed during its whole life would be the
equivalent of 34 days of its final delivery, and it would have
extracted food from a total weight of 80 kg. of water.
By weighing the preserved sponge alternately in alcohol
and water I found the total volume 0-31 ¢.cm., weight 0-59 gm.,
*. sp. gr. 1:9. Allowing 2-5 for the sp. gr. of the spicules and
1-4 for that of the dry protoplasm, this gives

Total Volume. Dry Weight.

c.cm. gm.
Spicules of Leucandra Seb uegE Wat: ue. bi 46 “38
Protoplasm . : « ALG “21
“31 59

Therefore from 80 kg. of water the sponge abstracted a total
weight of 0-38 gm. of carbonate of lime, or 0-005 gm. per kg.,
or about one-third of the amount of carbonate of lime which
can be dissolved in pure water free from carbonic acid, and
about one-twelfth of the total lime in an average sample of
sea-water. If we suppose the sponge’s life to have been longer,
or the osculum to have been more dilated (Note 6), the per-
centage proportion of lime extracted is proportionately less.

Notre 4. FAN-SHAPED SPONGES.

There is a possibility that these, like the Hexactinellids, are
tound always in a permanent current, on which they depend
for subsistence.

After this paper was read at the British Association, Sir W.
Herdman kindly informed me that in the deeps off Scotland
(where the sponges were found from the figures of which
Text-fig. 11 was made) there are many places where the
current sets only one way. And Professor Stanley Gardiner
added that in 30 fathoms off the Seychelles (where the sponge

316 G. P. BIDDER

was found of Text-fig. 12) there is a constant current down
the slope away from the shore. The hypothesis is possible that
the fan-shaped form only occurs in response to the stimulus
of a constant current (compare Note 2) across which its plane
is extended; but that if the current turns tidally from all
points of the compass, the fan grows up across each direction
in turn, so that an open cup is formed.

The advantage of the fan in still water is shown in the text,
but in a current turning tidally the efferent stream will be
driven back on the sponge for half the day. In such a position
the vertical oscular stream is the common form, because this
forms an equal angle with the supply from whatever point it
comes. Oyster-shaped sponges, with oscula on the edges,
are possibly from a channel where the tide runs alternately
from two opposite points ; they may be called * pectinate ’.

Sponges living always in the surf, or long flexible sponges,
such as Chalina oculata, which point downstream from
their stalks, have of course no need to do more than to lift the
outflowing water sufficiently from their surface for the current
down which they lie to bear it free of their more apical
parts.

The conditions discussed in this paper affect sponges which
are left long in tide-pools, and sponges which inhabit depths
where there is inappreciable wave-motion, and where currents
are feeble.

Note 5. CALCULATIONS OF PRESSURE AND OPTIMUM SIZE
OF OSCULUM.
(Mathematical basis of the paper.)
The loss of energy in a tube from resistance due to viscosity
in unit of time is
Sayud,
where w = velocity,
b = length of tube,
y= index of viscosity.

FORM OF A SPONGE 317

Therefore if # be the loss of energy per second in JN similar
tubes, E = 87p- Nu?b.*

Let g be the quantity of water passing per second through
the V tubes in parallel, the loss of energy per second is

2
— Bry. bE,
U

where «@ is the aggregate area of the cross-sections of the tubes.
In the sponge the whole of the water passes in succession
through
(1) Afferent canals ; (3) Efferent canals ;
(2) Flagellate chambers ; (4) The cloaca.

For the whole system, therefore, the loss of energy due to
resistance is the sum of the losses in these four systems, which

may be represented
NV.b

Y< = 81u( > — = ie
Let 8an( >>: PP Ff. (1)

Then the energy reaching the osculum per second is
i= Pq = Fq?, (2)
where P is the pressure maintained by the action of the flagella.

But if v be the velocity at the osculum, the energy of the jet

per second is 2

H=p.95> (3)

where p is the density of the water ; and if # be the diameter
of the osculum

* It was the late Professor Sir G. G. Stokes, in 1888, who supplied me with
this formula, and a clearly written exposition of its meaning, which could
be understood by the ignorant. I cannot allow my use of it to appear in
print without a tribute to his kindness to a then young man, unknown to
him, with no recommendation but a somewhat shameless request for
assistance.

318 G. P. BIDDER

Therefore from (2) and (3)

y
= P-7 Fav
poiee peas (5)
2 p p

If 7 be the length of the oscular jet, I tind by experiment
(p. 298 and Note 1) that

lL = Cve. (6)
* Substitute = = v in (5), therefore
A
/? + ol PP
ait 35 Oe =p” (7)
2 a F wl 2Px
a a ae a ; 8
C2 u Wp aC p (8)
Now J has a maximum or minimum when o =0(0. Differen-
tiating (8),
QLidl ar ae : gil 4Px
aaa t 30 btt gz) =
Therefore when a = 0, a = ue
da 2 pC pax
Lt se
C” 37Fa

Substitute this value in (7), therefore 7 has a maximum or
minimum when
SP? 19 eer
seh) “at - 3p z ioe

* T have to thank Mr. G.I. Taylor, F.R.S., who has very kindly read
the first proof-sheets of this paper, for giving me this simple demonstra-
tion of (9), from (5) and (6), to replace my own very clumsy differentia-
tion.—June 8, 1923.

FORM OF A SPONGE 319

=, eee
SH ew Soe 8 3
See ee 2
Ont F2°2P” 37% F?

or

(9)

*. when at

(Note that the value of C is not involved in this equation.)

P
F2°

The negative and imaginary roots do not concern us, for «
is necessarily rational and positive; and since P and F are
finite and positive, this equation gives a finite and rational
value for x, and therefore, from (5), / is also rational, positive
and finite.

But there is no jet from an aperture of infinite radius,
because the velocity is zero, and there can be no jet from
a closed aperture ;

= 1-081 p 5: or taking p = 1-025, a* = 1-11

therefore when c=0,7=0; and when 2=0, /=0;

therefore the value of « given by (9) corresponds to a greater
value of 2 than that when «= 0, or when « = ©; and as it is
the only positive and finite value for w for which = == 10);
therefore the corresponding value of / is the only positive
maximum of /, and the length of the oscular jet has its
greatest value when the diameter of the osculum has the
value X, where

4 iP

If a second sponge precisely similar to A. 11 were to have the oscular
end of its cloaca united with the oscular end of the cloaca of A. 11, to
make a twin sponge with a single osculum, we should have twice the
number of afferent and efferent canals, &c., and two cloacae ; so that in
the computation of F, N and a would both be doubled, with the result
that, comparing F, for the twin sponge with F of the original sponge,

by (1)
QNb\ F
Fy = 8p ( Da )=5:

2
Fane

(10)

320 G. P. BIDDER

With twice the number of flagellate chambers, g, will be approximately
equal to 29, therefore F',q, = Fg. The pressure in the flagellate chamber
depends solely on the structure and vigour of the flagellate cells, velocity
there being so slow that kinetic energy is always negligible, therefore
P, = P, whatever the number of chambers.

p vw

But, from (4), P= “ort Fq. (11)

. 2
i = + Fo q.

Therefore v* =», and the velocity from the osculum is the same in
sponges of similar canal systems. irrespective of size; that is, of the
number of similar units which are grouped to expel water by one
osculum ; and for Leucandra aspera gigantea in health, from

Note 1, L=12x85B=100B. (12)
Now by (10) X=103a/ VP.
F

Therefore for the twin sponge
DG = 1-03 GEE — 1 Os 2/1 = xe 2:
2 a . ey NI =

Similarly, if m similar units, for each of which the optimum oscular
diameter is X, be united to one osculum, and X,, be the optimum
diameter of this, then

Xm = X J/m.

With similar sponges the external volume may be taken as the
approximate measure of the number of similar units aggregated into
one individual ; or, more conveniently, the product of length, breadth,
and thickness may be taken as the measure. Calling this product MW, its
value for A. 11 is 4-1 ¢c.c., and we shall find in Note 6 that for A. 11
X = -25+-08.

Therefore, for Leucandra aspera gigantea of similar canal-
system, the optimum diameter of the osculum in centimetres is numeri-
cally

Wa =
€ — shy —— =e mI
Xn 25 Nie 12 ./M,
and generally for any one species and metamp
Xin = JM,

the area of the osculum varying as the volume of the sponge.
For a sponge like the bath-sponge, with N oscula, the sum of whose
diameters is > X, approximately,
iM

> xX=Na aly = VM

FORM OF A SPONGE 321

Nore 6. ARITHMETICAL TESTS, DATA, AND CONCLUSIONS.

From camera lucida drawings of the canals and their aper-
tures in the Leucandra aspera of Figs. 1 and7 (‘ A. 11’
of my records) F is computed in the table below to be
180 +30, the relative velocities being confirmed by the times
taken by litmus to pass through the walls of the sponge and
through its cloaca (p. 293). Using this value in (11), with
p = 1-025, v = 8-5+1-5, we find for A. 11 the equation

P = 37.0+ 14+ (1200+ 420)”, (13)

so that, if « = -20, the diameter of the osculum measured in
spirit, then
P= 85+30 =-9 mm.+38 mm. of water;

and by equation (10), the optimum diameter of the osculum
X = -235 +-025.

Note 2 shows the need of a probable correction in the
value of «% The Leucandra‘A. 11’ was 4 hrs. under ex-
periment before being preserved, and its velocity had sunk to
less than half its original value. If we may reason from the
observations on Leucaltis we should expect the diameter
of the osculum to have been reduced by 380 per cent., and
therefore that for a velocity of 8-5 cm. it had been -28 cm.
wide, instead of the -20 cm. measured after preservation.

With x% = -28,
P = 1381+45 = 1-33 mm.+-46 mm. of water,
AX = -26 4-03.

The conditions under which the diameter of the osculum is
equal to the theoretically best diameter are found by sub-
stituting the value of X in (10) for a in (11), giving the
relations

/ eva
(14)

ee its ala fo
t= 1-335,

822, G. P. BIDDER

So that, with F = 180,
i i= Be Xk ole eee
Ua Oe = a 256, ee

the latter being acceptable values. On the other hand, we
may ascertain what error is indicated in F, if we assume that
in healthy life « was -28 em. as suggested above, and that P
(cf. p. 306) was exactly proportional to the pressure of 370+ 25
found by Parker in the 35, spherical chambers of St ylo-
tella:

The chambers of Leucandra being cylindrical and 54y.
in diameter,

by assumption # = -28;

dy oth ney Sd. put Ane

The diminution of viscosity with warmth would reduce
F from 180 at 15°C. to 150 at 22°C. This is no unlikely
temperature for the Porto Militare in the summer, so thatthe
investigation shows the data in harmony with each other and
with the theories of the paper.

The arithmetical coincidence suggests that the probable
errors of the observations are overstated. The errors could
not be calculated statistically, and I could not estimate them
at smaller figures. I much regret having been unable to
claim greater exactitude.

FORM OF A SPONGE 323

TABLE OF DATA FOR CALCULATION OF VELOCITIES AND
RESISTANCE (L. aspera. ‘A. 11’).

M Aggregate | .. F
Number. ean | transverse N ,| Velocity.
length. : = -28 —+
5 area. e'! a) (2
Afferent
canals. . 81000 06 | 4-2 99 06 | -13
Subchoanal
space . . = — 200 — 0013) -0026
Flagellate i
chambers. 2,250000 019 x 3* 52-5 1 005 | -010
Efferent |
canals. . 5200 08 2:5 47 10 |-21
Cloaca .. 1 10 x 3* -21 3) tS) 4 Oe
= (1) 031)
Osculum. . 1 let 062 = 8 8.5
Whole
sponge. 180 |

Therefore, for A. 11, F = (180 +30) x {1—-024 (¢—15)}, where ¢° is the
temperature Centigrade.

* The factors } and 3 allow for the water entering by holes along the
walls.

+ At 15°C. From equation (1).

(1) Calculated for oscular diameter -20, area -031, delivery -26.

(2) Calculated for oscular diameter -28, area -062, delivery -53.

4%

A)

na

On the Development of the Hypobranchial,
Branchial, and Laryngeal Muscles of Cera-
todus. With a Note on the Development of
the Quadrate and Epihyal.

By
F. H. Edgeworth, M.D.

With 39 Text-figures.

As is well known, Wiedersheim stated that laryngeal muscles
exist in Lepidosiren and Protopterus, but are absent in Cera-
todus. He admitted, however, that the specimen investigated
was badly preserved.

In a recently published paper (1920) on the laryngeal muscles
of Amphibia, I suggested that, possibly, they might be found
in better specimens. Owing to the skill and perseverance of
Dr. Bancroft I have come into possession of some well-preserved
heads, and also of a series of embryos up to the stage of 80 mm.
in length for purposes of investigation. The material also
enabled me to examine the development of the hypobranchial
and pharyngeal muscles, which, together with other structures,
have been the subject of an elaborate memoir by Greil.
In the description given by Semon of the development of
Ceratodus the embryos were depicted in a series of stages
numbered from 1 to 48, the last mentioned and oldest stage
described being an embryo of 17-3 mm. Greil’s description is
based on Semon’s stages, and also extended to stage 48. The
embryos described in this paper had been fixed in formalin,
and their lengths are given in millimetres. Their relation
to Semon’s stages will be found in an appendix.

NO. 267 Z

326 F. H. EDGEWORTH

A tabular statement of the synonyms of the names employed
has also been appended.

Occipital Myotomes and Nerves. Fiurbringer
(1897) stated that, in the adult stage of Ceratodus, there are
two or three occipital and two occipito-spinal nerves.

X va eit avd bya
Vy Za ava bya
Vy zy aq bya
Vy zy avd bya ‘vad © coming out between

skull and vertebra.

The Plexus cervicalis is formed from x, y, z, or y, z. Sewertzoff
(1902) stated that in a 15-7 mm. embryo there are five occipital
myotomes in front of the first vertebral arch, the first two being
in process of reduction. There are ventral spinal roots corre-
sponding to the fourth and fifth. He identified the fourth
myotome with x of Firbringer, so that the first five are u, v,
w, X,y-

- Greil (1918) stated that no vertebral arches are present in
a 15-7 mm. embryo, and that they are developed in a 17-8 mm.
embryo. Five occipital myotomes are present in front of the
first vertebral arch in the latter stage. The first two have
no corresponding nerve-roots, the third has a (variable)
ventral root, the fourth a ventral root, and the fifth a ventral
and a (variable) dorsal root. The Nervus hypobranchialis
(which = the Plexus cervicalis of Firbringer) is formed from
the variable third, and the fourth and fifth nerve-roots. He
identified these five myotomes with v, w, x, y, z of Fiirbringer’s
terminology.

On comparison of these statements it would appear that
(1) Sewertzoff’s 15-7 mm. embryo was a little more advanced
in development than Greil’s 17-8 mm. embryo—vide infra.
The embryos I have examined at these stages agree with
those of Greil, so that, in all probability, Sewertzoff’s embryo
was somewhat shrunken. (2) Sewertzoff regarded the fourth
myotome—counting from before backward—and Greil the

MUSCLES OF CERATODUS 327

third myotome as myotome x of Firbringer’s classification.
The explanation of the difference of opinion is that Sewertzoff’s
embryo was one in which the variable nerve x was absent.
The variation certainly occurs, e.g. in an embryo of 20 mm.
Nerves x, y, Z were present; im one of 26mm. (vide Text-
fig. 26) there were only nerves y and z. I therefore follow
Greil’s nomenclature. Atrophy of myotomes takes place from
before backwards, as stated by Sewertzoff. Thus in a 20 mm.
embryo only a few fibres of myotome v persisted, in one of
24mm. myotome v has altogether disappeared and also the
ventral part of myotome w.

Coraco-hyoideus and Genio-coracoideus. Greil
stated that the Coraco-hyoideus is developed from downgrowths
of the third to the sixth myotomes, i.e. myotomes x, y, Z, a.
These downgrowths separate from the myotomes above, fuse
together, and form the Coraco-hyoideus, which extends from
the shoulder-girdle to the hyoid bar. The primordium of the
Genio-coracoideus separates from the ventral edge of the
third myotome (i.e. foremost) constituent of the Coraco-
hyoideus in a 13-9mm. embryo, and fuses with its fellow,
forming a median muscle which elongates forwards to the jaw
and backwards. The posterior end forks right and left, and in
a 17-8 mm. embryo—the latest stage investigated—reaches the
antero-posterior level of the first branchial arch. This method
of development of the anterior constituent of the hypobranchial
spinal muscles is not a usual one, and the initial stages were
not depicted. ‘The first figures given, i.e. Nos. 424-9, show
the Genio-coracoideus already developed as a median muscle,
partly in front of and partly underlying the anterior ends of
the Coraco-hyoidei.

I find that in a 9-5 mm. embryo (Text-figs. 1-5) the peri-
cardium extends forwards to the hyoid segment, and its anterior
end is ventral to the hinder edge of the thyroid outgrowth.
The primordium of the hypobranchial spinal muscles lies
laterally to the pericardium, and, as this lessens in size, approxi-
mates to its fellow. In front of the pericardium the two columns
come together and lie beneath the thyroid. The anterior part

Z2

328 F. H. EDGEWORTH

of the primordium consists solely of yolk-laden cells, the
posterior part of muscle-cells.
In a 10-5 mm. embryo (Text-figs. 6-9) the pericardium has

TExtT-Figs. 1-5.
Embryo 9-5 mm., transverse sections; Text-fig. 1 is the most anterior.

l brlateral pl

= . = ae) S -
> ae =; DY Te2
DBSZO OES rte). o
_ Cs 5 gos
woe io}

hypobr. sp. me?

ABBREVIATIONS TO 'TEXT-FIGURES.

aud, cap., auditory capsule. ceratobr., ceratobranchial. coraco-hy.,
Coraco-hyoideus. const. br., Constrictor branchialis. epi-br., epi-
branchial. genio-cor., Genio-coracoideus. g. petros. ix, ganglion
petrosum ix. hypobr.sp.ms., primordium of hypobranchial
spinal muscles. Jateral pl., lateral plate. Jev., Levator arcus
branchialis. mm. pl., muscle-plate. Ma., myotome x. WN2.,
nerve x. operc., opercular fold. parachord.c., parachordal
cartilage. pericard., pericardium. eric. perit. duct, pericardio-
peritoneal duct. thyr., thyroid body. Ist vert. arch, Ist vertebral
arch. Roman numerals, cranial nerves.

retreated a little, and its anterior end is 56 » behind the thyroid.
The primordium of the hypobranchial muscles has separated
into anterior and posterior parts—the Genio-coracoideus and

MUSCLES OF CERATODUS 329

TEXT-FIG. 2.

perwardiuum hypobr sp ms

TEXT-FIG. 3.

330 F. H. EDGEWORTH

TEXT-FIG. 4.

perwcard ;

TEXT-FIG. 5.

MUSCLES OF CERATODUS 331

the Coraco-hyoideus. The Coraco-hyoideus lies laterally to the
pericardium, and its anterior end is 56 » behind the thyroid.
The Genio-coracoidei form a A-shaped structure. The anterior
median part is ventral to the thyroid: it extends behind
this for 104 y, diverging into two lateral ends which lie beneath
the anterior ends of the Coraco-hyoidei. The Genio-coracoidei
consist of yolk-laden cells, the Coraco-hyoidei of muscle-cells.
In an 11 mm. embryo the anterior end of the Genio-coracoidei
extends a little farther forwards—in front of the thyroid, and
ina 12mm. embryo reaches Meckel’s cartilages. Its cells become
transformed into muscle-cells in a 18 mm. embryo.

The Genio-coracoidei extend slowly backward, diverging into
right and left halves. These ireach the level of the third
branchial arch in a 20 mm. embryo, and become attached to
the lateral edges of the median cartilage—called ‘ sternum ’
by Greil—which forms the ventral constituent of the cartila-
ginous shoulder-girdle, in the 28 mm. embryo.

The above may be summarized in the statement that the
primordium of the hypobranchial spinal muscles extends
forwards, laterally to the pericardium, and in front of this
joins its fellow. It then separates into Genio-coracoideus and
Coraco-hyoideus. The Genio-coracoidei, in contact with each
other from the first, form a median structure which extends
forwards to the jaw and backwards to the shoulder-girdle.

Fiirbringer stated that the Genio-coracoideus of the adult
form extends from the mandible to the shoulder-girdle (coracoid
and clavicula) and has one tendinous inscription on its inner,
i.e. dorsal surface. This, he said, gives rise to the idea that it
originally consisted of two myomeres, but this structure may
be secondary. This latter supposition is confirmed by the fact
that in a 28mm. embryo (in which the posterior end of the
muscle has reached the ‘ sternum’) there are no inscriptions
in the muscle. The same explanation applies to the three
inscriptions depicted in the muscle by Maurer.

Innervation.—Firbringer stated that the Coraco-
hyoideus and Genio-coracoideus are innervated by the Plexus
cervicalis, i.e. Nervi spinales x, y, z, or y, Z.

892) F. H. EDGEWORTH

TExt-Fies. 6-9.

Embryo 10-5 mm., transverse sections; Text-fig. 6 is the most anterior.

Enadodernr

TEXT-FIG. 7.

Endoderm

MUSCLES OF CERATODUS 333

TEXT-FIG. 8.
Endoderm

334 F. H. EDGEWORTH

Greil stated that the Coraco-hyoideus is innervated by the
N. hypobranchialis (derived from Ni. occips. x, y, z, or y, Z)
and by the R. hypobranchialis of N. occipito-spin. a, whilst
the Genio-coracoideus is innervated by the R. hypohyoideus
of N. posttrematicus ix. He did not refer to Firbringer’s
statement.

I find in a 27 mm. embryo that an anterior branch of the
N. hypobranchialis enters the posterior end of the Genio-
coracoideus, and have not found any branch of the [Xth nerve
entering the muscle.

The Genio-coracoideus and Coraco-hyoideus of Ceratodus
are homologous with the Genio-thoracicus and Coraco-hyoideus
of Protopterus and Lepidosiren. Agar has shown that the latter
is developed from myotomes x, y, z; but did not describe the
development of the Genio-thoracicus.

The Genio-coracoideus of Ceratodus resembles the Genio-
coracoideus s. Coraco-mandibularis of Selachii and the Genio-
branchialis s. Branchio-mandibularis of Ganoids in that it is
formed from the anterior constituent of the hypobranchial
spinal muscles and subsequently grows backwards overlapping
the posterior constituent—Coraco-hyoideus.

On the Source of the Branchial Muscles.—Greil
stated that in a 5-9 mm. embryo the mesoderm lateral to the
branchial region of the alimentary canal—between this and the
ectoderm—is continuous with the epithelium of the pericar-
dium, and is to be regarded as ‘ lateral-plate ’. In 6-6 to 9-8 mm.
embryos these lateral plates degenerate into connective tissue,
and their place is taken by downgrowths from the first and
second myotomes. The cells of these downgrowths are distin-
guishable from those of the lateral plates by the shape of their
nuclei and the later absorption of their yolk-granules. Pro-
cesses from the first myotomes penetrate the first three branchial
arches, whilst one from the second myotome forks over the
sixth gill-cleft into the fourth and fifth arches. In a 10-2 mm.
embryo these downgrowths separate from the myotomes above
and become the source of the branchial musculature.

I find that in an embryo of 9-5 mm. (Text-figs. 1-5) the cell-

MUSCLES OF CERATODUS 3385

columns in the first and second branchial segments, which
consist of yolk-laden cells, are not continuous with the myotome
above, but are continuous with the pericardial wall below.
I interpret them as lateral plates. In an embryo of 10-5 mm.
(Text-figs. 6-9) the pericardium has retreated a little, and its
anterior end is 56 » behind the thyroid, just in front of the lower
end of the first branchial segment. In this segment is the
first branchial muscle-plate, the lower end of which is detached
from the pericardial wall. In the second branchial segment
is the second branchial muscle-plate, the lower end of which
is continuous with the pericardial wall. The difference between
the two segments is owing to the slight retardation in develop-
ment from before backwards—from segment to segment.
I use the term ‘ muscle-plate’ to denote those cells of the
lateral plate which are obviously muscle-cells and the primordia
of the branchial muscles. Though still containing yolk-granules
they are distinguishable from the other cells of the lateral plate.

In explanation of the figures it should be added that in the
9-5 mm. embryos the branchial arches slope downwards and
slightly backwards, in the 10-5 and 12 mm. embryos they are
vertical, in the 16mm. embryo they slope downwards and
forwards.

I thus fail to find any continuity between the myotome above
and the cell-columns in the first and second branchial segments
in a 9-5 mm. embryo, i.e. at a stage when, according to Greil,
such a continuity exists. The same is true of a 9mm. embryo.
Further, in a 10:5 mm. embryo what is obviously the second
branchial muscle-plate is continuous with the pericardial wall.
The difference in length of these embryos are so slight that it
is improbable that—as is demanded by Greil’s theory—what is
‘lateral-plate > in the 9-5 mm. embryo is replaced by down-
growth from myotome in the 10-5 mm. embryo. Again, Greil’s
theory fails to explain why a muscle-plate derived from myo-
tome downgrowth should ever be continuous with the pericardial
wall. TI also fail to find any differences in the shape of the cell-
nuclei between the upper and lower parts of the cell-columns
of these segments in the 9-5 mm. embryo, as is stated by Greil.

336 F. H. EDGEWORTH

I am therefore of opinion that the evidences presented by
these embryos are sufficient to warrant a rejection of Greil’s
statement that the branchial muscles are derived from down-
growths of the myotomes above, and to show that they
are derived from the lateral plates—as is usual in Verte-
brates.

What is said above in relation to the first and second branchial
segments applies also to the third, fourth, and fifth.

Text-Fie. 10.

Embryo 1] mm., transverse section.

Constrictores branchiales and Levatores ar-
cuum branchialium.—Both these muscles are developed
in the first four branchial arches, but a Levator only in the
fifth. Their anatomy in the adult stage was first described by
Jaquet (1897), and subsequently—with more accuracy—by
K. Firbringer (1904). Greil stated that the Levators in the
first four arches are developed from the upper part of the
mesoblast (i.e. myotome downgrowth) in the arches, whilst
the Constrictors are developed from cells given off from the
muscles at their upper and lower ends. His words are ‘ welcher
jedoch keinen Rest des primiren den ganzen Bogen durch-

MUSCLES OF CERATODUS 337

ziehenden axialen Mesoderms bildet, sondern durch Ziige spin-
deliger Zellen, welche von der Dorsal- und Ventralseite
(Levatores und Interbranchiales) stammend, vorwachsen,
geschlossen wird’. ‘This occurs in the description of the
17-8 mm. stage (p. 1855). No figures were given illustrating this
derivation of the Constrictors.

In 10-5 and 12 mm. embryos, and as is additionally shown in

TExT-FIGs. 11-13.
Embryo 12 mm., sagittal sections ; Text-fig. 11 is the most external.

st as Nok
1 branch.m. ple

eS ball Gots pl

Text-figs. 11-13 taken from sagittal sections of a 12 mm.
embryo, the first and second branchial muscle-plates form
vertical strips through the whole extent of the arches. The
ventral end of the first branchial muscle-plate is detached from
the pericardial wall, and has grown a little forwards—towards,
but not yet reaching, the Ceratohyal. The ventral end of the
second branchial muscle-plate is continuous with the peri-
eardial- wall (Text-fig. 14). These muscle-plates, slightly
convex outwards, pass down external to the primordia of the
branchial bars.

338 F. H. EDGEWORTH

TExt-FIG. 12.

a% __-2“branch m pl.

0
i
e
\ :

a branch. m.pl

MUSCLES OF CERATODUS 339

In an embryo of 13-5 mm. (Text-fig. 15) the ventral end of
the first branchial muscle-plate has separated and grown
forwards to the Ceratohyal, forming the Branchio-hyoideus.
The rest of the muscle-plate persists. The lower end of the
second branchial muscle-plate has become detached from the
pericardial wall and grown inward, forming the Transversus
ventralis ii, and is detached from the vertical strip above.

TExtT-Fic. 14.

Embryo 12 mm., transverse section.

The development in the third and fourth arches is similar
to that in the second.

In an embryo of 16mm. (Text-figs. 18-20) the first four
branchial bars have separated into Epi- and Cerato-branchial
elements and have chondrified, the process being most complete
in the first. The upper end of the first branchial muscle-plate
has separated into an inner and outer portion—the inner is
the Levator and is inserted into the Epi- and Cerato-branchial 1,
the outer is the upper end of the Constrictor branchialis, which
is continued down through the arch to its lower end. In the
second branchial arch separation into these two constituents
is not complete, and in the third and fourth branchial arches
has barely begun, owing to the progressive retardation in
development from before backwards. In a 17-5 mm. embryo

340 F. H. EDGEWORTH

the separation has occurred in these also. In a 27 mm. embryo
the lower ends of the Constrictors have grown forwards. Hach
is attached to the lower end of the Ceratobranchial of the

Trext-FiIGs. 15-17.

Embryo 13-5 mm., transverse sections ; Text-fig. 15 is the most
anterior; Text-fig. 17 is 832 behind Text-fig. 16.

Subare,reckv Trans.vent.v perwara.

next anterior arch—the condition described in the adult by
K. Fiirbringer.

The above-recorded observations show that the Constrictores
branchiales are the direct descendants of the branchial muscle-

MUSCLES OF CERATODUS 341

TExtT-FIG. 17.

Myotome
Trans.vert.v
eric. pertton .

ye) perctton

Trans.vent v prolf*fom epithe.

TExtT-FIGs. 18-20.
Embryo 16 mm., horizontal sections ; Text-fig. 18 is the most dorsal.

lev. / eS . ee - aud. caps

<

Bo. C89,
-peog
B46 93Q°

mpl ii

m. pl.iv

mplv

NO. 267 Aa

342 f. H. EDGEWORTH

plate, and that the Levatores arcuum are separated from their
upper ends.

In the case of the fifth branchial arch Greil stated that the
mesoblast (i.e. myotome downgrowth) forms the Dorsgo-
pharyngeus. JI find that the development is similar to, but not
identical with, that of the more anterior arches. In the

TExtT-FIc. 19.

11 mm. embryo (Text-fig. 10) there is a lateral plate, the ventral
end of which is continuous with the pericardial wall. In the
13-5mm. embryo (Text-figs. 16, 17) the muscle-plate is formed ;
and its ventral end has separated from the pericardial wall
and grown inwards forming the Transversus ventralis v. In
the 16mm. embryo some few fibres of the muscle-plate are
attached to Ceratobranchial y., but the majority are continuous
with the Transversus as in the 13-5 embryo. The condition is
a more primitive one than in other arches. The upper part
might be called a Levator or a Constrictor branchialis. The

MUSCLES OF CERATODUS 343

Cucullaris is separated from it in a 14mm. embryo—as was
described by Greil.

It is not known whether the Constrictores branchiales and
Levatores arcuum are present in Protopterus and Lepidosiren.
Pinkus (1895) did not mention any individual branchial muscle

TExt-FIG. 20.

,
w22—— mpl. v
a

in his description of the cranial nerves of Protopterus. The
Constrictores are homologous with those of Selachi, the
Levatores with those of Ganoids and Amphibia.
Subarcuales recti and Cleido-branchialis.—
The Subarcualis rectus 1. s. Branchio-hyoideus was first described
by Furbringer (1897), who stated that it is a muscle passing from
the Ceratobranchial 1. to the Ceratohyal. It was also described
in the following year by Jaquet. Greil stated that it is developed
Aa2

344 F. H. EDGEWORTH

from the ventral end of the mesoblast (i.e. myotome down-
growth) in the first branchial arch. As stated above, I find it
to be developed from the ventral end of the first branchial
muscle-plate.

Behind this muscle are two others, also longitudinal in
direction—the Subarcualis rectus v. and the Cleido-branchialis.
The latter was first described by Firbringer under the name
Coraco-branchiales, of which he said five are present, passing
from the shoulder-girdle to the five Ceratobranchials. The
fifth is the broadest and some of its fibres are attached to the
skull. The others are slender. Greil described the development
of these muscles as follows. The Subarcualis rectus v. is
developed in a 17-8 mm. embryo by forward growth from the
mesoblast (i.e. myotome downgrowth) of the fifth branchial
arch. Its anterior end becomes attached to the ventral ends
of the branchial bars. The Cleido-branchialis is derived from
a process of the mesoblast of the fifth branchial arch which
grows forwards, separating into three or four pointed extremities
(‘ Zipfel ’) which reach the ventral ends of the branchial bars.
It forms an aberrant band of muscle which grows in the same
direction, but is separate from the Subarcualis rectus y. and
gains a secondary relation to the shoulder-girdle.

I find that these two muscles are developed from a single
primordium which appears in a 13-5 mm. embryo (Text-fig. 16)
as a slight forward growth from the junction of Levator v.,
and T'ransversus ventralis v. This primordium extends
forwards, reaching the antero-posterior level of the third
branchial arch in a 14mm. embryo and that of the second
branchial arch in a 16 mm. embryo (Text-fig. 28). In the last-
mentioned stage its hinder part has increased in vertical depth
and its postero-inferior angle is attached to the Cleithrum.
In a 28mm. embryo (Text-fig. 29) the primordium has fully
separated into the muscles it forms, viz. the Subarcualis
rectus v. passing from the fifth to the first Ceratobranchial,
and the Cleido-branchialis. This latter muscle is posteriorly
attached to the ventral surface of the Cleithrum and separates
into fascieuli which, passing dorsal to the Subarcualis rectus v.,

MUSCLES OF CERATODUS 345

are inserted into the ventral ends of all five Cerato-branchialia.
The fasciculus inserted into the first is confluent with the
anterior part of Subarcualis rectus v.

Innervation.—According to Firbringer and Greil, Sub-
arcualis rectus i. is innervated by the [Xth. I can confirm
this. According to Firbringer the Cleido-branchialis is
innervated by the Plexus cervicalis; according to Greil
both the Subarcualis rectus v. and the Cleido-branchialis are
innervated by the N. ultimus vagi (Quartus Vagi), 1.e. the nerve
to the fifth branchial arch. I can confirm Greil’s statement.

The Subarcualis rectus 1. s. Branchio-hyoideus is homologous
with the similar muscle in Lepidosiren, Protopterus, and
Amphibians. Its innervation is not known in these Dipnoans.
In Amphibia, as in Ceratodus, it is innervated by the [Xth.

The morphological nature of the hinder longitudinal muscles
is uncertain. Fiirbringer held that the Cleido-branchialis
is homologous with the Coraco-branchiales of Selachu. Greil
did not express any opinion other than that quoted above.

The muscles in Ceratodus are developed from a single
primordium which grows forward from the fifth branchial
arch. ‘The posterior end of the muscle gains a secondary
relation to the Cleithrum, and subsequently an almost complete
separation into two muscles takes place. The shoulder-girdle
is situated far forwards and has an oblique position—from
dorso-posterior to ventro-anterior—its lower part underlying
the branchial region. If this represents the phylogenetic
development of the muscles, as is probable, the original form
was probably a Subarcualis rectus v., passing from the fifth
to the first Ceratobranchial, and the Cleido-branchialis is
a secondary muscle. The developmental evidence thus leads
to rejection of Firbringer’s theory.

It is not known whether there is a Subarcualis rectus v. in
Protopterus and Lepidosiren, but Firbringer described a
homologue of the Cleido-branchialis in these Dipnoans and
stated that the innervation is, as in Ceratodus, from the Plexus
cervicalis.

The Subarcualis rectus v. is probably derived from a Sub-

346 F. H. EDGEWORTH

arcualis rectus v. passing from the fifth to the fourth Cerato-
branchial, and resembling the Subarcualis rectus iv a of Urodela
in its forward extension to the first Ceratobranchial. I do not
know of any homologue of the Cleido-branchialis in other groups.

Transversi ventrales ii, 111, 1vV, V are present—
Greil stated that they are developed from the ventral ends of
the mesoderm, i.e. myotome downgrowth, in the second, third,
fourth, and fifth branchial arches by inward growth. I find that
they are developed by inward growth from the ventral ends of
the branchial muscle-plates (vide supra). I would add that
Transversus ventralis iv. is not always developed, possibly
owing to the relatively great size of Transversus ventralis v.
Greil held that the Subarcualis rectus 1. s. Branchio-hyoideus,
developed in the first branchial arch, is serially homologous
with the Transversi ventrales of the hinder arches. But it is
difficult to think that a longitudinal muscle in one arch is
serially homologous with a transverse one in another when
neither changes its direction during development. In larvae
of Ichthyophis and Siphonops a Subarcualis rectus 1. and
a Transversus ventralis 1. are both developed in the first
branchial arch. In Ceratodus, too, a Subarcualis rectus and
a Transversus ventralis are developed in the fifth branchial
arch. ‘Transversi ventrales iv. and v. occur in Lepidosiren and
Protopterus. The Subarcuales recti and Transversi ventrales
are homologous or serially homologous with those of Ganoids
and Amphibia.

Transversus ventralis v. and Sphincter oeso-
phagi et laryngis.—The Transversus ventralis v., as stated
above, is developed in a 13-5mm. embryo, as an inward
growth from the ventral end of the fifth branchial muscle-
plate (Text-figs. 16,17). It extends back behind the posterior
edge of the sixth gill-clefts to a greater degree laterally than m
the median line, the distances being 96 w and 40 yp, 1. e. the pos-
terior edge of the muscle is concave from side to side. The
muscle in subsequent stages spreads backwards below the
anterior part of the oesophagus (Text-figs. 21-4). This is
concurrent with a backward shifting of the larynx (vide infra),

MUSCLES OF CERATODUS 347

so that in a 30mm. embryo—the latest embryonic stage
investigated—the larynx is not covered by the muscle.

TEXT-FIGs. 21-4.
Embryo 20 mm., transverse sections; Text-fig. 21 is the most anterior.

cesoph.

Greil represents the Transversus ventralis v. in his figure
of a model of a 17-8 mm. embryo (Taf. lxiv, fig. 3) with a
markedly convex posterior edge. But I have not found this
in embryos of 14, 15, 16, 17-5, 20, 24, 28, and 30mm. It 18
coneave from the first and remains so during the extension

348 F. H. EDGEWORTH

backwards of the muscle, i.e. the posterior extension takes
place as fast laterally as in the mid-line.

The posterior part of the Transversus ventralis v. forms the
ventral constituent of the Sphincter oesophagi et laryngis, but

TEXxtT-FIG. 23.

cesoph

ctl
ge

Cm
Lal
hes

Se Sn
VAS m6
EN ON

trachea

no distinction can be made between the muscles. They form
a continuous structure. The edges of the ventral constituent
slightly lap round the lateral edges of the oesophagus in a 16 mm.
embryo (Text-figs. 21, 22).

In a 24 mm. embryo (Text-fig. 25) a downgrowth takes place
on each side from myotome x—downwards, backwards, and
inwards, towards the upper part of the oesophagus, a little in

MUSCLES OF CERATODUS 349

front of the larynx. In one of 26 mm. (‘Text-fig. 26) this down-
growth has separated from its myotome, and there is a second

TExtT-FIa. 25.

Mex
an Mentest.x
sc down growth
\ Nae from Mx
YQ eee
cesoph . Trans, vert X

Embryo 24 mm., transverse section.

TExt-Fic. 26.

overt arch

perachord c ’ \ down-growth .
down growth

from My
from Mx ly i

Embryo 26 mm., sagittal section.

downgrowth from myotome y. In an embryo of 28 mm.
(Text-fig. 27) this, too, has separated from its myotome, and

350 F. H. EDGEWORTH

the two downgrowths form together a muscle-plate dorsal to
the oesophagus. The Sphincter oesophagi et laryngis is thus
developed from two constituents—a ventral derived from a
backward extension of Transversus ventralis v. and a dorsal
derived from the downgrowths of myotomes x and y.

The only changes which take place between this condition
and the adult form (as determined by transverse sections) is
that the Sphincter is completed by medial and lateral spread of
its dorsal constituents, and it extends a little farther back-
wards in the mid-ventral line so as to underlie the larynx.

Abide Bigies Pall

<e ” Y ie
a SRL SF ¥ bf
Sy a LE
= 5 ee vent.v
/ <<
aesoph..

Embryo 28 mm., transverse section.

The Transversus ventralis v. is said by Greil to be innervated
by the N. ultimus vagi, i.e. the nerve of the fifth branchial
arch. My observations confirm this, and I can add that no
additional branch of the vagus is developed in later stages
for its hinder part, i.e. the ventral part of the Sphincter
oesophagi et laryngis. This is in harmony with its develop-
ment. The dorsal part of the Sphincter—developed from
myotomes x and y—is innervated by branches of the Ni. ocecips.
x and y, or by the latter only when there is no N. occip. x.

Wiedersheim (1904) described the adult condition of the
Sphincter oesophagi et laryngis in all three Dipnoi,! and

' He gave figures of Protopterus and Lepidosiren, but not of Ceratodus.

MUSCLES OF CERATODUS 351

Géppert (1904), subsequently, in Protopterus. These writers
employed the term ‘ Constrictor pharyngis ’ on the theory that
it represents the musculature of atrophied hinder branchial
segments. But the theory receives no support from the
developmental phenomena in Lepidosiren, as described by
Agar (1907), nor could I see any evidence in its favour in
Ceratodus. I have therefore employed the term * Sphincter
oesophagi et laryngis ’.

Agar stated that the ventral portion of the Sphincter
oesophagi et laryngis of Lepidosiren is derived from cells budded
off from the inner walls of the pericardio-peritoneal ducts, and
its dorsal portions from downgrowths of the occipital myotome y
(‘ possibly, but not probably, the downgrowth extends to x
and z also’). These two constituents coalesce and form a
complete sphincter muscle.

As, however, the ventral part of the Sphincter is continuous
anteriorly with Transversus ventralis v. in all three adult
Dipnoi, and this is not excluded by the figures given by Agar
of the first stage of its development in Lepidosiren, it is possible
that this part is, as in Ceradotus, due to a backward extension
of Transversus ventralis v.

The ventral part of the sphincter of Lepidosiren is said by
Agar to be innervated by the N. muscularis vagi. This is
apparently the nerve of the fifth branchial arch, and the
innervation would thus agree with that of Ceratodus. But
the description given by Pinkus of the innervation of the
muscles in the branchial region in Protopterus is very vague,
and exact information is needed. The dorsal portion of the
Sphincter of Lepidosiren is innervated, in accordance with its
derivation, by Nervus occip. y (Agar); Lepidosiren and
Ceratodus are similar in this respect.

Larynx.—Neumeyer (1904) stated that the larynx is
developed in a 10-9mm. embryo immediately behind the
branchial region, Kellicott (1905) that it developed ina 11-6 mm.
embryo in the ventral wall of the pharynx near the commence-
ment of the oesophagus, Greil (1918) that it developed in a
11-6 mm. embryo at the level of the fifth branchial segment,

352 F. H. EDGEWORTH

i. e. behind the sixth gill-clefts. No observer made mention of
any subsequent backward shifting of the larynx.

I find that there is a slight trace of the laryngeal outgrowth
in a 10-5 mm. embryo, but it is quite clear in a 11 mm. embryo
(Text-fig. 10) as a median ventral outgrowth of the pharyngeal
and oesophageal epithelium, extending from the anterior border
of the fifth to 8 » behind the posterior border of the sixth gill-cleft.

TEXtT-FIG. 28.

Ceratobry

-)

Embryo 16mm. Model of Ceratobranchialia and Subarcualis
rectus v.!

In a 12mm. embryo the continuity of the laryngeal epithelium
with the pharyngeal and oesophageal epithelium extends from
16 w in front of the anterior border of the sixth gill-cleft to 72 u
behind it. In a 13-5 mm. embryo the continuity extends from
72 » behind the sixth gill-cleft to 152 » behind it. Ina 20 mm.
embryo it extends from 480 » behind the sixth gill-cleft to 600 u
behind it, and the oesophagus opens into the stomach 140 p

1 In Text-figs. 28 and 29 cnly the front parts of the Cleithrum and
Coraco-scapulare are represented. The epibranchialia i-iv are not shown.

MUSCLES OF CERATODUS 353

behind the posterior border of the larynx. The pre-laryngeal
region of the oesophagus at this stage is thus 460 yp, the laryngeal
region 120 yu, and the post-laryngeal region 140 in length.
The above can be summed up in the statement that the larynx

TEXxT-FIG. 29.

coraco-scap.

Embryo 27 mm. Model of Ceratobranchialia, Subarcualis rectus v,
Cleido-branchialis, and shoulder-girdle.

Text-Fic. 30.

———— ——— === Sph. oesoph.

(ventrat part)

1 Text-figs. 28, 29, and 30 were drawn by Miss Cross.

354 F. H. EDGEWORTH

is developed as a median ventral outgrowth. of the endoderm
in the last (fifth) branchial segment and anterior end of the
oesophagus, and subsequently shifts backward relative to the
sixth gill-cleft so as to be situated entirely in the oesophageal
region. During this period the larynx increases in size, though
to a far less extent. Thus in the 12 mm. embryo the epithelium
of the larynx is continuous with the endoderm above over
a length of 86 » from front and back; in the 20 mm. embryo
over a length of 120 w. The trachea begins to be formed in the
13-5 mm. embryo as a posterior outgrowth of the larynx.

The larynx of Ceratodus is thus developed in the same
position as in Menopoma, but unlike that of Menopoma it
subsequently shifts backwards. The backward extension of
Transversus ventralis v. is in relation to this. There is no
similar extension of Transversus ventralis iv. in Menopoma.

Oesophageal and Laryngeal Muscles.—tThe only
statement that Greil made in regard to these muscles is that
in a 13-9mm. embryo the free mesoderm cells increase in
numbers round the oesophagus. These cells are chiefly of
paraxial origm. They form a mantle round the gut and become
for the most part smooth muscle-cells.

I find that a Constrictor oesophagi and a Constrictor laryngis
are present in the adult, and that they are developed from
cells which are budded off from the epithelium of the pericardio-
peritoneal ducts in an 11 mm. embryo (Text-fig. 10). These
cells increase in numbers and spread round the larynx and
oesophagus, and in a 16 mm. embryo form the above-mentioned
muscles, which are better developed in a 20 mm. embryo from
which the figures are drawn (Text-figs. 22-4). The Constrictor
oesophagi extends from the level of the anterior edge of the
larynx to the posterior end of the oesophagus. In the laryngeal
region its fibres are interrupted ventrally and pass towards the
epithelium of the larynx through the fibres of the Constrictor
laryngis. Behind the larynx the fibres encircle the oesophagus.
The anterior part of the Constrictor oesophagi thus acts
additionally as a dilatator of the larynx, but this part is con-
tinuous with the posterior part, and remains so up to the adult
state. A separate Dilatator laryngis is not formed.

MUSCLES OF CERATODUS 355

The Constrictor laryngis consists of short muscle-cells
encircling the larynx in a horizontal plane. In transverse
sections they are most obvious as such just in front and behind
the larynx. The Constrictor laryngis gradually increases in
size, and in the adult forms a muscle of many fasciculi which
are penetrated by the dilatator fibres of the Constrictor
oesophagi.

Innervation.—Ina 27 mm. embryo the N. intestinalis x.
passes back dorso-lateral to the oesophagus just within the
upper edge of the ventral constituent of the Sphincter oesophagi
et laryngis. At the anterior edge of the Constrictor oesophagi
it gives off a ventral branch which passes to the larynx, the
Constrictor oesophagi and Constrictor laryngis.

The oesophageal and laryngeal muscles of Ceratodus thus
resemble those of Menopoma in that they are differentiated
from cells which are proliferated from the splanchnic layer
of the coelomic epithelium.

Pinkus described in Protopterus a fine twig from the N. intes-
tinalis x. to the mucous membrane of the pharynx and larynx.
This is the homologue of the laryngeal branch of the N. intes-
tinalis x. in Ceratodus.

The evidence hitherto available suggests the probability
that the Dilatator and Constrictor laryngis, and the Dilatator
laryngis (and also the Constrictor oesophagi if microscopical
examination shows it to be present) of Lepidosiren will be
found to be developed from the cells which were shown by Agar
to be proliferated from the pericardio-peritoneal ducts.

But, however this may be, comparison of the adult anatomy
of Ceratodus, Protopterus, and Lepidosiren shows that the
laryngeal muscles of Ceratodus—consisting as they do of a
Constrictor laryngis and the (unseparated) fibres of the Con-
strictor oesophagi which act as a dilatator—are the most primi-
tive existing in vertebrates.

I have the pleasure of thanking Dr. Bancroft for the
embryonic and adult stages of Ceratodus, and the Bristol
University Colston Society for defraying the expenses incurred
in the investigation.

356

F. H. EDGEWORTH.

RELATIONSHIP OF SEMON’S STAGES TO LENGTHS OF EMBRYO.

Stage 35= 5-9 mm.

» 36= 6-4
» 37= 66
3 £88 83
» 39= 9-0
, 40=.9:8
» 41=10-2
» 42=10-3

Ceratodus.
Genio-coracoideus.

Coraco-hyoideus.

Protopterus.
Genio-thoracicus.

Coraco-hyoideus.

Stage 43 =10-8 mm.

, 44 =109 ,,
» 45 =116 ,,
» 454=12-0_,,
» 46 =139 ,,
1 at be ee
» 48 =178=)

SYNONYMS.

Genio-coracoid, Humphry.

Coraco-mandibularis, Jaquet, Greil.

Coraco-mandibularis s. Coraco-cleido-
mandibularis, Fiirbringer.

Unnamed, Maurer.

Coraco-hyoid, Humphry.

Coraco-hyoideus, Jaquet.

Coraco-hyoideus s. Coraco-cleido-hyoi-
deus, Fiirbringer.

Cleido-hyoideus, Greil.

Unnamed, Maurer.

Genio-hyoideus, Owen.

Deep layer of superficial stratum of
ventral muscle, or Genio-hyoid, Hum-
phry.

Rectus inferior corporis (anterior part),
Jaquet.

Coraco-mandibularis s. thoracico-mandi-
bularis, Fiirbringer.

Unnamed, Maurer.

Retractor ossis hyoidei+ Coraco-hyoi-
deus, Owen.

Cervicalis profundus s. Coraco- or
Ventro-hyoid, Humphry.

Thoraco-hyoideus, Jaquet.

Coraco-hyoideus s. Coraco-cleido-thora-
cico-hyoideus, Fiirbringer.

Coraco-hyoid, Agar.

Unnamed, Maurer.

MUSCLES

Lepidosiren.
Genio-thoracicus.

Coraco-hyoideus.

Ceratodus.

Levatores arcuum branchia-
lium.

Constrictores branchiales.

Cucullaris.

Subarcularis rectus i. s.
Branchio-hyoideus.

Subarcualis rectus v.
Coraco-branchialis.
Transversi ventrales ii and iii.
Transversus ventralis ii.
Transversus ventralis iii.
Transversus ventralis iv.

Transversus ventralis v.

Sphincter oesophagi et
laryngis.

NO. 267

OF CERATODUS SOT

Gerade untere Stammmuskel (anterior
part of), Hyrtl.

Unnamed, Maurer.

Retractor ossis hyoidei and Coraco-
hyoideus, Hyrtl.

Coraco-hyoid, Agar.

Levatores arcuum branchialium, K. Fiir-
bringer.

Cranio-branchiales, Jaquet.

Levatores arcuum branchialium (first
four) and dorsal part of Dorso-
pharyngeus (fifth), Greil.

Branchiales, Jaquet.

Interbranchiales, K. Fiirbringez.

Branchiales septales, Greil.

Scapulo-branchialis, Jaquet.

Levator scapulae, Greil (1907).

Dorso-clavicularis s. Trapezius s. Clavi-
cularis, Greil (1913).

Cerato-hyoideus internus, Fiirbringer.

Grand abducteur du _ premier are
branchial, Jaquet.

Cerato-hyoideus, Greil.

Dorso-branchialis, Dorso-cerato-bran-
chialis, Greil.

Coraco-branchialis, Fiirbringer, Jaquet.

Coraco-branchialis, Cleide-branchialis,
Coraco-cleido-branchialis, Greil.

Muscle chiasmique, Jaquet.

Interbranchialis anterior.

Interbranchialis posterior. |

Interbranchialis iv. [ Greil.

Interbranchialis v. s. ventral |
part of Dorso-pharyngeus. ’

Ceratodus. Constrictor pharyngis,
Wiedersheim.

Protopterus. Constrictor pharyngis et
laryngis, Wiedersheim.

Constrictor pharyngis,

Goppert.

Lepidosiren. Constrictor isthmi fau-
cium, Hyrtl.

Bb

358 F.

Sphincter oesophagi et
laryngis.

Constrictor oesophagi.

Dilatator laryngis.

Constrictor laryngis.

H. EDGEWORTH

Lepidosiren. Constrictor pharyngis et
laryngis, Wiedersheim.
Constrictor pharyngis,
Agar.
Ceratodus, present.
Protopterus, ?
Lepidosiren, ?
Ceratodus. Not a separate muscle.
Protopterus. Dilatator, Wiedersheim.
Pharyngo-laryngeus,
Goppert.
Lepidosiren. Dilatator, Wiedersheim.
Ceratodus, present.
Protopterus. Sphincter laryngeus, Gop-
pert.
Lepidosiren, ?

LITERATURE.

’ Agar, W. E. (1907).—‘“* The Development of the Anterior Mesoderm and
Paired Fins with their Nerves, in Lepidosiren and Protepterus’’,
‘Trans. Roy. Soc. Hdin.’, vol. xlv, part ii.

Edgeworth, F. H. (1920).—‘‘ On the Development of the Hypobranchial
and Laryngeal Muscles of Amphibia ’”’, ‘ Journ. Anat.’, vol. liv.

Fiirbringer, K. (1904).—“‘ Beitriige zur Morphologie des Skeletes der

Dipnoer, nebst Bemerkungen iiber Pleuracanthiden, Holocephalen, und

99), 6

Squaliden ’’, ‘Semon’s Forschungsr.’, Bd. i, Lief. iv. .

- —— M. (1897).—‘‘ Ueberdie spino-occipitalen Nerven der Selachier und
Holocephalen ”’, ‘ Festschr. f. Gegenbaur ’, iii. ;

‘ Goéppert, E. (1904).—‘‘ Der Kehlkopf von Protopterus annectens”’,

‘ Festschr. f. Haeckel ’.

Greil, A. (1907).—‘‘ Ueber die Bildung des Kopfmesoderms bei Ceratodus
Forst.’’, ‘ Verhand. d. anat. Gesell.’

—— (1908 and 1913).—‘‘ Entwickelungsgeschichte des Kopfes und des
Blutgefasssystems von Ceratodus forsteri’’, ‘Erster und Zweiter Theil.
Semon’s zool. Forschungsr.’, Bd. i.

“’Humphry, G. M. (1872).—‘‘ The Muscles of Ceratodus”’, ‘Journ. of

Anat. and Phys.’, vol. vi.

Nerves ”’, ibid.

(1872).—‘‘ The Muscles of Lepidosiren annectens, with the Cranial

. Hyrtl, J. (1845).—‘ Lepidosiren paradoxa’, Prag.
Jaquet, M. (1897-9).—‘‘ Contribution a ’anatomie comparée des systémes
squelettaire et musculaire de Chimaera collei, Callorhynchus antarcticus,

MUSCLES OF CERATODUS 359

Spinax niger, Protopterus annectens, Ceratodus forsteri, et Axolotl’’,
‘ Archives méd. Bucarest ’, tom. ii—v.!

y Kellicott, W. E. (1905).—‘** The development of the Vascular and Respira-
tory Systems of Ceratodus”’, ‘New York Acad. Sc. Memoirs’, vol. ii,
part 4.

Maurer, F. (1912).—‘‘ Untersuchungen iiber das Muskelsystem der Wir-
beltiere ’’, ‘ Jenaische Zeitschr.’, Bd. xliv, Heft i.

yNeumeyer, L. (1904).—‘‘ Entwickelung des Darmkanales, von Lunge,
Leber, Milz, und Pankreas bei Ceratodus Forsteri’’, ‘Semon’s For-
schungsr.’, Bd. i, Lief. iv.

“ Owen, R. (1840).—“‘ Description of the Lepidosiren annectens”’, ‘ Trans.

¥

Linn. Soc.’, vol. xviii, part iii.
Pinkus, F. (1895).—‘‘ Die Hirnnerven des Protopterus annectens”’,

“Schwalbe’s Morph. Arbeiten’, Bd. iv.

» Sewertzoff, A. N. (1902).—‘‘ Zur Entwickelungsgeschichte des Ceratodus
forsteri’”’, ‘ Anat. Anz.’, xxi.

ySemon, R. (1901).—‘ Normentafel zur Entwicklungsgeschichte des
Ceratodus forsteri’, Jena.

; Wiedersheim, R. (1904).—‘‘ Ueber das Vorkommen eines Kehlkopfes bei
Ganoiden und Dipnoern sowie iiber die Phylogenie der Lunge ’’, ‘ Zool.
Jahrb.’, Suppl. vii.

NOTE.

On the development of the Quadrate and
Epihyal.

The Hyomandibula of Ceratodus was first described by
Huxley (1876), who stated that it is a small four-sided cartilage
with a short conical process at its antero-ventral angle. This
process is embedded in the dorsal and posterior part of the
hyosuspensorial ligament.

The cartilage was subsequently described by v. Wijhe (1882),
who identified it with a Hyomandibula, or possibly an Inter-
hyal.

Ridewood (1894) stated that the hyosuspensorial ligament
passes from the top of the Ceratohyal to a protuberance of

1 The only copy of this in the United Kingdom that I know of is in the
Library of the Royal College of Surgeons, Lincoln’s Inn Fields.
Bb 2

360 F. H. EDGEWORTH

the Quadrate. The Hyomandibula was always found dorsal
to the ligament. It is a rhombic cartilage applied at its
anterior edge to the ‘cranial cartilage’ (apparently the otic
process), and with a ventral process which is attached to the
upper surface of the suspensorial tubercle. The ventral
process may be a separate cartilage. In one specimen he found
a small accessory cartilage embedded in the ligament. Sewert-
zotf (1902) stated that the Quadrate is formed independently
of the chrondrocranium and fuses with it by three processes
—a ‘ palatobasal ’ uniting it with the trabecula, an “ ascending ’
uniting it with the alisphenoidal wall, an ‘ otic ’ uniting it with
the external wall of the auditory capsule. He did not describe
any pterygoid process. A Hyomandibula is present, lying just
dorsal to the ligament connecting the top of the ceratohyal
with the Quadrate.

K. Firbringer (1904), in the 17-3mm. stage, described
a cartilage which lies behind the Quadrate, and, internally,
is continuous with the anterior part of the auditory capsule.
At an earlier stage the cartilage is separate from the cranium.
This cartilage he identified as probably that described by
Huxley. It did not exactly correspond in position with the
Hyomandibula of Sewertzoff.

Krawetz (1911) stated that two cartilages are developed
behind the Quadrate, one, the ‘ Hyomandibula’ which les
in a cell-column (‘ Strang ’), passing from the Quadrate to the
auditory capsule, and a second, the Symplecticum, in close
association with the ligament passing from the ceratohyal
to the otic process of the Quadrate. These two cartilages. are
in connexion by a connective tissue, ‘Strang’. The first-
named of these cartilages is that represented by K. Firbringer
in his figures of the embryonic stage, the latter that repre-
sented by Sewertzoff. These two cartilages with the connecting
ligament represent the upper part of the hyoid bar which is
‘ zerfallen ’ into two parts.

The statement of Krawetz that the upper part of the hyoid
bar is ‘ zerfallen’ into two parts appears to be an inference
only, as he does not state that a single cartilage or structure

MUSCLES OF CERATODUS 361

had been seen in one stage and two cartilages in a succeeding
one; but the opinion he expressed is confirmed by the
observations described below.

Greil (1913), who did not refer to the observations of Sewert-
zoff, K. Firbringer, or Krawetz, stated that the Quadrate has
a ‘processus anterior (trabecularis) ’ which springs from the
trabecula and anterior portion of the parachordal, and a
Processus oticus connecting it with the otic capsule. It has
also (in a 11-5 mm. embryo) an inconstant, rudimentary, and
transitory pterygoid process at the point where the Processus
anterior springs from the trabecula.

In his figs. 421-3 (pp. 1154-6) he depicted a dorsal process
of the Quadrate, naming it ‘ Proc. asc. Qu.’ This, however,
is evidently the otic process, which passes upwards and back-
wards to the otic capsule. On p. 1390 and in fig. 587, in his
description of a 17-3 mm. embryo, he described and depicted
a ‘ Knorpelspange’ which ‘ stellt eine secundire Verbindung
des Sphenolateralknorpels mit der Pars anterior des Palato-
quadratus her’. This process is external to the N. ophthalmicus
profundus (v.), and is evidently the ascending process of
Sewertzoff.

In a 17-3mm. embryo, Greil described two or three egg-
shaped or oblong pieces of cartilage adjoining the dorsal edge
of the Ceratohyal and over which the N. hyomandibularis vu.
(which passes out of the skull above the vena capitis lateralis
and close under the otic process of the Quadrate) passes
outwards. In the figs. 543 and 544 he depicted two small
cartilages, the inner of which is ventro-external to the vena
capitis lateralis. These cartilages he called Epihyaha s.
Hyomandibularia. He also stated that a small cell-column
passes from the dorsal corner of the ceratohyal to the lateral
surface of the otic capsule and probably forms the primordium
of small cartilaginous collections.

I find that the primordia of the Quadrate and Meckel’s
cartilage are first visible in a 10-5 mm. embryo as a continuous
U-shaped column of cells on the inner side of the continuous
primordium of the masticatory muscles and Intermandibularis.

362 F. H. EDGEWORTH

It is not yet chondrified and not continuous with the trabecula
or parachordal. In an 11 mm. embryo the primordium has
separated into Quadrate and Meckel’s cartilage, which are
chondrified. The Quadrate is continuous with the junction of
the trabecula and parachordal by a cartilaginous basal process,
and also has ascending and otic processes which are continuous

TextT-FI@s. 31, 32.

Embryo 13 mm., transverse sections ; Text-fig. 32 is 424 behind
Text-fig. 31.

6) A
be Oye
o BP e

with the trabecular wall and otic capsule. The upper ends of
these processes are not chondrified until the 13 mm. stage.
I have not seen any pterygoid process of the Quadrate in any
one of several embryos of 11 and 12mm. in length. The
structures are depicted in Text-figs. 33 and 34 from a 13-5 mm.
embryo. The observations thus confirm those of Sewertzoff.
In a 13mm. embryo (Text-figs. 31, 32) the Ceratohyal passes
upwards and backwards in the hyoid segment. From the top
of the Ceratohyal a curved tract of cells, the Epihyal, passes

MUSCLES OF CERATODUS 363

upwards and slightly forwards, and then inwards below the
auditory capsule, but is not yet attached to this. The Epi-
hyal has separated in a 13-5mm. embryo into three parts
(Fext-figs. 35-7). The upper end—the primordium of the
oto-quadrate cartilage—is slightly chondrified and connected
by a narrow neck to the floor of the junction of the auditory

TExtT-FIG. 32.

capsule with the parachordal. The middle portion—the Inter-
hyal—is an aggregate of cells between the outer end of the
oto-quadrate cartilage and the hyosuspensorial ligament.
It is separated by a little gap from the former, but abuts
against the latter. The lower end, consisting of cells in a
slightly fibrillated matrix, is the primordium of the hyosuspen-
sorial ligament. It does not yet extend forwards to the
Quadrate, nor until the 17-5 mm. stage.

The outer end of the oto-quadrate cartilage joins the inner
surface of the otic process of the Quadrate in a 17-5 mm.
embryo. The Interhyal chondrifies in a 16-5mm. embryo
and subsequently enlarges (Text-fig. 38). In a 17-5 mm.

364 F. H. EDGEWORTH

TEXT-FIGS. 33-7.

Embryo 13-5 mm., transverse sections ; Text-fig. 33 is the most
anterior.

dorsal

basal pr of Quad.

TExtT-FIG. 34.

MUSCLES OF CERATODUS 365

Trext-Fia. 35.

oto-quadrate
cart.
Ceratobrs'

~oto-quadrate
cart,

Ceratobre

366 F. H. EDGEWORTH

embryo the suspensorial tubercle has developed on the back
of the otic process of the Quadrate and the anterior end of the
hyosuspensorial ligament joins it. These structures are shown

TEXxtT-FIG. 37.

°
o

ofoc® eet,
& S00,

Embryo 26 mm., sagittal section.

in Text-fig. 39 made from a model of a 27 mm. embryo. The
lateral end of the oto-quadrate cartilage turns forwards and
joins the Quadrate. The Interhyal is a <-shaped structure,
apex forwards, between the lateral edge of the auditory capsule

MUSCLES OF CERATODUS

367
and the hyosuspensorial ligament, and immediately behind the
Quadrate.

The ligament passes from the upper end of the
Ceratohyal to.the Quadrate.

The R. hyomandibularis vu.
passes outwards dorso-anterior to the oto-quadrate cartilage,

TEXxtT-FIG. 39.

|

Quadrate (cut across)

hyosuspers
Ceratohyale nterhyale
(clit across)

ip oto-quadrate
WY
i cart,
lq

Base of model chondrocranium of a 27mm. embryo.
and Ceratohyal have been cut across.

The Quadrate
is represented only on the right side.

The R. hyomandibularis vii.

dorsal to its lateral end, then between the Interhyal and the
Quadrate, and downwards external to the hyosuspensorial
ligament.

The oto-quadrate cartilage is the structure called ‘ Hyo-

mandibula ’ by K. Firbringer and Krawetz and probably the
inner of the two ‘ Epihyalia ’ depicted by Greil.

368 F. H. EDGEWORTH

The Interhyal is the cartilage called “Hyomandibula’ by
Huxley, Ridewood, v. Wijhe, and Sewertzoff, ‘Sympiecticum ’
by Krawetz, and the outer of the two ‘ Epihyalia’ of Greil.
The development, however, shows that no one of these names
is quite suitable, and the second suggestion of v. Wijhe is
adopted.

The hyosuspensorial ligament was so called by Huxley and
succeeding authors.

Comparison of the upper end of the hyoid bar with that of
the first four branchial bars suggests that the tract of cells
which separates into the above-mentioned structures is serially
homologous with the Epibranchialia, and it is accordingly
called Epihal.

The above can be summarized as follows. A precartilagimous

’ tract, the Epihyal, extends from the upper end of the Cerato-
hyal upwards and inwards beneath, though not at first attached
to, the auditory capsule. This Epihyal separates into three
portions—from above downwards the oto-quadrate cartilage,
the Interhyal, and the hyosuspensorial ligament. ‘The first-
named gains secondary attachments, its inner end to the base
of the chondrocranium and its outer end, subsequently, to
the Quadrate. The Interhyal chondrifies later than the
oto-quadrate cartilage. The hyosuspensorial ligament slowly
extends forwards to the Quadrate.

ADDITIONAL LITERATURE.

Huxley, T. H. (1876).—‘‘ On Ceratodus forsteri, with Observations on the
Classification of Fishes ’’, ‘ Proc. Zool. Soc.’

, Krawetz, L. (1911).—‘* Entwickelung des Knorpelschidels von Ceratodus ”’,
‘ Bulletin de la Société impériale des Naturalistes de Moscou’, Année
1910, Nouvelle série, tom. xxiv.

* Ridewood, W. G. (1894).—‘‘ On the Hyoid Arch of Ceratodus”’, * Proc.
Zool. Soc.’

van Wijhe, J. (1882).—‘‘ Ueber das Visceralskelet und die Nerven des
Kopfes der Ganoiden und von Ceratodus’’, ‘ Niederl. Arch. f. Zool.’,
vol, iii.

On Golgi’s Internal Apparatus in spontaneously
absorbing Tumour Cells.

By

C. Da Fano,
Reader in Histology, King’s College, University of London.

With Plates 19 and 20.

INTRODUCTORY.

THE results obtained by the study of the Golgi apparatus of
growing tumours of the mouse, rat, and guinea-pig were set
forth in a previous paper (7) in which also the literature of the
subject was summarized and discussed. The apparatus was
found to be constantly present in the healthy cells of all
tumours examined, in most of which it appeared with certain
characteristic features. These were maintained through the
successive regrafting of the same growths even when the general
histological picture of some of them had somewhat changed
and the tumour cells and their apparatus had become hyper-
trophic or had undergone partial degeneration.

It is now proposed to describe the behaviour of the apparatus
during the spontaneous absorption of some of the same tumours,
a point which could not be properly dealt with in the previous
work. Such an investigation has not, as yet, been carried out,
though it is important because it gives the opportunity for
studying the modifications of the apparatus in cells of either
epithelial or connective-tissue origin, undergoing regressive
changes after a period of active proliferation, a phenomenon
obtained with difficulty under different experimental conditions.

As shown in another paper (6), spontaneously absorbing

1 Part of the expenditure incurred for carrying out the present research
was defrayed by a grant from the Royal Society. The material was

obtained through the kindness of Dr. J. A. Murray, Director of the Labora-
tories of the Imperial Cancer Research Fund.

370 C. DA FANO

tumours of small laboratory animals are at first invaded by
elements of the lymphocytic type and then by connective-tissue
cells, the process generally ending in the complete disappearance
of the tumour cells and the production of a scar structurally
similar to that seen during the healing of certain inflammatory
lesions of the subcutaneous connective tissue. When studying
the apparatus of absorbing tumour cells, that of some of the
elements taking part in the production of the sear was observed
and will be also briefly described.

MATERIAL AND METHODS.

The following malignant new growths in various phases of
spontaneous absorption were examined :

Mouse.

Jensen. Alveolar carcinoma.

Twort. Alveolar carcinoma.

27. Papillary cystic adeno-carcinoma.

91. Haemorrhagic and cystic alveolar carcinoma.
113. Slightly haemorrhagic alveolar carcinoma.
155. Fissure forming adeno-carcinoma.

206. Alveolar carcinoma.

630. Squamous cell carcinoma.

37S. Spindle cell sarcoma.

Rat.
Rat 9. Adeno-carcinoma with dense hyaline stroma.

All tumours were investigated by the cobalt nitrate method
with the precautions suggested in the previous work. A certain
number of tumours were likewise investigated by the modifica-
tion of Veratti’s potassium antimoniate method previously
described. But this time the results obtained were not so
satisfactory as when investigating the apparatus of growing
tumours. This was probably due to the fact that absorbing
new growths often contain only a small number of healthy cells
and a great deal of detritus which becomes so intensely im-
pregnated by the potassium antimoniate method as to render

GOLGI BODIES OF TUMOUR CELLS onl

extremely difficult the interpretation of the histological pictures.
It is necessary to add here that a description of Veratti’s
original method was published by Barinetti in a paper which
had been previously overlooked (1).

No conclusive results were reached by the investigation of
the Rat 9 carcinoma either by the cobalt nitrate or by the
potassium antimoniate methods in spite of repeated attempts
made in different stages of the absorption process. Successful
impregnations were obtained only when most of the tumour
cells were in a healthy condition, and it is therefore proposed
not to include this tumour in the following description. How-
ever, the fact was worth mentioning because it has a parallel
in a previous observation regarding a transplantable lipo-
sarcoma of the guinea-pig. In that case satisfactory results
were obtained by Golgi’s arsenious acid method, but not
by others. The two observations taken together seem to
indicate that biological conditions, through which tissues may
be passing, have sometimes a decisive influence on the reaction
to which the silver impregnation is due. Ernst’s (10) recent
investigations on certain phenomena of adsorption are in
favour of this supposition.

Tur APPARATUS OF SPONTANEOUSLY ABSORBING CARCINOMATA
AND SARCOMATA OF THE MOUSE.

Jensen.—As previously described, the apparatus of the
healthy cells of this tumour has a perinuclear or juxta-nuclear
position, and generally consists of minor parts ring- or loop-
like in shape. This is also seen in the small group of unaltered
cells shown on one side of Pl. 19, fig. 1, which was drawn from
a zone of transition between a surviving nodule and an absorp-
tion area. In most of the degenerating tumour cells the
apparatus is recognized because of its characteristic aspect, the
ring- and loop-like shapes being scattered in the cytoplasm
instead of being collected in one juxta-nuclear formation.
Only in a relatively small number of cells an irregular fragmenta-
tion of the apparatus is observed, though this phenomenon
becomes more and more apparent and lastly predominates

372, Cc. DA FANO

when and where the absorption process is most advanced. In
such places, however, the degeneration of the tumour cells has
gone so far that they are hardly recognizable.

In transitional zones like that shown in PI. 19, fig. 1, multi-
nucleated protoplasmic masses are now and then observed,
apparently due to the conglutination of a variable number of
degenerating tumour cells. Next to almost each of the nuclei
included in these pseudo-giant cells, typical remnants of the
apparatus are frequently seen with arrangements similar to
that described in connexion with a polymorphous cell sarcoma
of the mouse (see Pl. xxi, fig. 29, of the previous work, 7).
A fusion of the apparatus of the single cells into a common
and centrally situated formation, as observed under different
conditions, was not met with in specimens from the Jensen
care¢ioma.

When the absorption process is nearing its end and the
connective-tissue proliferation leading to the formation of the
scar is very much advanced, the number of the carcinomatous
cells is very small, and these are recognized with difficulty,
particularly by the method used in this investigation. Never-
theless elements are now and then found which can be safely
considered as surviving tumour cells. Their identification is
chiefly based on the size and aspect of their nuclei and apparatus.
As shown by PI. 19, fig. 2, this has a much more robust appear-
ance than that of the other cells, most of which are large
wandering cells and fibroblasts mixed with a few leucocytes.
In some of the tumour cells the apparatus still possesses the
characteristic aspect to which reference has already been made ;
in others it is uncommonly large and looks like a somewhat
granular and disintegrating juxta-nuclear structure. The
apparatus of the connective-tissue cells is much smaller,
irregular in shape and structure, and often stretched in various
directions.

Twort.—The healthy cells of this tumour are provided
with an apparatus in the main smaller and more distinctly
reticular than that of Jensen’s carcmoma. The changes
observed during absorption are similar to those above deseribed,

GOLGI BODIES OF TUMOUR CELLS 373

though a simple disintegration of the apparatus into a strue-
tureless material was the prevalent feature of many specimens.
A transformation of the apparatus into a large juxta-nuclear
formation was sometimes observed (Pl. 19, fig. 3, tw.), but not
the fusion of various elements into pseudo-giant cells as in the
case of the Jensen tumour.

In advanced stages of absorption of the Twort carcinoma
accumulations of large macrophages were found. Most of them
contained only formless débris of argentophile material, but
some possessed a small and irregularly shaped apparatus
(Pl. 19, fig. 3, m.) situated close to the nucleus and on the
whole similar to that of the connective-tissue elements above
mentioned.

Tumour 27.—The apparatus of the healthy cells of this
adeno-carcinoma generally consists of short rods collected in
a bunch on that side of the nucleus which is turned towards an
existing or virtual glandular lumen. Only in some groups of
cells it appears formed of reticular portions irregular in size,
shape, and distribution. In absorbing tumours of the same
strain the typical rod-like aspect survives only in a rather small
number of cells. Most of them either show the picture occa-
sionally observed in growing tumours or convey the impression
that the rods forming the apparatus have swollen into elliptical
or roundish shapes within which a minute light space can be
detected. It has been impossible to decide whether these
spaces are really empty or contain a material which does not
take the silver. Pl. 19, fig. 4, is a good instance of this condition
which, in the specimens investigated, extended to wide areas,
easily distinguished from the unaltered tumour portions by
the pale colour of the nuclei. These showed in addition
a strong tendency to fuse into agglomerations in which the
boundaries between cell and cell could hardly be made out
even by very high magnifications.

These observations were at first found a little surprising.
Other absorbing tumours of the same strain were therefore
carefully investigated, but with results which did not
essentially differ from those already obtained. The number

NO. 267 ce

374 Cc. DA FANO

of cells provided with an apparatus of the characteristic type
described varied considerably from tumour to tumour and from
place to place of the same tumour ; but whenever the absorption
process had manifested itself, the apparatus showed in varying
degrees the changes shown in Pl. 19, fig. 4.

It is interesting to note that the same fact was even more
clearly observed in dividing tumour cells. Mitotically dividing
cells frequently occur in spontaneously absorbing tumours
even when these are much reduced in size and nearmg complete
disappearance. One of such cells is shown in a part of Pl. 19,
fig. 4, and another in the centre of Pl. 20, fig. 5. In both cases
the fragments of the apparatus (dictyosomes) to be subdivided
between the daughter cells no longer have the aspect generally
observed in growing tumours of small solid clumps of argento-
phile material, but appear as irregularly rounded or elliptical
shapes with a light central portion, as in most of the absorbing
cells at rest. Pl. 20, fig. 5, was drawn from a specimen of
a tumour in a very advanced phase of absorption, and the
apparatus of some cells show signs of the approaching com-
plete disintegration.

Tumours 91 and 206.—The apparatus of the healthy
cells of these two growths has a fine reticular structure, more
delicate in carcinoma 206 than in carcinoma 91. During
absorption the characteristic aspect of the apparatus is still
recognized for a time in the ¢ells of both of them. In the case
of tumour 91, however, the apparatus is soon reduced to a
eranular material with loss of the previous reticular arrange-
ment. ‘This is well shown in Pl. 20, fig. 6, which was drawn from
a place where the various phases of this degenerative process
could be seen one next to the other. In other areas of the same
absorbing tumour, appearances almost identical with that ex-
hibited by the central portion of Pl. xxu, fig. 32, of the previous
work were sometimes noticed. In such instances the similarity
between the disintegrative phenomena affecting the apparatus
of both the tumours now considered was rather striking, though
in carcinoma 206 the gradual fragmentation of the previous
fine network was plainer. As shown by PI. 20, fig. 7, a reticula

GOLGI BODIES OF TUMOUR CELLS 375

appearance still survived in spite of the breaking up of the
whole structure into small portions and threads, only now and
then united by thinner filaments.

Tumours 113 and 155.—The apparatus of the healthy
cells of these two carcinomata is very small and therefore less
suitable for observations of the kind considered in the present
paper. Inthe cells of areas undergoing absorption the apparatus
occurred, in both tumours, in the form of short rods or small
clumps closely arranged, one next to the other, on one pole of
the nuclei, but no structural details could be made out in the
intensely impregnated fragments. In more advanced phases
of absorption these fragments were still smaller and frequently
indistinguishable from formless débris.

In conditions of this sort elements provided with a small
and well impregnated apparatus were often seen, but accurate
observations, particularly of serial sections, led to the con-
clusion that such elements were connective-tissue cells similar
to those shown in part of Pl. xxu, fig. 32, of the previous work
and in PI. 19, fig. 2, of the present one.

Tumour 630.—The regressive changes exhibited during
absorption by this squamous cell carcinoma are identical with
those previously observed in keratinizing areas of the same
tumour and need no further description. When absorption is
very advanced the tumours consist of small nodules of a more
or less completely keratinized material surrounded by a shell
of proliferated connective tissue in which a considerable number
of multinucleated giant cells are found. As shown by Pl. 20,
fig. 8, some of them are provided with remnants of a centrally
situated and reticular apparatus, some only contain a finely
granular material, the granules being sometimes arranged in
such a way as to convey the impression that they also have
arisen from the fragmentation of a perhaps formerly reticular
apparatus. The presence of many giant cells in the above-
mentioned situation is probably due to the fact that im the
absorption phase in which they were noticed, nothing survived
of the old tumour but a hard substance in many respects similar
to a foreign body. The fragmentation of their apparatus may

cc2

376 Cc. DA FANO

be attributed to the process of involution through which also
the proliferated connective-tissue cells presumably pass before
disappearing with the remains of the tumour.

Together with the giant cells, either in the same situations
or between the keratinized nodules, elements of the large
wandering cell type were frequently observed. They were
provided with a small either reticular or compact apparatus
similar to that of elements of the same kind previously men-
tioned and described also by Verson (17), though in different
pathological conditions.

37 5 .—In the absorption areas of this sarcoma of the mouse
the apparatus undergoes a simple process of granular disinte-
gration, frequently manifesting itself at a period in which the
outward aspect and structure of the tumour cells is otherwise
almost unaltered. Apart from the size of the cells, the pheno-
menon is like that observed in the foreign body giant cells
described in connexion with Pl. 20, fig. 8. As easily deduced
from a comparison between PI. 20, figs. 8 and 9, in both cases
the apparatus soon becomes transformed into an accumulation
of small argentophile granules, the arrangement of which now
and then reminds one of a pre-existing reticular structure.
This observation is not without interest, because the cells
exhibiting the changes described have in common a connective-
tissue origin. PI. 20, fig. 9, was drawn from the peripheral portion
of a nodule, the central cells of which still appeared in a healthy
condition and possessed an apparatus identical with that
shown in Pl. xvi, fig. 19, of the previous work. Only a few of
the peripherally situated cells were provided with an intact
apparatus, though remains of it could be recognized even
in elemeats showing many signs of degeneration, such as loss
of a great part of the nuclear chromatin, liquefaction, and
fusion of the cell-bodies.

GENERAL CONSIDERATIONS.

One of the-principal facts resulting from the present investiga-
tion is the relative resistance of the apparatus to altered
biological conditions even when these lead to a complete

GOLGI BODIES OF TUMOUR CELLS B77

degeneration of the cells concerned. ‘The moment, in which the
disintegration of the apparatus begins, appears to vary within
wide limits and for causes which are beyond our present means
of observation. However, changes of the apparatus are on the
whole noticed when the external aspect and structure of the
cells seem otherwise unaltered. After this initial phase a period
follows during which portions and fragments of the apparatus
continue to be recognized until the extreme degree of cell
degeneration is reached. This is in agreement with previous
observations on growing transplantable tumours (7) and with
the results obtained by other authors in different fields of
investigation. It corroborates the suggestion resulting from
Bowen’s recent work (4, 5) that the apparatus plays perhaps
an important réle in the economy of the cell, though we
must confess with him our ignorance as to the exact nature
of this role.

A second point deserving brief discussion is the different
behaviour of the apparatus during the absorption of different
tumours. In some of them, as in certain connective-tissue
cells, it soon becomes transformed into a granular material
which, though persisting during cell degeneration, no longer
possesses any definite structure; in others this terminal
disintegration is reached through stages durmg which it either
swells into peculiar shapes or breaks into fragments and pieces
which are for a time endowed with certain distinctive features.
In certain cases it even passes through a sort of hypertrophic
condition which, in the material examined, could not be more
closely studied. These phenomena are probably influenced,
if not determined, by the different structure of the apparatus
in the healthy cells of the tumours investigated, and confirm
the opinion previously expressed as to the existence of a well-
defined relation between the apparent structural type and the
mode of being of the apparatus in the living cells.

The disintegration of the apparatus during the spontaneous
absorption of certain tumours has a parallel in the observations
made by other authors and myself under different pathological
or physiological conditions. For instance, the breaking up of

378 Cc. DA FANO

the apparatus in certain growths (Jensen’s carcinoma, tumours
206 and 91) has a striking similarity with Penfield’s retisper-
sion (15, 16) and with alterations of the same kind observed
in degenerating nerve-cells by Battistessa (2, 3), Marcora (14),
Da Fano (8), and others. Some of the changes exhibited by
the apparatus in tumour 27 have an almost surprising resem-
blance with the modifications exhibited by the apparatus
in germ-cells as described by Gatenby (11) and his collaborators
Woodger (12) and Ludford (18), and more recently by Bowen
(5). The same applies to the changes noticed by Da Fano (9)
during the involution of the mammary gland at the end of
lactation. The phenomena described by the above-mentioned
authors and in the present paper were noticed in tissues so
different that no detailed comparison is possible. However,
they were worth brief mention because they seem to indicate
that the fragmentation of the apparatus, whether under the
influence of physiological or pathological stimuli, is up to
a point determined by the as yet obscure part taken by the
apparatus in various cell activities.

SUMMARY.

In continuation of previous work the behaviour of Golgi’s
apparatus during spontaneous absorption of transplantable
tumours of the mouse is described. In some of them the
apparatus soon becomes transformed into a granular almost
structureless material ; in others this terminal phase is reached
through stages which are probably determined by the charac-
teristic structure exhibited by the apparatus in the healthy
cells of the corresponding tumours. The modifications of the
apparatus during such stages have a striking resemblance with
the changes observed in different tissues under various physio-
logical conditions. The connective-tissue cells, which invade the
tumours as they are absorbed, are provided with a small,
irregular, and sometimes reticular apparatus which is identical
with that described by other authors in similar elements though
in different pathological processes. In some of these elements

GOLGI BODIES OF TUMOUR CELLS 379

the apparatus undergoes changes not altogether different from
those observed in absorbing tumour cells. In general, the
fragmentation of the apparatus begins when other cell con-
stituents are still apparently unaltered, but its fragments seem
to possess a great power of resistance to degenerative conditions.

Lonpon,
March 1923.

REFERENCES.

tee

. Barinetti, C._—‘‘ L’apparato reticolare interno e la centrosfera nelle
cellule di alcuni tessuti’’, ‘ Boll. d. soc. med.-chir. di Pavia’,
vol. xxvi, 1912.

. Battistessa, P.—‘‘ Sulle alterazioni dell’ apparato reticolare interno
delle cellule nervose dei gangli spinali in seguito ad avvelenamento
da piombo e stricnina”’, ‘ Riv. Ital. d. neuropatol. psich. ed elet-
troter.’, vol. iv, 1911-12.

3. —— ‘‘Sulle alterazioni dell’ apparato reticolare interno di Golgi nelle
cellule nervose dei gangli intervertebrali nella demenza senile”,
‘Note e riv. d. psich.’, vol. iv, 1912.

4. Bowen, R. H.—‘‘ Studies on insect spermatogenesis, II”, ‘ Journ. of
Morphol.’, vol. xxxvii, 1922.

5. —— ‘‘ Studies on insect spermatogenesis, V’’, ‘ Quart. Journ. Micr.
Sci.’, vol. Ixvi, 1922.

6. Da Fano, C.—‘‘ Zellulire Analyse der Geschwulstimmunitatsreak-

tionen”’, ‘ Ztschr. f. Immunititsf. Orig.’, Bd. v, 1910.

bo

7. —— ‘On Golgi apparatus of transplantable tumour cells’, * Scient.
Rep. Imp. Cancer Res. Fund ’, vol. vii, 1921.

8. —— ‘‘ Changes of Golgi’s apparatus in nerve cells of the spinal cord
following exposure to cold”’, ‘ Journ. Nerv. and Ment. Dis.’, vol. liii,
1921.

9. ——‘‘ On Golgi’s internal apparatus in different physiological condi-

tions of the mammary gland ”, ‘ Journ. of Physiol.’, vol. lvi, 1922.

10. Ernst, P.—‘‘ Adsorptionserscheinungen an Blutk6rperchen und Bak-
terien”’, ‘ Beitr. z. path. Anat. u. z. allg. Path.’, Bd. Ixix, 1921.

11. Gatenby, J. B.—‘‘ The cytoplasmic inclusions of the germ cells,
Part X ’’, ‘Quart. Journ. Micr. Sci.’, vol. Ixvi, 1922.

12. Gatenby, J. B., and Woodger, J. H.—‘‘ The cytoplasmic inclusions ot
the germ cells, Part IX ”’, ibid., vol. Ixv, 1921.

13. Gatenby, J. B., and Ludford, R. J.—‘‘ Dictyokinesis in germ cells, or
the distribution of the Golgi apparatus during cell division”,
‘Proc. of the Royal Soc.’, B, vol. xcii, 1921.

380 C. DA FANO

14. Marcora, F.—‘‘ Sur les altérations de l'appareil réticulaire interne des
cellules nerveuses motrices, consécutives 4 des lésions des nerfs ”’,
‘ Arch. Ital. Biol.’, vol. liii, 1910.

15. Penfield, W. G.—‘‘ Alterations of the Golgi apparatus in nerve cells ”’,
* Brain’, vol. xliii, 1920.

“The Golgi apparatus and its relationship to Holmgren’s tropho-
spongium in nerve cells. Comparison during retispersion ’’, * Anat.
Rec.’, vol. xxii, 1921.

17. Verson, S.—‘‘Contributo allo studio delle cellule giganti tubercolari
e di altri elementi cellulari normali e patologici”’, ‘ Arch. Sc. Med.’,

16.

vol. xxxii, 1908.

DESCRIPTION OF PLATES 19 and 20.

All figures have been drawn by means of the camera lucida
at the magnification indicated for each of them, and are from
spontaneously absorbing tumours treated by the cobalt nitrate
method for the demonstration of Golgi’s internal apparatus.
They are fully described in the text.

REFERENCE LETTERS.

d.iu., degenerating tumour cell; /, fibroblast; g.c., giant cell;
ps.g.c., pseudo-giant cell; &., area of keratinization; m., macrophage ;
plc., polymorphonuclear leucocyte ; tw., tumour cell; w.c., wandering cell.

Fig. 1.—Jensen’s carcinoma (387 A). Moderately progressed absorption, —
( x 1160.)

Fig. 2.—Jensen’s carcinoma (3998). Advanced absorption ; surviving
tumour elements amongst proliferated connective-tissue cells. (x 1620.)

Fig. 3.—Twort’s carcinoma (123 A) at the end of the absorption process.
One surviving tumour cell; group of macrophages. (x 1620.)

Fig. 4.—Adeno-carcinoma 27 (122) in initial phase of absorption.
( x 1000.)

Fig. 5.—Adeno-carcinoma 27 (136 c). Group of surviving cells when
the tumour was nearing complete absorption. (x 1620.)

Fig. 6.—Alveolar carcinoma 91 (153 B). Internal apparatus in different
stages of fragmentation during spontaneous absorption. (x 1620.)

Fig. 7.—Alveolar carcinoma 206 (3314). Breaking up of reticular
apparatus during absorption. (x 1620.)

Fig. 8.—From connective-tissue capsule enveloping keratinized remains
of squamous cell carcinoma 630 (126 a). Connective-tissue cells and foreign
body giant cells with fragmented apparatus. (x 800.)

Fig. 9.—From the peripheral layers of an absorbing nodule of sarcoma

37S (140 a). ( 1240.)

W. Ca----+- 22 -> Bey
ao? >

2.

Quart. Journ. Micr. Sct. Vol. 67, NS., Pl. 19

se

a

‘1
Pe T
*
we be *
.
its
te
’
‘
*

ee

Val. 67, N.S., Pl. 20

The Golgi Bodies of a Coccidian.
By

Shana D. King and J. Bronté Gatenby,
School of Zoology, Dublin University.

With Plate 21.

For many months we have been collecting material of the
protozoan fauna of the gut of Lithobius forficatus.
Our object was to study the Golgi apparatus, if such were
present, and to examine also the behaviour of the mitochondria.
We chose the form Adelea because we believed that its life-
history was well known, and that there would be no special
difficulty in identifying the stages. In this we have been
disappointed—we found that Siedlecki and other workers
have left the matter in a confused state: we have been obliged
to spend many months trying to piece together the various
parts of the life-history : at present we are not sure of any of
the stages except those described in this paper.

Mr. Dobell (in literis) has answered some of our questions
as well as he was able considering that his own paper on the
subject of Adelea was written many years ago at the beginning
of his investigations on the Protozoa. In our first material
nothing except Adelea seemed to be present ; but subsequently
we found to our regret that the Coccidian (Eimeria schu-
bergi?) stages were present side by side with those of Adelea.
Minchin, in his ‘ Introduction to the Study of the Protozoa ’,
has brought out the fact that not only has there been a con-
fusion as to the stages in the life-history of many Coceidia, but
that in the case of Adelea, Eimeria, and Barroussia, there has
been even a confusion of species described under the one name.
That this confusion is not difficult to bring about can best
be understood by trying to identify the various species and
stages ; we ourselves have been much puzzled in the endeavour

382 SHANA D. KING AND J. BRONTE GATENBY

to identify Siedlecki’s stages of macro- and micro-schizonts.
We have no wish to enter into a purely protozoological field,
but later we shall have to come to some decision ; at present
we do not find any clear evidence for the macro- and. micro-
schizonts of Siedlecki, but wish to leave the matter open.
We feel sure, however, that the stages we definitely identify
in the present paper are schizonts and gametocytes. No
sporogony stages are described.

Several special methods, which show the Golgi apparatus
of the metazoan cell, have been used by us. In our material
they demonstrated a hitherto undescribed structure in Adelea.

Our reasons for identifying this intra-cellular structure as
the Golgi apparatus are as follows :

1. Its staining and fixing reactions are identical with those
of the Golgi bodies of the metazoa.

2. It occupies an excentric juxta-nuclear position and spreads
out in the cell cytoplasm, as does the Golgi apparatus in many
metazoan cells, e.g. the egg, and the nerve-cell.

3. It consists mostly of the very characteristically shaped
crescents and beads, known, in the case of the metazoan cell,
as dictyosomes.

4. As in the metazoan cell, these protozoan dictyosomes can
be found dividing by themselves in the ground cytoplasm.

An account of the details of cur technique and of our reasons
for making the statement in paragraph 1, above, will be given
in a later paper.

SCHIZOGONY.

The merozoite and sporozoite of certain Coccidia are said
to be much alike: small nuclear differences have been described :
in Himeria schubergi the nucleus of the merozoite is
said to have a distinct karyosome, absent in that of the sporo-
zoite. On this basis the stage drawn in Pl. 21, fig. 5, would be
a merozoite—it has a karyosome; at the present stage we
prefer to call the latter a nucleolus (nv). Such small crescenti¢
coceidians occur commonly in areas of the Lithobius gut,
where the asexual multiplication occurs. They are found

€ .

GOLGI BODIES OF A COCCIDIAN 383

among the cells of the gut, intra- and inter-cellular in position,
and often in the lumen.

In such stages one can impregnate a ring-like structure,
just near the nucleus, in exactly the same position, and of the
same general appearance as the Golgi bodies of many metazoan
cells. In Pl. 21, fig. 2, is drawn freehand at a very high magni-
fication, a number of these structures we now identify as the
Golgi apparatus ; the latter consists, as in sponge, coelenterate,
and many other cells, of dictyosomes (or Golgi crescents and
bent rods) lying upon the surface of a thickened protoplasmic
zone or centrosphere. Whether or not a centrosome exists in
Adelea, there is certainly a darker (denser) zone associated
with the sickle-shaped dictyosomes.

As the merozoite (sporozoite) grows and becomes definitely
a trophozoite, the Golgi apparatus develops into an important
part of the cell: in Pl. 21, fig. 11, a, is the Golgi apparatus of
a growing coccidian; the apparatus is seen to consist of a
number of bent rod-shaped structures, and it is produced by
the growth and fragmentation of the original Golgi fragment
of the younger cell; the nucleolus, hither excentric, becomes
more centrally disposed in the nucleus.

In Pl. 21, figs. 12 and 13, the Golgi apparatus is seen to have
fragmented partly and the separate dictyosomes are irregularly
scattered, though a main aggregation is found at G, in a juxta-
nuclear position.

Such cells proceed to division ; in Pl. 21, fig. 14, is a stage
with four nuclei (two shown), and the Golgi bodies had been
gathered into four rough groups, one around each nucleus (X).
The Golgi bodies at this stage were spherical, crescentic, or
granular; the next stage is shown in PI. 21, figs. 16 and 17,
where each one of the many nuclei had near it, its own quota
of Golgi beads, crescents, and granules. In Pl. 21, fig. 16, the
apparatus consisted of closely associated crescents, in fig. 17,
of much coarser rings and crescents. Such stages are followed
by ones showing the formation of cell-walls, and separation of
the merozoites and their subsequent scattering and growth again.

There were from 20 to 30 nuclei in the last stages of division :

384 SHANA D. KING AND J. BRONTE GATENBY

no centrosomes could be identified. Among our material one
finds all stages from cells like that in Pl. 21, figs. 16, 17, to
corps en barillet stages, one of which is depicted in Pl. 21,
fig. 18. This is a group of agametes or daughter individuals
produced by division of a trophozoite, and now ready to break
away into its component parts. Hach merozoite has a nucleus
with an excentric nucleolus (nm) which is always turned away
from the Golgi apparatus (@): no exception to this rule has
been found to occur. The Golgi apparatus is either formed of
several little crescents together making a little granule, or it is
elongate.

In Pl. 21, fig. 20, is a large trophozoite with a completely
scattered Golgi apparatus (G), which may be seen to lie among
other granules (vy) whose nature we do not care to examine
in this paper. Whether this is an individual of Eimeria or
Adelea we are not sure. Many trophozoites of the Lithobius
parasites have completely scattered Golgi bodies.

Occasionally one finds schizonts in which some body like the
Golgi apparatus (Pl. 21, fig. 15, ax) lies in the centre of the cell,
and appears to be taking no part in dictyokinesis. This is rare.

9 AND o7 GAMETOCYTES.

In Pl. 21, figs. 19 and 21, are two cells which we can positively
identify as ? gametocytes—the o7 gametocyte rests upon them.
The Golgi apparatus of the $ gametocyte is much lke that of
the trophozoites already described in PI. 21, figs. 11, 12, and 13:
in some Gases it is much more scattered. In the o7 gametocyte
we never found at this association stage a juxta-nuclear and
discrete apparatus as in the merozoites in PI. 21, fig. 18. Even
in preparations where the Golgi apparatus of the ° gametocyte
was beautifully marked as clear black rmgs and crescents on
a yellowish or grey background, the o7 gametocyte was found
only to contain a few black granules (@x) generally stuck im its
periphery and of doubtful nature. While we cannot positively
identify a Golgi apparatus in the o7 gametocyte at this stage,
the small granules which are present, and which might represent

GOLGI BODIES OF A COCCIDIAN 385

the Golgi bodies, are much fewer and smaller than those found
in the ? gametocyte.

Examination of a large number of cells which we believe to
be o7 gametocytes before association has enabled us, we think,
to throw some light on this matter: in Pl. 21, figs. 1, 3, 4, 6,
and 9, are what we believe to be o7 gametocytes. In them the
apparatus is rarely juxta-nuclear, but has fragmented, and its
elements seem to have passed to the periphery and are struck
beneath, or on, the cell-wall. We have found many o7 gameto-
cytes in association, with no signs of any Golgi bodies, except
a number of these peripheral black granules. We _ believe
that in the o7 gametocyte the Golgi apparatus 1s mainly extruded
or absorbed. We doubt very much if it takes any part in
fertilization. In Pl. 21, fig. 9, 1s a o7 gametocyte prepared in the
same manner as the cell in PI. 21, fig. 20. In PI. 21, figs. 7 and 8,
are two cells, which may be intermediates between o7 and @
gametocytes—such intermediates often occur in the metazoa :
in Pl. 21, fig. 7, there is certainly an apparatus at G, the nature
of the granules at Gx is more doubtful, but in both figures there
are elements which seem to be passing to the periphery.

The o7 gametocyte often stains so darkly that it is difficult
to make out much of its structure. In Pl. 21, fig. 10, is an
example of a o7 gametocyte containing a large granule which
did not appear to be taking any part in the activity of the cell
in which it lay: there were no clear dictyosomes.

DiIscussION.

In this paper we have described what we consider to be the
Golgi apparatus of the protozoan Adelea. This Golgi apparatus
is dissolved away by the same fluids, preserved by the same
reagents, and stained by the same methods as the Golgi
apparatus of the metazoan cell.

During growth, the coccidian Golgi apparatus, like that of the
metazoan egg, spreads out through the ground cytoplasm in
the form of curved banana-shaped rods or dictyosomes.
So far as we could observe, the Golgi apparatus of Adelea

386 SHANA D. KING AND J. BRONTE GATENBY

takes no part in the formation of the yolky bodies which are
to be found in both trophizoites and gametocyte.

In the asexual multiplication phase where cell-division takes
place, the Golgi apparatus behaves as is usual in metazoan
cells: it becomes sorted out into subequal groups around
each dividing nucleus, and each ultimate daughter nucleus
has gathered near it a part of the original apparatus. We
have therefore established that a true dictyokinesis takes
place in the protozoan Adelea.

The Golgi apparatus is found in every schizont and merozoite
we have examined, and strangely enough it always lies oriented
in a special manner with reference to the excentric nucleolus
(karyosome) of the oval nucleus: the Golgi bodies le always
away from the nucleolus. The exact significance of this we do
not know: it means possibly that this nucleolus does not
contain the body which during division is responsible for
shepherding the Golgi elements into groups around the syn-
cytial nuclei (Pl. 21, figs. 16, 17). Whether a true centrosome,
either intra- or extra-nuclear is present, we are unable to say :
previous workers are mainly against the view that an extra-
nuclear centrosome occurs in Adelea. Suffice to say at present
that within, or near, the coccidian nucleus is some body with
the power of attracting the dictyosomes, as occurs in the case
of the metazoan centrosome. ;

The interesting period of conjugation of the gametocytes
and of fertilization provided us with no facts worth recording
at length. We found not a jot or tittle of evidence for the view
that the Golgi apparatus of the male takes part in the process
of fertilization ; this period is undoubtedly the most difficult
to study, and we are not at present satisfied with our material
of these stages.

So far as the senior writer is concerned, this examination of
the Golgi bodies of a coccidian has been disappointing from
the point of view of the phylogenetic origin of the Golgi
apparatus; we have found a typically metazoan Golgi
apparatus, which acts just like that in the metazoan egg, during
oogenesis, and which, as in most cases of metazoan fertilization,

GOLGI BODIES OF A COCCIDIAN 387

takes no part in the process. We have an apparatus in the
coccidian which acts normally at dictyokinesis. It is evident
that in this protozoon we have the typical apparatus already
formed and established.

The search for a primitive apparatus must be carried out
amoug forms other than the Sporozoa, as Hirschler’s and our
work amply shows.

The Golgi bodies probably arose in connexion with the ter-
minal bead of the flagellum of some primitive flagellate. The
outer layer of the bead might have been differentiated to form
a lipoid store-house or elaborator of the energy-yielding
materials necessary for the nutrition of the locomotor organ.
From its primitive position in the metazoan cell, always
associated at some time with the centrosome-centrosphere com-
plex, we cannot but believe that in the early history of the
cell the Golgi apparatus and the centrosome were evolved side
by side, or thea pparatus from the centrosphere, in some way.
This speculation can only be tested when further evidence is
produced. An important field, altogether neglected hitherto, is
opened up to protozoologists. The latter more than any of
their fellow biologists are interested in the architecture of the
organisms which they study, and they should attack the
problem with the special methods explained elsewhere (5).

SUMMARY.

1. There is a true Golgi apparatus in the Coccidia (Pl. 21,
ese It 125°16,17, 18; 21):

2. It consists of separate dictyosomes or crescentic rods,
with the power of fission as in metazoa (PI. 21, figs. 2, 12, 17).

3. During growth the excentric Golgi apparatus (Pl. 21,
fig. 5) becomes larger and tends to spread out in the cell (PI. 21,
figs. 12, 13).

4. During division of the schizont the Golgi elements are
attracted into subequal groups of dictyosomes and granules
around each nucleus, as happens in most metazoan cell-
visions.

388 SHANA D. KING AND J. BRONTE GATENBY

5. No centrosome was identified—the Golgi elements are
probably attracted by some other body in the nucleus.

6. Each daughter schizont receives a part of the Golgi
apparatus of the mother cell.

7. The peculiar nucleolus (or karyosome) of the merozoite
(corps en barillet stage) always les at one end of the
nucleus. The Golgi apparatus always takes up its position
outside at the other end (PI. 21, fig. 18).

BIBLIOGRAPHY.

=

. Dobell.—‘“‘ Life History of Adelea ovata’’, ‘Proc. Roy. Soc.’, (B),
Ixxix, p. 155, 1907.

2. Minchin.—‘ Introduction to Study of Protozoa’, Arnold, 1912.

3. Hirschler.—‘‘ Uber Plasmastrukturen in den Tunicaten-, Spongien-,
und, Protozoenzellen ’’, ‘ Anat. Anz.’, Bd. xlvii, 1914.

4. Siedlecki—‘‘ Etude cytologique et cycle évolutif de Adelea ovata”,
‘ Annal. Instit. Pasteur ’, 13, 1899.

5. Gatenby.—Bolles Lee’s ‘Microtomist’s Vademecum’, 8th edition,
p. 292 et seq.

6. King and Gatenby.

‘Nature’, March 1923.

DESCRIPTION OF PLATE 21.

LETTERING, :

a, Golgi apparatus. Gx, bodies possibly either Golgi apparatus modified
or derived from the Golgi apparatus. N,nucleus. 7, nucleolus (karyosome).
y, yolk-like bodies.

Figs. 1, 3, 4, and 10.—Male gametocytes not containing a normal Golgi
apparatus. The granules Gx, impregnate like the Golgi apparatus, but
resemble them neither in position nor morphology.

Fig. 2.—The Golgi apparatus of a number of merozoites drawn free-
hand.

Fig. 5.—Young trophozoite just after separation of merozoites.

Fig. 6.—Male gametocyte containing an excentric juxta-nuclear Golgi
apparatus, and also numbers of the granules Gx, one near the Golgi
apparatus.

Figs. 7 and 8.—Cells too large to be normal male gametocytes, but
showing what seems to be the degeneration of the Golgi apparatus.

Fig. 9.—Shows yolky bodies in male gametocyte.

Figs. 11, 12, 13.—Trophozoite (schizont) showing Golgi apparatus.

<=

al.

Ca 7 *r

APS

be
&

-

aa

a a

> =

a®

Quart. Journ. Mier Sci. Vol.67NS8. P2121

GOLGI BODIES OF A COCCIDIAN 389

Figs. 14, 16, 17.—Stages in schizogony showing dictyokinesis of Golgi
elements. In fig. 17 the nuclei are not shown, all the dark rings being
Golgi bodies.

Fig. 15.—Four-nuclear stage of schizogony showing the Golgi apparatus
(Gx) apparently abnormally situated in the centre of the cell and taking
no part in division.

Fig. 18.—‘ Corps en barillet’ stage showing the merozoites, each with
a pale nucleus in which at one end lies the nucleolus, n (karyosome) ; at
the opposite pole, outside the nuclear membrane, is the Golgi apparatus.

Figs. 19, 21.—Association of gametocytes; in the o7 the granules qx
were the enly bodies which impregnated by Golgi-apparatus methods.

Fig. 20.—Coccidian trophozoite with scattered Golgi elements, and much
‘yolk’, y; the cytocyst lies around the cell.

NO. 267 pd

Some Observations upon Spirostomum
ambiguum (Ehrenberg).

By

Ann Bishop, M.Se.,

Victoria University, Manchester.

With Plates 22 and 23 and 9 Text-figures.

CONTENTS.

1. HISTORICAL .

Io ao -»

. MATERIAL AND METHODS

Fixation

Staining ; é 5 :
Observations on Living Specimens
Feeding Methods

. OBSERVATIONS ON THE MORPHOLOGY OF SPIROSTOMUM AMBIGUUM

Contractile Vacuole

Endoplasm and Nuclei

The Micronuclei : : : ;
Abnormalities in the form of the Meganucleus

. METHODS OF CULTIVATION

. Foop CycLE

. VARIETIES OF SPIROSTOMUM AMBIGUUM .
. REPRODUCTION

A. Observations on the Growth and Reproduction of Spiro-
stomum ambiguum during cultivation .

B. Fission

C. Conjugation

. LITERATURE .
. EXPLANATION OF PLATES 22 AND 23

pd2

PAGE
392
393
393
394
394
395
395
395
397
399
401
402
406
409
411

411
416
422
431
433

392, ANN BISHOP

1. Historicau.

THE genus Spirostomum was first mentioned by Ehrenberg,
but no definition nor description was given. Its systematic
position and the question of the number and identity of the
species contained in it was a subject for discussion for many
years. Later, Dujardin (7) gave a very satisfactory description
of the genus in the following words :

‘Corps cylindrique tres-allongé et trés-flexible, souvent
tordu sur lui-méme, couvert de cils disposés suivant les stries
obliques ou en hélice de la surface; avec une bouches située
latéralement au dela du milieu, a l’extrémité d’une rangée de
cils plus forts.’

He recognized, however, only Spirostomum ambiguum
as a true species.

It is to Dr. Stein (28) that we are indebted for a comprehen-
sive and beautifully illustrated description of the genus,
together with a detailed account of the vicissitudes of nomen-
clature through which it had passed since its discovery by
Ehrenberg. Stein recognizes two species of Spirostomum,
S. ambiguum (Ehrenberg) and 8. teres (Claparede et
Lachmann) (5). Previously Perty (24) had included another
form in the genus, and to it he gave the name §. semi-
virescens. His observations, which were founded on a
single specimen, are regarded by Stein as being of too
superficial a nature to justify the creation of a new species.
He believed it to be merely a variety of Spirostomum
ambiguum.

Pénard (28) recognizes S. ambiguum and §. teres
as true species, and to these adds 8. filum. This latter,
although described by Ehrenberg as Uroleptus filum,
was placed tentatively in the genus Spirostomum by Biitschli (8)
and Claparede et Lachmann. Stem does not classify it among
his species. None of the above workers, with the exception
of Ehrenberg, had seen it personally, but has relied upon
Ehrenberg’s figures for their data. Pénard has actually seen
it, and feels quite certain that it justifies the position he has
given it as a third species of the genus Spirostomum.

Ome

‘

SPIROSTOMUM AMBIGUUM 393

92. MATERIAL AND METHODS.

The Spirostoma from which the cultures were started were
obtained from ponds in North Cheshire, principally in the
neighbourhood of Ringway and Styal near Manchester. They
were most numerous in rather deep ponds, with muddy bottoms
covered with decaying vegetable matter, and with Lemna
covering the surface. A good supply was obtained during
the drought of the summer of 1921 whilst the ponds were low
and the water fairly concentrated, but all through the autumn
and winter, though a large number of ponds were visited,
including those visited in the summer, very few specimens were
obtained. In May and June they became plentiful again
and obviously were multiplying rapidly, since many dividing
forms were collected.

Fixation.—the fixation of Spirostomum ambiguum
is very unsatisfactory because the animal possesses very highly
developed powers of contraction. Attempts to narcotize
them with chloroform, ether, carbon-dioxide, or by the action
of Epsom salts proved unsuccessful, though the narcotics were
used in minute quantities and in very dilute solutions.

Bouin’s solution and hot or cold Schaudinn’s solution are
both good fixatives for the nuclear structures. For whole
mounts hot Schaudinn’s solution gives the best results, since
the contraction of the cell is less with this fixative. When
whole mounts were required, the animals, together with as
small a quantity of culture fluid as was possible, were placed
on a slide smeared with egg albumen. The hot Schaudinn’s
solution was dropped rapidly on the animals whilst they were
extended. Contraction of the whole body invariably took
place, but by this method there was no shrinkage of the
endoplasm from the ectoplasm. This contraction of the body
was not really very disastrous, since the main outline of the
meganucleus and its relative position were easily studied in the
living animal and fixed preparations were required for detailed
study.

It was found best, whenever possible, to starve the material

394 ANN BISHOP

for a few hours prior to fixation. This had the effect of removing
all the undigested food which otherwise would have obscured
the details of nuclear structure.

Animals destined to be sectioned were fixed in bulk in
a watch-glass with warm Schaudinn’s solution or with Bouin’s
solution. The fixative was washed away and the animals
were removed to a narrow tube where they were treated with
the different percentages of alcohol. When in 50 per cent.
alcohol they were lightly tinged with borax carmine to facilitate
their orientation in the paraffin wax. They were cleared
with xylol. After they had been cleared they were trans-
ferred to a watch-glass containing xylol in which paraffin
wax was gradually dissolved. By means of a warm, fine
pipette they were transferred to pure wax, contained in a
clean porcelain dish, and left in the embedding oven for
about two minutes. The animals, together with some of the
wax, were dropped upon a glass slide which previously had
been smeared with egg albumen. They were orientated with
a warm needle before the wax solidified. The solidified wax
was then shaved into small blocks.

Staining.—Borax carmine, alum carmine, paracarmine,
Delafield’s haematoxylin, aqueous iron haematoxylin, and
Dobell’s alcoholic iron haematoxylin all proved to be useful
stains. Aqueous iron haematoxylin gave the best results for
sections, but the alcoholic modification of the stain was
generally used for whole mounts, since in aqueous solution the
animals often became detached from the slide.

Methyl green in 1 per cent. acetic was used for fixing and
staining animals not needed as permanent preparations.

The method generally used for clearing whole mounts was
to soak the dehydrated preparations in clove oil for about
twenty minutes and to wash away the oil with xylol before
mounting in Canada balsam. Hairs were used to support the
coverslips because there is a tendency for the unsupported
coverslips to crush so large an animal as Spirostomum.

Observations on Living Specimens.—For isola-
tion of an animal for repeated observations it was found best

SPIROSTOMUM AMBIGUUM 395

to use a hollowed slide, which, when it was not under observa-
tion, was kept in a moist chamber. Although hanging drops
were used at first, they were abandoned later, when it was
found that when the animal moved to the edge of the drop,
which it normally did, it rapidly disintegrated there.

To facilitate observations on living Spirostoma, Caragheen
extract was used. This slows down their movements con-
siderably ; but, since their shape becomes somewhat distorted
with the density of the medium and disintegration often
follows, it is not advisable to use it when keeping the animals
under observation for a long period. It was very useful,
however, for the study of ciliary structures.

Feeding Methods.—For following the course of ingested
material, finely powdered carmine or Indian ink in culture
solution were both used. A dilute solution of milk in culture
solution was also used, as was also finely powdered yolk of egg.

3. OBSERVATIONS ON THE MORPHOLOGY OF SPIROSTOMUM
AMBIGUUM.

As a description of the general morphology and movements
of Spirostomum ambiguum Dr. Stein’s (28) excellent
account has not been improved upon. It will be sufficient
here, before passing on to a detailed account of the
various structures, to mention that Spirostomum ambi-
guum is a large, elongated ciliate belonging to the order
Heterotricha. The peristomial groove, which terminates in
the mouth, is lateral in position, but the distance of the mouth
from the anterior end of the body may vary considerably in
different individuals. The peristomial membranellae extend
from the extreme anterior end to the mouth, around which they
curve ina spiral manner. ‘The meganucleus is long and monili-
form. The numerous small micronuclei are situated close to
the meganucleus.

Contractile Vacuole.—_Spirostomum ambiguum
is bounded externally by a relatively thin layer of ectoplasm,
on the outer side of which is the thin cuticle. The ectoplasm

396 ANN BISHOP

has none of the coarsely vacuolated structure characteristic of
the endoplasm.

The contractile vacuole hes at the posterior end of the
animal. There is a long feeding canal stretching from the
anterior end of the animal to the contractile vacuole into
which it opens. When distended with fluid the contractile
vacuole fills almost completely the posterior end of the animal,
and only a very narrow band of endoplasm lies along one side
of it, down which the food passes to the median cytopyge (see
p. 408). The relative size of the vacuole to that of the whole
body, and also its shape, varies with the variety of Spiro-
stomum ambiguum, and this point will be dealt with
later.

Transverse sections of Spirostomum ambiguum show
that the outline of the contractile vacuole and its canal is
perfectly definite. The vacuole and its canal he immediately
below the ectoplasm (see Text-figs. 1 and 2). When full the
canal protrudes far into the endoplasm, by which it is almost
completely surrounded.

When contraction of the vacuole is about to take place
the liquid passes down the feeding canal, which normally
closes behind it, and into the vacuole proper, which becomes
very much distended. Normally the voiding of the contents
to the exterior immediately follows this, but in some cases,
particularly in partially narcotized animals, and also often
in animals kept for a long time in hollow slides with a small
quantity of liquid, evacuation does not immediately take place.
The animal continues to swim about with a large, closed vacuole
at the posteriorend. When the vacuole is about to be emptied an
opening is formed at the posterior end at the base of a slight
depression ; the body-walls surrounding the vacuole contract
from before backwards, the liquid is forced out of the opening
and the vacuole disappears. This complete contraction of the
contractile vacuole gives the posterior end of the animal a
compressed appearance. ‘The new vacuole and feeding canal
are formed in exactly the same position as was occupied by the
preceding one.

SPIROSTOMUM AMBIGUUM 397

Endoplasm and Nuclei.—The endoplasm consists of
large vacuoles separated by narrow meshes of fairly fluid
protoplasm. In the endoplasm les the long moniliform mega-
nucleus. In the living animal its form can be followed quite
easily, since its denser structure and greater powers of refrac-
tion readily distinguish it from the rest of the protoplasm.
Normally it consists of a single unbranched chain, extending
in a fairly straight or slightly zigzag manner from the anterior
end to the contractile vacuole. The lobes, which vary con-

TExtT-FieGs. 1 AND 2.

siderably in size, are joined together by commissures which may
be either almost as wide as the lobes themselves or very narrow.
The lobes vary also in number. The least number I have ever
seen in any member of the large variety was ten and the greatest
number was fifty.

In some animals, the nucleus, although it was normal in
length, had an unconstricted, vermiform shape and in places
was slightly coiled. In all other respects the individuals seemed
quite normal and the position of the mouth (see part on Fission)
did not point to any very recent or immediately approaching
fission. I found, however, by isolating these dividuals and
keeping them under observation for a number of hours which
varied with the individual, that lobation eventually did take
place and that it was in this case delayed for a much longer

398 ANN BISHOP

period after fission than usual. A similar phenomenon was
observed by Johnson (18) during his study of the Stentors.

Stein (28) describes cases where the meganucleus was only
a quarter of the body length, was not lobated, and lay in the
anterior end of the body. These were, I'feel certain, also stages
in fission.

The meganucleus is surrounded by a nuclear membrane
which adheres firmly to the nucleoplasm. It is best shown
in individuals which have been fixed and stained in methyl
green and acetic, especially if the cytoplasm has been teased
out prior to fixation.

In preparations well fixed and stained with iron haema-
toxylin the internal structure of the meganucleus is plainly
visible. It consists of numerous granules, which stain deeply
with iron haematoxylin, embedded in a fairly homogeneous
matrix,

Greenwood (9) terms these deeply staining granules macro-
somes, and describes in addition to these other minute granules
which do not stain deeply with haematoxylin but do so with
borax carmine. These latter she calls microsomes. In my
preparations with borax carmine the nucleus seems to have
a finely granular appearance, but the large granules (i.e.
Greenwood’s macrosomes) do not combine with this stain.

In preparations stained with iron haematoxylin the macro-
somes are seen to be present in both the commissures and lobes
of the nucleus ; but if the commissures are very narrow they
are confined to the lobes alone. The macrosomes vary in
size from minute dots barely visible at a high magnification
to masses up to 10 » or more in breadth (PI. 22, figs. 1, 2, and 3).
Often these granules are surrounded by lightly staining areas.
Since these are not always present it is possible that they are
due to the fixative and are not to be interpreted as part of
the normal nuclear structure. The macrosomes vary greatly
in shape; they are generally round, but may be oval, pear-
shaped, or even roughly oblong. The medium-sized granules
often show a single vacuole in the centre (Pl. 22, fig. 1, vac.),
whilst invariably within the large masses one or more vacuoles

SPIROSTOMUM AMBIGUUM 399

are present (Pl. 22, fig. 2, vac.). Occasionally as many as
five vacuoles have been seen in one large macrosome, each
vacuole being separated from those adjacent to it by strands
of the darkly staiming substance of which the macrosome is
composed.

The presence of small non-vacuolated granules, and medium
or large-sized ones containing one or more vacuoles, in the
same lobe of the nucleus is quite common. Sometimes all the
granules present in the nucleus are without vacuoles, whilst
in others they are all large with many vacuoles.

Collin (6) describes similar macrosomes and microsomes in
the nuclei of Acinetaria. The microsomes he believes to be true
chromatin grains, whilst the macrosomes he regards as true
nucleoli.

Owing to the fact that I have not had time or opportunity
as yet to study these structures in detail in Spirostomum, I
do not propose to offer any speculation as to their nature.
I should, however, like to add that since there exist all degrees
of vacuolation and non-vacuolation, and since whenever large
masses with numerous vacuoles are present the actual number
of masses is small, it seems to me very probable that the large
vacuolated masses (i.e. Greenwood’s macrosomes) are formed
by a flowing together of a number of the smaller macrosomes
and a subsequent vacuolation from several centres.

Animals having large multivacuolated macrosomes in their
meganuclei do not seem to be otherwise abnormal, and show
no signs of degeneration in the cytoplasmic structures. Since
the degree of vacuolation seems to be independent also of the
degree of growth after fission it seems quite probable, as Green-
wood suggests, that it is due to diet or to some temporary
condition of the culture medium.

The Micronuclei.—The micronuclei of Spirostomum
ambiguum are minute in size and difficult to find. They
completely escaped the notice of Stem (28). Maupas (17)
was the first to discover their existence. They lie close to,
but are not attached to, the meganucleus (PI. 22, fig. 1, M.N.).
In structure they consist of a central endosome, presumably

400 ANN BISHOP

composed of chromatin, since it stains darkly with the various
haematoxylin and carmine stains and with methyl green.
This endosome seems to be homogeneous and is surrounded by
a pale area or halo, around which there appears to be a definite
membrane.

In his summary of our knowledge of the multinucleate
cilates Calkins (4) says that ‘ Balbiani, in his earlier work at
least, held that the number of micronuclei is always the same
as the macronuclei, or in beaded forms, as many as there are
segments of the macronucleus’. He goes on to say that
Maupas (18), Gruber, Bitschli, and others disproved this view.
They found that the numbers were the same in some; in
’ other cases, of which Stentor is an example, the roieuenares
outnumber the segments of the macronucleus ; whilst in other
forms, including Spirostomum ambiguum, the seg-
ments of the macronucleus outnumber the micronuclei. I can
fully endorse the statement that the micronuclei do not
correspond in number to the lobes of the meganucleus, for
I have seen individuals in which a number of lobes had no
micronuclei near to them, whilst others have two, or in some
cases three, four, or five to each lobe. The micronuclei are
found opposite to the commissures as wellas opposite to the lobes.

In the majority of individuals which I examined, however,
the number of lobes of the meganucleus is greater than the
number of micronuclei. In a number of cases there have
been nearly twice as many lobes of meganucleus as micronuclei
present. On the other hand, quite an appreciable number of
individuals have been observed in which the micronuclei were
approximately equal in number to the lobes of the meganucleus
or slightly exceeded them. In one case where the number of
lobes in the meganucleus was only ten, twenty-six micronuclei
were present.

From these observations it seems clear that, subsequent
to fission, in the change from a vermiform to moniliform type
of nucleus, there 1s no correlation between the number of con-
strictions appearing in the meganucleus and the number of
micronuclei present in the daughter Spirostomum.

SPIROSTOMUM AMBIGUUM 401

Abnormalities inthe Form of the Meganucleus.
—A number of individuals from different cultures have been
found in which the meganucleus was abnormal. These observa-
tions include individuals of both the major and minor varieties
(see below).

One member of the major yariety was found whose mega-
nucleus had a short branch, consisting of two lobes and a
commissure, given off from one of the commissures. This was
the only case of a branched meganucleus met with.

A fairly common abnormality was the division of the mega-
nucleus into two pieces; in one case three pieces of mega-
nucleus were present. Such conditions might be brought
about by the snapping of a delicate commissure. Another
method was revealed, however, whilst watching a normal
individual divide. During the contraction of the meganucleus
towards the anterior end of the animal, the posterior end of
the meganucleus was seen to come apart from the rest and to
follow as a separate small oval fragment in the wake of the
rest. Hlongation subsequently took place, and, when the
constriction of the cell occurred, the anterior daughter cou-
tained a whole daughter meganucleus, whilst the posterior
daughter contained its share of the major fragment and the
separated minor fragment. Hach of these latter gave rise to
a piece of moniliform meganucleus.

The most interesting cases of abnormalities, however, were
found in a five months’ old culture of the minor variety of
Spirostomum ambiguum. The cilia of all were normal,
but the protoplasm seemed denser and more granular than
usual, although it showed no signs of the vacuolation usually
associated with degenerate forms. The meganucleus had lost
its moniliform appearance and lay collected in masses in the
endoplasm. In some cases it took the form of three or four
rounded masses separated from one another by quite distinct
gaps. In one case the meganucleus was represented by a big
sphere in the anterior end separated by a wide space from the
remainder, which took the form of four lobes of the normal
moniliform type.

402 ANN BISHOP

In three cases the meganucleus was broken up into three
or more rounded spheres lying in the endoplasm, the hinder-
most of which had passed down the body and lay as a small
refractive ball at the extreme posterior end. Stained prepara-
tions of two of these animals showed that their meganucleus
was composed of darkly staining granules packed more closely
together than they normally are but not vacuolated. Each
sphere of meganucleus was surrounded by micronuclei. The
remaining individual was isolated on a well-slide, supplied
with a little of the original culture medium and put into a moist
chamber. After about twenty-four hours it was again observed.
The spheres of the meganucleus were in practically the same
position in the anterior part, but the posterior sphere had
disappeared. Whether it had been absorbed or had passed
out of the body I cannot say ; but its position in relation to
the cytopyge on the previous day seemed to suggest that the
latter fate had befallen it. The animal was kept two more days
without any important changes taking place. At the end of
that time it died.

The culture in which these cases were found was an old
leaf one. The animals in it were very few and no ease of
division was observed while it was under observation. Frem
the lack of food vacuoles in the animals it was obvious that the
culture was in an impoverished state, and these abnormalities
were no doubt due to starvation.

4. MrtuHops oF CULTIVATION.

The first attempts to form cultures of Spirostomum
ambiguum were made with hay infusions. <A similar solu-
tion to Woodruff’s standard hay infusion (81) was made, the
formula being 10 grammes of chopped hay in 1 litre of tap-water,
and raised to the boiling-point for a few minutes. A culture
basin containing a quantity of this fluid was inoculated with
a few Spirostoma. In a few hours it was found that all the
animals had died. Experiments were then made with 75 per
cent., 50 per cent., and 25 per cent. dilutions of the fluid with

~~

ws

i »

SPIROSTOMUM AMBIGUUM 403

tap-water. In the stronger solutions the Spirostoma died
almost immediately, whilst although those in the weakest
solutions lingered for a few days, they showed no signs of
multiplying and ultimately locomotion was suspended and
disintegration followed.

During his work upon Spirostomum teres, Maupas (19)
fed the material upon a solution of flour in water, which he
added to their own pond water. In order to find out whether
& similar medium would suit Spirostomum ambiguum,
0-150 grm. of flour was added to 100 ¢.c. of tap-water and
boiled for ten minutes, these being the amounts used success-
fully by Calkins (4) for Uroleptus mobilis. Varying
volumes of this solution were added to culture dishes and
to test-tubes which contained Spirostoma together with
pond-water and a little débris. No marked success followed
this experiment. In two cases the test-tube cultures lived for
some days but no division was seen. Mixtures of flour, hay,
and pond-water in various proportions, and solutions of Lemco
and Vitmar were all in turn tried without any success.

Pond-water, together with the slimy, decaying leaves from
the bottom of ponds, was boiled for about ten minutes in order
to free it from any organisms which might be present. The
boiled leaves were placed into test-tubes with about 10 ¢.c.
of the water in which they had been boiled. The tubes were
filled up with pond-water boiled to free it from any organisms.
At first approximately 2¢.c. of Woodruff’s standard hay
infusion was added. This hay infusion was cooled and then
allowed to stand open to the air for twenty-four hours before
it was added to the cultures. By this means the hay infusion
was inoculated with a plentiful supply of bacteria. That it is
necessary to use newly made infusions is shown by the work of
Peters (25), who showed that the maximum development of
bacteria in a hay infusion is reached in approximately the first
three days. Further, Hargitt and Fray (10) have isolated from
old hay infusions many kinds of bacteria which are toxic to
Paramoecium. It is not improbable therefore that these old
infusions contain bacteria which are toxic to other ciliates as

404 ANN BISHOP

well. Later, the addition of hay infusion was found to be
unnecessary.

The tubes were allowed to stand for not less than four days
at a constant temperature of 20° C. or for a longer period at
a lower temperature in order that bacteria might multiply
and decomposition of the leaves and débris set in. Each tube
was then inoculated with a few Spirostoma.

In order to discover whether any one particular kind of leaf
is more suitable for the cultivation of these animals than
another, cultures were made of oak leaves, beech leaves,
rushes, and leaves of potomageton respectively. It was found,
however, that no individual leaf gave such good results as did
a mixture of several different kinds.

At first no multiplication of the Spirostoma took place,
although in the majority of the tubes the animals remained
alive. After a week had elapsed a smell of decay, in which
the odour of sulphuretted hydrogen could be detected easily,
issued from the tubes. The cultures darkened in colour, and
in a number of cases microscopical investigation revealed the
presence of Beggiatoa and numbers of minute green flagellates.
The Spirostoma then began to increase rapidly until, about
a month after the making of the cultures, they were present
in large numbers.

A comparison of these cultures with the ponds in which
Spirostomum ambiguum is numerous seems to show
that by this method conditions approximately similar to those
of their natural environment are obtained artificially. The
odour of sulphuretted hydrogen, so noticeable when collecting
in the type of pond in which Spirostomum ambiguum
is numerous, evidently arises from the decaying vegetable
matter. Lauterborn (14) believes its presence to be charac-
teristic of the environment necessary to what he calls a * sapro-
pelic’ fauna, and Spirostomum ambiguum figures in
his list of such sapropelic organisms.

There seems to be a direct relation between the presence
of sulphuretted hydrogen and a thriving condition of the
Spirostoma. When the amount of sulphuretted hydrogen 1s

SPIROSTOMUM AMBIGUUM 405

very great the Spirostoma die. Whether the Spirostoma are
directly dependent upon the sulphuretted hydrogen or upon
some product of protein decomposition during which sul-
phuretted hydrogen is liberated, or whether the kind of
bacteria upon which they flourish best is dependent upon it,
I am at present unable.to determine. The last supposition,
however, seems the most likely.

While these experiments were being made excellent cultures
of Amoeba proteus were being obtained by Sister Monica
Taylor’s wheat method (29). Since large ciliates were often
plentiful in such cultures it seemed probable that Spiro-
stomum ambiguum might be grown in a similar medium.
To test this assumption, two wheat grains, boiled to stop
germination, were put into a test-tube, which was then filled
up with aquarium water, previously boiled to free it from
any living organisms. The tubes were allowed to stand in
the incubator at a temperature of 20° C. from four to five
days to favour the development of a thick bacterial growth.
Spirostoma were then added. Excellent results were obtained
from this method, thicker cultures being obtained than from
the former cultures. In addition to the Spirostoma, Chilo-
monas, green flagellates, and pink bacteria often developed
in these cultures. The wheat cultures are not only easier
to prepare, but have the additional advantage that the bacterial
food-supply of the ciliata, and therefore the cultures themselves,
lasts longer than it does in the leaf-extract medium.

It is interesting to note that long, narrow test-tubes seem to
be necessary for the success of these cultures. All attempts to
cultivate Spirostomum in either leaf or wheat extract in
shallow, wide dishes were unsuccessful. Spirostomum evidently
thrives best in deep water where the surface area, and therefore
the amount of dissolved oxygen, is small. That it was not the
shallowness of the medium which killed the cultures was shown
by attempting to grow Spirostoma in wide, deep jars, when
the result, or rather the lack of result, was the same as in the
case of the wide, shallow: dishes.

NO. 267 Ee

406 ANN BISHOP

5. Foop Cycuz.

In Spirostomum ambiguum from a healthy culture
numbers of large, round food-balls can be observed throughout
the endoplasm. These food-balls may be formed of very small
green flagellates, clumped closely tagether to give them a
morula-like appearance; others are brownish in colour and
are composed of compact masses of bacteria. Some large,
pink-coloured balls are present owing to the animal having
fed upon pink bacteria present in the culture. On one or
two occasions I have seen Chilomonads in the endoplasm,
enclosed in a large fluid vacuole. That these Chilomonads
were not yet dead was shown by their undulating movements.
It is not usual, however, for Spirostomum ambiguum
to ingest anything so large.

In some cultures the food-bodies were entirely bacterial,
in others bacteria and flagellates mixed together formed the
food-balls, whilst in others the food-balls were composed
entirely of green flagellates. The animals were in a flourishing
state in all three cases.

Individual Spirostoma were examined for the presence of
fluid vacuoles surrounding the food-balls. In the case of
bacterial food-balls lying anterior to, or slightly posterior
to, the mouth, distinct fluid vacuoles could be seen encireling
them. These vacuoles were absent from balls close to the
posterior end. Vacuoles similarly encircled balls of bacteria
mixed with flagellates. In the case of flagellate balls, only
a very thin film of fluid could be detected, or in many cases the
balls seemed to be embedded in the coarsely vacuolated endo-
plasm, without the intervention of a vacuole, whatever their
position in the animal might be.

It was soon realized that a definite circulation of the food
took place in the endoplasm. An attempt to trace the cycle
was made by placing individuals in tap-water for a sufficient
length of time to allow all the food to pass out of the body,
and then isolating them in well-slides containing some culture
solution rich in green flagellates. Unfortunately all the

SPIROSTOMUM AMBIGUUM 407

animals treated in this way refused to feed. After many
unsuccessful attempts a few animals were persuaded to feed
from such a culture solution in small test-tubes.

The food is wafted down the long peristomial groove to the
cytostome by the peristomial membranellae. At the base of
the cytostome the food is gathered into a sphere, which varies
considerably in size from a ball only just visible under the
=" objective, to one-half the width of the animal’s body. The
food-ball then passes forward towards the anterior end of
the body. Its movement is, comparatively speaking, rapid. On
reaching the anterior end of the body it moves to the posterior
end in a course parallel to its former one. After regaining
a position approximately level with the cytostome its progress
becomes much slower, and with many halts it passes down the
side of the contractile vacuole, along the narrow strip of endo-
plasm found in this region, to the cytopyge. This cytopyge
appears at the base of a slight depression situated in the middle
of the posterior end. The undigested material is evacuated
slowly, one sphere at a time.

Since the movement of the food-balls at the posterior end
of the animal is so slow, there is often an accumulation
of these spheres in the neighbourhood of the contractile
vacuole.

The course followed by flagellate or bacterial food-bodies
is shown in Text-fig. 4.

In order to discover whether the course taken by food-balls
was similar in the case of substances of no food value, animals
were taken straight out of the culture and put directly into a
suspension of finely powdered carmine in culture solution. The
particles of carmine were wafted to the cytostome. The granules
of carmine were collected at the base of the cytostome exactly
as were the bacteria in the cases eited above, but the carmine
grains were packed much more loosely together. Like the
bacterial balls, the carmine balls moved from the base of the
cytostome forward, but instead of passing right to the anterior
end they passed inwards, as shown in Text-fig. 3, and then
continued backwards down the opposite side of the body to

i e.2

408 ANN BISHOP

the mouth. They were evacuated in the same manner as were
the remains of the nutritious particles. Smaller clumps of
carmine move more rapidly than larger clumps.

In cases where the animals were full of balls of nutritious
material before feeding, it could be seen that the carmine balls
moved backwards more rapidly than did nutritious balls.

A similar experiment in which Indian ink was used instead
of carmine gave identical results.

TEXT-FIGS. 3 AND 4.

The above observations differ from those of Lund (16) on
Bursaria, in that in Spirostomum the ingested material follows
a definite course through the body. When the material ingested
is of a nutritious nature, the balls travel to the anterior end
and then move backwards following a course parallel to the
meganucleus. In her work upon the food vacuoles of Car-
chesium, Greenwood (8) states that the food vacuoles travel
round the meganucleus. In Spirostomum ambiguum,
and in Paramoecium also, according to Metalnikoy (20),
substances of no nutritive value take a much shorter course,

SPIROSTOMUM AMBIGUUM 409

move much more rapidly, and are therefore expelled in a much
shorter time than nutritious ones.

A suspension of hard-boiled yolk of egg in water was fed
to some Spirostoma. The particles ingested were observed
to follow the course taken by carmine granules and Indian ink.
A similar course was followed by globules of raw milk. From
this it seems apparent that raw milk and yolk-granules are
not nutritious to Spirostomum ambiguum, the animal
evidently being unable to digest them.

6. VARIETIES OF SPIROSTOMUM AMBIGUUM.

During the observations made upon a number of different
cultures of Spirostomum ambiguum it soon became
evident that two distinct varieties were present. That these
were not stages in a developmental cycle was proved by making
pure cultures of each.

Stein (28) describes a number of varieties of which his
Pl. ii, fig. 10, and Pl. ui, fig. 8, show two chief types. These
two main varieties, corresponding to the ones present in the
cultures, are also recognized by Roux (26), and are termed
by him Spirostomum ambiguum major and Spiro-
stomum ambiguum minor. These two varieties differ
from one another in a number of important details, the most
striking of which is size.

The major variety is usually much broader in proportion
to its length than the latter. The average length of the ordinary
members of the major variety, when fixed with warm Schau-
dinn’s solution, is 800-900 », whilst that of ordinary members
of the minor variety, when treated in the same way, is 400-
500 uw. The posterior end of the minor variety is truncated,
whilst that of the major variety is rounded. The protoplasm
of the major variety 1s yellowish in colour, whilst that of the
minor variety is greyish white and the endoplasm is less
coarsely vacuolated and more granular in the latter than in
the former. In the major variety the peristomial membranellae
extend from the extreme anterior end to some point posterior
to the middle of the animal’s body. In some individuals of

410 ANN BISHOP

the major variety the mouth may be situated almost at the
extreme posterior end, about level with the middle of the con-
tractile vacuole. This variation in the position of the mouth
depends, as will be seen later, upon the degree of growth
to which the animal has attained since the last fission. In
the minor variety the peristomial membranellae extend from
the extreme anterior end backwards to the mouth, which
usually is in the anterior third of the body length.

The shape of the contractile vacuole also differs in each
form. In Spirostomum ambiguum major it is a
pear-shaped vessel, almost as broad as the animal, but never
occupying more than an eighth of the animal’s length. In
the minor variety it is much larger, generally filling, when fully
expanded, the posterior quarter or even third of the animal.
In some cases it has been observed to fill half the total length.
The shape of the contractile vacuole in the minor variety tends
towards an oblong.

The time between contractions of the vacuoles in the two
varieties varies. Thus, in the major variety the average
time obtained from repeated observations on a large number
of animals from four different cultures was every eight and
a half minutes. The maximum period between contractions
which was ever observed was ten minutes and the minimum
seven minutes. In the minor variety the average was sixteen
minutes. On two occasions thirty and thirty-four minutes
respectively elapsed between contraction of the vacuoles of
individuals of this variety. Since these two figures were
so different from the rest they were not included in the average.

The meganucleus in both varieties is similar in form and
structure, but I have never seen large, multi-vacuolated
macrosomes in the meganucleus of the small variety. Since
the, contractile vacuole is very large, the length of the mega-
nucleus in proportion to the body length in the small variety is
less than in the large variety.

The micronuclei in the small variety are similar to those
of the majority variety except that they are smaller and often
rather disc-shaped.

SPIROSTOMUM AMBIGUUM 411

The chief differences between the two varieties, therefore,
can be summarized as difference in size, in colour of proto-
plasm, in position of mouth and length of peristomial area,
and finally in the shape and relative size of the contractile
vacuole and in its periods of contraction.

ford

7. REPRODUCTION.

A. Observations on the Growth and Reproduc-
tion of Spirostomum ambiguum during Cul-
tivation. .

In many cultures of the large variety of Spirostomum
ambiguum (see p. 404) made by both the culture methods
described above, it was observed that the size of the individuals
in different cultures varied enormously. Observations made
upon animals grown in a rich wheat or leaf culture medium
and kept at a constant temperature of 16° C. in the incubator,
showed that they were larger than those grown in a similar
culture at a higher temperature. Even when the culture
was not very rich in food and was kept at a temperature of
about 16°C. the individuals were very large. That these
differences in size were not due to differences of race in the
Spirostoma was proved by reversing the conditions, when
the animals altered in size correspondingly.

Further, it was observed that individuals in the cultures
which were kept at 20° C. divided more rapidly than did those
in cultures kept at 16°C. or lower. This was confirmed by
starting two cultures similar in every way and containing the
same number of Spirostoma but keeping one at 20° C. and
the other at 16°C. The former increased more rapidly than
the latter.

This is probably the whole reason for the variations in size
found in Spirostoma ambiguum major in the various
cultures. When the cultural conditions favour rapidly repeated
divisions the individuals become smaller than the normal size
of the species. At each division their size is halved and the
intervals between successive divisions are so short that they

412 ANN BISHOP

are unable to reach normal size again before the next division
occurs. After this division they are therefore still smaller
than half the normal, and this decrease in size is progressive,
producing a culture full of small individuals.

On the other hand, in other cultures the stimulus to division
was evidently weak or in abeyance, although assimilation and
growth were in no way impaired. The Spirostoma divided,
therefore, infrequently, and grew to far beyond the average
size, sometimes attaining as much as two and a half times the
size of the ordinary individuals.

In starving cultures also, or in tap-water, they become
very small. The cytoplasm seems to decrease more rapidly
than the nucleus, since in such animals the nucleus is much
coiled together.

A curious phenomenon, observed in most cultures on certain
occasions, was the agglomeration of many Spirostoma into
balls and strings. In an undisturbed condition of such a
culture these clumps were suspended in the fluid, but they
sank to the bottom if they were agitated. This condition was
very marked at the time of conjugation in nearly all the cultures
which contained conjugants. Lebedew (15) describes a similar
massing together of individuals of Trachelocerea in the material
from which he obtained his conjugants. He believed that the
animals congregated around food-bodies. I have seldom been
able to find any food forming the nucleus of the balls occurring
in my cultures. Calkins (4) states that in Uroleptus
mobilis epidemics of conjugation ‘are invariably preceded
by a characteristic massing or agglomeration of individuals °.
By transferring these masses to other dishes containing fresh
culture medium, he invariably obtained epidemics of con-
jugation.

In many cultures Spirostoma were observed adhering to
the sides of the tubes, evidently by the mucous secretion
deseribed in this animal by Jennings (12). It is possible
gently to draw them away a little from the solid to which they
adhere without severing the mucus. I am of the opinion that
the suspended agglomerations are formed in a similar way, by

SPIROSTOMUM AMBIGUUM 413

the adhesion of many animals by a mucous secretion, and that
this condition is particularly prevalent at times of conjugation.
Possibly by its means the future conjugants first adhere one
to another before any protoplasmic connexion is established.

Some attempts were made to induce conjugation in the
large variety of Spirostomum by experimental means.

Maupas (19) produced conjugation in S. teres by sub-
jecting them to alternations of a rich diet and a period of semi-
starvation. A similar process was tried with Spirostomum
ambiguum but without any result.

In his work upon the conditions for conjugation in Para-
moecium, Hopkins (11) has produced conjugation by subjecting
the animals to a preliminary period of semi-starvation for
about two weeks and then adding food and a small percentage
of various solutions of inorganic salts. The salt solutions used
were :

0-00002 N solution of ferric chloride,
0-00025 N solution of potassium chloride,
0-001 N solution of sodium chloride,
0-0001—0-0004 N solution of calcium nitrate.

Somewhat similar experiments have been done by Zweibaum
(82). He, toc, has been able to produce conjugation, and has
found that ferric chloride gives the best results. I was unable,
however, to induce conjugation in Spirostomum by any of these
methods. Ferric chloride, in the quantities used by Hopkins,
was an excellent stimulant to division, and cultures treated in
this way gave rise to great numbers of very small Spirostoma.
which, some weeks later, presumably when the effects of the
salt had been lost, returned gradually to a normal size.

Conjugation was first observed in the large variety of
Spirostomum ambiguum on May 4, 1922. It occurred
in a wheat culture which had been kept at a constant tempera-
ture of 20° C. since March 18. On May 7 conjugating pairs
were observed in two other wheat cultures, both of which had
also been kept at the above constant temperature. One
culture dated from January 4 and the other from the end of

414 ANN BISHOP

March. All three cultures were quite normal, no experi-
ments having been performed upon them to induce con-
jugation. There was no appearance of a true epidemic such as
Mulsow (21) found in Stentor coeruleus and Stentor
polymorphus in May 1911, when he obtained some 2,000—
3,000 pairs of conjugants. A few pairs were observed each
morning in each culture for about a fortnight, and after that
the cultures again became normal.

For about a fortnight previous to conjugation being observed,
microscopic observation of a few individuals taken at random
from these cultures had shown that the protoplasm had
become dark in colour and somewhat granular in appearance.
But the protoplasm of all conjugants observed was of the normal
light colour. Whether the protoplasm darkens in colour
previous to conjugation and then grows light again when this
takes place, or whether individuals destined to conjugate
remain light, in which case we must suppose that all such
individuals had escaped my observation, I cannot say with
certainty. From the fact that in some leaf cultures (see below)
the protoplasm of all individuals observed was dark, and that
later almost all individuals in the culture conjugated, the former
suggestion seems to be the more probable.

Since all these cultures had been kept at a constant tempera-
ture, it was impossible that the sudden rise in room tempera-
tures, which took place about the above dates, could have
stimulated the Spirostoma to conjugate.

On May 24 a number of conjugating individuals were
observed in one wheat culture and in two leaf cultures which
had been kept always at the ordinary laboratory temperature,
then about 22° C. The period of conjugation lasted for about
ten days. In two of the cultures only a few pairs were found
to be conjugating, but in one leaf culture on an average five
or six pairs were removed each morning. This gave a fairly high
percentage of conjugating individuals. That the cultures were
in a good condition was shown by the fact that division took
place frequently in the non-conjugating organisms.

In his paper upon the conjugation of the Stentors Mulsow (21)

SPIROSTOMUM AMBIGUUM 415

concludes that it was unfavourable conditions in their environ-
ment which caused them to conjugate, since, when the con-
jugants were removed, all non-conjugants left in the cultures
died within a few days. That this did not apply to my cultures
of Spirostoma was shown by the fact that the non-conjugants
continued to live quite normally when left in the undisturbed
cultures after the period of conjugation had passed. The fact
that conjugation took place in both wheat and leaf cultures
was interesting, since it indicated that the raw material from
which the culture medium was made was not the factor inducing
conjugation. It is probable, however, that the physical and
chemical constitution of an extract of leaves is not so widely
different from that of an extract of wheat that it would affect
the behaviour of the organisms in this respect. This point also
seems worthy of more detailed investigation, since Baitsell (1)
found that in pedigreed cultures of Stylonichia pustulata
conjugation occurred on two occasions in animals kept in a
beef medium, whereas it never occurred in those forms kept
in hay infusion though they were identical in age to the
former. From this he concludes that conjugation is induced
by external conditions affecting the organism and that it
bears no relation, in this form at least, to a particular period
of a ‘ life-cycle ’.

Since all the cultures in which conjugation had so far taken
place had been stocked from the descendants of Spirostoma
obtained on one occasion from a pond near Styal (near Man-
chester) in July 1921, it was at first thought possible that the
length of time during which the animals had been cultivated
might be a factor influencing conjugation. It is a well-known
theory that Protozoa multiply for a long period without
conjugation, after which the rate of multiplication decreases
and a period of depression ensues in which the animals
degenerate and die unless they are stimulated to renewed
division by conjugation. This theory was first suggested by
Maupas (19).

From June 14 to June 22, however, conjugation was observed
in a leaf culture of Spirostomum ambiguum which had

416 ANN BISHOP

been collected from a pond at Hale in Cheshire only a fortnight
before. This new culture had been kept in the incubator and
had divided repeatedly. Moreover, the individuals which were
not conjugating were dividing actively during the period of
conjugation. The proportion of conjugants was greater in
this culture than in any other. The culture was started from
about ten individuals, and, since approximately forty pairs
of conjugants were removed during the third week, the con-
jugants must have been capable of repeated division imme-
diately prior to conjugation, in which case a senile condition
was impossible. That conjugation took place in a culture
where the division rate was high was found by Baitsell (1) in
Stylonichia pustulata, but, whereas this culture of
Spirostomum ambiguum continued to flourish after
the period of conjugation had passed, the non-conjugants in
the culture of Stylonichia pustulata became degenerate
and died out.

In one of two of the wheat cultures in which conjugation
was observed, numerous Colpidia and Paramoecia were present
in addition to Spirostomum ambiguum. Conjugation
was never observed among any members of the two former
species. Whatever the conditions might be inducing conjuga-
tion in Spirostomum ambiguum, they did not have
the same effect on the Paramoecia and Colpidia.

My experiments and observations have not, therefore, up
to the present time, thrown any new light upon the factors
causing conjugation. They seem to indicate that the seasonal
factor 1s an important one, in Spirostomum ambiguum
at any rate, and I must hope that further and more detailed
work will enable me to follow out such hints as I have so far
gained. An important part of such work would be the study
of this organism in its natural surroundings in its native ponds
and ditches.

B, -Fission.

It is to Stein (28) that we are indebted for the first deseription
and figures of fission in Spirostomum ambiguum. He
observed the phenomenon in four cases, three of which were

SPIROSTOMUM AMBIGUUM 417

in the small variety and one in the large variety. As he relied
upon freshly collected material this was not strange, for even
in healthy cultures individuals undergoing division often are
not numerous. This is, no doubt, due to the fact that a period
of two to three days elapses between divisions even in an
animal subjected, as far as it is possible for us to tell, to excellent
conditions. Divisions takes place during the night as well as
in the day time; for in the morning usually there are present
in the cultures a number of animals in the last stages of division,
or some which have recently divided.

As the process of division differs very little in the two varieties,
it is unnecessary to give a description of each ; it will be suffi-
cient to describe it as it occurs in the major variety and to note
any deviation which occurs in the minor variety.

In order to follow the entire process of division it was found
best to isolate individuals which showed the first signs of
coming fission and to observe them through the whole process.
Stained preparations were made at different stages.

The time taken to complete the phenomenon varies in
different individuals, but the average time is from seven to
eight and a half hours ; even then, the two daughters, though
separated, have not attained the normal form. ‘This is
especially true of the form of the meganucleus. Simpson (27)
gives one to two hours as the time required for division in
Spirostomum ambiguum; but this I take to mean the
actual division of the animal’s body into two parts irrespective
of nuclear changes.

In the major variety the mouth, in animals about to undergo
division, lies midway between the anterior and posterior ends.
Large numbers of animals belonging to the major variety have
been observed undergoing division, and in every case the
mouth of the parent was at the middle of the body length.
In his single observation on fission in this variety Stein describes
the old peristome as extending through the anterior two-thirds
of the body length, and the new peristome of the future daughter
as developing in the posterior third of the body. He further
states that after half an hour’s observation the beginning of

418 ANN BISHOP

the division of the body showed between the old and new
peristomes. In such cases division would be unequal; one
daughter would be two-thirds the length of the other. On no
occasion have I seen any such inequality in size between the
two daughters.

In the individuals of the small variety about to undergo
fission the old peristome lies in the anterior third and the
new peristome forms in the middle to posterior third of the body.

Animals about to divide are always much longer than the
average-sized individual of the same culture. This length is
reached by a gradual process of growth, principally in the
region behind the mouth.

The first indication of coming fission is cytoplasmic, not
nuclear. It consists of the formation of the peristomial mem-
branellae of the new cytostome. In the major variety the
anterior end of the new membranellae is almost level with the
posterior curve of the old. The first indication of the formation
of these daughter membranellae is a slight ridge in the posterior
half of the body, running parallel with the rows of the body
cilia. This ridge gradually becomes more pronounced, and
along it the new membranellae are formed. At first these are
very small but rapidly grow larger. The immature mem-
branellae are much shorter in proportion to their width than
are the mature ones. This gives them a somewhat leaf-like
appearance. The movements of these developing mem-
branellae, almost until the separation of the two daughters,
are very irregular, many of the individual membranellae moving
in different directions to their neighbours, which gives the
whole a ragged appearance. ‘This lack of co-ordination in
movement is very noticeable when compared with the steady,
undulating motion of the mature membranellae.

The formation of the membranellae is usually well advanced
before the nucleus shows any intimation of approaching fission.
The times given above were taken from the beginning of the
nuclear changes. The somewhat zigzag form of the meganucleus
becomes straightened, the lobation is gradually lost, and its
shape becomes vermiform. Stained preparations of this stage

SPIROSTOMUM AMBIGUUM 419

show that the granular structure of the nucleus is quite normal.
The micronuclei, too, are normal (Text-fig. 5).

The meganucleus next begins to contract and incidentally
to thicken. In some individuals contraction begins before the
lobation is completely lost, but in the majority of cases
the former is the normal method. In almost all individuals
the anterior part of the meganucleus remains in its normal

TEXT-FIGS. 5, 6, 7, 8, AND 9.

INT \177

position, and the posterior and middle parts contract towards
it in such a way that the almost completely contracted mega-
nucleus lies in the anterior end of the body (Text-fig. 6). Its
shape is still vermiform but much thickened and less than
a quarter of the length of the body.

All the sections which have been studied of animals at
this stage show darkly stained granules irregularly dispersed
over a mesh-work. These meshes do not combine with stains
so strongly as do the granules. Preparations of stages inter-
mediate between the expanded and the anteriorly contracted
stages of the meganucleus show that the micronuclei move

420 ANN BISHOP

forwards with the meganucleus, but that they become much
swollen and stain very lightly (Pl. 22, fig. 4, M.N.). At this
stage, when the meganucleus lies in the anterior end, the micro-
nuclei are almost three times their normal size. They are
surrounded by a distinct halo. Some are close to the mega-
nucleus, but others are a slight distance away from it.

The meganucleus moves backwards until it comes to lie in
the middle of the animal’s length, at the same time contracting
still further (Text-fig. 7). The movement backwards to the
middle of the body is fairly rapid. Here the meganucleus
remains for a relatively much longer period. In appearance
it is a roughly oval, dense structure. Its form is not rigid
but slowly changes, showing protuberances at the side which
cradually disappear and reappear at other places.

In many sections of animals with the meganucleus fully
contracted in the centre of the body it has not been possible
to find any micronuclei. In one or two specimens large,
pale-staining micronuclei have been seen; these evidently
have not yet undergone division. In other cases small, intensely
staining micronuclei have been found near the meganucleus.
It seems quite probable that these are the products of division,
and I hope to say more about them in a further paper. In all
further stages of division in Spirostomum ambiguum
the micronuclei are small and stain normally.

From the fact that, when the micronuclei of conjugants
are about to divide, they become swollen and similar in appear-
ance to the ones described above, it is to be concluded that
these micronuclei divide some time during the migration back-
wards of the anteriorly contracted meganucleus, or soon after
it takes up its central position. Unfortunately I have never
actually seen any division taking place in any preparations,
and conclude therefore that the division must take place very
rapidly. A similar difficulty was experienced by Johnson (18)
whilst working upon the division of Stentor.

During the time when the meganucleus is fully contracted
in the centre of the body, a slight dilation becomes visible
in the feeding canal of the contractile vacuole, on a level

SPIROSTOMUM AMBIGUUM 421

with the anterior end of the meganucleus. This is the beginning
of the daughter contractile vacuole.

A gradual elongation of the meganucleus next takes place
(Text-fig. 8). A curious phenomenon in the elongation of the
meganucleus, observable also in its contraction, is that this
process does not take place equally fast at each end of the
mass. Elongation takes place more rapidly in the posterior
part than in the anterior, so that while the posterior half of the
animal has quite a long developing meganucleus only a short
part projects into the anterior half. The elongating mega-
nucleus does not expand in a straight line but is often coiled in
its course. Both anterior and posterior ends have the shape
of a crook, which often persists even after the separation of the
two daughters.

During the elongation of the meganucleus a slight con-
striction of the body can be observed a little posterior to the
anterior cytostome. This gradually becomes more pronounced
and marks the point of the future separation of the two
daughters.

Coincident. with the development of this constriction is
the gradual enlargement of the dilation in the feeding canal
of the contractile vacuole. For some time it continues to
empty with the contents of the original vacuole, but a con-
siderable length of time before the separation of the two
daughters it becomes disconnected from the posterior part of
the canal and contracts independently of the posterior vacuole.
This separation of the anterior contractile vacuole from the
posterior one seems to take place as soon as the constriction
of the cytoplasm is sufficiently deep to allow the excretion
of the fluid through a pore in the median line of the con-
striction.

Although the meganucleus in the anterior half seems to
_ grow more rapidly immediately prior to the separation of the
two daughters than that in the posterior part, the length cf
meganucleus in each daughter at the time of separation is
still a trifle unequal (Text-fig. 9). This inequality seems to
be adjusted later. The meganucleus divides into the two

NO. 267 Ff

4929, ANN BISHOP

daughter meganuclei just before the cytological separation of
the two daughters takes place.

The newly separated daughters can be distinguished from
ordinary individuals by the fact that they are shorter; and
also the meganucleus is vermiform and not moniliform.
Further, the cytosome is always at the posterior end, and the
peristomial membranellae therefore extend practically the
whole of the animal’s length. Since growth takes place much
more rapidly behind the cytostome than in front of it, the
cytostome appears to move forwards gradually, so that in
animals about to undergo fission it is central in position.

Lobation of the meganucleus seems to take place at varying
times after the separation of the daughters. There seems to
be no correlation at all between the number of micronuclei
present and the number of lobes formed in the meganucleus.
It seems probable, as Collin suggested for the nuclei of Acine-
taria, that lobation is governed by the varying tensions of the
nuclear membrane.

C. Conjugation in the Major Variety of
Spirostomum ambiguum.

Stein (28) observed conjugation in Spirostomum ambi-
guum on July 28, 1857. Balbiani also observed it im this
species. Conjugation, as stated in Part 7, was seen by the
present worker in a number of individuals of the major variety
of this species during May and June 1922.

It was noticed that the conjugants were considerably
smaller than the ordinary individuals. Stein describes a
similar condition in his specimens. It was also noticed that the
pairs of conjugants in any one culture showed less variation
in size than non-conjugants of the same culture. The greater
amount of variation in size in the non-conjugants was probably
due largely to the fact that all stages in growth between newly
separated daughters and individuals about to divide were
present, whereas, from the fact that the cytostome was always
central in position with regard to the body-length, it was

SPIROSTOMUM AMBIGUUM 493

apparent that the conjugants were all at the final stage of
growth when division should commence. No information
throwing light on the means by which the smallness in size in
the conjugants is arrived at, was obtained. In his biometrical
study of conjugation in Paramoecium caudatum
Dr. Raymond Pearl (22) found ‘that conjugant individuals
when compared with non-conjugants were shorter and narrower
and less variable both in length and breadth ’. He also showed
that there was a high degree of correlation between the lengths
of the two members of conjugant pairs. He proved, by
numerous careful measurements, that such a high degree of
homogamie correlation was not due to the random pairing of
individuals in a ‘ homogeneous population of low variability ’.
Miss Watters (80) has obtained similar results with regard to
the relationship in size between conjugants and non-conjugants
in Blepharisma undulans. Such rough observations as
have been made during the present study of conjugation
seem to indicate that similar relationships exist between the
conjugants themselves and between the conjugants and the
non-conjugants in cultures of Spirostomum ambiguum.

Both conjugants are, as was noticed by Stein (28), attached
along the peristomial groove. This makes the taking in of
food during conjugation impossible. From the fact that
conjugants, even in the earliest stages of conjugation, are
rarely found containing ingested food, it would appear that
ingestion ceases some time prior to conjugation.

The peristomial membranellae are not absorbed during con-
jugation. Generally the anterior end of one individual of
the pair is attached to a point a little posterior to the anterior
end of the peristomial groove of the other. The attachment
ends at the cytostome. Since the cytostome is central in posi-
tion, it follows that conjugating imdividuals are attached for
half the body-length. Stein (28) in his figure of a pair of con-
jugants depicts them as being attached to one another from
the extreme anterior end of each. Quite a number of pairs
attached in this way have been met with during the present
observations, but the method of attachment in the majority

mie

424 ANN BISHOP

has been in the manner described above. When they are
attached im the manner described by Stein (28) the posterior
ends are level, since the conjugants are almost without exception
equal in size; when they are attached in the more common
manner, however, the posterior end of one individual projects
beyond that of the other. Observations of individuals just
beginning to conjugate showed that the anterior ends were the
first to become attached. Sections show that the conjugants
are jomed by a thin sheet of ectoplasm and that the endoplasm
of the individuals does not mingle.

The contractile vacuole seems to function m a normal
manner during conjugation, though the average time of its
contraction was not studied.

One of the difficulties encountered in studying conjugating
pairs 1s that newly attached individuals often become separated
in drawing them up a pipette. Such severed conjugants have
never been observed to become reattached but sooner or later
die. Maupas (19) experienced similar difficulties whilst working
upon conjugation in §. teres. Whilst working upon Para-
moecium, however, Calkins found that conjugants, if severed
before there had been any exchange of nuclear material,
would live and divide ina normal manner. Similar experiments
upon artificially severed conjugants were carried out by
Baitsell (1), but without exception the severed conjugants all
degenerated and died within the twenty-four hours following
the operation.

The time taken from the attachment of a pair of conjugants to
the time of their separation varies between sixty and seventy-two
hours. It is very difficult to be absolutely certain of the dura-
tion of conjugation in a pair, since it is practically impossible
to remove for observation a pair which are only just becoming
attached. They invariably become separated during the
removal from the culture to the depression slide.

In order to secure permanent preparations of as many stages
of conjugation as possible, the conjugants were removed from
the culture and placed into test-tubes, which contained some
of the culture solution, and fixed at different intervals. This

SPIROSTOMUM AMBIGUUM 495

unfortunately entailed a rather high death-rate, but it seemed
unavoidable. Further, the supply of material was meagre
and the technical difficulties considerable.

For an hour or two subsequent to the attachment. there
was no change in the meganuclei or micronuclei of either
conjugant.

The first big change to take place was the breaking up of the
meganucleus into isolated segments by the snapping of its com-
missures (Pls. 22 and 28, figs. 7and13). The greater number of
the segments of the meganucleus migrated towards the anterior
end of the conjugant’s body and came to lie in the area opposite
to the line of attachment. A few of the posterior segments
invariably remained in the neighbourhood of the contractile
vacuole and never migrated forwards. The fact that the mega-
nucleus became fragmented during conjugation was noticed
and figured by Stem (28). He did not deseribe it as actually
fragmenting, the pair of conjugants upon which he worked
evidently having passed this stage when he first observed them.

In some preparations made before fragmentation of the
meganucleus and in almost all the ones made afterwards,
vacuoles were observed in the substance of the meganucleus.
These vacuoles varied in size, sometimes attaining to half
the size of the lobes of the meganucleus. Sometimes only
one vacuole was present in each lobe, but in other cases three
or four were present. The vacuoles often projected, causing
the nuclear membrane to bulge outwards. In stained prepara-
tions these vacuoles were colourless, but lightly staining spheres
could be seen in the centre of some of them (PI. 22, fig. 5, vac.
and p.). Iam unable to offer any explanation of the nature
of these spheres. They were present at all stages from the
fragmentation of the meganucleus to the early stage of the
exconjugant. They were evidently a product of the degenera-
tion taking place in the meganucleus, and might result, from
the coagulation by the fixative of some fluid in the vacuole.
This suggestion unfortunately is not very plausible, since it
demands either that the spheres should be present in all the
vacuoles, which was not the case, or else that the contents

426 ANN BISHOP

of different vacuoles varies, which seems a very improbable
hypothesis.

These vacuoles, present in the meganucleus during conjuga-
tion, are not to be associated in any way with the vacuoles
inside the larger ‘ macrosomes ’ of a normal meganucleus. The
former were found in the interstices of the granular structure
and represent a vacuolation of the nuclear sap; whereas the
latter appeared within the substance of the ‘ macrosomes ’.
Although small macrosomes were always present in the mega-
nucleus of conjugants, no large nor vacuolated ones were ever
seen in the degenerating meganuclei of conjugants. It is
almost unnecessary to remark that this does not mean that the
presence of large vacuolated macrosomes can never be coin-
cident with conjugation, but merely that, in the comparatively
small number of conjugants that have been studied, they
have not been present. However, if the vacuolated-macrosome
condition is to be regarded as being caused by an unknown
factor in the culture, it may be that this same factor is not
suitable for inducing conjugation.

No further visible changes took place in the meganucleus
until after the separation of the conjugants.

In newly separated exconjugants the fragments of mega-
nucleus were present and stained as intensely as in conjugants.
These fragments contained vacuoles, as did those in the con-
jugants. In preparations made at a slightly later stage the
vacuoles had become more pronounced and now bulged out-
wards greatly. In some specimens they appeared to have
burst, for circular cavities could be seen in the fragments of
meganucleus.

Absorption of the old meganucleus took place during the
first few days subsequent to separation. Gradually the frag-
ments took up the stain less intensely, and often a clear space
could be seen in the cytoplasm surrounding them. Not all
the fragments were absorbed at the’ same time. Complete
absorption of the meganucleus did not take place until the
rudiments of the new meganucleus had attained to a con-
siderable size (PI. 22, fig. 6, Z. and A, A.).

SPIROSTOMUM AMBIGUUM 427

The first stages in the formation of the rudiments of the
meganucleus have not been seen. Preparation of the earliest
stages obtained showed two or more thin dises about the middle
of the exconjugant. Since these discs were denser than the
surrounding cytoplasm, they could be seen in living excon-
jugants. The ground substance of these disc-shaped rudiments
of the meganucleus stained very feebly, scarcely more intensely
than did the surrounding cytoplasm. They contained a number
of deeply staining granules. In the earlier stages these granules
were distributed through the interior of the disc, but later they
appeared to migrate to its periphery. Some of these granules
showed one or more vacuoles inside them and seemed to be
identical to the macrocomes of the meganucleus of normal
individuals of Spirostomum ambiguum (Pl. 22, fig. 8,
M.V.).

The normal number of these meganuclear discs present in
the exconjugants was two. Occasionally four, and in one
preparation six, were seen. Whether these large numbers
arose by division of an original two, or whether the normal
two were formed first, and later nuclei, which normally remain
as micronuclei, became converted into them, is not clear.

Although I succeeded in keeping exconjugants alive in test-
tubes until seventeen days after the separation of the con-
jugants, there were no signs of constrictions appearing in the
rudiments of the meganucleus to form the moniliform mega-
nucleus. The few individuals that remained alive so long died
after seventeen days, and the cessation of conjugation left me
without any material with which to make another attempt.

Owing to the small quantity of material at my disposal, the
smallness in size of the micronuclei, their great number, and
the difficulties of staining them whilst undergoing division,
my observations upon them are very fragmentary, a fault
which I hope to rectify in an additional paper.

The first change in the micronuclei during conjugation
occurred some time after the fusion of the conjugants and
before the severing of the commissures of the meganuclei.
This change consisted of the gradual swelling up of the majority

428 ANN BISHOP

of the micronuclei whilst still in their old position close to the
meganucleus. ‘Their staining powers decreased as their size
increased.

The increase in bulk of the chromatin sphere of the micro-
nucleus did not take place at the expense of the surrounding
halo, since this too increased in size and appeared as a wide,
clear area surrounding the swollen micronuclei (PI. 23, fig. 13).
Jt does not seem probable, therefore, that the increase in size
of the micronuclei was due to the absorption of fluid from the
surrounding halo, unless the latter obtains fresh liquid from the
surrounding cytoplasm. As they increase in size the muicro-
nuclei move a httle distance away from the meganuclei.

In a large number of the animals studied a few of the micro-
nuclei, particularly those situated towards the posterior end
of the conjugant, appeared to be unaffected by this change,
and they retained their minute size. Such micronuclei were
often near to isolated fragments of meganucleus at the posterior
end, even after the majority of the micronuclei were well
advanced in division. Their fate was not known, but since they
were never present in the exconjugant immediately after
separation, one may presume that they were subsequently
absorbed.

When the majority of the isolated segments of the mega-
nucleus migrated towards the anterior end of the conjugants
(see p. 425), the greater number of the swollen micronuclei
performed a similar migration and came to lie in the cytoplasm
between the scattered lobes of the meganucleus. A few of the
swollen micronuclei remained amongst the lobes of the mega-
nucleus at the posterior end of the body. These underwent
the same changes as did those at the anterior end.

In their new position the micronuclei continued to swell.
Since the amount of chromatin did not increase, but was
merely distributed through a greater bulk, the micronuclei
became very pale and difficult to study. When they had
attained to their greatest size the swollen spheres began to
stain unevenly as though the chromatin was becoming aggre-
gated at certain points (P]. 23, fig. 12). It then became apparent

SPIROSTOMUM AMBIGUUM 429

that the chromatin was gathering at one side of the halo
(Pl. 23, fig. 16), the rest of the micronucleus being almost colour-
less. At the opposite side of the micronucleus to the aggrega-
tion of chromatin a projection appeared, lengthening until it
touched the membrane at the edge of the halo. Threads
formed between the chromatin aggregation, which now became
broken up into definite granules, and the apex of the projec-
tion (Pl. 23, fig. 11). This I took to be the begmning of the
formation of a spindle ; but, in hig work upon Stentor, Mulsow
(21) states that such a condition may be preparatory to the
formation of the spindle or a stage in degeneration prior to
absorption of the micronuclei, both stages being remarkably
alike.

Thus it is not possible to state with certainty that all micro-
nuclei in this stage gave rise subsequently to spindles. No
stages in the formation of the second pole were discovered.

The formation of the spindle took place inside the micro-
nuclear membrane. The structure took up the entire space
and no halo was to be seen. Many of the spindles were almost
globular in shape (PI. 238, figs. 10 and 15) ; a few, however, were
more diamond-shaped (PI. 23, fig. 9). In the globular form the
apices were flattened, but in the others they were quite pointed.
Whether this difference in shape was accidental or was the
expression of different stages in the division process is not
certain, but the latter interpretation seems to me more probable.

No centrioles were ever seen. ‘The spindle-fibres were
usually quite obvious. They appeared to fuse and form groups
when approaching the apex of the spindle. In the diamond-
shaped spindle the fibres appeared to converge at one point,
but in the globular spindles the groups of fibres did not all come
together at the pole (Pl. 23, figs. 10 and 15).

Spindles were present at the same time in both conjugants,
but division did not take place simultaneously in all the
micronuclei of the conjugant, since, besides spindles at the
equatorial plate stage, micronuclei at the swollen stage prior
to division (see PI. 23, fig. 12) were present, as were also micro-
nuclei which were not so swollen or pale. Some of these

430 ANN BISHOP

very swollen micronuclei possibly were about to degenerate
and not to divide. No stages demonstrating the formation and
crossing over of the gamete nuclei nor the degeneration of
the surplus ones have been obtained.

In a preparation of an exconjugant, made soon after its
separation, a spindle was seen at the anterior end of the body.
This spindle (see Pl. 23, fig. 14) was more elongated than those
present in the conjugants. It was immediately surrounded by
denser protoplasm than that of which the rest of the body
was composed. It also was at the equatorial plate stage. A
very careful investigation of the rest of the animal did not
reveal the presence of other spindles nor of other micronuclei.
This spindle was therefore regarded as that of the zygote
nucleus, and it was concluded that the other micronuclei had
all degenerated. Subsequent stages in division of the nuclei
in the exconjugant were not seen.

No further stages of the division of the micronuclei were
discovered in the exconjugants. In two exconjugants four,
and in one preparation eight, micronuclei were discovered in
the neighbourhood of the two rudiments of the meganucleus.
In no preparations were micronuclei found attached to the edge
of the meganuclear discs as Mulsow found in exconjugants of
Stentor.

I should like to take this opportunity to thank the many
friends who have so kindly given me helpful advice and sugges-
tions. Particularly I should like to thank Professor Hickson
for the stimulating interest he has taken in the work. My
sincerest thanks are offered also to Dr. G. Lapage for the
help and encouragement he has given me throughout; to
Mr. Wadsworth for his advice particularly on many matters of
technique; and to Dr. Clifford Dobell for the assistance he
so kindly gave to a stranger. Without their help the omissions
and mistakes in this paper would have been considerably
greater.

SPIROSTOMUM AMBIGUUM 431

8. LitERATURE.

1. Baitsell, G. A. (1912).—‘‘ Experiments on the Reproduction of the
Hypotrichous Infusoria. I. Conjugation between closely related
individuals of Stylonichia pustulata’’, ‘ Journal of Experimental
Zoology ’, Philadelphia, vol. 13, no. 1.

(1914).—*‘ Experiments on the Reproduction of the Hypotrichous
Infusoria. II. A study of the so-called life-cycle in Oxytricha
fallax and Pleurotricha lanceolata ’’, ibid., vol. 16, no. 2.

8. Biitschli, O. (1889).—‘“‘ Protozoa. Part III. Infusoria’’, ‘ Bronn’s
Klassen und Ordnungen des Thier-Reichs ’, Leipzig.

4, Calkins, G. N. (1919).—‘‘ Uroleptus mobilis. I. History of the nuclei
during division and conjugation”, ‘Journal of Experimental
Zoology *‘, Philadelphia, vol. 27, no. 3.

5. Claparéde, E., and Lachmann, J. (1858).—‘‘ Etude sur les Infusoires
et les Rhizopodes ’’, ‘ Mém. Inst. Genévoise ’, 1858.

6. Collin, B. (1912)—‘‘ Etude monographique sur les Acinétiens”’,
‘ Archives de Zoologie expérimentale et générale’, tome 51.

7. Dujardin, F. (1841).—‘‘ Histoire naturelle des Zoophytes—Infusoires ”’,
‘Suite 4 Buffon’, Paris.

8. Greenwood, M. (1894).—“‘ On the constitution and mode of formation
of food vacuoles in Infusoria’’, ‘ Phil. Trans. Roy. Soc.’, London,
(B) clxxxv.

9. —— (1896).—‘“‘ On structural change in the resting nucleus of
Protozoa. Part I. The macronucleus of Carchesium polypinum ”’,
* Journal of Physiology ’, vol. 20. ‘

10. Hargitt, G. T., and Fray, W. (1917).—‘“* The growth of Paramoecium
in pure cultures of Bacteria’, ‘ Journal of Experimental Zoology ’,
Philadelphia, vol. 22, no. 2.

11. Hopkins, H. 8S. (1921).—‘‘ Condition for Conjugation in diverse races of
Paramoecium ”’, ibid., vol. 34, no. 3.

12. Jennings, H. S. (1915).—‘ Behaviour of the Lower Organisms’,
Columbia University Press.

13. Johnson, H. P. (1893).—‘‘ A contribution to the Morphology and
Biology of the Stentors ”’, ‘ Journal of Morphology ’, vol. 8.

14, Lauterborn, R. (1901).—‘‘ Die ‘Sapropelische’ Lebewelt”’, ‘ Zoologi-
scher Anzeiger’, Bd. 24.

15. Lebedew, W. (1909).—‘‘ Uber Trachelocerca phoenicopterus Cohn,
Ein marines Infusor’”’, ‘ Archiv fiir Protistenkunde ’, Bd. 13.

16. Lund, E. J.—“ Relations of Bursaria to Food ’’, ‘ Journal of Experi-
mental Zoology ’, Philadelphia, vol, 17, no. 1.

17. Maupas, E. (1879).—‘‘ Micronucleus of Stentor coeruieus and Spiro-
stomum ambiguum’’, foot-note, p. 1274, ‘Comptes rendus de
lV Académie des Sciences ’, Paris.

2.

432 ANN BISHOP

18. Maupas, E. (1885).—‘‘ Etude des Infusoires ciliés’’, ‘ Archives de
_ Zoologie expérimentale et générale’, sér. ii, vol. 1.

19. —— (1889).—‘‘Le Rajeunissement karyogamique chez les ciliés”
ibid., vol. 7.

20. Metalnikov, 8S. (1915).—‘‘Sur la digestion intracellulaire chez les
Protezoaires. La circulation des vacuoles digestives”’, ‘ Bulletin
de l'Institut Pasteur’, no. 15.

21. Mulsow, W. (1913).—‘‘ Die Conjugation von Stentor coeruleus und
Stentor polymorphus ”’, ‘ Archiv fiir Protistenkunde ’, Bd. 28.

22. Pearl, R. (1906).—‘‘ A Biometrical study of Conjugation in Para-
moecium’”’, ‘ Biometrika’, vol. 5. Abstract in ‘ Proc. of the
Royal Society ’, (B) 77.

23. Pénard, E. (1922).—‘ Etudes sur les Infusoires d’eau douce ’, Georg
et Cie, Genéve.

24. Perty, J. A. (1852).—‘ Zur Kenntniss kleinster Lebensformen ’, Bern.

25. Peters, A. W. (1907).—‘‘ Chemical studies on the cell and its medium.
Part II. Some chemico-biological relations in liquid culture media ”’,
“American Journal of Physiclogy ’, vol. 18.

26. Roux.— Fauna Infusorienne.’

27. Simpson, J. Y. (1901).—‘*‘ Observations on Binary Fission in the life-
history of Ciliata ’’, ‘ Proc. Roy. Soc. Edinburgh ’, vol. 23.

28. Stein, F. (1867).— Der Organismus der Infusionsthiere ’, Leipzig.

29. Taylor, M. (1920).—‘** Aquarium cultures for Biological Teaching ”’,
‘Nature’, April 22, 1920, no. 2634, vol. 105.

29a. -——‘‘ Notes on the collection and cultures of Amoeba proteus for
class purposes ’’, ‘ Proc. Roy. Soc. Edinburgh ’, vol. 20, part iv.

30. Watters, F. A. (1912).—‘‘ Size Relationship between Conjugants and
Non-conjugants in Blepharisma undulans’’, * Biolog. Bulletin’,
vol. 23, no. 3.

31. Woodruff, L. L. (1905).—‘‘ An experimental study of the life-history
of Hypotrichous Infusoria”’, ‘ Journal of Experimental Zoology ’
vol. 2, no. 4.

32. Zweibaum, J. (1912).—‘‘ La conjugaison et la différentiation sexuelle
chez les Infusoires. V. Les conditions nécessaires et suffisantes
pour la conjugaison du Paramoecium caudatum’”’, ‘ Archiv fiir.
Protistenkunde ’, Bd. 26.

SPIROSTOMUM AMBIGUUM 433

Booxs oF GENERAL REFERENCE.

Delage, Y., and Hérouard, E. (1896).—‘Traité de Zoologie Concréte,
I. La cellule et les Protozoaires ’, Schleicher Fréres, Paris.

Doflein, F. (1916).—‘ Lehrbuch der Protozoenkunde’, Gustav Fischer,
Jena.

Kent, W. Saville (1880).—‘ Manual of the Infusoria’, David Bogue,
London.

Minchin, E. A. (1912).—‘ An introduction to the study of the Protozoa’,
Edward Arnold, London.

9. EXPLANATION OF PLATES 22 AND 23.

Fig. 1—Part of the meganucleus of an ordinary individual. L., a lobe
of the meganucleus. C., a commissure. M.N., a micronucleus surrounded
by a halo. m., a non-vacuolated macrosome. vac., a Macrosome con-
taining a small vacuole. x 1200.

Fig. 2.—Part of the meganucleus of a normal individual. MV., large
macrosome. vac., vacuole inside the macrosome. 1200.

Fig. 3.—The terminal lobe of a similar meganucleus. x 1200.

Fig. 4.—Part of the meganucleus of an ordinary individual undergoing
fission, with the meganucleus contracted towards the anterior end of the
body. X., the anterior end of the meganucleus. M.N., swollen micro-
nuclei surrounded by halos. » 1200.

Fig. 5.—Isolated lobes of the fragmented meganucleus of a conjugant.
vac., vacuoles inside the lobes. p., pale-staining spheres appearing inside
some of these vacuoles. x 1200.

Fig. 6.—The nuclear apparatus of a young exconjugant. L., the darkly
staining fragments of the old meganucleus. A., the newly forming mega-
nuclei, as yet very pale. The presence of as many as five newly forming
meganuclei.in one exconjugant is unusual. 1200.

Fig. 7.—A preparation of a pair of conjugants showing the mega-
nucleus in a fragmented condition. 120. .

Fig. 8.—One single large sphere of the newly forming meganucleus
present in an exconjugant. MV., macrosome-like bodies. x 1200.

Fig. 9.—Pointed spindle ®f a micronucleus in a conjugant. x 1200.

Fig. 10.—Globular spindle of a micronucleus in a conjugant. 1200.

Fig. 11.—Swollen micronucleus with unevenly distributed chromatin
at one side and spindle-threads forming at the other. 1200.

Fig. 12.—Part of a conjugant. J/.N., micronuclei very much swollen.
The distribution of chromatin in the swollen micronuclei is very uneven.
M.x., a micronucleus which is less swollen. Z., isolated lobes of the old
meganucleus. 1200.

434 ANN BISHOP

Fig. 13.—Part of a conjugant. L., isolated lobes of the meganucleus.
M.N., micronuclei which have begun to increase in size but are still close
to the lobes of the meganucleus. x 1200.

Fig. 14.—The spindle of the dividing zygote nucleus in an exconjugant.
x 1200.

Fig. 15. A. and B., division spindles of micronuclei in a conjugant.
Spindle A. has flattened apices. 1200.

Fig. 16.—WM.S., a micronucleus approaching the spindle stage.

On some remarkable new Forms of Caryophyl-
laeidae from the Anglo-Egyptian Sudan, and
a Revision of the Families of the Cestodaria.
By
W. N. F. Woodland,

Wellcome Bureau of Scientific Research, London.

With Plates 24 and 25 and | Text-figure.

CONTENTS.

PAGE
INTRODUCTION . 3 . : é : 3 ‘ ‘ . 435
WENYONIA VIRILIS, GEN. ET SP. NOV. . A : : : 5 elas
WENYONIA ACUMINATA, SP. NOV. ; ; : 2 : . 444
WENYONIA MINUTA, SP. NOV. . : ‘ ‘ : : . 446
CARYOPHYLLAEUS FILIFORMIS, SP. NOV. : : : : . 447
THE GENERA AND SPECIES OF THE CARYOPHYLLAEIDAE : . 450

On THE CARYOPHYLLAEIDAE, GYROCOTYLIDAE AND AMPHILINIDAE,
AND CESTODARIA IN GENERAL : - . 460

On AN IMMATURE CARYOPHYLLAEUS SP. FROM e Aen oatnts OccI-
DENTALIS, Cuv. AND VAL., 1840 . ‘ : : ‘ 5 OT
SUMMARY - : é ; ; ‘ : : . 467
LITERATURE eee : : : : : ; , . 468
DESCRIPTION OF PLATES 24, 25. : : : ‘ 7 70

INTRODUCTION.

Up to the present time ten more or less well-defined species
of the Cestodarian family of the Caryophyllaeidae have
been described. These ten species have been named as follows :
Caryophyllaeus laticeps (syn. mutabilis, Rudolphi,
1801), Pallas (1781); C. tuba, Wagener (1854); C. fen-
nicus, G. Schneider (1902); C. syrdarjensis, Skrjabin
(1918); C. armeniacus, Cholodkovsky (1916); Mono-

436 W. N. F. WOODLAND

bothrium terebrans, Linton (1893); M. hexacotyle,
Linton (1898); Glaridacris catostomi, Cooper (1920) ;
Archigetes appendiculatus, Ratzel (1868), and A.
brachyurus, Mrazek (1908). Setting aside for the present
the question as to whether the first eight of these ten species
are rightly to be referred to the three genera named, it is to be
remarked that all these ten forms are very similar in their
general organization. Apart from minor differences in connexion
with the genitalia, more marked differences in the muscular
development and shape of the ‘ head ’, and the possession of
a ‘caudal’ appendage by Archigetes, all ten species agree m
general body-form, in bearing the sexual apertures in the last
quarter of the body-length and in therefore having the ovary
situated very near to the posterior end, and in the distance
between the median ‘isthmus’ of the ovary and the genital
openings being at most (in C. tuba) one-ninth the length
of the body (the length of the body being measured from
the hind end of the ovary to the anterior extremity in
not unduly contracted specimens) and usually much less ;
also in the presence of a very small group of vitellaria situated
posterior to the ovary.

The four new species of Caryophyllaeidae deseribed
in the present communication all differ from the ten species
above named in at least two of the features just stated; and
in three of these new species in all of these features, and since
all four species represent marked departures from the type of
Caryophyllaeus which, in the form of C. laticeps
(Pl. 24, fig. 10), has grown familiar to zoologists, they are of
more than usual interest.

My material consisted of specimens, already stained and
mounted in balsam, contained in the collection of siides which
belonged to the late Dr. A. J. Chalmers when Director of the
Wellcome Tropical Research Laboratories at Khartoum, and
kindly presented to the Wellcome Bureau by his suceessor,
Major R. G.. Archibald. The greater part of this material,
though sufficiently well preserved for all ordinary purposes,
is yet not good enough for minute histological observations,

CESTODARIA 437

despite much restaining and section-cutting on my part, and
I have therefore omitted descriptions of the finer tissue
structures ; I have also omitted to supply the details of struc-
ture of such organs as the cirrus and cirrus-sac, the uterus wall,
the ovary, testes, and vitellaria, not because it is impossible
to do so, but because I do not think that the information thus
to be gained is, at least at present, worth supplying,! in view
of the major differences separating these four new species from
all species hitherto described.

Wenyonia virilis, gen. et sp. nov. Woodland, 1923.

Of this species (PI. 24, figs. 1, 2), the most remarkable in form
of the four to be described in this paper and the type species
of the new genus, I possess altogether some twenty mature
specimens and five small immature specimens. This parasite
was found in the Nile Siluroid Synodontis schall, Bloch-
Schneider, 1801, common at Khartoum, presumably in the
intestine. Two of my specimens are much larger than the
remainder, one of the two measuring 52-5 mm. in length with
a maximum breadth of 8 mm., and the other (unmeasured and
now sectionized) being of about the same dimensions. The
other mature specimens range from 11 mm. to 16 mm. in length,
with maximum breadths of 15mm. to 1:-7mm. Apart from
the difference of size of body and differences in the lengths of
the several regions of the body relative to the length of the body
as a whole, the two large specimens are identical with the
smaller specimens, and I have no reason to believe that the
former belong to a distinct species or variety. In shape of
body Wenyonia virilis is very constant and characteristic
(Pl. 24, fig. 1). The Caryophyllaeid body is usually divided into
the three regions: (1) the ‘ Kopf’ or anterior extremity,
usually distinguished from the next region by expansion,
form, or muscular differentiation, or all three; (2) the ‘ Hals’,

' In most cases these inquiries would involve considerable additional
section-cutting of somewhat brittle material which has been flattened
and preserved in balsam for some years. Should this additional informa-
tion become necessary in the future, the material is always at hand.

NO. 267 Gg

438 W. N. F. WOODLAND

the short region intervening between the Kopf and the anterior
limit of the testes and vitellaria; and (3) the ‘ Rumpf’ or
remainder of the body containing the genitalia; but the
so-called neck region is very ill-defined in any species of
Caryophyllaeus and ceases to exist when the head is
contracted, so that in many cases the body is simply divisible
on this system into the head and the trunk regions. Since in
W. virilis a new and conspicuous post-ovarian region is de-
veloped (PI. 24, fig. 2)—a region inconspicuous in all previously
described Caryophyllaeidae—lI propose to dispense with
the old * trunk ’ region and subdivide the body of W. virilis?
into the four distinct regions which are obvious even to the
naked eye: (1) the head (including the ‘ neck’ when present) ;
(2) the testicular region, extending from the most
anterior testes to the genital openings and containing the testes
and anterior vitellaria ; (3) the dilated and therefore conspicuous
uterine region, extending from the genital apertures to
the hind end of the ovary and containing the uterus, ovary,
shell-gland, ootype, and the middle portions of the two elongated
vitellarian strands; and (4) the post-ovarian region
(Pl. 24, fig. 2, pPovr), composing the rest of the body and con-
taining the posterior more or less scattered vitellaria. The
broad ovoid uterine region is plainly visible with the naked
eye and is the region of maximum breadth.

In a specimen measuring 16mm. in length the head was
2mm. long and 1:3mm. broad at the base, the testicular
region was 2:2 mm. long and 1 mm. broad, the uterine region
3-5 mm. long and 1-5 mm. in maximum breadth, and the post-
ovarian vitellaria extended nearly to the end of the body,
i.e. for nearly 8mm. In my largest specimen (52-5 mm. long,
Pl. 24, fig. 1, a) the head was 2:1 mm. long and 1-1 mm. broad
at the base, the testicular region was 10 mm. long and 2 mm.
in maximum breadth, the uterme region 7mm. long and
2-5 mm. in maximum breadth, and the post-ovarian vitellaria
only extended about half-way down the remainder of the body,

1 This regional subdivision will, of course, apply to the body of all species
of Caryophyllaeids,

CESTODARIA 439

i.e. for about 16mm. Thus the ratios of the lengths of the
four regions to the length of the body in these two large and
small ppoormens an differed considerably, being 52°5
ae is we the head, 4% and 7% for the testicular region,
= and ?@ for the uterine region, and += and $% for the post-
ovarian dear 5 ; but it is noteworthy that the combined lengths
of the testicular and uterine regions in each of the two specimens
occupied in each case about the same fraction of the body
length, i.e. =4 and &%, both of which are roughly equivalent
to 4. The head and the tail thus appear to be the variable
regions, as might be expected (vide infra).

The head region is very variable in form. Pl. 24, fig. 2,
represents what may be regarded as its normal semi-contracted
shape, but it can be elongated so that its outline becomes
indistinguishable from that of the succeeding testicular region
(Pl. 24, fig. 3, a, b), or, on the other hand, contracted to form
a short globular mass (Pl. 24, figs. 3, d, e, 4). When semi-
contracted there is often a short space between the broad base
of the * head’ and the most anterior testes (Pl. 24, fig. 2), but
in many cases this space is absent. The surface of the head is
marked by deep longitudinal creases lined by thick cuticle
(Pl. 24, figs. 2, 3, 5, 6). These longitudinal creases vary in
number in different specimens—from 13 or 14 up to 25 or
26—and are perhaps constant in the same individual, beng
dependent upon the contraction of transverse musculature.
A transverse section through the base of the head (PI. 24, fig. 6)
shows well-marked longitudinal muscle-fibres scattered through-
out the parenchyma between the thick subcuticula and the
central muscle-free medullary area, and conspicuous transverse
fibres running from side to side : muscle-fibres running obliquely
and dorso-ventrally are not conspicuous. The excretory
system hes at the base of the subcuticula and not on the edge
of the medullary area. A longitudinal section through the
contracted head (Pl. 24, fig. 4) shows that the transverse band
often visible round and marking the base of the head (as shown
in Pl. 24, figs. 2,3, a) is due to a clustering of nuclei probably
belonging to muscle-fibres. The two main longitudinal nerves

Gg2

440 W. N. F. WOODLAND

are also seen to unite just below the extreme tip of the head
(Pl. 24, fig. 4). I could detect no ‘ faserzellenstringe ’, such as
have been described by Will (27) in C. laticeps, and by
Skrjabin (28) in C. syrdarjensis.

The testicular region is short relatively to the length of the
body as compared with the same region in C. laticeps
and other previously described species. It contains a central
core of testes (Pl. 24, fig. 2, rms) in, anda thin strand of vitellaria
(vit) on each side of, the central medullary area. The finer
branches of the vasa deferentia unite and eventually open into
a median main vas deferens (Pl. 24, fig. 7, vpEF) which forms
a stout cirrus surrounded by a large cirrus-sac (ctrs). The
main vitelline duct les on the inner side of each strand of
vitellaria.

The cirrus opening (Pl. 24, fig. 2, co) marks the boundary
between the testicular and uterine regions. Immediately
behind the male aperture lies another and somewhat smaller
opening—the vagino-uterine aperture (PI. 24, figs. 2, 7, vuo)—
which, as the name implies, serves both as an entrance to the
vagina and as an exit for the eggs from the uterus. An atrium
or depression surrounding the two apertures is absent. The
vagina (VAG) is a straight or slightly convoluted fairly wide
tube which runs in the median line on the ventral side of
the body direct from the vaginal aperture to the ovary. Open-
ing into the vagina on its dorsal side and from the right, at
a point immediately above the opening of the vagina to the
exterior, is the narrowed anterior end of the uterus (Pl. 24,
fig. 7, vop). From the point where it thus opens into the vagina
the uterus (ur) curves forward (thus rendering it impossible
for the cirrus to enter it) and bends to the left in front of the
male aperture, then turns posteriorly and crosses the vagina
dorsally to the right, and thence pursues its course posteriorly
as the dilated much convoluted canal shown in PI. 24, figs. 2
and 8. In all of my mature specimens the uterus is full of shell-
covered eggs. Posteriorly and immediately in front of the
ovary, the vagina develops a slight but constant dilatation,
the receptaculum seminis (Pl. 24, fig. 8, Rcps), then inclines to

CESTODARIA 441

the right, and just below the median isthmus of the ovary
dilates into the ootype (ooTp), a large oval chamber which
receives the openings of the oviduct (opo), common vitelline
duct (ovit), and the large shell-gland (suet). From the hind
end of the ootype arises the uterus (UT) as a slender convoluted
duct which, lying dorsal to and to the left side of the median
isthmus of the ovary, passes forward and, immediately anterior
to the isthmus, turns to the right and commences its zigzag
course anteriorly, attaining its wide lumen at a short distance
in front of the ovary. The vagina and uterus thus form a com-
plete * circuit ’ with a common opening to the exterior anteriorly,
the proximal part or vagina serving for the inlet of the sperma-
tozoa and the distal part or uterus serving for the exit of the
fertilized eggs. In no other type of Cestode do these two ducts
have a common external opening. The ovary or germarium is
bipartite and of the form shown in PI. 24, fig. 2. The two
halves are united across the middle line by a transversely
elongated receptacle—the isthmus (Pl. 24, figs. 2, 8, 10v)—
into each end of which open a number of fine ducts (opcTs)
from one-half of the ovary. In all my mature specimens the
isthmus is full of eggs. At the anterior end of the uterine region
the mass of testes, divided into two posteriorly in the testicular
region by the median vas deferens, ends as two diverging
strands, one on each side of the anterior uterus. The marginal
strands of vitellaria lie, one on each side, to the outside of the
uterus. Pl. 24, fig. 12, represents a transverse section through
the uterme region anteriorly, viewed from the hind aspect, in
which the uterus (ur, the median space containing two eggs)
is seen to be opening into the vagina (vag, the left extension of
the median space, both ducts having a common ventral opening
(the vagino-uterine pore, vuo) to the exterior. This transverse
section also shows the hind strands of the testicular mass (TEs),
the vitellaria (vir) external to these, the two main longitudinal
nerve-strands (N), the thick cuticle and subecuticula, the single
outer well-defined zone of longitudinal muscle-fibres (LMUs),
and the inconspicuous lateral dorso-ventral fibres.

The post-ovarian region—a region which is very short in

442, W. N. F. WOODLAND

C. laticeps and other previously described Caryophyllaeids—
is in W. virilis normally extremely long (PI. 24, fig. 2, povr)
and very characteristic of the species. The only part of the
genitalia it contains is an enormous posterior development and
extension of the vitellaria (vrr), the two marginal strands of
which these organs consist in the uterine and testicular regions
here merging immediately behind the ovary into a more or less
central core of scattered vesicles. The posterior extension of
the vitellaria differs in different specimens. In my largest
52-5 mm. specimen and in a specimen only measuring from
11 to 12 mm. the vitellaria only extend about half-way down the
post-ovarian region (Pl. 24, fig. 1, a), whereas in most other
specimens (Pl. 24, fig. 2) the vitellaria extend nearly to the
extreme end, but intermediate conditions doubtless exist,
though I do not possess an example. At the extreme posterior
end of this region there is developed a distinct muscular thick-
walled vesicle (Pl. 24, fig. 2, rpxB), into which open the terminal
excretory channels (Excv).

I must mention finally that in two of my specimens the curious
abbreviated condition of the post-ovarian region shown in
Pl. 24, fig. 9, was found. This condition may be due merely
to extreme local contraction, but since the concentration of the
vitellaria hardly seems to be sufficiently great on this supposi-
tion, it is possible that in these two specimens the hind end of
the animal had become detached during life and a new posterior
extremity regenerated on the stump.

A brief description of other features of W. virilis remains
to be supplied. The only portions of the nervous system my
material allowed me to make out are the two main longitudinal
marginal trunks (PI. 24, fig. 4, n) already described as joming
(JN) at the extreme anterior end of the head. I could not
detect the other longitudinal nerves which have been deseribed
by Will (27) in C. laticeps, though possibly restaiing of
my material by special methods would display them.

The excretory system (Pl. 24, fig. 13) posteriorly was well
observed in one of my specimens owing to accidental injection
with some foreign substance. It consists at the extreme

CESTODARIA 443

posterior end of some four or five main longitudinal canals
which unite into two or three to open into the excretory
bladder (TExB); more anteriorly these canals apparently
branch into ten or more (PI. 24, fig. 12, pxcv). All these main
longitudinal trunks are connected by a network of finer vessels.
In the head region (PI. 24, fig. 6), especially at the base, a large
number of peripheral canals are visible under the subcuticula,
apparently belonging to a complicated dense network (cf.
C. laticeps), and a few canals of about the same size are also
visible in the centre of the medullary region, 1.e. internal to
the zone of longitudinal muscle-fibres. I could observe no
distinction between ‘ascending’ and ‘descending’ vessels,
and my material was not sufficiently well-preserved to allow
me to detect flame-cells for certain.

The eggs (Pl. 24, fig. 11) contained in the uterus are ovoid and
thin-shelled, and measured, when mounted in balsam, 37-6—
41:3 microns in length and 20-1-25-6 microns in_ breadth.
They are thus nearly half the size of the eggs of C. laticeps.

In addition to mature specimens of W. virilis I possess
five immature specimens all about 5mm. in length (Pl. 24,
figs. 1, d, 14, 15). In the youngest of these (fig. 14) the only
signs of developing organs are a strand of thickened tissue
with expanded ends (vacut) representing the future vagina
and uterus (this latter arises as a straight tube), the genital
openings and the isthmus of the ovary, and cell-thickenings
representing the future testes (TEs). In an older stage
(fig. 15) the uterus has become convoluted, the genital openings
are distinct, the vas deferens is forming, and the rudiments
of the vitellaria are apparent. It is noteworthy in these
immature specimens that the post-ovarian region is evidently
not a mere secondary outgrowth, as is shown by the anterior
position of the rudiments of the hind ends of the vagina and
uterus. It is also noteworthy that these very immature forms
occur in the fish and not in a worm, and that they show no
signs of a ‘caudal’ appendage comparable with that of the
larva of C. laticeps. The form of the post-ovarian region
in the next two species of Wenyonia to be described clearly

444 W. N. F. WOODLAND

proves that this region cannot be homologized with a * caudal ’
appendage.

The definitions of the new genus Wenyonia—a genus
named in honour of my friend Dr. C. M. Wenyon—and the
species W. virilis will be stated below.

Wenyonia acuminata, sp. nov. Woodland, 1928.

Of this species (Pl. 24, figs. 16, 17) 1 possess four whole
specimens, fragments of two others, and a number of longitu-
dinal and transverse sections. The parasite is from the Nile
Siluroid Synodontis membranaceus, Is. Geoffr., caught
at Khartoum, and was presumably found in the intestine.
Thus W. virilis and W. acuminata are found in two
closely related species of Synodontis, and it also so happens
that these two parasites are also more closely related to each
other than to any other species of Caryophyllaeid, as will be
evident from the ensuing descriptions. My four whole speci-
mens measured 34:5 mm., 33:0 mm., 26-0 mm., and 17-5 mm.
respectively, with corresponding maximum breadths of 1-5 mm.,
1-:3mm., 1-2mm., and 1:2mm. In all the specimens the
greater part of the body was almost uniform in breadth and
not unlike that of certain Nematoda, while the head end
tapered to a fine point (hence the name of the species) and
the posterior extremity was ‘ stumpy ’and bluntly pomted. The
maximum breadth of the body occurs, if anywhere, in the
testicular region. ‘The head in most of my specimens was not
delimited from the rest of the body by any distinct base,
though in two of the specimens (Pl. 24, fig. 16) a slight circular
ridge or dark band appeared to mark a base, but this I am
inclined to think was a transitory contraction, since in both
of the specimens possessing a base the anterior testes and
vitellaria both extended well anterior to this. In the other
four whole specimens (of heads) and in my longitudinal sections
the head is quite continuous in outline with the rest of the
body (Pl. 24, fig. 17), and I shall consider of course that it ends
posteriorly at the point where the anterior testes and vitellaria
commence.

CESTODARIA 445

In my largest specimen (34-5 mm.) the head measured 3-5 mm.
in length, the testicular region 11 mm., the uterine region
13-5 mm., and the post-ovarian region 6-5 mm. In the 26 mm.
specimen the head was 4mm. long, the testicular region
measured about 10 mm. long, the uterine region 6-5 mm., and
the post-ovarian region 5-5mm. In the 17-5 mm. specimen
the head was 2-5 mm. long, the testicular region 7-5 mm., the
uterine region 5mm., and the post-ovarian region 2:5 mm.
There is thus some variation as regards the lengths occupied,
relative to the entire body-length, by the several regions.

The head region appears to be practically constant in form
and continuous with the testicular region. In transverse and
longitudinal sections (Pl. 24, figs. 18, 19) it is seen that the
subcuticula is not creased longitudinally, as in W. virilis,
both the cuticle and subcuticula being thin. The longitudinal
musculature (LMus) also is not nearly so pronounced as in
W. virilis, but, asin W. virilis, consists of a single outer
zone. ‘Transverse (TRMUS) and dorso-ventral (DvMs) muscle-
fibres are also easily seen.

In the testicular region there is nothing calling for special
comment, save that the testes end well in front of the cirrus
pore—they do not extend behind as in W. virilis—and that
the vitellaria do not extend so far forward as the testes. In
some of my preparations the vitellme duct, which lies internal
to the vitellaria on each side of the body, is very distinct.

In the uterine region it is seen (Pl. 24, figs. 17, 20) that the
uterus is not nearly so voluminous as in W. virilis, and is
much shorter, relative to the body-length. The ovary is more
attenuated than in W. virilis, and its exact limits are not
easy to ascertain; the vitellme strands are also very thin.
The detailed conformation of the ducts in the regions of the
genital openings and ovary are similar to that found in
W. virilis.

The post-ovarian region is much shorter relative to the body-
length (Pl. 24, fig. 17) than in W. virilis, and the vitellaria
largely retain their marginal strand arrangement, otherwise
it is similar to that in W. virilis.

446 W. N. F. WOODLAND

The nervous system is similar to that in W. virilis,
so far as the two main longitudinal trunks are concerned.
The excretory system is also similar but is much less extensive,
especially in the head region, when compared with that of
W. virilis. As Pl. 24, figs. 19 and 20 show, there are about
ten longitudinal excretory channels underlying the subcuticula
and two in the medulla, both at the base of the head and in the
uterine region.

The eggs (Pl. 24, fig. 21) are of the usual ovoid shape and
measure (when mounted in balsam) 34-7-36-6 microns in
length and 21-9-25-6 microns in breadth. They are distin-
guished from the smooth-surfaced eggs of W. virilis,
C. laticeps, the two species yet to be described, and, so
far as I know, all other species of Caryophyllaeidae,
by the fact that under a magnification of a 1,000 diameters
the shell is seen to be covered with minute spines (Pl. 24,
fig. 21).

The specific characters of W. acuminata will be stated
below.

Wenyonia minuta, sp. nov. Woodland, 1923.

Of this species (Pl. 24, figs. 22, 23) I possess one stained
mounted specimen only. Fortunately the specimen is well
stained, and I have been able to make out all the main features
of its structure without restaining or sectionizing. This parasite
was found in the Nile Siluroid Chrysichthys auratus,
Giinth., at Khartoum, presumably in the intestine. The
total length of the body is only 3-5 mm., and the maximum
breadth 0-3 mm. It is fully mature, the uterus being filled
with eggs. The body is distinctly more flattened than in either
of the two preceding species or in C. laticeps, the head is
distinctly more of the C. laticeps type, and, asin C. lati-
ceps, the post-ovarian region is short ; on the other hand,
as in the two preceding species, the genital apertures are
situated in the anterior half of the body and the uterus is
very long.

The head measured about 0-88 mm., the testicular region

CESTODARIA 447

about 0-7 mm., the uterine region nearly 1-5mm., and the
post-ovarian region about 0-46 mm.

Notable features concerning the genitalia are (1) the very
long uterus, extending over twice the distance covered by the
testis mass; (2) the position of the ootype (oorp) on the left
side and the initial coils of the uterus on the right side of the
ovary ; and (3) the opening of the uterus into the vagina from
the left side (Pl. 24, fig. 23, is viewed from the ventral aspect),
i.e. exactly the opposite of what occurs in the two preceding
species ;! (4) the breadth of the lateral rows of vitellaria ; (5) the
extension posteriorly of the testes to about the level of the
genital openings; and (6) the apparent anterior extension of
the ovary on each side to near the level of the genital apertures
(though of this I am not quite certain, owing to the difficulty
in my preparation of distinguishing the vitellaria from the
ovarian follicles).

The two main nerve-trunks are visible anteriorly, and the
excretory opening posteriorly.

The eggs are of the usual ovoid type and measure, when
mounted in balsam, 82-9-40-2 microns in length and 20-1-
25-6 microns in breadth.

The specific characters of W. minuta will be stated below.

Caryophyllaeus filiformis, sp. nov. Woodland, 1928.

Of this species (Pl. 25, fig. 24) I possess altogether thirty-one
entire mounted specimens, and several in transverse and longitu-
dinal sections. This parasite was found in the well-known Nile
fish, Mormyrus caschive, lL. (a Malacopterygian), at
Khartoum, presumably in the intestine. Most of my specimens
are mature, and these vary in length from 7-5 mm. to 24 mm.
One or two smaller than this (circ. 5mm.) do not appear to
have eggs in the uterus. The usual maximum breadth is 1 mm.,
but occasionally the body may locally expand to as much as

1 T at first concluded that I had viewed the specimen ‘ wrong way up’,
but re-examination proves that these statements are correct. I lay no
stress on these differences, since it is more than possible that the arrange-
ment of these ducts is variable in different individuals.

448 W. N. F. WOODLAND

2mm. The general shape of these parasites is mdicated in
Pl. 25, fig. 24, which in itself provides evidence that the body is
very contractile and therefore variable in form and length.
In the general topography (PI. 25, fig. 25) of the genitalia (and
in the structural details described below) these parasites are
closely related to C. laticeps and other previously described
species of Caryophyllaeus, since, as in these species, the
genital openings are situated very near to the posterior end
and the uterus is therefore relatively very short, and the
testes extend over a great length of the body. One feature,
however, in which C. filiformis markedly differs from
C. laticeps and all other species of Caryophyllaeus
is the total absence of vitellaria posterior to the ovary, and of
a post-ovarian region, the ovary in most cases extending
to the extreme end of the body.

In all specimens of C. filiformis the genital apertures are
situated well within the last one-seventh of the body-length
(from one-twelfth to one-seventh according to the length of the
head). In six typical examples the lengths of the three regions
were as follows: in a specimen measuring 21 mm. long, the
head measured 10mm. and the testicular region 9mm.; in
a specimen 23-5 mm. long, the head was 11-5mm. and the
testicular region 10 mm. ; in a specimen measuring 21-5 mm.
the head was 9mm. and the testicular region 10-5 mm.; in
a specimen measuring 20mm. the head was 8mm. and the
testicular region 10 mm. ; in a specimen measuring 11-5 mm.
the head was 5-5 mm. and the testicular region 5mm. ; and in
a specimen measuring 15-3 mm. the head was 5-5 mm. and the
testicular region 8-3mm. In five of the specimens (Pl. 25,
figs. 26, c, d) the testes extended much more anteriorly, almost
to the extreme front of the head—thus in a specimen measuring
22mm. the head was only 1-5 mm. long, while the testicular
region was 19 mm.—but this is not the normal condition, and
intermediate conditions exist between this and the normal.

The head, as already stated, is extremely contractile and
assumes the most various shapes (Pl. 25, fig. 26, a, b, ¢, d, e).
When uncontracted it is fairly flat and ribbon-shaped (Pl. 25,

CESTODARIA 449

fig. 25), but it may also be pointed (PI. 25, fig. 26, c, d) or exces-
sively flattened, with a crenulated margin (Pl. 25, fig. 26, e) ;
when contracted the head becomes transversely wrinkled and
oval in transverse section (Pl. 25, fig. 27) and more robust
(Pl. 25, fig. 26, a). In transverse section (Pl. 25, fig. 27) it is
noteworthy that the longitudinal muscle-fibres are disposed
into two distinct zones, an outer peripheral zone (scLMus)
underlying the subcuticula, and an inner more powerful
medullary zone (IbMus), the two zones being widely separated.
This condition of the longitudinal musculature (which is
equally well-marked in the testicular region) is also found in
C. laticeps, and it differs essentially from the single-zone
condition of W. virilis and W. acuminata and probably
W. minuta. Some transverse muscle-fibres lie immediately
external to the medullary longitudinal musculature. In the
anterior head region of some specimens I have observed
longitudinal thickenings (Pl. 25, fig. 25) which may be the
‘faserzellenstringe’ of Will and Skrjabin. I have not
examined them in detail.

The testicular region is also remarkable, not only on account
of its length relative to the uterine region—a feature in which
C. filiformis again resembles C. laticeps—but in the
fact that the vitellaria throughout the greater part of this
region practically form a ring round the testes, when viewed
in transverse section (Pl. 25, fig. 28). In C. laticeps there
is an approach to this annular arrangement, but, judging
from Will’s figures (27), it is not nearly so complete as in
Ga taliformis.

The uterine region (Pl. 25, fig. 29), as already stated, is very
short, the uterus itself being short and but loosely coiled.
The uterus opens anteriorly into the vagina from the left side
(Pl. 25, fig. 25, is from the ventral aspect) and the ootype les
on the left side of the ovary, asin W. minuta. The male
and female apertures are more separate from each other than
is the case in the preceding species, a space of about 25 microns
separating the hind border of the male aperture from the anterior
border of the female, and this distance is fairly constant. In

450 W. N. F. WOODLAND

C. laticeps, W. acuminata, and W. minuta the
borders of the two apertures are contiguous, as is also usually
the case in W. virilis. In all other respects the genitalia
of C. filiformis appear to conform to the plan charac-
teristic of Caryophyllaeus.

The two main nerve-trunks (Pl. 25, figs. 25, 27, 28, n) are
visible anteriorly, and I have seen them (with difficulty) i my
transverse sections.

The excretory channels (Pl. 25, figs. 27, 28, 29) are mostly
situated between the inner and outer zones of longitudinal
muscles, but a few also occur in the medulla. The vessels
cut across in transverse sections form a part of a complicated
network, and there appear to’ be no definite longitudinal
vessels ; posteriorly, however, some eight or ten longitudinal
excretory channels open into the excretory bladder.

The eggs are of the usual ovoid smooth-shelled type and
measure, when mounted in balsam, 62-2-69-5 microns in
length and 29-2-32-9 microns in breadth, thus being excep-
tionally long. j

The definition of this species, C. filiformis, will be
supplied below.

Tur GENERA AND SPECIES OF THE CARYOPHYLLAEIDAE.

Skrjabin (28) has, with his description of Caryophyllaeus
syrdarjensis, included a summary of the small differences
which distinguish C. laticeps, C. tuba, C. fennicus,
and C. syrdarjensis (all from Cyprinoid fishes) from each
other, and C. armeniacus (Cholodkowsky, 2) only differs
from these principally in its large size (55mm. in length),
truncated head, and large eggs (80x45 microns). Linton
(10, 11) has described two other species (also collected from
Cyprinidae), which, on account of the form of the head, he
refers to Diesing’s (6) gnus Monobothrium,? but, excepting

1 The “ Monobothrium serpentum n. sp.” of v. Linstow (‘ Arch.
ft. Mikr. Anat.’, Bd. Ixii, 1903, p. 108), 1-4 mm. long, from a cyst in the
mesentery of a snake, Patalung, cannot of course be referred to the
Caryophyllaeidae. It is almost certainly a Dithyridium larva,
several kinds of which have been described from snakes.

CESTODARIA 451

for the head, both these species also are essentially identical
with the five species of Caryophyllaeus above named.
Finally, Cooper (5) has described, under the name of *‘ Glari-
dacris catostomi’, a species (from a Cyprinoid fish)
which also ‘closely resembles’ the aforesaid five species of
Caryphyllaeus and the two species described by Linton
in general structure, and possesses a ‘ scolex’ which is * quite
similar at least in outward appearance to that of Archigetes
brachyurus, Mrazek’. Cooper appears to have been
unacquainted with the work of Lintonon* Monobothrium’,
since he supposes that his ‘Glaridacris’ is ‘the first
member of the group [Caryophyllaeidae] to be described
from America’. He bases his new genus ‘Glaridacris’
mainly on the characters of the ‘ scolex’. Now this ‘ scolex ’
of ‘Glaridacris’, ‘when not strongly contracted, has
somewhat the form of a truncated rectangular pyramid with
the longer diameter in the transverse direction . . . the edges
of the base and the apex protrude markedly, in this latter case
forming a terminal disc comparable to that of many of the
bothriocephalid Cestodes. The dorsal and ventral faces of
the organ are each divided by two ridges converging towards the
apex into three sucking grooves or loculi, of which the middle
is best developed and most efficacious during life. It is also
the last to become smoothed out with strong contraction of the
whole scolex. The lateral loculi are, furthermore, not in the
same plane with the medial one but inclined towards the corre-
sponding ones of the opposite surface so that the edges of the
scolex, especially just behind the terminal dise, are often not
much thicker than the ridges between the loculi . . . the scolex
of this form assumes a greater variety of shapes than that of
any other tapeworm I have yet examined, in which respect it
is comparable to the leaf-like anterior end of Caryophyl-
laeus.’ From which description and from Cooper’s figures,
it is evident that the ‘scolex’ of ‘Glaridacris’, besides
resembling that of Archigetes brachyurus, is extremely
similar to the head of Linton’s ‘Monobothrium hexa-
cotyle’, which also has six loculi and a central papilla
( which may project forward as a sharp conical elevation or be

452 W. N. F. WOODLAND

contracted to a low eminence’) and is similarly ‘ versatile ’
inform. If, then, Linton is right in referring his species ‘ hexa-
cotyle’’ and ‘ terebrans’ (which latter has a head ‘ variable,
subsagittate, wedge-shape or bluntly rounded ’) to the genus
Monobothrium, it is evident that Cooper's species must
likewise be referred to this genus, and ‘Glaridacris’
becomes a nomen nudum.

The genus Monobothrium was created by Diesing in
1863 (6) as one of three genera in his family of the Mono-
bothria, i.e. forms with one ‘ bothrium’; the other two
genera being Caryophyllaeus, Gmelin, and Diporus,
Diesing. The genus Diporus can at once be elimimated as
solely having reference to the ‘Caryophyllaeus trisig-
natus’ of Molin (14), a form! from the intestine of ‘Gadus
merlucius’ which, in view of Molin’s description and figures,
cannot be referred with any degree of probability even to the
Caryophyllaeidae and much less to any particular genus.
The genus Monobothrium, according to Diesing’s defini-
tion, only differs from Caryophyllaeus in that (a) the
body is not ‘ depressum ’, that (b) the head is ‘ subeylmdricum,
bothrio, terminali subcirculari’ mstead of being ‘ dilatatum
fimbriatum, bothrio terminali transverso bilabiato ’ (definition
of Caryophyllaeus head), and that (c) the male and female
genital apertures open into a common atrium (‘ apertura
genitalis unica ’) instead of opening separately on the surface
(‘apertura genitalis feminea pene postposita contigua °) as in
Caryophyllaeus. Diesing describes two species of Mono-
bothrium—M. tuba (the Caryophyllaeus tuba of
Wagener (15) and Monticelli (16)) from the intestine of Tinea
chrysitis, and M. punctulatum, the ‘Caryophyl-
laeus punctulatus’ of Molin (14) from the intestine
of Conger vulgaris—another indeterminate organism
deseribed by this author which probably is not even a
Caryophyllaeid.

The ‘Monobothrium tuba’ of Diesing was considered
both by Wagener previously and by Monticelli subsequently

1 Possibly a Tetrabothriid scolex, as Monticelli suggests.

i {
—

Bishop, del,

Quart. Journ. Micr. Sci.

Vol. 67, N.S.,

AG
ws

scl

Bishop, del.

i

oe

12.

¥ 13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

CESTODARIA 469

Gmelin, J. F.—‘‘ Caroli 4 Linné ’’, ‘ Systema Naturae ’, tom. i, pars vi,
1790, p. 3052.

. Haswell, W. A.—‘‘ On a Gyrocotyle from Chimaera ogilbyi, and on

Gyrocotyle in general’’, ‘ Proc. Linn. Soc. N. 8. Wales’, vol. xxvii,
1902, p. 48.

Hungerbiithler, Max.—‘‘ Studien an Gyrocotyle und Cestoden ’’, * Zool.
u. anthrop. Ergeb. einer Forschungsreise im westlichen u. zentralen
Siidafrika’, 1903-5, Bd. iv, pt. 3, p.495. ‘ Denkschr. d. Med.-Natur.
Gesell.’, Bd. xvi, Jena, 1910.

Linton, E.—‘‘ On Fish Entozoa from Yellowstone National Park ’’,
*‘U. S. Comm. of Fish and Fisheries, Part XVII. Commissioner’s
Report, 1889-91’, 1893, p. 545.

—— ‘‘ Notes on Cestode Parasites of Fishes ’’, ‘ Proc. U. S. National
Museum ’, vol. xx, 1898, p. 423.

Lénnberg, E.—“ Beitriige zur Phylogenie der parasitischen Plathel-
minthen”’, ‘ Centralbl. f. Bakt. u. Parasit.’, Bd. xxi, 1897, pp. 674,
(PAS.

Lithe, Max.—‘ Die Siisswasserfauna Deutschlands’, Heft 18:
** Parasitische Plattwiirmer. II: Cestodes’’, Jena, 1910.

Molin, R.—‘‘ Prodromus Faunae Helminthologicae Venetae”’’,
* Denkschr. der K. Akad. der Wissensch., Math.-Naturw. Cl., Wien’,
Bd. xix, 1860, p. 189.

Monticelli, Fr. Sav.—‘‘ Saggio di una morfologia dei Trematodi”,
‘Tesi per ottenere la privata docenza in Zoologia nella R. Uni-
versita di Napoli’, viii, 130 pp., 1888.

‘** Appunti sui Cestodaria’’, ‘Soc. Reale di Napoli. Atti della

Reale Accad. delle Scienze fisiche e mat.’, ser. 2, vol. v, 1893, no. 6.

Mrazek, Al.—‘‘ Ueber eine neue Art der Gattung Archigetes”’,
*Centralbl. f. Bakt., Parasit. u. Infekt.’, I. Abth., Bd. xlvi, 1908,
p. 719;

Miiller, O. F.—‘‘ Verzeichniss der bisher entdeckten Eingeweidewiirmer,
der Thiere, in welchen sie gefunden worden, und der besten Schriften,
die derselben erwihnen’’, ‘Der Naturforscher’, XXII. Stiick,
Halle, 1787, p. 79.

Pallas, P. S.—‘‘ Bemerkungen iiber die Bandwiirmer in Menschen und
Thieren *’, ‘Neue nord. Beytrage z. physik. und geogr. Erd- u.
Volkerbeschreibung, Naturgesch. u. Oeconomie’, Bd. i, 1781,
St. Petersb. and Leipzig.

Ratzel, F.—‘* Zur Entwickelungsgeschichte der Cestoden’”’, ‘ Arch. f.
Naturg.’, XXXIV, Jahrg., Bd. 1, 1868, p. 138.

Rudolphi, K. A.—‘‘ Beobachtungen iiber die Eingeweidewiirmer ”’,
*‘Wiedemann’s Archiv f. Zool. u. Zoot.’, Bd. 111, 1801.

Schneider, G.—‘‘ Caryophyllaeus fennicus, n. sp.’’, ‘ Archiv f. Naturg.’,
LXVIII. Jahrg., Bd. i, 1902, p. 65.

NO. 267 ite

470 W. N. F. WOODLAND

¥ 23. Skrjabin, K.—‘“‘ Fischparasiten aus Turkestan. 1. Hirudinea et

Cestodaria ”’, ‘ Archiv f. Naturg.’, LX XIX. Jahrg., Abt. A, Heft 2,
1913, p. 2.

24, Spencer, W. Baldwin.—‘‘ The Anatomy of Amphiptyches urna (Grube
and Wagener) ”’, ‘ Trans. Royal Soc. Victoria’, vol. i, 1889, p. 138.

25. Wagener, G. R.—‘‘ Die Entwicklung der Cestoden, nach eignen Unter-
suchungen’’, ‘Nov. Act. Acad. Caes. Leopold-Carol. Nat. Cur.’,
Bd. xxiv, Suppl., Breslau, 1854.

26.—-Watson, E. C.—‘‘ The Genus Gyrocotyle and its Significance for
Problems of Cestode Structure and Phylogeny ’’, ‘ Univ. California
Pub. Zoology ’, vol. vi, no. 15, 1911, p. 353.

27. Will, H.—‘‘ Anatomie von Caryophyllaeus mutabilis, Rud. Ein
Beitrag zur Kenntnis der Cestoden ”’, ‘ Zeit. f. wiss. Zool.’, Bd. lvi,
1893, p. 1.

28. Woodland, W. N. F.—‘‘ On Amphilina paragonopora, sp. n., and a
hitherto undescribed Phase in the Life-history of the Genus”’,
‘Quart. Journ. Micr. Sci.’, vol. Ixvii, N.S., 1923, p. 47.

DESCRIPTION OF PLATES 24 AND 25.

REFERENCE LETTERS.

BH, base of head; CIRS, cirrus sac; CO, cirrus opening; DVMs, dorso-
ventral muscle-fibres; Excv, excretory vessels; FzS, ‘ Faserzellen-
strange’ ?; H, head region; ILMus, internal layer of longitudinal muscle-
fibres on edge of medulla; tov, isthmus of ovary ; 3N, junction of the two
main longitudinal nerve-trunks anteriorly ; LcH, longitudinal creases or
grooves in subcuticula of head; Lp, local dilatation; tmus, longitudinal
muscle-fibres; M, medulla; N, main longitudinal nerve-trunk; NT,
nuclear thickening at base of head; opcts, ducts from ovary opening
into isthmus; OOTP, ootype; OPO, opening of oviduct into ootype: Oy,
ovary; OVD, oviduct; OviT, opening of vitelline duct into ootype;
POVR, post-ovarian region; RCPS, receptaculum seminis; SCLMUS, outer
layer of iongitudinal muscle-fibres underlying subcuticula; sHGI, shell-
gland ; TEs, testes ; TEXB, terminal excretory bladder ; TRMUS, transverse
muscle-fibres ; UopP, opening of uterus into vagina; UT, uterus; VAG,
vagina; VAGUT, rudiment of vagina and uterus; VDEF, vas deferens ;
vit, vitellaria ; virp, vitelline duct ; vuo, vagino-uterine opening.

N.B.—The magnifications given are those at which the figures were
drawn. The 5 cm. scales provided on the two plates show, when compared
with an actual 5 cm., the amount of reduction + which has occurred in repro-
ducing the figures. All figures drawn by means of the camera lucida.

1 Approximately 4

url

det

rssh it |

Woodland, del.

Vol. 67, N.S, Pl. 24

Quart. Journ. Micr. Sci.

TRMUS LMUS

PaSIAW

BVMS

TEXB

\

SSSR a aE

REA

VAG

CESTODARIA 471

PLATE 24,
Wenyonia virilis.

Fig. 1, a, b, c (x 2).—Adult specimens as mounted on slides and probably
somewhat flattened.

Fig. 1, d ( x 2).—Three immature specimens.

Fig. 2 ( x 24).—Entire specimen magnified (the two halves of the figure
should be continuous) and viewed from the dorsal aspect. The parts of
the drawing round the sexual openings and the ovary have been made
somewhat diagrammatic for the sake of clarity, and the uterus is in
actuality full of eggs. The vitelline and other fine ducts are not shown.

Fig. 3, a, b, c, d, e (x 24).—Figures showing the variations in form of
the head region.

Fig. 4 ( x 35).—Horizontal section through a much contracted head.

Fig. 5 ( x 35).—Transverse section through the anterior region of the
expanded head to show the longitudinal grooves (the thick cuticle and
excretory channels are not indicated).

Fig. 6 (x 78).—Transverse section through the posterior region of the
head to show the longitudinal grooves, musculature and excretory net-
work (thick cuticle not shown).

Fig. 7 (x 112).—Ventral aspect of the genital openings and associated
ducts.

Fig. 8 ( x 175).—Dorsal aspect of the ducts in the region of the ovary in
outline.

Fig. 9 (x24).—The abbreviated (contracted ? regenerated ?) post-
ovarian region in one of the two specimens showing this feature.

Caryophyllaeus laticeps.

Fig. 10 (x9).—C. laticeps from Tinca vulgaris, from the dorsal
aspect (for comparison with Wenyonia spp.). The structures in the
regions of the sexual apertures and ovary are represented somewhat
diagrammatically, but the figure is otherwise correct. In most specimens
of C. laticeps, however, the uterus does not extend anteriorly to the
cirrus opening, at least to any extent.

Wenyonia virilis.

Fig. 11 ( x 1060).—Egg (contained in the uterus).

Fig. 12 (x 78).—Transverse section through the uterine region, viewed
from the front aspect, showing the vagino-uterine aperture.

Fig. 13 ( x 78).—The posterior excretory system artificially injected.
Figs. 14, 15 ( x 24).—Immature specimens showing developing genitalia.

Wenyonia acuminata.

Fig. 16 (x 2).—Adult specimens (the dilated base of the head shown in
two of the specimens is probably not a constant feature).
mole

472 W. N. F. WOODLAND

Fig. 17 ( x9).—Entire specimen magnified and viewed from the dorsal
aspect. The parts of the drawing round the sexual openings and the
ovary have been made somewhat diagrammatic for the sake of clarity
and the uterus is in actuality full of eggs. The vitelline and other fine ducts
are not shown.

Fig. 18 ( x 55).—Longitudinal vertical section through head.

Fig. 19 (x55).—Transverse section through the anterior end of the
testicular region.

Fig. 20 ( x 55).—Transverse section through the middle of the uterine
region.

Fig. 21 ( x 1060).—Eggs in optical section and surface-view to show the
minute spinelets on the shell.

Wenyonia minuta.

Fig. 22 ( x 2).—The adult specimen.

Fig. 23 ( x 60).—The entire specimen magnified and viewed from the
ventral aspect. The parts of the drawing round the sexual opening
and the ovary have been made somewhat diagrammatic for the sake of
clarity and the uterus is in actuality full of eggs. The vitelline and other
fine ducts are not shown. The anterior extension of the ovary to the level
of the sexual openings is not certain, since the ovarian follicles and vitel-
laria are in this region almost indistinguishable in my preparation.

PLATE 25.
_ Caryophyllaeus filiformis.

Fig. 24 ( x 2).—Adult specimens.

Fig. 25 ( x 55).—Entire specimen magnified and viewed in optical section
from the ventral aspect (the three parts of the figure should be continuous),
The parts of the drawing round the sexual openings, and the ovary have
been made somewhat diagrammatic for the sake of clarity and the uterus

is in actuality full of eggs. The vitellaria and other fine ducts are not

shown. The vitellaria in actuality surround the testes in the testicular
region—vide fig. 28.
Fig. 26, a, b, c, d, e ( x 24).—Drawings showing variation in form of the
head and the variable extension of the testes and vitellaria anteriorly.
Fig. 27 (x 78).—Transverse section through the head region.
Fig, 28 ( x 78).—Transverse section through the testicular region.
Fig. 29 ( x 78).—Transverse section through the anterior uterine region.
Fig. 30 (x 78).—Transverse section through the median isthmus of the
ovary. The vagina has at this level become dorsal to the uterus.

Immature Caryophyllaeid from Auchenoglanis occidentalis.
Fig. 31 ( x 2).—The single specimen.
Fig. 32 ( x 35),—The external aspect of the specimen, viewed ventrally.

_

Quart. Journ. Micr. Sci. Vol. 67, N.S., Pl. 25

Dis

ene Unt aert

moe

rhea aS 30
5 3 & 5 coms ak OT iLmus

Woodland, del.

CESTODARIA 453

to belong to the genus Caryophyllaeus, the differences
from ‘C. mutabilis’ (=C. laticeps) only being specific
and not generic, and in view of the indefiniteness of such terms
as * subcylindrical’ when applied to a pleomorphic structure
like the head, of ‘ depressum ’ when applied to the body which
may be flat anteriorly and oval in section posteriorly even in
the same animal, and of ‘ single’ and ‘ separate’ when these
refer to genital apertures which show all degrees of opening
into a common atrium in different species which undoubtedly
are to be included in one genus, it is impossible not to agree
with Wagener and Monticelli. Further, as is demonstrated
in the present communication, the muscular cylindrical type
of head, which was the most characteristic feature of Diesing’s
genus Monobothrium, is to be found associated, indis-
criminately, with two very different types of body, whence it
is evident that the character of the head cannot serve in this
sroup as a feature of generic value. The name Monobothrium
thus also lapses.

Having thus concluded that eight of the ten Caryophyllaeid
species described previously to this communication belong to
the genus Caryophyllaeus, and provisionally allowing
Archigetes to stand as a separate genus on the sole ground
that it possesses a ‘caudal’ appendage in the sexual condition,
it is necessary to redefine the genus Caryophyllaeus and
to show reason for creating a new third genus, Wenyonia.

The genus Caryophyllaeus was apparently founded by
O. F. Miller in 1787 (18), but since I have not been able to
consult this work I will give the definition of this genus supplied
by Gmelin (7) in 1790, which is probably fully as comprehensive
as that of Muller, both being based on specimens of C. lati-
ceps. Gmelin’s definition of the genus is: ‘ Corpus teres :
ore fimbriato. . . . Habitat in piscium, aquas dulces inhabitan-
tium, potissimum cyprinorum, carpionis, tincae, jesis, bramae
intestimis, rarior, margaritaceus, ad pollicem longus, reliquis
intestinalibus vitae tenacior, fine posteriori rotundato, anteriori
latiori.’ This definition clearly applies to C. laticeps,
but it is too special to include other species subsequently

NO. 267 Hh

6~

454 WwW. N. F. WOODLAND

described, and Diesing (6) therefore, as aforesaid, created
another genus—Monobothrium—to include the ‘ Ligula
tuba’ of Wagener (25) and the ‘Caryophyllaeus pune-
tulatus’ of Molin. Diesing’s definitions of the two genera
Caryophyllaeus and Monobothrium are respectively
as follows: ‘I. Caryophyllaeus, Gmelin. Corpus con-
tinuum elongatum depressum, vesicula pulsatoria postica cum
poro excretorio. Caput dilatatum fimbriatum, bothrio ter-
minali transverso bilabiato. Os ... Collum nullum. Penis
lateralis conicus retractilis retro medium corporis ; apertura
genitalis feminea pene postposita contigua. In Piscium
fluviatilium praeprimis Cyprinorum intestinis. Evolutio
directa.’ [One species C. mutabilis, Rud.|,and ‘Il. Mono-
bothrium, Diesing. Corpus continuum elongatum, vesicula
pulsatoria postica, caput sybeylindricum, bothrio terminali
subcireulari. Os... Collum nullum. Apertura genitalis unica,
organo musculo et femineo communis, lateralis ventralis in
postico corporis triente. In Cyprmorum intestinis. Evolutio
ignota. (Animaleula bothri ope parieti intestini firmiter
adhaerent. Systema vasorum e truncis longitudinalibus
quatuor et vasculorum rete compositum).’ [Two species :
M. tuba, Wagener, and M. punctulatum, Molin.]
Other and more recent authors, e.g. Lihe (18) and Cholod-
kowsky (8), group the species of the Caryoph yllaeidae
into two genera—Caryophyllaeus and Archigetes—
and distinguish the former from the latter chiefly by the
absence of the * caudal’ appendage in the sexual stage of the
former.

Assuming the definition of the Caryophyllaeidae to
be that which I shall propose in the next section, I now
propose to divide these forms into three genera—Caryophyl-
laeus, Archigetes, and Wenyonia—with the following
definitions :

Caryophyllaeus, O. F. Miller, 1787 emend.

The sexual apertures are situated within the last quarter
of the body-length, and the ovary is near the posterior
extremity. The longitudinal extent of the uterus is at

CESTODARIA 455

most one-third of that of the testes and usually much less.4
Parasitic in the intestines of Malacopterygii and
Ostariophysi (Cyprinidae, Mormyridae, and
Siluridae?).

Archigetes, Leuckart, 1878.

The sexual apertures are situated within the last quarter
of the body-length; and the ovary is near the posterior
extremity. The longitudinal extent of the uterus is at
most one-third of that of the testes and usually much less.
The body of the mature parasite possesses a ‘ caudal’
appendage posteriorly. Parasitic in the body-cavity of
aquatic Oligochaetes.

Wenyonia, Woodland, 1923.
The sexual apertures are situated in the anterior half of
the body. The longitudinal extent of the uterus is at
least equal to that of the testes. Parasitic in the intestine
of Siluridae.

The species known up to the present contained in these three
genera are as follows :

C. laticeps, Pallas, 1781.
Length of body 11-80 mm.; breadth of body 0-5-2 mm.
Body posterior to the head oval in transverse section.
Hind end bluntly pointed or rounded. Head region

1 With the single exception of Linton’s ‘Monobothrium hexa-
cotyle’ (C. hexacotyle), which has eggs measuring only 38-40 microns
in length, all species of Caryophyllaeus, as enumerated further in
the text, possess eggs about 60 microns in length, or, as in the huge
C. armeniacus, 80 microns in length. On the other hand, in all three
species of Wenyonia the eggs average from 32 to 40 microns in length.
These differences are remarkable (especially in view of the body-sizes of
the parasites) and may possess a classificatory value. Another difference
which may distinguish these two genera is that in Wenyonia the
longitudinal muscles are solely contained in a peripheral subcuticular
zone, whereas in Caryophyllaeus the longitudinal muscles are disposed
in two zones—a subcuticular and an epi-medullary.

2 IT include this family because of the immature species of Caryo-
phyllaeus found by me in Auchenoglanis occidentalis and
described briefly at the end of this paper.

Hh2

456 W. N. F. WOODLAND

flattened, usually broader than the body and with fim-
briated edges, but very contractile and variable in form ;
longitudinal grooves absent on head. Genital apertures
in last fifth of body-length.1 Post-ovarian vitellaria are
present. Eggs measure about 66 microns in length.
Parasitic in intestine of Cyprinidae, Europe.

C. tuba, Wagener, 1854 (‘ Monobothrium tuba’).
Length of body 10-80 mm.; breadth of body 0-9-1 mm.
Body and head oval in transverse section. Hind end
slightly tapermg and pointed. Head ‘stumpy’, i.e.
bluntly rounded in outline and not broader than body ;
longitudinal grooves absent on head. Genital apertures
situated anterior to the last fifth of the body.? Post-
ovarian vitellaria are present. Eggs? Parasitic in intes-
tine of Tinca chrysitis, Italy.

C. fennicus,? Schneider, ee

Length of body 5-9-5 mm.; breadth of body 0-4-0-5 mm.
Body and head eit flat in transverse section.
Hind end bluntly pomted. Head rounded in contour
and not broader than body ; longitudinal grooves absent
on head. Genital apertures as in C. laticeps. Post-
ovarian vitellaria present. Eggs measure about 60 microns
inlength. Parasitic in intestine of Leuciscus erythro-
phthalmus, Finland.

C. syrdarjensis, Skrjabin, 1918.
Length of body 6-38-16 mm.; breadth of body 1-1-5 mm.
Body and head somewhat flat in transverse section.
Hind end bluntly rounded. Head slightly flattened and
slightly broader than body, with distinet ‘ faserzellen-
stringe’; longitudinal grooves absent on head. Genital
apertures apparently as in C. laticeps. Post-ovarian
vitellaria are present. Eggs measure about 63 microns in
length and 48 microns in breadth. Parasitic in intestine
of Schizothorax intermedius, Russo-Turkestan.

1 Statement based on measurements of four individuals in my posses-
sion.

* Statement based on the figures supplied by Monticelli (15) and
Wagener (25).

3 Skrjabin (28) is mistaken in supposing that in C. fennicus alone
do the coils of the uterus ever extend anterior to the vagino-uterine
aperture: I have seen this anterior extension in at least two specimens
of C. laticeps (PI. 24, fig. 10).

C.

4

C.

C:

CESTODARIA 457

armeniacus, Cholodkowsky, 1915.

Length of body reaches 55 mm., with breadth of 5 mm.
Body and head oval in transverse section. Hind end
slightly tapering and bluntly pointed. Head not broader
than body, truncated, with wedge-shaped anterior exten-
sion; longitudinal grooves absent on head. Genital
apertures in last fifth of body-length. Post-ovarian
vitellaria are probably present. Eggs about 80 microns
long and 45 microns broad. Parasitic in intestine of
Capoeta sp., Armenia.

terebrans, Linton, 1898 (Monobothrium tere-

brans’).

Length of body reaches 28 mm., with breadth about 2-5 mm.
Body oval in transverse secian: Hind end dilated at level
of sexual apertures and pointed posteriorly. Head broader
than body and subsagittate, wedge-shaped or bluntly
rounded; longitudinal grooves absent. Genital apertures
at about the anterior limit of the last fifth of the body.
Post-ovarian vitellaria are present. Hggs measure 60-65
microns in length and 30-85 microns in breadth. Parasitic
in intestine of Catostomus ardens, Wyoming, U.S.A.

hexacotyle, Linton, 1898 ( Monobothrium hexa-

cotyle’)

Length of body reaches 14-5 mm., with breadth of 1 min.
Body oval in transverse section. Hind end tapering and
pointed. Head not broader than body, broad at base,
pointed anteriorly and wedge-shaped, each of the dorsal
and ventral surfaces of the wedge bearing three broad
elongated loculi or triangular grooves which point to the
anterior extremity ; shape variable. Genital apertures in
last fifth of body-length. Post-ovarian vitellaria are
present. Eggs measure 88-40 microns in length and
20 microns in breadth. Parasitic in the intestine of
Catostomus sp., Arizona, U.S.A.

catostomi, Cooper, 1920 (‘Glaridacris cato-

stomi’).

Length of body measures 5-25 mm., with maximum breadth
0-4-1mm. Body. oval in transverse section. Hind end
tapering and pointed. Head region broader than body,
and similar in shape to that. of Archigetes brachy-
urus, i.e. wedge-shaped, the dorsal and _ ventral
surfaces of the wedge each bearing three broad loculi,
the end of the wedge being truncated when viewed dorsally

458 W. N. F. WOODLAND

or ventrally but pointed in lateral view ; shape variable.
Genital organs * like those of Caryophyllaeus’. Post-
ovarian vitellaria present. Eggs measure 54-66 microns
in length and 38—48 microns in breadth. Parasitic in the
intestine of Catostomus commersonii, Michigan,
U.S.A.

C. filiformis, Woodland, 1923.

Length of body measures 7-5-24 mm., with breadth (maxi-
mum) 1-2mm. The body is broadest in its posterior half,
the anterior half (or third) being attenuated and usually
filiform or ribbon-shaped ; the body posterior to the head
region is oval in transverse section. The hind end of the
body is bluntly rounded. The head region is highly con-
tractile, is more or less pointed in front, is narrow, always
more flattened than the body posteriorly, and varies im
form from a thin pointed filament or ribbon to a tapering
wrinkled digitiform structure ; longitudinal grooves are
absent. Genital apertures in last seventh of the body-
length ; the cirrus aperture is distinctly separated from
the vagino-uterine aperture, and there is no atrium.
Post-ovarian vitellaria are entirely absent. Eggs (in
balsam) measure 62-2-69-5 microns in length and 29-2—
32-9 microns in breadth. Parasitic in the intestine of
Mormyrus caschive, L., Anglo-Egyptian Sudan.

Archigetes appendiculatus, Ratzel, 1868, and A.
brachyurus, Mrazek, 1908.

Both of these species possess ‘ caudal’ appendages, the
former an appendage about equal in length to the rest of
the body, the latter an appendage only about one-sixth
the length of the rest of the body. A. appendi-
culatus varies from 2-5 to 3 mm. in length, and is parasitie
in the body-cavity of Limnodrilus hoffmeisteri
and Tubifex rivulorum, Europe; A. brachy-
urus is 5-6 mm. long and is also parasitic in the body-
cavity of Limnodrilus hoffmeisteri. The genital
organs of these two species apparently resemble those of
Caryophyllaeus.

Wenyonia virilis, Woodland, 1923.
Length of body varies from 11 to 52-5 mm. and maximum
breadth from 1-5to 3mm. The body is externally divisible
into four distinct regions-—a broad-pointed head region,
a narrow elongated testicular region, a dilated uterme
region, and a normally long narrow tapering post-ovarian

CESTODARIA 459

region ; the body and head are oval in transverse section.
The testicular and uterine regions, which are roughly equal
in length, together occupy ‘about. one-third of the total
body-length. The post-ovarian region is at least equal in
length to the combined testicular and uterine regions. ‘The
hind end of the body gradually tapers to a fine point.
The head region is normally sagittate in form, its base
only being slightly broader than the testicular region,
is oval in transverse section, highly muscular, and hears
thirteen to twenty-six deep longitudinal grooves which
are not suctorial or fixative in function; highly con-
tractile. Post-ovarian vitellaria about equal, as regards
area covered, to the anterior vitellaria. Eggs (in balsam)
measure 387-6-41-3 microns in length and 20-1—-25-6
microns in breadth, thus being very small compared with
eggs of Caryophyllaeus. Parasitic in intestine of
the Siluroid Synodontis schall, Anglo-Egyptian
Sudan.

Wenyonia acuminata, Woodland, 19238.

Length of body varies from 17 7-5 to 34-5 mm. and maximum
breadth (in testicular region) 1-2 to 1-5mm. ‘The body is
not externally divisible into regions but is nematodiform
and oval in transverse section. The testicular and uterine
regions, which are roughly equal in length, together
occupy about two- thirds of the total body- clength. The
post-ovarian region is never more than one-third the length
of the combined testicular and uterine regions, and: is
usually less. The hind part of the body ‘only slightly
tapers and ends in a blunt point. The head region 1s
tapering and finely pomted and narrower than the testicular
region ; it is not highly muscular, not specially contractile,
and bears no longitudinal grooves. ‘The post-ovarian
vitellaria are not so numerous as the anterior vitellaria.
Eggs (in balsam) measure 34-7-86-6 microns in length
and 21-9-25-6 microns in breadth, and are distinguished
from the eggs of all other (at present) known Caryo-
phyllaeidae by the egg-shell being covered with
minute spinelets. Parasitic in the intestine of Syno-
dontis membranaceus, Anglo-Egyptian Sudan.

Wenyonia minuta, Woodland, 1923
Length of body about 3-5mm. and breadth 0-5 mm.
The body is fluke-like m form, being flat, oval in outlme,
and with bluntly pointed extremities. The uterine region

460 W. N. F. WOODLAND

is at least twice the length of the testicular region and the
two regions together occupy about two-thirds of the
entire body-length. The post-ovarian region is incon-
spicuous. ‘The head region is, like the posterior regions,
flattened, and much narrower than the body, is bluntly
pointed, with irregular edges—in fact Caryophyllaeiform ;
it bears no longitudinal grooves and is very contractile.
The post-ovarian vitellaria form a small group continuous
with the anterior broad marginal strands. Eggs (in
balsam) measure 32-9-40-2 microns in length and 20-1-
25-6 microns in breadth. Parasitic in the intestine of the
Siluroid Chrysichthys auratus, Anglo-Egyptian
Sudan.

ON THE CARYOPHYLLAEIDAE, GYROCOTYLIDAE AND
AMPHILINIDAE, AND CESTODARIA IN GENERAL.

Having re-defined the genera of the Caryophyllaeidae,
it is necessary to re-define the family, smce previous definitions
not only alone refer to Caryophyllaeus and Archigetes
forms, but appear to me to be defective in emphasizing some
of the more important features displayed by these genera.

The family was apparently first defined, with some degree of
precision, by Claus in 1885 (4), stress being laid upon the
elongated, unsegmented body, wrinkled anterior border,
absence of hooks, eight or more main excretory canals, the
general features of the genital apparatus and simple develop-
ment, and subsequent authors have not improved upon this
definition to any extent. In re-defining the Caryophyl-
laeidae we have to distinguish them from the two other
families of the Cestodaria, viz. the Gyrocotylidae and
Amphilinidae, and therefore the features to be emphasized
are that the body is usually not flattened to the extent that
the Cestode strobila is, that calcareous corpuscles are
entirely absent, that they never possess true circular suckers,
that sucking grooves or bothria may be present (always different
in number, form, and arrangement to those found in merozoan
Cestoda) in some forms (C. catostomi, C. hexacotyle,
and Archigetes brachyurus), that the vagina and
uterus form a complete circuit with a single common opening

CESTODARIA 461

to the exterior, that the arrangement of the genitalia is very
distinctive, and that the larva of some species is hexacanth.
Taking these features into account, the Caryophyllaeidae
may be defined as

Cestodaria, usually with a slightly flattened cylindrical
elongated body but sometimes fluke-shaped, devoid of
calcareous corpuscles and cuticular spinelets or hooks,
with an anterior end extremely variable in form and size,
both in the individual and in different species, which never
carries circular suckers but may bear shallow elongated
crooves (different in number, form, and arrangement to
those found on the scolex of other Cestoda), with the cirrus
and vagino-uterine apertures contiguous or nearly con-
tiguous on the ventral surface of the body in the median
line and occasionally opening into a common shallow
atrium, with the testes situated anteriorly and entirely
in front of the uterus, with the vagina and uterus forming
a complete circuit with a common vagino-uterine opening
to the exterior, with a network of excretory channels, the
larger ones of which form irregular longitudinal canals
about eight or ten in number and all of which open
externally by a median posterior excretory bladder, and
with a larval form known in some cases to be hexacanth.
Parasitic in the intestine of Teleostome fishes (Malaco-
pterygii and Ostariophysi) and in the body-cavity of
aquatic Oligochaeta (Tubificidae).

Re-definition of the Caryophyllaeidae necessitates
re-definitions of the other two families of the Cestodaria. The
Gyrocotylidae, comprising four species of the well-known
genus Gyrocotyle (24, 8, 26), may be defined as

Cestodaria with a flattened elongated body, devoid of
calcareous corpuscles but possessing cuticular spinelets,
with an anterior ovoid sucker and a posterior ‘ rosette ’
in all known forms, with the cirrus and vaginal apertures
adjacent but not contiguous, and situated anteriorly on
the left margin of the body, with the testes situated
anteriorly and to the outer sides of the median uterus,
with the uterine aperture situated anteriorly on the ventral
surface in or near the median line, a short distance but quite
separate from the vaginal aperture, with a close network
of fine exeretory vessels devoid of main longitudinal

462 W. N. F. WOODLAND

channels and not opening by a posterior vesicle, and with
typical hexacanth larvae in some species but in others
a ten-hooked larva similar to that of Amphilinidae.
Parasitic in the intestine of Holocephali.

The Amphilinidae! may be re-defined as

Cestodaria with a flattened more or less elongated body,
possessing calcareous corpuscles but devoid of cuticular
spies and hooks, with a large anterior boring apparatus
(proboscis, boring muscle and anchor cells) but devoid of
suckers and bothria, with the cirrus and vaginal apertures
either in close apposition or only separated by a short
distance, both situated posteriorly on or near the edge of the
hind end of the body, with the testes extending over, with
the uterus, the greater length of the body, usually as two
marginal rows lying external to the coils of the uterus but
sometimes more ‘scattered, with the vagina lying posterior to
the ovary and the uterus anterior, with the uterine opening
at the extreme anterior end, with a very long uterus,

consisting of three limbs disposed like the letter N when
viewed ventrally, with two main lateral longitudinal
excretory channels opening medianly at the posterior
extremity, and with a larval form possessing ten hooklets.
Parasitic in the body-cavity of Acipenseridae,

Osteoglossidae (Arapaima), Haemulidae (Dia-
gramma), and Siluridae (Macrones).

Comparing these three families, it appears to me self-evident
that the Caryophyllaeidae and Gyrocotylidae are
much more closely allied to each other than is either of these
families to the Amphilinidae. In both of the former
families (a) the three sexual apertures are all grouped together,
and at least two of them (the uterine and vaginal) are, in both
families, situated on the (ventral ?) surface in or near the median

1 See the writer’s paper on Amphilina paragonopora recently
published in this journal (28). I may remark in this place that Poche’s
paper “Zur Kenntnis der Amphilinidea”’ (‘ Zoologischer Anzeiger ’,
Bd. liv, 1922, p. 276) appeared too late for me to refer to it in my own
paper, and that I am wholly unable to concur with the author’s grotesque
proposal to found separate families and sub-families for known species
of Amphilina. The best policy to adopt with suggestions of this kind is to
ignore them.

CESTODARIA 463

line and not on the edge of the body ; (b) in both families the
uterus and vagina run parallel to and in the same dorso-
ventral plane with each other along the longitudinal axis of
the body, and in both the testes occupy only the anterior end
of the body; (c) in both families the ovary is bipartite, and
(d) in both calcareous corpuscles are absent; (e) in both
families hexacanth embryos have been found ;1 and (f) a
reticulate type of excretory system is common to both. On the
other hand, in the Amphilinidae, (a) the uterme and
vaginal apertures are situated at opposite ends of the body and
are in all cases situated on or near the body edge ; (b) the uterus
and vagina run in opposite directions from the ovary, the former
anteriorly and the latter posteriorly ; (c) the ovary is uni-
partite ; (d) caleareous corpuscles are present ; (e) hexacanth
larvae have not been found, the larva being of the ten-hooked
kind ; and (f) the excretory system consists of two main longitu-
dinal canals only and minor regularly arranged looped canals.
Tam thus unable to agree either with Monticelli (15) in believing
that Amphilina and ‘Amphiptyches’ (Gyrocotyle)
are closely related forms, or with Loénnberg (12), who is of opinion
that the Caryophyllaeidae differ essentially from the
Gyrocotylidae in being secondarily monozoic forms, the
Gyrocotylidae being, in his opinion, primarily monozoic.
I cannot find that either of these authors have advanced any
valid reasons for these opinions. It is true that Gyrocotyle
differs from all Caryophyllaeidae in possessing an
anterior sucker and a posterior ‘funnel’ (with muscular and
nervous modifications to match), but the extremities of these
animals can develop almost any kind of process, as we see in
the Caryophyllaeidae, and this difference, as also the
minor differences in the conformation of the genitalia and in

1 Both Spencer (24) and Hungerbiihler (9) found in certain species of
Gyrocotyle ten-hooked larvae similar to those of Amphilina,
while in G. urna Lonnberg (‘ Biol. Féren, Férhandl., Verhandl. d.
biol. Ver. in Stockholm ’, vol. ii, 2, p. 55, 1890) found larvae apparently
altogether devoid of hooks. The significance of these facts is as yet
obscure.

F. WOODLAND

WwW. N.

464

‘oeprlpeydooorryyog oy} ur pure
VIUIVPOISO) OY} JO SOl[TUIVT 9o1Y} OY} UL VITVZIUEDS OY} Jo UOTZISOdsIp pUv UOTZLYUOTIO OY} 9}eIYSHTIE OT,

OQlIWHd 3 JOIN HLOE OINITVHdIWY AIMVALOSOYAD Olav MIAHdOAYW OD

3
5
.
:
}
4
i
¥
:
x
i
‘
b
?

“A
t
i,

reson eerie

0 6? 00% 04s 20060,00% 222% 0%0%e%s

See et ate

2,
ed Pd

*‘DIU- LX OY

CESTODARIA 465

the excretory system, cannot be considered of much weight
when the fundamental resemblances I have indicated are taken
into consideration. I therefore propose, in order to emphasize
these fundamental resemblances, provisionally to unite the
Caryophyllaeidae with the Gyrocotylidae into one
Order, the Paralinidea, in contradistinction to the
Amphilinidea, which contains only one known family,
the Amphilinidae.
The Paralinidea may be provisionally defined as

Cestodaria (a) with the three sexual apertures situated
close together; (b) with the uterus and vagina running
parallel to each other in the same dorso-ventral plane
along the median longitudinal axis of the body ; (¢) with
the testes restricted to the anterior end of the body ;
(d) with a bilobed ovary ; (e) without calcareous bodies ;
(f) with a reticulate excretory system; (g) with a hexa-
canth larva (?).

The Amphilinidea possess the opposite characters, viz.
(a) the uterme and vaginal apertures are situated at
opposite ends of the body, the former being at the anterior
and the latter posterior ; (b) the uterus and vagina run
in opposite directions from the ovary, the former anteriorly
and the latter posteriorly ; (c) the testes extend, with the
uterus, over the greater length of the body; (d) ovary
one-lobed ; (e) calcareous corpuscles present ; (f) excretory
system consisting of two main lateral longitudinal channels
with minor regularly arranged looped vessels; (yg) with
a ten-hooked larva.

Finally, it is evident that of the two Cestodarian Orders—
Amphilinidea and Paralinidea—the latter is much
more Closely related to the Cestoda merozoa than is the
former (Text-fig.). In both the Paralinidea and the
Bothriocephalidae the uterus and the vagina run
medianly in the same dorso-ventral plane, the vas deferens
opens from the anterior side of the proglottis (though the testes

? Benham (1) would place Wageneria in this Order, but this organism
is not yet known with sufficient accuracy to justify its inclusion in any
family.

466 W. N. F. WOODLAND

are necessarily displaced from their primitive anterior position),
the three sexual apertures are all close together, the ovary is
bipartite, and a hexacanth larva is present.

I would not, however, go so far at present as to follow Lihe
(18) in actually grouping the Caryophyllaeidae with the
Diphyllobothriidae in his Order Pseudophyllidea,
because for me the monozoan character of the Paralinidea
is a fundamental one, and to me it is inconceivable that the
Paralinidea have become secondarily monozoan. It is also
well to remember that the Amphilinidea may not be so far
removed from the Paralinidea as their existing structure
would seem to indicate, for the simple reason that this structure
is, in part at least, an obvious adaptation to the mode of life
of Amphilina. Amphilina is parasitic in the body-
cavity and not the gut of the fish, and it can only liberate its
larvae to the external world by boring through the body-wall
of its host (hence the huge boring apparatus I have described
(28) and the extreme anterior position of the uterine pore)
and has to retain its larvae until it has penetrated to the
outside of the fish (hence its enormously elongated uterus—
a duct which would not need to be so elongated if, as in the
Paralinidea, the parasite could continually shed its eggs
into the intestine). These bionomic necessities must have
caused a great transformation in the entire anterior end of the
animal and especially in the arrangement of all the contained
ducts, and it is therefore possible that even the posterior
position of the vas deferens and the vagina (the openings of
which must remain adjacent) is to be regarded as but another
indirect concomitant of the general rearrangement. In other
words, the original plan of the genitalia in Amphilina is
possibly largely masked by specialization and, in the light of
present knowledge, it may be incorrect to conclude that
Amphilina is more separate genetically? from the Para-

1 The fact that in both the Paralinidea and Amphilinidea
ten-hooked larvae have been found may be called to mind in this con-
nexion. I may here remark that Sanguinicola (still included in 1916
by Cholodkowsky (8) in the Cestodaria) is now known not to be a Cestode :

CESTODARIA 467

linidea than these are from the merozoan Cestodes. How-
ever, I concede that it is possible that this conclusion may prove
to be the correct one and that, as several authors have
supposed, the ‘ Cestodaria ’ may be polyphyletic in origin.

On AN IMMATURE CARYOPHYLLAEUS SP. FROM THE INTESTINE
oF AUCHENOGLANIS OCCIDENTALIS, Cuv. AND VAL., 1840.

I possess a single specimen of a species of Caryophyllaeus
(Pl. 25, fig. 31) which is immature and which was found in the
intestine of the Siluroid Auchenoglanis occidentalis
at Khartoum, Anglo-Egyptian Sudan. The specimen measures
2mm. in length and about 0-5 mm. in maximum breadth
(posteriorly). The body is relatively thick and has the wrinkled
appearance shown in PI. 25, fig. 32. I cut this specimen into
horizontal longitudinal sections and discovered that its internal
structure was on the same general plan as that of C. fili-
formis. The ovary and vitellaria were still rudimentary,
though the uterus, vas deferens, and their openings were well
defined. The testes were not fully developed. Apparently the
vagina was convoluted in form and not straight as is usual,
and it was not easily distinguishable in sections from the uterus.

In conclusion, I wish to express my thanks to Dr. Andrew
Balfour, Director of the Wellcome Bureau, for the opportunity
of examining the interesting new forms of Caryophyl-
laeidae above described, to Professor J. P. Hill for the loan
of five specimens of Caryophyllaeus laticeps, and to
Mr. C. A. Hoare for his kind aid in translating certain passages
in Russian.

SUMMARY.

1. Four remarkable new species of Caryophyllaeidae
from the Anglo-Egyptian Sudan are described, three of which
(from Siluroid fishes) are referred to a new genus, Wenyonia
—W. virilis, W. acuminata, and W. minuta—and
the fourth (from Mormyrus caschive) to the original genus

vide my paper on this form in the present volume of the ‘ Quart. Journ.
Micr. Sci.’

468 W. N. F. WOODLAND

Caryophyllaeus—C. filiformis. The chief character-
istics of the new genus are the situation of the sexual apertures
in the anterior half of the body, and the elongated uterus.
The family Caryophyllaeidae, after deletion of the
genera Diporus, Monobothrium, and Glaridacris,
thus contains three genera—Caryophyllaeus, Archi-
getes, and Wenyonia—all of which are re-defined, with
their known species.

2. Very young immature forms of Wenyonia occur in the
same (Siluroid) host as the adult and are devoid of a * caudal ’
appendage, whence it would appear that the life-history of
these new forms is different from that of C. laticeps.

3. The three families of the Cestodaria—Caryophyl-
laeidae, Gyrocotylidae, and Amphilinidae—are
re-defined.

4. The Cestodaria, after eliminating Sanguinicola
(a Trematode ?), are provisionally grouped into two Orders :
the Amphilinidea (with one family, the Amphilinidae)
and the Paralinidea (with two families, the Caryophyl-
laeidae and the Gyrocotylidae), the latter being closely
allied to the Bothriocephalidae. These two Orders are
defined.

LITERATURE REFERENCES.

1. Benham, W. B., in ‘ A Treatise on Zoology’, ed. by E. Ray Lankester,
“Part IV. The Platyhelmia, Mesozoa, and Nemertini’’, London,
1901, p. 97.

2. Cholodkowsky, N. A.—‘‘ Notes helminthologiques ’’, “ Petrograd Ann.
Mus. Zool. Ac. Sci.’, vol. xx, 1915, p. 164.

‘Explanatory Catalogue of the Collection of Parasitic Worms
in the Zoological Cabinet of the Imperial Military Academy of
Medicine, Petrograd ’, fascicle ii, 1916.

4, Claus, C.—‘ Lehrbuch der Zoologie’, 1885.

5. Cooper, A. R.—‘‘ Glaridacris catostomi, gen. nov., sp. nov. : a Cesto-
darian Parasite’’, ‘Trans. Amer. Micr. Soc.’, vol. xxxix, no. l,
1920, p. 5.

6. Diesing, K. M.—‘‘ Revision der Cephalocotyleen. Abtheilung :
Paramecotyleen’’, ‘ Sitzungsb. der Math.-Naturw. Cl. kais. Acad.
Wiss. Wien’, Bd. xlviii, I. Abth., 1863, p. 200.

3.

DIFFERENTIAL INHIBITION IN OBELIA 481

8, and 4). The stream of cells and débris can also be observed
directly under the microscope while resorption is going on in
the living organism.

The questions remain, how and under what conditions do
the resorbed elements start the migration, and where do they
eventually get to?

TEXxT-FIG. 2.

Campanularia. End of stage 2. The hypostome has disappeared,
the tentacles are represented only by minute knobs, some almost
resorbed. Nematocysts here and there project from the surface of
the tentacles, one discharged. (Camera lucida.)

We will first study a tentacle. In the normal zooid the
tentacle is almost twice as long as the hypostome, and its axis
of endoderm is composed of large cells with very definite walls
and flattened like a pile of dises. The endodermal axis occupies
more than three-quarters of the diameter of the tentacle
(Pl. 26, fig. 6).

In the first stage of resorption the tentacle has shrunk con-
siderably both in length and diameter (PI. 26, fig. 7). The endo-

482 J. S. HUXLEY AND G. R. DE BEER

derm cells are smaller, and the most distal ones have lost the
typical disc-like structure. They are rounded and have lost their
serial arrangement in a single row: at the tip it is difficult to dis-

Trxt-Fic. 3.

Campanularia. a, Stage 3. No trace of tentacles or hypostome.
Considerable new growth from the stem. 6, The same zooid
magnified to show its change of shape. The dotted outline was
drawn one minute later than the firm outline. c, Tip of the
new growth from the same specimen in expanded condition.
d, The same specimen contracted two minutes later. The
ectoderm near the tip is attached to the perisarc ; proximal to
this the contraction is clearly visible. ect., ectoderm ; en., endo-
derm ; p., perisarc. (Camera lucida.)

tinguishthem from the ectoderm cells, which have come to present
the same spheroidal appearance. PI. 26, fig. 7,1s a photograph
of a single section showing progressive loss of differentiation

OO

ee

DIFFERENTIAL INHIBITION IN OBELIA 483

in the endoderm axis of a tentacle as one approaches the distal
extremity. This loss of differentiation gradually extends
proximally, and the whole tentacle passes into the gastro-
vascular cavity after rupture of the mesoglaea (confirming
Thacher, 1903). In Pl. 26, fig. 4, the cavity is seen to be filled
with cells and débris, the result of resorption of the tentacles.

TExtT-FIg. 4.

Campanularia. The distal part of the zooid has separated from the
proximal part. Active cilia occur in both parts, also immigrated
cells, but those in the proximal part are sparser as they are
migrating down the stem. Some cells of the distal part are
migrating outwards. (Camera lucida.)

Resorbed elements from the hypostome are no doubt there
also, although indistinguishable.

Examined under high powers, it appears that the boundaries
of the loose cells in the gastrovascular cavity lose more of
their distinctness the farther the elements are from the
periphery. The spherical or ovoid bodies to be seen in the
centre of the zooids containing refractive matter (but not
staining very definitely) are nematocysts. Nearer the periphery

484 J. 8S. HUXLEY AND G. R. DE BEER

the cells appear as in Pl. 26, fig. 8, still with a definite cell
boundary. Nematocysts only appear in the gastrovascular
cavity in late stages of resorption, after rupture of the
mesoglaea at the base of the tentacles. The bulging in of
the basal mesoglaea before breaking is well shown in Pl. 26,
fig. 3, left-hand tentacle.

Appearances sometimes occur in Obelia which seem to show

TExtT-FIc. 5.

Campanularia. End of stage 4. The zooid is represented only by
a small stalked knob containing a mass of pigment, mostly
brown with some orange granules. (Camera lucida.)

that phagocytosis is taking place. This was corroborated on
a specimen of Campanularia (the one shown in Text-fig. 6)
which was accidentally ruptured: numerous cnidoblasts with
contained nematocysts were then seen; but in addition,
nematocysts were found inside cells much larger than enido-
blasts ; frequently two would be seen within a single large
cell (see Text-fig. 7).

Studies in Dedifferentiation. IV. Resorption
and Differential Inhibition in Obelia and
Campanularia.

By
J. S. Huxley, M.A.,
and

G. R. de Beer, B.A., B.Se.

With Plate 26 and 7 Text-figures.

INTRODUCTION.

Ir is well known that among the processes whereby an
organism is developed all are not constructive, but some may
involve ‘the demolition of certain structures which may either
have ceased to subserve a useful function, or may actually
hinder further development. A familiar example is the gill
or tail of Amphibian larvae at metamorphosis.

In normal circumstances this retrogression only affects
certain organs, but recently it has been found possible to
produce this effect experimentally in whole organisms (Driesch,
1906, &c.).

Protozoa, Planarians, Sponges, Ascidians, and Coelenterates
will under certain conditions give the retrogressive effect, as
evidenced by the work of Lund (1917), Child (1904), Maas
(1910), Huxley (1921 b), Loeb (1900), and others.

Following some observational work by one of us (J.5.H.),
it was thought that quantitative experiments involving the
subjection of organisms to different concentrations of poisons
might give interesting results. One of the authors (G.R. de B.)
accordingly performed some experiments on Obelia geni-
¢ulata which will be deseribed below.

AT4 J. 8S. HUXLEY AND G. R. DE BEER

The Hydrozoa are not virgin soil to the experimentalist in
this connexion. Loeb (1900) observed that under certain con-
ditions the zooids of a hydroid colony would lose all their shape
and structure and retreat into the hydrocaulus. He attributed
the cause of this to contact with solid objects, viz. the watch-
glass in which the organisms were kept ; but this explanation
is probably not correct.

Thacher (1903) investigated the process of retrogression in
hydroids and called it ‘absorption’. Cerfontaine (1902) says
of it: ‘les individus ... dégénerent et disparaissent.’ Gast
and Godlewski (1903) merely called it ‘ degeneration ’ (‘ Riick-
bildungsprozess ’). These terms are inadequate to designate
a process as specific as that which Huxley (1921 b) has described
in the case of Perophora. Strictly speaking, there are two
processes at work, viz. dedifferentiation and resorption. In
the following description of the experiments performed Resorp-
tion will be used to mean the process whereby the material
composing the zooid is transported, and Dedifferentiation
to mean other processes undergone involving a return of cells
or tissues to a simpler, less differentiated condition.

In the organisms chosen for experiment there is a coéxistence
of two sets of systems, the * zooid systems’ and the ‘ stolon
systems’. It is obvious that normally physiological equilibrium
must exist between them. But if circumstances can be found
whereby one system is adversely affected more than the other
there will occur differential inhibition associated with resorp-
tion or dedifferentiation or both.

The experiments were performed at the Biological Laboratory,
Woods’ Hole, in 1916 (Campanularia, J.S.H.), and at the
Marine Biological Laboratory, Plymouth, in 1920 (Obelia,
G.R. de B.).

EXPERIMENTAL.

Given the fact that mere subjection to unfavourable con-

ditions, viz. being kept in glass vessels in the laboratory, brings

about resorption of zooids in hydroids, it was to be expected
that if the toxicity of the water were increased, the process of

DIFFERENTIAL INHIBITION IN OBELIA 475

resorption would be accelerated and the differential imhibition
made more specific. Apart from their plentifulness, Obelia
and Campanularia are suitable material because :

(i) The zooids are conveniently far apart and attached to

the hydrocaulus by a fairly long stem ;

(ii) The zooids in their natural condition are highly differen-

tiated structures compared with the rest of the colony ;

(iii) The stages of resorption can be conveniently determined

by reference to the hydrotheca.

On the other hand there is the disadvantage that it has
poor viability, which means that resorption takes place even
in the controls in clean sea-water. This, however, occurs
long after it has done so in the toxic solutions.

Care was taken to ensure that the colonies were clean and
healthy and free from Diatoms and Protozoa, and that all the
polyps were normal and fully extended.

The experimental solutions are referred to in terms of
concentrations of KCN; but since the solvent was sea and

N
not pure water such an expression as 10 does not give us the

actual ionic concentration.

The solutions were made up in shallow glass dishes, in each of
which a small number of stems of Obelia each bearing eight
zooids (Table I) or one zooid (Tables II to IV) were placed.
In every case the material used was fresh from the sea. KCN
solutions were kept covered owing to the volatile nature of
KCN, and changed every day.

Resorption can be divided into five stages (Text-fig. 1).
Text-fig. 1, a, represents a normal zooid (also PI. 26, fig. 1).

First Stage. The tentacles are first affected. They may
become apposed to the hypostome, and may shrink so
as to become shorter than the hypostome. Adjacent
tentacles may fuse (Pl. 26, fig. 9), indicating an interesting
change in consistency. The mouth closes, but the hypo-
stome is still prominent. Ciliary action continues in the
enteron as in the normal zooid (Text-fig. 1, b, and Pl. 26,
figs. 2 and 3).

476 ; J. 8. HUXLEY AND G. R. DE BEER

Second Stage. The hypostome is completely resorbed and
the tentacles are represented only by a ring of tiny prominences.
The ‘ waist ’ in the body of the zooid has disappeared. The
whole is well within the margin of the hydrotheca. The stalk

TEXT-FIG. 1.

Diagrams representing: a,normal zooid; 6, first, c, second, d, third,
e, fourth, f, fifth stages of resorption remaining in the hydrotheca.

attaching the zooid to the hydrocaulus is very thin and within
it is a mass of cells and débris flowing slowly away from the

zooid (Text-figs. 1, c, and 2).
Third Stage. The zooid has shrunk towards the bottom

DIFFERENTIAL INHIBITION IN OBELIA AT7

of the hydrotheca ; its shape is roughly ovoid. No sign of
tentacles whatsoever (Text-figs. 1, d, and 3, a, and Pl. 26,
fig. 4).

Fourth Stage. Sometimes the distal portion of the zooid
may become separated from the rest (Text-fig. 4). The form-
determining properties of the zooid have become less powerful
than the surface tension acting upon it, and it has accordingly
become spherical, nowhere touching the hydrotheca and
connected by a thin stalk to the hydrocaulus. At this stage
and later the flow in the tube is irregular. It appears to be
maintained by pulsations of the stolon (see p. 479) (Text-
fig. 1, e, and Pl. 26, fig. 5).

Fifth Stage. The process has been continued until the zooid
is represented only by a tiny knob (often containing pigment)
smaller in diameter than the hydrocaulus (Text-figs. 1, f,
and 5). Occasionally the process is carried further and the
hydrotheca becomes empty. ‘This only occurs a considerable
time after stage 5, and is mainly a mere degeneration effect.

It should be noted that in those cases where the colonies
contained gonothecae, medusae were not liberated if resorption
had started. During resorption the zooids are perfectly
healthy and transparent. Dead tissues can always be distin-
guished (opacity, &c.). Small masses of dense pigment are
often found in the partly resorbed zooid, representing products
of degeneration.

As resorption goes on, the material derived from the zooid
passes into the hydrocaulus, and from the proximal (cut) end
of the latter a stolon begins to grow (Text-fig. 8, a@ and c).
It is very transparent and clear, and may grow to the length
of 10 mm. or more, affixing itself to the substratum. It some-
times happens that a small portion of what is left of the zooid
in stages 3-5 is completely nipped off (by surface tension)
from the hydrocaulus. It remains in the hydrotheca and dies
(Text-fig. 4).

In the earlier experiments the solutions used were too strong,
but even so a differential effect was obtained (Table I).

478 J. 8. HUXLEY AND G. R. DE BEER

TABLE [ (Obelia).

Strength of

KON.
N/2,000 Dead

|
N/4,000 Dead
N/8,000 Dead
|
N/16,000 Dead
|

N/32,000 I II Dead

N/64,000 I II
|

Controls Quite healthy I
Sea-water | |

3 To 1b 90 23 “Squheeem

The Roman figures denote the stage of resorption at a parti-
cular time in a solution of given strength. All solutions up to
N/32,000 are too strong and kill the organisms before any
resorption occurs: in the N/32,000 solution resorption pro-
ceeded a little way but the organisms were then overcome by
the poison.

It soon became apparent that the preparations used were not
wholly satisfactory, for of the eight zooids borne by the colonies
all had not been resorbed to the same extent at the same time.
To meet this difficulty portions of the hydrocaulus were used
bearmg only one well-expanded zooid.

TaBLE II (Obelia).

RE fter : 6 hours. 10 hours. 24 hours. 36 hours.
In

N/8,000 | All dead

N/16,000 | All dead

N/32,000 | Stage I All dead
N/64,000 | Trace of resorp- | Stages I and | All dead
tion i
Control Expanded and | Expanded and| Trace of resorp- | Stages I

motile motile tion and IT.

The results of this experiment were much more definite,
but the solutions used were still too strong.

DIFFERENTIAL INHIBITION IN OBELIA 479

TasBLE IIL (Obelia).

After: | Whours. | 24 hours. 40 hours. | 50 hours,

N/64,000 | Stage II. No me-| Dead
dusae liberated
N/128,000 | Stage I. No me-| Stages II and | Dead
dusae liberated | III
N/256,000 | Trace of resorp- | StagesIT and II | Stages III | Stage V and

tion. A few and IV Dead
medusae_liber-
ated

Controls | Fully expanded. | Traces of resorp- | Stages I and| Stage III.
Several medu-| tion IEE
sae liberated

Experiments in the N/256,000 solution were repeated several
times with the same results. These results indicate that a
N/256,000 solution of KCN inhibits the zooids without affecting
the hydrocaulus to any appreciable extent (at least for a con-
siderable time—fifty hours). The healthy nature of the
hydrocaulus is evidenced by movements of contraction and
pulsation, and by growth at the proximal extremity.

The contraction of the stem is of interest, since precisely
similar contraction occurs in Perophora and other Ascidians
(Huxley, 19216). In addition, the partly dedifferentiated
zooid also appears to contract at intervals (Text-fig. 3), although
it is possible that the contraction is a mere surface-tension
effect, exerted passively on relaxation of the walls of the stem.
It appears that the contraction of the cells of the stem occurs
when considerable internal tension has been produced through
the flow of liquid and cells from the zooid. In higher forms
embryonic cells which are destined to give rise to muscle appear
to start contracting before differentiation, also as a result of
tension (e.g. Carey, 1921 aand 1921 B, &c.).

It is possible that contraction produced by tension has no
normal function in hydroid stems. It is all the more interesting
to find that it occurs, being thus probably a general property
of all not too highly-differentiated cells. Possibly tension acts
also as a stimulus to outgrowth from the stem. (Cf. the well-
known fact of better growth in regenerating Tubularia in
hypotonic sea-water (Loeb, 1892).)

480 J. 8. HUXLEY AND G. R. DE BEER

In order to see whether any of the results obtained were due to
the specific effects of KCN, experiments were also made with
HgCl,. When solutions of N/1,000,000 and N/2,000,000 were
used, the resorption effects were identical with those in KCN.

TaBLE IV (Obelia).

After 24 hours.

E N
nA 128,000. - |Stages II and III.
N
256,000. - | Stage I.
HgCly za , :
3 10) 3 . | Half dead, remainder Stage ITI.
N
Cea . | Stage I,
Control 4 - | Fully expanded.

After varying periods in solutions of all the strengths, some
preparations were removed and placed in clean fresh sea-water
with a view to inducing them to cease resorption and reform
zooids. In no ease was this successful. The preparations ceased
resorption for a short time, but then continued. This was to
be expected from the behaviour of zooids in normal sea-water.

When the zooid is severed from the stem at the base of the
hydrotheca, the preliminary dedifferentiation occurs as usual ;
but after a certain number of cells have migrated into the
interior there is no room for more. The result is an ovoid
dedifferentiated mass, tightly packed with cells (Text-fig. 6).
Similar phenomena were seen in Perophora. Thus the degree
of resorption depends on the amount of space available. Resorp-
tion will only proceed to a limit when the migrating cells are
removed. A parallel is here provided with those chemical
reactions which will only proceed to a limit if the products of
the reaction are removed.

HistToLoey.

In the previous section the various stages of resorption were
briefly described. The actual route of migration of the cells
is, of course, through the gastro-vascular cavity (Pl. 26, figs. 2,

DIFFERENTIAL INHIBITION IN OBELIA 485

Presumably the large cells are endoderm cells which still
retain the phagocytic properties normally associated with
intracellular digestion.

Occasionally we have noticed in sections of normal specimens
appearances indicating the phagocytosis of small (presumably

TrExt-Fic. 6.

Campanularia. Zooid cut off at base of hydrotheca. Dedifferentia-
tion to an ellipsoid mass. Many nematocysts were to be seen in
the interior. (Camera lucida.)

TEXT-FIG. 7.

Diagram showing the manner of occurrence of nematocysts in
the interior of the zooid shown in Text-fig. 6. (Observations
in vivo.) Thenematocysts were loose, or still in their enidoblast,
or ingested in a large endoderm cell, or two inside one endoderm
cell.

interstitial) cells by normal endoderm cells. Professor G. C.
Bourne, F.R.S., has kindly informed us that he has seen
similar appearances in Hydra, which he also interprets as cases
of phagocytosis of one type of cell by another. This process,
however, is clearly much commoner in the specimens under-
song resorption than in normal zooids.

It is in any ease difficult to see on what occasions phago-

NO. 267 Kk

486 J. 8S. HUXLEY AND G. R. DE BEER

cytosis of the organism’s own nematocysts would occur in its
normal life-history. It would thus appear that im certain
circumstances the power of ingesting food-particles, possessed
by certam types of cells, results in typical phagocytosis of
other cells of the organism—a process unusual or abnormal
for these forms, but usual and normal in higher animals.

Loeb (1900) states that the tentacles fuse in some Cases.
Thacher (1903) considers this to be only the appearance caused
by their bemg crowded together and by the ectoderm being
thrown into folds by excessive contraction. After careful
study of serial sections of a zooid in this condition we can state
that this explanation will not suffice. The tentacles are, it is
true, contracted and crowded, but there is actual fusion in
several places (Pl. 26, fig. 9).

The ectoderm cells of the tentacles and hypostome become
more cubical during resorption.

Of the endoderm cells, the large mass of glandular cells in
the hypostome very soon disappears, the cells passing into the
cavity.

In the last stage of resorption before the hydrotheca is
evacuated altogether (stage 5), what is left of the zooid is still
bounded by a definite epithelium (PI. 26, fig. 8) of flattened cells,
beneath which are others losing their differentiation. The
behaviour of the ectoderm cells may be compared with that of
a rear-guard, continually retreating yet always maintaiming
an unbroken front. But as the zooid is resorbed, its volume
and surface decrease, and so the front diminishes.

At the start the cavity is not too congested and the cells and
débris pass down; but in later stages the cavity is almost
blocked up (PI. 26, fig. 8), and it is then that pulsation can be
observed. This obviously facilitates the evacuation. It 1s
presumably due to tension on the walls of the stolon.

In Obelia it is possible to observe the cells actually leaving
the tissues, a process which has often been taken for granted
in other forms (see Pl. 26, fig. 2, left side).

Where do the resorbed elements go when they pass down
into the hydrocaulus out of the zooid ? It appears that they

DIFFERENTIAL INHIBITION IN OBELIA 487

are ingested by the endoderm cells of the hydrocaulus, or break
down and in that condition are absorbed by those cells.

The walls of the hydrocaulus even in the last stages of resorp-
tion appear normal when seen alive under the microscope and
in sections, though its cavity may be filled up with cells and
débris.

But is it possible that the new growth which takes place at
the proximal cut end of the hydrocaulus consists of the very
elements derived from the resorbed zooids ? Loeb makes-the
observation that this growth is like the motion of a protoplasmic
mass, and such it certainly appears to be in our experiments.
But this would be the appearance presented by normal growth
proceeding at the rate at which this stolon was produced—
10 mm. in forty-eight hours or less.

In structure the new growth is similar to an ordinary piece
of hydrocaulus (‘Text-fig. 3).

This growth starts only after a certain stage of resorption
has been reached.

The new growth adheres to the substratum, thus resembling
the normal creeping stolon. We were unable to observe whether
it could give rise to buds in Obelia, as the preparations died.
In Campanularia, pieces with several zooids might give rise
to one or several buds during or after resorption of the original
zooid.

DISCUSSION.

Loeb (1900) attributes resorption to contact with solid
objects. According to him the transformation must be due to
liquefaction of the more solid constituents of the zooid. Con-
tact with the fluid, sea-water, makes for the production of the
more solid portion of the colony, the zooid; whereas con-
versely contact with a hard surface makes for the more fluid
stem. Accordingly if a zooid be subjected to the stimulus to
which the production of the stem-system is the reaction, the
result will be the conversion of the zooid into something
resembling the stem.

But resorption of zooids takes place even when the colony
is maintained in an erect position and no portion of the zooid

K ke

488 J. S. HUXLEY AND G. R. DE BEER

is allowed to touch any hard object (Thacher, 1903). Clearly
then, contact cannot be the only cause of resorption. It is
rather to be interpreted in terms of equilibrium between two
systems with different physiological reactions : if one wishes
to use Child’s phraseology one may say that they possess
different metabolic rates, and the one with the higher rate is,
normally, physiologically dominant over the other.

In the case of these hydroids there are two systems, the zooid
and the stem (hydrocaulus).

Normally, the more highly differentiated zooid is able to
maintain itself, but bemg more specialized and less plastic
than the stem the Zooid will not be able to maintain itself in
the face of conditions which, though adverse for the zooid,
do not appreciably affect the stem. Such are, e.g., a N/256,000
solution of KCN, or laboratory conditions after fifty hours.
The result of the adverse conditions is inhibition of normal
function ; and within limits it is differential, affecting the zooid
before and more than the stem.

By interfering with general metabolism, as is done by exposure
to toxic agencies, the output of energy is reduced. Energy is
needed to maintain differentiated form against surface-tension.
Thus one of the first results of non-lethal interference will be
the loss of typical form by cells and their reversion to a
spheroidal or cuboidal shape. ‘This is found im all cases of
dedifferentiation known, and often leads to the assumption of
spheroidal form by the whole organism (see Huxley, 1922).

The fact that exposure to laboratory conditions, to KCN and
to HgCl,, all bring about identical reactions indicates that the
effects of the poisons, &¢., are not specific, but that all act mm
a general way, by affecting the energy-production of the tissues.

It may be asked how the process we have called dedifferentia-
tion in Obelia differs from simple degeneration. The answer
is to be found in observation of the process. At no time can
it be said that the zooid is dead ; during the whole process of
resorption what 1s left of it is just as alive as the normal zooid
orstem. Ifthe zooid dies, as it does if the poisons are too strong,
the cells acquire a characteristic semi-opaque appearance

DIFFERENTIAL INHIBITION IN OBELIA 489

which cannot be mistaken. There is then no more resorption
and the cells macerate, and later disintegrate, without dedif-
ferentiation.

Resorption is a result of the process of migration and it
could not take place were the elements to be resorbed to remain
in their differentiated condition. Resorption then is consequent
on dedifferentiation. It occurs in many forms when a cavity
is present into which the migration may occur (Child, 1904 ;
Huxley, 1921 b).

In Obelia, as in Perophora and probably m many other
cases, if the cavity into which migration can occur be by some
method or other limited, dedifferentiation with no or slght
resorption may take place (Text-fig. 6, p. 485).

It is then dedifferentiation plus resorption that Loeb
means by ‘liquefaction’. But this is not a specific result of
contact with hard surfaces. Whether such contact by itself
can produce the effect we do not know, but as an unfavourable
condition it can and does accelerate it. Contact stimulates
tentacles to contract, and constant stimulation must be
unfavourable ; oxidation must also be reduced in proximity
to the substratum. Loeb’s analogy between the liquefaction
of the zooid and the clotting of blood, both due to contact with
solid objects, thus cannot stand.

It may be said that dedifferentiation implies potential sub-
sequent redifferentiation. There is, however, no reason why
dedifferentiation should be reversible any more than differen-
tiation. If we lay down that dedifferentiation is reversion to
a morphologically simpler state with lower energy-requirements,
the simpler condition being preserved for a considerable time
and not merely a stage in the process of dying, we have
a good working definition.

Dedifferentiation is the accepted term to denote the simplifi-
catory processes undergone by differentiated cells in tissue
culture ; and in this case there is usually no redifferentiation,
the tissues merely remaining alive for a longer or shorter time
in their simplified condition. Smooth muscle grown in culture
solutions dedifferentiates to a condition in which the cells divide

490 J. S. HUXLEY AND G. R. DE BEER

actively (no division normally occurs in adult smooth muscle).
If such a preparation be grafted back into its former position
it is just possible to arrest the mitoses, but redifferentiation
proceeds no further (Champy, 1913). (Strangeways mforms
us verbally that he has been more successful.) On the other
hand, redifferentiation of dedifferentiated tissue has been
obtaimed by Drew (1923) in vitro.

Redifferentiation of the zooid was not obtained in Obelia as
it was in Clavellina (Driesch, 1906; Huxley, unpub.), Pennaria
(Cerfontaine, 1902), or Sycon (Huxley, 1911). But in those
cases where redifferentiation does occur, we must ask whether
the new adult structure is formed from the redifferentiation of
the dedifferentiated cells, or from indifferent cells which have
all along retained the full hereditary potentialities. The study
of budding and asexual reproduction, especially Hadzi’s work
(1910) on Hydra, suggests that the latter is usually the case.
If this is so, then the failure to redifferentiate is in no way due
to the inactivity of the dedifferentiated elements, but of the
indifferent cells. Vandel (1921), however, shows that in
regenerating Planarians (Polycelis), the new pharynx is pro-
duced from cells of other organs which dedifferentiate and then
redifferentiate along new lines, scarcely any mitoses being
observed. This is a good example of pluripotent dedifferen-
tiation (Adami and Macrae, 1914; Huxley, 19214). The usual
process, however, in colonial forms, is for the dedifferentiated
tissues to provide material for new outgrowths of the nature
of stolons, from which later new zooids may arise (ef. Ascidians,
Huxley, 1921b; Hydroids, Miller, 1913).

When the metabolic requirements of the zooid have decreased,
the equilibrium between it and the stem is upset and the
balance is now in favour of the latter (Huxley, 1921b). In
the higher animals complete resorption of systems does not
usually occur. For one thing the cells are too solidly packed
in tissues, and they are usually attacked by phagocytes before
they have even had time to be resorbed. But a concurrence
of both processes is seen in the absorption of the tail in Ascidian
tadpoles. Here, according to Delage and Hérouard (1898), * ses

DIFFERENTIAL INHIBITION IN OBELIA 491

éléments se désagrégent,’ after which process the phagocytic
action commences. Most authors are also agreed that a process
of dedifferentiation initiates the resorption of the tail in Anuran
tadpoles, phagocytosis being secondary (cf. Naville, 1922).

Phagocytosis here only occurs after resorption, 1.e. after the
tissue elements have migrated from the tissues. It would appear
not to be a normal process in Hydroids, but to be a result of
(a) the power of the endoderm cells to ingest solid particles,
(b) the presence of abnormally situated cells which have migrated
out of the tissues in the neighbourhood of the endoderm cells.
Phagocytosis of this nature appears to occur both in the case
of emigrated endoderm cells of the zooid, and of normal endo-
derm cells in the walls of the stem.

Resorption (as a result of emigration of dedifferentiated cells)
may be regarded as the most primitive method of eliminating
tissues in Metazoa. Even at the outset it may be secondarily
accompanied by a low form of phagocytosis. Later the
function of phagocytosis is assigned to special cells, and the
dedifferentiating tissues are attacked by these at a much earlier
stage in the process. The limited extent of phagocytosis in low
forms is also to be seen in Planarians (Vandel, 1921).

Resorption is in the first instance a direct result of exposure to
unfavourable agencies, but may be utilized later as a method of
accomplishing normal processes of the life-history. This appears
to be the case in Echinoderm metamorphosis (Huxley, 1922).

The stimulus in the case of Obelia is a certain concentration
of toxic products in the water. It is the same stimulus which
causes dedifferentiation in Clavellima and Perophora, and also
in Echinoderm larvae, Planarians, Sponges, and Protozoa
(Lund, 1917).

Hunger is another stimulus which may cause dedifferentia-
tion; and of course may act differentially. Dedifferentiation
caused by hunger has been found in Hydra (Schultz, 1906),
Echinoderm larvae (Runnstrém, 1917), &¢. Starvation will
again act by interfering with general metabolic processes. As
an example of the differential action of hunger, it may be
mentioned that in starved tadpoles, localized dedifferentiation

492, J. S. HUXLEY AND G. R. DE BEER

often takes place in one or more places on the tail (unpublished
observations, J.S.H.). Dedifferentiation is not, however,
followed by resorption in this case, at least before death.

Miiller (1918, 1914) has conducted an elaborate series of
experiments with various species of Hydroids. He finds that
dedifferentiation and resorption may occur not only in
hydranths but also in gonophores and in portions of hydro-
caulus. There is a delicately balanced equilibrium between
various parts of a system; in a compound system, whether
gonophore, hydranth, or stem shall be resorbed depends
(a) on the relative sizes and (b) on the ages of the sub-systems
(cf. Perophora, Huxley, 1921 b). Wounds will induce gono-
phores to dedifferentiate and be resorbed.

The quantitative action of poisons in accelerating dedifferen-
tiation and resorption, and the fact that in severed zooids
dedifferentiation may proceed independently of resorption, are
points on which we would like to lay stress.

SUMMARY.

1. Confirmation is given of the results of Loeb, Thacher,
Godlewski and Gast, and others, in showing that the hydranths
of hydroids (in this case Obelia and Campanularia) when
exposed to unfavourable conditions proceed to dedifferentiate
and to be resorbed, wholly or mainly, into the stem.

2. Exposure to toxic agencies accelerates the process. Too
great concentration of poison kills the zooids before dedifferen-
tiation starts. Below the death-point, the acceleration is
proportional to the concentration.

3. The effect is non-specific, both KCN and HgCl, producing
the same result as prolonged exposure to laboratory conditions.

4. When zooids are separated from the stem, resorption is
impossible. Dedifferentiation, however, proceeds until an ovoid
undifferentiated body packed with cells is produced.

5. The tentacles are first affected, then the hypostome. In
early stages, separate tentacles may fuse locally. Stumps of
tentacles are, however, still present after the hypostome has
quite disappeared. The body becomes ovoid, then spherical,
and is finally reduced to a minute pigmented dot.

DIFFERENTIAL INHIBITION IN OBELIA 493

6. The surface tension of the dedifferentiated zooid causes
the emigrated zooid cells to flow into the stem. In later stages
spontaneous pulsations of the stem and of the zooid (these
possibly not spontaneous) occur.

7. Dedifferentiation of the tissues of the tentacles starts at
the tip. Progressive histological dedifferentiation of the endo-
derm cells can thus be clearly followed in a single section.

8. Only after the mesoglaea at the base of the tentacle has
ruptured can the contents be resorbed (confirming Thacher).

9. Cnidoblasts with nematocysts can be distinguished within
the gastrovascular cavity as resorption proceeds. They may
also be seen phagocytosed within large cells, presumably
immigrated endoderm cells.

10. The dedifferentiation is regarded as due to interference
with general metabolic processes, and especially with the pro-
duction of the energy needed to maintain form against surface-
tension.

11. Resorption is regarded as the natural result of dedifferen-
tiation when there are adjacent cavities into which the cells
can migrate. In higher forms it has been largely replaced by
phagocytosis.

EXPLANATION OF PLATE 26.

Acknowledgements are due to Mr. Chesterman, of the Anatomy Depart-
ment, Oxford, for assistance in the preparation of the microphotographs.

Obelia geniculata.

Fig. 1.—Longitudinal section through a normal zooid. x 100.

Fig. 2.—First stage of resorption. 115.

Fig. 3.—First stage of resorption (slightly later than Fig. 2). x 130.

Fig. 4.—Third stage of resorption. x 146.

Fig. 5.—Fourth stage of resorption. x 240.

Fig. 6.—A tentacle of a normal zooid showing the differentiation of the
endoderm. x 300.

Fig. 7.—A tentacle of a zooid in the first stage of resorption showing
beginning of dedifferentiation in the distal endoderm cells. Also note
especially the reduction in width of the endoderm cells. x 420.

Fig. 8.—Cells and nematocysts in a zooid in the fourth stage of resorp-
tion. x 420.

Fig. 9.—Fusion of tentacles. First stage. 270.

494 J. 8; HUKDLEY AND 1G, 22) (DEB VBEER

List oF LITERATURE CITED.

Adami and Macrae (1914).—‘ Text-book of Pathology’, London, 1914
(pp. 318-24).

Carey (1921a).—‘ Am, Journ. Anat.’, 29, no. 1.

(1921 b).—Ibid., p. 341.

Cerfontaine, P. (1902).—‘‘ Recherches expérimentales sur la Régénération
et l’Hétéromorphose chez Astroides calycularis et Pennaria cavolini”’,
‘ Arch. de Biol.’, 19.

Champy (1913).—‘ Arch. Zool. Exp.’, Notes et Rev., 53, 42; and
‘Comptes-rendus Soc. Biol.’, 85.

Child, C. M. (1904).—‘‘ Studies on Regulation, III”, ‘ Arch. Entw.-
Mech.’, 17.

Delage, Y., and Hérouard, E. (1898).—‘ Zoologie concréte’, tome viii,
p. 153.

Drew, A. H. (1923).—‘‘ Three lectures on cultivation in vitro”,
‘Lancet’, 204, and ‘ Brit. Journ. Exp. Path.’, 4.

Driesch (1906).—‘ Arch. Entw.-Mech.’, 20, p. 21.

Gast and Godlewski (1903).—‘‘ Ueber die Regulationserscheinungen bei
Pennaria cavolini’’, ibid., 16.

Hadzi, J. (1910).—‘‘ Die Entstehung der Knospe bei Hydra”, ‘ Arb.
Zool. Inst. Wien’, 18.

Huxley, J. 8. (1911).—‘‘ Some phenomena of Regeneration in Sycon, &c.”,
‘Phil. Trans. Roy. Soc. B.’, 202.

—— (1921 a).—“‘ Further Studies on Restitution bodies, &c.”, ‘ Quart.
Journ. Mier. Sci.’, 65.

—— (1921 b).—‘‘ Studies in Dedifierentiation, II’ (Perophora), ibid., 65.

—— (1922).—“‘ Dedifferentiation in Echinoid larvae, &c.”’, ‘ Biol. Bull’,
1922. (‘Studies in Dedifferentiation, LIT.”’)

Loeb, J. (1892).—‘* Untersuchungen zur physiologischen Morphologie der
Tiere II. Wiirzburg.

—— (1900).—* Transformation and Regeneration of Organs’”’, * Amer.
Journ. Phys.’, 4.

Lund, J. (1917).—‘‘ Reversibility of Morphogenetic processes in Bursaria ”’,
‘Journ. Exp. Zool.’, 24.

Maas, O. (1910).—‘‘ Uber Involutionserscheinungen bei Schwimmern”’,
‘Festschr. z. 70. Geb. R. Hertwig ’.

Miiller (1913 and 1914).—‘ Arch. Entw.-Mech.’, 37, p. 319, and 38, p. 283
(Hydroids).

Naville (1922).—‘‘ Metamorphic Changes in Tadpole’s Tail’, * Arch.
Biol ya2,04

Vol. 67, N.S., Pl. 26

Quart. Journ. Micr. Sci.

a ive

“*
Nas
spe
ae

DIFFERENTIAL INHIBITION IN OBELIA 495

Runnstrém, J. (1917).—** Analytische Studien iiber die Seeigelentwick-
lung’’, ‘ Arch. Entw.-Mech.’, 47, p. 223.

Schultz (1906).—Ibid., 21 (Hydra).

(1907).—Ibid., 24 (Clavellina).

Thacher, H. F. (1903).—‘‘ Absorption of the Hydranth in Hydroid
Polyps’, ‘ Biol. Bull.’, 5.

Vandel (1921).—‘ Comptes-rendus Ac. Sci.’, 178, p. 1614 (Polycelis).

The Cleavage of the Egg of Lepidosiren
paradoxa.

By

Agnes E, Miller, M.A,,

Department of Zoology, University of Glasgow.

With 12 Text-figures.

THE process of cleavage in Lepidosiren was first described
by Professor Graham Kerr in his paper on the development
of the external features of this animal in the ‘ Philosophical
Transactions of the Royal Society ’, B, vol. excii, 1900. Since
the date of Professor Kerr’s expedition additional material has
become available from which it has been possible to make up
a more completely graduated series of segmentation stages.
The following paper deals with this material and may be
regarded as supplementing Professor Kerr’s paper. In referring
to the various stages I use the same numbers as Professor
Kerr used in the paper already mentioned and also in his
‘ Normentafel ’." |

Stages 2—3.—Professor Kerr’s earliest cleavage stage (2)
showed the first meridional furrow bisecting the finer-grained
apical cap but not extending beyond its margin; while in
the second stage figured by him (8) the first meridional furrow
had spread down to the equator of the egg, while the second
meridional furrow at right angles to the first extended to just
beyond the margin of the apical cap.

Amongst the new material is an egg (Text-fig. 1) in which
the first meridional furrow has extended right to the abapical

' Keibel’s ‘Normentafeln zur Entwicklungsgeschichte der Wirbeltiere ’,
Zehntes Heft, 1909.

NO. 268 L |

498 A. E. MILLER

pole while the second furrow has not yet made its appearance.
Another egg shows the first meridional furrow extending
nearly to the abapical pole and a very faint second furrow

Trxt-rig. 1.

a. View of apical hemisphere. b. View of the abapical hemisphere.

TExt-Fic. 2.

Apical hemisphere.

intersecting it and extending as far as the margin of the
apical cap. A third egg (Text-fig. 2) is somewhat peculiar,
inasmuch as the first meridional furrow shows a break in
continuity in the region of the apical pole.

Three other eggs of the same stage of development agree
in the main with those already described, except that the

CLEAVAGE OF LEPIDOSIREN 499

first furrow is not strictly meridional but is displaced outwards
some little distance from the abapical pole.

On the whole the evidence of the seven eggs obtained of this
stage goes to indicate that the normal procedure of the egg of
Lepidosiren is that the first meridional furrow is com-
pleted before the second furrow at right angles to it begins to
make its appearance.

Stages 3—4.—Five eggs belong to the commencement of
this period during which the egg is normally sub-divided by the
completion of the first and second meridional furrows into four

TEXT-FIG. 3:

a. Apical view. b. Abapical view.

approximately equal parts, and the third set of furrows—
vertical—make their appearance. In two eggs out of the five
the first and second meridional furrows are alone present and
perfectly normal. Two other eggs exhibited a feature men-
tioned in an earlier stage, the continuity of the furrows being
very faintly marked and in some places quite obliterated.
In one of the two eggs one of the furrows in the region of the
apical pole shows quite faintly marked; in the other egg
(Text-fig. 3) in which neither the first nor the second furrows
had quite reached the abapical pole, the continuity of the
first meridional furrow is broken at one side and again at the
abapical pole.

The last specimen of this particular stage shows the tendency

tle

500 A. E. MILLER

before mentioned of displacement outwards of the meridional
furrow from the abapical pole.

Towards the end of this period the greater rapidity of the
segmentation at the apical pole becomes more marked, and the
four vertical furrows make their appearance.

Four eggs at about stage 4 were available and two showed
similar development, two of the new furrows showing the
normal ‘ vertical’ position and maintaining a vertical course,
while the other two furrows, instead of bemg vertical, have
their point of origin displaced towards the apical pole so that

TExtT-Fic. 4.

Apical view.

they have become meridional, bisecting two of the quadrants.
The study of sections show that all the eight nuclei are im the
metaphase stage of mitosis; therefore, up at least to this
stage, mitosis is synchronous throughout the egg. This paper
is not intended to deal with nuclear phenomena, but it will
be of interest to show an accurate drawing (Text-fig. 5) of a
nucleus from the segmenting egg of Lepidosiren as illustra-
ting the extraordinarily favourable nature of these nuclei for
cytological investigation.

In the egg shown in Text-fig. 6 the two meridional furrows
are complete. On one side of the egg (that shown in the upper
half of Text-fig. 6) two vertical furrows are present, exactly
in line with one another so that they have the appearance of

CLEAVAGE OF LEPIDOSIREN 501

one continuous furrow cutting across the meridional furrow ;
while on the other half of the egg (Text-fig. 6, lower half) two
small furrows, likewise vertical, are just beginning to develop,

TEXT-FIG. 5.

Nucleus of segmenting egg of Lepidosiren in early stage of

mitosis. The nucleus was traced in outline under a Zeiss 1”
homogeneous immersion objective with ocular 18, and then
worked up in detail under 3mm. apochromatic homogeneous
immersion ‘objective with Bitumi binocular eye-piece. An
attempt has been made to bring out the stereoscopic relief of the
original by shading the nucleus as if it were isolated.

The divisions of the scale represent hundredths of a millimetre.

taking their origin however from the other meridional furrow
than that from which the first two started.
In the fourth egg at this stage the segmentation furrows at

502 A. E. MILLER

the apical pole are for the most part obliterated, although the
fact that down at the abapical pole the two meridional furrows
quite visibly intersect indicates that the egg belongs to this
stage.

Stages 4-5.—Altogether six eggs were found to be
approximately at this stage, and five gave evidence of further
irregularities in development. One which belongs to the
beginning of this period (Text-fig. 7) showed, again, lack of
continuity of one of the primary meridional furrows in the
region of the apical pole. ‘Two secondary furrows arising

TExT-FIG. 6. TEXT-FIG. 7.

Apical view. Apical view.

from the same meridional furrow (see upper half of Text-fig. 7)
proceed in a latitudinal direction and reach the other meri-
dional furrow, while of the corresponding furrows of the opposite
side one is actually vertical in direction (see left-hand furrow
in lower half of Text-fig. 7), and the other, assuming a rather
latitudinal direction, runs into the meridional furrow of that
side,

A second egg similar to the one already described showed
again the discontinuity of the meridional furrows ; neither
at the apical nor the abapical poles is there any sign of inter-
section. Vertical furrows appear in one-half of the egg (lower
half of Text-fig. 8, a), and in the other (upper half in Text-
fig. 8, a) a short latitudinal groove grows out from the meri-

CLEAVAGE OF LEPIDOSIREN 503

dional furrow as seen on the right side of the figure. Before
this furrow has progressed very far another furrow * (Text-
fig. 8, a)* grows down from it towards the abapical pole
bisecting the quadrant; on the other side of that same
meridional furrow there is an indication of a similar latitudinal
furrow (* *),

The third specimen showed practically the same arrangement
and development of furrows as the egg shown in fig. 5 of
Professor Kerr’s ‘ Philosophical Transactions’ paper. One

TEXxT-FIG. 8.

* * —e *

a. Apical view. b. Abapical view.

meridional furrow is rather displaced, and one of the furrows
of the third set growing from it is definitely latitudinal (Text-
fig. 9, right lower quadrant).

To quote one more example illustrative of irregularity of
segmentation: in the egg represented in Text-fig. 10 one of
the two primary meridional furrows (the one which is horizontal
in the figure) is perfectly normal but the other shows a distinct
break at the apical pole, the portion next the apical pole in one
hemisphere (that which is above in Text-fig. 10) having under-
gone a distinct displacement. As seen in the figure the amount,
of this displacement increases as the apical pole is approached.

1 Which is however in this case uncomplicated by any branching or
distortion.

504 A. E. MILLER

The fifth specimen, illustrated by Text-fig. 11, is interesting
mainly from the fact, not shown in the drawing, that one of
the meridional furrows is markedly displaced outwards from
the region of the abapical pole.

TEXT-FIG. 9. Trext-Fric. 10.

Apical view. Apical view.

Texoeme. 1. Text-Fic. 12.

Apical view. Apical view.

The last eg¢ belonging to this period of development (Text-
fig. 12) illustrates very clearly the kind of irregularities in the
position of the furrows which do so much to obscure the
process of segmentation from now on. In the right half of the
figure the two furrows which would normally be vertical

CLEAVAGE OF LEPIDOSIREN 505

(¥ 2and ¥ 3) are seen to be displaced in each case so that their
outer ends reach one of the meridional furrows, and in the case
of one of them the furrow has become practically latitudinal.
Owing to the increasing frequency of such displacements
combined with increasing ‘loss of step ’ of the various blasto-
meres in their successive fissions, the regularity of the segment-
ing process becomes from now onwards completely obscured.

In conclusion I desire to thank Professor Graham Ixerr
both for providing me with the material on which these notes
are based, and also for his kind supervision during the course
of the work. I should like further to thank the Carnegie Trust
for providing the illustrations to this paper, and Mr. A. Kirk-
patrick Maxwell for the care and skill with which he has carried
them out.

os 4 ‘. FY SORTA

a oe viutt + bs nciepe pry ' ' wia ity Lig it oe qe orm ae »4

ire yh (> Hiis ey owed Walb ) ie
iy ie hy At Agel yeety ret Viiek ay wiees ah eta c
7 ' ‘fe 4 P ' i Di “) sti P| vets
| sce MEE Wee? icles toh talk” taieegern
‘ : >

= wf
> ’ 7 e414 Ciel y 1) FG
\. ’ } .
’
y
i ‘
: - i bee
;
,
’
€
:
f
.
i
4 =
oe 4
V

The Morphology of the Nudibranchiate Mollusc
Melibe (syn. Chioraera) leonina (Gould)

H. P. Kjerschow Agersborg, B.S., M.S., M.A., Ph.D.,

Williams College, Williamstown, Massachusetts.

IV.

By

With Plates DF 958) Bele

CoNTENTS.

INTRODUCTION
ACKNOWLEDGEMENTS
ON THE STATUS OF CHIORAERA G@ounn

MELIBE LEONINA (Ss. CHIORAERA LEONINA Gout)

1. The Head or Veil
(1) The Cirrhi

(2) The Dorsal Mentacles.« or ‘ Rhinophores °

2. The Papillae or aha k,
3. The Foot .
4. The Body-wall .
(1) The Odoriferous Glands
(2) The Muscular System .
5. The Visceral Cavity .
6. The Alimentary Canal
(1) The Buccal Cavity ;
a. Mandibles and Radula .
b. Buccal and Salivary Glands .
(2) The Oesophagus
(3) The Stomach
a. Proventriculus
b. Gizzard
c. Pyloric Diverticulum
(4) The Intestine
(5) The Liver
7. The Circulatory System
(1) The Pericardium
(2) The Heart and the Rees
(3) The Venous System
8. The Organs of Excretion
(1) The Kidney
(2) The Ureter :
(3) The Renal Syrinx
9. The Organs of Reproduction

(1) The Hermaphrodite Gland, a New Type

or Ov Or Ot Ot Ot Ot Ot Or Si Or Or Or St Or

t
ve
lS

Ot Sr St Ot

516

bo
~~

Wwe OW ow

~“1-1 Q Ot

WwWwWwWwWwwwwwnwnhd hd wv wv

SoS Oo Ol WS = © +]

Or Or Ot Ot Ot Sr Or OS

Or Ot or Or
DS Doc
Ne aD

5OS H. P. KJERSCHOW AGERSBORG

PAGE

(2) The Hermaphrodite Duct . e ; : « Joon

(3) The Oviduct : ; : : ; a” eOoT

(4) The Ovispermatotheca : : , : ; 1568

(5) The Male Genital Duct : : : ; . 569

(6) The Mucous Gland . : ; : : . 570

V. SumMMARY : ; : : P : : : «| (Dito
VI. Lirrerature CITED . : : d : . F « Sone
VII. Notre to ExpLanaTIon OF FIGURES i 2 : . B86
VIII. ExpLanation oF PLATES 27-37 : 2 . : . 586

I. INTRODUCTION.

A CONSIDERABLE part of this work was done under the
direction of Professor Trevor Kincaid, at the University of
Washington (Seattle), 1914-15. It was continued at Columbia
University, 1918-20, and at Woods Hole, 1919 (summer).
Dr. T. C. Frye, the Director of the Puget Sound Biological
Station at Friday Harbour, Washington, had some new material
collected during the summer of 1919, and sent to me at Woods
Hole. Upon this material, and the previous data together with
new material collected at Friday Harbour, during the summer
of 1921, this paper is concluded.

In three previous communications (1919, 1921) I described
the method of feeding, the kind of food, method of swimming
(1922), and the colour, of Melibe leonina (Gould) as well
as the distribution of the family Tethymelibidae? Bergh
(1890, 1892: 1039-43). I also made an extensive review of
the literature on the nudibranchiate molluscs, particularly the
Cladohepatica. Several misprints, relative to year, volume,
and page, occurred in the literature. These have been corrected.
I hope that the references as printed in my papers may be of
help to other workers in this field on the molluses.

* Attention should be called to the fact that the name Tethys as
applied to a Nudibranch is incorrect. This was established in 1895.
(Vide H. A. Pilsbry, ‘* Classification and Phylogeny of Tectibranchiata ”’,
‘Manual of Conchology’, 16: i-vii, 1-262, 74 plates.) Thename Tethys
was first given to a Tectibranch by Linnaeus. (Vide ‘Systema Natura’,
10th Edition, 1758, p. 653.) Consequently the family name Tet hy meli-
bidae must be rejected from the nomenclature of Nudibranchiata.
The Nudibranch ‘Tethys’ has apparently no name of its own.

MORPHOLOGY OF MELIBE 509

Il. ACKNOWLEDGEMENTS.

The writer wishes here to express sincere thanks to Professor
Trevor Kineaid for helpful suggestions during the first period
of this work; to Dr. H. L. Osterud for collecting and fixing
new material in 1916; to the Curator of Books and Literature,
Dr. R. W. Tower, of the American Museum of Natural History,
for unfailing kindness during the review of the hterature ; and
to the Director of the Puget Sound Biological Station, Dr. T. C.
Frye, for every co-operation and assistance while at the
station, in the summer of 1921.

III. On tHe Status og CHIORAERA GOULD.

Bergh’s description of various species of Melibe (1875),
Melibe capucina, M. rangii; 1879a, M. vexillifera;
1884, M. papillosa; 1888, 1890b, M. ocellata; 1902,
M. bucephala; and 1908, M. rosea Rang), emphasizes
the following as Melibean characteristics: * Bulbus pharyn-
geus cum mandibulis ut in Phylliroidis ; margo masticatorius
mandibulae fortiter dentatus’ (1875b: 3862). Perhaps the
only exception to this may be found in the species collected
at the mouth of the Columbia River, in the State of Washington
(1904), in which case the author is not sure of the mandibles.
He says: ‘Bulbus pharyngeus lingua destitutus. . . . Die
Mundrohre und der Schlundkopf schienen sich wie sonst bei
den Meliben zu verhalten....’ I have previously called atten-
tion (1919, 1921) to the possibility that this species may be the
same as the one described by Gould (1852) from the Puget
Sound region. Not all Melibes have the same characteristics,
as indicated by Bergh; this is also shown by Alder and
Hancock (1864), and substantiated by Eliot (1902). The generic
characteristics as enunciated by Bergh (1875b) do not neces-
sarily hold, even though this author thinks that Hancock’s
(Alder and Hancock, 1864) description is incorrect. Bergh
(1875b: 363, 364) says: ‘Es kann kaum bezweifelt werden,
dass die von Hancock untersuchte Form mit der von mir
besprochenen congenerischist. Es werden sich daher die bei

510 H. P. KJERSCHOW AGERSBORG

dem englischen Verf. vorkommenden, von den untenstehenden
abweichenden anatomischen Angaben wahrscheinlich als
unrichtig erweisen... Besonders wird solches wohl der Fall
sein, wo Hancock den Anfang des Verdauungskanals bespricht :
“The buccal organ is provided with neither tongue, jaws nor
collar ; it is not by any means very distinctly marked, formed
as it were by a mere enlargement of the oesophagus, and having
little or no increase of muscular power.” ’

But Eliot (1902) verifies Hancock’s claim when he writes :
‘LT also found Alder and Hancock’s description of the internal
anatomy correct, particularly as regards the absence of jaws.
... Mr. Crossland and I have . . . dissected several specimens
of Melibe fimbriata, and in all failed to detect any
trace of jaws.’

Gould’s Chioraera leonina (1852) corresponds very
closely in the general anatomy to that of Melibe fim-
briata Alder and Hancock, (1864); this is also true in
regard to the species discovered by Rang (1829) and sub-
sequently described by Bergh (18638, 1871, 1875b, 1879a
1884, 1888, 1890 b, 1902, 1904, and 1908). The only difference
is on the point in regard to mandibles. Some authors, Rang,
Gould, Pease, Cooper, and Fewkes, do not touch on this point,
and for that reason one cannot tell whether the particular speci-
mens with which they dealt actually had such organs. With
the exception of the mandibles, all the generic characteristics
as set forth by the earliest writers on this type of the molluses
agree (Rang, 1829; Gould, 1852; Pease, 1860 ; Cooper, 1863;
Alder and Hancock, 1864; de Filppi, 1867; Tapparone-
Canefri, 1876; and Fewkes, 1889; as well as the numerous
descriptions of Bergh, 1863-1908). The discovery of the genus
Melibe by Rang (1829) seems to have been unknown to
Gould (1852), who created a new genus (Chioraera) for this
type. Cooper (1863) and Fewkes (1889) employed the nomen-
clature of Gould. The generic characteristics as enunciated
by the original author for Melibe, Rang (1829), are practically
identical with those set forth by Gould twenty-three years
later for Chioraera. ‘Tryon, Jr. (1883: 328), without

MORPHOLOGY OF MELIBE 511

stating a reason, classifies Chioraera as a synonym of
Melibe. Owing to the fact that Gould and also Cooper were
ignorant of the actual discovery of the genus Melibe, the
name Chioraera was invented by Gould and subsequently
used by Cooper. The name is, in fact, a mythical term that
is related in meaning to the former. Bergh (1904), describing
a species from the territory of Gould, Cooper, and Fewkes,
does not hesitate to employ the nomenclature of Rang (1829),
so similar is this form to the Melibes from other parts of the
world. No other author, except Bergh, gives mandibles as
a generic characteristic. That is, this feature is not observed
by Rang (1829), Gould (1852), Pease (1860), Cooper (1863),
de Filippi (1867), Tapparone-Canefri (1876), or Fewkes (1889).
Although Melibe Rang (1829), and Chioraera Gould
(1852), differ somewhat in shape, they are very similar in most
other respects. Both have a series of papillae on each side
dorsolaterally ; a large hood, cowl, or veil; a pair of tentacles
(the so-called rhinophoria) on the veil; the veil frmged with
at least two rows of cirrhi; and a narrow grooved foot which
is blunt in front and pointed behind; the head distinctly
separated from the body, and in each case it is very large ;
the gizzard is lined with a ‘ keratinized’ secretion which
protects the delicate epithelium, the so-called stomach-plates
of Alder and Hancock, or ‘Magenzihne’ of Bergh; these
two types are carnivorous; both are pelagic; and both are
distinetly cladohepatic. Therefore the species of the American
west coast which falls within this description must be the same
genus, 1.e. Melibe. The effort, therefore, to build further
on the nomenclature of Gould, as was done by Cooper (1863),
Fewkes (1889), and more recently by Heath (1917), seems to me
to be indefensible. And, owing to the fact that the genus
Melibe may either possess mandibles (Bergh, 1875 b) or not,
(Alder and Hancock, 1864; de Filippi, 1867; Tapparone-
Canefri, 1876; Eliot, 1902), the generic deseription may be
modified to read, in part, Bulbus pharyngeus aut cum mandi-
bulis aut sme mandibulis; radula et lingua destitutus. None
of the authors (Gould, Cooper, Fewkes, Heath) who has not

512 H. P. KJERSCHOW AGERSBORG

employed the nomenclature of Rang for this type, has deseribed
mandibles, and O'Donoghue (1921) states: ‘The radula and
jaws or any representatives of such structures are entirely
absent.” Although O’Donoghue (1921) also employed the
nomenclature of Gould for the genus Melibe, ina subsequent
letter to me he states: ‘I have quite given up Chioraera
as a name.’ In recent publications by this author (1922:
125; 1922a: 165) and by O’Donoghue (1922: 134) it is
suggested in a foot-note that Melibe leonina would be
a ‘better’ name than Chioraera leonina. Neither
Cooper, Fewkes, nor Heath made an intensive study of the
type; this is evident from their descriptions. A careful study
of Gould’s Chioraera has brought out sufficient reason to
merge it with Melibe as indicated by Tryon, Jr. (1883),
Bergh (1908), and Agersborg (1921 a). ‘The structures and the
general characteristics of Chioraera leonina Gould,
correspond in many details with those of the Melibes of Rang,
Bergh, et al. For this reason I have adopted the name as
indicated by Tryon, Jr., and also suggested by my friend,
Professor Trevor Kineaid, viz. Melibe leonina (Gould)
as indicated in the title of this work. (Vide Agersborg, 1921 a,
1922a, 1923.)

IV. MELIBE LEONINA (SYN. CHIORAERA
LEONINA GOULD).

The type of the genus Melibe was discovered at the Cape
of Good Hope and was described by Rang in 1829. Since that
time a number of species (vide supra) have been added by
various authors.t In 1852 Gould described Melibe (s.
Chioraera) leonina from Puget Sound, founding for it
the genus Chioraera, now merged with Melibe. In 1914
I observed this animal at the Puget Sound Biological Station
(vide Agersborg, 1916, 1919, 1921, 1921 a, 1922, 1922a, 1928).

1 My designation of Melibaea australis Angas (1864), as
Melibe australis (Agersborg, 1919, 1921) is not justified, as indicated
by the description of Angas. His description seems to fit the genus
Doto (vide Kjerschow Agersborg, 1921).

MORPHOLOGY OF MELIBE 513

Since the descriptions of Gould and also of Cooper and Fewkes,
each of whom described a species from the American west
coast, are rather incomplete, and since the anatomy had not
been worked out, I felt that there was sufficient reason to
engage in such an investigation upon this very interesting
animal. As a result of this work I have succeeded in bringing
to light some points of considerable zoological interest.

The body-substance of Melibe leonina appears as a
mass of brown jelly, when the animal is alive or freshly caught :
in the aquarium it turns practically transparent ; when it
has been preserved in alcohol or formaldehyde it gradually
loses its brown colour, and becomes almost white. O’ Donoghue
(1922: 125) says: ‘ Hundreds of individuals of this species
have been seen, but there is practically no variation in colour.
In some forms the yellowish or whitish-grey jelly-like body
may be tinged with pale brown but hardly sufficient to notice.’
However, while this is really the general colour of Melibe
leonina, I have also seen it deep green in colour (vide
Agersborg, 1922a: 439-42). Gould (1852), in his original
description of this species, says: ‘ Body limaciform, smooth and
of a pearly and whitish colour, finely reticulate with orange.’
The colour of the animal is caused by an extensive ramifica-
tion of the brownish-coloured liver in the body-wall (Pls. 1 and
30, figs. 1, 2, 4, 7, 18-28, 25); a ramification which extends
to the hood, the tentacles, the papillae, and to the ectoderm
of the rest of the body. The colour of nudibranchs has some-
times been attributed to their food, Hecht (1895), Ehot (1910) ;
but Alder and Hancock (1845) ascribe the colour to the liver
and the gonads. Bergh (1879 a: 163, 165), describing Melibe
vexillifera, says in fact: ‘Durch die diinnen Korper-
winde schimmerten, besonders an den Seiten, die denselben
angehorenden, dicht an einander liegenden, schmalen, weiss-
lichen, parallellaufenden Liingsfasern hindurch ; ferner undeut-
lich die Eingeweide, besonders Theile der Leber und die
vordere Genitalmasse; ... Die Leber wie bei anderen Meliben
eine lose, gelblichweisse Masse, welche vorne an den Magen
reicht, hinten sich bis an das Ende der HKingeweidehdhle

NO. 268 Mm

514 H. P. KJERSCHOW AGERSBORG

erstreckt.’ From this one might infer that the author was
describing the species from preserved material; for the liver
is brown in colour as a rule, and from it Melibe obtains most
of its colour. The internal organs, however, may be partly
seen through the body-wall in living specimens of M.leonina
(PI. 27, fig. 2). Gould deseribes the brownish ramifications of
the papillae as vascular ; these, however, are branches of the
liver. He also says that the tentacles are ‘ destitute of vena-
tion’, referrmg no doubt to the hepatic branches which
are very abundant everywhere. The hepatic branches are
less abundant in the tentacles but they are not wholly
absent.

1; Tine Hiead or Won.

The head of M. leonina is very prominent because of its
exceedingly large veil or hood. This is modified by two rows
of cirrhi (Pls. 27, 28, figs. 1, 3, 10, 17) which fringe its edge, and
by a pair of ear-like tentacles (Pl. 27, figs. 1, 2, 3). The cirrhi
of the outer row are much larger and far less in number than
those of the inner row. The average, taken from a number of
specimens, was 48 in the outer row, and 123 in the inner row,
or 2-56 small cirrhi to 1 large. The outer row does not extend
entirely around the rim of the hood, but, in an animal about
6 centimetres long, terminates about 1 centimetre from the
mid-ventral lme. The inner row extends all around the peri-
phery of the hood, although the last three cirrhi, on each side
of the mid-ventral, are rudimentary (Pl. 28, fig. 17, Ter, Ter).
A large veil or hood, except in M. ocellata Bergh (18905),
is a common thing for Melibe Gould (1852), Alder and
Hancock (1865), Bergh (18756, 1902, 1904), Eliot (1902).
In M. bucephala, Bergh (1902), ‘ The edge of the head is
rather thin above and almost smooth ; its outer parts, however,
are thick, inwardly somewhat refolded or convoluted, and
provided with several, mostly perhaps about five, close-set
series of cirrhi, which are displaced among each other; these
cirrhi are conical, somewhat constricted at the base, the inner-
most ones are the larger, toward the outside they decrease

MORPHOLOGY OF MELIBE bla

regularly in length.’ Gould (1852), describing M. leonina,
says: ‘The mouth is inferior and surrounded by a series of
long cirrhi, each of which has an independent motion.’

(Ly Phe: Cire irs

Ghercrrhs (Pl. 27, figs..1, C, 10, Ie, Oc, Pl. 28, 17, Ire; Ore)
have an inner median axis (PI. 28, figs. 11, 12, Cg, 138, Cga) from
which radiating fibres pass to the periphery (Rf). The inner axis
seems to consist of a series of nerve ganglia from which fibres
radiate to the periphery of the cone-hke cirrhus, ending in the
basal layer of the epidermis (PI. 28, fig. 11, Oe, Ie). In other
words, a cirrhus is a conic structure having an epithelial wall
resting upon a delicate connective-tissue framework, with an
inner axis which passes from the apex to the base of the cirrhus,
and from this axis fibres radiate to the epithelial walls of the
cone. ‘These fibres have the resemblance of nerve-fibres,
particularly in their relation to the cells in the ganglionic
axis and to the epithelium of the epidermis. In one of the
angles, formed by the radiating fibres at the axis, is a large
bladder-like structure around which the fibres pass, and at the
periphery are placed large and small cells (Pl. 28, fig. 18, Pe)
the inner part being reticular in structure (Jr) and containing
apparently no cells, or a few only (Pl. 28, figs. 11, 12, 13, Ir).
The cirrhi are probably tactile or gustatory in function. Some-
thing to that effect was demonstrated experimentally (Agers-
borg, 1922a: p. 441). Sedgwick (1898: p. 3866) writing on the
gasteropods in general and opisthobranchs in particular says :
‘Tactile organs are represented by the tentacles, the edges
of the lips, which are often folded (labial palps), the tentacular
and lobe-hke prolongations which are found here and there
on the head, mouth, and foot.’ It should be remembered that
there are sometimes two pairs of tentacles in nudibranchs ;
the anterior pair being the one referred to here ; the posterior
is the so-called rhinophoria of many authors, which, however,
do not seem to be olfactory in function (Agersborg, 19224
493-44),

Mm2

516 H. P. KJERSCHOW AGERSBORG

(2) The Dorsal Tentacles or ‘Rhinophores’.

On the outside of the veil, in M. leonina, is a pair of
tentacles, supposed to be equal to the posterior pair in, e.g.,
Aeolidia Cuvier (1798). According to the original descrip-
tion of the genus Melibe by Rang (1829) the tentacles are ‘au
nombre de deux, situés a la base du voile, tres allongés, coniques,
terminés par une petite capsule, de laquelle sort un organe
conique et rétractile’; Gould (1852), in his description of
Chioraera leonina (s. Melibe leonina), points out:

. . tentaculae cephalicae fohatae, retractiles ;’ and Pease
(1860) for M. pilosa says: ‘Tentacles on the posterior
portion of the veil rather remote, small, ovate, closely and
transversely lamellated and retractile into long trumpet-
shaped sheaths, which are furnished with laciniated appendages.’
Again, Tapparone-Canefri (1876) for Jacunia (s. Melibe)
papillosa de Filippi, states: ‘ Tentacula (Rhynophoria)
laminata, tenuia, apice obtusiuscula, retractilia, e vagina
caliciformi angusta vix proeminentia.’ Cooper (1863) and
Fewkes (1889) are content with the description of Gould ;
neither of them mentions tentacles. Alder and Hancock
(1864) refer to these organs as dorsal tentacles. The largest
part of the tentacles in Melibe is the tentacular stalks ; they
are wedge-shaped bodies (Pl. 27, figs. 38, 8) somewhat rounded
at their base. The wedge is like a broad axe, the edge of which
is a little curved. They are arranged at right angles to each
other, and this angle would intersect posterior to their base
in the mid-dorsal line of the hood. Along the edge of the curved
wedges are slits, in one of which, on the inner part of the curve,
i.e. not on the apical part of the tentacle—except when it
is expanded—is a small organ (PI. 27, fig. 8, Rh), the real
tentacle, that may be retracted below the surface of the
wedge-shaped tentacular stalks (vide Agersborg, 1923, figs. 4,
5, pa). Gould, in describing these, says: * On the top of the
head are two foliate expansions destitute of venations, which
answer to the true tentacles ; on their anterior edge is an opaque
whitish papilla, presenting something of a spiral or lamellar

fond

MORPHOLOGY OF MELIBE 517

structure; they are sometimes wholly retracted within a
permanent sheath.’ At the base of these tentacular organs,
when they are retracted (Pl. 27, fig. 8 and Pl. 29, fig. 15, Ix), is
asmallknob. In sections (PI. 29, fig. 15) the lamellar structure
is indicated by certain lobations along the outer part (S). The
knob (i) seems to be made up of a mass of large and small
nerve-cells (Pl. 29, fig. 16), which fibres connect with similar
cells distal in position to the former (Ne, Nes, Nfb, Nfi), and
finally by innervation in the epithelium of the lamellar external
parts of the organ (Nfp). From the distal border of the knob-
like ganglion, fibres communicate with the base of the tentacle ;
these fibres are not made up of nerve-fibres only but also of
muscle-fibres (Nmf). This is known from their staining
reaction and also from the fact that some of these fibres
communicate with fibres within the organ which are decidedly
nervous, while the other fibres termimate on the organ. The
muscle-fibres help to retract the organ below the surface of
the tentacle, that is, to withdraw the tentacle within its stalk,
which in that case serves as a sheath; the nerve-fibres to
convey stimuli. There is no permanent sheath, as indicated
by Gould, save that part of the tentacular stalk which sur-
rounds the organ, and acts as a sheath when the tentacle
is retracted.

The function of these tentacular organs in nudibranchs
according to various authors is olfactory. Thus, Alder and
Hancock (1845: 19) say: ‘The dorsal tentacles are the
organs of smell, and, judging from their development, this
sense must be more acute in most of the nudibranchs than it
is in many other molluscs, with the exception, perhaps, of
Nautilus.’ Hancock and Embleton (1852: 242), discussing
a Doris, say: ‘ The dorsal tentacles, which have never been
observed to be used as tactile organs, we believe to be the seat
of the sense of smell; and this belief is strengthened when we
reflect that these sense organs are most highly developed and
minutely laminated; that they are plentifully supplied with
nerves from the ganglia placed in front of all the rest of the
cerebral masses; that they are externally covered with

518 H. P. KJERSCHOW AGERSBORG

vibratile cilia, and so placed on the head as easily to receive
impression from any odorous particles that may be mingled
with the circumambient water.’ Likewise, Jeffreys (1869)
claims olfaction for the dorsal tentacles, but he says: ‘... olfac-
tion in these animals probably is not so much to assist in the
discovery of alimentary matters, as to give warning of the
unhealthy state of the surrounding medium, arising from
putrescence or other causes. . . . and its outer surface, in all
the nudibranchs, is provided with vibratile cilia.’ The tentacles
in M. leonina, however, are not cilated. Bergh adopts
the term ‘rhinophoria’ for the dorsal tentacles, not only
indicating in that way the function, but he claims, in fact
olfaction to be their function. Later writers, Lang
(1896: 48, 103), Sedgwick (1898 : 366), seem to agree on this
point. Copeland (1918 : 177-227) demonstrated experimentally
that the monotocardiate prosobranchs Alectrion obsoleta
and Busycon canaliculatum respond to stimulations
by dilute food extracts and materials emanating from distant
food; he thinks that the snails do not find food by coming
upon it accidentally, but are directed to it by movements
brought about through stimulations of the olfactory organs
with odorous substances conducted to the receptor in varying
concentrations by the moving siphon. By means of an olfactory
apparatus consisting of a single organ of smell associated with
a siphon terminating in a shifting ‘ nostril’, for sampling the
surrounding water and its contents, the snail is as successfully
directed toward distant food as an animal which, like the
dogfish, possesses paired olfactory organs and fixed nostrils.
After the osphradium in Busycon was destroyed the snail
failed to respond to dilute food materials, but a year later,
when the lamellae of the organ were partly regenerated, the
scenting responses returned. The osphradium, therefore, is an
olfactory organ. ‘This author claims further that taste in the
snail (Busycon) is a diffused sense as compared with olfac-
tion, and that a large portion of the surface of the snail possesses
this sense. But Arey (1918: 531) distrusts the capability of
snails to analyse chemical stimuli as discrete sensations, and
thinks that it would be safer to avoid referring in their case

MORPHOLOGY OF MELIBE 519

to a sense of taste and smell at all, or even to a common
chemical sense, but rather to designate the particular senses
in question a general chemical sense. Moreover, Arey working
experimentally on several nudibranchs, Chromodoris
zebra, Facelina goslingi, Elysia ecrispa, and
Fiona marina, found that there was nothing in the tests
which he applied to these animals that connected the rhino-
phores with olfaction, with the exception perhaps of Facelina
whose non-retractile rhinophores react by a lashing withdrawal
and more vigorously than the oral tentacles when stimulated
by oil of pennyroyal, carbon bisulphide, and anilin oil. All
these forms, however, responded to tactile stimuli. Like
Copeland, he found that the general body-surface also responded
to chemical stimuli. Crozier and Arey (1919: 301) elaborated
on this by stating that in Chromodoris the rhinophores
and the oral tentacles are in a general way the parts most sensi-
tive to chemical stimuli. Agaim, as regards the rhmophores,
these authors (1919 : 278-81) found that while Chromodoris
zebra may creep in an entirely normal fashion after the
rhinophores have been removed, it loses its power of orientation
to the water current. In other words, these authors claim
that to currents of adequate velocity the nudibranchs are
negatively rheotropic and that the rhinophores are the prime
receptive organs for this kind of reaction. However, as I have
pointed out elsewhere (Agersborg, 1922a: 432, 439), the
dorsal tentacles (rhinophores) of Hermissenda opales-
cens (Cooper),' do not seem to have a rheotropic function,
because specimens with one or both of the dorsal tentacles
removed oriented as easily and moved against the current as
did the normal individuals. It did not seem to make any
difference whether the dorsal tentacles were present or not.
At any rate, the rhinophores or dorsal tentacles do not seem
to be ‘ rheotropic’ in Hermissenda. Copeland, Arey, and
Crozier found that the types with which they worked were
more sensitive around the anterior than elsewhere on the
body. Such a specialization of the integuments is also the case

1 The correct name is Hermissenda erassicornis Eschscholtz.
Vide C. H. O’ Donoghue (1922), ‘ Nautilus’, 35: 74-7.

520 H. P. KJERSCHOW AGERSBORG

in Hermissenda, Dendronotus, and Melibe (Agers-
borg, 1922a: pp. 428-44). Indeed, as recently brought to light
by Gross (1921), on Nereis virens Sars, the general
integuments of this organism are sensitive to chemical stimula-
tion with a localization or concentration of the chemical sense
in the palps and tentacles, a circumstance correlated with the
rich innervation of these appendages and the relation of their
nerves to the brain. However, I have not yet found any
specialized receptors either in the dorsal tentacles or in the
cirrhi of nudibranchs. Their function, therefore, as far as the
nudibranchs are concerned, may not be so definite as previously
indicated. For this reason, and because of the facts brought to
light by experimental evidence (Copeland, 1918; Arey, 1918 ;
Arey and Crozier, 1919; Agersborg, 1922a), I have used the
original name tentacles, as employed by Alder and Han-
cock (1845, 1864), Hancock and Embleton (1848), and Gould
(1852), rather than the suggestive *‘ rhinophore ’ as adopted by
Bergh and freely used by subsequent writers. For, although
the tentacles of the hood are highly specialized as indicated by
Alder and Hancock, Hancock and Embleton, and by my
drawings (Pl. 27, fig. 8, Pl. 29, figs. 15 and 16), it is now very
doubtful whether their function is olfactory per se, or even
shghtly so.

The remainder of the hood is apparently smooth, but upon
close examination it is found to be covered with tubercles,
a feature so common to the ectoderm all over the body of
Melibe leonina; these tubercles are macroscopic in
M. fimbriata Alder and Hancock (1864), Eliot (1902). The
ventral side of the cowl (PI. 27, fig. 1, Pl. 29, fig. 17) is concave in
M. leonina; muscle-fibres, radiating from the muscles of
the neck, support the veil. In the middle of the concave area
between the bases of the tentacles (PI. 29, fig. 17, R) isa marked
depression (Mdp). The ventral side of the cowl (PI. 27, fig. 6, En)
is tuberculate lke the external side, but it has no odoriferous
glands (Pl. 27, fig. 4, Oo, Pl. 30, fig. 25, Og): to be discussed
below. The head is set off distinctly from the body by a neck
(PI. 2, Hes. 2550).

or
bo
fond,

MORPHOLOGY OF MELIBE

2. The Papillae or Epinotidia.

Among the Cladohepatica, where the papillae are
mostly foliaceous or lobate structures, the denomination of
papillae is preferable to the term cerata, branchiae,
or gills. Many English authors have adopted this term :
Hancock and Embleton (1848), Jeffrey (1869), Gamble (1892),
et al. Others use a different nomenclature: Alder and
Hancock (1845), branchial papillae; Parona (1891), Vigwier
(1898), dorsal appendages; Parker and Haswell (1910),
secondary branchiae ; Lang (1898), dorsal respiratory appen-
dages (cerata); and still others, Herman and Clubb (1892),
Sedgwick (1898), Hertwig (1912), Arnold (1916), Pratt
(1916), use the term cerata for the cladohepatic nudibranchs,
and branchiae for the Holohepatica. ‘The followmg
authors, dealing with Melibe in each case, designate the
papillae as follows: Rang (the founder for the genus) (1829),
branchiae ; Gould (1852) (Chioraera s. Melibe), foliaceous
branchial expansions; Pease (1860), tuberculated lobes ;
Cooper (1863), branchiae ; Tapparone-Canefri (1876), branchial
lobes ; Fewkes (1889) (Chioraea leontina s. Chioraera
leonina, Melibe), branchial appendages; Bergh (1908),
epinotidia ; Heath (1917) (Chioraera dalli s. Melibe
leonina), lappets; and O’Donoghue (1921) (Chioraera
s. Melibe), branchial cerata. The last-named author employs
the term cerata for the following cladohepatic genera :
Dendronotus, Aeolidia, Coryphella, Hermis-
senda, and Doto; but he applies the word branchiae
to the Holohepatica. This is in keeping with the usage
of many authors (vide supra). Boas (1916, 1920) employs the
term gills for boththe Aeolidiidae andthe Dorididae,
while Bergh (1879¢: 73) says that he uses the term papillae
for the Aeolidiidae partly because it is a Linnean term,
partly because the organs do not exclusively serve for respira-
tion, which is partaken of by the whole surface of the skin,
that over the papillae as well as elsewhere, among all the
Nudibranchiata. ‘This fact was pointed out earlier by

592 H. P. KJERSCHOW AGERSBORG

Hancock and Embleton (1848: 103), who wrote: * The fune-
tion of respiration we believe to be performed by the whole
surface of the skin, including the papillae, the skin of the
back and of the sides between the papillae, and the entire
surface of the latter organs... .’ It may, therefore, be quite
incorrect to designate these ‘branchial lobes’ cerata;
moreover, this term stands nearer ctenidium or true gill in
meaning, and on that account, and for the reason stated by
Bergh, and the facts recorded by Hancock and Embleton,
papillae may be the most appropriate term. The Tethyme-
libidae being cladohepatic nudibranchs, of course, come
under this terminology. Bergh also uses the term papillae
for the Tethymelibidae, and is the most consistent
writer in this as well as in other respects relative to the
nomenclature he employs.

Melibe leonina has six pairs of papillae, alternating in
position (Pl. 27, figs. 1,3). They appear smooth to the naked
eye, but fundamentally they are tufted or fimbriated as
in M. fimbriata Alder and Hancock, M. bucephala
Bergh, and Tethys leporina Linnaeus, although in
M. leonina (Gould) the fimbriated condition of the papillae
is hardly distinguishable. The first pair is located dorsad and
a little posterior to the genital pores, and approximately in line
with the hepatic junction to the stomach. The arrangement, the
size, Shape, and structure of the papillae may be seen in PI. 27,
figs. 1, 3, and Pl. 30, figs. 18-25 respectively. Microscopically,
the papillae show two principal morphological constituents,
viz. (1) terminating branches of the liver (the brownish vascular
ramifications of Gould), and (2) smooth musele-fibres ; but
also vascular spaces (PI. 30, fig. 25, Osp), odoriferous glands
(Og), and a tubercular surface (Tbr). But the papillae are,
however, subject to variation in their structure, depending
on the age and the position of the papilla. The two anterior
pairs (Pl. 30, figs. 18-21) are far more profusely supplied with
hepatic diverticula and musele-fibres than are the remaiming
pairs (PI. 30, figs. 22-4). The last pair does not seem to have
any appreciable amount of musele-fibres or liver-branches.

MORPHOLOGY OF MELIBE 523

The muscles ramifying in the papillae form a sort of supporting
wall; underlying this are the hepatic branches (Pl. , fig. 25,
Cshb), with many transparent spaces between (Osp). On the
outside of the muscle-wall, just beneath the ectoderm, are
odoriferous glands (Og). The size of the papillae decreases
gradually from the anterior to the posterior pair, the last pair
being very small, and its hepatic as well as muscular contents
are reduced accordingly. Regenerating pairs of the anterior
papillae show a great number of hepatic branches and muscle-
fibres ; this is a striking contrast to old posterior papillae which
apparently have no muscle-fibres and no hepatic diverticula.
It has been observed quite frequently, by different authors,
that the papillae of Aeolidia when cast off swim through
the water like worms, propelled by the vibratile cia, and
occasionally by the spasmodic action of the muscles (Jeffrey,
1869). I have myself kept papillae of Aeolidia, at Woods
Hole, and of Hermissenda, at Friday Harbour, alive in
glass dishes for weeks at a time, the papillae being in constant
motion, swimming in a circle owing to ciliary action on their
curved surface. This phenomenon is not so extraordinary
as it seems. It is found in other invertebrates (Planaria,
Echinoderms, &e.).. Gamble (1892) reports that the
papillae of Lomanotus show remarkable co-ordinative move-
ments when they are touched gently. Autotomy and regenera-
tion of the dorsal appendices, according to Parona (1891), is
a common occurrence among Tethys and Aeolidia.
Pease (1860), referrmg to M. pilosa, says: ‘ When slightly
disturbed they would cast off one or all of their lobes ...: they
may be consequently reproduced, after being cast off.’ This
is also true for M. leonina, for it frequently throws off some
or all of its papillae, and yet I have kept specimens of this
species for a number of days without autotomy taking place ;
I have also kept preserved specimens for years without the
papillae having dropped off, even though they were subjected
to considerable handling (see PI. 27, figs. 1, 2,3). Lang (1896)
also records the fact that if the papillae fall off they are
regenerated. It is believed that the papillae serve as organs

524 ' H. P. KJERSCHOW AGERSBORG

of respiration, that is, they are at least partly respiratory in
function; for this purpose large intercellular sinuses are
present which communicate with the heart through the efferent
branchial veins (PI. 30, figs. 25, Osp, 54, Aur). Such sinuses,
however, are not present in the papillae only.

3, hey Moos.

The foot (PI. 27, fig. 1, Pl. 28, fig. 9) extends a little beyond the
trunk both anteriorly and posteriorly. Its general form is like
an inverted flat-bottom dory, with the anterior end wider than
the posterior, curved, and extended forward about 1 em.
from its base, the so-called keel. The posterior end, also
curved but narrower than the front, projects about 1-5 em.
from the base in an adult. The edges bend considerably
outward, making the width from rim to rim about twice
as wide as the base. Between the edges, on the ventral side,
is a depression so deep that the comparison with a dory is
quite fitting. This groove is highly tuberculate and ciliated
(Pl. 30, figs. 26, 27, Cil, Tbr). The internal structure of the
foot varies. The anterior end has a fine network of nerve-cells
spread throughout its length, and at the posterior end is an
aggregation of nerve-cells into a ganglionic centre (PI. 30,
figs. 26, 27, 28, 29). The anterior end of the foot has also
a great number of small glands which open to the outside all
along the foot by fine crypts through the ciliated ectoderm
(Pl. 30, figs. 28, 29, Mug) and decrease in number and size
toward the posterior end. The secretion of these glands is
perhaps of use to the animal in helping it to move over fronds
of marine vegetation and other solid objects. Very fine neural
fibres extend from the pedal ganglion (PI. 30, fig. 29, N#) to the
base of the ciliated columnar epithelial cells and to the glands.
This suggests that the cilia of the ectoderm of the foot and the
glands in the foot are under nerve control. And, as I have
previously recorded (1919, 1921), M. leonina may move
without any visible bodily contortions, the foot bemg then
either in touch with the surface tension of the water, or with
the fronds of marine vegetation or some other solid. In the

MORPHOLOGY OF MELIBE 525

laboratory it moves along the side of the glass aquarium as
other nudibranchs do, e.g. Aeolidia. Such movements
seem to be caused by the ciliary action of the foot. In this
respect my observations differ decidedly from those of Pease
(1860) on M. pilosa, who writes: * Their foot cannot be
used for creeping on a flat surface, but if is well adapted for
clasping sea-weed ;’ and O'Donoghue (1921: 194), who says
in part: ‘It does not creep about on the eel grass but only
seems to adhere for the purpose of laying its eggs. In the
laboratory, too, it does not creep on the sides of the aquaria
and only partly clings to them. It has not been observed
creeping on anything after the manner of other nudibranchs,
and if not entirely a pelagic form hke Phyllirhoe it is
beyond doubt very nearly so and is a most interesting form.’
Although O’Donoghue thinks Melibe is mainly pelagic, it
is quite evident, judging from its habitat, that the pelagic
habit is periodic at the most, i.e. its recurrence is spasmodic
(Agersborg, 1916, 1919, 1921, 1921a, 1922, 19224, 1928).
M. leonina occurs not only as a pelagic form, but may
be found at a considerable depth, which is perhaps its habitat
the greater part of the year. Gould’s specimen, 133 mm. long,
17 mm. high, and 32 mm. wide, was dredged at about 54 metres
depth ; Cooper’s, 70mm. long and 17 mm. high, at a depth
of 88 metres. Presumably, when it is at the bottom, it crawls
on the bottom, for it has a well-developed foot not only
for clinging to sea-weeds but also for actual creeping and,
indeed, for ‘ galloping’, to use a term employed by previous
writers for other forms (Agersborg, 1923).

On one occasion, as I was trying to feed M. leonina, it
dropped to the bottom of the aquarium and commenced
gliding along the bottom. I continued the feeding experi-
ment, when to my astonishment this nudibranch, Melibe
leonina, suddenly elongated to nearly twice its normal
length, showing a method of creeping similar to that described
by Parker (1917) for the sea-hare, Aplysia californica
Cooper (PI. iv, fig. 9). When elongating the body, the anterior
one-third of the foot was lifted above the substratum and

526 H. P. KJERSCHOW AGERSBORG

then let down; the posterior third then passed toward the
middle of the body which became much wider at the base
and along the sides, and then the forward stretching was
repeated. This sort of creeping was accomplished by a large
muscular wave which passed from the anterior to the posterior,
i.e. by direct monotaxic waves. In ordinary locomotion
(creeping) the cilia of the foot may play an important part
because the locomotor waves are almost indiscernible (Agers-
borg, 1923: 93-6). M. leonina is, indeed, pelagic, but it
is a poor swimmer as compared with Dendronotus gigan-
teus O'Donoghue (Agersborg, 1922: 264); it is less pelagic
than Phyllirhoe, which has lost its foot, and it is perfectly
able to use its foot both for clinging to sea-weed and other
solids and for creeping.

A ciliated foot is a common thing among the gasteropods
and other molluscs. This was recorded by Flemming as early
as 1869 for Helic hortensis; List (1887) for Tethys
fimbriata; later by Stempell (1899) for the lamellibranch
Solemya tagota Poli; and recently by Copeland (1918)
for Alectrion obsoleta, et al. I have myself examined
the foot of various cladohepatic nudibranchs (Aeolidia
olivacea’, “Ae. .coronata, ove... concinna, ieee
diversa,- Doto coronata, et al.) at Woods: Hole,
Massachusetts, and found a uniformly ciliated foot in each
case. Pedal glands are recorded by various authors: Leydig
(1876), List (1887), Lang (1896), Sedgwick (1898), Stempell
(1899), Lankester (1906), Parker and Haswell (1910), Hertwig
(1912), but no one has deseribed the pedal gland in Melibe.
Lankester (1906) comes the nearest to describing the condition
as it exists in this species. That is, the pedal gland is not an
aggregation of glands or a simple branched invagination of the
integuments opening in the mid-ventral line of the foot as im
Triton nodiferus in particular and other gasteropods in
general according to the records of Parker and Haswell, and
Hertwig, but consists of a number of unicellular glands,
apparently equally distributed all over the foot, that open by
small erypts through the ciliated columnar epithelial surface

MORPHOLOGY OF MELIBE 527

of the ventral side. List finds three kinds of glands in the
foot of Tethys fimbriata Linnaeus; two of these are
unicellular and the third is multinuclear. The unicellular
glands with one nucleus are located on the dorsal side of the
foot ; the multinuclear glands are found on the ventral side
of the foot. Some of the mononucleate unicellular glands
contain a fatty substance. ‘The foot is covered by a layer of
ciated columnar epithelium with basally placed nuclei.
Between these epithelial cells unicellular and multinucleated
glands open to the outside through an individual pore or
crypt. The structure of the foot of M. leonina conforms
very nearly to that recorded by List for Tethys. It is not,
however, my present intention to make a critical physiological
or morphological comparison of the foot in these two types.
The *‘ mehrkernige Driisen’ of List I am unable to recognize
in this species. Pl. 30, figs. 26, 27, 28, and 29, show the relation
of the unicellular mucous gland to the ectoderm. These
glands are highly granular in structure with a central nucleus
(Pl. 30, fig. 28, Gmug, Nu). They are basophil in their staining
and there is a marked contrast between them and the nerve-
cells which are scattered all through the foot as a net. The
latter, however, is firmly aggregated into a pedal ganglion
in the posterior end of the foot (Pl. 30, figs. 27, Gl, 29,
Pdgn).

The function of the pedal glands seems to be that of secreting
a mucus for the purpose of aiding the animal in its progressive
movements when creeping on the surface of any object. This,
in fact, 1s also practised among terrestrial gasteropods, some
of which may use the pedal secretion to spin themselves from
the limb of a tree or some other plant to the ground. In the
Aeolidia the activity of the pedal glands is so great that
a few specimens (circ. 1} ¢m. long) confined in a finger-bowl
for a few hours may produce a complete film of slime on the
surface of the water, to which the organisms adhere.

528 H. P. KJERSCHOW AGERSBORG

4. The Body-wall.

(1) The Odoriferous Glands.

The entire surface of the body of M. leonina is fimbriated
or tufted, although it is not recognizable to the naked eye,
but it is easily detectable with the aid of a lens (figs. 6, Hz,
25, Tbr, 26, 27, Tbr, 31, Pec). In this respect it resembles
M. rosea Rang, M. pilosa Pease, M. papillosa de
Filippi, M. fimbriata Alder and Hancock, M. bucephala
Bergh, and also Tethys fimbriata Linnaeus (s. fimbria,
Bohascht, Delle Chiais). Closely associated with the fimbriated
ectoderm are several kinds of glands (Pl. 31, fig. 30, Glo, Sm,
Um). There are at least three kinds of these glands: (1) odori-
ferous (Glo), (2) saccular mucous (Sm), and (3) unicellular
mucous (Um). The former is the largest of the three and
most numerous (PI. 27, figs. 4, Go, 7). In structure, the odori-
ferous and saccular mucous glands are similar, but the latter
react to Delafield’s haematoxylin very much hke that of the
mucous glands of the foot, while the former seem to be like
serous glands. The odoriferous glands, in addition to bemg
larger than the mucous glands, are also a little more complex,
i.e. compound saccular (Pl. 31, figs. 30, 31, Glo). One reason
for assigning the odoriferous function to the largest and most
numerous of the cutaneous glands is the fact that this species
exudes a rather strong odour, which in a previous paper (1921)
I have designated as a means of defence, when it is touched,
and at the same time mucus secretion is not noticeable. The
unicellular mucous glands are typical glands of its kind (PI. 31,
fig. 830, Um). The skin of M. vexillifera Bergh (1879a)
is also noted for its numerous ‘iahnliche Driisenzellen ’.
And according to Hancock and Embleton (1848: 103), referring
to Aeolidia, * The outer or dermal layer of the skin appears
to secrete the abundant tenacious matter that exudes from the
animal, and to be the seat of an exquisite sensibility. This
layer is thin but continuous with the next or muscular layer,
which might be called the cellular for its structure.’ But
Flemming (1870) thinks the subepithelial connective tissue in
Helix pomatia secretes slime.

MORPHOLOGY OF MELIBE 529

(2) The Muscular System.

The muscular system of M. leonina is one of the most
striking features of the animal. When the skin and the caecal
endings of the liver (PI. 27, fig. 3) are removed, the main arrange-
ment of the muscles may be seen to be like that of the imter-
woven fibres of a basket, the sides of the animal being supported
by a network of muscular fibres. One set runs parallel to the
median axis from the anterior to the posterior ends, terminating
anteriorly in the periphery of the hood. Dorsally these fibres
end by branching in the ridge of the back and in the papillae ;
posteriorly they end in the ridge and in the base of the papillae
of that part of the body; ventrally they run parallel to the
foot, ending anteriorly and posteriorly in the base of the foot ;
the last parallel fibres end in the groove of the foot. Another
set runs diagonally, also parallel to each other, and ends in
fine fibres anteriorly, posteriorly, dorsally, and ventrally.
Hartmann (1880: 11) describes the muscles for Tethys
fimbriata as follows: * Die Oberseite des Kopfsegels und
des Riickens zeigt auch 6fters gediipfelte, manchmal wieder
weiss gesiiumte Schrig- und Querbinder. Dies hat bereits
G. Cuvier recht gut abgebildet (Mollusques, Tab. VII, Fig. 1).’
It is then seen that Tethys and Melibe resemble each other
in the arrangement of the muscles of the body. In Melibe
the muscle-fibres are located midway: between the ectoderm
and the boundary of the visceral cavity (Pl. 31, fig. 80, Mb).
Between the ectoderm and the muscles are a great many
connective-tissue cells (Ct) and fibres (Pl. 31, fig. 34, Pme),
ends of the branching hepatic system (Pl. 27, figs. 4, 7, Hep),
and plasma (Pl. 31, fig. 31, Sp). The finer structure of the
muscles is of the type common to molluses : a circular arrange-
ment of finer fibrils in each individual muscle-cell (Pl. 31,
fig. 38, My). The beautiful picture, however, presented by
a transverse or longitudinal section through a muscle-bundle
needs to be commented on a little further.

Schneider (1908), describing the muscle structure of
Chiton siculus, says: ‘Ueber die feine Structur der
Muskelzellen ist nicht viel auszusagen. Die langen Fasern

NO. 268 Nn

530 H. P. KJERSCHOW AGERSBORG

sind Bindel von Myofibrillen ohne innere Sarcaxe. Der
lingliche Kern hegt der Faser dicht angepresst.’ Contimuing
on the same subject, but dealing with a different type, 1.e.
Helix pomatia, he says: ‘ Die Fasern sind von rundlichem
Querschnitt, langgestreckt, glattfibrilliér und zeigen den
Kern seitlich in einem geringen Sarcrest (Zellk6rper) anliegen.’
His comment on the muscles of Anodonta mutabilis
is much the same as for those already cited, i.e.: ‘ Die glatten
Muskelfasern zeigen gewohnlich nichts Auffallendes. Jede
Faser besteht aus parallel verlaufenden Fibrillen. Der lang-
liche Kern leet der Faser eimseitig an, innerhalb einer geringen
Saremenge die als Zellkérper zu bezeichnen ist.’

Schneider’s description of molluscan muscle is rather general-
ized and, in fact, fails to brmg out the facts concerning the
fmer structure. ‘The muscle-bundles of M. leonina are
surrounded by a thin membranous sheath, the perimysium
(Pl. 31, fig. 38, Ms), and each individual muscle-cell by an
exceedingly thin membrane, the endomysium (Mt). Between
the muscle-fibres, or muscle-cells, are primitive connective-tissue
cells (Inct). The muscle-cell or -fibre (PI. 31, fig. 83, Pl. 37, fig. 82,)
is differentiated into two sarcoplasmic regions : an outer finely
granular and an inner coarsely granular region with a trans-
parent ground substance. At the periphery of each region,
very coarse granules are arranged in such a way so as to
give the appearance of a granular network enclosmg each
(figs. 33, Sar, 82, Myf). The outer of these may be in close
relation with the sarcolemma if such a one is present; the
inner granular network holds a similar relation to the finely
granular region as the outer granular network holds to the peri-
phery of the cell, each enclosing, as it were, two different
cytoplasmic regions of the cell. There are thus four kinds of
sranules in the muscle-cell relative to their location and size.
These granules may be designated as myofibrillae because of
their linear arrangement. The nucleus is placed centrally
within the ground substance of the inner region surrounded by
the coarser myofibrillae, immediately, and ultimately by the
outer and more finely granular substance. In staining capacity,

MORPHOLOGY OF MELIBB 581

the ground substance (hyaloplasm or sarcoplasm) of the inner
region shows less affinity for cytoplasmic stain than that of the
outer region. The granules (macromeres) of the inner region
are farther apart than those (micromeres) of the outer region
of the muscle-fibre. This is perhaps also the reason why
the muscle-cell of molluscs appears as being, according to
Schneider, ‘innerhalb einer geringen Sarcmenge ’. Chromatin
bodies are distributed throughout the nucleus, either in the
meshes of the linin and around the nuclear periphery, or
around the peripheral part only (figs. 38, K, 82, Kar). The
peripheral granular net (sarcolemma ?) may be seen, in the
whole mounts of muscle-fibres stained by Congo red, as a fine
granular structure around the periphery of the fibre (Pl. 31
fig. 32).

The structure of the inner body-wall as found beneath the
basket formation (PI. 27, fig. 3) of the muscle arrangement in
the body-wall, shows apparently no regular arrangement of
fibres as in the case of the muscles ; the fibres here, which are
of connective tissue, seem to extend in every conceivable
direction (PI. 27, fig. 5), and this layer is continuous until some
visceral organ or the pericardial chamber is reached. In this
plexus of irregularly arranged connective-tissue fibres and cells
are the visceral bodies : the brain, the heart, the stomach, the
intestine, the organs of reproduction with their adjuncts, and
the renal organs. The amount of connective tissue in the body-
wall does not render it opaque. This is partly due to the loose
arrangement of the various kinds of tissues and to the presence
of numerous sinuses (PI. 31, fig. 31, Sp) containing a transparent
fluid. This fluid contains characteristic, primitive connective
tissue cells and strands (Pl. 31, fig. 84, Pne).

5. The Visceral Cavity.

According to Lang (1896: 211) the Mollusca are said
to have primary and secondary body-cavities. The former is
the system of lacunae and sinuses, into which the arteries
open, and out of which the veins, where these are present,
‘draw their blood. It has no epithelial walls of its own. Its

Nn2

joe H. Pw. KJERSCHOW AGERSBORG

boundaries are formed by connective, nerve, or muscle tissues,
or by epithelia, which, however, belong to other organs, such
as the intestine, the kidneys, or the body-wall. The latter,
the so-called secondary body-cavity or coelom, is in most
Mollusea very much reduced, usually consisting of only two
small cavities, the pericardium and the cavity of the gonads.
The coelom is always lined by an epithelium of its own, the
coelomic epithelium, and corresponds with the true coelom
of the Annelida, which also possesses such an epithelium.

The primary body-cavity of Lang corresponds with the
perivisceral cavity of Sedgwick (1898: 375), who says: ‘In
Gastropoda there is usually a well-developed perivisceral
cavity in relation with the alimentary canal or with the
anterior part of it.’ The secondary body-cavity of Lang
corresponds to the pericardial cavity of Sedgwick. ‘ There is
also another cavity, which has no connexion with the peri-
visceral, and is called the pericardial because it is related to
the heart. By most anatomists the perivisceral is regarded as
haemococlic in nature. It is part of the vascular system, and
therefore haemocoelic.’

In regard to the Tethymelibidae Bergh (1908: 97),
writing on M. rosea Rang, says: ‘The cavity of the body
reaches to the region of the last of the papillae. In M.
leonina Gould, I find that the cavity extends anteriorly
to the oesophagus, dorsally to the back, ventrally beyond the
genital ducts, posteriorly as far as the anus, back of which
the branching of the liver and the kidneys is so profuse, together
with the crossmg of connective-tissue fibres, as to render it
very difficult to tell whether the perivisceral cavity extends
beyond the anus. This cavity is not a true coelom. It corre-
sponds with the primary body-cavity of Lang or the peri-
visceral cavity of Sedgwick. There is no definite termination
of an inner body-wall, although the muscle-wall seems to
represent one, but that is really superficial. Beyond the
muscle-wall the connective-tissue fibres run in all directions,
all through the cavity. It is, therefore, not a well-defined
cavity. It is in this so-called cavity that all the visceral

MORPHOLOGY OF MELIBE 533

bodies are located, hence the visceral, or perivisceral, cavity.
The pericardial chamber, discussed below, is a true coelom.

6. The Alimentary Canal.
(1) The Buceal Cavity.

As in other organisms the buccal cavity of M. leonina is the
beginning of the alimentary canal. It corresponds very closely
with Alder and Hancock’s (1864) and Eliot’s (1902) descriptions
of M. fimbriata. The mouth is bounded by the two
lateral, shehtly furrowed lips (Pl. 27, fig. 1, L, Pl. 29, fig. 17, M,
Pl. 32, fig. 35); the furrows are not seen in fig. 1 as they are
smoothed out by the swelling of the lips, but they may be seen in
fig. 835 which is a photograph of a transverse section through the
buccal cavity. Within the mouth there is a uniform invagina-
tion of the ectoderm (fig. 35), and this imvagination pro-
duces a number of folds or corrugations which increase in
depth and finally merge with those of the oesophagus. The
food, as it is engulfed, passes directly through the oesophagus
into the proventriculus. No masticatory process is carried
on in the mouth for the simple reason that this species is
absolutely void of tongue, radula, or mandibles. The food is
swallowed whole, as evidenced by the contents of the alimentary
canal including the intestime, and is disintegrated by the
digestive processes only. The jawless condition of this species
is a character common with that of its relative, Tethys
Linnaeus. Jeffreys says: ‘Tethys has neither jaw nor
tongue.’ And Vayssiere (1901) finds for Tethys fimbriata
(5. leporina): ‘Buccal bulb absent. Large anterior
chamber; having exterior circular muscle-fibres probably
used in mastication. (Cette région offre a sa surface extérieure
un anneau ‘musculaire, auquel correspond intérieurement un
anneau de plis longitudinaux presque tendineux, que l’on peut
considérer comme un organe masticateur.)’ However, this
jawless condition does not prevent Tethys or Melibe from
being carnivorous, as is shown by the contents of their stomach
(von Jhering, 1876: 37; Berg, 1852; Vayssiére, 1901: 84-5:
Ehot, 1902: 69; Agersborg, 1919: 272; 1921: 228, 282).

534 H. P. KJERSCHOW AGERSBORG

a. Mandibles and Radula.

Bergh (1902: 207) reports the presence of mandibles for the
species M. bucephala, saymg: ‘The mandibles joing
above are of a form like that of other Melibes. . . . The masti-
catory edge is finely dentate in the upper part, in the lower
part provided with coarser, rounded teeth.’ For the species
M. pellucida (1904: 13) he reports: *... die gelblichgrauen
Mandibel ganz zerbréckelt.’. And for M. rosea (1908 : 94-9)
he writes: ‘... through the walls the outlines of the mandibles
were very distinctly visible (fig. 3b). The clear yellow mandibles
(fig. 5) resembling those of f.ex., the Tritoniadae or Pleuro-
phyllidiae;... very plump denticulated masticatory edge,
the denticles reaching a height (fig. 5) up to 0-08 mm. (The
bulbus pharyngeus with its mandibles agree very likely in other
species of Melibes ; in general with that of the typical species.) ’
It is thus seen that Bergh finds mandibles nm many of his
Melibes, in fact, in all species of Melibe which he deseribed,
or nearly so. He even took issue with Hancock’s deseription.
It really seems strange that Bergh should be so imsistent on
this pomt. I have my doubt as to the correctness of his
description of one of the species (M. pellucida), collected .
from the mouth of Columbia River in the State of Washington,
as no other authors (Gould, 1852; Cooper, 1863; Fewkes,
1889; Heath, 1917; Agersborg, 1916, 1919, 1921, 19214,
1922, 1922a, 1923; and O’Donoghue, 1921, 1922, 19224) who
have collected the species from the same coast, 1.e. off the
coast of Santa Barbara, at Monterey, South-eastern Alaska,
Puget Sound, and the Vancouver Island region, have recorded
mandibles for the types with which they dealt. A number of
other authors who have described several species from different
parts of the world also, do not record mandibles for this genus :
Rang (1829), M. rosea; Pease (1860), M. pilosa, * mouth
proboscidiform, and the orifice vertical’; de Filippi (1867),
Jacunia papillosa (s. Melibe papillosa de Filippi) ;
Tapparone-Canefri (1876), M. papillosa ‘... nel suo
interno ha né lingua, né radula, né mascelle.. On the

MORPHOLOGY OF MELIBE 535

findings of Bergh, however, Lankester (1906: 175) character-
izes Melibe as having mandibles; but Lang (1896: 180)
says: ‘Jaws are wanting or rudimentary in... many Nudi-
branchia (Tethys, Melibe, Doridopsis, Phyllidia).’
According to this, it would be best to state, at least in part,
as a generic characteristic for Melibe: the pharyngeal bulb is
either with or without mandibles ; radula and tongue always
absent.

The alimentary canal is remarkably straight (Pl. 28, fig. 9),
in fact there is no coiling or looping whatever ; only the intes-
tine curves a little from the median position and to the anus,
which opens on the right side, a little out of the median line.
In this way, the alimentary tract extends somewhat diagonally
through the body-cavity, and only the mtestine curves. This
corresponds with Lang’s (1896: 33) statement for the Nudi-
branchia: ‘The anus lies either dorsally in the median line,
or laterally to the right. In M. leonina, the anus is on
the right side of the body (PI. 27, fig. 2), a little ventral to the
base of the second papilla of that side. This is also its position
in M. bucephala Bergh (1902), though in this case it is
midway between the first two anterior pairs of the papillae.
In M. leonina the ureteric pore is laterodorsal to the anus.

b. The Buccal and Salivary Glands.

The buceal cavity is highly corrugated (PI. 32, figs. 35, 36, Oe),
but it is without jaws or radula. Specialized organs for
chewing are substituted by the foldmg and invagination of the
ectoderm. The folds are non-glandular, but just beneath this
ectodermal layer are numerous glands, even in the external
parts of the mouth. These glands correspond to the buccal
glands of various authors. Lang (1896: 185) distinguished
clearly between these glands and the salivary glands and says,
referring tothe Opisthobranchiata: ‘The salivary glands,
of which only one pair is almost always found, here vary in
size and shape still more than in the Pulmonata. Those
glands which enter the pharynx must not be confounded with

536 H. P. KJERSCHOW AGERSBORG

other glands which in many Opisthobranchiata enter
the buceal cavity, and are sometimes more strongly developed
than the salivary glands.’ Sedgwick (1898: 371), referring to
the same organs, says: ‘In addition to buccal glands, some-
times found round the buccal opening, there is always a pair
of salivary glands opening into the buccal cavity. The buccal
cavity leads into the oesophagus which is followed by a dilated
stomach, and is usually provided with a caecal appendage.’
In M. leonina the glands of the oesophagus consist of a
series of small, simple, saccular glands (Pl. 32, fig. 41, Sg)
arranged in rows along each side of the swallowing tube. They
open directly into the oesophagus by small crypts (V). Heath
(1917: 147) reports that the salivary glands are absent in
Chioraera dalli (s. Melibe leonina), but, of course,
he is mistaken about this. The activity of the salivary glands
in gasteropods was beautifully demonstrated by Lange (1902:
85-153), who showed there is a close relation between the
structure of the nucleus and cytoplasm and the physical
condition of the organism relative to starving and feeding of
the animals.

(2) The Oesophagus.

The length of the oesophagus in an animal circ. 10 em. long
is 8 mm. (Pl. 32, fig. 36, Oe). The oesophagus itself is simply
a narrow part of the alimentary canal between the mouth and
the proventriculus. Corrugations, which begin at the lips and
increase progressively in the mouth, deepen still more in the
oesophagus. The lining is still non-glandular, but the glands
in the underlying tissue increase until the anterior part of the
proventriculus is reached, when they end quite abruptly. The
corrugations of the oesophagus are largely longitudinal, which
suggests that the oesophagus is capable of expansion in case
the animal swallows some large food particle. In some nudi-
branchs, e.g. the Tritoniadae, according to Vayssiere
(1877), the oesophagus is very long.

MORPHOLOGY OF MELIBE 537

(3) The Stomach.

a. Proventriculus.

Following the oesophagus, the alimentary canal enlarges
into a chamber 3mm. long, 1:51 mm. in diameter, and is
constricted posteriorly by extensive evaginations of its epithelial
lining. This chamber, the glandular stomach or proventriculus,
constitutes the preliminary digestive cavity of the alimentary
tract. Its epithelial liming is distinctly different from that of the
oesophagus, by bemg highly glandular (Pl. 32, fig. 38, Gi).
The food is probably kept here until acted on by secretions
of the unicellular glands of its columnar epithelial lining.

b. Gizzard.

The remaining part of the stomach is the gizzard. Its
length in an animal 12 cm. long was 10 mm., with a diameter
of 4mm. at its widest part. The structure of the gizzard is
variable. Its walls consist of two coats (Pl. 32, figs. 37, 42, 43)
or layers, each of which may be divided into two parts. The
outer layer consists of a thin outer cover of connective tissue
and occasional muscle-fibres, which run longitudinally with the
organ (Cc), and of a thick median circular layer (Mus). The
inner layer consists of a single layer of tall columnar epithelial
cells of glandular nature (Hpt) and a false epithelial cell border
(Stpl), formed from the secretion of the underlying glandular
epithelium. ‘This secretion fuses nto a homogeneous mass
giving the appearance of a transitional epithelial layer of
cornified type (Pl. 32, fig. 42, T's, 48, Trp.pl). This cornified
border is about as thick (Pl. 32, fig. 37) on the dorsal side of the
middle part of the gizzard as the other two layers together,
and gradually decreases until, on the ventral side, .. .1t becomes
very thin (V). It is far more marked in the anterior part of
the gizzard than in the posterior. The glandular nature of
the epithelial border is well marked in the anterior part of the
gizzard but decreases in the ventral region of this part of the
stomach, where the cornified border also entirely disappears.

538 H. P» KJERSCHOW AGERSBORG

In the extreme posterior part of the stomach the epithelial
lining becomes ciliated, a feature which is continuous through-
out the remainder of the alimentary canal. In this respect the
alimentary canal of M. leonina differs greatly from that
of Neritina fluviatilis, a prosobranch, which according
to Lensen (1899) is ciliated from the oesophagus to the anus.
The stratum corneum of the epithelial border of the gizzard
represents in M. leonina the so-called stomach-plates of
various authors, which are supposed to be a common charac-
teristic of Melibe and related forms. ‘Thus, Alder and
Hancock (1864), in their description of M. fimbriata,
say : ‘ The stomach is a rather large pyriform pouch, with its
small extremity placed backwards. It hes diagonally across
the anterior portion of the visceral cavity and is divided into
an anterior and posterior chamber by a slight constriction near
the centre. The anterior of the lower portion of the chamber
is encircled transversely by an almost complete belt of horny,
compressed, lancet-shaped process, similar to those im the
gizzard of Scyllaea. Bergh (1875b), referrmg to Tethys
and Melibe, says: *... und das in der Tiefe des Kopf-
trichters hervorstehende Mundrohr leitet unmittelbar in die
Speiserdhre und in den ersten Magen hinein, der, wie bei den
Meliben, mit starken (Cuticular-) Falten bewaffnet ist,’ p. 346.
And again on p. 356: * Die Innenseite der mittleren klemeren
Abtheilung des Magens zeigt ...eme Masse von sehr klemen
und niedrigen Lingsfalten (fig. 1 ¢) oder eme geringere Anzahl
(15-20) von stiirkeren; die Falten sind mit emer horngelb-
lichen Cuticula tiberzogen, die im ersten Falle nur von germger
Dicke, im anderen viel stiarker, mit Leisten sich zu emer Hohe
von etwa 0-5-0-75 mm. erhebt; an den dickeren Cuticula-
Falten tritt eine (Taf. XLV, fig. 22) deuthche Querstreifung
oder mehr unregelmissige Theilung hervor. Die Unterseite
des Cuticula-Ueberzuges zeigt em fein k6rmiges Aussehen,
das auch an der Oberfliche der unterliegenden Schleimhaut
hervortritt und von den zahlreichen, dicht gedraéngten Papillen
derselben hervorgebracht wird.’ Again, on p. 366, referrmg to
M. capucina, he says: ‘ Magenziihne . . . 10 starken Kielen

MORPHOLOGY OF MELIBE 539

gebildet.’ Finally, referrmg to M. rangii, he finds: *.. . die
Magenzihne viel zahlreicher als bei der vorigen Art, schmiler
und im ganzen von einem Individuum 26, bei dem anderen 31.’
Also von Jhering (1876), writing on Tethys, says: * Der
Magen trigt nach innen von der Faser- und Muskelschicht ein
einfaches EXjpithel 0-02 bis 0-03 mm. grosser Zellen, und darauf
folgt nach innen eine oft mehr als 0-3mm. dicke Schicht,
welche aus einfachen, schlauchf6rmigen 0-014 mm. dicken
Driisenschliuchen besteht, deren histologischer Bau... nicht
sagen lisst ob es gestattet ist, sie mit der bekannten Cuticular-
schicht im Muskelmagen der Voégel zu vergleichen. Diese
Schicht erimnert sowohl in ihrem Aussehen als in ihrer Consis-
tenz an Knorpel, indem sie zwar weich und sehr elastisch
ist, aber doch durch diese an Gummi ermnernde Elasticitat
dem Magen denselben Schutz gewahrt, wie eine harte Ialk-
oder Chitinauskleidung.’ And Vayssiére (1877: 300), in his
description of a new genus of the family Tritoniadae,
writes : ‘ L’cesophage, qui est trés long, aboutit & une premiere
dilatation qui est le gésier : ¢’est dans lintérieur de cette cavité
que se trouvent pres de quarante dents cultriformes, placées
cote A cote et formant un anneau complet. Ce caractére ne
se montre parmi les Nudibranches que dans le genre Scyl-
laea.’ And again (1911: 102), working on the species
Bornella digitata, he records the followmg: ‘* Cette
poche dans son tiers antérieur posséde des parois musculaires
assez épaisses, muscles transverses et longitudinaux ; cette
forte musculature est destinée a soutenir a Vintérieur une
quinzaine de rangées longitudinales de longues épines un peu
recourbées, de nature chitineuse. Cette partie de la poche
constitue un véritable gésier dont l’armature sert a broyer les
aliments qui y arrivent.’ This is further considered by Bergh
(1879a: 165) in his description of M. vexillifera: * Der
hintere Teil des Magens ist kiirzer als der vorige und hat
festere Wande; an seiner Innenseite (fig. 9c) 14 starke
Magenplatten und zwischen denselben meistens eime (seltener)
kleinere und meistens kiirzere (fig. 9c) ; die Platten, die Falten
darstellen, an denen die Cuticula stirker entwickelt ist,...

540 H. P. KJERSCHOW AGERSBORG

schwach harngelber Faibe eine Liinge bis beilaufig 0-8 bei eimer
Hohe bis etwa 0-20, und einer Breite bis 0-25 mm. erreichend.
Das hintere Ende des Magens (fig. 2) mit von etwa der Mitte
eradurenden Falten, neben dem Pylorus hier eme taschen-
formige Erweiterung (fig. 2fy) (etwa wie in Tethys) mit
starken Falten der Innenseite.’ In 1888: 691, the same
author, describmng M. ocellata, writes: ‘.. . die Magen-
platten schimmerten undeutlich durch; diese letzteren fast
wie in der M. papillosa.’ And for the last-named species
(1884a): ‘.. . hinter der Mitte (des Magens) der Linge
nach schimmert der Zahngiirtel undeutlich hindurch.” And in
1902: 107, for M. bucephala: ‘ The belt of the stomach-
plates shmes through in about the first half of it, and
immediately before the belt the foremost liver-branch is
attached on either side somewhat upwardly. ... The belt of
plates consists of twenty-eight faint lemon-coloured firm
plates partly alternating in height.’ Eliot (1906), describing
Tritoniopsis, says in regard to the stomach: ‘.. . mto
a rather small membranous and fragile stomach, almost
entirely covered by the liver, and no trace of plates.’ And in
1910: p. 40 he writes: ‘ The liver secretions harden in the
stomach and form a protecting membrane which is found to
cover the stomach.’ On the previous page he stated: * Into
the posterior part open four or five liver-ducts and also a
pear-shaped gastric pouch, whose orifice in the stomach-wall
is closed with a more or less distinctly developed flap. This
pouch is often called the gall-bladder, but nothing indicates
that its functions correspond to this name. Its walls are
glandular, and appear to secrete globules of a glistening material
which is also found in the intestine. It is possible that this
secretion subsequently dissolves and forms a membrane
which is found to cover the walls of the stomach and intestine,
and probably serves to protect these delicate surfaces against
the spicules abounding in the sponges on which most Dorids
feed.’ Stomach-plates constitute a common feature among
species of Melibe and other related forms; something
similar to stomach-plates is present in other species (Eliot,

MORPHOLOGY OF MELIBE 541

1910). The origin of the stomach-plates in M. leonina,
which no one seems to have described, is not the same as
indicated by Eliot (1910: 39-40) for the Dorids. This may be
plainly seen in sections which are represented in my drawings
(Pl. 32, figs. 42, Trs, 43, Trp.pl). In the case of Melibe the
plates origmate as a secretion-product of the epithelial lining
of the gizzard. This secretion accumulates within the cell
and is voided, little by little, by the cell into the cavity of the
stomach. As these secreted droplets pass into the stomach
they become attached to their predecessors, harden, and form
into a continuous layer resembling the stratum corneum of the
human epidermis but, of course, arises differently. The
secretion droplets seem to origimate around the nucleus of the
epithelial cells; they then coalesce into larger droplets and
fmally break up into smaller ones that pass out of the cell
and into the stomach and then form by a keratinization (?)
process into the hard or protective linmg of the stomach.
This protective lming, the stomach-plates of Melibe, no
doubt serves a double function: to protect the living cells
against the spines of the crustaceous food ; to help in mastica-
ting the food before it passes into the intestine. This latter
use seems to be necessary, particularly in those cases in which
the organism is void of mandibles or radula of any kind.

é¢. The Pyloric Diverticulum.

The pyloric diverticulum (Pl. 27, fig. 9, Pd,) is deseribed
by Alder and Hancock (1845: p.14) as pancreatic in function.
Whatever function it may have, it seems logical to think that
it plays a specific part in the process of digestion because of its
internal structure. This part of the alimentary canal is situated
at the constricted posterior part of the stomach. Externally it
consists of an elaborate evagination into a number of folds,
beginning laterally and continuing ventrally, until meeting on
the opposite side. Internally (Pls. 832 & 33, figs. 89, 44) the
pyloric diverticulum is considerably corrugated, being thrown
into much larger folds than the remainder of the tract. Ciliation
of the mucous layer begins here. The corrugations are formed,

542 H. P. KJERSCHOW AGERSBORG

as in other animals, by the connective tissue which mtervenes
between an outer muscular layer (although in this case the
muscular layer is very thin) and an inner mucous layer. The
latter consists of ciliated columnar epithelium, and, as in the
stomach, is glandular in nature (PI. 33, fig. 44), but the secretion
of the intestine does not harden after it reaches the lumen of
the tract. The epithelium contains large and small vacuoles
which are formed in the neighbourhood of the nucleus but
in the distal region of the cell. As m the stomach, these
vacuoles or secretion droplets coalesce into larger ones
which either break up into smaller droplets before leaving the
cell, or pass directly into the lumen. but in no case does the
epithelium form into regular goblet-cells, nor does the secretion
form into protection-plates of the liming as in the stomach.

(4) The Intestine.

That the intestine is the principal digestive region of the
alimentary canal is well shown by the different conditions of the
food in the stomach and in the intestine. In the former, the
food seems to have undergone little or no disintegration,
while in the latter, only the skeletal parts of the food remain.
As the food passes through the pyloric diverticulum, it is
perhaps acted upon in such a way that the intestinal juices
more easily complete the digestion that takes place in the
posterior part of the alimentary canal. The slightly enlarged
part of the anterior portion of the intestime has parallel corruga-
tions externally, which may go to show that it is capable
of considerable enlargement. In specimens 7 cm. long the
corrugated enlargement was circ. 7mm. in diameter.

The absorptive surface of the mtestine is mereased very
greatly by the presence of a typhlosole which extends from the
pyloric diverticulum to the anus. The typhlosole is very large
in the anterior portion of the intestine, where it protrudes into
the intestinal cavity from the ventral side until there is little
free space left between it and the rest of the intestinal walls.
Posteriorly, the typhlosole gradually decreases in size, and just
before reaching the anus it 1s obliterated in the ventral part of

MORPHOLOGY OF MELIBE 543

the tract. Only the ordinary corrugated part of the intestinal
lining remains. The extent of the typhlosole, however, varies,
because one specimen, whose intestine was sectioned, had no
typhlosole in the smaller part of the canal.

The structure of the intestine is similar throughout. There
is an outer fibrous layer and an inner glandular one. Between
these are fine connective-tissue fibres and small cells and
colourless lymph. The glandular layer consists of tall ciliated
columnar epithelium (PI. 33, figs. 45-7). There is no difference
in the glandular layer of the typhlosole and of the remainder
of the intestine, or in these parts of the pyloric diverticulum.
A close study of the internal layer of the mtestine reveals some
interesting morphological facts, viz. this layer consists of very
tall columnar cells with the nucleus located, in most cases,
in the middle (PI. 33, figs. 45, 47) with vacuoles either around
the nucleus, on either side, or on the distal side only. The
vacuoles arise as a confluence of smaller vacuoles which arise
in the neighbourhood of the nucleus, and again break up into
smaller ones and then pass into the lumen of the intestine.
The epithelial secretion does not keratimize here as it does in
the gizzard. Eliot (1910) suggests that the hepatic secretion
hardens in the intestine and the stomach of Dorids. The
hard substance, however, which covers the endoderm of the
alimentary tracts of Eliot’s Dorids and of M. leonina, is
at least in the case of the latter of an entirely different origin.
Sometimes more than one nucleus may be present in the same
cell (PI. 33, fig. 46). The most striking feature of this epithelium
is the regular fibrillar structure, and the hnear arrangement
of the cytoplasmic granules from border to base of the cell
running parallel with the cilia; these do not converge on the
nucleus but pass to the base of the cell. The cell rests on
a distinctly granular basement membrane (PI. 33, fig. 45, Bm).
There are two distinct rows of basal granules, or terminal bar.
(desmochondria), the proximal being the larger of the two (5q).
The cilia (Cil) may be seen readily between the two terminal
bars. Beneath the basement membrane (Bm) is a loose
connective-tissue layer (cc) which is covered by a denser

p44 = H. P. KJERSCHOW AGERSBORG

fibrous coat with a few occasional muscle-fibres (Cc). This
same cover, in the pyloric diverticulum, is very loose (Pl. 33,
fig. 44, Vas), and suggests a possibility of interchange of body-
intestinal fluids in this part of the alimentary canal. The outer
coat of the remainder of the intestine, though much thicker,
is, however, so loose that it may allow ready interchange of
intra- and extra-intestinal fluids.

There are then, five distmct regions of the alimentary
canal each differing from the other in structure and function.
These are (1) the oesophagus with the non-glandular lning
below which are the oesophageal or salivary glands (Pl. 33,
figs. 36, Oe, 41, Sg); (2) the proventriculus with distinct
glandular lining (Pl. 33, figs. 36, G, 38, Gl); (8) the gizzard
with its stomach-plates (Pl. 38, figs. 37, 42-3, Stpl); (4) the
pyloric diverticulum with its glandular and ciliated imternal
surface, which secretions, as in the glandular stomach, do not
keratinize (PI. 33, figs. 38, 39, 44); (5) the mtestine with its
large typhlosole and glandular ciliated epithelial surface
(Pl. 83, figs. 40, 45, Pl. 34 figs. 56, 57).

(5) The Liver.

The gastro-hepatic apparatus, or hepatic caeca of M.
leonina, does not arise asin M. fimbriata, °...a little
in advance of the belt of horny processes ’ (Alder and Hancock,
1864), but it arises from the anterior portion of the gizzard
and consists of a very extensive arborization which passes to
all parts of the body (figs. 1, 2, Hp, 7, 9, He, 40, Hep).
As in other members of this genus, the liver consists of three
principal tubular trunks which start at the anterior end of the
gizzard and pass to the various parts of the body. Two of
these trunks, situated opposite each other, send out branches
as follows: the one on the left side runs m the main to the
gonads, branching very profusely in that region; it also sends
out other but minor branches, some to the papillae, some to the
body-wall, and one to the mucous gland. (The mucous gland
= (1) albuminous gland, and (2) nidamental gland.) The right
trunk sends out several major and minor branches. Of the

MORPHOLOGY OF MELIBE 545

former, one goes to the mucous gland and the other to the
gonads and prostata; of the latter, one goes to the veil and
others to the papillae and the body-wall. Both of these trunks
branch profusely in the posterior part of the body-cavity.
A third trunk situated in front of the other two on the left
anterior side of the stomach sends out two main branches,
one to the hood and the other to the left papilla of the first
pair. Besides these there are many minor branches which go
to the body-wall. The extreme branches, and particularly
those lying within the body proper and surrounding the gonads,
are highly glandular. The contents of the gizzard are affected
by the secretion of the liver, which is shown by the fact that
particles in the stomach near the hepatic openings stain very
similarly to the hepatic tracts through the stomach-plates.
The arrangement of the hepatic system in the different species
of this genus is considerably variable, and still in some of them
it is very similar to that of M. leonina. Thus Bergh
(1875b: 366) finds for M. capucina: * Die Leber scheint
eime lose, aus mehreren grossen, unregelmiissigen, lose mit
eimander verbundenen Lappen gebildete, etwas gelbliche Masse
zu sein, die sich durch die grésste Strecke der Hingeweidehohle
hinzieht, fast tiberall an den Lappen der Zwitterdriise ange-
heftet ist und, wie es scheint, auswirts gegen die Kéorperwand
(gegen die Papillae ?) kurze dicke Aeste ausschickt. Aus dem
vordersten Theile der Leber entspringt der ziemlich weite
Gallengang (fig. 19), der in die Riickenseite des (Cardeatheils
des) Magens eimmiindet.’ In 1879a: 165, for M. vexilli-
fera: *...die Leber wie bei anderen Meliben eine lose, gelblich-
weisse Masse, welche vorne an den Magen reicht, hinten sich
bis an das Ende der Eingeweidehéhle erstreckt ; sie ist eine
sehr stark verastelte, mit gerundeten, diimnwandigen Endkolben
und Ausbuchtungen versehene Driise (fig. 10), von welcher sich
aber zwei Lappen ganz abgelést hatten, die sich im ersten Papil-
lenpaare verbreiteten und in die vordere Abtheilung des Magens
einmiindeten. Diese Lebermasse ging vorn in einen ziemlich
weiten, kurzen, dimnwandigen, gemeinschaftlichen Gallengang
iiber.” The hepatic system in M. papillosa Bergh (1884)
NO. 268 00

546 H. P. KJERSCHOW AGERSBORG

is not so completely broken up as in M. leonina. Again
(1890b: 283) he calls attention to the fact that the liver of
M. ocellata opens into the stomach behind the stomach-
plates, and in that way it is seen to differ from M. leonina.
But the distribution of the hepatic branches is similar : * Dicht
hinter dem Giirtel der Magenplatten miimnden die dicken
Leberstimme cin, rechts der besonders dicke aus der ersten
rechten Papille, links der aus der entsprechenden linken, und
dicht neben demselben der grosse Hauptleberstamm, lings
der Riickenseite der Zwitterdrise und wher dieselbe hinaus
verlaufend.’ Finally, describing the species M. rosea Rang,
he shows (1908: 98) that it is quite similar to Chioraera
Gould, when he writes: ‘The three principal liver-branches
with their ramified hepatic ducts and the principal branchlets
to the dorsal epinotidia as usual... . Network of liver-branches
is interwoven with the much branched renal tubes (figs. 9, 10)
the branches reaching the root of the epinotidia, but did not
seem to ascend into them.’

Pease (1860: 34) refers to the liver in M. pilosa, only
by stating: ‘... body punctured with brown, which are most
conspicuous along the flank.’ And Alder and Hancock (1845:
13), writing on this subject, say in part: ‘In the greater
number of Eolididae (Aeolidiidae), however, the liver
has entirely disappeared from the abdomen and is broken up
into numerous: minute portions or glands which are thrust
into the branchial papillae. The delicate ducts from these
glands pass onward and unite to form great hepatic ducts or
trunk channels, which open into the stomach.’ Hertwig
(1912: 335) concurs in this by saying: ‘In Aeolidae
(Aeolidiidae) branches of the digestive tract enter the
cerata, expand distally to small sacs filled with nettle-cells
used for defence ; they are derived from hydroids on which
these animals feed.’ So also Lang (1900: 300) writes: ‘ Bei
zahlreichen Nudibranchiern lést sich die Verdauungsdriise in
sich veristelnde Darmdivertikel auf, die sich fast nach Art
der Gastrokanile oder Darmiste der Tubellarien in Ké6rper
ausbreiten und bis in die Riickenanhiinge des K6rpers empor-
steigen (cladohepatische Nudibranchier), wo sie mit den

MORPHOLOGY OF MELIBE 547

Nesselkapselsicken communiciren kénnen Diese Form der
** Leber’? macht wahrscheinlich, dass sie nicht etwa bloss ver-
dauende Secrete absondert, sondern sich auch selbst bei der Ver-
dauung und bei der Resorption der Producte der Verdauung be-
theiligen wird. In der That weiss man schon lange, dass bei
den Nudibranchiern Speisebrei in diese Verastelung des Darmes
hineingelangt ; aber auch fiir eme Form mit ganz compacter
Leber, naimlich fiir Helix pomatia, wurde kirzlich der
Beweis erbracht, dass in der That in der “‘ Leber” Aufsaugung
oder Resorption der verdauten Nahrung stattfindet.’
Quatrefages (1844, 1844a, 1848) maintained that the liver
in Nudibranchs is of a threefold function; hence his term
*Plebenterism’ to designate that species of gradation which
consists in the union of different functions in one system of
vessels. That is, he maintained the absence of anal opening,
heart, and blood-vessels, adopting the term gastro-vascular
system introduced by Milne H. Edward (1842, 1845) for the
digestive organs in the family Aeolidiidae, the true
significance of which has since been the subject of much
controversy. It is now, however, a well-established fact that
the group of molluscs with which de Quatrefages dealt
(Aeolidiidae) has a well-established circulatory system,
i.e. heart and blood-vessels, and alimentary tract with anal
opening. The liver branching off from the digestive tract forms
into many parts and ramifies to various parts of the body.
One unquestionable function of the liver, as far as Aeolidia
is concerned, is an exit for harmful and indigestible parts taken
in with food (Alder and Hancock, 1845; Glaser, 1903; Hert-
wig, 1912). Glaser describes the hepatic caeca as secondary
exits in nudibranchs which feed on hydroids whose nemato-
cysts produce indigestible formic acid ; the mollusc rids itself
of its useless stomach contents through these secondary
openings of the liver-branches which end in the dorsal papillae.
M. leonina does not feed on hydroids, but on crustaceans
(Agersborg, 1916, 1919, 1921, 1921a, 1922a, 1923); and,
although the hepatic system is tubular (PI. 30, figs. 25, Cshb,
31, Ch; Pl. 338, figs. 51, 58), it does not end with openings
002

548 H. P. KJERSCHOW AGERSBORG

through the ectoderm (Pl. 27, figs. 1, 2, 4, 7; Pl. 31, fig. 30),
but caecally between it and the muscle-wall (Pl. 27, figs. 2, 3).

Frenzel (1886 : 273) beleved with other authors that the liver
of molluses performs a double function: (1) as in Crustacea
it is a digestive gland, ‘d.h., dass sie ein Secret bildet und
ausscheidet, welches zur Verdauung der in den Darmkanal
aufgenommen Speisen verwendet wird.’ (2) In addition, this
gland is according to Max Weber (1880) for the Crustacea,
and according to Barfurth (1883) for the Gasteropoda, of
‘excretorische Function’. They think that the liver of these
forms is analogous to that of vertebrates. They describe cells
that have special functions, such as secretory and excretory.
Frenzel points out three kinds of epithelial cells of the liver
of Tethys: (1) ‘ Kornzellen ’, (2) ‘ Keulenzellen ’, (3) * Kalk-
zellen’. ‘To these different cells he ascribes the various fune-
tions of the organ. These cells are further described by Hecht
(1895: 675), as follows: ‘La présence de trois types bien
définis de cellules: (1) Cellules vacuolaires excrétrices carac-
térisées par leurs grandes dimensions et leurs grandes
vacuoles (Frenzel: Fermentzellen, Keulenzellen) contenant
chacune une granulation. ...; (2) Cellules exerétrices a grosses
sphéres brunes (Leberzellen, Kornzellen); (3) Cellules a
ferments. Leur coloration en gris par les réactifs osmiques ;
on y joindra; (4) Cellules indifférentes qui, je le suppose,
peuvent évoluer dans un sens ou dans l'autre.’ The structure
and function of the liver ina Doridiform cladohepatie
nudibranch, is still further commented on by. Ehot and
Evans (1908) as follows: ‘The cells which line the hepatic
lobules are columnar or cuboidal and highly granular. Some
are in a distended condition, others are attached to the wall of
the.lobule only by a strand or are free in its cavity. It would
seem, therefore, that some of the liver cells are excretory in
function, and are dropped into the follicle as they become
extended with excreted material.’ Eliot (1910: 39) attributes
to the liver the function which, in the case of M. leonina,
IT have shown to be the function of the epithelium of the
posterior chamber of the stomach, i.e. the gizzard, viz.:

MORPHOLOGY OF MELIBE 549

‘Into the posterior part open four or five liver-ducts and also
a pear-shaped gastric pouch, whose orifice in the stomach-wall
is closed with a more or less distinctly developed flap. This
pouch is often called the gall-bladder, but nothing indicates
that its functions correspond to this name. Its walls are
glandular, and appear to secrete globules of a glistening material
which is also found in the intestine. It is possible that this
secretion subsequently dissolves and forms a membrane which
is found to cover the walls of the stomach and intestine, and
probably serves to protect these delicate surfaces against the
spicules abounding in the sponges on which most Dorids feed.’
Finally, Arnold (1916: 353-4), referrmg to the Clado-
hepatica, thinks that the liver, which in most nudibranchs
is extremely large and completely surrounds the stomach,
in Dendronotus also extends into the dorsal cerata
(papillae), so that they may have some digestive function.
The hepatic diverticula of M. leonina consist structurally
of two main layers: an outer fibrous coat and an inner
columnar epithelial membrane. The fibrous layer, consisting
of connective tissue, is thrown into corrugations—inwardly—
so that the surface of the lumen of the hepatic appendages is
sreatly increased (PI. 38, figs. 51, 53). The cytoplasm of the
epithelium is highly glandular and one may notice considerable
variation in the contents of the cells. Some of the cells show
a similarity to the ‘ Keulenzellen’ of Frenzel, or ‘ Cellules
vacuolaires excrétrices’ of Hecht. In fact, as is shown in
Pl. 33, figs. 48, 49, 50, some of the cells have large vacuoles
containing granules of different sizes (Pl. 33, figs. 48-9), and
some are vacuolated and contain no granules, while others
have no vacuoles but their cytoplasm is highly granular.
Others, again, show a remarkable lnear arrangement of the
granules, basal to the nucleus (PI. 33, fig. 49). The greatest
activity of the cell seems to take place around or near the
nucleus with a progressive differentiation toward the border,
where the cell in many cases is more homogeneous than in the
remaiing part. Sometimes a large secretion vacuole may
contain threads of darkly stained substances with a darkly

550 H. P. KJERSCHOW AGERSBORG

staining cap toward the periphery. Such threads may be
present in the cell without the cell beimmg vacuolated. The
secretion product, at other times, aggregates from a number of
small vacuoles at the border of the cell and passes out en
masse, when it seems to be basic in staining reaction. That
is, the basophil reaction is shown in material which has been
fixed in an osmic acid mixture and stained with Heidenhain’s
haematoxylin. The nucleus, as a rule, is basal in position
and contains one or two nucleoli (Pl. 33, fig. 48). The morpho-
logical aspect of an actively functioning liver of Physa
gyrina Say, and Planorbis trivolvis Say, does not
differ much from that of Melibe leonina (Gould) (vide
Agersborg, 1923 a, figs. 35, 36, 38, 41, 42). It is not my purpose
at this time to discuss the function of the hepatic system of
M. leonina; but it is evident from the facts observed, and
as pointed out above, that at least some of its products passes
into the stomach, and for this reason it is secretory in function.
Perhaps, owing to the peculiar granular nature of the cytoplasm
of the epithelial layer, it may be absorptive in function also.
In that connexion it is interesting to note the extensive
distribution of the liver to the various organs of the body, the
significance of which at the present time may only be con-
jectured. The pecuhar granular nature of the cytoplasm of
the epithelium shows that its function is different from that
of the epithelium of the alimentary tube proper.

7 The Circulatory System.

The work of Milne Edwards (1842) on Aeolidiidae;
Alder and Hancock (1845) on the Aeolidiidae and
Tritoniidae; Hancock and Embleton (1848) on Aeolidia,
(1881) and (1882) on Doris; Hancock (1865) on Doris
tuberculata, D. reponda, D. bilamellata, Tri-
tonia hombergii, Bornella, and Scyllaea; Bergh
(1875) on Melibeand Tethys, (1884a) on M. papillosa;
Lansberg (1882) on Neritina; Boas (1886) on Pteropodes
(Limacina, and Cleodora acicula); Sedgwick (1888)
on Peripatus, (1898) on gasteropods ; Bouvier (1891) on

MORPHOLOGY OF MELIBE 551

Opisthobranchiata; lLankester (1893) on gasteropods
and other molluses ; Goodrich (1895) on nematodes, chitons,
Peripatus, &c¢.; Hecht (1895) on nudibranchs; Lang
(1896) on the Mollusea, (1900) nudibranchs ; Shipley and
Macbride (1915) on the molluscs; and others, have thrown
a great deal of light on the nature of the organs of circulation
in Invertebrata, particularly the works of Sedgwick
(1888), Lankester (1893), Goodrich (1895), and Lang (1900).

(1) The Pericardium.

According to Hancock (1864 : 513) the so-called pericardium
in Dorids lies immediately above the renal chamber and
directly below the dorsal skin in front of the branchial circle.
It is, with the exception of the opening leading into the
pyriform vesicle, a closed membranous sac, formed apparently
by what has been designated the peritoneum, and is just
sufficiently large for the accommodation of the dilated auricle
and ventricle. It is lined with its own proper membrane
which is closely adherent to and intimately confounded with
the peritoneal membrane, but can be observed reflected upon
the heart at the root of the aorta. It has just been stated that
this cavity is closed—previous communications to the contrary
were erroneous owing to defective material used.

Goodrich (1895: 484-6) maintains that the vascular system
or blood system is simply a liquefaction, as it were, of the
mesoblast (Lankester’s view). This corresponds with facts
as found in Diploblastica, e.g. (adult) Coelenterata,
where blood-spaces are entirely absent, *. .. while as to the
nematodes, .. . it seems probable the body-cavity is a blood-
space, corresponding in relation to the parenchyma of the
planarians,’ He shows, following Erlanger, how the peri-
cardium in Paludina arises as two coelomic sacs on either
side—a hollowing out of the mesoblast. These coelomic
cavities then fuse, and later by processes of special growth
form a peritoneal funnel that opens to the outside on either
side. ‘The gonad develops from the wall of the coelom ;
then together with the rudimentary left peritoneal funnel, it

552 H. P. KJERSCHOW AGERSBORG

becomes constricted off from the main division of the coelom
(the pericardium), forming a small genital sac. From the wall
of this sac the genital duct grows out, and joins an epidermal
invagination like the peritoneal funnel of the right side.’ In
Melibe the gonads are situated ventrally with openings on
the right side; the single kidney (nephrocoel) is situated
dorsally and communicates internally with the pericardium
through the renal syrmx and externally through the ureter
on the left side of the anal pore. The perigonadial coelom
(perigonadium) and pericardium are completely separated by
the visceral cavity or vascular (haemocoels) cavities. These
cavities are blood-spaces into which the blood percolates from
the atrial vessels bathing the visceral organs. According to
Goodrich (1895), in Chitons, a separation has taken place
in the genital region of the coelom from the renal ; the gonads
then require special ducts which may not be homologous with
the peritoneal funnels. In Peripatus, soon after the meta-
meric somites have been hollowed out from the coelomic
follicles, the upper half of each coelomic cavity becomes nipped
off from the lower half. From the wall of each of these lower
coelomic sacs a peritoneal funnel is formed as an outgrowth
which fuses with the epidermis. While these organs have
developed in this way, the dorsal or genital halves of the somites
in the posterior segments have become fused, forming two
genital tubes communicating posteriorly with the undivided
coelomic follicles of the last segment. The peritoneal funnels
of this segment retain their primitive function and develop
into the genital duct ; Sedgwick (1898: 375) finds that the
coelom of the Gasteropodaisin three sections. (1) The peri-
cardium ; (2) the nephridia ; (3) the gonads. The pericardium
is in relation with the heart; it normally communicates with
the nephridial system, and part of its lining is generally
glandular and forms the pericardial gland. It has no con-
nexion with the blood system. Finally, Lankester (1898:
428) points out the followmg: the perigonadie spaces and the
pericardial space are, then, the coelom of the Mollusea.
It is quite distinct from the haemocoel. In cephalopods, and

MORPHOLOGY OF MELIBE 553

in the archaic gasteropod Neomenia, the pericardial and
perigonadial coelomic remnants are continuous, and form one
cavity. There is strong reason to believe that in ancestral
molluses the haemocoel was more completely tubular and truly
vasiform than it is in living molluscs. In the later molluscs
the walls of the vessels have swollen out in many regions
(especially in the veins) and have obliterated the coelom,
which have shrunk to the small dimensions of the pericardium
and perigonadium. There are, however, many molluscs with
complete capillaries, arteries, and veins, in certain regions of
the body.

(2) The Heart and Arteries.

While the intestine in M. leonina, by its diagonal course
through the visceral cavity, disturbs the apparent bilateral
symmetry, the heart is situated in the median line, just anterior
to the anus (PI. 31, fig. 33), and to the left of the intestine. The
heart consists of two chambers, a dorsal and a ventral (Pl. 34,
fig. 55, Au, Vent). The dorsal chamber is the smaller of the
two; it is partitioned off into small spaces through which
the blood is returned by the efferent branchial vems (Au).
These chambers may be called auricular chambers ; they are
perforated (Av), and the partitions (Ar) which may serve as
valves may also close the perforations. The partitions or
valves with their apertures are so arranged that the openings
do not coincide with each other, and are therefore easily
closed. The ventricle or the larger of the two cardiac chambers
has a regular valve at its lower and constricted portion (PI. 34,
fig. 59, Valve). The heart, therefore, may be completely
closed upon the contraction by the valves of the two chambers.
The heart. is enclosed within the pericardium which also
encloses the efferent branchial veins (PI. 27, figs. 9, Au, V;
Pl. 34, fig. 54, Au, Per, Vent). The aorta passes from the
floor of the ventricle (figs. 9, 54, Ao) to the ventral region
of the visceral cavity where it divides into two anterior and
posterior trunks (fig. 54, da). Just ventral to the ventricular
valve is an enlargement of the wall, the structure of which

554 H. P. KJERSCHOW AGERSBORG

is like that of a lymph-node (PI. 34, fig. 59, Bgl). Within this
gland (node) are a number of free cells (Pl. 34, fig. 56), which
vary considerably in size and structure (Pl. 34, fig. 58, a, b, c).
Some of the cells of the inside of the gland are pseudopodie (a),
which is also perhaps the condition of the cells free in the
lumen (PI. 34, figs. 56, LZ, 58, c). The structure of these cells
may also be the same, but the cells within the gland remain
in a more stable environment during the time of death, and
on that account may be less subjected to physical shock than
the cells within the lumen of the blood-vessels at the time
of the killing. In fact, the cells from the lumen (PI. 34, fig. 58, c)
show a considerable morphological difference in that they are
highly vacuolated. It is known from the study of invertebrate
blood that the cells contained in the blood-fluid are exceedingly
unstable (Tait and Gunn, 1918). In fact, these men were
able to destroy the blood-cells of the circulation of the
fresh-water crayfish, Astacus fluviatilis, by imjecting
india ink into the circulation of the living animal; the cells
explode very easily upon contact with foreign solids. The
cells of the circulation of Melibe may not be as easily
destroyed or as sensitive as the blood-cells of Astacus.
Whether the difference in environment relative to that of the
blood-plasma and the node is sufficient to produce this differ-
ence in post-mortem structure by the same killing method
cannot be determined here.

Boas (1886) finds for Cleodora acicula that the ventricle
is constructed of a few, large, short, flattened, perhaps muscle-
cells, which touch each other by their edges. Each cell has
a nucleus which lies on the outside of the contractile substance,
surrounded by a small protoplasmic mass. The contractile
substance consists of fine fibrillae, which are visibly trans-
versely striped. In M. leonina the cardiac wall does not
consist of muscle-cells exclusively, but of large nucleated
fibrillated epithelioid cells (Pl. 34, fig. 57, Cfm), but I cannot
at this time tell definitely whether the fibrillae are striped
transversely ; it is quite evident, however, that these cells
are different from the muscle-cells of other organs and of the
body-wall of this animal.

qn

MORPHOLOGY OF MELIBE ire

(3) The Venous System.

The venous system seems to consist of a number of very
thin-walled sinuses so that the blood easily exudes through
them, bathing the surrounding organs. The efferent branchial
veins collect the blood from the sinuses of the papillae, and
perhaps also from the larger sinuses of the body-cavity. The
so-called pericardium lies closely below the mid-dorsum, and im
front of the intestine and the ureter, and above the anterior
branches of the kidney. I have not at this time determined
the exact nature of the pericardium and its relation to the
blood, whether it is a completely closed chamber or not ;
whether it is invested with its own peritoneal membrane, or
whether it is fenestrated, allowing blood to enter it from the
surrounding sinuses. It is, in fact, usually thought that
the pericardial space in the molluscs contains blood, and is in
free communication with veins; but Lankester (1893) has
succeeded in showing by observations on the red-blooded
Solenlegumen, and by more recent careful investigation on
Anodonta cygnea, ‘Patella vulgata, and Helix
aspersa, that the pericardium has no communication with
the vascular system and does not contain blood.

8. The Organs of Excretion.
(1) The Kidney.

According to Pelseneer (Lankester, 1896: 111) the kidney
is a compact mass, as a rule, without external projections,
but it is divided into lobes in Stenoglossa in general, and
in some Taemoglossa, viz. Paludina and Cypraea.
In a fairly large number of nudibranchs (Doridomorpha,
Janus, &c.) the kidney is divided into ramifications which
extend between the visceral organs of the greater part of the
body. Shipley and MacBride (1915) say the kidney is a
vesicle, into the cavity of which numerous folds project covered
by the peculiar cells which have the power of extracting waste
product from the blood, which flows in spaces in the kidney wall.
The kidney m Mollusea varies a good deal in structure, but

556 H. P. KJERSCHOW AGERSBORG

is always built on the same fundamental plan as that of the
snail. The excretory system of M. leonina consists of a
bilateral structure with two main renal trunks, ureter, and renal
syrinx. The trunks extend anteriorly and posteriorly, dividing
into two sub-trunks, each of which pass into primary and
secondary branches. The anterior trunk divides much earlier
than the posterior, and the spread of the anterior bifurcation
is much larger than that of the posterior (Pl. 35, fig. 60,
Ab; Pb).
(2) The Ureter.

The ureter (PI. 35, fig. 60, U) follows the mtestine very
closely and empties a little to the anterior and left side of the
anal opening. This corresponds to Pelseneer’s description (vide
Lankester, 1906), where it is recorded that the external opening
of the kidney is situated near the anus and sometimes the two
open together into a sort of common cloaca, as may be seen
in Gymnosomata and in certanm Pulmonata, such as
Limax. In rare cases, he says, such as in the nudibranch
Janus, the excretory aperture is distant from the anus.
The anterior bifurcations of the renal organ (Ab) extend just
beneath the pericardium with one sub-trunk on each side of
the aorta. The posterior renal trunk sends its branches among
the hepatic arborizations and connective tissues in the posterior
region of the animal, caudal to the ureter and the intestine.

(3) The Renal Syrinx.

On the side of the ureter, midway between the junction of
the ureter to the renal trunks and the ureteric pore, is a bilobed
and somewhat convoluted whitish body which empties into the
ureter (PI. 35, fig. 60, Rs). This body is described by Hancock
(1865) as the pyriform vesicle ; von Jhering (1876) as the * Peri-
cardialtrichter ’; Bergh (1884qa) as renal syrinx that drains the
pericardial chamber. The renal syrinx is quite peculiar in
structure (P1. 35, figs. 61, 64, 67). In sections, it is shown to be
extensively plicated, but its walls are not muscular as observed
by Hancock (1865) on Tritonia hombergii. The phieae are
strongly ciliated. The cilia of the individual cells are kept

MORPHOLOGY OF MELIBE 557

together in such a way that under a low magnification they
appear to form tufts which give to the lming the appearance
of flask-shaped cells (Pl. 35, figs. 61,64). Higher magnification
brings out their true nature, that they are moderately columnar
cells with large cilia nearly four times longer than the cell (PI. 36,
figs. 67, 68, 69). The renal syrmx communicates with the
pericardium by a cyncitial plate with the nuclei scattered,
but as a rule nearer the base, i.e. toward the syringeal
side (figs. 66, 67, Sypl). It communicates with the ureter,
however, by a rather wide opening (fig. 61). There is
then no reno-pericardial pore between the kidney and _ peri-
cardium, through the renal syrmx. This I have determined
by the study of serial sections of the organ. Hlysia, according
to Lankester (1906: 110), is exceptional in that the kidney is
placed below and partly surrounds the pericardium, and the
reno-pericardial orifices are multiple, some ten being present.
And, according to Shipley and MacBride (1915) there is a reno-
pericardial canal, a narrow ciliated passage, between the
kidney and pericardium in the molluscs. In Melibe leonina
on the pericardial side, the renal syrinx narrows into a neck
(Pl. 35, fig. 60, P), which internally is formed into two channels
by a plica or villus which extends from the cyneitial plate and
into the organ (PI. 35, figs. 64, 65, Pl. 36. fig,67, Pl). Only about
one-third of this villus is ciliated, that is, its tip, or the part
farthest away from the cyncitial plate. The sides of the syrmx
opposite the non-ciliated portion of the villus are also non-
ciliated. The non-ciliated part of the walls has a large number of
nuclei situated near the surface. The structure of the ciliated
columnar cells of the renal syrinx, or pyriform organ, in
M. leonina, shows the same remarkable feature as in the
intestine, viz. the individuality of the cilia as they pass into
the cell. In the renal syrinx there is first the plainly visible
terminal bars, but unlike the condition found in the intestine,
the terminal bars (basal granules) are shown only as one row,
or one for each cilium. From the terminal bar the cilium con-
tinues to the base of the cell as a distinctly granular fibrillar
structure. Asin the intestine (PI. 33, fig. 45, Bm), the basement

558 H. P. KJERSCHOW AGERSBORG

membrane is prominent, but unlike the condition here, where
it seems to be granular, the appearance of the basement
membrane in the ciliated cells of the renal syrmx is a con-
tinuous, non-granular line, or the granules if present are fused.
However, this may also be the condition in the intestinal
cells (Pl. 38, figs. 46, 47), reflecting, perhaps, the fact that the
bringing out of certain cytological features depends, at least,
on two things: (1) the physical condition of the organism
at the time of killing, (2) the method and kind of chemicals
employed in the killing. Another differential feature of the
ciliated cells of this organ, 1s, as pointed out above, the indepen-
dent arrangement of the cilia. That is, the cilia are not mingled
with the cilia of neighbouring cells, as in the case of the intestine.
Still another feature is the size of the cilia both in length and
diameter. This specialization of the cilia may poimt toward
a special function of the organ. For example, it may be that
of creating a suction within the organ in order to draw the
pericardial fluid toward and through the cyncitial plate. The
cyncitial plate, then, with the action of the specialized cilia
of the plicae may function as an extracting organ, and one may
expect this process to be that of ridding the pericardial fluid
of waste. This is also the opimion of Hancock (1864: 520)
for Dorids.

Von Jhering (1876: p. 49) applies the name ‘ Pericardial-
trichter’ to the renal syrinx. By this name its function is
indicated also. The case in question is that of Tethys,
for which the author finds that the syrmx communicates with
the lumen of the ureter: ‘... in weiter Communication steht,
andererseits durch eine kleinere runde Oeffnung mit der Peri-
cardialhéhle zusammenhingt. Die letztgenannte liegt in emer
Membran, welche quer zur Axe des Pericardialtrichters steht
und sein Lumen von dem des Pericardium trennt. In dieser
Membran legen um die Oeffnung herum zahlreiche rmgférmig
angeordnete Muskelfasern, die also emen Sphincter bilden durch
welchen die Communication zwischen Niere und Pericardium
nach Belieben aufgehoben werden kann.’

Bergh (1884a: 76) does not describe the exact relationship

MORPHOLOGY OF MELIBE 559

of the renal syrinx to the ureter and pericardium. Dealing
with a number of types: Phylliroe, Acura, Rizzolia
australis, Bornella, Tritonia challengeriana,
and Marionia (pp. 3, 8, 30, 41, 47, and 51, respectively),
he only describes the shape and size and a few of its finer
structures, and also that it opens into the pericardium on the
one hand and the ureter on the other. For the holohepatic
form, Chromodoris striastella, he says: the renal
syrinx is bulb-shaped of 0-75 mm. greatest diameter ; the folds
of the interior can easily be seen from the outside ; the ciliated
cells are as usual. The duct of the renal syrinx is about
1-5 mm. long, opening into the chamber; in the anterior are
the usual villi and papillary outgrowths.

Pelseneer (1893: 458): * Je crois que c’est le plus antérieur
ou ventral qu’on doit considérer comme tel: les Nudibranches
les plus voisins de Elysia (Hermaea, Cyerce) m’ont en
effet montré Vorifice réno-péricardique a la méme_ place,
ventralement et a gauche. Les autres conduits seraient
secondaires ou cénogénétiques et résulteraient vraisemblable-
ment de la multiplicité des points de contact entre le péricarde
et le rein, celui-ci entourant plus ou moins le premier.’

Stempell (1899: 142) finds for Solemya togata, Poli:
‘ Von histologischem Interesse ist zunichst die Beschaffenheit
der Nierenspitzen. Dieselben besitzen nimlich nicht wie
diejenigen der Nuculiden ein flaches, mit langen Geisseln
besetztes Epithel, sondern ein gewodhnliches mittelhohes
Cylinderepithel, welches nur miissig lange Cilien tragt.’

And MacFarland (1912: 527), for Dirona picta, writes:
‘The reno-pericardial opening is found in the renal syrinx,
a conspicuous pyriform body situated midway of the animal’s
léngth, upon the right dorsal surface of the visceral complex.
It communicates below with the pericardial cavity, opening
through the floor of the right side. Its lumen is divided by
numerous folds of the wall, many of which in turn bear
secondary folds. The complicated opening thus formed is
lmed in its upper portion by high columnar cells, bearing very
long cilia, which are directed downward.’

560 H. P. KJERSCHOW AGERSBORG

Thus it is seen that, while the communication between the
kidney and the pericardium differ, the internal structure of
the renal syrinx seems to be similar as far as being lined with
ciliated columnar epithelium.

The cilia of the renal syrmx of M. leonina discontinue
in the region of junction with the ureter (PI. 35, figs. 61, A,
62,4). The cells are indistinct in this region (PI. 35, figs. 61, B,
63). The structure of the ureter is unique in itself. The
epithelium shows a cyncitium (PI. 36, figs. 71, 72), some parts
being conspicuous by the presence of large vacuoles which seem
to have formed by the confluence of smaller ones that arise
around the nucleus. These larger ones, then, as in the case
of the intestine, pass into the lumen of the organ. The ureter
is covered by an exceedingly fine fibrous cover (PI. 36, figs. 71,
72, Hx). This cover as well as the epithelium vary according
to their position. That is, nearer the renal syrinx it is more
glandular in its feature (PI. 36, fig. 72) than near the end of the
ureter (PI. 36, fig. 71). The renal chamber proper consists of
a corrugated glandular lining with a small amount of fibrous
tissue covering it (Pl. 36, fig. 70, Pr. nph, Ct). The structure of
the epithelium suggests that the organ is one of periodic
function. According to Stempell (1899: 142), ‘ Das Epithel
der Nierenschliuche selbst hat eme ziemlich typische Form.
In den distalen Abschnitten der vergleichsweise hohen Zellen
finden sich helle Vacuolen, welche . . . regelmissig kleme
Concrement - Klumpen enthalten. Nach Conservirung mit
Flemming’scher Flissigkeit gelang es mir auch, deutliche
Cilien auf den Zellen nachzuweisen. The kidney of M.
leonina, as far as [ have seen up to the present time, is not
ciliated. But the epithelium is highly glandular in structure,
which also is the nature of the lining of the ureter, but there is
considerable difference in structure, nevertheless.

The function of the kidney is supposed to be that of extracting
waste from the blood (Shipley and MacBride, 1915). Ward
(1900: 152) finds that among the variable types of excretory
cells two appear to be constant: the first absorbs indigo-
carmine and refuses ammonium-carminate, while the second

MORPHOLOGY OF MELIBE 561

precisely reverses this action. Rarely excretory cells do both,
but even then an excretory cell absorbs the one substance
more freely than the other, or vice versa. “These two types
are associated with voluminous organs. The indigo kidneys
produce urea, uric acid, and urates, while in carminate kidneys,
thus far known, none of these substances are formed, though
some non-indigo excretory cells contaim urates. Referring to
special cases he states: ‘In two groups of molluscs the nephridia
instead of being lined throughout their entire extent by a
single type of excretory cells present noteworthy differences :
in Amphineura the reno-pericardial ducts of acid reaction
eliminate actively carminate and litmus ; while the rest of the
nephridium, formed of different cells, and with alkaline reaction,
eliminate indigo. Both nephridia of Patella eliminate
equally indigo. ‘The most numerous non-ciliate eliminate
indigo; the others, ciliated, eliminate only carminate—the
single nephridium being thus a physiological equivalent of two
nephridia in the Diotocardia (Trochus, &¢.).’

9. The Organs of Reproduction.

It was pointed out by Lankester (1881, 1893), Sedgwick
(1888, 1898), Goodrich (1895), and Lang (1896), that the true
coelom in molluscs is much reduced, being divided into three
parts: (1) the pericardium, (2) the perigonadium, and (8) the
nephrocoel; the remaining body-cavities being haemocoels
which are derived in part from a system of spaces which arise
between the ectoderm and the entoderm (Sedgwick, 1888 :
383). Itis not my purpose to discuss here the homology of these
cavities in M. leonina, but only to call attention to their
relation to one another and their relative duct systems. As
stated aboye the pericardium is a closed cavity which com-
municates with the nephrocoel by the renal syrinx which seems
to be closed at its pomt of communication by what I have
called a cyncitial plate. The perigonadial coelom lies directly
below the nephrocoel. It was shown by Erlanger, 1891-2,
for Paludina that the pericardium arises as two coelomic
sacs on either side by a hollowmg out of the mesoblast.

NO. 268 Pp vf

562 H. P. KJERSCHOW AGERSBORG

These coelomic cavities then fuse and later by a process of
special growth form a peritoneal funnel that opens to the outside
oneither side. The gonad develops from the wall of the coelom ;
then, together with the rudimentary left peritoneal funnel, it
becomes constricted off from the main division of the coelom
(the pericardium), forming a small genital sac. From the wall of
this sac, the genital duct grows out, and joins an epidermal
invagination like the peritoneal funnel of the right side.

In M. leonina the perigonadium is situated caudoventrally
in the perivisceral cavity. It has a complex duct-system which
opens on the right side near the anterior end of the trunk of
the body (Pl. 27, figs. 2, 8, P; Pl. 37, fig. 81). The single
kidney (nephrocoel) is situated dorsally and communicates
internally with the pericardium through the renal syrmx and
externally through the ureter which opens on the left side of,
and close to, the anal pore. The perigonadial coelom (peri-
gonadium) and pericardium are completely separated by
the visceral cavity or vascular (haemocoels) cavities. These
cavities are blood-spaces into which the blood percolates from
the arterial vessels, bathing the visceral organs. In Chitons,
“A separation has taken place in the genital region of the
coelom from the renal; the gonad then acquires special ducts
which may not be homologous with the peritoneal funnels ’
(Goodrich, 1895: p. 486).

(1) The Hermaphrodite Gland.

It is a well-known fact that the gonads among nudibranchs
and many other molluscs are hermaphroditic, but the duet
system of the two different functional regions varies relative
to their complete development. Some authors who have
worked on the Elysiadiidae (Allmann, Hancock, Souleyet,
Gegenbaur, &c.) maintained that the male and female parts
of the hermaphrodite gland are separate as are the ducts
(vide Pelseneer, 1891). It has been shown by Pelseneer (1891),
who worked on several genera of the group in question, that
the same part of the gland is both male and female ; some of
the follicles of the hermaphrodite gland among certain

MORPHOLOGY OF MELIBE 563

Doridiidae and Aeolidiidae, he found, were distinctly
male or female; among the Elysiidae all the follicles, in
fact, contaimed two genital products. Mazzarelli (1891 a) found
that, while the hermaprodite gland is divided into lobes with
subdivisions somewhat deeply placed, each lobe presents a
great number of acini contrary to that which up to that time
had been observed in tectibranchs—referring in particular
to the works of Lucaze-Duthier on Pleurobranchus,
of Moquin-Tandon on Umbrella, of Vayssiére on Aphala-
spidea, and by himself on the Aplysiidae—each acinus
produces at the same time and contains ova and spermatozoa.
In Pleurobranchiaea, however, there are both male
acini and female acini.

Lang (1896) characterizes three types of genital ducts in the
molluses as follows :

Type I.— The hermophrodite gland has a single un-
divided efferent duct opening through a single aperture—
Gastropteron, Pteropoda, Cephalospedae (Bulla,
Dorium).’

Type I1.— The hermaphrodite gland gives rise to a
hermaphrodite duct which soon divides into two parts, the
vas deferens or-seminal duct, and the oviduct. The former
runs to the male copulatory apparatus, the latter to the female
genital aperture. The male aperture and the penis lie in front
of the female . . . both le on the right. This second type may
be deduced from the first, if we assume that the common
duct of the hermaphrodite gland divided into a male and female
duct, but also that the seminal furrow closed to form a canal
in continuation of the male duct. To this type belong :

1. A few species of Dendebardia)\
2. Basommatophora,

3. Oncida, and

4. Vaginulidae

of the Pulmonata.’

Type I1I1.— In all Nudibranchia and a few Tecti-
branchia (e.g. Pleurobranchiaea), the hermaphrodite
gland gives rise to a hermaphrodite duct, which, as in the

Pp2

564 H. P. KJERSCHOW AGERSBORG

second type, sooner or later divides into a male and female
duct. These, however, do not open through distinct apertures,
but again unite to form a common atrium genitale or genital
cloaca.’

The first type is also demonstrated by Bonnevie (1916)
for the pteropod Cuvierina columnella Rang, where
the hermaphrodite duct arises first as a groove (Rinne) and
later forms into a thin duct (Rohr) which makes its exit from
the left dorsal edge of the hermaphrodite gland.

Type LV.—tThis is a new type. It is represented by the
reproductive system of Melibe leonina (Gould) and is
equivalent to types 11 and 111 plus a more complete duct system.
The organs of reproduction in this species (Pls. 28 and 37, figs. 9,
81) consist of a well-defined pair of gonads, each consisting of
many lobes or acini; an oviduct, a ‘ prostate gland’ (con-
voluted portion of the female duct), a uterus (enlarged distal
portion of the oviduct), a spermatotheca, and a vagina ; a vas
deferens with its ampulla, a penis, and a mucous gland. The
male duct is further modified into vasa efferentia. Hach of
these parts is peculiarly modified, and together furnish a
unique system of reproduction. The relative position of these
organs in the body-cavity is shown in Pl. 28, fig. 9.

The male and female ducts n M. leonina do not unite
to form a common genital atrium as set forth by Lang for
‘all Nudibranchia.and a few Tectibranchia’, but
open close together through separate apertures (Pl. 37, fig. 77,
Mgp); that is, the penis lies in front of the vagina (PI. 28,
fig. 9, P); in that way, it resembles the second type of Lang.
Both branch, i.e. a vas efferens and an oviduct pass to the
same acinus, which is to say: the hermaphrodite gland gives
rise to a double genital duct system which passes from the
respective male and female germ-cell area of the various
acini. In this respect the reproductive system of M. leonina
differs from all the three types of Lang, and for this reason I
have designated the genital duct system of this molluse as
constituting a fourth type. However, since the oviduct
is still connected to the ampulla of the vas deferens by a duct,

MORPHOLOGY OF MELIBE 565

it is evident that this type is derived from the third type
of Lang.

The hermaphrodite gland lies in the caudo-ventral region of
the perivisceral cavity and consists of a great number of
bilobed acini (PI. 37, fig. 81, Ot). The eggs and the spermatozoa
are situated in different regions of the same acinus, which is
easily demonstrable in the neutral phase, that is when the
male or female germ-cells are in a regressive or progressive
stage of growth ; as soon as the one has gained the ascendency
either a male or a female phase appears which seems to take
over the entire acinus. At such a time the small duct system
which leads from the indifferent or resting region of the acinus
is nearly crowded out by the actively employed ducts, and then
the acinus may give the appearance as having only a single
duct arising from it. That is, during the ripe male phase, the
female germ-area with the ducts of any acinus may be crowded
to the periphery or to one side in such a way so as to give the
appearance of only one duct system leading out from that
acinus. However, in any stage of either male or female phases,
the two duct systems may be discerned in some of the acini.
I cannot determine with any certainty whether this species is
protandrous or not, as I have not studied sufficiently young
individuals on this point. All the individuals, whose glands
I have sectioned, have shown ripe spermatozoa in the acini
and also ova.

Pelseneer (1895: 31) states that protandry ought to be
regarded as a general phenomenon in EKuthyneurous
gasteropods. ‘That this is notoriously the case in pul-
monates, and that it has been recognized in various opistho-
branchs which have been studied from this point of view, viz.
Lohiga, the Thecosomatous pteropods, e.g. Clio-
striola, &¢.; nudibranchs, among which he observed it in
Aeolidia and Elysia; and lastly Clione limacina
(Gymnosomata), m which he noticed that individuals of
a length of 15 mm. (or less) do not as yet show any ova in their
genital glands, bué stages in the development of spermatozoa
only. He also found that the ovogenous and spermatogenous

566 H. P. KJERSCHOW AGERSBORG

regions of the acini in Onchidiopsis are not demarcated
with any regularity, that in the middle portion male and female
acini can be seen in sections lying side by side. Also, that
the products of the two sexes either do not arise in the
same caecum or they do not arise in the same region of
a caecum.

Bonnevie (1916) finds that Cuvierina columnella is
protandric ; the spermatozoa pass into the ‘ Zwittergang ’
and then the eggs develop; there being only one ripening of
the germ-cells in the life of the individual. If this be also the
case among individuals of Aeolidia (Coryphella) lands-
burgii, as reported by Pelseneer (1895: 23), that the acini
of the hermaphrodite gland produce ova in their distal portion
and spermatozoa in their proximal portion, it is not so difficult
to understand how the product of the peripheral region of the
acinus with only one duct system may gain access to the duct,
the ripe spermatozoa simply passing out of the way, giving
its former position to the oncoming mass of the ripening ova.
Pelseneer reports that the same condition as noticed in
Aeolidia landsburgii has been recognized as general
in all the Elysioidea (Cyerce, Hermea, Elysia, and
Limapontia). Mazzarelli (1891 a) found male and female
germ-products present at the same time in the acini of Aply-
slidae, and in Pleurobranchaea separate male and
female acini. ‘In the former there are only spermatozoa and
spermatids in varying stages of development, and spermatozoa
with sheaths (fascetti) seem jomed together by cytophors.
In the latter (female acini) there occur only ova. The tiny ova
in diverse grades of development lie distributed all around the
internal aspect of the wall of the acinus as an epithelium.
Of the original germinative epithelium there remains slight
traces. The ova which are larger and nearly being expelled
are found more or less in the centre of the lumen. This fact,
that of the formation of spermatozoa and of ova in separate
acini in the male-female gland which is abnormal in the
Tectibranchs, is seen to be true or ordinary in effect in
many Nudibranchs.’

MORPHOLOGY OF MELIBE 567

(2) The Hermaphrodite Duct.

Mazzarelli (1891 a) found that the hermaphrodite duct which
leaves the male-female gland is very minute in diameter at its
beginning, but later, after a certain descent, dilates abruptly
to a greater lumen. From this point it gradually narrows again
only to bifureate, giving origin to two minute ductuli. From
this point one passes straight to the penis (oedagus), in which
it constitutes the deferent channel. The other ductule which
is the oviduct enlarges rapidly after its origin and here presents
a tiny caecum, then the oviduct contracting gradually, only
again abruptly to dilate, develops the first ampulla. In
M. leonina, however, I find that the hermaphrodite duct
has separated into two distinct male-female ducts, i.e. from the
time an exit is formed in the acinus it 1s double. It should be
pointed out, however, that the male duct caudad of the ampulla
is larger than the corresponding female duct.

(3) The Oviduet.

The oviduct, after it leaves the last acinus (PI. 37, fig. 81, Od),
passes into the prostata (Pr), which, in fact, is a part of the
female duct. It stands in relation with the ampulla of the
vas deferens by a biluminate duct (Bil.dpr) (PI. 37, figs. 73, 79,
Pro) ; the so called prostate gland is a much-coiled portion of
the oviduct, consisting of two kinds of coils, a large and a small
coil system coiled upon itself. The ampulla-prostate duct shows
that the female-duct system formerly was in functional relation
with the male-duct system at the ampulla; the genital-duct
system from the ampulla and to the gonads constituted the
hermaphrodite duct. At the point of exit from the prostata
the oviduct dilates into a much saccular portion (Pl. 37, figs. 79,
Gis. 80.5F 81, Ut), which after some distance passes into a
narrow portion. From the distal part of this (Pl. 37, fig. 81, Osp)
there is a large sac, the spermatotheca, and from this point
to the orifice the duct dilates a little (Va) forming into what
I have called the vagina. The most distal portion of the
duct lies in communication with the mucous gland (Mg).

568 H. P. KJERSCHOW AGERSBORG

The prostate portion of the oviduct is not very glandular.
The wall of the uterus consists mainly of fibrous and of some
muscular tissue. Its lining consists of glandular epithelium
(Pl. 37, fig. 80, Ms, Gl).

(4) Ovispermatotheca.

The ovispermatotheca has a most unique internal structure
(PI. 37, fig. 82). Its outer part consists of a loose vascular con-
nective tissue (L) and a muscular layer of circa two cells in
depth. Its middle part is a connective-tissue layer (Sm)
upon which rests a papillated epithelial layer. The muscle-
cells of the muscle-layer show the interesting structure already
described. The epithelium retains Delafield’s haematoxylin
stain very well. The papillated epithelium consists of cells
that are free (Hpt) at their two-thirds distal portion, abutting
(Pl. 37, fig. 74, Spt) into the cavity as finger-like processes
(Pl. 37, fig. 82, Spt). The larger of these seem to be supported
by the underlying basement layer (Sm, Bm). In the cytoplasm
there is a distinct micromeric network (Mic). The nucleus is
situated at the base in the smaller cells, and midway between
the base and the free end in the larger cells. The cells of the
larger papillae are wider at their free or distal portion, so that
they actually approach each other. The spermatotheca
contains both semen and ova, and, since in some cases this
organ is filled with eggs, I have called it ovispermatotheca.

Mazzarelli (1891 a) found that the structure of the spermato-
theca was plical. He says: ‘ Indeed the entire aspect of the
wall of this presents a great number of longitudinal folds
(or plicae) highly developed and disposed in such a manner
as to constitute a series of correlated passages or channels
(rooms) (‘ concameragioni’), on the periphery of the lumen
of the ampulla which are commonly engorged with the sperm.’

Eliot and Evans (1908: 287) write: ‘The walls of the
spermatotheca in Doridoides gardineri are thick and
produce a secretion. In some specimens small clumps of
spermatozoa are imbedded in this secretion. In others all the
spermatozoa form a central mass in the main cavity of the

MORPHOLOGY OF MELIBE 569

spermatotheca. It is possible that the secretion serves to
form small pockets of spermatozoa or spermatophores.’
MacFarland (1912) found, in the Dironidae, that the
spermatotheca was almost rudimentary, and that the dilated
oviduct seemed to have assumed in part the function of a sperma-
totheca, ‘ for it is frequently crowded with spermatozoa, while
the spermatotheca itself contains relatively few.’ On the
contrary, ‘in Diron albobineata the oviduct is short
and slender: spermatotheca very large, reaching a diameter
of 1:3mm., total length being 4mm. in a large specimen ’.
In this connexion it is well to note that, in M. leonina,
which has a large spermatotheca, the male germinal product
also frequently passes into the distal regions of the female
genital duct system even as far as and including the prostata.

(5) The Male Genital Duct.

The male genital duct starts in the hermaphrodite gland as
small tubules (Pl. 37, fig. 81, Ve), which join into a common
median duct that enlarges into a round bulb-like part (Amp) at
the anterior region of the ovitestes (gonads). This enlargement
T have called ampulla. From the ampulla the male duct passes
anteriorly as a large organ of fibrous tissue (Pl. 28, fig. 9, Vd).
It is surrounded by a sheath of its own (PI. 37, figs. 78, Iglp,
74, 76, 77, 79, 88, Oc). Intervening between the penal sheath
and the penis itself, is lymph or mucus, a colourless, structure-
less substance (/). The penis is covered with cells of epithelial
structure, perhaps both of these, and the cells of the liming of
the penal chamber just described, secrete the mucous substance
also mentioned. The organ itself is made up mainly of fibrous
connective tissue (Pl. 37, figs. 75, 78, 88, Int, Ml). In fig. 75
the biluminate effect as shown here is due to the coiled condi-
tion of the organ at the point of section. The main bulk of the
organ consists of this fibrous tissue (Pl. 37, fig. 76, P). The
lumen of the organ is lied with ciliated cuboidal epithelium
(Pl. 37, fig. 83, Iepl, cil). The ampulla of the penis seems to be
quadroluminate ; anteriorly these lumina converge into three,
then two, and finally into one, which becomes the seminal tube

570 H. P. KJERSCHOW AGERSBORG

of the penis. The structure of the seminal vesicle and of the
ampulla is alike; it consists of a heterogeneously arranged
cell-mass, so well welded together that the whole structure is
quite compact. The penis is, indeed, so large that when it is
withdrawn it fills a large part of the body-cavity. The penis
is simply an extension of the seminal vesicle. The sheath
of its anterior portion is firmly lodged on the mucous gland
(Pl. 37, fig. 81, Cl.p) with the penis in its pore. The penal pore
merges with the body-wall, adjacent to the vaginal orifice
(Pl. 37, fig. 77, Mga, Fv). The penis is sometimes extended
to the outside, and is then curved like a screw (PI. 27, fig. 3, P).
In copulation the penis, which is long, twisted like a screw
and of tough musculature, is inserted into the posterior (female)
genital pore of the mate, and so firm is the union that separation
may not occur even though the couple be dipped from their
natural abode and placed in a vessel (Agersborg, 1921: 238).
I have not found mutual coitus effected at the same time
in this species, although it is supposed to be a common practice
among nudibranchs, according to different authors: Alder
and Hancock (1845: 25), Mazzarelli (1891a: 287), Crozier
(1919) et al

(6) The Mucous Gland.

The mucous gland constitutes the albuminous and nidamental
glands (Pl. 37, figs. 74, 76, 79, Mg, 81, Mgl, and 84, A, B,C, D).
It consists of laminations so arranged that sections through
the side of it have a six-layered aspect ; continuous sections
soon bring out the true conditions. The gland is made up of
simple, tall, ciliated columnar epithelium, highly glandular in
nature, and which rests on thin connective-tissue fibres with
cells which connect. the gland to the body-wall. The gland
extends almost to the outside of the vaginal orifice (Pl. 37,
figs. 74, 76, 77). It functions when the animal spawns; the
mucus and the capsulated eggs pass out together.

Mazzarelli (1891), m Pleurobranchaea, finds that the
vagina at the back of the oviduct’s terminal point prolongs
itself remarkably dorsad becoming sacculated, its walls bemg

MORPHOLOGY OF MELIBE 571

formed of robust folds and studded with glands which are
the glands of the nidamento; near the opening of the
oviduct into the vagina but still more dorsad opens the gland
of the albume. This is contrary to that noted in other
tectobranchs in which the albume gland terminates in a
vast number of minute ductuli with blind origins arranged in
such a way as to constitute un fitto gomitolo. The
albume gland of Pleurobranchaea resembles much in
structure the albume gland of some nudibranchs, e. g. Hreo-
lania (? Hercolania) as described by Trinchese.

The albuminous gland of M. leonina seems to be more
uniform in its physiological condition relative to that of the
mucous gland. Unfortunately, I have not at the present time
worked out its exact relation to the oviduct, but in general it is
somewhat like that described by Mazzarelli for Oscanius
and Acera, i.e. the nidamental gland is nearer the orifice
of the vagina than is the albuminous gland. The epithelium
of the nidamental gland which secretes a great deal of mucus
at the time of oviposition shows some very interesting things
relative to its activity during such a time (PI. 37, fig. 84, A-C) :

1. The nucleus (6) presents no visible membrane, the nucleo-
plasm being filled with almost uniformly sized granules which
seem to be formed by the nucleolus and then pass as a liquid
into the cytoplasm of the cell where from small, minute
micromeres the cell becomes entirely filled with darkly staining
macromeres which seem to have grown from these smaller
ones so that the ordinary and less stainable cytoplasm is
practically obliterated, i.e. obscured by these granules.
These granules are very strongly basophil in their staining
reaction. The macromeres then liquefy and pass out of the cell
and into the lumen of the gland.

2. This leaves the cell in a condition strongly contrasting
to the one before the liquefying of the mucus. The cell is now
very vacuolated, contaiming a non-stainable, or rather oxyphil,
substance (A, Rs) with relatively few granules. These granules
are micromeric, and are aggregated in the meshes of the reticular
net-work of the cytoplasm.

572 H. P. KJERSCHOW AGERSBORG

3. In the proximal region of the cell may be seen the nucleus,
small and shrunken, containing a small nucleolus,
and lodged at the base of the cell.

4. A state of refilling now ensues (C). The nucleus begins to
enlarge and then a homogeneous cytoplasmic substance is
accumulated around it, which then spreads throughout the
cell. At the same time the micromeres of the vacuolated
cytoplasm of the cell increase in size.

That the nucleus takes a very active part in the formation
of the mucinous substance is clearly evident, but the micro-
meres of the cytoplasm seem also to take an active part in
the refilling of the cell as indicated by the growth of the
micromeres formed in the meshes of the reticular network
of the cytoplasm before the nucleus has assumed its normal
condition. In semi-vacuolated cells, i.e. cells which are in
the state of refillmg, the nucleus is mtermediate in size and
granulation to those described under (B). It is of mterest to
note some of the findings of Lange (1902) from his studies of
the structure and function of the ‘ Speicheldriisen ’ of gastero-
pods. This author found that the cells of the glands showed
great differences during feeding as compared with periods of
starvation. Some of these phases of activity of the cell are
quite similar to those recorded above for the mucous gland
of M. leonina. Lange says :

In allen Stadien der Fitterung und auch des Hunger-
zustandes finden sich nie simtliche Sekretionszellen auf dersel-
ben Sekretionsstufe. Ks kommen in jedem Stadium der
Fiitterung und des Hungers alle Stadien vor, doch in ver-
schiedener Haufigkeit. Der Kern nimmt innigen Anteil an
der sekretorischen Thitigkeit, mdem im Anfang seme Membran
sich auflést und sein Inhalt sich mit dem Protoplasma ver-
mischt, sodass der erste sekretorische Vorgang sich am Kern
bemerkbar macht. Die von Barfurth als ‘ Speichelkugeln ’
bezeichneten Gebilde sind Sekretvakuolen, welche angefiillt
sind mit muciginer Substanz. Man kann deutlich verfolgen,
wie sich das Mucigen in diesen Sekretvakuolen zu Mucin
umwandelt. Ist das Mucin gebildet, so verlert der Kern
seinen Turgor wohl durch Austritt von Kernflissigkeit. Dabei
steht das Kerninnere stets im offenen Zusammenhang mit dem

MORPHOLOGY OF MELIBE 573

Protoplasma. Es lassen sich zwei Teile an dem Zellich unter-
scheiden: der den Kern umgebende protoplastische Teil und
der periphere parablastische mit den Sekretvakuolen ; dieser
letztere wird mitsamt dem in ihnen gereiften Mucin bei der
Sekretion ausgetossen. Es bleibt in der Bindegewebskapsel
allein der protoplastische kernhaltige Teil ibrig, von dem aus
die Neuproduction des Zelleibis vor sich geht, sodass die
Sekretion zunichst in der Bildung von Sekretionsvakuolen,
sodann aber in der Ausstossung des ganzen peripheren Teils
der Zelle mit dem gebildeten Sekret aus der bindegewebigen
Kapsel besteht.

In the epithelium of the mucous gland of M. leonina,
as in the epithelium of the intestine and of the renal syrinx,
basal granules or desmochondria are demonstrable; but in
the mucous gland the linear fibrillar arrangement of the micro-
macromeres of the cytoplasm is absent. The reason for this is

obvious ; the cilia also seem to end on the distal basal granular
border (Pl. 37, fig. 84, Bg, Fcb).

V. SUMMARY.

1. It is evidenced by the work of various authors that
Gould’s Chioraera (1852) is identical with Rang’s Melibe
(1829); Chioraera, therefore, is a synonym of Melibe;
Gould, at the time of his description, did not know of the
genus discovered by Rang. For these reasons I have con-
sistently named it throughout all my works on this species
Melibe leonina (Gould) in spite of the attempt of certain
authors to build on the nomenclature of Gould.

2. Melibe leonina is absolutely void of masticatory
organs ; the generic description of Gould may be augmented,
therefore, to read in part, Bulbus pharyngeus aut cum mandi-
bulis aut sine mandibulis ; radula et lingua destitutus.

3. The anterior end of M. leonina is formed into a large
cowl ; this has a pair of stalked foliaceous tentacles which may
be retracted below the edge of the stalk which then acts as
a sheath. The tentacles are very complex in structure, being
innervated with nervous tissue. The tentacles are not ciliated,
as claimed by Jeffreys (1869) for all nudibranchs. The cowl

574 H. P. KJERSCHOW AGERSBORG

is fringed with two rows of cirrhi which also are highly complex
in structure. From an inner ganglionic axis, nerve-fibres
radiate to the peripheral ectoderm of the cirrhus. The exact
function of the tentacles, as well as of the cirrhi, is not known.
The tentacles are commonly called rhinophoria but, since the
exact function is not known, I have employed the original
term tentacles (dorsal tentacles) instead of the commonly
used term ‘rhinophoria’. ‘The cirrhi are more sensitive to
tactile stimulus than are the dorsal tentacles (Agersborg,
1922a: 441-3).

4. The body-surface of M. leonina appears smooth, but
upon close examination it is found to be everywhere tuberculate,
including the sole of the foot and the ventral side of the hood ;
in that way this species corresponds to other members of
this genus.

5. The dorsal appendages, which in M. leonina consist
of six pairs of foliaceous lobate structures, I have called by
the Linnean term papillae, mstead of cerata, or branchial
papillae for the reason as stated by Bergh (1879 c): * Respira-
tion takes place all over the surface in Nudibranchs,’ &e.
The papillae alternate in position ; they are subject to varia-
tion in structure relative to position and age (Pl. 27, figs. 1, 2,
3, and Pl. 30, figs. 18-25).

6. The foot projects in front of, and behind, the main body.
It is highly tuberculate and ciliated. Internally, a fine nerve net-
work is seen spread throughout its length and breadth, and at
the posterior end it aggregates into a ganglionic centre. Fine
nerve fibrils are seen to pass to the ciliated ectoderm. A great
many mucous glands are present all through the foot, which
open independently through small crypts between the ecto-
derm cells. These glands are the pedal glands which are
scattered all through the foot (figs. 1, 2, 3, 9, 26, 27, 28, 29).

7. There are three kinds of glands in the body-wall: (1) the
largest and most numerous are the odoriferous glands; (2)
the next in size and number are the saccular mucous glands ;
and (8) the unicellular mucous glands (PI. 31, figs. 30, 31).

8. The muscle system hes below the glandular fimbriated

MORPHOLOGY OF MELIBE 575

ectoderm ; the muscle-fibres are arranged in a fashion like the
fibres in a basket (Pl. 27, fig. 3).

9. The muscle-cell consists of two sarcoplasmic regions each
containing an abundance of micromeres: those in the inner
region are larger (Pl. 31, fig. 33, Ca) than those in the outer
region. The larger are called in the text macromeres. Each
of these micro-macromeric sarcoplasmic regions is invested
by a coarsely granular net-work (My, Sar). The nucleus is
placed centrally within the cell. Its chromomeres (I) are
scattered differently, i.e. sometimes around the periphery of the
nucleus (PI. 31, figs. 82, 33, Pl. 37, fig. 82) and sometimes less so.

10. There is no definite body-cavity. The body-cavity as it
exists corresponds to the primary body-ecavity of
Lang (1896) or the perivisceral cavity of Sedgwick
(1898).

11. The alimentary canal consists of five regions :

(1) The oesophagus with the non-glandular lining, and back
of it the oesophageal glands or salivary glands (Pl. 32,
figs. 36, Oe, 41, Sq).

(2) The proventriculus with distinct glandular lining
(Pl. 32, figs. 36, G, 38, Gl).

(3) The gizzard with its stomach-plates which are formed
by the secretion of the epithelial lining (vide supra),
(Pl. 82, figs. 87, 42, 43, Stpl).

(4) The pyloric diverticulum with its glandular and ciliated
and much corrugated surface, which secretion does
not keratinize as that of the gizzard (Pl. 32, figs. 35,
39, 44).

(5) The intestine with its large typhlosole protruding into
the cavity from the ventral side, and glandular ciliated
surface (Pls. 82 and 33, figs. 40, 45, 46, 47). The
structure of the intestinal lining is unique as concerns
the fibrillar nature of the cytoplasm, the clearly
visible terminal bars, and the basement membrane ;
also, the non-convergence of the cytoplasmic portion
of the cilia on the nucleus.

12. The liver ramifies all the parts of the body. Its secretion

576 H. P. KJERSCHOW AGERSBORG

into the gizzard does not harden in the alimentary canal.
The glandular structure of the epithelium of the liver exhibits
that it has an active function in vivo, owing to the presence
of a variable series of granules and vacuoles in the adjacent
cells fixed at the same time in the same way (PI. 27, figs. 1, 2,
4, 7, Pl. 33, 48-53).

13. The heart consists of two chambers enclosed within the
pericardium. These chambers are separated by valves from
the efferent branchial veins and the afferent aortic trunk-
vessel. In the aorta, just below the valve of the ventricle,
is a blood-gland or node which contains pseudopodic cells.
The cells found on the outside of this node, i.e. within the
lumen of the aorta, are different in structure from those found
within the node (PI. 34, figs. 54-9).

14. The wall of the heart consists of epithelioid and some
semi-musculofibrilloid cells (Pl. 34, fig. 59).

15. The kidney is much branched, and is situated between
the pericardium and the gonadium. It communicates with the
pericardium through the renal syrinx which is closed at the
point of junction with the pericardium by what I have called a
cyncitial plate. The liming of the kidney is glandular ; so
is also that of the ureter, but neither is ciated. Therenal syrinx,
however, is ciliated. The cells of the renal syrinx are peculiar
in that the cilia are very large and independent in position,
i.e. they do not mingle with those of adjacent cells. The renal
syrinx is plicated, and from the cyncitial plate a ciliated villus
protrudes into the organ (Pls. 85 and 36, figs. 64, 65, 67, Pl).

16. The organs of reproduction represent an additional type
to the three types enumerated and described by Lang (1896).
The male and female ducts n M. leonina do not unite to
form a common atrium genitale as set forth by Lang for all
nudibranchs and a few tectibranchs, but open close together
through separate apertures (Pl. 37, fig. 77, Map); the penis
lies in front of the vagina (Pl. 28, fig. 9, P); im that way it
resembles the second type of Lang. Both genital ducts pass
independently to the same acinus; in this respect it differs
from all three types of Lang, and, for this reason, I have

MORPHOLOGY OF MELIBE 577

designated the genital duct system in this molluse as con-

stitutng a Fourth Type. During any ripe phase of the

gonads the duct leading from the imactive area of an acinus
may be quite obscured by the ripening mass of germ-cells.

The organs of reproduction are represented in Pl. 37, fig. 81.
17. The spermatotheca I have called ovispermatotheca

because it is frequently filled with eggs from the oviduct.

The structure of the ovispermatotheca is quite peculiar owing

to the plicated nature of its linmg. The cells lining this organ

are flask-shaped, the neck beimg longer than the body and

abutting into the cavity (Pl. 37, fig. 82).

18. The mucous gland is relatively rather large. A great
deal of mucus is formed by this gland at the time of oviposition
(Agersborg, 1919, 1921, 19234). Sections of the gland which
I have studied show that during the act of mucus-formation
the nucleus takes an active part. The nuclear membrane is
then very obscure or apparently absent, the nucleus goes
through fragmentation, the smallest cytoplasmic granules of
basophil nature are the nearest to the nucleus; after the
mucous granules have liquefied and passed into the lumen or
cavity of the gland, the nucleus of the gland-cells is small,
shrunken, and non-granular, with a small nucleolus, and basal
in position within the cell. The cell then passes through a
period of refilling during which time the nucleus first grows in
size, and at the same time the micromeres of the cytoplasm,
which are lodged in the meshes of the reticular structure of the
cell, also grow.

VI. Lirzerature Crirep.

Agersborg, H. P. Kjerschow (1916).—‘ A study of the nudibranchiate
mollusc, Melibe leonina (Gould).’ A thesis submitted for the
degree of Master of Science. University of Washington, Seattle.
120 pp. of text and 53 plates.

—— (1919).—** Notes on Melibe leonina (Gould).’’ ‘Pub. Puget
Sound Biol. Sta.’, 2: 269-77, 2 plates.

——— (1921).—*‘ Contribution to the knowledge of the nudibranchiate
mollusc, Melibe leonina (Gould)”, ‘Amer, Nat.’, 55: 222-53,
12 figures.

—— (1921a).—“‘ On the status of Chioraera (Gould)”’, ‘ Nautilus’,
35: 50-7.

NO. 268 Qq

578 H. P. KJERSCHOW AGERSBORG

Agersborg, H. P. Kjerschow (1922).—‘‘ Notes on the locomotion of the
nudibranchiate mollusc, Dendronotus giganteus O’ Donoghue ”’,
‘ Biol. Bull.’, 42 : 257-66, 4 photographs.

(1922 a).—‘‘ Some observations on qualitative chemical and physical
stimulations in nudibranchiate molluscs with special reference to the
role of the ‘ rhinophores’ ”’, ‘ Journ. Exper. Zool.’, 36: 423-44, 4 tables

~ and 2 figures.

(1923).—‘* A critique on Professor Harold Heath’s Chioraera

dalli, with special reference to the use of the foot in the nudibranchiate

mollusc, Melibe leonina (Gould)”’, ‘ Nautilus’, 36: 86—96, 4 plates.

(1923 a).—‘ Studies on the effect of parasitism upon the tissues, I.
With special reference to certain gasteropod molluses”’, ‘ Quart. Journ.
Micro. Sci.’ (In press.) 10 plates.

Alder, J., and Hancock, A. (1845).—* A monograph of the British nudi-
branchiate Mollusca.’ The Ray Society, London. pp. 1-54; 48 plates
with descriptions ; Appendix, pp. i-xiii.

(1864).—** The similarity between the species Melibe fimbriata
and Tethys fimbriata. Notice of a collection of nudibranchiate Mollusca
made in India, by Walter Elliott, Esq., with description of several
genera and species ’’, ‘ Trans. Zool. Soc.’, 5: 136-9, 1 plate.

Angas, G. F. (1864).—‘‘ Melibaea australis. Description d’espéces nou-
velles appartenant a plusieurs genres de mollusques nudibranches des
environs de Port Jackson (Nouvelle-Galles du Sud) accompagnée de
dessins faits d’aprés nature’’, ‘Journ. de Conch.’, 3e sér., 4: 43-90,
3 coloured plates.

Arey, Leslie B. (1917).—‘‘ The sensory potentialities of the nudibranch
‘rhinophore’”’, ‘ Anat. Rec.’, 11: 514-17.

(1918).—‘** The multiple sensory activities of the so-called rhino-
phores of nudibranchs’’, ‘ Amer. Journ. Physiol.’, 46 : 526-32.

Arnold, A. F. (1916).—‘ The sea-beach at ebb-tide.’ 490 pp. New York.

Bergh, R. (1863).—‘‘ Campaspe pusilla, en ny slegtsform af Dendronoti-
dernes gruppe, samt bemerkninger om Dotidernes familie ’, Professor
J. C. Schiddte ‘ Naturhistorisk Tidsskrift ’ (1861-3), 3 Rekke, 1: 471-
83, 1 tab.

—— (1871).—“ Beitrige zur Kenntniss der Mollusken des Sargasso-
meeres”’ (Doto Oken, Scyllaea pelagica Linnaeus, Fiona Alder
and Hancock, Melibe Rang, &c.), ‘ Verh. zool.-bot. Gesell. in Wien ’,
21: 1273-1308, 3 Taf.

—— (1873).—‘‘ Beitrige zur Kenntniss der Aeolidiaden, I”’, ibid., 23:
596-628, 4 Taf.

—— (1874).—‘‘ Beitrige zur Kenntniss der Aeolidiaden, II”, ibid.,
24: 395-416, 3 Taf.

(1875).—‘ Beitrage zur Kenntniss der Aeolidiaden, III”, ibid.,

25: 633-58, 3 Taf.

MORPHOLOGY OF MELIBE 579

—— (1875a).—‘‘ Neue Beitraige zur Kenntniss der Phylliaden”’, ibid.,
25: 659-74, 1 Taf.

— (1875b).—Carl Semper, ‘‘ Reisen im Archipel der Philippinen, II ”’,
*Malacologische Untersuchungen’ (IX Heft, Tethys, Melibe), 2:
345-76, 3 Taf.

(1876).—“ Beitrige zur Kenntniss der Aeolidiaden, IV”, ‘ Verh.

zool.-bot. Gesell. in Wien’, 26: 737-64, 4 Taf.

(1877).—‘ Beitriige zur Kenntniss der Aeolidiaden, V”’, ibid., 27:

807-40, 3 Taf.

(1877 a).—‘‘ Notizen iiber Tethys leporina Linnaeus ”’, ‘ Jahrb.

mal Gesell.’, 4: 335-9.

(1878).—“ Beitrage zur Kenntniss der Aeolidiaden, VI”, * Verh.

zool.-bot. Gesell. in Wien’, 28: 553-84, 3 Taf.

(1879).—‘ Beitrage zu einer Monographie der Polyceraden, I”’,

ibid., 29: 599-652, 6 Taf.

(1879 a).—*‘ Beitrige zur Kenntniss der Japanischen Nudibranchien

i’, ibid., 80°: 155-200, 5 Taf.

(1879 b).—*‘ Beitriige zu einer Monographie der Polyceraden, II”’,

ibid., 30: 629-68, 6 Taf.

(1879 c).—‘‘ On the nudibranchiate gasteropod Mollusca of the North

Pacific Ocean, with special reference to those of Alaska, I”’, ‘ Acad.

Nat. Sci. Philad.’, 3 ser., 31 : 71—F32, 8 plates.

(1880).—‘‘ On the nudibranchiate gasteropod Mollusca of the north
Pacific Ocean, with special reference to those of Alaska, II”’, ibid.,
32: 40-127, 8 plates.

— (1881).—*‘ Beitrige zur Kenntniss der Japanischen Nudibranchien,
II”, ‘ Verh. zool.-bot. Gesell. in Wien’, 31: 219-254, 5 Taf.

——— (1882).—“‘ Beitrage zur Kenntniss der Aeolidiaden, VII’’, ibid.,
32: 7-74, 6 Taf.

(1883).—“‘ Beitrage zu einer Monographie der Polyceraden, III”’,

ibid., 33: 135-80, 5 Taf.

(1884).—‘‘ Beitrage zur Kenntniss der Gattung Melibe Rang
(Melibe papillosa De Filippi)’’, ‘Z. wiss. Zool.’, 41: 142-53, 1 Taf.

—— (1884a).—‘“* Report on the Nudibranchiata’’, ‘ Zool. Challeng.’.
10: 1-151, 14 plates.

(1885).—“‘ Beitriige zur Kenntniss der Aeolidiaden, VIII’’, ‘ Verh.

zool.-bot. Gesell. in Wien’, 35: 1-60, 7 Taf.

(1888).—*‘ Beitrage zur Kenntniss der Aeolidiaden, IX (Melibe
ocellata Bergh) ’’, ibid., 38: 673-706, 4 Taf.

—— (1890).—‘** Report on the results of dredging, under the supervision
of A. Agassiz, in the Gulf of Mexico (1887 and 1888), and in the Carib-
bean Sea (1879 and 1880) by the U.S. Survey steamer ‘ Blake’, Lieut.,
Commander C. D. Sigsbee, U.S.N., and Commander J. R. Bartlett,
U.S.N., commanding ”’, * Bull. Mus. C. Z. Harvard College’, 19: 155-81,
3 plates.

Qq2

~

580 H. P. KJERSCHOW AGERSBORG

Bergh, R. (1890a).—“‘ Die cladohepatischen Nudibranchien’”’, ‘ Zool.
Jahrb.’, 5: 1-75.

—— (1890 b).—‘* Die Nudibranchien des Sunda-Meeres”’, C. Semper,
* Reisen im Archipel der Philippinen, IL’, 2: 873-991, 4 Taf.

—— (1880-92.—‘‘ System der nudibranchiaten Gasteropoden”’, ibid.,
3: 995-1165.

—— (1892 .a).—‘‘ Die Nudibranchiata holohepatica porostomata ”’, * Verh.
zool.-bot. Gesell. in Wien’, 42: 1-16.

(1894).—“‘ Reports on the dredging operations off the west coast of

Central America to the Galapagos, to the west coast of Mexico, and in

the Gulf of California, in charge of Alexander Agassiz, carried on by

the U.S. Fish Commission steamer ‘ Albatross’, during 1891, Lieut.-

Commander Z. L. Tanner, U.S.N., commanding. XIII, Die Opistho-

branchien’”’, ‘ Bull. Mus. C. Z. Harvard College’, 25: 125-233,

12 plates.

(1897).—* Ergebnisse einer zoologischen Forschungsreise in den
Molukken und Borneo. Opisthobranchien”’, “ Abb. Senckenb. naturf.
Gesell.’, 24 : 93-130, 2 plates.

—— (1902).—‘‘ Den Danske Expedition til Siam, I (1899-1900). Opistho-
branchiata”’, ‘D. Kgl. Danske Vidensk. Selsk. Skrift., 6 Rekke,
12: 153-218, 3 plates and map.

—— (1904).—** Nudibranchiata kladohepatica (Melibe pellucida, Bergh,
on the Columbia River, Washington) ”’, C. Semper, ‘ Reisen im Archipel
der Philippinen’, wiss. Resultate, IX, 6: 1-55, 4 Taf.

—— (1908).—‘** The Opisthobranchiata of South Africa (Tethymelibidae) ”.
‘Trans. S. Africa Phil. Soc.’, 17: 1-144, 14 plates.

Boas, J. E. V. (1886).—*‘ Spolia atlantica. Bidrag til Pteropodernes
Morfologi og Systematik samt til Kundskaben om deres geografiske
Udbredelse. Avec résumé en francais”, ‘D. Kgl. Danske Vidensk
Selsk. Skrift.’, 6 Rekke, 4: 1-248, 8 tavler.

(1920).—‘ Lerebog i zoologien.’ 5te udgave. 725 pp., 655 figurer.
Nordisk Forlag, Kristiania.

Bonnevie, Kristine (1916).—‘* Mitteilungen iiber Pteropoden. I. Beobach-
tungen iiber den Geschlechtsapparat von Cuvierina columnella, Rang ”’,
‘ Jenaische Zeitschr. f. Naturwiss.’ (N.F., 47 Band), 54: 245-76, 4 Taf.

Bouvier, E. L. (1891).—‘** Recherches anatomiques sur les Gastéropodes
provenant des campagnes du yacht l Hirondelle. Rapports de l'apparei
circulatoire artériel avec le systéme nerveux. (Opisthobranches.) ~
* Bull. Soc. Z. d. Fr.’, 16: 53-6:

Cooper, J. G. (1863).—‘‘ On new or rare Mollusca inhabiting the coast of
California, II’, ‘ Proc. Calif. Acad. Nat. Sci.’ (1863-7), 3: 60.

Copeland, M. (1918).—‘‘ The olfactory reactions and organs of the marine
snails Alectrion obsoleta Say, and Busycon canalicu-
latum Linnaeus ”’, ‘ Journ. Exper. Zool.’, 25: 177-227.

MORPHOLOGY OF MELIBE 581

Crozier, W. J., and Arey, L. B. (1919).—‘‘ Sensory reactions of Chromodoris
zebra’, ibid., 29: 261-310.

Cuvier, G. v. (1798).—‘ Tableau élémentaire de histoire naturelle des
animaux.’ 710 pp., 14 plates. Paris.

(1817).—*‘ Le régne animal, etc.”’, ‘ Les Mollusques ’, 3: 351-494.

—— 1844).—** Observations nouvelles sur quelques Mollusques ’’, ‘ Bull.
d. sci., Soc. Philom.’, 2: 52-3.

Edwards, H. Milne (1840).—** Observations sur les organes sexuels des
divers Mollusques et Zoophytes”’, ‘Ann, Scienc. natur.’, 2 sér., Zoologie,
13: 375-6.

(1842).—‘‘ Observations sur la structure et les fonctions de divers

Mollusques ’’, ibid., 2 sér., Zoologie, 18: 321-50.

(1845).—‘‘ Observations et expériences sur la circulation chez les

Mollusques ’’, ‘C. R. Acad. Sci. Paris’, 20: 261-77.

(1847).—‘*‘ De l appareil circulatoire des Thetis”, ‘Ann. Science.
natur.’, 3 sér., Zoologie, 8: 64-71, 1 plate.

Egede, Hans (1745).—‘ A description of Greenland (translated from the
Norwegian).’ 220 pp., 12 plates.

Eliot, C. N. E (1902).—‘‘ On some nudibranchs from Zanzibar ’’, ‘ Proc.
Zool. Soc. London’, 2: 62-72, 2 plates.

(1906).—“‘ The Nudibranchiata of the Scottish National Antarctic

Expedition ”’, ‘ Trans. R. Soc. Edin.’, 41: 519-32.

(1910).—* A monograph of the nudibranchiate Mollusca.’ The Ray
Society, London. 197 pp., 8 plates.

Eliot, C. N. E., and Evans, T. J. (1908).—“‘ Doridoides gardineri:
a doridiform cladohepatic nudibranch”, ‘Quart. Journ. Micr. Sci.’
(N.S.), 52: 279-99, 2 plates.

Fewkes, W. J. (1889).—‘* New Invertebrata from the coast of California’.
Printed for the author. 50 pp. Boston.

Filippi, F. de (1866—7).—(An extract from de Filippi’s letter addressed to
the Academy at Turin, containing among other things reference to
Jacunia papillosa, s. Melibe papillosa, &c.) ‘ Atti della R. Accad. delle
Scienze di Torino, pubblic. dagli Accademici sec. del due Classi,’ 2: 233.

Flemming, W. (1869).—‘* Die haartragenden Sinneszellen in der Oberhaut
der Mollusken ”’, ‘ Arch. f. mikr. Anat.’, 5: 425-44, 1 Taf.

(1870).—*‘ Untersuchungen iiber Sinnesepithelien der Mollusken ”’,
ibid., 6: 439-71, 2 Taf.

Frenzel, J. (1886).—‘* Mikrographie der Mitteldarmdriise (Leber) der
Mollusken. JI. Allgemeine Morphologie und Physiologie des Driisen-
epithels ’, “Acta Acad. Caes. Leopold.-Carol. Germanicae’, (2), 48:
83-296, 3 Taf.

—— (1894).—** Micrographie der Mitteldarmdriise (Leber) der Mollusken.
II. Specielle Morphologie des Driisenepithels der Lamellibranchiaten, Pro-
sobranchiaten und Opisthobranchiaten”, ibid., (3), 60 : 321-404, 4 Taf.

582 H. P. KJERSCHOW AGERSBORG

Gamble, F. W. (1892).—‘‘ Observations of two rare British nudibranchs
(Lomanotus genei, Verany, and Hancockia endactylota, Gosse)”’,
‘Ann. Mag. Nat. Hist.’, 9: 374-85, 1 plate.

Glaser, O. C. (1903).—‘‘ The nematocysts of nudibranch molluscs ”’, ‘ Johns
Hopkins Univ. Cire.’, 22: 22-4.

Goodrich, Edwin 8. (1895).—‘‘ On the coelom, genital ducts, and nephri-
dia’’; ‘ Quart. Journ. Micr. Sci.’, 37: 477-510, 2 plates.

Gould, A. A. (1852).—‘ Mollusca and shells (Chioraera leonina s. Melibe
leonina, Gould).’ United States Exploring Expedition, during the
years 1838, 1839, 1840, 1841, 1842. Under the command of Charles
Wilkes, U.S.N. Boston. 12: xv, 510, Pl. 52, 1 Atlas.

Gross, A. O. (1921).—‘‘ The feeding habits and chemical sense of Nereis
virens Sars”’, ‘Journ. Exper. Zool.’, 32: 427-42.

Hancock, A. (1865).—‘‘ On the structure and homologies of the renal organs
in the nudibranchiate Mollusca”, ‘Trans. Linn. Soc.’, 24: 511-30,
6 plates.

Hancock, A., and Embleton, D. (1848).—‘‘ On the anatomy of Eolis,
(not a new genus of molluscs) of the order Nudibranchiata ’’, ‘ Ann. Mag.
Nat Hist.’, (2), 1 : 88-105, 2 plates.

(1851).—‘* On the anatomy of Doris”’, ‘ Report Brit. Assoc.

Adv. Sci. held in 1850’, p. 124.

(1852).—‘** On the anatomy of Doris ”’, * Phil. Trans. Royal Soc.
London’, p. i, 142: 207-52, 7 plates.

Hartmann, R. (1880).—‘‘ Ueber die Organisation von Tethys fimbriata,
Linnaeus (s. fimbria)”’, ‘ Sitzber. naturf. Freunde’, 1880: 9-12.

Hecht, E. (1895).—‘‘ Contribution a Pétude des Nudibranches”’, ‘ Mém.
Soc. Zool. France’, 8: 537-711, 5 plates.

Heath, H. (1917).—‘‘ The anatomy of an eolid, Chioraera dalli” (s.
Melibe leonina), ‘ Proc. Acad. Nat. Sci. Phila.’, 69 : 137-47, 3 plates.

Herdman, W. A., and Clubb, J. A. (1892).—** On the innervation of the
cerata of some Nudibranchiata ”’, ‘ Quart. Journ. Micr. Sci.’, 33 : 541-58,
3 plates.

Hertwig, R. (1912).—* A manual of zoology,’ translated by J. 8. Kingsley,
New York. 675 pp.

Jeffreys, J. G. (1869).—** History of naked marine Gasteropoda ”’, ‘ Brit.
Conchol.’, 5: 1—258, 102 plates.

Jhering, H. v. (1876).—** Tethys: ein Beitrag zur Phylogenie der Gastro-
poden ”’, ‘ Morph. Jahrb.’, 2: 27-62, 1 Taf.

(1877).—‘ Vergleichende Anatomie des Nervensystems und Phylo-
genie der Mollusken.’ Leipzig, 290 pp., 8 Taf., 16 Holzschnitte.

Lacaze-Duthiers, Henri de (1859).—‘‘ Histoire anatomique et physio-
logique du Pleurobranche orange ’’, ‘ Ann. Sci. Nat.’, 4 sér., Zoologie,
11: 199-302, 12 plates.

Lang, Arnold (1896).—‘ Text-book on comparative anatomy.’ Translated
by H. M. and M. Bernard. Pt. ii. London. 618 pp.

MORPHOLOGY OF MELIBE 583

—— (1900).—** Lehrbuch der vergleichenden Anatomie cer wirbellosen
Thiere”’, 2. umgearb. Aufl., 1. Lief. ‘Mollusca.’ Bearb. von K.
Hescheler. Jena. 509 pp., 410 figures.

—— (1903).—*‘ Beitrige zu einer Trophocéltheorie. Betrachtungen und
Suggestionen iiber die phylogenetische Ableitung der Blut- und Lymph-
behalter, insbesondere der Articulaten. Mit einem einleitenden Abschnitt
iiber die Abstammung der Anneliden ”’, ‘ Jenaische Zeitschrift Naturw.’,
N.F., 31, 38: 1-376, 6 Taf.

Lange, Arthur (1902).—‘‘ Ueber den Bau und die Funktion der Speichel-
driisen bei Gastropoden ”’, ‘ Anatomische Hefte ’, 19: 85-153, 1 Taf.
Lankester, E. Ray (1881).—‘* On the originally bilateral character of
the renal organs of Prosobranchia, and on the homologies of the yolk-sac

of Cephalopoda ’’, ‘ Ann. Nat. Hist.’, (5), 7: 432-7, with a woodcut.

(1893).—‘‘ Note on the coelom and vascular system of Mollusca and

Arthropoda ’’, ‘ Quart. Journ. Micr. Sci.’, 34: 427-32, 4 diagrams.

(1906).—‘ A treatise on zoology’, pt. v, “Mollusca”, by Paul
Pelseneer. London. 355 pp.

Lansberg, B. (1882).—‘‘ Ueber das Herz und die Niere von Neritina
fluviatilis”, ‘Zool. Anz.’,5: 661-4.

Lenssen, J. (1899).—** Systeme digestif et systéme génital dela Neritina
fluviatilis”, ‘La Cellule’, 16: 179-232, 4 plates.

—— (1899a).—“ Anatomie de la Neritina fluviatilis’, ‘ Anat. Anz.’,
16: 401-4.

Leydig, F. (1876).—‘‘ Die Hautdecke und Schale der Gastropoden nebst
einer Uebersicht der einheimischen Limacinen”’, ‘ Arch. Naturgesch.’,
42: 209-92, 7 Taf.

List, J. H. (1887).—** Zur Kenntniss der Driisen im Fusse von Tethys
fimbriata Linnaeus,” ‘ Zeitsch. wiss. Zool.’, 45: 308-26, 1 Taf.

Lo Bianco, 8. (1888).—‘‘ Notizie biologiche riguardanti specialmente il
periodo di maturita sessuale degli animali del golfo di Napoli’’, ‘ Mitth.
Zool. Stat. Neap.’, 8: 385-440. (‘ Mollusca’, pp. 414-24.)

MacFarland, F. M. (1912).—“* The nudibranch family Dironidae ”’, ‘ Zool.
Jahrb.’, Suppl. 15, 1 : 515-36, 3 plates.

Martens, E. v. (1879).—‘* Uebersicht der von W. Peters von 1842 bis 1847
im Mossambique gesammelten Mollusken’’, ‘ Monatsber. Ak. Wiss.
Berl.’, pp. 727-49.

Mazzarelli, G.. F. (1889).—“ Intorno alle secrezioni della glandola opalina
(Vayssiére), e delle glandole dell’ opercolo branchiale nelle Aplysiae del
Golfo di Napoli”’, ‘ Zool. Anz.’, 12: 580-3.

—— (1889 a).—“‘ Intorno all anatomia e fisiologia dell’ apparato ripro-
duttore delle Aplysiae del Golfo di Napoli’’, ‘ Boll. Soc. Nat. Napoli’.
3: 120-8.

— (1890).—‘‘ Sul valore fisiologico della vescicola di Swammerdam delle
Aplysiae (Tasca copulatrice di Meckel) ’’, ‘Zool. Anz.’, 13: 391-9.

— (1891).—“ Ricerche sulla morfologia e fisiologia dell’ apparato

584 H. P. KJERSCHOW AGERSBORG

riproduttore nelle Aplysiae del Golfo di Napoli”’, ‘ Atti Reale Acc. Sci.
fis. e Mat. di Napoli’, Sec. Il, append. no. 5,4: 50 pp., 4 plates.

Mazzarelli, G. F. (1891 a).—‘‘ Intorno all’ apparato riproduttore di aleuni
Tettibranchi (Pleurobranchaea, Oscanius, Acera)’’, * Zool. Anz.’, 14:
233-43, 7 figures.

Meckel, J. Fr. (1808).—‘‘ Anatomie der Thetis leporina”, ‘ Beitr. z.
vergl. Anat.’, 1: 9-25, 2 Taf.

Meyer, H. A., und Moébius, K. (1865),—“‘ Fauna der Kieler Bucht”’, ‘ Die
Hinterkiemer oder Opisthobranchiata, Leipzig’, 1 : xxx, 1-88, 26 plates.

Moquin-Tandon, M. G. (1870).—‘* Recherches anatomiques sur !}Om-
brelle de la Méditerranée ”’, ‘ Ann. Sci. Nat.’, 5 sér., Zoologie, 14: 1-135,
8 plates.

O’ Donoghue, C. H. (1921).—‘‘ Nudibranchiate Mollusca from the Van-
couver Island region”’’, ‘Trans. Roy. Canad. Inst., Toronto’, 13:
147-209, 5 plates.

—— (1922).—** Notes on the nudibranchiate Mollusca from the Vancouver
Island region. J. Colour variation ”’, ibid., 14: 123-30, 1 plate.

O’ Donoghue, C. H., and O’ Donoghue, E. (1922).—*‘ Notes on the nudi-
branchiate Mollusca. Il. The spawn of certain species’, ibid., 14:
13143, 2 plates.

O’ Donoghue, C. H. (1922 a).—‘* Notes on the nudibranchiate Mollusca from
the Vancouver Island region. III. Records of species and distribution ”’,
ibid., 14: 146-67, 2 plates.

Parker, G. H. (1917).—‘* Pedal locomotion of the Sea-hare Aplysia
californica Cooper ’’, ‘ Journ. Exper. Zool.’, 24: 139-45.

Parker, T. J., and Haswell, W. A. (1910).—* A text-book of zoology ’.
New York, 1: 779 pp.

Parona, C. (1891).—‘‘ L’autotomia e la rigenerazione delle appendici dorsali
(Phoenicurus) nella Tethys leporina”’, ‘Zool. Anz.’, 14: 293-5.
Pease, W. H. (1860).—‘*‘Melibe pilosa. Description of a new species
of Mollusca from the Sandwich Islands’, ‘ Proc. Zool. Soc., London ’,

pt. 28: 34.

Pelseneer, P. (1891).—** L’hermaphroditisme des Nudibranches sacoglosses
(Elysiens, etc.) ’’, “ Ann. Soc. Malacol. Belgique ’, 4 sér., 6: 28-30.

—— (1891a).—** Sur quelques points d’organisation des Nudibranches
et sur leur phylogénie ’’, ‘ Bull. Soc. Malacol. Belgique ’, pp. 58-71.

(1893).—‘‘ Les appareils excréteur et reproducteur de Elysia”’,

‘Zool. Anz.’, 16: 458-60.

(1895).—** Hermaphroditism in Mollusca”, ‘ Quart. Journ. Micr.

Sci.’, 37: 19-46, 3 plates.

(1895 a).—*‘ L’hermaphroditisme chez les Mollusques”’, ‘ Arch. Biol. ’,

14: 33-62, 3 plates.

(1895b).—‘* Recherches sur divers Opisthobranches”’, ‘Mém. cour

Acad. Roy. Belgique’, 53: iii and 1-157, 25 plates.

MORPHOLOGY OF MELIBE 585

Pratt, H. 8. (1916).—* A manual of the common invertebrate animals,
exclusive of insects.’ Chicago. 739 pp.

Quatrefages, A. de (1844).—‘‘ Sur les Phlébentéres ordre nouveau de la
classe des Gastéropodes ”’, ‘C. R. Acad. Sci., Paris ’*, 18: 13-14.

—— (1844a).—* Observations générales sur le Phlébentérism”’, ibid.,
19: 1150-2.

~ (1848).—‘** Résumé des observations faites en 1844 sur les Gastéro-
podes phlébentéres ”’, ‘ Ann. Sci. Nat.’, 3 sér., Zoology, 10: 121-43.

Rang, 8S. (1829).—Melibee, Melibe Rang: Manuel des Mollusques,
Paris, pp. 129-30.

Sars, G. O. (1878).—‘* Bidrag til kundskaben om norges arktiske Fauna’,
“T. Mollusca regionis Arcticae Norvegicae”’. Christiania. xiii [iii]
466 pp., 52 plates. 1 map.

Schneider, K. C. (1902).—‘ Lehrbuch der vergleichenden Histologie der
Thiere. Jena. xiv and 988 pp.

—— (1908).—‘ Histologisches Practicum der Thiere fiir Studenten und
Forscher.’ Jena. ix and 615 pp.

Sedgwick, A. (1888).—‘‘ The development of the Cape species of Peri-
patus’’, pt. iv, ‘ Quart. Journ. Micr. Sci.’, 28 : 373-396.

(1898).— A student’s text-book of zoology’, ‘‘ Gastropoda ”’,
London. 1: xii and 619 pp., 472 figures.

Shipley, A. A., and MacBride, E. W. (1915).—‘ Zoology’, 3rd ed. Cam-
bridge. Ch. xiii.

Stempell, W. (1899).—** Zur Anatomie von Solemya tagota Poli”,
‘Zool. Jahrb. (Anat. u. Autogenie)’, 13: 89-170, 3 plates.

Tait, J., and Gunn, J. D. (1918).—** The blood of Astacus fluviatilis:
A study in Crustacean blood, with special reference to coagulation and
phagocytosis’, ‘ Quart. Journ. Exper. Physiol.’, 12: 35-80, 3 plates.

Tapparone - Canefri, C. (1876).—‘‘Genere Melibe Rang; Melibe
papillosa de Filippi”’’, “Mem. della Reale Accad. delle Sci., Torino’,
2 sér., 28: 219-20. :

Trinchese, 8S. (1881).—‘ Aeolididae e famiglie affini del Porto di Genova,
pt. i., Atlante’. Bologna, 1877-9. Large 8vo, 40 plates.

—— (1883).— Aeolididae e famiglie affini del Porto di Genova,’ pt. ii.
Testo Bologna. 109 pp., 81 plates.

—— (1894).—*‘ Nervi motori e nervi sensitivi del Phyllobranchus bor-
gninii”’, ‘Th. Rend. Acc. Sci. fis. e mat., Napoli’, 2 sér., 8: 190-1.

Tryon, Jr., G. W. (1883).—‘ Structural and systematic conchology ’,
Philadelphia, 2: 430 pp., 69 plates.

Vayssiére, A. (1877)—‘‘ Sur un nouveau genre de la famille des Tri-
toniades ’’, ‘C. R. Acad. Sci., Paris’, 85: 299-301.

(1880).—** Recherches anatomiques sur les Mollusques de la famille

des Bullides ’’, * Ann. Sci. Nat.’, 6 sér., Zoologie, 9: 1-123, 12 plates.

(1885).—** Recherches zoologiques et anatomiques sur les Mollusques

586 H. P. KJERSCHOW AGERSBORG

opisthobranches du Golfe de Marseille. re partie, Tectibranches ”’,
‘Ann. Mus. Marseille’, 2 Mém., 3: 181 pp., 6 plates.

Vayssiére, A. (1898).—‘‘ Monographie de la Famille des Pleurobranchides”’.
‘ Ann. Sci. Nat.’, 7: 209-402, 16 plates.

——— (1901).—‘*‘ Recherches zoologiques et anatomiques sur les Opistho-
branches du Golfe de Marseille. 3éme partie, Nudibranches ”’, “Ann.
Mus. Marseille *, 6: 1—130, 7 plates.

(1901 a).—‘‘ Etude comparée des Opisthobranches des cétes frangaises
de l Océan atlantique et de la Manche avec ceux de nos cétes méditer-
ranéennes ”’, * Bull. Sci. France et Belgique’, 34: 281-315.

—— (1911).—*‘ Recherches zoologiques et anatomiques sur les Opistho-
branches de la Mer Rouge et du Golfe d’Aden”’, ‘Ann. Fac. Sci., Mar-
seille ’, 20: 5-157.

Viguier, C. (1898).—‘‘ Recherches sur les animaux inférieurs de la Baie
d’ Alger. V. Contribution a l'étude du développement de la Tethys
fimbriata’”’, ‘ Arch. Zool. Expér.’, 3 sér., 6: 37-62, 3 plates.

Ward, H. B. (1900).—‘‘ Excretion in Mollusca’, ‘Amer. Nat.’, 34: 151-4.
(Review of: L. Cuénot’s (1899), “‘ L’Excrétion chez les Mollusques”’,
‘ Arch. Biol.’, 16: 49-96, 2 plates.)

Weber, Max (1880).—‘* Ueber den Bau und die Thitigkeit der sogenannten
Leber der Crustaceen ”’, ‘ Arch. mikr. Anat.’, 17: 385-457, 2 plates.

VII. Note to EXPLANATION OF FIGURES.

Figs. 11, 12, 18, 14, 15, 16, 46, 47, 48, 49, 50, 52, 53, 56, 57,
59, 61, 62, 63, 64, 65, 66, 67, 69, 72, and 74 were drawn by the
aid of Spencer’s compound microscope, tube length 16 mm.,
and camera lucida, angle of mirror 35. The others are either
free-hand drawings, microphotographs, or photographs from
preparations.

n.p.r.=low power with front lens removed. 1.p.=low power.
h.p.=high power. 0.i.=oil immersion. ¢.].=camera lucida.

VIII. EXPLANATION OF PLATES 27-87.

Fig. 1.—Photograph of Melibe leonina (Gould) from the ventral
side. Specimen preserved in 70 per cent. alcohol. C, cirrhi; F, foot;
Hp, hepatic diverticula; ZL, lips; M, mouth; Ml, marginal edge of
hoodaessaie

Fig. 2.—Photograph of a preserved specimen from the right side showing

MORPHOLOGY OF MELIBE 587

the profuse arborescence of the liver. Ao, anal pore; Ap, anterior right
papilla; F,foot; P, penis; R, tentacle. x1.

Fig. 3.—Photograph of a preserved specimen from which the ectoderm
and hepatic arborizations of the body-wall have been removed to show the
arrangement of the muscles of the body-wall. P, penis. 1.

Fig. 4.—Microphotograph of the skin to show the odoriferous glands
(Go); Hep, hepatic branches ; M, muscle-fibres.

Fig. 5.—Photograph of a whole-mount of the skin with the underlying
fibrous tissue to show the heterogeneous arrangement of the connective
tissues.

Fig. 6.—Photograph of a trans-section of the hood to show the tufted
or tuberculate surface of both the external side (Hx) and ventral side (Hn)
of the hood. In the ectoderm of the external side of the hood are also
shown the odoriferous (...) glands, which are absent from the ventral
side. x 24.

Fig. 7.—Microphotograph of the body-wall seen from the inside. It
shows the caecal endings of the hepatic diverticula. Contrast with
fig. 4, which is of the skin, photograph taken from without; note the
relative position of the odoriferous glands in the body-wall.

Fig. 8.—Photograph of dorsal tentacle with the sense organ (papilla of
Gould) (Rh) retracted. C, lamellated part; K, neural knob. See also
K in fig. 15. (Vide Agersborg, 1923, figs. 4, 5, pa.)

Fig. 9.—Schematic drawing of a dissected adult to show the general
arrangement of the visceral organs. A, anus; Aw, efferent branchial
veins; Br, brain; hc, hepatic trunks; Ft, foot; G, proventriculus ;
In, larger part of the intestine ; 7, mouth; Mg, mucous gland ; Oe, oeso-
phagus ; Od, oviduct; Ospt, ovispermatotheca ; Ot, ovitestes ; P, penis ;
Pc, pericardium ; Prg, prostate gland; Si, smaller part of the intestine ;
Vd, vas deferens; V, ventricle; Vg, vagina; Sto, gizzard; Pd, pyloric
diverticulum.

Fig. 10.—Part of the rim of the hood to show the arrangement of the
cirrhi. Jc, inner row; Oc, outer row; M, muscle-fibres.

Fig. 11.—Cross-section of a large cirrhus, showing the axis at Lem
from which fibres radiate (Rf) to the sub-epithelial (Ze) layer of the peri- _
phery. Oe, outer epidermal layer. (I.p., c.1.)

Fig. 12.—The axis of the cirrhus shown in fig. 11. Cem, smaller central
cell-masses with a reticular structure; Cg, central ganglion; Lem, large
central cell-mass with a few cells scattered ; Rf, radiating fibres. (h.p. c.1.)

Fig. 13.—Same as fig. 12. Jr, inner reticular network of the large central
cell-mass (Lem); Pc, peripheral cells; Cga, central ganglion. 1,013-4;
(o.1., ¢.1.).

Fig. 14.—Part of the periphery of cirrhus as shown in fig. 13. Je, sub-
epithelial layer with which the radiating nerve-fibres (Rf) communicate ;
Oe, super-epithelial layer. »1,013-4; (0.i., ¢.1.)

Fig. 15.—Longitudinal section through the sense organ of the dorsal

588 H. P. KJERSCHOW AGERSBORG

tentacle (lamellar papilla of Gould). N/fb, nerve-fibres of a highly gan-
glionic knob; Nfi, the same fibres in the lamellated part; N/fp, nerve-
fibres which communicate with the peripheral ganglion below (S);
Ne, small nerve-cells surrounding central fibres; Nes, ganglionic knob,
consisting of nerve-cells only ; Nm/, neuro-muscular fibres which inter-
communicate between the sense organ and the base of the tentacle ;
Owr, the outer wall of the tentacle which serves as a sheath to the organ ;
S, lamellae. 75.

Fig. 16.—Nerve-cells from * K ’ in figs. 8and 15. Nc, larger nerve-cells ;
Nes, smaller nerve-cells. Notice the large granular nucleus. » 1,013-4.

Fig. 17.—Front view of Melibe leonina. Fh, rim of hood; Fy,
front; Ft, foot; Jrc, inner row of cirrhi; Jrc', mid-ventral rudimentary
cirrhi; L, lip; M, mouth; Mf, muscle-fibres ; dp, mid-dorsal depres-
sion; Orc, outer row of cirrhi; Ff, base of tentacle. 14.

Figs. 18-24.—Papillae. Figs. 18, 19, first pair; 20, one of the second
pair; 21, one of the third pair; 22, one of the fourth pair; 23, one of
the fifth pair; 24, the sixth pair. Hep, caecal terminal branches of the
liver. The parallel lines represent muscle-fibres. x 1.

Fig. 25.—Longitudinal section of an anterior papilla. Cshb, cross-
section of a hepatic branch; Wf, muscle-fibres ; Og, odoriferous gland ;
Osp, vascular space or sinus; Tbr, tubercles. 2.

Fig. 26.—Cross-section through the anterior region of the foot. Cil,
cilia; Lv, liver; Mc, muscle-bundle; Og, odoriferous gland; Tbr,
tubercles. 8.

Fig. 27.—\Cross-section through the posterior region of the foot. Be,
connective tissue ; Gl, pedal ganglion. The other labelling as the preceding.
x 8.

Hig. 28. Section of a single tubercle from the anterior end of the foot.
Cec, ciliated columnar ectoderm; Cul, cilia; Gmug, granules of mucous
glands; Grc, granular border; Mf, muscle-fibres; Nf, nerve-fibres ;
Ng, nerve-cells ; Nu, nucleus.

Fig. 29.—Section through the pedal ganglion of the foot (vide fig. 27).
Cec, ciliated columnar ectoderm; Cuil, cilia; Gl, ganglionic cells; Mf,
muscle-fibres ; Mug, mucous glands with crypts passing through ectoderm ;
Nf, nerve-fibres; Ng, nerve-cells; Nt, neural fibrillae ending on to ciliated
ectoderm ; Pdng, pedal ganglion.

Fig. 30.—Section through the body-wall. bv, body-wall; Cr, crypt;
Ct, connective tissue; Hc, ectoderm; Sm, saccular mucous gland; Jb,
muscle-bundle ; Wf, muscle-fibres; Glo, odoriferous glands; Um, uni-
cellular gland.

Fig. 31.—Section from part of the wall of a large papilla (see fig. 25).
Ce, connective-tissue cells; Ch, cross-section of hepatic branches; Ct,
connective-tissue fibres; Glo, odoriferous gland; Pec, ectoderm of
fimbriated surface.

MORPHOLOGY OF MELIBE 589

Fig. 32.—From a whole mount of muscle-cells taken from the body-wall.
The fine granular appearance along the periphery of the cells is represented
in the next figure by ‘ Sar’; Nu, nucleus.

Fig. 33.—Cross-section of muscle-bundle showing section of a few fibres
only. Ca, axial sarcoplasm; Jntc, inter-fibre connective-tissue cell ;
K, chromatin granules; Lin, linin; Mc, muscle-cell cut through the centre ;
Mcp, peripheral micromeric region of cell; JJ/fe, muscle-cell cut near its
tapering end; Ms, perimysium; M/, endomysium; My, myofibrillae ;
Nuco, nucleolus; Nu, nucleus; Nus, nuclear sap; Sar, sarcolemma.

Fig. 34.—A few connective-tissue cells and muscle-fibres from the body-
wall. Jf, muscle-fibres ; Pnc, granular connective-tissue cells.

Fig. 35.—Photograph of a cross-section of the pharynx near the mouth
to show the corrugations (Cor) of the lining.

Fig. 36.—Microphotograph of a median sagittal section of the oeso-
phagus. Oe, oesophagus; G, proventriculus; S, gizzard; br, brain. x 8.

Fig. 37.—Microphotograph of a cross-section through the gizzard.
Hep, hepatic canals into the gizzard; V, ventral side. x 14.

Fig. 38.—Microphotograph of a cross-section of the posterior part of the
proventriculus showing the end of the glandular lining. F, remnants of
tood; Cm, circular muscle-layer; Gl, glandular mucous lining; 1,
stomach-contents (mucus ?). 22.

Fig. 39. Microphotograph of a cross-section through the anterior part
of the pyloric diverticulum. Mu, mucous coat; Mus, muscle-layer ;
Ty, typhlosole. x18.

Fig. 40. Microphotograph of a cross-section of the larger portion of the
intestine. Hx, mucosa; Ty, typhlosole; Veb, ventral blood-vessel.

x 37.

Fig. 41. Drawing of a longitudinal section of the oesophagus. Ac, con-
nective-tissue cells; Hpt, epithelial lining; Mst, muscle-fibres; Nw,
nucleus ; Sg, salivary glands; V, crypts of salivary glands.

Fig. 42.—Drawing of a cross-section of the gizzard. bm, basement
membrane of the endoderm lining of the gizzard; Cc, external cover ;
Ept, epithelium of the gizzard ; Mus, circular muscle-layer ; Stpl, stomach-
plates; Ts, transitional parts of the stomach-plates, a product of the
epithelial lining.

Fig. 43.—Longitudinal section of the gizzard. Ct, connective tissue,
and a few ‘muscle-fibres; Cs, circular muscle-layer cut transversely ;
Ept, epithelial lining showing several secretion vacuoles in the cytoplasm
and nuclei in various conditions of the so-called resting stage: Kar,
chromatin granules; Lin, linin; Nwuco, nucleolus; Nu, nucleus; Ol,
external cover; St.pl, stomach-plates; Sv, secretion vacuoles ; T'rp.pl,
transitional part of stomach-plates.

Fig. 44.—Drawing of one of the corrugations from the pyloric diverti-
culum (vide fig. 39). Bg, basal granules (terminal bars); Bm, basal

590 H. P. KJERSCHOW AGERSBORG

membrane ; Cuil, cilia; Sb, sub mucosa; Vas, highly vascular cover of
the organ; that is, the tissue is very loose and seems to contain many
spaces or sinuses; Vac, mucus vacuoles.

Fig. 45.—Cross-section of the wall of the smaller intestine. 5g, terminal
bars; Bm, basement membrane; Cc, connective-tissue cover; Cuil,
cilia; Kar, chromatin granules; Lcc, loose connective-tissue layer or
sub mucosa with many vascular sinuses; Nwco, nucleolus; Nu, nucleus ;
Nus, nuclear sap; Sv, secretion vacuoles.

Figs. 46, 47.—The same as fig. 45. The secretion vacuoles are not so
large. Bg, terminal bars; Cul, cilia; Sv, secretion vacuoles. x 1,013-4;
(o-2:¢e21))

Fig. 48.—Longitudinal section of one of the main branches of the
hepatic caeca. Nu, nucleus; Sc, secretion cap (product). x 1,013-4;
(Ostes cel5)

Fig. 49.—Longitudinal section of hepatic branch, showing the very
variable condition in adjacent cells. (h.p., ¢.1.)

Fig. 50.—Cross-section of hepatic branch in the body-wall of the hood.
Sg, secretion product. 1,013-4'; (0.i., ¢.1)

Fig. 51.—Cross-section of hepatic branch from the body-wall. Dsc,
darkly staining cells; Gc, vacuoles; L, lumen; Nw, nucleus.

Hig. 525 Hrom* x 7/in fig, Sa. “(h-p:, cL)

Fig. 53.—Cross-section of a branch of the liver in the tentacle. 107.
(ips.e:1)

Fig. 54.—Schematic drawing of the heart and part of the pericardium.
Aa, anterior arteries ; Ao, descending aorta; Aosh, very fine transparent
sheath of the aorta; Au, efferent branchial veins; Per, pericardium ;
Vent, ventricle.

Fig. 55.—Longitudinal section through the auricular part of the heart.
Au, efferent branchial veins; Av, auricular apertures in the auriculo-
ventricular valves (Ar); Bc, blood-cells; Hnd, epicardium; Vent,
cardiac wall of the ventricle.

Fig. 56.—Blood-gland from the wall of the aorta just below the ventri-
cular valve (vide fig. 59). Aosh, sheath of aorta; LZ, lumen. x 107.
(Ep: sne:15)

Fig. 57.—Longitudinal section of the aortic wall opposite the blood-
gland seen in figs. 56 and 59. Cm/f, cardiac muscle-fibres. Section
15 micra thick. x1,013°4; (0.1., c.1L.)

Fig. 58.—Blood-cells from the lumen of the aorta. a, 6, from the
inside of the blood-gland (node), fig. 56; c, from the lumen close to the
gland. 1,013-4; (0.1., c.L)

Fig. 59.—Longitudinal section of the lower part of the ventricle of
the heart. Bgl, blood-gland (node); Pc, pericardium; Valve, valves
between the aorta and the ventricle; Vw, ventricular wall; Wda, wall
of the aorta. (h.p., c.1.)

MORPHOLOGY OF MELIBE 591

Fig. 60.—Schematic drawing of the kidney. Ab, anterior branches ;
Pb, posterior branches ; P, point of communication with the pericardium ;
Rs, renal syrinx; U, ureter; Up, uretero-connexion with the renal
syrinx.

Fig. 61.—Section through the renal syrinx showing its connexion with the
ureter (Ur) and pericardium (Rs). Note the arrangement of the cilia of
the renal syrinx ; Pw, pericardium. (n.p.r., c.1.)

Fig. 62.—From 4A, in fig. 61, showing the transition of the epithelium
from ciliated to non-ciliated. (h.p., c.l.)

Fig. 63.—From B, in fig. 61, showing the cyncitial relation of the cells.
(hp :c.1.)

Fig. 64.—Section through the renal syrinx showing complete connexion
with the pericardium, A—A. Crs, cavity of renal syrinx; Pl, plica or
villus of the cyncitial plate; Ur, ureter. (n.p.r., c.1.)

Fig. 65.—A-—A in fig. 64. Pl, villus of the cyncitial plate. 107.
(ip., c.l.)

Fig. 66.—From B in fig. 64. Pl, plica or villus. (h.p., ¢.1.)

Fig. 67.—The cyncitial plate with villus (Pl). (h.p., ¢.1.)

Fig. 68.—From the wall of the renal syrinx. Cil, ciliated columnar cells
lining the organ ; Sw, syringeal wall covering the epithelium. (h.p., c.l.)

Fig. 69.—A single cell from the wall of the renal syrinx showing its
remarkable structure. Bg, basal granules; Bm, basement membrane.
-O13-4 = (0:1;,.c.1.)

Fig. 70.—Cross-section of a renal branch. Ct, connective-tissue capsule ;
Pr.neph, periodically functioning cells. 267.

Fig. 71.—Section from the wall of the ureter. nt, internal border :
Ex, external cover; Nu, nucleus; Nuco, nucleolus.

Fig. 72.—Longitudinal section of the ureter nearer to the kidney than
that part shown in fig. 71. Note the glandular condition. »1,013-4;
(oxi; ¢-1:)

Fig. 73.—Microphotograph of a cross-section of the body in the region
of the prostate coils of the female genital tube. Hep, hepatic branch :
Int, intestine; Pro, large coils of the prostata; St, stomach; Vd, vas
deferens. x 14.

Fig. 74.—Microphotograph of a cross-section through the region of the
mucous gland. Fd, food in the stomach; Hep, section of the liver;
Mg, mucous gland; P, penis; Sept, ovispermatotheca; Vg, uterus.
x 14.

Fig. 75.—Microphotograph of a cross-section of the posterior part of the
penis. Lm, lumen. Two lumina are shown owing to the coiling of the
organ in this region. 15.

Fig. 76.—Microphotograph of a trans-section of the body in the region
of the brain. Cg, cerebral ganglia; P, penis; Mg, mucous gland. x 14.

Fig. 77.—Microphotograph of a trans-section of the body in the region

592 H. P. KJERSCHOW AGERSBORG

of the genital pores. Fv, female genital vestibule; Mga, male genital
aperture; St, stomach. This section is of particular interest
since it shows the two separate openings. X15.

Fig. 78.—Microphotograph of a cross-section of the penis. lg.lp,
glandular lining of the penal cavity; Lm, lumen; Og.lp, glandular cover
of the penis; M/s, musculo-fibrous part of the penis.

Fig. 79.—Microphotograph of a cross-section of the body in the region
of the female genital duct, showing the many corrugations of the uterus
(Ut). Mg, mucous gland; Pro, prostata; St, stomach. 14.

Fig. 80.—Microphotograph of a cross-section of the uterus, showing
semen in its lumen (Sem). F, uterine pocket; Gil, glandular lining ;
Ms, muscular wall. x42.

Fig. 81.—The organs of reproduction in Melibe leonina. Bil.dpr,
biluminate ampullo-prostate duct; Amp, ampulla; Cl.p, penal cleft in
the mucous gland; Mgl, mucous gland; Od, oviduct; O.sp, ovisperma-
totheca; Ot, ovitestis; P, penis; Pr,prostata; Sv, seminal vesicle of
the penis; Ut, uterus; Va, vagina; Ve, vasa efferentia. 2.

Fig. 82.—Drawing of portion of a cross-section of the ovispermatotheca.
Bm, basement membrane ; Hpt, tall columnar epithelium lining the organ ;
Mf, muscle-cell cut longitudinally; Nu, nucleus; Nuco, nucleolus ;
Kar, chromatin granules; Lin, linin; Mic, micromeres; Myf, myo-
fibrillae ; Ret, reticular structure or axial myofibrillae; L, loose con-
nective-tissue cover.

Fig. 83.—Drawing of part of a cross-section of the penis. Cul, cilia ;
Cc, connective-tissue cells ; Hxcpl, external epithelial cover of the penis ;
Tepl, internal epithelial lining of the organ ; Jnt.ml, internal fibrous layer ;
M, mucous substance ; Oc, lining of the penal cavity. » 36.

Fig. 84.—Drawing of a few cells from the mucous gland, showing the
gland to be composed of tall columnar epithelium with a short ciliated
border. Bg, basal granules; Bm, basement membrane; Ctc, connective-
tissue capsule; cb, free ciliated border of the internal surface of the
gland; Mic, micromeres; Nu, nucleus, small and shrunken; Penu,
nucleus of actively filling cells; Rs, reticular structure of cells in the state
of refilling. The cells in A are in a state of exhaustion. They stained
poorly with Delafield’s haematoxylin. B represents a condition prior
to liquefaction of the granula. Note the nuclear membrane seems to be
wanting and there are a number of small chromatic granules near the
nucleus. The nucleus (Pcnu) is filled with uniformly sized chromatic
bodies. The nucleolus being in some cases irregular in shape. C represents
cells like A and B, but in C the micromeres (mic) are larger than in A,
and the nucleus not quite so shrunken. That is, these cells are in a state
of refilling. D, cuboidal epithelium from a loop of a non-glandular part
of the organ.

7
>
.
>

4g

rt 5 S 1
= e
‘
1 a - .
‘
- 7
'
a 7) = |
~ <
2
‘
,
x b
ime ;
a, é ~
i Se: 3

wa

“hy

1 a i I

Quart. Journ. Micr. Sci. Vol. 67, N.S., Pl. 27

~!

»
AR ie, dees.
s bathe a
‘s Arye
.: ja"
»

Ft Fi an 2
ae that

Hes fy is 1

S

&,

Vol. 67, N.S., Pl. 28

Cl.

Si

Quart. Journ. Micr

Ni

Eh pen, IR em eae a ms

13

aE at

TATE ET
nce,

A . 7”
‘ e
; m t* rs (he 8 ‘ T
: az f ‘
me = BF oD oo
A fod ‘
#1,
fal
é.
= ue
4
a
a ~
r .

Quart. Journ. Micr. Sci. Vol. 67, N.S., Pl. 29

nal 7 u/ a
~ ~ i
: = ie te
+
=5 2
L i mat vio es 2 ¢ Fin
aw es itet a wy ork t ~
7
~ .
‘
-
4
)
\ ’
«= 3
«
-
-
‘
‘
*
be
)
,
/

Quart. Journ. Mier, Sci. Vol. 67, N.S., Pl. 3

re)

G
oO Sk

Say

Wy,

Mi,

cies

$5-2N

a
“)
- a
;
’
ey + +“ hy
Pla”
+ ' he
» ; € ei
-
?
.
J
.
‘
'

Quart. Journ. Micr. Sci. Vol. 67, N.S., Pl. 31

[cco OCC

Quart. Journ. Micr. Sci. Vol. 67, N.S., Pl. 32

ee
= BEER. SS =

=

Caines
aN rai

i ‘I

Hi A VN i Ny
| i cae
ae -.

=—— =

iN Mr i

Wil

I ql a

Mt

SS

Vol.107, INOS. Bese

Quart. Journ. Micr. Sci.

NG
‘

re Thee ere as :
en VA ws le

Wek h
iit D
A BB Min
ul 1 iid

Ai | Ws
B { | q

Aosh

Pa -@

Volv 67, NOS) Piast

Quart. Journ. Mier. Sci.

Valve— — — — +

Po--- — ---%&

‘

AY
d
Be Qasr?

60

Vol. 67, N.S., Pl. 35

t,

Quart. Journ. Micr. Se

ee
Sg eee,

Vol. 67, N.S., Pl. 36

Quart. Journ Mier. Sci.

S., Plaae

r

, 67, N

Vol

Quart. Journ. Mier. Sci.

" ye ~

\ Kar \ Nuco.

4 >
And Re
ee BE E54 \ 4
: ou) A a*
Fae a < SS
> Benn

82

Ctc Bm Nu

Poenu

¢ - =
} 5
| .
.
*
SS

Observations upon the Behaviour and Structure
of Hydra.

By
Sheina Marshall, B.Se.,

Assistant Naturalist, Scottish Marine Biological Station, Millport.

With 4 Text-figures.

CONTENTS.

PAGE
INTRODUCTION . : ; : ; 3 : - : . 593
FEEDING . : é ; : ; : : ‘ : . 596
REACTION TO STIMULI ; : ‘ : P 3 ‘ 2 599
Eea FoRMATION : : ‘ ‘ : ‘ : : . 600
NEMATOCYSTS . ; : ; ; ‘ - . , 5 (G0R:
NeERyous SYSTEM : . ‘ : 2 : : : . 608
SYMBIOTIC CELLS Z j i F ; : , , ee Ollie
TAXONOMY : : , : ‘ ‘ : : ; 3) Gollllss
REFERENCES . F P : ; ; : : : . 6116

INTRODUCTION.

THe Hydras upon which the following observations were
made were obtained from various sources, chiefly through the
kindness of Dr. Monica Taylor, from the Convent of Notre-
Dame. They were kept in covered, half-pint glass tumblers,
in water from a large tank in which there was a fair quantity
of weed and a variety of animal life (Isopods, Cladocerans,
Planarians, &c.). This was used because the Hydra would
not live for more than one or two days in tap-water. The
water was ‘changed and the tumblers cleaned when necessary.
This was about once a week in summer, as food-remains became
foul very quickly then, but less frequently in winter. The
water was never aerated artificially. The Hydras were fed
on a culture of Daphnia twice a week, and the remains and
excreta removed, as far as possible, the followmg day. Under

NO, 268 RE

594 SHEINA MARSHALL

these conditions the Hydras lived and remaimed healthy for
months. Occasionally one or two, for no apparent reason,
would decrease in size and finally degenerate and die, but in
only one case was a whole tumbler attacked by a ‘ depression
period’. Even here a fair proportion remained healthy
throughout. When well fed the Hydras budded actively.
None of my specimens carried more than four or five buds
at a time and the number was usually less.

Sexual reproduction took place in autumn and in early
summer. The animals were hermaphrodite. With the excep-
tion of four specimens, not differmg outwardly from the rest,
all those which produced eggs produced testes at the same
time. ‘Testis formation usually began before egg formation
in any individual. Many specimens showed testes without
eggs, but as egg production entails a considerably larger
expenditure of energy and food material than testis produc-
tion, this is not surprismg. The four exceptional, apparently
female, specimens were kept under observation for about eight
weeks, but died before undergoing another sexual period.
Three or four eggs were sometimes formed at one time, although
they might not all attain full size and break through the
ectoderm.

One large and healthy specimen showed four eggs developing
and a number of testes. Three of the latter had developed in
the ectoderm of three tentacles. One testis was just at the
root of a tentacle, another a little further out, and the third
ata distance from its base of about quarter the length of the
tentacle. The testes were ripe and spermatozoa were swarming
inthem. In this animal interstitial cells must have been present
in the tentacles, which is not usually the case.

The period from the freeing of the egg to the hatching of the
young Hydra varies with the season. A batch of eggs set
free in November hatched in January, while some produced
in summer took only about three weeks to develop. In many
cases the eggs failed to hatch owing to the attacks of bacteria
or fungi. The young Hydra emerges by a crack in the shell,
usually equatorial. It is oval and almost colourless, but. in ©

OBSERVATIONS ON HYDRA 595

a few minutes it stretches out and extends little projections
which develop into tentacles during the next twenty-four hours.
It is then able to feed. The rupture of the shell suggests the
formation of a hatching ferment, such as has been described
in Lepidosiren and Teleostean fishes (15).
Abnormalities are not uncommon among newly hatched
Hydras, possibly due to injuries during hatching. Several
two-headed specimens appeared, the double part varying in
length. Such specimens are not infrequent among adult
Hydras. In one or two cases the division appears to be growing
gradually deeper, so that eventually the compound would split

Trxt-riG. I,

x 1 :

a b € dd e

into two separate Hydras, but in most cases the specimens
remained without change for weeks and died without further
division. In no case was there any suggestion that a process
of fusion was going on. In two specimens the hypostome only
was double and several of the tentacles belonged to both rings.

The origin of a double Hydra can sometimes be observed
(Text-fig. 1). One normal specimen produced two buds close
together (Text-fig. 1, a), and in the course of development
these grew out on a common stalk. Although one of the
buds was a day or two younger than the other, they soon grew
to equal size. This double individual then separated (Text-
fig. 1, b) from the normal parent, and a week later both of its
limbs produced normal buds which separated off. Two days
later a third bud was formed near the junction of the two
limbs (Text-fig. 1, c). The next day a curious pointed projection
grew up between the bud and one limb (Text-fig. 1, d). Hven-
tually this grew into a second bud joined to the first (Text-

Rr2

596 SHEINA MARSHALL

fig. 1, e), and the whole was separated as a double Hydra.
Unfortunately this specimen died before reproducing itself
further. It should be noted that the mode of origin of the
second half of the compound is distinctly abnormal.

Leiber (8) reports a case in which one component of a
double Hydra, just before the division had reached the
foot and the two were about to separate, itself divided again
at the hypostome. It died before further observations could
be made. ‘These cases indicate that a tendency towards
doubleness may be inherent in some Hydra stocks. From
the infrequency and origin of these abnormalities and the
frequent death of the compounds, it does not seem probable
that longitudinal fission is a normal method of reproduction.

I have seen transverse fission take place only in obviously
unhealthy specimens.

Hydras are sometimes found in which two or more tentacles
are in process of fusion from the base upwards. This seems to
be a method of regulating the number of tentacles to the size
of the Hydra, for it is found chiefly in animals which are
decreasing in size, or in those which have an exceptionally
large number of tentacles for their size. The commonest
numbers of tentacles were five, six, or seven, but some
specimens showed as many as ten.

The appearance of an animal undergoing ‘ depression’ has
been so often described that it is unnecessary to do so here.
In the early stages the Hydra may take on an appearance
very different from its usual, the differences bemg sometimes
those described as characteristic for another species.

For several months the Hydras I had were overrun with
Kerona and Trichodina, but they seemed none the
worse for it. This is contrary to the observations of P. Schulze
(9), who states that Kerona caused depression in his
animals.

FEEDING.

The Hydras were usually fed on a culture of Daphnia,
but were occasionally given Cyclops, Cypris, or small

OBSERVATIONS ON HYDRA 597

insect larvae, all of which they ate readily. Simocephalus
was also tried as a food, but the Hydra seemed unable to
lnllit. The Simocephalus were frequently caught and held
struggling for an hour or more, but in the end they freed
themselves and escaped uninjured. If killed and presented
to the Hydra, they were eaten as readily as Daphnia.
On three occasions a Simocephalus was captured and
eaten by a Hydra. In the first case the animal remained alive
inside the Hydra for a considerable time and could be seen
moving its antennae. A second was caught and digested
immediately after ecdysis. I have, however, seen a Hydra
catch a Simocephalus which then underwent ecdysis
and was immediately recaptured, yet eventually freed itself.
When caught, the contrast between the behaviour of Daph-
nia and of Simocephalus is remarkable. The Daphnia
struggles violently for a few minutes, then the heart stops
beating and the animal soon succumbs although the antennae
may keep up a quivering movement for some time longer.
If freed from the Hydra, it does not recover. Simo-
cephalus, on the other hand, continues to live and to
struggle at intervals until it frees itself. The heart continues
to beat the whole time. The cuticle of Simocephalus is
not appreciably thicker than that of Daphnia. It may be
more resistant to the entry of the nematocysts, or the tissues
of the animal to the action of their poison. When both
Daphnia and Simocephalus are immersed in dilute
solutions of poisons (such as formic acid, or chloroform)
Simocephalus succumbs first.

Schulze (9) mentions that in a culture of H. circum-
cincta a Daphnia sometimes stuck on to the tentacles
and was dropped again uninjured. It is perhaps possible that
these were really Simocephalus also, for the two genera
are closely similar and may be mistaken for one another
unless carefully examined.

The capture of the food generally seems to be a more or
less passive action, any small object presented to the Hydra
being seized and carried to the mouth. Indifferent substances

598 SHEINA MARSHALL

are usually dropped after a few seconds, although a few
Hydras were induced to swallow pieces of white of egg (which
were returned undigested in the course of twelve hours) and
pieces of blotting-paper when soaked in blood.

~ If Daphnia is left in a weak solution of litmus for eighteen
to twenty-four hours, the alimentary canal in the head-region
usually takes on a pink tinge. If these stained animals are fed
to Hydra, a colour-change from pink to blue takes place
in those animals which show nematocysts sticking into the
head-region. This may indicate an alkaline reaction for the
nematocyst poison. The change takes place slowly, often after
the death of the Daphnia, but does not occur when the
animal is killed with a needle. Simocephalus does not
take up the colour well, and often dies if left in the lhtmus
solution for twenty-four hours.

The digestive juices of Hydra are alkaline, but have
no effect on Hydra itself. I have seen one Hydra com-
pletely ingested by another, in whose cavity it remained for
more than twelve hours. It was returned again none the
worse. A tentacle is frequently swallowed along with the
food to which it is stickmg, and remains in the coelenteron
for some hours, but comes out quite unaltered. One Hydra
even went as far as to engulf about half its own body, beginning
at the foot, where a Daphnia had stuck.

In many cases, particularly where the animal was attempting
to swallow something exceptionally large, the hypostome was
turned inside out over the tentacle-bases and remaimed so for
some time. It rarely went further than this, but in one small
regenerating specimen the process went on till the whole
animal had turned inside out. It righted itself in the course
of an hour. The converse also takes place sometimes, and the
hypostome is turned inwards till it hangs down into the coelen-
teron. Both of these performances are of interest, masmuch
as a condition is assumed which has become permanent and
normal in other types of Coelenterata, e.g. the trumpet-
shaped hypostome of Obelia and the invaginated hypostome
of the Actinozoa.

OBSERVATIONS ON HYDRA 599

REACTION TO STIMULI.

Hydras were tested to find out their reaction to mechanical
and chemical stimuli.

Mechanical Stimulation was carried out with a glass
rod drawn to a fine point.

Slight stimulation of a tentacle leads to contraction of that
tentacle only. The tentacle often adheres to the rod for a few
seconds.

Strong stimulation of a tentacle leads to its contraction over
the mouth, when the other tentacles bend up till they meet
and the head turns to one side, exactly as when catching prey
(capture response). This response is sometimes obtained by
rubbing on the insides of the tentacle bases, and sometimes by
touching the oral cone. Strong stimulation of the oral cone
leads to contraction.

Stimulation of the outsides of the tentacle bases and of the
body in that region sometimes leads to contraction, but often
has no result.

Gentle stimulation of the body generally has no result, but
sometimes leads to contraction, as strong stimulation always
does.

Stimulation of the foot always leads to contraction of the
body (not of the tentacles, unless strong). This is probably
an adaptive reaction to movements of the object to which the
Hydra is attached. When contracted, the resistance to water
will diminish, and the animal will be less liable to be torn off.

Gentle touches repeated at tervals of five seconds or two
and a half seconds sometimes have no effect even when carried
on for several minutes. More often they lead, first, to a
swinging away of the body, and finally to contraction after
a few minutes. In swinging away the body is bent just above
the foot region. As the body remains quite straight the
action must be confined to a small number of muscle-fibres
in the region of the bend, which is usually in a different part
of the body from the stimulation. This shows the existence
of a conducting mechanism.

Chemical Stimulation.—When food, such as a piece

600 SHEINA MARSHALL

of Daphnia, is made to touch a tentacle tip, the capture
response immediately takes place. The food is thus brought
in to the hypostome ; and the tentacles, bending over, would
prevent escape were it alive.

Chemical stimulation was also carried out with weak acetic
acid coloured with methylene blue. The strength used was
0-025 per cent., but this was probably much weaker by the
time it actually reached the Hydra. It did not injure the
tissues appreciably.

Stimulation applied to Tentacle.—lIn about 60 per
cent. of the cases the tentacle contracted and the capture
response followed in about twenty seconds. In 30 per cent.
of the cases the whole Hydra swung away after forty
seconds, and in the remainder the animal contracted after
about thirty seconds. In one case a repetition of the stimulus
immediately after the response led to a second response
‘without any pause.

Body.—tThe body, when stimulated, either contracted or
swung away after thirty to forty seconds. In several cases
there was no result.

Oral Cone.—When the oral cone was stimulated the
body contracted, sometimes immediately, sometimes after
thirty to forty seconds.

Foot.—Stimulation of the foot resulted m a general con-
traction of the body, but it was very insensitive, and out
of eighteen stimulations twelve had no effect. In other cases
the body contracted after an interval varying from fifteen to
thirty seconds.

From the above it will be seen that the head and foot regions
are much the most sensitive parts of the Hydra to mechanical
stimuli, while the foot appears to be comparatively insensible
to chemical stimuli, at least to acetic acid.

Eaa ForMATION.

When the Hydra is about to form an egg the interstitial
cells are multiplied enormously and form a mass bulging
out the ectoderm. Characteristic changes take place in their

OBSERVATIONS ON HYDRA 601

nuclei. The chromatin collects round the periphery of the
nucleus, leaving the nucleolus in the centre of a clear space
(Text-fig. 2, a). The nucleolus sometimes becomes difficult
to stain, or only part of its periphery stains. Wager (18)

TEXT-FIG. 2.

4—* “8

coe ee ee ee oe
C.

Sections of ectoderm showing development of ovum.

believes that vacuoles form in it at this stage. Secondary
nucleoli may make their appearance, often in large numbers.
These may be droplets of food material as Wager suggests,
for the cytoplasm is also full of darkly staining particles. At
a stage just before the definite egg-cell becomes recognizable

602 SHEINA MARSHALL

several interstitial cells, much larger than the rest, may be
seen undergoing nuclear changes. The chromatin has formed
a thick spireme thread, and in some cases the nuclear mem-
brane has disappeared. ‘The nucleolus takes no part in this,
but lies unchanged in the midst of, or to one side of, the
spireme. This is in accordance with the observations of
Wager, who states that ‘the egg begins its growth by the
coalescence of a group of the primitive ova; this process
is frequently attended by a peculiar nuclear degeneration ’.

Later the egg appears as a hemispherical mass of protoplasm
with lobed edges, which les with its base in contact with the
mesogloea. The protoplasm is reticular, and at this stage
contains none of the so-called ‘ pseudo-cells’. The nucleus
corresponds im size with the cell and contains a large nucleolus
and a number of smaller secondary nucleoli varying in size.
The egg grows by the absorption of other interstitial cells,
but at this stage the nuclei of the latter break down com-
pletely before absorption. The protoplasm of the egg is filled
with minute deeply stamimg dots, probably, at least in part,
the remains of the chromatin of the absorbed cells.

At a later stage (Text-fig. 2, b) the egg protoplasm is filled
with a mass of degenerated cells (‘ pseudo-cells *) and inter-
stitial cell nuclei. Intermediate steps can be traced showing
the process of degeneration of an interstitial cell. The process
goes on both inside and outside the egg. In groups of inter-
stitial cells, usually at some distance from the egg itself, the
nucleus is seen in the centre of the cell, surrounded by a dense
mass of protoplasm. The nuclear membrane is hardly visible
or has disappeared entirely (‘Text-fig. 2, c, 1 and 2). This mass
moves to one side and applies itself to the cell-wall as a densely
staining mass (Text-fig. 2, c, 83 and 4). The nucleus is at first
visible as a darker body, but eventually the whole mass stais
so deeply that the constituent parts are indistinguishable
(Text-fig. 2, c, 5). Groups of these degenerate cells are found
in the ectoderm after the egg has separated and are possibly
used up by a subsequent egg.

When the degenerative process takes place within the cyto-

OBSERVATIONS ON HYDRA 603

plasm of the egg, it may go on much as described above
(Text-fig. 2, b, 1), or the cytoplasm may, to all appearance, be
absorbed directly into that of the egg and the nucleus alone
undergo visible change. Wager (18) describes such degenerative
products formed from whole cells, from nuclei, and from
nucleoli. I have not seen any of the last named in process of
formation, but from the small size of some of the masses it
seems probable that this is the case im my specimens also.
In several cases a nucleolus seemed to be breaking through the
nuclear membrane of an interstitial cell.

These degenerative cells are looked on as stores of energy
to be used up by the embryo during development. Tann-
reuther (11) states that they divide by amitosis after their
absorption into the egg-cell, but I have seen no signs of this.
It has been stated that they are all used up before the young
Hydra hatches, but this is not the case, for large numbers
are present in the tissues of newly hatched Hydras and they
do not entirely vanish till some time after hatching.

In my specimens the egg-shell is formed by a kind of vacuole
formation at the surface of the egg. The edges of the vacuoles
come in contact with one another, and their limiting membranes
then harden to form the spicules. This is in accordance with
the observations of Brauer (2) and Kleinenberg (7). The eggs
then drop off and fall to the bottom.

NEMATOCYSTS.

The nematocysts have been much studied, and of late
have been used largely to distinguish the different species of
Hydra. There are four main types of nematocyst, large
pear-shaped, small pear-shaped, and two types of cylindrical
nematocysts, which differ from each other in size and in the
way the thread is coiled inside. As the two last are not always
strictly cylindrical, and as the names in other respects are not
suitable, Schulze (9) has proposed to name the different types
from their functions as Penetrant, Volvent, and Glutinant.
Otto van Toppe (12) was the first to study the functions of all
three types in detail.

604 SHEINA MARSHALL

The large pear-shaped nematocyst is, as the name penetrant
implies, used for piercing the chitinous exoskeleton of the
prey. When the hard shell has been pierced, the thread is
everted into the soft tissues beneath. Usually only that part
of the nematocyst immediately above the large barbs succeeds
in penetrating, but as the thread itself has a spiral coil of small
hairs or barbs upon it, it is firmly held in place. The large
barbs, which are so conspicuous a feature of the exploded

TEXT-FIG. 3.

NEN

B

Discharged nematocyst.

nematocyst, are never found imbedded in the tissues of the
prey. Iwanzoff (6) says that the three, just at the moment of
their eversion, form a stiletto which pierces the cuticle of the
prey, making a hole by which the thread can enter. With the
further eversion of the thread the barbs swing outward to their
final position. Some chemical action is exerted on the chitin,
as can be seen from a study of sections. Immediately around
the point of entry is a deeply staining area, irregular in form
(Text-fig. 8, 1). This probably corresponds to an outpouring
of the poisonous fluid contained in the nematocyst, for the
thread stains in the same way. Outside this dark patch is

OBSERVATIONS ON HYDRA 605

a bowl-shaped space which stains less deeply than the normal
chitin (Text-fig. 3, 2). The thread can be seen lying in the soft
tissues, the first part at right angles to the external surface.
The distal part usually curves to one side. It is filled with
a darkly staining mass which is extruded either at the end or
along its course (Text-fig. 3, 3). In preparations of unfixed,
exploded nematocysts stained in methylene blue, the fluid
inside the penetrants stains deeply and can be seen partly
inside the capsule and partly in the form of tiny drops on the
outside of the thread. The thread has, apparently, rows of
minute pores or permeable areas through which the fluid can
escape, as well as by the opening at the end.

The threads of the small pear-shaped nematocysts, or vol-
vents, when exploded, wind tightly round any protruding hairs
or bristles on the prey, and so hold it captive till it has been
kailled by the poison of the penetrants. The thread is coiled
inside the capsule in two loops which lie on one another so that
they appear in section like one thick rmg. There is a row of
small hairs on the thread arranged in a very open spiral. When
exploded, the thread coils up tightly, and in optical section
there is seen to be a narrow space along the axis of the coils
which is closely beset with hairs. This forms an efficient
mechanism for grasping the bristle, and: there is some evidence
that the secretion in the capsule and thread, which stains deeply
with methylene blue, is sticky. Van Toppe (12) states that the
stimulus for the explosion of this type of nematocyst is different
from that exploding the penetrants, as the latter explode when
their enidocils come in contact with flat surfaces, while the
former do not.

The third and fourth types of nematocyst, the glutinants,
are usually cylindrical. One type is usually larger than the
other. In the former the thread shows four or five turns
almost at right angles to the long axis of the capsule and below
this is wound irregularly. In the latter the thread is wound
in an irregular figure of eight. Schulze (9) therefore calls
them streptoline and stereoline respectively. In some Hydras
(e.g. H. attenuata, van Toppe, H. circumcinecta,

606 SHEINA MARSHALL

Schulze, H. stellata, Schulze, and Pelmatohydra
braueri, Schulze) the streptoline is not cylindrical but pear-
shaped. When Hydra sticks on to glass or to any other
surface by means of its tentacles or hypostome, it uses these
nematocysts. If one of the adherent tentacles 1s examined,
there are seen numerous exploded glutinants whose threads
are firmly attached to the glass. They are so firmly fixed that
if any pull is exerted on them the cell protoplasm of the
Hydra is drawn out into a thread with the nematocyst at
its tip. Zygoff (16) was the first to notice these, and looked on
them as pseudopodia by which the animal moved. Toppe
discovered their true nature. These processes can withstand
a considerable stram. I have seen a Hydra apparently
trying to free a tentacle which was held at the tip by one
of these nematocysts only. The tentacle was given several
tugs, was twirled round rapidly, first m one direction and then
in the other, the animal contracted tightly once or twice, and
finally the tentacle was torn away. It was strikmg to watch
an animal like the Hydra exhibiting such apparently pur-
poseful movements. The twirlmg movements are much more
complex than any the Hydra usually shows, and must have
called into play a different mechanism. The nematocyst is
always left sticking to the substratum while the protoplasmic
process is gradually withdrawn into the cell. A similar process
is sometimes drawn out when a tentacle is pulled away from
some bristle on which a volvent has wound itself.

In unfixed, exploded glutinants stamed with methylene
blue, numerous droplets can be seen on the outside of the
thread, as in the case of the penetrants. This secretion is
probably sticky, and possibly hardens in contact with water.
When used, it is extruded not only by the pore at the end of
the thread but also by the side pores, for the thread can often
be seen adhering at a point about half-way down its course.

It seems probable that the secretion of the other types of
nematocyst has also to some degree the property of sticking
firmly. I have observed tentacles adhering both by the
penetrants and by the volvents, though not so firmly.

OBSERVATIONS ON HYDRA 607

The size of the nematocysts has been cited as a charac-
teristic difference for the various species. Most authors,
however, give relative sizes only, and all measurements are
given for the penetrants alone, which, as will be seen, are
much the most variable in size of all the types. Steche (10)
states that the penetrants of H. fusca are at most 8-8-5y,
and those of H. grisea at least 10-5 and usually 13-13-5p.
Toppe (12) says that those of H. fusca are the smallest,
H. grisea next, and those of H. attenuata the largest.
Schulze (9) gives 25y for his H. attenuata (which is not
the same species as Toppe’s) and 13 for Pelmatohydra
oligactis.

In my Hydra the size of the penetrant varies greatly.
Some forms measure 10 or 11 and others may be as large
as 224. Hven in one individual the sizes may vary by as much
as 8u. The small penetrants are found in the tentacles as well
as in the body and are not merely incompletely developed.
It is difficult to speak with certainty, but the most frequent
size seems to be about 154; 12-13, and 18-19, are com-
moner than the imtermediate sizes. In small Hydras the
smaller size seems to be more frequent than the large.

The glutinants vary much less, the ranges of individual
variation not bemg more than 3 or 4u. The larger type
(streptoline) measures about 11-5 (maximum 13 and minimum
10), and the smaller (stereoline) about 9 (7-11 1).

The volvents measure from 5-10. The usual size is about
8p and the smaller sizes are more numerous than the larger.

In order to see whether the size of the nematocysts varied
from time to time with changes in the size of the individual,
I took a large well-grown Hydra, and after removing two of
the tentacles in order to measure the nematocysts, starved it
for seventy-three days. The measurement of the width of the
foot when the animal was fully extended was, before starvation,
about 0-275 mm., and, after, about 0-100 mm. It had therefore
decreased to almost a third of its original width, and this gives
a rough indication of the general effect. Whereas, before
starvation, the nematocysts were of normal size (penetrants

608 SHEINA MARSHALL

19), the penetrants now measured only 11-15y. This was due,
in the first place, to the disappearance of the large type. The
other nematocysts, however, also showed a similar though
slighter decrease in size. The same experiment with another
Hydra (starved forty-seven days) yielded similar results.
If the size of the nematocysts varies with the size or condition
of the individual it is obvious that this cannot be used ag
a trustworthy characteristic in differentiating between the
varlous species.

In Hydra viridis the penetrants measure 8-l0u; the
streptoline glutmants are kidney-shaped and are as large
as the penetrants, bemg 10-lly. The stereoline glutinants
are much smaller and rarer, being only 6-7, and the volvents
are about 5p.

To obtain the nematocysts freely the Hydra was at first
macerated in weak chloroform water. Latterly I used Schulze’s
phenol-glycerme mixture (phenol, crystallized, 1 grm., glycerine
200 ¢.c., distilled water 200 ¢.c.), as giving the same results
and being more convenient to handle.

NERVOUS SYSTEM.

The distribution of the nervous elements is such as one
would expect from the reactions of the living animal. They
are most numerous in the foot and in the head and tentacles,
and are much scarcer in the middle part of the body.

The nervous cells are best seen in maceration preparations.
I have found the most useful method to be maceration in
Hertwig-Schneider’s mixture (0-02 per cent. osmic acid,
1 part: 5 per cent. acetic acid, 4 parts) for fifteen to twenty
minutes and subsequent staining in a strong filtered solution
of methylene blue for three-quarters of an hour or longer.
The methylene blue is dissolved until a deep blue solution 1s
obtained. The specimen is then well macerated and several
gentle taps on the coverslip are sufficient to separate the
cells. By dividing the Hydra before maceration has begun
a fair idea of the relative distribution of the nervous cells may
be obtained.

OBSERVATIONS ON HYDRA 609

The maceration method has been largely used by Hadzi (4)
in his work on the nervous system of Hydra. I cannot,
however, confirm all his findings.

The nervous cells may be roughly divided into two types, (1)
the ganglion cells and (2) cells which are more or less inter-
mediate in form between the ganglion cell and the epithelial cell.

The former are of various shapes according to the number
of processes they possess (Text-fig. 4, a, c, d, e). The nucleus
stains much less deeply with methylene blue than that of the
ordinary interstitial cell, as does also the nucleolus. There
may be two of the latter. The cell-body usually consists of
a thin layer of protoplasm surrounding the nucleus, but it is
frequently prolonged at either end into a short thick process
before giving off the long nerve-threads. These latter are
long thin processes, often branching. Where they branch
there is usually a slight swelling, and these swellings are seen
along the course of unbranched processes as well (Text-fig. 4,
aand d). The processes of one cell can often be seen to unite
with those of another, so that the cell-body appears to lie in
a network of interlacing threads. Many of the processes
apparently end freely, sometimes in a small knob, while others
are attached to the muscle-fibres of myoepithelial cells from
which they cannot be detached by tapping on the cover-glass
or by irrigation under it. The ganglion cells of the body are
usually simpler and less branched than those of the head
or foot.

The second type of nervous cell is long and narrow in shape,
with the nucleus about the middle of the cell-body, and has
a nervous process only at one end (Text-fig. 4, b and f). This
may branch and sometimes it comes into connexion with
another nervous cell. The other end is flattened (Text-fig. 4, b)
or knob-like. These cells correspond to Hadzi’s * sensory
cells ’ or ‘ sensory nerve-cells ’, but I have, as a rule, not been
able to find a short projecting hair on the flattened end.
In one case a knob-shaped end bore two fine hairs each ending
in a little swelling, and in several other cases there were one
or two short hairs (Text-fig. 4, f). On examining the surface

NO. 268 Ss

610

SHEINA MARSHALL

TEXT-FIG. 4,

b.

Nerve cells.

ries 1y4

OBSERVATIONS ON HYDRA 611

of a deeply stained Hydra with an oil-immersion lens,
I have not been able to find any number of projecting hairs
(apart from the cnidocils) such as one would expect if sensory
cells bearing hairs played a large part in the stimulation of the
Hydra.

I have also tried Hadzi’s vital methylene blue method (4)
but without success. Hadzi states that his method gives good
results only with H. viridis and in sunshine. In my
specimens the nematocysts took up the stain strongly and the
ectoderm faintly, but there was never any differential staining
of the nervous elements.

It is considerably more difficult to recognize nervous cells
in section, since the processes are cut short. Hadzi figures,
and describes as nervous, cells which stain more deeply than
the other interstitial cells and lie basi-epithelially, sending
processes to the surface and in other directions. I have seen
such appearances in section, but find it impossible to say
whether these apparent threads are not merely strands of
protoplasm belonging to the myoepithelial cells, or cut edges
of cells. I have not found the sensory apparatus which he
describes at the surface.

The nerve-cells probably originate from interstitial cells.
Some of the latter may often be found connected to one
another by short strands, and interstitial cells with short
or long processes are not uncommon especially in the tissues
of newly hatched Hydras. These differ little from the nervous
cells except in their nuclei and in the larger amount of cell
protoplasm which they possess.

The effect of various nerve poisons was tried.

Chloroform.—A weak solution of chloroform in water
(which anaesthetized a Daphnia completely in a few
minutes) caused a curious rhythmic contraction. The animal
contracted down into a tight spiral quickly, and then slowly
straightened out again. This was repeated at intervals which
gradually increased from about two minutes up to fifteen or
twenty minutes. Eventually it became motionless and insensi-
tive to contact. This occurs in two hours or longer, after which

S82

612 SHEINA MARSHALL

the Hydra can recover if removed to pure water. The
vhloroform has, however, a macerating effect, which begins to
act at the tentacles.

Chloretone.—the effect of this was to make the Hydra
throw out masses of endodermal cells by the mouth. It
usually died.

Cholin.—Weak solutions (1: 1,000) of this were used but
were ineffective. Stronger solutions led to a half-contracted
state, but the animal soon recovered in fresh water.

Curare.—A weak solution of curare was prepared by
erinding up 0-1 grm. curare with 20c¢.c. water and filtering.
A clear yellowish solution was obtamed. A Hydra was put
in a small dish and this solution added till it was about half
strength. For some hours the Hydra remained quite
normal. After that a stimulus on any part of the body was
responded to by a general contraction, and the tentacles
remained half contracted. It was left in the solution over
night and in the morning was found to have eatena Daphnia
which had been present; but it was very much contracted
and did not respond at all to stimulation. Removed to fresh
water it expanded but remained very insensitive. On examina-
tion the tentacles were found to be degenerating from their
tips downwards.

The experiment was repeated with another Hydra and
the same result obtained much more quickly, for in two hours
the animal ceased to react to stimulation and its tentacles
began to degenerate.

It is noteworthy that necrosis always begins at the tentacle
tips and works down gradually, the body remaining apparently
normal for some time after the tips of the tentacles have
disappeared entirely.

In one ease the whole head suffered necrosis and the animal
remained as a closed tube for some days. It then produced
two buds about the middle of the body and a third near the
head region but slightly to one side. These developed and
constricted off and a fourth bud appeared, again to one side
of the head. As it developed it swung round so that eventually

OBSERVATIONS ON HYDRA 613

it was in line with the mam axis of the body, and acted as
the true head. The curare thus appears to destroy the power
of regeneration in the affected parts: possibly the interstitial
cells are killed off.

SYMBIOTIC CELLS.

Since the differences between the various species of brown
Hydra seem less marked than thei resemblances, it is
of interest to know whether, under any circumstances, a
H. viridis deprived of its green cells would grow to resemble
a brown Hydra.

Goetsch (8) obtained brown Hydras showing pathological
features, which, when fed with algae, turned green. As they
did so they diminished in size, budding ceased, and under
natural conditions they died. Some which were fed with
freshly killed Daphnia lived, and produced testes or ovaries.
The symbiosis was easily lost, disappearmg after four weeks
in darkness. Goetsch suggests that this Hydra is a new
mutant, capable of receiving the alga, which is a large form
of Chlorella.

Whitney (14) describes a method of ridding H. viridis
of its green inhabitants. He kept his specimens in a weak
solution of glycerine (0-5 per cent.) for a few weeks. During
this period the endodermal cells swelled up and extruded the
algae, which were then thrown out by the mouth. Eventually
he obtained several colourless Hydra which lived normally
in the aquarium for some time without being reinfected.
They retained all the features typical of H. viridis except
the colour.

On January 26 I set five Hydras in a jar of 0-5 per cent.
glycerine, where they were fed as usual. They budded actively.
On February 13 they had increased to nine and were removed
to 0-75 per cent. glycerine as the’ weak solution had had no
effect. On March 1 there were twenty-three Hydras, still
quite green, and they were removed to 1 per cent. glycerme.
One was fixed and sectioned. The endoderm appeared to be
quite normal in size, and the green algae were arranged as

614 SHEINA MARSHALL

usual all through the cell and were not collected at the distal
end. On March 9 thirty-one Hydras were removed to 1-5 per
cent. glycerine. In the middle of April they were still quite
ereen.

Several Hydras were kept in the dark for the same purpose,
but although the colour grew somewhat paler they all died
before the green cells had been entirely lost. Hadzi (5) has
also noted that they do not survive in darkness. LHggs,
apparently colourless, which were produced in the dark, died
before hatching.

Some brown Hydras were induced to swallow pieces of
H. viridis by slipping the latter mside the carapace of
Daphnia, but they were ejected along with the remains of
the food and had no effect.

Daphnia were also fed on a pure culture of Chloro-
sphaera and were then given to the Hydras, but with no
effect.

There have been many attempts to make a pure culture of
the green organism inhabiting H. viridis. In Beyerinek’s (1)
paper on the culture of algae and lichens he states that he has
been unable to obtain a pure culture of the zoochlorella from
Hydra, but he adds a foot-note to the effect that he had
obtained such a culture, and that the organism was indistin-
guishable from Chlorella vulgaris.

Later writers have made damp cell-cultures and have seen
division taking place, but so far as I know there has been no
large culture obtained.

I washed H. viridis in several changes of sterile water
and then teased it up with needles till practically all the green
cells were freed. They were then sown on Miguel solution in
tubes and sporulation dishes, on Amoeba-agar, and on agar
made up with Miguel solution, but in none of these was any
culture obtained.

In one tube of Miguel, Chlorosphaera limicola
(Beyerinck) appeared. The Daphnia on which the Hydras
were fed were themselves fed on a mixed green culture which
proved to contain Chlorosphaera, and the organism may

OBSERVATIONS ON HYDRA 615

have remained alive inside the Hydra after the Daphnia
had been eaten.

In damp chambers and on sterile slides the green cells
remained alive for a week or more and some divided, but
eventually they all died off or bacteria appeared. The cells
divided either into two or three. Radais states that in a
culture of Chlorella vulgaris the cells divided into four
when healthy, but as the culture grew older the rate of division
slowed down, and the cells divided into three or two.

When stained, the organism from H. viridis shows a large
and distinct pyrenoid. The nucleus is less distinct and usually
appears as an irregular ring of darkly staining material. No
division stages were seen.

TAXONOMY.

The number of species of Hydra has been much discussed
ever since the foundation of the genus. Schulze (9) has
lately divided it into three genera and about ten species.
The Hydras on which I worked do not exactly correspond to any
of Schulze’s species but come nearest to his H. attenuata,
from which they differ in bemg hermaphrodite. It seems
to me improbable that the genus Hydra is justifiably
divided up into so many definite species. Some of Schulze’s
species are founded on the examination of preserved specimens
only. The general habit, colour, size, and so on, are used as
differentiating characters, while the nematocysts are always
treated as important diagnostically. The Hydras on which
I worked varied considerably in size and habit, but all possessed
the same kind of nematocysts. In some the egg was stuck on
the side of the glass and in some it fell freely to the bottom.
Considering the great variation in appearance which may take
place within the lifetime of one individual, it seems unsafe to
separate off as distinct species animals whose whole life history
has not been completely followed through.

This work was done during my tenure of a Carnegie scholar-
ship from 1920-2, in the Natural History Department of

— 616 SHEINA MARSHALL

Glasgow University. I should like to express my gratitude
to Professor Graham Kerr for the help he gave, and the interest
he took in my work.

REFERENCES TO LITERATURE.

. Beyerinck (1890).—** Kulturversuche mit Zoochlorelia, Lichengonidien

und anderen Algen ”, ‘ Botan. Zeit.’, Jahrg. xviii.
Brauer (1891).—‘* Uber die Entwicklung von Hydra”, ‘ Zeit. f. w.
Zool,’, li.

. Goetsch (1922).—‘ Die Naturwissenschaft.’
. Hadzi (1909).—‘‘ Uber das Nervensystem von Hydra”, ‘ Arbeiten

a. d. Zool. Instit. Wien’.

. —— (1906).—‘‘ Vorversuche zur Biologie von Hydra”, ‘ Archiv fiir

Entwicklungs-Mechanik’, Bd. xxii.

. Iwanzoff (1896).—‘ Anat. Anzeiger’, Bd. xi.

Kleinenberg (1872).—‘ Hydra: anatomisch-entwicklungsgeschicht-
liche Untersuchung.’ Leipzig.

. Lieber (1909).—‘ Zool. Anz.’, Bd. xxxiv.
. Schulze (1917).—‘‘ Neue Beitriige zu einer Monographie der Gattung

Hydra ”’, ‘ Archiv f. Biontologie’, Bd. iv.

. Steche (1911).—‘ Hydra u. die Hydroiden.’ Leipzig.
. Tannreuther (1908).—‘‘ The Development of Hydra’”’, ‘ Biol. Bull.’,

vol. xiv.

. Toppe (1910).—* Bau und Funktion d. Nesselkapseln”’, ‘ Zool. Jahrb.,

Abt. £. Anat.?, Bd. xxix.

. Wager (1919).—‘‘ Oogenesis and early development of Hydra”,

‘ Biol. Bull.’, vol. xviii.

. Whitney (1907).—*‘ Artificial removal of the green bodies of Hydra

viridis ’’, ibid., vol. xiii.
Wintrebert, P. (1912).—‘*‘ Le mécanisme de léclosion chez la truite
arc-en-ciel”’, ‘C. R. Soc. de Biologie’, Ixxii, p. 724.

. Zygoff (1898).—‘‘ Bewegungder Hydra fusca”’, ‘ Biol. Zentralblatt’,

Bd. xviii.

Head Length Dimorphism of Mammalian
Spermatozoa.
By

A. S. Parkes, B.A. (Cantab,), Ph. D.,
Department of Zoology, University of Manchester.

With 3 Text-figures.

CoNTENTS.

PAGE
1. IyTRODUCTORY : : : : 3 : ; . 4 (ollizs
2. MetHops AND MATERIAL : ' , 5 : : . 618
3. SPERMATOZOA OF Man, Rat, Cat, AND MOUSE . : , » 62h
4. CONCLUSION . : ; : : : : : ; . 623
5. SUMMARY . : : : : : 2 : : O24:
6. BIBLIOGRAPHY ; : e ; : F 3 ; 2625

INTRODUCTORY.

Work on mammalian spermatogenesis has in a large nuniber
of cases shown that the spermatozoa are of two types, one
type possessing the accessory chromosome, whilst the other
type has no sex chromosome or a mere vestigal complement.
As the spermatozoon head is constituted almost entirely of
nucleus it might be expected that the additional chromatin
possessed by the one type would slightly merease the size
of the head. In three cases this correlation has been found.
Wodsedalek (5 and 6) has shown that in the horse and bull
the spermatozoa are of two types, and that in each case a
frequency polygon of the head lengths shows distinct
dimorphism. In the case of the dog, Malone (2) found an
unpaired accessory chromosome in spermatogenesis, and

618 A. S. PARKES

Zeleny and Faust (8) have demonstrated dimorphism of the
head lengths. The work recorded in this paper was an attempt
to extend the application of this correlation to other mamma-
lian spermatozoa for which chromosomal dimorphism has been
shown.

MrtuHops AND MATERIAL.

The new work recorded here deals with man, the rat, the
cat, and the mouse. I have to thank Mrs. R. Sellars of the
Manchester Medical School for procuring the human material
for me from the Manchester Royal Infirmary. In the other
cases the material was obtained by dissection of the epididimis.
In each case smear slides were made, as this method has advan-
tages compared with using testis sections. Some difficulty
was at first experienced in making satisfactory smears owing
to the tendency for the spermatozoa to drop off. Increased
experience in manipulation, however, was found to surmount
this. By teasing out the epididimis in salt solution and fixing,
it was found possible to make the spermatozoa adhere without
using egg-albumen cement. For fixmg Zenker’s fluid was
used to start with, as recommended by Zeleny and Faust (8),
but finally the ordinary corrosive and aceti-solution (90 per
cent. saturated solution corrosive sublimate and 10 per cent.
glacial acetic acid) was found to be quite efficient. Various
stains were tried, but Delafield’s haematoxylin was eventually
found to be by far the most satisfactory.

The measurement of the spermatozoa was found to present
sreat difficulty, especially in the case of the rat and mouse
where the head-piece is sickle-shaped. This fact, together with
the minute size of mammalian spermatozoa, makes measure-
ment with an ocular micrometer almost impossible. Both
these difficulties were alleviated by using the Zeiss-Greil
drawing apparatus possessed by the department. This consists
of a lantern throwing light through a horizontal photo-micro-
scope and projecting the image on to a screen. By placing
a mirror at 45 degrees in front of the eye-piece the image can
be thrown down on to a table. This apparatus can be used

DIMORPHISM OF SPERMATOZOA 619

with an oil immersion, and the resulting image thrown on the
table, even of such a small object as a spermatozoon head, is
sufficiently large to admit of measuring round a curve with
a pair of compasses. The co-efficient was worked out previously
by putting a stage micrometer in the microscope and finding
out how many centimetres on the table corresponded to 10
on the slide. The subsequent calculation was as easy as that
necessary when using an ocular micrometer. By this means
the unavoidable margim of error in the measurements was
very greatly reduced. On one occasion when the drawing
apparatus was out of order some measurements were made
with the aid of a camera lucida. In both cases extreme
dimensions were marked on paper, connected by a line, and the
whole number measured afterwards. An 18 Zeiss ocular could
be used with the camera lucida, but not with the drawing
lantern, owing to the excessive diffusion of light, and so the
nett magnification came to about the same in both cases.

Wodsedalek measured the spermatozoa in testis sections
by the camera-lucida method, and Zeleny and Faust used
smear slides with the ocular micrometer. It will be seen,
therefore, that the method described above is a combination
of the two, and it was found to be the most satisfactory. The
error involved by the use of the ocular micrometer is avoided,
as is the possibility of getting immature spermatozoa on the
slide.

As the whole value of the work depends on the degree of
accuracy which can be achieved, the following remarks may
not be out of place. Three sources of error arise :

(a) Distortion of the spermatozoa during fixing and preparing.
(b) The personal factor in measuring.
(c) The unavoidable error in measuring.

With regard to the first point, general shrinkage of the
spermatozoon head must almost inevitably occur ; but only one
slide was used for each set of measurements, and no attempt
has been made to mix measurements from different slides.
As the spermatozoa on one slide would all be affected in the

620 A. S. PARKES

same manner, this precaution should remove the first source
of error.

Secondly, bias particularly easily arises in such work and
may almost unconsciously detract from the accuracy. A con-
scious attempt to discount this bias may lead to the opposite
extreme and cause an equal inaccuracy. Also, the work is
very trying and strain seriously disturbs the measurements.
In general, however, the personal factor was discounted, as far
as possible, by only working for very short periods at a time.

Trxt-Fria. 1.
co

0
30

20

k yo) 5 5S ‘6 66 HT] “15 8 eS 9
Frequency polygon of head lengths of human spermatozoa.

Thirdly, the unavoidable margin of error must be considered.
Fortunately, however, this is a calculable quantity, and the
following tests were made. In the first case the same sperma-
tozoon head was measured under three different magnifications,
the co-efficients of which were known, and the results com-
pared. The following table sums up the results :

Ob. and Oc. Size in p.
Bausch and Lomb ?” and Zeiss 18 : 4-09
Koristka ;'>” and Bausch and Lomb 10 4-10

Koristka 7:” and Zeiss 18 . E i 4:14

DIMORPHISM OF SPERMATOZOA 621

It will be seen that the range of the variation in the three
measurements is only 0-05 or 35h.

The second manner of testing accuracy consisting in measur-
ing the same spermatozoon several times under the same
magnification, and the results gave a range of variation in six
measurements of 0-104 or ;5. Both these margins of error
are very much smaller than the least fraction of » usually
dealt with.

I should like to take this opportunity of acknowledgmg my
great obligation to Mr. J. 'T. Wadsworth for his invaluable
assistance in the technique of this work.

SPERMATOZOA OF Man, Rat, Cat, anp Mousz.

Von Winiwarter (4) has described chromosome dimorphism
of human spermatozoa according to the presence or absence
of an unpaired accessory. In my material, measurement of
the head lengths was found to give the following results :

TaBLE I. FREQUENCY OF HEAD LENGTHS OF SPERMATOZOA

oF Man.
Head lengths Number Head lengths Number
x 1,460 (in cms.) found. «1,460 (in cms.) found.

0-4 1 0-7 52
0-45 2 0:75 31
0°5 2 0:8 Pal
0:55 20 0-85 7
0-6 49 0-9 6
0-65 28

In the case of the rat Allen (1) has demonstrated an accessory
chromosome in spermatogenesis, the spermatozoa having
eighteen or nineteen chromosomes. The dimorphism again
appears to communicate itself to the head sizes of the sperma-
tozoa, for dimorphism was found in the head lengths of the
spermatozoa of the rat. The frequency polygon given below
(Text-fig. 2) was made from measurements of the head lengths
( < 4,000) of 155 spermatozoa.

622 A. 8S. PARKES

TaBLE II. FreQquENCY oF Heap LENGTHS OF SPERMATOZOA

OF Rat,
Head lengths Number Head lengths Number
«4,000 (in cms.) found. «4,000 (in ems.) found.
3:2 I 3-7 10
3:25 2 3:75 8
3:3 3 3-8 17
3°35 5 3°85 19
3-4 a 3°9 df
3:45 6 3-95 6
3°5 7 4-0 4
3:55 13 4-05 3
3:6 15 4+] 3
3°65 14 4:15 1

TEXT-F G. 2.

© & FH

9-2 3 35 445-5 55-6 OS 7 45 6 85 4 540-05 1 15
Frequency po'ygon of head lengths of spermatozoa of rat.

Von Winiwarter (8) has shown that the spermatozoa of the
cat are cytologically dimorphic, one type having eighteen
chromosomes and the other seventeen. In measuring the
head lengths, however, no clear dimorphism was found. The
results were as follows :

DIMORPHISM OF SPERMATOZOA 623

TaBLeE III. Frequency or Heap LENGTHS OF SPERMATOZOA

oF Cat,
Head lengths Number Head lengths Number
«4,000 (in cms.) found. «4,000 (in cms.) found.
1:2 2 1:7 32
1-25 2 1:75 20
1:3 4 1:8 28
1:35 2 1-85 13
1-4 7 1-9 22
1:45 9 1:95 6
1-5 17 2-0 5
1:55 20 2:05 1
1-6 31 2-1 1
1-65 26 2-15 il

Yocum (7) has shown that the spermatozoa of the mouse are
cytologically dimorphic, one type having nineteen and the
other twenty chromosomes. I found this reflected in the
head lengths, as the following results show :

TaBLeE IV. FREQUENCY oF HEAD LENGTHS OF SPERMATOZOA

oF Mouse.
Head lengths Number l Head lengths Number
x 4,000 (tn cms.) found. «4,000 (in cms.) found.

2-15 1 2-65 25
2-2 il 2-7 2

2-25 2 2-75 12
2:3 6 2-8 il
2-35 6 2-85 9
2-4 9 2-9 6
2-45 13 2-95 5
2-5 25 3-0 5
2-55 4 3:05 I
2-6 10 3:1 1

CoNCLUSION.

It would thus appear that in three more species of mammals
the spermatozoa show dimorphism in the head length, while
in a fourth species, the cat, the evidence is uncertain. The
chief interest in these conclusions lies in the bearing which the
size dimorphism of the spermatozoa might have in determining
the proportion of the sexes at conception. If the potentially
male and the potentially female-producing spermatozoa are
of different size, their activity and vigour may also be relatively

624 A. S. PARKES

different, causing more of one type than of the other to survive
the severe journey through the female organs to the ova
(as ‘I’. H. Morgan has suggested, ‘ Physical Basis of Heredity ’,
1919). It is most probable that the disproportion which exists
between the sexes at conception in most mammals is con-
nected with this point.

TExtT-Fic. 3.

20

2:15 2°35 255 2-15 2-95
Frequency polygon of head lengths of spermatozoa of mouse.

SUMMARY.

1. Chromosome dimorphism of the spermatozoa has been
shown for a variety of mammals, and in some cases this has
been shown to be correlated with dimorphism in the head
lengths of the spermatozoa.

2. In the present paper this correlation has been extended to
the spermatozoa of man, the mouse, and the rat, m which
chromosome dimorphism of the spermatozoa had previously
been shown, and in which head length dimorphism seems to
exist.

3. The interest of these results lies in the probability that
the histological difference in the X- and Y-spermatozoa may
account for the inequality of the sexes at conception in
mammals.

DIMORPHISM OF SPERMATOZOA 625

BIBLIOGRAPHY.

1. Allen, E.—‘“‘ Studies in cell division in the Albino rat. III. Spermato-
genesis: the origin of the first spermatocytes and the organization
of the chromosomes including the accessory’, ‘Journ. Morph.’,
vol. xxxi, 1918.

2. Malone, J. Y.—‘‘ Spermatogenesis in the Dog”, ‘Trans. Am. Micr.
Soc.’, vol. xxxvii, 1918, 97-110.

3. Von Winiwarter.—‘ Arch. Biol.’, vol. xxiv, 1909.

4. “Etudes sur la spermatogenése humaine. II. Hétérochromo-
somes et mitoses de l’épithélium séminal ”’, ‘ Arch. de Biol.’, vol. xxvii,
1912.

5. Wodsedalek, J. E.—‘‘Spermatogenesis in Horse”’, ‘ Biol. Bull.’,
vol, xxvii, 1914.

“ Studies on the cells of Cattle with special reference to spermato-
genesis, oogonia, and sex determination’, ibid., vol. xxxviii, 1920,
290-316.

7. Yocum, H. B.—‘‘ Some phases of spermatogenesis in the Mouse”’,
‘ Berkeley Uni. Cal. Pub. Zool.’, vol. xvi, 1917.

8. Zeleny, C., and Faust, E. C.—‘‘ Size Dimorphism in Spermatozoa ”’,
‘Journ. Exp. Zool.’, vol. xviii, 1915.

NO. 268 Tt

—
ri
by
bs
*
SF
e
Pe
=
ba
"
—
a
ey
=
iA
ik

ead | a
LEAN OSs a
ae ' et |
rl 6 op? lep ipes tee Nk

ee | ics

waved 4 se pista bt. Geb it cqeule
: a
'

a
—
‘

' i.
’ y a oy

On Brinkmann’s System of the Nemertea
Enopla and Siboganemertes Weberi, n.g.n.sp.

By
Dr. Gerarda Stiasny-Wijnhoff, Leiden.

With 26 Text-figures.

Aut text-books of zoology give the Nemertean system
as Birger published it several times in his monographs
on this subject (6, 7, 8). The old systems of Schultze,
MelIntosh, Hubrecht have been forsaken, and our text-
books do not divide the Nemertini any more in Anopla
and Enopla, or Palaeonemertea, Hoplonemertea, and Schizo-
nemertea. The armature of the proboscis and the arrangement
of the cephalic slits are believed to be of secondary importance,
and Birger divided the Nemertini into four orders, three otf
which, his Protonemertini, Mesonemertini, and Metanemertini
are supposed to be closely related by having a body-wall that
consists of an epithelium, a basement membrane, and two mus-
cular layers. Benham,in Ray Lankester’s ‘Treatise on
Zoology ’, unites them as Dimyaria, and his Trimyaria consist
of one order only, the Heteronemertini Birger. The Proto-
nemertin’ are McIntosh’s family Carinellidae, Meso-
nemertini are his family Cephalotricidae and the genus
Carmoma (Hubrecht); Metanemertini is a new name for
Hoplonemertini (Hubrecht) or Enopla (Max Schultze), and
Heteronemertini are Hubrecht’s Schizonemertini and his
families Eupoliaidae and Valenciniaidae. Everybody agrees
that the last two families are more nearly related to Lineids
and Cerebratulids than to the Protonemerteans, and their
enclosure in the order Heteronemertini seems to be well

1 ti?

628 GERARDA STIASNY-WIJNHOFF

founded. The following schema gives the synonyms in the
different systems.

Max Blaxland
Schultze. McIntosh. Hubrecht. Biirger. Benham.
Enopla fam. amok Hoplonemertini. Meta-

poridae nemertini
fam. Cephalo- | fam. Cephalotri-
tricidae cidae Meso- Dimyaria.
gen. Carinoma nemertini
fam. Carinel- Palaeo- roto-
Anopla ~ lidae ~ nemertini nemertini |
| | fam. Carinel- emo Wupoliaidae
lidae fam. Valencini- Hetero- Trimyaria.
e nemertini |
fam. Lineidae Schizonemertini

The combination of Biirger’s three other orders in the
Dimyaria was not well founded. Biirger meant them to be
two stages in the development of the third, the Metanemerteans
or Hoplonemertini; the Protonemertini with their epithelial
nervous system (Text-fig. 1, a) being the most primitive ones.
Carinoma and Cephalothrix, the Mesonemertini, give the
development that leads to the central position of this system
in the body parenchyma (Text-fig. 1, c). Now Bergendal,
in his treatise on Carinoma armandi (2), made it clear
that this genus is closely related to Carinella and the Hetero-
nemerteans (Text-fig. 1, b), and has no affinity at all to Cephalo-
thrix nor to the armed Nemerteans. In another Swedish
article he shows (1) that Biirger’s order Mesonemertini is
quite unnatural, and that both genera Cephalothrix and
Carinoma are true Protonemertini; so Hubrecht’s order
Palaeonemertini ought to be restored. A study on Cephalo-
thrix taught me (16) that this genus is not nearer related to
the armed Nemerteans than to the unarmed, and so we get
instead of Proto- and Meso-nemertini the old order of Palaeo-
nemertini, nearly related to the Heteronemertini and without
any special relation to these groups the armed Nemerteans.
In 1912 (17) I proposed to return to the old division of Max
Schultze. The class Nemertini is subdivided into two
sub-classes: Anopla and Enopla. Hach sub-class contains
two orders: the Anopla with Palaeonemertini (Hubrecht) and
Heteronemertini (Birger), of which diagnoses were given and
schemes are given here in T'ext-fig. 1, aand b; and the Enopla

NEMERTEA ENOPLA 629

TExtT-Fic. 1.

Schemata of the musculature, nervous system, and body-wall in:
a, Palaeonemertini; b, Heteronemertini; c, Enopla. In trans-
verse section.

List oF ABBREVIATIONS.

acc.st., accessory stylet; an.bl.v., unpaired anal blood-vessel; an. /.,
anal commissure of blood-vascular system; atr., atrium ; bas.,
base of stylet; bas.m., basement membrane; b.g., entodermal
blind-gut; bl.v., blood-vessel; bod.par., body parenchyma ;
br., brain; cer.bl.c., cerebral blood commissure ; cer.can., cerebral
canal; cer.sac., sac of cerebral organ; circ.m.f., circular muscle-
fibres; c.m., new circular musculature of proboscis; d.b.v.,
dorsal blood-vessel; d.g., dorsal ganglion; dig., digestive tract ;
d.n.comm., dorsal nerve commissure; e¢j.d., ejaculatory duct ;
gangl., ganglionic part of cerebral organ; gastr., gastric cavity ;
gl., glands; g.p., gonopore; g.s., gonadial sac; 7.c.m., inner
circular muscle-coat; i.l.m., inner longitudinal muscle-coat ;
int., integument; int.p., intestinal pouch; lat.n., lateral nerve-
cord; J.b.v., lateral blood-vessel; /.m.f., longitudinal muscle-
fibre; m., mouth; m.bl.c., metamerical blood-vessel commis-
sure; musc.sept., muscular septum; nephr., nephridium ;
0.c.m., outer circular musculature; oes., oesophagus; 0.l.m.,
outer longitudinal musculature ; ov., ovary; p.e., proboscidian
epithelium ; p.end., proboscidian endothelium; p.p., proboscis
pore; prob., proboscis ; prob.cav., proboscidian cavity ; prob.n.,
proboscidian nerve; prob.w., proboscidian wall; pyl., pylorus ;
rad.m., dorsoventral musculature ; rh.c.m., normal circular coat
of rhynchocoelomic wall; rh.l.m., normal longitudinal coat of
rhynchocoelomic wall; rhynch., rhynchodaeum; rhynch.bl.v.,
rhynchocoelomic vessel; rhynch.cav., rhynchocoelomic cavity ;
rhynch.div., rnynchocoelomic diverticula ; rhynch.end., rhyncho-
coelomic endothelium ; rhynch.m., rnynchocoelomic musculature ;
rhynch.w., rhynchocoelomic wall; sac, sac with accessory stylets ;
st., stylet; ¢., testis; v.g., ventral ganglion; v.g.s., V-shaped
gonadial sac ; v.n.comm., ventral nerve commissure of the brain.

630 GERARDA STIASNY-WIJNHOFF

with Bdellomorpha (Verrill) and Hoplonemertini (Hubrecht)
(Text-fig. 1, c).

Class Nemertini.

Sub-class 1. Anopla. Sub-class 2. Enopla.
= ns es
Order 1. Order 2. Order 3. Order 4. a

Palaeonemertini. Heteronemertini. Bdellomorpha. Hoplonemertini.

The sub-class Enopla (Text-fig. 1, c) shows a tendency to have
the digestive system and the proboscidian apparatus opening
to the exterior by one aperture. In the Anopla both openings

TExtT-FIG. 2.

sa aie ayes . : y
me .
WUSEUSHERTCNSUGEETS aaa ae ee a
Skee, - 5
eee ane, : : : P -. ee
it Sauegees Seer

Schematic longitudinal section of an unarmed Nemertean after
Birger (6, Pl. xxi, fig. 1, Cerebratulus marginatus).

are always widely separated, as shown in Text-fig. 2; im the
Enopla the common aperture is obtained in different ways.
The Bdellomorpha, containing the parasitic genus Malacobdella,
which Birger included in his Metanemertini, though it lacks
an armed proboscis and shows great differences in the structure
of almost every organ, has its proboscis inserted in the wall of
the stomodaeum (Text-fig. 3). In the Hoplonemertea the same
result, one common mouth, is developed in two other ways, as
will be shown afterwards. As Biirger’s Metanemertini are
only a newer name for Hoplonemertini his subdivision of this
order into Pro- and Holo-rhynchocoelomia might be followed
in our system as well. The main difference between these

NEMERTEA ENOPLA 631

sub-orders exists in the length of the proboscis sheath, which
in the first group is developed in part of the body only, in the
Holorhynchocoelomia, however, is present from the snout to
the tail. That this division is unnatural Brinkmann
showed in his monograph on the pelagic Nemerteans (4).
In this most interesting paper, that contains the minute
anatomical description of eighteen genera of pelagic Nemer-
teans with thirty-two species, the greater part of which are
new, Brinkmann describes nearly related species of one
genus, Balaenanemertes, that might be types of Biirger’s

TEXT-FIG. 3.

rhynch. Ww.

d.ncomm.

-
D eas
22> G

OTT Tr

Birger (6, Pl. xviii, fig. 2).

two sub-orders. In other genera the difference is less great but
still exists. This fact alone is sufficient to demonstrate the
unnaturalness of the subdivision of Biirger’s Metanemertini.
Another fact of interest was that all pelagic forms are nearly
related, and show a peculiarity im the armature of the proboscis
that we knew only from the genus Drepanophorus. This
genus is one of Biirger’s Holorhynchocoelomia. Though
at first meceluded in the family Amphiporidae, the family
Drepanophoridae was later established; and Birger
believed this genus with its paired rhynchocoelomic diverticula
to be the most highly specialized one of his sub-class. In his
study on Uniporus, a nearly related genus, Brinkmann (8)
came to the conclusion that the Drepanophoridae in many
respects are very primitive forms of Hoplonemerteans, which

632 GERARDA STIASNY-WIJNHOFF

conclusion I share. All these facts show evidently that
Birger’s system of Metanemertini does not give the real
relationship of the genera. Brinkmann gives another
system, that seems to suit much better our present state of
knowledge. The armature of the proboscis is the distinctive
character. In most armed Nemerteans the proboscis has one
stylet on the top of a somewhat pear-shaped handle (Text-
fig. 4). The only known exception to this rule was till fifteen
years ago Drepanophorus ; then the Valdivia material showed
that some of the pelagic Nemerteans have a crescent-shaped
handlelike Drepanophorus with many small stylets (Text-fig. 5)

TEXxT-FIc. 4.

Armature of the proboscis of Stichostemma eilhardi after
Montgomery, 1894 (‘ Zool. Anzeiger’, Jahrgang 17, fig. 3)
(Monostilifera).

and Brinkmann (4) confirmed this discovery of Birger
for all pelagic forms. He divides the Hoplonemertini ito
two sub-orders, Polystilifera and Monostilifera (4, p. 145).
The Monostilifera contain all genera of Metanemertini (Birger),
with the exception of (1) Malacobdella (=Bdellomorpha,
Verrill), (2) the pelagic genera, and (8) Drepanophoridae.
The Polystilifera consist of all pelagic Nemerteans and the
genera Drepanophorus (Hubrecht), and Uniporus (Brinkmann).
There can be no doubt as to the naturalness of these sub-orders.
Both contain a great number of genera and species, which
are widely different in structure, but still are more closely
related to each other than to any other form. This is shown
by the position of the mouth, which in the Anopla hes behind

NEMERTEA ENOPLA 633

the brain. The proboscis pore is found in front of it, as a rule
at the tip of the snout (Text-fig. 2). I remarked already that
in Enopla both structures stand in connexion with each other ;
in the Bdellomorpha the rhynchodaeum is absent and the
proboscis cavity opens into the digestive tract (T'ext-fig. 3).
In the Hoplonemertini, it is said, the digestive tract opens into
the rhynchodaeum (Text-fig. 6). This last fact is only true as
far as concerns the Monostilifera. That this connexion of the
two systems is not primitive is shown by the embryology.

TEXT-FIG. 5.

Armature of the proboscis of Drepanophorus spectabilis
(Polystilifera) after Birger (6, Pl. viii, fig. 2). a. Base and
stylets of D. crassus.

In my article on the proboscidian system in Nemertines (18)
I put the facts together in the following way (p. 304) : ‘ Drepa-
nophorus, the genus in which oesophagus and rhynchodaeum
open separately, shows no connexion at all between the two
systems, not even in embryology ; for here the blastopore is
closed, the narrow endoderm part giving rise to the blind gut
by bemg removed forward. The primary ectodermic oeso-
phagus invaginates near the proboscidian system, but perfectly
separately. . . . In all other Hoplonemertea the primary
oesophagus originates in exactly the same way; the mouth
closes afterwards and the primary oesophagus gets a new
opening to the exterior through the rhynchodaeum ’ (Text-
fig. 7).

634 GERARDA STIASNY-WIJNHOFF

What has been said of Drepanophorus is not true for all
Polystilifera. We know species in which mouth and proboscis
pore are widely separated and the mouth even lies under the

TEXT-FIG. 6.

rob.
te le rhynch. W d. Ge GC: Pieces
LTT SUCASRERUAITERPZATEGI NORIO
ne!
; TI
Sot tet be ; ([T77
: mesreeaee ROS = Ses oe PP.
) E = SS Se eS -f “as oS
/ ek eee : ; | eee, ee et i

F eae ; .
= A e 5 q ‘ at | ; Fi ee ee
AUSCGURGheeeeedesamnamnneraunesiesecesnseeessamnneseeen

i}

pyl gastr VRC. O€s.

Schematic longitudinal section of a Monostilifer after Biirger
(6, Pl. xv, fig. 1, Nemertopsis peronea).

TExtT-FIGc. 7.

gt.

oes atr SIAL, rhynck.

Longitudinal section through Prosorochmus after Salensky
(‘““ Uber die embryonale Entwicklung des P. viviparus,”
‘ Bull. Acad. Imp. Sciences ’, Petersb., 1909, fig. 8).

brain. But we also know species in which they communicate
by one pore; we even know all the stages between these
extremes in the Polystilifera. Brinkmann (4) showed the
development of an ectodermal atrium, in which rhynchodaeum
and mouth open separately in the Pelagica. He also describes

rhyrch
Cay.

pyt.

NEMERTEA ENOPLA 635

(8) a large ectodermal atrium in Uniporus, one of his Reptantia ;
Drepanophorus lankesteri (Text-fig. 8) exhibits the
same feature, and so do some genera of the East Indian archi-
pelago. In the Drepanophorus species of the Channel, known
as D. spectabilis, these openings lie quite near to each
other—I might say, they touch each other ; the species known
under the same name from Naples has a large space between
the two, but never does the mouth open into the rhynchodaeum,
nor vice versa. So there is another distinctive character
between Poly- and Mono-stilifera. A difference in habits and

TEXT-FIG. 8.

prob. rhypnehw

Bo oe eee a oe Sa SSREr os Ssa nt nasTeey,
[TT
- aa

SS SSS Sore aSSosESawma SSaea
SSS
SS)

SSS, va,
rT1t1 SS eee SSa5 Th wreh

x = —
ooSsofo S S a
Saco S Daa aaa RS hae Sp

ae)
; =, arr
<, QE -S

= SS
. 7 =
= =

ETI I

5S ae Ry
: << U é>
b. rw rek LL
. crs> Lr <5,
= < s 5 we

ou SS SI
DOUAUHTUUEEUULAUEUSIVIFGUEIUICEOOEIONIUIINNGULOELIUIINaNeSUInGTalaccscscsseeeee eters eHREBREDEE

gastr oes v.N. comm.

Schematic longitudinal section of Drepanophorus lankesteri
from sections.

manner of life accompanies the great differences of structure
in Brinkmann’s divisions of the Polystilifera. The
Pelagica are free-swimming or hovering pelagic Nemerteans
that live at a great depth, without eyes, without the, for
Nemerteans, so characteristic cerebral organs, without a
nephridial system, without rhynchocoelomic diverticula,
without metamerical vascular communications. They might
be considered related to the Monostilifera as well as to the
Polystilifera with all these negative characters if we had not
known the structure of the proboscis armature and of its
sheath. Another positive character is the place of the male
gonads. Though the ova develop in metamerically placed sacs
between each pair of intestinal diverticula, the sperm-cells

636 GERARDA STIASNY-WIJNHOFF

develop only in front of this region, at the side of or directly
behind the brain, a unique fact in Nemerteans. The Reptantia,
containing the genera Drepanophorus and Uniporus, and,
as the Siboga material shows, quite a number of other genera,
that crawl about at the bottom of the sea and its coasts,
have as a rule eyes and metamerical blood-vessels, but always
they possess cerebral organs, nephridia, rhynchocoelomic
diverticula, and metamerically placed o7 gonads. Especially
the cerebral organs are different from those of the Monostilifera
and the development of a sac in this organ as well as the
presence of diverticula of the proboscis sheath show that the
Reptantia are widely different from the Monostilifera. Almost
all Polystilifera that the Siboga expedition brought home
belong to the Reptantia. About one form only there can be
any doubt, as it is collected by the deep-sea trawl to the south
of Timor at a depth of 883 metres. This is the depth in which
most pelagic forms occur and, as occasionally pelagic Nemer-
teans can and have been caught by the trawl, we might be
in doubt as to the manner of life of this Nemertean. Moreover
the inner structure of Siboganemertes weberi reveals
such peculiarities that we cannot with certaimty decide
anything. It has no eyes, but Uniporus, a genus of Reptantia
of the Norwegian sea, living in the dis- and aphotic regions,
lacks them as well. It possesses cerebral organs, but they
are quite minute and of a much more primitive structure
than anything known. Nephridia are present, but metamerical
blood-vessels fail as in Polystilifer@) and Uniporus. Rhyn-
chocoelomic diverticula are present, but instead of lying
peripherically at the outside of all organs as in all Reptantia
(Text-fig. 10) they le inside between the proboscis sheath and

the digestive tract (Text-fig. 9). The testes are placed meta-

merically, but they display features that we do not know in
other Polystilifera. The mouth hes under the brain, which
in its structure shows a great resemblance to the Pelagica and
differs greatly from the Reptantia. The digestive tract, which
lacks an oesophagus in the pelagic forms and has a well-
developed one in the Reptantia, has quite a short balloon-like

NEMERTEA ENOPLA 637

oesophagus that does not reach the brain; on the other hand
it differs greatly from that in both groups, as the different
parts of the stomodaeum, that as a rule gradually pass into
each other (Text-fig. 6), are sharply separated. The narrow

TEXT-FIG. 9.

t. rhynch.din jpg. 1
-i.fm

Seah cle
A) ES rte STR

a ex :

a De

ey

Section through Siboganemertes weberi, n. gen. n. sp.

Trxt-Fric. 10.

rhynch. w

thypch A
ay

lat n. Ce dg.
Section through Drepanophorus albolineatus after Biirger
(6, Pl. xvii, fig. 10).

pylorus opens by a hole in the wall of the gastric cavity ;
oesophagus and gastric cavity communicate by a narrow
opening (Text-fig. 11). It seems evident that, though the
presence of rhynchocoelomic diverticula, cerebral organs, and
metamerically placed genital organs might suggest the enclo-
sure of Siboganemertes weberi in the Reptantia, the

6388 GERARDA STIASNY-WIJNHOFF

structure of the digestive system and the arrangement of the
diverticula of the proboscis sheath separate them.

The most interesting feature seems to be the structure of the
cerebral organ that is so highly developed in the Reptantia.
This sense organ consists in the Monostilifera (Text-fig. 12, a)
of three different parts: a channel, a ganglion, and glands.
These are joined together to a more or less rounded organ with
its own neurilemma. In the Reptantia the same constituents
are present, but as a rule the different parts are more free from
each other and partly le outside the rounded circumference
of the organ, as in Drepanophorus cerinus and wil-

Text-Fic. 11.

Schematic longitudinal section of the digestive tract of Siboga-
nemertes weberi after transverse sections.

leyanus (14, 15) and in Uniporus (8). Moreover, the duct
that leads from the cephalic furrows into the cerebral organ
bifurcates in the organ; one part gives rise to a more or
less spacious sac, characteristic of Reptantia, and the other
part ends as a narrow duct in the glandular portion of the
organ (Text-fig 13). Both sac and glandular tube can lie
embedded in the body parenchyma. In Siboganemertes there
is no bifurcation of the cerebral canal (Text-fig. 12, b). When
the channel gets to the ganglion two small bunches of glands
open into it which lie quite free. The epithelhum is sensory
and never gets glandular. ‘The channel is as primitive as
possible ; it turns backward at the contact with the ganglion
and at its end bends upward and forward on its first part,
where it ends blindly. This is the most primitive cerebral

NEMERTEA ENOPLA 639

organ we know in Enopla, and makes it very probable that
we stand near the origin of this organ.

As to the brain it displays very primitive features too.
Brinkmann writes in his monograph (4, p. 165): ‘ Die den
meisten Drepanophoriden so characteristische starke Vergrés-
serung der dorgalen Gehirnganglien, die dazu fiihrt, dass sie

TExT-FIG. 12.

Cerebral organs of a, Prostoma cephalophorum after Birger
(6, Pl. viii, fig. 28); 6, Siboganemertes weberi (Schema),

wie grosse, kuglige Gebilde den klemen, ventralen, birnférmigen
Ganglien aufsitzen und bei weitem die grisste Masse des
Gehirns bilden, kommt bei den pelagischen Nemertinen nicht
vor. Die hier stattgefundene Reduktion, die dazu fihrt,
dass die dorsalen Ganglen héchstens nur wenig grésser sind
als die ventralen, ja gar nicht selten kleiner als diese werden
konnen, ist zweifelsohne durch das Verschwinden der Cerebral-
organe verursacht, denn es sind ja diese Organe, die vor allem
von den dorsalen Ganglien aus innerviert werden.’ Siboga-
nemertes weberi exhibits the same structure of brain as
certain Pelagica, though a small cerebral organ is present.

640 GERARDA STIASNY-WIJNHOFF

It is comparable with those in which the dorsal ganglia lie
quite outside the ventral. Both have the same length, but the
position of the dorsal lobe is somewhat more forward than the

Trext-Fic. 13.

Cerebral organ of Drepanophorus spectabilis after Birger
(6, Pl. viii, fig. 23), Schema.

TeEextT-Fic. 14.

Transverse sections through the brain in the hinder region of the
ventral ganglion: a, of Siboganemertes weberi; Bb. of
Drepanophorus albolineatus of Birger (6, Pl. xvii,
fig. 3), Schema. The two crosses give the boundary of the
ganglia and the corresponding places in the two genera; c, Dre-
panophorus latus, just after the origin of the lateral nerve-
cord (6, Pl. xxiv, fig. 43).

ventral. In Text-fig. 14, a, the ventral lobe obtaimed its greatest
diameter, as the nerve-cells of the nerve-cord are partly
included. The two crosses give the limit of the lobes. In

NEMERTEA ENOPLA 641

Reptantia the dorsal lobe attains its greatest development in
its himder part, when the ventral lobe has disappeared. A
section of Drepanophorus albolineatus (Text-fig. 14, b)
to be compared with Text-fig. 14, a, reveals the difference
between the two. The proportions have changed ; instead of
the ventral the dorsal ganglion exceeds in Drepanophorus,
and if we compare a section of another form, in which the
nerve-cord is seen instead of the ventral lobe, the contrast is
still more obvious (Text-fig. 14, c). I cannot agree with
Brinkmann that the disappearance of the cerebral organs
caused a reduction of the dorsal brain-lobe in Pelagica.
Brinkmann considers these to be descendants of true
Drepanophoridae that have lost eyes, cerebral organs, the great
development of the brain, the nephridia, the rhynchocoelomic
diverticula, the anastomosing blood-vessels. As to the eyes
I might agree with him ; it seems quite plausible that species
or even genera that live in the aphotic regions of the sea lose
their eyes, as most Pelagica, Siboganemertes, and Uniporus
hyalinus, though for Uniporus acutocaudatus and
U. borealis this reason cannot exist as they live in the dyspho-
tic zone just as well. The presence of atrophied eyes in some
pelagic genera, however, makes it probable that they got lost
in the other. But certainly the Pelagica never possessed cere-
bral organs. These have developed in different ways in armed
and unarmed Nemerteans. In both sub-classes we know
genera without them, and these in both are primitive forms.
Callinera, Carinesta, Cephalothrix belong to the most primitive
Palaeonemerteans and they have no cerebral organs. In the
Enopla this sense organ is absent in the Pelagica and in Mala-

cobdella. The parasitic genus Gononemertes has them and in ,)\

Carcimonemertes they seem to fail. Why must it have got
lost in Malacobdella and Carcinonemertes, when it is present
in the third parasitic genus? As to Carcinonemertes, that
belongs to a non-parasitic family of Monostilifera with well-
developed cerebral organs, it seems natural to consider the
parasitic habits of the genus as the cause of their absence,
though nothing is less certain. In Bdellonemertea this
NO. 268 Loy Wl

642 GERARDA STIASNY-WIJNHOFF

reason is quite absent, and so it is in pelagic forms. When we
can, moreover, follow the development, as is the case (1) in
the Anopla, from stages like Tubulanus pellucidus,
Procarinina, and other Tubulanus species through Hubrechtia
to the Heteronemerteans, and on the other hand in Enopla
from Siboganemertes (Text-fig. 12, b) to several Monostilifera
(Text-fig. 12, a) and to Drepanophorus (Text-fig. 18), from
(2) the irregular organ with partly free and simple constituents
through the composite and irregular organ of several Reptantia
as Uniporus (8, Pl. i, figs. 6 and 7) and Drepanophorus
cerinus, willeyanus, indicus 15, 14), to the well-

Text-Fic. 15.

Transverse section through the cerebral organ of Emplectonema
gracile after Birger (6, Pl. xxvi, fig. 41).

defined organ of D. spectabilis (Text-fig. 13), or from
(3) Siboganemertes (Text-fig. 12, b) through stages as Em-
plectonema (Text-fig. 15) and Prostoma cruciatum to
Prostoma cephalophorum (Text-fig. 12, a), 1t seems to
be rather probable that this organ has developed in Nemerteans
and has not been inherited from now extinct ancestors. The
great development of the dorsal brain-lobe is a characteristic
feature of the Reptantia, but need not have been a possession
of all Polystilifera. In other forms with a well-developed
cerebral organ, as in Amphiporus, the difference of the propor-
tions of the brain-lobes is not great, and I am rather melmed
to think that the development of the sac caused the different
structure of the dorsal braim-lobe of the Reptantia. Paired
rhynchocoelomic diverticula are absent in all Nemer-

NEMERTEA ENOPLA 643

teans with the exception of Reptantia and Siboganemertes
and in these they developed in different ways.’ As in the

1 Here seems to be the right place to mention the fact that in one
species of Monostilifera the presence of paired rhynchocoelomic diverti-
cula is noted by Birger, i.e. Amphiporus stannii. However,
these diverticula are quite other structures, or at least much more primi-
tive; they are present in the nephridial region alone and are very small
though wide. The musculature of the rhynchocoelomic cavity widens out
at certain places. These are the sacs that have a wide lumen and open
with a wide mouth into the rhynchocoelomic cavity. The muscular walls
of these sacs are the regular continuations of, and just as thick as, the
rhynchocoelomic wall, and a difference of structure seems not to exist.
The wall of these diverticula in Polystilifera is as a rule much thinner,
even when contracted, and we know that the mouth is provided with
a sphincter, that is absent in A. stannii. In some unarmed genera
other rhynchocoelomic diverticula, paired and unpaired, exist, but they
never are comparable with those of the Polystilifera. Amphiporus
stannii is a Monostilifer, as its stylet is well known. In Drepano-
phorus valdiviae (Birger), which species has exactly the same
rhynchocoelomic diverticula as Amphiporus stannii, the stylet is
unknown ; in both species these structures are restricted to the nephridial
region, and in other characteristics they are very much alike too: they
have no eyes, both possess small cerebral organs without a sac, behind or
partly behind the brains, both have lateral nerve-cords (not ventral as in
Drepanophorus), both have a layer of glands in the head that fails in all
Polystilifera and is present in Amphiporus, both have brains that are
quite different from all Polystilifera, with a very small dorsal and larger,
perfectly separated ventral ganglia, in both the vascular system differs
from that of Polystilifera by the presence of a dorsal loop over the brain,
&c., &c. Birger says in his monograph of the Valdivia expedition
(9, p. 174): * Leider ist aber der Riissel nicht vorhanden, und die Organisa-
tion weist einige Ziige auf, die mehr auf Amphiporus als auf Drepano-
phorus hindeuten ; indessen ist dieses Stiick dem Genus Drepanophorus
zuzurechnen, weil sein Rhynchocélom, wenigstens im vordersten Abschnitt,
laterale, einander gegeniiber entspringende seitliche Aussackungen besitzt,
die bisher nur. von Drepanophorus bekannt sind.’ He forgets, however,
that he himself described them in 1895 in Amphiporus stannii
in the monograph of the Nemerteans of Naples, p. 571, and Pl. xvii,
figs. 5, 13,and 14. A comparison of these figures with those of the Valdivia
(Pl. xxxi) gives the striking resemblance of the above-discussed species,
which certainly belong to one genus, which I might mention Valdivia-
nemertes. The presence of the only stylet in Valdivianemertes
stannii (Grube) makes it certain that both V. stannii and V. val-

uu 2

644 GERARDA STIASNY-WIJNHOFF

Pelagica no traces of a reduction in the proboscidian system
are to be found, as emphatically stated by Brinkmann,
we must consider this tribe as the more primitive one in the
Polystilifera.

The structure of the muscular wall of the rhyncho-
coelomic cavity seems to prove this. In 1914 I tried to demon-
strate (18) that the proboscis and its sheath together are an
invagination of the body-wall, and that all parts of the body-

TeExt-Fic. 16.

:
oO
ce)
\
ery: ‘oF \
S AW 2 SQA
h W\\ 2 poco
aS 0 o WN
' yo oo >
pe oS) ac
ayehee 3
.

Section through the proboscis of Amphiporus pulcher after
Birger (6, Pl. xxiii, fig. 3).

wall are to be found in situ, either in the proboscis or in
the wall of the sheath. In the Anopla this seems quite clear,
but in the Enopla several difficulties arise. The presence of
the third or inner muscular layer of the body-wall, which in
Palaeonemerteans is characteristic, the inner circular muscle-
coat (Text-fig. 1, a), has never been demonstrated, though in
Drepanophorus, as I know now, its presence is quite clear
in the stomodaeal region. Also the dorsoventral muscles show
the same peculiarities as in the Anopla, where they are derivates
of this musculature. So it was not certain whether we had
a right to look for this layer in the proboscidian system of the

diviae (Biirger) belong to the Monostilifera. Through this conclusion
we have excluded Drepanophorus valdiviae from the Polystilifera,
in which it might cause much trouble by the different structure of almost
every organ.

NEMERTEA ENOPLA 645

armed Nemerteans. The other difficulty was that the spot
where delamination took place seemed to be different in the
Enopla. For in the unarmed Nemertini the inner longitudinal
muscle-layer was split into two parts that enclose the rhyncho-
coelomic cavity, which is lined by an endothelium. However,
in the Hoplonemertini a circular muscle-layer lies between
the longitudinal fibres and the endothelial linmg of the pro-
boscis (Text-fig. 16). I then suggested that these circular

Trext-Fic. 17.

Section through Emplectonema gracile after Birger
(6, Pl. xv, fig. 27).

fibres are a new acquisition and do not belong to the primary
proboscis, as the proboscis sheath itself is built like that of the
Anopla (Text-fig. 17). Chuniella, one of the most primitive
Polystilifera, seems to prove this supposition, as the proboscis
has no circular muscles beneath the endothelium, and in
Monostilifera the genus Zygonemertes (19) shows the same
feature, as sections from South African species reveal. We
knew nothing then about Polystilifera with the exception of
the genus Drepanophorus, in which the wall of the rhyncho-
coelomic cavity consists of longitudinal and circular muscle-
fibres, that are interwoven (T'ext-fig. 18). In many Pelagica

646 GERARDA STIASNY-WIJNHOFF

this is the case too, as in Siboganemertes and all Reptantia ;
but in Chuniella, Nectonemertes, Natonemertes, Para-, Pro-,
and Balaenanemertes no traces of interlacing of these fibres
are found, and the longitudinal layer lies next to the endothe-
lium exactly as in the Monostilifera (Text-fig. 17). Brink-
mann considers this kind of rhynchocoelomic musculature
not as primitive, because the layers show another arrangement
at the place of insertion of the proboscis. We know, however,
from the Anopla that exactly in this part of the proboscidian

Trext-Fic. 18.

Birger (6, Pl. xxiii, fig. 37).

system the first traces of the new outer longitudinal layer appear,
when the middle part holds the older structure; that here
other layers disappear first, that mm other species fail absolutely
and in more primitive genera are present in all parts; that
at this spot new constrictors and retractors can develop, that
in most species are unknown, to be short, that all changes
start in this part of the proboscidian system. When we remem-
ber, moreover, that this spot is the place where originally the
invagination of the whole system took place, that by the
development of the precerebral region, as will be discussed
later, the continuity with the body musculature was broken,
‘Muskelseptum, Riisselfixatoren’ originated, * Seitenstamm-
muskeln’ developed; that the imner circular muscle-layer
is and must be present in the rhynchocoelomic wall, but almost
disappeared in the body-wall; that originally here the central

NEMERTEA ENOPLA 647

nervous system was found embedded in the longitudinal
musculature, as seems still to be the case in some species
(Brinkmann, 4), and that its place changed in the different
genera, then we understand that we cannot look for primary
conditions in this part of the proboscidian system. That the
longitudinal musculature of proboscis and sheath are in
contact with each other at the place of insertion is quite natural
(Text-fig. 19), while they both are part of originally one layer
and from one place take their origin.

TExt-FiG. 19.

prob. rhynch.
eer i musc.sept

te \Oaee :
ANC. /

d.n.cominv
cy7

rhynch
saeee

SE

rh.c.m.

Dorsal part of a longitudinal section through the anterior region
of Balaenanemertes musculocaudatus (Brinkmann)
with protruded proboscis (4, Pl. 15, fig. 10).

Brinkmann’s statement that the brains of Pendo-
nemertes and Balaenanemertes are situated in the middle of
the musculature of the proboscidian system is of the greatest
importance ; for we know from Drepanophorus that the
longitudinal musculature is m contact with the same parts
of the body-wall by a muscular septum, which separates the
precerebral or head-region from the brain and the body and
always expands just before the gangha. In the Pelagica this
septum as a rule is broken up into several muscles which
Brinkmann calls ‘ Risselfixatoren’ and that as a rule lie
outside and before the nerve-ring. The brain lies at the same

648 GERARDA STIASNY-WIJNHOFF

place as the nerves of all EKnopla, on the inner side of the inner
longitudinal musculature. So we have to look for the inner
circular muscles of these genera behind the brain, and not
before it. Jf we ask where the brain lies exactly in Pendo-
nemertes and Balaenanemertes, Pl. v, fig. 1, Pl. xiv, fig. 19,
Pl. xv, figs. 8 and 4, Pl. xvi, figs. 17 and 18, of the monograph
(4) show us that it really is found between two layers of longi-
tudinal musculature. Whether the outer layer consists of
strands that go to the body-wall (‘ Risselfixatoren ’) or of the
inner parts of them that still have to join the wall of the cavity,
is not clear. But in any case it is certain that we can expect
the inner circular muscle-layer only behind the braim and not
before it. Wherever Brinkmann describes the exact rela-
tions of the muscle-layers in these parts, it invariably is men-
tioned that the first traces of the outer circular musculature
of the proboscis sheath are found behind the brain. This cannot
always be so, for I know cases in the Reptantia where the remains
of the circular musculature of the body-wall are found around
the hinder parts of the brain, and in this case it is evident that
the contact with the rhynchocoelomic part of this layer must
be looked for before the brain. In such cases we must expect
the outer circular musculature of the sheath to be in the
nerve-ring. In others I noted the same beginning of this layer
as Brinkmann, 1.e. behind the nerve-ring.

The wall of the cavity before the brain is built differently
in different cases. It is interesting to note that in one Malayan
species all circular muscle-fibres are absent in front of the brain,
in another all longitudinal, and always the interlacmg begins
behind the nerve-ring. Brinkmann describes m all his genera
of Pelagica the presence of an inner circular muscle-layer
(as the direct continuation of the new proboscidian layer)
before the brain and outside of it a longitudinal layer which,
he says, passes through the circular musculature behmd the
brain and so comes to lie inside (Text-fig. 19). If, however,
this really was a passing through we should find an interlacing
of fibres at this place. Though Brinkmann is very exact
in his statements he never mentions this, and his figures show

NEMERTEA ENOPLA 649

everywhere a very distinct border of the musculature (Text-
fig. 19). Pro- and Parabalaenanemertes, Balaenanemertes,
Nato-, Nectonemertes, and Chuniella show us this type of
rhynchocoelom, as found in all Monostilifera too, and so does
the greater part of the sheath of Pelagonemertes. The develop-
ment of a third layer, which might be mdicated where the
proboscis is affixed, never took place, as I will now demonstrate.

The interlacmg of circular and longitudinal muscle-fibres,
which in Pelagonemertes took place in the hinder part of the
rhynchocoelomic wall, is found in the genera Protopelago-
nemertes, Plotonemertes, Pendonemertes, Mergonemertes,
Paradino-, Phallo-, Crasso-, and Planktonemertes, in all
Reptantia and in Siboganemertes weberi. Whether
the new circular muscle-coat of Brinkmann is present or
not we cannot decide im these genera ; in Pelagonemertes it
certainly is absent, and in the Malayan species referred to
above the interlacing took place between the two original
layers, as I will describe in Siboganemertes. The proboscis
has a thin outer circular muscle-coat and a thick longitudinal
coat, but the new layer is absent. The precerebral septum
which connects the longitudinal musculature of the body with
the rhynchocoelomic wall and proboscis les exactly im front
of the brain. Inner circular muscle-fibres between this septum
and the endothelial liming of the cavity of the sheath fail
absolutely ; a great quantity of fine mesenchymatic fibres,
which stain quite differently and are found at many places in
the body parenchyma too, les inside the endothelium and
between these the first longitudinal fibres are embedded.
Outside of these the first circular fibres appear behind the
nerve-ring and they are very few. The whole muscular wall
is thin and, in the ventral part, disappears but for a few
longitudinal fibres. It is, however, quite clear that in the
proximal part an interlacing of fibres takes place, and here
certainly the new circular layer has not developed. It is absent
in the proboscis too. From the results obtained in Siboga-
nemertes, and from similar facts in some Drepanophoridae and
in Pelagonemertes, I might conclude that the development of

650 GERARDA STIASNY-WIJNHOFF

the wall of the sheath in Pelagica took place in the same way.
Primary are the stages with two muscle-layers, as Chuniella,
&e. The first stages of interlacing are given in Armaueria and
Dinonemertes (Text-fig. 20). In these genera started the
penetration of the circular muscle-coat by the longitudinal
fibres, or vice versa ; but the longitudinal musculature has not
yet reached the outside. We see an interlaced inner wall,
a longitudinal middle coat, and outside of it a circular layer.

Text-Fic. 20.

Transverse section through the rhynchocoelomic cavity of Dino-
nemertes alberti (Brinkmann) (4, Pl. vii, fig. 13).

As Dinonemertes seems to be connected with Mergo-, Phallo-,
and Paradinonemertes and Armaueria with Pendonemertes,
I might rather solve the problem of these genera as the begin-
ning of the interlacing, which is completed in Protopelago-,
Ploto-, Pendo-, Crasso-, Plankto-, Mergo-, Phallo-, and Para-
cdinonemertes.

Pelagonemertes kept the more primitive stage in the proximal
part of the sheath, which fact seems to point to the hinder part
ot the cavity as the place of origin of the interlacmg. Bir-
geriella gives still a higher development that confirms these

NEMERTEA ENOPLA 651

views, as it is the most specialized genus of the Pelagica in
many respects. The distal part acquired an inner circular
layer, as the originally outer circular coat after the stage of
interlacing, shown still in the proximal part, lost its longitu-
dinal fibres which all he outside of it. This is im accordance
with the other anatomical facts, that make us look for the
nearest relations of this aberrant genus among species with
a basket-like structure of the sheath; proximally it is
basket-like in Biirgeriella too; the distal part has an inner
circular and an outer longitudinal layer. Compared with
Text-fig. 20 of Dinonemertes the position of the fibres is this,
that the whole circular muscle-coat traversed the longitudinal
one (in Dinonemertes only the inner parts) and so came to
le inside. Other traces of an inner circular muscle-coat fail,
and as Biirgeriella evidently is a highly specialized genus
it would be rather incomprehensible why it should be the only
one that had beheld this primitive feature on Brinkmann’s
explanation of facts.

The other support of Brinkmann’s theory of a third
mus¢le-layer of the wall of the sheath is Protopelagonemertes,
in which genus the interlacing of fibres is already found in
the nerve-ring. However, if we suppose, as I do, that the
interlacing starts in the hinder part of the cavity and goes on
from behind forwards, as shown in Birgeriella and Pelago-
nemertes, every reason fails why the interlacing should stop
with the brain as the nerve-ring lies in the muscular septum.
Protopelagonemertes bears its name quite undeserved, as
Pelagonemertes seems not to be related to it and is also to
a certain extent more primitive.

The result of this discussion is that the genus Chuniella,
which after, Brinkmann’s description must be one of the
most primitive genera if not the most primitive genus of the
Polystilifera, perhaps even of the Hoplonemertini, has a pro-
boseis with exactly the same muscular layers as all primitive
Anopla, Malacobdella, and some Monostilifera, and as Siboga-
nemertes weberi, the most primitive genus of the
Reptantia ; that also the wall of the sheath in this genus is

652 GERARDA STIASNY-WIJNHOFF

built like that of all Anopla and of all Monostilifera, and that
this wall in the Polystilifera is found in the pelagic genera Nato-
and Nectonemertes and the family Balaenanemertidae as well.
That the interlacing of these two original muscle-coats, which
is characteristic of all Polystilifera Reptantia and elsewhere
is unknown, is also found in many Pelagica ; that the process
of interlacing seems to start in the hinder part of the rhyncho-
coelomic cavity and proceeds proximally, as shown in Pelago-
nemertes. That we see the penetration of the two layers go
on in Dinonemertes and Armaueria, and that the interlacing
is completed in all other genera of Pelagica and in the Reptantia.
That in one genus this process resulted in the inversion of the
original layers, i.e. the aberrant genus Biirgeriella, where the
proximal part of the sheath has the basket-like structure
characteristic of Polystilifera, and the distal part, as in Pelago-
nemertes, shows the result of this process.

If we look at the digestive tract three remarkable
points are to be distinguished. The position of the mouth
under the brain was stated to be very primitive in armed Nemer-
teans and even in Polystilifera to be quite unusual. As to the
oesophagus we have the statement of Brinkmann that this
part of the stomodaeum is absent in all Pelagica with the only
exception perhaps of Planktonemertes. His fig. 23, Pl. xin (4),
gives no right to compare this small oesophagus with that of
Siboganemertes (‘Text-fig. 11); after his description on p. 24,
however, we can hardly speak of an oesophagus, and truly can
say in Pelagica the oesophagus is absent, as in the unarmed
genera. But in the Reptantia a well-developed oesophagus is
always, in the Monostilifera, as a rule present. We know its
absence in Amphiporus marmoratus (Birger) (6,
Pl. xvi, fig. 1), though in Amphiporus marmoratus
(Joubin) it is well developed as in most other species (12,
p- 564, fig. 4). This figure interests us still more because the
different parts of the stomodaeum with the exception of the
oesophagus show about the same features as Siboganemertes.
The pyloric tube of Joubin’s species is much wider than in
our specimen, but it opens into the gastric cavity at the same

NEMERTEA ENOPLA 653

place. What is the contimuation of the oesophagus in
A. marmoratus (Joubin) is the ventral unpaired diverti-
culum of Siboganemertes, and the true gastric cavity les
enclosed between the pyloric tube and the ventral diverticulum.
All parts of the stomodaeum are unpaired in Siboganemertes
but the intestinal blind-gut shows at the side of ventral
unpaired pouches paired lateral diverticula that are longer
than the blind-gut itself, as m A. marmoratus (Joubin).
Nothing of this kind was ever found in Polystilifera, though
some very interesting features are known from Brinkmann’s
studies on the bathypelagic species. Not only do they show the
absence of an oesophagus, but the whole stomodaeum is very
short and much less differentiated. In almost all his figures
the pylorus is already shown beneath the brain, and as a rule
the blind-gut extends till here. In fig. 9, Pl. xv, a longitudinal
section shows the short and narrow gastric cavity of Balaena-
nemertes musculocaudatus; fig. 13, Pl xu (Text-
fig. 26), gives the same features in Nectonemertes primi-
tiva, and Brinkmann states that in N. minima the
epithelium of the gastric cavity is unfolded, the cavity still
narrower and shorter. Brinkmann takes these forms as
the most reduced ones. However, how can we explain these
differences in the same structure within a monophyletic group,
as the Polystilifera certainly are, the highly differentiated gastric
cavity of Siboganemertes, the quite differently but not less
highly developed structures of the Drepanophoridae and the
more or less simple stomodaeum of the Pelagica, if we do not
suppose these to be primitive ?

The stomodaeum in armed Nemerteans is a structure different
from that in the Anopla, as is shown by its development and
by the presence of an entodermal blind-gut in the first. We
know it to be a simple structure, a mere narrow tube in Oto-
typhlonemertes, in Zygonemertes it is not much more; we
know that the oesophagus is absent in Geonemertes, in Sticho-
stemma. Why then must the Pelagica, that have the same
peculiarities, have developed from highly differentiated forms
as Drepanophorus ? On the contrary we see here how the simple

654 GHRARDA STIASNY-WIJNHOFF

invagiating stomodaeum with a few glandular cells only
develops into a gastric tube that opens only a short distance
behind the beginning of the enteron, asin Balaenanemertes
musculocaudatus or Nectonemertes. With the greater
development of the stomodaeum the different parts become
better differentiated, in the first place gastric cavity and
pylorus, and this can be obtaimed in different ways, as shown
by Siboganemertes (Text-fig. 11) and Drepanophorus (Text-
fig. 8), or by A. marmoratus (Joubm) and A. mar-
moratus (Birger); the Pelagica provide us with a whole
series of stages in this development. ‘Thirdly an oesophagus is
differentiated, which all Pelagica lack and often even other
groups. So Siboganemertes with its diverging structure of the
digestive system is not as primitive as the pelagic forms ; but
it cannot be included in the Drepanophoridae either, as the
development goes in a different direction.

The vascular system shows a very primitive type, as
in Siboganemertes there are no anastomozing vessels with the
exception of the cephalic loops. ‘The cephale vessels bend
into the nervous ring of the brain in the ordinary way and
a dorsal vessel is present, as far as sections were made ; but it
is not in contact with the rhynchocoelomic cavity. The absence
of metamerically situated vascular loops is known in Uniporus,
in Reptantia, and in all Pelagica. The occasional presence of a
double anal loop in an abnormal individual of Pendonemer-
tes levinseni (Text-fig. 21, b) and the existence of a blind
dorsal median vessel in the tail arising from the anal loop, makes
it at first sight rather plausible that the reduction series as given
by Brinkmann on p. 163 of his monograph (4) gives a true
account of the facts. But when we know that in primitive
Anopla the anastomozing vessels are absent, that they fail in
Uniporus (in which genus even the anal loop should be absent)
(3), and that they fail in Siboganemertes, we become sceptical
to the explanation of their absence in Pelagica. Moreover the
reconstruction of the tail of Pendonemertes on p. 20 (Text-
fig. 21, b) and the scheme on p. 1638 (Text-fig. 21, ¢c) are rather
different, and it seems not at all certain that the hinder vessel

NEMERTEA ENOPLA 655

is the continuation of the lateral vessel; the course of the
blood-vessels seems too irregular to decide anything from this
single abnormality.

Another interesting fact in Siboganemertes is the absence of
any connexion between the dorsal vessel and the rhynchocoelo-
mie cavity. ‘This is known from one form of Polystilifera,
Armaueria fusca. The Monostilifera exhibit the same
feature in some Prostoma species (P. amphiporoides,
duboisi, antarcticum, gulliveri (Birger)), and it
is found in Malacobdella. Wherever a dorsal blood-vessel exists

TExtT-FIGc. 21.

Blood-vascular system in the tail of Pendonemertes levin-
seni. a, Schema of normal individual after Brinkmann
(4, p. 163, Text-fig. 29, IL); 6, abnormal individual (4, p. 20,
Text-fig. 4); c, schema of this abnormality (4, p. 163, Text-fig.
29, III).

in the unarmed forms, it is in connexion with the rhyncho-
coelomic cavity (Text-fig. 20), though other special rhyncho-
coelomic vessels may be present. In most Palaeonemerteans,
however, the dorsal vessel is quite absent. This is a rare
case in Hoplonemertini, and, as far as I know, it has been
described in Pelagonemertes moseleyi, Balaena-
nemertes chuni, and Carcinonemertes carcino-
phila. Brinkmann, guided by the opinion that the
Pelagica are reduced Drepanophoridae, considers these forms

656 GERARDA STIASNY-WIJNHOFF

as the most advanced ones; I, however, believe that the
formation of a dorsal blood-vessel takes place in the Pelagiea ;
that it has been formed in two ways, either in relation to the
rhynchocoelomic cavity, or quite separately. This last way
is represented in three aberrant genera, Armaueria in the
Pelagica, Malacobdella, the Bdellonemertean, and Siboga-
nemertes, the representative of a new group of Reptantia.
Perhaps other genera passed this stage to acquire a rhyncho
coelomic vessel afterwards, as might be plausible in Prostoma.
Some genera, however, seem to have got this rhynchocoelomic¢
vessel directly, as isshownin Pelagonemertes rollestoni
and the nearly related genus Natonemertes with a short,
blind-ending proboscidian blood-vessel, or in the family of the
Balaenanemertidae, where a dorsal vessel is absent in
B. chuni, and in other species of the same genus a blind
rhynchocoelomic vessel is present as well as m Probalaena-
nemertes and Parabalaenanemertes. Another fact of imterest
in the blood-vascular system, and which seems to demonstrate
how the organisms of this group try to obtain a certain result in
different ways, is the development of what Brinkmann ealls
‘ Ovarialschlingen ’. He demonstrates in Dinonemertes
investigatoris that the lateral blood-vessels im the
gonadial region make large, irregular loops between the
ovaries and the entodermal sacs. These loops that convey
the nutrition from the sacs to the ovaries are absent in all other
forms ; but I found them also in Siboganemertes. It is supposed
that the vascular loops between the dorsal and lateral vessels
of other Nemerteans have the same purpose, and Brink-
mann remarks that this fact states Dollo’s law of irrever-
sibility, as the regular loops that once disappeared did not
return, but another structure took on the same function.
What we find m Dinonemertes and Siboganemertes can
perhaps just as well be the beginning of what results in vascular
loops between the vessels. So everywhere I reach the same
result ; the Pelagica show the development of every organ, from
the primitive stages known in Palaeonemerteans to the
specialized features of Drepanophoridae and Monostilifera ; the

NEMERTEA ENOPLA 657

development supposed by Brinkmann seems to have
taken place in the reverse direction; what he calls highly
reduced, I call primitive, and vice versa.

This disagreement does not extend to the nephridial
system; for this is present in all Nemerteans without
exception that do not belong to the Pelagica. That it has
not yet been found in Prosadenoporus must partly be due to
the highly developed head-glands that extend into the nephri-
dial region, partly to the smallness of the canals or the preserva-
tion. As all Platyhelminths possess a well-developed nephridial
system, we are obliged to explain its absence in the Pelagica
by reduction. In Siboganemertes nephridia are present, but
of a type that differs from that of the Reptantia. A large
efferent duct is present at each side, extends behind the real
nephridium, and has a more caudal, lateral mouth. The
nephridia lie at the side of the dorsal brain-lobes and the
cerebral organs, and just behind these obtain their greatest
development. The ducts open laterally behind the end of the
pyloric tube. This type is known from primitive Anopla,
a well-localized system of canals with a long efferent duct,
quite different from the other types of nephridia that extend
through the whole body in the same way and have one or more
short ducts. In the Reptantia also it is much less circumscribed,
extends as a rule from the end of the brain along the stomodaeal
tract, and has one efferent duct that can take its origin in any
part of the system and opens directly to the exterior. Our
knowledge of Monostilifera is as vet too incomplete to under-
stand the value of these facts.

The gonads, however, seem to be much more interesting.
The only individual of Siboganemertes happened to be a male
with well-developed testes, a fact of the greatest importance,
as the Pelagica exhibit an extraordinary feature in the position
of these glands that is characteristic of the group.

As a rule the gonads le, be they ¢ or o7, m the intestinal
region in armed and unarmed Nemerteans. The only exception
are the testes of the Pelagica that never are developed in this
region, and always le before it, directly behind, at the side of,

NO. 268 eK

658 GERARDA STIASNY-WIJNHOFF

or even before the brain. The ovaries are developed in the
usual way. These facts can easily be understood as we know
the testes of Platyhelminths to be developed all over the body.
The arrangement of the gonads in primitive Anopla without
intestinal pouches is absolutely irregular, as shown by Birger
in Tubulanus (8, Pl. iv, fig. 2). Two interesting facts are to
be mentioned: the gonads are placed in several irregular rows,
and the gonopores le on the dorsal surface. In the Anopla
we can follow the development from this stage without intes-
tinal diverticula to the pseudo-metameric arrangement in
Lineus and Cerebratulus, where always one gonadial sac hes
between two intestinal pouches, opening to the exterior by
one row of dorsal gonopores.

In Enopla the Bdellomorpha display the same irregular
position, as for instance Tubulanus polymorphus, and
have dorsal gonopores. The Hoplonemertini show a great
variety of arrangement. Tirst we have to look at the Mono-
stilifera, of which Geonemertes, Nemertopsis (Text-fig. 22, a),
Progadenoporus, Prosorochmus have a number of gonads
between two following intestinal pouches, the first stage of
arrangement that follows on the above-described displacement
of Tubulanus polymorphus in the Anopla. All these
worms are more rounded than the flat Malacobdella and the
Tubulanidae. In consequence the gonopores partly lie more
laterally, but always above the nerve-cords. The next stage is
the reduction of the number of gonads per pseudomere to one
on each side, as in Prostoma coronatum (8, PI. ii, fig. 5)
and Amphiporus species. At last we get a still greater reduction
of this number as in A. pulcher (8, Pl. xin, fig. 6).

In the Polystilifera we know two genera of Reptantia,
Uniporus and Drepanophorus. Uniporus (Text-fig. 22, b)
has in each pseudomere two to five gonadial sacs with dorsal
pores and exhibits im consequence a very primitive feature.
In Drepanophorus we know the great regularity m which
intestinal pouches and gonads alternate, one sac between two
pouches. But the gonopores lie laterally (Text-fig. 28, a)
as in D. willeyanus, cerinus, indicus, or ventrally

NEMERTEA ENOPLA 659

(Text-fig. 24, D. albolineatus). Moreover, these sacs have,
as Punnett (14, 15) showed, a peculiar form. In more
rounded species, as the first, they are V-shaped (Text-fig. 23, a)
with the two legs above and beneath the intestine and the
lateral gonoduct at the place of junction of the legs. In
Siboganemertes this is found too, but the sperm is not developed
in this large sac alone. Peripherally small sacs are found

TEXT-FIG. 22.

a. b.

Sections through gonadial region: a, of Nemertopsis peronea
after Birger (6, Pl. xv, fig. 5); b, of Uniporus hyalinus
after Brinkmann (8, Pl. 1, fig. 5).

(Text-fig. 9) which open into the central V-shaped sac (Text-
fig. 23, b). Burger found the same development of ovaries
in Drepanophorus albolineatus (Text-fig. 24), one
egg in each sac; but in other species the central sac itself is
filled with eggs or sperm (Text-fig. 28, a). It seems to me
that the small sacs are comparable with the gonadial sacs of
Uniporus, Geonemertes, Tubulanus, Malacobdella, and that
the central V-shaped pouch is a new growth in the Reptantia
and Siboganemertes. The epithelium of the central pouch,
whether it is simply a new gonoduct or the invaginated ectoderm
with the original gonopores, acquires later the function of the
gonadial epithelium and the original gonads disappear. In
x K/2

660 GERARDA STIASNY-WIJNHOFF

each case we see the tendency of the gonads, whether testes or
ovaries, to arrange themselves metamerically and become
reduced in number.

In the Pelagica the differences of sexes are greater. The

Trext-Fic. 23.

pte)
Bs MN
Pee es
yf Js aS
WY

ed SE
« “Pr AO

Sections through the gonadial region: a, of Drepanophorus
willeyanus (Punnett) (14, Pl. lix, fig. 20); b, of Siboga-
nemertes weberi.

TEext-Fic. 24.

Section through Drepanophorus albolineatus with an
ovary after Biirger (6, Pl. xxvii, fig. 52).

ovaries are metamerically developed, never more than one
between two intestinal pouches, but the gonopores le on the
ventral side. This may be the case in Drepanophorus too
(Text-fig. 24). In broader and flattened forms the body becomes
thin with flat edges, outgrowths of the body-wall and paren-
chyma, that in certain species contain no organs at all, in others
rhynchocoelomic diverticula only. In some of these species it
is quite obvious that the outgrowth took place above the

NEMERTEA ENOPLA 661

lateral gonopores, that in this way became ventral (Drepano-
phoridae-Siboga expedition) ; the V-shaped gonadial sac still
is quite clear. In D. albolineatus (Text-fig. 24) the
gonad acquired another, third part, that lies in this ‘ Seiten-
rippe ’, and the ventral primary leg was somewhat reduced ;
but all three parts are present. In the Pelagica the develop-
ment of the ventral ovaries cannot have taken place in this
way, as all traces of the V-shaped sac or of many ovaries are
absent, and real dorsal gonopores are unknown in both sexes.
The only cases, as far as our present knowledge goes, in which
the gonopores are not ventral but lateral seem to be the testes
of Armaueria and Parabalaenanemertes, and even in these the
testes open partly ventrolaterally. As to the ovaries they are
rather uniformly built, and there is a reduction of the number
of eggs, which grow very large and contain much yolk. The
ovaries are so small that it seems unnatural to derive them from
the large sacs of true Reptantia ; it is more justifiable to com-
pare these gonads with the smaller sacs of Uniporus and other
Hoplonemertini that are reduced to one pair per pseudomere,
or even to less as in Pelagonemertes and other genera. That
the ovaries secondarily migrate into a more central position is
shown by Brinkmann in one case; the young ovaries le
outside the peripheral lateral nerves and when they become
older grow inside and become a tube. This tube may bend
over the nerves to the inside of them; but always the origin
of the sacs seems to be at the outside of the nerves. The
medioventral gonopores of Balaenanemertes, Probalaena-
nemertes, Pelagonemertes, and Armaueria may take their
origin from the inner leg of such forms, though it may just as
well be possible that in Pelagica the more central position of the
nerve-cords as compared with those of other Hoplonemertini
for the first time make the displacement of the gonopores to
the middle line possible.

The development of the testes proceeds in two distinct ways.
Instead of being present in the middle and hinder parts of the
body like the ovaries, they are found only in the anterior
part from the brain to the enteron, and in some cases even

662 GERARDA STIASNY-WIJNHOFF

before the bram. Brinkmann showed in his monograph
that this characteristic seems to be a most important fact in
the propagation of the species, coinciding with the development
of copulatory organs. In this part of the body the pseudomeric
arrangement is not so well developed or is even absent as it is
in the stomodaeal part, and only the entodermic blind-gut can
show metamerically arranged diverticula. In the genera in
which testes are known these are arranged in two ways : (1) with
a tendency to metamerical arrangement (only behind the brain)
in Plotonemertes, Paradinonemertes, Phallonemertes, Chuniella
and Dinonemertes, Biirgeriella. Chuniella, with its long
irregular rows of testes, seems to be the most primitive; the
influence of the diverticula of the blind-gut is seen here as well
as in Burgeriella, where they he more irregularly but are less
in number. Plotonemertes represents the next stage, and the
regularity seems to be perfect im Phallonemertes, Paradino-
nemertes, and perhaps in Dinonemertes alberti.
(2) In the other group the irregularly placed testes show a ten-
dency to discharge the sperm as near to the head as possible
and to form clusters. The effect of this arrangement is shown
by Brinkmann. In Nectonemertes with its tentacles the
testes lie in two irregular rows behind and at the side of the
brain, but long gonoducts have developed (Text-fig. 25) that
all point to the head ; they are still more forward and irregular
in Armaueria ; in Natonemertes a pair of irregular clusters lies
just beneath the brain, and in Para- and Balaenanemertes
the clusters lie at the side of and before the brain and have their
gonopores all directed to the proximal edge of the body.
Pelagonemertes shows the same features as Balaenanemertes.
So the gonads of the Pelagica developed quite differently from
those of all other Nemerteans.

The result of the examination of the anatomical features of
the Pelagica, the Reptantia, and Siboganemertes in these
pages is :

1. That the division of the Enopla into Polystilifera and
Monostilifera is well founded, as not only the differences in the
armature of the proboscis exist, but also the way in which the

NEMERTEA ENOPLA 663

connexion between proboscis pore and mouth can be formed
differs in these sub-orders.

2. That the Polystilifera exhibit the more primitive features,
as in most genera the proboscidian and digestive systems are
quite separate from each other, and the mouth even can be
found underneath the brain as in Siboga- and Paradino-
nemertes.

3. That in the Polystilifera the Pelagica are the more
primitive, though a specialized tribe.

TExtT-FIG. 25.

DS:

rhunch cov

Position of testes and ejaculatory ducts on the ventral side of
Nectonemertes minima (Brinkmann) (4, p. 104, Text-fig. 23).

4. That the absence of cerebral organs, of a highly differen-
tiated brain, of a long much-developed stomodaeum with
oesophagus, of rhynchocoelomic diverticula, of metamerically
arranged vascular loops, are primitive features, and that
reduction did not cause them.

5. That the development of the musculature of the proboscis
and its sheath in all Hoplonemerteans is in perfect accordance
with our knowledge of the anatomy of the Anopla and of their
embryology, and that every stage of this development from
the Anopla stage to the interlacing of Drepanophorus is found
in the different genera of Pelagica.

6. That the blood-vascular system in the different genera of
Pelagica shows the development of the dorsal blood-vessel
from a short blind rhynchocoelomic vessel, that in some species

664 GERARDA STIASNY-WIJNHOFF

still is absent, to a large vessel, that communicates in the tail
with the lateral vessels.

7. That the reproductive system shows quite a number of
characteristic features that are unknown in all other Nemerteans
and cannot have developed from stages known in Monostilifera
and Anopla, nor from those in the Reptantia ; that the influ-
ence of the pelagic habits, which caused reduction of the
number of eggs as these grew larger, and copulation, cannot
account for all these facts, though the ventral position of the
gonopores may be due to it.

Here I might call attention to another fact of importance.
If we look at the figures of the pelagic Nemerteans it must be
evident to everybody, especially on comparison with illustra-
tions that give the anatomy of the whole animals, that the
region we call the head in other Nemerteans, or the precerebral
region, 1s absolutely absent. As already stated above, the
insertion of the proboscis and the muscular septum before the
brain mark the place where originally the mvagination of
the proboscidian system took place. This we see actually in the
pelagic forms ; the rhynchodaeum is only very short and there
is no true head region (Text-fig. 26). Brinkmann states
several times that the rhnynchodaeum may be extremely short,
and only in this way can we understand a dorsal migration of
the proboscis pore that comes to lhe above the brain as in
Armaueria or Parabalaenanemertes. In the primitive genus
Siboganemertes the precerebral region is extremely short too,
and the broad line which demarcates the proximal end of the
animal reminds one of the same feature in the headless Pelagica.
The brain les directly behind the septum, 1. e. quite terminally,
as seen in all the illustrations of Brinkmann. A comparison
of Text-figs. 25 and 26 with Text-figs. 2, 3, 6, and 8, shows
this very clearly. In the armed Nemerteans a displacement
of the mouth goes hand in hand with the development of the
head, and in consequence of this the development of the
stomodaeum and oesophagus. If we understand the head
region of Nemerteans in this way the difference in the structure
of the body-wall before and behind the brain at once becomes

NEMERTEA ENOPLA 665

clear, and the presence of muscle-strands in the snout can be
understood also. The newly developed region could only
get some small outer layer of the musculature as the bulk of
the muscular coat is used by the formation of the proboscidian
system. So we find the greater part of the longitudinal muscula-
ture as the septum before the brain or as proboscidian muscles,

TEXxtT-FIG. 26.

muse. sepe,

Longitudinal section through the anterior region of Necto-
nemertes primitiva (Brinkmann) (4, Pl. 12, fig. 13) to show
the total absence of a head and rhynchodaeum. The proboscis
is lost and the stomodaeum protrudes through the mouth. The
tip of the snout is indicated by a cross.

and only a very thin layer of longitudinal fibres is seen to
accompany the epithelium and the few circular fibres under-
veath. That in this process of division of the musculature some
longitudinal fibres are found in the parenchyma of the snout
that connect the musculature of body-wall, rhynchodaeum,
and septum seems to be quite natural. The aberrant structure
of the head of all Nemerteans as far as concerns the musculature
can only be understood in this way.

666 GERARDA STIASNY-WIJNHOFF

Even in this feature the Pelagica are very primitive im the
absence of a true snout.

With regard to Siboganemertes we have to state the following
facts :

The presence of cerebral organs, rhynchocoelomic diverticula,
an oesophagus, nephridia, metamerically arranged male
gonads brings it in near relation to the Reptantia.

However :

1. The position of the mouth under the brain is more primitive
than in any of these genera.

2. The absence of a snout, present in other Reptantia, recalls
this characteristic of the Pelagica.

3. The proboscis has but two muscular layers, and not three
as in Reptantia and Monostilifera.

4. The rhynchocoelomic diverticula he on the inside of the
entodermal diverticula and never peripherally as in the
other Reptantia.

5. The brain is most primitive, as in Pelagica, without large,
free, dorsal lobes.

6. The cerebral organs are the most primitive we know in
Hoplonemertini, consisting of free mdependent parts,
without a bifurcation of the canal characteristic of the
Drepanophoridae.

The digestive tract has a short bulb-lke oesophagus.

8. The stomodaeal parts are much more highly developed
than in the other Reptantia, displaying the same features
as in certain Amphiporus species, and all parts are distinetly
and sharply separated from each other, as is never the case
in other known Polystilifera.

9. The entodermal blind-gut has unpaired diverticula.

10. The nephridia are different from those of the Reptantia.

11. The lack of metameric blood-vessels is a primitive feature
in common with all Pelagica and Uniporus.

12. The dorsal blood-vessel never lies in the rhynchocoelomic
cavity, a rare feature known in Armaueria in the Pelagica,
in Malacobdella and in some Prostomas in the Monostili-
fera, but unknown in Reptantia.

NEMERTEA ENOPLA 667

18. The testes consist of many small peripheral sacs that open
into a large V-shaped sac as known in the Drepanophoridae
only, and representing probably a more primitive stage
than that of most Drepanophorus species.

Every organ of Siboganemertes is either more primitive than
in the other Reptantia or quite differently developed (rhyncho-
coelomic diverticula, digestive system, nephridia, dorsal
blood-vessel). We must include it in the well-defined group
of Reptantia as given by Brinkmann. On the other hand
we cannot include this genus in his family Drepanophoridae,
nor in the Uniporidae or any other family of the Siboga material.
The real relationship between the known Drepanophoridae
and Siboganemertes we can only indicate by dividing the
tribus Reptantia (Brinkmann) into two subtribus, the Archi-
reptantia and the Eureptantia, of which the first contains the
family Siboganemertidae and the other the different groups of
Drepanophoridae as yet known.

The diagnoses of the different systematic divisions of Knopla
are as follows :

Sub-classis Hnopla (Max Sehultze).

The body-wall consists of a one-layered epithelium, a
basement membrane, a circular muscle-layer, and an inner
longitudinal muscle-coat. The nervous system is embedded
in the body parenchyma. Cerebral organs, where present,
separated from the brain. Proboscidian and digestive system
show a tendency to acquire a common mouth. Blood-
vascular system without lacunae.

Ordo I. Bdellomorpha (Verrill).

Parasitic Nemerteans with a sucker. The probos¢is is in-
serted in the wall of the digestive system ; without armature.
Digestive tract a more or less winding tube without diverti-
cula and blind-gut. Blood-vessels highly branched.

Ordo II. Hoplonemertini (Hubrecht).
Proboscis armed. Digestive system with blind-gut and
paired diverticula ; straight. Vascular system without tree-

668 GERARDA STIASNY-WIJNHOFF

like branching ; as a rule with metamerically arranged loops
between the three longitudinal vessels.
Sub-ordo I. Polystilifera (Brinkmann).

Hoplonemertini with many stylets on a crescent-shaped
base. Proboscis pore and mouth are separate or open
separately in a common atrium. The muscle-coats of the
rhynchocoelomic cavity interlace and become complicated
as a rule.

Sub-ordo IT. Monostilifera (Brinkmann).

Hoplonemertini with one stylet on a handle-shaped base.
The mouth opens into the rhynchodaeum. The rhyncho-
coelomic wall never shows interlacing, and consists of an
inner longitudinal and an outer circular muscle-coat.

The sub-ordo Polystilifera contains the following groups :
Tribus I. Pelagica (Brinkmann).

Pelagic Polystilifera without a distinct snout. Cerebral
organs, nephridia, rhynchocoelomic diverticula, meta-
meric blood-vessels, and oesophagus absent. Testes only
in stomodaeal region. Gonopores ventral.

Tribus IT. Reptantia (Brinkmann).

Polystilfera with cerebral organs, rhynchocoelomic
diverticula, nephridia, and oesophagus, and with meta-
merically situated gonads in the intestinal region.
Sub-tribus I. Archireptantia.

Without a snout. Central rhynchocoelomic diverticula.
Small dorsal ganglia and a primitive cerebral organ.
Different parts of stomodaeum sharply separated and well .
developed. Nephridia with a large, distal efferent duct.
Without metamerical vascular loops.

Sub-tribus II. Eureptantia.

With a snout. Peripheral rhynchocoelomic diverticula.
Cerebral organs with a sac. Large, free dorsal ganglia.
Different parts of stomodaeum continued ito each other.
Nephridia with, as a rule, proximal efferent ducts. With
metamerical vascular loops.

LEYDEN,
March 1, 1923.

—
.

A)

ies)

12.

13.

14.

15.

16.

17.

18.

19.

NEMERTEA ENOPLA 669

LITERATURE.

D. Bergendal.—* Bor ordningen Palaeonemertini Hubrecht uppdelas
itvanne ordningar Protonemertini och Mesonemertini ? ” ‘ Ofversigt
af Kongl. Vetensk. Akad. Forhandl. Stockholm ’, 1900, no. 6.

*“Studien iiber Nemertinen.

Ill. Beobachtungen iiber den

Bau von Carinoma Oudemans nebst Beitriigen zur Systematik der
Nemertinen ”’, “ Lunds Univ. Arsskrift’, Bd. 39, Afd. 2, 1903.

. Brinkmann, A.—‘‘ Uniporus, ein neues Genus der Familie Drepano-
>

phoridae Verrill”’, ‘ Bergens Museums Aarbok’, 1914-15, no. 6.
—— “Die pelagischen Nemertinen”’, ‘ Bergens Museums Skrifter.

Nyrekke’, Bd. 3, no. 1.

Biirger, O.—*‘ Untersuchungen iiber die Anatomie und Histologie der
Nemertinen nebst Beitrigen zur Systematik ’’, ‘ Zeitschr. fiir wiss.

Zool.’, Bd. 50, 1890.

—— “ Die Nemertinen des Golfes von Neapel”’, ‘ Fauna und Flora’,

Bd. 22, 1895.

—— ‘“‘ Nemertini’’, ‘ Das Tierreich ’, Lief. 20, 1904.

—— “ Nemertini”’, ‘ Bronn’s Klassen und Ordnungen des Tierreichs ’,
Bd. 4, Supplement, Leipzig, 1897-1907.

— ‘Die Nemertinen”’, ‘ Wissensch. Ergebnisse der deutschen
Tiefsee-Exped. Valdivia’, Bd. 16, Jena, 1909.

. Hubrecht, A. A. W.—*‘ The genera of European Nemerteans revised ”’,

‘Notes from the Leyden Museum ’, Note 44, 1879.

—— “ Nemertea’”’, ‘ Report Scientif. Res. Challenger ’, vol. 19, 1887.

Joubin, L.—*‘ Recherches sur les Turbellariés des cétes de France
(Némertes) ”, ‘ Arch. de Zool. expér.’, sér. ii, tome 8, 1890.

McIntosh, W. C.—‘ A Monograph of the British Annelids. Part I,
Nemerteans ’, Ray Society, 1873-4.

Punnett, R. C.—‘‘ On some South Pacific Nemertines, collected by
Dr. Willey ’’, ‘ Zool. Results’, Cambridge, 1900, part v.

—— “ Nemerteans ”, ‘ Fauna and Geography, Maldive and Laccadive
Archipelagoes ’, vol. 1, part i, 1901.

Wijnhoff, G.—‘* Die Gattung Cephalothrix, I”’, ‘Zool. Jahrb., Abt.

Anat.’, Bd. 30, 1910.

—— “Die Systematik der Nemertinen’’, ‘ Zool. Anzeiger’, Bd. 40,

E912...

—— ‘The Proboscidian System in Nemertines’’, ‘ Quart. Journ.

Mier. Sci.’, vol. 60, part ii, 1914.

1916.

“ Die Gattung Zygonemertes ”’,

‘Zool. Anzeiger’, Bd. 47, no. 1,

ort Wiatag 2.

‘gt rs : iQ) de a
! al 7

>. oe

INDEX: TO” VOL. 67

NEW SERIES

Agar, W. E. The Male Meiotic
Phase in two Genera of Marsu-
pials (Macropus and Petauroides),
183-202, pls. 12-14.

Agersborg, H. P. K. The Morpho-
logy of the Nudibranchiate Mol-
lusc Melibe (syn. Chioraera
leonina, Gould), 507-592, pls.
27-37.

Amia, hypophysis, de Beer, 257.

Amoeba proteus, nuclear divi-
sion, Taylor, 39.

Amphilina paragonopora,
sp.n., Woodland, 47.

Astraeidae, histology, Matthai, 100.

Aurelia, strobilization, Percival, 85.

Badham, C. On Centropygus
joseensis, a leech from Brazil,
243-256.

Beddard, F. E. Some Observations
upon the Development of the
Teeth of Physeter macro-
cephalus, 1-32.

Beer, G. R. de. Some Observations
on the Hypophysis of Petro-
myzon and of Amia, 257-292.

— see Huxley, 473-496.

Bidder, G. P. The Relation of the
form of a Sponge to its Currents,
293-324.

Bishop, A. Some Observations
upon Spirostomum ambi-
guum, 391-434, pls. 22, 23.

Brazil, leech from, Badham, 243.

Campanularia, resorption and in-
hibition, Huxley, 473.

Caryophyllaeidae from
Woodland, 435.

Sudan,

Centropygus joseensis, a
leech from Brazil, Badham, 243.

Ceratodus; development of muscles,
Edgeworth, 325.

— development of quadrate and

epihyal, Edgeworth, 359.

Cestodaria, revision, Woodland,
435.

Coccidian golgi bodies, King, 381.

Corals ; histology, Matthai, 100.

Da Fano, C. On Golgi’s Internal
Apparatus in spontaneously ab-
sorbing Tumour Cells, 369-380,
pls. 19, 20.

Dedifferentiation, resorption and
inhibition in Obelia and Cam-
panularia, Huxley, 473.

Dogiel, V. A. On Sexual Differen-
tiation in the Infusoria, 219-232,
ple V7:

Edgeworth, F. H. On the Develop-
ment of the Hypobranchial,
Branchial, and Laryngeal Mus-
cles of Ceratodus. With a note
on the Development of the
Quadrate and Epihyal, 325-368.

Egg; Frog, Meek, 33.

— Lepidosiren, Miller, 497.

Embryology and development :

— Ceratodus, muscles, Edge-

worth, 325.
— — quadrate and epihyal, 359.
— Lepidosiren, egg cleavage,
Miller, 497.
— Physeter, teeth, Beddard, 1.
See also under Marsupialia.

672

Flynn, T. Thomson. The Yolk-Sac |

and Allantoic Placenta in Pera-
meles, 123-182, pls. 9-11.

Frog; ‘Segmentation cavity’ of
egg, Meek, 33.

Gatenby, J. B. See King, 381.
Golgi bodies :
of tumour cells, Da Fano, 369.
of a Coccidian, King, 381.
Greenwood, A. W. Marsupial Sper-
matogenesis, 203-218, pls. 15, 16.

Histology ; Corals, Matthai, 100.

Huxley, J. S., and de Beer, G. R.
Studies in Dedifferentiation.
IV. Resorption and Differential
Inhibition in Obelia and Cam-
panularia, 473-496, pl. 26.

Hydra, Marshall, 593.

Hypophysis, Petromyzon,
Amia, de Beer, 257.

Infusoria ;
Dogiel, 219.

—Spirostomum ambiguum,
Bishop, 391.

King, S. D., and Gatenby, J. B. |

The Golgi Bodies of a Coccidian,
381-390, pl. 21. ’

Leech from Brazil, Badham, 243.

Lepidosiren paradoxa, egg
cleavage, Miller, 497.

meiotic phase,

Macropus, male

Agar, 183.

Mammalian spermatozoa, Parkes, |
| Petauroides, male meiotic phase,
Observations upon |

617.
Marshall, S.
the Behaviour and Structure of
Hydra, 593-616.
Marsupialia; embryology
meles, Flynn, 123.
— Chromosomes, Agar, 183.

Pera-

and |

sexual differentiation,

INDEX

Marsupialia ; Spermatogenesis,
Greenwood, 203.

Matthai, G. Histology of the Soft
Parts of Astraeid Corals, 100-
122, pls. 7, 8.

Meek, A. ‘The ‘ Segmentation
Cavity’ of the Egg of the Frog,
33-38, pl. 1.

Melibe, a nudibranchiate Mollusc,
Agersborg, 507.

Miller, A. E. The Cleavage of the
Egg of Lepidosiren para-
doxa, 497-506.

Mollusca, Melibe, Agersborg, 507.

Muscles, Ceratodus development,
Edgeworth, 325.

Nemertea Enopla, Stiasny-
Wijnhoff, 627.

Nuclear division, Amoeba, Taylor,
39.

Obelia, resorption and inhibition,
Huxley, 473.

Parasites :

— Cestodaria, Woodward, 47,
435. ;

— Sanguinicola (? Trematoda),

Woodland, 233.

Coccidia from Lithobius, golgi
bodies, King, 381.

Ophryoscolex in Ungulata,
Dogiel, 219.

Parkes, A. S. Head length Dimor-
phism of Mammalian Spermato-
zoa, 617-626.

Perameles, placenta, Flynn, 123.

| Percival, E. On the Strobilization

of Aurelia, 85-100, pl. 6.

Agar, 183.

Petromyzon, hypophysis, de Beer,
257.

Physeter; development of teeth,
Beddard, 1.

Placenta in Perameles, Flynn, 123.

INDEX

Sanguinicola from
Woodland, 233.
‘Segmentation cavity’, Frog, Meek,

33
Sex in Infusoria, Dogiel, 219.
Siboganemertes weberi,ng.,
n.sp., Stiasny-Wijnhoff, 627.
Spermatogenesis, Marsupial, Green-
wood, 203.

the Sudan,

673

| Taylor, Monica. Nuclear Divisions

Spermatozoa, Mammalian, Parkes, —

617.
Spirostomum, Bishop, 391.

Sponge form in relation to currents, |

Bidder, 293.

Stiasny-Wijnhoff, G. On _ Brink-
man’s System of the Nemertea
Enopla and Siboganemertes
we beri, n.g., n.sp., 627-669.

Strobilization of Aurelia, Percival,
85.

Sudan :

Sanguinicola, Woodland, 233.
Caryophyllaeidae, Woodland, 435.

NO. 268

ey

in Amoeba proteus, 39-46,

pl. 2.
Teeth; Physeter development,
Beddard, 1.

Tumour cells, Da Fano, 369.

Wenyonia, n.g. Cestodaria, Wood-
land, 435.

Woodland, W. N. F. On Am-
philina paragonopora,
sp.n., and a hitherto undescribed
Phase in the Life-history of the
Genus, 47-84, pls. 3-5.

—Sanguinicola from the Sudan,
233--242, pl. 18.

— On some remarkable new Forms
of Caryophyllaeidae from the
Anglo-Egyptian Sudan, and a
Revision of the Families of the
Cestodaria, 435-472, pls. 24,
25.

New Series. No. 265.] APRIL, 1923. [Volume 67. Part I.

THE

QUARTERLY JOURNAL

MICROSCOPICAL SCIENCE.

HONORARY EDITOR:

siz RAY LANKESTER, K.C.B., M.A., D.Sc., LL.D., F.B.S.

EDITOR ?

EDWIN S. GOODRICH, M.A., F.RS.,

LINACRE PROFESSOR OF ZOCGLOGY AND COMPARATIVE ANATOMY IN THE UNIVERSILY OF OXFORD ;

WITH THE CO-GPERATION OF
SYDNEY J. HICKSON, M.A., F.RBS.,
BEYER PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF MANCHESTER ;
GILBERT C. BOURNE, M.A., D.Sc, FRS.;
J. GRAHAM KERR, M.A., F.B.S.,

REGIUS PROFESSOR OF ZO@LOGY IN THE UNIVERSITY OF GLASGOW ;

E. W. MACBRIDE, M.A., D.Sc., LL.D., F.R.S.,

PROFESSOR OF ZOOLOGY AT THE IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY ;

G. P. BIDDER, M.A., Sc.D.

WITH LITHOGRAPHIC PLATES AND TEXT-FIGURES.

OXFORD UNIVERSITY PRESS,
HUMPHREY MILFORD, AMEN CORNER, LONDON, E.C. 4.

100 brinces Street, EpInBuRGH. Cathedral Buildings, MELBOURNE.
104 West George Street, GLASGow. Markham’s Buildings, Carr Town.
St. Kongensgade 40 H, CoPENHAGEN. . 17-19 Elphinstone Circle, BomBay.
35 West 32nd Street, New York. Temple Chambers, CALCUTTA.
25-27 Richmond Street West, Toronto. 10 Esplanade, George Town, Mavras.
C 445 Honan Road, SHANGHAI.

Subscription price per Volume, £4 4s. net. Single parts, £1 is. net.

CONTENTS OF No. 265.—New Series.

Volume 67. Part I.

MEMOIRS: ; PAGE

Some Observations upon the Development of the Teeth of Physeter
macrocephalus. By Frank EK. Bepparp, D.Sc., M.A., F.R.S.

(With 13 Text-figures) . : ; : é 3 1
The ‘Segmentation Cavity’ of the Egg of the Frog. By ALEXANDER

MEEK. (With Plate 1 and 1 Text-figure) ? : : ° . 933
Nuclear Divisions in Amoeba proteus. By Monica TAYLor,

S.N.D. (With Plate 2) . 2 . , : : ; : = isa ee

On Amphilina paragonopora, sp.n., and a hitherto unde-
scribed Phase in the Life-history of the Genus. By W. N. F.
WooDLAND, D.Sc. (London). (With Plates 3-5) . ; ; “ E

On the Strobilization of Aurelia. By E. PErctvat, B.Sc., Department
of Zoology, The University, Leeds. (With Plate 6 and 3 Text-

figures) . : : ; : ; : ‘ : 5 oc] oe
Histology of the Soft Parts of Astraeid Corals. By G. MATTHAT, M.A.
(With Plates 7 and 8) : : : : : meas * . 101

The Yolk-Sac and Allantoic Placenta in Perameles. By T. THomson
Fiywn, D.Sc., Ralston Professor of Biology, University of Tasmania.
(With Plates 9-11 and 4 Text-figures) . : : : : . 123

Mr. Milford, Oxford University Press, Amen Corner, London, E.C. 4, will be glad to

receive offers for back numbers of the Quarterly Jowrnal of Microseopical Sezence.

H. K. LEWIS & CO. Ln, TECHNICAL BOOKSELLERS.

Text-books and Works in General, Technical, and Applied Science of all Publishers:

' Orders by Post or Telephone promptly attended to.

Large Stock of SECOND-HAND Books at greatly reduced prices (wusebar|
at 140 Gower Street. Catalogue Post Free on Application, t 4031.

SCIENTIFIC AND TECHNICAL CIRCULATING LIBRARY.

Annual Subscription, Town or Country, from One Guinea.
LIBRARY, READING anp WRITING ROOM (First Froor) OPEN DAILY TO SUBSCRIBERS.

136 GOWER STREET and 24 GOWER PLACE, LONDON, W.C. 1.
Telegrams: ‘PUBLICAVIT, EUSROAD, LONDON.’ Telephone : MUSEUM, 1072.

THE MARINE BIOLOGICAL ASSOCIATION

OF THE

UNITED KINGDOM.

Patron—HIS MAJESTY THE KING.
President—Sir RAY LANKESTER, K.C.B., LL.D., F.R.S.

io

Tur AssocrIATION WAS FOUNDED ‘TO ESTABLISH AND MAINTAIN LABORATORIES ON
THE COAST OF THE Unirep Kinapom, WHERE ACCURATE RESEARCHES MAY BE CARRIED ON,
LEADING TO THE IMPROVEMENT OF ZOOLOGICAL AND BoTanicaAL SCIENCE, AND TO AN
INCREASE OF OUR KNOWLEDGE AS REGARDS THE FOOD, LIFE CONDITIONS, AND HABITS
or BRITISH FOOD-FISHES AND MOLLUSCS’.

The Laboratory at Plymouth

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

Naturalists desiring to work at the Laboratory

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

Works published by the Association

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

The Journal of the Marine Biological Association

is issued half-yearly, 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.

Awnnuat MemBeErs . “ 4 - £1 1 O per annum.
Lire MemBerRs: . : i é 15 15 0 Composition Fee.
FounDERS : 100 0 0 *

99

GOVERNORS (Life Members of Council) 500 0 0

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.

Quarterly Journal of Microscopical
Science.

The SUBSCRIPTION for Vol. 67 is 84s. for the Four Numbers;
for this sum (prepaid) the JouRNAL is sent Post Free to any part
of the world. Separate Numbers are sold at 21s. net each.

BACK NUMBERS of the JourNAL which remain in print, up
to and including Vol. 64, are sold at 15s. net each.

OXFORD AT THE CLARENDON PRESS
LONDON HUMPHREY MILFORD
OXFORD UNIVERSITY PRESS, AMEN CORNER, E.C.4

NOTICE TO CONTRIBUTORS.

Contributions to the JouRNAL should be addressed to the Editor,
Professor E, 8. GoopricH, University Museum, Oxford.

Contributors are requested to send typewritten manuscripts,
and to conclude each paper with a short summary and the list of
works to which reference is made. The position of figures in the
text should be indicated, and words to be printed in spaced type
should be underlined. Manuscripts should be sent in fully corrected;
an allowance of ten shillings per sheet of sixteen pages is made for
alterations in the proof, contributors being responsible for any
EXCESS.

Monochrome lithographic plates will be provided where neces-
sary; but further colours only in exceptional cases and after
consultation with the Editor.

‘Line and dot’ drawings in black process ink should be used
whenever possible, instead of those requiring lithography or half-
tone. The best results are obtained by drawing the Agu on
a scale to be afterwards reduced.

The size of a text-page is 61x33 inches, of a single plate
725 inches, and of a double plate 73x11 inches. The placing
of figures across the fold is to be avoided.

The lettering of figures should be inserted in pencil or written
on accurately imposed outlines on tracing-paper. ;

Authors receive 50 copies of their contributions gratis, and
may buy additional copies, up to 100, if they apply to the Editor
when they return the corrected proofs.

~ Printed in England at the Oxford University Press by Frederick Hall.

New Series. No. 266.] JULY, 1923. [Volume 67. Part II.
THE

QUARTERLY JOURNAL
MICROSCOPICAL SCIENCE.

HONORARY EDITOR:

Str RAY LANKESTER, K.C.B., M.A., D.Sc., LL.D., F.B.S.

EDITOR ?

EDWIN S. GOODRICH, M.A., F.BS.,
LINACRE PROFESSOR OF ZOOLOGY AND COMPARATIVE ANATOMY IN THE UNIVERSITY OF OXFORD;
WITH THE CO-OPERATION OF

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

BEYER PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF MANCHESTER;

GILBERT C. BOURNE, M.A., D.Sc, F.RS.; ,
J. GRAHAM KERR, M.A., F.RB.S.,

REGIUS PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF GLASGOW ;

E. W. MACBRIDE, M.A., D.Sc. LL.D., F.B.S,

PROFESSOR OF ZOOLOGY AT THE IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY .

G. P. BIDDER, M.A., Sc.D.

WITH LITHOGRAPHIC PLATES AND TEXT-FIGURES.

\

OXFORD UNIVERSITY PRESS,
HUMPHREY MILFORD, AMEN CORNER, LONDON, E.C. 4.

100 Princes Street, EDINBURGH. Cathedral Buildings, MELBOURNE.
104 West George Street, GLASGOW. Markham’s Buildings, CarE Town.
St. Kongensgade 40 H, CopENHAGEN. 17-19 Elphinstone Circle, BomBay.
35 West 32nd Street, New York. 1 Garstin Place, CALCUTTA.

25-27 Richmond Street West, Toronto, Anjunan Buildings, Mount Road,Mapras.
C 445 Honan Road, SHANGHAI.

Subscription price per Volume, £4 4s. net. Single parts, £1 1s. net.

CONTENTS OF No. 266.—New Series.

Volume 67. Part II.

MEMOIRS: PAGE
The Male Meiotic Phase in two Genera of Marsupials (Macropus and
Petauroides). By W. HE. Acar, F.R.S., Professor of Zoology in the

University of Melbourne. (With Plates 12-14). : : . 183

Marsupial Spermatogenesis. By A. W. GREENWOOD, B.Sc., University
of Melbourne. (With Plates15and16). . ; : ; . 2038

On Sexual Differentiation in the Infusoria. By Prof. V. A. DoarEt,
Zootomical Laboratory, University of Petrograd. (With Plate 17
and 1 Text-figure) . : i : : : : 4 ‘ . 2

Sanguinicola from the Sudan. By W.N. F. WoopLAnD, Wellcome
Bureau of Scientific Research. (With Plate 18) . : , . 233

On Centropygus joseensis, a Leech from Brazil. By CHARLES
BapuAam, B.Sc., M.B., Ch.M. (With 10 Text-figures) . : . 248

Some Observations on the Hypophysis of Petromyzon and of Amia.
By G. R. pE Beer, B.A., Bc.C. (With 34 Text-figures) . ‘ . 257

The Relation of the form of a Sponge to its Currents. By G, P.
BIDDER, Sc.D. (With 12 Text-figures) . aye Es at 5B . 293

Mr. Milford, Oxford University Press, Amen Corner, London, E.C’. 4, will be glad to
receive offers for back numbers of the Quarterly Journal of Microscopical Science.

SCIENTIFIC AND. . -
H. K. LEWIS & CO. LtD., TECHNICAL BOOKSELLERS.
Text-books and Works in General, Technical, and Applied Science of all Publishers.

Orders by Post or Telephone promptly attended to.

Large Stock of SECOND-HAND Books at greatly reduced prices
at 1440 Gower Street. Catalogue Post Free on Application.
Phone: MUSEUM, 4031.

SCIENTIFIC & TECHNICAL CIRCULATING LIBRARY
Annual Subscription, Town or Country, from One Guinea.
LIBRARY, READING anp WRITING ROOM (First Fitoor) OPEN DAILY TO SUBSCRIBERS.

136 GOWER STREET and 24 GOWER PLACE, LONDON, W.C.1.
Telegrams: ‘PUBLICAVIT, EUSROAD, LONDON.’ Telephone : MUSEUM, 1072

THE MARINE BIOLOGICAL ASSOCIATION

OF THE

UNITED KINGDOM.

Patron—HIS MAJESTY THE KING.
President—Sir RAY LANKESTER, K.C.B., LL.D., F.R.S.

Tue ASSOCIATION WAS FOUNDED ‘TO ESTABLISH AND MAINTAIN LABORATORIES ON
THE COAST OF THE Unitep Kin@pom, 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., 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, 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,

poth 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 . 3 : - £1 1 O per annum.
Lire MEMBERS : - 2 : 15 15 0 Composition Fee.
FounDERS - 100 0 0

? 9

GOVERNORS (Life Members of Council) 500 0 0

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

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

The DIRECTOR,
The Laboratory,
Plymouth.

eS E—EEEEaar a anna

Quarterly Journal of Microscopical
Science.

The SUBSCRIPTION for Vol. 67 is 84s. for the Four Numbers;
for this sum (prepaid) the JOURNAL is sent Post Free to any part
of the world. Separate Numbers are sold at 21s. net each.

BACK NUMBERS of the Journat which remain in print, up
to and including Vol. 64, are sold at 15s. net each.

OXFORD AT THE CLARENDON PRESS
LONDON HUMPHREY MILFORD
OXFORD UNIVERSITY PRESS, AMEN CORNER, E.C.4

NOTICE TO CONTRIBUTORS.

Contributions to the JOURNAL should be addressed to the Editor,
Professor E. 8. GoopricH, University Museum, Oxford.

Contributors are requested to send typewritten manuscripts,
and to conclude each paper with a short summary and the list of
works to which reference is made. The position of figures in the
text should be indicated, and words to be printed in spaced type
should be underlined. Manuscripts should be sent in fully corrected ;
an allowance of ten shillings per sheet of sixteen pages is made for
alterations in the proof, contributors being responsible for any
eXcess.

Monochrome lithographic plates will be provided where neces-
sary; but further colours only in exceptional cases and after
consultation with the Editor.

‘Line and dot’ drawings in black process ink should be used
whenever possible, instead of those requiring lithography or half-
tone. The best results are obtained by drawing the figures on
a scale to be afterwards reduced.

ee ae Se ea ee ee ee oe

The size of a text-page is 64x32 inches, of a single plate
7% x5 inches, and of a double plate 73x11 inches. The placing
of figures across the fold is to be avoided.

The lettering of figures should be inserted in pencil or written
on accurately imposed outlines on tracing-paper.

Authors receive 50 copies of their contributions gratis, and
may buy additional copies, up to 100, if they apply to ge Editor
when they return the corrected proofs.

A sabe

Printed in England at the Oxford University Press by Frederick Hall.

ee

New Series. No. 267.] OCTOBER, 1923. [Volume 67. Part III.
THE

QUARTERLY JOURNAL
MICROSCOPICAL SCIENCE.

HONORARY EDITOR ¢

Str RAY LANKESTER, K.C.B., M.A., D.Sc., LL.D., F.BS.

EDITOR :

EDWIN S. GOODRICH, M.A., F.RBS.,
LINACRE PROFESSOR OF ZOOLOGY AND COMPARATIVE ANATOMY IN THE UNIVERSITY OF OXFORD;
WITH THE CO-OPERATION OF

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

BEYER PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF MANCHESTER;

GILBERT ©. BOURNE, M.A., D.Sc, F.RS.;
J. GRAHAM KERR, M.A., F.BS.,

REGIUS PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF GLASGOW ;

E. W. MACBRIDE, M.A., D.Sc., LL.D., F.B.S.,

PROFESSOR GF ZOOLOGY AT THE IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY ;

G. P. BIDDER, M.A., Sc.D.

WITH LITHOGRAPHIC PLATES AND TEXT-FIGURES.

OXFORD UNIVERSITY PRESS,
HUMPHREY MILFORD, AMEN CORNER, LONDON, E.C. 4.

100 Princes Street, EDINBURGH. Cathedral Buildings, MELBOURNE.
104 West George Street, GLASGOW. Markham’s Buildings, CAPE Town.
St. Kongensgade 40 H, CopENHAGEN. 17-19 Elphinstone Circle, BoMBAY.
35 West 32nd Street, New York. 1 Garstin Place, CALCUTTA.

25-27 Richmond Street West, Toronto, Anjunan Buildings, Mount Road, MADRAS.

C 445 Honan Road, SHANGHAI.

\

Subscription price per Volume, £4 4s. net. Single parts, £1 1s.

net

CONTENTS OF No. 267.—New Series.

Volume 67. Part III.

MEMOIRS: PAGE
On the Development of the Hypobranchial, Branchial, and Laryngeal
Muscles of Ceratodus. With a Note on the Development of the
Quadrate and asec By F. H. Epgzwortu, M.D. (With 39
Text-figures) . : ‘ : : ‘ en 3 . 825 .

On Golgi’s Internal Apparatus in spontaneously absorbing Tumour
Cells. By C. Da Fano, Reader in Histology, King’s College, Uni-
versity of London. (With Plates 19 and 20) . ; : F . 369

The Golgi. Bodies of a Coccidian. BySuana D. Kine and J. BRontTEé
GATENBY, School of Zoology, Dublin University. (With Plate 21) 381

Some Observations upon Spirostomum ambiguum (Ehren-
berg). By Ann BisHop, M.Se., Victoria University, Manchester. ‘
(With Plates 22 and 23 and 9 Text-figures) , : i 5 . Ook

On some remarkable new Forms of Caryophyllaeidae from the Anglo-
Egyptian Sudan, and a Revision of the Families of the Cestodaria.
By W. N. F. WoopLanD, Wellcome Bureau of Scientific Research,
London. (With Plates 24 and 25 and 1 Text-figure) . ; .” 435

Studies in Dedifferentiation. IV. Resorption and Differential Inhi-
bition in Obelia and Campanularia. By J.S. Huxuey, M.A.,, and
G. R. pE Beer, B.A., B.Sc. (With Plate 26 and 7 Text-figures) . 473

Mr. Milford, Oxford University Press, Amen Corner, London, E.C. 4, will be glad to
receive offers for back numbers of the Quarterly Journal of Microscopical Science.

H. K. LEWIS & CO. Lrp., Tectnicat sooxsettens.

Text-books and Works in General, Technical, and Applied Science of all Publishers.

Orders by Post or Telephone promptly attended to.

Large Stock of SECOND-HAND Books at greatly reduced prices
at 140 Gower Street. Catalogue Post Free on Application.
Phone: MUSEUM, 4031.

SCIENTIFIC & TECHNICAL CIRCULATING LIBRARY
Annual Subscription, Town or Country, from One Guinea.
LIBRARY READING anp WRITING ROOM (First Fioor) OPEN DAILY TO SUBSCRIBERS.

136 GOWER STREET and 24 GOWER PLACE, LONDON, W.C.1.
Telegrams: ‘PUBLICAVIT, EUSROAD, LONDON.’ Telephone : MUSEUM, 1072.

— rt

THE MARINE BIOLOGICAL ASSOCIATION

OF THE

UNITED KINGDOM.

Patron—HIS MAJESTY THE KING.
President—Sir RAY LANKESTER, K.C.B., LL.D., F.R.S.

——

Tur ASSOCIATION WAS FOUNDED ‘TO ESTABLISH AND MAINTAIN LABORATORIES ON
THE COAST OF THE Unitep Kin@pom, 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
or 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., 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, 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 . é - : £1 1 O per annum.
Lire MEMBERS 5 - : - 15 15 0 Composition Fee.
FounDERS - 10008 0

2 9

Governors (Life Members of Council) 500 0 0

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.

aera

Quarterly Journal of Microscopical
Science.

The SUBSCRIPTION for Vol. 67 is 84s. for the Four Numbers;
for this sum (prepaid) the JOURNAL is sent Post Free to any part
of the world. Separate Numbers are sold at 21s. net each.

BACK NUMBERS of the Journat which remain in print, up
to and including Vol. 64, are sold at 15s. net each.

OXFORD AT THE CLARENDON PRESS
LONDON HUMPHREY MILFORD
OXFORD - UNIVERSITY PRESS, AMEN CORNER, E.C.4

NOTICE TO CONTRIBUTORS.

Contributions to the JouRNAL should be addressed to the Editor,
Professor E. S. Goopricu, University Museum, Oxford.
Contributors are requested to send typewritten manuscripts,

and to conclude each paper with a short summary and the list of

works to which reference is made. The position of figures in the
text should be indicated, and words to be printed in spaced type
should be underlined. Manuscripts should be sent in fully corrected;
an allowance of ten shillings per sheet of sixteen pages is made for
alterations in the proof, contributors being responsible for any
eXcess.

Monochrome lithographic plates will be provided where neces-
sary; but further colours only in exceptional cases and after
consultation with the Editor.

‘Line and dot’ drawings in black process ink should be used
whenever possible, instead of those requiring lithography or half-
tone. The best results are obtained by drawing the figures on
a scale to be afterwards reduced.

The size of a text-page is 64x33 inches, of a single plate
72x45 inches, and of a double plate 73x11 inches. The placing
of figures across the fold is to be avoided.

The lettering of figures should be inserted in pencil or written
on accurately imposed outlines on tracing-paper.

Authors receive 50 copies of their contributions gratis, and
may buy additional copies, up to 100, if they apply to the Editor
when they return the corrected proofs.

Printed in England at the Oxford University Press by Frederick Hall.

ee ee

New Series. No. 268.] DECEMBER, 1923. [Volume 67. Part IV.

THE

QUARTERLY JOURNAL
MICROSCOPICAL SCIENCE.

HONORARY EDITOR :

Str RAY LANKESTER, K.C.B., M.A., D.

EDITOR :

EDWIN 8. GOODRICH, M.A.,

LINACRE PROFESSOR OF ZOOLOGY AND COMPARATIVE ANATOMY IN THE UNIVERSITY OF OXFORD;

WITH THE CO-OPERATION OF

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

BEYER PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF MANCHESTER;

GILBERT C. BOURNE, M.A., D.
J. GRAHAM KERR, M.A., F.RB.S.,

REGIUS PROFESSOR OF ZOOLOGY IN THE UNIVERSITY

E, W. MACBRIDE, M.A., D.Sc., LL.D., F.B.S.,

PROFESSOR OF ZOOLOGY AT THE IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY ,

G. P. BIDDER, M.A., Sc.D.

WITH LITHOGRAPHIC PLATES AND TEXT-FIGURES.

OXFORD UNIVERSITY PRESS,
HUMPHREY MILFORD, LONDON, E.C. 4.
100 Princes Street, EDINBURGH. Cathedral Buildings, MELBOURNE.
104 West George Street, GLASGOW. Markham’s Buildings, Carr Town.
St. Kongensgade 40 H, CopENHAGEN. 17-19 Elphinstone Circle, BomBay.
35 West 32nd Street, New York. 1 Garstin Place, CALCUTTA.

25-27 Richmond Street West, Toronto. Anjunan Buildings, Mount Road, MApRas.

C 445 Honan Road, SHANGHAI.

Se).7-L, D5, Fie

F.RB.S.,

Sc, F.R.S.;

OF GLASGOW ;

Subscription price per Volume, £4 4s. net.

Single parts, £1 1s. net

CONTENTS OF No. 268.—New Series.

MEMOIRS: PAGE
The Cleavage of the Egg of Lepidosiren paradoxa. By Acnes E.

MriuterR, M.A., Department of Zoology, University of Glasgow.
(With 12 Text-figures) . ; . : . g ; ; ‘497

The Morphology of the Nudibranchiate Mollusc Melibe (syn. Chioraera
leonina Gould. By H. P. Kserscoow Acerssorg, B.S., M.S., M.A,
Ph.D., Williams College, Williamstown, Massachusetts, (With Plates
27 to 37) . 5 5 , . ; ; : ; ; ‘ . 507

Observations upon the Behaviour and Structure of Hydra, By SHEINA
MARSHALL, B.Sc., Assistant Naturalist, Scottish Marine Biological
Station, Millport. (With 4 Text-figures) . ; 5 - : . 593

Head Length Dimorphism of Mammalian Spermatozoa. By A. S. PARKEs,
B.A. (Cantab.), Department of Zoology, University of Manchester.
(With 3Text-figures) . : ; : : 3 : : . 617

On Brinkmann’s System of the Nemertea Enopla and Siboganemertes

Weberi, n.g.n.sp. By Dr. GERARDA STIASNY-WisNHOFF. (With
26 Text-figures) - , 5 5 : : : : : . 627

Mr. Milford, Oxford University Press, London, E.C. 4, will be glad to receive offers
for back numbers of the Quarterly Journal of Microseopical Science.

°

H. K. LEWIS & CO. Ltp., technicat sooxseuuens.

Text-books and Works in General, Technical, and Applied Science of all Publishers.

Orders by Post or Telephone promptly attended to.

Large Stock of SECOND-HAND Books at greatly reduced prices
at 1440 Gower Street. Catalogue Post Free on Application.
Telephone: MUSEUM, 4031.

SCIENTIFIC & TECHNICAL CIRCULATING LIBRARY
Annual Subscription, Town or Country, from One Guinea.
er a ee

LIBRARY, READING anp WRITING ROOM (First Froor) OPEN DAILY TO SUBSCRIBERS.

136 GOWER STREET and 24 GOWER PLACE, LONDON, W.C.1.
Telegrams: ‘PUBLICAVIT, EUSROAD, LONDON.’ Telephone : MUSEUM, 1072.

THE MARINE BIOLOGICAL ASSOCIATION

OF THE

UNITED KINGDOM.

Patron—HIS MAJESTY THE KING.
President—Sir RAY LANKESTER, K.C.B., LL.D., F.R.S.

——_——-—__ —_——

Tur ASSOCIATION WAS FOUNDED ‘TO ESTABLISH AND MAINTAIN LABORATORIES ON
THE COAST OF THE Unirep Kin@DOM, 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., 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, 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 . 5 : 5 £1 1 O per annum.
Lire MemBers : : E : 15 15 0 Composition Fee.
FounDERS ; LOO 10" >-0 + 7

GovERNoRs (Life Members of Council) 500 0 0

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,
3 The Laboratory,
. Plymouth.

Quarterly Journal of Microscopical
Science.

The SUBSCRIPTION for Vol. 67 is 84s. for the Four Numbers;
for this sum (prepaid) the JOURNAL is sent Post Free to any part
of the world. Separate Numbers are sold at 21s. net each.

BACK NUMBERS of the JourNnat which remain in print, up
to and including Vol. 64, are sold at 15s. net each.

OXFORD AT THE CLARENDON PRESS
LONDON HUMPHREY MILFORD
OXFORD UNIVERSITY PRESS, E.C. 4

NOTICE TO CONTRIBUTORS.

Contributions to the JouRNAL should be addressed to the Editor,
Professor E. S. Goopricu, University Museum, Oxford.

Contributors are requested to send typewritten manuscripts,
and to conclude each paper with a short summary and the list of
works to which reference is made. The position of figures in the
text should be indicated, and words to be printed in spaced type
should be underlined. Manuscripts should be sent in fully corrected;
an allowance of ten shillings per sheet of sixteen pages is made for
alterations in the proof, contributors being responsible for any
eXcess.

Monochrome lithographic plates will be provided where neces-
sary; but further colours only in exceptional cases and after
consultation with the Editor.

‘Line and dot’ drawings in black process ink should be used
whenever possible, instead of those requiring lithography or half-
tone. The best results are obtained by drawing the figures on
a scale to be afterwards reduced.

The size of a text-page is 64x33 inches, of a single plate
73 x5 inches, and of a double plate 73x11 inches. The placing
of figures across the fold is to be avoided.

The lettering of figures should be inserted in pencil or written ~
on accurately imposed outlines on tracing-paper.

Authors receive 50 copies of their contributions gratis, and
may buy additional copies, up to 100, if they apply to the Editor
when they return the corrected proofs.

Printed in England at the Oxford University Press by Frederick Hall.

<=

;
A
:
‘
2
t
3