ees ew paw

SO See

ee ee ee cae ae
pe Men = ‘
omnes —= sea te iat lin = a ee ee
ss a oe a a

SIO en ee Te be De Soe “ Des
> > Ne et ae eles Raia ia

caduantaiabede inteiieenndndidenemnmadeeee eee

HARVARD UNIVERSITY.

LIBRARY

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY.

—
Bought
ihe IL 900 Soptim ber [9o4

it 7 ‘
oh ‘|
1 uy

QUARTERLY JOURNAL

MICRKOSCOPICAL SCIENCE.

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

HONOKARKY FELLOW OF EXETER COLLEGE, OXFORD; CORRESPONDENT OF THE INSTITUTE OF FRANCE
AND OF THE IMPERIAL ACADEMY OF SCIENCES OF ST. PETERSBURG, AND OF THE ACADEMY
OF SCIENCES OF PHILADELPHIA; FOREIGN MEMBER OF THE ROYAL BOHEMIAN
SOCIETY OF SCIENCES, AND OF THE ACADEMY OF THE LINCEI OF ROME;
ASSOCIATE OF THE ROYAL ACADEMY OF BELGIUM; HONORARY MEMBER
OF THE NEW YORK ACADEMY OF SCIENCES, AND OF THE
CAMBRIDGE PHILOSOPHICAL SOCIETY, AND OF THE ROYAL
PHYSICAL SOCIETY OF EDINBURGH } HONORARY
MEMBER OF THE BIOLOGICAL SOCIETY
OF PARIS;
DIRECTOR OF THE NATURAL HISTORY DEPARTMENTS OF THE BRITISIL MUSEUM; LATE FULLERIAN
PROFESSOR OF PHYSIOLOGY IN THE ROYAL INSTITUTION OF GREAT BRETAIN.

WITH THE CO-OPERATION OF

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

FELLOW AND TUTOR OF TRINITY COLLEGE, CAMBRIDGE 5

W. F. R. WELDON, M.A., F.RB.S.,

LINACKE PROFESSOR OF COMPARATIVE ANATOMY AND FELLOW OF MERTON COLLEGE, OXFORD}
LATE FELLOW OF 8T. JOHN'S COLLEGE, CAMBRIDGE 5

AND

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

BEYER PROFESSOR OF ZOOLOGY IN THE OWENS COLLEGE, MANCHESTER.

VOLUME 44.—New Sertes.
With Rithographic Plates and Engrabings on Wood

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

CONTENTS.

CONTENTS OF No. 173, N.S., NOVEMBER, 1900.

MEMOIRS:
The “Sexual Season” of Mammals and the Relation of the ‘‘ Pro-
cestrum” to Menstruation. By Watrer Hearse, M.A., Trinity
College, Cambridge

A Description of Ephydatia blembingia, with an Account of
the Formation and Structure of the Gemmule. By Ricnuarp
Evans, B.A. (With Plates 1—4) .

On a Collection of Nemerteans from Singapore. By R. C.
- Punnett, B.A. (With Plates’5—8) :

On the Protostigmata of Molgula manhattensis (De Kay).
By ArruuR Wittry. (With Plate 9) :

CONTENTS OF No. 174, N.S., MARCH, 1901.

MEMOIRS::

The Development and Succession of Teeth in Hatteria pune-
tata. By H. Spencer Harrison, B.Se.(Lond.), A.R.C.Sc.,
Demonstrator and Assistant Lecturer in Biology, University
College, Cardiff. (From the Zoological Laboratory, Royal Col-
lege of Science, London, and University College, Cardiff.)
(With Plates’10—12)

The Anatomy of Pleurotomaria Beyrichii, Hilg. By Martin
F. Woopwarp, Demonstrator of Zoology, Royal College of
Science, London. (With Plates 13—16)

PAGE

71

1a!

14]

161

215

Dolichorhynehus indicus, n. g., n. sp.. a New Acraniate.

By AxtHuR WILLEY

269

Heteropleuron Hectori, the New Zealand Lancelet. By W.

Biaxianp Brnnam, D.Se., M.A., F.Z.S., Professor of Biology

in the University of Otago. (With Plate 17)

273

lv CONTENTS.

On some Parasites found in Echinus esculentus, L. By
ARTHUR E. Surptey, M.A., Fellow and Tutor of Christ’s Col-
lege, Cambridge, and Lecturer in the Advanced Morphology of
the Invertebrata in the University. (With Plate 18)

The Scottish Silurian Scorpion. By R. I. Pocock. (With Plate
49) : : : :

CONTENTS OF No. 175, N.S., MAY, 1901.

MEMOIRS:
The Life-History of Nucula delphinodonta (Mighels). By

Gitman A. Drew, Professor of Biology, University of Maine,
Orono, Me. (With Plates 20—95)

On the Structure of the Hairs of Mylodon Listai and other
South American Edentata. By W.G. Ripewoop, D.Sc, F.L.S.,
Lecturer on Biology at the Medical School of St. Mary’s Hos-
pital, London, (With Plate 26)

On the Structure and Affinities of Saccocirrus. By Kpwin 8.
Goopricu, M.A., Fellow of Merton College, Oxford. (With
Plates 27—29) ‘ B :

On the Question of Priority with regard to certain Discoveries
upon the Adtiology of Malarial Diseases. By GrorcE H. F.
Nurraut, M.A., M.D., Ph.D., University Lecturer in Bacteri-
ology and Preventive Medicine, Cambridge

Studies in the Retina: Rods and Cones in the Frog and in some
other Amphibia. By H. M. Bernarp, M.A.Cantab. (from the
Biological Laboratories, Royal College of Science, London).
Part LI. (With Plates’30 and’31)

Staining with Brazilin. By Sypney J. Hickson, Beyer Professor
of Zoology in the Owens College, Manchester

CONTENTS OF No. 176, N.S., AUGUST, 1901.

MEMOIRS:
On Two New Species of Onychophora from the Siamese Malay

States. By Ricuarp Jivans, M.A., B.Se., of Jesus College,
Oxford. (With Plates 327) . : j ;

PAGE

281

291

313

393

413

429

448

469

473

CONTENTS.

Eoperipatus Butleri (nov. sp.). By RicHarp Evans, M.A.,
B.Sc., of Jesus College, Oxford. (With Plate’38) : - 539

On Two New British Nemerteans. By R. C.
(With Plates’39 and 40)

Punnett, B.A.

The Celomic Fluid in Acanthodrilids. By W. Biaxtanp BenuamM,

D.Sc., M.A., F.Z.S., Professor of Biology in the University of

Otago, New Zealand. (With Plate’41) .

565

The Crystalline Style of Lamellibranchia. By S. B. Mirra, of

Calcutta, late of University College, London.

TrrLE, INDEX, AND CoNTENTSs.

VoL. 44.—NEW SERIES.

(With Plate42) 591

ee eo yee
Ley, apes ine a
7 - P
nd 7* VR aah

a oe %
4 {=e a A ‘eo ae em of vA "er tee bee Sa 4 7
¥i a ¢ bl Ty 2. oe

athens

ct. (Cae sally Hh a ‘ i
2 | aa ies a

{ ay ar ay ¥

:

ime tes

‘ if he " 7 iy ; oi fi
Ppl} BNA, a 6 suri :

} say dues

if

|
>
‘Hi vhs
Pgs

New Series, No. 173 (Vol. 44, Part 1). Price 10s.
eee NOVEMBER, 1900. |
NOV 22/ 1900 ee

QUARTERLY JOURNAL

OF
4

MICROSCOPICAL SCIENCE. |

EDITED BY

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

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

DIRECTOR OF THE NATURAL HISTORY DEPARTMENTS OF THE BRITISH MUSEUM} FULLERIAN PROFESSOR

OF PHYSIOLOGY IN THE ROYAL INSTITUTION OF GREAT BRITAIN.

WITH THE CO-OPERATION OF

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

FELLOW AND TUTOR OF TRINITY COLLEGE, CAMBRIDGE 3

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

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

AND

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

BEYER PROFESSOR OF ZOOLOGY IN THE OWENS COLLEGE, MANCHESTER

WITH LITHOGRAPHIC PLATES AND ENGRAVINGS ON WOOD.

LONDON: |
J. & A. CHURCHILL, 7, GREAT MARLBOROUGH STREET
1900.

| ACE Pe ee eS ee iE mee SET oT
Adlard and Son,] [Bartholomew Close

CONTENTS OF No. 173.—New Series.

MEMOIRS :

The “Sexual Season’? of Mammals and the Relation of the ‘“ Pro-
cestrum”? to Menstruation. By Water Heaprz, M.A., Trinity
College, Cambridge ;

A Description of Ephydatia blembingia, with an Account of the
Formation and Structure of the Gemmule. By Ricnuarp vans,
B.A. (With Plates 1—4)

On a Collection of Nemerteans from Singapore. By R. C. Punnerr,
B.A. (With Plates 5—8)

On the Protostigmata of Molgula manhattensis (De Kay). By
ARTHUR WitLuy. (With Plate 9) :

PAGE

71

13d

141

NOV 22 1900

The “Sexual Season’? of Mammals and the

Relation of the “ Pro-cstrum” to Menstruation.

By

Walter Heape, M.A.,
Trinity College, Cambridge.

ContTENTS.

PAGE
Introduction i
The sexual season of male animals 11
The breeding season of female mammals 13
The sexual season of female mammals 1b

The periodicity of the sexual season in moncestrous mame in the
absence of the male 21

The duration of the sexual season in oly@atrous manomlss in vdlie ah:
sence of the male . 5 5

The duration of the cwstrus in panearraie a nelyeateous Tieminale
in the absence of tle male : " 40
The effect of maternal influences on the featule season and on cestrus 42
The pro-cestrum 45
The period of cestrus 54
Summary and conclusion 56
Literature . : ; ' ; : 66

INTRODUCTION.

Tue following paper is concerned with certain phenomena
which affect reproduction and which occur in female mammals

prior to the fertilisation of the ovum.

The times of propagation of the species and the behaviour
of many female mammals during certain portions of the

vou. 44, PART 1,.—NEW SERIES. A

2 WALITTER HEAPH.

breeding season have been noted by zoologists, but the
changes which take place in the female generative system
prior to gestation require examination which it is impossible
to extend, with few exceptions, to mammals in the wild
state, and almost all that is known on this matter is derived
from a study of domesticated mammals and of some few wild
animals kept in captivity. But little attention, however, has
been paid to the subject at all by scientific students, while the
only attempt, so far as I am aware, which has been made to
treat it from a comparative point of view is that of Wiltshire,
whose “ Lectures on the Comparative Physiology of Menstrua-
tion” were published in the ‘British Medical Journal’ (1883).

The subject is of importance in proportion to the light it
may throw upon the evolution of the functional phenomena
of breeding. To attack such a subject by means of data
obtained from the highest groups of animals may seem to
many to be beginning at the wrong end of the story, and
there is, of course, much truth in that view ; but knowledge of
the physiology of the lower animais is at present very limited,
and information regarding the habits of most of them at
times of reproduction extremely scanty ;! on the other hand,
we have available some knowledge of the habits of many
classes of mammals and of the variety of sexual phenomena
exhibited by them.

‘he data for a comprehensive comparative account even in
mammals does not exist; at the same time there is sufficient
material at hand, in my opinion, to permit of a foundation
being laid, upon which it will be more easy to arrange facts
in the future. It is, therefore, not with any idea of finality,
but with the purpose of suggesting a wide field of inquiry,
and with the hope of assisting therein, that this chapter on the
comparative physiology of breeding has been written.

In the first place, with regard to the terms to be used; at
present there is great confusion regarding those used by
breeders ; the same terms are used for both male and female

1 In this relation Dr. Lo Bianco’s papers in the ‘ Mitth. Zool. Stat.
Neapel,’ vol. viii, 1888, and vol. xiii, 1899, are of great value.

tHe © SEXUAL SEASON’’ OF MAMMALS. 3

animals when they should not be so used, the same terms are
used for different processes and conditions in female mam-
mals, and it is necessary for a clear understanding of the
subject that the limits of their use should be defined, and
where needful new terms adopted.

One of the most fertile sources of confusion 1s, disregard
of the fact that the history of the generative phenomena
exhibited by female mammals is different when reproduc-
tion takes place and when it does not take place; it is
essential that this fact should be kept in mind.

The remainder of this Introduction I have devoted to a
definition of the terms used in the following pages, and to an
endeavour to show wherein they differ or are in accord with
those now in use.

Reproductive Period.—lI have used this expression to
denote the whole of that period in the life of a mammal,
whether male or female, during which its generative organs
are capable of the reproductive function ; and in contrast to
the Pre-reproductive and Post-reproductive periods
which severally precede and follow it, during which the gene-
rative organs are either not fully developed or are degenerate.

The bearing of young, however, is not possible at all times
during the reproductive period. In the course of that period
there are intervals during which the generative organs of all
mammals exhibit, on the one hand special activity, and on
the other hand a fallow condition. This variation is periodic,
and is due, not to a periodic degeneration from a stable con-
dition, but to the periodic accession of a special stimulus,
culminating in sexual desire, and resulting in coition and in
gestation in the female when conception takes place.

The periodicity of this stimulus is very variable, and is
influenced by many factors of both extraneous and internal
origin.

Breeding Season is adopted to denote the whole of that
consecutive period during which any male or female mammal
is concerned in the production of young, and it is not apph-
cable to any isolated portion of that period.

4. WALTER HEAPE.

Although the part which the male takes in breeding 1s
confined to the insemination of the female, while the whole
of the rest of the process is carried on by the female, and
in spite of the fact that the word “breed ” carries with it,
in its original sense, as I understand it, the giving of nourish-
ment, and might perhaps in that sense be confined to the
female, it is impossible to avoid including the male. The
extent of the breeding season of a male depends upon the
length of time during which he is preparing for, and is capable
of, inseminating a female; while the extent of the breeding
season of a female mammal depends upon the extent of the
sexual season, during which her generative organs are pre-
paring for conception, plus the time occupied by gestation,
or the gestation period.

The term has been used to describe specially the season
when mammals copulate, or, again, it has been used to specially
designate the period of gestation in the female; but it is not
applicable as a definitive of either of these periods separately,
and must be used for the whole consecutive breeding period
experienced, and in this sense is applicable to both male and
female mammals.

The term “ breeding” is also frequently used in connection
with the rearing of the young after birth, and this has given
rise to confusion, inasmuch as while the mother is providing
nutriment for young already born, she may or may not be
bearing others. It is obviously inconvenient to include the
period of suckling in the “ breeding season,” for which reason
I have called it the Nursing Period.

Sexual Season is aterm I have used to designate, for
both male and female mammals, the particular time or times
of the year during which their sexual organs exhibit special
activity.

2

Some mammals experience only one sexual season each year,
some experience more than one; in some it is a brief period,
in others it occupies a much longer time ; in some the sexual
season of the female may be interfered with by gestation, in
others it is not.

THE ‘* SEXUAL SEASON’’ OF MAMMALS. 5

It is a convenience to be able to use one term for this
phenomena in both sexes, but it is to be noted in the first
place, that the sexual season of an individual male and indi-
vidual female of the same species is not necessarily coincident,
either with regard to time or with regard to extent; and in
the second place, that the phenomena exhibited are different
in the two sexes. [or this reason special terms are used for
each sex.

The Male Sexual Season.

Rutting Season.—This term is used to describe all
seasons of special activity of the generative organs of the
male, during which he is desirous of coition and normally
capable of inseminating the female. In some animals these
seasons are of short duration and at long intervals ; in others
the intervals may be shorter or the duration of the season
longer; while in others, again, there would appear to be little
or no cessation of the generative power.

It is necessary here to remark that the term “rut”
(German “ Brunst,’’? French “rut’’) is used by German and
French authors frequently, and by some English writers, to
designate the conditions obtaining in both male and female
mammals during the sexual season. ‘This is an error; it is
essentially a word which should be confined to the phenomena
exhibited by the male; it has its origin in the Latin word
“pugire,’ to roar or bellow, and is, I believe, strictly appli-
cable only to such animals as stags and boars. There are,
however, other male animals to which the term may be applied
in its original sense, as, for instance, the bull elephant in a
condition of ‘‘ must,” and it will be convenient to extend the
use of the term “rut” to the males of all animals which
exhibit seasons of special generative activity ; to those, on
the other hand, who are capable of inseminating the female
at all times of the year, the term is not applicable.

The Female Sexual Season.

In the case of the female the activity of her generative

6 WALTER HEAPE.

organs and the form which that activity takes is modified by
conception, and it is necessary to consider the subject under
two heads: (1) when reproduction does not take place, that
is in the absence of the male, or when coition does not
result in conception; and (2) when reproduction does take
place.

Under either of these circumstances the changes which
take place in the generative system are both complex and
variable, and for purposes of comparative study must be
identified.

(1) When Reproduction does not take place.—In
the first place we will consider the changes which take place
in the simplest form of the female sexual season, and after-
wards indicate the nature of the more complicated processes.

Pro-cestrum, or the Pro-wstrous Period, is the term
I have adopted to describe the first phases of generative
activity in the female mammal at the beginning of a sexual
season ; it is identical with the period spoken of by tlhe more
accurate breeders as the time during which an animal is
“coming on heat,” or “coming in season.” During this
period certain changes take place in the generative organs
which, while in some animals they are more drastic, in some
more complete than in others, are based on the same plan,
have the same object, and the same effect in all. They result
in a condition which I have called—

(istrus.—This is the climax of the process ; it is the special
period of sexual desire of the female; it is during cestrus,
and only at that time, the female is willing to receive the
male and fruitful coition rendered possible in most, if not in
all, mammals.

(istrus may be a brief period and exist for only a few
hours, or it may extend for days, or apparently even for weeks;
it is possible, however, normally, only as a result of the active
changes which take place in the generative organs during
pro-cestrum.,

The period of cestrus is referred to by various writers
as ““Brunst,”’ “rnt;” “heat siseasone 75) “brim «amor

THE ‘* SEXUAL SEASON”? OF MAMMALS. a

“cestrum ; ”

as I have before remarked, some of these terms
are used also to designate the rutting season of the male, and
most of them are used indiscriminately for both the periods
of pro-cestrum and cestrus, which I seek now to establish
for the female. In comparing, therefore, the writings of
former investigators with the statements made in the follow-
ing paper, it must be recollected that the various terms
hitherto used are not necessarily homologous with those used
by me, and are not necessarily descriptive of the sexual
season or of the breeding season of female mammals, as I
understand these processes. Much of the confusion and
misunderstanding which exists, regarding breeding phe-
nomena, is due to the indiscriminate use of the terms above
noted, and it is essential that their use should be restricted,
or given up altogether.

There is one point which I should briefly refer to here. I
have said above that cestrus, the period of desire, normally
follows pro-cestrum ; but there are times when the females of
certain, probably of many, mammals are anxious to receive
the male without the pro-cestrum having taken place.

This condition may occur in various mammals during
pregnancy, and has frequently been noticed in most species
of domestic mammals during that period, while it is evident
in a considerable number of animals also at other times. This
may be called abnormal cestrus. Normal estrus, as we
shall see below, occurs in conjunction with certain changes
in the uterine tissue, and this is accompanied by congestion
and stimulation or irritation of the copulatory organs;
whether the congested condition of the latter organs is in
itself sufficient to induce cestrus, or whether the presence of
some peculiar substance in the blood, or other influences, are
essential for that purpose, is not known ; however that may
be, congestion is invariably present, and is an essential
condition.

So also in abnormal cestrus, congestion of the copulatory
organs takes place, but the changes in the uterus which are
evident in normal cestrus are apparently absent. When

8 WALTER HEAPE.

cestrus occurs during pregnancy it is probably due to a
temporary diversion of a superabundant supply of placental
blood ; when it occurs at other times, the highly nutritious
food, with which the animals which experience it appear to
be generally supplied, or the condition resulting therefrom,
is possibly largely responsible for it.

Metcestrum, or the Metcstrous Period.—lIf concep-
tion does not take place during cestrus the activity of the
generative organs gradually subsides during a definite period,
which I have called the metcestrum ; and this is followed, in
the simple form which we are now considering, by a long
period of rest.

Ancstrum, or the Anestrous Period, is the name I
have given to this period of rest. It may last two or three
or eleven or possibly more months, and during that time the
generative organs lie fallow in the non-pregnant female. It
is eventually succeeded by a new pro-cestrum, and the four
periods, pro-cestrum, cestrus, metcestrum, and ancestrum, con-
stitute what I have designated as the anestrous cycle.

By some this period of ancestrum is called the non-breed-
ing season, but this is not correct, for although conception
cannot take place during this period it may be occupied
partially or wholly by the period of gestation, and inasmuch
as gestation is included in the breeding season, the ancestrum
cannot be considered as a non-breeding season.

We now have to consider a more complicated form of
sexual season. In this case the sexual season is ushered in
as before, with the pro-cestrum, cestrus follows, and is
succeeded by metcestrum, but instead of the ancestrum, a
short quiescent period now occurs which I have called the—

Dicstrum, or the Diwstrous Period.—This is a brief
period lasting only a few days, at the most probably not
more than twelve or fourteen days, while in some animals
four to six days may be its duration. It is followed at once
by a new pro-cestrum, and the four periods, pro-cestrum,
cestrus, metoestrum, and dicestrum, I have designated the
dicestrous cycle.

THE “SEXUAL SEASON”? OF MAMMATS. 9

In those animals which experience the dicestrous cycle the
sexual season (when conception does not take place) consists
of a series of such cycles, two or more; and may occupy any
length of time from one month to the whole year: In the
former case it is limited to a definite portion of the year
only, while in the latter case it may be coincident with the
whole reproductive period [human female, under certain
conditions]. But when the recurrence of the dicestrous
cycle is limited to a definite portion of the year, the sexual
season is, of course, also limited to that period, and there is
consequently a period of rest, which is the ancestrum.

In such cases the non-pregnant female experiences a series
of dicestrous cycles during the sexual season, at the end of
which, instead of dicestrum following metcestrum, the latter
is succeeded by ancestrum, which persists until the next
sexual season occurs.

In order to distinguish between the two classes of female
mammals into which the occurrence or absence of dicestrum
divides them, I have called those which experience a single
oestrus during each sexual season, or in other words those in
which the ancestrous cycle only occurs, moncestrous
mammals; while those whose sexual season is occupied by
a series of dicestrous cycles, or in other words those who
experience a series of recurrent cestri, | have called poly-
cestrous mammals.

The complication into which an otherwise simple story is
thrown is due, therefore, to variation in the quiescent period.
In some animals this may bea very brief period, never lasting
more than a few days; in others it may occupy from two to
eleven months each time it occurs ; while in others again both
these conditions are experienced at different times of the year.

Functionally, no doubt, these two varieties of the quiescent
period are homologous, the one is a modification of the other ;
and the modification is probably due, as will be shown below,
to an increased or decreased power of reproduction. At the
same time, for the purposes of the present paper, the difference
between them is essential, and their relation to the sexual

10 WALTER HEAPE,

season renders it necessary to discriminate clearly between
them.

The result of the foregoing may be summarised thus:
when the male has not access to the female during the sexual
season, or when insemination at that time does not result in
the fertilisation of an ovum, pro-cestrum and cestrus are
followed by metcestrum and, if the animal be polycestrous,
dicestrum is followed by another pro-cestrum, and such di-
cestrous cycles continue so long as the sexual season lasts ;
whereas if the animal be moncestrous, or if the dicestrous
cycles of the polycestrous animal be ended, ancestrum follows,
and persists until a new sexual season occurs.

A few examples will render the foregoing somewhat more
clear. Among moncestrous mammals is the wolf, which, in
the wild state, experiences only one sexual season at a
particular time each year; in her case pro-cestrum and cestrus
are followed, when conception does not take place, by met-
cestrum, and the whole of the remainder of the year is
occupied by ancestrum. She therefore experiences a single
ancestrous cycle each year.

Another moncestrous animal is the domestic bitch ; but in
her case, in the absence of gestation, the ancestrous cycle
may recur two, three, or even four times each year.

Among polycestrous mammals the mare may be taken as
an example; during a certain portion of the year, of variable
extent, she undergoes a series of dicestrous cycles when she
is not pregnant; this portion of the year is her sexual
season; when it is over ancestrum occurs and lasts until the
commencement of the same time the following year.

The human female, who is also a polycestrous mammal,
under certain circumstances has a continuous series of
dicestrous cycles throughout the year when she is not preg-
nant, and is thus subject to a sexual season during the whole
of her reproductive period.

(2) When Reproduction does take place.—In this
ease the pro-cestrum is followed by cestrus, during which
period insemination occurs and the ovum or ova are fer-

THE ‘‘ SEXUAL SEASON’’ OF MAMMALS. 1]

tilised; gestation results and persists until parturition takes
place.

After parturition there may be a considerable interval of
rest; this interval may occupy only what remains of the
ancestrous period which the same animal would experience
in case it had not borne young, or it may persist during a
nursing period which extends beyond the normal limits of
such ancestrous period, or it may be even still further pro-
longed. On the other hand, parturition may be followed
almost immediately, and in spite of the nursing period, by
pro-cestrum, cestrus, insemination, and renewed gestation.
While finally, the same animal may at one time of the year
exhibit a recurrent gestation, while at another time of the
year its generative organs may continue fallow for the re-
mainder of that interval which represents the ancestrous
period.

Such briefly are the different types of breeding phenomena
exhibited by female mammals during their reproductive
period; the following account will show that they all con-
form to one plan, and that the variability, which altered
conditions of hfe induce therein, clearly indicates the origin
of these types. On this account the subject is likely to be
of considerable interest to students of variation, and the
collection of facts which bear thereon is urgently needed.

The Sexual Season of Male Mammals.

It is unnecessary to do more than mention here that males
may be divided into two classes: those which rut (stag), and
those which do not rut (dog). Rutting males have a special
sexual season; those which do not rut experience sexual
capability all the year round.

The sexual season of some males in captivity! is capable

1 Information regarding wild animals in captivity, unless otherwise stated,
has been obtained from certain keepers in the Zoological Gardens of London,
whose statements appear to me to be reliable. The reference given in the
text is (Zoo.),

12 WALTER HEAPE.

of modification similar to that of certain females under the
same conditions; for instance, wapiti stags under natural
conditions have a special limited rutting season, but in
captivity (Zoo.) they rut all the year round except during the
season when they cast their antlers and until those structures
grow again.

When rutting exists it is probably excited by similar
influences to those which induce the advent of cestrus in the
female; on the other hand, when the sexual season of a male
is a permanent characteristic, either all the females of that
species have a sexual season all the year round or individual
females have different times for their sexual season.

As examples of these two conditions it may be pointed out
that the camel in the Zoological Gardens of London ruts at
much the same time as the female camels experience cestrus
in Mongolia, namely, early in spring (Prejevalsky, 1876),
although in the Gardens there are no female camels ; while
the sexual passions of the dog, on the other hand, are excited
by oestrus of the bitch and may be called forth at any time
of the year.

At the same time the proximity of the two sexes may
stimulate both cestrus and rutting. The stimulation of
cestrus is noticed in some of the larger carnivora in the
Zoological Gardens by the presence of the male, while I have
noticed rut in Semnopithecus entellus, in the Calcutta
Zoological Gardens, stimulated by the female; and rut in
the domestic rabbit stimulated by a doe under the influence
of cestrus.

It is interesting to observe that while the sexual activity
of domestic mammals (Miller, 1858) and of wild animals in
captivity (Heck, 1899) may be more frequently exhibited,
it is not so violent as is shown by animals in the wild
state.

For the purposes of this paper, this is all that need be said
specially, regarding the generative phenomena exhibited by
the male; although the activity of his generative organs may
be to some extent influenced by the presence or absence of the

THE “SEXUAL SEASON’? OF MAMMALS. 13

female, the general scheme of his reproductive period, and
breeding, sexual or rutting season, remains the same.

The Breeding Season of Female Mammals.

The breeding season of mammals should rightly be con-
sidered after the sexual season has been discussed, but,
owing to the fact that the term “ breeding season ” has been
so universally used to designate the sexual season as well
as the gestation period of breeding mammals, it is necessary
to say a few words here in order to make the following
account clear.

The occurrence of a breeding season depends upon the
occurrence of a sexual season, and those factors which
influence the former, influence also the latter, and will be
treated under that head. ‘The same is true for the recurrence
of both the breeding and the sexual seasons.

The recurrence of the sexual season may be interfered with
by the bearing of young, both gestation and nursing may
so interfere, but that does not remove a consideration of the
question, under such circumstances, out of the realm of the
sexual season; the effect of these processes, of bearing young
and of nursing young, on the sexual season, must be con-
sidered in relation to that period, and must not be supposed to
have relation only to the remainder of the breeding season.

Questions regarding the breeding season of mammals con-
cern what happens during both the sexual and the gestation
periods jointly, and, as I have before stated, the expression is
a term used to define the period passed through by an animal
which experiences both these processes ; it is not applicable as
an expression or term which may be used for the occurrence
of either of them separately, nor for the effect one of these
processes may have upon the other.

A breeding season may include only one sexual season and
one gestation period; this is true for all moncestrous mam-
mals, of which the bitch will serve as an example, and it may

14 WALTER HEAPE.

be also true for certain polycestrous mammals, as, for im-
stance, the mare, under certain circumstances. On the other
hand it may include several sexual seasons and several gesta-
tion periods, a condition to which only certain polycestrous
mammals can attain, of which the rat is an example.

The time occupied by a breeding season is very variable,
from a few weeks (bitch) to several months (mare), and even
more than a year (elephant). ‘There may be only one
breeding season in the course of several years, as shown by
the walrus (Bell, 1874), elephant, and probably rhinoceros
(Willoughby, 1889). There may be one breeding season each
year (mare) or more than one (domestic bitch and cat).

The result of a breeding season may be the birth of one
young one (mare usually), one litter of young ones (bitch), or
many litters (rat).

There may be great variation in the period of gestation of
different species of the same order of mammals. For in-
stance, among Rodents, the rat goes twenty-one days in
young, the rabbit thirty-two days, the guinea-pig sixty-three
days. Among Hquide, mares carry their foals eleven months,
asses from three hundred and fifty-eight to three hundred
and eighty-five days, and Burchell’s zebra over thirteen
months (Tegetmeier and Sutherland, 1895). Among Ovidee
the Barbary wild sheep goes from twelve to fifteen weeks in
young (Zoo.), while the domestic sheep in this country
averages about twenty-one weeks.

‘There may even be variation in the period of gestation in
varieties of one species ; for instance, Merino sheep average
150°3 days’ gestation, while Southdowns average 144-2 days
(Darwin, 1875), and different breeds of cows apparently vary
from 277 to 288°75 days’ gestation (Varigny, 1892).

The supply of food available may influence the length of
time occupied by gestation. A correspondent who is a
sheep: breeder informed me that his ewes, when run on poor
land, experience an appreciably longer gestation period than
those run on rich land; and I am strongly inclined to think
investigation will show that the supply of food, and the

THE ‘°* SEXUAL SHASON’’ OF MAMMALS. 15

quality of that food, have very marked effects not only upon
breeding seasons and gestation periods, but upon fertility
generally, upon the mother and upon the foetus (Latarte, 1891).

It is with such questions as these that the consideration of
the breeding season as a whole is concerned; with them the
following paper does not deal, and it is obvious that before
they can be profitably discussed, not only the sexual season,
but the gestation period must be examined separately.

The Sexual Season of Female Mammals.

In dealing with this subject we have to discriminate be-
tween mammals under three different conditions: namely,
wild animals in a state of nature, wild animals in captivity,
and domesticated animals ; and this is necessary, because the
generative system of wild animals is affected by the condi-
tions attending captivity, because the effect of captivity is
not necessarily the same as the effect of domestication, and
because wild animals cannot be examined so closely as the
others and less is known about them.

In dealing with wild animals in captivity it 1s necessary to
bear in mind the fact that good food, warmth, and shelter
have a very great effect on the increase of the generative
powers of some animals, while on others a strange climate,
confinement, want of violent exercise, and probably the
absence of opportunity for periodic gorging of freshly killed
food, or of a sufficient variety of food, have the opposite
effect.

As an example of the former the deer and cattle in the
Zoological Gardens may be quoted, as an example of the
latter some of the larger carnivora will stand.

In dealing with domesticated animals we do not know
what the original conditions were, and we have to take the
facts as they stand. At the same time we may assume that
animals which do not show themselves to be prolific under
domestication are rarely domesticated, and that a very long

16 WALTER HEAPE.

course of artificial selection, added to the plentiful supply of
food, with warmth and shelter inseparable from domestica-
tion, has no doubt greatly increased their power of repro-
duction.

As has been already stated, there are two forms of sexual
season evident in female mammals ; the moncestrous, in which
there is only a single cestrus at one or more particular times
of the year (bitch), and the polycestrous, in which there are
two or more concurrent dicestrous cycles at a particular time
of the year (mare).

The sexual season may be influenced by the climate of the
region in which the animal lives, by the seasons of the year
when these are of marked variation, and by the supply of
food, or possibly by the nature of the food, obtainable. These
may be called climatic influences.

It may also be influenced by special nervous, vascular, and
secretory peculiarities of the individual and by its habits of
life. These may be called individual influences.

It may also be influenced by the length of gestation, the
claims of the newly-born offspring on the mother (i.e.
nursing), and by her powers of recuperation. These may be
called maternal influences.

Such influences may affect the time of year when the
sexual season occurs, its recurrence, and its duration. The
influences which affect the time of year when the sexual
season occurs, concern both moncestrous and polycestrous
mammals, and are essentially governed by climatic considera-
tions, including the supply of food. The recurrence and
duration of the sexual season on the other hand are affected
either by climatic, individual, or maternal influences, and
are also experienced both by moncestrous and polycestrous
mammals, though in a somewhat different way by each.

In order to understand this difference it is necessary to
examine briefly what occurs in these two classes of animals.
Among moneestrous animals there are a variable number of
sexual seasons each year, one or more, each composed of a
single cestrous of variable duration. So that the result of the

THE ‘* SEXUAL SEASON’? OF MAMMALS. 17

different influences which affect the sexual season may be
either to increase or decrease the periodicity of that season,
or to increase or decrease the duration of each one.

Among polycestrous animals there is usually one sexual
season per annum, which is composed of two or more dices-
trous cycles, and the result of these influences on such animals
may be, either to increase or decrease the number of con-
secutive dicestrous cycles in any one sexual season, or to
increase or decrease the duration of each cycle.

The effect of these influences in both cases is to increase or
decrease the reproductive power of the animals, and they act
in moncestrous animals by affecting both the periodicity and
duration of the sexual season, in polyoestrous animals chiefly
by affecting the duration, but in two different ways, namely
by increasing or decreasing both the number of consecutive
dicestrous cycles and the duration of the cestri which occur
therein.

Modification of the periodicity of the sexual season, there-
fore, is chiefly found among moncestrous animals ; while modi-
fication of its duration is common to both moncestrous and
polycestrous animals. It would seem possible to simplify
these conditions if it were assumed that the polycestrous
arose from the moncestrous condition ; if it were assumed, in
point of fact, that polycestrum is simply a condition arrived
at by the concentration of several moncestrous sexual seasons.

There might seem to be some reason for this when such
animals as the red deer, for instance, are considered ; in the
wild state this animal is apparently moncestrous, while in
captivity it is polycestrous, at any rate in this country.

But it may equally plausibly be argued that moncestrum is
simply decentralised polycestrum. There are instances among
domesticated animals of moncestrous animals with a tendency
to polycestrum (bitch), and of polycestrous animals with a
tendency to moneestrum (mare). So also among wild animals
there are instances of animals which are moncestrous in one
climate and apparently polycestrous in another (Sciurus
vulgaris) (compare Bell, 1874, and Lataste, 1887).

vou. 44, part 1.—NEW SERIES. B

18 WALTER HEAPE.

I doubt if, in the present state of our knowledge of the
subject, it is possible to determine which is the original of
these two conditions. The behaviour of animals in captivity
and under domestication inclines one to believe that monces-
trum is the original form ; then, again, it is the simplest form,
and on that ground may be thought the more archaic. But,
on the other hand, it is the lower animals which are the most
prolific breeders, and, for many reasons, we may perhaps
expect the ancestral mammal to have been more prolific than
wild animals are now.

If this should be true, the increased capacity for reproduc-
tion, shown by domesticated animals, would indicate rever-
sion to ancestral powers, due to the removal of such obstruc-
tions as must be inseparable from the struggle for existence.
Thus all we can be certain of is the close similarity between
these two forms of sexual season.

A further complication is introduced by certain breeds of
domesticated sheep and by pigs; these are polycestrous ani-
mals when domesticated, and they may also exhibit more than
one sexual season each year. Sucha condition appears to be
exceptional, and I have not included this form of variation
in the foregoing account for that reason; but I am quite
prepared to believe a more exact knowledge of what takes
place among domesticated animals will show a similar varia-
tion among individuals of other classes of animals.

Variation in the periodicity of sexual seasons is brought
about by an increase or decrease in the persistence of the
anoestrum, and has nothing to do, necessarily, with variation
in the cestrus cycle itself; while, on the other hand, variation
in the duration of a sexual season is brought about by an
increase or decrease in the number of consecutive dicestrous
cycles (polycestrous mammals), or by an increase or decrease
in the persistence of the cestrus (moncestrous and polycestrous
mammals), the ancestrum being only secondarily affected in
consequence thereof.

The effect of an increase in the periodicity of sexual
seasons may be twofold ; it permits of reproduction at differ-

THE ‘© SEXUAL SEASON’? OF MAMMALS, 19 -

ent times of the year, and, when gestation is of sufficiently
short duration, of reproduction more than once a year, An
increase in the duration of the sexual season may also have
two effects; it gives increased opportunity for successful
coition, highly advantageous to those animals which live an
isolated life, while, among animals which experience a sufli-
ciently short period of gestation, it gives them opportunity
for reproducing several times in each season. On the other
hand, a decrease in the periodicity or duration of the sexual
season has an opposite effect.

It would seem highly probable that the method of increas-
ing or decreasing the opportunities of reproduction varies
with the habits of animals, the claims of maternity, and the
climate in which they live. The different methods are not
necessarily peculiar to particular groups or classes of animals,
and they may vary, within limits, in the same species in
different localities and in the same individual under different
circumstances. Climatic and maternal influences may be
observed in wild, captive, and domesticated animals; but
individual influences can only be noted in the two latter
classes, and especially in domestic mammals.

It has been freely stated, originally by Aristotle and sub-
sequently by numerous biologists, that the sexual seasons
are governed by the requirements of the newly born young ;
in other words that the season for conception is regulated
by the length of gestation and the time of year which is
most favourable for the birth of the young; and it is argued,
that the different times of the year during which the sexual
seasons of similar animals occur is sufficient ground for that
view.

I cannot agree with this opinion ; if it were so, why should
some bats experience a sexual season in the autumn and not
produce young until about June (Beneden, 1880, and others,
see below), although not more that two months are required
for gestation and these animals are active for that length of
time in the spring, before parturition takes place? Again,
why should roe deer, in Germany, have their sexual season

20 WALTER HEAPE.

in early autumn, when the embryo does not develop beyond
the segmentation stage until the following spring? (Bischoff,
1854). Why should the seal take eleven to twelve months for
gestation, while a large dog only requires three months? If
there was a great difference in the size of these animals the
variation might to some extent be accounted for, but it is
not so. Itis true that the newly born seal is a far more perfect
animal than the newly born puppy, but it cannot be that the
whole of the difference in the time of gestation, namely, eight
to nine months, is required for the extra development of |
the more perfect seal embryo, other factors being equal.

Again, how is it that an unusual change of climate will
affect the sexual season of an animal? This is constantly
observed among domesticated animals, and a very marked
case is recorded of cows, in Skye, after an exceptionally
severe winter (Youatt, 1834).

And how is it that the sexual season, for instance, of the
fox (Bischoff, 1863) and red deer (Cameron, 1900), is modified
in accordance with the nature of the country in which it lives,
whether high or low ground, in accordance with the age of
the animal, and probably also in accordance with its bodily
condition ?

There seems to me ample reason for the belief that the sexual
season is governed directly by the influences detailed above—
climatic, individual, and maternal; and that variation in the
rate of development of the embryo, in the length of gestation,
and in the powers of nursing, are quite sufficient to provide
for the launching of the young at a favourable time of the year.

The origin of the sexual season is another matter; for a
solution of this question a comparative study of the phe-
nomena in the lower animals is necessary.

That it is the result of a stimulus which may be effected
through the alimentary canal is demonstrated by the effect
upon ewes of certain stimulating foods; the sexual season
of ewes may be thus forced by “flushing” them, as it is
called by flockmasters.

In the same way it is stated that a quart of milk, drawn

HE ‘* SEXUAL SEASON’? OF MAMMALS. 21

from a cow “in season”’ (i. e. during cestrus), but which has
not had the bull, will, if drunk by another cow, bring on the
sexual season of the latter (Youatt, 1834).

That it is associated with a stimulus which is manifested
by exceptional vigour and exceptional bodily ‘condition ”
is demonstrated by the pugnacity of the males at such times,
by the restless activity of the females, by the brilliant
colouring of such widely divergent animals as, for instance,
annelids, amphibia, birds, and mammals, by the condition of
the plumage of birds, and of the pelage or skin of mammals.

That itis associated with nutrition, and thatitis a stimulus
which is gradually collected, is indicated by the increased
frequency of the sexual season among domesticated mammals
as compared with nearly allied species in the wild state.

That it is manifested by hypertrophy and by congestion
of the mucous tissue of the generative organs, and of various
other organs, such as the wattles and combs of birds, the
crest of the newt; and by the activity of special glands, the
affection of all of which may be exceedingly severe, is true.

These,and many other similar facts,are well known, but they
do not assist in the elucidation of the origin of the function.

The most that they do is to show that the sexual instinct is
seasonal, and that nutrition, whether affected by external or
internal factors, plays an important part in its manifestation.

The Periodicity of the Sexual Season in
Monecestrous Mammals in the Absence of the Male.

In the absence of gestation most mammals would appear to
experience at least one sexual season per annum, under natural
conditions, but there is great variation in the periodicity of
the sexual season in captive and domesticated mammals, the
variation being extended not only to varieties of a species,
but also to individuals of that species under domestication.

Among certain wild animals which are known to undergo
parturition only during a very circumscribed time, the
moneestrous condition may be assumed as probable, and the
periodicity of the sexual season calculated ; but it must be

WE) WALLER HEAPE,

recollected that without accurate observation, during the ab-
sence of the male, it cannot positively be asserted that such
animals are moncestrous.

In the case of the bitch, in Danish Greenland the bitch
generally experiences only one sexual season per annum,
though sometimes she may have two (Rink, 1877).

In this country, as a rule, the bitch has two sexual seasons
each year, one in the spring and one in the autumn, but there
are wide variations to this rule. She may have only one
sexual season per annum, or if may occur every eleven, ten,
nine, eight, seven, six, five, or four months (Stonehenge,
1887). It seems probable the sexual season recurs less
frequently in breeds of large dogs as a rule; a correspondent
(Dr. Inmann), who breeds St. Bernard dogs, informs me this
is his experience, and I have had information from other
breeders of large dogs, mastiffs and bloodhounds, which shows
there is an obvious tendency in this direction; at the same
time it does not appear to be by any means a universal rule.
Again, while the spring and autumn are the usual times when
the sexual seasons of dogs occur, the sexual seasons of each
bitch have a more or less exact periodicity pecuhar to herself.

Finally, the sexual seasons of any bitch may be interfered
with to the extent even of complete cessation, if she is sys-
tematically prevented from breeding.

The bitch may be considered a case of extreme variation in
the periodicity of the sexual seasons of a moncestrous domes-
ticated mammal. The normal two sexual seasons experienced
in this country are reduced usually to one in Danish Greenland,
probably owing to the effect of climatic influences, while the
variations which exist in this country indicate the effect of
individual influences, which are largely stimulated by artificial
selection and domestication.

The wolf, jackal, and fox are moneestrous like the dog, and
in captivity in the Zoological Gardens they show two sexual
seasons per annum, like the normal dog in this country.

Bears are also moncestrous, but they have only one sexual
season per annum in the Zoological Gardens.

THE “°° SEXUAL SHASON’”? OF MAMMALS. 23

Badgers also are probably moncestrous, but there is great
uncertainty regarding their breeding (Harting, 1888; Den-
wood, 1894).

The same is true for the Barbary wild sheep ; they are said
to be moncestrous and to have one sexual season per annum
in captivity in this country (Zoo.).

The red deer, fallow deer, and roe deer are probably
moncestrous in the wild state; they have only one sexual
season of very limited duration (Bell, 1874). The same
may be said for the ibex, Markhor, Barasingh, and
Hemitragus jemlaicus in Cashmir (Laurence, 1895),
possibly also the American bison (Allen, 1876), and various
other species of Bos, Ovis, and Capra (lydekker, 1898) ;
also the black-tailed deer in Montana (Roosevelt, 1893), and
several antelopes (Sclater and ‘Thomas, 1900).

The truth regarding these animals is not, however, known ;
their moncestrous condition is rendered probable from the
known very limited sexual and calving seasons, but it is by
no means certain.

The genus Sorex, some of the Mustela, Myoxus avel-
lanarius, Arvicola amphibius, and Sciurus vulgaris,
in this country (Bell, 1874) are probably moncestrous in the
wild state, as are also the wild cat and the fox, and they have
only one sexual season.

Phoca vitulina, P. hispida, P. greenlandica, Cysto-
phora cristata, and Halichcerus gryphus have all a
very limited sexual season, occurring once only in the year,
and it is highly probable they also are monestrous (Bell,
1874, and Turner, 1875).

Variation in the periodicity of the sexual season of various
domesticated animals, in comparison with nearly allied
species in the wild state, has been observed in a few cases.

The cat in the wild state has one (Hamilton, 1896)—some
say two (Mivart, 1881), though this seems doubtful—sexual
season per annum, while the domestic cat may have three
or four sexual seasons each year.

The sow has only one sexual season in the wild state in

24, WALTER HEAPE,

France (Beever, 1870), but it is not clear whether she is
moncestrous or polycestrous; when domesticated, however,
she is polycestrous (Fleming, 1878; see also Aristotle).

Certain wild sheep, O. argali, O. burrhel, O. poli,
have only one sexual season per annum, and are probably
moncestrous (Prejevalsky, 1876) ; whereas domesticated sheep
are polycestrous, and may have such an extended series of
dicestrous cycles that they are capable of producing young
almost at any time of the year; such, for instance, are Dorset
Horns in the south of Hngland and Hampshire Downs in
some parts of Ireland (compare also Aristotle). As a rule,
however, sheep in this country have a much more limited
polycestrous season,—as, for instance, the Scotch black-faced
sheep, which has only two recurrent periods of cestrus
(Cameron, 1900).

Variation in the periodicity of the sexual season of wild
animals, as compared with individuals of the same species
in captivity, has been noted in but few cases. Some of the
large carnivora in the Zoological Gardens exhibit great irre-
gularity in their sexual seasons, but too little attention has
been paid to the subject in these animals to allow of more
being said than that, in some cases, their generative activity
appears to have been stimulated, in others checked.

The wolves in the Zoological Gardens have two sexual
seasons, while the Tibet wolf (lL. chanco) has only one (Pre-
jevalsky, 1876) in a wild state; in New Mexico also, I am told
by a keen sportsman familiar with the country, wolves bear
young only once each year (W. Ruston). ‘The same is true
tor the foxes in the Zoological Gardens; they have two sexual
seasons, while the Tibet fox (Prejevalsky, 1876) and the
Knglish fox (Bell, 1874) have only one in the wild condition.

The wild cat, on the other hand, in captivity does not
experience more sexual seasons than when in a feral state,
namely one (Hamilton, 1896); and the tame cat, when it
becomes wild, has apparently only one sexual season, whereas
the same animal under domestication has from two to four
sexual seasons per annum.

THE ** SEXUAL SEASON’? OF MAMMALS. 25

Among the deer in the Zoological Gardens their generative
activity appears to have been universally stimulated—they will
be referred to under the heading of “ Duration of the Sexual
Seasonin Polycestrous Mammals,”—for it wouldseem that their
normal (as I have considered it) moncestrous sexual season is
increased by the conditions of captivity until it may become a
continuous polycestrous sexual season.

The Barbary wild sheep, on the other hand, does not appear
to be affected by captivity ; it exhibits a single moncestrous
sexual season only, each year (Zoo.), and that is probably its
condition in the wild state if we may judge from what is known
of O. argali, and O. burrhel in Jibet (Prejevalsky, 1876).

It is with some hesitation I have included among monces-
trous mammals deer, sheep, and pigs in the wild state ; their
retention in this class is doubtful; but if these animals were
omitted there remains a remarkable series of examples of the
variability of the sexual season of moncestrous mammals
under various conditions.

The Duration of the Sexual Seasonin Polyestrous
Mammals in the Absence of the Male.

‘he duration of the sexual season in these animals depends
upon two factors, the length of the dicestrous cycle and the
number of times it recurs. Both factors may be different in
different species of animals, and either may be different in
different individuals of some species, or variable in the same
individual at different times.

Knowledge of polycestrum in animals in a wild state in
this country is lmited to certain rodents. The rat (M.
decumanus), mouse (M. musculus), and the rabbit in this
country are known to experience a recurrence of the dices-
trous cycle. It is probably recurrent also in M. minutus,
M. sylvaticus, M. rattus, Arvicolaagrestis, A. glare-
olus, and Lepus timidus; while possibly Mustela vul-
garisand Lepus variabilis, under favourablecircumstances,
may also experience a recurrence of the dicestrous cycle,
judging from the account given of them by Bell (1874).

26 WALTER HEAPE.

In Southern Hurope and Algiers polycestrum is apparently
usual amongst rodents (Lataste, 1887). It appears to be
ascertained for Sciurus vulgaris living in that part of the
world—though the same animal is probably monoestrous in
thiscountry—for Eliomys quercinus, Gerbillus hirtipes,
Dipodillus campestris, D. simoni, Meriones shawi,
M. longifrons, Mus musculus, M. rattus, M. decu-
manus, and to be probably true also for several other species.

The animals on which these observations were made by
Lataste were kept in captivity, but there is good reason
to think that the conditions under which they were kept did
not interfere with their habits in this respect.

Among domesticated animals polycestrum occurs in horses,
cattle, sheep, and pigs. While for wild animals in captivity
it has been observed (Zoo.) in the gayal and bison; in
wapiti, axis and red deer; in the gnu, eland, nilghau, and
waterbuck ; in Gazelle dorcas, in giraffes, in elephants, and
probably in kangaroos.

In its most complete form polycestrum occurs in certain
monkeys and in the human female; probably most monkeys
are similarly affected, and possibly also lemurs; in these
animals there is a regularly recurrent series of dicestrous
cycles throughout the year.

The Duration of the Diwstrous Cycle varies from
five days (exceptional in rodents, Lataste, 1887) to as much
as two months (exceptional in mares, and in various wild
animals in captivity from time to time, Zoo.).

The usual length of the dicestrous cycle for rodents is ten
to twenty days, and in other animals in which the phenomena
has been observed from three to four weeks. In the rodents
observed by Lataste (1887) the dicestrous cycle was usually
ten days, and in the rat and mouse in this country the same
may be said to be approximately true. In the domestic
rabbit, however, I find great variability ; while some indivi-
duals exhibit cestrus every three weeks fairly regularly, others
do so every ten days; on the whole I think ten to fifteen days
is the usual length of their dicestrous cycle.

THE ‘* SEXUAL SEASON’? OF MAMMALS. 27

In the domestic mare and cow three to four weeks, and in
the domestic sheep and pig two to four weeks is said
(Fleming, 1878) to be the length of the dicestrous cycle,
while another authority (Ellenberger, 1892) regards three
to four weeks as the usual time for all these animals.

In wild cattle, deer, and antelopes in captivity (Zoo.) three
weeks is the usual time. In monkeys it appears to be about
one month in duration (Heape, 1894, 1897, Keith, 1899). In
the human female, while twenty-eight days is the normal
length of time occupied by the dicestrous cycle, it is fre-
quently experienced every three weeks or every five weeks,
while occasionally even shorter or longer periods are known.
Aristotle is represented to have made the extraordinary state-
ment that few women menstruate every month, while most
menstruate every three months. It would seem possible that
the opposite is what he meant; at the same time it should be
remarked that various observers (Wiltshire, 1883) have re-
corded their opinion that the women of certain tribes in
different parts of the world menstruate only at long intervals
(see also Ellis).

The recurrence of the Dicestrous Cycle is also very
variable ; exact knowledge on this point is not possible for
wild animals; only those under observation, captive or
domestic, can supply the requisite information.

The known limitations of the sexual season among certain
wild animals, however, admit of a fairly accurate idea being
gained of the recurrence of their dicestrous cycles, although
not accurately enough to enable one to determine with cer-
tainty whether an animal is moncestrous or polycestrous.
For instance, the American bison (Allen, 1876) experiences
a sexual season from some time in July until some time in
August. In the Cashmir ibex it persists during parts of
November and December. In the Markhor and Hemitra-
gus jemlaicus in Cashmir it occurs in December, while in
the “Barasingh” in that country from September 20th to
November 20th it has been observed (Laurence, 1895),

In Scotland the red deer’s sexual season lasts three weeks,

28 WALTER HEAPE.

during September and October, according to Cameron (1900)
six weeks, while in this country September is the sexual
month for the fallow deer, and July and August the time
when the roe deer will receive the male.

In all these cases there can be little over three weeks
during which copulation takes place, and the extremely
limited period during which parturition occurs strongly
corroborates the view that this is the extent of the usual
time during which sexual intercourse is possible. The fact
that in captivity three weeks is the usual period which
intervenes between two cestri in such animals, and the
extreme probability that individual females do not all ex-
perience cestrus at exactly the same time (Cameron, 1900),
predispose one to believe that they are moncestrous in the
wild state ; but, if the limit of time for coition is three weeks,
there is still just time for the females to undergo two dices-
trous cycles, and it is this possibility which prevents positive
assertion on the matter.

Among captive animals (Zoo.) not more than two dicestrous
cycles have been observed in the gnu during one sexual
season. Gazelle dorcas has two or three; the giraffe about
three; while the eland, nylghau, and waterbuck have a
series of dicestrous cycles, each lasting three weeks, during
May, June, and July each year.

‘The gayal and bison, the axis and wapiti deer, on the
other hand, experience a continuous series of dicestrous cycles
all the year round, at intervals of about three weeks.

The hippopotamus at present in the Gardens is an old
animal ; for long she showed no signs of asexual season, but
lately she has done so at irregular intervals ; no doubt in her
case captivity has checked the generative function, for a
former specimen which bore three young while there is said
to have exhibited monthly sexual excitement (Wiltshire, 1883).

Among wild rodents in this country, recurrent dicestrous
cycles last about two months, probably,in Lepus variabilis;
about three months, probably, in Arvicola agrestis; from
four to six months, probably, in Mus minutus; about nine

THE ** SEXUAL SEASON’’ OF MAMMALS. 29

months in Mus rattus; and even longer, perhaps, in Mus
musculus and M. decumanus.

Bell (1874) appears to think that, under favourable circum-
stances, the dicestrous cycles may continue all the year round
in these latter animals and in the rabbit, but I am inclined to
think such a condition is unusual in this country among wild
rodents, since it is exceptional to find any of them pregnant
during the winter months.

Among domesticated animals the period during which the
dicestrous cycles recur, in the absence of the male, lasts from
one month to as many as eight months for the mare, about
five to six months for the rabbit, from one to three months for
the sheep (with certain exceptions), and about two months
for the pig. So far as the domestic rabbit is concerned, no
doubt, if they are kept warm, carefully fed, and their breed-
ing carefully regulated throughout the spring and summer,
they may exhibit cestrus also in winter, but it must be recol-
lected that here we are treating of cestrus independent of
pregnancy, which is a very different matter.

Among certain monkeys, probably in most of them, the
dicestrous cycle recurs all the year round (Geoffroy, St. Hilaire
and Cuvier, HKhrenberg, 1833, Numan, 1838, Heape, 1894,
1897, Keith, 1899 ; compare also Rengger, 1830, Sutton, 1880,
and Ellis). In the human female, as a rule, this is also the
case; there appear, however, to be exceptions to this rule, for
instance, the women of the Hsquimaux peoples living between
the seventy-sixth and seventy-ninth parallel do not always
menstruate during the winter months. It is said (Cook, 1894)
that not more than 10 per cent. of these women menstruate
during the long dark winter months, and itis possibletoimagine
that the peculiar conditions of life they experience during that
time may well be responsible for their peculiarity. If this be
so, a true ancestrous period may be experienced by women.

Rink’s (1877) account of the origin of these people, if
correct, precludes the probability that the occurrence of an
ancestrous period is a racial characteristic, and emphasises
the view that it is a variation due to climatic conditions.

30 WALTER HEAPE.

It is held by some writers, several of which are quoted by
Wiltshire (1883), that the women of various savage tribes
exhibit the menstrual flow only at intervals of several
months ; and the same author remarks on the fact that girls at
puberty in this country menstruate only at intervals of three,
four, or six months; and that it may be this condition is an
indication of an ancestral habit. Hllis also quotes various
authors who state that menstruation takes place at long
intervals in women of Lapland, Greenland, the Faroe Islands,
Tierra del Fuego, and among the Guaranis of Paraguay.

The effect of captivity or domestication on the
duration of the sexual season in mammals is very re-
markable.

As has been already pointed out, wild sheep have only a
very limited sexual season per annum (O. argali, burrhel
and poli, in Tibet, Prejevalsky, 1876), a condition which is
maintained by the Barbary wild sheep in captivity in this
country (Zoo.); whereas the domestic sheep has a much
longer sexual season, and in addition has for many centuries
(Aristotle) been capable of reproducing twice in each year
under favourable circumstances.

Again, the wild goat has a very limited sexual season
(Lydekker, 1898), whereas the domesticated goat will receive
the male at almost any time (Low, 1845). A more remark-
able example is that of certain deer in captivity (Zoo.). Wild
red deer have a special sexual season, extending little over
three weeks (Bell, 1874), and including certainly not more
than two dicestrous cycles; whereas in captivity (Zoo.) the
sexual season of these animals extends over most of the year,
and consists of an extensive series of dicestrous cycles.

A similar condition prevails with the wapiti deer in the
wild state (Roosevelt, 1893), while in captivity (Zoo.) the
possibility of pregnancy at any time of the year is only pre-
vented by the fact that the male does not rut during the
casting and growth of his antlers; and it is asserted that
park-fed wapiti stags in America are able to beget offspring
even after their horns are shed (Caton, 1881),

THE ** SEXUAL SEASON’’ OF MAMMALS. 31

Wild cattle in captivity (Zoo.) are also capable of repro-
duction at any time of the year, and they also experience a
remarkable increase in the recurrence of their dicestrous
cycles, from what we are led to infer, by the limited
calving season, is the case among similar animals in the
wild state.

Among domesticated mammals similar modifications are
evident, not only in animals of different species, but in indi-
viduals of the same species, as, for instance, in cattle and
horses.

Mares may have only one period of cestrus in the year, in
which case they are purely moncestrous animals, but this is a
rare condition; rarely, also, they may have two dicestrous
cycles, but usually they have many. In the latter case cestrus
may recur every three weeks, or the interval may be longer.
As a rule among thoroughbred mares the history of the
sexual season shows aseries of dicestrous cycles, each occupy-
ing about three weeks and recurring throughout the spring
and often until the early autumn, as many as seven or eight
months being in some cases thus occupied.

Although these animals—horses, cattle, and deer—either
in captivity or under domestication, experience such an ex-
tensive increase in the consecutive recurrence of the dicestrous
cycle, it is not a condition natural to them ; it is due, in all
probability, to the care and attention paid to them by man; in
the same way, it may be argued, that the stimulated power
of reproduction evinced by certain rodents is also due to the
advantages derived from their intimate relations with the
luxuries of civilisation (rat and mouse).

The only animals, so far as is at present known, which
experience a continuous series of dicestrous cycles in a state
of nature are certain monkeys.

The fact that it is possible to induce such an enormously
increased capacity for cestrus in any animals, prepares one to
consider the regular recurrence of the dicestrous cycle in
monkeys, and in the human female also, as a very slight step
in advance ; and when the whole of the evidence is considered,

32 WALTER HBAPE.

it will, I believe, be found that the regularly recurrent
dicestrous cycles of the Primates are strictly homologous with
the more or less regular dicestrous and ancestrous cycles of
the lower animals.

The Sexual Season in Monkeys.—The consideration
of this subject introduces a further complication, and that is,
while monkeys may have a continuous series of dicestrous
cycles, they are not all of them fitted for reproduction at all
times of the year.

Some monkeys in tropical countries may be in a condition
to become pregnant at all times of the year; though this is by
no means certain it is not an impossible fact, but others are
certainly not so. For instance the chimpanzee and gorilla
are said to have a special sexual season in West Africa
(Garner, 1896).

Semnopithecus entellus, from the jungles on the south
bank of the Hugli, has a definite time for reproduction
(Heape, 1894); and Macacus rhesus, the area of whose
geographical distribution is very large, apparently produces
young at different and definite times in different districts
(Heape, 1897).

There is every reason to believe, however, that these ani-
mals experience regular recurrent dicestrous cycles through-
out the whole year.

If the dicestrous cycle of a monkey is homologous with the
anoestrous cycle of a dog—and that this homology exists will
be apparent when the question is considered from a histo-
logical point of view—it is obvious that we are naturally led
to suppose that an increased number of cestri should result
in an increased number of opportunities for pregnancy, pre-
cisely as in the case of the mare, deer, etc. But this is not
so, and the result is that there exist certain mammals which,
while they exhibit a continuous recurrence of the dicestrous
cycle, have a circumscribed season for conception.

Ag I have shown elsewhere (Heape, 1894, 1897), this result
is due to the fact that, although menstruation recurs regu-
larly, ovulation does not; or, in other words, that ovula-

THE ** SEXUAL SEASON”? OF MAMMALS. 33

tion is not necessarily coincident with the cestrus in these
animals.

This opens up a wide question, which I hope to treat of in
a separate paper, but it is necessary to refer to it here in
order to point out, that the limited season for conception in
some monkeys is no reason for regarding their dicestrous
cycle as in any way different from that of other animals.

Briefly, we may say that both ovulation and cestrus are due
to stimulating influences. But they are not necessarily co-
incident in the lower animals, and they are not necessarily
both induced by the same means, nor at the same time.

In the virgin domesticated rabbit I find that ovulation
does not occur in consequence of cestrus alone; while various
observers have shown that in the bat ovulation may occur at
quite a different time of year from cestrus, in some cases
probably as much as six months may intervene between the
two functions in this animal (Benecke, 1879; Eimer, 1879;
Fries, 1879; Beneden and Julin, 1880).

Again, as I have already noted, there may be abnormal
cestrus in many animals, it may occur during gestation and be
independent of ovulation ; while finally, it is quite certain
that many animals which usually experience ovulation during
cestrus, sometimes fail to become pregnant at that time in
consequence of the failure of the function of ovulation.

Such being the case, it may truly be said the period of
cestrus is not invariably identical with the period of ovula-
tion; the two are separate functions, possibly closely asso-
ciated, but also possibly widely divergent.

In monkeys we have an instance of animals in which the
rhythm of ovulation may be different from the rhythm of
cestrus, but it must not be supposed, on this account, that
either of these processes is not homologous with the same
process in other animals in which the rhythm may be
identical. It would seem as if the sexual activity of these
animals had been developed more than, and out of propor-
tion to, the ovarian activity; or, in other words, that their
sexual powers were greater than their powers of reproduction.

vou. 44, part 1,—NEW SERIES. Cc

34 WALTER HBAPE.

The ideas on this subject which have for so long prevailed
and which even now are taught, namely, the identity of
“menstruation”! and of cestrus with ovulation, would make
this view impossible; but since it is known that, in various
animals, either “menstruation” or cestrus may take place
without ovulation, and that ovulation may occur without the
coincidence of “ menstruation”? (Leopold and Mironoff, 1894)
or of cestrus (hat), the possibility of isolating these func-
tions is demonstrated. Thus it is no longer impossible to
suppose that, while they are both due to similar stimulating
influences, one of them may be developed in excess of the
other. In this respect monkeys stand in an intermediate
position between the lower mammals and man.

The Sexual Season in Man.—In the human female
this question of the simultaneity of ovulation and cestrus
(“ menstruation,” as it has been wrongly called) has given
rise to wide discussion. I have referred to the question
somewhat fully elsewhere (Heape, 1894, 1897, 1898), and
have shown that the majority of modern writers on the sub-
ject are in favour of the view that the two functions are not
necessarily coincident in the human female, the correctness
of which conclusion it seems to me impossible to doubt.

With regard to the existence of a special limited sexual
season or seasons, it is interesting to note that there is some
evidence of such in the human female; evidence both of a
time in the past when such special seasons were common to
all, and of a time in the present day to which certain peoples
confine such matters and during which most peoples exhibit
special generative activity.

Here again we are upon the edge of a very wide field of -
research which it is impossible to do more than touch. I
may, however, briefly draw attention to certain facts which
in my opinion throw some light upon the matter.

Feasts, similar to the erotic feasts which were indulged
in by the ancients—Babylonians, Phoenicians, Egyptians,
Greeks, and Romans (Ploss, 1887),—were still practised to

1 « Menstruation ” is used here wn its original sense,

THE °*SHXUAL SEASON’? OF MAMMALS. 35

some extent in the sixteenth century in Russia (Kowalewsky,
1890 and 1891), and in some parts of India at a much more
recent date (Rousselet, 1876), while such customs as “ gwneyd
Bragod”’ (Owen, 1886) and possibly our own “ bean feasts”
may not improbably be the modern representatives of these
ancient customs in our own country.

Again, it is worthy of note that the erotic feasts of more
civilised peoples are not greatly dissimilar to the sexual feasts
and dances of the savage peoples of Australia, Polynesia,
West Africa, South America, New Britain, and West Asia
(Ploss, 1887). Apart from the fact that many of them surely
have some reference to phallic worship, as in the case of the
maypole, the origin of these feasts—shrouded as they are in
the mists of ancient customs now but little understood, and
of laws long since forgotten, complicated as they are by
customs, religions, and laws of a later growth—is not de-
finitely known.

It is indeed a matter of the greatest difficulty to trace,
justly, the true relation and likeness of any one of these
customs to another, however similar they apparently may be.
At the same time the universality of such customs is very
remarkable, and may, I think with some justice, prepare one
to believe that in prehistoric times man was impelled to
indulge, if not wholly, at least more freely, in sexual inter-
course at certain seasons rather than at other times of the
year.

Hill (1888) attempts to trace the apparent survival of a
human pairing season, by the customs of the Hindus and the
proportions of births in each month of the year; while
Westermarck (1891) records customs and statistics which
certainly point even more strongly in the same direction.
Ploss (1887) also gives many valuable statistics for Russia,
Germany, Italy, and France; and Haycraft (1880) does the
same for Scotland. It is remarkable that the statistics
brought forward by these authors in all cases show a con-
siderable rise in the birth rate at certain seasons. In Scot-
land, Haycraft oints out that from 1866 to 1875 a marked

36 WALTER HEAPE.

increase of births occurred with striking regularity in April,
showing that a maximum of conceptions takes place in July.

Hill says that ten years’ statistics of the district in which
he lived in India show that the maximum of conceptions
occurs in December, when food is cheapest and the salubrity
of the country at its best; while the minimum of conceptions
occurs in September, towards the end of the hot season,
when food is most scarce and malaria rife.

Ploss shows that in Russia the maximum of conceptions
takes place in autumn, in Germany during May and December,
while in Italy and France May is the month responsible for
most conceptions. This author also points out that in Russia
religion affects the birth-curve, and he traces the cause to
fasting seasons.

Westermarck goes very fully into this matter, and has
collected a great many facts bearing upon it which are of
great interest. The sexual instinct in civilised man, he
concludes, has two special seasons of activity—spring and
autumn, but it is most active towards the end of spring as a
rule, in the south of Europe this activity being most marked
somewhat earlier than it is in the more northern countries.

Illegitimate births, it is remarked, are comparatively more
numerous in early spring, and this, it is suggested, is due to
an increase of sexual instinct during May and June.

These conclusions are interesting inasmuch as they indicate
a season or seasons which may be the original sexual seasons ;
but it is the evidence he produces of the sexual seasons of
more savage peoples which is of special interest here.

Some of the Indians of California are stated to have a
regular sexual season, spring being a literal St. Valentine’s
Day with them.

The Watch-and-dies of West Australia and the Tasmanians
have sexual feasts in the middle of spring-time.

The Hos, an Indian hill tribe, have a similar feast, which
becomes a saturnalia during which absolute sexual freedom
is indulged in, in the month of January; while among the
Santals, another hill tribe, marriages mostly take place in

THE ** SEXUAL SEASON’? OF MAMMALS. BY |

January. Among the lower castes of the Panjas in Jeypore
a festival in January is kept up for a month, during which
promiscuous sexual intercourse is allowed. The Kotars, a
tribe in the Neilgherries, have a similar feast marked by
similar licence and debauchery; and the same may be said
for the Keres in New Mexico, the Hottentots, and some
tribes near Nyassa.

In New Caledonia November (that is late spring) used to
be the time when marriage engagements were made, and
among the Rajputs of Mewar the last days of spring are
dedicated to the god of love.

Among the Kaflirs of Cis-Natalian Kafirland more children
are born in August and September than in any other month,
and it seems probable this is due to certain feasts during
which there is unrestricted intercourse between the un-
married people of both sexes.

Among the Bateke—Stanley Pool—most children are born
in September and October—the season of the early rains,—
and it is said the same is the case among the Bakongo.

In Chili the maximum of births occurs in September.

Dalton (1872) gives an account of the Miris, an Indian
hill tribe, which shows that at one season of the year
sexual relations between the unmarried are specially counte-
nanced and indulged in.

My friend Mr. Caldwell tells me that the Queensland
natives with which he was brought in contact have a distinct
sexual season in September (that is spring), and that they
cannot be prevailed upon to do any work for some weeks at
that time of the year.

Cook (1894) records that the Esquimaux which inhabit the
country lying between the seventy-sixth and seventy-ninth
parallels, exhibit a distinct sexual season, which recurs with
great intensity at the first appearance of the sun, and that
little else is thought of for some time afterwards: an account
which is in agreement with statements made by Bosquet
(1885) regarding other Hsquimaux.

Finally Man (1882) notes that the children of the natives

38 WALTER HEAPE.

of the Andaman Islands are said to be born mostly at a
particular time of the year—during the rains.

I have not done more here than simply to indicate the
bearing of a very considerable literature which deals specially
or incidentally with this subject ; one section of this literature
demonstrates by means of statistics, for countries where such
are available, an excessive birth rate in special seasons; the
other shows that the habits and customs of the less civilised
peoples indicate that their sexual and reproductive functions
are specially stimulated at definite times of the year.

While there is some variation in the season for special
sexual activity indicated by the above statements, spring is
obviously the most usual time. Hutchinson (1897) seeks to
show that the time of marriage among certain widely diver-
gent people is largely governed by times of agricultural
plenty; for economic reasons this might reasonably be
expected, though the evidence he brings forward is not at
all conclusive. But it doesnot seem to me to be an important
point. Many reasons, religious or otherwise social, may have
arisen to interfere with such a rule, supposing it ever was a
rule. The importance of the evidence consists in the proof
that any time is or was specially conducive to sexual dis-
turbance, and this, I think, has been proved. (See also
Laycock [1840] and Ellis’s very interesting résumé of this
question.)

The wide variation in the time of the year during which
the sexual season of the lower mammals occurs in different
parts of the world, renders it not surprising that there
should be wide variation in man also in this respect, in
different geographical areas.

However that may be, the fact remains that there is much
evidence in favour of the view that special sexual seasons
were, at one time, universally experienced by the various
races of man, a fact of great importance from a comparative
point of view.

But not only is there evidence of a circumscribed period
for reproduction in the ancestral human being, and in those

THE ** SEXUAL SEASON’? OF MAMMALS. 39

peoples who occupy a low position in the scale of civilisation,
but there is also evidence that the latter produce smaller
families.

In some cases this is ascribed to the practice of infant
marriage, to the strain of child-bearing on a mother who
requires for herself all the nourishment she is capable of
assimilating ; but comparatively small families are usual in
many savage peoples whose women do not become mothers
at a very early age (Westermarck, 1891).

In these cases the result is probably due not only to pro-
longed lactation, or to infant mortality, but to inability to
produce more children; for, as the practice of polygamy
shows, the advantage of large families is fully recognised,
and each individual woman will be required to reproduce as
frequently as possible.

It would seem highly probable, therefore, that the repro-
ductive power of man has increased with civilisation, precisely
as it may be increased in the lower animals by domestication ;
that the effect of a regular supply of good food, together
with all the other stimulating factors available and exercised
in modern civilised communities, has resulted in such great
activity of the generative organs, and so great an increase in
the supply of the reproductive elements, that conception in
the healthy human female may be said to be possible almost
at any time during the reproductive period.

We have come to believe that it is to the regular monthly
menstrual periods, which the human female generally ex-
periences, that this great reproductive power is due. But
the evidence of a regular menstruation with a limited con-
ception period in monkeys, shows that this is certainly not
so. As in monkeys, so in man, these two functions are not
necessarily equally developed.

I think it may fairly be stated that an increase in the
frequency of menstruation is not necessarily a sign of an
increased power of reproduction among women, and that
there is no indication that women who menstruate every two
or three weeks are more prolific than those who menstruate

40 WALTER HEAPE.

every month; in fact, the reverse is probably true, and the
excessive activity of the menstrual organ, if it is not developed
at the expense of the reproductive power, in many cases
results in lessened fertility.

We are here doubtless in the region of pathological condi-
tions, since when there is a considerably increased menstrua-
tion, either by increase of the amount of the menstrual flow
or by decrease of the intra-menstrual period, it 1s accom-
panied by exhaustion and the evils which result therefrom.

If the above be true, it would appear that civilised woman
has reached the limits of reproduction compatible with her
mode of life, and it may be concluded that increased repro-
ductive power will not arise until her powers of assimilation
are increased to a sufficient extent, and until the products of
that assimilation are devoted more exclusively to the repro-
ductive function.

The Duration of the Qstrus in Monostrous and
Polycestrous Mammals in the Absence of the
Male.

There is very little known regarding this point except in
certain domesticated animals. The cestrus of moncestrous
mammals may last a short or a long time. In the Barbary
wild sheep in captivity (Zoo.) it only lasts a few hours. In
the bitch it lasts a variable time, variable both in different
individuals of the same species and in the same individual at
different times. The winter cestrus of the bitch does not last
so long as the summer cestrus in certain breeds ; a well-known
breeder (Dr. Inman) has assured me this is the case with his
bitches. The usual time is probably from seven to nine days.

A most careful observer, however, tells me that a bitch
which he had for many years usually remained in a condition
of cestrus for nine days, but sometimes it persisted in her
for fourteen days. Other breeders have informed me they
have had bitches undergoing cestrus for even a longer period
than this, but it is undoubtedly an exceptional experience.

THE ** SEXUAL SEASON’? OF MAMMALS. Ad

In certain bloodhounds a well-known breeder (Mr. Brough)
has observed cestrus to last twenty-one days, but only very ex-
ceptionally,and not asa characteristic of any particular bitch.

There can be little doubt the persistence of cestrus in
bitches may be influenced by their temperament, by their
food, and by the particular conditions of existence expe-
rienced by each bitch.

Wolves, jackals, and foxes in the Zoological Gardens have
about the same duration for cestrus as the average bitch,
from seven to nine days. In the cat cestrus lasts nine to ten
days (Hamilton, 1896); in tigers in captivity (Zoo.) for
eight days at the longest. In bears, on the other hand (Zoo.),
cestrus appears to last continuously for two to three months ;
it must be recollected, however, that this occurs with females
kept together with males under conditions which, while they
may very probably excite sexual feelings, do not result in
gestation.

Among wild animals the duration of the cestrus can only
be assumed by comparison with other individuals of the
Same species in captivity; although the duration of the
sexual season may be inferred from the known season during
which parturition takes place, the duration of the cestrus
cannot be so determined.

Among polycestrous mammals there is not such great varia-
tion in the duration of cestrus, since, instead of a long period
of cestrus, these animals exhibit a recurrence thereof ; still
there is some difference apparent: the domestic sheep has
cestrus for only a few hours, say twelve hours; the cow for
not more than twenty-four hours as a rule; while antelope,
deer, and wild cattle in captivity (Zoo.) closely imitate
domestic cattle and sheep in this respect.

The mare endures cestrus probably for a slightly longer
period as a rule, but this depends very much on the tempera-
ment of the individual mare, and the conditions under which
she is kept.

The elephant in the Zoological Gardens has persistent
cestrus for probably three to four days.

42 WALTER HEAPH.

In monkeys the cestrus has not usually been carefully
noticed, but | am assured that the Moor macac in confinement
(Zoo.) has a definite cestrus which lasts two or three days ;
and inafew other monkeys a similar condition has been from
time to time noticed (Ellis).

In the human female there is not wanting evidence of a
similar condition (Aristotle; Martin, 1888; Haycraft, 1880),
and on this point information has been supplied to me by
various experts, which leads me to think it will probably be
found that those women who are most robust, and who suffer
least from the enervating effects of civilised life, experience
a condition comparable to that of cestrus in the lower mam-
mals (confer also Hllis).

The Effect of Maternal Influences on the Sexual
Season and on (Ustrus.

Maternal influences may or may not completely disorganise
the sexual season; this depends upon whether or not they
interfere with its recurrence or with that of cestrus.

Gestation.—Gestation may or may not interfere with the
recurrence of one or other of these factors. In the dog it does
not do so, because the dog has only one cestrus during each
sexual season, and the period between two sexual seasons,
i.e. the ancestrum, is longer than the period of gestation.
In the elephant it does do so, because the gestation period is
longer than the ancestrous period. So also with badgers
this appears probable (Denwood, 1894). In camels, whose
gestation occupies thirteen months, the sexual season is inter-
fered with by gestation, and is on that account put off for
another year. ‘The camel conceives every two years (Swayne,
1895). Inthe rat, on the other hand, gestation does not
interfere with the recurrence of the sexual season, but does
interfere with that of cestrus, because the rat has a series of
dicestrous cycles in each sexual season, and she may also
undergo a series of gestation periods during that time, and
because the maternal generative cycle (twenty-one days) is
longer than the dicestrous cycle (ten days).

THE “‘ SEXUAL SEASON” OF MAMMALS. 43

But whenever gestation occurs if encroaches upon, if it
does not entirely absorb, the ancestrum ; that is to say, it re-
duces the period during which the generative organs would
lie fallow if the sexual season were a barren one. Thus in
the case of the mare, a barren sexual season may consist of
a series of dicestrous cycles lasting for as long as six months,
in which case the ancestrum lasts six months also, after which
another sexual season again begins.

A reproductive sexual season, however, results in a period
of eleven months’ gestation; interfering not only with the
dicestrous cycles which would have recurred if conception
had not taken place, but also absorbing practically the whole
of the ancestrum ; for, nine days after parturition, the ma-
jority of mares again experience cestrus.

Nursing.—Nursing also may or may not interfere with
the recurrence of the sexual season and of cestrus. The
rat suckles her new-born litter of young while pregnant
with another litter; so also does the domestic rabbit and
guinea-pig, and probably many rodents. ‘lhe mare also,
as a rule, readily becomes pregnant while suckling her
newly born foal. Here, however, there is some evidence of
variation, for I am informed, by a breeder of large shire
horses in the west of England, that many of the mares in his
stud become pregnant only once every two years; the drain
on the system, in consequence of gestation and nursing, in
these large animals being, apparently, too great to admit of
the immediate recurrence of another sexualseason. Another
breeder of shire horses, however, assures me that he gets a
foal each year from his mares.

On the whole there is some reason to believe that, unless
these large mares are exceptionally carefully tended, they
are lable to miss bearing annually, from time to time.

A few instances may be given here of animals in the wild
state which do not bear young every year. ‘The grizzly bear
in British Columbia bears young only every second year
(Somerset, 1895). The wild yak in the Tibetan desert only
produces a calf every second year (Prejevalsky, 1876), and

44, WALTER HEAPE.

the same is probably true for the Greenland musk ox
(Lydekker, 1898); while the walrus, which goes nearly
twelve months with young, nurses her calf or provides it
with food for two years (Bell, 1874), and during that time
ancestrum appears to persist.

Similar evidence of variation is to be found in the human
female. Among the Esquimaux in high latitudes children
are nursed from four to six years, and women bear children
about every four years (Cook, 1894). It is not uncommon to
hear of women of various tribes purposely prolonging the
nursing period in order to avoid too frequent breeding. The
Waganda women uurse their children until two years of age,
and live apart from their husbands from the time of con-
ception until the child is weaned (Felkin, 1885). The
Andaman Island native women nurse their children as long as
they can (Man, 1882). On the other hand, it is recorded
that among the North-west Central Queensland natives
nursing may go on for three, four, or five years, and a mother
is frequently seen with two children of different ages at the
breast (Roth, 1897). Among more civilised women menstrua-
tion is frequently in abeyance during the nursing period,
nevertheless many women menstruate while lactation is still
possible. Such a possibility is not confined to women among
menstruating animals. I have seen a monkey, Macacus
cynomolgus, in the gardens of the Zoological Society at
Calcutta, which menstruated regularly while still suckling a
young one.

The whole question of lactation and its relation to sexual
phenomena, more especially gestation, is of great interest,
all the more perhaps when it is remembered that virgin
bitches frequently secrete milk in sufficient quantities to
interfere with their work (foxhounds), while mules have
been known to nurse successfully the foal of a mare; but for
our present purpose sufficient has been said, and in conclusion
it may be argued that when nursing encroaches upon the
sexual season, the recurrence of the latter depends upon the
vigour of the mother and her powers of recuperation.

THE ‘°* SEXUAL SEASON’? OF MAMMALS. 45

The Pro-cestrum,

The pro-cestrum, as I have already stated, is the forerunner
of cestrus. Evidence of it is to be seen in each of the large
groups of the Vertebrata, fishes, amphibia, reptiles, birds, and
mammals (Wiltshire, 1883), and it must be regarded in all of
them as a sign of the preparation of the generative system
for the sexual act.

Pro-cestrum is usually associated, in the minds of breeders,
with reproduction, to an extent which entails the supposition
that the same stimulus which incites the former also causes
the latter ; but the fact that pro-cestrum may occur normally
without the concurrent production of ova shows that the two
functions are not always interdependent, and that what
serves as sufficient stimulus for sexual desire is not necessarily
sufficient for reproduction.

A consideration of these relationships belongs rather to the
study of ovulation than to the subject-matter of the present
paper. I would merely remark here that while the ovary
probably does participate to some extent in the excitement
evidenced by pro-cestrum, this function in mammals must be
considered as evidence mainly of sexual rather than of repro-
ductive power.

Pro-cestrum is first evident in the tissue of the external
generative organs and the surrounding parts, and while it
increases in intensity there, it extends to the uterus; during
this time certain changes (to be mentioned below) take place
in the uterine tissue, and they are followed by the subsidence
of the disturbance, first in the uterus and finally in the
external generative organs.

The length of time during which pro-cestrum lasts is ex-
tremely difficult to determine ; there seems to be considerable
variation in different animals, and in the same animal at
different times; but that may be due to variation in the
intensity of the external evidence rather than to variation in
the duration of the pro-cestrum itself.

In the rabbit I have observed this period lasts, probably,

46 WALTER HEAPE.

one to four days; in the bitch seven to twelve days (Stone-
henge, 1887); in the chimpanzee six to eight days (Keith,
1899). In cattle and sheep the external evidence of pro-
cestrum is difficult to determine, and cestrus appears to follow
very quickly upon the former, about one day after or less.
Pigs, on the other hand, exhibit external signs of pro-cestrum
somewhat longer, while mares are very variable in this respect.

A further consideration of the subject is divided into the
external and internal evidence of pro-cestrum.

The External Evidence of Pro-cstrum in Mam-
mals.—The first sign of pro-cestrum noticed, in the lower
mammals, is a swollen and congested vulva, and a general
restlessness, excitement, or uneasiness. There are other
sions familiar to breeders of various mammals, such as the
congested conjunctiva of the rabbit’s eye, and the droop-
ing ear of the pig, which are considered by some as even
more reliable indications of the probability or capability
of conception than is afforded by the vulva alone. Many
monkeys (Heape, 1894, 1897, Keith, 1899) exhibit conges-
tion of the face and nipples, as well as of the buttocks, thighs,
and neighbouring parts; sometimes they are congested to a
very marked extent, and in some species a swelling, occa-
sionally prodigious, of the soft tissues round the anal and
generative openings, which is also at the time brilliantly
congested, indicates the progress of the pro-cestrum.

The Pro-cstrous Discharge and Menstrual Flow.
—Following the swelling and congestion of the external
eenerative organs, there is, in most animals, a discharge from
the generative canal. The discharge may consist merely of
mucus from the uterine glands and from the glands of the
cervix and from those in the neighbourhood of the vaginal
orifice, of the products derived from the breaking down of
epithelial tissue, and of fragments or small masses of pave-
ment epithelium from the vagina; such a discharge is usually
to be seen in the rat and mole.

In addition, fragments or small masses of columnar uterine
epithelium may be observed in various animals. Again, to

THE ** SEXUAL SEASON’? OF MAMMALS. 47

the above, blood may be added for a large number of animals,
some of which rarely, some frequently, and some always
suffer from a loss of blood. While, finally, more or less com-
pact masses of uterine stroma tissue are included in the
discharge of the Primates and some of the lower mammals.
Blood has been observed in the discharge during pro-cestrum
in the mare, ass, cow, sheep, goat, pig, cat, rabbit (Aristotle ;
Ellenberger, 1892; Fleming, 1878; Wiltshire, 1883), and rat
(Lataste, 1887) ; it is also recorded as having been observed
in marsupials (Wiltshire, 1883); in the bitch it is almost
invariably present, and so also it would appear to be in
Pachyuromys duprasi, Dipodillus simoni, Meriones
shawi (Lataste, 1887), and in Tupaja javanica and
Tarsius spectrum (Stratz, 1898). In mostof these animals
there is only enough blood to tinge the discharge more or
less, but in the bitch, and probably T. javanica and T.
spectrum, there is a flow of blood almost as concentrated
as that recorded for monkeys (Heape, 1894, 1897).

It has been recorded for a large herd of highly bred
Alderney or Jersey cattle in the south of England, that a
discharge of blood is of regular recurrence among’ them
(Wiltshire, 1883); but so far as I can learn this is excep-
tional, although its occurrence in individuals is by no means
rare. It has been suggested that bleeding in the lower
mammals during pro-cestrum is confined to domesticated
species, but this is not true (Lataste, 1887; Stratz, 1898;
Wiltshire, 1883) ; at the same time it is not improbable that
the circumstances attending domestication tend to increase
the flow of blood from the uterus, and that highly bred
domesticated animals are more liable to experience it than
those which are hardier, less carefully attended to, and less
luxuriously fed.

The pro-cestrous discharge, then, varies in quality in dif-
ferent animals, and not only is this true, but it varies at
different times in the same animal, both as to quantity and
quality. There is ample evidence of this in various human
tribes (Holder, 1892, Hillis) and in individuals. Among

48 WALTER HEAPE.

domesticated animals, mares, cows, sheep, and rabbits do not
always experience a loss of blood ; further, individual animals
of these species sometimes experience a much more profuse
flow than at other times, or they may experience a profuse
flow only rarely or not at all.

The Internal Phenomena of Pro-cstrum.—lIt will
be convenient first to abstract the account I have given else-
where (Heape, 1894) of the changes which take place in the
uterus of the monkey during pro-cestrum, and then to com-
pare these changes with those which occur at that time in the
human female on the one hand, and in the lower animals on
the other.

a. Period of Rest.—Stage I. The resting stage. This is
the period before pro-cestrum occurs, and at that time the
uterine mucosa is a shallow bed, opaque, white, and anemic.

B. Period of Growth.—Stage II. The growth of stroma.
It isnow that pro-cestrum first becomes apparent; the uterine
stroma thickens, hypertrophy takes place, and it becomes
semi-transparent, soft, and flabby.

Stage III. The increase of vessels. The growth of the
stroma tissue is rapidly followed by an increase in the number
and size of the vessels of the stroma, the whole becomes
richly supplied with blood, and the surface is flushed and
highly vascular. This process goes on until the whole of
the uterine stroma becomes tense and brilliantly injected
with blood.

c. Period of Degeneration.—Stage 4. The breaking
down of vessels. The walls of the superficial vessels now
break down, and the blood contained therein is extravasated
throughout the superficial portion of the mucosa.

Stage V. The formation of lacune. The extravasated
blood becomes gradually collected in lacune, which at first
lie within the stroma, but gradually become enlarged and
project as rounded hillocks, bounded superficially by the
uterine epithelium, into the cavity of the uterus.

Stage VI. The rupture of lacunez. The superficial mucosa
cells, isolated or in patches, now begin to degenerate ; they

THE ‘SEXUAL SEASON”’ OF MAMMALS. 49
are cut off, as it were, by the extravasated blood, from the
deeper mucosa cells, and they shrivel up and die. Soon the
uterine epithelium follows suit and, with the degeneration of
its cells, loses its continuity and ruptures, thus allowing the
blood hitherto contained to pour into the uterine cavity.

Stage VII. The formation of the menstrual clot. With
the blood which is poured out from the ruptured lacune is
mixed also degenerated epithelium cells, isolated or in strings ;
and as the tissue below is laid bare, the extravasated blood in
the deeper parts of the mucosa, together with stroma tissue
and the superficial portion of uterine glands, also collects in
the uterine cavity, and the whole forms therein a more or less
dense clot. Some of the blood and degenerate uterine tissue
oozes out through the os uteri to the vagina and thence to
the exterior while the process is in progress, but there is fre-
quently left behind until a later stage a clot, which in some
cases entirely fills the uterine cavity.

Dp. Periodof Recuperation.—Stage VIII. The recupera-
tive stage. While the clot is still within the uterus, a new epi-
thelium begins to grow over the, now much reduced, uterine
stroma. At the same time new capillary vessels are formed, the
extravasated blood which stillremains in the tissues is collected
therein, and brought back into the circulatory system.
During this period the clot is expelled, and subsequently the
uterus assumes again the appearance first described, and
eventually becomes again at rest. It is at or towards the
close of this period that cestrus normally occurs.

For the human female the histology of pro-cestrum (men-
struation) has never been so fully worked out in healthy normal
uteri. Many observers have described isolated specimens,
and most of them have had recourse to material which has
either been obtained some time after death, or from indi-
viduals suffering from diseases which may well have produced
pathological changes in the uterine tissue. Then, again, the
extent of menstruation varies in different peoples and indi-
viduals, and in the same individual at different times. The
amount of the menstrual flow and the quality of that flow also

VoL. 44, PART [.—NEW SERIES. D

50 WALTER HEAPE.

varies, to such an extent, indeed, that, while some women lose
a large amount of blood at each pro-cestrum, others sometimes
and some never lose any at all. It is not surprising, there-
fore, to find that while some observers hold that no change
takes place in the uterine tissue during pro-cestrum, others
state that highly specialised decidual tissue is formed at that
time; while some deny that even a portion of the uterine epi-
thelium is lost by denudation during pro-cestrum, others
maintain that the whole of the mucosa layer is discarded
during that process.

The question has been somewhat fully discussed by me in
a former paper (Heape, 1894), where an account is also given
of the more important literature of the subject. Here it is
only necessary to add the conclusions arrived at, which are
that in all essential points the menstruation or pro-cestrum of
the human female is identical with that of monkeys. More
recently I have described (Heape, 1898) two menstruating
human uteri, the first of which shows congestion and is closely
comparable to Stage IV of the monkey, while the second
shows denudation, and appears to be practically identical
with Stage VII of the monkey.

A slightly earlier condition of denudation in the human
uterus has been described and figured by Minot (1892), and
again supports the view above expressed.

Among lemurs, Stratz (1898) has described what he calls
bloody “‘ menstruation” for Tarsius spectrum. I gather
that, in this animal, denudation of the epithelium of the uterus
takes place and that Stage VII exists; but there is no proof
that denudation extends to the stroma tissue, and therein possi-
bly les the difference between lemurs and monkeys, otherwise
there can be little doubt of the homology of the process in
these two animals.

Stratz has also described the existence of a blood-clot and a
“menstrual”! flow in Tupaja javanica, and here again the

' The use of the term “ menstrual” flow, as it is here used, to denote a

flow of blood from the uterus, without regard to the periodicity of that
flow, is to be deprecated.

THE ‘* SEXUAL SEASON’? OF MAMMALS. 51
tissue contained in the clot apparently consists only of
desquamated epithelium.

Retterer (1892) has given a more detailed account of what
happens during the pro-cestrum of the bitch. During period
A, of rest, Stage I, the mucosa of the uterine horns is firm,
pale, and of a thickness of °3 to*5 mm.; but with the com-
mencement of pro-cestrum, period B, there is a well-marked
Stage, II,in which the mucosa grows rapidly to three or four
times its original thickness, and becomes soft and spongy.
Stage III is also well marked, and the mucosa becomes
injected with numerous vessels distended with blood. ‘Then
period ¢ occurs, and Stage IV is marked by the breaking down
of the vessels and extravasation of the blood in the mucosa
tissue. Lacunze are formed, Stage V, which, during Stage
VI, rupture, and pour the contained blood into the uterime
cavity.

So far the similarity of the progress of the pro-cestrum in
the bitch is practically identical with that of monkeys, but
there is no blood-clot formed, and Retterer’s account renders
it doubtful whether any denudation, even of epithelium, takes
place. He himself thinks not. I have myself worked out
the history of the pro-cestrum of the bitch to some extent,
and have satisfied myself that Retterer’s account is true in all
essential details.

I have also failed to find any area of the uterine mucosa
which has been denuded of epithelium, and do not believe
that this process occurs to any extent; at the same time,
where lacune rupture there must be loss of epithelium, though
I think denudation is confined to these spots.

The pigmentation of the uterus, described by Altmann
(1878), is further evidence of the probability that much of
the extravasated blood is not discharged into the uterine
cavity, but is retained in the uterine tissue and absorbed from
thence.

‘The homology of this process in the bitch with that already
described for monkeys is absolutely certain, and if nothing
more were known, would establish the identity of the pro-

52 WALTER HBEAPE.

cestrum in these animals; or, in other words, the homology
between the pro-cestrum (so-called “ heat”) of the lower
mammals and the menstruation of the Primates.

he absence of Stage VII, the menstrual clot, is not to be
wondered at in a large bifid uterus ; the denudation of tissue
in sufficient quantity to form a clot would be a very severe
operation in such a comparatively large organ.

The only other paper dealing with this subject, for the
bitch, with which I am acquainted (Johnstone, 1888), treats
of what the author calls the “corpuscular development ” of
the mucosa of the bitch during the pro-cestrum, but I do not
gather the author has satisfactorily demonstrated the truth
of the view he advocates (see also Johnstone, 1895).

Pouchet’s description of the changes in the uterus of the
sow during pro-cestrum shows the existence of Stages II and
III (Wiltshire, 1883) ; he does not describe the breaking down
of vessels or the formation of lacunz, but his description of
the histology of the uterine discharge shows that it contains,
besides mucus, both blood and uterine epithelium. Stage IV,
therefore, is assuredly represented, and there can be little
doubt that Stages V and VI are also passed through, since
there must have been rupture of the uterine tissue in order
that pieces of it should be contained in the discharge.

Ellenberger’s (1892) account of the changes which take
place during pro-cestrum in domestic mammals includes
Stages II and III; he also does not describe Stages IV, V,
and VI, but he records the presence of both blood and epi-
thelial cells in the discharge, and these stages must therefore
have been passed through, although denudation is in all
probability very slight. Fleming (1878) adopts the view
that, among ruminants, the blood which finds its way to the
exterior exudes from the cotyledons; while Hllenberger
describes pigmentation there, and states it is caused by
the blood left behind in that tissue after pro-cestrum has
occurred.

Bonnet (1892) also describes Stages II and III in various
domestic mammals during pro-cestrum, but he also adds

THE ‘© SEXUAL SEASON’’ OF MAMMALS. 53

Stage IV for ruminants, horse and pig, and where external
bleeding is seen in these animals the occurrence of the
equivalent of Stages V and VI is essential.

Kazzander (1890) notes the existence of extravasated
blood (Stage IV) in the sheep’s uterus during pro-cestrum, at
a period before external bleeding is noted ; so that when the
latter occurs, a condition equivalent to Stage VI is passed
through by thisanimal. Both this author and Bonnet (1880,
1882), whom he quotes, describe pigmentation in the uterine
mucosa of the sheep, and consider it is due to the extrava-
sated blood which is not discharged during the pro-cestrum.

Lataste (1887) describes desquamation of uterine epithelium
in several Muridz, and states that it takes place independently
of pro-cestrum (or, as he callsit, “ rut ”), during which Stages
If and III are noted, and at the close of which a bloody dis-
charge (which he calls “ menstruation”) is evident. Stages
IV, V, and VI are therefore probably also passed through
in the case of these animals.

I find it difficult to determine exactly what this author
means, but I gather it is his opinion that in these animals
there is a periodic shedding of uterine and superficial vaginal
epithelium, and that this precedes and is independent of the
pro-cestrous discharge (p. 163) ; if this be so it is quite dif-
ferent from anything which has been already described for
any other of the lower mammalia, and is comparable only to
that somewhat rare phenomenon, exfoliation of the vagina, in
the human female. The same author declares (1898) there is
a rhythmical transformation of the epithelium of the vagina in
certain of the lower mammalia, which is in connection with
rhythmical generative changes; he describes the denudation
of this epithelium, and its recuperation from the lower layers.

The subject has been very rarely investigated in the lower
mammals, and still more rarely has it been attacked from a
histological point of view; isolated specimens have been
described with more or less detail, but no attempt has been
made to work out the history of the process by any one, so
far as I know, but Retterer.

54 WALTER HEAPE.

For this reason the evidence available is fragmentary, but —
it is remarkably consistent ; and although further researches
may, and probably will, show variations in detail, the broad
fact of the homology of the internal process of pro-cestrum
in all mammals is sufficiently demonstrated.

This we may summarise as follows: the uterus of all
mammals during the quiescent period is comparatively
anemic, and its mucosa is a thin layer; it has at that period
the appearance of lying fallow.

During pro-cestrum hypertrophy of the mucosa first takes
place, and is followed by congestion, which results usually in
the rupture of the superficial vessels and consequent ex-
travasation of the blood into the surrounding tissue ; in some
cases this extravasated blood finds its way into the cavity of
the uterus and thence to the exterior, with either more or
less denudation of the superficial mucosa, while in other
cases there is no external hemorrhage, and the extravasated
blood is absorbed in situ. While, therefore, neither the
discharge of blood nor the extravasation of blood is an
essential feature of the pro-cestrum, the hypertrophy and con-
gestion of the mucosa is invariably present in all mammals,
a condition which we may confidently expect to find also in
the lower Vertebrata.

The Period of @strus.

The period of normal cestrus, as I have stated in the intro-
duction to this paper, occurs as a result of pro-cestrum.

As arule breeders regard cestrus (the period of desire) as
an attendant condition of pro-cestrum rather than as a result
thereof; where there is no discharge evident there is some
excuse for this .view, especially as, even when a discharge
does occur, cestrus may happen before the discharge com-
pletely ceases. Cistrus, however, is possible only after the
changes due to pro-cestrum have taken place in the uterus.
A wave of disturbance, at first evident in the external
generative organs, extends to the uterus, and after the various
phases of pro-cestrum have been gone through in that organ,

THE ‘° SEXUAL SEASON’? OF MAMMALS. 55

and the excitement there is subsiding, it would seem as if
the external organs gain renewed stimulus, and it is then
that cestrus takes place. If the uterine changes are confined
to Stages II and III, that is simply hypertrophy and conges-
tion of the mucosa, cestrus rapidly follows the first external
signs of pro-cestrum ; but if more elaborate disturbance takes
place in the uterus, the period of cestrus is delayed.

Thus it is during the subsidence of the uterine disturbance
that cestrus takes place. The period during which the dis-
charge continues is not, however, a true indication of the
permanence of the uterine disturbance. In comparatively
large uteri, especially in those which extend as long horns
from the corpus uteri, the area of denudation or hemorrhage
may be situated far from the vagina; and the products of
that hemorrhage and of denudation may take a considerable
time to find their way to the exterior; this is especially the
case where there is little blood and much mucous discharge.

We have seen above that in the monkey, Stage VIII, a
new epithelium is formed over the surface of the newly
denuded uterus before the blood-clot is evacuated; and in
the same way, before the discharge from long-horned uteri
reaches the exterior, the uterine disturbance is largely
allayed, and renewed stimulus may be supplied to the ex-
ternal generative organs.

In all animals which have been investigated, coition is not
allowed by the female until some time after the swelling and
congestion of the vulva and surrounding tissue is first
demonstrated, and in those animals which suffer from a
considerable discharge of blood, the main portion of that
discharge, if not the whole of it, will be evacuated before
sexual intercourse is allowed.

Thus in Pachyaromys duprasi, which experiences
hemorrhage, coition is not allowed during the flow (Lataste,
1887).

Bitches, except rarely, receive the dog only after bleeding
is over (Stonehenge, 1887), although a mucous discharge,
which frequently continues after the discharge of blood

56 WALTER HEAPE.

ceases, may be still in progress at the time coition is per-
mitted by the bitch (Millais).

The Moor mace in the Zoological Gardens has a definite
cestrus, which always occurs shortly after the menstrual
discharge ceases, and which lasts for two or three days; and
there is strong reason for believing this is also the case with
various other monkeys, as, for instance, the orang-utang
(Hillis).

The human female frequently experiences cestrus with
marked strength after menstruation is over (Martin, 1888),
more especially, it would appear, in those individuals who
do not suffer from excessive menstruation,—in other words,
in those whose generative system is least disturbed by the
consequences of civilisation and social life. ‘This special
time for cestrus, in the human female, has very frequently
been denied, and no doubt civilisation and modern social
life do much to check the natural sexual instinct where
there is undue strain on the constitution, or to stimulate
it at other times, where extreme vigour is the result.

For these reasons a definite period of cestrus may readily be
interfered with, but the instinct is, I am convinced, still
marked. Ellis quotes various authors who hold a similar
view, but they do not all agree as to the time when estrus
occurs ; if, therefore, the views which I have advocated here
are correct, it would seem probable that abnormal cestrus
has been mistaken for true cestrus in many of these cases.

Summary and Conclusion.

Introduction.—After criticismg the terms commonly
used to denote the various stages of the ‘sexual season” of
mammals, I have defined the terms used in the present
paper.

Female mammals are divided into two classes, “ mones-
trous”’ and “ polycestrous”” mammals, and I have explained
that, in the absence of the male, “ pro-cestrum,” “ cestrus,”
and “‘metcestrum”’ are followed by “ dicestrum” in polyes-
trous mammals, during the recurrence of the “ dicestrous

THE ‘“* SEXUAL SEASON”? OF MAMMALS.
cycles,” and by “ancestrum” in monstrous mamma s
always, and in polycestrous mammals at the close of the
sexual season.

The difference between the dicestrous or ancestrous cycles
in the absence of the male, and the “ maternal generative
cycle” when cestrus is followed by insemination, fertilisation
of the ovum, and gestation, is drawn attention to.

The occurrence of abnormal cestrus is noted.

The Breeding Season of Mammals is merely touched
upon; inasmuch as it concerns what happens during both the
sexual season and the gestation period jointly, its full con-
sideration is not possible in this paper.

The Sexual Season of Male Mammals.—Males are
divided into two classes, those which have a special sexual
season, “rut,” and those which are sexually capable all the
year round. The influence of captivity is touched upon, and
it is shown that, while sexual activity is not so violent in
captive animals as in those in the wild state, it may be
much more frequently or continuously exhibited.

The Sexual Season of Female Mammals.—This is
considered in wild mammals in a state of nature, in those
which are captive, and in domestic mammals ; and the effects
of climatic, individual, and maternal influences are drawn
attention to.

Among moncestrous mammals the effect of these influences
may be to increase or decrease the periodicity or the duration
of the sexual season, while among poly ‘* °° xammals the
effect may be to increase or decrease the 1numper of dicestrous
cycles in each sexual season or the duration of each cycle;
the effect in both classes of animals being to increase or
decrease their reproductive power.

It is pointed out that the knowledge at present available
throws no light on the origin of the sexual season; but that
it is due to astimulus which appears to be gradually collected,
that itis associated with nutrition, and is manifested by excep-
tional vigour and bodily ‘ condition ” seems assured.

The Periodicity of the Sexual Season in Mones-

58 WALTER HEAPE.

trous Mammals in the Absence of the Male.—This
is shown to be affected by climatic and by individual influ-
ences, to be more frequent in domesticated than in wild
animals of the same species, and to be variously affected by
captivity.

The Duration of the Sexual Seasonin Polyestrous
Mammals in the Absence of the Male.— The sexual
season in these animals is affected by the duration and the
recurrence of the dicestrous cycle; as in moneestrous mam-
mals, it is shown to be affected by climatic and individual
influences, by domestication and captivity.

It is here that we are first brought into contact with monkeys
and man, and I have endeavoured to show that the sexual
season which undoubtedly exists in monkeys, exists also in
certain human peoples in the present day, while there is some
evidence that, in the past, all peoples were similarly affected,
and that a definite sexual season was the rule. The fact
that, in spite of the regular recurrence of the cestrus, monkeys
have only a limited season during which conception is possible,
is drawn attention to. It is pointed out that this is due to
the fact that the ovary is not active all the year round,
and evidence is brought to show that the function of ovula-
tion is also not necessarily coincident with cestrus in various
other mammals. This condition is apparently due to the
want of sufficient energy for both cestrus and ovulation.

The Duration of the Gstrus in Monestrous and
Polyestrous Mammals in the Absence of the Male.—
Knowledge of this point is practically confined to domesti-
cated mammals and to certain animals in captivity, and the
evidence, which a study of these animals renders available,
shows that the duration of cestrus is very variable, not only
in different species, but also in different individuals of the
same species, and in the same individuals at different times.

There is greater variation in this respect among moncestrous
than among polycestrous mammals, as a rule.

The Effect of Maternal Influences on the Sexual
Season and (istrus,—These may or may not completely

THE ‘© SEXUAL SEASON’? OF MAMMALS. 59

disorganise the sexual season, and this depends on whether
or not they interfere with its recurrence or with that of
cestrus.

The above is true for both moncestrous and polycestrous
mammals, for both gestation and nursing; but whereas
gestation interferes with the recurrence of the cestrus, only
if it extends over the time which would otherwise be a sexual
season, the interference of nursing depends upon the vigour
of the mother and her powers of recuperation.

The Pro-cestrum.—Evidence of pro-cestrum is to be seen
in all Vertebrata, and is the forerunner of cestrus. It is
first noticeable in mammals in the external generative
organs, and extends thence to the uterus.

The essential manifestations thereof are first hypertrophy,
and secondly congestion of the tissues affected, and this is
very usually, indeed probably always, followed by a
discharge.

The discharge always consists partly of mucus from the
uterus, and partly of desquamated vaginal epithelium and
the products of broken-down epithelial tissue.

In some animals always, and in others sometimes, blood is
also evacuated, which has its origin from the uterine mucosa,
in which case there is always more or less of uterine tissue
also contained in the discharge.

There is very considerable variation in the extent of both
hypertrophy and congestion of the tissue in various mammals,
but it is essential to note that these phenomena are to some
extent always present, and are frequently combined with
the rupture of the congested vessels in the mucosa, and also,
more rarely, with a discharge of blood from, and still more
rarely a denudation of, the superficial uterine mucosa.

The evolution of the pro-cestrum in its most advanced form,
that is to say the menstruation of the Primates, from the
simplest form, as it appears in such animals as the mole, is
traced, and menstruation is shown to be identical with
“heat.”

The Period of Gistrus.—This is possible only after

60 WALTER HEAPE.

the active changes due to pro-cestrum have taken place in the
uterus ; it is always present, under normal conditions, in the
lower mammals at that time, and is much more frequent then
in the Primates than is generally supposed.

Conclusion.

The conclusions I draw from the evidence detailed above
are then, very briefly, as follows :

A sexual season is common to all female mammals; its
recurrence may be interfered with in consequence of climatic,
individual, or maternal influences, and it may be modified by
the influences attending captivity, domestication, or civilisa-
tion.

The modification brought about by one or other of these
various influences is not necessarily the same in different
species of the same genus, nor in different individuals of the
same species, nor even in the same individual at all times ;
but whatever differences there may be, they are merely
modifications of the same plan.

The sexual season of all mammals is evidenced by a series
of phenomena which constitute, in the absence of the male,
one cestrous cycle (moncestrous mammals) or a series of
cestrous cycles (polycestrous mammals); animals usually
moncestrous may, under certain circumstances, show a
tendency to polycestrum ; in the same way animals usually
polycestrous may show atendency to moncestrum. These two
conditions are very closely related, and the main difference
between them is the method by which the reproductive power
is increased.

The various constituent parts of an cestrous cycle are in-
variably demonstrated in all mammals; there is in all of them
a period during which the generative organs are hypertro-
phied and congested (pro-cestrum), followed by a period of
desire for coition (cestrus), which, in the absence of the male,
gradually dies away (metcestrum), and results in a period of
rest (dicestrum or ancestrum). When this period of rest
merely separates two recurrent dicestrous cycles it is brief,

THE ‘© SEXUAL SEASON” OF MAMMALS. 61

and I have called it the dicestrum; but where it serves to
separate two sexual seasons it persists for a considerable
length of time, and I have called it the ancestrum.

The pro-cestrum is always associated with hypertrophy and
congestion of both external and internal sexual organs and
the uterus, and with a discharge from the generative orifice.
These phenomena are common to all mammals; they may,
however, be further complicated. These complications may
include rupture of the congested vessels of the hypertrophied
superficial uterine mucosa, and extravasation of the blood
contained therein ; they may include a discharge of this blood
into the uterine cavity, and from thence to the exterior; and
even more or less denudation of the mucosa may take place,
leading to the formation of a menstrual clot.

The rupture of the vessels of the mucosa and the subse-
quent phenomena are not experienced by all mammals; they
are supplementary to the essential factors of pro-cestrum, and
occur in part rarely in some animals, in part always in some
animals, and in a complete sequence only, so far as is known,
in Primates.

That the pro-cestrum of Primates is identical with the pro-
cestrum of other mammals does not, however, admit of any
doubt ; there is ample evidence of this in the various inter-
mediate conditions of other mammals, by means of which,
and bearing in mind the influence of domestication and
civilisation on polycestrum, the evolution of the menstruation
of monkeys and of the human female from the pro-cestrum of
the lower mammals can be surely traced. A further evidence
of this is the time of the occurrence of cestrus. It is mani-
fested at a certain period after pro-cestrum, and has a certain
relation to it—that is, it follows and is not coincident with
pro-cestrum in the lower mammals, as is usually supposed.

In some monkeys the same relation of cestrus to pro-cestrum
obtains, and in others it is probably so, while in the human
female there is evidence of a similar condition, especially,
probably, among normally strong individuals who lead a
healthy life.

62 WALTER HEAPE.

Thus the human female may exhibit a sexual season, a
pro-cestrum, and a period of cestrus, precisely like any other
mammal, and the homology of these processes in all mammals
is, in my opinion, established.

A review of the literature which treats of the relation be-
tween ‘‘heat” or “rut,” as it is usually called, and men-
struation, resolves itself practically into an enumeration of
those who deny there is any ground for comparison, and those
who assert they are identical processes. I do not propose to
enter into a detailed criticism of the voluminous literature
which bears upon the subject, but will content myself with
quoting the essence of the most frequent assertions which
are made for and against the homology of these processes,
and with briefly replying to them.

Those who uphold the homology do so because—

I. There is congestion of the generative organs during
both “heat” and menstruation.

II. There may be a recurrence of ‘“‘ heat” as there is a re-
currence of menstruation.

III. The discharge during ‘ heat” may be of a menstrual
character.

IV. From a phylogenetic point of view the homology is to
be expected.

These statements may be disposed of together; so far as
they go they are true enough, but they are not in themselves,
separately nor collectively, conclusive evidence.

Those who deny the homology do so because—

1. The discharge during “heat” in the lower animals is
said to be mucus, while in the human female it is mostly
blood.

2. The time of “heat” is said to be the only time the
lower animals will permit of coition, while sexual union
during menstruation is a very rare occurrence.

3. “Heat” or “rut”? is said to occur in both males and
females in the lower animals and to depend upon the seasons,
whereas in the human species it is said to be not so.

4, After “heat” the female of the lower animals is said

THER ** SEXUAL SEASON’? OF MAMMALS. 68

to refuse the male, whereas in the human female sexual
desire is not confined to the time of menstruation.

5. “Heat” is necessary to the production of the species in
the lower animals, while in woman “desire” is said to be
not essential to conception.

6. In the lower animals the ovaries are said to contain
ripe ova only during the time of “heat,” whereas ripe ova
are said to be found in the human ovary at all times with-
out reference to menstruation.

7. There is said to be no proof of the identity of the two
conditions.

I think these propositions fairly cover the ground over
which those who deny the relationship of what they call
“heat” to menstruation have hitherto travelled.

It will be seen at a glance that the denials originate, in
most instances, in misconception of the facts, and that many
of the errors are due to the misuse of terms.

It will be worth while, however, to answer each of them
separately, and the following replies are numbered to corre-
spond with the numbers of the above objections.

1. The discharge in many animals during the pro-cestrum
contains blood and sometimes uterine tissue ; it is not always
solely mucus, and when blood is absent it has been shown
that its absence is due to a modification of, and not to any
radical difference in, the process.

2. The term “heat” is here wrongly used; it is made to
include both the pro-cestrum and the cestrus in the lower mam-
mals, and is compared in that extended sense with the term
menstruation, which is an error. The time the lower animals
will permit of coition is not during pro-cestrum, which is
synonymous with menstruation, but during cestrus, which
immediately follows the pro-cestrum. I have shown above
that there is not wanting evidence that the same may be true
for the human female.

3. Although the time for sexual intercourse among human
beings is not universally confined to particular seasons, I
have shown that in some cases this is so, and that in all

64 WALI'ER HEAPE.

peoples there is a marked disposition to indulge in sexual
intercourse at particular times of the year, which are un-
doubtedly comparable to the so-called “ breeding seasons ” of
the lower mammals. Further, in certain domesticated animals
and certain wild animals kept in captivity the males do not
“rut” only at certain times of the year, but are prepared to
propagate at all times (dog) or almost at all times (captive
cattle or deer) throughout the year.

4, There is some truth in this objection; but it must not be
forgotten that, among the lower mammals, while captivity
and domestication reduce the violence of the sexual passion,
they increase its frequency; and that in civilised woman, in
all probability, it is this variation of the function still further
exaggerated which is responsible for the difference (see
also 2).

5. Here again the objection is largely due to a mistaken
use of the term “heat,” which in this case is used to denote
cestrus.

Menstruation, that is pro-cestrum, in women is as necessary
to the production of the species as pro-cestrum in the lower
animals can be; the fact that cestrus is less pronounced in
the former is true, but it is not altogether absent, and has
already been referred to in the replies to propositions num-
bered 2 and 4.

6. This objection has reference to the question of ovulation,
which has not been treated of in this paper; with regard to
it I would merely say, that ovulation in certain of the lower
mammals is not necessarily coincident with cestrus, while in
some of them cestrus and ovulation are quite separate
functions. Ripe ova are not found at all times in each
human female, and the fact that they may be found at times
which are not coincident with menstruation, is merely further
evidence that these functions are independent also in women.
Further, the degree of independence which these two functions
assume is apparently variable in the human female.

7. The answer to this objection is contained in the foregoing

paper.

THE ** SEXUAL SEASON’ OF MAMMALS. 65

In spite of the fact that the evidence I have brought
forward is fragmentary, and notwithstanding that only the
fringe of a vast subject has been touched upon, I venture to
hope enough has been said to show that the wide variations
in the sexual functions exhibited by various mammals are
variations in degree, not variations in kind; and I venture to
think that the evidence of the homology, not only of pro-
cestrum and menstruation, but of each of the various sexual
phenomena dealt with in the various types, is incontrovertible.

One word with regard to the future development of the
subject. It is the cause of the sexual season which requires
determination.

Much stress has been laid upon the rhythmical nature of all
breeding processes ; this has been carried furthest by Lataste
(1887 and 1891), and by Beard in a very suggestive paper on
gestation (1897). So far as the sexual season is concerned,
its rhythm is no explanation of its origin. It may, I suppose,
be asserted that all forces are exerted rhythmically, that is a
condition ; whereas what is required here is knowledge of
the nature of the force itself, and the causes which govern or
limit its rhythm.

These are questions for the comparative physiologist, in
whose hands, as it seems to me, lie so many of the great bio-
logical problems of the day.

Speaking generally, the rhythm of the sexual season and
the power of breeding is seasonal, it is governed by external
forces which are exerted in consequence of seasonal change,
and by internal forces which are dependent upon individual
powers; further there is abundant evidence that nutriment,
and the capacity for storing nutriment, and the energy result-
ing therefrom are essential factors.

I differ from those who, like Beard, consider the ovary the
seat of the governing power of the breeding function ; ovula-
tion and the cestrus cycle are not necessarily coincident, the
stimulus sufficient to induce the one is apparently not suffi-
cient in all cases to induce the other, and it would appear
that the requisite initiative is independently produced.

vou. 44, pany 1.—NEW SERIES. E

66 WALTER HEAPE.

I am tempted to suggest the probability that there is
present in the blood from time to time what may be called
an estrus toxin, to suggest that its presence is due to the
external and internal forces mentioned above, and to relegate
to it the power which stimulates the activity of the sexual
season, and brings about the actual production of those
generative elements which nutrition has enabled the animal
to elaborate.

It appears to me that research in this direction would be
likely to be rewarded; it would not only be of great theo-
retical interest, but might well lead to increase of knowledge
regarding some of the causes of sterility, and prove of enor-
mous practical value.

LITERATURE.

ALLEN.—“‘ The American Bisons, Living and Extinct,” ‘Mems. of Museum
of Comp. Zool. at Harvard Coll., Cambridge, Mass.,’ 1876.

Aurmann.—Article in ‘Jahresbericht tber die Fortsch. der Anat. und
Physiologie,’ Hoffmann and Schwalbe, vol. vi, 1878.

ARIStOTLE.—‘ History of Animals,’ Bohn’s Library, 1883.

Bearp.— The Space of Gestation and the Cause of Birth,’ Jena, 1897.

Berver.—‘ Notes on Fields and Cattle,’ 1870.

Brtit.—‘ A History of British Quadrupeds,’ 2nd edit., 1874.

Benecke.—“ Ueber Reifung und Befruchtung der Kies bei den Fledermausen,”
‘Zoologischer Anzeiger,’ vol. ii, 1879.

BENEDEN and JuLtin.—“ Observations sur la maturation, la fécondation, et la

segmentation de l’ceuf chez les Cheiroptéres,” ‘ Archives de Biologie,’
vol. 1, 1880.

BriscHuorr.—‘ Entwicklungsgeschichte des Rehes,’ Giessen, 1854.

Biscuorr.— Ueber die Ranzzeit des Fuchses und die erste Entwicklung
seines Kies,” ‘Sitz. der Math.-phys. Classe vom 13 Juni, vol. ii,
1863.

Bonnet.—Articles in ‘ Deutsche Zeitschrift fiir Thiermedicin,’ vol. vi, 1880
vol. vii, 1882.

Bonnet.—Article in Kllenberger’s * Vergleichende Physiologie der Hatssau-
gethiere,’ 1892.

THE °° SEXUAL SEASON” OF MAMMALS. 67

Bosqunr.—* Note on a Case of Abseuce of the Uterus,” ‘ Obstetrical Trans-
actions,’ vol. xxvil, 1885.
Cameron.—‘‘ Notes in regard to the Breeding of Wild Red Deer,” MS., 1900.

CampBELL.—“ Notes on the Mode of Capture of Elephants in Assam,’ ‘ Proe.
Zool. Soc.,’ 1869.

Caton.—‘ The Antelope and Deer of America,’ 1881.

Coox.—“ Medical Observations among Wsquimaux,” ‘New York Journal of
Gynecology and Obstetrics,’ vol. iv, 1894.

Datton.—‘ Descriptive Ethnology of Bengal,’ Calcutta, 1872.

Darwin.—-‘ Animals and Plants under Domestication,’ 2nd edit., revised.

1875.
Denwoop.—“ The Badger, its Period of Gestation,” ‘ Zoologist,’ vol. xvill,
1894.

Enrenserc.—‘ Abhandl. d. Akad. zu Berlin,’ 1833.

Kimer.—“ Ueber die Fortpflanzung der Fledermause,” ‘ Zoologischer Anzel-
ser, vol. 11, 1879.

ELLENBERGER.—‘ Vergleichende Physiologie der Haussaugethiere,’ 1892.

Etirs.—* Psychology of Sex: vol. ii, ‘Phenomena of Sexual Perio-
dicity.”

Fevtxin.—“ A Contribution to the Determination of Sex derived from
Observations made on an African Tribe,” ‘Trans, Edinburgh Obstet.
Soe.,’ vol. xi, 1885-6.

!'LeminG.—‘ Veterinary Obstetrics, 1878.

Frizrs.—‘‘ Ueber die Fortpflanzung der einheimisclhen Cheiropteren,”
‘Gottingen Nachrichten,’ 1879.

GARNER.—‘ Gorillas and Chimpanzees,’ 1896.

GrorFrroy Sarnt-HiLatre and Cuvier.—‘ Hist. natur. des Mammileres.’

GraBuamM.—‘‘On the Breeding of Sorex araneus,” ‘ Zoologist, vol. xx,
1896.

Guise.—‘ On the Tribes inhabiting the Mouth of the Wanegela River, New
Guinea,” ‘Journ. Anthropological Instit.,’ vol. 1, 1899.

Hamitron.— The Wild Cat of Europe,’ 1896.

Hartine.—“ The Badger,” ‘ Zoologist,’ vol. xii, 1888.

Haycrarr.—‘‘ On some Physiological Results of Temperature Variations,”
‘Trans. Royal Society Edinburgh,’ vol. xxix, 1880.

Heare.—“ The Menstruation of Semnopitheeus entellus,’ ‘Trans,
Royal Society,’ 189+.

Hearre.—“ The Menstruation and Ovulation of Macacus rhesus,” ‘Trans.
Royal Society,’ 1897.

68 WALTER HEAPE.

Heape.—* The Menstruation of Monkeys andthe Human Female,” ‘Trans.
Obstetrical Society,’ vol. xl, 1898.

Hecx.—Zoologischen Gartens, Berlin, private letter, 1899.

Hitt.—‘ Life Statistics of an Indian Province,’ ‘Nature,’ July 12th, 1888.

Hotprr.—“< Gynecological Notes among American Indians,” ‘Amer. Journ.
of Obstetrics,’ 1892.

Hurcuinson.— Miarriage Customs in Many Lands,’ 1897.

* British Central Africa,” 1897.

JounsrenE.—‘‘'The Kndometrium in the Cycle of the Rat,’ * British Gyne-
cological Journal,’ vol. iu, 1888.

JOHNSTON.

. Jounstons.— Relation of Menstruation to other Reproductive Functions,”
‘Amer. Journ. of Obstetrics,’ vol. xxxil, 1895.

KazzanpEr.—“ Ueber die Pigmentation der Uterinschleimhaut des Schafes,”
‘ Archiv fiir mikroskop. Anat.,’ vol. xxxvi, 1890.

Kerru.— The Relationship of the Chimpanzees to the Gorilla,” ‘ Proc, Zool.
Soc.,’ 1899.

KovaLevsky.—‘ Modern Customs and Ancient Laws of Russia,’ 1891.

KowaLewsky.—‘ Marriage among the Karly Slavs,’’ ‘ Folklore,’ vol. i, 1890.

Larasre.—‘ Recherches de Zooéthique sur les Mammiferes de l’ordre des
Rongeurs,’ Bordeaux, 1887.

Laraste.—‘ Des variations de durée de Ja gestation chez les Mammiferes,”
‘Comp. Rend. et Mém. de la Soc. de Biologie,’ Ser. 9, vol. ii, 1891.

Lataste.— Rytlime vaginal des Mammiferes,” ‘Actes de la Soc. Scient. du
Chili,’ vol. iii, 1893.

Laurencre.—‘ Valley of Kashmir,’ 1895.

Laycocx.—‘ A Treatise on the Nervous Diseases of Women,’ 1840.

Lerorotp and Mrronorr.—“ Beitrage zur Lehre von der Menstruation und
Ovulation,” ‘Archiv fiir Gynakologie,’ vol. xlv, 1894.

Lo Branco.—‘ Notizie biologiche riguardanti specialmenti il periodo di
maturita sessuale degli animali del Golfo di Napoli,” ‘ Mitth. Zool. Stat.
Neapel,’ vol. viii, 1888; vol. xiii, 1899.

Loprr.—“ Note on the Period of Gestation of the Indian Antelope,” ‘ Proc.
Zool. Soc.,’ 1894.

Low.— The Domesticated Animals,’ 1845.

LypEKKER.—‘ Wild Oxen, Sheep, and Goats of all Lands,’ 1898.

Man.—* On the Aboriginal Inhabitants of the Andaman Islands,” ‘Trans. of
Anthropological Institute,’ vol. xii, 1882-3.

« Martin.—* The Physiology and Histology of Ovulation, Menstruation, and
Fertilisation,” Hirst’s ‘System of Obstetries,’ vol. i, 1888.

THE ‘* SEXUAL SEASON’’ OF MAMMALS. 69

Mituats, the late Sir Evrrett.—‘ Letters.’

Mtnor.—‘ Human Embryology,’ 1892.

Mrvart.—‘ The Cat,’ 1881.

Mi.ter.—‘ Elements of Physiology,’ translated by W. Baly, London, 1838.

Numan.—“ Over de periodische ontlasting van bloed, etc., de menstruatie der
vrouw,” ‘ Froriep’s Notizen,’ No. 150, 1838.

OweEN.—‘ Old Stone Crosses of the Vale of Clwyd and Neighbouring
Parishes,’ 1886.

Pross.—‘ Das Weib,’ 1887.

PrEJEVALSKY.—‘ Mongolia, the Tangut Country, and the Solitudes of Northern
Tibet,’ translated by Morgan, London, 1876.

Renecer.—‘ Naturgeschichte d. Saugethiere von Paraguay,’ Basel, 1830.

RETTERER.—“ Sur les modifications de la muqueuse utérine a l’époque du
rut,” Société de Biologie, July 15th, 1892.

Rinx.—‘ Danish Greenland, its People and Products,’ 1877.

Rinx.—‘ The Eskimo Tribes,’ 1887.

RoosEvELT.—‘* The Wilderness Hunter,’ 1893.

Rora.— Ethnological Studies among the North-west Central Queensland
Aborigines,’ 1897.

RovussELET.—‘ India and its Native Princes,’ 1876.

Sciater and 'l'Homas.—‘ The Book of Antelopes,’ 1900.

SomerseT.—‘ The Land of the Muskeg,’ 1895.

SouTHWELL.—“ On the Winter Breeding of the Otter,” * Zoologist,’ 1888.

STONEHENGE.—‘ The Horse in the Stable and the Field,’ 1877.

SronEvENGE.—‘ The Dog in Health and Disease,’ 4th edit., 1887.

Stratz.— Die Geschlechtsreife Siugethiereierstock,” ‘ Haag. Mart. Nijhoff,’
1898.

Surron.—‘* Menstruation in Monkeys,
1880.

Swayne.— Seventeen Trips through Somaliland,’ 1895.

26

srit. Gynecological Journal,’ vol. ii,

TEGETMETER aud SUTHERLAND.— Horses, Asses, Zebras, Mules, and Mule
Breeding,’ 1895.

Turner.—‘On the Placentation of Seals,” ‘Trans. Royal Society Edinburgh,’
vol. xxvii, 1875. i

Varieny.—‘ Experimental Evolution,’ 1892.

Wave.—“ Ingleby Lectures,” ‘ Lancet,’ June, 1886.

WesTERMARCK.—‘ The History of Human Marriage,’ 1891.

WiLLoucHBY.—‘ Hast Africa and its Big Game,’ 1889.

70 WALTER HBEAPE.

. Wittsuire.—“ Lectures on the Comparative Physiology of Menstruation,”

‘ British Medical Journal,’ 1883.
Youarr.— Cattle,” ‘ Library of Useful Knowledge,’ 1834.
Youatr.—‘‘ The Sheep,” ‘ Library of Useful Knowledge,’ 1837.

Yovatr.—‘ The Pig,’ 1860.

DESCRIPTION OF EPHYDATIA BLEMBINGIA, 71

A Description of Ephydatia blembingia, with
an Account of the Formation and Structure
of the Gemmule.

By

Richard Evans, M.A., B.Sc.

With Plates 1-—4.

ContTENTS.

PAGE

Part I.—TuHeE Morpnotoecy, ptc., or HPHYDATIA BLEMBINGIA.
I. Introduction 72
II. Description Siuaieecine hlembine va. ‘ 72
(1) Colour, habits of growth, and external form 72
(2) Skeleton 73
A. Spicules 5 73
B. Arrangement of earulee to [orn fibres 76
c. Spongin 76
(3) Canal system 77
(4) The structure of the mature ge sneole 78
ILI. Affinities fer eae blembingia 79
TV. Summary : 8]

Part IJ.—THe Formation oF THE GEMMULE OF HPHYDATIA
BLEMBINGIA.

I. Introduction ; , : d . : iol

II. Historical review cr
III. Descriptive account of the ae welonment of ihe penile af Ephy-

datia blembingia . . ge

(1) Origin, ete., of the reproductive art of the comin . 89

(2) Origin, ete,, of the cells which form the chitinous layers, etc. 91

(3) Origin, ete., of the seleroblasts and amphidises 92

(4) Structure, ete., of the “ es 5 : ; . 94

(5) General conclusions . 95

IV. Critical review of previous accounts . ' 96

V. Bibliography : : : : : . 103

Description of Plates. ; ; , . ¢ 105

U2 RICHARD EVANS.

Part I—The Morphology, etc., of Ephydatia blembingia.

I. Inrropvuction.

Ephydatia blembingia is a fresh-water sponge which
Mr. Annandale came across in a small pool of water while in
search of snails: It was collected and preserved by me on
the 25rd of July last year.

The specific name blembingia has been applied to it on
account of its locality. Blembing is a small Malay village
which was visited by the members of an expedition sent out
by Cambridge University to the Siamese Malay States, and
which is situated on a small river of the same name. The
river Blembing is a tributary of the Pergau, which in its turn
empties itself into the Kelantan River.

The pool of water in which the sponge, now described for
the first time, was found, was situated in a comparatively
dense jungle at a distance of a few yards from the bank of
the river. The trees growing around it were so big, and
their foliage so thick, as to admit of only a small amount of
light ever passing through them. Consequently the pool of
water in which Ephydatia blembingia was found was
always in a deep shade.

The material which I collected was preserved in the following
reagents :

(1) Flemming’s solution (weak fluid).

(2) Saturated solution of corrosive sublimate (92 volumes)
and glacial acetic acid (8 volumes).

(3) Absolute alcohol.

(4) Rectified 70 per cent. spirits.

II. Description oF HPHYDATIA BLEMBINGIA.

(1) Colour, Habits of Growth,and External Form.—
Ephydatia blembingia is almost colourless, or to use a term
which was used many years ago by Professor Lankester (11)
to describe the colour of Spongilla from the Thames, it is
“pale flesh-coloured.” Knowing as we do that Spongilla

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 73

lacustris and Ephydatia fluviatilis, denizens of our
own rivers, are green in colour only when they grow in bright
sunlight, this is what we would have expected in the case of
a sponge which grew in a pool of water scarcely ever brightened
by direct sunlight.

The habits of growth of Ephydatia blembingia are
peculiar. In reality it is an encrusting sponge, though some
specimens have a massive appearance. But this is due to the
habit of growing on such supports as blades of grass and
branching weeds of various kinds which inhabit the same
pool of water as the sponge. It never seems to produce
independent branches, which, when present, give a sponge a
kind of bush-like appearance, as Spongilla lacustris does.
If, at first, a specimen appears to branch, on closer examina-
tion the apparent branching reveals itself as the result of
creeping over a branched support. Consequently, in spite of
its massive appearance, Ephydatia blembingia is an
encrusting sponge. The biggest specimens measure no more
than about an inch across (Pl. 1,-fig. 1).

The surface texture of the preserved sponge’is Somewhat
woolly, an appearance caused by the spicule fibres which sup-
port the otherwise smooth dermal membrane. The fibres
often penetrate the membrane, owing undoubtedly to its being
rubbed off their extreme points.

To sum up, Ephydatia blembingia may be described as
a pale flesh-coloured sponge, with encrusting habits, creeping
over branched vegetable supports, and consequently irregular
in shape and woolly in texture.

The oscula, not to speak of the dermal ostia, are so small
as to be invisible without the aid of the microscope. The
openings represented in fig. 1 are those of the inhalant canals
seen through the dermal membrane.

(2) Skeleton.—The skeleton consists almost entirely of
spicules, which I shall now proceed to describe.

A. Spicules.—tIn order to facilitate the description of this
most important element of the skeleton, I shall arrange the
spicules under three heads.

74, RICHARD EVANS.

(a) The first group of spicules consist of diactinal monaxons
or amphioxea, which are usually curved, though straight
specimens are occasionally seen (PI. 1, fig. 3, a—e).

(b) The second group also consists of curved amphioxea,
but for reasons which will be stated further on they are sepa-
rated from the first group (PI. 1, fig. 3, f).

(c) The third group consists of amphidiscs, which may be
present in a fully developed or in an immature form (PI. 1,
figs. 3, g—m, 4, a—c).

(a) The amphioxea belonging to the first group taper
gradually to a sharp point. ‘They are never provided with a
swelling at the middle point of the shaft, and scarcely ever
are they malformed or modified inany way. In both respects,
therefore, they differ most strikingly from the spicules of
Spongilla moorei, a description of which was published im
this journal a year and a half ago. They appear to be in-
variably covered with small spines.

(b) The amphioxea belonging to the second group are
invariably curved and covered with small spines. In fact,
they present the same characters as the spicules of the
first group, but differ from them in being only half as long
and less than half as thick. They are not found in the
general tissues of the sponge or in its membranes, but are
grouped together round small bodies! which are embedded in

1 The bodies above mentioned seem to possess a definite outline, and to
lie in cavities of their own, much in the same way as the gemmules (PI. 4,
fig. 17). Ihave no conception what these bodies are, but several solutions
have suggested themselves. Unfortunately I have been unable to find them
in thin sections, and consequently cannot speak of their internal structure.

The first suggestion, with regard to their nature, to present itself was that
they were a second kind of gemmule. ‘The arrangement of the spicules round
them reminds us of that of the spicules round the gemmuiles of Spongilla
lacustris, and is, so far, in favour of the supposition that these bodies are
some kind of gemmules. But apart from the fact that Ephydatia blem-
bingia possesses another kind of gemmule, these structures are much more
transparent than ordinary gemmules are at any stage in their development or
when they are mature. If they were gemmules their basket-like shape could

be easily explained as the result of contraction under the action of preserving
reagents, owing to the cuticular coat being extremely thin. Apart from the

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 75

the deeper tissues, and which present a kind of basket-
shaped form.

(c) The last class of spicules to be considered consists of
the amphidises (PI. 1, fig. 4, b). The diameter of the hat-
shaped disc is about three times that of the shaft. The two
discs are exactly similar in size and shape. ‘The surface
situated away from the shaft is smooth and convex, while the
other surface is concave. Their margins are very finely
serrated. The shaft is covered with spines which are conical
in shape and placed at right angles to the axis of the amphi-

discs.
In addition to the fully mature forms, all the stages of
development are represented, from the simple rod slightly

fact that, if the supposition here made were true, it would cut at the very
root of the system of division into sub-families, now adopted, of the so-called
Spongillide, the reasons given above seem to be sufficiently weighty to compel
us to lay aside this possible view of tle nature of these enigmatical bodies.

The second supposition that suggests itself as a solution of the problem is
that these structures are a kind of symbiotic or parasitic sponge. This
supposition is not so unreasonable as it would at first appear, for we already
know that Spongilla bohmii is parasitic on Spongilla nitens (17). Be-
sides, it must be remembered that all the spicules in connection with these
bodies are quite different from those which form the sponge skeleton, being,
as has been stated already, only half as long and less than half as thick. It
is no argument to say that they are incompletely developed, for they are all
of equal size, which would not be the case if they were merely young spicules.
However, if these bodies are of the nature of a parasitic sponge, there are, at
present, no data by which its position among the Spongillidg can be determined,

There is still left another possible solution of the problem, namely, that the
bodies here discussed are the result of parasitism on the part of some animal
other than a parasitic sponge. If this supposition were true, these bodies
would have to be considered as a kind of gall, by means of which the sponge
endeavoured to protect itself from the action of an unwelcome intruder. But
there are two facts which go against this view. In the first place, though I
have examined several of these bodies, I have so far failed to find any animal
inside them. In the second place, though there are many parasites in the
sponge, not one of them has as yet been found to possess such a coat as these
bodies would provide.

Though it must be left an open question what the nature of these bodies
are, for the reasons given above I am inclined to adopt the view that they are

parasitic sponges,

76 RICHARD EVANS.

swollen at both ends to the fully-formed amphidiscs. Their
development, however, will be considered along with that of
the gemmule.

B. The Arrangement of the Spicules to form
Fibres, etc. (Pl. 1, fig. 2).—The spicule fibres are poorly
developed, and consequently stand in a most marked contrast
with those of some other fresh-water sponges. I have never
seen more than three spicules situated side by side in a
spicule fibre, and scarcely ever saw more than two. As often
as not, the spicules seem to be arranged end on in a single
file. In the deeper parts of the sponge, fibres are almost
non-existent, the spicules lying about freely and presenting
no particular arrangement. Nearer the surface, however, the
fibres are better developed, and traverse the strands of tissue
which separate the various compartments of the sub-dermal
cavity from one another. On the outer ends of the fibres is
situated the dermal membrane, which is often pierced owing
merely to the wear and tear of the life which the sponge
lives. Owing to the absence of flesh spicules or microscleres,
the skeletal fibres formed of megascleres present an evident
tendency to run in the vicinity of the membranes which line
the canals and cavities of the sponge.

As has been stated above, spicules of the class b take no
part in the formation of the skeleton, but this is not true of
those belonging to class c, 1.e. the amphidises. The latter
are found in all stages of development scattered about in the
general tissues of the sponge, while the former are limited to
the walls of the enigmatic bodies described above. Special
stress must be put on the fact that the developing stages of
the amphidiscs have been seen in the sponge tissues, and
not in the gemmule wall.

c. Spongin.—lIt is scarcely necessary to mention spongin
in connection with Ephydatia blembingia, for it is almost
completely absent. In this respect the sponge here described
strongly contrasts with some fresh-water sponges. In
Spongilla moorei the spicule fibres and the dermal mem-
brane are covered with this substance (7), but in Ephy-

DESCRIPTION OF EPHYDATIA BLEMBINGIA, ae

datia blembingia there is no spongin on the surface, and
the spicule fibres are, at most, provided with a very small
amount at the junction of the spicules.

This difference is explained by the dissimilarity in the con-
ditions of life. On the one hand, Ephydatia blembingia
lives in a small pool of water which probably dries up for the
ereater part of the year; while Spongilla moorei, on the
other hand, lives at the bottom of Lake Tanganyika. ‘There-
fore, the former may be described as an annual, while the
latter—so to speak—is a perennial spongilla. If this differ-
ence in the conditions of life under which these two sponges
live were to have any effect at all, we would naturally expect
the spongin part of the skeleton to suffer most.

(3) The Canal System.—Owing to the presence of gem-
mules in all stages of development, the canal system could
hardly be in such a condition as to be capable of minute
description, for the formation of gemmules is accompanied by
the breaking down of the sponge tissue. Besides, we know
of no preserving fluid that does not admit of a considerable
amount of disassociation of the tissue cells of the Monaxo-
nida. Though they be preserved with the greatest care, and
with the best reagents known, free cells are found in great
abundance in the interior of the sponge tissues. The presence
of so many amoeboid cells is conducive to this state of things.
Consequently our remarks on the canal system must be meagre
at best.

As has been stated above, the dermal ostia are micro-
scopically small but comparatively numerous. They open as
usual into the subdermal cavities, which are large and ex-
tensive (fig. 2), and which are lined by cells which possess
eranular nuclei. These in their turn open into the inhalant
canals, which are also well developed, but decrease in size
towards the surface of fixation of the sponge. The flagellated
chambers are small and numerous, lying about in the extremely
loose tissues of the sponge. The exhalant canals, though at
first of fine calibre, assume comparatively huge proportions.
The oscula, however, by which they open to the exterior are

78 RICHARD EVANS.

small. The membranes which line the canals are not pro-
vided with special spicules, but are supported by the spicule
fibres, which are situated close to the lining membranes.

4. The Structure of the Gemmule.—The gemmules
are scattered about singly throughout the whole tissue of the
sponge. They are found, on the one hand, near the surface,
and on the other hand, quite close to the vegetable supports
of the sponge. They are never found in groups. Hach gem-
mule occupies its own cavity (Pl. 1, fig. 2, gem.).

I shall here describe only the structure of the mature gem-
mule, the development of which will be described in Part II.
Nevertheless, it must be remembered that the sponge con-
tained gemmules in all stages of development at the time it
was collected.

The gemmule is oval in shape, being, as a rule, shghtly
flattened on the side on which the opening is situated. The
external opening or pore is placed at the bottom of a small
depression surrounded by a rosette-lke structure, which is
raised up, and into the composition of which all the layers of
the gemmule coat enter (Pl. 1, fig. 7).

The contents of the gemmule consist of a number of glo-
bular cells which are full of oval-shaped food granules. The
cells are all alike, and the whole mass possesses no membrane
of any kind save the gemmule coat, which I shall now proceed
to describe.

The gemmule coat consists of three layers which differ
from one another, to a considerable degree, both in structure
and extent of development.

The inner layer of the gemmule coat completely surrounds
the cells which are situated in the interior. It presents the
general shape of the gemmule and is prolonged round the
aperture to form a kind of a tube, the passage through which
is interrupted by a chitinous membrane situated about the
middle. The cellular contents of the gemmule extend into
the inner half, and the second layer of the gemmule coat to
the outer half of thistube. Instructure this layer is chitinous,
and resists the action of all ordinary reagents, save the mineral

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 79

acids. It often happens, that, in sections, it splits in two, a
result brought about by the weakening produced through the
inner ends of the amphidiscs being embedded in it, and not
lying upon it as is usually described in the gemmules of
Ephydatia. The lne along which the splitting takes place
is that in which the discs are situated (Pl. 4, fig. 17, a).

The middle layer of the gemmule coat is by far the thickest
and approximately extends over the whole length of the shafts
of the amphidiscs. It is clear in structure, and presents in
section the appearance of ordinary parenchyma with very
small granules at the nodes. In the fully developed gemmule
there are no lines of division indicative of the different cells
out of which it was originally formed (Pl. 4, fig. 17, a).

The outer layer of the gemmule coat is thinner than either
of the other two, and in it are embedded the outer ends of
the amphidises. It consists of the same substance as the
inner layer but is much more granular. In the mature gem-
mule it is often rubbed off, and consequently the outer discs
of the spicules protrude from the gemmule coat (PI. 4, fig.
Neca).

The amphidises he partly in the three layers. The imner
disc hes in the corresponding layer, the shaft im the middle
layer, and the outer disc in the thin outer layer. They are
so closely packed that the discs overlap one another and
consequently are not on the same level. Their shafts never
seem to cross one another, but lie approximately parallel.

III. Tae Arrinities oF EPHYDATIA BLEMBINGIA.

The presence of gemmules in the material at my disposal
made the task of determining the systematic position of the
fresh-water sponge here described a comparatively easy one.
The possession of gemmules excludes it from the sub-family
Lubomirskine, which is a sub-family created for the pur-
pose of grouping together a number of fresh-water sponges
in which the gemmule, if it does exist, has not yet been
discovered. Further, the existence of the thick coat which
surrounds the gemmule cells and which contains, embedded

80 RICHARD EVANS.

in it, a thickly-set layer of amphidiscs separates it, on the
one hand, from the sub-fanily Spongilline, and on the
other hand places it among the Meyenine. Again, its
generic position 1s not difficult to determine. The equality
of size of the amphidisc rotules separates it from both
Tubella and Parmula, the serrated edge of the rotules from
Trochospongilla, the equality in length of all the amphi-
discs from Heteromeyenia, and the absence of any kind of
filament or appendage, attached to the chitinous tube, from
Carterius. Consequently the sponge, which is described in
this paper, belongs to the genus Ephydatia. Of the species
contained in this genus, the sponge to which the name Ephy-
datia blembingia has been given seems to approach
Ephydatia plumosa (Carter, 2) more closely than it does
any other well-marked species. Several species of the genus
Ephydatia are provided with amphioxea, which are covered
with small spines, and are the constituent elements of the
skeletal fibres. In Ephydatia fluviatilis (17) both smooth
and spined spicules occur together. It follows, therefore,
that the presence or absence of small spines on the skeletal
spicules is not distinctive as a specific character. Potts (17)
seems to consider this difference so unimportant that he
describes an American sponge, to which he has given the
name palmeri, as a mere variety of the Indian sponge
plumosa; though the skeletal spicules in the former are
covered with small spines, while in the latter they are smooth.
The skeletal spicules of Hphydatia blembingia agree
with those of palmeri, and not with those of plumosa.

The amphidiscs seem to be closely similar in plumosa,
palmeri, and blembingia, though the rotules appear to be
more deeply notched in the two sponges mentioned first than
they are in blembingia. If these were all the differences
that could be enumerated the sponge now discussed would
have to be considered a slight variety of the species plumosa,
if, indeed, not actually identical with the variety palmeri.
However, there still remains to be mentioned another most
important difference, namely, the absence from blembingia

DESCRIPTION OF EPHYDATIA BLEMBINGTA. 81

of the flesh spicules so characteristic of both plumosa and
palmeri. Though this is a negative character, combined with
the other differences it seems to be a sufficient reason for the
formation of a new species, to which I have given the name

blembingia.

IV. Summary.

Ephydatia blembingia is an encrusting sponge which
grows on vegetable supports. It is pale flesh in colour, and
loose in texture. The skeletal spicules are covered with small
spines. Flesh spicules are absent unless the small amphioxea
(b) be considered to belong to such a category. The spicule
fibres are poorly developed, and in the deeper parts of the
sponge the spicules, as a rule, he about irregularly arranged
in the tissues. Spongin is present only in very small quan-
tities. The gemmules are numerous, but not aggregated in
eroups. ‘They are situated—each one occupying a cavity of
its own—near the surface as well as deeper down in the
tissues of the sponge. They are oval in shape, and possess
an opening resembling that of a bottle, which is obstructed
by a chitinous septum. They are provided with a thick and
well-developed coat, in which amphidises of equal lengths are
arranged in a single layer. The shaft of the amphidiscs is
furnished with conical spines, large in size and situated at
right angles to the longitudinal axis. The outer surface of
the dises is convex, and the margin is shehtly serrated. Amphi-
dises, in all stages of development, are scattered about in the
sponge tissue where they are formed.

Part II—The Formation of the Gemmule of Ephydatia
blembingia.

I. [wrropuction.

When I took the description of HEphydatia blembingia
in hand I had no intention of describing the development of

von. 44, pART 1.—NEW SERIES. P

82 RICHARD EVANS.

the gemmule ; but when I saw that the material at my disposal
contained gemmules in all stages of development I thought
it would be a mistake not to describe it. Further, I was en-
couraged to do so by Professor Weldon, to whom I am greatly
indebted both for the free use of his laboratory and all its
resources, and for much invaluable assistance, especially in
connection with the literature on the subject. I shall first
give a summary of what is already known of the gemmule.
I shall then proceed to describe my own observations, the
method followed being that of tracing the origin and subse-
quent changes of the various cells which take part in the
process, this method being considered simpler and more intel-
ligible than that of giving a complete description of the differ-
ent stages of development. The reader can easily make out
for himself, by examining the figures 8, 9, . . . . 17, the true
relation of the changes in the different parts of the developmg
gemmule much better than by reading the best possible
description. Finally, I shall review previous accounts and
compare my own conclusions with them.

Il. Historicat REvIEew.

Carter (2), who was the first to attempt an explanation of
the origin of the gemmule, which he terms the seed-like body,
writes as follows :— At the earliest period of development im
which I have recognised the seed-like body it has been com-
posed of a number of cells, united together in a globular or
ovoid mass (according to the species) by an intercellular
substance. In this stage, apparently without any capsule,
and about half the size of the full-developed seed-lke body,
it seems to le in a cavity formed by a condensation of the
common structure of the sponge immediately surrounding it.
It passes from the state just mentioned into a more cireum-
scribed form, then becomes surrounded by a soft, white,
compressible capsule ; and finally thickens, turns yellow, and
develops upon its exterior a firm crust of siliceous spicules.”
He says with regard to the origin of the gemmule, “T do not

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 83

wish it to be inferred that I am of opinion that the seed-like
body is but an ageregate of separate sponge-cells;” and
further, after describing certain cells of the sponge, he says:
“Tt may, perhaps, be one of these cell-bearing cells which
becomes the seed-like body.”

Lieberkiihn (12), in the year 1856, published an account
of the origin and structure of the gemmule. He found in
the deeper parts of the sponge shiny white gemmules, which
on the whole appeared like ordinary brown gemmules, and
which possessed exceedingly plainamphidiscs. He also found
other gemmules, distinguished by their very delicate trans-
parent shells, also possessing very obvious amphidiscs. These,
he said, had a superficial layer of a substance feebly refrac-
tile, and a central mass brilliantly refractile. The feebly refrac-
tile cells separated easily, while the others only did so with
difficulty. In these bodies he was not able to find the delicate
transparent encrusting layer, which he had seen round the
white gemmule; but found a layer of cell-like spherules which
resembled the ordinary sponge-cells in the arrangement of
their granules and of their nucleolus; while others contained
the amphidiscs. Some of the enclosed amphidises had exactly
the shape of those found surrounding the ordinary gemmule.
Others, he said, did not possess the two discs, but in the
interior of each cell-like structure there was a delicate rod
with a slight knob-like swelling at each end. In others a
series of very fine spicules radiate from the terminal swelling.
He derived the amphidises by imagining these spicules to
become broader, and the axial rod to become thicker. The
contours of the cells containing the spicules were described
as being as sharp as those of ordinary sponge-cells. He could
find no nuclei in these cells. He finally concluded that these
bodies were incompletely developed gemmules. He also
found certain bodies which he deseribed as white aggregations
of sponge-cells, possessed of the same size and shape as
ordinary gemmules. In the same year he published a second
paper, in which he summed up as follows (13) :—‘‘ That the
gvemmules are derived from a heap of ordinary sponge-cells

84. RICHARD EVANS.

we can very plainly see in that branched sponge which has
gvemmules with smooth shells. In a longitudinal section of a
suitable piece we find—(l1) Gemmules which are completely
developed, and possess a smooth shell containing a large
number of the rounded masses accurately described by Meyen.
Each of these masses is spherical, and contains in its interior
an albuminous fluid and many strongly refractive spherules.
It is about as large as a sponge-cell, and quickly disintegrates
in water. (2) Gemmules with an obvious shell, which con-
tains Meyen’s spherical masses and also contains bodies which
have Meyen’s masses, but are distinguished from these by
sending pseudopodia like the ordinary sponge-cells. (3)
Gemmules in which the shell and the pore are obvious, con-
taining only cellular bodies which send out pseudopodia.
Some of these contain a nucleus and a nucleolus like sponge-
cells, and are distinguished from these only by the fact that
they contain in their interior the refracting spherules already
alluded to. (4) Spherical heaps corresponding in size to the
gemmules which consist of the above-mentioned bodies,
sending out pseudopodia, and of undoubted sponge-cells. The
sponge-cells have an obvious nucleus and nucleolus, and they
contain besides a mass of very fine granules, which may be
scattered through the whole cell-body or may be collected
in small spherical masses. These spherical masses are of the
same size as the refracting spherules already described, and
one or two such spherules are often found in the sponge-cells.
Round some of these spherical heaps of cells a very fine
structureless membrane can be recognised. The spherical
masses. of Meyen which are commonly found in gemmules
are nothing else than altered sponge-cells ; by compressing the
contents of the gemmule under the cover-slip we can find a
nucleus and a neucleolus in every such mass; but nucleus and
nucleolus are so hidden by the strongly refractile contents of
the Meyen’s masses that they can only be demonstrated by a
process of pressure. These nuclei and nucleoli do not espe-
cially differ from those of ordinary sponge-cells.”

Lieberkiihn again, in a third paper (14), speaks as follows

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 85

of the gemmule :—“'The gemmules are not eggs, but a sort of
cyst or capsule, out of which the same individual which built
them ultimately creeps through the pore.”

In a later publication (8) Carter describes the seed-like
bodies as being globular in shape, and consisting of a cori-
aceous membrane enclosing a number of delicate, transparent,
spherical cells, more or less filled with ovules and granular
matter, while an incrustation of gelatinous matter charged
with small spicules peculiar to the species surrounds the ex-
terior of the coriaceous membrane. “ It has also been shown,”
he adds, “that at an early period of development the spherical
masses, Which we shall henceforth call ovi-bearing cells, are
polymorphic—identical, but for the ovules, with the ordinary
sponge-cells—and surrounded by a layer of peculiar cells:
equally polymorphic, which I have conjectured to be the chief
agents engaged in constructing the capsule.”

_ Again, in a later publication (4), he speaks of the “ova”—

preferring the term “ovum” to “seed-like body”—of Spon-
gilla as follows:— At an early period of the ovum the
spherical cells, though already filled with the refractive
granules, are few in number and sub-polymorphic; hence it
may be reasonably inferred that their multiplication as the
ovum increases in size is produced by fission ; the younger the
ovum the more polymorphic and resistent are these cells, while
the older it becomes the more they are attenuated, and the
more rapidly they burst by endosmose after liberation.”

In the year 1874 he further writes of the gemmules as
follows (5) :—“It may be a question whether the entire body
may not be the ovarium of a Spongozoon in the first place;
while, as in hundreds of instances of the same kind in the
animal kingdom, all the other parts have perished, their
function having ended when sufficient nutriment had been
gathered and assimilated to support the reproductive elements
until they could do this for themselves.” Further on he adds,
“Tt is an assemblage of ova which are at once developed to-
gether into a young Spongilla.”

In his final communication (6) on the gemmule, he views it

86 RICHARD EVANS.

“as a simple ovum with modified form to meet the require-
ments of the case.”

It seems Carter was always uncertain as to the origin of
the gemmule, and at one time or another he appears to have
had four views. First, that the gemmule was a mere aggre-
gation of sponge-cells; secondly, that it was an aggregation
of cells produced from one cell, the “ovi-bearmg” cell;
thirdly, that it was a single ovum, which was his final view ;
and fourthly, that it was a single “ovarium” of a dead
“ spongozoon.”

In the year 1884 Marshall published an account of the de-
velopment of the gemmule of Spongilla lacustris (15).
He says that the first sign of the gemmule consists of a
number of amoeboid cells, which are found in the neighbour-
hood of the inhalant canals and the ciliated chambers, and
which he terms the “trophophores.” They fill themselves
with reserve material, and wander together in groups. They
become round and give up water, so that they look lke
masses of reserved food material. Very early round the
pseudomorula formed in this way there appears a delicate
structureless membrane, a cuticle, the matrix of which should
be probably looked for on the surface of the pseudomorula
itself. The “mesoderm” outside this cuticle builds at first
an endothelium which deposits on the cuticle further layers
of horny substance and delicate siliceous structures, in this
case spiny tangential needles.

In the year 1886 appeared Goette’s account of the develop-
ment of the gemmule of Spongilla fluviatilis (10). He
says that the first rudiment of the gemmule is formed by an
ageregation of ordinary parenchyma cells in a nearly spheri-
cal area of 36—44 4 in diameter; really, the flagellated
chambers and canals of this region become enclosed in the
aggregation, which is produced through hypertrophy of the
cells. In this aggregation of cells the formation of two
layers quickly takes place; a central mass of cells, con-
taining a great number of yolk-granules, and an outer sheet
of cells, which become club-shaped and form a kind of

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 87

columnar epithelium round the central mass. This sheet
secretes a cuticle round the central mass, and its cells form
the amphidiscs. Subsequently the club-shaped cells migrate
outwards, and secrete a second cuticle outside the amphi-
discs.

In the same year as Goette, and independently of him,
Wierzejski described the development of the gemmule (19).
He describes the first rudiment of the eemmule as a group of
naked amoeboid cells. He says that the cells of the mother-
sponge can migrate to the body of the gemmule and thus
increase its volume. The heap of cells brought together
through migration from the sponge tissue become differen-
tiated into a central mass and a peripheral layer. Shining
spherules and granules are deposited in the cells of the
central mass, those of the peripheral layer becoming co-
lumnar. The amphidises are not developed in the peripheral
cells, but in the surrounding tissues, and only subsequently
migrate to the columnar layer.

In the year 1892 Zykoff published an account of the de-
velopment of the gemmule (21). This account adds little, if
anything, to what was known before of the formation of the
eemmule. He found, among the ordinary amceboid cells of
the parenchyma, cells which contained a number of refractive
eranules of a very definite form, which he describes as boat-
shaped. He considers the appearance of refractive yolk-
substance in a few amoeboid cells of the mesenchyme as the
first step in the development of the gemmule. These amceboid
cells have the protoplasmic structure of Fiedler’s amoeboid
“ Fresszellen,’” but the nuclear structure of his “ Nahrzellen.”
He disagrees with Goette and supports Wierzejski on the
question of the origin of the first rudiment of the gemmule.
He denies Goette’s statement that the flagellated chambers
and the epithelial lining of the canals participate in the
formation of the gemmule. The rudiment of the gemmule
soon becomes differentiated to a central mass of yolk-cells,
among which amoeboid cells of the mesenchyme occur, and a
peripheral stage which consists of one or two concentric layers

8&8 RICHARD EVANS.

of mesenchyme cells of the sponge. The peripheral cells be-
come club-shaped and not columnar. ‘This change takes place
eradually, not all at once. The club-shaped cells secrete the
inner cuticle, and the amphidises migrate from the sponge
and take up their position among the club-shaped cells, which
subsequently migrate outwards, secrete the outer cuticle, and,
finally losmg their club-shaped form, gradually become re-
sorbed.

In the year 1895 Weltner published a short paper (18), in
which he brings together the different views expressed as to
several important points m connection with the structure and
development of the gemmule, and from his own observations
draws his own conclusions. Having discussed the use of the
protective coat; the presence of a thin membrane, which he
does not believe to exist, round the reproductive portion of
the gemmule; the number of nuclei in each cell, of which he
has seen more than one in several cases, he finally deals with
the question of the origin of the cells of the gemmule, in the
first rudiment of which he finds three kinds of cells, namely
cells which have yolk-bodies alone, cells which display fine
eranules of equal size and a distinct nucleolus, and cells
which have large granules of unequal size. The third class
of cells are different from the cells with granules of unequal
size found in the parenchyme.

He comes to the conclusion that the development of the
eemmule is not yet sufficiently known, and that a fresh
inquiry should be instituted as to two main points: first,
the origin and nature of the cells which form the first rudi-
ment of the gemmule; secondly, to ascertain the fate of
these cells.

He suggests that their origin and nature should be exa-
mined with a view to the following possibilities :

Is the first rudiment of the gemmule formed from a single
cell which has the value of an egg? Then the gemmule
should be a group of segmenting cells.

Or, does the inner mass of the gemmule arise from one class
of cells derived from the previous mesoderm ?

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 89

Or does it arise from more than one class of mesoderm
cells ?

Or, finally, is it built from different germ layers (two or
three) ? The gemmule should then be considered a bud.

In Section III of the second part of this paper I shall give
an account of the development of the gemmule in Ephy-
datia blembingia, reserving criticism of whatever kind to
Section IV. In Section III I shall include nothing but a
simple description, followed by a few conclusions. This
course will be pursued in order to make the account more
available and more intelligible to the reader than it would be
if it were mixed up with critical remarks and conclusions
scattered about throughout the paper.

II. Descriprive Account oF THE DEVELOPMENT OF THE
GEMMULE OF HPHYDATIA BLEMBINGIA.

(1) Origin and Further Development of the Repro-
ductive Part of the Gemmule.—The first sign of prepara-
tion for the formation of the gemmule consists in the presence
of single cells or small groups of cells scattered about chiefly
in the dermal membrane ; the strands of tissues which support
the dermal membrane ; and in the tissues situated immediately
below the subdermal cavity.

The protoplasm of the cells in question is uniformly clear,
and the nucleus is granular and not vesicular (Pl. 2,
fig. 8). I have been unable to detect a karyokinetic figure in
any of these cells. Consequently I am of opinion that the
constituent cells of these groups seldom divide during the
early stages of formation of the gemmule, which is contrary
to what must have been the case if the cells of the repro-
ductive part of the gemmule were derived from one mother-
cell.

The cells in virtue of their power of wandering travel
through the dermal membrane, and strands of tissue which
support the membrane, and become aggregated in groups
situated either deep in the tissues of the sponge or even in
the strands of tissue above mentioned (PI. 2, fig. 8).

90 RICHARD EVANS.

The protoplasm soon loses its uniformly clear appearance
and becomes unevenly granular (PI. 2, fig. 9), a feature which
rapidly becomes more accentuated (Pl. 2, fig. 10). The con-
tained granules or irregular blotches at this stage le in round,
clear spaces in the protoplasm, but they soon increase in size
to such an extent as to fill the spaces above mentioned. At
the same time they acquire an oval or spherical form and
exhibit a certain amount of internal structure, in the form of
unevenly distributed granules of very small size (PI. 2, fig. 11a).
The subsequent change in the interior of the spherical
eranules or yolk bodies, as they may be termed henceforth,
consists in the differentiation of a peripheral layer or coat
which sometimes, though not always, contains fine granules,
from a centre which invariably seems to possess a finely
eranular structure (Pl. 3, fig. 13d). The yolk bodies have at
this stage attained their ultimate structure, and fill the cell in
which they have been formed.

While these changes are going on a curious change takes
place in the character of the nucleus. At first granular, 1t now
becomes vesicular, or perhaps more correctly it presents an
appearance intermediate between the typical vesicular nucleus
with a solid nucleolus and a granular nucleus (Pl. 3, fig.
13d, nu.). The cells seem never to possess more than one
nucleus.

The yolk cells, as they may be termed henceforth, have
increased slightly in size during the changes above described.
However, they retain their individuality, though owing to the
pressure which they exert on one another they are often
polygonal in shape. In the fully developed gemmule they
are so pressed against one another that their individual outline
can be seen only with difficulty, which is in no way a remark-
able thing seeing that at no stage do they possess a definite
cell wall though having a well-defined cell limit.

The yolk cells collectively, or the reproductive part of the
gemmule, as they may be termed, at no stage possess a
membrane, though in the fully mature gemmule they are so
pressed against the inner chitinous layer of the protective

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 91

coat as to present a perfectly smooth and membrane-like
appearance.

(2) The Origin and Subsequent Changes of the
Cells which produce the Ground Substance of the
Protective Coat of the Gemmule.—These cells, after
having wandered from the general sponge tissues, appear in
the neighbourhood of the gemmule as a loosely arranged layer
situated outside the future yolk cells. Fig. lla (Pl. 2)
shows how they travel towards the developing gemmule and
how they become concentrated to form a layer.

Their general protoplasm is clear, but they contain a number
of granules or yolk bodies which resemble those of the yolk
cells. In addition, they often contain a much bigger spherical
body which seems to be of the same nature as what I have
described in my account of the structure of the larva of
Spongilla lacustris as nutritive vacuoles. The cells which
develop to yolk cells seem never to contain either of the
above bodies at their first appearance. At first they are
spherical in shape, but soon become columnar, though never
club-shaped. However, their outer end may be round and
not flat during certain stages (PI. 3, fig. 15, and PI. 4, fig. 15a).
They assume the columnar form, at first, only on one side of
the reproductive mass of cells, the columnar layer so formed
gradually increasing in extent until it completely surrounds
the yolk cells. The poimt at which the columnar layer is
finally completed marks the position of the future pore of the
gemmule.

Subsequent to the assumption of the columnar form, these
cells begin to secrete the inner chitinous layer, which in its
formation follows the same course as the columnar layer did,
which is a proof that the layer in question is secreted by the
columnar cells and not by yolk cells (Pl. 3, fig. 13; Pl. 4, figs.
14 and 15).

Soon after the amphidiscs have taken up their position
among the columnar cells—a phenomenon which takes place
soon after the formation of the columnar layer—the latter
erow out and before long appear outside the outer ends of

92 RICHARD EVANS.

the amphidiscs (PI. 4, fig. 16). While this is going on their
inner ends situated between. the amphidiscs become trans-
formed to the parenchyma-like substance situated in the
mature gemmule between the inner and outer chitinous coats.
During the elongation of the columnar cells outwardly the
nucleus is carried along. After their inner moiety has been
modified and the nucleus has passed to the outer portion they
secrete the outer chitinous layer and ultimately break off, and
so becoming liberated they pass back again to the sponge
tissue (Pl. 4, fig. 16a). The nucleus at the close of these
changes, as at the beginning, is vesicular.

The outer chitinous coat is much thinner and less homo-
geneous than the inner. In the fully mature gemmule the
greater part of it is lost, so that the outer ends of the
amphidises are uncovered.

(3) The Origin, Migration, and Final Modification
of the Scleroblasts, inside which the Amphidiscs
are developed, and their Migration from the Sponge
Tissue into the Columnar Layer.—At the outset special
emphasis must be laid on the poimt that incompletely deve-
loped amphidises were never seen in the protective coat of
the gemmule, whether during the early or later stages. The
amphidiscs situated in the gemmule coat are always fully
developed, while in the sponge tissues incompletely developed
stages as well as fully developed ones are plentiful.

The first stage observed in the formation of the amphidiscs
consists of a rod-like structure swollen at both ends (PL. 1,
fig. 3, m, and fig. 6,a), in which respect they differ essentially
from the young stages of the amphioxea, which are always
pointed (PI. 1, fig. 5). Both kinds make their first appear-
ance in cells with vesicular nuclei, which soon become trans-
formed and become granular, especially in the mother-cells
of the amphidiscs. The next change consists in the deve-
lopment of a more or less conical form by the ends of the
above-mentioned rods, the cone-shaped end at the same time
becoming surrounded by a rim (PI. 1, fig. 6, 6). The cone-
shaped end, together with its slightly developed rim, ulti-

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 93

mately grows to the hat-shaped dise. Throughout the pro-
cess of formation of the amphidise both ends are of the
same shape. If one end is incompletely developed, the other
is equally so. The spicules retain their position inside the
scleroblast until they have reached their definitive form, and
there seems to be no reason for supposing that, were the
scleroblast in any way injured, the spicule could ever attain
full development. The amphidiscs thus described assume
their ultimate form while yet in the general tissues of the
sponge. It is important to remember that they are developed
in cells which are essentially amceboid. When gemmules are
being developed, the scleroblasts in virtue of their inherent
power of locomotion move towards them. They travel along
the strands of tissue which have been described above as
passing from the general sponge tissue to the somewhat loose
membrane which surrounds the gemmule. Ultimately they
make their way among the columnar cells which surround
the gemmule (Pl. 4, figs. 14 and 15d). Fig. 15b (PI. 4) is
particularly interesting in that it shows the last spicule that
has entered the columnar layer as well as one situated in a
strand of tissue close by. The latter is on its way to take up
its position alongside the former spicule among the columnar
cells. When the amphidises, still situated inside the sclero-
blasts, have reached their final position, at first they are
longer than the columnar cells, which he completely inside
their outer ends. At this stage the scleroblasts, though
already considerably modified, can be distinctly seen. In the
fully grown gemmule, however, they are indistinguishable
from the parenchyma-like substance produced from the modi-
fied inner ends of the columnar cells.

The scleroblasts with their contained amphidiscs first push
their way in among the columnar cells at that point where
the columnar layer and the inner chitinous coat made their
first appearance. 'I'hey become more numerous and gradually
increase in number until finally they envelop the whole
gvemmule (Pl. 4, figs. 15 and 16). There are, therefore, three
distinct structures at least which first appear on the same

94 RICHARD EVANS.

side of the central cells, i.e. on the side opposite the point
which later on will be occupied by the pore, and all three
increase in extent in a similar way. They ultimately form
complete layers, though one of them, viz. the columnar layer,
is no longer found in the mature gemmule.

The migration of scleroblasts, or cells that would become
scleroblasts, is not a new idea to zoological literature. Mr.
Bourne described such migration of the calicoblasts in Helio-
pora coerulea (1), and Professor Minchin has given a full
account of the migration of the epithelial cells in the Ascons
to the interior, and the subsequent formation of spicules inside
them (16). It is true that in both these cases the migration
to the interior is previous to the formation of spicules, while
in Ephydatia blembingia the amphidiscs are fully formed
before the change of position takes place. This difference
does not in any way tend to minimise the importance of the
facts described above. The amphidiscs are so small as com-
pared with ordinary spicules, and their ends are rounded, con-
sequently there is no inherent improbability in the view that
they are carried from one place to another by the scleroblasts.

(4) The Origin, Structure, and History of the
Trophocytes.—The trophocytes are large round cells with
vesicular nuclei, the chromatin of which is for the most part
aggregated in small granules either round the spherical
central corpuscle or against the nuclear membrane, the inter-
vening space being, as a rule, occupied by only a few small
granules. In the immediate neighbourhood of the nucleus
there are innumerable small and irregularly shaped granules
which give the cell a dirty-looking appearance, the peripheral
portion being exceptionally clear and devoid of granules of
any kind. A negative feature of these cells is seen in the
absence of both yolk bodies and nutritive vacuoles.

The trophocytes originate from the sponge as a separate
class of cells, hke the three other classes which have been
already considered. They migrate from the sponge tissue at
the same time as, and alone with, the cells which become
columnar. While the columnar cells always remain outside

DESCRIPTION OF EPHYDATIA BLEMBINGTA. 95

the yolk-cells, the trophocytes pass in among them. They
are incapable of passing through the columnar layer after it
has been completely formed, but seem to be able to push
their way through when the cells in question are arranging
themselves and becoming elongated. Not all of them pass
among the yolk-cells, some, as it appears, only enterimg among
the developing columnar cells and turning back. The majority
of them, however, seem to pass among the yolk-cells. Asa
rule, they pass through the developmg columnar layer singly,
but occasionally groups of several cells are witnessed making
their way in. After the trophocytes have entered among
the yolk-cells they distribute nutritive material to them, pro-
bably in solution. They take no part in the formation of the
reproductive portion of the gemmule further than to supply
it with nutritive material which the yolk-cells store up in the
yolk-bodies. When the inner chitinous layer is about half
formed (PI. 3, fig. 151), the few remaining trophophores are
seen travelling towards that part of the gemmule where the
pore will appear. They pass out and become scattered about
round the gemmule (Pl. 3, fig. 15c). It is not difficult to
understand why the trophocytes travel all in the same direc-
tion, i. e. away from the portion that is already formed of the
inner chitinous layer, for it is undoubtedly the direction of
least resistance.

5. Summary of Conclusions.—(1) Four classes of cells,
each of which is derived independently from the sponge,
take part in the formation of the gemmule ; first, the mother-
cells of the yolk-cells which, alone, constitute the reproduc-
tive portion of the gemmule ; secondly, the mother-cells of
the columnar cells which pass back to the sponge; thirdly,
the mother-cells of the amphidises, “ scleroblasts,’ which

1 The sections represented in figs. 13—13d (Pl. 3) were cut from material
preserved in Flemming’s weak solution, while those represented in ail the
other figures were from material preserved either in absolute alcohol or in a
mixture of 92 parts of saturated solution of corrosive sublimate and 8 parts
of glacial acetic. This explains the absence of the dirty-looking granules
from all the trophocytes except those represented in figs, 13—18d (PI. 8),

96 RICHARD EVANS.

become modified and form a part of the intermediate layer of
the protective coat of the gemmule ; and fourthly, the tropho-
cytes, whose function is to supply both the columnar and the
yolk cells with food material, and which, hke the columnar
cells, pass back to the sponge.

(2) The yolk-cells and the columnar cells draw their food
material in solution from the trophocytes; the yollk-cells
storing it up as a reserve in the yolk bodies; the columnar
cells using it in such a way as to enable them to secrete the
inner chitinous layer, to grow and pass out between the
outer ends of the amphidises, their inner ends being modified
to form the greater part of the ground substance of the pro-
tective coat of the gemmule, and finally to secrete the outer
chitinous layer ; processes which mean that there is an enor-
mous amount of metabolism going on.

(3) The amphidises are developed in cells, the scleroblasts,
which carry them through strands of the sponge tissue to
their ultimate position in the protective coat of the gemmule.

TV. CriticaL Review or Previous Accounts.

On perusal of the historical section of this paper it will be
seen that the views which have been expressed as to the first
appearance of the gemmule are numerous and conflicting.
The only thing certain is that a group or aggregation of cells
is formed. How it is formed and whence it is derived no one
seems to know, though every one has a theory to put forward,
Again, it is equally uncertain whether the gemmule is formed
from the group which first appears, or whether this group in
order to bmld up the gemmule structure acquires recruits
from among the sponge cells and tissues.

Probably the first question that should be discussed is
whether the group of cells above mentioned is the product of
cell migration to one spot, or of cell division either of a single
cell or of a group of cells.

In his first attempt to explain the origin and structure of
the gemmule in the year 1849 (2), Carter expressed himself

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 97

in favour of the view that the gemmule is derived from what
he calls an “ovi-bearing cell.” In the year 1886 Goette
supports the view that this group of cells is the product of
cell proliferation (hypertrophy) (10).

In reply to both of these views it will suffice to point out
that at no stage during the early development of the gem-
mule are there any signs of cell division. Though during the
very earliest stages the cells are absolutely clear (Pl. 2,
fig. 8) I am totally unable to find the least sign of nuclear
division, not to speak of fragmentation. In all cases the
nuclei seem to be well formed, and in no way modified.
Carter had not the facts required to support his view, while
Goette seems to have merely figured a piece of ordinary
sponge, indifferently preserved, as the first rudiment of the
gemmule. For fig. 31 (10) can hardly be explained in any
other way. It must be admitted as certain that he saw
flagellated chambers and canals in the specimen represented
in the above-mentioned figure, but it seems almost equally
certain that what he saw was not the rudiment of a gemmule,
for the gemmule at its first appearance offers no points of
comparison with Goette’s representation. I'rom the con-
sideration of the absence of cell division, the view that the
gemmule rudiment is formed by that means may be set aside—
to say the least

The second view of the origin of the gemmule rudiment to

as a most highly improbable one.

be considered is the one according to which it contains collar
cells and flat epithelium cells, or, as Weltner expresses it, that
it consists of cells from two or three germ layers. his view
has its most influential advocate in Goette. It was held by
Carter also at one time, and probably by Lieberkiihn, who says
of the spherical heaps of cells he found in the sponge tissue,
that, besides containing Meyen’s masses, they also contain
undoubted sponge cells.

If the explanation given above of Goette’s fig. 31 (10) is
correct—and it seems that it must be—it easily explains how
he arrived at the conclusion that all the sponge layers parti-
cipate in the formation of the gemmule. Besides, it is quite

VoL. 44, part 1.—NEW SERIES. G

98 RICHARD EVANS.

possible that those who hold this view of the origin of the
gemmule are mentally dominated by the principles of the
“Germ Layer Theory.” If so, this would be a splendid
example of an otherwise good theory leading to false conclu-
sions. Further, it is more than probable that Carter, Lieber-
kithn, and Goette never saw the first signs of the formation
of the gemmule. This is undoubtedly the most charitable
view to take of the conclusions they arrived at.

Now that the above-mentioned views have been disposed
of, there remain for consideration two more views, one of

which can be set aside after only a few remarks. The view ©

in question is the one according to which the gemmule is
derived from a group of cells, all of which are alike. This
view has not found favour with those who have investigated
the structure and formation of the gemmule. Carter at one
time held it (5), thinking that the gemmule was an ovarium
of a “Spongozoon,” a name which he gave to his imaginary
sponge-animal. However, in a later publication he gave his
support to another view. In fact, a single glance at a good
section of the gemmule during some of the early stages is
enough to cause one to recoil from the idea that only one
class of cells take part in its formation. Consequently there
remains only one view, namely, that the gemmule originates
from a number of cells belonging to various classes. This
view, in one form or another, is supported by Marshall (15),
Wierzejski (19), Zykoff (21), and Weltner (18).

Marshall’s account does not concern us as much as those of
the other authors above mentioned do, for the reason that he
worked on the gemmule of a species belonging to a different
venus. There is, however, one poit which must be men-
tioned. The point in question is that he derives the gem-
mules from two classes of cells at least; namely, the cells

> and which give rise to the

which he terms “ trophophores,’
contents of the gemmule, that is the reproductive part, as
well as to the delicate structureless membrane surrounding
it; and the “mesoderm” cells, which give rise to the outer

shell as well as to the spicules. The importance of this dis-

(a3

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 99

covery lies in the fact that the reproductive part of the
gemmule and its protective coat are respectively formed from
classes of cells which are absolutely different, a conclusion
which is endorsed in the present paper, though the existence
of Marshall’s “ delicate membrane” round the central mass is
here denied.

Wierzejski, evidently, has observed some phenomena which
he did not understand. According to this author a first heap
of naked amoeboid cells becomes differentiated to a central
mass of yolk contaiming cells, and a peripheral layer of
columnar cells. But he also states that the cells of the
mother sponge can even migrate to the body of the gemmule,
and thus increase its size. In the light of the facts which
have been described in the foregoing section of this paper, it
seems certain that Wierzejski discovered the migration of
cells, on the one hand, to form the columnar layer, and on the
other hand to feed the mother-cells of the reproductive cells

of the gemmule. Wierzejski in describing the first group of

ce 2

cells uses the term “ pseudomorula,’ and, probably knowing
that a true morula always becomes differentiated to two
classes of cells, he comes to the conclusion, as it appears, that
his “pseudomorula” must do the same. Consequently he
commits the mistake of describing the columnar layer of cells
as originating by differentiation from his pseudomorula
instead of by further migration from the sponge tissue. Not
only this, he was also unfortunate in not being able to discover
the true nature of the cells which migrated to the interior of
the gemmule, as he says, to increase its size. His failure was
probably due to the method of preservation he used.

Before proceeding any further, it is necessary to refer to
Fiedler’s account of the cells which he found during his in-
vestigations of Ephydatia fluviatilis (9). Fiedler de-
scribes and figures two kinds of cells (9, pl. x1, figs. 3 and 4,
and pl. xii, figs. 36 and 37). One kind, which he terms
“amoeboid Fresszellen,” has granules of equal size in its
protoplasm, and a nucleus the chromatin of which is arranged

in a network. The other kind, which he terms “amoeboid

100 RICHARD EVANS.

Nahrzellen,’” has granules of unequal size in its protoplasm,
and a nucleus with a distinct nucleolus.

Zykoff, who writes in the light of Fiedler’s discoveries,
considers the appearance of refractive yolk substance in a few
amoeboid cells of the mesenchyme as the first development of
the gemmule. He finds these cells belong to neither of
Fiedler’s classes of cells, for they have the protoplasm of
the “amceboid Fresszellen” and the nucleus of the “amoe-
boid Nahrzellen.” These cells, together with others like
them, but without yolk substance, are described as creeping
together to form a spherical heap of cells, which differen-
tiates to a central mass which consists of yolk-cells, amongst
which here and there are scattered amoeboid cells of the
mesenchyme, and to a peripheral sheet of mesenchyme cells
without yolk, which pass to the general mesenchyme of the
sponge. Zykoff has described Wierzejski’s figures as being
diagrammatic and far from the truth. His figures, however,
might with a certain amount of propriety be described in the
same terms, and Weltner’s criticism that they are not natural
is quite true. Zykoff, however, is in error when he says the
cells of the peripheral sheet above mentioned do not contain
yolk. It is true that there are cells among them without
spherical bodies in them, the trophocytes of the present
paper; but it is equally true that the greater number of them
contain bodies which are in all respects similar to the “re-
fractive yolk substance” which Zykoff professes to have seen
“in a few amoeboid cells of the mesenchyme,” which he de-
scribes as the appearance of the first development of the
gemmule. Zykoff has here failed to distinguish between two
classes of cells, and consequently the description he has given
of them is not true of either. The cells which he found to
contain yolk-bodies in the sponge tissue, and which, he
assumes, become the yolk-cells, develop, as it appears, to the
columnar cells, and do not fall under the category of “ amoe-
boid Fresszellen,” or that of “amoeboid Nahrzellen.” They
form a separate class, while the other cells found among them
as well as among the yolk-cells must be placed in a class by

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 101

themselves. In the present account they have been termed
“trophocytes,” and seem to be identical with Fiedler’s
“Nahrzellen.” YZykoff found the cells in question among the
yolk-cells, but does not properly account for their absence
from that position in later stages. It seems certain that
Zykoff never saw the first stages in the development of the
gemmule. His first figure has not the remotest resemblance
to the first rudiment of the gemmule. If he never saw the
first stages, this explains how he missed the cells with nuclei
the chromatin of which is arranged in a network at first, but
later on presents the appearance of a modified vesicular
nucleus. However, there seems to be little doubt but that
these cells form a different class from the above-mentioned
classes, and correspond when they are coming together in all
respects to Fiedler’s “Fresszellen.” Consequently, at the
time the peripheral layer of cells appears the whole group
consists of three classes of cells: first, the mother-cells of the
yolk-cells; secondly, the mother-cells of the columnar cells ;
and thirdly, the “trophocytes.” The first class consists of
Fiedler’s “amoeboid Fresszellen,” the third class of his
“amoeboid Nahrzellen,’ while the second class consists of
those cells which, according to Weltner, belong to neither of
Fiedler’s classes, and, according to Zykoff, occupy a position
between the two.

Now that the somewhat difficult questions of the origin and
fate of the cells above discussed seems to have been solved,
there remain but few points to be considered in connection
with the formation of the protective coat of the gemmule.

At no stage ‘in the formation of the gemmule was a delicate
coat or membrane, situated internally to the inner chitinous
layer, found to exist. It often happens, however, that the
outer limit of the reproductive portion of the gemmule is
sharp, smooth, and well defined. But there is no membrane,
the sharpness of contour being merely the result of the pres-
sure exerted by the mass of cells on the inner chitinous layer.

The cells of the outer layer are columnar in form, and not
club-shaped. This, however, is a small point hardly worthy

102 RICHARD EVANS.

of all the importance attached to it by Zykoff. The columnar
cells during their transference from the inner to the outer
side of the external ends of the amphidiscs grow out rather
than migrate out. The result is that the spaces between the
amphidiscs are partly occupied by the inner moiety of the
cells, which moiety, being more or less cut off by the outer
ends of the amphidiscs, becomes transformed to the paren-
chyma-like substance which occupies that position m the
mature gemmule. That this is true can be easily seen on
examination of fig. 15b (Pl. 4), where the inner ends of the
columnar cells are already undergoing the above-mentioned
transformation, though the amphidiscs are not yet in position.
Consequently the origin of this layer need no longer be con-
sidered unknown, as has been done by Goette and Zykoff.

The next question to be considered is the origin of the
amphidises. Lieberkiihn describes the amphidiscs as bemg
developed in some of the cells of the peripheral Jayer (see
p. 83). Goette figures a developing amphidisc in one of
these cells, and describes these spicules as being formed from
within outwards. As has already been pointed out, incom-
pletely developed amphidises are never seen in the gemmule
coat (p. 92), but are abundant in the sponge tissue. They
seem to be invariably symmetrical in form, one end being the
exact counterpart of the other. Goette seems to have been
in error on both these points.

Zykotf merely confirms Wierzejsk’’s view that the amphi-
discs are formed outside the gemmule, but neither of them
was able to find the scleroblast, which is most surprising,
seeing that Lieberkiihn says that the outlies of the cellular
structures containing the amphidiscs are as sharp as those of
ordinary sponge cells. Zykoff discusses at considerable length
the mode of migration of the amphidiscs from the sponge
tissue to the gemmule coat, and arrives at the somewhat
amusing conclusion that they are pushed from one position to
the other by the sponge cells, much in the same way, I should
imagine, as a colony of ants carries away bits of food which
are too heavy a bundle for one. The presence of amphidises

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 103

in the scleroblasts, both in the sponge tissue and in the gem-
mule coat, disposes of the necessity of such a supposition. It
seems that it may be considered as finally established that the
amphidiscs are carried to their ultimate position by the sclero-
blasts which secrete them.

In conclusion I wish to offer my sincerest thanks to Pro-
fessor Weldon for the free use of his laboratory and all its
resources, as well as for much invaluable assistance in relation
to the literature of the subject; to Professor Minchin for
reading the proof sheets; to the Government Grant Com-
mittee of the Royal Society for their kind and timely assist-
ance; and to the Principal and Fellows of Jesus College,
Oxford, for further help.

Tue DepartTMENT of COMPARATIVE ANATOMY,

Tur Museum, OXForD.

V. BreLioGRAPHy.

1. Bournz, G. C.—“On the Structure and Affinities of Heliopora
cerulea, Pallas, with some Observations on the Structure of Xenia
and Heteroxenia,” ‘ Phil. Trans. Roy. Soc. Lond., vol. elxxxvi (1895),
B, pp. 455—483, Plates 10O—138.

2. Carter, H. J—A Descriptive Account of the Fresh-water Sponges
(genus Spongilla) in the Island of Bombay, with Observations on the
Structure and Development,” ‘Ann. Mag. Nat. Hist.’ (2), vol. iv,
pp. 81—100, Plates 3—5.

3. Carter, H. J.—‘On the Ultimate Structure of Spongilla and Additional
Notes on Fresh-water Infusoria,’” ‘Ann. Mag. Nat. Hist.’ (2), vol. xx,
pp. 21—41, Plate 1.

4, Carter, H, J—‘On the Identity in Structure and Composition of the
So-called Seed-like Body of Spongilla with the Winter Egg of the
Bryozoa, and the Presence of Starch Granules in each,’ ‘Ann. Mag.
Nat. Hist.’ (3), vol. iii, pp. 381—348, Plate 8.

5. Carrer, H. J—*“ Development of the Marine Sponges from the Earliest
Recognisable Appearance of the Ovum to the Perfected Individual,”
‘Ann. Mag. Nat. Hist.’ (4), vol. xiv, pp. 388-406.

6. Carter, H. J.—“ On the Nature of the Seed-like body of Spongiila, on
the Origin of the Mother Cell of the Spicule, and on the Presence of
Spermatozoa in Spongida,” ‘Ann. Mag. Nat. Hist.’ (4), vol. xiv, pp.
97—111, Plate 10.

104 RICHARD EVANS.

7. Evans, R.—‘* A Description of ‘'wo New Species of Spongilla from Lake

8

9

10.

11.

12.

13.

14.

15.

16.

Lv fe

18

19

Tanganyika,” ‘Quart. Journ. Micr. Sci.,’ N.S., vol. xli, pp. 471—488,
Plates 37 and 38.

. Evans, R.—“The Structure and Metamorphosis of the Larva of Spongilla
lacustris,” ‘Quart. Journ. Mier. Sci.,’ N.S., vol. xlii, pp. 363—476,
Plates 35—41.

. Fiepuer, K. A.—‘‘ Ueber Ei- und Spermabildung bei Spongilla

fluviatilis,” ‘ Zeits. f. wissen. Zool.,’ Bd. xlvii.
Gorttr, A.—‘* Untersuchungen zur Entwicklungsgeschichte von Spon-
gilla fluviatilis,’ Hamburg and Leipzig, 1886.

Lanxester, E. R.—‘ On the Chorophyll Corpuscles and Aniyloid Deposits
of Spongilla and Hydra,” ‘Quart. Journ. Mier. Sci.,’ N.S., vol. xxii,
pp. 229—254, Plate 20.

Lieserktiun, N.—‘‘ Beitrage zur Entwicklungsgeschichte der Spon-
gillen,” ‘Arch. Anat. Phys.,’ J. Muller, 1856, pp. 1—19.

Lizsperktun, N.—“ Zur Entwicklungsgeschichte der Spongillen,” ‘ Arch.

Anat. Phys.,’ J. Muller, 1856, pp. 399—414, Taf. xv.

Lizperkinn, N.—‘ Zusatze zur Entwicklungsgeschichte der Spongillen,”
‘Arch. Anat. Phys.,’ pp. 496—514, Taf. xviii, figs. 8—18.

Marsnatt, W.—*“ Vorl. Bemerkungen iiber die Fortpflanzungs-Ver-
haltinisse von Spongilla lacustris,” ‘Sitzungsb. Naturf. Ges.,’
Leipzig, Jahrg. xi, 1884.

Mincury, EK. A.—* Materials for a Monograph of the Ascons. I. On the
Origin and Growth of the Triradiate and Quadrirate Spicules in the
Family Clathrinide,” ‘Quart. Journ. Mier. Sci.,’ N.S., vol. xl, pp.
469—587, Plates 388—42.

Porrs, E.—* Contributions towards a Synopsis of American Freshwater
Sponges, &.” ‘ Proc. Acad. Nat. Sci.,’ Philadelphia, 1887, p. 158.

. Wettner, W.—“ Bemerkungen uber den Bau und die Entwicklung der
Gemmula der Spongilliden,” ‘Biol. Centralbl.,’ vol. xiii, pp. 119 —126.
Abstract in ‘ Zool. Record,’ 1892, also in ‘Journ. Roy. Micr. Soc.’
(1893), p. 492.

. Wierzessx1, A.—‘‘ Le developpement des gemmules des Eponges déau
douce d’Europre,” ‘ Archives slaves de Biologie’ (1886), p. 26, Taf. I.

20. Zyxorr, W.—“‘Die Entwicklung der Gemmula der Ephydatia
fluviatilis,’ Auct. ‘Zool. Auz.,’ xv, Jahrb. pp. 95—96.
21. Zyxorr, W.—“ Die Entwicklung der Gemmula bei Ephydatia

fluviatilis, Auct.” ‘Bull. Soc. Imp. Natur., Moscow’ (1892), pp.
1—16, ‘Taf, I, 11; Abstract in ‘Zool. Record,’ vol. xxix (1892); also
in ‘Journ. Roy. Mier. Soe.’ (1892), p. 378.

DESCRIPTION OF KPHYDATIA BLEMBINGIA. 105

EXPLANATION OF PLATES 1—4,

Illustrating Myr. Richard Evans’ paper on “ Ephydatia
blembingia, and the Development of the Gemmule in
the same Species.”

All the figures from 2—17 @, both inclusive, have been drawn with the
camera lucida.

SIGNIFICANCE OF THE LETTERING.

am. wand. cell. = fut. yo. cells. Amceboid wandering cells which later on in
the development become the yolk-cells. dm. wand. cells = fut. col. cells.
Ameboid wandering cells which later on in the development become the
columnar cells. Amphid. Amphidises. chit. sept. Chitinous septum. col.
cells. Columnar cells. col. day. Columnar layer. @. m. Dermal membrane.
in. chit. lay. Inner chitinous layer. xz. Nucleus. ow. chit. lay. Outer chiti-
nous layer. mod. sclerob. Modified scleroblast or spicule cell. sclerob. Sclero-
blast. s.d.ec. Sub-dermal cavity. spice. Spicule. sp.fib. Spicule fibre.
tropho. Trophophore. v.s. Vegetable support of the sponge. yo. body
Yolk-body. yo. cell. Yolk-cell.

PLATE 1.

Fie. 1 (x 14).—Ephydatia blembingia growing on vegetable supports,
v. 8.

Fic. 2 (xX 28).—A section showing diagrammatically the structure of the
sponge, especially the dermal membrane (d.m.) and its supports, which are
traversed by the spicule fibres; the large sub-dermal cavities (s. d.¢.); the
poorly developed spicule fibres (sp./id.), the great number of spicules, both
amphioxea and amphidiscs, scattered about more or less loosely in the
tissues, and the gemmules (gem.) which are never found aggregated together
in groups.

Fic. 3 (x 225).—A representation of the various kinds of spicules. a—e
represent the amphioxea which on the one hand form the spicule-fibres, and
on the other hand lie about loosely in the sponge-tissues. / is one of the
spicules which are seen grouped together in fig. 7. g—i are the amphidises.
g and A are fully formed; ¢ and & are intermediate in size, while 7 and m
represent the early stages.

Fig. 3 was drawn from specimens cleaned witli nitric acid.

Fig. 4 (x 575).—A more highly magnified representation of some of the

106 RICHARD EVANS.

spicules shown in Fig. 3. a represents m of Fig.3; dis an enlarged drawing
of g of Fig. 3; and ¢ shows the end of the amphidise when looked down
upon. Note the serrated edge.

Fie. 5 (xX 665).—A young amphioxea shown inside the scleroblast. It
should be specially compared with the Fig. 6, a, which represents the
early stage of the amphidisc. Both are drawn on the same scale, and on
comparison it will be clearly seen that the spicule represented in Fig. 6, a,
cannot be a young amphioxea.

Fic. 6 (x 665).—A representation of an early stage, a; intermediate stage,
4; and a fully-grown stage, ec, of the amphidisc. Each spicule is situated
inside the scleroblast which produced it. The cells themselves are amceboid
in character.

Fie. 7 (Xx 225).—This figure represents a side view of the somewhat pro-
blematic body discussed in the footnote on p. 74, and described as having a
basket-like form. It is covered with spicules (sp.) of the amphioxea type be-
longing to group & (compare Fig. 3, /).

PLATE 2.

Fie. 8 (xX 565).—A representation of the cells which later on in the
development become the yolk-cells of the gemmule. On the left side of the
figure the cells are seen coming together in virtue of their power of wander-
ing. On the right side a portion of a much bigger group of cells is seen.
The group in question is situated in the interior of one of the columns of
tissue which support the dermal membrane. Note that the protoplasm of the
cells is absolutely clear, and that the nucleus is not vesicular.

lic. 9 (X 950).—A representation, more highly magnified, of a slightly
later stage than that shown in Vig. 8. Note that the protoplasm is becoming
slightly granular.

Fre. 10 (x 950).—A representation of a slightly later stage than that
shown in Fig. 9. Note that the protoplasm has become still more
granular.

Vie. 11 (x 180).—A representation of the gemmule at a stage slightly
later than that shown in Fig. 10. Note that the granules in the cells have
increased to a considerable extent in size, but that their internal structure is
not so dense as at later stages. Also note that a great number of cells
possessing different characters are aggregated round the central cells (yo.
cells), ‘These cells seem to have been derived from the sponge ata later stage
tuan the yolk-cells, and constituie a different class both as regards their
origin and fate.

Fie. lla (X 950).—A representation, more highly magnified, of a portion
of the section shown in Fig. 11. Note the yolk-cells (yo. ced/s) with their

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 107

yolk-oodies (yo. body), also the columnar cells which are as yet only anaggre-
gation of amceboid wandering cells with fuod vacuoles and yolk-bodies, also
the “trophophores ” which possess clear protoplasm, but no yolk-bodies or
food vacuoles.

Fie. 12 (X 180).—A representation of a stage slightly later than that
shown in Fig. 11. Note that the amcboid wandering cells outside are
becoming columnar, especially on one side, also that the outer cells are
becoming separated, from those which are becoming columnar, to form a kind
of loose membrane (cf. Fig. 14).

Fic. 12a (xX 950).—A representation, more highly magnified, of the
ameeboid wandering cells, which are becoming columnar on one side of the
section shown in Fig. 12, and also, on the left side of the figure, of the cells
which begin to form the loose membrane.

PLATE 3.

Fie. 13 (x 1380).—A representation of a stage slightly later than that
shown in Fig. 12. Note that the columnar layer is complete except over a
small portion at which, later on, the pore will appear, and at which alone the
“trophocytes”’ (¢roph.) are found; also that the inner chitinous layer is
being formed from the same position, i.e. from the bottom, as the columnar
layer of cells was formed.

Fie. 13a (x 1150).—A representation of a portion of a section similar to
the one shown in Fig. 13. Note the group of six trophocytes (¢ropho.)
which are making their way to the interior of the gemmule, also the “ tropho-
cyte” which has already reached that position. This group is an unusually
large one, and was found opposite one of the strands of tissues which pass
from the sponge tissue to the loose membrane which surrounds the gemmule.
The “‘ trophocytes” appear to travel chiefly along these strands of tissue.

Fig. 136 (x 1150).—A representation of a similar portion to that shown
in Fig. 18a. Note that the “trophocytes” (¢roph.) are scattered about
among the amoeboid wandering cells, which Jater on become the columnar
cells. On the left of the figure is seen the end of one of the strands of tissue
along which the trophocytes travel. Also note the trophocyte on the right
of the figure. This cell is just passing among the yolk-cells (yo. cell).

Fie. 13¢ (xX 960).—A representation of a portion from the top of a section
similar to the one shown in Fig, 13. Note that the ‘ trophocytes” (¢roph.)
are arranged chiefly outside the columuar cells, but that there are some
situated still among the yolk-cells. Those outside have already travelled out
while those inside are in the process of doing so, In the lower part of the
section there were no ‘‘ trophophores.”’

Fie. 13d (x 1150).—A representation of a yolk-cell from the same section
as Fig. ldc. Note the large, central, vesicular nucleus (xw.).

108 RICHARD EVANS.

PLATE 4.

Fie. 14 (x 180).—A representation of a stage slightly later than that
shown in Fig. 13. Note that the columnar layer is complete, and that the
loose membrane which surrounds the gemmule is almost complete and is con-
nected by strands of tissue, in which amphidises are found, with the general
sponge structure. Further, note that there are neither amphidiscs nor
trophophores among the columnar cells.

Fig, 14a (x 960).—A representation, more bighly magnified, of a portion
of the columnar layer and loose membrane shown in Fig. 14.

Vic. 15 (xX 130).—A representation of a stage slightly later than that
shown in Fig. 14. Note that there are amphidiscs (ampAz.) over half the
extent of the columnar layer of cells, while the other half is as yet free of
them. Also note that the inner chitinous layer is still incomplete at the
point where the pore will be formed; but the yolk granules in the interior
are fully formed. The outer end of the amphidiscs extends beyond the
columnar cells.

Fic. 15a (x 665).—A representation, more highly magnified, of the amphi-
dises (amphid.) lying inside the modified scleroblast (mod. scleroS.) which has
lost its nucleus, and of the columnar cells lying between the amphidiscs.
Note that the inner ends of the columnar cells are becoming clear, as is shown
in Fig. 15.

Fig. 154 (x 665).—A representation of the inner chitinous layer (¢z. chit.
lay.), the columnar layer (cod. ced/.), the loose membrane, and one of the
strands which pass to the membrave, the whole being taken from the region
intermediate between the one occupied by amphidises and the one devoid of
them in a gemmule similar to that shown in Fig. 15. Both amphidises are
shown inside their scleroblasts, and the one in the strand of tissue outside is
being carried to its position alongside the other amphidise among the columnar
cells.

Fie. 16 (xX 180).—A representation of a stage slightly later than that
shown in Fig. 15. Note that the spicular layer and the inner chitinous layer
are complete. Further, note that the columnar cells have passed out and are
situated externally to the outer end of the amphidises, their inner ends having
been modified to form the parenchyma-like substance situated between the
amphidiscs.

Fic. 16a (x 665).—A more highly magnified representation of the gem-
mule coat shown in Fig. 16. Note that the columnar cells are forming the
outer chitinous layer and are becoming separated preparatory to their passing
back to the sponge.

Fic. 17 (x 1380).—A representation of the fully-developed gemmule, show-
ing the contents passing up the pore as far as the chitinous septum (cAzé.

DESCRIPTION OF EPHYDATIA BLEMBINGIA. 109

sept.) ; the inner chitinous layer (2x. cAz¢. /ay.); the amphidises (amphid.) ;
and the not strongly developed outer chitinous layer (ow. chit. lay.). Note
that the columnar cells are no longer present.

Fig. 17a (x 665).—A more highly magnified representation of the gem-
mule coat, showing the several parts indicated in the description of Fig. 17.
Note the parenchyma-like structure of the substance situated between the
amphidises (¢véer. /ay.), that is, the intermediate layer of the protective coat
of the gemmule.

wa aa ih 4
Se iy
7 -
wed 7 ra } =
Rss 4h : > .

. » ’ : . ~

\ Bie) POA ee YP Ue
a gs fos San ndeS Serl Sale’ se |
“a ® PP sks : La
=e — oa Pete, (ees ey
i: . oe Rew of

Piee¥ Ly “4

COLLECTION OF NEMERTEANS FROM SINGAPORE. 111

On a Collection of Nemerteans from Singapore.

By

R. C. Punnett, B.A.

With Plates 5—S.

Tur Nemerteans described below were collected by Mr.
F. P. Bedford and Mr. W. F. Lanchester during a year’s stay
in and near Singapore. They comprise ten species, nine of
which have not hitherto been described. All belong to the
order Heteronemertini, whilst eight out of the ten come into
the family of the Lineide. It is worthy of note that no speci-
mens of Drepanophorus were found, seeing that in collections
from adjacent seas, previously described by Birger (1) and
myself (7), they form a marked constituent of the Nemertean
fauna. It will be noticed that I have assigned all the Lineidz
here described to the genus Cerebratulus. This I have done
not because I am convinced that they approximate more
closely in every case to the forms already placed in that
genus, but because of the difficulties of distinguishing, in
preserved specimens only, the three genera, Lineus, Cerebra-
tulus, and Micrura. This difficulty was dealt with in Hu-
brecht’s case (5) by recognising in the family only the single
e@enus Cerebratulus. As, however, the family already contains
more than eighty species, with the probability of large addi-
tions in the near future, it seems expedient to adhere to the
present arrangement of genera, in spite of its unsatisfactory
character, until the accumulation of anatomical data be
sufficient to warrant revision. The caudal appendage forms
the chief distinction between the genus Lineus on the one

112 R. ©. PUNNETT.

hand and Cerebratulus and Micrura on the other (the small
genera Borlasia and Langia being sufficiently well marked to
be left out of account here). Though its presence shows its
possessor to be either a Cerebratulus or a Micrura, its absence
does not necessarily establish the specimen in question as a
Lineus, unless a large amount of material is forthcoming,
since it is a delicate structure which may easily be broken off.
In his monograph Birger (3) attempts to place these genera
on a firmer basis by taking into account certain anatomical
features, such as the sidefolds usually characteristic of Cere-
bratulus, the presence or absence of a diagonal muscle layer,
and of neurochord cells. Based mainly on a study of the
Neapolitan forms, the attempt is for them fairly successful.
For exotic forms, however, it is less so. Thus C. natans,
which closely resembles many of the Mediterranean forms in
its general form, its well-marked side-folds, and its swimming
habits, differs markedly from them in the absence of neuro-
chord cells and of a diagonal muscle layer. Again, C. brun-
neus agrees with the genus Micrura in the absence of side-
folds and a diagonal muscle layer, whilst its comparatively
large size and stoutness of build incline one to associate it
rather with the genus Cerebratulus. The small and slender
C. erythrus, again, with its absence of a diagonal muscle
layer, might be relegated to the genus Micrura, were it not
for the presence of small side-folds and of neurochord cells.
Other instances might be taken, but the above are enough to
show that Biirger’s system is not altogether satisfactory.
However, before we can hope to improve upon it more ana-
tomical evidence must be forthcoming. And, indeed, on many
important points much yet remains to be made out in the
species inhabiting our own shores. Especially is this the case
with the excretory system, its range, topography, and the
position and number of its ducts. We are only acquainted
with these details in two of the thirty odd species of Cerebra-
tulus, and in six or seven cases in nearly the same number of
species of Lineus. The figures at the end of this paper will
give some idea of the diversities shown by this system, even

COLLECTION OF NEMERTEANS FROM SINGAPORE. 118

in members of the same genus. Other features which may
prove of taxonomic value, and to which I venture to draw
attention in the following list, are :

(1) The presence or absence of a frontal organ.

(2) The presence or absence « 4 cephalic vascular loop.

(3) The extent of the rhynchoccelom.

(4) The minute structure of the cutis.

(5) Structure of the cerebral organ.

Moreover it would greatly facilitate the labours of the
systematists in this group if collectors were able to make a
coloured sketch of the living animal, and to make particular
note as to the presence or absence of a caudal appendage
whilst the animal is still hving. A word on preservation may
not be out of place here.

Excellent results are to be obtained by stupefying the worms
in a 1 per cent. solution of chloral hydrate in sea water, and
subsequent treatment with saturated corrosive sublimate.
Particular care should be taken that the anterior few centi-
metres of the animal are preserved as straight as possible
by killing the animal on a glass slide or some other such
means. Since this region must be cut into serial sections the
labour-saving importance of this precaution can hardly be
exaggerated.

HETERONEMERTINI.
EupPoLiup”.

Kupolia melanogramma, mihi (= EH. quinquelineata,
Birger).

Five specimens of this large and striking worm were obtained,
showing considerable variation in the completeness of the five
dorsal black lines.

(1) Specimen, 32 xX 6 mm., having the five lines well
marked for the whole length of the body except near the
posterior end, where the outer two are occasionally broken,
All lines of nearly equal thickness.

voL, 44, PART 1,—NEW SERIES. H

114 R. C. PUNNETT.

(2) Specimen, 52 cm. X 6 mm., agreeing very closely with
last specimen.

(3) Specimen, 36 cm. X 4 mm., with the three median
lines well marked, of equal thickness and unbroken. The
two outer ones are present only along the anterior half of the
body, and even here are finer than the others and much
broken.

(4) Specimen, 15 cm. X 5 mm., with only three broad
inner lines except for the anterior 2 cm., where there are
two fine outer lines.

(5) Specimen, 54 cm. x 7 mm., with three dorsal lines
only, except for an exceedingly fine trace of an outer one for
a distance of about $ cm. on the right side only.

From which it may be seen that these five form a series, at
one end of which there is a specimen with five well marked
lines of equal thickness throughout almost the whole length,
whilst at the other there is a specimen which shows only three
lines (excluding the faint trace mentioned above). In this
connection it is interesting to note that Birger (2) has
described a single specimen with seven dorsal lines on which
he founds a new species, H. septemlineata, but concerning
whose inner organisation he gives no details. With almost
equal justice the specimen last described above might be
christened EH. trilineata. Since, in the light of the above
facts, it seems reasonable to look upon Biirger’s new species
as a variety, I would venture to suggest the name melano-
eramma for the species as a whole, restricting the terms
quinquelineata, septemlineata, etc., to denote varieties where
such a proceeding is thought desirable.

With regard to the inner organisation of this form, I am
able to confirm the special features given by Biirger (2), and
also to add a few points of interest omitted by him.

The epithelium is high, and its external portion contains
a number of anucleate (fig. 5) unicellular glands containing
minute yellowish bodies. Beneath the epithelium is a struc-
tureless basement membrane. Below this, again, is a layer of
longitudinal muscle fibrils, mixed up with which, in the

COLLECTION OF NEMERTEANS FROM SINGAPORE. 115

region of the black lines, is the pigment, which in sections is
seen to be brown. The cutis glands are well marked and
their contents stain deeply with thionin, though not at all
with carmalum. The gelatinous layer of the cutis is rich in
circular fibrils.

The vascular system in the snout shows the wide hori-
zontal lacunee characteristic of the genus. These bend round
beneath the dorsal ganglion at its commencement, and, taking
an upward course, come to lie between the latter and the
proboscis sheath at the level where the proboscis is attached
(fig. 1, a). These lacunz then extend dorsally, and at the
posterior limit of the ventral commissure unite to form the
dorsal vessel, and also envelop the top of the dorsal ganglion
(fig. 1, 6). In this region a small diverticulum projects into
the tissue between the dorsal and ventral ganglia. This
diverticulum becomes separated off, and a few sections later
is seen to unite with its fellow of the opposite side, forming
the median ventral cesophageal vessel (Schlundgefass) (fig. 1,
c). At this level the lacuna which earlier enveloped the
dorsal portion of the dorsal ganglion is now seen to lie upon
the dorsal surface of the cerebral organ. Still more dorsally
a portion of the lacuna has separated off, and this later unites
with the median vesophageal vessel on each side (fig. 1, d),
and the common lacuna so formed is continued into the
cesophageal lacune.

The alimentary canal is characterised by the great
thickness of the glandular tissue beneath the epithelium of
the cesophagus. In the intestinal region the lumen is wide.
The intestinal diverticula alternate in depth, every other one
being nearly twice the depth of those directly behind and in
front. The deeper set are not so deep as the width of the
lumen of the intestine. There is no ventral cutter.

The proboscis shows a layer of diagonal muscles between
the circular and longitudinal layers. The longitudinal fibres
are characteristically arranged, recalling the condition seen
in transverse sections of Lumbricus.

The proboscis sheath extends rather less than one third

116 R. C. PUNNETT.

of the length of the body. In a specimen 36 cm. in length,
it came to an end 10°5 em. from the anterior end.

The excretory system commences soon after the mouth,
and extends as far as the commencement of the intestine.
Anteriorly small portions of it show a tendency to become
isolated from the rest (cf. E. multiporata [7]). It possesses
a number of openings to the exterior, and also presents the
unique condition of ducts opening into the cesophagus (fig. 2).
In this way the lumen of the alimentary canal is indirectly
placed in communication with the exterior medium. That
such a condition is not pathological but normal in this species
is shown by the fact that it was found in another specimen,
the anterior portion of whose body was cut. The subjoimed
tables show the position of the various ducts (see pp. 118 and
119). The sections were of a uniform thickness of 8 pu, and
commence with the anterior extremities of the animals.

It will be noticed that the ducts are in some cases rudi-
mentary, i.e. they do not form a communication with the
excretory system, but end blindly when they reach the
circular muscle layer. It is also interesting to notice that
in several cases they pass through the ganglion-cell layer of
the side stem (fig. 5); and also that in one instance the duct
passed beneath the side stems, the only instance, as far as
IT am aware, of such a condition occurring in the group.

The genital sacs were devoid of sexual cells.

The main features of the nervous system have been
touched upon by Biirger. An interesting point, however, is
to be made out with regard to the anal commissure. In this
species it is very strong and ventral to the gut. In this case
the pigment lines afford an infallible criterion of orientation,
and there can be no question of a twisting of the body such
as described by Hubrecht in the case of EH. delineata, though
he states that the commissure is also ventral in this case (5)
(p. 11). Oudemans (6) (p. 41) mentions the commissure of EH.
curta as being dorsal. In the species of Hupolia next de-
scribed in this paper there is no commissure. How far such
divergent conditions may be due to injury and subsequent

COLLECTION OF NEMERTEANS FROM SINGAPORE. 117

regeneration seems to the writer a subject better fitted for
experimental inquiry than for morphological speculation.

Eupolia pholidota, n. sp.

A single specimen of this worm was taken, and in the
preserved state measured 19 cm. in length. The posterior
two thirds or so were very slender, being barely half as thick
as the anterior portion of the worm, which was about 3 mm.
in diameter. The record of its appearance in lifetime shows
it to have been “white with reddish-black spots.” The only
indication of such markings left are faint yellowish blotches
(fig. 7). These do not extend beyond the head furrow
anteriorly.

The epithelium is high and in its outer portion lacks the
small granular cells which characterised that of the preceding
species. The basement membrane (fig. 8, bm) is exceedingly
thick, and beneath it is a well-marked layer of longitudinal
muscle fibrils. The cutis glands are of the usual Hupolia
type.

The muscular system presents no peculiarities.

The vascular system resembles that described above for
EK. melanogramma. The lacune round the cerebral organ
are, however, more complete, and the vessels forming the
buccal commissure (fig. 9, bbvc) are seen in section to pass
round the outer side of the cerebral organ instead of the inner
side as in the preceding species.

In the alimentary canal the glandular tissue beneath
the cesophageal epithehum is comparatively thinner than is
usual in most members of the family. The cesophagus ter-
minates about 4°25 mm. from the tip of the snout. In the
intestinal region the lateral diverticula are small and there is
a deep ventral gutter.

The proboscis is extremely slender.

The generative sacs are full of spermatozoa, most of
which are ripe. Ducts are present.

The excretory system commences at about ‘85 mm. from

118 R. C. PUNNETT.

Specimen No. 1.

RIGHT SIDE. LEFT SIDE.
External Internal External Internal
openings. openings. openings. openings,
231—257 { 220
295
270—352 332
a 317
(
413
430 |
| 450
458
| 485 493
529
548
| 564
256—860 J
| 355 — 860 4
| 598 (through n.c.))
617
| a
| 638
647
681 |
678
685
706 706 702
717
720
732
| 751
753 (incomplete)
756 762
| 775 761 | 769
772
| 791
807 | 828
829 (below n.c.)
830 (through n.c.)
834 (through n.c.)
84.2 |
| |855 (incomplete) | |859

COLLECTION OF NEMERTEANS FROM SINGAPORE. 119

Specimen No. 8.

RIGHT sIpDE. LEFT SIDE.
External Internal External Internal
openings. openings. openings. openings.
c
| ee
319 —364.4 362—375
L
425 —457 :
455
if 513 (fF
525 |
571 H
583 |
| | 615
476—- ¢ 485-— 4
| 749 |
| 816 ; WD
839 : 786
| | | 800
| | | 829

120 R. C. PUNNETT.

the tip of the snout, and the ducts commence about this level
also. The system is interrupted in many places. These ducts
are exceedingly numerous, far exceeding in number those of
any other Nemertine known. On one side I was able to count
nearly a hundred ducts, and even then the excretory system
had not come to an end. Another feature peculiar to this
worm, and one which, so far as I am aware, occurs in no other
member of the group, is the backward extent of the excretory
system into the intestinal region, where the generative sacs
and their ducts also occur. Their extent may be more easily
recognised by a glance at the table below giving the levels
at which various systems commence or terminate. All sections
are 5 uw in thickness.

Commencement of brain : . Section 31
Commencement of cerebral organ : 5 4160
Termination of cerebral organ . ‘ aj enna
Commencement of excretory system 5 EO
Termination of cesophagus : : » 800
Commencement of generative region . », 1250
Excretory system still found in : 3, 2b20

In the generative region the excretory tubules have almost
disappeared, though traces of them may still be recognised
(fig. 10 *). The excretory ducts are in this region often
incomplete, not penetrating the circular muscle layer. Their
external opening, however, appears to be always present. As
these ducts and the gonads are found in the same region, it
seemed possible that the former might, where present, act as
the ducts of the latter. Such a view, however, appears to be
negatived by the following considerations :

(1) The ducts which are seen to be m connection with the
gonads show a different histological structure, being composed
of long fibrillated cells with slender rod-like nuclei (fig. 10,
gd.). The excretory ducts do not present this fibrillated
structure, and their nuclei are smaller and slightly oval

fig. 10, ead.).

COLLECTION OF NEMERTEANS FROM SINGAPORE. Lt

(2) The excretory ducts generally show a certain amount
of expansion in the circular muscle layer, the gonadial ducts
never.

(3) Spermatozoa may often be detected in the gonadial ducts
though not in the excretory ducts.

(4) The excretory ducts generally pass out just over the
side stem, and take a more or less horizontal course through
the body-wall. The gonadial ducts, on the other hand, are
usually given off somewhat more dorsally, and their later
course 1s directed more vertically than is the case with the
former (fig. 9).

Such considerations appear to show conclusively that,
although these two sets of ducts co-exist in the same region
in this particular worm, there is no connection between them ;
moreover, they point to the improbability of any homology
being established between the two sets of ducts, whose chief
feature in common appears to be that of repetition.

With regard to the nervous system the brain is small
for the size of the animal. The ventral ganglion is small
compared with the dorsal (fig. 6). The cesophageal commis-
sure is well marked and contains ganglion-cells (fig. 6 a, onc.).
The side stems end blindly near the anus without forming a
commissure above or below the rectum. Possibly this may be
owing to injury and subsequent regeneration, but it is impos-
sible to say without more material.

The cerebral organ is small. The gland cells form a
peculiar process directed inwards and ventralwards (fig. 6a,
gcorg).

The cerebral canal opens to the exterior laterally and some-
what ventrally.

Hyes are apparently absent, as careful search over sections
through the precerebral portion of the groove failed to reveal
any structures which might be construed as such. With the
exception of HK. rugosa (7) all the other species of the genus
known possess these sense organs.

The head glands are well marked (fig. 6, hg.) and stretch
back over and under the brain well into the cesophageal region.

122 R. C. PUNNETT.

LINEIDZ.
Cerebratulus natans, n. sp.

Five specimens of this worm were procured from shallow
water during the night time by the use of the tow-net. Mr.
Bedford informs me that he has observed what was probably
the same species swimming with eel-like movements near the
bottom during the day time.

C. natans is from 8—10 cm. long. The anterior 2 cm. are
rounded and about 4mm. in diameter. Posteriorly it becomes
much flattened dorso-ventrally, and the width increases to as
much as 8 mm., which is about six times the depth at this
portion of the body. <A caudal appendage is present (fig. 11).
The snout is blunted, and is marked by a black patch above
and below the proboscis pore (fig. 12). Mr. Bedford informs
me that in life dorso-lateral longitudinal reddish lines were to
be seen. As there is no trace of pigment to be seen in this
region after preservation, either in the epithelium or the
cutis, it is probable that such an appearance was due to the
coloration of the intestinal pouches viewed through the pale
whitish-brown integument of the living animal (ef. Verrill
[8], p. 435). The head slits are well marked, and are about
2°5 mm. in length. The mouth commences rather before
the posterior end of the head slits, and is about 2°5 mm. long.
A shallow ventral groove runs along the flattened portion of
the body, commencing about 15 mm. from the anterior end of
the worm (figs. 11 and 15, vgr.).

The cutis in the esophageal region shows a well-marked
nervous layer directly beneath the epithelium (fig. 21, nep.).
Below this is found a thin layer of circular muscle fibres
(mcc.). The longitudinal muscle layer of the cutis is very
highly developed, and in it are imbedded the large cutis
glands (cgl.), which have a strong affinity for staining re-
agents such as carmalum. The connective tissue layer (c7.)
is fairly well developed though it is to a great extent invaded

by fibres from the external longitudinal muscular coat (mlo.).|

COLLECTION OF NEMERTEANS FROM SINGAPORE. 123

The muscle layers in the cesophageal region are well
developed, the external longitudinal layer (mlo.) being more
than double the thickness of the circular layer (mc.), whilst
this in turn is about twice as thick as the internal longitu-
dinal layer (mli.). Further back the circular layer becomes
relatively smaller, and near the posterior end the outer longi-
tudinal layer becomes very sparse. The horizontal layer
over the mouth is well marked. The dorso-ventral muscles
between the intestinal diverticula are strong. There is no
diagonal muscle layer.

The vascular system anteriorly forms a well-marked
head loop. There is in addition a small dorsal portion sepa-
rated off. The rest of the vascular system is formed on the
usual type for the family as described by Oudemans (6) and
Coe (4). The proboscis vessel leaves the proboscis sheath
about 6°5 mm. from the anterior end.

The alimentary canal is of the usual type. The intes-
tinal diverticula are very deep, and with a very narrow
lumen (fig. 15). There is no ventral gutter posteriorly.

The proboscis shows the usual three layers of muscles,
1.e. outer longitudinal, circular, and inner longitudinal, the
last named being the thickest. Outside the circular muscles
is a well-marked nerve layer. There are two muscle crosses
(fig. 13, mer.) dorsally and ventrally. The epithelium either
dorsally or ventrally (which it is impossible to determine) is
considerably thickened and highly glandular. The outer
borders of the cells are here rendered conspicuous by the
presence of a dark greenish pigment.

The proboscis sheath extends to within a millimetre of
the posterior end, though it is here much reduced.

The generative organs appear as thin-walled sacs
between the intestinal diverticula. There is no trace of
sexual products.

The excretory system commences about 3 mm. from the
tip of the snout, and extends over the same distance. The
tubules, except at the anterior portion, are exclusively ventral
to the level of the side stems. Anteriorly there is a single

124 R. C. PUNNETT.

duct on each side opening just over the level of the side stems
(fig. 37).

The nervous system exhibits the usual type. The
median dorsal nerve is not greatly differentiated from the
nervous sheath enclosing the circular musculature.

The cerebral organ is well developed, and is rounded in
shape though somewhat flattened dorso-ventrally. The
ventral glands are well developed, but dorsally they are
almost absent. The dorsal lobe of the dorsal ganglion ceases
before either the ciliated canal effects a junction with the
ventral lobe or the appearance of gland cells round the
former.

A frontal organ of the typical threefold arrangement is
present at the tip of the snout.

Eyes are absent.

The head glands are well marked, reaching ventrally
beyond the commencement of the brain. Both dorsally and
ventrally they merge into the cutis glands.

Cerebratulus brunneus, n. sp.

The single specimen procured measured about 12 cm. in
leneth and 4 mm. at its greatest breadth. In shape it was
rounded anteriorly, though somewhat flattened posteriorly.
The snout was much blunted. The mouth was small and
rounded. A caudal appendage was present. The colour was
darkish chocolate in life, slightly paler ventrally. It persisted
to a great extent after preservation.

The epithelium is very thick, and near the basement
membrane are a number of deeply staining, small, unicellular
glands. The circular muscle layer of the cutis is thin, and
beneath it is a comparatively thick though sparse layer of
longitudinal fibres (fig. 40, m/c.). The cutis glands form a
compact, thick, and continuous layer, beneath which may be
seen traces of the connective-tissue layer which is almost
absent in this species.

The muscle layers are strong, the outer longitudinal layer

COLLECTION OF NEMERTEANS FROM SINGAPORE. 125

being more than double the thickness of the circular layer,
whilst this in turn is about four times as thick as the inner
longitudinal layer. There is a layer of longitudinal fibres
between the proboscis sheath and alimentary canal (cf.
Lineus versicolor, Birger [8], p. 638). There is no dia-
gonal muscle layer.

The vascular system anteriorly shows a network of small
lacunze instead of a head loop. From this network the main
cephalic vessel, just before the brain, is situated dorsally to
the proboscis sheath (as in the Eupoliide). The cesophageal
vessels are merged into a single large lacuna on each side in
the posterior region of the cesophagus, about 8 mm. from the
tip of the snout.

The alimentary canal presents no special features. A
ventral gutter is present in the intestinal region, which com-
mences about 1°5 cm. from the tip of the snout. It is very
conspicuous.

The proboscis is extremely thin. It contains no muscle
crosses.

In the proboscis sheath the circular layer is extremely
thick in the mouth region. The proboscis sheath is not found
in the posterior 5 cm. of the body.

The generative sacs contain developed ova.

The excretory system commences 3°4 mm. behind the
snout, and extends over 3°3 mm. The tubules do not extend
much, either dorsally or ventrally, over the level of the side
stems. There are ten ducts on the left side and five on the
right (fig. 55). They open to the exterior just above the level
of the side stems.

The nervous system is built on the usual type. The
median dorsal nerve is well marked.

The cerebral organ is well developed and rounded in
transverse section. The dorsal glands are more strongly
developed than the ventral (fig. 27, a—d). he connective
tissue contains a bright yellow pigment (pg.). The dorsal
lobe of the dorsal ganglion stretches over the anterior portion
of the organ. The head shts are deep, though not reaching

126 R. @. PUNNETT.

to the brain. They are continued beyond the level where
the ciliated canal is given off.
Frontal organ and eyes are both absent.

Cerebratulus robustus, n. sp.

A single specimen only was procured. It isa short, thick
form 7 em. long and about 7 mm. broad in the middle, with
a depth of about 4 mm. The snout is very blunted. The
mouth is small and the head slits rather short. No caudal
appendage could be observed, but this may have been due to
an accident. Unfortunately no record was made of the colour
in the living form. In the preserved specimen it is a uniform
pale buffish brown.

The epithelium is separated by a fine line (possibly the
expression of a delicate membrane) into an outer portion con-
taming a few unicellular glands but no nuclei, and an inner
nuclear portion also containing unicellular glands (fig. 20, ep.).

The cutis, which is separated from the epithelium by a
fine basement membrane, possesses a thin layer of circular
and longitudinal muscle fibres, contains a moderately well-
developed layer of glands. Beneath these, again, the connec-
tive-tissue layer is well marked in the cesophageal region,
being thicker than the rest of the cutis.

The muscular system is of the normal type. A diagonal
layer is present between the circular layer and the nervous
layer.

The vascular system anteriorly forms a well-marked
head loop. The blood lacunz in the cesophageal region are
extremely small, and a single one is present in each of the
ventral cesophageal folds. There is no well-marked anasto-
mosis among them, as in most cases. Otherwise the arrange-
ment is typical.

The alimentary canal in the cesophageal region (fig. 16)
presents a well-developed glandular layer beneath the epithe-
hum ventrally—recalling somewhat the condition found in
the Kupolide. The intestine commences about 1°3 cm. from
the anterior end,

COLLECTION OF NEMERTEANS FROM SINGAPORE. 127

The proboscis is very thin and without muscle crosses.

The proboscis sheath reaches nearly to the posterior
extremity of the animal. The rhynchodzeum is very small.

The generative sacs contain immature ova,

The excretory system commences about 2°5 mm. from
the tip of the snout, and extends over about 5 mm. It lies
wholly ventral to the level of the side stems. There are no
excretory ducts opening to the exterior. Moreover, I have
been unable to discover any openings into the alimentary
canal, and as the series of sections was quite complete, I can
only conclude that in this instance the excretory system is
without any openings (cf. (7) Eupolia hemprichi).

The nervous system is built on the usual type. The
upper lobes of the dorsal ganglion are considerably larger
than is general. The median dorsal nerve stands out well
from the nervous layer beneath the outer longitudinal muscle
layer. The ventral commissure is very strong, and where
they separate the ventral ganglion is nearly as large as the
dorsal.

The cerebral organ is of moderate size and rounded in
shape, though posteriorly its outline becomes more oval (fig.
31,d). The dorsal gland cells are few, though the ventral
are well developed. The organ is further removed from the
ventral nerve cord than is usually the case (fig. 31,6). The
dorsal lobe of the dorsal ganglion ends before the canal enters
the ventral lobe.

The head slits reach nearly to the brain, but cease very
abruptly externally at the level of the anterior limits of the
brain.

A frontal organ is present, but it is small.

Eyes are absent.

The head glands are well developed, and extend back-
wards nearly as far as the brain.

Cerebratulus erythrus, n. sp.

Several specimens of this worm were procured, the average
length being between 6 and 7 cm, The shape is rounded

128 R. C. PUNNETT.

throughout, though small side-folds could be made out. The
snout is rather sharply pointed. The head slits do not extend
beyond the commencement of the mouth, except as very
shallow depressions. A small caudal appendage is present.
The colour in life was bright red, and is fairly well preserved
in a glycerin specimen.

The epithelium contains small, non-staining, unicellular
glands in its outer portion, and also small deeply-staining ones
nearer the basement membrane. A circular muscle layer
cannot be made out in the cutis, and the longitudinal layer
(fig. 22, mle.) is feebly developed. The composite glands of
the cutis (cgl.) are small and rounded, and do not stam
appreciably with carmalum. The connective-tissue layer is
thick, and contains a few longitudinal muscle fibres. In front
of the brain this last layer is small.

Of the muscle layers, the external longitudinal in the
cesophageal region is about twice as thick as either the cir-
cular or the internal longitudinal layer, which are of approxi-
mately equal thickness. There is a well-marked horizontal
layer over the mouth. In the snout the longitudinal fibres
are very diffuse. There is no diagonal muscle layer.

The vascular system is of the usual type, a well-marked
head loop bemg present. The dorsal vessel leaves the
proboscis sheath about 4 mm. behind the tip of the snout
(fig. 42).

The alimentary canal presents no special features. The
intestine commences about 8 mm. behind the tip of the snout.
The ventral gutter is large and the gut pockets small.

The proboscis is in all cases missing.

The proboscis sheath reaches very nearly to the pos-
terior extremity.

The generative sacs contain no sexual cells.

The excretory system commences at the level of the
posterior border of the mouth, and extends some way above
and below the level of the side stems. A single excretory
pore occurs on each side near the hind end of the system.

The nervous system is of the ordinary type, the dorsal

COLLECTION OF NEMERTEANS FROM SINGAPORE. 129

commissure being, as usual, much thinner than the ventral.
The latter is found somewhat behind the former. The median
dorsal nerve is well marked.

The cerebral organ is small. Both dorsal and ventral
glands are well developed. The dorsal lobe of the dorsal
ganglion (fig. 6a, ddg.) does not end until the glands of the
cerebral organ have made their appearance.

The head slits are wide and shallow, not reaching more
than halfway to the brain.

The frontal organ is exceedingly well developed, being
very conspicuous in sections through the snout. At the base
of each of its three divisions are a number of small cells. with
large nuclei, bearing a great resemblance to the ganglion cells
of the brain (gf.), though whether they are of such a nature
it is impossible to state definitely. A nervous connection be-
tween the frontal organ and the brain could not be traced.

Hyes are not present.

The head glands are well marked, falling anteriorly into
three divisions round the frontal organ. Posteriorly these
merge into one surrounding the head loop. They take a
hght stain with carmalum, and do not become continuous
with the cutis glands. Neither do they reach backwards as
far as the brain.

Cerebratulus sordidus, n. sp.

The shape is rather rounded anteriorly, more flattened be-
hind. The length is about 6 em., but the tail was somewhat
damaged, for which reason it is not possible to say whether a
caudal appendage was present. The head slits extend beyond
the mouth region as shallow depressions. ‘The colour in life
was a dirty yellow-brown.

The epithelium contains greenish unicellular glands
superficially, and small more deeply staining ones near the
basement membrane. The cutis presents a fine layer of
delicate longitudinal muscle-fibrils, but no circular ones. A
layer of composite lightly-staiming cutis glands occurs (fig.
23, cgl.), beneath which there is a layer of connective tissue

voL. 44, parr ]1,.—NEW SERIES. 1

130 R. C. PUNNETY.

containing muscle-fibrils. These glands become much smaller
posteriorly.

In the muscular system the horizontal layer over the
mouth is well marked. There is no diagonal layer.

The vascular system is of the regular type, except that
the head loop is replaced by a network of lacune.

The alimentary canal shows no special features. The
intestine commences about 6°5 mm. from the tip of the snout.
Ventral gutter very small.

The proboscis is missing.

The genital sacs contain ova, which, however, are not
mature.

The excretory system reaches forwards to the point
where the cerebral organ ends. It extends some way dor-
sally and ventrally to the level of the side stems, the tubules
of the opposite sides nearly meeting at places in the mid-
ventral line. The paired ducts open from the middle of the
system, and the position of the pores is somewhat dorsal.

The nervous system is of the usual type.

The cerebral organ is considerably elongated dorso-
ventrally (fig. 32, ¢.) owing to the large development of both
the dorsal and ventral glands. The dorsal lobe of the dorsal
ganglion ends before the organ starts, and the ventral lobe
becomes greatly diminished in size before entering into its
composition (fig. 32, d.). The head slits reach nearly to the
brain.

The frontal organ is well developed.

Eyes are not present.

The head glands are very small, and rapidly merge into
the cutis glands.

Cerebratulus bedfordii, n. sp.

A single specimen of this species was obtained by Mr.
Bedford, with whose name I have much pleasure in associating
it. It was about 16 cm. long and 3 mm. broad, being rounded
in shape all through. The anterior end is sharply pointed.
The head slits are about 2°5 mm. long, and the mouth, which

COLLECTION OF NEMERTEANS FROM SINGAPORE. fod

is small and round, commences after their termination. No
record was preserved of the colour in lifetime. In the
preserved specimen it is of a pale dirty brown throughout,
and the animal resembles, both in this respect and in its
shape, the smaller C. erythrus. No caudal appendage was
observed, the posterior end being damaged.

The epithelium is almost destitute of the small unicellular
glands so characteristic of the group. A few deeply staining
ones occurred near the basement membrane. ‘The basement
membrane itself differs markedly from that of all the other
species here described in that it is of great thickness, and
almost unaffected by staiming reagents (fig. 24, bm.). Both
circular and longitudinal muscle layers are present in the
cutis. The cutis glands are small, and stain deeply with
carmalum. The connective-tissue layer is well marked,
becoming very strong posteriorly. Its fibres are to a great
extent oblique.

Of the muscle layers the outer longitudinal in the
cesophageal region is more than double the thickness of the
circular layer, which in its turn is thicker than the imner
longitudinal layer. The horizontal musculature over the
mouth is very feeble, and there is a trace of longitudinal
fibres between the alimentary canal and the proboscis sheath.
Circular fibres surround the cephalic lacunz, and a ventral
muscle-cross occurs (fig. 18). There is no diagonal layer.

The vascular system presents no head loop. Instead
there is a network of lacunz separated by fibrillated tissue
containing very large oval nuclei. The two lateral vessels
are not formed until the brain is reached. The dorsal vessel
does not leave the proboscis sheath until 1*2 em. behind the
tip of the snout.

The alimentary canal is of the usual type. The intes-
tinal diverticula commence about 2 em. from the tip of the
snout. <A ventral gutter is well marked.

The proboscis is about 12 cm. long and very stout.
Two muscle-crosses are present (fig. 17, mer.). The epi-
thelium is in part much thickened, and contains a number of

182 > 1 nid hd W
Os R. C. PUNNETYT.

mushroom-shaped bodies conspicuous by their deep green
pigment. They appear to contain numbers of small “rhab-
dites.” To the naked eye the proboscis shows two longi-
tudinal deep green bands in this position. The nervous layer
is considerably thickened here.

The proboscis sheath reaches to within 5 cm. of the
posterior end.

The genital sacs contain nearly ripe ova.

The excretory system commences about 5 mm. from the
tip of the snout, and extends over 7mm. The excretory
ducts are numerous, there being twenty-two on one side and
eighteen on the other. As they pass through the circular
muscle layer they show a bladder-like expansion, which
disappears when they emerge into the outer longitudinal
layer.

The nervous system is of the usual type, with the
exception that the cephalic nerves are very strong and well
marked, and are accompanied with a few ganglion cells
externally (fig. 18). The median dorsal nerve stands out
distinctly from the nervous layer.

The cerebral organ is very small, and to a great extent
overlapped by the dorsal lobe of the dorsal ganglion, which,
however, 1s not well marked, and does not make its appear-
ance until after the ciliated canal has entered the brain
(fig. 30, c.). The gland cells are not numerous, and a separa-
tion of a dorsal and ventral portion does not occur. The
head slits are very shallow and wide.

Frontal organ and eyes are absent.

The head glands are small but compact, and appear to
be merely a rather specialised portion of the deeper cutis
elands.

Cerebratulus insignis, n. sp.

Two specimens of this species were procured, one of which
was about 7 cm. and the other about 10 cm. long. The
shape is rounded, the side folds being small but marked.
The mouth is small, and the head slits are continued beyond

COLLECTION OF NEMERTEANS FROM SINGAPORE. 133

it as shallow depressions. A caudal appendage is present. The
colour is olive-green above, becoming shehtly paler below. A
broad white transverse band occurs round the snout near the
tip. The tip of the snout is nearly black (fig. 14).

The epithelium contains numerous greenish unicellular
elands (fig. 25, ep.). The cutis contains the usual circular
and longitudinal layers of muscle fibrille. Cutis glands are
absent except for two longitudinal streaks anteriorly at the
level of the head slits (fig. 25, a. and begl.). These elands
do not stain appreciably with hemalum, but take a vivid
stain with eosin. The connective-tissue layer is well marked,
and contains a few muscle-fibrils except beneath the cutis
glands. Behind the cesophageal region there are no gland
cells in the cutis.

The muscle layers are of the usual order, the outer
longitudinal layer being considerably thicker than the other
two together. A horizontal layer occurs over the mouth.
There is no diagonal layer.

The vascular system is of the usual type, and there is a
well-marked head loop.

The alimentary canal presents no special features.
The dorso-ventral musculature is very weak. The ventral
gutter is very small.

The proboscis is missing in both cases.

The generative organs contain immature ova.

The excretory system commences before the termina-
tion of the cerebral organ, and lies, for the whole of its short
extent, entirely dorsal to the level of the side stems. The
excretory duct is given off about the middle (fig. 40).

The nervous system shows no special features except
that the brain lobes are rather high and short. The median
nerve is indistinguishable.

The cerebral organ is large compared with the size of
the brain, its extent from before backwards being rather
greater than that of the brain. ‘The dorsal lobe of the
dorsal ganglion ends just as the cerebral organ commences.
Dorsal and ventral glands are both well developed (fig. 29,

134 R. C. PUNNETT.

b—e). The organ is ovoid in shape, its dorso-ventral dia-
meter being the larger. The head slits reach nearly to the
brain.

A frontal organ is present.

Hyes are absent.

The head glands are diffuse, and soon become continuous
with the cutis glands laterally.

Cerebratulus ulatiformius, n. sp.

The single specimen obtained is rather flattened throughout
and about twice as wide as deep. ‘The side folds are marked.
The length is just over 6 em. The mouth commences just
behind the head slits, which end abruptly shortly before
the commencement of the former. The proboscis pore is
not terminal, but is found on the ventral surface 1 mm.
behind the tip of the snout. No caudal appendage was ob-
served. Colour a uniform orange-red. The name bestowed
on this worm is derived from the native word “ ulat,”
which Mr. Bedford informs me is a term applied to such
creatures.

The epithelium contains sniall unicellular unstaining
gland cells. The cutis shows the usual circular and longi-
tudinal muscular fibrille. The composite cutis glands (fig.
26, egl.) are well marked in the cesophageal region, but almost
cease posteriorly.

The muscle layers are of the usual type, the outer
longitudinal being considerably the thickest, especially later-
ally, where the side folds are developed. There is no
diagonal muscle layer.

The vascular system shows a large well-marked head
loop. The dorsal vessel leaves the proboscis sheath 2°5 mm.
behind the tip of the snout.

The alimentary canal is of the usual type, the intestinal
diverticula commencing about 5 mm. from the tip of the
snout. ‘he animal is remarkable in having food inside its
intestine, but the remains are too problematical to make it

COLLECTION OF NEMERTEANS FROM SINGAPORE. 135

worth venturing a suggestion as to their origin. <A ventral
gutter is not present.

The proboscis is lacking.

The proboscis sheath ends rather more than 1 cm. from
the posterior extremity.

The generative sacs contain nearly mature ova.

The excretory system commences before the end of the
cerebral organ. The system is short, and opens by a duct on
each side which passes a little dorsal to the side stem.

The nervous system is in no way remarkable. The
dorsal lobe of the dorsal ganglion is large, but does not reach
as far as the cerebral organ.

The cerebral organ is short, and the dorsal glands are
not strongly developed (fig. 34, a—d).

The frontal organ is well marked.

Eyes are not present.

The head glands are fairly strong, and reach backwards
about halfway to the brain, where they end abruptly.

Before bringing this paper to a close it may be as well
briefly to summarise the points of more general interest which
have arisen in connection with this collection.

(1) A considerable amount of variation occurs in the
striking and characteristically marked Eupolia melano-
gramma, and such variation appears to be continuous rather
than discontinuous.

(2) The excretory system of E. melanogramma possesses
the hitherto unique feature of ducts opening into the alimen-
tary canal.

(3) In E. pholidota the excretory ducts reach back into
the intestinal region, thus co-existing in the same region with
the gonidial ducts. Histological evidence tends to show that
these two sets of ducts are not homodynamous.

(4) In the same genus (Hupolia) the lateral nerve-stems
may form a commissure either above or below the anus, or
else may end blindly.

136 R. C. PUNNETT.

(5) In the Lineide great variations occur in the range and
topography of the excretory system, and also in the number
of ducts. Neither here nor in the Eupoliide is to be found
that incipient metamerism of the ducts which Oudemans
claims to have shown.

(6) The vascular system of the Lineidee shows but little
variation in the different species, except in so far as in the
precerebral region there may be either a well-marked head
loop or else a vascular network—a fact already pointed out
by Birger. It is worthy of note that there appears to be
some correlation between the extent of the excretory system
and the point of exit of the dorsal vessel from the proboscis
sheath. The vessel in all the species here described (except
in one case where the preservation was not sufficiently good
to determine this point) leaves the sheath within a few micro-
millimetres of the hinder termination of the excretory system
wherever the ducts of the latter may be situated (cf.
figs. 35—42),

(7) The frontal organ which characterises the Lineide is
not always present.

(8) The structure of the cerebral organ and of the cutis is
characteristic for the different species.

GatBy Marini LaBoraTory,
Sr. AnprEws, N.B.;
March, 1900.

List oF PAPERS REFERRED TO.

Sai

Bircer, O.—* Untersuchungen iiber die Anatomie und Histologie der
Nemertinen nebst Beitragen zur Systematik,” ‘ Zeit. wiss. Zool.,’
Bad. 1, 1890.

2. Birnerr, O.—‘Siidgeorgische und andere exotische Nemertinen,”

*Z. Jahr. Abth. Syst.,’ Bd. vii, 1893.

3. Birerr, O.—‘ Die Nemertinen des Golfes von Neapel,’ Berlin, 1895.

4. Cor, W. R.—*“On the Anatomy of a Species of Nemertean,” ‘Trans.

Connect. Ac.,’ vol. ix, 1895.

5. Husrecut, A. A. W.—* Report on the Nemertea,” ‘Rep. Challenger,
Zool.,’ vol. xix, 1887.

COLLECTION OF NEMERTEANS FROM SINGAPORE. 137

6. Oupremans, A. C.—“ The Circulatory and Nephridial Apparatus of the
Nemertea,” ‘ Quart. Journ. Mic. Sci.’ (Supplement), 1885,

7. Punnett, R. C.—“On some South Pacific Nemertines collected by Dr.
Willey,” ‘Dr. Willey’s Zoological Results,’ part v, 1900. (In the
press.)

8. Verritt, A. E.—‘‘ The Marine Nemerteans of New England and Ad-
jacent Waters,” ‘Trans. Connect. Ac.,’ vol. viii, 1892.

EXPLANATION OF PLATES 5—8,

Tllustrating Mr. R. C. Punnett’s paper “On a Collection
of Nemerteans from Singapore.”

Reference Letters.

ac. Anal commissure. a/. Alimentary canal. d/v. Buccal vessel. dlve.
Commissure of buccal vessels. 4/c. Blood lacuna of cerebral organ. dm.
Basement membrane. ¢. Cutis. cap. Caudal appendage. cc. Ciliated canal
of cerebral organ. cg/. Cutis glands. corg. Cerebral organ. cu. Connec-
tive-tissue layer of cutis. dv. Dorsal blood-vessel. dc. Dorsal com-
missure. ddg. Dorsal division of dorsal ganglion. dy. Dorsal ganglion.
ep. Epithelium. evd. Exeretory duct. eadi. Internal excretory duct. ez.
Excretory tubule. fc. Fibrous core of cerebral organ. fue. Fibrous core
of side stem. fr. Frontal organ. g. Ganglion cells. ge. Gelatinous
connective-tissue layer of cutis. gceorg. Gland-cells of cerebral organ.
gd. Genital duct. go. Glandular layer of cesophagus. 2/. Head furrow.
hq. Head glands. 2. Cephalic lacuna. 4s. Head slit. id. Intestinal
diverticulum. /év. Lateral blood-vessel. m. Mouth. me. Circular muscle
layer. mec. Circular muscles of cutis. mez. Muscle-fibres and connective
tissue. mcr. Muscle-cross. m/. Longitudinal muscles of proboscis sheath.
mic. Longitudinal muscles of cutis. mi. Inner longitudinal muscle layer.
mlo. Outer longitudinal muscle layer. xc. Lateral nerve-cord. zd. Dorsal
median nerve. ep. Subepithelial nervous layer. 24. Cephalic nerves.
us. Nervous sheath. o/. Cisophageal lacune. ps. Proboscis sheath. +.
Rectum. rhe. Rhynchocelom. 7. Testis. ¢by. Transverse blood-vessel.
tg. Glandular (?) tissue round cephalic lacune. vc. Ventral commissure.
vdy. Ventral division of dorsal ganglion. vg. Ventral ganglion. vgr. Ventral
eroove. vgt. Ventral gutter of intestine.

138 R. OC. PUNNETT.

Fie. 1, a—d.—Eupolia melanogramma, Transverse sections, illustrat.
ing the arrangement of the blood-vessels in the region of the brain. x 45.

Fig. 2.—E. melanogramma. Portion of a transverse section, showing an
excretory tubule opening into the cesophagus. x 110.

Fic. 3.—E. melanogramma. Showing an excretory duct piercing the
lateral nerve-cord. x 110.

Fie. 4.—K.melanogramma, Transverse section, showing ventral position
of anal commissure. X 60.

Fie. 5.—EH.melanogramma. ‘Transverse section of body-wall in cso-
phageal region. xX 110.

Fic. 6.—E. pholidota. Transverse section through brain just behind the
commissures. X 45.

Vic. 6a.—E. pholidota. Transverse section through cerebral organ and
neighbouring structures. X 45.

Fic. 7.—E. pholidota. Dorsal view of anterior end. x 2.

Fic. 8.—E. pholidota. Transverse section through outer body-wall in
the cesophageal region. xX 160.

Fic. 9.—l. pholidota. ‘Transverse section through anterior portion of
intestinal region. 45.

Fie. 10.—E. pholidota. Portion of a section from about the same level
as the preceding, but more highly magnified. x

Fic. 11.—Cerebratulus natans. From spirit specimen. Natural size.

Fic. 12.—C. natans. Anterior extremity, showing pigment patches.
From a formol specimen. Natural size.

Fig. 13.—C. natans. ‘Transverse section of proboscis. x 45.

Fie. 14.—C. insignis. Anteriorend. xX 3.

Fig. 15.—C. natans. Transverse section through intestinal region. x 32.

Fie. 16.—C. robustus. Section through one of the cesophageal folds.
x 110.

Fic. 17.—C. bedfordii. Transverse section through proboscis. x 45.

Fie. 18.—C. bedfordii. Section through snout. x 45.

Fics. 19—26. Sections through the outer portion of the body-wall in the
cesophageal region. All x 160.

Fig. 19.—C. brunneus.
Fig. 20.—C. robustus.
Fig. 21.—C. natans.

Fig. 22.—C. erythrus.
Fig. 23.—C. sordidus.

COLLECTION OF NEMERTEANS FROM SINGAPORE. 139

Fig
Vig

Fig

. 24.—C
. 25.—C
. 26.—C

. bedfordii.
. insignis.

.ulatiformius.

Fies. 27—34.—Sections taken at intervals of 50 p» through the cerebral
organs of the above species. All x 45.

Fig. 27.—C. brunneus.
Fig. 28.—C. natans.

Fig. 29.—C. insignis.

Fig. 30.—C. bedfordii.
Fig. 31.—C. robustus.
Fig. 32.—C. sordidus.
Fig. 33.—C. erythrus.
Fig. 34.—C. ulatiformius.

Fics. 35—42.—Diagrammatic reconstructions of the various systems in the
anterior of the body in the above members of the Lineide. All x 10. The
proboscis and sheath are omitted.

Fig.
Fig. 36.—C
Fig
Fig. 38.—C
Fig. 39.—C
Fig. 40.—C
Fig
Fig. 42.—C

35.—C. brunneus.

. bedfordii.

.o7.—C. natans.
.robustus.

. ulatiformius.

. insignis,

. 41.—C. sordidus.

.erythrus.

PROTOSTIGMATA OF MOLGULA MANHATYENSIS. 141

On the Protostigmata of Molgula manhat-
tensis (De Kay).!

By

Arthur Willey.
With Plate 9.

INTRODUCTION.

In the case of several species of simple Ascidians with
relatively alecithal eggs, it is known that the larva, at or
about the time of fixation, possesses two pairs of gill-cletts or
primary branchial stigmata. The period during which the
stigmatic system of the young Ascidian is represented by
these two pairs of clefts alone may be one of considerable
duration, thus constituting a definite stage of develop-
ment. This is succeeded by another well-marked stage,
characterised by the presence not of three but of four stig-

1 Molgula manhattensis is one of the commonest Ascidians on the
coast of New England. Whereas other Ascidians are eclectic in their dis-
tribution, this species, as Veriill and Smith (‘ Report upon the Invertebrate
Animals of Vineyard Scund,’ Washington, 1874) lave shown, is at home in
all kinds of places. In New Bedford it occurs or occurred in large clusters
on the piles of a wharf. Isent specimens to Professor Herdman, who has kindly
informed me that he is satisfied that they are true Molgule. ‘The species is
the Molgula manhattensis of American authors, and the generic name
Bostrichobranchus, given to it by Traustedt, must rest upon some mis-
understanding which ought to be cleared up. Professor Herdman suggests
(in lit.) that the specific name manhattensis may be misapplied. It may
be so, but I think the burden of proof rests upon the author of the genus
Bostrichobranchus.

There should be no doubt concerning such a common and easily obtainable
Ascidian.

142 ARTHUR WILLEY.

mata on each side of the pharynx. In the next stage there
are five pairs of stigmata; and, finally, the last stage of the
infantile or nepionic period of development is heralded by
the appearance of a sixth pair.

The primary branchial stigmata become greatly drawn out
in the transverse direction (i. e. dorso-ventrally), and even-
tually become subdivided to form secondary branchial
stigmata, which, in their turn, undergo elongation in the
longitudinal direction, and again become subdivided to form
the tertiary branchial stigmata of the adult.

In some cases, therefore, itis highly probable that the
entire stigmatic system of the adult is derived by subdivision
from six pairs of primary branchial stigmata.

Thus the nepionic period of development comprises four
stages, which it is necessary to particularise, because each of
them requires or may require separate morphological treat-
ment; and, moreover, young Ascidians in one or other of
these stages may be collected in the open, and so constitute
specimens requiring diagnosis.

The stages may be named and tabulated as follows, atten-
tion being again drawn to the singular absence of a stage
presenting three pairs of stigmata :

1. Distigmatic stage.
2. Tetrastigmatic stage.
3. Pentastigmatic _,,

Nepionic period. . .
ee Hexastigmatic __,,

Young Ascidians may often be collected in the hexastig-
matic stage. It is often difficult to know what is meant by
stages of development, but it is not difficult in this case, the
stages being quite definite.

The above enumeration of stages refers to species belonging
to the family Ascidiidx, and also, as I shall proceed to show,
to at least one member of the Molgulide. The table does
not apply to such forms as Perophora, Clavelina, Dis-
taplia, etc., where, as shown (and the observation is capable
of ready confirmation) by the researches of Giard (1872),

PROTOSTIGMATA OF MOLGULA MANHATTENSIS. 143

Seeliger (1885), and Lahille (1890), the rule is that the
definite stigmata arise each by separate perforation of the
pharyngeal wall.

A method of formation of the stigmata intermediate in
character between that followed by a form lke Ciona on
the one hand, and Clavelina on the other, is afforded by
Botryllus, as shown by Mr. Garstang, to whose researches
I shall again refer.

Meanwhile I pass on to a brief recapitulation of my pre-
vious

OBSERVATIONS ON CIONA INTESTINALIS.!

1. Distigmatic Stage.—lI could not detect any differ-
ence in the time and manner of formation of the two proto-
stigmata which characterise this stage. They appeared to me
to arise simultaneously, and to resemble one another in all
respects. Accordingly, to account for this similarity and
simultaneity, | suggested a certain interpretation (namely, the
precocious subdivision of a single branchial pouch) which I do
not relinquish, but do not now wish to press further.

Ina recent publication Professor Julin? dissents from my
version of affairs regarding this stage. As he reserves full
details for a future occasion it is not desirable to enter upon
an extended discussion, and I will only touch upon one point.
M. Julin states definitely that the stage now under considera-
tion is preceded by a stage in which only one pair of primary
stigmata is present; in other words, that the two pairs of
stigmata which characterise this stage do not arise simul-
taneously, but one after the other.

These two pairs of stigmata arise before the change of axis
of the body of the larva (which ensues after fixation) has

‘See this Journal, vol. 34, January, 1893, p. 317; also in ‘ Proc. Roy.
Soc.,’ vol. li, 1892, p. 518.

® Charles Julin, 1899, ‘* Contribution & histoire phylogénétique des
Tuniciers ;” ‘ Recherches sur le développement du péricarde, du cceur, et les

transformations de l’épicarde chez les Ascidies simples ;?? ‘ Miscellanées
biologiques,’ Trav. Stat. Zool. Wimereaux, t. vii, see p. 336,

144. ARTHUR WILLEY.

taken place; and the consequence is that they do not at first
occur one behind the other in the longitudinal direction along
lines parallel to the longitudinal axis of the larva, but they
lie somewhat obliquely in a subvertical or dorso-ventral
direction, one being dorsad of the other, i. e. one of them is
nearer the oral or dorsal side of the body, while the other is
nearer the ventral side (cf. my Pl. 30, fig. 2, op. cit.).

The figures 1 to 5 on pl. xxi of M. Julin’s memoir (op. cit.)
represent transverse sections through a tailed larva of Ciona
intestinalis, preserved fifteen hours after hatching. This
larva is described as possessing one pair only of primary
branchial stigmata, in the form of a nearly transverse bran-
chial tube on each side, opening by an internal aperture
into the pharynx and by an external orifice to the exterior.
This branchial tube is regarded by M. Julin as the primor-
dium of the corresponding half of the peribranchial chamber.

In M. Julin’s fig. 1 the right branchial tube is seen to open
internally at the extreme dorso-lateral angle of the pharynx,
whereas in fig. 2, which we are told represents the next
section in the series, it passes through the external (atrial)
orifice of the right side, and the same branchial tube opens
internally through the middle of the wall of the pharynx
1.e. at quite a different (lower) level. I cannot understand
how the right internal orifice of the branchial tube shown in
M. Julin’s fig. 2 can be identical with the right internal orifice
of fig. 1. Remembering the primary axial relations of the
two first-formed stigmata, and their slightly oblique position,
one being below and somewhat behind the other, I should
rather interpret M. Julin’s figs. 1 and 2 as passing respectively
through the two branchial stigmata of the right side, these
stigmata in their turn communicating with the exterior by
way of the right atrial chamber.

It will soon be evident that the two first-formed stigmata
have identical properties and identical fates, thereby creating
a strong presumption in favour of the identity of their origin.

The distigmatic stage of Phallusia mammillata was

described by Krohn (‘ Miiller’s Archiv,’ 1852), who also ad-

PROTOSTIGMATA OF MOLGULA MANHATTENSIS, 145

mitted the simultaneous origin of the two first-formed gill-
clefts.

2. Tetrastigmatic Stage.—This stage has been observed
by Krohn (1852) for Phallusia mammillata, by Ed. van
Beneden and Julin (1884) for Ascidiella scabroides, and
by me (1892) for Ciona intestinalis, and again (1893) for
Molgula manhattensis (see below).

The inauguration of this stage or the transition from the
preceding to this had not been observed before my account
of the process in Ciona.

The ventral ends of the two primordial stigmata of each
side bend towards one another, and their walls even come into
contact; each of them, then, forms by constriction a small
diverticulum, that from the first stigma being directed back-
wards, and that from the second stigma forwards. The two
diverticula from the primordial stigmata thus come to le in
the primary interstigmatic space and are finally cut off from
the parent stigmata. In this quite unexpected way is the
tetrastigmatic stage of Ciona intestinalis introduced.

The second stigma of the distigmatic stage becomes, after
the abstriction of the two intercalary stigmata, the fourth of
the series of protostigmata in the present stage.

3 and 4, Penta- and Hexastigmatic Stages.—These
can be considered together, because in Ciona intestinalis
the fifth and sixth protostigmata arise successively by inde-
pendent perforation. A pentastigmatic stage was described
by Krohn for Phallusia mammillata and a hexastigmatic
stage by van Beneden and Julin for Ascidiella scabroides.

The description of the hexastigmatic stage of the last-named
species given by van Beneden and Julin was the only one with
which [ could compare the corresponding stage observed by
me in Ciona. In making the comparison I was influenced by
Krohn’s statements regarding Phallusia mammillata, and
consequently assumed too great a uniformity in the method
of formation of the protostigmata of simple Ascidians.

Van Beneden and Julin found that the protostigmata of
Ascidiella scabroides were of unequal and irregular sizes,

you. 44, part 1.—NEW SERIES, K

146 ARTHUR WILLEY.

and thought it probable that the size of the stigmata was an
indication of their order of formation. It may or it may not
be so in this particular case, but nothing can be affirmed
without special investigation.

In his valuable “Etude de la branchie chez Ascidiella
scabroides,” M. de Selys Longchamps! criticises my state-
ments concerning the imdependent perforation of protostig-
mata V and VI in Ciona. He says (p. 121):

“Hn fait, rien n’est plus difficile que d’établir qu’un
stigmate est né par perforation; je me demande méme sur
quel fait positif Willey se base, pour attribuer une origine
indépendante aux protostigmates V et VI chez Ciona.”

To this I can reply that, apart from the general conviction
resulting from the comparison of numerous specimens, the
positive observation which demonstrated to me the fact of the
independent perforation of protostigma V is illustrated im my
Pl. XXX, fig. 14 (op. cit.). Here it will be seen that the
ventral ends of the protostigmata I and IV are inclined
respectively towards their derivatives II and III, while V
appears as a minute perforation apart, and is later followed
by VI without any indication of abstriction.

Moreover Molgula manhattensis will afford collateral
evidence in corroboration of my former account of Ciona.

5. Adolescent Period.—If it be thought useful to
recognise different periods of development it 1s necessary to
be precise. Fortunately this is comparatively easy with the
simple Ascidians, which exhibit five well-marked periods,
namely :—l. Embryonic. 2. Larval. 3. Nepionic. 4. Ado-
lescent. 5. Adult.

The nepionic period obviously commences with the fixation
of the larva, but there is a sheht difficulty as to its termina-
tion, due to the fact that there is no such abrupt contrast
between the nepionic and adolescent periods as there is
between the former and the larval period. It is a question
whether it should be regarded as ending with the establish-
ment of the six protostigmata or with the subdivision of these

1 * Archives de Biologie,’ xvi, 1899, pp. 117—171.

PROTOSTIGMATA OF MOLGULA MANHATTENSIS, 147

to form six rows of secondary branchial stigmata. I take the
latter view, while recognising its arbitrary nature.

The adolescent period would then be maugurated by the
change in the long axis of the secondary stigmata (cf. my
pl. xxxi, fig. 18, op. cit.), and characterised by the for-
mation, by subdivision from the secondary stigmata, of the
definitive or tertiary branchial stigmata; and by the develop-
ment of the reproductive organs from the genital primor-
dium.

It is satisfactory to note that collateral evidence in confir-
mation of the account which I gave (1) as to the increase in
the number of stigmata in a transverse row, and (2) as to the
increase in the number of transverse rows in Ciona, has been
forthcoming on the part of M. de Selys Longchamps! in
respect of Ascidiella scabroides:—* La régle générale est
la méme ici que chez Ciona intestinalis; tout nouveau
stigmate d’une rangée n’est autre chose qu’une
partie séparée dun stigmate préexistant.” And
again :—“ L’augmentation du nombre des rangées_ trans-
versales de stigmates ne résulte pas de apparition de rangées
nouvelles entre les rangées préexistantes, mais du dédouble-
ment de ces rangées préexistantes, ainsi que Willey l’a décrit
et représenté fig. 22, chez Ciona.”

The formation of the intercalary stigmata in a transverse
row of Ascidiella scabroides as described by de Selys
Longchamps differs in detail from the method which preyails
in Ciona in that minute diverticula are budded off from the
ends of the pre-existing stigmata in a manner which strikingly
recalls the mode of origin of protostigmata II and III in
Ciona.?

OBSERVATIONS ON MOLGULA MANHATTENSIS.

My observations on Molegula manhattensis were made
during the long vacation of 1893, im the Marine Biological
Laboratory at Woods Holl, Mass., where I was privileged to

1 Op. cit., pp. 124—126.
De Selys Longchamps, op. cit., pl. viii, figs. 5—7.

148 ARTHUR WILLEY.

occupy a table. I collected the material myself in the
company of a friend from the piles of a wharf at New
Bedford.

All the nepionic stages of Ciona, which I observed at
Naples, were reared in jars from artificially fertilised eggs ;
those of Molgula manhattensis were collected in the
open. I gave the gist of my observations on the protostig-
mata of the latter species in the book entitled ‘Amphioxus
and the Ancestry of the Vertebrates,’ ' which I prepared for
the Columbia University Biological Series, and intended to let
that suffice. For various reasons, however, I have come to
the conclusion that it would be desirable to furnish further
details, and that is the object of the present contribution.

As previously noted by Kingsley,” this species is viviparous,
in so far that the embryonic development, during which the
embryo is surrounded by the follicular membrane, takes place
within the peribranchial chamber of the parent.

In the urodele larva the organ of fixation assumes the
form of a hollow lobe at the anterior end of the body, muchas
in Ciona, where I identified it as a preeoral lobe. ‘The main
body of the larva is so opaque that the internal structures
cannot be made out in any detail without sections. I could
not see the buccal and atrial orifices in the larva before
fixation ; but after fixation there appeared, in addition to the
buccal orifice, a-single median atrial siphon. It would be
interesting to have it established exactly at what moment the
mouth breaks through, since this crucial event appears to
vary in relative time of occurrence in different species. In
another Moleulid, Lithonephria eugyranda (nard, Julin®
finds that the mouth is formed long before the hatching of
the larva.

In the larva of M. manhattensis, at the time of fixation,
the preoral lobe elongates as a stolon-like tube, and at the

1 1894, see pp. 2382—233.

2 J. S. Kingsley, “Some Points in the Development of Molgula man-
hattensis,” ‘Proc. Boston Soc. Nat. Hist.,’ xxi, 1882, pp. 441—451. The

account of the segmentation stages was erroneous.
3 Op. cit., see ppe 355-6,

PROTOSTIGMATA OF MOLGULA MANHATTENSIS. 149

same time three or four hollow ectodermic processes grow out
at different points from the main body, and project into the
test (cf. Pl. 9, fig. 5). These processes are contractile, and
their somewhat dilated extremities may sometimes be seen to
execute a prominent systole; they are still to be seen at the
end of the nepionic period (fig. 12).

Under certain imperfectly known circumstances the larva
may undergo a portion of its metamorphosis while still en-
closed within the follicle, so that it passes through no free-
swimming’ stage.!

I am quite unable to say whether such precocious metamor-
phosis is a normal variant, but even if it is a teratological
phenomenon it is of some interest, and would have a more
definite significance if (1) the exact circumstances under
which it takes place were determined, and (2) it were ascer-
tained whether any of the individuals attained maturity.

In fig. 2 a larva is shown whose tail is in regression
within the follicle. One point of special interest in this
larva is the great development of the preoral lobe (organ of
fixation), which extends as a hollow tube containing loose
mesenchyme, close beneath the follicular membrane between
the latter and the embryo. In this position it appears likely
that it would fulfil a respiratory function, since I observed it
were two round pigmented bodies in the larval brain. Often
only one spot is present; one only was observed by Kingsley
in this species, and one by Julin in Lithonephria. I found
reason for supposing that the presence of two spots might be
due to the fragmentation of a single one. What has been

to contract and expand. In the same larva (fig. 2) there

1 This indication of precocious metamorphosis is of importance in com-
parison with the development of other Molgulids, in which, according to the
observations of Lacaze-Duthiers and Kupffer, the urodele stage is omitted
from the ontogeny. It also recalls to my mind some experiments which I
made on species of Styelide some years ago, when I succeeded in hastening
the preliminary stages of the metamorphosis by cutting off the tails of the
larve. [A. Willey, ‘‘ Report on the Occupation of the Table of the British
Association at the Marine Biological Laboratory, Plymouth, during August
and September, 1891,” ‘ Rep. Brit. Ass. (Edinburgh),’ 1892.]

150 ARTHUR WILLEY.

said is enough to show that the larva of Molgula manhat-
tensis is well worth further study, and the teratological side
should not be neglected. I now pass on to the description
of the nepionic stages, which is the special purpose of this
paper.

1. Distigmatic Stage (Pl. 9, fig. 3)—Owing to the
opacity of the larve the earliest stage at which the two
first-formed stigmata could be recognised is that which 1s
represented in fig. 3. In this case the details of the neuro-
hypophysial system could not be made out in the object
viewed as a whole. It is noteworthy that we have here a
median dorsal atrial aperture in place of paired apertures, an
observation which accords with that of Julin on Litho-
nephria.

Now if we follow the growth of the two stigmata, which we
may distinguish by the letters A and B, we find that they
increase in size principally in the dorso-ventral direction, A
being generally somewhat in advance of B (fig. 4). Having
attained a certain transverse diameter, their ventral ends
(i.e. the ends directed towards the endostyle) become bent
backwards towards the fundus of the pharynx (fig. 5). Next
the recurved ventral ends of the stigmata A and B double
round in the dorsal direction so that each of the originally
simple clefts become biramous, having a larger anterior and a
smaller posterior arm (figs. 6 and 7).

2. Tetrastigmatic Stage (fig. 8)—A young Molgula
manhattensis collected in the open is shown in fig. 8,
illustrating the commencement of the tetrastigmatic stage.

Here we note that the posterior arm of stigma A_ has
become separated by constriction from the parent cleft, while
in the case of B the abstriction of its posterior arm has not
yet taken place. This specimen was fortunately in a state of
complete expansion when the sketch was taken, and it pre-
sents the facies of a typical Ascidian. I took special pains to
note that the stigmata were in precisely the same condition
on the right side as that in which they are here represented
to be on the lett.

PROTOSTIGMATA OF MOLGULA MANHATTENSIS. 151

The fact that the growth of stigma A is in advance of that
of B is a detail of some interest 1m comparison with the
behaviour of the homologous stigmata in Ciona intestinalis,
where B tended to take the priority of A.

Thus stigmata A and B produce by abstriction two new
stigmata, C and D. These latter arise, as the figures show,
not in the form of minute diverticula of A and B, as is the
case in Ciona, but as large reduplications equal in longitu-
dinal diameter to the parent clefts, though less in their trans-
verse (dorso-ventral) diameter. Moreover, whereas in Ciona
the stigma D was the third in the definitive series of proto-
stigmata, here it becomes the fourth of the series, owing to
the fact that its abstriction from B takes place in the opposite
direction to that observed for Ciona.

3. Pentastigmatic Stage—I regret that among the
drawings which I still possess relating to M. manhattensis
I can find none illustrating this stage, so that I must refer to
the composite diagram reproduced on Pl. 9, fig. 10, which
was drawn at the time that the observations were made in
1893.!

After the establishment of the tetrastigmatic stage by the
abstriction of C and D from A and B, a fifth protostigma, E,
arises by independent perforation. In M. manhattensis,
by observing many specimens, noting the curvature of the
already formed stigmata (cf. fig. 11), and seeing a new stigma
at its first appearance as a minute perforated disc, it is possible
to demonstrate beyond reasonable doubt that the fifth proto-
stigma is formed independently of the rest. This being the
case, it is the more singular to find that the fifth protostigma,
HK, subsequently behaves in exactly the same manner as A
and B did, namely, it doubles round upon itself at its ventral
end, and the recurved arm is the primordium of the sixth
and last protostigma, F.

4. Hexastigmatic Stage (Pl. 9, figs. 11 and 12).—The
appearance of the protostigmata shortly after the abstriction of
the sixth from the fifth (F from E) is shown in fig. 11. Here it

1 Ci. ‘ Amphioxus and the Ancestry of the Vertebrates,’ 1894, pp. 232-8.

152 ARTHUR WILLEY.

is to be noted that while the primitive relations of each couple
of stigmata are indicated by the slight curvature of their
ventral ends, yet the actual order of their formation could
not be deduced from their appearance and size. For example,
D appears larger than B, from which it in reality was derived.

A later condition of the protostigmata is shown clearly in
the specimen represented in fig. 12. Here we still see the
long-drawn-out dorso-ventral clefts, but the first one has
undergone subdivision into four approximately equal secon-
dary branchial stigmata. ‘The method of division by the
growth of a tongue-like projection across the primitive cleft
is shown in fig, 13.

The vascular trabecule, which traverse the wall of the
pharynx in the longitudinal direction, pass across the proto-
stigmata independently of the tongue-like bridges which
subdivide them. In the specimen represented in fig. 12, only
the first protostigma is subdivided, but there were four
longitudinal vascular trabeculee passing from the buccal region
to the fundus of the pharynx, crossing over the stigmata, and,
in the case of the first stigma, not corresponding with the
true subdivisions of the cleft. They have been omitted from
the lithographed drawing for the sake of clearness.

It should further be noted that at the time when the first
protostigma is subdivided, the sixth is still quite small. I
am unable to say whether any fresh perforation occurs after
the formation of the series of protostigmata, but certainly the
bulk of the secondary branchial stigmata arise by subdivision
of and abstriction from the protostigmata.

The transverse rows of secondary stigmata do not long
retain the simple form which they present in fig. 12. They
soon become arched in the manner shown in fig. 14, where it
is seen that the concavities of the arches face one another.
Next occurs the abstriction of small tertiary stigmata, which
take part in the formation of the intercalary rows (fig. 15).
Finally, by differential growth, the stigmata gradually assume
the spiral disposition which is so characteristic of the Molgu-
loid pharynx.

PROTOSTIGMATA OF MOLGULA MANHATTENSIS. 153

I desire to lay special stress upon the protostigmata A, B,
C, and D. It is often difficult to distinguish between an
actual observation and an interpretation of one. There is not
the shghtest doubt in the world that the origin of the proto-
stigmata C and D from A and Bin Molgula manhattensis
takes place in the manner described in the preceding lines.

There are perhaps two main points of interest which are
brought out by comparing the protostigmata of M. manhat-
tensis with those of Ciona intestinalis. he first is a
question of homology, and the second is one of homodynamy.
The stigma Dis the third protostigma in C. intestinalis and
the fourth in M. manhattensis. It arises from abstriction
from B in both cases, but in reversed senses. We have, there-
fore, here a special instance of enantiomorphism.

The point of homodynamy relates to the two protostigmata
which are the first to be formed, namely A and B. In my
judgment the behaviour of these clefts both in Ciona and
Molgula, and also in Phallusia, shows them clearly to be
homodynamous formations. Protostigma A is obviously homo-
dynamous with protostigma B. This is my contention, but it
is not the opinion of Professor Julin.

In the work to which reference has already been made
(above, p. 143) M. Julin, in developimg an idea previously
expressed by him in conjunction with Professor EK. van Beneden
(1887), maintains that the first primary branchial stigma is
quite distinct in its origin and fate from the second, being
essentially concerned, in its capacity of branchial pouch, with
the formation of the peribranchial chamber. According to
Julin,’ the peribranchial chamber is essentially formed by
dilatation of the walls of the first pair of primary branchial
stigmata, and the part taken by the atrial involutions is quite
subordinate, if not altogether negligible.

"It should be stated that M. Julin’s observations relate principally to
Ciona intestinalis, Lithonephria eugyranda, and Styelopsis
grossularia. In the last two species there is at no stage a paired atrium,
the cloacal aperture being unpaired and median from the beginning. I do

not attempt to contradict his observations on these species, but his interpre-
tation is open to discussion.

154 ARTHUR WILLEY.

M. Julin’s work overlaps mine in regard to Ciona intes-
tinalis, and it follows inevitably from his statements, although
he does not say it in so many words, that protostigma A. is
not homodynamous with protostigma B. I think these two
protostigmata are homodynamous structures, and that 1s how
the matter rests.

The comparison of the protostigmata in M. manhattensis
and C. intestinalis may be tabulated as follows, the stig-
mata bemeg denoted by Roman numerals in their serial, and
by capital letters in their individual capacities :

Protostigmata. Ciona intestinalis. Molgula manhattensis.

5 Ae ae
1 A eee

A
C
B
D
i
F

—
<
ene Srey r

GARSTANG’S OBSERVATIONS ON Borryiius.!

If it is desired to form any idea as to the probability of one
method of formation of the branchial stigmata of Ascidians
having a more primitive character than another, it is necessary
to take note of Garstang’s observations on the formation of
the stigmata in the oozoid and blastozoid of Botryllus.
Reference has been made above (p. 142) to the known fact
that in Clavelina and other Ascidians the definitive stigmata
arise by separate perforations, without the intervention of
protostigmata. Here, therefore, we have two different methods
to consider, and to decide between them which is the more
primitive. The question can only be approached from a basis
of probability, and in this sense it has been answered un-
equivocally by Garstang to the effect that the formation of
protostigmata is a primitive character. Garstang found that

1 W. Garstang, “On the Development of the Stigmata in Ascidians,”

‘Proc. Roy. Soc.,’ vol. lii, 1892, p. 505. In this paper the term protostig-
mata was first introduced.

PROTOSTIGMATA OF MOLGULA MANHATTENSIS. 155

in the oozoid of Botryllus the definite stigmata were pre-
ceded by protostigmata, while in the blastozoid protostigmata
were never formed in any generation of buds, but the definite
stigmata arose separately as small rounded perforations.

The earliest stage observed by Garstang was the tetrastig-
matic stage. He describes (loc. cit., p. 509) the pharynx as
being provided with ‘four pairs of transversely elongated
stigmata, whose transverse diameters are nine times as great
as their antero-posterior diameters. .... They are not
exactly of equal size, but decrease slightly in width in regular
order from before backwards.”

In the next stage observed “the pharynx possesses, in
place of the four pairs of transversely elongated stigmata,
four transverse rows of small stigmata on each side.” Even-
tually a fifth row isadded. Garstang further remarks (p. 510)
that “the primary stigmata—if an inference may be drawn
from their relative sizes—arise one after another in regular
order from before backwards, and that they are subsequently
subdivided in the same order.” If this is so—and it seems
probable that it is so—we should have the following sequence
in Botryllus to be compared with the table given above
on the preceding page:

if A
II B
Ill C
Ly D
Vv .

Having decided that protostigmata are primitive, the dis-
cussion becomes narrowed down to a consideration of the proto-
stigmata themselves. The logic is simple, but of course not
demonstrative. Protostigmata are primitive ; the origin of the
definite stigmata by subdivision of protostigmata is indirect as
compared with the direct method of independent perforation ;
the formation of protostigmata by abstriction is indirect as com-
pared with the direct perforation in regular order from before
backwards, which probably occurs in Botryllus; therefore

156 ARTHUR WILLEY.

the formation of protostigmata by abstriction is probably
primitive. From this point of view it may be concluded that
subdivision of protostigmata has dropped out of the
ontogeny of Clavelina, while abstriction of protostig-
mata no longer occurs in Botryllus.!

RECTIFICATION.

In my former studies on the Protochordata, published in
this Journal,? the main object of which was to obtain an inde-
pendent basis of comparison between the Ascidians and
Amphioxus, I referred incidentally to certain points which
did not fall within the strict limits of my thesis, among these
being the pericardium and heart of Ciona. With regard to
the cavity of the heart, it seemed to me to arise by a splitting
apart of the two layers of what I described as a_ pericardial
septum. In the memoir quoted above, Professor Julin denies
the existence of this septum, and says that the heart arises in
Ciona, as in all Ascidians, by invagination of the wall of the
pericardium. In a form suchas Clavelina, where a longitu-
dinal invagination of the dorsal wall of the pericardium
occurs, the lips of the involution finally come together and
coalesce to form a longitudinal cardiac raphe. Now in
Ciona there is a ventral raphe as well as a dorsal one, as
shown in my fig. 30, Pl. 31, vol. 34, and in some of Julin’s
figures. If, as M. Julin states, the heart of Ciona arises by
invagination and not by the splitting of a septum, then it 1s
necessary to give an explanation of the double cardiac raphe
which has been observed in this Ascidian.®

The question as to the exact origin of the heart in Ciona
is not, however, the special matter to which the heading of this
section of the present paper refers.

1 Garstang states (p. 511) that in the Polystyelid Thylacium sylvani
“eight protostigmata arise on each side of the pharynx, and become sub-
divided, in regular order from before backwards, to form a corresponding
number of rows of secondary stigmata.”

2 Vols. 34 and 35, 1893.

3 See note on p. 158.

PROTOSTIGMATA OF MOLGULA MANHATTENSIS. 157

The point which requires rectification relates to the epi-
cardium. With regard to this important structure I wrote
as follows :'—“I have seen the epicardium in Clavelina, which
was described by van Beneden and Julin, but no trace of such
an organ in Ciona, although it was held as probable by these
authors that it occurred in all Tunicates.’ A few months
later, in vol. 35 of this Journal, appeared a paper by Mr.
Newstead describing the communication of the capacious
perivisceral cavity of Ciona with the cavity of the branchial
sac by means of a pair of retro-pharyngeal orifices.*
More recently M. Damas* has shown (1) that the perivisceral
cavity of Ciona is paired, and (2) that it is the homologue
of the epicardium of Clavyelina. Finally, Professor Julin
(op. cit.) has described the development of this structure from
the beginning to the end.

When it is known that the epicardium (or rather the homo-
logue of the epicardium) is larger in Ciona than in any other
Ascidian, it may seem unaccountable that I should not have
seen it. As a matter of fact, however, Damas finds nothing
surprising therein, since, according to his observations, the
epicardial diverticula occur at a later stage than those
examined by me.

Damas and Julin have shown that the epicardial diverticula
in Ciona are the direct extensions of the fundus of the
pharynx on either side of a ridge called the retro-pharyn-
geal crest, which is in continuity with the right lip of the
endostyle (Julin). Throughout the nepionic period the epi-
cardial diverticula are little more than the somewhat deepened
pharyngeal fundus; in fact, the latter is shown by Julin to be
the primordium of the epicardium.

In Clavelina, as shown partly by Seeliger (1885), and

1 This Journal, vol. 34, pp. 351, 352.

2 A. H. L. Newstead, ‘‘On the Perivisceral Cavity of Ciona,” ‘Quart.
Journ. Mier. Sci.,’ vol.35, July, 1893, pp. 119—128, Pl. 8. The author confirms
van Beneden’s and Julin’s suspicion that the perivisceral cavity of Ciona
represents the epicardiac tubes of Clavelina.

3 PD. Damas, “ Les formations épicardiques chez Ciona intestina is,”
‘ Archives de Biol.,’ vol, xvi, 1899, pp, 1—25, pi. 1—8,

158 ARTHUR WILUEY.

finally by van Beneden and Julin (1887), the pericardium and
epicardium are developed from a common primordium.
Damas, who saw the origin of the epicardial diverticula in
Ciona, found them in complete independence of the peri-
cardium.

Thus Damas describes the origin of the epicardium apart
from the pericardium, while I referred to the latter im igno-
rance of the former.

It seems quite clear that the epicardium of Ascidians is a
very important morphological entity, and, in fact, this has
already been recognised in a practical manner by M. Guard,
who has instituted an interesting comparison between the
Tunicate epicardium and the post-branchial bodies of Verte-
brates!

Norr.—As a matter of fact my original account of the
double nature of the pericardial primordium in Ciona, and of
the origi of the heart by splitting of the pericardial septum,
has been fully confirmed quite recently by Dr. Mare de Selys
Longchamps (“Développement du cceur, du péricarde, et des
épicardes chez Ciona intestinalis,” ‘Bull. Ac. Belgique,’
No. 6, June, 1900). In another article in the same bulletin,
by Selys Longchamps and Damas (‘‘ Recherches sur le
développement post-embryonnaire et Vorganisation de Mol-
gula ampulloides, P. J. Van Beneden”), the following
passage occurs :—“Ce stade a six protostigmates que nous
avons observé chez Vampulloide rappelle un moment corre-
spondant du développement des autres Ascidies simples.
Son origine nayant pas encore été mise complétement hors de
doute, nous ferons remarquer la compléte ressemblance de
la présente description avec celle fournie par A. Willey (1)
chez la Molgule des iles Manhattes.”

' Alfred Giard, “Sur VPhomologie des thyroides latérales (corps post-
branchiaux, Verdun) avec l’épicarde des ‘luniciers,’’ ‘Comptes rendus Soe.
Biol.,’ April, 1898,

PROTOSTIGMATA OF MOLGULA MANHATTENSIS. 159

EXPLANATION OF PLATE 9,

Illustrating Mr. Arthur Willey’s paper “ On the Protostig-
mata of Moleula manhattensis (De Kay).

Letlers of Reference.

at. Atrial or cloacal orifice. Boj. Organ of Bojanus (renal organ). ¢. Cardio-
pericardiac vesicle. cc. Central canal of medullary tube. ev. Cerebral vesicle
with pigment-globule. d.¢. Dorsal tubercle. exd. Endostyle. ep. Ecto-
dermal processes. e.v. [Eudodermic vesicle. /. Follicle. g. Ganglion. m.
Mouth. xz. Notochord. o.e. Gisophagus and its orifice. p. 6. Peripharyngeal
band. p./, Preoral lobe (organ of fixation). sz.g. Subneural gland. - sé.
Stomach. ¢. ‘Tail. ¢.. Tongue dividing cleft. ¢.c. Testa cells. v.¢. Vascular
trabecula.

All the figures refer to Molgula manhattensis,

Fic. 1.—Anterior portion of urodele larva (tadpole) after hatching.
Zeiss 3 D.

Fig. 2.—Urodele larva undergoing precocious metamorphosis within the
follicle. ‘The tail is seen in regression at its insertion into the body, where
there is a mass of histolytic residua. 3 D.

Fis. 3.—Distigmatie stage, after fixation. A and B are the two primordial
stigmata. ‘The pigmented globule (eye-spot) marks the position of the cerebral
vesicle now in regression. 3 D.

Fics. 4—6.—Successive stages in the growth of the primordial stigmata of
the right side.

Fig. 7.—Condition of protostigmata rather later than that shown in Fig. 6,
from the left side.

Fic. 8.—Commencement of the tetrastigmatic stage from the left side.
Between the mouth and the peripharyngeal band are seen four pericoronal
tentacles. 3 A.

Fie. 9.—Protostigmata of preceding.

Fic. 10.—Composite diagram illustrating the mode of origin of the six
protostigmata. [From Willey, ‘ Amphioxus and the Ancestry of the Verte-
brates,’ 1894, p. 233. ]

Vic. 11.—Protostigmata of the right side at the hexastigmatic stage.

Fie. 12.—Entire individual viewed as a transparent object at the hexastig-
matic stage, The first protostigma has become subdivided into four secondary

160 ARTHUR WILLEY.

stigmata, The organ of Bojanus contains a renal calculus. 2 C, reduced by
one third.

Fic. 13.—First two protostigmata at a slightly earlier stage than preceding,
showing the first in process of subdivision. From the right side.

Fic. 14.—Change of shape of stigmata in the early adolescent period. The
figure represents a portion of the third and fourth rows of the right side of an
individual possessing eight rows on the right side, and six on the left.

Fic. 15.—Formation of interealary stigmata by abstriction (the figure does
not show the actual process of abstriction). From left side; about ten rows
of stigmata present in specimen.

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

Quarterly Journal of Microscopical

Science.

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

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

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

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

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

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

TO CORRESPONDENTS.

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

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

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

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

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

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

For every four pages or less—

25 copies : : ‘ ' 5/-
a) eae : : : 4 6/-
BOOS, ‘ 7/-

Plates, 2/- per 25 if uncoloured; if coloured, at the same rate for
every colour.
Prepayment by P.O. Order is requested.
ALL COMMUNICATIONS FOR THE EDITORS TO BE ADDRESSED TO THE CARE
or Mussrs. J. & A. CuurcHILL, 7, Great Marieorovucu Srreet,
Lonpon, W.

THE MARINE BIOLOGICAL ASSOCIATION

OF THE
UNITED KINGDOM.

Patron—H.R.H. the PRINCE OF WALES, K.G., F.R.S.
President—Prof. E. RAY LANKESTER, LL.D., F.R.S.

0%

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

The Laboratory at Plymouth

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

Naturalists desiring to work at the Laboratory

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

Works published by the Association

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

The Journal of the Marine Biological Association

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

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

Specimens of Marine Animals and Plants,

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

TERMS OF MEMBERSHIP.

ANNUAL MEMBERS , : : . £1 1 Operannum.

Lire MemBeErs . ‘ ; 5 . 15 15 O Composition Fee.
FOUNDERS . : 7) LOOM OREO eS Bs
GOVERNORS (Life Menivers ae Coune . 500 O O ag a

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

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

The DIRECTOR,
The Laboratory,
Plymouth.

New Series, No. 174 (Vol. 44, Part 2). Price 10s.
Sqr) MARCH, 1901.

THE

QUARTERLY JOURNAL

OF

MICROSCOPICAL SCIENCE.

EDITED BY

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

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

DIRECTOR OF THE NATURAL HISTORY DEPARTMENTS OF THE BRITISH MUSEUM; LATE FPULLERIAN

PROFESSOR OF PHYSIOLOGY IN THE ROYAL INSTITUTION OF GREAT BRITAIN-

WITH THE CO-OPERATION OF

ADAM SEDGWICK, M.A., F.RS.,
FELLOW AND TUTOR OF TRINITY COLLEGE, CAMBRIDGE 3

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

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

AND

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

BEYER PROFESSOR OF ZOOLOGY IN THE OWENS COLLEGE, MANCHESTER.

WITH LITHOGRAPHIC PLATES AND ENGRAVINGS ON WOOD.

Ay, ONO Ne:
J. & A. CHURCHILL, 7, GREAT MARLBOROUGH STREET.
SOK.

| \ aS
TTL oT OT IE I EIN ee PGT EIEN ye

Adlard and Son,] [Bartholomew Close

CONTENTS OF No. 174.—New Series.

MEMOIRS:

PAGE
The Development and Succession of Teeth in Hatteria punctata.

By H. Srpnczr Harrison, B.Se.(Lond.), A.R.C.Se., Demonstrator
and Assistant Lecturer in Biology, University College, Cardiff.
(From the Zoological Laboratory, Royal College of Science, London,
and University College, Cardiff.) (With Plates 1O—12) : sie Od

The Anatomy of Pleurotomaria Beyrichii, Hilg. By Martin F.
WoopwarD, Demonstrator of Zoology, Royal College of Science,
London. (With Plates 13—16) . : : : é é oe

Dolichorhynehus indicus, nu. g., n. sp., a New Acraniate. By
ARTHUR WILLEY . : : - : : : : 5 . 269

Heteropleuron hectori, the New Zealand Lanecelet. By W.
BriaxLtanD Brenuam, D.Se., M.A., F.Z.8., Professor of Biology in
the University of Otago. (With Plate 17) . i : : -- 273

On some Parasites found in Echinus esculentus, L. By ArtHur
E. Suretry, M.A., Fellow and Tutor of Christ’s College, Cam-
bridge, and Lecturer in the Advanced Morphology of the Inverte-
brata in the University. (With Plate 18) . ; : : 281

The Scottish Silurian Scorpion. By R. I. Pocock. (With Plate 19) 291

APR 5 1901

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 161

The Development and Succession of Teeth in
Hatteria punctata.

By

H. Spencer Harrison, B.Sc.(Lond.), A.R.C.S8e.,
Demonstrator and Assistant Lecturer in Biology, University College, Cardiff.
(From the Zoological Laboratory, Royal College of Science, London,
and University College, Cardiff.)

With Plates 10—12.

INTRODUCTION.

Tue dentition of Hatteria presents certain peculiar and
characteristic features, which have always attracted atten-
tion, and many of which have not yet received adequate ex-
planation. I need only mention here such points as the
structure of the premaxillary and anterior mandibular teeth,
the irregularity in occurrence and number of vomerine teeth,
and the part taken by the edges of the jaws when the teeth
have become worn down. Other matters on which uncer-
tainty exists are the presence or absence of enamel, and the
extent of the tooth change, which is usually stated not to
occur at all. Owing to the great difficulty of obtaining
embryos and young, our knowledge on these points is in-
complete. However, Professor Dendy has recently to a
large extent overcome this difficulty, and most of the ma-
terial on which this investigation is based was sent to Eng-
land by him. In my endeavours to clear up the points at
issue, I have been greatly indebted to Professor Howes and

VoL, 44, PART 2,—NEW SERIES. L

162 H. SPENCER HARRISON.

Mr. Swinnerton, who have generously placed at my disposal
all the slides and preparations made for their research on the
development of the skeleton, and who have also granted me
the privilege of examining their manuscript and figures.
My thanks are also due to Professor Howes for suggesting
this research and for the use of a table in his laboratory,
and to Mr. M. F. Woodward for much friendly criticism and
advice.

Historical.

The first account of the adult dentition of Hatteria is that
given in 1867 by Giinther (1), where he emphasises the acro-
dont character, and points out the important part played by
the edges of the dentigerous bones when the teeth have be-
come worn down. He gives a full account of the macro-
scopic features of the dentition of the adult, but appears
to have examined no individuals with vomerine teeth. He
also describes a young specimen (seven inches long) in
which each of the premaxillary teeth of the adult was repre-
sented by two, and each of the large anterior mandibular
teeth by three separate teeth. In this young specimen he
could distinguish no polished surface on the “alveolar”
edges, but does not speak positively on this point.

The discovery of the occurrence of vomerine teeth was
made by Baur (2) in 1886. His specimen was 210 mm. in
length, and had a tooth on each vomer.

In 1890 Professor Howes (8) examined nine specimens for
vomerine teeth. In five of these no such teeth were to be
found; in one they were bilaterally symmetrical, in one the
last traces could be detected, in one there was a tooth on the
right side only, and in another the right tooth was larger
than the left. In no case did they project into the cavity of
the mouth. Even when teeth were wanting he found a ridge
of the mucous membrane in the proper position on the vomer.
He suggested finally that the vomerine teeth are vestigial
structures, and are undergoing suppression from left to

right.

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 1638

Siebenrock (4) in 1893 described a specimen having two
vomerine teeth on the right side and one on the left. ‘This
was the only instance of teeth occurring on the vomer out of
nine specimens examined. He states that the tips were
covered with enamel.

In the same year Rése (5) stated that he could find no
tooth replacement in Hatteria, although he apparently found
an “Hrsatzleiste.’ He says “ .. . trotzdem findet sich,
besonders im Oberkiefer, hinter den functionirenden Zihnen
eine wohl entwickelte Zahn- oder Ersatzleiste. Am hinteren
Ende derselben findet zeitlebens eine fortwihrende Neubil-
dung von Ziihnen statt.” It is, however, difficult to say from
the context whether he is referring to Hatteria, or Chameleo,
or both. Later, also, he says that in Chameeleo, “und ver-
mutlich auch bei Hatteria,” there is behind the functional
teeth “eine giinzlich functionlose! aber wohl entwickelte
Zahn- oder Ersatzleiste.”

Burckhardt (6) in 1896, in a review of our knowledge of
the dentition of the Sauropsida, expresses his belief that the
form of the anterior teeth of Hatteria is due to concrescence.

About the same time Baur (7) examined the skull of a
young Hatteria (25 mm. from premaxilla to occipital con-
dyle), and found two separate teeth in each premaxilla, an
outer larger and an inner smaller one, the latter having at
its base a replacing tooth, which, however, he says does not
become functional. In speaking of the maxilla also, he says,
“Der fiinfte Zahn besitzt wie der Innere des Premaxillare
einen Ersatzzahn, und der letzte steht frei in einer Alveole.”
He found a well-developed tooth on each vomer, and in
addition an anterior small one on the left side. He states
that the second tooth in the mandible has an ‘ Hrsatzzahn,”
and concludes that Hatteria has an incomplete set of succes-
sional teeth, which, however, are never functional. He does
not give his reasons for the latter part of his conclusion.

Osawa (8), in 1897, stated that the teeth of Hatteria con-
sist chiefly of dentine. He could find no enamel, and to its

1 The spacing is mine.—H. 8. H,

164 H. SPENCER HARRISON.

presumed absence he attributes the early wearing down of
the teeth. He found no true cement; but bone corpuscles
and dentine tubules occur side by side, so that at this point
he considers we have to do with osteodentine.

In 1899, in his paper on the “ Outlines of the Development
of Hatteria,’ Dendy (9) first described the presence of three
separate teeth in each premaxilla and on each side of the
mandibular symphysis, at a period just before hatching. He
suggests that fusion takes place later, the result being the
formation of the four anterior teeth of the adult.

Howes and Swinnerton (10), in a paper at present in the
press, point out that the middle premaxillary tooth on each
side ceases to grow after a certain stage (Stage S), and is
in all probability shed. hey also figure the replacing tooth
lying lingually to it, but owing to lack of material they
do not venture to decide whether this does or does not be-
come functional. ‘They discuss Baur’s observations, and
draw attention to the alternation in size of the teeth in the
embryos and young. ‘They found no calcified vomerine
teeth in any of the embryonic stages investigated. As
regards the substance on the exposed surfaces of the denti-
gerous bones, they accept Tomes’s (11) view that it consists
of true bone, and they have therefore introduced a new term
—‘“‘hyperacrodont”—for describing the secondarily theco-
dont condition produced in some of the teeth by the dis-
position of this bone round their bases.

Material and Observations.

I have examined complete series of sections taken in
various planes, through embryos and young of several stages,
ranging from about the third month of incubation (Q) to a
period some months after hatching. I have also examined
preparations, dissections, and specimens of young and adult
animals. In describing the stages I shall make use of letters
and numbers in the same sense as Dendy, and as Howes
and Swinnerton, adding such other data as may be desirable,

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 165

Stage Q.

Embryo Q, 524, in about the third month of incubation.
The skeleton is mainly cartilaginous, but along the upper
border of Meckel’s cartilage, and in several other regions, a
considerable amount of ossification has taken place. Length
of head as measured from sections 5 mm.

The epithelium along the upper and lower jaws has grown
down into the mesoderm in the form of a well-marked
dental lamina, consisting of an outer limiting layer of
columnar cells, and a central core of more irregular ones
(fig. 1, d.l.). The line of ingrowth is marked on the
surface of the epithelium by a shallow groove, the dental
furrow (fig. 1, d.f.), such as has been described in many
other vertebrates. The dental lamina is on the whole less
developed in the lower than the upper jaw, and in the
former is quite indistinguishable for some little distance in
the middle line anteriorly ; in the latter, also, it is not con-
spicuous in this region.

The most interesting feature of this stage is the presence
of minute teeth, which are perhaps homologous with the
early (embryonic) dentition described by Rose (12, 18) in
the Crocodile, by Leche (14) in Iguana tuberculata, and
by others in various fishes and Amphibians.

The best developed tooth of this first series is shown in
fig. 2. Here we see that the tooth is formed just below the
epidermis, and has a well-defined enamel organ, although I
have been unable to detect any enamel. The presence of
the tooth produces a prominent papilla. The dentine is
formed by the activity of odontoblasts, which do not be-
come so elongated as do those of the next set of teeth, but
the process of dentine formation is apparently the same
in the two cases. In fact, except for the minute size and
the position immediately below the epidermis, the members
of the first series of teeth are very similar to those of the
functional set. Referring to fig. 1, we see that at the labial
attachment of the dental lamina to the epithelium there is

166 H. SPENCER HARRISON.

an enamel organ, in which the deepest layer of cells is
beginning to assume the character of an enamel epithelium,
part of the cells belonging to the epithelium of the jaw at
this point, while the others belong to the labial side of the
dental lamina. There is as usual a great accumulation of
mesodermal cells below the epidermal cap, and also in the
region towards which the dental lamina is growing. ‘This
enamel organ belongs to the first series, and its position is
obviously labial to the dental lamina; with the exception
of the palatine teeth of this series the dental lamina always
occupies this relative position. As Rése (12) says of the
Crocodile, “‘ liegen die verkalkten Ziahnchen der ersten Serie
nach aussen von der Zahnleiste.” Leche also (14) says of
the embryonic teeth of Iguana tuberculata, “ Ks ist
besonders hervorzuheben dass die fraglichen Zaihne nur im
Bereiche der Schmeltzleiste entstehen.”’ In this specimen
of Hatteria I found a pair of calcified teeth of the first series
in both upper and lower jaw; in each case these were the
second in order, the anterior pair above and below being
represented by enamel organs. Of the latter, I found on
each side eight above and seven below, a number which
corresponds fairly well with the number of calcified teeth
occurring at a later stage. This is considerably more than
have been found in Iguana or Crocodilus.

The presence of an enamel organ usually produces an
epidermal papilla, which is probably what Rose (18) found
in the Crocodile and described as follows:—‘“ Die ersten
Zahnanlagen der Krokodile zeigen sich ganz ahnlich wie
die Placoid-schuppen und ersten Zahne der Selachier in
Form von frei tiber die Kieferschleimhaut hervorstehenden
Papillen.” He says also, “Die ersten Zihnchen entstehen
ganz ahnlich wie bei Selachiern durch Verkalkung von frei
iiber die Schleimhautoberflache hervorragenden Schleim-
hautpapillen.” Although I should scarcely describe the pro-
cess of formation of the first teeth in Hatteria in these
terms, I am inclined to think there is no fundamental point
of difference between this and Crocodilus. Rose in the same

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 167

paper refers to ‘ Epithelwucherungen,” which are formed
before the ingrowth of the dental lamina and which pro-
ject above the surface of the mucous membrane; it is not
easy to be sure when he is speaking of these, and when of
the papille produced by the presence of an enamel organ of
the first series! In a recent paper, Laaser (15), in de-
scribing the development of the dental lamina in Spinax
niger, says: “ Die ersten Zihne liegen also meist am
Uebergang vom dusseren Zahnepithel zur Zahnleiste ;”? and
his fig. 4, showing the enamel organ of such a tooth, is
almost a fac-simile of my fig. 1. Here, then, we have a
Selachian whose first teeth originate in the same way as do
those of Hatteria, and in neither case can we accurately
speak of development ‘ durch Verkalkung von frei iiber die
Schleimhautoberflache hervorragenden Schleimhautpapillen.”

In addition to the marginal teeth of the upper and lower
jaws, there are present in a position slightly posterior and
external to the choane, two calcified teeth of the first series,
one on each side. As is seen from an examination of later
stages, this situation corresponds to the anterior extremity
of the line along which the palatine dental lamina is subse-
quently formed. It is interesting that each of these teeth is
situated within a prominent papilla, but owing to the lack of
sections showing their development I am unable to say what
is the time of origin of the papilla, i.e. whether before or
after the tooth is formed.

Taking all points into consideration, the teeth of this
series seem to resemble more closely those of Iguana tuber-
culata than those of the Crocodile. ‘The latter in particular,
according to Rose, form a “‘Cementsockel,” and are eventually
removed by absorption within the mesoderm. In Hatteria,
on the contrary, no such substance is produced, and instead
of passing deeper into the mesoderm the teeth are moved

! In Hatteria the apex of most of these papilla is certainly in the form of
a small group of cells (not shown in my fig. 1), which suggest “ Epithel-
wucherungen.” These may correspond to what Rose describes, but my
specimens do not enable me to say whether they arise before the enamel
organs are formed.

168 H. SPENCER HARRISON.

into the epidermis, and are finally shed about the time of
hatching (see later stages and fig. 12). In these features
Iguana appears to agree with Hatteria.

The general resemblance to the functional teeth which is
exhibited by these embryonic teeth in their development
and structure, causes me to doubt their special homology
with the placoid scale, and I prefer to describe them as
simply belonging to the first or embryonic dentition, which is
no longer functional, but which, no doubt, was so when the
incubation period of Hatteria was shorter than it is at present.

At Stage Q the teeth of the future functional series have
not yet begun their development.

Stage R.

Embryo R, 142, in about the fifth month of incubation.
Length of head from occipital condyle to premaxilla 6°5 mm.

This embryo shows a very considerable advance on the last.
The teeth of the first series are now all calcified, and show
some degree of degeneration, inasmuch as many of them have
become solid nodules of dentine, lying immediately below the
epidermis, and having lost their pulp. The enamel epi-
thelium in some cases is still distinct, and an unmistakable
tendency to the formation of stellate tissue is occasionally to
be seen. The teeth number about fourteen above and fourteen
below, in addition to the palatine pair. We found in Stage
( that there were only six of these teeth in a calcified con-
dition, the others being represented by a slightly greater
number of enamel organs than there are teeth at this stage.
In dealing with vestigial structures of this kind, little
morphological importance can probably be attached to such
differences in time of development, but it is interesting to
note that the first marginal teeth to calcify are those four which
hie near the anterior angles of the jaws, and which, in a
young animal, would be most useful in catching small prey.

The teeth and enamel organs belonging to the future
functional series of the young animal are at this stage con-

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 169

spicuous structures. Many of them are well advanced in
calcification, and a deposit of enamel has commenced. In fig.
3 is Shown the enamel organ of the first premaxillary tooth of
the right side. The mesodermal papilla indents the deepest
portion of the dental lamina at this point, and there is no
indication of a lingual continuation of the latter. The
enamel epithelium is conspicuous, and evidences of the stel-
late tissue are seen, while the mesoderm shows a tendency to
form a capsule. In this tooth no calcification has yet
occurred, and the odontoblasts are still undifferentiated meso-
derm cells. In the second premaxillary tooth, however, there
is a considerable degree of calcification, and an indication of
a prolongation of the dental lamina on the lingual side can
be detected. In both these points the third premaxillary
resembles the first. Comparing these three teeth with the
corresponding anterior three on each side in the lower jaw,
we find that similar relations obtain, i.e. that the second is
more advanced than the other two, and shows indications of
a lingual prolongation of the dental lamina. I shall have
more to say on these points in the sequel. Hach maxilla has
ten teeth, most of which are more or less calcified; but here,
again, there is a well-marked alternation in the degree of
calcification. As sagittal sections only of this embryo were
at my disposal, | am unable to give the precise relations of
the dental lamina in the lateral regions of the jaws. The
mandibular teeth posterior to the first three (eight in number
on each side) are, as regards the first six, well calcified, the
two posterior only slightly so.

The anterior end of the palatine tooth-bearing regions is
indicated by the presence, on each side, of the tooth of the
first series, mentioned in the last section. The dental lamina
bears three enamel organs, which show no calcification.

Embryo R, 162, in about the eighth month of incubation.
Although this is apparently so much older than the last
embryo, the stage in development is actually not much more
advanced, for Dendy (loc. cit.) has pointed out that, for a
long period during the winter months, very little progress

170 H. SPENCER HARRISON.

made, and embryos from the fifth to the ninth or tenth month
of incubation may all be included in his Stage R.

The teeth of the first series have become less tooth-like in
form, and some of them are enclosed in the epidermis (fig.
4, pal. 1). They are not so numerous as in the last embryo
described, and perhaps some of them have already been shed.
They occur as far back in the jaw as the ninth maxillary
teeth of the future functional series.

As regards the future functional teeth, there are in each
premaxilla three teeth, the first being smaller and less
developed than the other two, and the second being the
most advanced.

Hach maxilla has ten teeth, and, as in the last embryo,
these exhibit an alternation similar to that already described,
but in this case I am able to give amore detailed description.
Before doing so I wish to refer to fig. 11, where the
alternating series of the maxilla at Stage S is shown. Here
a large tooth alternates with a small one, and the larger differ
somewhat in shape from the smaller. A comparison of figs.
6 and 7 will illustrate the condition found in frontal
sections at Stage R. Fig. 6 shows a tooth having a consider-
able development of dentine, a lingual prolongation of the
dental lamina (7. d. l.), stellate tissue much broken down and
only separated from the exterior by two or three layers of
flattened cells. Fig. 7, on the other hand, shows a tooth in
which the amount of dentine is much less, though the enamel
organ as a whole is larger; there is no lingual prolongation
of the dental lamina; the stellate tissue is more intact, and is
separated from the exterior by several layers of cells; the
whole structure is obviously in an earlier stage of deve-
lopment. Fig. 6 is the eighth maxillary tooth of the right
side, and fig. 7 is the tooth immediately anterior to it; the
former represents one of the smaller teeth of the alternating
series shown in fig. 11, the latter one of the larger. This
alternation of character, as seen in sections, prevails through-
out the maxillary series, and it is difficult to avoid
coming to the conclusion that we are here dealing

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 171

with two dentitions, the smaller tooth representing
an earlier set, and the largeralater one. ‘he same is
true of the premaxillary teeth, where only the second belongs
to the earlier set. The interpretation of the epithelial strand
lingual to the smaller teeth (fig. 6,7.d.1.; fig. 4, 7. d. 1.) as a
residual dental lamina seems to me to be justifiable, although
at later stages, and at this stage in the lower jaw (fig. 5,
man. lin.), epithelial ingrowths of a different nature arise,
which produce some uncertainty. The fifth pair of maxillary
teeth form an exception to the regular alternation, inasmuch
as though they resemble the larger teeth in most respects,
they have an apparent residual dental lamina. ‘This fact has
an interest in view of the development of a successional tooth
at this point at a much later stage.

Labially to the enamel organs the epidermis is beginning
to grow down into the mesoderm in the form of a band of
cells more or less closely connected with the enamel organs
or with the dental lamina in the intermediate regions (figs.
4, 5, 6, 7, m. lab.). This is the labio-dental strand, and its
further development will be seen in later stages.

The most anterior indication of the palatine dental
lamina occurs at the level of the sixth pair of maxillary teeth.
The epidermal ingrowth is anteriorly a solid roundish mass
of cells, which, traced backwards, is found to assume in
section a crescentic form, the attachment to the epithelium
being by the convex side of the crescent. The inner limb
is the one more closely related to the enamel organs, and
represents the dental lamina. The outer limb, on the other
hand, is more independent (figs. 4 and 5, p. lab.), and repre-
sents a glandular ingrowth, the relations of which will be
presently described. The most anterior tooth of the palatine
belongs to the first (embryonic) dentition, and is situated
lingually to the dental lamina (fig. 4, pal. 1). It is a more
or less triangular fragment of dentine, completely enclosed in
epidermis, and on the point of being shed. Its position is
exceptional, as the teeth of this series are usually labial to
the dental lamina; but I do not attach much importance to

172 H. SPENGER HARRISON.

this divergence from the rule, as this particular tooth is
formed before the dental lamina, and may easily have
reached its present relative position through the growth of
the epidermis. The palatine dental lamina of each side has
four enamel organs, of which the first two only have partially
calcified teeth (fig. 4, p. e. o.; this section passes through the
anterior portion of the first enamel organ). The second is
the only one having an epidermal strand lingual to the
enamel organ. Between the teeth the dental lamina is con-
tinued as a well-marked strand of cells (fig. 5, d. l.) in both
palatine and maxilla: in the maxilla it terminates without
becoming separated from the epidermis; in the palatine
the dental lamina and the lingual epidermal ingrowth run
together behind (as they arise together in front), and form a
roundish knob of cells attached to the epidermis, which,
again, is continuous posteriorly with a groove in the roof of
the mouth.

In the mandible of the last embryo described we saw
that of the first three teeth on each side the second was more
advanced than the other two, and this is the case in the
present specimen. It has the same claim to be considered
a member of an earher series than the others as has the
second premaxillary. There are now ten pairs of calcified
cheek teeth, and a pair of enamel organs without calcification
on the posterior free prolongations of the dental lamine. <A
reference to fig. 11 will show that the mandibular teeth
exhibit an alternation in size similar to, though not so con-
spicuous as, that seen in the maxilla. I have been unable,
however, to convince myself that there is a corresponding
difference in their period of development ; and as the mandible
is In a more advanced stage than the maxilla as regards the
epidermal ingrowths which are so conspicuous at a later
stage, the evidence of a possible residual dental lamina
cannot be relied upon. As a matter of fact, each of these
mandibular teeth has a well-marked epidermal ingrowth on
its lingual side (fig. 5, man. lin.), but the interpretation
of this structure as a residual dental lamina is hardly justi-

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 173

fiable, since at a later stage a similar ingrowth is found
opposite each tooth in the upper jaw, with the position and
appearance of a dental lamina, though the result of its further
development is mainly a glandular groove. On the whole I
regard the alternation in size of the mandibular teeth as
indicative of the presence of teeth belonging to the two
dentitions, which I have shown to be probably represented in
the premaxille, maxille, and palatines. I shall have occasion
to refer to this point again in the sequel.

Posteriorly the dental lamina (which is probably already
involved in the ingrowth mentioned above) leaves the epi-
dermis and runs back for some distance deep in the meso-
derm, until it terminates between the coronoid and dentary.
This portion of it is probably entirely dental in character,
and bears one enamel organ.

In view of the complexity of the epidermal down-growths
into the mesoderm at this and following stages, it will be
conducive to clearness if a short account of the arrangement
of the parts of the mouth in the adult in the neighbourhood
of the teeth be given here.

In both upper and lower jaws the lips and teeth are, as is
usual, separated by a deep groove, for which we may use the
term labio-dental groove. In the lower lip there is a sort of
glandular shelf running along the internal face, and the
groove itself has glands on the face looking towards the
teeth (figs. 14 and 19, J. d. gr.). The groove is developed
from the labio-dental strand (figs. 4—9, m. lab., man. lab.).
Running along the inner or lingual face of both jaws is
a prominent fold or ridge of mucous membrane, reaching
almost as high as the apices of the teeth. That of the lower
jaw is the crista gingivalis inferior (8), that of the
upper the crista gingivalis lateralis. The deep groove
between the crest and the mandible may be called the
mandibular (internal) dental fossa, and that in the corre-
sponding position in the upper jaw the maxillary (internal)
dental fossa. In Hatteria these grooves are formed from
epidermal ingrowths, situated lingually to the teeth (and

174 H. SPENCER HARRISON.

already mentioned), which are superposed on the denta:
laminee in such a way as to lead to great difficulty in dis-
tinguishing the respective limits (see figs. 5, 8, man. lin. ;
9, m. lin.; 14 and 19, lon. gr. and r.d.1.). As in the labio-
dental groove, the side facing the teeth is glandular.

Lying between palatine and maxillary series is a ridge
similar to the others described (crista gingivalis media-
lis), and the internal face of the palatine is covered by
another ridge (crista gingivalis palatine), the two
being connected anteriorly. The palatine has, therefore,
erooves on its outer and inner faces, and they are continuous
anteriorly. We may call these the internal aud external
palatine fossee. The latter corresponds to the labio-dental
eroove, and is formed from an independent ingrowth of the
epidermis (figs. 5 and 9, p. lab.), whereas the former is
developed from an ingrowth of cells bearing the same relation-
ship to the teeth and dental lamina as I have just described
in the case of the mandibular and maxillary lingual in-
growths.

As in the two following stages it is not always possible to
distinguish between the dental laminz and the superposed
ingrowths, I have made use of descriptive terms having no
morphological significance.

Stage R—S.

‘This embryo was intermediate between the Stages Rand 5,
as defined by Dendy, and was 63 cm. in length. It was
found dead in the egg, and its preservation therefore leaves
something to be desired.

In all three tooth-bearing regions there are now present
both external and internal (labial and lingual) epidermal
ingrowths, having extremely varying and perplexing relation-
ships to the enamel organs. It is no longer possible to
indicate with certainty any of the dental lamin, although
they are no doubt represented in the lingual strands
(fig. 8, man. lin.; fig. 9, m. lin., p. lin.). In some cases
these are attached to the lingual sides of the enamel organs

DEVELOPMENT OF TEETH IN HATTERIA PUNOTATA. 175

(fig. 9, ». lin.), and in others they are quite independent of
the latter (fig. 8, man. lin.; fig. 9, m. lin.).

In the maxillary region we find that the labial ingrowth
appears in sections at or near the middle of any one of the
posterior teeth, in the form of a more or less lobed projection
from the side of the enamel organ (fig. 9, m. lab.). In the
space between two teeth it arises independently from the
epidermis, and this is also the case opposite the anterior
maxillary teeth. Its deepest portion is close to the developing
maxilla, and the intervening mesoblast is denser than else-
where. The lingual ingrowth appears anteriorly in the form
of a simple band of cells projecting from the sides of the
enamel organs, and having here the appearance of a residual
dental lamina. Posteriorly it is free from the enamel organs
(fig. 9, m. lin.). The alternating character of the maxillary
teeth is now expressed chiefly in their sizes.

The anterior indication of the palatine ingrowths is found
as a mass of cells attached to the epidermis, and already un-
mistakably glandular in arrangement. These cells can be
traced backwards into the labial and lingual ingrowths, and
in the space between these the enamel organs and _ teeth
appear. ‘The relation is, as before, closest between the lingual
ingrowth and the enamel organs, and fig. 9, p. lin., represents
the condition for all four palatine teeth. The labial ingrowth
is more conspicuous than the lingual, and is throughout
independent of the enamel organs. ‘lhe two fuse posteriorly,
and are continued backwards for a little distance as an
epidermal groove, as in Stage R. In the maxilla there is a
short continuation of the ingrowths posteriorly, free from
the epidermis.

The teeth are well calcified, and have a conspicuous layer
of enamel (figs. 8 and 9, en.). They are not yet fused with
the bone, but are in close proximity to it.

In the lower jaw the teeth are more advanced than in the
upper, and are separated from the bone by a very short
interval, especially on the outer side. ‘The line of subsequent
union is indicated by a band of mesodermal cells, On each

176 H. SPENCER HARRISON.

side of the teeth are the labial and lingual epidermal ingrowths
(fig. 8, man. lin., man. lab.), which are now broad bands of
cells projecting down into the mesoderm beyond the deepest
point reached by the enamel organs; in fact, as seen in
section, they embrace between them the edge of the dentary
as well as the base of the teeth. In the section figured
(fig. 8) the relations of the epidermal ingrowths are very
simple, but in some sections we may find anterior and poste-
rior portions of two enamel organs, the labial ingrowth
appearing as a process of one of them, and the lingual as a
process of the adjacent one. In the anterior and middle
regions of the Jaw the enamel organs are mainly free from
the lingual ingrowth, but the posterior younger enamel
organs are found with an internal connection with it.

Posteriorly the labial and lingual strands fuse, separate
from the epidermis, and pass backwards, as in Stage R,
into the space between coronoid and dentary. The fifteenth
mandibular tooth, the last at this stage, makes its first
appearance on this portion.

Stage 8.

This stage embraces the later period of incubation.

In a skull measuring 13°5 mm. from premaxilla to occipital
condyle the condition of the teeth is as follows (see also figs.
10 and 11):

The premaxillary teeth (fig. 10) are three in number on
each side, the third being considerably the largest. The
maxillary teeth (fig. 11) show a conspicuous alternation in
size, Nos. 1, 3, 5, 7, 9, on the left, being larger than the
intermediate teeth, although the first is smaller than the
others of the larger series. On the right side the first three
are small, the middle one being the largest, while Nos. 4,
6, 8, 10, correspond to the larger teeth of the left side. We
see, then, that although the alternation is regular, there is a
tendency for the first two or three to be small, whichever
series they may belong to, and also that the two sides do not
correspond in this specimen,

DEVELOPMENT OF TEETH IN HATTERIA PUNOCTATA. 177

Hach palatine has four teeth, three anterior larger and
one posterior smaller.

In the mandible, number 3 is much larger than the
first two; in the middle line over the symphysis is left a gap
into which the two first premaxillary teeth are received. The
third mandibular “ bites” posterior to the third premaxillary
tooth. In the cheek teeth (10 and 11 respectively on the two
sides), the alternation in size is less regular and conspicuous
than in the maxilla, and the average size is also less.

The teeth are not simple cones, but are elongated along the
line of the jaw, and overlap one another to some extent, The
large maxillary teeth are markedly rounded on their external
faces, the internal one being flattened; the converse is the
case with the palatine teeth, the flattened faces being turned
towards the maxilla. The mandibular teeth more closely
approach the conical form, as do the smaller ones of the upper
jaw.

All the teeth are more or less triangular in side view; the
apices of the larger ones in the upper jaw, however, point
slightly backwards.

Embryo §, 2.—7°5 cm. in length, and was prematurely
hatched. The length of skull as measured from sections was
11°5 mm., shghtly smaller than the last.

There are very few teeth of the embryonic series now
remaining, and these are evidently being forced to the
exterior by the growth of the epidermis. The epidermal
cells form a more or less regular capsule, and the tooth
usually shows no signs of absorption, being probably shed
entire (fig. 12, 7. 1).

In the functional series, we find the first premaxillary
tooth smaller than the other two and further removed from
the surface.

The maxillary teeth are about equal in number to those
in the skull just described, and show a similar alternation.
Posteriorly the dental lamina takes a curious upward then
forward course (fig. 16, d.1.), previously losing its direct
connection with the epidermis. The eleventh and twelfth

VoL. 44, PART 2.—NEW SERIES, M

178 H. SPENCER HARRISON.

teeth are borne on this portion (fig. 16,¢.0.f.; this is the
eleventh), as is shown in the figure. The twelfth is only
indicated in the section drawn by a slight swelling anterior
to the eleventh.

In number and size the palatine and mandibular teeth
show no special features, and correspond with those of the
skull described. The mandibular dental lamina (fig. 16, d. J.)
has a free posterior continuation, as in earlier stages, and this
bears a small, slightly calcified tooth and the enamel organ of
another (neither shown in fig. 16).

The most interesting observation I have made in this
embryo relates to the vomerine dentition. I have shown in
the historical portion of this paper that the occurrence of
vomerine teeth is comparatively rare, and that considerable
variation and inconstancy obtains in their relative size and
even in their number.

In the present embryo I find on the left side, very near the
middle line, a longitudinal thickening of the epithelium of the
palate, situated below the posterior portion of the vomer.
Indenting the middle portion of this thickenmg is a meso-
dermal papilla and a well-marked early enamel organ formed
from the overlying epithelium (fig. 13, V. e. 0.). Between
this structure and the bone there is a dense mass of flattened
mesodermal cells. In a corresponding position on the right
side is a slight epidermal thickening, but only a faint though
unmistakable indication of an enamel organ and papilla.
The condition on this side is, I believe, degenerate as com-
pared with the other, and it is not a case of later deve-
lopment.

Returning to the teeth of the future functional series in
the upper and lower jaw, we find that the labial and lingual
epidermal ingrowths are now very conspicuous structures,
and the former at least is beginning to form large cavities
and to show a lobed character, indicative of the future
glandular structure of a portion of its constituent cells. The
cavities are the first stages in the formation of the labio-dental
groove, which is found in the next embryo (fig. 14, /. d. gr.).

DEVELOPMENT OF TEETH IN HATTERIA PUNOCTATA. 179

The teeth themselves have not yet cut the gum, but they
cause conspicuous projections from the surface of the jaws.

In the lower jaw fusion with the bone is on the point of
taking place in some cases and has already occurred in others.
Tomes, as the result of the study of several reptiles (17, 18),
stated that the occurrence of cement in this class is com-
paratively rare. Santa Sirena (19) mentions a “ falsche”
cement in Lacerta agilis, and in Bronn’s Tierreich (20)
cement is described as differing from dentine in the absence
of dentinal tubules and the presence of “so-called” bone cells.
I have not succeeded in finding a satisfactory account of the
development and structure of cement in reptiles, and Hatteria
in this respect confirms Tomes’s results in the slowworm
and green lizard. Up to the stage immediately before fusion
the respective territories of the osteoblasts and the odonto-
blasts are well defined. When dentine and bone are almost
in contact there is a layer of flattened osteoblasts intervening,
and these become included in the bone or pushed to one side,
fusion taking place without the production of any third
substance whatever. The bone is very simple in structure, and
the dentine at the base of the tooth usually has its dentinal
tubules indistinctly shown, so that, as we shall see in the next
embryo, the transition from tooth to bone is not an abrupt
one as far as appearance is concerned, though the presence of
bone corpuscles is of course diagnostic (fig. 15). I do not
agree with Osawa’s use of the term ‘‘ osteodentine” for the
transitional region.

Another very interesting feature shown in this embryo is
the fusion of teeth in the lower jaw. Most of the cheek
teeth have here begun to fuse with their anterior and pos-
terior neighbours, the enamel organs however apparently
remaining distinct from one another. The odontoblasts are
congregated in great numbers at the points of fusion. We
have here an undoubted case of concrescence in
the wider sense of the term. This process does not
occur, at any rate to the same extent, in any other tooth-
bearing region, and we shall see that it is the probable cause

180 H. SPENCER HARRISON.

of an important difference between the dentary and the other
dentigerous bones as regards their tooth supply.

Embryo S, 19a4.—This was a newly-hatched individual
and considerably more advanced than the last. In transverse
sections through the lower jaw the fully formed teeth are
seen to be firmly united with the jaw, and many have cut
the gum (fig. 14). The internal surface of the tooth is lmed
by a layer of slightly flattened odontoblasts (fig. 15, od.).
The labio-dental and the internal dental grooves are now
formed, and the glandular structure is evident (fig. 14,
lL. d. gr., lin. gr., gl.). At the base of the internal groove is a
ridge of cells which is non-glandular in structure, and prob-
ably represents the persistent dental portion of the labial
ingrowth (r. d. l.). The crista gingivalis (c. g.) has also
become cut off by the formation of the internal fossa. At
the junction of tooth and bone the attachment of the inner
limb of the V-shaped sides of the labio-dental groove to the
calcified tissue 1s very intimate (fig. 15), and the same is the
case to a less extent with the corresponding limb of the
internal groove, although in the section (fig. 15) it had
broken away. There is a thin layer of mesoderm between
the epithelium and the bone, below the attachment of the
former, and this is concerned in the formation of new bone
lamelle at later stages.

The teeth are at this stage typically acrodont,
and in the lower jaw their fusion in a longitudinal direction
can clearly be seen.

The enamel has been completely removed by decalcifica-
tion in the tooth figured.

Stage T and later.

Stage 'l' includes young individuals, from a few weeks to a
few months old, measuring 15—17 em. in length.

In a skull measuring 19 mm. from premaxilla to occipital
condyle, on the right side the condition of maxillary and
palatine teeth is much the same as in the skull of Stage S,
but the general size is somewhat larger. he chief interest

DEVELOPMENT OF THETH IN HATT'ERIA PUNCTATA. 181

lies in the premaxillary teeth (figs. 17 and 20). Here we find
that the second tooth has remained of about the same size as
at Stage S, whilst the first has grown considerably. The
third is still the largest. A stained and mounted preparation
of the premaxilla shows that No. 2 is evidently about to be
shed, and that there has been a large absorption of bone at
its base internally. From the evidence of sections of this
stage (see below), there was no doubt a successional tooth at
this point, unfortunately removed in making the preparation
(fig. 20).

In the lower jaw the teeth have increased to sixteen on
each side. The alteration in size is fairly regular, except in
the case of the first five teeth. There is a short gap ante-
riorly between the teeth of the two sides over the mandibular
symphysis.

Sagittal sections of the premaxilla of another individual
of this stage show that the labio-dental groove has been
formed ; the internal dental fossa is not typically developed
in this region (fig. 18). From the deepest portion of the
epithelium of the internal fossa a dental lamina projects into
the mesoderm. This is found in the whole premaxillary
region, and opposite the second functional tooth it bears a
well-developed enamel organ with a calcified tooth (fig. 18,
t.s. 1). A similar condition of the dental lamina is found in
the lower jaw at this stage (fig. 19, 7. d.l.), but there are
at present no signs of successional teeth. The second pre-
maxillary is shown to be separated from the inner table of
the bone, absorption having taken place at the base of the
tooth. In some sections a portion of bone is seen to be still
attached to the latter, and is probably cast away with it,
perhaps representing what ‘Tomes (‘Manual of Dental Ana-
tomy’) calls “bone of attachment.” On the labial side also
bone absorption has commenced. As is seen in fig. 20 the
absorption on the lingual side extends to the bases of the
first and third teeth, and in addition, sections show that
there is a separate small area of absorption at the base of
the third. I shall prove later that each of these three teeth

182 H. SPENCER HARRISON.

is shed, and that successional teeth arise at different times to
fill their places.

The number of teeth in the maxilla at this stage is eleven,
of which the last two are behind the others in development,
and are attached to a posterior free prolongation of the
dental lamina. The labio-dental and the internal maxillary
grooves with their glands are now obvious structures, and
the dental lamina arising from the deepest part of the
internal groove has a posterior prolongation, to which the
last tooth is still attached. The palatine has six teeth, the
last two resembling the corresponding maxillary teeth in
their state of development.

The fifteen teeth of the mandible are succeeded poste-
riorly by two enamel organs, the posterior of which is in a
very early stage of development, the anterior being well
calcified.

My observations had reached this stage when, through the
further kindness of Professor Howes, I was enabled to
examine two fine young specimens, measuring respectively
17°8 cm. and 21:2 cm. in total length. These, which filled
the gaps in the Howes-Swinnerton series, were by a fortu-
nate coincidence received at the very moment I was in need
of them. They were the gifts of Professor A. W. P. Thomas,
of Auckland, New Zealand, and from their study I am able
to show that there is not only a development of successional
teeth in Hatteria, as was first stated by Baur, but also an
actual tooth replacement, denied by him and others. We
may consider first the smaller specimen, 17:8 cm. total
length, 7-2 cm. from anterior border of pelvis to snout.

In place of the three premaxillary teeth on each side,
as seen in Stage T' (figs. 17 and 20), we find in this specimen
only two. It is clear that at any rate the second tooth
(counting from the middle line) has been shed. Not only so,
but the first has also fallen out, and the successional tooth
(figs. 18 and 20, ¢. s. 1.) has replaced them both, as seen in
an entire preparation of the right premaxilla (fig. 21). In
the same preparation (and figure) we see that No. 3 is

DEVELOPMENT OF TEETH IN HATYTERIA PUNCTATA. 183

becoming detached from the bone by absorption at its base
internally, and also that there is a further absorption of bone
at some depth, a developing tooth (#. s. 2) being found at
this point. Serial sections through the left premaxilla
confirm all these observations, and show that in all essential
respects the second successional tooth resembles the first
(fig. 18, ¢. s. 1) in its relations to the neighbouring parts.
Comparison of figs. 20,21, and 22, will make clear the order of
tooth change in the premaxilla up to the stage of my largest
young specimen.

The number of maxillary teeth on each side, visible with-
out dissection, was thirteen, regularly alternating in size
throughout, although the three anterior differed somewhat in
size from the following teeth. Microscopic examination
shows that the first tooth is one which has only recently been
formed (fig. 23, t. s. 1). Its pulp cavity is wide, its enamel
is almost undamaged, it is not yet fused with the bone, and it
has evidently just grown up from the area (indicated in the
figure by lighter shading), where bone absorption had been
proceeding. ‘There is no doubt as to the successional cha-
racter of this tooth, although I have not found its early
stages in any of my specimens. Iam unable to say whether
it replaces two or only one of the preceding teeth.

The palatine teeth are six in number on each side, some
being large and others small, but they show a less definite
alternation than do the marginal teeth.

In the lower Jaw we find on each side three anterior teeth,
the outer being a little larger than the other two. Posterior
to these are thirteen teeth alternating in size, but all of them
small, and further back still are three or four larger teeth
of uniform size. Here, then, we see that whereas in the
maxilla all the teeth belong to the alternating series, in the
mandible there is already a commencement of a posterior
uniform series. In this specimen I found no indication of
replacing teeth in the lower jaw.

In the middle line of the roof of the mouth, about opposite
the ninth maxillary tooth, and just anterior to the line joining

184 H. SPENCHR HARRISON.

the first palatine pair, I found a conspicuous longitudinal
oval swelling, about 3 by 1 mm., having a whitish appear-
ance in the spirit specimen. Closer examination showed
that in the anterior portion of this swelling were two small
teeth, one on each side the middle line, and about 1 mm.
apart, the left being slightly anterior to the right. Both had
their apices directed backwards and inwards, and they had
not broken through the oral epithelium. These were, of
course, the vomerine teeth, of which the only other indi-
cations I have found occurred in a specimen of Stage § (see
fig. 13).

Fig. 24 represents a section passing through the middle of
the right vomerine tooth, and through the posterior part of
the left. Considering the former, we see that it is a well-
developed tooth firmly fused with the underlying bone and
showing no signs of degeneration. As regards the left
tooth, however, the structure is different. In the figure it
appears in two portions, this being due in part to the back-
ward curve of the tooth; but its shape in this section is
also influenced by the fact that there is a large aperture
on the outer side, through which the surrounding meso-
derm is in free communication with the pulp cavity. In
other sections, also, the tooth is seen to have a somewhat
irregular shape and imperfect construction. These facts
point to a condition of degeneracy, and it seems possible
that the gap in the dentine arose from a similar imperfection
of the enamel organ. I have not found any enamel on either
of the teeth, but this may have been entirely removed by
decalcification. Siebenrock, who described a specimen with
vomerine teeth, stated that the tips were covered with enamel,
and this may very well have been the case in the present
instance. Reference to fig. 24 will show that there are large
ingrowths of epidermis closely applied to the internal faces
of the teeth. ‘These are, however, continued anteriorly and
posteriorly into more or less glandular structures. How far
we should be safe in assuming the possibility of a tooth
change in this case may be inferred from the previous

DEVELOPMENT OF TEETH IN HATYERIA PUNCTATA. 185

observations on the subject of such ingrowths. Here on
the vomers there are several of these structures at various
points. Anterior to each tooth is a thin, flat band of epi-
thelial cells lying freely below the epidermis, and continuous
posteriorly with the epidermal cells in contact with the teeth.
I am inclined to regard this as dental in its nature, and as
perhaps giving rise to the occasional anterior small vomerine
tooth described (on one side only) by Baur (7) and Siebenrock
(4). There are no indications of this tooth in my specimen.

The larger specimen of the last two examined measured
21:2 cm. in total length, and 9 cm. from tip of snout to ante-
rior border of pelvic girdle.

In this individual, as in the last, each premaxilla has two
teeth projecting beyond the gum. A microscopic preparation
of the right side, however, shows that they are not the same
in the two specimens (fig. 22). The following is my inter-
pretation of the tooth change here, although it is possible
that other successional teeth may have intervened between
the conditions shown in figs. 21 and 22.. I believe that No. 3
of the newly-hatched animal has now been shed, and is re-
placed by the second successional tooth (¢. s. 2). The first
successional is still in place, but it is on the point of being
shed to make way for the third (¢. s.3), which is nearing the
surface. I have confirmed the appearances seen in this
entire preparation by sections through the left premaxilla.

In the maxilla there are thirteen and fourteen teeth
respectively on the two sides (fig. 27, for the left side).
Throughout the whole length a large alternates with a small
tooth. An entire preparation of the left side shows that the
anterior portion is undergoing tooth change. In addition
to the first successional (¢. s. 1) seen in the last specimen, the
second in order is now also a successional tooth (¢. s. 2).
Judging from its size and extent it has replaced two teeth
of the earlier set. Finally, there is a third successional tooth
(t.s.3) lying deeply seated above the fourth in order, and
pointing with its apex towards the fifth. I have not found
any indication of a tooth change posterior to this point up to

186 H. SPENCER HARRISON.

this stage. In sections of the right maxilla I have found
that at the base of the first successional tooth a successor
has commenced to form. The indications of an active tooth
change in this portion of the maxilla are therefore very
marked. ‘The third successional mentioned here is no doubt
the one discovered by Baur (7) in the specimen in which he
found also the first successional tooth in the premaxilla, and
the first in the mandible (to be presently described). He
appears, however, to have worked from dissection only, and
it is difficult to see how he arrived at the conclusion that
these teeth never become functional. The “alveolus” he
alludes to as containing the teeth was no doubt the cavity
caused by bone absorption. His specimen was evidently
slightly younger than the one under consideration.

From the foregoing it is clear that there are at least four
successional teeth developed in the anterior part of the
maxilla. Sections through the posterior maxillary tooth
show that this is large, well developed, and completely fused
with the bone. The dental lamina is continued beyond this
point for a short distance, but has no traces of enamel organs.
From this fact, and other considerations to be presently
adduced, I conclude that there is a halt in the formation of
maxillary teeth, commencing when about sixteen! of the
alternating series have been produced, and lasting for a con-
siderable period beyond the stage of the specimen now
described. It follows that these teeth (all of the alternating
series) or their vertical successors, when present, are functional
during a large part of the period of growth of the animal.
Subsequently new teeth are added from behind forwards,
and these are uniform in size. At this stage they have not
commenced their development.

Each palatine has seven teeth, which show differences in
size but no very regular alternation. Ina series of sections
through this region I find the enamel organ of a succes-
sional tooth situated lingually to the second functional, the
usual absorption of bone being in progress. The dental

1 Probably more, 18—19. See Addendum.

DEVELOPMENT OF TKEVH IN HATTERIA PUNCTATA. 187

lamina dies out immediately posterior to this point. The
seventh palatine tooth is, like the last maxillary, fully formed
and fused with the bone. In this case the dental lamina,
which reappears here and is continued some little distance
backwards, bears two enamel organs, both in the earliest
stages. Here, again, we have an indication of a cessation of
tooth formation, but from the condition in the adult, and
from the above observations, this is much less prolonged than
in the maxilla. The two early enamel organs are doubtless
destined to form teeth of the uniform series, which are in
the adult proportionately much more numerous here than in
the maxilla, commencing their formation as we see at an
earlier period.

I have found no signs of vomerine teeth in this speci-
men.

The number of teeth in the two halves of the mandible are
twenty and twenty-one respectively. Although we may still
single out three more conspicuous front teeth on either side,
these are not the same as those of the newly-hatched animal,
or even of the last specimen. ‘lhe second is clearly a succes-
sional tooth (fig. 26, t.s. 1), and a comparison shows that it
has replaced Nos. 2 and 3. In Baur’s specimen it was still
below the gum. Inaddition to this, as fig. 26 shows, I found
a second successional tooth, deeply seated below the fourth
and fifth in order. I have confirmed these points in sections,
and find that the first tooth in order shows signs of being
deciduous. There is a conspicuous dental lamina attached
to the side of the internal fossa nearest the tooth. It extends
round the anterior curve of the jaw, and dies out just posterior
to the fifth in order. It has aswelling between the second and
third teeth, which is perhaps an early enamel organ. I have
found no other indications of tooth change in the mandible.
Returning to the macroscopic features of the mandibular
teeth we find that, posterior to the first two, there are thirteen
teeth belonging to the alternating series. Posterior to these,
again, are five larger teeth uniform in size. In the last
speciinen we found also thirteen alternating teeth, and three

188 H. SPENCER HARRISON.

or four of the uniform series. We see, then, that, whereas in
the maxilla the uniform series of teeth has not yet commenced
its development, and is represented by two early enamel
organs in the palatine, the teeth are already being rapidly
produced in the mandible as the jaws lengthen. he only
evidence of the cessation of tooth formation in the mandible
was found at Stage T’ (see p. 180), when there were sixteen
teeth of the alternating series, followed posteriorly by one
early enamel organ. Stated briefly, then, the position is as
follows :—at a stage (differing secondarily, as I believe, in
point of time) in the dentition of the three bones under
consideration, when the mandible and maxilla have in the
alternating series about sixteen teeth each and the palatine
six or seven, there is a more or less prolonged cessation of
tooth development, this being longest in the maxilla, shortest
in the mandible, and intermediate in the palatine. Up to
this stage the teeth have shown the alternation in size so
often mentioned here, this characteristic being least marked
in the palatine. After this period of inactivity, there is a
renewed formation of teeth from behind forwards,! com-
mencing earliest in the mandible, latest in the mavyilla, and
continuing during the whole period of growth of the animal.
I shall attempt to explain these facts in the next section.
There remain to be discussed a few points of special
interest with regard to the succession of teeth. In my
description of Stage R, 1 put forward the view that the
smaller teeth of the alternating series represent an earlier set
than the larger. ‘lhe evidence for this was more conclusive
in the upper jaw and in the anterior part of the lower than in
the posterior part of the latter, where, in my specimens, size
alone suggested any difference in the nature of the teeth. I
also found that the second palatine gave marked indications
of belonging to the earlier dentition. The mandibular alterna-

1 The posterior teeth of the alternating series are also, as we have seen,
formed on backward prolongations of the dental lamin. In the case of this
series, liowever, there is no cessation of tooth formation until the set is com-
plete.

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 189

ting teeth are subject to great modifications in later stages,
and this has no doubt affected their earlier development, so
that we may well consider them as originally similar to the
rest. At Stage R, then, No. 2 premaxillary, Nos. 2, 4, 6, 8,
10 maxillary (though here this is not quite constant as far as
the first five are concerned), No. 2 palatine, and Nos. 2, 4, 6,
8, 10 mandibular, belong to the earlier set of smaller teeth.
It is, I consider, a support to this view, that amongst the first
successional teeth to be developed are those formed lingually
to No. 2 premaxillary, and No. 2 mandibular, the first
palatine successional tooth being also lingual to the second
in order. In the maxilla the relations are not so clear,
though the third successional is lingual to one of the smaller
teeth, the first and second giving no indications of their exact
place of origin in any of my specimens. The second succes-
sional of the mandible lies at the base of two teeth, though
from its relations I believe it to have originated lingually to
the smaller one. I have already noted that these successional
teeth replace, undoubtedly in three cases, and very probably
in two others, two deciduous teeth of the alternating series.
Now according to my view, in the premaxille of Hatteria the
following dentitions are represented :

First, or embryonic—those shed during incubation.

Second,—the smaller teeth of the alternating series (i.e.
the second in order).

Third,—the larger teeth of the alternating series (i.e.
the first and third).

Fourth,—the first and second successional.

Fifth,—the third successional (which, as is required by
my theory, evidently replaces only one of the preceding
dentition).

I have chosen the premaxillz for illustration, because they
retain their successional teeth in greater number than do the
other dentigerous bones, as far, at least, as my specimens
enable me to decide.

I have endeavoured to correlate and account for the fore-
going anomalies of tooth succession by means of the following

190 H. SPENCER HARRISON.

theory :—During the phylogeny of Hatteria there was pro-
bably a period when emergence from the egg occurred much
earlier than it does at present. The presence in ontogeny of a
complete though degenerate set of embryonic teeth, reaching
their maximum of development so long as nine months before
the end of incubation, has already suggested this to us. Let
us now proceed on the assumption that at one time incubation
terminated at this stage, and that the young animal was set
free with a set of these small teeth. In the ordinary course
these would be followed by a second set, which are now
represented by the smaller teeth of the alternating series, and
these would in their turn be displaced by the third set, now
appearing as the larger teeth of the alternating series. After
these would follow, as in more normal reptiles and lower
types, a constant succession of teeth. For some reason the
period of incubation became greatly lengthened—perhaps
owing to difficulty in obtaining food during the winter months
or to changed climatic conditions,—and the embryo, instead
of remaining about five months in the egg, had the period
prolonged to thirteen months, as at present. The embryonic
dentition, together with the two dentitions now represented
by the alternating series, were thus developed during the
intra-capsular period. The first set of teeth became use-
less, and were shed before hatching. ‘The second set, con-
sisting of much larger teeth, were retained, owing probably
to the disadvantage of shedding or absorbing them within the
egg, as would be necessary if the third set were to take their
places. ‘These latter, being thus unable to take their proper
positions, were forced to occupy the intervening spaces. ‘The
animal therefore left the egg with a set of teeth
(called in this paper the alternating series) repre-
senting the second and third dentitions.! The pos-
terior members of both dentitions, however, suffered a
retardation in development, especially in the maxilla, so that
several of them do not now appear till some time after hatch-

1'The number of teeth of the first dentition, as compared with those of
the second and third (as seen in the alternating series), is shown jn the

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 191

ing. The crowding together of these two dentitions on a
small length of jaw, caused in the mandible that early inti-
mate fusion with each other which we observed at Stage 8.
Probably, also, this process was associated with an upgrowth
of bony tissue round the bases of the teeth such as we find
at the present time in many of them. All this would render
the process of tooth change more expensive, and it has
therefore become much reduced, confined to the anterior
region of the jaws, and restricted to the period of imma-
turity, when the amount of bone absorption is much less
than it would be in the adult. ‘The reason for the retention
of an active tooth change in the anterior regions of the jaws
is perhaps due to the necessity for an efficient dental arma-
ture in the front of the mouth.

The following diagrams will help to explain my views of
the relations of the tooth change in the premaxillary region
to the normal types of succession. Incidentally they suggest
a possible explanation for that obliquity of position notice-
able in the youngest mandibular and maxillary successional
teeth, where these are seen to lie at the base of a small tooth
(second dentition), and point towards a large one (third den-
tition), both of which they probably replace. The explana-
tion suggested is, of course, that the teeth of the third
dentition being forced to come up between two of their pre-
decessors, gradually came to develop with their apices
pointing in this direction, and that this change also affected
the teeth of the fourth dentition. It must be understood that
the condition found in the first diagram is never realised, the
second set at the present time always alternating with the
first, from the earliest stages. It will be seen that, accord-
ing to my view, the premaxilla had primitively only one

following table. Only the marginal teeth of one side are given, the pre-
maxillary and anterior mandibular being included.

Upper jaw. Lower jaw
Stage Q (enamel organs and teeth of first dentition) 9 : 8
Alternating teeth (second and third dentitions) , W/e=) 16

192 H. SPENCER HARRISON.

tooth,’ with a constant series of successors, the outer one in
b and ¢ being really a maxillary tooth. The dotted tooth of
the fifth dentition has not been found in any of my specimens,
but I believe it will be found to occur in slightly older
individuals.®

We are now called upon to account for the presence of
the posterior uniform series, and for the lack of synchronism
in its first appearance in the different dentigerous bones.

That the uniform series must be regarded as a later ac-
quirement than the other teeth, I consider has been suf-
ficiently proved by the facts adduced as to the occurrence of
a cessation of activity when the alternating series have

Diagrams to illustrate the dentitions represented in the premaxillary teeth up
to the stage of my largest specimen (compare with figs. 20, 21, and 22).
Dentitions :—I, black; II, horizontal lines; IIT, perpendicular lines;
IV, dotted; V, blank. J/z, maxilla. Pmz, premaxilla. J/, middle line
of jaw. a. Purely theoretical, showing all the successional teeth at one
time. 6, Modified from figs. 20, 21, and 22 of Stage T, and a little
later, ¢. As seen in my largest specimen, with hypothetical tooth added
(dotted outline).

been completed. What was the cause of the development of
this new set of teeth can only be conjectured. It may have
been due to a general increase in size of the animal, or to the

' In this connection it is interesting to note that in Zittel’s ‘ Paleontology ’
(23) Homceosaurus of the Jurassic, which is placed next to Hatteria in
the family Sphenodontide, is stated to have the “ Zwischenkiefer paarig

am Vorderrand mit einem einfachen sclneidenden Zahn besetzt.”
* See Addendum,

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 193

modifications caused by the lengthening of the incubation
period.

The reason for the early appearance of the uniform teeth
in the mandible as compared with the maxilla and palatine
‘is undoubtedly to be found in the concrescence described in
Stage S; this fusion causes the whole of the mandibular
alternating series (not including the first three on each side)
to assume the character of a single multicuspid tooth, thus
preventing any adaptation to the growth in length of the
jaw. As will be seen from the comparative measurements
given below, the region in question does not increase in
length from Stage T onwards. In the adult it is represented
by a short apparently toothless ridge immediately posterior
to the large anterior tooth. When this fusion occurred
during the phylogeny of Hatteria, the formation of the
uniform teeth was hastened, and hence their early appearance
at the present time. I have not observed any concrescence
in the palatine teeth, but it probably occurs to some extent,
as here the uniform series begins its development much
earlier than in the maxilla, though later than in the mandible.

The following is a comparative table of the number of,
and space occupied by, the alternating teeth or their vertical
successors in maxilla and mandible (three anterior teeth not
included) at various stages. It is usually impossible to detect
the separate teeth of this series in the adult mandible, so that
the number inserted is only an approximation. ‘The other
figures are averages where possible. It will be found that
the number of teeth of this series is less in the adult than in
the young. This is partly due to the fact that successional teeth,
in some cases at least, replace two of their predecessors, and
partly to some of the teeth having become indistinguishable.
In the anterior mandibular region also the alternating series
is encroached on by the formation of the two compound
teeth. I have added to the table the number of uniform
teeth.

Comparison of dentition of young with that of
adult.—We are now in a position to consider the relation

VoL, 44, PART 2,—NEW SERIES, N

194. H. SPENCER HARRISON.

MAxXILLA. MANDIBLE.
STAGE. —
Alternating.| Uniform. ||Alternating.|Uniform.

S Number of teeth . . | 9—10 0 10o—11 |] 0
: “| On length of jawof . | 6mm. | — | 5° mm.}; —
i Number of teeth . . .| 10—11 0 12—138 0
: "(On length of jaw of . | 7mm. | — 7 mm. —
Specimen of f Number of teeth . . Pe 0 13 3
17°8cem. lLOnlength of jawof . 10° mm) — 6mm. | —
Specimen Leer of teeth . . || 14(16)t) 0 13 5
21°2cm. LOnlength of jawof . | 12mm.} — 6mm. | —
Adult Number of teeth . . | 11 5 7 12
; "On length of jawof . | 15 mm.}| — 5 mm. —

between the dentition of the last specimen described and that
of the adult.

The great majority of adult Sphenodons show two large
“cusps” to each incisor. In some cases there are in addi-
tion one or two minute projections, which have sometimes
been taken for cusps. I believe, however, that these are
merely portions of an overgrowth of bone that occurs round
the base of many of the adult teeth. It appears certain that
there are only two teeth concerned in the formation of each
adult premaxillary tooth, Knox (21), it is true, stated that
he found a specimen with three cones to each, but his de-
scription is very brief and not at all clear. I have shown
that in my largest young animal the third successional tooth
is on the point of replacing the first. All the three teeth of
the newly-hatched animal are lost. In the absence of other
young specimens I prefer to reserve judgment as to whether
the two main teeth of my specimen (i.e. the second and third
successional, see figs. 21, 22) are the two components of the
adult tooth. On the whole, I think it is very probable that
there may be a constant succession of teeth in this region

1 Tf the two successional teeth have each replaced two of the alternating

series, 16 would be the correct number. All the figures given are approxi-
mate, as sometimes the two sides differ slightly. (See also Addendum.)

DEVELOPMENT OF TEETH IN HATTERIA PUNOTATA. 195

until the jaws have practically stopped growing. However,
most of the teeth of Hatteria retain their odontoblasts in an
active condition during the greater part of the life of the
animal, and it is possible that no further tooth change takes
place.! Burckhardt’s opinion that the form of the
premaxillary teeth is due to concrescence is quite
erroneous, if by this term he means the fusion of dentine
with dentine before the tooth is fully formed ; still more so
-¢ he intends to imply the fusion of one enamel organ with
another to produce a compound tooth, as has been described
in the Chameleon. The anterior teeth of Hatteria at the last
stage I have investigated are successional teeth developed
quite independently of one another, and their fusion is pro-
duced by the formation of bone round their bases, which
even passes some little distance down their sides. The
premaxillary bones also grow down, so as to cause their
apical tooth-bearing edge to project considerably beyond the
eum.

In the maxille of a full-grown specimen now before me
there are anteriorly on one side three, on the other four, well-
defined teeth. ‘Ihe insertion of these on the jaw is not at the
same level as that of the teeth posterior to them. The appear-
ance suggests that a small strip of the tooth-bearing edge has
been removed, and the teeth inserted along the exposed
surface. This is in effect what has happened. As may be
seen from several of my figures, when a tooth is shed a large
absorption of bone takes place, and the new tooth is inserted
at a different level from the old one. The assumption that
each of these anterior teeth is a successional one seems quite
justifiable, im view of the fact that I have myself found
four successional teeth in various stages of development in
this region. Assuming, also, that each of these teeth has
replaced two others of the alternating series, the number of
teeth now left will agree approximately with the original
number (say sixteen) of the latter in the maxilla. Posterior
to these anterior teeth there is, as T have intimated, a distinct

1 See Addendum.

196 H. SPENCER HARRISON.

step in the dentigerous edge of the bone, and on this higher
level are the much worn members of the original alternating
series. The first three on the one side, and two on the
other, judging by the space occupied by a smooth ridge, are
worn down to the bone, and posterior to this are four teeth
in each case alternating in size. Allowing each successional
tooth to equal two of the original set, and including the
teeth that are obviously worn away, the totals are thirteen
and fourteen respectively. ‘The relationships are not always
so simple as this, but I believe it will be found that tooth
change is confined to the anterior third of the jaw, the first
four maxillary teeth being usually successional. This skull
has on the two maxille respectively four and five teeth of
the uniform series. My remarks on the subject of the con-
tinuance of tooth change in the premaxille during further
growth will also apply to the maxilla. It is possible, though
not probable, that successional teeth are formed further back
than I have described, but at my latest stage there is no
independent residual dental lamina in the middle region of
the maxilla. This argument, however, in the case of Hatteria
has little weight in view of what we have seen in other
regions, where the dental epidermal ingrowth may remain
latent, as it were, in the face of the internal groove which is
attached to the teeth.

The right palatine of the same adult specimen has a large
anterior tooth, showing the same indications of its succes-
sional nature as do the first three or four maxillary. The left
palatine, on the other hand, terminates in a slightly sinuate
ridge. ‘The successional tooth of the right side no doubt
corresponds with the one shown in fig. 28, and has replaced
Nos. 1 and 2 in order. In another adult skull I find that the
large anterior palatine tooth is present on both sides. It is an
interesting example of the greater variability of the palatine
dentition, that whereas in the first of the adult skulls I have
described the proportion of uniform teeth in maxilla and
palatine respectively is 5: 7, in the other it is 6:5. I be-
heve that the tooth-change in the palatine will be found to

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 197

be restricted to the replacement of Nos. 1 and 2 by a succes-
sional tooth, and that even this may not always occur.
Passing on to the teeth of the lower jaw, we find that the
anterior region, which in the adult is occupied by a large
tooth on either side, has in the newly-hatched animal three
separate teeth in the corresponding positions. We have seen,
also, that the posterior two of these are replaced by a succes-
sional tooth, and that there are signs of the deciduous nature
of the first. My material does not take me far enough to
decide whether any of the teeth of the last young individual
described take part in the formation of the adult tooth. The
presence of a well-developed residual dental lamina in this
region at this stage proves that further change may occur.
Concerning the second successional, it is quite possible that
it may be a component of the adult compound tooth, which
I believe is formed partly at the expense of the anterior teeth
of the alternating series, their vertical successors taking part
in its constitution. In any case I believe the constituent
teeth to be three in number on each side, though in the adult
this is rarely evident.1. Newman (22) appears to have had a
specimen in which the middle cusp was especially distinct.
The “fusion” is produced in the same manner as in the case
of the large incisors. In the adult jaw the large anterior
tooth is succeeded by a short ridge, which is the representa-
tive of the posterior portion of the alternating series. The
teeth are practically worn down to the bone, which is there-
fore, as in some parts of the upper jaw, the functional equi-
valent of the teeth. From the attachment of the large
anterior teeth to the jaw there is a marked step to the level
of this ridge. I have observed that in the falling out of
deciduous teeth in the lower jaw, the amount. of bone that is
separated along with the tooth by absorption is very great.
The alternating series of the lower jaw very soon cease
growth, and are hence worn away at an early stage, the
bone here growing very rapidly and becoming dense and
solid. ‘This cessation of growth is no doubt due to their
1 Two. See Addendum.

198 H. SPENCER HARRISON.

early concrescence, which by preventing separation at the
base materially interferes with any increase in size. There
are in all probability no successional teeth represented in
this portion of the jaw. The posterior uniform teeth we
have already considered, and a reference to the table on
page 194 will show the number present at various stages. I
believe that these teeth have never possessed a successional
set, and they have no vertical predecessors.

Leaving out of account the uniform series, we are justified
in stating that there are at least three dentitions
represented in the functional teeth of the adult,
viz. a second and third in the persisting members of the
alternating series, and a fourth or fifth, or even some later
dentition, in the anterior teeth of both jaws. Perhaps
members of the later dentitions exist side by side in the
adult (as they do in the premaxilla of the young), and if so,
there may be four or more dentitions represented in the
adult.

The vomerine teeth are undoubtedly in course of sup-
pression, as suggested by Howes (8). ‘This is evident from
the infrequency of their occurrence (especially as paired
structures), and their small size in the adult when present:
this view is supported by the evidently degenerate condition
of one of the teeth in the specimen I have described. When
they occur they seem to begin their development about the
end of the incubation period, and I have found no indications
previous to Stage 8. It is possible that they represent the
surviving members of two longitudinal series, as in two cases
one vomer has been found to possess two teeth, a smaller
one anterior to the larger; there are also traces of an anterior
prolongation of the dental lamina on both sides in one of my
specimens.

Contrary to the usually accepted statement, the teeth
of Hatteria have a well-developed outer coat of
enamel. I have seen this in its earliest stages of develop-
ment, and I give a figure of a portion of the fully formed
tooth, showing the dentinal tubules branching near their

DEVELOPMENT OF TKETH IN HATTERIA PUNCTATA. 199

terminations, and being in many cases continued into tlie
enamel as far as its outer limit (fig. 28). This continuation
of the tubules into the enamel has been described in many
vertebrates. Ifa young tooth of Hatteria is treated with 3
per cent. HNO, in alcohol, the enamel is in the course of an
hour or two completely dissolved, and the dentine is then seen
to be surrounded by a fringe of short processes, representing
the ends of the dentinal tubules, perhaps still containing the
odontoblast processes. As will be seen in fig. 28, the enamel
is characterised by the presence of alternate light and dark
lines, running approximately parallel to the long axis of the
tooth. Ihave not isolated the enamel prisms, but from the
course of the dentinal tubules, from the marked tendency to
fracture at right angles to the surface, and from the ap-
pearances in early stages of development, I believe these are
at right angles to the surface of the tooth, if, indeed, they
exist at all as elements possible of isolation.

The first teeth to fuse with the bone show the typically
acrodont condition, as seen in fig. 14 of a mandibular tooth
of Stage 8. At later stages, however, secondary over-
growths of bone occur round the bases of many of the teeth,
which thus come to be firmly fixed in a socket with which
they are fused at all points externally, as well as at their base.
Howes and Swinnerton have used the term ‘‘ hyperacrodont”’
for this condition, and it is a very convenient one, inasmuch
as we are dealing with a type of attachment which is un-
doubtedly derived from the acrodont ; it would be inaccurate
to use here the term “ thecodont,” since although the base of
the tooth is in a socket, the origin of the socket and its rela-
tions to the tooth do not indicate a morphological identity
with the thecodont type. The bony upgrowth is of a rather
peculiar structure. It consists of a number of thin, closely-
applied lamellee, which in transverse sections of the decalci-
fied jaw give the appearance of a number of fibres converging
on the base of the teeth. The extent to which the upgrowth
actually surrounds the latter varies in different regions. It
may- be conspicuously seen embracing a considerable portion

200 H. SPENCER HARRISON.

of the tooth in the palatines, for instance. This bony over-
growth assists in the “fusion” of the anterior teeth of both
upper and lower jaws, and its outer layers are similar in struc-
ture to the specialised bone described in the next paragraph.

A conspicuous feature of a prepared skull of Hatteria is
the polished appearance of the edges of the jaws for a con-
siderable distance above or below the insertion of the teeth.
Tomes found this outer layer to be true bone, but gives no
details of his preparations. Ina ground section of the lower
jaw (transverse) I found the superficial layer to present the
appearance shown in fig. 29. The similarity to enamel is
striking, and this is in some places increased by its tendency
to fracture at right angles to the surface. I have found,
however, that weak acids do not dissolve it, and in some cases
I have been able to trace it into continuity with bone lamella
containing bone corpuscles. These are wanting in the fully-
formed material. In addition to this, it is not so highly
refractive as enamel, though more so than bone, and it seems
to be more affected by weak acids than does the latter. I
therefore conclude that it is a specialised layer of bone,
owing its similarity in appearance (in ground sections) to
enamel to a similar method of deposition by osteoblasts. It
apparently contains more calcareous matter than true bone,
and by its hardness enables the edges of the jaws to
functionally replace the teeth when these are worn away.
As this material is formed by the mesoderm (the process
commences at Stage 'l'), the overlying epithelium moves
further from the teeth, until its line of attachment to the
jaws is separated from the bases of the teeth by a very con-
siderable interval. The intervening bony surface is covered
with this outer specialised layer. It is of interest to note
that in this retreat of the epithelium from the bases of the
teeth, the potential dental Jaminz are removed to a great
distance from the latter, so that the succession of teeth in
the adult could only be effected at the cost of an immense
absorption of bone, which is here very dense, particularly in
the mandible. Whether this has been a factor in the partial

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 201

suppression of tooth replacement, or whether it has only
been concomitant with it, I do not venture to decide.

Summary and Conclusions.

1. The first dentition of Hatteria consists of a
set of minute teeth, about thirty-six in number,
which originate immediately below the epidermis,
labial to the dental lamina. They subsequently pass
into the epidermis, and are shed about the time of hatching.
The period of their best development is about the fifth month
of incubation. ‘They are never functional, and form no con-
nection with the bone. The anterior end of each palatine
dentigerous region is first indicated by the development of
one of these teeth, formed before the dental Jamina. Amongst
other reptiles Crocodilus porosus and Iguana tuber-
culata possess a somewhat similar, though less extensive,
functionless dentition, the members of which closely resemble
the first teeth of some Selachians as regards place of origin.

9. he dental lamina arises as in other Vertebrates, and
its position is at first indicated externally by a dental furrow.

3. At an early stage (R) the dental laminz and the enamel
organs (of the future functional teeth of the newly-hatched
animal) in all the main tooth-bearing regions, become in-
volved in epidermal ingrowths, which much obscure the
primary relationships of the parts. In each case there is a
lingual and a labial ingrowth, and the former carries with it
the dental lamina, which is thus for a long period indis-
tinguishable as an independent structure. Towards the end
of incubation the lingual ingrowth is mainly converted into
a glandular groove; the side attached to the base of the
tooth, however, remains in the condition of a stratified epithe-
lium, and may project as an epithelial ridge beyond the
deepest part of the groove ; it is from the cells of this portion
that the enamel organs of the successional teeth are deve-
loped in the young animal. The labial ingrowths (labio-
dental strands) of the premaxille, maxille, and mandible,

202 H. SPENCER HARRISON,

form the labio-dental grooves, that of the palatine an external
groove bounded labially by a prominent ridge (crista medi-
alis) between maxilla and palatine.

4, The teeth which are developed on the dental
lamina during the incubation period, and which
function during the early life of the young animal,
are almost certainly the members of two distinct
dentitions (the second and third), the later teeth
instead of displacing the earlier coming to alter-
nate with them. The chief arguments in favour of this
view are—(a) In the upper jaw, at Stage R, every alternate
tooth is small, is further advanced in development, and has
a lingual prolongation of the dental lamina; these belong to
the second dentition. The other teeth, on the contrary, have
larger enamel organs, are in an earlier stage of development,
and have no lingual prolongation of the dental lamina (third
dentition). In the lower jaw it is chiefly the alternation in
size which indicates the two dentitions. (b) ‘he number of
teeth in the alternating series is eventually approximately
double that of the first set. (c) The first successional teeth
of the young animal, in several cases, each replace a large
and a small tooth of the alternating series. That is, a member
of the fourth dentition displaces not only a member of the
third, but also one of the second. On the other hand, as
seen in the premaxilla, a member of the fifth has only one
predecessor of the fourth.

5. Successional teeth make their first appearance in the
premaxille some months after hatching. In this region
there are in all five distinct sets of teeth repre-
sented during development toa length of 21:2cm,,
commencing with the first or embryonic series. Hach pre-
maxilla originally bore one tooth only, with a constant suc-
cession ; the presence of three in the newly-hatched animal
is due to the second and third dentitions functioning at the
same period, the premaxilla then having also a tooth of the
third dentition belonging to the maxilla. ‘The number is
reduced to two by the displacement of the premaxillary teeth

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 208

of the second and third dentitions by one of the fourth
dentition. In my oldest specimen the fourth is the
functional dentition in the premaxilla, though there is one
of the fifth on the point of displacing its predecessor of the
fourth. It is quite possible that more successional teeth
will be found in this region on examination of later stages.'
Kach adult “incisor” tooth has two components,
but the union is not due to concrescence, being caused
by intimate fusion with the bone, which embraces the bases
of the teeth and also grows down beyond the gums, thus
increasing the apparent size of the compound tooth. Hatteria
does not foreshadow the heterodont condition of higher
types, but is truly homodont.

6. In the maxille the anterior alternating teeth are sub-
ject to replacement by members of the fourth dentition, each
of these replacing, in some cases at least, two of the alter-
nating series. I have found one successional tooth which I
believe to represent the fifth dentition in the maxilla. ‘The
first four teeth of the adult maxilla are usually successional
(i.e. belong to the fourth or a later dentition).

7. The only successional tooth I have found in the pala-
tines was situated lingually to No. 2 (second dentition). It
replaces in all probability both Nos. 1 and 2, and may be seen
in many adult specimens as a large tooth at the commence-
ment of the palatine series, if, indeed, this does not represent
a still later dentition, e. g. fifth or sixth.

8. The vomerine teeth are in course of suppression. They
may be paired or unpaired, and one may show signs of
degeneration in structure. Insome cases two teeth have been
found on one side. In most specimens they are completely
absent, probably during the whole of development. The
dental lamina is not a conspicuous structure, but appears to
have a short prolongation anterior to the teeth. We are
perhaps here dealing with the surviving members of two
longitudinal series.

9. The anterior portion of the mandible at one stage

' See Addendum.

204 H. SPENCER HARRISON.

possesses successional teeth belonging to the fourth dentition,
replacing, in one case at least, two of the alternating teeth.
The number of components of the large front teeth is pro-
bably three in each case,! the union being produced in the
same way as in the premaxillary. J am uncertain how many
or which dentitions are represented here in the adult, as I
believe further change takes place in larger young indi-
viduals than I have examined. In the adult there is
posterior to the large front teeth a ridge of bone frequently
to all appearance edentulous. This represents the posterior
teeth of the alternating series, which, owing to concrescence
at Stage S, are unable to adapt themselves to the increasing
length of jaw, and lose their power of growth, becoming,
therefore, worn down at an early stage. There are probably
no teeth of the fourth or later dentitions represented here,
those which are formed anteriorly in the young animal taking
part in the compound anterior teeth.

10. In the main dentigerous regions there is a
more or less prolonged cessation of tooth develop-
ment when the alternating series is complete.
Subsequently a renewed formation of teeth takes
place from behind forwards, and these are uniform
in size. In the lower jaw this halt is of short duration, and ~
the uniform teeth make their appearance long before those
of the palatine and maxilla. In fact, there are already five
uniform teeth in the mandible, when in the maxilla this
series 1s not even represented by enamel organs, and in the
palatines only by the first stages of two dental germs.
This set of teeth is no doubt a more recent acquisition, and
has, I believe, hke the permanent molars of mammals, no
vertical predecessors or successors. The want of synchron-
ism in its appearance in the different regions is due to
the fact that the concrescence of the alternating series in
the lower jaw prevents the teeth from adapting them-
selves to the growth in length, as do those of the maxilla,
and of the palatine (to a less extent). During the period

' More probably two. See Addendum.

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 205

of growth of the animal there is a constant suc-
cession of these teeth from behind forwards, the
series having once commenced.

11. All the teeth except those of the first dentition have
an outer coating of enamel, which, however, does not extend
as far as the base of the fully formed teeth. It contains
prolongations of the dentinal tubules. The odonto-
blasts of many of the teeth remain in activity for a con-
siderable period, probably at least till the animal is full-
grown.

12. The teeth of the very young animal are typically
acrodont in their attachment. Later, through a secondary
formation of bone round their bases, many of them come to
lie in shallow alveoli, apparently similar to those described
in some fossil Rhynchocephalians, such as Proterosaurus and
Champsosaurus. Since the teeth are firmly fused with the
base and sides of the alveoli, we are not justified in calling
them thecodont, and the condition is best described by the
use of the term ‘‘ hyperacrodont,” suggested by Howes and
Swinnerton.

The sides of the jaws are covered with a specialised layer
of bone, apparently more highly calcified than the rest. In
the adult a large part of the jaw is exposed, and when the
teeth are worn down this outer layer assists the exposed
portion to functionally replace the teeth. In some of its
macroscopic and microscopic features it resembles enamel.

13. The evidence of the tooth development and
succession suggests that the long incubation
period of Hatteria (thirteen months) is a compara-
tively recently acquired character.’ This view affords
us an explanation of the possession, after about five months
of incubation, of a complete (though degenerate) dentition,

1 This view is also supported by several of Dendy’s discoveries. The
presence of a characteristic pattern on the skin, after about five months’
incubation, and its almost complete disappearance before hatching, is very
suggestive. So also is the very early development of the epidermal shell-

breaker on the snout.

206 H. SPENCER HARRISON.

- which is shed before hatching, as well as of the fact that the
succeeding set is approximately double the number of the
first, and therefore (as is strongly supported by many other
arguments) represents two dentitions, which in consequence
of their development within the egg have come to form an
alternating series. The crowding together of these on too
short a length of jaw explains the concrescence which takes
place in the mandible: this, again, is the cause of the early
appearance of the uniform teeth in this region and their
greater number in the adult. It is possible also that the
origin in phylogeny of the uniform series, and the other
irregularities of tooth succession, may be traced to this
lengthening of the incubation period.

ADDENDUM.

Some time after the foregoing paper was sent to press, I
was enabled, through the kindness of my chief, Professor
W.N. Parker, to examine the prepared skull of a young
Hatteria, whose total length was 246 cm. The skull
measures 3°5 cm. from premaxilla to occipital condyle. The
dentition forms an interesting intermediate stage between
the last described and that of the young adult, and a study
of it has caused me to slightly modify some of my previous
opinions.

The character and disposition of the teeth is shown in the
following table, one side only being given. ‘The smaller
teeth of the alternating series are so worn, even at this
stage, as to be difficult of identification.

Successional. Alternating. Uniform.
Premaxilla . 2 5 — ats ==
Maxilla . ; 3 we PO==12 0 ~..: —
Palatine . : 1 ae 4A—5 aM 2
Mandibular. 2 es < nee 8

In the premaxilla the outer tooth has obviously only

DEVELOPMENT OF TEETH IN HATTERTIA PUNCTATA. 207

recently cut the gum, whereas che inner is older and more
worn. The latter is no doubt the third successional, shown
still below the gum in my fig. 22, whilst the former is a
fourth successional, belonging, like the third, to the fifth
dentition. It is indeed the tooth represented as
hypothetical in my diagram c. The bone has not yet
grown down with the teeth to any extent.

Of the three successional teeth of the maxilla, the
anterior is the least worn and the middle one the most. The
posterior is the third successional shown deeply seated in
fig. 25 (t. s. 3), the middle one is the second successional
(t. s. 2), whilst the anterior is a fourth successional (which I
have mentioned but not figured), and represents the fifth
dentition, having replaced the first successional (fourth den-
tition). Posterior to these three anterior teeth are from ten
to twelve of the alternating series, but there are as yet no
indications of the uniform series. The number of teeth in
the upper jaw belonging to the second and third dentitions is
slightly greater than I had supposed from my earlier speci-
mens.

Taking the maximum of twelve as the number remaining
at this stage, and adding two for each of the successional
(maxillary and premaxillary), we arrive at twenty-two as the
number of alternating teeth represented on each side of the
upper jaw. This would give us eleven each of the second
and third dentitions, whereas the first has only nine in the
corresponding region. ‘I'he discrepancy does not, however,
constitute any serious objection to the view that the ances-
tors of Hatteria at one period possessed a dentition consist-
ing of a definite number of teeth, about ten (marginal)! on
each side above and below, and that these were replaced
regularly by vertical successors, without increase in number.
As I have already said, the uniform series is almost certainly
a later acquisition.

1 The number of palatine teeth was probably 3—4 on each side, though
the evidence here is not so satisfactory as in the case of the marginal teeth.
The number of vomerine teeth at this period is quite uncertain.

208 H. SPENCER HARRISON.

The anterior palatine tooth is clearly the successional
tooth I mentioned in describing my last specimen, and judg-
ing from the space it occupies, has replaced at least two of
the alternating series. Of the latter there are four to five,
posterior to which are two of the uniform series, these being
no doubt the representatives of the two enamel organs
described in my 21:2 cm. specimen.

The left vomer has a small though well-developed tooth.

In the mandible the anterior region is beginning to
assume the appearance characteristic of the young adult,
that is to say, the teeth and bone are becoming related to
one another in such a way as to foreshadow the two large
anterior “teeth.” At this stage the components are clearly
two in each case, the anterior being the first successional
(fig. 26, t. s. 1), while the posterior is the second successional
(t.s.2). The latter is a very large tooth, and, from comparison
of this stage with adult individuals, I have come to the con-
clusion that it eventually forms practically the only actual
dentinal portion of the adult compound tooth, the remainder
consisting of bone. In the body of this paper I have stated
that I consider these front teeth to have each three compo-
nents. [I am now of opinion that there are only two
concerned, and that the anterior of these is worn away at
a stage very little beyond that we are now considering. ‘The
most anterior tooth of fig. 26 is not to be found in the present
specimen, and has probably been shed.

Posterior to the large successional tooth are the remaining
members of the alternating series, now only seven in number.
Since there were originally sixteen alternating teeth on each
side, it is evident that the space formerly occupied by nine
teeth is now taken up by two only. From this comparison,
and from the space in the jaw occupied by the second suc-
cessional, the latter has obviously displaced six or seven of
the alternating series. This would at first sight appear to be
very contradictory to my view that a tooth of the fourth
dentition usually replaces two teeth, one each of the second
and third. ‘The conspicuous irregularity in this case is, how-

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 209

ever, probably due to two causes: (1) the fusion of the alter-
nating teeth, which has caused them to occupy less than
their proper space on the jaw; and (2) the large size of the
successional tooth in question, due to its importance as the
main constituent of the anterior compound tooth.

Posteriorly are eight fully formed teeth of the uniform
series, and one in course of development below the gum.

From a comparison of numerous adults with the specimen
just described, I believe that very little further tooth change
takes place. In most adults, however, the first four maxil-
lary teeth are clearly successional. The fifth is the latest
dentition I have been able to trace in any of my specimens,
and this is represented in the premaxille and maxille of the
present one.

LITERATURE,

GintHEeR.—“‘ On the Anatomy of Hatteria,” ‘Phil. Trans.,’ 1867, pt. ii.
. G. Baur.—‘ Osteologische Notizen,” ‘Zoolog. Anzeig.,’ 1886, p. 685.

. G. B. Howrs.—“ Remarks on Hatteria,”’ ‘Proce. Zool. Soc.,’ 1890.
. Srepenrocx.—“ A Contribution to the Osteology of the Head of Hat-
teria,’ ‘Ann. and Mag. of Nat. Hist.,’ 1894 (translated from ‘ Sitzber.
der k. Akad. der Wissensch. in Wien,’ 1893),
5. C. Roése.— Ueber die Zahneutwickelung von Chamaeleon,” ‘ Anat.
Anz.,’ Bd. viii, 1893.
6. Burckyuarpt.—‘ Das Gebiss der Sauropsiden,”’ ‘ Morph. Arbeit.,’ v,
1896.
7. G. Baur.—“ Das Gebiss von Sphenodon (Hatteria), ete.,” ‘ Anat. Anz.,’
vol. x1, 1895, p. 436.
8. Osawa. —“ Beitrage zur Lehre von den Hingeweiden der Hatteria
punctata,” ‘ Archiv fiir mikros. Anat.,’ Bd. xlix, 1897.

0 De

9. A. Denpy.—“ Outlines of the Development of the Tuatera, Sphenodon
(Hatteria) punctatus,” ‘Quart. Journ. Micros. Sci.,’ N.S., No.
165, 1899.

10. Howszs anp Swinnerton.—‘‘ On the Development of the Skeleton of
the Tuatera, Sphenodon (Hatteria) punctatus,” ‘Trans. Zool.
Soc.,’ 1901.

vou. 44, PART 2.—NHW SERIES. fe)

210 H. SPENCER HARRISON.

11. C. S. Tomzs.—‘ Manual of Dental Anatomy,’ 1894, 4th edition, p. 253.

12. C. Rosz.—‘‘ Ueber die Zahnentwickelung der Krokodile,” ‘ Morpholog.
Arbeit.,’ Bd. iii, 1894.

18. C. Roszt.—‘‘ Die Zahnentwickelung der Krokodile,” ‘Anat, Anz.,’ Bd.
vii, 1892.

14. Lecnue.—‘ Ueber die Zahnentwickelung von Iguana tuberculata,”
‘ Anat. Anz.,’ Bd. viii, 1898.

15. Laaser.—‘ Die Entwickelung der Zahnleiste bei den Selachiern,” ‘ Anat.
Anz.,’ Bd. xvii, Nos. 24, 25, 1900.

16. A. Cartssen.—‘‘ Ueber den Zahnersatz bei Agama colonorum,”
‘Anat. Anz.,’ Bd. xi, 1896.

17. C. S. Tomses.—* The Development of the Teeth of the Newt, Frog, -
Slowworm, and Green Lizard,” ‘ Phil. Trans.,’ part i, 1875.

18. C. S. Tomes.—“ On the Structure and Development of the Teeth of
Ophidia,” ibid.

19. Santa Srrena.— Ueber den Bau und die Entwickelung der Zahne bei
den Amphibien und Reptilien,” ‘ Verhand. d. Phys.-med. Gesellsch. in
Wiirzburg,’ Bd. ii, 1872.

20. Bronn’s ‘Klassen u. Ordnungen des Thierreichs,’ Bd. vi, Abth. iii
(Reptilien).

21. Kwox.—‘‘ On the Tuatara,”’ ‘ Trans. New Zealand Inst.,’ vol. ii, 1869.

22, Newman.—“ Notes on the Physiology and Anatomy of the Tuatara,”
‘Trans. New Zealand Inst.,’ vol. x, 1877.

23. ZirtEL.—‘ Handbuch der Palaeontologie,’ vol. iii, Vertebrata, 1887,
1890.

DESCRIPTION OF PLATES 10—12,

Illustrating Mr. H. Spencer Harrison’s paper on “The De-
velopment and Succession of Teeth in Hatteria
punctata.”

List oF REFERENCE LETTERS.

Ant. Anterior. ad. Adamautoblasts. 6.v. Blood-vessel. cav. Artificial
space. c.g. Crista gingivalis inferior. d. Dentine. dex, Dentary. d./.
Dental lamina. ex. Enamel. e. 0.1. Enamel organ of tooth of first (embry-
onic) series. ep. Epithelium of buccal cavity. gl. Gland. Lad. Labial side
of section. Liz. Lingual side of section. /.d.gr. Labio-dentai groove.

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 211

lin. gr. Mandibular (internal) dental fossa. man. Mandible. maz. lab. Epi-
thelial ingrowth labial to mandibular teeth (labio-dental strand). maz. lin.
Epithelial ingrowth lingual to mandibular teeth. maz.f. Maxillary tooth of
functional series. mes. Mesoderm. ms. Meckel’s cartilage. m.lab. Epi-
thelial ingrowth labial to maxillary teeth (labio-dental strand). m. din. Epi-
thelial ingrowth lingual to maxillary tooth. m.y. Mesodermal dental papilla.
p. lab. Epithelial ingrowth labial to palatine teeth. p./iz. Epithelial in-
growth lingual to palatine teeth. pmaz. Premaxilla. post. Posterior. 7. d./.

Residual dental lamina. s¢. Stellate tissue of enamel organ. ¢.s. Successional
tooth.

(Outlines of all sections drawn with camera lucida.)

Fic. 1.—Section of the enamel organ of a tooth of the first series (first,
left, upper), with dental lamina on the lingual side. d.f. Dental furrow.
(Embryo Q, 52a. Sagittal section. x 240.)

Fie. 2.—Section of the second tooth of the first series, which is already
well calcified. (From same series of sections as fig. 1. x 240.)

Fic. 3.—Section through the enamel organ of the first premaxillary tooth
(right), belonging to the future functional series. (Embryo R, 142. Sagittal
section. x 240.)

Fie. 4.—Section through enamel organ of the right anterior palatine tooth,
and the sixth maxillary, of the functional series. The small palatine tooth of
the first series is seen lying lingually to the palatine, enclosed in epidermis.
The epidermal ingrowth labial to maxillary (labio-dental strand) is just
becoming evident. p.e.o. Enamel organ of palatine tooth of functional series.
pal.\. Palatine tooth of first series. (R, 162. Frontal section. x 60.)

Fie. 5.—Section through left maxillary, palatine, and mandibular dental
regions, immediately posterior to first palatine tooth, and through the seventh
mandibular. (From same series as fig. 4. x 60.)

Fic. 6.—Section through eighth maxillary tooth of right side. This is one
of the smaller teeth of the alternating series (cf. fig. 11), and is more
advanced than the seventh (cf. fig. 7). (R, 162. Frontal section. x 240.)

Fie. 7.—Section through seventh maxillary tooth of right side, for com-
parison with fig. 6. (R, 162. Frontal section. x 240.)

Fig. 8.—Section through the eighth mandibular (left). (R—S. Frontal
section. x 60.)

Fie. 9.—Section through right palatine and maxillary teeth to show the
relations of the epidermal ingrowths at this stage. pal. f. Palatine tooth of
functional series. max. Maxilla. (Embryo R—S. Frontal. x 60.)

Fic. 10.—View of premaxillary and anterior mandibular teeth at Stage S,
from a prepared skull. x 5.

Fic. 11.—Lateral view of teeth at Stage S, to show the conspicuous alter-

212. H. SPENCER HARRISON.

nation in the maxillary and the less marked one in the mandibular teeth.
Palatine teeth not shown. xX 5.

Fic. 12.—Section of a lower tooth of the first series at Stage S. 7.1. Tooth
of first series. ep.s. Capsule of flattened cells. (S.2. x 240.)

Fie. 13.—Section of palate passing through the enamel organ of the
vomerine tooth of left side. Vo. Vomer. V.e.o. Enamel organ. (S.
Sagital. x 240.)

Fic. 14.—Section of mandibular tooth, with bone of jaw. In this section
are also figured the labio-dental groove (/.d.gr.), formed from the corre-
sponding strand of cells by splitting ; also the mandibular dental fossa, formed
in the same way from the lingual ingrowth, cutting off a ridge of mucous
membrane (with a core of mesoderm), which runs between the inner face of
the jaw and teeth and the tongue (cf. also figs. 5, 8,19). The glands on the
sides of the grooves and on the lower lip (/./.), together with the shelf of the
latter, are shown. The enamel of the tooth has been removed in the process
of preparation. At the base of the internal dental fossa a solid ridge of
epithelial cells (7. d.¢.) remains non-glandular, and no doubt represents the
dental lamina, which has been carried down by the glandular ingrowth (ef.
figs. 19, 7.d.1., and 18, ¢.s.'). (Embryo 8, 19a. Frontal section. x 60.)

Fie. 15.—Section of junction of tooth and bone at Stage 8. On its
internal surface the dentine has a more recently formed layer (unshaded).
On the labial side the close attachment of the epithelium to the tooth and
bone is shown. 6.c¢. Bone-corpuscles. ep. gr. Epithelium of inner limb of
Jabio-dental groove. (S, 19a. Frontal section. x 240.)

Fie. 16.—Sagittal section through posterior part of right upper and lower
jaws at Stage 8, to show the posterior continuation of the dental lamine of
maxilla and dentary. Three large teeth have been cut through, and one small
one on the upper dental lamina. e. o. 7 Hnamel organ of tooth of alternating
series. (S2. Sagittal. x 60.)

Fic. 17.—Anterior view of premaxillary and anterior mandibular teeth at
Stage T, to show the relative reduction in size of the second premaxillary on
each side. x 5.

Fic. 18.—Sagittal section through the second premaxillary tooth of one
side, to show the developmeut of a successional tooth, which arises from the
epithelium near the deepest part of the internal dental fossa (diz. g7’.). The
glandular portion of the latter is not evident, and the labial side of the
section is damaged. ‘The functional tooth is seen to be attached to the bone
on the labial side only. pm. ¢. Second premaxillary tooth. ¢. s. 1. First
successional tooth. (Stage T. Sagittal section. x 60.)

Fie. 19.—Section through third mandibular tooth of one side, showing a
well-defined mass of epithelial cells at the base of the internal dental fossa.
The interpretation of this as a dental lamina is justified by the formation of

DEVELOPMENT OF TEETH IN HATTERIA PUNCTATA. 2138

successional teeth on its lingual side at a later stage (see also fig. 18).
(Stage T. x 60.)

Figs. 20, 21, 22.—Three stages in the development and succession of teeth
in the premaxilla (right) of specimens 17 em., 17°8 em., and 21:2 em. in
length. 1, 2,3 are the teeth of the newly-hatched animal; 7. s. 1, 2, and 3,
are the successional teeth in the order of their formation. (¢. s. 1 in fig. 20
was missing in the preparation figured, but was inserted from sections).
x 10.

Fic. 23.—Inner view of the anterior maxillary teeth of left side, from a
specimen 17°8 cm. in length. ¢.s. 1 is the first successional tooth of the
maxilla, which has cut the gum but is not yet fused with the bone. x 14.

Fic. 24.—Transverse section across the vomers (vo.), cutting the two
vomerine teeth (/.v. ¢., 7. v. ¢.) more or less longitudinally. The left tooth
had a considerable backward curve, which partly accounts for its division into
a free and an attached portion. From the same specimen as the last figure.
50.

Fie. 25.—Inner view of anterior maxillary teeth of left side, from speci-
men 212 cm. in length. ¢.s.1,2, and 3 are the successional teeth, the
last being still deeply seated. x 14.

Fic. 26.—Inner view of anterior mandibular teeth of left side, showing
two successional teeth. ¢. s. 1 no doubt replaced two of the earlier set (from
same specimen as last figure). x 14.

Fig. 27.—Outline of marginal teeth of 21:2 em. specimen. The uniform
series in the lower jaw already occupies considerable space, whereas all the
maxillary teeth are of the alternating series or their vertical successors (tlhe
same may be said of the palatines at this stage). x 5.

Fie. 28.—Portion of ground longitudinal section through the base of a
young tooth, to show the branching dentinal tubules (d. ¢.) with their pro-
longations into the enamel (d. @’.). x 300.

Fig. 29.—Portion of a ground section through the specialised outer layer
of bone, on the lingual side of the adult lower jaw, to show the close
resemblance to enamel. 4. c. Bone-corpuscles. x 300.

»
ial,
_ [=>

el (iiey 7 one te *

- =) a, -
pte ~

>

THE ANATOMY OF PLEUROTOMARIA BEYRICHII.

bo

15

The Anatomy of Pleurotomaria Beyrichii,
Hilg.

By

ry

Martin F. Woodward,
Demonstrator of Zoology, Royal College of Science, London.

With Plates 13—16.

THE vast antiquity which characterises the genus Pleuro-
tomaria—for no one can doubt the identity of the living and
fossil shells which are customarily grouped together under
this name—has justly endowed this molluse with great
interest for those studying the ancestry of the Prosobran-
chia. When, therefore, a living example was obtained by
Agassiz in 1871, and later in 1879 several specimens of both
P. Quoyana and P. Adansoniana were dredged by the
United States steamer “ Blake,” the result of an investiga-
tion of the anatomy of these forms was awaited with great
interest. Unfortunately, however, the specimens all turned
out to be in a bad state of preservation, and although falling
into such skilled hands as those of Dr. Dall, it was found
impossible to make out much of their anatomy. Dall, how-
ever, published ! figures and descriptions of the external cha-
racters of the animals, of the radulz and of some few points in
connection with the pallial complex, the rest of the body
being too much decomposed for investigation.

During the last few years a further examination of one of

1 “ Report on the Mollusca dredged by the United States steamer ‘ Blake,’ ”
‘Bull, Mus. Comp. Zool., Harvard,’ vol. xviii, 1889.

216 MARTIN F. WOODWARD.

the specimens of P. Quoyana obtained by the “ Blake ” has
been made by Fischer and Bouvier,! and these authors have
made a still more detailed examination of the radula of this
form, and also of the nervous system, which had not previ-
ously been examined. These investigators have published a
most exhaustive history of the genus, giving in addition a
complete list of the recent specimens obtained up to the year
1898, and they further append a full literature relating to
this mollusc. Since it is not my intention to enter into these
branches of the subject, I must refer the reader to Messrs.
Bouvier and Fischer’s paper.®

Through the kindness of the Director of the Natural
History Museum I had placed at my disposal an example of
the animal of P. Beyrichii obtained off Boshu, in Japan.
The animal was beautifully preserved, but unfortunately it
declined to part company with its shell save in pieces, so that
my first investigation was much retarded by having to build
up the anatomy of the animal from these fragments. Fortu-
nately, however, two more specimens were obtained by the
Museum and handed over to me by Professor EH. Ray Lan-
kester, whom I have to thank for entrusting their examina-
tion to me. Without these additional specimens my results
would have been very incomplete, since the first specimen,
though in much the best state of preservation, was in so
many fragments that it was extremely difficult if not impos-
sible to make out the exact relations of some of the organs.
Although my investigations are largely based on an examina-

1 Etude monographique des Pleurotomaries actuel,” ‘Archiv. Zool. éxp.
(3), vol. vi, 1898. Reprinted in ‘ Bull. Mus. Comp. Zool., Harvard,’ vol.
XXXil.

2 Since the publication of Bouvier and Fischer’s monograph at least five
new specimens of P. Beyrichii have been obtained. ‘These all came from
the Boshu, Japan, being captured alive in nets set at the bottom at a depth
of seventy to eighty fathoms; they were preserved in spirit with the animal.
One of these specimens has been described by Rolle (‘ Nachrbl. Deutsch.
Malak. Ges.,’ 1899) as a new species under the name of P. salmiana. I
think, myself, that is only a variety of P. Beyrichii. An additional shell of
P, Adansoniana has also been obtained. ;

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. yal AF

tion of the fragmentary specimen, they have in every case
been verified by a comparison with the two complete speci-
mens.

The External Characters.—The first example which
came into my hands was the identical specimen whose ex-
ternal characters were described by Professor Mitsukuri in
the ‘Annotat. Zool.,’ Japan, vol. i, p. 67; consequently I
cannot do better than quote his description in extenso.

“The animal was not very lively and could not be per-
suaded to extend itself fully. At the utmost we were able
to see the foot and a part of the head. The sole of the foot
was straw-yellow. The side of the foot and the throat were
mottled with large and small patches, and streaks of deep
carmine-red on the ground colour of reddish yellow. The
proboscis was uniformly deep carmine-red. The left tentacle
had a small branch near the tip. On the sides and the
posterior aspect of the foot we were able to make out two
lobes, one standing up from each side of the foot and applied
to the shell. It seemed probable to me that when fully
extended these lobes enveloped the shell to a greater extent,
a supposition which is strengthened, as was first pointed out
by Mr. Namiye, by the fact that the shells of Pleuroto-
maria hitherto found are all extremely clean, and have
never barnacles, worm-tubes, etc., attached to them. The
mantle was not at all visible, and we were thus not able to
see how it is related to the slit on the outer lip.”

It will be seen from the above account that Mitsukuri
makes no mention either of the presence or absence of an
operculum—a strange omission when we remember that an
operculum had been described by Dall as present in both
P. Quoyana and P. Adansoniana. When I received the
specimen I found that it had no operculum, nor could I
find, after a careful examination, any suggestion that the
operculum had been torn away. The only indication of the
possible presence of this organ was a minute lobe (fig. 2,
op. l.) situated on the dorsal side of the foot in the position
of, the opercular lobe of Trochus, ‘The arriyal of the

218 MARTIN F. WOODWARD.

second specimen, however, showed that P. Beyrichii, like
the two other species mentioned above, possessed a fairly
stout though somewhat small operculum attached to the foot
by a large circular lobe (figs. 3 and 4). We are, however,
still unable to determine whether the first specimen had lost
its operculum during its free life, or if it had been born
without one. Judging from the presence of the opercular
lobe, I should be inclined, in spite of its small size, to suggest
that the operculum had been present, but accidentally lost
either through disease, or mishap, early in life,

The operculum (fig. 4) is nearly circular in outline,
measuring, in the largest specimen, 14°5 mm. in diameter ; in
character it is trochiform, consisting of about twenty closely
coiled whorls, strongly marked with line of growth. It is
apparently composed solely of dark brown horny (chitinous)
matter, and for its size is very thick and strong, retaining its
thickness quite to the margin.

The mouth of the shell from which the operculum was taken
measured 40 mm. in transverse and 50 mm. in vertical dia-
meter. Hence it will be seen that the operculum can be of
very little use in closing the aperture, and thus protecting
the retracted animal; it may, however, be of some service in
protecting the upper surface of the foot from mechanical
injury which might be caused by the rubbing of the shell
when the animal was fully extended, since under these con-
ditions the shell rests, as in the Trochide and Turbinide,
directly upon the operculum.

Compared with the opercula of P. Quoyana and P. Adan-
soniana, the operculum of P. Beyrichii appears to most
nearly resemble that of the first-named species, although
Dall in his description does not mention what is such a
striking feature in P. Beyrichii, the thickness of the oper-
culum. In P. Adansoniana the operculum is very much
larger and thinner, and still more closely resembles the
opercula of the Trochide.

The small size of the operculum in two of the three speci-
mens, and its absence in the third, suggests that this organ

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 219

is of very little importance to P. Beyrichii, and that pos-
sibly it will disappear in the near future.

The Foot.—The foot, although contracted in my speci-
mens, is still very large, and is evidently capable of great
extension. As is the case with many Prosobranchs its ante-
rior margin is double (figs. 1, 2, and 7), the upper surface of
the foot being separated from the sole by a well-marked
transverse groove. We are quite at a loss to account for
this structure, which is evidently of great importance since
it is present in so many Gastropoda.

The lateral surfaces of the foot are finely rugose, being
closely beset with minute papille (figs. 1 and 2). These
papille are wanting on the dorsal surface, which is separated
from the lateral surfaces by the paired epipodial folds (ep.).
At the anterior extremity of the dorsal surface is situated
the opercular lobe (figs. 2 and 8, op.l.); this in its func-
tional condition is circular and nearly as large as the oper-
culum. On the right side it is produced out into a little lobe,
which is in turn attached to the upper surface of the foot,
and marks the growing point of the multispiral operculum.
Behind the opercular lobe a median longitudinal groove
leads to the posterior end of the foot; on either side of this
is a modified area due to the presence of numerous transverse
grooves originating from the median one; some of these are
symmetrically arranged, but others are unpaired (fig. 2).

This somewhat y-shaped modified area is bounded in front
by a couple of longitudinally-placed bands running back
from under the opercular lobe; these, however, only extend
for about one third of the length of this area, which is else-
where bounded by a groove marking the commencement of
the epipodium. A similar modified area was found by Dall
in P. Adansoniana, but strangely enough this appears to
be quite wanting in P. Quoyana, a point upon which Dall
lays some stress. This is a very curious fact, for in other
respects, notably in the operculum and in the radula, as we
shall see later, P. Beyrichii is more closely related to P.
Quoyana than to P. Adansoniana, a relationship which

220 MARTIN F. WOODWARD.

had already been noted by Crosse and Fischer from a study of
the shells, and expressed by the institution of the section
Perotrochus for the first two species.

This peculiar specialised area is also to be met with in the
Trochide (notably in T. [Gibbula] magus and T. [Callio-
stoma] zizyphinus); but though so commonly present, I
am unable to offer any suggestion as to its function.

The Epipodium.—This structure, which is so charac-
teristic of the majority of the Diotocardia, or of that sub-
division for which Fischer proposed the name Thysanopoda,
is not conspicuously developed in P. Beyrichii. It takes
the form of a couple of folds, one on either side of the body.
They start a short distance behind the head and attain their
maximum development in the region of the operculum ;
whence they extend back in the posterior extremity of the
foot, practically meeting in the middle line behind the median
dorsal groove. These folds, which are evidently somewhat
contracted in the spirit specimen, are like the rest of the
body covered with minute papille, and are entirely devoid of
those accessory lapets and tentacles so characteristic of the
epipodia of the Trochidz, Haliotide, and other Thysanopoda.
Judging by the figures given by Dall (op. cit., pl. xxx, figs.
1, 4, and 5) of the living animal of P. Adansoniana, the
epipodium would be more conspicuous in the living animal
in P. Quoyana; Fischer and Bouvier even speak of it as
being largely developed. In comparison with the Trochide
and Haliotidee, however, I should rather conclude that the
epipodium was feebly developed in Pleurotomaria,

I do not think there is any evidence to support the view
advanced by Mitsukuri that these lobes partly envelop
the shell, although they are apparently closely applied to its
base, and I would rather account for the clean nature of the
shell by the habitat of the animal being in deep water—
seventy to eighty fathoms,—where life, both animal and
vegetable, is not so abundant asin the littoral zone inhabited
by the Trochidz, whose shells are so generally encrusted
with foreign matter.

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. Bae!

The Head.—The head is large and produced into a
somewhat cylindrical snout, bent downwards as in the Tro-
chide. On the under surface, which is abruptly truncated,
is situated the mouth, surrounded by a horseshoe-shaped lip,
the gap in the lip being placed ventrally. The head and
anterior part of the body are much smoother than the foot,
being practically devoid of papille. The tentacles are situated
low down on either side of the head. In form they are cylin-
drical, with roughly pointed extremities ; they are highly mus-
cular and evidently much contracted. The only peculiar
feature about them is the tendency which they show to be-
come branched at their free ends; this is most marked in the
left tentacle of the first specimen examined (fig. 1), but the
second specimen also showed indications of a similar condi-
tion in both tentacles, although to a lesser degree, while the
right tentacle of the third specimen shows no less than two
accessory branches.

The Eyes.—The eyes, which are small and inconspicuous,
are situated, each on a slight elevation, at the posterior side of
the base of the tentacle. Examination with a lens shows the
cornea to be perforated, as in Trochus and most Dioto-
cardia.

An examination of sections (fig. 20), however, shows that
they are simple in constitution; hke those of Trochus the
central cavity is only partially filled by a vitreous body, the
rest of the space being occupied by sea-water.

The eyes, as may be supposed, are not specially well pre-
served, but one can see that the retina consists of a series of
pigmented rods, turned towards the optic cup, and an external
layer of ganglionic cells. I was unable to make out the clear
distal segments of the rods, such as are figured for other
Diotocardia. The retina is bounded by a delicate capsule,
outside which we see the optic nerve and a few mesodermal
pigment cells. The retinal pigments extend out through the
perforation in the optic cup into the adjacent epidermis.
he structure of the vitreous body suggests that is secreted
by the individual rod-cells.

Doe MARTIN F,. WOODWARD.

The Mantle.—In the contracted state of the dead animal
the mantle-slit, so characteristic of this genus, is inconspi-
cuous, and appears more like a broad, shallow sinus than a
deep narrow slit ; even in the living animal Mitsukuri was un-
able to observe its relation to the shell-slit. Its true relation
is, however, well seen in Dall’s figures of P. Quoyana and
P. Adansoniana taken from the living animal. In these
forms we see that the margins of the mantle-cleft are closely
applied to the margins of the shell-slit, through which they
may slightly protrude. ‘The free edge of the mantle is thick-
ened and closely beset with numerous small papillae, which
are evidently slightly protrusible, although not to the extent
seen in Haliotis. The mantle-fold completely encircles the
body, but is only feebly developed behind, and in this region
its margin is quite smooth.

The Pallial Complex.—Owing to the bad state of
preservation of the specimens collected by the “ Blake,”
Dall was unable to give us much information concerning the
organs falling under this category ; he was further unfortunate
in his attempts to identify these badly preserved parts, and
consequently, beyond a slight knowledge of the gills, we were
quite in the dark as to the relations of the kidneys and the
genitalia, since the structures to which Dall, and after him
Fischer and Bouvier, applied these names, have quite different
significances.

The Ctenidia.—The gills are very large and conspicuous,
and possess the form characteristic of the Scutibranchia
(figs. 5,6, 7 and 14). The two gills, though symmetrically
placed, are not equally developed, that on the left side being
very much larger than that on the right (cf. figs. 5 and 7).
This is a very interesting feature, which is obviously con-
nected with the dextral coiling of the shell, and one which is
of great significance when studying the phylogeny of the
Azygobranchia. Tach gill is characteristically bipectinate,
consisting of an axis which takes the form of a long and some-
what stout septum, containing the efferent and afferent bran-
chial vessels, and two sets of gill-filaments, which have the

THE ANATOMY OF PLEUROTOMARKIA BEYRICHII. 223

form of triangular plates, whose surfaces bear a number of
fine plications (fig. 14). In each gill the inner or under set
of plates are somewhat smaller than the outer set, a con-
dition leading towards the more specialised one seen in the
Trochide. As in other Scutibranchs, the anterior end of
each gill is not attached to the mantle, but projects freely
into the mantle-cavity, and, in the contracted state of the
mantle, almost beyond the anterior margin of that fold.

Structure of the Gill-plates.—A careful study of
sections of the gill taken through the three principal planes,
i.e. transverse to the long axis, longitudinal sections parallel
to the gill-septum, and horizontal sections, enables us to
construct a diagram (fig. 15) showing the circulation of the
blood in the gill-plates. ‘The afferent branchial vessel (a. b.),
as we have already seen, is situated at the ventral edge of
the gill-septum under a thickened ridge of glandular epi-
dermis ; this vessel gives off on either side small branches,
which enter one into each of the very thin gill-plates. The
vessel then spreads out as a delicate film between the two
laminee which together constitute the plate. After the blood
is aérated by being brought into such close proximity to the
sea-water it leaves the plate near its dorsal attachment, this
efferent channel joins across the septum with the corre-
sponding vein from the opposite plate, the conjoint vessel runs
up the septum and enters the efferent branchial vessel (e. b.),
which lies at the junction of the septum and mantle. It must
not, however, be supposed that the space between the two
lamin of a gill-plate is as simple as represented diagrammati-
cally on the left of fig. 15; such is not the case, the space
being broken up into numerous small channels by the pre-
sence of great numbers of interlaminar connections (2. J. c.,
figs. 15, 16, and 18), extending across the space and joining
the two laminze which compose the gill-plate. The blood thus
takes a very sinuous course among these connections. A
somewhat larger channel is, however, present all round the
margin of the gill-plate.

The extremely delicate nature of the gill-plates is well seen

224 MARTIN F. WOODWARD.

in fig. 16, which represents vertical sections across two gill-
plates. The ventral and dorsal margins are seen to be
dilated, as also are the blood-spaces nearer the dorsal margin,
in which region also the interlaminar connections are larger
and fewer, whence the dotted appearance seen in fig. 14. The
curious crumpling shown in the ventral part of these plates
(fig. 16) represents transverse sections through the folds seen
in the surface view of gill-plate (fig. 14); this appearance
suggests that the margin of the gill-plate is too short to
surround the central area without the latter becoming puck-
ered. The gill-plates present another interesting feature in
the presence along their outer margins of a couple of sup-
porting rods (s. 7.), the relations of which are well shown in
figs. 15 and 18. Froman examination of a transverse section
of these rods (fig. 18) it will be seen that they are flattened
structures, closely applied to the base of the epidermal cells,
and enclosing between them portions of the blood-space of the
gill-plate. A section taken parallel to the gill septum and
passing through the dorsal junctions of the gill-plates with
the septum (fig. 17) shows that the two rods in each plate are
perfectly independent of one another, and that each rod is
related to two gill-plates. In other words, each rod is a
U-shaped structure which embraces the space between two
gill-plates, one limb extending into each of these plates, a
condition which at first sight reminds one of the relation of
the gill-skeleton in the Lamellibranchia. he epithelium
covering the margin of the gill-plates is thickened and special-
ised ; that covering the ventral margin is ciliated (fig. 16), so,
too, is that covering the outer border of the plate (fig. 18).
Specially long cilia are present near the outer margin in two
bands, one on either side of the plate a short distance from
the free margin (figs. 16 and 18); the cells bearing these
cilia are particularly large, and are closely related to the
supporting rods.

Comparison with the Lamellibranch Gill.—The
general structure and relation of the gill as seen in the
zygobranchiate Diotocardia is highly suggestive of that

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 229

met with in the more lowly: Lamellibranchs, viz. the Proto-
branchia (Mitsukuri!). Consequently one might naturally
expect to find a similar resemblance in the finer structure of
these two organs. In this, however, one is doomed to disap-
pointment, for although at first sight there appears to be a
great resemblance (cf. Mitsukuri, op. cit., pl. xxxiv, figs. 6
and 8, and my figs. 17 and 18), yet when we examine this
more carefully we find that instead of a resemblance, there 1s
in reality a very marked difference. Thus in Nucula the
supporting rods lie along the ventral border of the gill-plates
and meet along the ventral edge of the gill-septum, whereas
in Pleurotomaria, as we have seen, the gill-skeleton is
situated along the outer or dorso-lateral margin of the plates,
and the connections between the adjacent rods take place at
the dorsal attachment of the plates to the septum. Simi-
larly, the modified ciliated epithelium, which is closely related
to these rods, is dorso-lateral in the Gastropod, and ventral in
the Lamellibranch. We thus see that there is a very strik-
ing and fundamental difference in the relation of the gill-
skeleton in the two groups, and one which must tend to throw
back the common ancestor of the two to a still earlier period
than that generally assigned to it.

The dorso-lateral position of the supporting rods is, how-
ever, found in another great Molluscan order, the Cephalopoda.
Thus in Sepia Burne” has described cartilaginous rods, one to
each gill-plate, strengthening the supporting lamella, in a
position corresponding to the outer and dorsal margin of the
gill-plate of Pleurotomaria. ‘This is an interesting point,
for, as we shall see later, the Diotocardia appear to approach
the Cephalopoda further in the relation of the spiral stomach-
cecum. ‘The skeletal difference between the gills of the
Diotocardia and the Protobranchia is, however, far more
surprising than the resemblance of the former to the Cepha-
lopoda, for, in addition to the general form and relation of

1 “On the Structure and Significance of some Aberrant Forms of Lamelli-
branchiate Gills,” ‘ Quart. Journ. Mier. Sci.,’ vol. xxi, 1881.
2 «Proc. Mal. Soc.,’ vol. ili, p. 53.

VoL. 44, PAR! 2.—NEW SERIES, P

226 MARTIN F. WOODWARD.

the gill, Nucula approaches the Diotocardia in so many
other respects that one would naturally have expected a very
close agreement on this point. In the face of the unexpected
difference one feels some doubt as to the full value of
generally accepted views on the relations of these forms.

The Branchial Ganglia.—On the outer side of each
gill and close to its anterior point of attachment is situated
avery conspicuous hemispherical swelling (figs. 5, 6, and 7,
bn. g.). These protuberances, which were described by Dall
as blood-sinuses, are caused by the presence of a large
ganglion, situated on the branchial nerve. ‘I'he branchial
gangha are the most conspicuous ganglionic swellings on
the nervous system. In section (fig. 19) they exhibit a great
accumulation of nerve-cells, arranged in two layers round
the periphery of the ganglion, a narrower outer and a broader
inner layer, the two being separated by a very narrow band
of fibrous tissue. ‘The centre of the ganglion is occupied by
a great mass of fibrous tissue, the bundles of which run in
various directions. Near the periphery of the central mass
are some curious dim bodies, which at first sight suggest
large ganglionic cells; but the entire absence of nuclei and
the want of sharpness of outline lead me to conclude that
they are in reality bundles of fibre, rather more closely
packed than usual.

A very large nerve is given off from the ganglion to the
gill, and from this is derived that very conspicuous layer of
nerves (fig. 15, 1. J.) following the course of the efferent
branchial vessel.

The Osphradium.—Dall has figured a small hemispheri-
cal structure, situated somewhat nearer the middle line than
is the branchial ganglion (blood-sinus of Dall), which he
thinks may represent the osphradium. His description of
the position of this organ is, however, not very clear; and a
comparison of his figure (op. cit., pl. xxx, fig. 2) with my
fig. 6 suggests that the gill he represents is the right gill
seen from below, in which case his osphradium would in
reality be situated externally to the branchial ganglion, its

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 227

more median position being only an apparent one due to the
displacement of the mantle in the dissection. However this
may be, I have utterly failed to find any such structure,
either internally or externally to the branchial ganglion ; and
if I am correct in my identification of the osphradium I
cannot help doubting the osphradial nature of the structure
to which Dall assigns this significance in P. Adansoniana.

While examining a series of transverse sections through
the free end of the gill I noticed a thickened patch of epithe-
hum situated on the ventral or external border of the gill,
and extending from near the point of attachment to the free
end (figs. 5—7, os.). An examination of these sections
shows that this thickened patch of epithelium overlies the
great branchial nerve, and receives numerous branches from
it. The epithelium itself has all the character of a sensory
one, consisting as it does of delicate fusiform cells supported
by more columnar ones, and also exhibiting a few pigment
cells. Although a few gland cells are to be seen they are
much less numerous here than in the adjacent epithelium.
At times there is even a suggestion of the bunching together
of these cells into oval masses, not unlike the taste bulbs of
the Vertebrata and the sensory organs of Acavus Waltoni,
as described by the Sarasins.

The identity of this strip of sensory epithelium with the
osphradium is confirmed by a comparison with the latter
organ, as seen in Haliotis. In this genus the osphradium,
as described by Spengel, has precisely the same relationship
and form as the strip of sensory epithelium found in Pleuro-
tomaria Beyrichii, which last we may, I think, safely
identify as the osphradium.

At the posterior end of each gill the afferent and effer-
ent branchial vessels may be seen (figs. 5 and 7, a. b. and
e. b.). The former, springing from a sinus (figs. 7 and 23,
v. 8.) situated ventrally to the rectum and ureter, run forward
near the former structure and diverge outward to the gills;
while the latter, passing on either side of the mantle-cavity,
converge on the heart. As we have already seen, the afferent

228 MARTIN F. WOODWARD.

vessel lies near the free margin of the gill-septum, while
the efferent vessel is situated at its base.

The Hypobranchial Mucous Glands.—One of the
most striking features in the mantle-cavity is a large oval
glandular structure, which, occupying a median position,
extends from the posterior limit of the mantle-slit along the
roof of the mantle-cavity to about the level of the posterior ~
end of the right gill (figs. 5, 6, and 7, m. g.).

In one specimen this gland still retained a pinkish colora-
tion. This gland is partially divided by a median longitudinal
furrow into two halves, each of which is marked by a number
of more or less interrupted grooves which converge on the
median one. The whole structure presents an appearance
not unlike the venation of a leaf. Anteriorly, however, the
two halves of the gland slightly separate from one another,
and end in a couple of pointed structures, in which Dall
thought he could perceive openings which he took to be the
renal apertures. In this supposition he was mistaken, for
the renal organs have a perfectly normal position, and the
gland, as may be seen from a microscopic examination, is a
true mucous gland: further, if we examine Haliotis we
shall find a gland, the hypobranchial mucous gland, occu-
pying a precisely similar position; and I think there can
be no question as to the homology of these two structures
and of the similarly named gland of the Monotocardia.

T'wo additional mucous glands are found in the roof of the
mantle-cavity behind the large hypobranchial gland (figs. 5
and 7, mg’. and mg”.) ; these are situated one at the base of
each gill, that on the left being much the largest, a further
example of the asymmetry which we have already seen fore-
shadowed in the gills, and which affects the whole pallial
complex.

The rectum is situated somewhat to the right side, and
extends forward over the hypobranchial gland in a variable
manner (cf. figs. 5 and 7); but in neither of my specimens
does it extend so far forward as in the P. Adansoniana
figured by Dall.

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 229

The Kidneys.—As in the majority of the Diotocardia,
there are two kidneys, a right and a left one, which exhibit
very different structure and perform different functions. The
left kidney or papillary sac is situated in the left-hand upper
corner of the manitle-cavity. In form it is somewhat oval,
and it opens by a wide slit-like aperture into the mantle-
cavity near the rectum (figs. 7, 23, and 26, l.k.a.). On
cutting open this sac it is found to have a large central cavity
bounded by thick walls, whose epithelium is thickened and
forms numerous rounded papille ; the outer wall is, moreover,
folded. As in Trochus and Haliotis, this left kidney
alone communicates with the pericardium, the reno-peri-
cardial pore taking the form of a long canal, which runs
along the floor of the papillary sac and opens into it near its
external aperture by a ciliated slit (figs. 24—26, r. p.c.).
The structure of this kidney and the relation of the reno-
pericardial pore closely resemble that seen in Trochus, the
only difference being that in the latter the reno-pericardial
canal is distinctly shorter, about half as long, but otherwise
it has the same relation. Unlike the condition seen in
Patella (Goodrich!), it is the aperture leading into the
kidney which is ciliated in Trochus, and not that leading
into the pericardium.

A microscopic investigation of the papillary sac shows that
this organ is highly vascular, but I have been unable to
ascertain whence this blood-supply is derived. According to
Perrier? the vascular system of the left kidney of the Dioto-
cardia is directly connected with the auricle or auricles, is,
we may fairly assume, a similar condition for Pleurotomaria.
The folds and papille which project into the central cavity
are invariably supplied with conspicuous blood-lacune, which
break up into a rich capillary system ; this lies embedded in
a connective-tissue framework containing large quantities of

1 “On the Reno-pericardial Canals in Patella,” ‘ Quart. Journ. Micro.
Sci.,’ vol. xli, 1899.

2 Recherches sur |’Anatomie et PHistologie du rein des Gastéropodes
prosobranches,” ‘ Ann, Sci. Nat.,’ (7) Zool., tom. viii, 1889,

230 MARTIN F. WOODWARD.

leucocytes. The whole papilla is covered by an epithelium
whose cells are somewhat conical; the free expanded bases of
these cells are crowded with yellowish granules, which, since
they are also seen in some of the leucocytes, are probably
waste matter taken up by phagocytes in different parts of
the body, and carried to the papillary sac to be discharged.

The right kidney is very large and complicated, and
probably forms the more important excretory organ, beside
serving to transmit the genital products. ‘This kidney opens
into the mantle-cavity through a glandular tube, which, from
a situation to the right of, has now come to lie almost ven-
trally to the rectum. This thick-walled glandular tube passes
behind into a thin-walled funnel-shaped structure, which
may be termed the ureter (w.), but is really the commence-
ment of the kidney-chamber (k. c.) ; this passes back beneath
the pericardium, and enlarges behind this structure to form
a wide chamber with thick walls, the posterior portion of the
right kidney (p. 7. k.). The walls of this chamber project into
the cavity in the form of a series of deep semilunar folds,
covered with glandular epithelium and richly supplied by a
plexus of blood-vessels containing venous blood. ‘This,
however, only forms a part of the right kidney, a very large
portion running forward below the floor of the mantle-cavity
between the crop and the intestine as far forward as the point
where the brown tint stops, and marked a.r. k., fig. 7. We may
speak of this portion as the anterior lobe of the right kidney
(figs. 23, 25, and 26, a.7r. k.) ; its cavity communicates with the
kidney-chamber near the anterior boundary of the pericardium.
Like the posterior lobe it is richly supplied with venous blood,
since it receives all the blood coming from the anterior part
of the body and from the foot on its way to the gills.

The right kidney has thus a very complicated form, and
one that will be best understood by an examination of the
diagrams given in figs. 25 and 26.

The Genital Organs.—Of the three specimens examined
two were females and the third a male. The genital gland,
which presents a similar appearance in both sexes, forms as

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. De

in other Prosobranchia a fairly thick investment to the dorsal
surface of the liver (fig. 7, g. g.), extending to the top of the
spire.

The origin of the efferent duct from the gland was not to
be made out owing to the upper part of the body being slightly
damaged in removal from the shell, but I think there can be
no doubt that the genital products are shed into a series of
thin-walled remnants of the true ccelom, which in turn unite
to form the somewhat thickened duct (g.d.) shown in figs.
23—26. This duct, which is present in both sexes, runs on
the inner side of the spire, and communicates by a slit-like
opening (g.a@.) with that portion of the right kidney-chamber
which we termed the ureter (u.). The conclusion that this
is the genital duct is supported by a comparison with
Trochus, where the undoubted genital duct has precisely
the same relationships.

In the male this constitutes the whole of the genital system,
there being no accessory organs, the genital products passing
out directly through the unmodified right ureter. In the
female, however, the distal portion of the ureter which serves
to transmit both the excretory and genital products becomes
much modified, owing to the enormous development of glan-
dular tissue in its walls; the latter become so much thickened
that it is by no means easy to find the lumen of this tube,
which may now be called the oviduct (ov. d.).

The presence of this modified oviduct places Pleuroto-
maria about on the level with the Trochidz, and indicates a
somewhat more specialised condition than that met with in
many Diotocardia, for in these latter the genital products are
discharged into the mantle-cavity through the unmodified
right kidney duct,—in some cases, it is thought, without the
intervention even of the simple genital duct seen above.

The Alimentary Canal.—The mouth communicates with
a thick-walled buccal cavity situated in the free portion of
the head. This buccal mass, which is slightly constricted by
the nerve-ring, is closely attached to the body-wall by nu-
merous short radiating muscle-fibres (figs. 6, 7, and 8), which

DSTA MARTIN F. WOODWARD.

are, however, less developed posteriorly where the salivary
gland (fig. 8, sl. g.) occupies the roof of this structure.

On opening the buccal cavity a couple of laterally placed
folds, covered with horny matter, will be seen (fig. 9,j.).
These folds, which undoubtedly correspond to the jaws of
other Gastropods, are but feebly developed in Pleuroto-
maria, and probably serve, as Dall suggested, to protect
the soft wall of the buccal cavity from the scraping action of
the radula. The structure of one of these is shown in fig.
54. In front of the horny jaws a number of small flattened
papille, also covered with horny matter, are to be seen (fig. 9,
h. p.). Between and behind the jaws the ventrally placed
odontophore may be seen bearing the chitinous radula (rd.),
the functional portion of which when at rest appears some-
what V-shaped when viewed from above, and thus only covers
the central portion of the odontophore, the sides of which are
covered by the lateral extension of the radular membrane.
Thus the whole of the buccal cavity is more or less protected
by a lining of chitin.

The Salivary Glands form a compact mass in the roof
of the alimentary canal, at the junction of the buccal mass
and the crop (figs. 6—8, s. g.) ; their ducts (sl.d.), which are
closely related to the buccal nerves, run forward within the
thickness of the wall of the buccal mass, and open into the
buccal cavity just above the odontophore (fig. 10). The
gland, which is a much branched one (fig. 8), was not well
enough preserved to enable me to study its histology.

The odontophore is enormously developed, being highly
muscular, and further strengthened by the odontophoral car-
tilages. When at rest it forms a comparatively slight projec-
tion into the buccal cavity, but, on the other hand, it projects
as a great muscular mass into the hemoceele (fig. 7, od.).
Between it and the crop the enormous radular sac (r.s.) will
be seen extending back for two or more inches, and becoming
involved in the anterior lobe of the right kidney.

The Musculature of the Buccal Mass.—As in other
Odontophora, the muscles of the buccal mass can be divided

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 230

into the extrinsic and the intrinsic muscles; the former
being concerned more in the movement of the mass as a
whole, while the latter are specially related to the movements
of the odontophore.

Extrinsic Muscles.—Curiously enough, these seem to be
mainly protractor muscles, the retractors being but feebly
developed.

(1) The lateral protractors. Three laterally placed
vertical sheets of muscle arising from the side wall of the
head, and inserted towards the posterior end of the buccal
mass (fig. 30 A, l. pr.).

(2) The ventral protractors. A large paired muscle
arising from the region of the lower lip, and inserted upon
the basal cartilages (fig. 30 H, v. pr.).

(3) The lateral retractors (? divaricators of the carti-
lages). Five or six small strands of muscles arising from the
side wall of the head, and inserted upon the main odonto-
phoral cartilage just below the edge of the radular membrane
(figs. 9 and 30 4, l.r.).

(4) The ventral retractors (?). A pair of short longi-
tudinal sheets of muscle arising from the body-wall just
above the pleuro-pedal cords, and inserted upon the radular
sac as it emerges from the odontophoral mass (figs. 29 and
a0 A, vor.) .

(5) The depressor muscle (figs. 30 Band D,d.m.). A
small muscle inserted upon the main cartilage, just in front
of the insertion of No. 3, and passing down to the ventral
side of the head.

I have called Nos. 3 and 4 retractors because when the
buccal mass is protruded their fibres would be on a stretch,
but I think that this is probably only part of their function.
Thus if the right and left portion of No. 3 contracted to-
gether they would separate the main odontophoral cartilages,
and No. 4 may also function to prevent too great a displace-
ment of the growing part of the radula.

Intrinsic Muscles.—These, again, fall under two heads:
those concerned in the movements of the radula itself by

234 MARTIN F. WOODWARD.

acting directly upon it, or upon the infra-radular membrane ;
and those concerned in the movements of the odontophoral
cartilages.

On examination of the odontophore from the side, after
removal of the extrinsic muscles, three muscles will be seen
(fig. 830 B). One of these (d.l.m.) runs from the outer edge
of the infra-radular membrane to the upper border of the
main odontophoral cartilage, the fibres being arranged some-
what obliquely to the length of the cartilage. This muscle
must by its contraction serve to flatten, 7.e. expand, the
radula, and at the same time slightly pulls it back over the
odontophoral cartilages. It is the largest and most powerful
of the intrinsic muscles, and may be termed the dorsal or
postero-dorsal longitudinal muscle. Three muscles are an-
tagonistic to this; one of these (v.l.m.) 1s a small ventro-
lateral band attached in front to the antero-ventral edge of
the infra-radular membrane, and behind to the accessory
basal odontophoral cartilage. This muscle, which we may
term the ventral or antero-ventral longitudinal muscle, serves
to pull the radula over the odontophoral cartilage, and also
to flatten the anterior part of the radula. The second of
these muscles is not seen in this dissection, since it lies on the
inner side of the main cartilage; it is, however, shown in the
median, the ventral and dorsal aspects (fig. 80 H, #, G,7.l.m.).
This muscle is attached to the under side of the radula and
to the infra-radular membrane, where it underlies the middle
functional part of the radula, its insertion forming an oblique
line, starting near the median ventral line, and passing up-
wards and outward until it ends on the edge of the basal
membrane of the radula; posteriorly this muscle is attached
behind to the accessory basal cartilage. The contraction
of this muscle causes the radula to assume once more its
V-shaped grooved character, and in addition it acts as a
powerful retractor. It may be termed the internal longi-
tudinal muscle. The third muscle is a very small one (fig.
30 D, x.) attached to the infra-radular membrane laterally,
and running forwards it is inserted up the anterior portion

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 235

of the odontophoral cartilage; it pulls the radula forwards
and inwards.

The remaining muscles only act indirectly upon the radula
through the movements of the odontophoral cartilages. One
of these lies at the side of the main cartilage, to which it is
attached by a long fleshy insertion (fig. 30 C, l.1.m.) ; it then
runs back as a flat muscular band, and takes its origin from
the outer border of the basal cartilage. The contracture of
this pair of muscles causes the anterior ends of two main
cartilages to diverge, and so tends to flatten the anterior
part of the radula. The second of these two muscles is
situated ventrally, and is the only unpaired muscle in the
buccal mass ; it consists of a transverse band of fibres running
from the outer border of one main cartilage to the corre-
sponding surface of the other, and thus by its contraction
approximates the cartilages (fig. 30 G and H, v. t.m.).

The odontophoral cartilages are fourin number. Of these
two are very large and laterally compressed, constituting the
main cartilages which support the radula. The remaining
two are the small basal plates presenting concave surfaces
for articulation with the former. In spite of the small size
of the basal plates, they appear to be the relatively fixed
points for insertion of the majority of the muscles of the
buccal mass.

The radula itself will be considered later.

Owing to the complicated nature of the movements of the
radula we commonly find that the muscles of the odontophore
are similarly complicated. Unfortunately it is not easy to
ascertain with any degree of precision the exact nature of the
movements produced by the contraction of a given muscle,
and consequently it is inadvisable in the present state of our
knowledge to give them very precise names.

It is interesting to find that the arrangement of the odonto-
phoral muscles of Pleurotomaria compares very closely
with that described as occurring in Patella by Geddes.!

* On the Mechanism of the Odontophore of certain Mollusea,” ‘Trans,
Zool. Soc.,’ vol. x, 1879,

236 MARTIN F. WOODWARD.

Thus we find in both forms similarly placed lateral and
ventral protractors among the extrinsic muscles, while
among the intrinsic the dorsal, ventral longitudinal muscles
connected with the infra-radular membrane are similar, as
also is the transverse ventral muscle. The remaining
muscles, however, differ, as, moreover, do the cartilages,
since there are three pairs of cartilages in Patella and only
two pairs in Pleurotomaria,

The Crop.—As in most Diotocardia, the first portion of the
cesophagus is much dilated and saccular, and may be thus
spoken of asacrop. Itis closely connected with the body-wall
by fine bundles of muscle-fibres, making the removal of the
latter very difficult, and giving the crop a villous appearance.
Its internal structure also is very characteristic, its walls, as
in Haliotis, being thickly covered by numerous papillee
(figs. 9 and 10). These papilla are, however, wanting in
front where the crop and buccal cavity join, and in the
morphological dorsal and ventral middle line. The epithelium
covering these papillz is highly glandular, and the centre of
each papilla is a blood-lacuna. The presence of the papille
thus causes an increase of the secretory epithelium.

Situated immediately behind the odontophore is a somewhat
oval thickening. At a little distance from this structure we
find on either side a slit-like depression (fig. 10, lp. lp'.),
which we may term the lateral cesophageal pouch. Hach of
these depressions is bounded by a couple of folds (lettered in
fig. 10, 1 and 2 on the left side, and 3 and 4 on the right).
Tracing these structures back, we find that by their enlarge-
ment and rotation they cause the crop to assume a very
complicated form. Thus the two ventral folds 2 and 3,
enclosing between them the ventral median area, pass first to
the right side and then gradually ascend until they assume a
dorsal position; this causes a corresponding displacement of
the two dorsal folds (1 and 4), which pass down the left side
until they attain a ventral position. At the same time the
lateral pouches become enlarged, and undergo a corresponding
displacement ; thus the original left pouch (/p.) now constitutes

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 237

the ventral and right half of the crop, while the right pouch
(Ip}.), which is somewhat smaller, forms the dorsal and left
portion of the same.

The non-papillate median dorsal and ventral areas remain
small, and are practically reversed in position. ‘The position
of these folds and pouches is shown in figs. 10 and 11, the
latter being a diagrammatic transverse section of the crop,
from which it will be seen that the enlargement of the crop is
practically confined to the lateral pouches.

Tracing the crop still further back it is found to gradually
diminish in size and complication, until it assumes the form
of a simple tube with rather thick walls, which we may speak
of as the cesophagus (fig. 7, 0. ¢.).

The Stomach.—The cesophagus extends back a little
behind the heart, and then suddenly debouches into the
stomach, which, as a large U-shaped cavity, lies below and
behind the right kidney (figs. 7 and 12, st.) The cavity of
the stomach is large and divided by a marked constriction
into a right and left portion, the former receiving the
cesophagus, while the latter receives the bile-duct (6. d.), and
gives origin to the intestine and the spiral caecum (sp. c.).
The cesophageal aperture is very narrow and guarded by a
sphincter, while the intestinal orifice is large (fig. 12). The
bile-duct (b. d.) opens by a wide slit situated immediately to
the left of the semilunar fold which grows in from the floor
of the stomach and separates the two chambers.

The spiral cecum (sp. c.) opens on the dorsal wall imme-
diately above the bile-duct, but the structure which may be
described as the columella of the spiral caecum is prolonged
down to the floor of the stomach, and forms the anterior lip
of the constriction between the two stomach-chambers. The
ceecum itself forms a perfect helicoid spiral situated dorsally,
and overlying the two halves of the stomach.

A spiral stomach-czecum is a very characteristic feature of
those Diotocardia possessed of a spiral shell, being specially
well marked in Trochus (fig. 13), Turbo,and Phasianella;
it is also developed in Haliotis, and in a very much modified

238 MARTIN F. WOODWARD.

form in Scutum and Fissurella. Traces of this organ are
also found in that primitive tenioglossan, Nassopsis.
The almost universal occurrence of this organ in the Dioto-
cardia suggests that it is a structure of great antiquity and
functional importance, although we are unable to ascribe any
special physiological function to it.

This caecum is in most cases connected with the postero-
dorsal wall of the stomach (postero-ventral in Phasianella),
and its lips are invariably related to the opening of the bile-
duct. Regarding the stomach as a U-shaped structure
composed of an cesophageal and an intestinal chamber, the
cecum invariably arises close to the junction of the two, but
essentially belonging to the intestinal chamber, and is closely
associated with the bile-ducts.

This structure has no homology with the crystalline style sac
of other Gastropoda or of the Lamellibranchia ; the two struc-
tures are undoubtedly co-existent in Nassopsis (Moore),
and possibly in some Diotocardia. It is, however, extremely
suggestive of the spiral caecum present in the Cephalopoda,
which, like the caecum described above, is a postero-dorsal
outgrowth from the stomach, closely related to the bile-ducts
and to the point of origin of the intestine.

An attempt to homologise the spiral ceecum found in two
such distinct orders of Mollusca as the Gastropoda and the
Cephalopoda may at first sight seem unjustifiable, but the rela-
tions of the two organs to the alimentary canal are so precisely
alike that one cannot help being struck with their similitude.

It is, moreover, generally accepted that the Cephalopoda
and Gastropoda are descended from a common ancestor, so
that presence in the two groups of a spiral stomach-caecum
is not so surprising, and would only suggest that this struc-
ture was present in that ancestral form. Unfortunately we
know nothing of the connecting type, which is not astonishing
when we remember that both the Cephalopoda and the Dioto-
cardia extend back to the Cambrian epoch. The only group

1 «The Molluses of the Great African Lakes. IV. Nassopsis and Bytho-
ceras,” ‘Quart. Journ. Micro. Sei.’ vol. xiii, 1899.

HE ANATOMY OF PLEUROTOMARIA BEYRICHII. 239

which is sometimes regarded as representing the primitive
molluscan stock, viz. the Amphineura, does not exhibit this
organ; but, on the other hand, they do not extend back
so far in time, the earliest chiton being only found in the
Ordovician ; and further, the components of this group, while
retaining many primitive features, are obviously specialised
along a particular line, so that I do not think the absence of this
spiral cecum in the Amphineura can be regarded as disprov-
ing the homology of the two ceca seen respectively in the
Cephalopoda and Diotocardia.

From a consideration, therefore, of the similar structural
relations of the spiral caecum in these two groups, I conclude
that the two structures are homologous.

The intestine (figs. 7 and 12, inf.) is very simple. It runs
forward until within about half an inch of the salivary glands,
and then forming a U-shaped bend, it passes back towards
the stomach, whence it curves dorsally, perforating the
pericardium and the ventricle, and bending once more on itself,
it enters the mantle-cavity, to the roof of which it is attached,
at first slightly to the right of the middle line, but gradually
assuming a more median position (figs. 5 and 7,7.). It is
attached below the hypobranchial gland, and opens into the
mantle-cavity by the anal orifice situated some considerable
distance from the posterior limit of the mantle-slit, and there-
fore very differently from the condition observed by Dall in
P. Adansoniana.

The Vascular System.—tThe heart, which is enclosed in
a spacious pericardium (figs. 7 and 23—26), is that of a
typical Zygobranch, consisting of a muscular ventricle (v.)
surrounding the rectum, and a pair of thin-walled auricles (U.
au. and . aw.), which receive the blood from the long efferent
branchial vessels.

A common aorta springs from the posterior portion of the
ventricle, and soon divides into an anterior and a posterior
artery ; the former (figs. 6, 7, and 28, a. a.) is distributed to
the anterior and ventral parts of the body, while the latter
supplies the stomach, liver, and genital gland.

240 MARTIN F. WOODWARD.

The venous system takes the form of series of more or less
well-marked canals and sinuses, which are specially conspicu-
ous in the region of the right kidney. The blood from the
foot and anterior parts of the body is apparently collected
into a series of channels, which run in close connection with
the excretory epithelium of the anterior lobe of the kidney,
while that from the liver and stomach passes through the
posterior lobe. These various renal veins eventually open
into a large sinus situated ventrally to the ureter, genital
duct, pericardium, and rectum (figs. 7 and 23, v. s.), from
which the afferent branchial vessels arise.

The body-cavity of the adult is very inconspicuous, owing
to the great development of the crop with its radiating
muscle-fibres. This cavity represents part of the venous
system, and is of the nature of a hemoccele. The true ccelom is
only represented in the pericardial, renal, and genital cavities.

The Nervous System.—An examination of fig. 27 will
show at a glance that the nervous system of P. Beyrichii
presents all the essential features of that of a typical Dioto-
cardian, this being especially noticeable in the practical
absence of distinct ganglia; for although on the removal of
the dense connective-tissue sheath a certain amount of orange
colour is noticeable in the cerebral and pedal centres, thus
indicating an accumulation of nerve-cells, yet an examination
of a series of sections through these regions and the inter-
vening connectives shows (fig. 22) that while the nerve-cells
are more abundant in these coloured areas, yet they are not
confined to these regions, but are distributed, though in
smaller numbers, throughout the whole length of the connec-
tives, commissures, and even many of the nerves. ‘The
orange-coloured areas where the nerve-cells are more abun-
dant correspond with the cerebral centres (cb. g.), the points
of origin of the visceral loop (pl. c.), and the anterior portion
of the pleuro-pedal cords. This distribution of the nerve-
cells along the connectives makes it extremely difficult to
localise the individual ganglia, and forces us to rely rather
upon the points of origin of certain nerves than upon the

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 241

definite accumulation of nerve-cells met with in other Proso-
branchs.

Fischer and Bouvier seem to have been misled, either by
this coloration or by the slight swelling of the cords in
certain of these regions, into the belief of the existence of
definite ganglionic areas, and they indicate such regions by
means of dots in their figures; they appear to have over-
looked the presence of nerve-cells along the connectives, and
the still more important, though slight, accumulation and
coloration at the point of origin of the visceral nerve.

The Cerebral Ganglia.—The circum-cesophageal nerve-
ring is much enlarged on either side of the anterior part of
the buccal cavity, and since the tentacular and certain other
nerves which are characteristic of the cerebral ganglion of
other Prosobranchs arise from this region, we may regard it
as representing that ganglion. The cerebral ganglia are,
then, a pair of elongate band-like structures widening out
below ; they are connected together above the buccal mass by
a slightly narrow region (cb. c.), which represents the
cerebral commissure of more specialised forms, but which
here is indistinguishable from the ganglia themselves, since
both in its size and in the number of its ganglionic cells it
passes imperceptibly into the laterally placed ganglionic
areas. The cerebral ganglia give origin to five pairs of
nerves supplying the lips (figs. 21, 22, and 29), and to a pair
of laterally placed tentacular nerves (¢. .), from which in
turn the optic nerve arises. Arising with the most ventral of
these labial nerves is a broad nerve which runs downwards
and below the buccal mass (figs. 21 and 29); this nerve
gives off a sixth lip-nerve, and is then continued ven-
trally to the mouth and close to the lips, to meet and fuse
with a similar nerve from the opposite side of the body, thus
constituting the labial commissure (J. c.) so characteristic of
the Diotocardia and archi-Tenioglossa.

Yet another nerve arises from the ventral continuation of
the cerebral ganglion, but in order to see this properly the
mesial aspect of the ganglion must be examined. Such a

VOL. 44, PART 2.—NEW SERIES. Q

242 MARTIN F. WOODWARD.

view (fig. 21) shows a nerve arising just between the fourth
and fifth lip-nerves; this nerve, the buccal nerve (b. n.),
curves sharply up over the muscular odontophore, giving off
branches on its course. After ascending for some distance
it bends sharply back and becomes greatly enlarged, and
may now be spoken of as the buccal ganglion (figs. 8 and 29,
b. g.). This ganglion is a curiously elongate structure, and
gives off branches anteriorly and ventrally ; while the main
mass is continued back under the radular sac, where it unites
with its fellow from the opposite side (fig. 27). A very con-
spicuous branch arises from the middle of the dorsal border of
the ganglion, which curving upwards and backwards runsalong
the salivary duct and supplies the salivary gland (figs. 8and 29).

From the posterior border of the cerebral ganglia two very
large cords arise, these represent the cerebro-pleural and
cerebro-pedal connectives; of these the former and more
posterior cord is as usual much the larger. As with the
other parts of the nervous system, ganglionic cells are
scattered along the length of these cords, more especially at
the periphery, and more abundantly in the cerebro-pleural
than in the cerebro-pedal connective (fig. 22).

The cerebro-pedal connective passes back, taking at first a
somewhat horizontal position, but eventually curving down-
ward to join the great scalariform pleuro-pedal cords. It is
closely followed by the cerebro-pleural connective, and the
two become reunited near the posterior border of the great
odontophoral muscular mass. The combined pleuro-pedal
mass then enters the foot, where it becomes connected by
a transverse commissure with the corresponding structure
from the opposite side (figs. 21 and 27). This transverse
commissure contains elements derived from both the pleural
and pedal systems (fig. 21).

The Pleuro-pedal Cords.—Although the pleural and
pedal cords are now closely connected, they can still be dis-
tinguished from one another by the presence of a groove
which runs along the whole length of the combined pleuro-
pedal cord (figs. 28 and 29).

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 243

This separation is probably only an external one, since
sections through the cords fail to reveal any continuous
layer of connective tissue separating the two, and bundles of
nerve-fibres apparently pass from the ventral to the dorsal
moieties of the cord, and vice versa.

The right and left pleuro-pedal cords now diverge slightly
from one another and run back within the substance of the
foot, each cord lying ina slight blood-sinus situated below
and on either side of the main pedal sinus.

The cords extend to the posterior extremity of the foot,
and are furthest apart near the middle of the foot ; toward the
posterior end they become somewhat approximated (fig. 27).
These cords, along which ganglionic cells are fairly evenly
distributed, are, as we have seen, equally derived from the
pedal and pleural systems: in width each cord at its anterior
extremity is but very slightly if at all larger than the
cerebro-pleural and cerebro-pedal connectives when closely
approximated, i.e. there is no marked swelling indicative of
a great accumulation of ganglionic cells (fig. 22). In fact,
there is but a slight increase in number of these cells in
this region, and that mainly in the ventral or pedal portion.
It becomes then very difficult if not impossible to speak of a
pedal, and, as I shall endeavour to show later, inadvisable to
attempt to identify a pleural ganglion in this pleuro-pedal cord.

These ganglionic pleuro-pedal cords are connected at inter-
vals by transverse commissures: the first of these, as already
mentioned, is derived from both the pleural and pedal
moieties ; but the posterior ones, of which there are at least
twelve, at first sight would be considered as derived ex-
clusively from the pedal portion of the cord. An examina-
tion of sections, however, reveals the fact that a bundle of
nerve-fibres comes down from the dorsal portion of the
pleuro-pedal cords and enters the commissure, which is there-
fore derived equally from both portions of the cord.

The laterally placed pedal nerves arise like the above from
both portions of the cord. ‘The double root of these nerves
is often very conspicuous (fig. 21, p?.), somewhat resembling

244 MARTIN F. WOODWARD.

the dorsal and ventral roots of a vertebrate spinal nerve. In
addition to these large latero-ventral nerves there are present
certain small nerves, which apparently arise from the pleural
portion of the cord and pass to the dorsal pedal muscles.
When, however, we remember that the distinction between
these two portions of the cord is practically only an external
one we Shall probably be right in concluding that all the
nerves derived from these cords are connected with both
subdivisions.

In connection with the apparent separation of the pleuro-
pedal cords into two distinct portions by means of a longi-
tudinal groove it is interesting to note that Haller’ had
already come to the conclusion that this groove has no mor-
phological significance; thus he found in other Rhipidoglossa,
as I have found in Pleurotomaria, that transverse sections of
this cord failed to reveal any line of separation between the
pleural and pedal portions of nerve tracts running from one
into the other.

Visceral Commissure.—As suggested by Bouvier and
Fischer from the study of an imperfect specimen, Pleuro-
tomaria exhibits a typical streptoneurous condition in its
visceral loop (figs. 6, 27, and 28); but at the same time this
mollusc is most peculiar among the Diotocardia in the point
of origin of its visceral nerves.

If the cerebro-pleural connective on either side of the
body be examined, it will be seen that between its origin
from the cerebral ganglion and its fusion with the pedal
system it gives rise to a very large nerve, whose relations at
once identify it with the visceral nerve, that on the right side
being the supra- and that on the left the sub-intestinal nerve.

As already mentioned, the points of origin of these nerves
appear, after the removal of the thick nerve-sheath, slightly
orange-coloured, owing to the presence of a considerable
number of nerve-cells which are continued, though in smaller
numbers, from this point up to the branchial ganglion. ‘The

1 «Untersuchungen tiber marine Rhipidoglossen. II. Textus des Central-
nervensystem und seiner Hillen,” ‘ Morph. Jahrb.,’ Bd. xi, 1886.

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 245

origin of the visceral nerves alone would suggest that we
were here dealing with the pleural centres—a view which is
greatly strengthened by the presence of an accumulation of
nerve-cells, and to which we shall refer again.

The right half of the visceral loop arises fairly close to the
cerebral ganglion, whereas the left half originates very much
closer to the pedal ganglion (cf. figs. 21, 22, and 27), thus
producing a marked asymmetry.

The supra-intestinal nerve passes over the alimentary canal
immediately behind the salivary gland (figs. 6, 10, and
29), whereas the subintestinal passes below the crop and
radular sac. Both of these nerves perforate the muscular
body-wall, and come to lie in the mantle close to its junction
with the former (figs. 6 and 28). Hach nerve then bifurcates,
one portion running back parallel with the gill to complete
the visceral loop by uniting with its fellow in a blood-sinus
just below the right kidney duct (figs. 23 and 27), while the
other portion runs out and gives rise to the immense bran-
chial ganglion, which forms a large round swelling close to
the point of attachment of the gill, which it innervates.
There is no distinct abdominal ganglion such as was surmised
by Fischer and Bouvier.

In addition to the great visceral nerves a number of smaller
nerves arise from the pleural connectives, the presence of
which strengthens the view that the pleural ganglia are not
yet condensed in Pleurotomaria, and that this mollusc is
a fairly primitive one, for in most other Gastropods there is
a tendency for these nerves supplying the muscles of the side
wall of the head and posterior parts of the mantle to take
the form of one large nerve arising in fairly close connection
with the pleural ganglion.

So far as I can ascertain about four moderately conspicuous
nerves arise from the two cerebro-pleural connectives to be
distributed to the side walls of the head (fig. 21). From
the connective behind the origin of the visceral loop, to
which I apply the term pleuro-pedal connective, and before
it fuses with the cerebro-pedal connective, there arise two

946 MARTIN F. WOODWARD.

fairly large nerves (figs. 21, 22, 28, 29); one of these runs
forwards between the cerebro-pleural and cerebro-pedal con-
nectives (fig. 22) to the muscles of the side of the neck,
while the other runs up to the body-wall above the crop to
the floor of the mantle cavity; this last may be Bouvier and
Fischer’s pallial nerve, although it does not arise at the same
spot. I do not feel at all certain about the identity of these
nerves, since I have not been able to trace any of them to
the free mantle-fold, and consequently am not inclined to call
any of them pallial nerves. Still less am I satisfied concern-
ing the presence of the primary pallial nerves of these
authors, and I take it that they rather assume that such
nerves must be present. An examination of their fig. C, op.
cit., p. 170, will show two large nerves arising from the upper
part of the pleuro-pedal cords, the anterior of these corre-
sponding with the nerve marked with an asterisk in my figs.
21 and 22; the nerve runs up parallel to my pleuro-pedal
connective, branching repeatedly, and is eventually lost in
the muscle of this portion of the body: it is possible that
some of its finer fibres may penetrate into the mantle. With
regard to the second, which they represent as co-extensive
with the pleuro-pedal cords, I can only say that it does not
exist in P. Beyrichii. Behind the last-mentioned nerve a
series of four small nerves are seen to arise from the pleural
portion of the pleuro-pedal cords, between the point of origin
of the first and second pedal nerves (figs. 21 and 22, p.! and
p.2). These nerves, which are distributed to the muscles on
the dorsal surface of the foot where the latter joins the body,
—i.e. to the commencement of the columella muscle (fig. 28)
—are the only nerves which occur in the region corre-
sponding to that whence Bouvier and Fischer’s great hypo-
thetical pallial nerve springs; they are, however, quite small,
and I have not been able to trace them beyond the columella
muscle. Since I cannot think that there is likely to be any
great difference between the different species of Pleuroto-
maria in this respect, I can only conclude that the great
primary pallial nerve of Bouvier and Fischer does not exist

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 24:7

in Pleurotomaria. Even in Trochus I think they greatly
overrate the size and importance of this nerve, which so far
as I can find is a small nerve distributed to the posterior
portion of the mantle, and Eepab ls to the secretory epithe-
lium of that region.

A few small nerves arise from the cerebro-pedal commis-
sure, these being distributed to the muscles between the
under lip and the foot.

The Sense-organs.—The eyes and osphradia having
been already described, it only remains to draw attention to
the otocysts. These latter take the form of a pair of large
vesicles, situated just above and in front of the pleuro-
pedal cords (fig. 27). The actual otocyst is not very large,
but it is surrounded by a very tough, thick, concentric ar-
ranged sheath of connective tissue (fig. 31 4). The otoconia
are small and numerous; typically they are spherical bodies,
varying much in size and often fusing together to form
reniform structures (fig. 51 B).

The Radula (figs. 32—52).—The radula of Pleuroto
maria Beyrichii is extremely complex, and exhibits the
same type as that described for P. Adansoniana by Dall,
and P. Quoyana by Fischer and Bouvier. In the nuraber
and character of its teeth it more closely approximates to the
latter species—a fact which strongly supports the view ad-
vanced by Crosse in 1882, from the study of the shells, that
these two species should be grouped together as a section or
sub-genus of Pleurotomaria, for which group P. Fischer
proposed the name Perotrochus.

The radula is very large, one example measuring 62 mm.
long by 5mm. wide. The greater portion of the radula is of

course not functional, but lies buried in the radular sac,
which extends up to the anterior lobe of the right kidney.
In fig. 32 I have given a view of half a transverse row,
which, as mentioned by Fischer and Bouvier, does not run
straight across the radula, but has a somewhat V-shaped
course.
The number of teeth in a transverse row is 223; one of

248 MARTIN F. WOODWARD.

these being unpaired occupies the centre of the row, and on
either side of this are situated 111 teeth. So far as I can see
from the examination of many rows this number is quite con-
stant. The lateral teeth exhibit a number of different types,
at least five, which, however, merge imperceptibly into one
another. For convenience’ sake we may, however, follow
Fischer and Bouvier, and divide them into the following
groups:—(1) The central teeth, (2) the lamellate teeth, (3)
the hooked teeth or uncini, (4) the brush or tufted teeth, (5)
flabelliform teeth.

The Rhachian or Unpaired Tooth.—This tooth (figs. 32
and 33) hasa very curiousform. Viewed from above (fig. 32),
it appears to consist of a somewhat pointed oval or lanceolate
lamella, which overlaps the adjacent central teeth. When,
however, the rhachian tooth is isolated and viewed from the
side (fig. 33), it is seen that this more or less horizontally
placed lamella is attached to a longitudinally placed vertical
plate, the posterior half of which is thickened, and forms the
base of attachment of the tooth to the basal membrane. The
tooth thus consists of two pieces—a flat horizontal lamella,
and a vertical plate strengthening and attaching the former
to the radular membrane.

The Central Pairs.—On either side of the rhachian are
situated three large teeth (fig. 382), which, while asymmetrical
in form, nevertheless approximate somewhat in structure to
the symmetrical rhachian tooth, forming a gradation between
this tooth and the more laterally placed lamellate teeth. It
is very difficult, if not impossible, to draw a line between
these central pairs and the lamellate teeth, and we only separate
them for the convenience of description. In the central
teeth (figs. 34—36) the vertical plate has greatly increased
in size, while the horizontal lamella, so characteristic of the
rhachidian, is much reduced, and only present on the outer
side of the vertical plate near the base of attachment ; it still,
however, overlaps the tooth immediately external to it (fig.
32). The portion corresponding to the vertical plate of the
rhachian is now no longer placed vertically, but has become

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 249

inclined to the basal membrane, and its anterior extremity is
much widened. ‘The form of these teeth may be best under-
stood by an examination of figs. 32—36.

The Lamellate Teeth.—lIt will be seen from figs. 32 and
36 that the central teeth are not sharply marked off from the
more laterally placed lamellate teeth, but pass imperceptibly
into them. ‘Thus, owing to its size, the first lamellate tooth
might be almost equally well classified with the central
teeth.

There are twenty-one teeth which may be grouped under
this head. Though varying in size, they are, on the whole,
the smallest teeth in the radula, and present very uniform
characters. ‘I'he free end of each of these teeth is abruptly
truncated, and the upper border is generally straight, while
the under margin is either convex or angulated, and the base
of attachment is small (figs. 32, and 37 A, B,C, D). The
lamellate teeth are far more complicated than would appear
at first sight, so much so that it is very difficult indeed to
gain any idea of their form from a written description. I
have therefore thought it better to give several drawings of
one of these teeth in different positions (see fig. 37 A, B, C,
D),and to these I must refer the reader who desires to obtain
an idea of the form of these very characteristic teeth.

The first five or six of these teeth presenta slightly concave
free border, and thus approach the central pairs, which they
further resemble in the greater development of the outwardly
flexed free margin, which evidently represents the last trace
of the overlapping lamella of the central teeth.

As we pass outward the lamellate teeth increase in size,
and thus approximate to the hooked teeth (figs. 32 and 38).

The Hooked Teeth or Uncini.—The gradation between
the lamellate and the hooked teeth is completed by the
twenty-fifth tooth (fig. 38), which, though only slightly larger
than the preceding tooth, approximates in form to the more
lateral hooked teeth. It will be seen to present a double
curvature in its free margin, and a shghtly hooked free
extremity. The twenty-sixth tooth (fig. 89) is much larger

250 MARTIN F. WOODWARD.

than the above, and shows a well-marked hook at its
extremity, and a small cusp a little below this; this cusp is
still visible on the twenty-seventh (fig. 40) but completely
disappears on the succeeding teeth, which have the form of
long massive hooks (figs. 41 and 42), and constitute the
largest teeth on the radula. After the thirtieth, the teeth,
while still remaining long, become much slighter (fig. 42),
and soon (about the thirty-seventh) begin to show signs of
the development of two additional cusps (fig. 43), which
attain their full development on the forty-ninth tooth (fig.
44). The teeth have now the form of long delicate sickles,
the free end of which exhibits two deep notches. The last
one or two hooked teeth are somewhat shorter than the earlier
ones, and thus lead to the distinctly shorter brush teeth.

I have somewhat arbitrarily drawn the line separating the
hooked and the brush teeth between the forty-first and forty-
second tooth, thus making seventeen hooked teeth.

The Brush Teeth.—The forty-second tooth at first sight
does not appear to differ materially from the forty-first, but a
more careful examination shows that it possesses on either
side on a level with the lowest cusp two minute bristles (fig.
45). On the next tooth these bristles are longer, and one or
two more are appearing (fig. 45); and if we examine the
feature as we pass outward in the row of teeth (figs. 47—
52) we find that the bristles steadily increase in number and
length, until by the forty-ninth tooth they form a consider-
able brush reaching to the free end of the tooth. Thus it will
be seen that it is impossible to separate the forty-second
tooth from the true brush teeth; and although it is more
closely approximate in general appearance to the hooked
teeth, yet, in the presence of the two minute hairs on either
side, it already shows the essential feature of the brush
teeth.

The brush teeth are sixty-three in number, and they form
the most characteristic feature of the radula of Pleuroto-
maria.

A tooth taken from the middle of this series (fig. 50) shows

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 251

a decided reduction in the length of the hooked portion of
the tooth, the three cusps being now somewhat closely
crowded near the free end of the tooth. The tuft of bristles
now appears to be inserted lower down on a ridge placed
transversely to the axis of the tooth, and the bristles them-
selves, forming a compact brush, extend considerably beyond
the free end of the tooth (figs. 5|0—52). With the reduction
of the hooked portion of the tooth the two sets of bristles meet
behind, and now form a horseshoe-shaped brush embracing
the end of the tooth. Passing outward the teeth become
still more delicate, and the cusps smaller and smaller, until they
completely disappear. ‘Traces of the cusp-bearing lobe are,
however, still distinctly visible on the 101st tooth, although
it is now only a narrow slightly notched process. The same
structure, but still smaller and devoid of notches, may be seen
on the 102nd and 103rd teeth, and I think on all the
remaining teeth, in the form of a slight process on upper
border of the teeth.

As the upper tooth-bearing lobe becomes reduced, the two
sets of bristles run together and form a single clump, and
gradually approach the free upper border of the tooth. This
latter condition is accelerated by the development on 101st—
104th of a lamina springing from the back of the tooth, and
foreshadowing the flabelliform tooth (fig. 53).

The bristles remain well developed after the disappearance
of the cusps, and even the 103rd tooth possesses a good brush.
The 104th, however, shows a marked reduction in its bristles,
and this is the last of the brush teeth, since the 105th tooth
is entirely devoid of these structures. In other respects the
difference between these two teeth is slight, their general
form being very similar.

The Flabelliform T’eeth.—There are seven of these
teeth (the 105th to the 111th inclusive), which have the form
of delicate narrow lamelle, arranged like the rays of a fan;
they all bear a shght notch at their free end, corresponding
to the point of attachment of the bristles in the brush teeth,
and possibly representing the tooth-bearing lobe,

252 MARTIN F. WOODWARD.

The Radula as a Whole—If we recognise the five
divisions described above, we may express the arrangement
and number of teeth on the radula by the following numerals :
—7, 68, 17, 20, 3; R. 3, 20, 17, 63, 7; there being, as we
have seen, a single rhachian, 3 central pairs, 20 lamellate, 17
hooked, 63 brush, and 7 flabelliform teeth.

One of the most noticeable features in this radula, how-
ever, is the great difficulty which its teeth offer to our
attempts to arrange them in groups, this being due to the
presence of intermediate forms between each two adjacent
groups of teeth, thus causing them to merge into one ano-
ther, and making it almost impossible to draw any sharp
line between them. Nevertheless there are a number of very
marked types of teeth in this radula, notably the lamellate,
the hooked, the brush, and the flabelliform teeth; of these
the lamellate and the brush teeth are very striking and
peculiar, and not apparently met with in any other mollusc.

It is somewhat difficult to understand the function of the
lamellate and brush teeth, especially the former, and in order
to do so we require to know more about the habits and the
nature of the food of Pleurotomaria. An examination of
the contents of the stomach of two specimens revealed a large
quantity of sponge spicules, both megascleres and micro-
scleres, belonging to one of the Halichondrina (? a species of
Amphilectus). From the fact that many of these spicules
appeared to be bound together by tissue, I conclude that
Pleurotomaria feeds on the living sponge. Tor this pur-
pose the hooked teeth would be useful in tearing away great
pieces of the sponge, and the brush teeth might at the same
time rasp away some of the flesh from the spicules ; but one
is still at a loss to understand the action of the lamellate
teeth.

Another peculiar feature in this radula is the presence of
what Bouvier and Fischer term the accessory basal
plates. ‘These structures take the form of little chitinous
plates, somewhat of the same shape as the basal plates of the
teeth themselves, which are attached to the radular mem-

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 253

brane. These accessory plates, of which there are about
thirty-seven on each side of the middle line in each row of
teeth, are situated about in the middle of each half-row of
teeth, commencing with the twenty-seventh and extending
out as far as about the sixty-fourth tooth. ‘They appear to
alternate with the true bases of the teeth in front, whereas
posteriorly they underlie them. The row represented in fig.
32 would underlie the next posterior row of teeth.

Comparison with the Radula of P. Quoyana and
P. Adansoniana.

The radule of the three species of Pleurotomaria, of
which the animals have been examined, stand apart from
those of all other Diotocardia in the absence of that sharp
division into regions which is so characteristic of the ma-
jority of this group. They are further to be distinguished
by the character of their central teeth, and in the possession
of brush teeth.

Of the two species, P. Quoyana much more nearly ap-
proaches P. Beyrichii in the character of its radula than
does P. Adansoniana. ‘lhe radula of the former, accord-
ing to Bouvier and Fischer, may be expressed as follows :—
R. 3, 24, 138, 63, 6, there being 109 teeth on either side of the
rhachian. Except in the number of the teeth in the different
groups there is very little difference indeed between the two
species, the resemblance being so close that one might almost
match the individual teeth in the two radule ; thus the rhachi-
dian, the central pairs, and the lamellate are very similar,
the only difference being in the greater number of lamellate
teeth (twenty-four) in P. Quoyana. The thirtieth tooth of
the latter species forms an exact match with the twenty-
sixth of P. Beyrichii, and the fiftieth with the forty-third.

This close resemblance between the radule of these two
species is strong argument in favour of the retention of
these two forms in a sub-section of the genus Pleuroto-
maria (section Perotrochus, Fischer). Since P. Bey-

254 MARTIN F. WOODWARD.

richii so closely resembles P. Quoyana in its radula, it
differs equally with the latter species from P. Adanso-
niana, the radula of which, according to Dall, shows a
rhachian, 15 laterals, 5 tufted uncini, 4 denticulate uncini,
and 45 simple uncini, or 69 teeth on either side of the
rhachian. An examination of his figure would lead us to
interpret the teeth somewhat differently, but since the num-
bering of the teeth in the plate and account given in the text
are at variance a detailed comparison becomes difficult. It
is obvious, however, that this form differs considerably in its
radula from the section Perotrochus, and thus justifies
Fischer’s creation of the section Entemnotrochus.

A comparison of the Pleurotomarian radula with that of
other Diotocardia is almost impossible, for while the former
is a typical rhipidoglossate radula, yet it is so peculiar that
we can find no other living form which at all approximates
to it. This is, perhaps, not so surprising when we consider
the great antiquity of this form, on which account we might
expect that Pleurotomaria would show either a very
primitive type, or if, on the other hand, the radula had
undergone much change, that it would show a very spe-
cialised one.

When we attempt to decide the question as to the primi-
tive or specialised nature of this radula, we are at once at
fault, for we have not one particle of evidence to show us
what the nature of the pro-rhipidoglossate radula was. All
the evidence we possess tells us that the Diotocardia are un-
doubtedly the most primitive of living Prosobranchia, and
that they all possess the highly developed rhipidoglossate
type of radula. Of the early Diotocardia, Pleurotomaria
is the only form of which we have any knowledge, all the
other living zygobranchiate Diotocardia being comparatively
modern forms, and this genus also shows us a rhipidoglos-
sate radula. It is true that the radula of Pleurotomaria
differs from that of all other Diotocardia in the absence of
those sharply marked regions which are so characteristic of
the majority of the rhipidoglossate radule. The question

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 255

then arises as to whether, taking into consideration the anti-
quity of Pleurotomaria, we are justified in regarding this
feature as a primitive one.

Considerable stress has been laid by Troschel, Moore, and
others upon the breaking up of the rhipidoglossate radula
into zones, three on either side of the rhachian tooth, and on
the occasional replacement of the great group of marginal
laterals by one large tooth, which, however, generally re-
tains sufficient traces of the individuals which it replaces
to suggest that it represents a fusion of teeth, a view which
is supported when we find that this takes place in undoubtedly
specialised forms (notably in Addisonia and Cocculina
among the Rhipidoglossa, and certain Cyclophoride among
the archi-Tznioglossa). Such a condition has led some to
suppose that the tooth arrangement met with in the Tenio-
glossa might be derived from the Rhipidoglossate radula
by a fusion of the elements of the three zones, thus giving a
formula of 1.1.1.1.1.1.1, a view which the condition of the
archi-Teenioglossate Cyclophoride seems to support.

If, then, this subdivision of the row of teeth into sharply
marked zones is a foreshadowing or a tendency in the direc-
tion of the condition met with in the 'Tenioglossa, it seems
only natural to conclude that this in turn was derived from
a radula in which all the teeth in a transverse row were
sunilar. Such a stage has not been preserved to us, but in
Pleurotomaria we have an approximation to this condition,
inasmuch as all the various specialised tooth areas merge
imperceptibly into one another, and this in my opinion is a
very primitive character.

I therefore conclude that, in spite of its very specialised
brush teeth, the radula of Pleurotomaria exhibits the most
primitive type among all existing Gastropods.

Considerations regarding the Primitive Nature
of Pleurotomaria.—lIf we are justified in concluding, as I
have done above, that in its radula Pleurotomaria is a
most primitive form, then we might naturally expect to find
indications of this primitive character in other of its organs.

256 MARTIN F. WOODWARD.

We have already in dealing with the different organs
suggested that certain of them presented primitive characters,
—for example, the eye, the spiral ceecum in the stomach, and
still more notably the characters of the nervous system.

The morphology of the nervous system has already been
dealt with at length by Bouvier and Fischer, especially with
reference to the relation of the pleuro-pedal cords to the
origin of the visceral connectives from the conditions seen in
Chiton. Personally, however, Ido not think Pleurotomaria
throws any fresh light on this branch of inquiry. This,
nevertheless, does not rob the nervous system of the mollusc
of all interest, for, as we have already seen, in the very
uniform distribution of the nerve-cells through the connec-
tives and commissures, and the consequent practical absence
of distinct ganglia, we have retained in Pleurotomaria, no
matter what view we take of the origin of the molluscan
nervous system, a very primitive feature.

The second point of interest, which taken in connection
with the above yields to no other feature in the anatomy of
Pleurotomaria in its importance, concerns the position of
the pleural centres. In the Gastropoda the pleural ganglia
may be defined as the accumulations of nerve-cells related
to both the cerebral and pedal ganglia, and giving origin to
the visceral connectives, these being the only constant
features presented by the pleural centres. In position the
pleural centres may vary from one close to the cerebral
ganglia, as exemplified by the majority of the Monotocardia,
to one close to the pedals as in Haliotis and Trochus, but
in each case the visceral nerves arise direct from these
ganglia. In Pleurotomaria, however, these nerves arise
as we have seen from the connective joining the cerebral with
the pleuro-pedal cords, so that if Bouvier and Fischer are
correct in their localisation of the pleural centres at the
anterior end and on the upper surface of the pleuro-pedal
cords, we should have the very peculiar and absolutely unique
condition of the visceral nerve arising from the cerebro-
pleural connective quite independent of the pleural centre ;

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 257

but, as I have pointed out above, although there is a certain
amount of concentration of nerve-cells in this region, yet it
is mainly in the ventral half of the pleuro-pedal cords, i. e. in
the region of the pedal ganglion. Moreover, while the nerve-
cells are distributed along the whole pleural connective, yet
they are distinctly concentrated to a small extent round the
origin of the visceral nerve sufficient to give it a slightly
orange colour, an appearance which distinctly suggests the
localisation of the pleural centre at this point, a condition
which would be in harmony with what we find in many other
Gastropoda. From these considerations I am forced to the
conclusion that the pleural ganglion, such as figured by
Fischer and Bouvier, does not exist, and that a distinct pleural
ganglion has not yet evolved in Pleurotomaria. Never-
theless we can distinguish a pleural centre in the point of
origin of the visceral nerves, and it is here that a pleural
ganglion would form by an aggregation of nerve-cells, sup-
posing a form were to arise from Pleurotomaria possessed
of distinct ganglia.

The above conclusion is of great importance in considering
the phylogeny of the T'zenioglossa, for with the exception of
Cyclophorus and Ampullaria—two very aberrant archi-
Tzenioglossa, all the remaining members of the great tenio-
glossate group exhibit a condition in which the pleural
ganglia are more nearly approximated to the cerebral ganglia
than to the pleuro-pedal cords. The connection of these
forms with the typical nervous system of the Diotocardia has
been sought in the Cyclophoride and in the Trochida,
but a careful consideration of both these well-known types
of nervous system will show that they are both specialised
along a different line from that characteristic of the Tzenio-
glossa, by a tendency of the pleural ganglion to mount up the
visceral nerve (see Bouvier and Fischer’s diagram, figs. p and
£). On the other hand, in Paludina, the form which, so far
as its nervous system is concerned, appears to me to be the
only true archi-tenioglossan, the pleural ganglion giving
origin to the visceral nerve is little more than a swelling

von. 44, PART 2,—NEW SERIES. R

258 MARTIN F. WOODWARD.

along the course of the posterior connective joining the cere-
bral to the pleuro-pedal cords. This condition is practically
the same as that seen in Pleurotomaria if we imagine a
crowding together of the ganglionic cells at the point of
origin of the visceral nerve, aud is the natural outcome of
that tendency towards a shortening of the nerve-tracts and
concentration of the nerve-cells into ganglia which is so
characteristic of the Gastropoda.

From the condition seen in Paludina it would be very
easy to derive, by a shortening of the cerebro-pleural con-
nective, the condition of all other Tzenioglossa, with the
possible exception of the Cyclophoride and Ampullaria,
which are probably special and independent derivatives of
more specialised Diotocardia.

From the above consideration I conclude that Pleuro-
tomaria in its nervous system, as in some other points in
its anatomy, is the most primitive of existing Diotocardia,
and presents a condition from which that of the majority of
the Tzenioglossa may be derived,—possibly also that of the
other Diotocardia, the form in the latter being attained by a
shortening of the pleuro-pedal connective, thus causing the
pleural centres to be approximate to the pedal ganglia; thus
the condition seen in Haliotis, Trochus, Fissurella, and
Patella would be a derived and not a primitive one.

While it is fairly easy to derive the Monotocardian type of
nervous system, radula, gill, and reproductive system from
the corresponding organs of existing Diotocardia, yet in the
conformation of the kidneys we meet one of the greatest
stumblingblocks in our attempt to derive the former group
from the latter.

All the Diotocardia with the exception of the aberrant
Neritinoid group possess two kidneys, and in the majority
these two organs differ markedly in structure and function.!

1 In Fissurella and Patella, both of which, however, are specialised
forms, the two kidneys, though differing in size and relationship, are both
excretory in function; but the left kidney, as in other Diotocardia, derives its
blood-supply from the auricles.

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 259

The left characteristically forms the papillary sac, and alone
(save in Patella) communicates with the pericardium: this
organ is not truly excretory, but serves as a reserve organ,
and only removes foreign matter by phagocytosis ; thus if in-
soluble powder-hke carmine be injected into the body it is
removed by phagocytes which discharge through the papil-
lary sac (Pelseneer!). This left kidney has alsoa peculiar and
characteristic blood-supply, being directly connected with
the auricle or auricles (Perrier, op. cit.), and thus receives
arterial blood.

The right kidney, on the other hand, is very large, and
characteristically situated between the pericardium and
stomach, being also at times extended below the former into
the anterior part of the hemocele. This kidney, which
receives the venous blood on its way to the gills, is the true
excretory organ, since it alone removes the soluble waste
products. The right kidney further serves to transmit the
genital products, its duct being frequently modified and
elandular in this connection.

In the adult Monotocardian a single kidney alone is present.
The position occupied by this gland is somewhat intermediate
between that of the two seen in the Diotocardia, being placed
in the majority between the pericardium and stomach. It
opens normally (where no secondary ureter is developed)
by a slit-hke orifice between the rectum and gill near
the posterior limit of the mantle-cavity, much as does the
left kidney of the Diotocardia, and it further resembles that
organ in the fact that its cavity communicates with the peri-
cardium ; but at the same time it is a true kidney, and fune-
tions like the right kidney of the Diotocardia. Closely
pressed between this organ and the pericardium is a glandular
mass, often spoken of as the renal gland; the last organ has
the peculiar blood-supply found in the papillary sac of the
Diotocardia, which it further resembles in function (Perrier),

We see, then, that the kidneys of these two great sub-

1 “ Les reins, les glandes génitales, et leurs conduits dans les Mollusques,”
‘ Zool, Anz,,’ Bd. xix, 1896,

260 MARTIN EF. WOODWARD.

divisions of the Streptoneura are very differently developed,
and it is consequently not surprising that a considerable
diversity of opinion has been expressed concerning the
homology of the single kidney of the Monotocardia.

A consideration of the position of the orifice of this organ
in the Monotocardia and its possession of a reno-pericardial
pore at once suggests a comparison with the left kidney or
papillary sac; but on the other hand in the position of the
gland itself, in the nature of its activity, and in the actual
presence of the peculiar renal gland, it more nearly approxi-
mates to the corresponding right organ of the Diotocardia.

Ray Lankester appears to have been the first to seriously
attempt to seek for the homology of the single kidney of the
Monotocardia in the left kidney of the Diotocardia, and this
view, which is now practically universally accepted, has been
further supported by the embryological works of v. Erlanger
on Paludina,! and on comparative grounds by Pelseneer.
Practically the only opponents of this view in recent years
have been Haller? and Perrier.

_ The first view, which is based mainly upon the considera-
tion of the relation of the kidney to the rectum and the pre-
sence of a reno-pericardial pore, receives additional support
from v. Erlanger’s ontogenetic observations. An examina-
tion of these latter shows them to be much less satisfactory
than one would gather from the account given in the average
text-book, since the only trace of the supposed missing
kidney, the adult right, is an angulation and faintest indi-
cation of an outgrowth from the pericardium (ccelom) on the
opposite side to that at which the functional kidney is deve-
loping. This vestigial structure disappears very speedily
without ever attaining any characters which would stamp it
as a kidney, so that the support afforded to this theory by
v. Hrlanger’s ontogenetic researches is very meagre. A
similar unsatisfactory condition is attached to his surmise—

1 « Zur Entwicklung von Paludina vivipara,” ‘ Morph. Jalirb.,’ Bd. xvii,
1891.

2 « Beitrage zur Kenntniss der Niere der Prosobranchia,” ‘ Morph.
Jahrb.,’ Bd. xi, 1886.

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 261

for it is no more—that the genital duct of Paludina arises
from the right secondary ureter, a structure which is not
known to be present in any living mollusc, and whose
existence we have no reason to presuppose.

If, however, we may rely upon these ontogenetic researches,
then the single kidney of the adult Monotocardia would be
the left kidney or papillary sac of the Diotocardia, a view
which is supported by the presence of a reno-pericardial pore.

Against this view we have the position of the kidney in
relation to the stomach and pericardium in the majority of
the Monotocardia (Paludina being an exception), and the
necessity, if we accept it, to seek our ancestral Monoto-
cardian in some very archaic Diotocardian, one in which the
left kidney has not attained the specialised character of a
papillary sac. Moreover the acceptance of this view does
not explain the presence of the peculiar renal gland in the
Monotocardia, which has much the character of, and which
possesses the peculiar vascular relation of the papillary sac.

Perrier, who made a very exhaustive investigation on the
molluscan kidney, believes that the single kidney of the
Monocardia contains representatives of both the kidneys of
the Diotocardia, and he sees in the renal gland of the former
group the representative of the papillary sac of the latter
group. This view, which is an extremely suggestive one,
has not met the consideration which it deserves, most z0o-
logists apparently accepting Hrlanger’s statements on the
development of these organs in Paludina as conclusively
proving that the monotocardian kidney is the papillary sac.

It is, however, possible to approach this subject from ano-
ther standpoint, and to endeavour to reconstruct the stages
which must have occurred in the displacement of the kidney
following upon the disappearance of the right gill and the
consequent displacement of the heart and pericardium, and

1 It might be thought that Fissurella or Patella among living Dioto-
cardians presented us with the condition we want, but these forms are too
obviously specialised in other respects to serve as the ancestors of the Mono-
tocardia.

262 MARTIN F. WOODWARD.

it appears to me if this view be carefully followed out that it
is possible to derive the Monotocardia from such an existing
Diotocardian as Pleurotomaria.

If we examine the condition of these organs in one of the
Azygobranchia we shall find that with the loss of the right
auricle and gill the pericardium becomes displaced to the
left, and consequently the two kidneys approach one another
very nearly, so much so that Haller thought he found a com-
munication between the two. It appears, however, doubtful
if such a connection was present in the forms he examined,
but at the same time it seems extremely probable to me that
such a condition was attained in the early Monotocardia, as
the pericardium shifted further to the left to take up a posi-
tion at the end of the left gill, and the two kidneys conse-
quently came into close contact. Supposing a perforation to
occur in the wall intervening between the two kidneys, a
condition would be attained that would be of the greatest
advantage to the mollusc, as it would enable it to discharge
the secretion of the right kidney through the cavity of the
left, while the old right kidney-duct would now serve to
transmit the genital products unmixed with excreta.

By a diminution in size of the glandular portion of the
papillary sac, and a complete severance of the right kidney
duct as a genital duct, we arrive at the condition of the
Monotocardia, in which we find a kidney situated in the
position of the right kidney, but whose cavity communicates
with the pericardium, and whose aperture suggests that of
the papillary sac; while packed in between this kidney and
the pericardium is the degenerate glandular portion of the
papillary sac forming the renal gland.

I would thus regard with Perrier the single kidney of the
Monotocardia as representing the excretory part of the right
kidney of the Diotocardia plus the cavity, external aperture,
and reno-pericardial pore of the papillary sac; while the
glandular part of the latter structure persists as the renal

gland, and the duct of the right kidney becomes the genital
duct.

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 263

If these conclusions regarding the homology of the kidney
of the Monotocardia have any truth in them, then it would
be quite possible to derive the Monotocardia from a Dioto-
cardian having the type of kidney seen in Pleurotomaria,
Trochus, or Haliotis.

As I have already pointed out, the nervous system of
Pleurotomaria would serve as an excellent starting-point
from which to derive that characteristic of the Tenioglossa,
better by far than that of Trochus, which in the character
of its gills more nearly approaches the Monotocardia. The
general lowly character of Pleurotomaria, especially of its
nervous system and radula, and slight reduction of the right
gill, taken together with its great antiquity, justifies us, I
think, in regarding it as a very primitive form, and one from
which the great monotocardian group may very possibly
have arisen, and possibly also some of the subdivisions of the
Diotocardia.

The following is a brief summary of some of the conclu-
sions at which I have arrived.

SUMMARY.

1. Pleurotomaria is a typical example of a zygobran-
chiate Diotocardian.

2. In the absence of sharply marked specialised regions in
the radula Pleurotomaria Beyrichii and P. Quoyana
are distinctly primitive among the Rhipidoglossa.

3. Inthe reduction of the right gill Pleurotomaria tends
to approach the azygobranchiate Diotocardia.

4, In the uniform distribution of the ganglionic cells
through the connectives, the commissure, and even the large
nerves, and the consequent absence of distinct ganglia,
Pleurotomaria is extremely primitive.

5. In the position of the point of origin of the visceral
loop (roughly halfway between the cerebral and _ pedal
regions) Pleurotomaria approaches the archi-tenioglossate
Paludina and Nassopsis.

6. The pleural ganglion probably arises at the point of

264 MARTIN F. WOODWARD.

origin of the visceral loop by a further concentration of the
ganghionic cells.

7. There is no special concentration of the ganglionic cells
just above the future pedal ganglion, such as Bouvier and
Fischer identify as the pleural ganglion.

8. That in the position of the supporting skeleton of the
gills and the possession of a spiral stomach-ceecum Pleuro-
tomaria shows signs of a common ancestry with the
Cephalopoda.

9. That Perrier is correct in regarding the single kidney
(including the renal gland) of the Monotocardia as represent-
ing both the right and left kidney of the Diotocardia.

10. That Pleurotomaria may be regarded as a form very
closely related to the stock from which the Monotocardia
originated.

July, 1900.

DESCRIPTION OF PLATES 138—16,

Mlustrating Mr. Martin F. Woodward’s paper on ‘'lhe
Anatomy of Pleurotomaria Beyrichii, Hilg.”

List of Reference Letters.

a. a. Anterior aorta. a. d. Afferent branchial vessel. a. p. Pedal
artery. a. 7. &. Auterior lobe of right kidney. 4. d. Bile-duct. 4. g. Buccal
ganglion. d. 2. Buccal nerve. dz. g. Branchial ganglion. cb. ¢c. Cerebral
commissure. cb. g. Cerebral ganglion. cb. p. Cerebro-pedal connective. cd.
pl. Cerebro-pleural connective. ev. Crop. d./.m. Dorsal longitudinal muscle.
d. m. Depressor muscle. e. 4. Efferent branchial vessel. ep. Epipodium.
g',g'. Right and left gills. g. a. Genital aperture. g. d. Genital duct. g. g.
Genital gland. 4. p. Horny buccal papilla. ¢. 2. c. Interlaminar connections.
i. 2. m. Internal longitudinal muscle. ézé. Intestine. 7. 7. m. Infra-radular
membrane. j. Jaw. &. c. Kidney chamber (right). @. Liver. /. aw. Left
auricle. /.c. Labial commissure. /. *. Left kidney (papillary sac). /. 4. @.
Left renal aperture. /. 2. m. Lateral longitudinal muscle. 7. m. Longitu-

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 265

dinal muscle of gill-septum. 7. p., 7. p. Lateral cesophageal pouches. /. pr.
Lateral protractor. J. 7. Lateral retractor. . Mouth. ma. Mantle. m.c.
Mantle cavity. m.g. Mucous (hypobranchial) gland. m.g'.,m.g". Accessory
mucous glands. m. s. Mantle-slit. 2. 7. Nerve-layer in gill-septum. od.
Odontophore. od. c'., od. ce’. Odontophoral cartilages. @.(@sophagus. 0. x.
Optic nerve. op. 7. Opercular lobe, os. Osphradium. o¢. ~. Otocyst nerve.
ovd. Oviduct. p’'., p. First and second pedal nerves. p.7. &. Posterior lobe
of right kiduey. pe. Pericardium. pl. c. Pleural centre. pl. p. Pleuro-pedal
connective. rd. Radula. r. Rectum. 7. az. Right auricle. 7. 4. a. Right
kidney aperture. r. 4. d. Right kidney duct. +. p. c. Reno-pericardial canal.
ry. s. Radular sac. s.7. Supporting rods. sd. ix¢. Subintestinal nerve. s/. d.
Salivary duct. s/. g. Salivary gland. sp. a. Supra-neural artery. sp. ¢.
Spiral cecum. sp. iat. Supra-intestinal nerve. s¢. Stomach. ¢. x. Tentacular
nerve. w. Ureter (right kidney duct). vc. Ventricle. v. 6. Vitreous body.
v. l. m. Ventral longitudinal muscle. v.pr. Ventral protractor. v. 7. Ventral
retractor. v.s. Venous sinuses. v. ¢. m. Ventral transverse muscle.
The figures, unless otherwise stated, are of the natural size.

PLATE 13.
Figs. 1—12. Pleurotomaria Beyrichii.

Fic. 1.—Anterior part of the body viewed from the left side, showing the
bifid left tentacle.

Fic. 2.—Anterior part of the body viewed from above, to show dorsal
surface of foot. This specimen had lost its operculum.

Fic. 3.—Opercular lobe of normal specimen.
Fic. 4.—Operculum.

Fic. 5.—Dorsal wall of mantle cavity, with gills, mucous glands, and
rectum ; viewed from below.

Fic. 6.—Dissection of the anterior part of the body; mantle divided and
reflected, floor of mantle cavity and dorsal surface of head removed, to show
the relations of the anterior viscera.

Fic. 7.—General dissection from the right side, showing the mutual
relations of the alimentary canal, nervous system, heart, and pallial complex.
‘The forward extension of the right kidney is indicated by a brown shade.

Fic. 8.—Side view of the buccal mass, showing the salivary gland with its
duct, the cerebral ganglia, and the buccal nerves. Enlarged.

Fig. 9.—Dissection of the buccal cavity, showing the radula, jaws, and
horny papilla.

Fie. 10.—Dissection of the buccal mass and crop: 1 and 2 the left, 3 and
4 the right cesophageal folds.

Fie. 11.—Diagrammatic transverse section across the crop.

266 MARTIN F. WOODWARD.

Fig. 12.—Dissection of the stomach with its spiral cecum, from above.

Fie. 18.—Dissection of the stomach of Trochus zizyphinus, from
above. Enlarged.

PLATE 14.
P. Beyrichii.

Fie. 14.—Transverse section of the gill and branchial ganglion, showing
the gill-plates in surface view. Somewhat diagrammatic. x about 12.

Fic. 15.—Transverse section of the gill, showing the circulation of the
blood in the gill-plates. Diagram constructed from sections. x about 13.

Fie. 16.—Section across two gill-plates, taken along the line @ 4, Vig. 14.
x about 36.

Fic. 17.—Section parallel to the last, but passing through the dorsal
junction of the gill-plates with the septum. x 170.

Fie. 18.—Horizontal section across the outer margin of a gill-plate. x
500.

Fre. 19.—Section through the branchial ganglion at the origin of the
branchial nerve. The ganglion is slightly contracted away from the connece
tive tissue of the mantle. x 50.

Fic. 20.—Section through the eye. x 60.

Fre. 21.—Anterior portion of the nervous system from the left side of the
body viewed from within, showing the cerebral ganglion and its connections
with the pleuro-pedal cords, together with the origin of the subintestinal
nerve and the principal nerves to the head and side-walls of the anterior
part of the body. The first transverse pedal commissure is seen to be
formed from both the pleural and pedal cords, as also is tlie second pedal
nerve. X 23.

Fic. 22.—The corresponding portion of the nervous system from the right
side of the body, viewed from without. Drawn from a dissection with the
portion of the nerve-cells indicated diagrammatically from microtome
sections. X 6.

PLATE 15.

Fic. 28.—Dissection of the kidneys and pericardium, showing the extent
of the right kidney and its relation to the genital duct; also the great venous
sinus.

Fic. 24.—Dissection showing the relation of the left kidney to the peri
cardium.

Fic. 25.—Semi-diagrammatic representation of the two kidneys, the peri-
cardium, reno-pericardial canal, and genital duct.

Fic. 26.—Schematic representation of the same.

THE ANATOMY OF PLEUROTOMARIA BEYRICHII. 267

Fic. 27.—Diagram of the nervous system viewed from above.

Fic. 28.—Dissection of the head, foot, and mantle, showing the relations
of the nervous system on the right side of the body; also the anterior aorta
and supra-neural and pedal artery.

Fic. 29.—Enlarged view of the brain, cerebro-pleural, and cerebro-pedal
connectives, and the relation of the nerves to the lips and buccal mass.

Fie. 30.—A—H. Light views of the musculature of the buccal mass. 2.
Dorsal view. B. Side view, with the lateral protractor (/. pr.) reflected. C.
Ditto, with the ventral longitudinal muscle (v. /. m.) reflected. D. Ditto, after
the removal of the lateral longitudinal, the radula, and greater part of the
infra-radular membrane (2.7.m.). #. Ditto, after the removal of the main
odontophoral cartilage, exposing the internal longitudinal muscle (7. /. m.).
F. Dorsal dissection, showing the cartilages on the left, and the muscles and
infra-radular membrane on the right. G. Viewed from below, showing the
transverse muscle (v.¢.m.). H. Diagrammatic transverse section,

Fic. 31.—A. Entire otocysts. x 50. B. Isolated otoconia. x 400.

PLATE 16.
Fie. 32.—Half a transverse row of teeth, including the rhachidian. x 50.
Fre. 33.—Rhachian tooth, side view. x 66.
Fic. 34.—First central tooth, side view. x 66.
Fic. 35.—Second central tooth, side view. x 66.
Fie. 36.—Third central, with first and second lamellate tooth. x 66.

Fic. 37.—Four views of a typical lamellate tooth, from the left side of
the radula. 4. Viewed from the outer side. 2B. Normal view. C. Edge on.
D. Viewed from below. x 140.

Fig. 38.—Twenty-fifth tooth. x 60.
Fie, 89.—Twenty-sixth tooth. x 60.
Fic. 40.—Twenty-seventh tooth. x 60.
Fie. 41.—Thirtieth tuoth. x 60.

Fig. 42.—Thirty-fourth tooth. x 60.
Fic. 43.—Thirty-seventh tooth. x 60.
Fie. 44.—Thirty-ninth tooth. x 60.
Fie. 45.—Forty-second tooth. x 120.
Fie. 46.—Forty-third tooth. x 120.
Fie. 47.—Forty-fourth tooth. x 120,
Vig. 48.—Forty-seventh tooth. x 120.
Fic. 49.—Forty-ninth tooth. x 120.
Kie. 50.—Type of tooth between the fiftieth and the sixtieth. x 120.

268 MARTIN F. WOODWARD.

Fie. 51.—Seventy-fourth tooth. x 120.
Fig. 52.—Ninety-seventh tooth. x 120.

Fic. 53.—Last four brush teeth and the seven flabelliform teeth, 1. e. the
101st to the 111th tooth, from the left side.of the radula. x 120.

Fie. 54.—Left jaw, from its inner side. x 9.

DOLICHORHYNCHUS INDICUS, A NEW ACRANIATE, 269

Dolichorhynchus indicus, n. g., n. sp.
A New Acraniate.

By

Arthur Willey.

In the collection of Polycheeta made during the voyages of
H.M.S. “Investigator,” in the Indian Ocean, under the
direction of Dr. Alcock, there is a tube containing several
specimens of an Amphioxus, which on inspection has
proved to be the type of a new sub-genus of the genus Bran-
chiostoma. |

Not one of the specimens appears to be in a condition of
sexual maturity, in spite of the fact that the largest attains a
length of 25°75 mm. The body is elongated, slender, late-
rally compressed, and tapering gradually towards the posterior
end. There are seventy-one,myotomes, and the formula is
42—14—15.

The feature which at once differentiates it from all other
known forms of Amphioxus is the great length of the pra-
oral lobe, close upon 2 mm. measured from the anterior
termination of the neurochord, or equal in length to the first
six myotomes (Fig. 1). The metapleural folds terminate sym-
metrically some distance behind the atriopore on either side
of the ventral fin, a fact which denotes the systematic posi-
tion of the species in the absence of data afforded by the
gonads (Fig. 2). There are about forty-five ventral fin cham-
bers behind the termination of the metapleural folds, and
four or five in front of this point. In the specimen figured
the tentacular cirri (buccal cirri) are mostly concealed within
the vestibule of the mouth, but the ends of several are pro-
jecting from beneath the oral hood in front.

The dorsal fin is well marked, being about one fifth the

270 ARTHUR WILLEY.

total height of the body. In the single specimen cut for the
examination of the ventral fin rays they do not appear as
paired structures, but as massive median expansions of the
hyaline laminar tissue.

It will be noticed that the modification which characterises
this species, namely, the prolongation of the notochord and
cephalic fin in front, is of an exactly opposite nature to that

Anterior region of D. indicus, comprising the oral hood and pre-oral
lobe from the left side. The anterior end of the neurochord with the
eye-spot projects in front of the first myotome.

ne, OH,

Region of the atriopore of D. indicus in ventral view, to show the
symmetrical termination of the metapleural folds on either side of the
ventral.fin behind the V-shaped atriopore.

which distinguishes Asymmetron, where the notochord
and caudal fin extend far behind the posterior limit of the

myotomes. :
Locality.—Off Black Pagoda, Orissa Coast ; 11 fathoms ;

January 15th, 1889.

DOLICHORHYNCHUS INDICUS, A NEW ACRANIATE. 271

The following tabulation of the genera and sub-genera of
Amphioxus will serve to show the systematic position of
the new form.

Genus I.—Branchiostoma, Costa, 1834.
With biserial gonads.
Sub-genus 1.—Amphioxus, Yarrell, 1836.
Type, A. lanceolatus (Pallas).
Sub-genus 2.—Dolichorhynchus, n. g.
Type, D. indicus, n. sp.
Genus II.—Heteropleuron, Kirkaldy, 1895.
With uniserial gonads.
Sub-genus 1.—Paramphioxus, Haeckel, 1893 [in
Semon’s ‘ Forschungsreise,’ Bd. i, p. xiii].
Type, P. bassanus (Giinther).
Sub-genus 2.—Hpigonichthys, W. Peters, 1876.
Type, E. cultellus, Peters.
Sub-genus 3.—Asymmetron, H. A. Andrews, 1893.
Type, A. lucayanum, Andrews.

Of the above sub-genera the three which are most peculiar
in external form, namely, Dolichorhynchus, Epigonich-
thys, and Asymmetron, are monotypic if we consider
Asymmetron caudatum, Willey, 1896, to be merely of
subspecific rank, as would seem to be the case.

fay 8

)
ie
.
i) =
* ‘
& og- ur
’ - z
a> é
.
*
;
v
ce! te
‘
i
x ‘
= e
>
: fs
’
fT 3
i
I ~
s *
y 4
‘
18> @t
t
4
5 i
i , x
‘
r
‘

ee Phe an ene re as
? fe Pat

asrite

Pais oe lind ak eae
Sepa, oh sy tel hon

me oe

J
14

al SP

“Wt

e427" a5 Le

aa Py ai tre es
ae 7 ‘
ie |b: Se ge sent eee
rc
SAP =
a = ¥
1 ee, ’ 7 2 mee
1 _ 4
‘ §
Pa re
j \
i
ay. s aa
; :
rs :
f ar ww oc
‘ ;
7 vs
Pans = Sie AEA
{ Pe
& wes ; 7 A J
. =
f is
" a iP eG: .
‘ if
-
meth Hy
F he re
¥ i
— * 7
5 4 ay 4 ;
r? 2 aS
j
7
® + @ i we
: t
i) = -
i
(3
’
i
'
e
z

HETEROPLEURON HECTORI, NEW ZEALAND LANCELET. 278

Heteropleuron hectori, the New Zealand
Lancelet.

By

W. Blaxliand Benham, D.Sc., I.A., F.Z.S.,
Professor of Biology in the University of Otago.

With Plate 17.

By the kindness of Sir James Hector I have been able to
examine a couple of specimens of an “ Amphioxus”’ that have
been for some years past in the Colonial Museum at
Wellington, N.Z. The specimens are referred to by Captain
Hutton in his ‘Catalogue of the Fishes of New Zealand,’
published by the Colonial Museum and Geological Survey
Department in 1872.

On p. 88 of this catalogue, under the title of ‘ Bran-
chiostoma lanceolatum,” a brief series of measurements are
given, but without any details to enable one to judge that
they are different from the type of the family. At that
period, and for some years later, indeed, even Dr. Giinther
believed that the various specimens from extra-Huropean
seas belonged to this same species, as is evident from the
account of the lancelet in his ‘Study of Fishes’ (1880).
Since Hutton’s reference to them they seem to have been
entirely overlooked by recent writers, for no mention is made
of any New Zealand representative of the family either by
Andrews, by Willey, or by Kirkaldy, in their respective
accounts of this animal.

This New Zealand Lancelet is the type of a new species of

VOL. 44, pART 2,—NEW SERIES. S

274 W. BLAXLAND BENHAM.

the genus Heteropleuron, for which I propose the name
H. hectori.

"he specimens had apparently been preserved in osmic
acid, for they are dark grey ; and though they had received
various slight injuries to the side of the body and to the
fins, they are in sufficiently good condition to enable me to
make out all the important specific characters. One of the
specimens I was permitted to open, doing as little damage to
it as possible ; and afterwards I cleared it in oil of cloves for
more detailed examination of certain parts.

The extreme tips at both ends of the body in each specimen
were more or less injured, but by comparing the two I have
been able to reconstruct, with some confidence, these ends.
The injury affects the tip of the caudal fin, and part of the
rostral, which was folded round the side: in the case of the
latter the outlnes are in the drawings represented by
dotted lines, as there is some doubt as to the exact form of
the fin; while the general curvature of the upper and lower
margins of the caudal fin, preceding the injury, suffices to
show that probably the fin is naturally of the form shown in
the drawing.

Heteropleuron hectori, n.sp., has a length of about two
inches, the actual measurements being 48 mm. and 49 mm.
respectively.

The myotomes number 84 or 85, and the myotome
formula is 55 + 19 (20) + 12,1.¢e. there are 55 myotomes
from the anterior end to the hinder margin of the atriopore,
19 or 20 thence to the posterior margin of the anus, and the
remaining 12 are post-anal. The last two or three are very
small, and the usual difficulties in deciding as to the exact
number of myotomes between atriopore and anus were en-
countered. But from careful examination of the two speci-
mens, I believe that the above formula (which, as Kirkaldy’s
paper ! shows, is subject to individual variation in all species)
1s correct.

The dorsal fin is very shallow over the greater part of its

' « Revision of tle Branchiostomide,” ‘Quart. Journ. Mier. Sci.,’ xxxvii.

HETEROPLEURON HECTORI, NEW ZEALAND LANCELET. 275

extent, though rather higher over its anterior quarter ; while
shortly before the level of the atriopore it again gradually
rises to form the caudal fin. The fin rays cease at the level
of the anus. The rostral fin—as I have remarked—is only
indicated in the drawing with some hesitation ; but it appears
to be rhomboidal in outline, rising suddenly from the dorsal.
The ventral fin, or pre-anal region of the median fin, is short
and low; it contains about a dozen unpaired fin-ray boxes,
which are, however, without fin rays, as is the case, too, in H.
cultellum. At first these boxes are quite distinct, but
after twelve or fourteen complete ones the outlines become
less and less distinct, and soon disappear altogether.

The caudal fin rises quite gradually from the dorsal, at
about the level of the atriopore, without any abrupt angle,
such as is seen in H. bassanum and H. cingalense;
but in this respect it resembles H. cultellum. Its lower
moiety is, however, deeper than the upper, and rises rela-
tively far forwards—about the seventh myotome behind the
atriopore.

It attains its greatest height at about the fourteenth
post-atrioporal myotome, that is some distance anterior to the
anus. Posteriorly the upper and lower margins slope
gradually, and equally and regularly backwards, and appear
to pass in the same curve to a point a short distance behind
the notochord.

On examining the transparent specimen I noted a series
of short brownish chitinoid rods along the ventral base of
the caudal fin, extending outwards from the lower ends of
the muscles for a distance equal to about one fourth the
depth of the fin (fig. 5). Hach rod spreads out slightly near
its distal end, and becomes thinner and more transparent—
losing itself in the tissue of the fin. These rods I traced
backwards to the end of the body, whilst forwards there is
a great gap between them and the ventral fin-ray boxes, the
walls of which have quite a different aspect from the rods,
which do not appear to be optical sections of transverse
walls, but appear to be definite, solid, rod-like structures.

276 W. BLAXLAND BENHAM.

Along the upper base of the caudal are also a few shorter
and less distinct rod-like structures, but I do not feel sure
that in this case they are not the transverse walls of empty
fin-ray boxes ; the animal was lying in an awkward position,
and though there was a gap between the dorsal rod-like
structures and the hindermost distinct fin-ray boxes, yet the
higher surface of the body was here slightly injured and
torn, so that it was not possible to trace the continuity of the
two series.

At first I imagined that we had in this species an interest-
ing vestige of the peculiar caudal fin rays of the early larva,
as figured in Lankester and Willey’s memoir! (pl. xxix,
fig. 1). But on examining the tail of a specimen of H.
bassanum I found that it was unnecessary to explain the
appearance in this temptingly interesting manner; for in
H. bassanum [I find that the fin rays of the ventral fin
are continued past the anus along the under surface of the
body to the antepenultimate myotome (fig. 6) : it is true they
and their “‘ boxes” are smaller here than in the true ventral
fin, anterior to the anus, but they are perfectly distinct right
along the base of the caudal fin.

These ventral rays extend further backwards than do the
dorsal fin rays, which are here only represented by a series
of empty boxes and irregular “lymph-spaces ;” in fact, the
“rods” in H. hectori are the shallow transverse walls of
the empty fin-ray boxes. This post-anal extension of the
fin rays does not appear to have been noted” in any other
member of the group; and in A. lanceolatus Lankester
states definitely that they cease in front of the anus, and I
have examined mounted specimens myself and can confirm
this statement.

In connection with H. bassanum I have to correct
what appears to be an error in Kirkaldy’s diagnosis of this
species, where it is stated (p. 314) that the ventral fin-ray

| Lankester and Willey, ‘“‘ Development of the Atrial Chamber in Amphi-
oxus,” ‘Quart. Journ. Mier. Sei.,’ xxxi.

2 «Contributions to ihe Knowledge of A. lanceolatus,” ‘ Quart. Journ.
Mier. Sci.,’ xxix, p. 373.

HETEROPLEURON HECTORI, NEW ZEALAND LANCELET. 277

chambers contain “ paired fin rays.” In specimens collected
in Port Phillip, and presented to me by Prof. Dendy, I find,
on the contrary, most definitely only a single series of
fin rays in the ventral fin in this species.

This continuity of the fin rays post-anally seems to show the
probability that the ‘ventral fin” is a part of the ‘ median
fin,’ as is suggested by Lankester and Willey (p. 456), in
opposition to the earlier view by the former author that the
ventral fin is the result of fusion of a paired structure
(8, p.373). It becomes more evident that the double fin rays
of A. lanceolatus are secondary, arising perhaps as a result
of splitting of single rays.

The pre-oral hood is much deeper on the right than on the
left side, so that when viewed from the latter aspect both
margins and their cirri are visible (fig. 2), and the vestibule
opens distinctly on the left side of the animal. This is even
better seen in a ventral view (fig. 4), where the right hood is
seen passing obliquely forwards to be continued into the
ventral fin, while the left margin disappears from view as it
curves dorsally upwards.

In the drawings of Heteropleuron and Asymmetron given
by Kirkaldy and Andrews! the vestibule and its opening are
represented as being quite symmetrical. It is of interest
that in this new species a condition is retained which isa
distinct reminiscence of the larval state of affairs. Further,
the cirri on the right side are somewhat shorter than those
on the left. These cirri number nineteen on each side, with
one median ventral, which is shorter than the lowest of the
lateral series; these commence as long filiform structures,
and gradually diminish in length as the series approaches
the dorsal termination.

The gonads, present only on the right side, appear to be
about eighteen in number, but as they dropped away from
the body-wall as it was turned aside there may have been a
few more.

1 «An Undescribed Acraniate—Assymmetron lucayanum,” ‘Stud,
Biol. Lab., J. H. Univ.,’ v,

278 W. BLAXLAND BENHAM,

It is scarcely necessary to state that the right metapleur
is continuous with the ventral fin, as that is one of the
characters of the genus.

Locality.—EHast coast of the North Island of New
Zealand.

From this brief but sufficient survey of the external
characters of the New Zealand species, it will be seen that it
differs from each of the previously known species of Hetero-
pleuron.

In form this new species seems, from the drawings avail-
able, to be somewhat stouter than other species, while the
tapering anteriorly is comparatively sudden (fig. 1). The
greatest height, measured from the upper edge of the dorsal
fin to the lower margin of the metapleure, is 5 mm., which is
about one tenth of the total length. These measurements
are, of course, lable to variation according to the condition
of preservation ; my specimens, however, are not shrunk in
the way that occurs when living specimens are plunged into
strong alcohol, but have retained a form similar to that of
A. lanceolatus preserved in picric acid, and though some-
what soft are not in any way “rotten.” Therefore I think
the form given in the plate, which is drawn to scale, is as
nearly as possible true to life. It must be borne in mind,
however, that the pre-oral hood is retracted, while in the
most reliable drawing of Amphioxus, viz. that given by Pro-
fessor Lankester, this hood hangs down as a nearly semi-
circular membrane. The left metapleural ridge can be
traced in my specimens right forwards above the hood,
which has shrunk upwards below it (see fig. 2).

The distance between the atriopore and the anus is one
sixth of the total length, and the distance of the anus from
the end of the body is one sixteenth of the total.

In size it exceeds the largest, which is H. bassanum,

1 According to a verbal communication by Sir James Hector, these two
specimens were collected at Awanui, just south of the East Cape; whilst
Hutton in the ‘Catalogue’ gives as the locality ‘* Poverty Bay,” which is a
little further south,

HETEROPLEURON HECTORI, NEW ZEALAND LANCELET. 279

with a length of 43 mm.; in total number of myotomes, too,
it exceeds any Amphioxid hitherto described—the nearest
approach being seventy-nine in “ A. elongatum™” of San-
deval, and seventy-eight in H. bassanum.

It is perhaps worth noting that this excess is chiefly due
to an increase in the number of pre-atrioporal segments; for
the post-anal segments in other species are from eight in H.
cingalense to fourteen or even seventeen in H. bassanum,
with an interporal number of ten to seventeen in the
various species.

In regard to the caudal fin, there is equally sufficient evi-
dence of distinctness, for whereas in H. bassanum it com-
mences behind the anus, in H. cingalense it arises
immediately in front of it, and is very short; while in H.
cultellum, though it begins at a point further forward, yet
this point is some distance relatively behind its point of
origin in H. hectori, while the position of the greatest
depth is behind the anus in all three, instead of being ante-
rior to it, as in the present species.

The sea surrounding Australia and the neighbouring
islands is evidently rich in species of Branchiostomide, for
already four species belonging to each of the three known
genera have been recorded,—viz. Amphioxus belcheri,
Gray, from Torres Straits, as well as from the coast of
Borneo; Heteropleuron cultellum, Peters, from Torres
Straits and further east coast of Australia; Asymmetron
caudatum, Willey, from the Louisiade Archipelago,! due
east of the Torres Straits; and H. bassanum, Giinther,
from the south of Australia, from Bass’s Straits. The
present species thus makes the fifth in these southern seas ;
its habitat is two thousand miles or more distant from each
of these localities. These seas appear to be the home of
the asymmetrical species, and one is tempted to think that
these may be the more primitive of the family, especially as
my species presents one, perhaps two survivals apparently
of a larval condition.

1 “Quart. Journ. Mier. Sei.,’ vol. xxxix, p. 210,

280 W. BLAXLAND BENHAM.

I regret that I have been unable to make a fuller exami-
nation of this interesting species, and look forward to obtain-
ing living specimens; but it appeared worth while to rescue
these individuals from their obscurity when it was found
that they differed from those of the neighbouring seas,

DUNEDIN;
August 25th, 1900.

EXPLANATION OF PLATE 17,

Illustrating Professor Blaxland Benham’s paper on
“ Heteropleuron hectori, the New Zealand Lancelet.”

Fic. 1.—Heteropleuron hectori, n. sp. (x 2). a. Metapleur. 4.
Floor of atrium. c¢. Atriopore. d. Anus.

Fig. 2.—Heteropleuron hectori. Side view of the anterior end (x
8). a. Left metapleur. J. Floor of the atrium. e¢. Right metapleur. The
outline of the rostral fin is dotted, as there is some doubt as to its true shape
and size.

Fic. 3.—Hinder end of thesame (x 8). a@. Left metapleur. &. Atriopore.
e. Ventral fin. d. Anus.

Fre. 4.—Ventral view of the anterior end of the same, showing the vesti-
bule opening distinctly on the left side of the animal. a. Left metapleur, J.
Floor of atrium. c. Right metapleur. d. Ventral fin.

Fie. 5.—A portion of the ventral base of the caudal fin of a transparent
specimen, seen under a low power. a. Muscles, 4. Fin. ¢. Rod-like
structures, the walls of empty fin-ray boxes.

Fie. 6.—View of the tail of H. bassanum, cleared in clove oil, showing
the post-anal continuation of the ventral fin rays. The myotome marked 3
is the antepenultimate segment, beyond which the fin rays are absent. a.
Ventral caudal expansion of median fin, the edge of which is folded. 6. Fin
rays. ¢. Dorsal portion of the caudal fin, along the base of which empty
“‘hoxes ” and lymph-spaces are seen.

Fic. 7.—A view of a portion of the per-anal part of the ventral fin of H.
bassanum. The lower part of the body was cut off, cleared and mounted ;
it is seen from below, and shows a single series of fin rays.

PARASITES FOUND IN ECHINUS ESCULENTUS, L. 281

On some Parasites found in Echinus
esculentus, L.

By
Arthur E. Shipley, M.A.,

Fellow and Tutor of Christ’s College, Cambridge, and Lecturer in the
Advanced Morphology of the Invertebrata in the University.

With Plate 18.

I. TURBELLARIA.

nu interesting parasite, Syndesmus, was first observed
by Patrick Geddes,! who, however, beyond drawing attention
to its partly Turbellarian, partly Trematode characters, made
no attempt to describe it. He found it in the perivisceral
cavity of Echinus esculentus, L.

W. A. Silliman? in the following year gave a description
of the external features of the parasite, and suggested the
generic name Syndesmus. He found it living on a large
green nematode, which seemed to him to bea parasite of KH.
esculentus taken at Roscoff.

Five years later Ph. Frangois* in the same Proceedings
records the occurrence of this animal in the intestine of
Strongylocentrotus lividus, Lam., and of EH. acutus,
Lam., at Banyuls. His description, however, differs mate-
rially from that of Silliman,—so much so, indeed, that Braun*

1 “Arch. Zool.,’ exp. 1, ser. vili, 1879-80, p. 488.
2 ¢C. R. Ac. Sci,’ xciii, 1881, p. 1087.

3 T[bid., cili, 1886, p. 752.

4 «Central. Bakter.,”’ v, 1889, p. 41.

282 ARTHUR E. SHIPLEY.

remarks that one might think that the two authors were
observing different species. Francois suggests the specific
name echinorum.

The animal is again recorded in 1892-3 by L. Cuénot,}
who draws attention to the fact that it corresponds well with
the character of the family Vorticipa of von Graff, and
indeed to his sub-family Vorricina Parasirica, which
includesanother parasitic genus of Turbellaria, Anoplodium,
also found in echinoderms. Silliman, Frangois, and Cuénot
all promise full accounts with figures of the anatomy, but as
far as I can find out these have not yet appeared.

Last autumn my friend Mr. W. F. Cooper brought me
eleven specimens of this parasite which he had found, ten
in the alimentary canal, and one lying on the genital gland
in the perivisceral cavity of a specimen of Hchinus escu-
lentus, L., that he was dissecting at the Marine Biological
Laboratory at Plymouth. This winter I have worked out the
anatomy of the form and made numerous drawings. After
I had completed the work I discovered that Professor
Russo? had been over verv much the same ground ; he has,
I believe, anticipated me in many details, but as the parasite
is very interesting, and is now recorded from the British
area, and as the periodical in which Professor Rosso’s full
paper appears is very inaccessible—I have not been able to
find a copy in any of our libraries—I have thought it not
useless to publish the following general account of the
anatomy of Syndesmus echinorum, Frang.

Anatomy: External Features.—The eleven specimens
vary from 1 mm. to 2 mm. in length, and their greatest
breadth is one half of their length. ‘These dimensions are
considerably less than those recorded by Frangois. His
specimens were nearly twice this size. In shape the animals
are leaf-like, and havea tendency to be hollowed out ventrally.
The anterior end is more rounded than the posterior, but in

1 * Rey. biol. Nord France,’ v, 1892-3, p. 1.

2 § Rie, Labor. anat. Roma,’ v, 1895, abstracted in ‘ Monit, Zool. ital,,’ vii,
1896, p. 6.

PARASITES FOUND IN ECHINUS ESCULENTUS, L. 283

some cases the latter is produced into a small papilla caused
by the evagination of the penis.

The mouth leads into a well-marked sucker-like pharynx.
It is situated on the ventral surface in the middle line, about
one eighth of the body-length from the anterior end of the
animal. The opening of the vas deferens, the vagina, and
the uterus are all at the posterior end of the body, and open
by a common pore (figs. 4 and 5).

There are no tentacles, or papille, or hooks, or spines, and
as far as I could observe no skin-glands.

Histology.—The whole body is covered with a thin but
distinct cuticle of uniform thickness (fig. 6). This is continued
into the mouth and genital openings, but soon disappears.
Externally this cuticle bears numerous small processes, very
minute, but sufficient to give a rough appearance to the out-
side surface under a high power of the microscope. These
are almost certainly the cilia described by all authors who
have observed the animal alive.

The cuticle is secreted by a single layer of ectoderm cells
with large, clear, spherical nuclei (fig. 6). In some sections
these ectoderm cells showed fairly definite cell limits, and in
that case each cell was about as broad as it was long; in
other cases the limits of the cell could only be guessed by
observing the nuclei placed at regular intervals. In all the
specimens I cut the ectoderm of the dorsal and ventral
surfaces had separated from the subjacent tissues, leaving a
considerable space, but it had retained its normal position
along the edges of the animal.

Beneath the layer of ectoderm cells is a basement mem-
brane, which seems, however, to belong rather to the under-
lying parenchymatous cells than to the ectoderm; it gives a
smooth and clearly defined outline to the body where the
ectoderm has broken loose from it.

The muscular system described by some authors was not
visible in iny sections.

The parenchyma of the body presented different appearances
in accordance with the different state of preservation of the

284. ARTHUR E. SHIPLEY.

specimens. It consists in the more typical form of a
number of large, more or less cubical ceils, full of a densely
granular protoplasm. ‘The celis take every variety of shape,
owing to mutual pressure and the various strains and stresses
which affect them. In life their outline cannot remain
constant for any length of time. In the more poorly pre-
served specimens the granular protoplasm had shrunk away
from the firmer exterior of all but one surface of the cell,
leaving a large but irregular vacuole. The firm external
part of the cell, with from time to time patches of contracted
protoplasm adhering to it, gives the parenchyma the appear-
ance of a network with considerable vacuoles. When this
firmer exterior is a little more emphasised it forms the base-
ment membrane, which underlies the ectoderm and surrounds
the various parts of the reproductive system ; it is, however,
very noteworthy that no such basement membrane surrounds
the alimentary canal or intestine.

The Digestive System.—The mouth is ventral, in the
middle line and situated about the distance of one eighth or
one tenth of the body-length from the anterior end of the
body (fig. 4). It leads by a very short passage, lined by
cuticle, and bearing as far as I could make out no glands of
any sort, into a spherical pharynx. ‘This organ is of the
type found in Vortex or Plagiostoma. The minute lumen
is lined by a uniform cuticle, and the bulk of the thick
wall is built up of radial muscle-fibres, among which a few
large nuclei stand out in stained sections (fig. 7). From the
inner end of the pharynx a very short cesophagus provided
with numerous glands—the so-called salivary glands—leads
to the digestive sac.

The stomach or intestine, or, as I prefer to call it, the
digestive sac, is a rod-like organ extending along the middle
line of the animal, and so close to the dorsal surface that
there is practically none of the parenchymatous tissue which
serves as a packing for the various organs of the body be-
tween it and the epidermis (fig. 2). The axis of the lumen of
the mouth, pharynx, and cesophagus is a dorso-ventral one, but

PARASITES FOUND IN ECHINUS ESCULENTUS, L. 285

where the last-named passage joins the digestive sac it forms
a right angle with the lumen of the alimentary canal. Ante-
riorly the digestive sac extends a little way in front of the
level of the entrance of the cesophagus, and when looking
through a series of transverse sections it comes into view
before any trace of the pharynx makes its appearance.
Posteriorly the digestive sac extends to near the end of the
body, coming to an end at a distance of perhaps one tenth or
one twelfth of the total body-length from the end.

The digestive sac is lined by a very definite layer which
is in the main a plasmodium, though it shows here and there
traces of division into cell areas (fig. 6). The lmit of the
tube is clearly defined, but the basement membrane is very
thin, and in places the outer edges of the endoderm plas-
modium rests against the packing cells of the body. Inter-
nally the lining is produced into many apparently amoeboid
processes or pseudopodia, which project loosely into the
cavity, and the free ends of which often are cut off and lie as
isolated pieces of stained protoplasm in the sections. It is
along the inner boundary from which these processes arise
that evidence of cell structure is most evident, since the
chinks between the bases of the pseudopodia are continued
by fine lines, which pass a little way into plasmodium,
dividing it as 1t were into cell areas.

Throughout the plasmodium deeply staining nuclei are
distributed, and numerous vacuoles are scattered; some
apparently contain drops of fluid, probably oil or fat; others
contain uniformly staining spheres of unknown nature.

I have not been able to find any trace of a secretory
apparatus; neither canals nor pore could be made out in any
of my sections.

Nervous System.—The nervous system consists of a
well-marked ganglion, situated just anterior to the mouth ; it
is somewhat rectangular in outline, and a nerve is given off
from each angle. The anterior pair of nerves soon disappear ;
the posterior, which bend backwards, are stained, but I failed
to follow them very far down the body. In one,stained

286 ARTHUR E. SHIPLEY.

specimen a median nerve seemed to leave the ganglion between
the anterior two nerves. It is probable that this nerve
divides into two branches.

Reproductive System.—The external opening of the
vas deferens and of the uterus lie side by side, close to one
another, at the posterior end of the body.

The male reproductive organs consist of paired branching
testes. Hach half presents some ten or twelve twigs lying on
either side of the anterior end of the digestive sac, and ex-
tending in front of the mouth (fig. 5). These twigs fill up most
of the sides of the body, from in front of the mouth to the region
of the yolk-glands. The several branches of each half of the
testis unite and open intoa pair of tubes, which may be termed
the vasa efferentia. ‘hese soon fall into one another, and
form a long median and anteriorly much-coiled tube. This
vas deferens makes a well-marked loop forward to the left of
the mouth (fig. 4). In its hindermost part, however, the
tube is straight, and is provided with thick muscular walls
lined with a cuticle.

Russo describes a complicated penis. I have not been able
to follow all his details, but there is undoubtedly a protrusible
intromittent organ present.

The histology of the male reproductive organs presents little
worthy of notice. The branches of the testis were outlined
by a very thin basement membrane, but beyond this they
presented no special investiture. heir contents were cells
of some size with large nuclei and conspicuous chromatin.
Near the end next the ducts, bundles of tailed spermatozoa
are to be seen. The vas deferens is a long and much-coiled
duct, so that, as a rule, portions of it are seen several times in
any one section. It has a smooth internal wall or cuticle,
and apparently a thin muscular lining; at the posterior end
the wall of the tube is very much thickened by a stout
muscular sheath, and this portion is protrusible, and indeed in
one specimen is protruded as a penis.

The ovary, like the testis, is double and branched ; each half
is compared by Francois to a hand with the fingers extended.

PARASITES FOUND IN ECHINUS ESCULENTUS, L. 287

Each branch of the ovary contains, as a rule, a single row of
large angular ova, with very large spherical nuclei. The ova
are mostly bounded by flat sides. They show some tendency
to squeeze one another out of the single row, and when this is
the case the row appears double. The ova at the end of each
branch next the outlet are markedly bigger and more rounded
than those near the top, where they are very small, and
apparently it is here that they arise.

The coating of the ovary is thin, and it is continued in
each side into a short duct which unites with its fellow, and
at or near the point of union the ducts of the yolk-glands
open.

The yolk-glands are large and branching ; they he on each
side of the body between the testes and the ovary,—on the
whole, more dorsal than the ovaries (fig. 2). The tissue of the
yolk-glands is dense, and stains deeply near the tips of the
branches; but it becomes much vacuolated and stains less
deeply near its opening, which leads into the duct of the yolk-
glands,

The two oviducts of each side and the two ducts of the
yolk-glands open into acommon chamber of somewhat angular
shape.

The shell-glands are paired, and occupy much of the
posterior end of the body. The numerous little glands which
constitute the organ are unicellular and generally somewhat
angular in shape, packed away as they are amongst the
interstices of the parenchyma. Each is crowded with fine
granules, and leads by a very delicate duct, which, converging
towards each side of the uterus, does not open into the yolk-
gland ovary complex, but as far as I can make out into the
uterus.

In each specimen the uterus contained a beautiful golden
egg, oval in outline and continued posteriorly into a long
filament. This filament is bent and curved so as to form a
tangled skein in the centre of the body ; gradually it becomes
finer, and its end, which is of extreme tenuity, lies in the
neighbourhood of the external opening of the uterus. The

288 ARTHUR E. SHIPLEY.

contents of the golden egg-shell stained uniformly and deeply,
so that no nucleus could be detected. In bulk the egg in
the egg-shell surpassed the ripe ova in the yolk-gland ovary
complex by some five or six times; this is almost certainly due
to the addition of the yolk. On the other hand, the golden
case may have been an egg capsule, and contained more than
one egg. I rather gather that Russo takes this view.

Il. Nematopa.

In 1854 Dr. Leydig! described some nematodes belonging
to the genus Oncholaimus which he had found in the
alimentary canal of Echinus esculentus. The parasites
were 4 mm. long, thread-like and pointed at both extremities.
The oral cavity was provided with a certain toothed and ridged
armature in the shape of thickenings of the cuticle prolonged
from the firm cuticle covering the body. The cesophagus
was long, and posteriorly enlarged, but nowhere did it form
a bulb. The intestine ran in a straight line to the anus at
the base of the tail, and had a brown colour due to pigmented
granules which crowded the cells. The ovary had an anterior
and a posterior branch, and each branch terminated in a line
which doubles back and ends near the genital opening about
the middle of the body. The ripe egg was of considerable size
and of oval shape. The oviducts united to form a sharply
defined vagina. The cuticle had longitudinal striations.

Dr. Leydig suggested the name Oncholaimus echini
for this parasite.

The only other nematode that I find mentioned as coming
from within the body of Echinus esculentus is the large
green nematode of Silliman, which presumably—it 1s not quite
certain—came out of one of these creatures taken at Roscoff.

A year or two ago Mr. A. J. Smith, assistant at the
Marine Biological Laboratory at Plymouth, found two or
three very long nematodes in the perivisceral cavity of an H.
esculentus at Plymouth, which he sent to me for investiga-

1 «Miuller’s Archiv,’ Jalrgang 1854, p. 291.

PARASITES FOUND IN ECHINUS ESCULENTUS, L. 289

tion. Unfortunately I did not undertake this at once, and
when I came to look at the specimens a short time ago I
found that, owing to a piece of iron being in the bottle in
which they were preserved, the nematodes had become coated
with rust, and in freeing it from rust their structure was so
injured that nothing of their histology could be made out.
The larger worm had further been injured in extracting it
from the shell of the host.

Beyond the facts that the longest nematode is some 46 cm.
in length, a little under 1 mm. in average diameter, and the
smaller specimens were some 6 cm. in length; that both ends
of the animals taper, but more particularly the anterior ; that
the posterior end is recurved, as is so usual amongst male
nematodes ; and that the alimentary canal was visible through
the skin in the line specimens as an opaque strand, I can say
nothing. Enough is not known to warrant the suggestion of
any specific characters, and I mention the parasite only in the
hope that it may attract attention to it and lead to its being
found again, carefully preserved, and investigated.

ZOOLOGICAL LABORATORY, CAMBRIDGE ;
April, 1900.

KXPLANATION OF PLATE 18,

Illustrating Mr. Arthur E. Shipley’s paper “On some
Parasites found in Echinus esculeutus, L.”

List or ABBREVIATIONS.

e. Cuticle. c.g. Cerebral ganglion. egg. Egg in uterus. ep. Mpidermis.
g.d. Genital duct. y.p. Genital pore. yp. Ameeboid plasmodium lining
intestine. @. Intestine. om, Mouth. ov. Ovary. par. Parenchyma. se.
Sucker. s.g/. Salivary gland. sh. gl. Shell-gland. ¢. Testis. v. Vagina.
v.d. Vas deferens. y.g. Yolk-gland.

Fic. 1.—A longitudinal horizontal section through Syndesmus echi-
norum near the ventral surface.

VOL. 44, PART 2.—NEW SERIKS. a

290 ARTHUR KE. SHIPLEY.

Fic. 2.—A transverse section through about the centre of the body of
Syndesmus echinorum.

Fic, 3.—A longitudinal horizontal section through Syndesmus echino-
rum. This section is cut in a somewhat oblique plane, the anterior end being
nearer the dorsal surface, the posterior nearer the ventral surface.

Fic. 4.—A sketch of a stained and mounted specimen of Syndesmus
echinorum. The parts shown can be identified by a reference to Fig. 5.

Fie. 5.—A diagram to explain the anatomy of Syndesmus echinorum.

Fie. 6.—A small portion of the epidermis and intestinal wall of Syn-
desmus echinorum, very highly magnified to show the nature of the
plasmodium lining the alimentary canal.

Fie. 7.—A transverse section of Syndesmus echinorum through the
region of the mouth and pharynx. ‘To the left the anterior loop of the vas
deferens is shown.

Fic. 8.—Large nematode extracted from the ceelom of Echinus escu-
lentus. x 1.

THE SCOTTISH SILURIAN SCORPION. 291

The Scottish Silurian Scorpion.
By
R. I. Pocock.

With Plate 19.

1. Inrropucrory REMARKS.

Our knowledge of the existence of scorpions in marine beds
of Upper Silurian age dates from the publication of an an-
nouncement to this effect in the ‘Comptes rendus de
Académie des Sciences,’ Paris, in December, 1884, wherein
Professor Lindstr6m and Dr. Thorell gave an account of the
discovery of the well-preserved remains of a fossil scorpion at
Gotland in Sweden, proposing for the new form the name
Paleophonus nuncius. This important find in paleon-
tology attracted wide-spread interest, and was discussed in
various journals, scientific and popular. In 1885 it was fol-
lowed by an exhaustive memoir on the fossil by Lindstrém
and Thorell (‘Kongl. Sv. Vet.-Akad. Handl., xxi, No. 9,
1885). Prior to the appearance of this memoir an article
entitled “Ancient Air Breathers,” by Mr. B. N. Peach, was
printed in ‘Nature’ (vol. xxxi, pp. 295—298, 1885). In this a
preliminary description was given of a second Upper Silurian
scorpion, which had been unearthed in the summer of 1883
at Lesmahago, in Lanarkshire, and formed part of the rich
collection of fossils belonging to Dr. Hunter. ‘he value of
this second specimen was enhanced by the circumstance that
it fortunately hes with its ventral surface exposed, and is thus
the complement, as it were, of the Gotland fossil, of which

992 R. I. POCOOK.

the dorsal surface, at all events, of the anterior half of the
body is uppermost.

For those who hold that the terrestrial Arachnids are
descended from marine ancestors allied to Limulus and the
Kurypterida, and recognise genetic affinity instead of “ for-
tuitous coincidence ” and “convergence” in the many deep-
seated structural resemblances between the two groups, these
archaic scorpions have, since their discovery, been vested
with a peculiar interest, largely in view of the possibility of
their supplying fresh evidence in support of this relationship.
Little in this direction was yielded by the memoir on the
Gotland scorpion; and Peach’s description of the Scotch
specimen, although containing many important anatomical
observations, was by no means exhaustive, and the figure that
accompanied it not all that could be desired. Hence it has
“for many years been felt that a complete and properly illus-
trated account of this unique fossil would make a valuable
addition to zoological literature.

In July of last year Prof. Ray Lankester wrote for the loan
of the specimen to the authorities of the Kilmarnock Museum,
where it has been preserved since the death of Dr. Hunter.
The authorities not only kindly and promptly acceded to the
request, but most generously permitted the specimen to be
kept for three months at the Natural History Museum. I
eladly avail myself of this opportunity to express my sincere
thanks to Professor Lankester for placing the specimen in
wy hands for investigation. Iam also indebted to Miss G,
M. Woodward for the trouble and time she devoted to the
lithograph, her skill and experience in interpreting fossils
being most helpful in the present instance.

2. DESCRIPTION OF THE SPECIMEN.

So far as the disposition of the various members is con-
cerned, my restoration agrees with that of Mr. Peach in most
particulars. I think, however, that the second leg on the
right side lies distally across the anterior portion of the

THE SCOTTISH SILURIAN SCORPION. 293

“hand ” of the chela, and not across its posterior portion as
shown in the figure in ‘Nature.’ One or two other particu-
lars in which I differ from him are referred to in the following
pages.

The specimen gives the following measurements in milli-
metres :—Total length on stone 32:5, actual total length when
extended 35:5, trunk 16°5, tail 19.

The Gotland specimen is considerably larger, measuring
62 mm. in total length, the tail being at least 26 mm.

Prosoma.—Owing to the outward displacement of the
chele the anterior portion of the carapace is visible between
the basal segments of these appendages, and in front of those
of the first pair of legs. Its surface is thickly granular, its
anterior border lightly concave, as is the Gotland specimen,
and its antero-lateral angles subquadrate.

Kyes.—In the Gotland specimen no trace of eyes, either
median or lateral, is discernible, though the median ocular
tubercle of recent scorpions is represented by a relatively
large and longitudinally oval elevation, situated in the ante-
rior third of the carapace, and separated from its anterior
edge by a space equalling about one half the length of the
elevation. Judging from the figure, this tubercle is pre-
served in its entirety; hence there is no reason to doubt
that if eyes had been borne upon it, some trace of them at
least would have been preserved.

In the Scotch specimen also there is no sign of the lateral
eyes. If, however, as is possible, these organs existed, and
were placed behind the level of the median eyes, as is the
case in the normal Pedipalpi, and, as is alleged, in the Car-
boniferous Anthracoscorpii, they would be concealed from
view beneath the basal segments of the anterior legs, which
on each side overlie that portion of the carapace immediately
behind the median eyes. The median eyes are very distinctly
represented by a pair of elliptical impressions situated close
together, one on each side of the middle line, and scarcely
more than their own long diameter from the anterior border
of the carapace. There is no evidence that these eyes were

294, Ree SeOCOCK,.

elevated upon a tubercle. If, indeed, such a tubercle existed
as is exhibited in the Gotland specimen, the eyes must have
been situated on its extreme anterior border. The presence
of these median eyes, and the probable absence of the tubercle,
are two important structural differences to distinguish the
Scotch specimen from the Swedish.

Appendages.—The six pairs of prosomatic appendages
(i—vi, Pl. 19) are preserved in a state of greater or less com-
pleteness, those on the left side being on the whole more
clearly defined than those on the right.

The chelicere or mandibles are, as in the Gotland
specimen, very large as compared with those of recent scor-
pions. The left chelicera, crushed out of shape and position,
shows no recognisable feature but a portion of the immove-
able digit. The right, on the contrary, is well preserved
and occupies its normal position, projecting straight for-
wards from the fore-part of the prosoma. The immoveable
digit is slender, pointed, and nearly straight ; the moveable is
equally slender and pointed, but is lightly curved and armed
in the middle of its lower edge with a single tubercular tooth.

It is noticeable that the digits of the chelicera are thinner,
and overlap at the apex to a much greater extent than in the
Gotland fossil.

Owing to the distortion and displacement of the left che-
licera a portion of the matrix is displayed between the bases
of the two appendages just in front of the middle line of the
anterior border of the carapace. Presumably it is this por-
tion of matrix which Mr. Peach describes—I think errone-
ously—as “a fleshy labrum (camerostome) between the bases
of the cheliceree.”

Chelz.—As in the Gotland specimen, these appendages
do not appear to differ in any essential respects from those
of recent scorpions. Their basal segments are too badly
preserved for delineation—a particularly regrettable circum-
stance in view of the fact that in the Gotland specimen they
are concealed from view. Hence it is impossible to surmise
whether they took a greater, less, or an equal share in masti-

THE SCOTTISH SILURIAN SCORPION. 295

cation as compared with those of existing forms. The
second segments project on each side of the antero-lateral
angles of the carapace, and are granularly sculptured. The
anterior surface of the third segment is apparently normally
crested above and below, and the fourth segment of the left
side shows traces of the basal prominence so noticeable in
living species. Granules are observable along the anterior
side of both these segments. The fifth segment (hand) of
the left side differs in shape from that of the right, being
more oval in form, with its posterior border in approximately
the same straight line as that of the distal segment, the
bulge being confined to the anterior surface as in the Got-
land specimen and recent species. On the right side the
hand is unusually globular, its posterior surface, probably
owing to crushing, being abnormally swollen. The fingers
are thinner, more taper, and straighter than in the Gotland
specimen and recent scorpions. No distinct joint between
the finger and hand is discernible, although presumably it
is the under side of the hand and of the moveable finger
that is exposed to view, both on the right and left sides.
It is possible that the shallow median longitudinal groove
observable on the finger of the right chela represents the
line along which the two fingers meet when closed. The
finger of the opposite side is similarly marked with a fine
sculptured ridge.

Legs.—So far as can be ascertained the legs resemble
those of the Gotland specimen in length, strength, and seg-
mentation. As in other scorpions, and typically in all orders
of Arachnida, they increase in length from before back-
wards, the fourth pair being nearly half as long again as the
first. They consist, moreover, of what is doubtless the primi-
tive number of segments—namely, seven. Primitiveness of
segmentation is also shown by the subequality in length of
the individual segments—a character which, in conjunction
with the sharply pointed, practically clawless terminal
segment, serves to distinguish the legs of Paleophonus
from those of all other scorpions, living or fossil. I say

296 Rai. LPOCOCK.

practically clawless because Thorell detected a minute claw-
like structure at the tip of the seventh (tarsal) segment in

Fie. 1.—Restoration of Paleophonus nuncius.
Dorsal view (after Thorell).

the Gotland specimen. Although no trace of such a struc-
ture was found in the Scotch fossil, no great value must be

THE SCOTTISH SILURIAN SCORPION. 297

attached to its apparent absence, in view of the chances
against the preservation of an organ so delicate.

Nor was I able to detect a sign of the presence on the
fifth segment of any of the legs of that spur so clearly shown
on the first, second, and third pairs in the Gotland fossil, and
described and figured by Thorell (see cut, p. 296). The
interest invested in this spur depends upon the probability
of its direct homology with the so-called “ tibial spur” found
upon the arthrodial membrane at the distal end of the fifth
segment in some recent Buthoid scorpions. Certain genera
of this family (e.g. Buthus, Lychas) possess it upon the
third and fourth legs, one alone (Babycurus) retaining it
only on the fourth leg. Assuming that the spurs in the
genera just mentioned are homologous to those found in the
Swedish Paleophonus, their presence upon the third and
fourth, or upon the fourth leg in the former, and upon the
first, second, and third legs in the latter, suggests that
scorpions primitively possessed them upon all four legs.
In that case the absence of the spur from the fourth leg
in the type of Paleophonus nuncius may be a natural
characteristic of the species, or may be due to a mere
accident of preservation. The same may be said of the
apparent total absence of this spur from the legs of the
Scotch specimen.

There is, however, a still deeper interest attached to this
spur, on account of its apparent presence upon the fourth
leg (sixth prosomatic appendage) of Limulus. The first and
second appendages of this animal agree in structure and in
the number of segments with those of scorpions, the former
consisting of three and the latter of six segments. But the
third, fourth, and fifth appendages of Limulus also consist
apparently of six segments, resembling in all particulars those
of the second pair. In the scorpions, on the contrary, these
appendages, as well as the sixth pair, consist of seven seg-
ments, the distal being furnished with a pair of moveable claws.
Careful examination of these appendages in Limulus, how-
ever, shows that the fourth segment is encircled in its basal

298 2 tls POOOCK:.

half with a sutural impression, which represents, I believe,
the line of union between two segments, the portion on the
proximal side of the line being the fourth, that on the distal
side the fifth segment of the appendage. If this interpre-
tation be correct there is the same number of segments in
these appendages in both Limulus and the scorpions. Now
in the fourth leg of Limulus (exceptin L. rotundicauda)
the fifth segment, according to this new method of enumera-
tion, is furnished beneath distally with a spur like those
described above in the scorpions. Again, at the extremity
of the sixth segment in Limulus there are four moveable
lobate sclerites, which spread out like the fingers of a hand
when the leg is plunged into the mud. At the extremity of
the sixth segment in the scorpion’s leg, or rather on the
arthrodial membrane between it and the seventh, there are
either one or two “ pedal” spurs, which represent, I suggest,
the lobate sclerites in the same position on the leg of
Limulus. Lastly, there is attached to the distal extremity
of the seventh segment in Limulus a pair of short moveable
sclerites, forming a small nipper. Similarly there is a pair
of moveable sclerites or claws articulated to the distal
extremity of the seventh segment in the scorpion’s leg. The
annexed figure (Fig. 2) will make these suggested homologies
clear.

Whether Paleophonus possessed any structures com-
parable to the pedal spurs of recent scorpions and to the
lobate sclerites of Limulus is doubtful. I can detect
nothing comparable to them in the Scottish specimen, but
the figure of the Gotland specimen suggests the possibility
of the presence of one or more spurs at the distal end of the
sixth segment.

It is a matter for regret that the exact structure of the
basal segments of the legs, and the relation of these segments
to one another, are not with certainty interpretable, owing to
the crushing and displacement of the parts composing the
ventral area of the prosoma, and of the anterior somites of
the mesosoma. Hence too much reliance must not be placed

THE SCOTTISH SILURIAN SCORPION. 299

upon the accuracy of the attempted restoration of these
structures,

Fig. 2.—a. Fourth leg of Limulus molueccanus. B. Fourth leg of a
recent scorpion (Buthus australis). c. Third leg of Silurian scorpion,
Paleophonus nuncius, after Thorell.

1—7. Segments. s. Suture between fourth and fifth segments of the leg
in Limulus. sp. Spurs and lobate sclerites. ? sp. Processes possibly re-
presenting the point of attachment of spurs in Palewophonus. cl, Claws
in the scorpion, and pair of sclerites forming a nipper in Limulus.

In existing scorpions the basal segments (coxee) of the
legs of the first and second pairs are furnished with a for-

300 R. I. POCOCK.

wardly directed sterno-coxal or maxillary process, the coxe
of the second leg meeting each other in the middle line in
front of the prosomatic sternum, and sending forwards these
processes, which are in contact throughout their length, to

cn
NY:

Fig. 3.—Restoration of Paleophonus ILunteri (ventral view).

underlie the mouth. The coxe of the third and fourth legs,
on the contrary, are devoid of sterno-coxal processes, and are
separated from each other in the middle line by the sternal
plate, against the sides of which they abut.

THE SCOTTISH SILURIAN SCORPION. 301

A very different state of things appears to obtain in
Palewophonus. No trace of a sterno-coxal process is dis-
coverable upon the first leg. On the second, however, a
small one seems to be present. This lies transversely, and
meets its fellow of the opposite side in the middle line. On
the third leg a process similar in its form and relations is
also indicated, and the segment that bears it, instead of
abutting against the sternum, is mesially in contact with its
fellow. The probability of the correctness of this conclusion
is enhanced by its tallying with Peach’s opinion. I cannot,
however, quite agree with this author in believing that the
legs of the fourth pair are basally separated by the sternum
as in recent scorpions. On the left side of the specimen,
where the leg is well preserved, the segments seem to be
traceable right up to the middle line, the basal segment
being sharply defined. On the right side, however, this is
not so clearly indicated, on account of a displacement which
has resulted in the overlap of the proximal end of the fourth
leg by that of the third.

The sternum (st., P]. 19) does not stand out as a sharply
defined plate with clean-cut edges, but is merely represented
by the subpentagonal area that lies between and behind the
two proximal segments of the fourth leg of the left side,
and those of the third and fourth legs of the right side.
It shows a faint central circular depression answering
presumably to the similarly shaped sternal depression
in Cherilus, and to the median groove in other recent
scorpions.

The above-described arrangement of the skeletal pieces,
forming the ventral surface of the prosoma, offers many
points of morphological importance in view of the differences
that obtain in this particular between the recent scorpions
and Limulus or one of the Eurypterida. The relations of
the sternum to the coxee and the coxe to each other in the
scorpions have already been described. ‘Those of Limulus
and the Kurypterida may be stated in a very few words. In
the latter the basal segments of all the appendages, ex-

302 R. I. POCOCK.

cepting those of the first pair, acted as jaws, and were
frequently armed with teeth, the greatest share in crushing
and masticating food falling to the coxe of the fourth pair,
which were especially enlarged for the purpose. Behind,
and partially concealing them from the ventral side, lay a
large plate, the so-called “ metastoma,” the homologue of
the scorpion’s sternum. ‘l’o all intents and purposes the same
arrangement is found in Limulus, except that the coxe of
the fourth are less masticatory in function, and the “ meta-
stoma” is represented by a pair of moveable sclerites, the
“chilaria,” set immediately behind and between the bases
of the legs of the fourth pair.

In Paleophonus the sternal plate of the prosoma lies
apparently behind the basal segments of the fourth legs as in
Limulus, and, as in the latter and in the Hurypterida, the
basal segments of all the appendages were in contact or
capable of meeting in the middle line. On the other hand,
the coxee of the fourth were small and functionless so far as
the mouth was concerned, and food was probably crushed by
those of the chele as in recent scorpions, the sterno-coxal
sclerites. of the second and third pairs assisting in this
process, and preventing the escape of nutritive juices. Thus,
so far as the parts now under discussion are concerned, this
archaic scorpion presents a condition of things intermediate
in many particulars between that of the typical scorpions and
of Limulus or Kurypterus.

Mesosoma.—The ventral portion of the first somite of
the mesosoma is represented by a relatively short but wide
area lying behind the sternal region of the prosoma. This
area is marked in the middle line with a short longitudinal
groove (gen., Pl. 19), representing in all probability the divi-
sional line between the right and left halves of the genital
operculum. On each side this area is impressed with a
shallow but conspicuous indentation, which from its position
seems hollowed out for the reception of the third segment of
the fourth leg, perhaps in order that this portion of the
appendage might be insunk to the level of the generative

THE SCOTTISH SILURIAN SCORPION. 303

orifice, so that its prominence should offer no obstacle to the
act of copulation. .

A short distance behind the genital cleft a similar but
larger and more conspicuous median cleft is visible. This is
flanked on each side by a narrow longitudinally elongate
plate or lobe (end., Pl. 19), somewhat resembling one half of
the genital operculum of recent scorpions. On the outer
side of the right-hand lobe lies a bisegmented appendage
(pect., Pl. 19), which may be regarded as the homologue of a
recent scorpion’s pecten or comb. Along the posterior
border of this appendage are traceable a number of fine
striz occupying the position of the pectinal teeth. Similar
striz are traceable upon the left-hand side, although the
pecten itself is obliterated.

Peach regarded the cleft between the two above-described
lobes as the generative aperture, a conclusion it is impossible
to accept in view of the improbability of the backward move-
ment of this aperture on to the somite that bears the pec-
tines. ‘The opinion, which I here put forward, that the
generative aperture is represented by the slit which, although
not mentioned by Peach, appears on his published figure
immediately behind the pentagonal prosomatic sternite,
seems on morphological grounds far more likely to be correct.
Thorell, moreover, suggested that the pair of lobes lying
between the pectines correspond to the small, sometimes
longitudinally grooved pectinal sternite of recent scorpions.
This may be the true interpretation; but the shape of the
lobes, the length and depth of the groove that separates
them, and their relations to the pecten, suggest that they
have another significance, and are probably to be regarded
as the inner branches of an appendage of which the pecten is
the outer branch. From this standpoint the appendage may
be compared with the mesosomatic appendages of Limulus,
and of the archaic spider Liphistius. In the former the
appendages (except in the case of the genital operculum of
the Hastern species) consist of a broad foliaceous trisegmented
external branch, and of a slender trisegmented internal

304. Re , Z2OCOOK.

branch. In Liphistius also there are two branches, the
inner slender and unsegmented, the outer stout and composed
of two principal segments. Although in general form the
inner lobes (end., Pl. 19) of Paleophonus resemble those
of Limulus, they differ from the latter, and approach those
of Liphistius in being unsegmented. The outer branch is
broad and flattened, and is somewhat like that of Limulus,
except that it is relatively smaller, and les with its axial
line directed, not longitudinally, but obliquely outwards and
backwards like the comb of a typical scorpion. It shows,
however, no signs of segmentation into so-called ‘ fulcra ”
and ‘‘intermediate lamella,” such as are found in the combs
in the majority of species. Structurally, in short, it is inter-
mediate between a typical comb and the outer branch of one
of the mesosomatic appendages of Limulus. Furthermore,
the fine strie which fringe its posterior edge are, in my
opinion, too delicate to be the remains of teeth comparable
in shape and size to those of recent scorpions. Rather would
I suggest that they are portions of the edges of branchial
lamellae which were affixed like those of Limulus to the
posterior side of the appendage, with their lines of attach-
ment lying at right angles to its longitudinal axis.

‘These appendages overlie and almost completely conceal
the sternite of the third mesosomatic somite. The stern-
ites of the fourth, fifth, and sixth somites, however, are
fully exposed and well preserved. They are granular, and
resemble the corresponding plates in recent scorpions but
for the absence of the muscular impressions and, so far as my
observations go, of the stigmata. Peach, however, declares
most emphatically that “all four sterna exhibit on the right
side undoubted slit-like stigmata at the usual places.” It is
true that the sternites are somewhat wrinkled laterally, and,
as shown on PI. 19, exhibit certain shallow impressions,
which, especially in the case of the fourth and fifth sternites,
might be mistaken for stigmata; but it is hard to believe
that slits as conspicuous as the stigmata of recent scorpions
should be so indistinctly preserved on sternites in such an

THE SCOTTISH SILURIAN SCORPION. 305

admirable state of conservation that even their granulation is
still apparent. Nevertheless it must be borne in mind that
Peach’s opinion on this point is in complete agreement with
Thorell’s regarding the Gotland specimen.

According to Thorell this specimen exhibits on its right
side a portion of a displaced sternal plate, upon which a
distinct stigma is visible. This sternal plate he assigns to
the third somite of the mesosoma; but a glance at his draw-
ing shows that the greater part of it les at the sides of and
beneath the tergite of the second somite, and that at all
events a large part of the third sternite is situated on the
left-hand side beneath its corresponding tergite. To hold
that this third sternite has been fractured and displaced to
the extent that Thorell’s interpretation demands appears to
me to be an opinion based on an improbability. From the
position of the fragment that protrudes on the right-hand side,
I judge that it belongs to the second mesosomatic somite—
a somite which in all known scorpions bears the pectines but
is without stigmata,—and that it is part of its pleural mem-
brane. ‘This interpretation, if correct, nvolves the conclusion
that the “spiraculum” described by Thorell is a fortuitous
crack in the integument. There is one other point, too, bear-
ing indirectly upon the question of the presence or absence of
stigmata, in which, without further evidence, I find it im-
possible to accept Thorell’s decision. The Swedish specimen
is broken in two by a transverse fracture, crossing the fourth
somite of the mesosoma. The posterior half thus contains
the fifth and sixth mesosomatic somites and the metasoma. It
is admitted—and there is no reason to doubt—that the ventral
surface of the metasoma is exposed. According to Thorell,
however, the two mesosomatic somites which go to make up
the severed portion of the body le back uppermost. This
supposition implies the belief that the severed portion of
the specimen was itself completely divided into two at the
junction of the mesosoma and metasoma, that the latter was
overturned, and was so accurately fitted into place that perfect
continuity between it and the mesosoma was restored. That

vou. 44, PART 2.—NEW SERIES. U

306 Bet. POCOCK.

the uninterrupted outline presented by the somites in question,
which imparts so natural an appearance to this region, is thus
the result of pure accident I find hardly credible. In fact,
there is, I think, no reason to doubt that the fifth and sixth
mesosomatic somites were united to the metasoma, and shared
its unmistakable inversion. Hence the plates in question are
sternites. The important point attached to this conclusion
is the absence of stigmata on these sternites. Perhaps it was
this fact which led Thorell to his decision as to their tergal
character.

The above-given reasons justify a sceptical attitude towards
the alleged existence of stigmata in the Gotland Paleopho-
nus, at all events until a further examination of the specimen
settles the points now under dispute. And since I found no
distinct traces of stigmata in the Scotch specimen, I am in-
clined to believe that Peach fell into error on this point
perhaps influenced in part by the alleged presence of stigmata
in the Gotland example, perhaps in part by the assumption
that a form so closely resembling recent scorpions in other
structural details must also resemble them in the nature of
its respiratory organs.

To the belief in the presence of stigmata, implying the
existence of organs fitted for aérial respiration, coupled with
the knowledge of the terrestrial habits of all living scorpions,
is traceable the conviction evinced by most previous writers
that these Silurian scorpions lived on the land. This belief
is less easy to reconcile with the facts that both the known
specimens are relatively inanadmirable state of preservation,
and were met with in strata of undoubted marine origin,
containing abundance of admittedly marine organisms, than
the belief, which I hold, that Paleeophonus lived in the sea,
probably in shallow water, its strong, sharply pointed legs
being admirably fitted, like those of a crab, for maintaining
a secure hold amongst the seaweed or on the jagged surface
of rocks, and for resisting the force of the rising and falling
waves.

Respiration, as already suggested, may have been effected

THE SCOTTISH SILURIAN SCORPION. 307

by means of the appendages of the second mesosomatic
somite, although it must be admitted they appear too small to
have performed this office for the whole organism without
help from other organs. It is possible that there were such
organs in the form of small appendages bearing branchial
lamellae attached to the mesosomatic sterna. But if so, no
definite trace of such has been preserved. Or, indeed, it is
possible that the ventral plates, above regarded as meso-
somatic sternites, may have been broadly laminate mesosomatic
appendages, closely pressed down against the ventral surface
of this region, and bearing branchial lamelle on their pos-
terior surfaces. This suggestion gains some support from the
fact that the laminate mesosomatic appendages of the
Kurypterida are generally indistinguishable from sternal
plates.

Metasoma.—This region of the body in the Scotch speci-
men closely resembles that of the Swedish specimen, the same
surface, namely the ventral, being in each case uppermost.
Peach, however, states that the dorsal surface of the posterior
caudal segments is in part exposed. According to my inter-
pretation, on the contrary, in all the segments it is the area
lying between the inferior lateral keel on the left side (af.
lat. and sup. lat., Pl. 19) and the superior lateral keel on the
right that is exposed. Both of these keels are granular. As
in most recent scorpions, a pair of median keels (inf. med.,
Pl. 19) lie along the lower surface of the tail, between the
inferior lateral keels on the first four segments of the tail.
Keels corresponding to these four inferior medians and
inferior laterals are traceable upon the first metasomatic
sternite, and also, I think, upon the sixth (fifth caudal
segment). ‘his last fact, if true,is of some interest, inas-
much as it shows a more primitive arrangement than is found
in recent scorpions, where the two median keels have invariably
coalesced into one. ‘The inferior median keels on the pos-
terior caudal segments appear to be smooth. In the Swedish
specimen they are granular. ‘The lower side of the vesicle is
granular in both, but the aculeus in the Scotch example is

308 BR. I. POCOCK:

apparently less curved, less circular in section, and more tri-
angular than in recent scorpions and the Swedish specimen.

3. DxEscRIPTION OF THE Species, with Nores ON THE OTHER
KNOWN SILURIAN SCORPIONS.

The preceding description of the Scottish fossil, and the
comparisons that have been made between it and the Swedish
specimen, have revealed some noticeable structural differ-
ences between the two, which leave no other course open
than to regard the former as the representative of a distinct
and undescribed species. ‘This I propose to dedicate to Dr.
Hunter, and to diagnose as follows :

Paleophonus Hunteri, sp. n.

Differing from P. nuncius in its much smaller size, being
35°5 mm. as compared with 62 mm. in total length, in
possessing a pair of median eyes set close to the anterior
border of the carapace, in having the digits of the cheliceree
longer and thinner, and the moveable more curved, and the
cheles very much lighter in build, with the digits nearly
straight; in the absence of a spur from the fifth segment of
the first, second, and third pairs of legs, and in the smooth-
ness of the inferior median keels on the posterior segments
of the tail.

In addition to the specimens discussed in the preceding
pages, two other scorpions have been recorded from Silurian
strata, namely, Proscorpius Osborni and Paleophonus
loudonensis. The first was described by Whitfield
(‘Science,’ vi, p. 88, 1885; ‘ Bull. Amer. Mus. Nat. Hist.,’ i,
No. 6, pp. 181 to 190, 1885), and was based upon a fairly
well preserved specimen, with the dorsal surface exposed,
from rocks referred to the middle of the Upper Silurian.
Like the Swedish and Scotch specimens, it was associated
with fossil remains of Pterygotus, Hurypterus, and other
marine organisms.

The specimen was examined by both Whitfield and

THE SCOTTISH SILURIAN SCORPION. 309

Scudder. The latter (Zittel’s ‘ Handbuch der Palaontologie,’
ll, p. 739, 1885) classified it with the Carboniferous scorpions
on account of the alleged presence of a pair of claws at the
extremity ot the anterior leg of the left side. ‘This classifi-
cation was endorsed by Whitfield, who based the genus
Proscorpius mainly upon these claws, declaring them to
be very similar to those of living forms. His figure shows
no such similarity. The apical segment of the leg is simply
bifid at the tip, a feature which may be due to fracture, or
may represent a pair of sclerites like those borne at the tip
of the distal segment of the fourth leg of Limulus; or may
be explained on the supposition that the end segment ter-
minated in a sharp point as in Paleophonus, and was
furnished near the tip with a moveable spine or spur. Since,
however, there is no agreement between Scudder and Whit-
field as to whether the segment stated to possess these claws
is numerically the third or sixth from the base, it seems idle
to discuss the matter further. If Scudder’s interpretation
of the numbers of the segments is correct, these ‘claws’
are situated at the end of the third or fourth segment, and
cannot be compared with the tarsal claws of other scorpions.

Apart from the leg, the chief points of interest connected
with Proscorpius Osborni are the presence of a pair of
eyes ona median ocular tubercle, and of a row of lateral eyes
(not shown in the figures, by the way) on each side of the
carapace. ‘lhe rounded median tubercle projects in the
middle line of the fore border of the carapace, the lateral
angles of which are also rounded. Hence the trilobate
appearance of the anterior border of this plate, which forms
such a contrast to the even emargination seen in the Swedish
and Scotch Palwophonus. It is further to be noticed that
the dorsal integument is smooth, and not granular as in
Paleophonus nuncius,

Along the right-hand side of the specimen, both Scudder
and Whitfield agree that six (five mesosomatic and one
metasomatic) abdominal sternites are exposed. The first of
these belongs to the second mesosomatic somite, which in

310 hws, ~ROCOOK.

recent scorpions bears the pectines and has no appreciable
sternal area.

But there appears to me to be no reason for regarding this
so-called sternal area other than as the pleural membrane of
the second somite of the mesosoma.

Mr. Whitfield could find no satisfactory evidence for the
existence of stigmata, and infers from this fact, and from the
nature of the strata in which the specimen was preserved,
that the species was ‘‘ aquatic in habits,’ and furnishes a
“link between the true aquatic forms like Hurypterus and
Pterygotus and the true air-breathing scorpions of recent
periods.”

Of Palwophonus loudonensis, described by Laurie,
from the Upper Silurians of the Pentland Hills (‘Tr. Royal
Soc. Edinb.,’ xxxix, p. 576, pl. 1, fig. 1, 1889), little need be
said, the specimen being too imperfectly preserved to yield
satisfactory data for discussion. That the specimen was
specifically distinct both from P. nuncius and P. Hunteri
cannot be doubted if the great length of the carapace and
the slenderness of the tail in the fossil are not attributable to
imperfection of preservation. Asin P. Hunteri, there are a
pair of median eyes close behind the fore border of the cara-
pace, which is emarginate.

No genuine stigmata were discovered, but on some of the
mesosomatic somites a curved ridge running obliquely out-
wards and backwards on the sides of the segments was
traceable. ‘he ridge on the second somite Laurie inter-
prets as the impression of the outline of the pecten, those
on the others as the outline of a plate-like guill-bearing
appendage.

4, RECAPITULATION.

From a morphological point of view, perhaps the most
important results obtained by the examination of this fossil
are those connected with the structures of the basal segments
of the prosomatic appendages, and their relation to the
sternal area of this region, and those connected with the

THE SCOTTISH SILURIAN SCORPION. a1 ET!

structure of the appendage of the second somite of the
mesosoma.

If the above-given interpretation of the arrangement of
the parts constituting the ventral side of the prosoma is
correct—and [| do not think it is likely to be very far wrong
—Palzeophonus occupies an intermediate position between
Limulus and the Eurypterida on the one hand, and recent
scorpions on the other, standing, if anything, rather nearer
to the former than to the latter.

The same may be said of the structure of the second
mesosomatic appendage, which with its outer and inner
branch is like the corresponding appendage in Limulus;
while the outer branch itself, although in general form and
size resembling the pecten of a scorpion, offers some inte-
resting structural features in which it differs from that organ,
and resembles the outer branch of a mesosomatic limb of
Limulus.

On the whole, it must be admitted that Paleophonus
Hunteri supplies a few more links to the chain of evidence
pointing to the descent of the scorpions from marine Limu-
loid ancestors.

EXPLANATION OF PLATE 19,

Illustratmg Mr. Pocock’s paper on “The Scottish Silurian
Scorpion”? (Paleophonus Hunteri).

The figure gives a magnified view of the specimen of Paleophonus
Hunteri, Pocock, from the Upper Silurian of Lesmahago, Lanarkshire. It
was formerly in the collection of Dr. Hunter, and is now in the Kilmarnock
Museum, This lithograph was executed from the specimen itself by Miss
G. M. Woodward, under the supervision of Mr. R. I. Pocock, in October,
1900.

I—vI. Prosomatic appendages. sf. Sternal area of prosoma. gen. Genital
cleft. pec¢. Pecten, or external branch of appendage of second somite of
mesosoma. ezd. Internal branch of appendage of second somite of meso-
soma, sup. lat., inf. lat., inf. med. Superior lateral, inferior lateral, and in-
ferior median crests of fourth somite of metasoma.

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

Quarterly Journal of Microscopical

Science.

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

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

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

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

Each Number is sold at 10/-, and Four Numbers make upa
Volume.

London: ee & 7 CHURCHILE: ci Great ee Street.

TO CORRESPONDENTS.

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

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

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

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

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

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

For every four pages or less—

5 /

25 copies : ‘ : : Ee
T1V bale , ; ; : 6/-
LOOe* 5. Tie

Plates, 2/- per 25 if uneolonred ; if coloured, at the same rate for
every "colour.
Prepayment by P.O. Order is requested.
ALL COMMUNICATIONS FOR THE EDITORS TO BE ADDRESSED TO THE CARE
or Mussrs. J. & A. CuurcHitt, 7, Great Martporovuen Srreer,
Lonpon, W.

THE MARINE BIOLOCIGAL ASSOCIATION

OF THE
UNITED KINGDOM.

Patron—HIS MAJESTY THE KING.
President—Prof. E. RAY LANKESTER, LL.D., F.R.S.

ALO)g

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

The Laboratory at Plymouth
was opened in 1888. Since that time investigations, practical and scientific, have

been constantly pursued by naturalists appointed by the Association, as well as by
those from England and abroad who have carried on independent researches.

Naturalists desiring to work at the Laboratory

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

Works published by the Association

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

The Journal of the Marine Biological Association

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

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

Specimens of Marine Animals and Plants,

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

TERMS OF MEMBERSHIP.

ANNUAL MEMBERS . ‘ » #l- 1 Oper annum,

Lire MEMBERS . : : : . 15 15 0 Composition Fee.
FOUNDERS . : LOOM omO 3 -
GOVERNORS (Life Meniber: ae Conneil) 500 O O oe ef

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

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

The DIRECTOR,
The Laboratory,
Plymouth.

New Series, No. 175 (Vol. 44, Part 3). Price 10s.
| a Si), MAY, 1901. |

| THE

QUARTERLY JOURNAL

OF

MICROSCOPICAL SCIENCE.

EDITED BY

E. RAY LANKESTER, M.A., LL.D.; F.B.S.,

HONORARY FELLOW OF EXETER COLLEGE, OXFORD; CORRESPONDENT OF THE INSTITUTE OF PRANCE,
AND OF THE IMPERIAL ACADEMY OF SCIENCES OF ST. PETERSBURG, AND OF THE ACADEMY
OF SCIENCES OF PHILADELPHIA; FOREIGN MEMBER OF THE ROYAL BONEMIAN
SOCIRTY OF SCIENCES, AND OF THE ACADEMY OF THE LINCEI OF ROME}
ASSOCIATE OF THE ROYAL ACADEMY OF BELGIUM; HONORARY MEMBER
OF THE NEW YORK ACADEMY OF SCIENCES, AND OF THE
CAMBRIDGE PHILOSOPHICAL SOCIETY, AND OF THE ROYAI
PHYSICAL SOCIETY OF EDINBURGH ; HONORARY
MEMBER OF THE BIOLOGICAL SOCIETY
OF PARIS;
DIRECTOR OF THE NATURAL HISTORY DEPARTMENTS OF THE BRITISH MUSEUM; LATE PULLERIAN
PROFESSOR OF PHYSIOLOGY IN THE ROYAL INSTITUTION OF GREAT BRITAIN.

|
|

WITH THE CO-OPERATION OF

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

FELLOW AND TUTOR OF TRINITY COLLEGE, CAMBRIDGE}

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

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

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

BEYER PROFESSOR OF ZOOLOGY IN THE OWENS COLLEGE, MANCUESTER,

WITH LITHOGRAPHIC PLATES AND ENGRAVINGS ON WOOD. ‘

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

1901.

EE TTS A AE Ee OO Tea She BS ST,

Adlard and Son,]} [Bartholomew Close.

CONTENTS OF No. 175.—New Series.

MEMOIRS:

The Life-History of Nucula delphinodonta (Mighels). By GiL-
MAN A. Drew, Professor of Biology, University of Maine, Orono,
Me. (With Plates 20—25) .

On the Structure of the Hairs of Mylodon Listai and other South
American Edentata. By W.G. Riprwoop, D.Sc. F.L.8., Lecturer
on Biology at the Medical School of St. Mary’s gen London.
(With Plate 26) :

On the Structure and Affinities of Saccocirrus. By Epwin §S.
GoopricH, M.A., Fellow of Merton College, Oxford. (With Plates
27—29) ; . : 3

On the Question of Priority with regard to certain Discoveries upon
the Aitiology of Malarial Diseases. By Groner H. I. Nurtatt,
M.A., M.D., Ph.D., University Lecturer in Bacteriology and Pre-
ventive Medicine, Cambridge

Studies in the Retina: Rods and Cones in the Frog and in some other
Amphibia. By H. M. Bernarp, M.A.Cantab. (from the Biological
Laboratories, Royal College of Science, London). Part IJ. (With
Plates 30 and 31) . :

Staining with Brazilin. By Sypnny J. Hickson, Beyer Professor of
Zoology in the Owens College, Manchester

PAGE

313

393

413

429

443

469

JUN 27 1961

HE LIFE-HISTORY OF NUCULA DELPHINODONTA. 3138

The Life-History of Nucula delphinodonta
(Mighels).

By

Gilman A. Drew,
Professor of Biology, University of Maine, Orono, Me.

With Plates 20—25.

Tur material upon which these observations were made
was secured at Casco Bay, Maine, during the summers of
1897 and 1898. Nucula delphinodonta is a small form,
seldom growing to be more than 4 mm. in length, and as it
lives below low-tide mark it is not very well known by col-
lectors. By using a sufficiently fine dredge, however, un-
limited numbers of adult and young specimens may be
procured. Individuals may be found living under very
different conditions; in inlets and protected places, and ex-
posed to the open sea, and from near low-tide mark to a
depth of several fathoms. The principal habitat, however,
is in the shallow inlets and near the heads of sounds, where
the bottom is composed of fine mud, mixed with some sand,
broken shells, and decaying vegetable matter. Individuals
are most numerous just outside of the eel grass which skirts
the shore where the bottom is of this character, in water
which at low tide is from one to three fathoms deep. The
mud in which they live is much like that inhabited by
Yoldia limatula, except that it is not so free from shore
débris. Although some specimens may be obtained where
Yoldia is most abundant, they are generally more numerous

VOL. 44, PART 3.—NEW SERIES. x

314 GILMAN A. DREW.

somewhat nearer the shore, and they may be very numerous
at considerable distances from places where Yoldia is known
to thrive.

In picturing the conditions under which these animals live
along the coast of Maine, the reader should not fail to take
into account the average tide of about ten feet, which keeps
the water very pure over a comparatively foul bottom. The
fauna and flora of these bottoms are very abundant and di-
versified, but have not been carefully catalogued. Diatoms
of several species abound, and form a large part of the food
of Nucula. Other Algz, Ostracods, Foraminifers, small
Lamellibrauchs, and Gastropods are also very abundant, and
small individuals of most of these forms are occasionally
found in the stomachs of preserved specimens.

While I have never succeeded in getting individuals to
form brood-sacs in captivity, they live well in aquaria, and
may be kept for several weeks either in vessels containing
the mud in which they normally live, or in vessels without
this mud. It is not even essential that the water be changed
very frequently.

When placed in vessels containing mud they bury them-
selves, and seem never to come to the surface to stay for any
considerable time. ‘They are at all times comparatively
sluggish, and seem to wander around in the mud by slow
thrusts and retractions of the foot, which is a very perfect
burrowing organ. When placed in mud that is just sufli-
ciently deep to cover them, their movements can be followed
fairly well by the movements of the mud. ‘To see them
feeding it is necessary to use only a very thin layer of mud.
The action of the palp appendages can then be observed.
They perform the same function that is performed by similar
appendages on the palps of Yoldia (1), that is, they are
food collectors. Nucula delphinodonta seems normally
to feed beneath the surface of the mud, so feeding cannot be
observed as easily as it can be in the case of Yoldia (Text-
ne).

The movements of the foot are best observed by placing

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 315

specimens in shallow dishes of sea water. When specimens
are placed on mud they bury themselves so promptly that
the movements of the foot cannot be carefully followed.
The movements are al] such as would be of service in bur-
rowing in mud. Although specimens have been kept under
observation under different conditions for long periods of
time, I have never known one to execute movements that
could be interpreted as creeping. In 1853 Forbes and
Hanley, in describing Nucula nucleus (4), made the fol-
lowing statement :—“The foot is white, and as if peduncu-
lated and deeply grooved, so as to expand into a broad leaf-
shaped disc with serrated margins; by means of this organ
it can creep like a Gasteropod, and we have seen it walk up
the sides of a glass of sea water.” This seems to be the
only observation of this kind on record, although many
students have worked on this and related forms. The
authors who have adopted the view that the foot functions
as a creeping organ in members of this group have, in nearly
every case, had only preserved material to work upon, and
perhaps have been influenced by finding so many characters
that seem to them to denote generalised structure. Some
Lamellibranchs are able to pull themselves over smooth
surfaces, but my observations lead me to believe that the
form and structure of foot found in this group is especially
poorly adapted for such a purpose (8). ‘The expanded foot
of Nucula delphinodonta is relatively very large, and
the almost spherical shell is frequently turned from one side
to the other, but nothing comparable to creeping has been
observed.

Although many Lamellibranchs carry their eggs and de-
veloping embryos, I think this is the first case reported
where a special external sac is formed for the purpose. This
sac (fig. 1) is composed of a mucus-like material, mixed with
foreign bodies, and is attached to the posterior ends of the
valves of the shell. Although the process of making the
sac has never been observed, it seems probable that the
mucus-like material is secreted by the hypobranchial glands.

316 GILMAN A. DREW.

This material is probably passed posteriorly by the action of
the cilia on the mantle, and very likely the respiratory cur-
rents of water swell it into a sort of bubble that remains
attached to the posterior ends of the shell-valves, and, while
still soft, adheres to the foreign particles with which it comes
in contact.

That the hypobranchial glands are concerned in the forma-
tion of the material from which the brood-sac is formed is
indicated by their appearance before and after the sacs have
been formed. In females in which the ovaries are still full
of eggs, the cells of the hypobranchial glands are large and
gorged with secretions, while in females that have formed
the brood-sacs the cells are shrivelled and almost devoid of
secretions.

The eggs are deposited in the brood-sac (fig. 1), and in it
the embryos are carried until they reach an advanced stage
in development, probably for a period of three or four weeks.

The eggs of this species are brown, opaque, few in number,
and correspondingly large. From about twenty to seventy
may be found in a sac, and they average about ‘21 mm. in
diameter. Each egg is enclosed in a membrane that is pro-
bably secreted by the egg, but its formation has not been
observed. Fertilisation is probably accomplished in the
brood-sac. Eggs and young embryos do not live well after
they are removed from the brood-sacs, so the ages of the
various stages have not been determined. Processes of
maturation and cleavage proceed slowly. ‘he time between
the appearance of the first and the second polar body is
frequently as much as two hours, and the time between
cleavages seems to be nearly or quite as long. It is not
beyond doubt, however, that the removal of the eggs from
the brood-sacs influenced the length of time. That develop-
ment is slow is not to be doubted. Embryos taken from the
brood-sacs of specimens kept under as nearly natural condi-
tions as possible for a month, had only reached the stage
where two gill-lobes were formed.

It seems probable that the polar bodies may be formed by

THE LIFE-HISTORY OF NUCULA DELPHINODON'TA. 317

eggs that have not been fertilised. Egos were sometimes
obtained that formed polar bodies and developed no further.!

Just before each polar body is formed, a more or less
distinct, and frequently a very pronounced swelling, makes

C) 446 8

FIGA
Ale = SA

Text-rics, A, B, C, and D.—Early stages in the development of
Nucula delphinodonta.

its appearance on the side of the egg opposite the point

* Most of the eggs of an isolated specimen of Nucula proxima, a form
that throws its eggs free into the water, formed the polar bodies, and a few
eggs cleft the first time. It is possible that some sperm were in the water,
but the water had not been changed for nearly twenty-four hours before the
eggs were laid, and sperm of this species do not seem to retain their vitality
for nearly so long a time.

318 GILMAN A. DREW.

where the polar body will appear. In the preparation for
the first cleavage a similar swelling is formed on the side
opposite the polar bodies. When the egg divides, the divid-
ing wall passes to one side of this swelling. The two blasto-
meres are accordingly rather unequal in size. ‘The difference
in the size of the two blastomeres seems to depend upon the
size of the swelling that precedes their formation. Cleavage
into four and eight cells (Text-figs. C and D) are typical.
The polar bodies retain their position on the animal pole
until the embryo acquires cilia, when they are rolled around
on the inside of the membrane. No attempt has been made
to follow out the fate of the individual cells.

In the sixteen-celled stage, figs. 2 and 3, a small cleavage
cavity is present. Later this becomes slightly more pro-
nounced. The cells on one side of the blastula divide more
rapidly than those on the other side, and push over them in
the form of a cap (fig. 4). A pocket appears between the
large cells at such a point as is indicated by the asterisk in
fig. 4. Just how this pocket is formed is still a matter of
some doubt, but it seems to be formed by the separation and
division of some of the larger cells. This pocket (fig. 8) can
now be compared with the invaginate portion of a gastrula.
It represents the first appearance of the gut.

About the time that the pocket is formed most of the
smaller surface cells acquire cilia (fig. 6), and the embryo
begins to roll around in the membrane. ‘The cilia are all
short, similar in appearance, and seem to be evenly scattered
over the surfaces of the cells. In whole mounts the
boundaries of the surface cells are not very distinct, but the
cells do not seem to have a very definite arrangement.

From these small surface cells, that at this stage appear
very much alike, the test,' the apical plate, and the cerebral
ganglia are formed.

The large cells near the blastopore do not bear cilia, at

1 T use the term “test” here, as in former publications, to designate the

surface cells that bear cilia and may be homologised with the velum of other
forms.

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 319

least none could be found on preserved specimens. They
are concerned in the formation of the shell-gland.

The embryo is still nearly spherical, and so opaque that,
while alive, internal changes cannot be followed. A few
cells, probably the beginning of the mesoderm, he above
and by the sides of the gut. About this time some of the
surface cells around the blastopore divide, and push in to
form a stomodeum. Other cells near the blastopore become
enclosed by the surface cells, and together with cells probably
derived from those forming the stomodeum, finally form a
portion of the new ectoderm, that soon covers the body of
the embryo inside of the test. When the ectodermal layer
is complete it joins, but does not enclose, the stomodzum.
In position as well as origin the stomodzum is ectodermal.

Before the ectodermal layer is complete the embryo begins
to elongate, and the surface cells close in over the shell-
gland from the sides and anterior end, At the same time
the surface cells become arranged in rather definite rows.
It is very difficult to get satisfactory views of these cells in
whole mounts, but there seem to be five rows, beside a group
at the anterior end that forms the apical plate. ‘T'wo or three
of the posterior rows are interrupted in the region of the
shell-gland, but this interruption disappears as the shell-
gland becomes closed in. Closing is never complete. A
small opening is left dorsal to the blastopore, separated from
it by the width of one test cell (fig. 15). The anus comes to
lie near this opening at a later stage (fig. 24). Before the
shell-gland is covered the gut turns towards the dorsal side
(fig. 11, mg.), and the mesoderm cells take up a position near
the posterior end of the embryo. Two of the mesoderm
cells are large, and have very large and conspicuous nuclei.
These cells are far posterior, and lie side by side.

Soon after the shell-gland is covered, the gut begins to
grow posteriorly, almost, if not quite, in contact with the
shell-gland dorsally, and separated from the stomodeeum
ventrally by a few mesoderm cells (fig. 15). A small space
appears among these mesoderm cells that later becomes con-

320 GILMAN A. DREW.

nected with a space that is formed between the gut and the
shell-gland.

At no stage in its development is the shell-gland invagi-
nated. From the time of its formation it arches dorsally to
some extent (figs. 7 and 9, sg.). Just before it becomes
covered by the test it flattens somewhat (figs. 11 and 12), but
it soon arches dorsally again and becomes quite convex (fig.
L73.804)\.

The cells that give rise to the cerebral ganglia are few in
number (fig. 15, cg.), and lie ventral to the anterior end of
the stomodeum. They frequently come to the surface, but

Trxt-ric. E.—Surface view of a young embryo of
Nucula delphinodonta.

they may be entirely covered by test cells. A more or less
distinctly recognisable test cell lies between the cerebral
ganglia and the apical plate, but beneath this test cell the
cerebral ganglia and the apical plate are in contact. The
two cerebral ganglia seem to originate from a single mass of
cells. There is no indication of the formation of cerebral
pouches, as in Yoldia (Text-fig. V). The position occupied by
the developing body of Nucula does not make it necessary
for the cerebral ganglia to shift their position from the point
of their formation until the test is shed.

The apical plate is composed of a number of cells, the walls

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 321

of which are rather indistinct (fig. 15, ap.). They bear cilia
that in size and distribution resemble those that cover the
test cells.

Under favourable conditions the test cells can be seen to
be arranged in five rows; occasionally part of a sixth row is
present. As in other stages, the boundaries between the
test cells are poorly marked, and it is quite impossible to
sketch them accurately. Text-fig. E shows their general
arrangement, but it must be understood that this 1s quite
diagrammatic. The cilia on the test cells of this species are
not collected into bands as they are in Yoldia (Text-fig. F),

Tpxr-ric. F.—Surface view of a forty-five hour embryo of Yoldia limatula.
ac. Apical cilia. 62. Blastopore. . Depression where the cells that
form the cerebral ganglia come to the surface.

but are evenly scattered over their surfaces. The embryo
becomes free from the egg membrane about the time that
the shell-gland becomes covered by the test, but the cilia are
barely powerful enough to slowly move the embryo on the
bottom of a dish. The absence of the bands of cilia, and of
the long tuft of apical cilia, is probably due to the protected
life of the embryo. Nucula proxima lays its eggs free in
the water, where they are fertilised and develop. These
embryos have to shift for themselves, and are very active.
Here, as in Yoldia, the cilia on each of the three intermediate
rows of test cells are long and collected into a band (Text-

DREW.

GILMAN A.

B22

“puels iets
‘ayeid jeoidy

bs
‘dp

“Say, “2 ‘wWneBpowoyg “pys
qns-ply “4m “pawtoy o1v erSues yeiqasao 9y4 yom wo. s]jag 49 ‘asodoyseig “79
‘B[ NABI] VIP[OR Jo odtqua anoy xis-£4I1Y} B JO UOTYOas [e44yISes ULIPe|{—'H ‘9Id-LXA,

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 323

fig. H). Sometimes part of a fourth band is present. The
end rows of test cells have the cilia evenly scattered over
their surfaces. The apical cilia are long and bunched into a
sort of whip that precedes the embryo when it swims. In
fact, the embryo resembles that of Yoldia so closely that,
except fora difference in size and a slight difference in shape,
a description of the surface appearance and movements of
one will do very well for the other also.

The cilia on the embryos of Nucula delphinodonta
may then be regarded as arrested in their development.
Life in the protecting brood-sac makes active locomotion
unnecessary and even dangerous, inasmuch as active embryos
would be likely to find their way out of the brood-sac, and
so be exposed to outside dangers.

The embryos continue to elongate and begin to flatten
slightly laterally (fig. 25). In the living embryo, viewed by
transmitted light, this stage is marked by the appearance of
a light spot near the dorsal margin. A smaller, much less
distinct ight spot has been present near the ventral margin
for some time, and corresponds in position to the cavity that
was mentioned as appearing in the mesoderm, ventral to the
gut. This space has enlarged considerably (fig. 24), but is
covered laterally by rather thick walls of ectoderm and by
some mesoderm, so it is not very distinct. ‘lhe dorsal space
is formed by the arching up and flattening out of the cells of
the shell-gland, which are now beginning to form the mantle
lobes (fig. 20). It is bounded dorsally, laterally, and poste-
riorly by the mantle, anteriorly by the mantle and the apical
plate, and ventrally by the gut and by the body-wall. A
few cells, apparently mesodermal, lie in this space, generally
attached to the mantle or to the gut.

At a little later stage (fig. 25) two fibre-like cells stretch
.from the anterior end of the gut posteriorly and dorsally to
the mantle. ‘They are quite conspicuous in living embryos,
and they retain their position until after the test is thrown
away.

About this stage the gut, which has grown posteriorly,

DREW,

GILMAN A,

324

“qSay, "2 ‘wnepouloyg ‘pys “purys-ljayg “4s “qnS-prpy “Fu -aqed jeordy -dv
‘eurxodd epnonyy jo ofiqwa inog aAy-fyU9My & JO MOTOS [e4ILSeS ULIpo|{—' F “N1I-LX a],

(HE LIFE-HISTORY OF NUCULA DELPHINODONTA. 3825

acquires an auus (fig. 24). The anus is not directly applied
to the pore that opens between the test cells, but it opens
into a cavity that is continuous laterally with that portion of
the embryo that, as the mantle continues to grow, becomes
the mantle chamber. This communication will be described
in a later stage.

The embryo flattens laterally until its thickness equals
about two thirds of its dorso-ventral width, and the dorsal
space becomes considerably enlarged (fig. 25). Near the
anterior end of this space the anterior adductor muscle (aa.)
makes its appearance. At first it consists of a very few
fibres, and is not conspicuous. The anterior enlarged
portion of the gut takes on the distinctive characters of the
stomach (sto.), and the liver grows out as paired right and
left pouches (l.). The anterior end of the stomach is carried
dorsally, and a more or less distinct bend is made where it
joins the intestine.

The relationship of the various cavities in the embryo to
each other, and of the anal pore in the test to the mantle
chamber, can be best understood by comparing the sagittal,
horizontal, and transverse sections of embryos, represented
on Plate 22, with the reconstruction of an embryo at the same
stages of development (Plate 21, fig. 25). The position of the
horizontal and transverse sections are indicated on fig. 25 by
numbers that correspond to the numbers of the figures.

The dorsal cavity is separated from the ventral cavity by
the gut (fig. 28). Insome sections the two cavities communi-
cate around the sides of the gut. ‘This may be due to
shrinkage, but it seems more likely that the two portions are
parts of a single cavity. Itis just possible that the cleavage
cavity never entirely disappears, and that this cavity can be
traced back tothe blastoccele, but I am of the opinion that it
is a later formation, and represents a schizoccele. Its fate is
of interest, and will be referred to in later stages.

The lobes of the mantle are now well formed, a distinct
shell-cuticle has been secreted, and some lime salts have been
deposited. ‘The stomodeum for most of its length is joined,

DREW.

GILMAN A.

326

“UOITSUBS [B190ST A.
‘ba "ysay, "7 ‘wnepowojyg “pys “]]aG ‘s ‘“eSaeS [eaqaseo ay} 0} ovjans oy WoIy Surpeay yonog *
‘uorfsues [epeg ‘4d -‘ajosnut sojonppe sorsaysog ‘md *4sk00jQ “70 “puelS aatysaSip ayy Jo aqoy 4Jory +77
‘oulysayuy “7az “goog fF “uolpsues yerqoiaD *49 ‘arodoyseig "79 ‘eiylo Jeordy ‘9m ‘ajosnut s0jonppe
LOMayWY “7 “eTp1O [wolde ayy ynonyIA Suc] "wu Z. aw suawioadg apts 4ja] ayy Woy Woes st OLIquia oI] y,

"HO 4svo ST 480} ayy o10joq qsnf osvys v ye B[NyVUI] BIp[OX jo okIqua Ue jo UOTONIysUOIey— | ‘OId-LXTY,

‘HE LIFE-HISTORY OF NUCULA DELPHINODONTA. 327

but is not enclosed by the body ectoderm (fig. 28), which in
this region forms the walls of the foot (f.). Near its external
opening the stomodzum has become free, and is more or less
closely jointed to the test cells.

The relation of the anal test pore to the mantle chamber
can now be understood. As shown by a sagittal section
(fig. 26), this pore opens into a small cavity that receives the
anus. ‘This cavity is bounded anteriorly by the posterior
wall of the foot, and ventrally either by the stomodzeum or
by cells covering the dorsal portion of the stomodeum.
‘'ransverse (fig. 27) and horizontal (fig. 31) sections show that
this cavity spreads out laterally, and becomes continuous with
that portion of the mantle chamber posterior to the foot. At
this stage the foot is very imperfectly formed, and contains
the cavity that has been referred to as the ventral cavity.
The cavity soon disappears, and the ectoderm on the two sides
of the foot fuse ventrally, dorsal to the stomodeum. The foot
is still very small, and shows no sign of its future activity.
At a corresponding stage the foot of Yoldia is quite well
developed (Text-fig. I). his is about the condition of the
embryo when the test is thrown away.

It takes several hours for embryos of this species to cast
the test, a process that with Yoldia limatula and Nucula
proxima is completed within a very few minutes after it is
begun. The test cells in the region of the anal pore break
apart, and the whole mass is frequently pushed forward to
the region of the apical plate. his stripping forward
carries the outer end of the stomodeum forward to some
such position as is shown by fig. 34. The cilia on the test
cells remain feebly active for a considerable time. While the
test cells, stomodeum, and apical plate still adhere to the
embryo, the stomach and liver pouches are drawn some
distance dorsally into the schizoccele (fig. 34, sto. and 1.).
Whether the fibres extending from the stomach to the mantle
are important in effecting this movement is not known. Their
position is suggestive, but I have no direct evidence that
they contract. The position now occupied by the stomach

328 GILMAN A. DREW.

causes the bend where the intestine joins the stomach to
become quite abrupt.

At the same time that the stomach moves dorsally, the
cerebral ganglia (fig. 34, cg.), which are still a mass of rather
undifferentiated cells, are carried up, and come to lie poste-
rior and a little ventral to the anterior adductor muscle (aa.).
The foot (f.) retains its position beneath the intestine and

; Ze
\ Xe
|

\
std \
ot ‘p

Texr-r1g. J—Reconstruction of an embryo of Yoldia limatula at a stage
during casting. Represented as seen from the right side, with the right
shell-valve and mantle lobe removed. aa. Anterior adductor muscle.
eg. cerebral ganglion. /. Foot. g. Rudiment of gill. iz¢. Intestine.
ot. Otocyst. pa. Posterior adductor muscle. pg. Pedal ganglion. +.
Pouch that leads to the cerebral ganglia. 7r/. Right lobe of the digestive
gland. séd. Stomodeum. ¢. Adhering test cells. vy. Visceral ganglion.

stomach, and in the general dorsal movement is carried a
little further from the margin of the shell. A similar stage
for Yoldia is represented by Text-fig. J. At the end of several
hours the stomodeum (fig. 34, std.) breaks across near the
tip of the foot, and together with the apical plate and the
remnants of the test cells is thrown away. From appear-
ances I am inclined to believe that the whole of the apical

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 329

plate is thrown away, but this may not be the case. The
test cells may or may not remain attached to the apical plate
and stomodzum until these are thrown away. Generally
many of them break loose or go to pieces before this change
occurs, but some of them nearly always remain.

After casting is completed (fig. 35) the stomach (sto.) and
the liver lobes (/.) are drawn further into the schizoceele, and
the liver lobes begin to be flattened out against the mantle.
The cerebral ganglia (cg.) lie almost directly posterior to
the anterior adductor muscle (aa.), and the ectodermal
thickenings that result in the formation of the pedal and
visceral ganglia soon begin to form (fig. 36).

Thus far in the development of the animal the shell-valves
have remained gaping, but after the removal of the apical
plate and the stomodzeum they are free to close. ‘his is
effected by the contraction of the anterior adductor muscle, and
materially diminishes the space between the shell-valves.

The closing of the shell is accompanied by important
changes in the liver pouches, changes similar to those that
have been described for Yoldia (1). Apparently as the result
of the mechanical pressure the liver pouches go to pieces,
and the large cells of which they were composed become
rounded and scattered through most of the schizoccele
(fig. 36, z.). The posterior portion of the schizoccele is not
filled by the scattered liver-cells. This persists and finally
becomes the pericardium.

The foot (fig. 39, f.) grows and soon executes feeble move-
ments. The pedal gangha (pg.) and visceral ganglia (vg.)
take on definite form; the posterior adductor muscle (pa.)
appears ; and the invaginations that result in the formation
of the otocysts are formed. Very possibly commissures
connect the ganglia at this time, but I have not been able tu
distinguish them from the surrounding tissue until a some-
what later stage. A thickening on the inner surface of the
posterior end of each lobe of the mantle indicates the begin-
ning of the formation of the gill (fig. 59, g.).

About this time a little invagination on the mid-line of the

VoL. 44, PART 3,—NEW SERIES, Y

330 GILMAN A. DREW.

ventral portion of the foot, just anterior to the heel-like
projection, makes its appearance (fig. 39, bg.). This develops
into the byssal gland. It grows rapidly until it becomes
proportionately very large (fig. 45, bg.), then ceases to grow,
and possibly shrinks somewhat. In the adult it is compara-
tively insignificant (fig. 48). No signs of byssal threads have
ever been observed, nor have the secretions ever been seen
to protrude from the duct of the gland.

The foot grows rapidly, and the projection that looks like
a heel becomes more marked (fig. 40, f.). Anterior to this

Bask,
i
~
wes a
IN 2%
y OS

Text-Fic. K.—Reconstruction of a ten-day embryo of Yoldia limatula.
Represented as seen from the right side with the right shell-valve and
mantle lobe removed. aa. Anterior adductor muscle. cg. Cerebral
ganglion. f. Foot. g. Gill. iz¢. Intestine. //. Left lobe of the di-
gestive gland. of. Otocyst. pa. Posterior adductor muscle. pg. Pedal
ganglion. ri. Right lobe of the digestive gland. sfo. Stomach. vg.
Visceral ganglion.

projection the sides grow ventrally faster than the interme-
diate portion, and finally from the side flaps that are so
characteristic of the foot of the adult. Movements of the
foot now become energetic.

The gill (fig. 40, g.) becomes more pronounced, and soon

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 331

unequal growth causes it to be divided into two lobes. The
dorsal wall of the stomach re-forms, and the liver-cells begin
to be rearranged. The commissures between the ganglia are
distinctly visible. The otocysts (ot.) are quite large, and
contain granules. Although adults have canals leading from
the otocysts to the exterior, I have not been able to demon-
strate their existence in this or somewhat older stages. The
presence of the otocystic canal had been explained (18) as
the persistent opening of the otocyst, which was formed as
an invagination from the surface of the body. This seems
to be the natural explanation, but if canalsare present at this
stage they are certainly very small. I am inclined to regard
the exceedingly small size or absence of these canals as
evidence against the view that the otoliths are foreign
particles,

Thus far most of the embryos have been carried in the
brood-sacs, but many of them now become free. They are
not set free by any act of the mother, but they individually
find their way into the mantle chamber of the mother and so
to the exterior.

Frequently younger embryos become free, but they
generally do not live long. Many embryos remain in the
brood-sacs until a much later period, but they do not seem to
be in need of its protection after the stage that has just been
described. The brood-sacs frequently remain intact after all
of the embryos have left them.

The more dorsal of the gill lobes elongates into a finger-
like process, and the ventral lobe broadens and becomes
divided into two lobes (fig. 41, g.). New lobes are thus
formed as the result of unequal growth of the most ventral
lobe.

About the time that the third lobe of the gill begins to
form a few papillae appear along the margins of the side flaps
of the foot (fig. 41, f.). The liver lobes also become hollowed
out and lose most of the rounded cells. Part of these cells
seem to go to pieces much as if digested (fig. 43), and it
seems quite possible that this is the case,

332 GILMAN A. DREW.

The heart (fig. 41, i.) is apparently formed from meso-
dermal tissue that collects to form a strand, that runs across
the pericardium from one side to the other. I have found no
indication of its being formed as paired pouches, as described
by Ziegler (20) for Cyclas cornea, nor have I found any
evidence that it originates as two masses that grow toward
each other. Its first appearance seems to be in the form of a
mesodermal strand of tissue that soon hollows out and
encloses the intestine. ‘The fact that the heart forms around
the intestine, and not dorsal to it, is of interest, and will be
discussed under the head of the Circulatory System.

The growth of the kidneys, which are now present as
sminall tubes, seems later to force the sides of the heart up
around the intestine (fig. 68), so that the ventral portion of
the ventricle becomes drawn out into a trough in which the
intestine lies. As the kidneys grow the trough becomes
deeper. By gradually closing in dorsal to the intestine at
the anterior and posterior ends, the trough is shortened, and
the intestine finally becomes free from the heart and lies
ventral to it (fig. 69). This is accomplshed by a very slow
process, and is not completed until after the animal has
become sexually mature.

I am inclined toward the opinion that the kidneys are
formed by the differentiation of mesodermal tissue. When
they first appear each is a very narrow tube, and extends
from its external opening in the mantle chamber to the mid-
line of the body. I have not succeeded in demonstrating
the inner pericardial openings of the kidneys in this or in
later stages. The cells soon become large and vacuolated,
and the kidneys grow rapidly and crowd anteriorly ventral
to the pericardium, where they become coiled and_sac-
culated.

With the formation of the fourth lobe of the gill (fig. 45)
processes make their appearance on the bases of the lobes,
between them and the mantle lobe to which the gill is
attached. ‘hese processes grow to form what have been
called the outer gill plates, but in this species their position

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 333

is better described as posterior than outer. The lobes, at
the bases of which the outer plates are formed, develop into
the inner plates. Viewed from the side, both sets of plates
are visible. The gills of Yoldia hang so that in a side view
the outer plates hide the inner plates, which lie directly
behind them (Text-fig. L). When viewing Nucula from the
side we see that portion of the gill that corresponds to the
ventral portion in Yoldia.

The labial palps appear as patches of ciliaon embryos with
three gill lobes (fig. 41). The outer palps soon begin to grow

Trext-rie. L.—Adult specimen of Yoldia limatula. Represented as seen
from the right side. Reconstructed to show internal organs. Fully
grown specimens may be 6 cm. long. aa. Anterior adductor muscle.
afm. Anterior foot muscles, 4g. Byssal gland. cg. Cerebral ganglion.
es. Exhalant siphon. f£ Foot. g. Gill. #%. Heart. zt. Intestine. ts.
Inhalant siphon. /p. Labial palp. of. Otocyst. pa. Posterior adductor
muscle. pap. Palp appendage. je. Posterior expansion of the margin
of the mantle. p/m. Posterior foot muscle. py. Pedal ganglion. _ sé.
Siphonal tentacle. s¢o. Stomach. vg. Visceral ganglion.

out as flaps (fig. 45, Ip.), and by the time that the fifth pair
of gill plates are formed the inner palps are present as folds.
The formation of the ridges on the ciliated surfaces of the
outer palps begins with embryos having six pairs of gill
plates, and the palp appendages are formed soon after

334 GILMAN A. DREW.

(figs. 55 and 56). The development indicates that each palp
appendage (fig. 56, pap.) is to be regarded as a pair of ridges
with an enclosed groove, developed and modified so that it
may be extended beyond the edges of the shell.

Little remains to be described in this general sketch of the
development, further than to call attention to the formation
of the loops of the intestine, that are indicated in different
stages of development by Text-figs. M to 8; to the forma-
tion of the cartilage pit and teeth on the valves of the shell ;
to the formation of more gill plates and foot papille as these
organs continue to grow; to the appearance of the otocystic
canals about the time that the sixth pair of gill plates are
formed ; and to the formation of the genital organs.

Mention should be made of a peculiar closed pouch (figs.
40, 48, and 63, v.), of unknown function, that lies just anterior
to the anterior adductor muscle. It makes its appearance in
embryos that are just getting the second gill lobes, and is
fairly conspicuous in adult animals.

Germ Layers.

An almost spherical embryo is formed as the result of the
first few cleavages (fig. 5), the cells on one side of which are
much larger than those on the other side. The large cells
extend far into the interior of the embryo, and the smaller
cells form a cap over the larger ones (fig. 4).

In reaching this stage of development the embryo has
passed through a blastula stage, in which the cleavage cavity
was very small (fig. 3). As the cells become arranged in the
manner described, the greater part of the cleavage cavity
disappears. It has not been determined whether any of it
remains or not. A depression appears near one side of the
group of larger cells at a point corresponding to the asterisk
in fig. 4. This depression seems to be formed by the separa-
tion and further division of some of the large cells, and
results in the formation of the gut (fig. 8, mg.).

The surface cells may now be regarded as ectoderm, and
at least two kinds may be distinguished: small ones, which

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 335

finally form the test, the apical plate, the cerebral ganglia, the
stomodzum, and a considerable portion of the future ecto-
derm of the embryo; and large ones that form the shell-gland.

The endodermal pouch is carried further into the interior
by the division and pushing in of ectodermal cells in the
region of the blastopore (figs. 9 and 11, mg.). In this way a
long stomodeum is formed on the ventral side of the de-
veloping embryo. The ectodermal covering of the later
embryo, exclusive of that derived from the shell-gland, seems
to be formed in connection with the formation of the stomo-
dzeum, by cells that wander in from the region of the blasto-
pore, and perhaps from cells derived from the stomodeum
itself.

About the time that the stomodeum begins to form, a few
cells, two of which are quite large and conspicuous, make
their appearance by the sides of the endodermal pouch, and
extend between it and the shell-gland. ‘These are meso-
dermal cells. Their exact origin has not been traced. As
the embryo elongates, the two large cells come to lie near
the posterior end of the embryo (fig. 19). They probably
correspond to similar cells that have frequently been de-
scribed for other forms. Similar cells are found in Yoldia in
a corresponding position.

Tost.

As the result of the first few cleavages a number of large
cells become covered on one side by a cap of smaller cells
(fig. 4). A part of the smaller cells become covered with cilia,
about the time that the gut is formed (fig. 8) ; others near the
blastopore divide rapidly and form the stomodeeum (figs. 9
and 11); still others form the cerebral ganglia ; while others
in the region of the blastopore wander in and form a part of
the future ectoderm.

The cells that bear cilia are concerned in the formation of
the test and apical plate. These cells soon cover the surface
that is not occupied by the shell-gland and the cerebral
ganglia. Both the apical plate and the cerebral ganglia are

336 GILMAN A. DREW.

small at this stage, consist of a very few cells, and can hardly
be distinguished from the surrounding cells. ‘he apical
cells acquire cilia about the time that the test cells do (figs. 9
and 11), and for some time they cannot be distinguished from
them. Later the apical plate may be told by its shape and
position (figs. 15 and 24, ap.).

As development proceeds the test begins to close in over
the shell-gland from the sides and anterior end (figs. 10—18).
Five rows of test cells can now be seen under favourable
conditions, but their outlines are very hard to determine.
Until the shell-gland is covered, two or three of the posterior
rows are incomplete dorsally. A small pore is left near the
posterior end, separated from the blastopore by the width of
one test cell (fig. 15). The anus comes to lie near this open-
ing (fig. 24).

The five rows of cells are now arranged much as shown in
Text-fig. h. From the formation of the test until its ultimate
disappearance its cells are evenly ciliated with short cilia.
In this respect the embryos differ from those of Yoldia
limatula (Text-fig. F) and Nucula proxima (Text-fig. H).
Both of these forms have the cilia on each of the three inter-
mediate rows of test cells collected into a band. Sometimes
a fourth more or less complete band is present. ‘The cilia on
the end rows of the test cells of all of the forms are short
and evenly scattered over the surfaces of the cells.

In this connection it is of interest to observe that the cilia
on the apical plate of Nucula delphinodonta are short
and independent, while those on the apical plates of Yoldia
limatula and Nucula proxima are long and bunched
together. They all seem to have a rather scattering origin,
and when animals are killed the cilia become separated from
one another.

In both species of Nucula the embryo differs from that of
Yoldia limatula in having a posterior opening in the test,
dorsal to the blastopore (fig. 15, and Text-figs. G and H).
This difference might easily be accounted for by a slight dif-
ference in the closing in of the test over the shell-gland.

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 837

The ciliated embryos of Nucula delphinodonta, unlike
those of the other two forms, are not able to swim freely in
the water. At the most they are barely able to move on the
surface of a glass dish. This is probably the result of their
being carried in a protecting brood-sac. It seems but
natural that the bands of strong cilia and the apical tuft of
cilia would not be developed by embryos such as these,
were there no need for active Jocomotion, and where active
locomotion would be dangerous. It is for the best interest
of embryos that they remain in the brood-sacs, where they
are protected from many enemies. Were they capable of
active movement, many would probably escape and perish.
In the two related forms, Nucula proxima and Yoldia
hmatula, the embryos have to depend on their own
activities for their existence.

It is highly probable that the embryos of the ancestors of
Nucula delphinodonta led an active, free-swimming
existence. ‘The rearing of embryos in protecting brood-sacs
is very possibly connected with the present life of the animal
beneath the surface of the mud, and, in any case, has prob-
ably been acquired at a comparatively recent day. Again,
the test in its present condition is of no appreciable value to
the embryo, and no doubt is to be regarded as a vestige of a
once functional organ.

Young embryos of Nucula delphinodonta when taken
from the brood-sacs do not live well, and it is accordingly
difficult to determine how long the test is retained. As near
as could be judged, it seems to be retained about two weeks.
Its cells then begin to break apart near the posterior end of
the embryo, and many of them move toward the anterior
end, where they remain attached to the apical plate and the
stomodeeum (fig. 34). Sometimes most of the cells of the test
seem to thus accumulate at the anterior end, but they fre-
quently become detached and go to pieces before reaching
this position. In any case they, together with the apical
plate, and the stomodzeum, to the position of the future
mouth, are finally thrown away (fig. 35). In many cases the

338 GILMAN A. DREW.

process of casting occupies several and sometimes as many
as fifteen hours. The process is much more rapid for both
Yoldia limatula and Nucula proxima (2). It is quite
possible that the difference in the length of the time occu-
pied by the different embryos is connected with the differ-
ence in the conditions under which they develop.

Further study has tended to confirm my view that the test
should be regarded as the homologue of the velum of other
forms. In a former publication (1) I made the statement
that “in either Dentalium or Patella, if we imagine the
velum to be stretched posteriorly over the shell-gland dor-
sally, and the foot ventrally, so as to enclose the body, the
cesophagus will be pulled out into a narrow tube ventral to
the foot, and the position of the blastopore will correspond
to its position in Yoldia. Furthermore the position of the
foot and shell-gland will correspond, and the alimentary
canal will be bent in the same way.” This states the case
backward, and may be a little confusing. If we begin with
the condition found in Yoldia and Nucula, and imagine the
test to shrink until it consists of a band of ciliated cells sur-
rounding the embryo anterior to the mouth, the condition
would be comparable to that shown by embryos of Den-
talium and Patella.

As in the case of Yoldia, the closest resemblance to the
test, outside of the group,is shown by Dondersia. Although
Pruvot’s (15) account of these embryos is very short, and
only three figures are given, there is quite a striking external
resemblance. In both cases the surface cells are arranged
in five rows, all of which bear cilia. They are both provided.
with apical plates, and with both the test is finally thrown
away. The bodies of the embryos of Dondersia protrude
posteriorly during development. A. slight posterior protru-
sion of the body of Nucula sometimes takes place through
the opening dorsal to the blastopore.

The resemblances shown by embryos of Dentalium
(8 and 9) and Patella (12) are not so striking, but they are.
somewhat similar. The apparent posterior protrusion of the

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 339

body in each of these forms is such as might be produced if
the body of Nucula were to grow and protrude to a corre-
sponding extent. In such a case the test of Nucula would
occupy a corresponding position to that occupied by the
velum in the other forms.

Apical Plate.

At an early period the cells of the apical plate cannot be
distinguished from those that form the test, but as develop-
ment proceeds they become marked off as a rather definite
plate at the anterior end of the embryo (figs. 11, 15, and 24,
ap.). This plate is relatively large and thick, and extends
posteriorly as far as the stomach. The cells from which the
cerebral ganglia are formed le ventral to it (figs. 15 and 24,
cg.) Beneath the test the cerebral ganglia and the apical
plate are in contact.

The cells of the apical plate are evenly ciliated with short
cilia, like those borne by the test cells (fig. 15). In this
respect this species differs from both Nucula proxima and
Yoldia limatula. Both of these forms have long apical
cilia (Text-figs. G and H) that during life are bunched
together (Text-fig. F). Nucula proxima has an apical plate
that in extent may be compared to that of Nucula del-
phinodonta, but the apical plate of Yoldia is comparatively
very small. The short, diffuse cilia on the apical plate of
Nucula delphinodonta are probably the result of the
conditions that make active locomotion at this stage both
unnecessary and dangerous. (See what is said regarding this
under the head of Test.) Certainly most of the apical plate,
and probably all of it, is cast away when the test is shed (figs.
34 and 35).

Shell.

Some lime salts are deposited soon after the cuticle of the
shell begins to be secreted, which takes place about the time
that the lobes of the mantle begin to form (fig. 20). When
the test is shed (figs. 34 and 35), the shell-valves are white,

340 GILMAN A. DREW.

glossy, and quite transparent. They do not correspond to the
adult valves in shape (fig. 50), and they do not have the long,
straight hinge-line of the prodissoconch of Yoldia (‘Text-fig. K).
‘he hinge-line is not very definitely marked off from the rest
of the shell, but it can be distinguished as a nearly straight
or slightly curved portion on the dorsal margin (fig. 36). The
difference in the shape of the prodissoconches of Nucula and
Yoldia is quite marked, more marked than might have been
expected for forms so closely related, when there is so much
resemblance between the prodissoconches of many Lamelli-
branchs (6). They both conform to the same type, how-
ever.

At first the valves are thin and have neither cartilage pit
nor teeth. Soon after casting, a little knob of cartilage (fig.
36, ca.) makes its appearance near the middle of the hinge-
line. The teeth do not form until a much later stage (fig. 46).
About the time that the fifth pair of gill plates are formed, a
tooth appears on each valve in front of the cartilage pit.
This is soon followed by another, which is added anteriorly.
The teeth posterior to the cartilage pit begin to appear about
the time that the third tooth anterior to the cartilage pit is
formed. New teeth in the posterior series are added pos-
teriorly. Only about half as many teeth are formed posterior
to the cartilage pit as anterior to it. Apparently as long as
the shell continues to grow in size new teeth are added.
Shells of fully grown specimens are about 4 mm. long, but
they sometimes occur nearly 5 mm. long.

Each shell-valve is very convex (figs. 50 and 51), slightly
oblong, and moderately thick. The beaks are directed pos-
teriorly and placed far back on the shell. This gives an
appearance quite the reverse of most Lamellibranch shells,
which have the beaks nearer the anterior than the posterior
ends, and directed forward. The cuticle of the shell differs
in different specimens from horn colour to dark brown or
nearly black. It may be considerably broken near the beaks,
but it is generally quite perfect and smooth. Unlike most
Lamellibranchs, the shells of this species contain so much

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 341

animal matter that they retain their forms after the lime salts
have been dissolved away. ‘I'he material is quite tough, and
frequently causes much trouble in cutting series of sections.
Each valve of the shell of fully adult animals has from ten
to twelve teeth in the series anterior to the cartilage pit and
five or six posterior to it (figs. 50 and 51). All of the teeth
are more or less conical, pointed, curved toward the dorsal
margin of the shell, and distinctly grooved on the side away
from the cartilage pit. [Hach series of teeth forms a ridge
some distance from the dorsal margin of the shell, which
disappears dorsal to the adductor muscle-scar. The teeth of
the two valves interlock so completely that it is frequently
quite impossible to separate the valves without breaking
some of them. The cartilage pit is large and deep. The
adductor muscle-scars and pallial lines are faintly marked.

Mantle.

The shell-gland is formed early. About the time that the
gut is formed it consists of a number of large cells that lie
near the blastopore, on what may be distinguished as the
dorsal side of the embryo. Its cells do not seem to bear
cilia, but only preserved material was at hand for the deter-
mination of this point. The surrounding ciliated cells, those
that form the test, begin to grow over the shell-gland from
the sides and anterior end (figs. 10, 11, 12, and 13, sg.). At
the same time the shell-gland flattens slightly, and the cells
along its margins push up and form a slight ridge, that keeps
the surface of the shell-gland separated from the overgrowing
test. Soon after the shell-gland is covered by the test, it
arches dorsally, and the two come to he close together (figs.
17 and 18, sg.). As the embryo flattens laterally the shell-
gland arches dorsally still more (fig. 20), and a space appears
between it and the intestine. This space seems to be formed
by the multiplication and flattening of the cells of the shell-
gland, which arches dorsally and becomes separated from the
intestine. Lateral folds (fig. 20, m.), the beginnings of the

342 GILMAN A. DREW.

mantle lobes, are soon formed. About this time the shell
cuticle is secreted and some lime salts are deposited.

Soon after casting has been completed, swellings, the
beginnings of the gills (fig. 39, g.), are formed near the
posterior margin of each lobe of the mantle. The gills are
thus formed as appendages of the mantle.

The mantle now has the adult structure and appearance,
except that at a later stage a portion of its inner epithelium,
and of the epithelium covering the suspensory membranes of
the gills, becomes converted into the hypobranchial glands.
These glands are present in both sexes, but just before the
breeding season they are much better developed in the
females than in the males, and there is considerable evidence
that they furnish most, if not all, of the material from which
the brood-sacs are formed. The margins of the mantle lobes
remain thickened and contain the glands that secrete the
cuticle of the shell. Some cells along the ventral and pos-
terior borders of the mantle lobes bear cilia. Pallial muscles
are attached to the shell-valves, and extend out to the margins
of the mantle. ‘These serve to retract the margins of the
mantle when the shell is tightly closed.

Foot.

At a stage such as is represented by figs. 14 and 15, a
group of cells lie between the gut and the stomodeum.
These cells, together with the ectodermal side walls, are
concerned in the formation of the foot. The side walls of
the foot are continuous with the general ectodermal covering
of the body beneath the test. The cells lying between the
gut and the stomodzum are apparently mesodermal, and en-
close a small space (figs. 15 and 24). The shell-gland spreads
out, arches dorsally, and folds laterally to form the mantle,
and a large space is left between it and the stomach and in-
testine (figs. 20, 24, and 26). In some transverse sections
the space between the stomodeum and the intestine, and the
space dorsal to the intestine, are more or less connected.
This connection may be due to shrinkage caused by pre-

THK LIFE-HISTORY OF NUCULA DELPHINODONTA. 343

servatives, but it seems probable that the two spaces are
naturally more or less definitely connected around the sides
of the stomach and intestine, and that they may be regarded
as a single cavity-—a schizoceele.

The side walls of the foot join the stomodzum, and are not
continuous with each other ventrally (figs. 20 and 28). Just
before the test is cast away they begin to unite dorsal to the
stomodeum, and the stomodeum becomes comparatively
free. This change begins at the posterior end of the foot
and works forward.

The process of casting is slow, and includes a large part
of the stomodeum. When it is completed, the foot consists
of a small mass of tissue, lying ventral to the stomach and
intestine (figs. 34 and 35, f.). It is not capable of executing
movements, and for a period of about a day, or even longer,
the embryo lies perfectly quiet with the shell-valves tightly
closed. At first I supposed that this comparatively im-
mature condition of the foot at the time of casting was con-
nected with the protected life of the embryo. The foot of
Yoldia executes movements before the test is shed, and bur-
rowing is begun almost as soon as the process is completed.
It seemed natural to conclude that the greater development
of Yoldia at this time depended upon the necessity for self-
preservation. It was surprising, then, to find that at a cor-
responding time the foot of Nucula proxima is no better
developed than is the foot of Nucula delphinodonta.
This seems very remarkable to me, for Nucula proxima
inhabits muddy and shelly bottoms over which flow quite
strong tidal currents. Under these conditions it would seem
that such perfectly helpless embryos would surely perish.

The foot of Nucula delphinodonta grows rapidly, and
by the second day (fig. 36) performs feeble movements, but
it is not thrust out of the shell for some time. It becomes
provided with cilia (figs. 39 and 40), but they are not as
powerful as those on the foot of Yoldia (Text-fig. K), and
they are of but little service in locomotion.

‘he first movements of the foot are feeble twitches. These

344 GILMAN A. DREW.

in time become more frequent and powerful. Finally the foot
is thrust out of the shell, stretched ventrally and anteriorly,
swelled up at the end, and held more or less rigid while the
cilia vibrate. After being held in this position for a few
seconds it is withdrawn, either to remain quiet for some time,
or to be immediately thrust out again. The earlier move-
ments are not very energetic, and as the side flaps have not
been formed, they are not like the movements of the adult.

The first indication of the side flaps consists of a slight
longitudinal groove on the mid-line of the ventral surface of
the foot. On each side of this groove the foot grows to form
flaps (figs. 40 and 61) that le side by side.

Soon after the test is shed, a rounded knob develops on
the postero-ventral portion of the foot (fig. 36). This grows
quite rapidly, and forms the prominence that appears like a
heel (figs. 40 and 41). It soon stops its rapid growth, and
in the adult is comparatively small (figs. 48 and 49). In
this species it is comparatively much larger in the adult
animal than in any of the other species that I have studied.

The side flaps at first have smooth margins (fig. 40), but
papillae soon begin to be formed (fig. 41). The anterior
papillae are formed first, and new ones are added posteriorly
as the foot grows, until as many as thirteen pairs have been
formed (fig. 48). ‘The number differs with the size of the
individual. Sexually mature specimens may be found with
no more than eight pairs. The papille are large, conical,
more or less pointed, and very sensitive to mechanical stimu-
lation. ;

The movements of the foot of this species when compared
with the movement of the foot of Yoldia are very deliberate,
but the foot is so large, and the muscles so powerful, that
burrowing is quite rapid. Individuals of this species seem
normally to live entirely covered by mud, in which they
wander around by slow thrusts and retractions of the foot.
Specimens do not seem to come to the surface of the mud to
remain for any considerable time, and it seems probable that
the greater part of the lives of individuals are passed beneath

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 345

the surface of the mud. Observations made on specimens
kept in dishes of sea water in which there was no mud show
that individuals of this species execute movements very
similar to those executed by Yoldia (1), but that in all cases
they are much more deliberate (3). Leaping movements are
absent, but slow thrusts with the flaps extended may fre-
quently be observed. In former publications attention has
been called to the characteristic movements of the foot, and
they need not here be redescribed (1 and 8). Asin the case
of other members of this group, the movements of burrowing
are very effective. The somewhat spherical shape of the
shell, and the relatively large size of the foot, make it
possible to raise the shell from the bottom of a dish, and
occasionally to keep it balanced for a few seconds over the
expanded foot. My observations lead me to believe that the
animals never creep.

As in Yoldia, the foot is supplied with complicated and
powerful muscles (1 and 8). It is attached to the shell by
three pairs of muscles, and by a few fibres that lie ventral to
the genital mass and liver. The posterior pair of foot
muscles is very powerful. These muscles are attached to
the shell at the bases of the teeth, just anterior to the pos-
terior adductor muscle, and extend along the sides of the
foot in an anterior and ventral direction. They are the
powerful retractor muscles of the foot. Fibres from them
are extended into the muscular flaps, and are important in
spreading them apart. .

The two anterior pairs of foot muscles correspond to the
three anterior pairs of foot muscles in Yoldia. - They are in-
serted on the shell close together along the bases of the
teeth, just posterior to the anterior adductor muscle. he
most anterior pair has much the same distribution as the two
anterior pairs in Yoldia, and in some cases each muscle seems
to be slightly separated into two near its origin. They spread
out along the sides of the foot, and are distributed to its
posterior and ventral portions. These muscles seem to be
closely connected with the muscle-fibres that are attached

‘VOL. 44, PART 3.—NEW SERIES. Z

346 GILMAN A. DREW.

along the sides of the shell ventral to the genital mass and
liver. The more posterior of the two anterior pairs of foot
muscles passes between the pair just mentioned, and is dis-
tributed to the anterior and ventral portions of the foot.

In the foot all of the muscles are closely bound together
by their own fibres and by interlacing fibres, so that many
movements occur that cannot be explained by direct pulls of
one or more muscles. It should constantly be borne in mind
that the attachments of the fibres are all along the sides of
the foot, and that many, if not most of the muscle-fibres pull
from one portion of the body-wall to another, without chang-
ing the relation of the body to the shell. ‘Thus the flaps can
be spread apart after the shell has been removed. By com-
pressing the blood contained in the large spaces of the foot,
many movements, especially those connected with protruding
the foot, may be performed.

As in the case of Yoldia, the foot muscles are so large that
they are attached along a considerable portion of the dorsal
surface of the shell. I regard this as the result of the size
of specialised muscles, and do not agree with Pelseneer (18)
that it should be regarded as a primitive character.

Byssal Gland.

The byssal gland is formed as an invagination, just ante-
rior to the posterior projection of the foot, about the time
that the side flaps of the foot begin to form (fig. 39). Although
there is but a single external opening, the gland at first con-
sists of right and left pouches that extend into the foot near
its posterior side. ‘The cells forming the upper portion of
the gland soon become somewhat swollen, and do not stain
very well with hematoxylin. The lumen of the gland soon
shows traces of a secretion, but the secretion has never been
seen protruding from the duct.

The gland soon enlarges to a remarkable extent, becomes
quite irregular, and the paired appearance disappears. At
this stage, which extends from about the time that the gill
acquires its third lobe (fig. 41) until about the time that it

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 847

acquires its fifth pair of plates (fig. 45), the byssal gland
extends through a considerable portion of the foot, and in
the posterior side‘ of the foot it may extend to a position
somewhat dorsal to the pedal ganglia. The cells of the gland
during this stage are greatly swollen and vacuolated, and
have thin, almost indistinguishable walls. They are crowded
together so as to almost obliterate the lumen of the gland.
The result is that stained sections of the gland have the
appearance of a fibrous or reticular mass that is so mixed
up as to be hardly intelligible.

As the embryo gill begins to acquire its fifth pair of plates
the byssal gland generally becomes less extensive. In the
adult it is reduced to a small pouch (fig. 48, bg.) that opens
in the median groove of the foot, just anterior to the heel-
like projection. ‘The dorsal, blind end of the pouch cousists
of comparatively large cells with small nuclei, and seems to
contain some secretions. They are not generally distended
with secretion, and the duct is generally quite empty. Nothing
comparable to byssal threads have been observed. Towards
the opening of the gland the cells become smaller and bear
cilia.

I have described the adult condition that seems most fre-
quently to prevail. In a few specimens the gland cells are
much shrunken, and seem to contain little or no secretion.
In some specimens of Nucula proxima the gland is more
extensive and the cells are greatly distended. This would
seem to indicate that the gland is functional, but not as an
organ for the formation of threads. The present use of such
a secretion is problematical.

It is very natural to compare this gland to the mucus-
secreting glands of Gastropoda, but there seems to be little
direct evidence that they are homologous.

Alimentary Canal.

There is a stage when the embryo resembles an epibolic
gastrula (fig. 4). A pouch appears between the large cells, at
a point corresponding to the asterisk, that seems to be formed

348 GILMAN A. DREW.

by the separation of some of the larger cells, accompanied
by their division into smaller cells. This pouch is the first
indication of the alimentary canal (fig. 8, mg.). Partly by
the division of cells forming it, and partly by the addition of
ectodermal cells around the blastopore, the gut is carried
further into the interior (figs. 9 and 11), and comes to le at
the end of a narrow tube, the stomodeeum (fig. 15, std.). The
blastopore never closes, so from its first appearance the
stomodgeum is connected with the gut.

The blind end of the gut turns dorsally beneath the shell-
gland (fig. 11, mg.), and soon begins to grow posteriorly
(fig. 15, int.). It finally comes to the surface at the posterior
end of the embryo at a point ventral to the shell-gland and
dorsal to the blastopore (fig. 24), where the anus is formed.
The anus does not open directly to the exterior, but opens
into the mantle chamber near an external opening in the test.

The alimentary canal at this stage consists of three distinct
parts (fig. 24) : aslender tube, the stomodzeum (std.), opening
through the blastopore and extending forward nearly to the
apical plate that is formed from the ectoderm ; a rather thick-
walled stomach (sto.) that lies dorsal to the anterior end of
the stomodewum, and ventral to the shell-gland; and the
intestine (int.), which joins the posterior end of the stomach,
and at first has rather thick walls.

Dorsal to the stomach and intestine, between them and the
shell-gland, a cavity makes its appearance that communicates
by lateral passages with another cavity that lies ventral to
the stomach and intestine, between them and the stomodzum.
The ultimate fate of these cavities has been referred to in the
sketch of the life-history, and in connection with the foot,
and will again be referred to in connection with the peri-
cardium. For some time they are rather large, and a portion
of the alimentary canal is left quite free from surrounding
tissue, except where it seems to rest on the walls of the
developing foot (fig. 28). A short time before the test is shed
the liver pouches make their appearance (fig. 25, l.). These
are formed from the sides of the anterior end of the stomach.

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 349

The cells of the epithelial walls of the stomach are of two
kinds. Those at the anterior end of the stomach carry com-
paratively few cilia, and those at the posterior end carry
many cilia. At this stage some of the cells on the dorsal
side of the stomach, near its anterior end, begin to secrete a
mucus-lke material that extends posteriorly in the lumen
of the stomach as a small rod that probably represents the
crystalline style (fig. 26). Later the posterior portion of the
whole dorsal division of the stomach (the part that at this
stage is the dorsal part of the anterior portion) is given over
to secreting this material, but a definite rod may not be
present.

About the time that the embryo casts its test the stomach
grows dorsally into the space above it, so that a ventral bend
is formed where the stomach joins the intestine (fig. 26).
This is the beginning of the abrupt bend that marks this
portion of the alimentary canal in later life. Two fibre-like
cells stretch across the dorsal space from the anterior end of
the stomach to the mantle (fig. 25). Their position suggests
that they may aid in moving the stomach into the more
dorsal position, but there is no direct evidence that this is
the case.

When the test is cast away and the adductor muscle pulls
the shell-valves together, the stomach is crowded further into
the dorsal space, and the bend in the intestine becomes
more pronounced (figs. 34and 35). The same pressure appa-
rently causes the liver pouches to go to pieces. Their cells
become more or less separated, and fill the larger part of the
cavity dorsal to the stomach (figs. 836—89, z.). The same
changes have been noticed in embryos of Yoldia limatula
and Nucula proxima. In all of these forms the changes
occur in connection with the closing of the shell. Until the
test is shed, tissue lies between the valves of the shell so that
they cannot be shut together. When the tissue is removed,
and the shell is closed, there is no longer room for the liver
pouches to le on the sides of the stomach and retain their
original shape. ‘They are accordingly flattened and pressed

350 GILMAN A. DREW.

into the unoccupied space dorsal to the stomach. The cells
are no longer arranged to form definite walls (figs. 36—839, z.),
but later some of them seem to form liver pouches again
(figs. 42—44). A small portion of the space into which the
stomach and liver are crowded is not filled, and finally forms
the pericardium (figs. 39—41).

The rupture of the liver pouches leaves the dorsal part
of the stomach without side walls, and the dorsal wall is
commonly broken (figs. 37 and 39). The dorsal wall is formed
again before the liver pouches regain their cavities (fig. 40).
Some of the separated liver cells find their ways into the
open stomach (figs. 37 and 38), and together with mucus
practically fill it. For a period of two or three days after
casting, the animal is not active, and it is doubtful if it
feeds. At the end of this time the walls of the stomach
begin to re-form, and the mass of material that has filled the
stomach has largely disappeared. For a number of days
the liver does not form definite pouches. The rounded and
scattered cells are finally collected into two masses (figs. 40
and 42) that finally form new liverlobes. The left is slightly
larger than the right mass, but the masses are more nearly
equal in size than is the case with Yoldia. In both cases it
seems that the difference in the size of the two liver lobes
causes the developing loops of the intestine to take up a posi-
tion on the right side. Cavities gradually extend out into the
liver pouches from the stomach (figs. 43 and 44). In the
formation of the cavities some of the rounded cells seem to
go to pieces in much the same way as they would if digested.

The elongation of the intestine that results in the forma-
tion of the loops begins about the time that the embryo
acquires its fourth pair of gill plates (fig. 45). This elonga-
tion carries the portion of the intestine that lies dorsal to the
posterior adductor muscle toward the posterior wall of the
stomach and nearer the dorsal margin of the shell. The end
of the loop is forced over to the right side, and is extended
anteriorly nearly to the anterior wall of the stomach. At
this stage (Text-fig. O) the loop of the intestine is much like

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 351

the loop that occursinadult Yoldia limatula (Text-fig. L).
The dorsal bend now begins to elongate and project ante-
riorly (Text-fig. P). This continues until the loop is ex-
tended between the limbs of the loop that was made first
(Text-fig. Q). The lower limb of the loop begins to elongate
(Text-fig. R), and the adult condition is soon reached (Text-
fig. S and fig. 48).

‘The heart makes its appearance some time before the loops

fia M fieN 4120

41g S

Text-ries. M, N, O, P, Q, R, anp 8.—Stages in the development of the
loops of the intestine in Nucula delphinodonta.

of the intestine begin to be formed (fig. 41). From the first
appearance of its cavity the heart surrounds the intestine.
This condition continues for a long time, until the loops of
the intestine have been formed, and, in fact, until after the

352 GILMAN A. DREW.

animal has reached sexual maturity. At first the intestine
passes through the middle of the heart (fig. 67). The sides
of the heart seem later to be forced dorsally by the growth
of the kidneys, and the intestine becomes applied to the
ventral wall of the heart. By the continued growth of the
kidneys the ventral portion of the ventricle is drawn out into
a trough, in which the intestine lies (fig. 68). The growth is
continued until the trough is considerably deeper than the
width of the intestine. By gradually closing in dorsal to
the intestine at the anterior and posterior ends the trough is
shortened, and the intestine finally becomes free from the
heart and lies ventral to it (fig. 69).

In the adult animal (fig. 48) the esophagus is a rather
broad and long, nearly cylindrical tube, that opens between
the palps just posterior to the anterior adductor muscle. I
find no indication of anything that can be interpreted as
salivary glands at any stage in the development (18).
Throughout its length it is evenly ciliated and quite devoid
of ridges. The corners of the mouth are continuous with
the groove between the two labial palps. The stomach is
large, somewhat spindle-shaped, and extends from near the
dorsal margin of the shell to the level of the pedal ganghia.
Near its middle there is a nearly complete ridge of elongated
epithelial cells, and frequently a more or less well-marked
external groove that divides it into a dorsal and a ventral
portion. The posterior and part of the lateral walls of the
dorsal portion of the stomach are formed by long and slender
epithelial cells that stain but slightly. They secrete a
mucus-like material that stains deeply, and probably corre-
sponds to the crystalline style. In adults this secretion
seldom takes the form of a rod, but in embryos a rod is
commonly present (figs. 26, 28, 30, and 64). The remaining
cells in the dorsal portion of the stomach are short, stain
deeply, and are evenly ciliated. The ducts from the liver
open in the dorsal end of this portion of the stomach. The
epithelial cells of the ventral portion of the stomach are
short, stain deeply, and carry a quantity of short cilia.

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 353

Leaving the ventral end of the stomach, the intestine bends
dorsally, and follows the posterior walls of the stomach nearly
to its dorsal end. Here the loops already described are
formed. From the loops the intestine passes posteriorly
ventral to the heart, bends around the posterior side of the
posterior adductor muscle, and opens in the mantle chamber.
It is composed of short ciliated cells that stain deeply. Its
lumen varies greatly in size, according to the amount of
matter it contains.

Labial Palps.

Soon after the embryo acquires its second gill lobe the
epithelium around the mouth, and for a short distance along
the sides of the body, becomes ciliated (fig. 41). This cilia-
tion precedes the formation of the palps, and, to a certain
extent, marks out the region where they will form. The
cilia are more numerous immediately anterior to the mouth
than they are immediately posterior to it, and they soon extend
along the sides of the body for about half the width of the
foot. The position of the ciliated patches on the body-wall
is such that the dorsal portion of each tends to lie horizon-
tally, and the ventral portion tends to the vertical position
(fig. 62, lp.). The groove thus formed becomes the groove
between the outer and the inner palps. The portion above
the groove forms the outer palp, and that below the groove
the inner palp. This is accomplished by the growth and
folding of the body-wall. The outer palp begins to grow
first, and in such a way that the line marking the dorsal
limit of the cilia becomes the free margin of the palp. This
leaves the inner surface of each outer palp covered with cilia,
and the outer surface unciliated. The two outer palps are
continuous anterior to the mouth, where they form a slight
ridge (figs. 54 and 63),

For some time after the outer palps form folds, the inner
palps are represented by ciliated ridges (fig. 54), that reach
some distance beyond the posterior ends of the outer palps.
These ridges grow so that the lines marking the ventral limit

354 GILMAN A. DREW.

of the cilia become the free margins of the inner palps.
The two inner palps are continuous posterior to the mouth,
where they form a slight ridge (figs. 55 and 63). Like
the elevation anterior to the mouth, this never becomes
prominent.

The inner surface of each outer palp becomes folded near
its anterior end to form ridges and grooves (fig. 55), and the
postero-ventral portion protrudes to forma lobe. This lobe
is the beginning of the formation of the palp appendage.
The edges of this lobe soon begin to thicken, and a groove
is left between the ridges thus formed. ‘l'his is accompanied
by a considerable growth in length (fig. 56, pap.). At this
stage of development the palp appendage is seen to corre-
spond to two of the ridges on the general surface of the palp,
with a groove enclosed between them.

Posterior and dorsal to this appendage another smaller
appendage is formed (fig. 56). This is also on the outer palp,
and consists of two ridges with a groove between them. It
never grows to be very long, but resembles the large appen-
dage that lies ventral to it in its formation.

As development proceeds the larger appendage (fig. 56,
pap.) twists, so that its groove opens dorsally and posteriorly
(fig. 57, pap.), and the smaller appendage twists so that its
groove opens ventrally. This double twisting brings that
portion of the small appendage that was dorsal nearly or
quite in contact with that portion of the large appendage
that was ventral, so that for a short distance the two grooves
together form a tube that opens anteriorly between the two
palps (fig. 57). During the development of the palp appen-
dages both outer and inner palps have grown to be quite
large, and their ciliated surfaces have been thrown into
series of ridges and grooves.

The palps on each side of an adult animal consist of two
large, somewhat triangular folds of tissue (fig. 48, lp.), united
to each other along their dorsal margins, and suspended
from the body-wall by a thin membrane. The onter palps
on the two sides of the body are connected in front of the

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 350

mouth by a small ridge that occupies the position of an
upper lip. In the same way the inner palps are connected
by a ridge posterior to the mouth that is comparable to a
lower lip. The corners of the mouth are continuous with the
space between the two palps of each side. The opposed
surfaces of the palps are densely ciliated, and thrown into a
series of ridges and grooves that tend to lie opposite each
other on the two palps. Near the free margins this arrange-
ment may be considerably broken. Large blood-spaces
follow along these ridges. Each outer palp is supplied with
two grooved appendages that originate near its dorsal
margin. The most ventral of these appendages (fig. 48, pap.)
can be extended far beyond the margin of the shell, and is
used to elevate mud with the contained food. The dorsal
appendage sets over the anterior end of the groove of the
ventral appendage, and with it forms a short tube that opens
between the palps. Hach ventral palp appendage is supplied
with longitudinal muscles (fig. 66, Jm.), that are continued in
from the body-wall; with a large nerve (pz.) that originates
in a cerebral ganglion, and runs posteriorly along the united
dorsal margins of the outer and inner palps; and with a
continuous blood-space (bs.). The epithelium lining the
groove of the appendage is very thick, and is densely covered
with cilia. The nuclei of these epithelial cells are very long
and slender. The muscles in the palp appendages are so
placed that their contraction causes the appendages to curl,
as shown in fig. 48, pap.

It is not easy to observe individuals of this species while
they are feeding, as they normally live entirely covered by
the mud. If specimens are placed in a dish of sea water,
in which there is only a thin layer of mud, the action of
the palp appendages may be observed. It is well to use as
much mud as possible without affording the animals an
opportunity to bury themselves, and to use specimens that
have not been in mud for several days and are accordingly
hungry. The mud is passed along the grooves of the palp
appendages by the action of the cilia, and finally conducted

356 GILMAN A. DREW.

between the palps, where the cilia carry it to the mouth.
Very few specimens have shells that are transparent enough
to allow observation of processes carried on inside of the
shell, but there can be no doubt as to the path taken by the
mud after it has started up the grooves in the palp appen-
dages.

Feeding is much more easily observed in the case of
Yoldia limatula. In this species the animal has fre-

Trxt-Fic. T.—An adult specimen of Yoldia limatula as it appears while
feeding. es. Exhalant siphon. ¢s. Inhalant siphon. pap. Palp appen-
dages. sf. Siphonal tentacle.

quently as much as one third of the posterior end of the shell
above the mud while feeding (Text-fig. T). The palp appen-
dages are protruded, and one at least bends over and inserts
its tip in the mud. By the action of the cilia in the longi-
tudinal groove, large quantities of mud and food are elevated.
There is no reason to suppose that the palp appendages of

THE LIFB-HISTORY OF NUCULA DELPHINODONTA. 397

Nucula are not as effective as those of Yoldia, but the method
of life makes observation more difficult. As suggested by
Mitsukuri (11), it seems probable that the large palps with
their numerous large blood-spaces may be important respira-
tory organs.

Gills.

A short time after the embryo sheds its test, a portion of
each lobe of the mantle near its posterior border begins to
thicken (fig. 39, g.) and then to project anteriorly. These
thickenings are the beginnings of the gills. They grow
rapidly, acquire cilia, broaden dorso-ventraliy, and each
begins to divide into two lobes (fig. 40, g.). The formation of
the lobes is due to unequal growth more than to constriction.
Kach lobe is at first a little knob that is flattened slightly
laterally. As growth proceeds the ventral lobe broadens
and flattens along its anterior border preparatory to the
formation of another lobe. Coincident with these changes
in the ventral lobe, the dorsal lobe grows anteriorly, and
forms a rather long finger-like process or filament, that
closely resembles the filaments of the developing gills of other
Lamellibranchs (fig. 41, g.). New lobes are added to the
gill by the unequal growth and division of each ventral lobe
in its turn, and as the new lobes are formed the more dorsal
lobes lengthen.

Throughout life the gill occupies a decidedly dorso-ventral
position, but growth carries the ventral end some distance
toward the posterior end of the animal, so that the adult gill
lies somewhat diagonally (fig. 48,9.). In Yoldia (Text-fig. L)
the gills lie more nearly parallel to the long axis of the body.

The chitinous support of the gill makes its appearance
when the gill is still in the two-lobed condition. At first it
consists of a thin plate lying just beneath the epithelium on
the anterior border of the gill, and is continued from one
lobe into the other. Its ends lie near the anterior extremity
of each lobe. As the ventral lobe flattens the chitinous
plate is extended along its anterior border, so that with the

358 GILMAN A. DREW.

formation of the third lobe the plate is extended into it. In
this way, as new lobes are formed, the chitinous plate
is extended into each, and continues to be connected
throughout the length of the gill. As the lobes grow to
form filaments, the chitinous plates extend with them, and
each becomes trough-shaped with the open side of the trough
directed away from the corresponding lobe of the mantle.
Later the free edges of the trough are brought near to-
gether, and the support in each filament practically assumes
the form of a tube that extends out nearly to the tip of each
filament. The tubes that support the different filaments are
united at their bases, so the chitinous support is continuous
throughout the gill.

As the lobes elongate to form filaments, the cilia on each
becomes restricted, so that the side that is turned away from
the lobe of the mantle to which it is attached becomes quite
free from them. On the remaining sides the cilia are long
and powerful.

About the time that the fourth division of the gill is
formed the mantle begins to thicken at the bases of the fila-
ments, between them and the shell (fig. 45). These thicken-
ings are generally opposite the bases of the filaments, and
connected with them, but as there are sometimes more plates
on one side of the gill! of the adult animal than on the other,
the thickenings are probably not always formed in this
position.

They represent the beginnings of the outer plates of the
gill. he filaments, at the bases of which these thickenings
are formed, form the inner plates of the gull.

For a considerable time the outer plates remain much
smaller than the inner plates, and they never quite equal
them in size (fig. 53). As the outer plates of the gill are
formed, the chitinous support is carried out into them as
branches from the portion that runs lengthwise of the gill.

1 The term gilli is for convenience applied to the respiratory organ on one
side of the animal, although writers agree that it probably corresponds to tae
two gills found on each side of most Lamellibranchs.

YHE LIFE-HISTORY OF NUCULA DELPHINODONTA. 309

These branches become trough-shaped, with the open part of
the trough directed away from the inner plates. Finally, the
free edges of the troughs come close together, as described
in connection with the other set of filaments or plates.

The chitinous material at the bases of the two sets of
plates also becomes trough-shaped, and has the open portion
of the trough directed away from the plates. Thus the
chitinous support of the gill consists of two series of troughs,
bent so as to form tubes, each of which is connected by one
end to the side of a larger trough that runs lengthwise of the
gill. The whole might be compared to a large trough with a
series of spouts leaving each side, the individual spouts of
the two series being placed opposite each other. Later,
bridges are built across the main trough in the intervals
between the side spouts. ‘The whole system is in direct
communication with the blood-spaces of the gill, but
probably is not concerned with the circulation of the blood.

The two sets of plates do not lie parallel to each other,
but they grow away from each other at an obtuse angle.
The inner plates grow almost in an anterior direction, and
the outer plates grow laterally and a little posteriorly, so that
the angle formed by the two sets of plates on the two sides
of the gill is visible when the animal is viewed from the side.
The suspensory membrane, formed by the growth of the
mantle at the base of each gill, makes it possible for the gill
to take up this position.

The filaments begin to grow into flattened triangular plates
about the time that the fourth division of the gill is formed.
This is accomplished by slow, unequal growth, and throws
no light on the phylogeny of the gill. It seems to be a matter
of individual opinion whether each of the plates should be
considered to be homologous with a descending filament of
an ordinary Lamellibranch gill, or whether it should be con-
sidered to be homologous with both a descending and an
ascending filament.

The adult structure of the gill of Nucula has been so care-
fully and accurately described by others, that were it not for

360 GILMAN A. DREW.
the sake of completeness, it would not be necessary to
describe it here. Mitsukuri’s (11) description of the gill of
Nucula proxima holds good in all essentials for the gill of
this species, and since his description was published others
have verified and supplemented his results (7, 13, and 16)
until our knowledge of the structure is comparatively com-
plete.

The adult gill of Nucula delphinodonta is suspended
from the body-wall by a fold of tissue, the suspensory mem-

iim

Text-Fic. U.—A pair of plates from a gill of Yoldialimatula. 4s. Blood-
space. cr. Chitinous rod. //m. Lower longitudinal muscle. sw, Sus-
pensory membrane. «adm. Upper longitudinal muscle. v. Cut surface
of achitinous rod. y. Cut wall of the gill plate where it bends to join
the plate anterior to it.

brane (fig. 58, gs.), that was originally a fold on the inner
surface of the mantle lobe. The suspensory membrane
contains between its walls a large blood-space that communi-
cates near its anterior end with the auricles of the heart, and
throughout its length communicates with blood-spaces in the
mantle. At intervals it communicates with similar spaces in
the body proper. Unlike the suspensory membrane of Yoldia
(Text-fig. U), this membrane is not very muscular, but some

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 3861

muscle-fibres are always present. The epithelium covering
the outer surfaces of the suspensory membranes, those
surfaces that are turned away from the mid-line of the body,
is modified to form a portion of the hypobranchial glands.

Hach suspensory membrane bears two series of gill plates
that generally lie opposite to each other. Occasionally a
gill occurs in which there are more plates on one side than
on the other, but even in these cases the order is interrupted
only for a short distance. The number of plates differs with
the size of the individual, but about twenty pairs seems to
be common for well-grown specimens. Lach plate is thin
and triangular, and is composed of epithelial walls, between
which there are loose connective tissue, large blood-spaces,
and the chitinous framework. The epithelial walls on the
edges of the plates that are directed away from the suspensory
membranes are thickened and covered with strong: cilia.
This thickened ciliated epithelium extends between the plates
for a short distance, but most of the epithelium is quite thin
and destitute of cilia. he wall of each plate is continuous
with the wall of the plate that lies in front of it, with the
wall of the plate that hes behind it, and with the plate on the
other side of the gill that lies opposite to it. Near the border
furthest from the suspensory membrane, the opposing walls
of the two series of gill plates are separated so as to form a
large blood-space (fig. 55, bs.), that runs the whole length of
the gill. This space is continued as a narrow slit to the base
of the suspensory membrane. Thus the blood-space in the
suspensory membrane is in direct communication with the
blood-space of each plate, and in the gill the blood is free to
flow from one part to another.

The chitinous framework consists of a bridged trough that
occupies the bottom and part of the sides of the blood-space
that lies between the two series of plates, and of two series
of side spouts that project into the plates on the two sides,
and le in contact with the thickened epithelium. Although
the chitinous framework is arranged as a system of troughs
and spouts that, from their position, must be filled with

VoL. 44, PART 5,—NEW SERIBS. AA

362 GILMAN A. DREW.

blood, they are probably not directly concerned in the circu-
lation of the blood.

Between the chitinous trough and the suspensory mem-
brane there is a small bundle of muscle-fibres that are con-
tinued the whole length of the gill (fig. 53, Im.). This
bundle lies in the open part of the chitinous trough, and
probably corresponds to the large bundle that occupies a
similar position in the gill of Yoldia (Text-fig. U, llm.). A
second longitudinal bundle of muscles is found in the gill of
Yoldia (wlm.), but this does not seem to be present in this
species. A few of the muscle-fibres in the suspensory mem-
brane seem to be continued into the plates. They are not
numerous, and they have not been carefully followed. The
gill of this species is so small that it is not favourable for the
determination of minute details.

The gills probably act as respiratory organs, but their
small size, together with the blood-supply of other parts,
makes it seem probable that other organs, such as the mantle
and the palps, are also concerned in respiration. The opaque
character of the shells of adult animals makes it quite im-
possible to observe the normal movements of the gills. They
can be seen to move slightly, however, and it seems probable
that the suspensory membranes contract slightly at intervals.
Such movements would be useful in causing movements in
the contained blood, but they are not sufficient to cause
strong currents of water. The shape of the gills is not such
as would make them efficient pumping organs. (Compare
fig. 58 and Text-fig. U.) Inasmuch as these animals live
entirely covered by mud, the production of strong currents
of water could not be beneficial. As the animal wanders
around in the mud the feeces naturally drop out of the open
mantle chamber.

It would be a matter of some interest if the exact relation-
ship of the gills of Nucula and Yoldia could be determined.
It would seem to be a comparatively easy task to account for
the changes in the shape and structure of the gill of Yoldia
if we were to start with a gill such as has been described for

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 363

Nucula. The habits of Yoldia are such as to render the
formation of strong currents of water absolutely necessary,
for otherwise the mantle chamber would become clogged
with feeces and dirt. The gill of Yoldia might have been
perfected for pumping water from a Nucula-like gill. It
would, however, be equally easy to account for the reverse
modifications when we consider what the formation of strong
currents of water by an animal entirely covered by a soft,
slimy mud would mean. If we follow the generally accepted
theory of the gill, the former change would seem more likely
than the latter, though it is quite possible that nothing lke
a direct change from one to the other has taken place. The
generally accepted theory of the gill has grown up as the
result of structural and embryological considerations, and
but scant attention has been given to probable modifications
for the special purposes of the animals. Until we have
much more detailed knowledge regarding the habits of most
of the Lamellibranchs that have plate gills, and of some of the
supposed near relatives of these Lamellibranchs, it seems to
me that we lack the necessary data to give the derivation of
the gill with anything like accuracy. There is much in the
structure and embryology of Nucula that points to a gene-
ralised type, and in this much it seems natural to look at the
gills as primitive ; but the gills of Yoldia—its undoubted near
relative—are so remarkably well adapted for the performance
of a special function, that it hardly seems safe to regard them
as slightly modified gills until there are more careful observa-
tions on the habits of other forms. I recognise fully the
mass of evidence in favour of the primitive form of the plate-
like gill. My only plea is for caution.

Hypobranchial Glands.

The epithelium on the inside of the posterior end of each
lobe of the mantle, and on the outer side of a corresponding
portion of the suspensory membrane of each gill, is glandular,
and has been termed the hypobranchial gland. When these
glands are present in Lamellibranchs, their secretions seem to

364 GILMAN A. DREW.

correspond very closely to mucus, and they are generally
reterred to as mucus glands. During the greater part of
the year the hypobranchial glands of both sexes of Nucula
delphinodonta are rather small and inconspicuous. They
contain rounded or oblong masses of a refractive material
that takes no stain. The cells themselves are small, and do
not seem to be secreting actively. The hypobranchial
glands of specimens of males seem to have the appearance
that has been described, no matter what time of the year
they are collected. As the breeding season approaches, the
hypobranchial glands of the females become greatly dis-
tended with secretions. The rounded or oblong masses that
are common at other seasons of the year are now seldom
found, and the cells are packed full of rather large granules.
Immediately after the brood-sac is formed, the cells of the
hypobranchial glands appear shrunken and free from
granules, and the glands have the appearance of having
discharged their secretions. After examining a large number
of specimens, I have become convinced that the hypobranchial
glands furnish nearly all of the material from which the
brood-sacs are formed. Specimens kept in aquaria do not
form brood-sacs, and accordingly the processes of their
formation have not been observed, but it seems probable that
the secretions from the glands are passed posteriorly by cilia
on the mantle, and probably swelled out into a bubble by
the respiratory current of water. While the material is still
soft it adheres to the foreign bodies with which it comes in
contact.

Well-developed hypobranchial glands are present in only
a limited number of Lamellibranchs, and their special func-
tion is hard to determine. It is interesting to find that they
are concerned in the formation of the brood-sacs in this
species, but this is the first instance that has been reported
where such a sac is formed. It may be that other forms
that possess especially large hypobranchial glands will be
found to form similar brood-sacs, but this will not hold true
for all. Nucula proxima has rather large hypobranchial

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 365

glands, and I find that it does not form brood-sacs. Such a
case as this, where it is known that brood-sacs are not
formed, seems to indicate either that the glands have some
function to perform other than providing the material for
the formation of brood-sacs, and that Nucula delphino-
donta has adapted them to this purpose; that they are
retained from forms that originally formed brood-sacs, in
which case we must suppose that the arfcestors of all forms
that possess hypobranchial glands formed brood-sacs; or
that in forms where brood-sacs are not formed the glands
are, when present, mere vestiges, and are not now functional.

The latter explanation seems unlikely, as the glands of
Nucula proxima are better developed than vestiges are
likely to be. If the second explanation is accepted, we must
regard the rearing of embryos in brood-sacs as more primi-
tive, for this group at least, than throwing the eggs in the
water where the embryos have to take care of themselves.
From the standpoint of specialisation this seems to be very
unlikely, and the fact that the embryos of Nucula delphi-
nodonta possess tests that seem to serve no purpose, while
similar tests function as organs for locomotion in other forms,
points clearly to a condition when all of these embryos de-
pended on their own activities for protection. It seems
most likely that Nucula delphinodonta has made use of
already existing glands to furnish the secretions for the
formation of its brood-sacs, and that they may have other
functions to perform.

Pericardium.

A short time before the shell-gland begins to fold at the
sides to form the lobes of the mantle, a space appears
between the stomodeum and the gut, and a little later a
space begins to form between the shell-gland and the gut
(fig. 24). These two spaces are separated by the gut, but in
preserved material they are frequently connected around the
sides of the gut. While these connections may be due to

366 GILMAN A. DREW.

shrinkage caused by the treatment with preservatives, it
seems most likely that the spaces are normally connected
with each other. It is just possible that these spaces may be
traced back in their formation to the blastoccele, but it is
more probable that the blastoccele entirely disappears, and
that they represent a schizoceele, At first the space ventral
to the gut is larger than that dorsal to it, but the latter
grows as the mantle arches dorsally, and the ventral space
remains practically unchanged.

As the foot begins to take form the ventral space becomes
quite small, and about the same time that the embryo sheds its
test it disappears altogether. A short time before the test
is shed the dorsal space reaches its greatest size (figs. 25 and
26). About the time that the test cells begin to break apart,
the stomach is carried dorsally some distance into this space
(fig. 54). Two fibres, that in shape suggest muscle-fibres,
extend from the anterior end of the stomach to the mantle.
Their position suggests that they may aid in moving the
stomach dorsally, but of this I have no proof. As the
stomach moves dorsally they become shorter and thicker, but
there is no evidence that they are moving factors. Until
casting is completed, the apical plate and the stomodzeum lie
between the edges of the shell-valves, and keep them from
bemg closed. When they are removed, the contraction of
the adductor muscle closes the shell, and the body, which has
until now been lying between gaping valves, is made to
change its shape and position. The stomach and liver
pouches are forced into the dorsal space until the dorsal end
of the stomach comes in contact with the mantle.

‘his divides the space into anterior and posterior parts
(fig. 35). There is no longer room for the liver pouches to
retain their form and position, and as the body continues to
move dorsally they are flattened out and soon go to pieces
(fig. 36). Most of the cells that formerly composed their walls
become scattered and rounded, and the anterior space
becomes entirely filled (fig. 39). The posterior space, some-
what diminished in size, persists, and finally becomes the

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 367

pericardium (figs. 40 and 41). The pericardial space is not
smooth as it has been necessary to represent it in the figures.
Mesoderm cells project into it from the surrounding tissue,
and others lie comparatively free within it. As yet it lies
almost wholly dorsal to the intestine, but just before the
heart is formed it is extended beneath the intestine, and
begins to have’a rather definite epithelial lining. The epi-
thelial lining seems to be formed by the change in shape and
position of cells in the immediate vicinity. I find no indica-
tion that the pericardium originates as a pair of pouches, as
has been described by Ziegler for Cyclas cornea (20).

Vascular System.

Small connected cavities are present throughout the body
from an early time, but a true vascular system, with a heart
and anything like a definite circulation, is not to be distin-
guished until much later, and a closed system of vessels with
capillaries is never present.

The heart is formed about the time that the gill becomes
well divided into two lobes, or just before the third lobe is
formed. It seems to be formed by the hollowing out of a
strand of mesoderm that stretches across the pericardial
cavity. I have seen nothing that would indicate that the
heart has a double origin, as Ziegler has described for
Cyclas cornea (20). Mesoderm cells in the pericardial
cavity and along its walls arrange themselves to form a
strand that becomes hollow and begins to pulsate. From the
first appearance of its cavity the heart surrounds the intes-
tine (figs. 41 and 67). Most specimens show the heart col-
lapsed with its walls in contact with the intestine, but some
specimens have it distended with blood. In all cases it is
easy to determine that the heart is perforated by the intes-
tine, but it is especially evident in specimens where the heart
is distended. In most of these cases the intestine lies nearer
the ventral than the dorsal wall of the heart, and in many
cases it lies directly in contact with this wall. At this stage
the heart is not separated into auricles and ventricle (fig.

368 GILMAN A. DREW.

67,h.). It isin the form of a bent spindle, the two ends of
which communicate with the blood-spaces of the gill. The
larger median portion arches dorsally and surrounds the
intestine. Anterior and posterior aorte leave the heart, but
no attempt has been made to follow them, until the adult
stage is reached.

For a considerable time after its formation there is no
appreciable change in the heart. About the time that the
eighth pair of gill plates are formed it begins to be separated
into ventricles and auricles. The auricles are at first very
small and narrow. ‘They extend only a short distance from
each gill, and are separated from the ventricle by slight
constrictions. There has been no change in the relative
positions of the heart and intestine. At a slightly later
stage, when the gill has about ten pairs of plates, the ven-
tricle of the heart begins to change its shape. This seems to
be due to the growth of the kidneys, which push anteriorly
ventral to the pericardium. As the kidneys grow, the two
sides of the heart are pushed dorsally, while the middle part
of its ventral wall is held in its original position by the
intestine. Jn this way the ventral wall is pulled out into a
sort of trough in which the intestine lies (fig. 68, h.). Con-
tinued growth deepens the trough until it is considerably
deeper than the intestine is wide. ‘The heart gradually
closes in, dorsal to the intestine, at the anterior and posterior
ends of the trough, until it becomes free from the intestine,
and lies dorsal to it (fig. 69). This is a very slow process,
and is not completed until after the animal has reached
sexual maturity.!

The adult heart consists of a ventricle and a pair of auri-
cles, separated from each other by constrictions that are
much deeper on the dorsal than on the ventral surface
(fig. 69, h.). The openings between the auricles and the
ventricle are so small that they must be quite obliterated

1 Every specimen of Nucula proxima that I have examined has its
heart perforated by the intestine. ‘The specimens are all of good size, and
many of them are the same ones from which I obtained eggs and sperm.

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 369

during contraction. A band of muscle occurs near the end
of each auricle, that keeps the blood from flowing back into
the spaces of the gills. Hach auricle is somewhat conical,
small where it joins the gill, and considerably enlarged at
the end next to the ventricle. The ventricle is swollen at the
ends next to the auricles, and flattened over the intestine.
The swollen ends of the auricles and the corresponding ends
of the ventricle make right and left enlargements that
superficially might be mistaken for two hearts.

A blood-vessel leaves the anterior end of the ventricle on
the left side of the intestine, and not in contact with it.
Another blood-vessel leaves the posterior end of the ventricle
above the intestine and in contact with it. The anterior
vessel is somewhat larger than the other. It runs forward
over the dorsal end of the stomach and sends branches to
the liver and genital organs, to the stomach and loops of the
intestine, to the foot, to the labial palps, and to the anterior
portions of the lobes of the mantle. The vessel that leaves
the ventricle posteriorly is at first dorsal to the intestine, but
it soon becomes ventral to it, and is distributed to the pos-
terior part of the body.

All of the blood-channels seem to end in rather large
connected spaces, that ramify throughout the body. The
course of the blood cannot be traced in these spaces. The
blood-spaces of the foot, beside providing for the ordinary
blood-supply, serve as reservoirs in which blood can be
forced to extend the foot. By suppressing some channels
and squeezing blood into others different results may be
obtaied. Blood must undergo respiratory changes in the
gills, the mantle lobes, and the palps.

The opinions of writers on Lamellibranch morphology,
regarding the primitive form and position of the heart, are
very different. Milne-Kdwards (10) thought that the double
appearance of the heart of Nucula and Arca pointed toward
a primitive condition in which the heart was double. Thiele
(19), basing his conclusions on Ziegler’s observations on the
formation of the heart of Cyclas, holds that the heart was

370 GILMAN A. DREW.

probably originally a double organ, and that upon uniting in
the median line it has taken up the various positions in regard to
the intestine. Grobben (5) considers the single heart primi-
tive, and thinks that the double condition is the result of
changes in the position of retractor muscles. Pelseneer (13)
and others, depending largely upon the position of the heart
in Nucula and Arca, have considered the dorsal position of
the heart to be the primitive position. Stempell (17) rightly
holds that the ventral position of the heart of Malletia
chilensis destroys the oundation of Pelseneer’s reasoning,
inasmuch as Nucula and Malletia are closely related forms.
Stempell apparently considers the perforated heart to be
the most primitive. From this position the heart may become
dorsal or ventral to the intestine by a comparatively simple
process.

The development of the heart of Nucula seems to indicate
that the perforated heart is more primitive than the dorsal
heart in this group. While,as Stempell holds, it seems most
reasonable to consider a perforated heart that may become
either dorsal or ventral by comparatively simple changes as
more primitive than either a dorsally or a ventrally placed
heart—where, in order to reach the opposite extreme, the
heart would have to enclose the intestine, and then become
free on the other side,—there is still nothing to prove that the
ventral position of the heart is not primitive. The develop-
ment of the heart of Malletia would accordingly be of con-
siderable interest.

As neither the pericardium nor the heart of this Lamelli-
branch seems to be formed as a paired structure, there is
nothing here to further the view of Thiele (19) that the
position of the heart in regard to the intestine depends
simply upon the position of two lateral hearts, that may, as a
matter of convenience, fuse dorsally, ventrally, or around the
intestine.

Nervous System.

The cerebral ganglia are formed in direct contact with the

apical plate. The cells from which they originate can first

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 371

be distinguished as a group soon after the surface cells that
form the test become ciliated (fig. 15, cg.). They frequently
remain as surface cells for some time, and they doubtless
originate as surface cells in all cases. The group of cells is
not distinctly paired, and does not invaginate as it does in
Yoldia. Each cell becomes much larger at the inner end than
at the end that comes to the surface. Although a test cell les
between the cerebral ganglia and the apical plate they still

T'ext-rig. V.—Transverse section of a forty-five hour embryo of Yoldia
limatula, taken through the cerebral pouches. eg. Cerebral pouches.
mg. Anterior wall of the mid-gut. ¢. Test.

remain in contact beneath this cell. Little change occurs in
the appearance, size, or position of the cerebral ganglia until
the test is cast away, and until then no other part of the
nervous system can be distinguished.

When the test cells break apart and accumulate near the
anterior end of the embryo (fig. 34) a portion of the body of
the embryo is carried dorsally at the expense of the large

$72 GILMAN A. DREW.

dorsal space. This dorsal movement includes the cerebral
ganglia (cg.). When casting is completed, and the valves of
the shell are closed, a further dorsal movement occurs, that
results in the filling of the greater part of the dorsal space.
This movement places the cerebral ganglia in position poste-
rior to the anterior adductor muscle (figs. 85 and 36, cq.).
The foot now begins to grow quite rapidly, and the pedal and
visceral ganglia begin to form (fig. 36, pg. and vg.). Both
pairs of these ganglia are formed as thickenings of the
surface ectoderm. The thickenings that give rise to the
pedal ganglia begin to form first, but both pairs of ganglia
are in process of formation at the same time. Owing to the
character of the embryonic tissue it is very difficult to
determine how the commissures that connect the ganglia
arise. ‘They are first found very close to the surface, almost,
if not quite, in contact with the ectoderm. Later they sink
deeper into the body. ‘The cerebro-visceral commissures are
quite thick, and differ from the cerebro-pedal commissures in
having much the same structure as the ganglia themselves.
In the earlier stages I have been able to demonstrate only a
single cerebral origin for each cerebro-pedal commissure.
This may be due to the difficulty of tracing commissures in
embryonic tissue. Later stages show two separate origins
very distinctly.

The double origin of the cerebro-pedal commissures has
been regarded by Pelseneer (18) as an indication of the
presence of cerebral and pleural ganglia in each anterior
nerve-mass. Furthermore, Pelseneer and others find that
each mass is divided by a constriction into two rather distinct
parts. I have not been able to satisfy myself that there is a
distinct separation into cerebral and pleural ganglia, either
in this or the other forms that I have studied.

The cerebral and pedal ganglia are about equal in size, but
they differ in shape (fig. 48). The visceral ganglia are
smaller than the cerebral ganglia, but compare pretty well
with them in shape. Hach cerebral ganglion is large at its
anterior end, and tapers posteriorly into the cerebro-visceral

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 373

commissure. The commissure that connects the two cere-
bral ganglia is a broad, somewhat flattened band, that passes
between the cesophagus and the anterior adductor muscle,
and joins their anterior ends. The cerebral ganglia give rise
to a number of nerves. A large nerve leaves the ventral side
of each near its anterior end, passes ventrally along the
posterior and ventral surfaces of the anterior adductor
muscle, to which it sends branches, and is distributed to the
anterior and ventral portions of the corresponding lobe of
the mantle. Just posterior to the origin of the pallial nerve,
and a little closer to the median line, another nerve, about
equal to the pallial nerve in size, leaves each cerebral
ganglion. This nerve follows along the fold of tissue that
suspends the labial palps and is continued into the palp
appendage. Other nerves from these ganglia are distributed
to the visceral mass and to the dorsal portions of the foot
muscles. Posterior and still further toward the median line
than the palp nerve, the two portions of each cerebro-pedal
commissure leave each cerebral ganglion, one a little ante-
rior and ventral to the other. The two portions run poste-
riorly a short distance, and join to form a single commissure
that is continued to the pedal ganglion of the same side. A
nerve leaves each cerebro-pedal commissure dorsal to the
corresponding otocyst, and is continued to it. This nerve is
generally supposed to have its origin in the cerebral ganglion,
and the angle at which it issues from the commissure indi-
cates that this is probably the case. The otocystic nerve is
about equal in size to the posterior division of the cerebro-
pedal commissure. Stempell (18) finds that each otocystic
nerve of Solemya togata leaves the cerebral ganglion
direct, and runs an independent course to the otocyst. He
also finds that each cerebro-pedal commissure leaves the
cerebral ganglion as a single strand. He thinks that this is
a double commissure, because it receives fibres from what he
considers cerebral and pleural ganglia.

It seems more likely to me that the nervous systems of all
molluscs have been derived from some such a generalised

374 GILMAN A. DREW.

type as is found in Chiton, and that each class has developed
ganglia according to its needs, than that the ancestors of
Lamellibranchs possessed the comparatively complex system
of ganglia found in Gastropods. If this is true, it is easy to
understand why Gastropods with their complicated head
apparatus should develop ganglia for which Lamellibranchs
have no need. Accordingly the necessity to homologise all
of the ganglia in the two classes disappears.

In most Lamellibranchs the otocystic nerves spring from
the cerebro-pedal commissures, and they are supposed to
originate in the cerebral ganglia. In Solemya togata,
Stempell finds that the otocystic nerves leave the cerebral
gangha direct, and are not included in the cerebro-pedal
commissures in any part of their length. Is it not possible
that the posterior root of the cerebro-pedal commissure, in
forms where there are two roots, is the central end of the
otocystic nerve ?

The pedal ganglia (fig. 48, pg.) are rounded and nearly
equal to the cerebral ganglia in size. ‘They lie close together,
and they are connected by a moderately large commissure.
The nerves from the pedal ganglia supply the muscles of
the foot. They need no special mention.

The visceral ganglia (fig. 48, vg.) are the smallest of the
three pairs of ganglia. In shape they resemble the cerebral
ganglia, but they are turned in the opposite direction. Hach
visceral ganglion is elongated, and gradually tapers ante-
riorly into the cerebro-visceral commissure. ‘The two ganglia
lie far apart, and are connected near their posterior ends by
a long and rather thick commissure. A rather large nerve
leaves the posterior end of each ganglion, runs posteriorly
ventral to the posterior adductor muscle, and, besides giving
branches to this muscle, supplies the posterior and ventral
portions of the corresponding lobe of the mantle. Anterior
and ventral to the posterior pallial nerves another rather
large nerve leaves each ganglion. This nerve runs along the
inner side of the suspensory membrane of the corresponding
gill nearly to its posterior end,

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 375

Otocysts.

The otocysts are formed soon after the embryo sheds its
test. They originate as invaginations in the body-wall, a
httle posterior and dorsal to the pedal ganglia. The invagi-
nations deepen and close over to form what seem to be closed
sacs, that soon come to lie near the pedal ganglia in the inte-
rior of the foot. As in the case of Yoldia, these sacs are
apparently entirely closed. Soon after the otocysts are
formed, before the gills acquire their second lobes, otoliths
appear. ‘The otoliths have the appearance of little crystalline
fragments, but I am inclined to think that they are formed
in the otocysts, and are not introduced through the otocystic
canals, as has been held by some writers. ‘The particles seem
to be too large to have been introduced through canals that,
at this stage, lam unable to find. Again, the otocysts never
seem to contain diatoms. Diatoms are very abundant in the
brood-sacs in which the embryos are carried, and form a
large part of the animal’s food. Many of them are well
shaped to pass through small openings, and one would
expect to find them occasionally in the otocysts, if the con-
tained material consists of foreign bodies that have gained
access through the otocystic canals.

About the time that the gills acquire their sixth pair of
plates the otocysts can be seen to be connected with the sur-
face of the foot (figs. 46 and 64, ot.). At first the connection
seems to be solid, but a little later openings can be traced
from the otocysts to the exterior. ‘These tubes, the otocystic
canals, are quite slender near the otocysts, but widen toward
the surface of the foot. From each otocyst the canal passes
anteriorly, laterally, and a little dorsally to open to the
exterior (figs. 46 and 64, ot.).

The position of the external opening is not just what
might be expected if the otocystic canals are remnants of the
invaginations that formed the otocysts. The otocysts are
formed just posterior and alittle dorsal to the pedal ganglia.
As they develop, they sink into the interior of the foot and

376 GILMAN A. DREW.

become permanently settled near the ganglia at points nearly
opposite their points of origin. As the same relation between
organs in this region is retained during the whole of the
development, there is no reason to think that growth is more
from one side than from another. If, then, the otocystic
canals are remnants of the original invaginations, we might
expect them to run almost perpendicular to the surface
instead of opening so far anterior and dorsal. It might be
thought that the development of the anterior foot muscles
has crowded the stomach, ganglia, and otocysts posteriorly,

Text-rig. W.—Horizontal section of the foot of an adult Nucula delphi-
nodonta. The otocystic canals leave the dorsal side of the otocysts, so
that in this section only the dorsal wall of the otocyst is seen on the side
where the canal is present. afm. Anterior foot muscles. zz¢. Intestine.
oc. Otocystic canal. of. Otocyst. pg. Pedal ganglion. sfo. Stomach.

and caused the otocystic canals to. take up this position, but
reference to fig. 64 will show that before these muscles
become very large the otocystic canals open further toward
the anterior than when these muscles become highly deve-
loped (Text-fig. W). This seems to show that the muscles, as
they develop, project anteriorly, and do not affect the organs
lying behind them.

In a former publication (1) I have described the imperfect

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 877

otocystic canals of Yoldia limatula, The canals are short
and do not reach the surface of the foot. Frequently a number
of closed pouches are connected with the end of each otocystic
canal by astrand of fibrous material (Text-fig. X). Stempell
reports that the otocystic canals of Leda pella are very
rudimentary. He has never been able to trace them with
certainty to the surface skin (17). The conditions illustrated
by these two forms can easily be explained as the result of
degeneration, but there is no direct evidence that this is the
case. If the canals are not remnants of the original invagina-
tions, the imperfect canals may be structures that have never
been perfect canals, or they may indicate the degeneration
of canals that have at some time been perfect.

The nerve-supply of the otocysts has been discussed in
connection with the nervous system. It seems possible that

Trext-Fic. X.—Otocyst of Yoldia limatula. cpe. Cerebro-pedal commissure.
oc. Otocystic canal. o/. Otolith. oz. Otoeystic nerve. op. Otocystic
pouch. ot, Otocyst.

the dorsal roots of the cerebro-pedal commissures may be the
central ends of the otocystic nerves.

Muscular System.

For convenience in treating the subject the muscles may
be grouped into those concerned in shutting the shell, in
moving the foot, in propelling blood, in retracting the
margins of the mantle, in retracting the palp appendages,
and in raising the gills. Beside these muscles, there are
many scattered and interlacing fibres that are concerned in
making many of the movements.

VOL. 44, PART 3,—NEW SERIES, BB

378 GILMAN A. DREW.

The anterior adductor muscle (figs. 25 and 36, aa.) is
formed somewhat earlier than the posterior adductor muscle,
and throughout life the former is larger than the latter (fig.
48). When the anterior adductor muscle is first formed (fig.
25, aa.), it occupies a position at the anterior end of the
dorsal space, very near the apical plate. Soon after the test
is shed, it becomes surrounded by tissue that is drawn up
around it (figs. 35 and 39, aa.). The posterior adductor
muscle is formed soon after the test is shed (fig. 39, pa.). It
lies ventral to the intestine, and posterior to the visceral
oanglia, and from its first appearance is surrounded by other
tissue.

In the adult the adductors are attached to the shell, with
their dorsal borders very near the ends of the rows of teeth.
The function of these muscles is simply to close the shell.
The contraction of the muscles, and the consequent closing
of the shell, compresses an elastic pad known as the cartilage,
that hes in the cartilage pit. As soon as the adductor
muscles relax, the expansion of this piece of cartilage opens
the shell. The epidermis is not thickened to form a pro-
minent external ligament.

The foot is attached to the shell by three pairs of well-
developed foot muscles, and by a number of fibres that form
a more or less connected series on each side, ventral to the
genital organ and liver. Of the three pairs of foot muscles,
one is posterior and two are anterior. The posterior muscles
are inserted on the shell along the bases of the teeth,
anterior and dorsal to the posterior adductor muscle. They
extend anteriorly and ventrally along the sides of the foot,
and form the strong retractors of the foot. The two pairs of
anterior foot muscles are attached to the shell close together,
along the bases of the teeth, posterior and dorsal to the
anterior adductor muscle. In distribution, the anterior pair
of these muscles correspond to the two anterior pairs in
Yoldia. They spread out along the sides of the foot, and are
distributed to its posterior and ventral portions. The more
posterior of the two pairs of muscles passes between the pair

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 379

just mentioned, and is distributed to the anterior and ventral
portions of the foot.

All of these muscles are closely bound together by their
own fibres and by interlacing fibres, so that many movements
occur that cannot be explained by direct pulls of one or more
muscles. It should be remembered that the attachments of
the fibres are all along the sides of the foot, and that many,
if not most of the muscle-fibres, pull from one part of the
body-wall to another, without changing the relation of the
body to the shell. Thus the muscular side flaps of the foot
can be spread apart after the animal has been removed from
the shell.

Between the muscles, loose connective tissue and large
blood-spaces occur. Many of the movements, especially
those that result in the protrusion of the foot, seem to
depend on the action of muscles on the fluids of the body,
more especially upon the blood contained in the spaces of the
foot. By obliterating some channels and forcing blood into
others, different results may be obtained.

The muscle-fibres that are attached to the shell along the
ventral border of the genital mass and liver are distributed
to the body-wall. They are not as numerous as they are
in Yoldia. I have found no indication of a special muscle
at the posterior end of each series, as is the case with
Yoldia (3).

The heart is largely made up of interlacing muscle-fibres.
Each auricle is separated from the ventricle by a constriction
(figs. 68 and 69,h.). It seems probable that, when the ventricle
begins to contract, the contraction of the muscles in these
constrictions closes the openings between the ventricle and
the auricles so that the blood cannot flow back into them.
Where the auricles join the blood-spaces of the gills and
mantle lobes, the muscles probably act in the same way.

There are some muscle-fibres in the suspensory membranes
of the gills that probably contract at intervals. The opaque
shells make it impossible to watch the movements of the gills,
but it will be seen that such movements as are made must

380 GILMAN A. DREW.

force some of the blood out of the blood-spaces of the sus-
pensory membranes. The movements are not enough to form
strong currents of water, such as are formed by Yoldia (1).

The margins of the lobes of the mantles are never pro-
truded far beyond the margins of the valves of the shell, and
the pallial muscles are accordingly not excessively developed.

Each of the large palp appendages is supplied with a
rather large muscle that is continued into it from the body-
wall. It occupies the ventral (morphologically outer) side of
the appendage (fig. 66, /m.), and is continued to its tip. This
muscle serves to retract the appendage. Its position in the
appendage is such that when the appendage is strongly
retracted it is curled as shown in fig. 48. The muscle seems
to be homologous with fibres that extend into the membrane
that suspends the palps from the body-wall.

Excretory Organs.

Just before embryos reach the stage where the second gill
filaments begin to flatten, preparatory to forming the third
gill filaments, a pair of narrow tubes appear just anterior to
the visceral ganglia and ventral to the pericardium. The
two tubes touch each other on the median line of the body,
but their cavities do not seem to communicate. Laterally
they are extended to the surface of the body, where they open
into the mantle chamber. This is the earliest stage in which
I have been able to distinguish the kidneys. I have not
succeeded in determining whether the external openings are
present from the beginning, or whether they are formed later.
I am inclined toward the view that the kidneys are meso-
dermal in their origin; but this view is based simply on the
length and narrowness of the tubes when they can first be
distinguished. They may be formed as invaginations from
the surface.

The cells forming the walls of the kidneys soon become
large and vacuolated. This character is retained throughout
the life of the animal, and makes the tracing of their cavities

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 381

in some places a very difficult matter. Near the outer end of
each kidney the cells are smaller, and the lumen is more
easily traced. As the kidneys grow, they extend anteriorly
and crowd dorsally beneath the pericardium and heart. As
growth continues they become bent into loops, and numerous
side pouches are formed.

Although much time has been spent in trying to find the
inner, pericardial openings of the kidneys, I have not suc-
ceeded in placing them. Cavities leading from the peri-
cardium have frequently been traced nearly to the kidneys,
but the vacuolated condition of the cells that compose their
walls makes it very difficult to trace cavities with accuracy.
I have no reason to suppose that the pericardial openings do
not exist. I have simply been unable to find them.

In the adult, the ducts of the genital organs pass close to
the lateral extremities of the pericardium. Near its end each
duct turns toward the median line, meets the outer end of the
kidney on the same side of the body, and opens with it into
the mantle chamber. This connection is easy to demonstrate.
Whether the genital ducts also communicate with the peri-
cardium, or with the inner ends of the kidneys, I am not
prepared to say.

Genital Organs.

The genital organs appear after the animal has become
adult in most other respects. Each genital organ consists,
at first, of a short and rather narrow tube that lies close to
the pericardium, for the most part in contact with it.
Whether this tube originates from the pericardium, or
whether it is formed in some other way, has not been deter-
mined. The genital organs grow rapidly, and extend ante-
riorly and dorsally over and among the lobules of the liver,
which are now very numerous. Soon the eggs and sperm
begin to be formed, and the sexes can be distinguished. The
eggs are few in number, but they are large and brown. The
sperm are very numerous, of moderate size, and pale yellow.

382 GILMAN A. DREW.

These colours are imparted to the genital organs. As their
products begin to mature, the genital organs become very
extensive and crowd between and around other organs, until
all available space is filled.

The genital ducts of the adult, as in the young, connect
with the outer ends of the kidneys, and with them open into
the mantle chamber.

Summary.

The young embryos of Nucula delphinodonta and
Yoldia limatula resemble each other in most respects.
They differ considerably in appearance, because of the dif-
ference in the size and distribution of the surface cilia. In
the case of Yoldia the apical cilia are long and bunched to-
gether, and the cilia on the three intermediate rows of test-
cells are collected into bands (Text-fig. F). In Nucula
delphinodonta all of the cilia on the surface of the embryo
are short and evenly scattered (Text-fig. H). The embryos
of Yoldia swim freely in the water, and have to depend on
their own activities for safety. The embryos of Nucula
delphinodonta develop in a protecting brood-sac (fig. 1).
It is to the advantage of these embryos to remain in the
brood-sac, so active locomotion would not only be of no
value, but it would be a positive danger. ‘I'he possession of
a test that is not functional as an organ of locomotion pro-
bably indicates that the embryos of the ancestors of Nucula
delphinodonta were free-swimming. They then probably
corresponded closely in appearance to the embryos of Yoldia
limatula and Nucula proxima, both of which have the
apical tuft and the bands of cilia.

The presence of a separate anal opening in the test, an ex-
tensive apical plate, and the formation of the cerebral ganglia
without invaginations (fig. 24), are points in which Nucula
delphinodonta differs from Yoldia. Nucula delphino-
donta sheds its test when the foot is very immature.

The development of many of the organs of Yoldia has not

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 383

been traced. The following is a brief review of the organs
of Nucula delphinodonta.

Test.—The test consists of five rows of flattened cells,
that together cover the greater portion of the body of the
embryo. The cilia on the test are short and evenly dis-
tributed. The size and distribution of the cilia are probably
the result of the protection afforded the developing embryo
by the brood-sac. The test is finally thrown away. It is
probably homologous to the velum of most Lamellibranch
embryos. (See pp. 385—339, and figs. 11, 24, 25, 34, and 35.)

Apical Plate.—The apical plate is quite extensive, and
bears short diffuse cilia that resemble the cilia on the test
cells. The size of the apical cilia is probably also the result
of the protection afforded the developing embryo by the
brood-sac. ‘he apical plate is thrown away with the test.
(See p. 339, and figs. 11 and 24.)

Shell.—The shell begins to form some time before the
test is shed. The prodissoconch has a rounded outline and a
short straight hinge-line. The adult shell is very robust.
(See pp. 339—341, and figs. 20, 36, 50, and 51.)

Mantle.—The mantle lobes are formed by the growth
and folding of the shell-gland. There are no tentacles on
the margins of the mantle, and no siphons are formed. (See
pp. 341, 342, and figs. 8, 17, 20, 48, and 69.)

Foot.—The foot is formed by the growth of tissue that,
at first, lies between the stomodeeum and the gut. At the
time the test is shed it is very small and cannot be moved.
The side flaps are developed as the result of unequal growth
of the ventral side of the foot. The foot is a remarkably
good burrowing organ, and it seems never to be used in
creeping. (See pp. 8342—346, and figs. 25, 28, 34, 36, 39, 40,
41, 48, 49, and 69.)

Byssal Gland.—The byssal gland is formed as an in-
vagination on the ventral surface of the foot soon after the
test is shed. It becomes very extensive, but in the adult is
quite small. It seems never to form fibres. (See pp. 346,
347, and figs. 39, 40, 41, 45, and 48.)

384 GILMAN A. DREW.

Alimentary Canal.—The primitive gut is formed by the
separation and division of cells on one side of the embryo.
It is carried further into the intericr by the addition of cells
around the blastopore. These cells form the stomodeum.
Later the gut grows posteriorly, beneath the shell-gland,
and forms the stomach and intestine. ‘The anus opens into
the mantle chamber near the anal pore in the test. The
future shape of the intestine seems to depend upon the posi-
tion of certain organs during its elongation. (See pp. 347—
308, and figs. 8, 9, 11, 15, 24, 25, 34, 36, 40, 45, 46, 47, and
48, and 'Text-figs. M to 8.)

Labial Palps.—The labial palps are marked out as
patches of cilia about the time that the third lobe of the gill
begins to form (fig. 41). The ciliated patches along the sides
of the body are bent so as to form grooves (fig. 62, lp.) ; the
dorsal portions of the patches form the outer palps, and the
ventral portions the inner palps. The palp appendages are
formed by unequal growth of the posterior portion of the
outer palps, and each corresponds morphologically to a pair
of ridges with a groove between them. ‘They can be ex-
tended beyond the margins of the shell, and they are used as
food collectors. (See pp. 353—357, and figs. 41, 45, 47, 48,
54, 55, 56, 57, 58, 59, 60, 62, and 66.)

Gills.—The gills are formed as folds on the inner sides of
the lobes of the mantle. The folds form lobes that grow to
form filaments and finally plates. The inner plates are
formed first. The outer plates are formed by growth at the
bases of the inner plates. A study of their development
throws no light on the phylogeny of the gills. (See pp.
357—363, and figs. 39, 40, 41, 45, 48, 52, and 53.)

Hypobranchial Glands.—The hypobranchial glands
are formed about the time that the animals become sexually
mature. They seem to furnish the secretions from which
the brood-sac is formed, and they may have other functions.
(See pp. 363—365.)

Pericardium.—The pericardium is a remnant of a cavity
that probably represents a schizoccele. Its epithelial lining

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 385

seems to be formed by the rearrangement of mesodermal
cells. (See pp. 365—367, and figs. 24, 26, 28, 34, 35, 36, 39,
41, and 48.)

Vascular System.—The heart is formed as a mesodermal
strand that stretches across the pericardium from one side to
the other. There is no indication that it is formed by the
fusion of either a pair of pouches or two masses of cells. It
seems to be single in its origin. It is formed around the in-
testine, but later becomes free and dorsal to it. This seems
to show that for this group, at least, the dorsal position of
the heart as found in the adult is not primitive. The vas-
cular system consists largely of spaces that occur in all parts
of the body. (See pp. 367—870, and figs. 41, 48, 67, 68, and
69.)

Nervous System.—The cerebral and pedal ganglia are
about equal in size, and the visceral ganglia are considerably
smaller. ‘The cerebro-visceral commissures are very large,
and contain many scattered nuclei. The cerebro-pedal com-
missures show ordinary structure. It is suggested that the
smaller, dorsal roots of the cerebro-pedal commissures may
be the central ends of the otocystic nerves. (See pp. 870—
374, and figs. 24, 34, 36, 40, 46, and 48.)

Otocysts.—The otocysts are formed as invaginations from
the body-wall soon after the test is shed. They seem to be
entirely closed off, but canals connecting them with the
surface are present in the adult. The otoliths are irregular
bodies, but they are probably formed in the otocysts them-
selves. (See pp. 375—877, and figs. 40, 46, 48, and 64, and
Text-fig. W.)

Muscular System.—The muscular system is well devel-
oped, and resembles the muscular system of Yoldia in most
respects. The extensive attachments of the foot muscles to
the dorsal portion of the shell is accounted for by the great
development of these muscles. (See pp. 377—380.)

Excretory Organs.—The vacuolated character of the
cells of the excretory organs makes it difficult to trace some
portions of the cavities of these organs. The inner peri-

386 GILMAN A. DREW.

cardial openings are hard to find. The outer end of each
excretory organ opens into the mantle chamber, in connec-
tion with the genital duct of the same side. (See pp. 380, 381.)

Genital Organs.—The genital organs are formed after
the animal is adult in most other respects. They can first be
distiaguished as short tubes that le very close to, or in
contact with, the pericardium, and open into the mantle
chamber in connection with the outer ends of the kidneys.
The genital organs become very extensive in the adult. The
sexes are separate. (See pp. 381, 382.)

Most of the work necessary for the preparation of this
paper was done in the Biological Laboratory of the Johns
Hopkins University. To many that are now or formerly
were connected with that laboratory, and especially to Pro-
fessor W. K. Brooks, I am indebted for suggestions and
encouragement. I also desire to express my appreciation of
the courtesies extended to me by Professor C. O. Whitman,
at the Marine Biological Laboratory. I am_ particularly
indebted to my wife, who has, among other things, performed
a great share of the work counected with the care and pre-
servation of material. Beside the work at the sea-shore,
embryos obtained in June were kept alive in Baltimore from
October 1st until January Ist, with water sent from the sea.

LITERATURE.

1. Drew.—“ Yoldia limatula,’ ‘Memoirs from the Biol. Lab. of the
Johns Hopkins Univ.,’ vol. iv, No. 3, 1899.

2. Drew.— Some Observations on the Habits, Anatomy, and Embryology
of Members of the Protobranchia,’”’ ‘Anat. Anz.,’ Bd. xv, No. 24,
1899.

8. Drew.—‘ Locomotion in Solenomya and its Relatives,” ‘ Anat. Anz.,’
Bd. xvii, No. 15, 1900.

4. Forses ann Haniry.—‘ History of British Mollusca and their Shells,’
1853. ;

5. Gropren.—‘ Die Pericardialdrise der Lamellibranchiaten,’ ‘ Arb.
Zool. Inst. Wien,’ Bd. vii, 1888.

6. Jacxson.— Phylogeny of the Pelecypoda,” ‘Mem. Boston Soc. Nat.
Hist.,’ vol. iv, No. 8, 1890.

TEE LIFE-HISTORY OF NUCULA DELPHINODONTA. 387

7. Kerxttoce.—“ A Contribution to our Knowledge of the Morphology of
Lamellibranchiate Mollusks,” ‘ Bull. U. 8. Fish. Com.,’ vol. x, 1890.
8. Kowa.evsky.—< Etude sur Vembryogene du Dentale,” ‘Anu. Musée
d’Hist. nat. de Marseille, Zool.,’ tome 1, 1883.
9. Lacaze-Durniers.—“ Histoire de Vorganisation et du développement
du Dentale,” ‘ Ann. des Sci. Nat.,’ series 4, tome vii, 1857.
10. Muitnze-Epwarps.— Lecons sur la physiologie et ’anatomie comparée.’
11. Mirsuxuri.—‘‘On the Structure and Significance of some Aberrant
Forms of Lamellibranchiate Gills,” ‘ Quart. Journ. Mier. Sei.,’ vol.
xxi, 1881.
12. Parren.—* The Embryology of Patella,” ‘ Arb. Zool. Inst. Univ. Wien,’
Bd. vi, 1886.
13. PrtseneerR. “ Contribution a l’étude des Lameillibranchs,’’ ‘ Arch. de
Biol.,’ tome xi, 1891.
14. Prtsengrr. ‘ Recherches morphologiques et phylogénétiques sur les
Mollusques archaiques,’ 1899.
15. Pruvor.— Sur le développement d’un Solénogastre,” ‘Compt. rend.
Acad. Sci.,’ Paris, tome exi, 1890.
16. Kicr.—* Die systematische Verwertbarkeit der Kiemen bei den
Lamellibranchiaten,”’ ‘Jen. Zeit. fiir Naturwiss.,”’ Bd. xxxi, 1597.
17. Sremprtt.—“ Beitrage zur Kenntniss der Nuculiden,” * Zool. Jalirb.,’
Suppl. 4, Fauna Chilensis, Heft 2, 1898.
18. Sremprecy.—‘‘ Zur Anatomie von Solemya togata,” ‘ Zool. Jabrb.,’
Bd. xiii, 1899.
19. THreLe.—“ Die Stammesverwandtschaft der Mollusken,” ‘Jen. Zeit.
fiir Naturwiss.,’ Bd. xxv, 1891.
20. Ziecier.—‘ Die Entwickelung von Cyclas cornea,
Zool.,’ Bd. xli, 1885.

2 ¢

Zeit. fiir wiss.

EXPLANATION OF PLATES 20—25,
Illustrating Mr. Gilman A. Drew’s paper on ‘ The Life-
History of Nucula delphinodonta (Mighels).”

Reference Letters.

aa. Anterior adductor muscle. aas, Anterior adductor muscle-scar. ap.
Apical plate. dg. Byssal gland. 4s. Blood-space. ca. Cartilage. cg.

388 GILMAN A. DREW.

Cerebral ganglion. cp. Cartilage pit. cs. Chitinous support. ec. Eetoderm.
f. Foot. g. Gill. gs. Suspensory membrane of gill. 4. Heart. iné. Intes-
tine. ip. Inner plate of the gill. dp. Inner labial palp. 4. Kidney. J.
Liver. Jm. Longitudinal muscle. Jp. Labial palp. m. Mantle. mg. Mid-
gut. mo. Mouth. oes. @sophagus. o/p. Outer labial palp. op. Outer
plate of the gill. of. Otocyst. pa. Posterior adductor muscle. pap. Palp
appendage. pas. Posterior adductor muscle-scar. pg. Pedal ganglion. pz.
Palp nerve. sg. Shell-gland. sd. Stomodeum. sto. Stomach, ¢. Test.
tc. Cavities in the mantle caused by teeth on the shell. 7. An organ of un-
known function. vg. Visceral ganglion. y. Cut wall of gill plate. z.
Scattered cells of the disorganised liver.

PLATE 20.

Fic. 1.—Adult specimen with the brood-sac attached. The brood-sac is
torn open to show the eggs inside. x 10.

Fie. 2.—Sixteen-celled stage. x 150.

Fic. 3.—Section of an embryo in the sixteen-celled stage. x 275.

Fic. 4.—Section of a later cleavage stage that corresponds to an epibolic
gastrula. The asterisk (*) marks the position where the gut is formed. xX
275.
Fic. 5.—An embryo that is slightly older than the one represented in sec-
tion by Fig. 4. x 150.

Fic. 6.—Lateral view of an embryo in which the gut has been formed,
represented as a slightly transparent object. From the study of preserved
material I am inclined to think that the shell-gland does not bear cilia, but
this has not been determined on living material. The line marked 7 indicates
the plane in which the section, Fig. 7, was taken. x 150.

Fic. 7.—Transverse section of an embryo in the stage represented by Fig.
6. The line 7, on Fig. 6, indicates the plane of the section. Xx 275.

Fic. 8.—Sagittal section of an embryo in the stage represented by Fig. 6.
x 275.

Fic. 9.—Sagittal section of an embryo slightly older than the embryo of
which Fig. 8 is a section. It represents the beginning of the formation of the
stomodeum. X 275.

Fic. 10.—Dorsal view of an embryo in which the test is growing over the
shell-gland. The lines numbered 11, 12, and 13 indicate the planes of sec-
tions represented in corresponding figures. X 150.

Fic. 11.—Sagittal section of an embryo in the stage represented by Fig.
10. The line 11 on Fig. 10 indicates the plane of the section. X 275.

Fics. 12 and 13.—Transverse sections of an embryo in the stage repre-
sented by Fig. 10. ‘The lines 12 and 18 on Fig. 10 indicate the planes of the
sections. xX 275.

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 3889

PLATE 21.

Fic. 14.—Lateral view of an embryo in which the test is fully formed,
represented as a slightly transparent object. The lines numbered 16, 17, 18,
and 19 indicate the planes of sections represented in corresponding figures.
50:

Fic. 15.—Sagittal section of an embryo in the stage represented by Fig.
14. The anterior ends point in opposite directions in the two figures. xX
275.

Fics. 16, 17, 18, and 19.—Transverse sections of an embryo in the stage
represented by Fig. 14. The lines 16, 17, 18, and 19 on Fig. 14 indicate the
planes of the sections represented by these figures. Xx 275.

Figs. 20, 21, and 22.—Transverse sections of an embryo older than the one
represented by Fig. 14, and a little younger than the embryo represented by
Fig. 23. The lines 20, 21, and 22 on Fig. 23 represent planes that correspond
to these sections. X 275.

Fic. 23.—Lateral view of an embryo in which the mantle is beginning to
form, represented as a slightly transparent object. The lines 20, 21, and 22
indicate the planes of sections represented in corresponding figures, but the
embryo represented in Fig. 28 is slightly older than the one from which these
sections were obtained. x 150.

Fic. 24.—Sagittal section of an embryo in the stage represented by Fig. 23.
x 275:

Fie. 25.—Lateral view of an embryo that would soon shed its test. The
test cells, indicated in outline, are very indistinct, and are not accurately
drawn. Cilia have been indicated along the margins only. They cover the
whole of the surface. The organs are more clearly shown than in preceding
figures of embryos. They are not visible in whole mounts, but have been
reconstructed from sections. The Jines 27, 28, 29, 30, 31, 32, and 33 indicate
the planes of sections represented by these figures. (See Plate 22.) x 150,

PLATE 22.

Fic. 26.—Sagittal section of an embryo in the stage represented by Fig. 25,
Pi. 21. x 275.

Fies. 27, 28, and 29.—Transverse section of an embryo in the stage repre-
sented by Fig. 25, Pl. 21. The lines numbered 27, 28, and 29 on Fig. 25
indicate the planes of sections represented by these figures. x 275.

Figs. 30, 31, 32, and 33.—Horizontal sections of an embryo in the stage
represented by Fig. 25, Pl. 21. The lines numbered 80, 31, 32, and 33 on Fig.
25 indicate the planes of the sections represented by these figures. x 275.

PLATE 23.

Fie. 34.—Lateral view of a reconstruction of an embryo that has just com.

390 GILMAN A. DREW.

pleted the first step in the process of casting. The test cells, apical plate, and
stomodzeum still adhere to the anterior end of the embryo. x 150.

Fie. 85.—Lateral view of a reconstruction of an embryo that has completed
the process of casting. x 150.

Fre, 36.—Lateral view of a reconstruction of an embryo in which the liver
pouches have begun to go to pieces. x 150.

Fie. 37.—Sagittal section of an embryo in the stage represented by Fig. 36.
x 275.

Ite, 38.—Transverse section of an embryo in the stage represented by Fig.
36, taken through the stomach just posterior to the pedal ganglia. x 275.

Fig. 89.—Lateral view of a reconstruction of an embryo that is just begin-
ning to form the gills. x 150.

Fie. 40.—Lateral view of a reconstruction of an embryo in which each gill
is beginning to form two lobes. x 150.

Fie. 41.—Lateral view of a reconstruction of an embryo in which each gill
is beginning to form its third lobe. x 150.

Fre. 42.—Horizontal section of an embryo in a stage represented by Fig.
40, taken through the dorsal end of the stomach and the re-forming lobes of
the liver. xX 200.

Fic. 43.—Horizontal section of an embryo in the stage represented by Fig.
41, taken through the dorsal end of the stomach and the re-forming lobes of
the liver. x 200.

Fre. 44.—Horizontal section of an embryo in the stage represented by Fig.
46, Pl. 24, taken through the dorsal end of the stomach and the re-formed
lobes of the liver. x 150.

PLATE 24.

Fre. 45.—Lateral view of a recoustruction of an embryo in which each gill
has four pairs of plates. x 125.

Fie. 46.—Lateral view of a reconstruction of an embryo in which each gill
has six pairs of plates. x 125.

¥re, 47.—Lateral view of a reconstruction of an embryo in which each gill
has eight pairs of plates. x 110.

Fre. 48.—Lateral view of a reconstruction of an adult specimen. x 30.

Fie, 49.—Adult specimen with the foot protruded and the side flaps spread
apart. x 10.

Fie. 50.—View of the inside of au adult left shell-valve. x 15.
Fie. 51.—Left shell-valve seen obliquely from the dorsal margin. x 15.

Fic. 52.—A nearly horizontal section of an embryo in the stage represented
by Fig, 46, cut to show the developing outer plates of the gills, x 150,

THE LIFE-HISTORY OF NUCULA DELPHINODONTA. 391

Fic. 53.—A pair of adult gill plates. The suspensory membrane, the
continuous chitinous trough, the longitudinal muscle, and the walls of the
plates that join the plates next in succession have all been cut across in
removing the plates from the gill. (Drawn from a study of sections.)
x 250.

PLATE 25.

Fres. 54, 55, and 56.—Stages in the development of the labial palps. Phe
palps have been carefully drawn, but for the sake of clearness they have in
each case been represented with the outer palp on the right side turned away
from the corresponding inner palp. The foot is represented as cut off, and the
specimen is turned so that the mouth can be seen between the palps. x 125.

Fie. 57.—The posterior portions of the right outer and inner palps of an
adult specimen. ‘he two palps are represented as spread apart and placed in
a position that corresponds with Fig. 56. x 65.

Fies. 58, 59, and 60.—Successive sections of the labial palps of a specimen
that has six pairs of gill plates. The sections are taken transverse to the
embryo. The stage is much the same as is represented by Fig. 54. Fig.
58 is near the mouth, Fig. 59 is near the posterior end of the outer palp, and
Fig. 60 is posterior to the outer palp. x 150.

Fie. 61.—Transverse section of an embryo with four pairs of gill plates
(see Fig. 45, Plate 24) taken through the mouth. x 200.

Fic. 62.—Transverse section of an embryo with four pairs of gill plates,
taken just anterior to the stomach. x 200.

Fie. 63.—Sagittal section of the antero-dorsal portion of an embryo that
has eight pairs of gill plates. x 150.

Fie. 64.—Horizontal section of the foot of an embryo that has six pairs of
gill plates, taken just ventral to the mouth. x 150.

Fic. 65.—Horizontal section of the foot of au embryo that, has six pairs of
gill plates, taken through the mouth. x 150.

Fia. 66.—Transverse section of the palp appendage of an adult specimen.
x 200.

Fig. 67.—A nearly transverse section of an embryo that has five pairs of
gill plates, taken through the heart. x 200.

Fria. 68.—A diagonal section of au embryo that has nine pairs of gill plates,
taken through the heart. x 90.

Fre. 69.—A diagonal section of an adult specimen, taken through the heart,
x 45.

STRUCTURE OF THE HAIRS OF MYLODON IISTAI. 393

On the Structure of the Hairs of Mylodon
Listai and other South American Edentata.

By

W. G. Ridewood, D.Sc., F.L.S.,
Lecturer on Biology at the Medical School of St. Mary’s Hospital, London.

With Plate 26.

THE interest attaching to the discovery of portions of well-
preserved skin of a great Ground Sloth, very closely allied to
if not identical with Mylodon, was considerably increased
when it was found that the hairs do not agree in their minute
structure with those of the Tree Sloths, Bradypus and
Cholepus. While agreeing with the latter in the absence
of a definite medulla, they are destitute of the extra-cortical
layer which characterises the hairs of Bradypus, and have
not the fluted surface which is such a distinctive feature of
the hairs of Cholcepus. The characters of the hairs have
been commented upon by several authors in the course of
their remarks upon the remains of this ground sloth, but the
subject has never been treated exhaustively ; and Professor
Ray Lankester suggested to me that the matter was worthy
of further inquiry, and that it was desirable to compare the
newly discovered hairs not only with those of Bradypus
and Cholcepus, but also with those of the ant-eaters and
armadillos.

The order Edentata, as at present constituted, will proba-
bly prove to be an unnatural assemblage of animals, and it
may become necessary, when our knowledge is more complete,

vou, 44, PART 3, —NEW SERIES. co

394 W. G. RIDEWOOD.

to remove the Old World forms Manis and Orycteropus,
to constitute two new orders by themselves. For the present
purpose, however, the relationships are not material, except
for the fact that the late traveller Ramon Lista saw and shot
at a curious animal in South America, which he likened to a
hairy and scaleless Pangolin. It is generally denied
(Ameghino [1], Lonnberg [%7, p. 168]) that this “‘ pangolin”
was the Mylodon of which the skin and bones have more
recently been found.

Accounts of the various pieces of skin discovered have been
published by Ameghino (1), Lénnberg (7), Roth (14), Smith
Woodward and Moreno (18), and Smith Woodward (19).
The locality from which Dr. Ameghino obtained his specimen
is not stated, but the other pieces of skin were found on
different occasions in the loose earth of a cavern near
Consuelo Cove, Last Hope Inlet, Patagonia. The deposit in
which they were found is regarded as of Pampean age, and
there can be no doubt that these ground sloths were con-
temporaneous with man, if not actually living in the cavern in
a state of domestication.

Concerning the generic name, there appears to be no valid
reason why Mylodon should not be used. The genus was
first established by Owen in 1840 (11, p. 68), the type species
being Mylodon Darwinii. ‘Two years later Owen (12) de-
scribed a nearly complete skeleton of a ground sloth which he
called robustus, and referred to the same genus. Reinhardt
(13), writing in 1879, showed that the two species were
generically distinct, and renamed the earler specimen
Grypotherium. If, however, the rules of priority are to be
observed at all, the term Mylodon should be retained for the
species Darwinii, and the robustus should be accorded a
new generic name. ‘Theargument that the species robustus
was fully described, whereas Darwinii was represented only
by a fragment of jaw, is obviously inadmissible, for if the
fragment is sufficiently perfect to enable Reinhardt’s specimen
and those recently discovered to be regarded as generically
identical with it, it is sufficiently perfect and important to

STRUCTURE OF THE HAIRS OF MYLODON LISTAI. 395

act as the type of the genus. Glossotherium, a genus
founded by Owen in 1840 (11, p. 57), is admitted by himself
(12, p. 154) to be identical with Mylodon Darwinii, and
this genus may thus be dismissed as asynonym of Mylodon.
The same fate befalls Neomylodon, since the. newly dis-
covered specimens to which the name was applied are widely
regarded as generically, if not specifically (Nordenskjéld [10]),
identical with Owen’s MylodonDarwinii. Until, however,
the specific identity has been more fully established it is
preferable to retain Ameghino’s specific name of Listai for
these remains.

The hairs of Mylodon Listai have been described by
Lénnberg (7), Jacob (4), and Smith Woodward (18 and 19),
and transverse sections have been figured by the first two
authors. The descriptions presuppose a knowledge of the
hair structure in Bradypus, the ant-eaters and armadillos,
and so in the present communication the consideration of
the Mylodon hairs and the criticism of the views of these
three authors are left till the last.

The method adopted for the examination of the hairs was
in all cases the same. The hairs were arranged with the
roots pointing one way and the free ends the other; they
were tied up in bundles, stained with a weak alcoholic
solution of magenta, washed and dehydrated, The bundles
were then soaked in xylol, and transferred to hard paraffin.
After cooling the paraffin was cut into convenient blocks,
and the sections were made by hand with a sliding motion of
the razor. It was found that better results were obtained in
this way than by the employment of any form of microtome.
Some of the sections were mounted in glycerine jelly, but the
majority in Canada balsam, since the former medium has the
disadvantage of dissolving out the stain. A few hairs of
each of the species studied were stained and mounted whole.
For the Mylodon hairs I am indebted to the kindness of Dr.
F. P. Moreno; the hairs of the other Edentata were obtained
from dried specimens in the Natural History Museum,
London.

396 W. G. RIDEWOOD.

Bradypus tridactylus.

The hairs of Bradypus are oval in section, and exhibit a
central clear area and a darker marginal area (fig. 8). The
central area stains very faintly if at all with magenta, and
being brittle is apt to crack in the cutting. It is marked by
a small number of minute air spaces, the true shape of which
is fusiform. The long axis of each spindle is parallel to the
length of the hair, and consequently the transverse sections
of the spaces are larger or smaller according as they are cut
through the middle or near the ends of the spindles. The
outer substance stains deeply, and is thickly marked with
dark granules, and exhibits at the same time two sets of
radiating lines—a set of very fine and closely set lines around
the outer edge, and a set of coarser and more irregular lines
branching out from the central mass. The average size of
the transverse section is 240, x 145 p.

The outer substance is a layer not represented, or at least
not in this form, in the hair of any other mammal. It does
not extend the full length of the hair, but stops short near
the free end, and is absent from the basal third of the hair.
In optical section (fig. 4, upper part) it exhibits an oblique
striation. The terminal portion of the hair (the “ Endfaden ”
of Welcker [17]) has the normal structure of a non-medullate
hair with a scaly cuticle, but at a certain distance from the
point the diameter increases quite suddenly by the addition
of this new layer (fig. 2). The diminution in the width of
the central core at this point is probably not real, but an
optical effect due to the refrangibility of the external layer.
‘he figure represents an optical section, not an actual slice
taken from the middle of the hair. The basal third of the
hair is thin as compared with the distal part, and measures
only 64 4 across (fig. 6) ; it appears transparent when the hair
has been clarified and mounted whole. In addition to the
minute fusiform air spaces it frequently has larger air-filled
cavities, blunt ended, and about 60, long and 6 uw broad.
The transverse section of this part of the hair is nearly

STRUCTURE OF THE HAIRS OF MYLODON LISTAI. 397

circular (fig. 5). In all parts of the hair of Bradypus the
cuticular scaling can be seen by suitable staining and accu-
rate focussing.

That the central part of the hair of Bradypus is a cortex,
and not medulla, as Kble supposed (2, Bd. 11, p. 440, and
Taf. x, fig. 111), and that the peripheral part is extra-cortical,
has been ably shown by Welcker (17), who applied the name
‘* Belegschicht ” to it. The relation which the extra-cortical
layer bears to the normal cuticle is very difficult to determine.
A careful examination of the part of the hair where the
transition occurs between the normal terminal portion and the
part provided with the extra-cortex (fig. 2) shows that the
arrangement of the imbricate scales of the cuticle is con-
tinued without interruption upon the exterior of the extra-
cortical layer, thus seeming to show that the cuticle is con-
tinued over the outer surface of this layer. The extra-cortex,
however, is very friable in old hairs, and comes away readily,
leaving the central column of cortical substance bare; and it
is then seen that the surface of the column is marked by lines
taking a more or less transverse course, and suggesting
forcibly that the cuticular scaling is continued on the surface
of the cortex beneath the extra-cortical layer. There is yet
a third possibility, which may eventually prove to be the
correct interpretation, since it accounts for both sets of
appearances. It is that the extra-cortical layer is the cuticle
itself, enormously thickened and distinctly cellular, instead
of more or less homogeneous and _ structureless. The
arrangement of the cells would account for the markings on
the external surface of the hair, and the scaly appearance of
the cortical rod when laid bare would be due to the impress
left by the extra-cortical cells. ‘The appearances presented
by that basal part of the hair where the extra-cortex is just
dwindling away certainly favours the third supposition. ‘The
cells of the extra-cortex get thinner and thinner, and come to
resemble the scales of the cuticle. ‘They become more firmly
adherent to one another and to the cortex, they appear more
homogeneous, and they stain less deeply. The figure given

398 W. G. RIDEWOOD.

by Welcker (17, Taf. ii, fig. 11) of the young hair in its follicle
at a time when the extra-cortex is forming would appear
to allow of no alternative proposition. Yet Welcker was
disposed to regard the Belegschicht as a new tissue inter-
calated between the cuticle and the cortical rod (17, p. 44) ;
and the effect obtained by macerating the hair in water, and
thus causing a thin cuticular layer to peel off (17, Taf. i, fig.
14), lends support to his view. But this effect is very possi-
bly due to the excessive cuticularisation of the outer parts of
the external cells, and not to any morphological distinction
of layers.

Leydig (6, p. 687) took the extra-cortex, or at least a part
of it, to be the cuticle, for he observed that, contrary to the
generalisation made by Reissner and Reichert, the hair cuticle
does contain pigment granules in one mammal, namely,
Bradypus. Waldeyer (16, p. 186) supported, in the main,
Welcker’s contention, and regarded the “ Rindenmantel ” as
a layer peculiar to the sloths, and lying below the cuticula ;
aud Leche (5, p. 934) is probably only adopting Welcker’s
suggestion when he remarks of the ‘‘ Umkleidungsschicht ”
that “sie besteht aus einer zwischen Cuticula und Rinden-
substanz gelegenen pulpdsen, lufthaltigen Zellenschicht.”
Maurer (8, p. 278), on the other hand, holds that the thicken-
ing of the distal part of the hair of Bradypus is mainly
effected by the cuticle (Oberhautchen). His account, how-
ever, is very confusing, since he speaks of a medulla extend-
ing two thirds of the length of the hair, and of the cortical
cells being pigmented; and although he gives the title of
Welcker’s classical paper in his bibliography, he fails to
contrast bis own observations with those which this author
had already placed on record.

The biological significance of the extra-cortical layer is full
of interest, and has been made known by the writings of
Welcker (17) and Sorby (15). The layer has a tendency to
crack in a transverse direction, and in the cracks there come
to lodge unicellular alge, to which Kihn (17, p. 66) has
given the name Pleurococcus Bradypi. The moisture of

STRUCTURE OF THE HAIRS OF MYLODON LISTAI. 3899

the climate in which Brady pus lives enables the alga to live
and propagate in this curious position, and the sloth acquires
a general green tint, which must render it very difficult to
distinguish as it hangs among the green foliage. In thick
transverse sections of the hair these algal bodies show up
very clearly, since they stain deeply, and have a sharply
defined, circular or slightly oval outline. Unless the hair is
much broken they are confined to the outer parts of the extra-
cortical layer.

In addition to the larger hairs just described, Bradypus
has a set of shorter and much finer hairs, constituting the
under-fur. These hairs have a diameter of 24 4, and consist
of a column of cortical substance traversed by fine fusiform
air spaces, and covered by an imbricated cuticle (fig. 7).
Like the larger hairs of the body, they have no medulla.

Cholepus didactylus.

The hairs of Cholcepus are no less remarkable than those
of Bradypus, but in a totally different way. The bulk of
the hair is composed of cortex, the surface of which is fluted
orchannelled. The grooves, as is well known, are occupied by
strands of extra-cortex, in which lives an alga—the Pleuro-
coccus Cholepi of Kihn (17, p. 66). Even from the hairs
of dried museum specimens a green solution, giving the
absorption bands of chlorophyll, can be obtained by boiling
first in water and then in alcohol.

Wheu seen in transverse section (fig. 8) the outline is oval,
and measures about 1504 x 904. The cortical substance is
in some cases quite clear and hyaline, but in others it is
marked by brown spots—differences presumably related to
the age of the hairs. In both cases this cortical substance
does not stain with magenta. But running throughout,
except towards the summits of the superficial ridges of the
hair, are irregular branching lines, which stain deeply, and
are discernible in unstained sections by reason of their
different refrangibility. In very thin sections these lines are

400 W. G. RIDEWOOD.

seen to be empty tubes, with a deeply staining lining. These
conditions do not appear to be paralleled in the hair of any
other animal. 'The branching tubes may possibly represent
a diffused medulla, for in most hairs the medulla stains deeply
and becomes largely infiltrated with air. This is the view
taken by Waldeyer (16, p. 187), who writes that the hair
shows “einen grossen centralen Markstrang, der aber durch
Balken von Rindenschicht vielfach durchsetzt ist,’ and by
Weicker (17, p. 55), according to whom “diese Markrohre
ist, wie bereits Erdl[8] erwahnt, innerhalb des dickeren Thiels
des Haars nicht circumscript, sondern in eigenthiimlicher
Weise mit Rindenschicht untermischt.” Maurer’s account of
the hair structure in Cholepus (8, p. 278) is as unintelligible
as his description of that of Bradypus. He speaks of the
cortex being thin in the broad part of the hair, thereby
implying that a compact central medulla is present.

The cuticle is present, and it is imbricate, as can be seen
by the serrated appearance of the edge of the hair when
viewed in optical section. By staining rapidly, and washing
before the deeper parts of the hair have become affected, the
edges of the scales can be seen when the surface of the hair
is in focus. This is particularly the case with the hairs of
the under parts of the body, which have fewer longitudinal
grooves than those on the back. On the summit of the
ridges the cuticle is thick and highly refractive, but how the
cuticle is continued from one ridge to the next it is difficult
to determine. In very thin sections the cuticle can be traced
down the sides of the groove, becoming thinner and thinner,
and disappearing at the bottom. The grooves would thus
seem to be morphologically outside the hair. Yet it can be
seen in many places that the grooves are not perfect, as if
made with a plough, but are discontinuous ; and each portion
is canoe-shaped, open to the exterior at its middle, but covered
in at the two ends. Sections taken across the end of such a
segment of the groove show a continuous cover of cuticle
(see a, fig. 8), and in surface view, with carefully stained
specimens, the edges of the cuticular scales can be traced

STRUCTURE OF THE HAIRS OF MYLODON LISTAI. 401

across. It is no uncommon thing to find a ragged flap of
cuticle overhanging the groove, as at 5 in fig. 8. These facts
tend to show that the grooves are subcuticular ; Welcker, in
fact, goes so far as to state (17, p. 56) that the cuticle bridges
over the grooves except in certain places, and his fig. 21
lends support to this view. And yet there is no denying the
fact stated above, that the cuticle can be traced down the
side of the groove. The logical conclusion, therefore, to
which these facts point is that the grooves are morpho-
logically intra-cuticular, a view which is in complete accord
with the third suggestion offered in the case of the extra-
cortex of Bradypus—that the cells are those of the cuticular
layer, more numerous and less cuticularised than usual.

The hairs of Cholcepus are as a rule coarse, and with a
single curve extending over the greater part of the length,
while the basal fourth or so is wavy ; but in young specimens,
and in some apparently adult specimens from Costa Rica, the |
hair is very delicate and soft, and sinuous from base to point.
The differences may be specific,| or due to age, season, or
sex. However, in these forms the hairs are only about 42 n
across, and have only two or three furrows instead of the
more usual nine, ten, or eleven. The alge, also, are quite
absent from many of the grooves. When such an empty
groove is examined in optical section (fig. 12) it exhibits the
outlines of obsolete extra-cortical cells, the edges of which
are conterminous with those serrations of the margin which
indicate the edges of the cuticular scales. In baby specimens
more than half of the hairs are slender, non-medullate cylin-
ders, with very distinct scaly cuticle, and no grooves on the
surface. They are only slightly shorter than the two- or
three-grooved hairs just referred to, and constitute the
nearest approach to an under-fur found in Cholepus.

' The species didactylus and Hoffmanni were supposed to differ in the
number of cervical vertebra. Although this distinction has broken down,
Cholepus Hoffmanni may still prove to be a good species. Until more
accurate knowledge is available concerning the geographical range and internal
anatomy of the so-called species of Cholepus the point must remain open.

402 W. G. RIDEWOOD.

In Cholcepus, as in Bradypus, the hairs are very thin
at their basal ends (60). The flutings of the surface die
away on the basal sixth of the hair, and here the structure is
that of a normal non-medullate hair (figs. 10 and 11). The
cortex is not marked by the deeply staining branched tubes,
but is rendered shghtly granular by the presence of a number
of fine air spaces, some spherical and scattered, some sphe-
rical and arranged in series of five or six, like strings of
beads, and some fusiform, as though formed by the coales-
cence of such series of smaller cavities. ‘The cuticle is thin
and distinctly imbricate.

There are in Cholcepus no fine hairs to constitute a
proper under-fur, and Welcker has remarked (17, p. 70),
“Der Gegensatz von Stichelhaaren und Wollhaaren fehlt bei
Cholepus;” but de Meijere (9, p. 361) has described some
flattened, stiff, and slightly curved hairs, much shorter than
the ordinary hairs, and possessed of large medullary cells,
surrounded by avery thin cortical layer. These hairs I have
searched for in vain.

Myrmecophaga jubata.

‘he hairs of the great ant-eater are much flattened, and
resemble a ribbon which is thinner in the middle than toward
its edges. ‘lhe actual measurements are—breadth 400 uy,
thickness in the middle 110», thickness near the edge 170 p.
The cuticle is thin for the size of the hair, and exhibits,
rather indistinctly, the usual imbricate or serrate appearance,
according as a surface view or an optical section is taken.
The cortex is full of air spaces (fig. 14), which are provided
with a deeply staining lining after the manner of the branch-
ing tubes which permeate the cortex of the Cholcpus hair.
These spaces, however, can hardly be regarded as a diffused
medulla, since a true medullary region is here differentiated ;
and the suggestion made to this effect in the case of Cho-
lee pus thus receives by analogy a partial refutation. When
the hair is examined from the side the cortical vacuoles are

STRUCTURE OF THE HAIRS OF MYLODON LISTAI. 403

seen to consist of rows of six or seven spherical spaces
arranged in the direction of the length of the hair. ‘lhe
middle spaces of each series are the largest, and the terminal
ones the smallest, so that the general effect is that of a seg-
mented spindle. In the pigmented parts of the hairs the
pigment granules are disposed mainly around the smallest
air spaces at the ends of the spindles. The central part of
the hair is occupied by a slit-like air space partially filled
with a highly refractive substance, which shows no traces of
its cellular origin as the medulla so frequently does. The
basal part of the hair is more cylindrical in shape than the
middle part; and the medullary cavity dwindles gradually
away, to disappear altogether in the part of the hair within
the follicle, or just outside it. The basal parts are trans-
parent, owing to the reduction in the number and size of the
air vacuoles. A section of the hair taken about 3 mm. out-
side the follicle is shown in fig. 15.

Tamandua tetradactyla.

In this ant-eater the hairs are less coarse than in Myr-
mecophaga, and have the form of slightly compressed
cylinders. The transverse section is oval in form, measures
140 nw x 90 mw, and exhibits a solid, non-medullate area of
cortex, marked with numerous brown spots arranged in
groups (fig. 16). The cortex is enclosed within a thick and
tangentially stratified cuticle of clear, highly refractive
aspect. Hxamined from the side the cuticle shows the usual
imbricate markings. ‘The basal part of the hair is more
circular in section; it 1s free from the brown granules, and
contains only a few scattered air spaces of minute size.

Cyclothurus didactylus.

The two-toed ant-eater has in addition to the principal
hairs of the body a well-developed under-fur of much finer
hairs. ‘The whole pelage is soft and fluffy. The principal
hairs, although much smaller than those of Tamandua, do

404 W. G. RIDEWOOD.

not differ from these in any essential respect. They have a
fairly thick cuticle, but no medulla. They are broadest at
about one sixth of their length from the free end, and in this
part the cortex is coloured brown by numerous granules ;
whereas in the basal half or more these are wanting, and the
hair appears quite clear, with just an odd air vacuole here
and there.

The scaling of the cuticle is very strongly marked on the
basal part of the hair, but in the pigmented portion it is less
easy to distinguish. In the fine hairs of the under-fur the
cuticular scaling is the most obvious feature. The greatest
width of the larger hairs is 70 w; that of the supplementary
hairs 20 n. ‘There is, however, no rigid distinction between
the two kinds of hair, and transitional forms are fairly
common.

Chlamydophorus truncatus.

The soft fur of Chlamydophorus is made up of fine
non-medullate hairs, the average breadth of which is 17 p.
The cortex is transparent and unpigmented, and contains
only a few scattered granular markings. ‘he scales of the
cuticle project considerably, and give a ragged appearance to
the surface of the hair (fig. 17).

Dasypus sexcinctus and villosus.

In both species the hairs are coarse, brown, and oval in
section. When examined from the side they show a fine
and close longitudinal striation, due to the arrangement of
highly refracting granules in fusiform series. The cuticular
scaling is close, and can be made out only with difficulty.
In Dasypus villosus (fig. 19) the section is less perfectly
oval than in Dasypus sexcinctus (fig. 18), since it tends
rather towards the rectangle in shape. ‘here is a distinct
slit-like medullary cavity in D. villosus, but this is wanting

~

STRUCTURE OF THE HAIRS OF MYLODON LISTAI. 405

in D. sexcinctus;! the cuticle, also, is considerably thicker.
In both species the granules in the cortex are most thickly
set at some little distance from the margin, thus leaving a
central part and a peripheral part of the cortex relatively
clear. The long diameter of the oval measures about 230 u
in both species.

Tolypeutes conurus.
y

Tolypeutes has solid hairs provided with a thin, finely
scaled cuticle. The minute structure very closely resembles
that of the Dasypus hairs; in fact, except for their lighter
colour, these hairs might be considered as of intermediate
character between those of the two species of Dasypus
examined. The sections are oval in shape (fig. 20), and
there is a central clear area suggesting a medulla such as
occurs in Dasypus villosus, but it has no cavity, and does
not stain differently from the cortex. The cortex contains
bright granules, not of a brown colour, disposed most thickly
around the central clearspace. Nearer the base of the hair
the section is circular (fig. 22), and has no central clear area.
The width of an average hair at its broadest part is 200 pu.

Tatusia novemcincta and pilosa.

The hairs of ‘atusia are clear, solid, and non-medullate,
with a sharply marked cuticular scaling and a very faint
longitudinal striation. In transverse section the cortex
appears very clear, and contains only a few highly refractive
colourless granules (fig. 23). These are uniformly distributed,
and are particularly scarce in Tatusia novemcincta. The
sections of the hairs of T. pilosa are oval, and measure 115 4
x 95 w; while those of T. novemcincta are circular in shape,
and measure 130 « across.

Mylodon Listai.
The hairs of Mylodon Listai are solid, and without any

1 Lonnberg (7, p. 162) speaks of D. sexcinctus as though its hairs
possessed a central pith,

406 W. G. RIDEWOOD.

trace of medulla. The width is very uniform, and measures
170 » throughout the middle six eighths of the hair. The
basal eighth is slightly narrower, and the free end tapers
gradually to a blunt point, which is missing from most of the
hairs. A perfect hair measures about 6 cm.in length.’ The
cortex would be quite clear and homogeneous but for the few
short, fusiform air spaces, which are visible both from the
side and in transverse sections. The vacuoles are uniformly
distributed in all my preparations, and I have been unable to
discover the peripheral clear zone of the cortex mentioned and
figured by Jacob (4, p. 62 and fig. 2). Transverse sections
from different parts of the hair are all similar in character
(fig. 24).

The cuticle is moderately thick, and stains deeply. When
the hair is examined from the side the cuticular scaling is
very clearly observable on the basal third (fig. 25), but
cannot be seen over the rest of the hair. This fact, together
with a certain anxiety to make this ground sloth conform in
its hair structure with the tree sloths, has led Lonnberg (7)
to conclude that the hairs of Mylodon Listai, as we know
them, are but the central cores of hairs which were provided,
like those of Bradypus, with a more perishable extra-
cortical layer. The fragments of adhering material, however,
which he alludes to as the remains of the extra-cortex, are,
judging by my own preparations, nothing but foreign matter
such as dried mud or portions of the shrivelled root-sheath.
On the basal part of the hair of the human head organic
cellular substance, probably derived from the inner root-
sheath, is commonly found attached to the cuticle long after
that part of the hair on which it is found has emerged from
the follicle. The fact of the cuticular scaling showing
only on the basal part of the hair appears at first sight to
support Lénnberg’s view, for in Bradypus the extra-cortex

1 The observations were made upon the specimen described by Smith
Woodward and Moreno in 1899 (18). In amore recently discovered speci-
men, less well preserved, the hairs are much longer. See Smith Woodward

(19).

STRUCTURE OF THE HAIRS OF MYLODON LISTAI. 407

is wanting on the basal part, and the scaling there is particu-
larly clear. But in Mylodon the transition is very gradual,
the scaling becoming fainter and fainter, and giving place to
a uniformly stained external layer, whereas in a damaged
Bradypus hair the scaling disappears in sharply outlined
patches which do not stain at all.

Lénnberg writes (7, p. 162) that in tangential sections of
the skin he saw around some of the hairs two rings, the outer
of which was the epithelial hair-sheath, while the inner he
took to be the “loose outer bark of the hair ’’—the extra-
cortex, in fact. Since, however, in the fully grown hairs of
neither Brady pus nor Cholepus is the extra-cortex found
within the follicle or anywhere near it, the argument fails to
carry as much weight as he evidently intended it to do. His
remarks on the following pages concerning the rings of dried
epithelial substance around the exposed bases of the hairs
are equally unfortunate. On the drying of the skin the hairs
become protruded, and that region of each which in life was
flush with the surface and closely surrounded by the general
epidermis is pushed some distance out, a millimetre or so,
dragging up the stratum corneum into the form of a cone.
As the cone dries it cracks horizontally at a little below the
summit, leaving an annulus attached to the hair. Thisis the
kind of thing which can be seen by examining with a lens
almost any museum skin of a mammal with a thick skin and
stiff hair.

Again, a glance at the transverse section of the Bradypus
hair taken through its broad part (fig. 3) is sufficient to show
that, were the extra-cortex removed by disintegration or
rough treatment, the central core would present a very
ragged, unstainable edge, quite unlike the uniform, smooth,
and deeply staining cuticular border of the transverse section
of the Mylodon hair (fig. 24). And lastly, the reappearance
of the cuticular scaling at the tip of the Bradypus hair (fig.
2) is not paralleled in the hair of Mylodon. The bulk of the
evidence appears, therefore, to be against Lénnberg’s theory,
and we must consider the hairs of Mylodon to be preserved

408 W. G. RIDEWOOD.

to us in their completeness. ‘They will bear a very close
comparison with the hairs of Tatusia (fig. 23), and are not
remarkably different from those of Tamandua (fig. 16) and
Dasypus sexcinctus (fig. 18).

It is a curious fact that in all the American Hdentates exa-
mined any characteristic features which each kind of hair
may possess is absent from the basal portion. The basal
parts of the hairs of Bradypus (figs. 5 and 6), Cholepus
(figs. 10 and 11), Myrmecophaga (fig. 15), and Toly-
peutes (fig. 22) agree with one another, and furnish a gene-
ralised type of hair structure, with which the whole hairs of
Tatusia, Tamandua, and Dasypus sexcinctus conform.
Since the hairs of Mylodon are in such close agreement with
this generalised type, it seems wiser to accept them as of
primitive and generalised structure than to attempt to estab-
lish a parallelism between them and those of the tree sloths,
especially in view of the fact that these latter are extremely
aberrant, and differ so remarkably inter se. ‘There is no
need to conclude that Mylodon is the less a sloth, and more
related to the ant-eaters and armadillos, because its hairs fail
to possess an extra-cortex.

As regards the arrangement of the hairs in the skin there
is not much to be said. The hairs are, as has been pointed
out by Smith Woodward (18, p. 149) and Lénnberg (7%, p.
164), all of one kind, there being no under-fur, and they are
uniformly distributed, without any signs of symmetrical
grouping. In Bradypus the follicles of the small hairs are
clustered around those of the principal hairs, as Leydig (6, p.
707), Welcker (17, pp. 68—70, and pl. i, fig. 4), and de
Meijere (9, p. 361) have shown ; while in Choten us the hairs
are arranged in bundles of two, though occasionally ao
and there is no proper under-fur.

Appended to Dr. Sorby’s remarks on Bradypus (15,
p. 339) is a foot-note by the editor! of the ‘ Linnean Society’s
Journal,’ which reads, “There is a small sloth, however, in
which the larger hairs are smooth and solid.” It is much to

' The late Mr. K, R, Alston,

STRUCTURE OF HE HAIRS OF MYLODON LISTAI. 409

be regretted that he did not mention the species he had in
mind. The existence of a sloth with such hair would, of
course, be of the greatest interest in the present connection,
and so I examined the hair of every species of sloth avail-
able at the Natural History Museum. The results, however,
do not enable me to confirm the editor’s remark.

10.

iT;

12.

13.

List or WoRrKS REFERRED TO.

. AMeGHINO, F.—“ An. Existing Ground-Sloth in Patagonia,” ‘ Natural

Science,’ xiii, London, 1898, pp. 324—326.

. Este, B.—‘ Die Lehre von den Haaren,’ Wien, 1831.
. Erpi, M.—“ Vergl. Darstellung des inneren Baues der Haare,” ‘ Abhandl.

Akad, Wiss., Munchen,’ ili, 1840—1843, pp. 413—453, three plates.

. Jacos, C.—‘‘ Examen microscdpico de la pieza cutanea del mamifero mis-

terioso de la Patagonia, Grypotherium domesticum,” ‘ Revista
del Museo de la Plata,’ 1899, x, pp. 61, 62.

. Lecut, W.—Bronn’s ‘ Klassen und Ordnungen des Thierreiclhis,’ * Sauge-
> D > BD

thiere,” Bd. vi, Abth. v, Lief. 45 u. 46, Leipzig, 1897.

. Leypic, F.—‘ Uber die ausseren Bedeckungen der Saugethiere,”’ ¢ Arch.

f. Anat. u. Phys.,’ Leipzig, 1859, pp. 677—747, two plates.

. LonnBerc, E.—‘‘ On some Remains of Neomylodon Listai,” ‘ Wiss.
y >

Ergebnisse der schwedischen Exped. Magellansland.,’ Bd. 11, ‘ Zool.,”
Stockholm, 1899, pp. 149—170, three plates.

. Maurer, F.—‘ DiesEpidermis und ihre Abkommlinge,’ Leipzig, 1895.
. Meere, J. C. H. pe.—‘ Uber die Haare der Siugethiere, besonders

iiber ihre Anordnung,” ‘ Morph. Jahrb.,’ xxi, Leipzig, 1894, pp. 312—
424.

NorpENskJOLD.—“ Jakttagelser och Fyndi Grottor vid Ultima Esperanza
i sydvestra Patagonien,” ‘ K. Vetensk.-Akad. Handl.,’ vol. xxxiii, No.
3, 1900.

Owen, R.—‘ The Voyage of H.M.S. Beagle,’ Part I, ‘ Fossil Mammalia,”
London, 1840.

Owen, R.—‘ Description of the Skeleton of an Extinct Gigantic Sloth,
Mylodon robustus,’ London, 1842.

REINHARDT, J.—“ Beskrivelse af Hovedskallen af et Kempedovendyr,
Grypotherium Darwinii,” ‘ Vidensk. Selsk. Skr., 5 Rekke. na-
turvid. og math. Afd.,’ xii, 4, Copenhagen, 1879, pp. 353—380, two
plates.

voL. 44, PART 3.—NEW SERIES. DD

410 W. G. RIDEWOOD.

14. Rotu, 8.—‘‘ Descripcion de los restos encontrados en la caverna de
Ultima Esperanza,” ‘Revista de] Museo de la Plata, ix, La Plata,
1899, pp. 421—458.

15. Sorsy, H. C.—“‘On the Green Colour of the Hair of Sloths,” ‘Journ.
Linn. Soc.,’ xv, London, 1881, pp. 337—341.

16. WaLpEYER, W.—‘ Atlas der Mensch. und Thierischen Haare,’ Jahr
1884.

17. Wewcxer, H., and Ktun, J.—* Ueber die Entwicklung und den Bau der
Haut und der Haare bei Bradypus,” ‘Abhandl. naturf. Gesell. zu
Halle,’ ix, 1864, pp. 17—72d, two plates.

18. Woopwakp, A. Smitu, and Morrno, F. P.—“On a Portion of Mam-
malian Skin, named Neomylodon Listai, from a Cavern near
Consuelo Cove, Patagonia,” ‘Proc. Zool. Soc.,’ London, 1899, pp.
144—156, three plates.

19. Woopwarp, A. Smirw.—‘‘On some Remains of Grypotherium
(Neomylodon) Listai and associated Mammals from a Cavern near
Consuelo Cove, Patagonia,” ‘Proc. Zool. Soc.,? London, 1900, pp.
64—79, five plates.

EXPLANATION OF PLATE 26,

Illustrating Dr. W. G. Ridewood’s paper “ On the Structure
of the Hairs of Mylodon Listai and other South
American Kdentata.

al. Alga. co. Cortex. cw. Cuticle. eco. Hxtra-cortex. m. Medulla.

Fic. 1.—Bradypus tridactylus. Hair, one and a half times natural
size.

Fig. 2.—Terminal portion of hair (a in Fig. 1). x 100.

Fig. 3.—Transverse section taken through the thickest part of the hair
(Bin Fig. 1). x 100.

Fie. 4.—Corresponding part of hair seen longitudinally ; the upper part in
optical section, the lower part with the surface in focus. x 100.

Fies. 5 and 6.—Transverse section and side view of basal part of hair
taken in the position indicated by c in Fig. 1. x 100.

Fie. 7.—Portion of one of the fine hairs of the under-fur, and a transverse
section of the same. xX 100.

STRUCTURE OF THE HAIRS OF MYLODON LISTAI. 411

Fie. §8.—Cholepus didactylus. ‘Transverse section through the
middle of the length of the hair. The superficial grooves are occupied by
shrivelied extra-cortex and alge. At a the groove is completely, and at 4
partially closed over by cuticle. x 150.

Fie. 9.—Portion of the hair seen from the side. x 150.

Fies. 10 and 11.—Transverse section and side view of the hair at a point
one eighth of the total length from the base. x 150.

Fig. 12.—Optical section of a hair of a soft-furred specimen of Cholepus
didactylus, showing in the groove, which is in focus on the right side of the
figure, the cell outlines of the extra-cortex. x 150.

Fic. 13.—Transverse section of the same hair. x 150.

Fic. 14.—Myrmecophaga jubata. Transverse section through the
middle of the length of the hair. x 100.

Fie. 15.—Transverse section through the hair at about 3 mm. above tlie
surface of the skin. x 100.

Fic. 16.—Tamandua tetradactyla. ‘Transverse section through the
middle of the length of the hair. x 150.

Fic. 17.—Chlamydophorus truncatus. Hair seen in transverse sec-
tion and from the side. x 600.

Fie. 18.—Dasypus sexcinctus. ‘Transverse section through the middle
of the hair. x 100.

Fig. 19.—Dasypus villosus. ‘Transverse section through the middle
of the hair. x 100.

Fic. 20.—Tolypeutes conurus. ‘Transverse section through the middle
of the hair. x 100.

Fig. 21.—Middle part of the hair seen from the side; upper part in optica.
section, lower part with the surface in focus. x 100.

Fie. 22.—Transverse section through the hair at about one quarter of its
length from the base. x 100.

Fig, 23.—Tatusia pilosa. Transverse section through the middle of
the length of the hair. x 150.

Fie. 24.—Mylodon Listai. ‘Transverse section through the middle of
the hair. x 100.

Fic. 25.—Portion of the hair about one quarter of the total length from
the basal end; surface in focus to show the cuticular sealing. x 100.

os a :
nets bike it plied 2.40 ais eh ihe

ce oe

XL hi) rat eh

a o

es. | at). Writs oe seie. 390 f Wit
= WL x) ae ee ae

a» 7 - ;
Ait 3 3a SP ka IT 9 7. ke th sey a Ml

‘ in :
Ray # ta, oe att i al wd- ton Mie Min. wy aed, rust a ws ie if .
i aS sau yas sein AP) ok lary gia x

1 - .

Pt eo Fal) Ree. oe gee Seen i fe ~ ®
oe ie i as Te ee gah phre Oma, x3ht

ee

2 a5 & so aan

tie f WEED te Ae Bust yi
ay . . = ¢
r , ;

se Bese ot 1) SP - fine SS A) eA

a: ' 7 ; p » §Ak
Je ee id Oat 7 Oe ee

if eee te pe oe . '

i a | :
rk eT ae pide ha f
‘ ’ 4,
s
( } ait an y
a 5: L
t i j ars
3 ‘ % oP i Ge
Ee 7 Tye j p
a ¥ ”
: : x
(Hie Fs
- ° y !
4
ay oF +f
4 a
i Aa
7 = - a ol wae
' pats - - + . <a
'
/ aS
2 ~e th 4 ' = 2 fs
A
i
é ‘ s 7 |
re - = © 4 ,
’ i 1
' > ey =
4
a *” » % ‘ r ia : “oy
vars ots
‘
: +s
; 4
4 Pia y ei ae ia)
. = =
\ 4 he
5
-
5 I 2 ; 22" re i A
AGQE ae
| .
- oo be % + 7 ee
- 4 : & \ Shee
i Wi <5 infty ‘ ¥ ae ' rr -
; erly “secds : ott
cP w -
wr
= }

THE STRUCTURE AND AFFINITIES OF SACCOCTRRUS. 413

On the Structure and Affinities of Saccocirrus.
By

Edwin 8. Goodrich, M.A.,
Fellow of Merton College, Oxford.

With Plates 27—29.

Wuen studying the morphology of the excretory organs
of the Polycheeta, I thought it advisable to investigate those
few cases of worms in which the true nephridia have been
alleged to function as genital ducts. Of these Polychetes,
Saccocirrus papillocercus, Bobr., is one of the most
interesting; and whilst occupying for a short time last
winter, at the Zoological Station at Naples, a table kindly
placed at my disposal by the committee of the British
Association, I had an opportunity of studying its structure.

Unfortunately the results are somewhat disappointing,
since I have not been able to discover any facts of such
decisive significance as to settle the difficult question of the
morphological value of the organs which contribute to form
the complex genital apparatus of this httle worm. However,
as I was able to complete and to correct the accounts of
previous authors in many points of detail, | publish this
paper as a small contribution to the subject, and have
appended some general remarks concerning the validity of
the group ‘“ Archi-annelida.”’

Marion and Bobretzky,! in 1875, published an excellent
account of the structure and habits of Saccocirrus, from
material studied at Marseilles (10). Since then Langerhans

} Bobretzky’s original paper (1) I have not been able to consult.

414 EDWIN S. GOODRICH.

(8) has mentioned this worm in a paper on the fauna of
Madeira, and Fraipont has made a detailed study of its
nervous system (8).

External Characters.—I have nothing to add concern-
ing the external morphology of Saccocirrus, excepting with
regard to its parapodia and chet. Unlike what has been
described by Marion and Bobretzky, the parapodia in my
specimens do not extend all the way from the second! to the
last segment; but there is at the posterior end of the animal
a variable number of segments (from ten to twelve), on which
neither parapodia nor chetz are present. The transition
from the region with parapodia to the region without is
marked by one or two segments in which the parapodia and
cheetee are rudimentary. In every bundle of chete, besides
the ordinary chetz described with blunt ends, I find one
long needle-like cheeta, the tip of which is divided into three
sharp prongs (fig. 9).

It is of course possible that the Saccocirrus of Naples is
not of the same species as the worm described by Marion
and Bobretzky from Marseilles ; but on the whole this seems
unlikely, since Saccocirrus papillocercus has also been
found in the Black Sea by its original describer, and at
Madeira by Langerhans. The discrepancies in the descrip-
tions may vanish on a closer inspection of specimens from
the other localities.

Nervous System.—It is well known that the central
nervous system consists of a brain in the prostomium, from
which two nerve-cords run along ventrally above the epi-
dermis on either side of the body. Fraipont also found,
extending along the cesophagus, two strands, which he con-
jectured must be of nervous nature, although he was unable
to follow them to the brain. In my series of sections these
two nerves can be plaiuly seen to arise from the ventral sur-
face of the brain, more towards the middle line, but quite
near to the origin of the main nerve-cords (fig. 2). They
pass backwards along the roof of the buccal cavity (fig. 16)

1 The ‘first segment” is probably formed of two fused segments,

THE STRUCTURE AND AFFINITIES OF SACCOCIRRUS. 415

and the sides of the cesophagus (figs. 17, 18, and 19), below
which they join to form a complete loop just behind the
muscular pharyngeal sac (figs. 20 and 2). A few nuclei,
especially in this region, indicate the presence of ganglionic
cells. Obviously this loop represents the stomatogastric
system.!

A small knob with sensory hairs is situated behind each
parapodium.

Alimentary Canal.—Hitherto the alimentary canal of
Saccocirrus has been described as quite simple, without
special muscular pharynx. In my specimens, however, I
find a well-developed muscular pharyngeal sac below the
cesophagus,” opening forwards into the buccal cavity (figs.
1 and 2), and extending backwards into the third segment.
The roof of this diverticulum is thin, and the floor is
thickened into a sort of muscular pad (figs. 1, 19, and 20).
Special muscles, passing from the hinder lip of the mouth
backwards round the sac, and then forwards to be attached
dorsally to the wall of the cesophagus, serve no doubt to
push the sac and its pad forwards and outwards; but I have
never seen this apparatus in action.

The inside of the sac is lined with rather thick cuticle, and
is not ciliated like the rest of the digestive tract. Behind
this organ is the glandular region with muscular walls, repre-
senting the digestive stomach. It reaches, as already de-
seribed by Marion and Bobretzky, to about the fourteenth seg-
ment. Following on this is the long sacculated intestine, the
absorptive region, covered externally with chloragogen cells.

Vascular System.—Of the blood-vascular system,
Marion and Bobretzky only found the dorsal vessel; Frai-
pont figured a ventral vessel (3). Langerhans had previously
described this ventral vessel, and a vessel passing into the

1 Surely it is to such a stomatogastric system that the subcesophageal
ganglion of Rotifers is to be compared, and not to the ventral nerve-cords of
the body-wall as urged by Hisig (“ Zur Entwicklungsgeschichte der Capitel-
liden,” ‘ Mitth. Zool. Stat. Neapel,’ vol. xiii, 1898).

* The sac is absent in two out of a dozen series of sections,

416 EDWIN S. GOODRICH.

tentacles (8). I can confirm this author’s discovery of the
two main longitudinal vessels, but not that of a tentacular
vessel. The dorsal vessel divides below the brain into two
branches, which run down on either side of the cesophagus
to join below, and behind the pharyngeal sac to form the
ventral vessel (figs. 17 to 21).

Ccelom.—Marion and Bobretzky described the spacious
ccelom, subdivided by median dorsal and ventral mesenteries
and by transverse septa, and traversed by the oblique muscle
strands. Fraipont has figured the ccelom as filled with a fine
reticulum of stellate cells (2 and 8). It may be the case that
at a certain age, or at a certain time of the year, the cavity
of the body is so filled; but I have never observed this con-
dition myself, and I am inclined to believe that Fraipont has
been somewhat misled by appearances brought about by the
coagulating action of fixatives, or by studying sections taken
in the region of a septum, or close to the head, where nu-
merous tissue strands extend from the anterior end of the
cesophagus to the body-wall, as in worms generally.

The wall of the intestine is covered externally with large
irregular cells projecting far into the ccelom, and filled with
coloured granules (fig. 5). The ccelomic epithelium lining
the other parts of the body-cavity is composed of cells, which
often are so filled with vacuoles that they project consider-
ably into the coelom (fig. 9). When a Saccocirrus is viewed
compressed under a cover-glass the body-cavity may appear
to be almost obliterated, but this appearance is deceptive.

Head Cavity.—Amongst the most peculiar characters of
Saccocirrus may be reckoned the possession of a special
closed cavity in the head, consisting of a right and left
ampulla, situated in the peristomial segment, passing forwards
into canals which run up the prostomial tentacles (figs. 1, 2,
and 17). A transverse communicating canal runs across
beneath the posterior edge of the brain (fig. 16). The walls
of this system are provided with muscles—longitudinal in the
tentacular canal, circular as well in the ampulla. Its cavity
is filled with a fluid and lined by an epithelium.

THE STRUCTURE AND AFFINITIES OF SACCOCIRRUS. 417

Marion and Bobretzky correctly described the head cavity,
and believed its function to be erectile; contraction of the
ampulle driving the contained fluid into the tentacular
canals, and so straightening out the tentacles. This inter-
pretation is very probably correct, but I have never myself
observed any very manifest signs of such action. Langerhans
added the observation that the wall of the head cavity is
provided with an outer cuticular lining. Staining with
nigrosin shows the covering coloured dark blue like the
cuticle. Fraipont has figured transverse sections of the
system, the lumen being drawn as if filled with a reticulum,

Whilst fully confirming most of these authors’ statements,
it may be added that the contents seemed to me to be not
colourless, as described, but pale pink; and further, that the
“ fluid” is really composed chiefly of large cells with granular
and very fluid contents closely packed together (figs. 1 and
17). No doubt there is some intervening liquid, but it does
not appear in sections.

The morphological significance of this head cavity is not
easy to determine. It seems to me probable that it represents
the specialised ccelomic cavity of the peristomial segment
(see p. 404).

Excretory and Genital Organs.—Our knowledge of
these organs is due entirely to the observations of Marion
and Bobretzky, some of which Langerhans confirmed. This
description, on the whole, is remarkably good, but it is
incomplete and incorrect in a few points.

These authors found that the sexes are separate; that the
testes and ovaries are developed from the ccelomic epithelium
on the posterior surface of the septa; that in the male there
is, in the region behind the cesophagus, on either side of each
segment a protrusible penis with sheath, leading into a
ciliated duct which enlarges to form a vesicula seminalis,
continued into an open funnel, ‘ disposé d’aprés la forme
typique des organes segmentaires”’ (10). In the esophageal
region they found ordinary nephridia opening at the same
level as the penis. Naturally they believed that the male

418 EDWIN S. GOODRICH.

ducts were merely modified nephridia: ‘‘ Nous croyons que
le Saccocirrus est jusqu’ici le premier exemple, parmi les
Annélides Polychétes, de vers dans lequels les organes
segmentaires du male se transforment et deviennent de
véritables appareils de copulation.”

In the female they described nephridia, as in the male,
which they believed to open to the exterior near the para-
podium, and to function as oviducts in the genital region,
Bobretzky having seen the nephridial canal dilated with
eggs in ripe specimens. Besides these, a pair of sperma-
theese were found in every genital segment opening ventrally
by a narrow duct, with cilia producing a current inwards.
To explain how fertilisation takes place, Marion and
Bobretzky added that “il faut supposer que ces vésicules
[spermathece] s’ouvrent dans la grande chambre ot les
ovules s’accumulent.”

Now when this investigation was begun I hoped to find
this internal opening, and to interpret the excretory organs
of the female as true nephridia, and the spermathece as
ccelomostomes (5). But such an opening does not appear to
exist. The spermatheca consists of a pear-shaped sac, with
a long duct passing straight through the ventro-lateral
longitudinal muscles (figs. 14 and 21) to open to the
exterior. The wall of the duct is formed of ordinary ciliated
epithelium (figs. 3 and 13); but the cilia do not reach far
into the sac, where the epithelium soon becomes altered in
character. Near the base of the sac, surrounding the
entrance of the duct, is a cup-shaped region where the lining
is formed of very large granular cells containing yellow
granules, and at their inner ends large irregular angular
bodies of yellow refringent matter (figs. 3,4,and 13). Similar
bodies are distributed in the epithelium lining the swollen
end of the sac, and together with the granules give the
spermatheca its yellow tinge. What the function of these
bodies can be it is difficult to guess; possibly they serve as a
reserve of food material for the spermatozoa. There is no
reason to consider them as of an excretory nature,

THE STRUCTURE AND AFFINITIES OF SACCOCIRRUS. ‘419

The nephridium in the female is a slender tube running
in the angle between the oblique muscles and the body-wall,
to open forwards into the ccelom through a septum (figs. 3
and 22). In the genital segments (intestinal region), when
it reaches a point nearly opposite the spermathecal pore, the
nephridium turns sharply downwards, passing between the
epidermis and the longitudinal muscles, until it touches the
spermathecal duct, into the very base of which it opens (figs.
3, 6, 21, and 14). This downward limb of the nephridium
becomes very narrow, having a much diminished lumen
towards the minute opening. ‘The wall of the nephridium is
formed of granular, much vacuolated protoplasm ; a few cilia
are seen in the lumen (fig. 3). There is no projecting lip
to the funnel, which merges rapidly into the epithelium
covering the septum (in my specimens, which were not quite
mature). Round the opening are long cilia passing down
into the lumen of the canal (fig. 3).

In the male the cavity of the penis is lined with granular
cells, and its wall is strengthened by a number of delicate
refringent cuticular rods pointed at both ends (figs. 8 and 9).
The sperm-sac and nephridium do not form one continuous
duct, as described by Marion and Bobretzky. Coming off
from the penis is a ciliated duct, which soon widens out into
a pear-shaped sac lying, unlike the spermatheca, entirely in
the lateral chamber of the ccelom (figs. 9, 10, 11, and 14).
This sperm-sac contains ripe spermatozoa, and ends blindly
at its swollen extremity. In the genital segments the
nephridia, quite similar organs to those of the female in
their general structure, open into the duct of the sperm-sac
near its entrance into the penis. ‘lhe funnel in these seg-
ments is much enlarged, richly ciliated, and spreads for a
considerable distance over the anterior face of the septum
(figs. 9, 12, and 14).

Small-funnelled nephridia are found in both sexes in the
segments of the cesophageal region, beginning after the first
bundle of chete (fig. 15). But whereas in the male they
open dorsally, at the level of the penis in the more posterior

420 EDWIN S. GOODRICH.

segments, in the female they open ventrally as in the genital
segments.

It will be seen that the general relations of the genital
organs is really much the same in both sexes; that the
sperm-sac can be compared to the spermatheca; and that if
an invagination similar to that which has given rise to the
penis took place in the female at the mouth of the sperma-
theca, a condition would be brought about almost identical
with that which obtains in the male. Whether the sacs
themselves are formed from invaginations of the skin can
only be decided by a study of their development; on the
whole, it seems probable that this is the case.

As to the morphological value of the excretory organ,
whether it be a true nephridium with a nephridiostome, or a
nephromixium formed of a ccelomostome grafted on to a
nephridium (5), only a knowledge of development can here
again help us to decide. The structure of the male organ
especially seems to lend itself readily to the latter interpre-
tation. The main part of the canal appears to be undoubt-
edly of nephridial nature, comparable to the very similar
nephridia of the Polygordiide (2 and 5). On the other
hand, the funnel has very much the appearance of being
derived from the coelomic epithelium (figs. 9 and 12).

Summary and Conclusion.

The chief additions to our knowledge of the structure of
Saccocirrus recorded in the foregoing pages may be briefly
enumerated as follows :—the parapodia and cheetze are absent
from the last ten or twelve segments; in each bundle of
chete, besides the ordinary blunt bristles, is one long
slender cheta, ending in three prongs; there is a stomato-
gastric nervous system, consisting of two nerves passing
backwards from the brain, and joining below the cesophagus ;
the buceal cavity is prolonged into a muscular diverticulum
below the cesophagus, which forms a ventral pharyngeal
pouch, lined with cuticle, and probably eversible ; the dorsal
blood-vessel divides below the brain into two branches,

{THE STRUCTURE AND AFFINITIES OF SACCOCIRRUS. 421

which-pass down on either side of the cesophagus to join
below it in the ventral vessel; in the male genital region
the nephridia which open into the ccelom by wide funnels
spreading over the front of the septa, pass backwards to
open by a narrow tube into the ducts of the sperm-sacs
before their entrance into the penes; the penis is provided
with a number of supporting cuticular rods in its wall; in the
female the nephridium of the genital region opens in front
by a small ccelomic funnel, and runs backwards to open by a
minute pore into the base of the spermathecal duct.

Although we cannot hope, in the present state of our
knowledge, to definitely determine the affinities of Sacco-
cirrus, yet a general review of the question may be made
with some profit,

Marion and Bobretzky (10) considered Saccocirrus to be
allied to Polygordius ; Hatschek (6), followed by Fraipont
(2), believing the Polygerdiude to represent an Archi-an-
nelidan group outside the Polychxta, places Saccocirrus at
the beginning of the Polychetes, between the Archi-annelids
and the Opheliide.' It is obvious that to determine the
position of Saccocirrus, we must first of all inquire into the
affinities of the Polygordiide. The Polygordiude comprise
the two genera Polygordius and Protodrilus. ‘The cha-
racters on which Hatschek founds his opinion that these
worms form a group ancestral to the remainder of the Anne-
lida are the following:—the absence of parapodia and
cheetze ; the homonomy of the segments, and the fact that
segmentation is chiefly internal (a statement which scarcely
agrees with the structure of Protodrilus, however); the
restriction of the pharynx to the buccal segment; the close
connection of the nervous system with the epidermis; the
absence of ventral ganglia; the simplicity of the muscula-
ture, there being no circular muscles (except in Polygor-
dius Villoti, described by Perrier [11] ) ; the presence of
dorsal and ventral mesenteries, and the simplicity of the

' Perrier, in lis ‘ Traité de Zoologie’ (1897), places the Polygordiide near
the Phyliodocide.

422 EDWIN S. GOODRICH.

vascular system. ‘l'o these characters "raipont added the
simple primitive structure of the nephridia running in the
body-wall, below the ccelomic epithelium.

That the general organisation of Polygcrdius and Proto-
drilus is simple cannot be denied ; but that ihis simplicity is
necessarily of an archaic nature remains to be proved. To a
great extent it may be due to a general tendency tc simplifi-
cation shown in the smaller representatives of many families
of Polychetes, especially amongst sand-inhabiting forms.
It is well known that in the smaller Syllids, Ophelids,
Hunicids, etc., the nerve-cords are closely connected with the
epidermis. The intimacy of the connection between the
epidermis and the nerve-cords in the Polygordiide appears
to me to have been a little exaggerated by Fraipont (2).
Protodrilus I have not studied ; but in Polygordius the dis-
tinction between the two does not seem to me to be much
less marked than in many small Polychetes (in Saccocirrus a
line of demarcation can always be made out separating the
nerve-cord from the epidermal cells). However, the absence -
of ganglionic concentrations may certainly be archaic; but
the presence of ganglia in such forms as Dinophilus and
Histriodrilus, animals considered by some authors to be
allied to the Polygordiide, tends to show that it is a cha-
racter of no very fundamental importance.

Again, with regard to the absence of circular muscles ; not
only have they been expressly stated to exist in Polygor-
dius Villoti by Perrier (11), but surely if we consider how
well developed they are in Nemertines and Platyhelminths,
the assumption that their absence is primary and not secon-
dary does not seem to be justified. ‘lhe peculiar develop-
ment of the oblique muscles in the Polygordiide is much
more like what we find in many Polycheta where the para-
podia are reduced (Arenicola, Capitellide, Opheliide, etc.),
than anything we know of in the lower classes of Coslo-
mata.

Moreover little importance can be attached to the position
of the nephridia; in almost all Polychaeta these organs are

THE STRUCTURE AND AFFINITIES OF SACCOCIRRUS. 423

primarily covered by the ccelomic epithelium—even in the
adult ; and in Polygordius the nephridia are not always in the
body-wall, as stated by Fraipont, but may run (as in Poly-
ophthalmnus, for instance) along the oblique muscles crossing
the celom, as shown in fig. 7. The persistence of both
mesenteries is probably a truly archaic character. This and
the absence of the parapodia and chetz are the only points
of first-rate importance common to Polygordius and Proto-
drilus in the list of alleged primitive characters. It is, then,
to the absence of parapodia and chet that we must turn
our attention.

Let us try, for the sake of argument, to conceive what sort
of creature the common ancestor of the Annelida (Polycheta,
Oligocheta, Hirudinea, Hchiuroidea, and Myzostomaria)
must have been. It may safely be conjectured that it was a
creeping segmented worm, with a skin, whether ciliated or
not, covered by a cuticle ; with metameric bundles of chetz,
longitudinal ventral nerve-cords ; circular, longitudinal, and
dorso-ventral muscles, septa, and mesenteries; with seg-
mental nephridia (probably not opening into the ccelom),
ccelomic cavities and ccelomostomes (genital ducts) leading
to the exterior. Such may have been approximately the
structure of the Annelid common ancestor in pre-Cambrian
times. This primitive Annelid must have been itself derived
from a form which, we may presume, to some extent ap-
proximated to the plan of structure now elaborated along
the Nemertine and Platyhelminth lines of descent. In other
words, it was probably derived from worms in which the
muscles and parenchyma were well developed, but in which
the ccelom and internal segmentation were not so well differ-
entiated, and which were provided with nephridia ending in
flame-cells.

Now this is just the position the Polygordiide occupy
according to the Archi-annelid theory. Can it be truly said
that they fit in the place assigned to them? Moreover can
it be believed that these little modified ancestral forms have
persisted to the present day, and live happily together with

494, EDWIN 8S. GOODRICH.

the highly modified Chetopods, which are supposed to be
derived from them, in the Bay of Naples? The absence of
parapodia and of chaete are the only characters which seri-
ously entitle the Polygordiide to such a bold claim. There-
fore, if it can be shown that both chate and parapodia have
been reduced or even lost in other cases; and if, further, it
can be shown that the Polygordiide are quite nearly related
toa form possessing parapodia and chete, the “ Archi-annelid
theory ” will be severely shaken.

That parapodia may disappear more or less completely is
shown in many families of Polychetes. One may mention
especially the Scalibregmide, Chlorhamide, Sternaspide, and
the Opheliide, to which family McIntosh (9) and Giard (4)
believe Polygordius to be closely related. The Oligocheta
exhibit an example of the total loss of chate (Anacheta
and the Discodrilide ?), and of course in almost all the Hiru-
dinea they have entirely vanished; but amongst the Poly-
cheeta it may be pointed out that in Arenicola a considerable
region of the body is devoid of both chate and parapodia.
In Tomopteris also no cheiz are present on the trunk seg-
meuts.

More important still is the near relationship which un-
doubtedly exists between the Polygordiide and Saccocirrus.
Already Marion and Bobretzky, as mentioned above, remarked
on this affinity, and Fraipont has further insisted on it. But
the most convincing evidence seems to have been brought
forward by Langerhans (8); curiously enough it appears to
have escaped the notice of later writers on this subject. The
piece of evidence to which I refer concerns the very remark-
able contractile head cavity of Saccocirrus, hitherto generally
supposed to be unique.

Describing Protodrilus (Polygordius) Schneideri,
Langerhans says, ‘“Innen von den Gefassen liegt, ganz wie
bei Saccocirrus, in den T'entakeln ein grésserer Hohlraum
(Fig. 47), welcher im Kopf zwischen Quergefaéss und Hirn
mit dem der anderen Seite zusammenhiangt, von diessem
Verbindungstiick geht ein kleiner dorsaler Forsatz ab

THE STRUCTURE AND AFFINITIES OF SACCOCIRRUS. 425

(Fig. 48). Durch die lebhaften Bewegungen der Filer
werden in diesen Hohlraumen lose Zellen hin und her getrie-
ben” (p. 126 [8]). Moreover Uljanin evidently figures a
similar cavity in the head of Protodrilus flavocapitatus
(figs. 5,h.,h’., and 15, sch. [12]). Hatschek himself, in describ-
ing Protodrilus Leuckartii, mentions a cavity in the
peristomial segment which runs up the tentacles, and is filled
with reddish fluid (the blood is colourless). There seems to
be no doubt, then, that in Protodrilus, as in Saccocirrus, there
exists a special ‘‘ head cavity’ in connection with the ten-
tacles, a cavity which, in fact, 1 believe to be the modified
ccelom of the first segment.

All known species of the genus Protodrilus are provided
with a ventral, muscular, pharyngeal sac below the esophagus.
As further evidence in favour of the close affinity advocated
above, one may bring forward the discovery of a very similar
sac, lined with cuticle and with muscular walls, in Saccocirrus.
In the details of its structure the pharyngeal sac of Sacco-
cirrus seems to be more like that described in Protodrilus
Schneidert than like that of Protodrilus Leuckartii.

In the presence of these facts we are forced to the conclu-
sion that in some important respects Protodrilus is much
more closely allied to Saccocirrus than to Poly-
gordius! Wherever we put Protodrilus in our system of
classification, there also we must place Saccocirrus. On the
other hand, that Polygordius and Protodrilus are nearly
related seems to be almost equally certain. There remains,
therefore, no alternative but to unite all three genera in one
group. And if we do this, many of the list of supposed
primitive characters of the ‘‘Archi-annelida” (Polygordiide)
are struck off at one blow. No longer can we enumerate as
characters of this group the absence of parapodia and chete,
of external segmentation, of circular muscles, or even of
muscles in the wall of the gut, for these statements do not
apply to Saccocirrus. And, moreover, it may be added that in
this worm the segments are not homonomous, and the pharyn-
geal sac extends into the third segment.

' T have observed a similar dorsal process of the wall in Saccocirrus.

VOL. 44, PART 3.—NEW SERIES. EE

426 EDWIN S. GOODRICH.

This group, containing the genera Polygordius, Protodri-
lus, and Saccocirrus, to which Lankester’s name Haplodrili
might still be applied, we should regard not as ancestral to
all the Annelida, but as composed of specialised offshoots of
the Annelid stem (probably, indeed, of the Polychete stem
itself), in some respects primitive, in other respects highly
specialised.! As specialised characters we may reckon the
development of a contractile “ head cavity ” (Protodrilus and
Saccocirrus), of a very complex genital apparatus (Saccocir-
rus), and the partial (Saccocirrus) or total loss of parapodia
and cheetz (Polygordius and Protodrilus). All three genera
possess two prostomial tentacles.

If it be objected that in the foregoing pages I have
tacked much theoretical speculation on to very little fact, my
excuse must be that I have endeavoured not so much to add
to the number of existing theories as to diminish it, and so to
bring back the question of the affinities of Polygordius and
Saccocirrus to the position it occupied before the “ Archi-
annelid theory”? was put forth.

The remarkably close affinity which has been shown to
exist between Saccocirrus and Protodrilus seems to force on
us the conclusion that the absence of parapodia and cheete in
the Polygordiide is not primitive, but secondary.

List oF REFERENCES.
1. Boprerzky, N.—‘‘Saccocirrus papillocercus, nov. gen., n. sp.,”
‘“Mém. Soc. des Naturalistes de Kiew,’ 1871.

2. Frarpont, J.—‘‘ Le genre Polygordius,” ‘Fauna u. Flora des Golfes
von Neapel,’ vol. xiv, 1887.

3. Fratpont, J.—‘‘ Le systeme nerveux . . . des Archiannélides,” ‘ Arch.
Biol.,’ vol. v, 1884.

4, Giarp, A.—“ Sur les affinités du genre Polygordius,”’ ‘ Comptes rendus,’
vol. xci, Paris, 1880.

1 Giard seems to me to have been very near the truth when he wrote of
Polygordius, “ C’est un type d’Annélide archaique et aberrant” (4).

THE STRUCTURE AND AFFINITIES OF SACCOCIRRUS. 427

5. Goopricu, EH. S.—“ On the Nephridia of the Polychaeta,” Part 3, ‘ Quart.
Journ. Micr. Sci.,’ vol. 43, 1900.

6. HatscHex, B.—‘‘ Studien wher Entwickl. der Anneliden,” ‘Arb. Zool.
Inst., Wien,’ vol. i, 1878.

7. Hatscuer, B.—“Protodrilus Leuckartii,” ibid., vol. iii, 1881.
8. Lancernans, P.—“ Die Wurmfauna von Madeira,” ‘ Zeit. f. wiss. Zool.,’
vol. xxxiv, 1880,
8. McIntosu, —.—“ Note on Linotrypane apogon,” ‘Ann. and Mag.
Nat. Hist.,’ ser. 4, vol. xvi, 1875.
10. Marion, A. F., and Bopretzky.—“ Annélides du Golfe de Marseille,”
‘Ann. Sei. Nat. Zool.,’ sér. 6, vol. ii, 1875.
11. PERRIER, E.—‘‘ Sur un nouveau type. . . Polygordius Schneideri,”’
‘Comptes rendus,’ vol. xc, Paris, 1875.
12, Unsanin, B.—“ Observations on Polygordius, etc.,’”’ in Russian. ‘ Bull.
des Naturalistes de Moscou,’ vol. i, 1877.

EXPLANATION OF PLATES 27—29,

Illustrating Mr. Edwin S. Goodrich’s paper ‘‘ On the Struc-
ture and Affinities of Saccocirrus.”

All the figures, excepting No. 7, are of Saccocirrus papillocercus.

Fic. 1.—Nearly median sagittal section of the anterior end, showing the
root of the tentacle and the pharyngeal sac. Cam. D, oc. 2.

Fie. 2.—Somewhat diagrammatic representation of the anterior end, drawn
from living specimens and preparations. The internal organs are seen by
transparency ; the nervous system is coloured yellow.

Fic. 3.—View of a portion of a genital segment of a female. The neplii-
dium and spermatheca are shown in optical section. From the living. Oil
im. 75-

Fic. 4.—Highly magnified view of a small portion of the wall of the sper-
matheca in optical section.

Fig. 5.—Similar view of a portion of the wall of the intestine.

Fie. 6.—Highly magnified view of a small region of the ventral surface of
a genital segment of a female, showing beneath the transparent skin the
*unction of the spermathecal and nephridial ducts. Oil im. 3.

428 EDWIN 8. GOODRICH.

Fic. 7.—Portion of a transverse section of Polygordius neapolitanus
showing the nephridium lying on the oblique muscles. Cam. D, oe. 3.

Fie. 8.—One of the supporting rods of the penis. ,

Fic. 9.—Portion of three genital segments of a male, seen under pressure
from above. The internal organs are represented in optical section. From
the living. Oil im. #5.

Fic. 10.—Portion of a transverse section passing through a penis. Cam.
Doerr 2:

Fie, 11.—Part of a similar section passing through the nephridium and
sperm-sac. Cam. D, oc. 2. ies

Fic. 12.—Section through the septum and edge of the funnel im a male
genital segment. Cam. Oil im. 4, oc. 3.

Fic. 13.—Section through the spermatheca and its duct from a longi-
tudinal series. Cam. Ap. 4 mm., oc. 3. ;

Fic. 14.—Diagram of the male and female excretory and genital ducts,
seen in transverse section.

Fie. 15.—Portion of a sagittal section of several segments in the cesopha-
geal region of a female, showing the nepbridia opening into the celom. Cam.
D, oc. 2.

Fics. 16, 17.—Two transverse sections of one series, the first taken
through the posterior end of the brain, the second across the mouth. Cam.
ID Nocw2:

Figs. 18 —20.—Three transverse sections of one series, showing the posi-
tion and structure of the ventral pharyngeal sac. Cam. D, oc. 2.

Fic. 21.—Transverse section of a genital segment of a female. Cam. D,
oc. 2.

Fie, 22.—Small part of a similar section, showing the position and struc-
ture of the nephridium. Cam. Oil im. 3, oe. 3.

THE ETIOLOGY OF MALARIAL DISEASES. 429

On the Question of Priority with Regard to
certain Discoveries upon the Attiology of
Malarial Diseases.

By

George H. F. Nuttall, M.A., M.D., Ph.D.,
University Lecturer in Bacteriology and Preventive Medicine, Cambridge.

T'HouaH it has long been a popular belief in certain countries
that malaria is communicated to man by means of mosquitoes,
experimental proof was lacking until a recent date. The
history of the mosquito-malaria theory has been amply dis-
cussed elsewhere by the writer, to whose papers the reader
is also referred for a detailed description of the experimental
work on the part played by mosquitoes in the propagation of
malarial diseases.!. It is not the object of this paper to
discuss these matters in detail.

Persons who read the medical literature of but one country
will naturally become biassed in their judgment. This ac-
counts for the fact that at present different investigators
receive the credit of having definitely established the part
played by mosquitoes in malarial diseases. In view of the
confusion which will naturally result from the claims made

1 Nuttall, G. H. F. (1899-1900). I. “ Onthe Role of Insects, Arachnids,
and Myriapods as Carriers in the Spread of Bacterial and Parasitic Diseases
of Man and Animals: a critical and historical Study ;” ‘Johns Hopkins
Hospital Reports,’ vol. viii, pp. 1—154, 3 plates (Bibliography). II. *‘ Die
Mosquito-Malaria-Theorie,” ‘Centralbl. f. Bakteriologie,’ vol. xxv, pp. 162—
170, 209—216, 245—247, 285—296, 337-346 (Bibliography). III.
“Neuere Forschungen iiber die Rolle der Mosquitos bei der Verbreitung der
Malaria : Zusammenfassendes Referat ;” ‘ Centralbl. f. Bakteriologie,’ vol. xxv,
pp. 140—147, and vol. xxvii, pp. 193—196, 218—225, 260—264, 328 —-340
(exhaustive Bibliography).

430 GEORGE H. F. NUTTALL.

in various quarters, it seems eminently desirable to give a
brief impartial summary of the experimental work which has
been done, relying solely upon published researches, these
being cited in their chronological order. With the facts
thus marshalled before him every reader is at liberty to draw
his own conclusions.

The study of the hemocytozoa begins with the discovery
by Ray Lankester in 1871 of Drepanidium ranarum.
Human malarial parasites were seen, but their significance
not comprehended until Laveran published his investigations
in November, 1880. Following upon the fundamental work
of Laveran, the most important discovery was that of Golgi
(November, 1885), who demonstrated the relationship exist-
ing between the life-cycle of the parasites within the human
body and the occurrence of the febrile attack. With regard
to these investigations there has never been any dispute on
the question of priority, but this is far from being the case
with the discoveries which followed. Any further disputes
regarding the priority of subsequent discoveries should be
disposed of by such a chronological record as that which
follows, in which not only the year, but also the month and
even day of publication are given.

Chronology relating to certain of the more Im-
portant Recent Researches on Malaria.

1893 and 1895, Sacharoff demonstrated the presence of
chromatic substance within the “flagella” of
certain avian parasites by means of the Romanowsky
stain.

December 17th, 1895, Ross observed the process of
“flagellation” of crescentic parasites to occur in the
stomach of mosquitoes (species not determined) fed
on the blood of a malarial patient.

1896, Bignami and Dionisi report the negative results of two experi-
ments made in 1893-4 with mosquitoes (species uncertain) collected in

malarious localities, the insects being permitted to bite healthy per-
sons. ‘They attribute the failure of the experiment to the dispersion of

THE ATIOLOGY OF MALARIAL DISEASES. 431

the insects in the room where they were liberated, and to the experiment
not having been continued long enough. They cite Calandruccio as
having observed the degeneration of malarial parasites in the stomach
of mosquitoes (species not stated).

November 13th, 1897, MacCallum, in Baltimore, found
that the “flagella” of Halteridium and of
estivo-autumnal parasites constitute the male
element, and serve to impregnate the “pig-
mented spheres” or female element. In the case
of Halteridium the impregnated spheres became con-
verted into motile “vermicules.” This transforma-
tion was, however, not observed in the human parasites.

December 18th, 1897, Ross fed mosquitoes upon human
blood containing crescentic parasites. The ex-
periments were made at Secunderabad, and were
reported upon at the time as follows :

After examining hundreds of mosquitoes fed on
malarial blood, always with negative results, he obtained a
few which belonged to a species with spotted wings, which
he had hitherto not used. As Ross distinctly describes
the egg of this species, there is no doubt whatever
but that he was dealing with a species of Ano-
pheles. The insects were bred from larve, and fed
with blood containing crescentic parasites. Four to five
days later peculiar pigmented cells were observed lying
within the walls of their stomachs. These cells were
round or oval ; they measured 12—16 , on the fourth, and
20 » on the fifth day after feeding, and the pigment they
contained was similar to that within the malarial para-
sites in the blood upon which the insects had been fed.
Such bodies could not be found in control mosquitoes.
Ross concluded that he had found the m osquito
which served asa host for the parasite.

February 26th, 1898, Ross refers again to his experiments
with crescentic parasites. After examining some
scores of ‘ dapple-winged”’ mosquitoes unfed or fed
with healthy blood, all the results were negative until

432

GEORGE H. F. NUTTALL.

“at last two of this species were persuaded to feed on a
patient with crescents. One of them was killed next
day; no pigmented cells could be found. The second
was killed forty-eight hours after feeding ; numerous
pigmented cells were present. They were all small,
much smaller than epithelial cells, ovoid, about
7 win the major axis, and each contained about
twenty granules of typical pigment, which were
often arranged circumferentially, just as in the malarial
parasite.” Though it is not stated in this publication
that he raised these mosquitoes from larvae, reference to
Ross’s previous paper (p. 1786) will show this to have
been a part of the method he employed.

Experiments with Tertian Parasites.—A hundred or more grey or

‘barred-back’ mosquitoes, unfed or fed on healthy or crescent blood,
have been dissected without finding the pigment cells. At last one was
observed feeding on a patient whose blood that morning had been seen
to contain numerous mild tertian.parasites.” Killed on the third
day, the insect contained many pigmented cells measuring
8—25 p. (Ross subsequently discarded this experiment, as it was
possible that the insect which was not raised from the larva had become
infected with some other parasite.)

May 21st, 1898, Experiments on Proteosoma.—Work-

ing in Calcutta, Ross observed the development of
Proteosoma in a species of Culex (subsequently
determined as C. fatigans, Wied.), the insects being
fed on the blood of infected crows, larks, and sparrows.
The parasites found in the external coat of the insects’
stomachs measured 6 yw after thirty hours, 60 uw after six
days. ‘‘Successive feeds by the same mosquito on the
same bird are followed by fresh crops of young coccidia.

Similar pigmented cells”? had been previously
observed in mosquitoes fed on human parasites. Ninety-
four per cent. of the mosquitoes fed on blood containing
mature Proteosoma became infected.

September 24th, 1898.—Manson reported to the British

THE ATIOLOGY OF MALARIAL DISEASES, 433

Medical Association Meeting at Edinburgh (July) on
behalf of Ross regarding further experiments with
Proteosoma. ‘These observations showed that the
encapsulated parasites, on reaching a certain size, rup-
tured and emptied their contents into the ccelom of the
insect. The contents of the ruptured capsules consisted
of minute spindle-shaped bodies, and these bodies sub-
sequently accumulated in the salivary gland of the
insect. When this had occurred the insects were
capable of communicating the proteosomal infection to
healthy birds. Of twenty-four sparrows exposed to the
bites of insects fed on mature parasites, twenty-two
became infected.

October Ist, 1898, Grassi reported that he had reason for
suspecting three species of Culicide as being carriers of
malarial infection, claiming that they were confined in
their geographical distribution to those regions where
malaria was prevalent in Italy. The three species were
Culex penicillaris, Anopheles claviger (syn.
A. maculipennis), and a purported new species,
Culex malariz.! It has since been proved that only

' In his paper in the ‘ Policlinico’ (October 1st, 1898), Grassi writes: ‘In
conclusione, io sono d’ avviso che il Culex penicillaris e l’ Anopheles
claviger o per lo meno il Culex penicillaris, fors’ anche il Culex
malariz, nella malaria si comportano come le zecca nella febbre del Texas.”

Grassi therefore makes a misstatement in a later paper (December Ist, 1900)

when he writes, “ Proclamai come indiziati due specie di culex, ma sopratutto

? Anopheles claviger.” It is curious that Grassi should subsequently have

continued to lay stress upon the geographical coincidence having led him to

the discovery of Anopheles claviger being a host of malarial parasites,
for two out of three species which he for this reason supposed must be hosts
were afterwards proved not to be such, He certainly considered A. claviger
at first to be of quite secondary importance; we have his own words for it:

“‘Certi casi di malaria sviluppatisi in Settembre a Locate Triulzi, nei quali

gli Anopheles di certo o non punsero 0 soltanto rarissime volte, denunciano

decisamente come trasmissore il Culex penicillaris, enorma-
mente comune in tutti iluoghi malarici.” (The italics are Grassi’s.)

It is bnt fair to Ross to state here that Grassi in his paper of the 1st of

October refers to the experiments made by Smith and Kilborne upon Texas

fever, and by Ross upon avian malaria as having been a “ forte argumento ”

434 GEORGE H. F. NUTTALL.

the second of the three species named can serve as a
host for human malarial parasites. The coincidence in
the geographical distribution of ague and malaria-bear-
ing mosquitoes in Italy, as claimed repeatedly by Grassi,
has been disproved by Celli. The claim that this geo-
graphical agreement would probably be found to hold
in other parts of the world has been disproved by Nut-
tall, Cobbett, and Strangeways-Pigg (1901) in England.
We cannot, therefore, accept Grassi’s statement that he
discovered the malarial mosquito because of its geo-
graphical distribution, pretty and ingenious as_ the
hypothesis seemed in the beginning. It seems certain
that Grassi was after all entirely guided by Ross’s pub-
lication of December 18th, 1897, in which he describes
an insect with spotted wings and eggs like those which
characterise Anopheles.

November 6th, 1898, Infection Experiment on Man.—
Grassi mentions that Bignami had made an infection
experiment by means of mosquitoes (the three species
above named were employed) collected at Maccarese,
a malarious locality. The result was positive in this
case, the person acquiring estivo-autumnal fever.
(Several infection experiments were subsequently car-
ried out by Bignami, Bastianelli, and Grassi in colla-
boration, these being reported in various papers of
later date. The first experiment did not prove which
species harboured the parasites, and of itself was insuf-
ficient to establish the theory on a firm basis.)

December 4th, 1898, Bastianelli, Bignami, and Grassi
observed the development of crescentic parasites
in Anopheles claviger, the appearances correspond-
ing to those described by Ross for Proteosoma on the

in favour of the mosquito-malaria hypothesis. In the paper read on the next
day at the Accademia dei Lincei, under the same title as that which appeared
in the ‘ Policlinico,’ Grassi omits to mention Ross, though he refers to what
was known regarding Texas fever. The paper, published in the ‘ Transac-
tions’ of the Accademia, differs in several respects from that which appeared
in the ‘ Policlinico.’

THE AMTIOLOGY OF MALARIAL DISEASES. 435

fourth day in Culex. Referring to his experiments with
human parasites, they write, ‘ Verisimilmente i due
mosquitos coli ali macchiate nei quali il Ross in India
trovo stadi di sviluppo simili a quelli del proteosoma (3°
giorno circa) appartenevano pure alla specie Anopheles
claviger, Fabr.” (This statement is of interest in view
of Grassi’s subsequent claim that Ross might very well
have been working with: insects belonging to the genus
Culex, and not with Anopheles at all.) They, more-
over, consider that Ross had not certainly determined
the development of the crescents in his mosquitoes, for
his observations had been broken off at too early a date ;
besides which the insects might have infected themselves
with hematozoa from some other animal. We have seen
that the latter supposition is unwarranted, because Ross’s
Anopheles were raised from larve. Moreover they
themselves neglect to state that they raised their Ano-
pheles from larve, so we must presume that they did
not.

Infection Experiment on Man.—In a foot-note to the
above publication it is reported that the authors had
successfully infected a person with tertian fever by
means of infected A. claviger, collected at Maccarese.

December 22nd, 1898, Grassi, Bignami, and Bastianelli
follow the development of crescentic parasites
in Anopheles claviger to the formation of ‘ sporo-
zoites,’’ the escape of the latter into the ccelom of the
insect, and their accumulation in the salivary gland.
The development was found to be slower at 20° to 22°
than at 80° C. The fully developed capsules measured
70 , the sporozoites measured 144. The process of
development, the size of the fully developed capsules,
and of the sporozoites, were the same as Ross had
observed in Proteosoma.

The development of tertian parasites was observed to

_ take place in A. claviger up to the fifth day.

February 2nd, 1899, Koch published a preliminary note

436 GEORGE H. F. NUTTALL.

upon the results of the investigations conducted by the
German Malaria Commission, consisting of himself, R.
Pfeiffer, and H. Kossel. Further details will be found
in a publication which appeared September 8th, 1899.
The Commission observed the development of Pro-
teosoma in Culex nemorosus, from the formation of
the ‘ vermiculi” described by MacCallum for Halteri-
dium to their appearance*in the salivary gland of the
insect. The process of fertilisation was found to occur
in Proteosoma, as MacCallum had found for Halteri-
dium and human crescentic parasites. Healthy birds
were successfully infected by means of infected insects.
The later publication, which is illustrated by excellent
microphotographs, completely confirms the observations
of Ross and others.

February 5th, 1899, Grassi, Bignami, and Bastianelli
observe the development of quartan parasites in
A. claviger. Ross (September 2nd, 1899) observed
the development of quartan parasites in a species of
Anopheles in Sierra Leone.

January 25rd, 1899, Daniels reported to the Royal Society
that he had been able to confirm Ross’s observations
with Proteosoma. He followed their development in
a species of Culex, and successfully infected healthy
birds by means of infected insects. He added nothing
to what Ross had already found.

April 19th, 1899, Bastianelli and Bignami reported further
studies upon the development of tertian parasites in
Anopheles claviger, and describe three successful
infection experiments on man by means of A. claviger
previously fed on tertian parasites.

May 7th, 1899, Grassi, Bignami, and Bastianelli report
to the Accademia dei Lincei that they had observed the
development of tertian and crescentic parasites in
Anopheles bifurcatus.

June 18th, 1899, Grassi observed the development of
tertian and crescentic parasites in Anopheles

THE MIIOLOGY OF MALARIAL DISEASES. 437

pseudopictus, but not in various species of
Culex. The latter result again obtained later (October
4th, 1899).

June 28th, 1899, Ross stated that Proteosoma scarcely
developed in Culex at 21°, and that the growth of the
parasites was already slowed at 27° C. in Calcutta. The
development of tertian parasites in spotted-winged
mosquitoes raised from larve was also observed (letter
dated February 22nd, 1899, to Nuttall; see ‘ Centralbl.
f. Bakteriologie,’ vol. xxv, p. 908).

September, 1899, Bastianelli and Bignami give a de-
tailed description of tertian and crescentic parasites, the
publication being accompanied by the best coloured
plates hitherto published, illustrating their development.
They prove that a single infected Anopheles claviger
may communicate malaria (tertian) to man.

May 4th, 1900, Ziemann, working in Cameroon, observes the
development of the parasites of tropical malaria in
two species of Anopheles, as also the development of
tertian parasites in one species of Anopheles. He
followed the development to the appearance of sporozo-
ites in the salivary glands of the insects. He subse-
quently (November 22nd, 1900) found that the parasites
would not develop in Cimex lectularius nor in sand-
flies.

September, 1900, van der Scheer and van Berlekom, in
Holland, observe the development of tertian parasites
in A. claviger.

September 29th, 1900, Manson reported a positive infection
experiment with tertian-infected Anopheles (spec. ?)
imported from Rome, the insects being permitted to
bite his son in London.

October 6th, 1900, Rees reports a similar experiment to
the former.

After perusing the above chronology, and remembering
the question most disputed—the discovery of the develop-

438 GEORGE H. F. NUTYALL.

ment of human parasites in Anopheles, we must conclude
that the pigmented encapsulated bodies observed by Ross in
“ spotted-winged mosquitoes” at Secunderabad were cres-
centic parasites in early stages of development. In his first
paper Ross definitely states that he raised the imagos from
larve kept in bottles ; that the parasites which subsequently
developed within them contained a pigment similar to that
of the parasites in man; and his description of the insects’
eggs leaves no room for doubt but that they were Ano-
pheles. (In their paper of December 4th, 1898, Bastianelli,
Bignami, and Grassi even made the statement that it is
extremely likely that Ross’s spotted-winged mosquito was
A. claviger!) The work done subsequently on Proteo-
soma quite rightly confirmed Ross in his belief. We are,
however, indebted to the Italian investigators named for
completing the study of the further development of human
parasites in various species of Anopheles, these studies
being subsequently pursued by still other investigators in
other countries.! Ross is perfectly justified in laying stress
upon the fundamental importance of his discoveries in the
development of Proteosoma, and there’ can be no doubt
whatever about his work having served as a guide to sub-
sequent investigators. here is no denying that both the
human and avian parasites referred to offer great points of
similarity throughout. The assumption was, therefore, per-
fectly justified that the further stages in the development of
crescentic parasites such as Ross had observed at Secunder-
abad would be identical with what he saw in the case of
Proteosoma afterwards in Calcutta. ;

In conclusion we must not forget to mention the name of
Patrick Manson, who until recently took no part in the
experimental solution of the problem, but who throughout
Ross’s investigations, which he stimulated, did much to
further the studies which in one direction at least have
reached such a satisfactory conclusion.

' Tt has not heen deemed necessary to refer to all of these.

THE ATIOLOGY OF MALARIAL DISEASES. 439

PUBLICATIONS CITED.

DASTIANELLI, G., and Bignami, A. (dated April 19th, 1899).—*‘Sullo
sviluppo dei parassiti della terzana nell’ Anopheles claviger,”
‘Bullettino della R. Accad. Med. di Roma,’ anno xxv, 1898-9,
fasc. 3 (reprint 28 pp.).

BastianzeLul, G., and Brenami, A. (published end of September, 1899).—
I. “Sullo sviluppo dei parassiti della terzana nell’ Anopheles
claviger.” IJ. “Sulla struttura dei parassiti malarici e in specie
dei gameti dei parassiti estivo-autunnali,” ‘Annali d’ Igiene speri-
mentale,’ N.S., vol. ix, fase. 3, pp. 272—293 (1 col. plate), pp. 245
—257 (2 col. plates).

BasTIANELLI, G., Biegnami, A., and Grassi, B. (meeting of December 4th,
1898).—‘‘ Coltivazione delle semilune malariche dell’ uomo nell’
Anopheles claviger, Fabr. (Sinonimo, Anopheles maculi-
pennis, Meig.); Nota preliminare,” ‘R. Accad. dei Lincei,’ vol.
vii, 2 sem., ser. 5a, fasc. 11 (reprint 1 page).

Bienami, A. (1896).—‘‘ La ipotesi dei parassiti malarici fuori dell’ uomo,”
‘Policlinico,’ No. 14. Also transl., ‘ Hypothesis as to the Life-
History of the Malarial Parasite outside the Human Body (apropos
of an article by Dr. Patrick Manson),” ‘Lancet,’ vol. ii, pp. 1363—
1367, 1441—1444.

Ceuui, A. (November 5tli, 1900).—* Beitrag zur Erkenntniss der Malaria-
eptdemiologie vom neuesten aetiologischen Standpunkte aus,”
‘ Centralbl. f. Bakteriol.,’ vol. xxviii, pp. 530—535.

Danipts, C. W. (read March 16th, 1899).—*‘‘On Transmission of Proteo-
soma to Birds by the Mosquito: a Report to the Malarial Committee
of the Royal Society,” ‘ Proc. of the Royal Soc.,’ vol. Ixiv, pp. 443—
454,

Gore (1886).-—“‘ Sull’ infezione malarica,” ‘Arch. per le Scienz. med.,’ vol.
x, pp. 109—185. See also ‘ Arch. Ital. de Biol.,” vol. viii (1887).

Grassi, B. (published October Ist and 2nd, 1898).—‘* Rapporti tra la
malaria e peculiari insetti (Zanzaroni e Zanzare palustri); Note
preliminare,” ‘ Policlinico,’ vol. V—M (reprint 10 pp.). Also
(under the same title though altered) ‘ Rendiconti della R. Accad. dei
Lincei ’ (meeting of, October 2nd), vol. vii, 2 sem., pp. 163—172.

Grassi, B. (June 22nd, 1899).—‘* Ancora sulla malaria,’ ‘R. Accad. dei
Lincei,’ ser, 5a, vol. vili, 1 sem., fase. 12 (reprint 3 pp.).

Grassi, B. (October 4th, 1899).—‘‘ Osservazioni sul rapporto della seconda
spedizione malarica in Italia presieduta dal Prof. Koch, etc.,” ‘ Rendi-
conti della ik. Accad. dei Lincei,’ Cl. di sc. fis., mat., e nat.

440 ‘GEORGE H. F. NUTTALL.

Grassi, B. (December Ist, 1900).—“ Per la storia delle recenti scoperte
sulla malaria,” ‘Il Policlinico,’ anno vii, No. 23, pp. 593—600.
Grassi, B., Bienami, A., and BAsTIANELLI, G. (dated December 22nd, 1898).
—‘‘ Ulteriori ricerche sul ciclo dei parassiti malarici umani nel corpo

del zanzarone,” ‘ Rendiconti della R. Accad. dei Lincei.’

Grassi, B., Brenami, A., and BasTIANELLI, G. (meeting of February 5tb,
1899).—* Resoconto degli studi fatta sulla malaria durante il mese di
gennaio,”’ ‘ Rendiconti della R. Accad. dei Lincei,’ vol. viii, 1 sem.,
ser. 5a, fasc. 3 (reprint 4 pp.).

Grassi, B., Brena, A., and BAsTIANELLI, G. (meeting of May 7th, 1899).—
**Ulteriori ricerche sulla malaria,” ‘Rendiconti della R. Accad. dei
Lincei,’ Classe di sc. fis., mat., e nat., vol. viii, 1 sem., ser. 5a, fase.
9 (reprint 5 pp.).

Kocn, R. (February 2nd, 1899).—‘* Ergebnisse der wissenschaftlichen Expe-
dition des Geheimen Medicinalraths Professor Dr. Koch nach Italien zur
Erforschung der Malaria,” ‘ Deutsche med. Wochenschr.,’ Jahrg. 25,
pp. 69, 70. (Preliminary Report ; see following.)

Kocu, R. (September 8th, 1899).—‘* Ueber die Entwickelung der Malaria-
parasiten,”’ ‘ Zeitschr. f. Hygiene u. Infektionskr.,’ vol. xxxil, pp. 1—24
(4 Tafeln).

LankKESTER, Ray (1871).—‘‘ Observations and Experiments on the Red
Blood-corpuscle,” ‘ Quart. Journ. Mier. Sci.,’ vol. x1, p. 389.

LanxkesteR, Ray (1882).—“ On Drepanidium ranarum,” ‘ Quart. Journ.
Mier. Sci.,’ vol. xxi, p. 53.

Laveray, A. (November 23rd, 1800).—‘* Note sur un nouveau parasite, etc.,”
‘Bull. de Acad. de Méd. de Paris.’

Laveran, A. (December 24th, 1§80).—‘ Un nouveau parasite trouve dans le
sang des malades atteints de fievre palustre ;” ‘‘ Origine parasitaire
des accidents de |’impaludisme,”’ ‘ Bull. et Mém. de la Soc. Méd. des
Ho6pit. de Paris.’

MacCauium, W. G. (November, 1897).—** On the Hematozoan Infections of
Birds,” ‘Johns Hopkins Hospital Bulletin,’ No. 80; also ‘ Lancet,’
1897 (Preliminary Note).

MacCatium, W. G. (January, 1898).—‘‘On the Hematozoan Infections of
Birds,” ‘Journ. of Experimental Medicine,’ vol. iii, No. 1, pp. 117—
136 (one coloured plate).

Manson, P. (September 24th, 1898).—‘ The Mosquito and the Malaria Para-
site,” ‘ Brit. Med. Journ.,’ vol. ii, pp. 849—853, illustrated by figures
obtained from Dr. Ronald Ross. (Reported at the meeting of the
Brit. Med. Assoc., Edinburgh, in July.)

Manson, P. (September 29th, 1900).—“‘ Experimental Proof of the Mosquito-

THE ATIOLOGY OF MALARIAL DISEASES. 441

Malaria Theory,” ‘Brit. Med. Journ.,’ vol. ii, pp. 949—951 (two
figures and one chart); also ‘ Lancet,’ vol. ii, pp. 923—925,

Nutrta.t, G. H. F., Copsett, L., and Srranceways Pice (January, 1901).—
“Studies in Relation to Malaria. I. The Geographical Distribution
of Anopheles in Relation to the Former Distribution of Ague in
England ” (two maps), ‘Journ. of Hygiene,’ vol. i, pp. 4—44.

Rees, D. C. (October 6th, 1900).—“ Experimental Proof of the Malaria-
Mosquito Theory,” ‘ Brit. Med. Journ.,’ vol. ii, p. 1054.

Ross, R. (December 17th, 1895).—* Proe. of the South Indian Branch of the
Brit. Med. Assoc. ;’ see also abstracts of Ross’s letters in Manson’s
article, ‘ Lancet,’ 1896.

Ross, R. (Report to the Director-General of the Indian Med. Service, dated
September 19th, 1897, published 18th December, 1897, in England).—
“On some Peculiar Pigmented Cells found in Two Mosquitoes fed on
Malarial Blood” (with a note by Surgeon-Major Smyth), ‘ Brit. Med.
Journ.,’ vol. il, pp. 1786—1788.

Ross, R. (February 26th, 1898).—‘ Pigmented Cells in Mosquitoes,” ‘ Brit.
Med. Journ.,’ vol. i, pp. 550, 551.

Ross, R. (May 21st, 1898).—‘ Report on the Cultivation of Proteosoma,
Labbé, in Grey Mosquitoes’ (Office of the Superintendent of Govern-
ment Printing, Calcutta, India), 4to, 21 pages, 9 plates.

Ross, R. (September 2nd, 1899).—*‘ The Cultivation of the Malarial Quartan
Parasite in Anopheles,” ‘ Brit. Med. Journ.,’ vol. ii, p. 608.

Sacuarorr, N. (December, 1893).—“ Recherches sur les hématozoaires des
oiseaux,’ ‘ Annales de |’Inst. Pasteur,’ vol. vii, pp. 801—811 (une
planche colorée).

Sacuarorr, N. (1895).—‘‘ Ueber die selbstandiche Bewegung der Chromo-
somen bei Malariaparasiten,” ‘Centralbl. f. Bakteriologie, vol. xvii,
pp. 874—380 (2 ‘Tafeln).

Van DER ScuHEER, A., and BERDENIS VAN BeERLEKoM, J. J. (reprint re-
ceived October, 1900).—‘* Malaria en Muskieten in Zeeland,” ‘ Ned.
Tijdschrift voor Geneeskunde,’ vol. ii, No. 14 (reprint 16 pp., 1 plate).

ZiEMANN, H. (June 21st, 1900).—‘* Ueber die Beziehungen der Mosquitos
zu den Malariaparasiten in Kamerun,” ‘ Deutsche med. Wochenselir.,’
Jahrg. xxvi, No. 25, p. 399.

ZirMaNN, H. (November 22nd, 1900).—* Zweiter Bericht iiber Malaria und
Moskitos an der afrikanischen Westkiste,”’ ‘ Deutsche med. Wochen-
schr.,’ Jahrg. xxvi, Nos. 47 and 48, pp. 753—756, 769—772.

9

voL, 44, PART 3.—NEW SERIES, EF

asst

“ei a

Seite A, Leal itt h
ee tt | ee
re od cay * a> AS

é
.
ay =bt ‘
.
>
»
1
i mijepe A
'
, ‘
‘ 3
«
.
: ‘
%
“
:
°

aan W
Vike, sano bg
oe by 4

en a

HOM Sse 5

5 ey ee, “Page
Gee Wet sod ieee» § dene

Pit :
ie IE
b, i~- 7

{'
“7 4

STUDIES IN THE RETINA. 443

Studies in the Retina: Rods and Cones in the
Frog and in some other Amphibia.

By

H. M. Bernard, M.A.Cantab.

(From the Biological Laboratories, Royal College of Science, London.)

With Plates 30 and 31.

Part II.}

Tue first part of this paper was devoted to showing that
the structures called “ cones”? in the amphibian retina were
the earlier stages in the development of the new rods required
by growth, and that they force their way in wherever there
is room for them between already existing rods. The forms
of these elements with the positions of their nuclei? were alone
dealt with. In this paper it is proposed to give some account
of the intimate structure of the amphibian rod. The minute
details to be described will necessitate some discussion of the
physiological processes which, so far as I have been able to
interpret them, underlie their forms.

Little success has so far attended the attempts of naturalists
to unravel the finer structure of the rods. Indeed, the sub-
ject seems to have been temporarily closed by the classical
researches of Max Schultze in the sixties, for since that
time little or no advance has been made. The earlier

1 The final revision of this MS. was kindly undertaken by my friend Mr.
Martin Woodward, during my temporary absence from England.

2 On page 44 I inadvertently attributed to Borysiekiewitz an observation
of my own, This will be fully dealt with in Part IIT.

444, H. M. BERNARD.

literature is, however, full of scattered observations, and it is
possible that some of them may have been overlooked by me
or far too briefly noticed. I do not pretend to have mastered
the whole of the literature on the retina. I do not wish,
therefore, to make any claims to priority, but simply to
describe my observations, referring briefly to those of former
students, so far as I know of any covering the same ground.

And here I should add that, while confining myself in this
paper solely to the Amphibia,! these researches have extended
over other Vertebrates, and that the general conclusions
arrived at are not drawn solely from the facts here described.

Enveloping Membranes.—While the existence of the
membrane investing the inner limb of the rod requires no
demonstration, it has been much disputed whether the outer
limbs possess any such envelope or not. Apart from the fact
that such a covering is difficult to demonstrate, it is possible
that the conception of the rod as a cuticular structure may
have strengthened the doubt. It has long been known that
the outer limbs of the rods can be made to divide up trans-
versely into discs, and that on such a dissolution no investing
membrane can be seen (cf. Max Schultze’s figures, ‘Arch.
mikr. Anat.,’ Bd. iii, pl. xiii, figs. 11, etc.).

Merkel? found membranes wherever he looked for them
except in the Amphibia, while Landolt*® figured very thick
homogeneous membranes covering the rods (frog and newt)
and the cones (newt). Iam not aware that these have ever
been confirmed, and I doubt their existence. He figures
them even passing in between outer limb and ellipsoid.
Something like what he figures may be seen in my fig.
13, c—i, the significance of which will be discussed later on.
In the meantime I may state that I do not regard the thick
rind there shown as an outer covering.

1 The following forms have been examined:—Rana temporaria, Bufo
vulgaris, Molge cristata and M. vulgaris, Salamandra maculosa,
and Siredon pisciformis.

2 © Arch. Anat. u. Phys.,’ 1870, pl. xiv.

2 «Arch. mikr. Anat.,’ Bd. vii, 1871, p. 81, pl. ix,

STUDIES IN THE RETINA. 445

That enveloping membranes occur in the outer limbs of
Amphibian rods is certain, both on theoretical grounds and
because they can be demonstrated.

As we saw in Part I of this paper (this Journal, p. 29), the
rods are primarily protoplasmic vesicles protruded from the
retina. ‘The walls of the vesicles are of extraordinary delicacy
and transparency, and it will be a triumph of microscopic
technique when retinas can be fixed so as to show them
intact. They are best seen in their very earliest stages of
protrusion, before any rods are formed and the pigment is
only just being forced away from the retina by their increase
in number and size. From this early stage we traced them
through their principal form-phases till they became normal
rods, and all these phases were not only consistent with their
being long, membranous sacs, but even confirmatory of this
conception of their essential structure. Lastly, the persist-
ence of the membrane covering the inner limbs has, as we
have seen, long been an established fact.

But granted that in all the earlier stages of the rods we
have a wall to the vesicle—a wall which persists in the inner
limb,—we have still to ask whether that is the case with the
outer limb when the rod is complete. May not the proto-
plasmic wall merge in the substance which fills up the
interior of the outer limb and lose its individuality, so that it
would be impossible to speak of any investing membrane?
This is, of course, quite possible ; and, moreover, it is certain
that even if it preserved its individuality one would rarely
expect to demonstrate the existence of such a delicate film of
transparent protoplasm round the outer limb of the amphibian
rod, with its usually refractive contents. Actual observa-
tions, however, show clearly that the protoplasmic wall does
retain its individuality, and that to the last the rod is a thin
protoplasmic vesicle filled up with matter, the origin and
nature of which will be discussed in the following pages.

As demonstration of this persistence of the protoplasmic
wall of the vesicle, I will call attention to Pl. 30, fig. 2, which
is taken from the retina of a toad. All the rods in this

446 H. M. BERNARD.

retina are obviously bags which have, under pressures and
strains, lost their normal cylindrical shapes, and are now
pulled out or crushed together into every variety of form,
from short, rounded sacs to long, thin clubs with round knobs
at the tips.! Hndless, too, are the instances in which the
inner and outer limbs have been pulled somewhat apart, and
the stretched or torn membrane becomes visible under good
microscopic powers. Important, also, in this connection is
Pl. 30, fig. 1, from a newt, in which in one spot all the rods
were broken away, but their basal portions persisted emptied
of all contents except the remains of the ellipsoids. This
last is a fortunate observation, because it shows that, in
essence, the inner and outer limbs are simply two sacs
separated by a thin wall, and that the great differences seen
between them must be referred entirely to their contents.
To this we shall return later.

Lastly, I have sections of a newt’s retina in which the thin
coverings of the rods have taken stain, and are quite demon-
strable in optical sections.

External Markings.—The longitudinal striation of the
outer limbs of the rods has long been seen, but its nature
has never been satisfactorily settled. Max Schultze regarded
it as a furrowing of the surface, and figured the cross-sections
of the rod as having an outline lke that shown in PI. 30,
fig. 8.

With regard to the inner limbs of the rods in Amphibia,
exact records of striation are few. The well-known “ Faser-
korb” of Max Schultze was found by him, “ essentially the
same,” in all the classes of Vertebrates, including the Am-
phibia. He records finding it in the axolotl, and he figures
it in the newt.” Hoffmann also figures the upper ends of
these or similar threads round the bases of the inner limbs of
amphibian rods, and forming a ring of needle-like points
similar to those figured by Max Schultze as projecting from

1 Max Schultze gives a somewhat similar figure, viz. the sac-like rod of a
pike, produced artificially (‘ Arch. mikr. Anat.,’ Bd. iii, pl. xiii, fig. 18 a).
2 ¢Arch, mikr. Anat.,’ Bd. v, p. 379, pl. xxii, fig. 2a.

STUDIES IN THE RETINA. 447

the membrana lim. externa when the rods and cones are
broken away (cf. Hoffmann’s figures [Bronn’s ‘ Thierreich ;
Amphibien,’ pl. xxii, 11—18, and pl. xxiv, 2—8] with
Max Schultze’s [‘ Arch. mikr. Anat.,’ Bd. v, pl. xxii]). The
latter author traced the threads of his ‘‘ Faserkorb ” proxi-
mally into the connective tissue of the outer nuclear layer,
but inasmuch as distally they ran on to the outer limbs of
the rods, he clearly wished to see in them the ends of the
nerves (cf. Stricker’s ‘Handbuch’). Hoffmann, who figures
the basal threads as running only a short way down the inner
limbs of the Amphibia, and then only loosely applied, appa-
rently regarded them as nothing more than hair-like prolon-
gations of the membrana lim. externa. Further, as stated,
Max Schultze described a continuation of his “ Faserkorb ” a
short way down the outer limbs. Hoffmann (loc. cit.) also
figures somewhat similar threads running on to the outer
limbs of amphibian rods; these he could not explain because
his basal threads were not supposed to run the whole length
of the inner limbs.

There seems, then, to have been a distinct tendency to
attribute to the inner limbs in the Amphibia a system of
longitudinal fibres, though apparently not so pronounced or
complete as the “ Faserkorb” of the inner limbs of the
human rods and cones.

We may say, then, that the rods are thought to be longi-
tudinally striated, but while the inner limbs are externally
striated with fibrils the outer limbs are marked by furrows.
My own observations entirely confirm the existence of longi-
tudinal striz ; but those on the inner limb and those on the
outer limb are not distinct in kind from one another, but are
parts of one system.

Long before I had succeeded in discovering the true rela-
tions of these striations to one another, I had noticed that
the markings on the outer limbs consisted far more of longi-
tudinal rows of dots than of furrows. The rows, though
mostly continuous, are not always strictly parallel ; and the
dots only occasionally fall into circular series running nearly

448 H. M. BERNARD.

evenly round the rod. I find in my notes that at times two
series of dots at right angles to one another are recorded as
marking the exterior of the rods. The dots were usually
shightly drawn out longitudinally. Fig. 3, a, b,! are from
my earlier drawings. It was noticed that the dots appeared
almost as if they raised the surface of the rod, and that,
hence, between the rows there were slight furrows, but on
this point I have never satisfied myself; if any furrowing
exists, it must be very slight. While these longitudinal rows
of dots on the outer limbs were clear with any well-preserved
retinas stained in Ehrlich’s hematoxylin, it was not till I
employed the iron-alum hematoxylin method of staining
that I saw any striation of the inner limbs, and then, while
that on the outer limbs was very strong and regular, that on
the inner limbs was hardly ever regular, often indeed not
recognisable as a system of striz at all. Further, I then
found, as stated, that the two are not distinct phenomena,
but that the fine staining threads which run down in the
walls of the inner limbs are continued on to the outer limbs,
as Max Schultze observed: but they do not stop short, as he
supposed ; on the contrary, they run down the whole way,
swelling into small clumps of staining matter at short dis-
tances from one another, these clumps being the rows of dots
I had seen all along.

Fig. 11 shows diagrammatically the arrangements of this
system of threads, while figs. 13, a—d, and 29, a, b, e—j, are
from actual preparations of retinas from different Amphibia.
Beginning usually faint near the nucleus, and seldom as a dis-
tinct system, the arrangement gets more pronounced distally.
It may be very pronounced indeed near the ellipsoid (e. g. in
the toad, fig. 13, a, b). Here it passes on to the outer limbs,
and, where inner and outer limbs are stretched a little apart,
may be seen as a nearly regular ring of smooth, thin threads,

1 Cf. Max Schultze, ‘Arch. mikr. Anat.,’ Bd. in, pl. xiii, fig. 11, where
he shows a rod covered with ‘‘ pigment granules ;” another figure occurs in
Bd. v, pl. xxii, fig. 17 a. The dots above referred to are quite distinct from
pigment granules, one of which I have drawn in fig. 3, d.

STUDIES IN THE RETINA. 449

not free, hke Max Schultze’s needle-like prolongations of his
“Faserkorb,” but rather as thickenings of the stretched
membrane. On these threads clumps of staining matter soon
appear (fig. 13, c). In the diagram, fig. 11, the system is
drawn very symmetrically from the nucleus outward, but this
is not by any means usually the case. The nearest approach
to it has been found in the axolotl, preparations of which
inspired this diagram. Fig. 29, b, represents more truly the
ordinary conditions. We have a gradual formation of the
symmetrical system of striz towards the distal ends of the
inner limbs (though usually quite irregularly), and when
formed it passes on to the outer limbs. There is some evi-
dence that this is also what takes place in the human rods
and cones, for the “ fibrillation” is said to be limited to the
outer portions of the inner limbs (cf. ‘Quain’s Anatomy,’
1894, vol. iii, part 3, p. 49, fig. 52, after Schwalbe).

Some variation seems to occur in the numbers of the longi-
tudinal threads on the outer limbs; they are sometimes very
numerous (e. g. newt, fig. 30), at others very sparse ; and this
is not only the case in different Amphibia, but in different
specimens of the same. Figs. 3 and 6 are from different
frogs; in one case the threads are crowded, and in the other
quite far apart : the rods in this latter case have been greatly
stretched, but one does not see why that should lessen the
number of striz. The significance of some of the irregularities
of this system of striz! will be better understood when we
have described the connection between these threads and the
contents of the vesicles in whose walls they occur.

The rods, then, are delicate protoplasmic vesicles, in the
thin walls of which staining threads occur. In the walls of
the outer limbs these threads are usually more or less
beaded with clumps of staining matter. The claim made by
Max Schultze and Hoffmann (see the figures and plates re-

' The spiral twist of the striz oa the outer limbs has been rightly attributed
to torsion. Ihave only seen it, and then very marked, on rods broken off
like that shown in fig. 13, ¢.

450 H. M. BERNARD.

ferred to above) that the outer limbs of the cones are also
striated will be discussed later.

The Contents of the Rods.— According to Max
Schultze the outer limbs of the rods are built up of discs
joined together by some cementing substance. This descrip-
tion, propounded by so great an observer, seems to have had
the effect of turning away attention from Hensen’s figures of
cross-sections of rods of the frog,! which clearly showed some
definite internal structure. It must, however, be admitted
that Hensen’s cross-sections differed among themselves ;
there were two kinds (see PI. 30, figs. 7, a, b, which reproduce
them), and they were not easy to reconcile with one another.
Nevertheless I think it cannot be doubted that the discs of
Max Schultze, which are, I believe, artificial phenomena,
helped to consign them to temporary oblivion. As a matter
of fact, Hensen’s figures, which were optical sections and
hence hazy, come near the truth, and are, as we shall see
presently, reconcilable with my own observations. It seems
fairly clear, for instance, that his two sections may compare
with my own figs. 9, b, 138, g, and 12, 13, k, respectively.
Hensen, however, was too anxious to discover nerve-endings,
and was therefore prepared to see fibrils in any clear space
or small refractive portion of the section. In the case of
fig. 7, a, he thought the meshes of the reticulum round the
periphery of the sections were fibrils of doubtful significance,
but in fig. 7, b, those in the centre were regarded as nerves,—
three, he thought, in the centre of each rod.

With regard to the contents of the inner limb of the rod,
its most conspicuous element, the “ ellipsoid,” has long been
known; it has been regarded as the organ in which the
nerves end (cf. fig. 23, on the right), and deserves a separate
section. This is readily accorded, inasmuch as it admits of
being described separately, and what follows will be clearer if
we temporarily ignore it. At the same time we shall find it
necessary to discuss the contents of both outer and inner
limbs together, passing by for the present this particular body.

1 Virchow’s ‘ Arch. path. Anat.,’ Bd. xxxix, 1867, pl. xii, figs. 7 and 8.

STUDIES IN THE RETINA. 451

For a clear understanding of the description and figures
relating to the contents of the rods to be here given, it is
worth while turning once more to their development, and
noting that, in essence, they are protoplasmic vesicles ex-
truded from the retina. As seen in the first part of this paper,
the early stages of these vesicles are seldom found intact,
but when they are they usually appear clear, and apparently
with only fluid contents. Faint traces of delicate proto-
plasmic networks may occasionally be seen (see Part I, Pl. 3,
fig. 16). Networks are, again, found in well-preserved and
properly stained preparations in the large basal vacuoles of
the cones (see Pl. 31, figs. 28, 27, 28). Later we find dis-
tinct networks in the inner limbs of cones and rods, with
usually a certain number of very pronounced threads running
down in their delicate walls (see above and figs. 29, a, h,
z,) ; so also in the outer limbs—which, as we saw in Part I
of this paper, began as fluid vesicles at the tips of the cones—
a protoplasmic reticulum ultimately appears. The staining
reticulum in the outer limbs is not often found as a simple
meshwork, but this is sometimes the case, and we may assume
that it first appears as such. ‘lwo instances are shown in
the figures (4, b, and 6). We gather from these cross-sections
that the clumps on the longitudinal threads running down
the rods are the points of attachment of this internal reti-
culum to the walls of the vesicle. As a rule this reticulum
is not evenly distributed ; we find a tendency for it to be
compressed into the axis of the rod, always, however,
remaining attached by its threads to the wall fibrils. As
this compression increases the threads of the internal
axial portion get very thick, coarse, and matted together.
The compression may go so far that the reticulum merely
consists of an axial strand with a few meshes in it, while
the attaching threads are lengthened so as, in _cross-
section, to look like the spokes of a wheel (see figs. 12 and
13, k, and also cf. Hensen’s optical section reproduced in
my fig. 7, 6).

So far, then, the rods are protoplasmic vesicles, each

452 H. M. BERNARD.

divided into two compartments by a cross-membrane ;! and
as they assume their definitive shapes they become gradually
filled with a staining reticulum, which, omitting the ellipsoid,
develops especially strongly in the outer and, in the adult
Amphibian rod, more important of the compartments.

This account seems to justify the description of the rods as
prolongations of the “‘ visual cells.” It is obvious that each
may be regarded as a prolongation of the cytoplasm belong-
ing to each rod nucleus, a prolongation at first filled with
fluid, but sooner or later containing also the usual reticulum
which ramifies through the cytoplasm of ordinary cells. My
only objection to this description is to the term “ visual
cells.”” My researches long ago compelled me to abandon
the usual conception of the retina as composed of cells, and I
now regard it as a syncytium, in which the nuclei
are arranged in layers, not as fixed morphological
units, but solely as centres of physiological acti-.
vities which may at times require them to migrate
outwards, ultimately, if life lasts long enough, to
become rod nuclei. The evidence for this is, te my mind,
so convincing that I have no hesitation in making the state-
ment, even though it stands in such startling contrast to the
conclusions of nearly all the most recent workers on the
retina, such as Ramon y Cajal, Dogiel, and others, and though
a criticism of the method and results of these authors is here
out of the question. In the first part of this paper, p. 43, I
referred to the migration of nuclei from the middle nuclear
layer to the outer nuclear layer, and showed that, even if we
could not see evidence of it in our sections, it would be
necessary to assume it; and I here add figures of nuclei
passing through the outer reticular layer in different Am-
phibia (figs. 21—28, 25, 26) ; while, again, in fig. 24 one or
perhaps two nuclei have moved outwards together, leaving a
space vacant in the middle nuclear layer, and apparently

1 T have not yet been able to ascertain for certain the time of appearance
of this membrane. As we shall see below, it probably appears before the
ellipsoid,

STUDIES IN THE RETINA. 453

dragging the cytoplasmic reticulum after them. Such
figures might be multiplied indefinitely, and, moreover,
taken from nearly every retina that is closely enough exa-
mined. I reserve full discussion of this somewhat revolu-
tionary conception of the retina as a syncytium for another
communication. But in the meantime I feel compelled to
state my conviction that the rods are not the prolon-
gations of “visual cells,” but protrusions of the
cytoplasm of the retinal syncytium, each, at least
in the Amphibia, dominated by a nucleus,

Passing on from this digression, and regarding it for the
moment as indifferent how we describe the rods in their
relations to the nuclei, the evidence is abundant, as I shall
now endeavour to show, that these nuclei are the centres of
the physiological activity which gives rise to the rods.

In the first place, a great part, if not all of the fluid or
hyaline matter, here always spoken of as fluid, which first
causes the vesicle to protrude, comes from the associated
nuclei.

Fig. 17 can hardly admit of any other interpretation than
that fluid is extruded by the nuclei into the inner limbs of rods.
If it is objected that these figures might as easily be inter-
preted as representing phenomena due to the stimulation of
fixing agents, this argument will not apply to fig. 28,
where we see a “‘ double cone,’”! in which one nucleus is still
large and vesicular, while the other is collapsed, because its
fluid contents have been discharged into the base of the cone
belonging to it. Indeed, a study of cones with their basal
vacuoles makes it very evident that the fluid of these vacuoles
has been derived from their nuclei. Large vesicular nuclei
in the position of cone-nuclei, i.e. well within the membrana
limitans externa, are very common and in striking contrast
to the more condensed rod-nuclei (figs. 16, a, b, and 18).
The same contrasts may also be found in the other nuclear
layers, but here, again, it is impossible to give in this paper

For the correct interpretation of “double cones” in the Amphibia see
Part I, p. 33.

454 H. M. BERNARD.

an extended account of the observations made relating to this
subject. Selecting one more instance, I would refer to
fiz. 20, in which a large fluid vesicle has been discharged
from its associated nucleus, and apparently has not found a
way down as a young cone between the adjoining rod-nuclei,
or, if part of it has succeeded in doing so, that part did not
come into the optical field. Lastly, fig. 19 shows a rod
thrust outwards by an increase in size of its basal vacuole.

In the second place, the staining reticulum of each rod is
also certainly derived from its associated nucleus. Not only
can the reticulum of the inner limbs be seen in direct con-
nection with the linin network of the nucleus (see figs. 29,
a, i,j), but a thick stream can be seen descending from the
nucleus on to the ellipsoid (figs. 10,23, 27), a phenomenon to
which we shall refer more fully later on. Indeed, if the form
of the cone or young rod (figs. 13, d, 15, b, 23, 29, a!) with its
nucleus surmounting its narrow basal neck be kept in mind,
it is difficult to conceive of any other origin than the nucleus
for the large amount of staining material which finds its way
outwards into what was certainly originally a fluid vesicle, with,
at the most, a few delicate reticular strands. The longitudinal
fibrils running down the outer limbs are, in their shape and
arrangement, evidence for this outward movement, while the
clumps of staiming matter along the whole length of their
threads, and the density of the reticulum in the interiors of
the rods, are witnesses of the immense quantity of this staining
matter required.

Actual demonstration of the derivation of this reticulum
of the outer limb from that of the inner limb, and both from
the nuclear reticulum, can be seen in the figures. For in-
stance, there occur, in different parts of the inner lmb,
often in the wall low down and partly apparently embedded
in the ellipsoid, deeply staining refractive bodies, usually
globular, and, what is more important to note, always sur-
rounded by clear zones as if they were the centres of small
fluid vacuoles (figs. 15, a, and 29,c—g). These are certainly

1 Many more are figured in Part I, Pl, 3.

STUDIES IN THE RETINA. 455

chromatin globules, and are usually found in young rapidly
growing retinas.! In well-stained preparations it is common
to find that, from these bodies, fine threads run down the
walls of the outer limb. In one figure of a developed rod
this thread was the only one which took the stain (fig. 29, c).
In another figure, two staining and rather straggling threads
came from one of these bodies, which had apparently been
flattened out against the membranous partition between inner
and outer limb (fig. 15, a). To this phenomenon, i.e. this
membrane acting as a barrier between inner and outer limb,
we shall return.

Even where there are no such bright globules of chroma-
tin, the derivation of the reticulum of the outer limb from
that of the inner can be at once seen if we study the figures
of the developing cones shown in fig. 29,e—j. These figures
are merely a selection, and might be multiplied indefinitely.
They show quite clearly that the staining material within the
outer limb appears where the thin threads from the inner limb
come down on its wall. This fact shows that the striation of
the outer limbs of the cones figured by Max Schultze and
Hoffmann may exist, not as a complete system as they repre-
sented it, but as the first beginnings of the subsequent stria-
tion of the rods.

Unfortunately none of these figures (29) seem to show the
true tips of the cones; still, enough is here seen to demon-
strate the point we have immediately in hand.

Lastly I would refer to fig. 26, which is by no means an
uncommon phenomenon. A nucleus is seen passing through
the outer reticular layer and about to join the outer nuclear
layer (that of the rods and cones). It is preceded by a fluid
space, while from it a very delicate reticulum streams out-
wards. This I interpret as representing a very early stage
in the formation of a rod, being still entirely within the

‘The only other figure I know of which shows such a body is one by
Hensen (I. c., fig. 7, c), who, as we have seen, came so near discovering the
structure of the rods, having failed apparently for the want of better micro.
scopic technique.

456 H. M. BERNARD.

retina. The fluid vesicle in the ordinary course of things
would, on approaching the mem. lim. externa, form the usual
conical protrusion, and into it the staining reticulum would
follow. On the other hand, it is only fair to note that
streams of very delicate staining reticulum occur elsewhere ;
one other, for instance, is shown running up from the left-
hand rod-nucleus in fig. 27. The explanation of this must
be deferred until I can at the same time give the evidence
in full on which it rests, and this I hope to be able to do in
the near future.

Further Contents of the Rod.—So far, then, we have
described the origin and structure of the rod as a protoplas-
mic protrusion from the retina, containing the usual staining
network very strongly developed in the outer limb, and with
some clear fluid in the meshes or interstices.

This network and this fluid are not, however, the sole
contents of the normal rod, and the striking difference
between inner and outer limbs, apart from the difference
in shape and density of the reticulum, is found in the fact
that while the former remain protoplasmic vesicles, with
apparently soft, flexible walls filled with these elementary
constituents which we have described (passing over for the
moment the ellipsoid), the outer limbs become filled with
some highly refractive substance, which renders them turgid.

The change from the loose, long terminal bag found at
the tip of the advanced cone (c,) to the outer limb of the
rod (7) (see Part I, Pl. 3, fig. 4) 1s seen to consist not only
in the squeezing outwards of the staining matter to the
distal end of the inner limb, but also in the filling up of
the outer limb. Now while we have traced to its source
some of the matter which helps to fill the outer limb, viz.
the staining reticulum, this will not account for the refractive
contents which now seem to make them turgid and cylindrical.
Further, we saw that the outer limbs of the rods lengthened
(from 7, to 7r,),and hence apparently continued to take in more
of this refractive constituent of their contents ; and not only
lengthened, but as a rule became also much thicker, I have

STUDIES IN THF RETINA. 457

noticed also that the outer limbs of Schwalbe’s rods (7, and r,)
were in most cases rather more deeply stained than the longer,
thicker definitive rods, although I lay no great stress on this.
The accidents which can never be eliminated from our technical
methods are too numerous to allow conclusions to be based
upon mere variations in diffuse staining. I mention the point,
however, just because it is possible that the proportion of the
refractive matter to the staining reticulum might be expected
to be less in an outer limb, just beginning to fill up, than in a
large swollen rod. It is this refractive matter which gives the
rods their characteristic appearances, and which has led to
their being classed among cuticular structures.

The source of this refractive matter is to be seen in the
pigment epithelium into which the tips of the rods are plunged,
and it is largely composed of pigment granules, probably with
some portion of the protoplasm of the epithelial cells. At
least the absorption of cytoplasm as well as pigment by the
rods can actually be shown to take place under special cir-
cumstances, as we shall presently see.

In the first place, dealing for the moment with general
considerations, I again refer to the development of the rod;
a fluid vesicle is thrust into the pigment layer, and slowly
becomes filled with refractive matter. Both the vesicle and
the epithelial cells are, so far as we can see, naked proto-
plasm in the very closest contact with one another,—indeed,
tightly interlocked, the pigment cells constantly foreing a
passage up between the packed rods.!_ Between these some
interaction is almost certain to take place. This interaction
is, I contend, in part at least an absorption of pigment by
the rods. The pigment of the epithelial cells is constantly
recruited by an outward streaming of granules from the
choroidal layer adjacent to it, a streaming which can be seen
in every successful preparation. So that we may conclude
that pigment is being used up and as constantly replaced.
The only other alternatives to this view are either that the
refractive matter in the outer limbs of the rods comes from

' For the evidence that the rod layer is normally compact see Part I.
vou. 44, PART 3,—NEW SERIES, GG

458 H. M. BERNARD.

the retina, or that it is manufactured in situ within the
rods.

That it does not come from the retina, from which we can
easily trace the fluid and the staining network, we gather
from the total absence of any refractive matter in the inner
limb except in the ellipsoid; and, as we shall presently see,
the position of this body forms additional evidence that the
source of the refractive matter is from without inwards to-
wards the retina, and not from the retina outwards.

That the matter is not manufactured in situ we gather
from the microscopic appearances, which show very clearly
that it is forced in through the walls. The evidence for this
is to be seen in the changes already described, which take
place in the character of the reticulum within the outer limb
of the rod. Figs. 4, b, and 6, b, show this reticulum simply
diffused equally across the section; figs. 13, c—k, and 12
show different stages in its compression towards the axis
of the rod. Now it is difficult to explain this compression
except on the assumption of some matter passing in through
the walls and crushing it inwards, stretching, or perhaps
merely lengthening the threads which attach it to the walls.
Fig. 138, 7, shows the process as being irregular, while fig. 14
shows that it may take place locally, i.e. along one side of
a rod and not on the opposite side. This observation is im-
portant, because it is in keeping with the fact that the
tongues of the pigment cells run up lengthwise between the
rods. Fig. 13, 7, shows that at times the reticulum, though
compressed towards the axis, may retain some of its concentric
threads, the refractive matter passing them by. ‘The refrac-
tive layer was here 1°5 » thick, the whole rod being 9 p.!

Again, in eyes in which, after exposure to light, the pig-
ment has been forced up to the membrana limitans externa,
individual granules can be seen remaining behind after the
general retreat of the pigment, and sticking to the clear
protoplasmic walls of the inner limbs. Many of them can

1 Zenker (‘ Arch. mikr, Anat.,’ iii, 1867, p. 259) discovered that the outer
layer of the rod is more highly refractive than the axial portion,

STUDIES IN THE RETINA. 459

then be seen obviously fading away, the shape being re-
tained, but the bright colour and sharpness of contour have
disappeared, and the whole appearance suggests their being
slowly absorbed. Although, as above stated, with the excep-
tion of the ellipsoid (and the oil globule in thecones of the frog),
I have never found refractive matter in the inner limbs in
Amphibia, cases occur elsewhere in the animal kingdom in
which large inner limbs become filled with it, but in a manner
entirely confirmatory of my argument that its source is the
pigment epithelium. The clinging of pigment granules to
the protoplasmic walls of cones was noted in Part I.

Again, in a series of sections of retinas of animals which
had been exposed for three hours to the light of an arc lamp,!
the heat rays being screened off as far as possible, one
interesting result is conspicuous. The pigment epithelium
is here and there disorganised, and isolated pigment cells
have forced their way up to various heights among the rods.
These can be found in all stages of losing their pigment;
some appear as nuclei still thickly enveloped in pigment,
others with only a trace of pigment, while here and there
nuclei alone persist from which all the pigment and the
protoplasm have disappeared. Fig. 12 shows in a tangential
section, selected because of the cross-sections of the rods,
such a nucleus, bereft of all its pigment, embedded among
rods, and in these latter the reticulum has been compressed
into the axis, which, as above suggested, indicates the absorp-
tion of extraneous matter through the walls.

Other effects of this exposure to such a fierce light have
still to be studied. For instance, the contents of the rods
have a singularly blotchy appearance, but I cannot satisfy
myself whether this lies in the object or in the accidents of
staining.

While these arguments are, I think, sufficient for the

1 T am indebted to my friend Mr. George Newth, of the Royal College of
Science, not only for the use of the necessary apparatus, but also for indis-
pensable advice and assistance in making a series of experiments with pure
monochromatic light, the results of which are still being worked out.

460 H. M. BERNARD.

present demonstration that the refractive matter within the
outer limbs is absorbed by the rods from the pigment, I
should like to mention two points on which I am in great
uncertainty. It has appeared to me more than once as if
the pigment granules could pass bodily into the rods, and,
at least for a time, maintain their individuality. I do not
see why this should not occasionally happen ; indeed, I cannot
explain some of the phenomena on any other hypothesis.
Still, the evidence shows conclusively that this is not the
normal method, but that the pigment granules are absorbed
as a colourless or nearly colourless refractive and amorphous
matter. The occasional finding of retinas in which the
colour of this refractive matter within the rod is the same as
that of the pigment granules without (I have seen this in
sections of the retinas of the pigeon and of frog tadpoles,
etc.) may be mentioned, in passing, as additional evidence of
the origin of the former from the latter.

One appearance suggestive of pigment granules within the
rod seen in osmic acid preparations must be familiar to all
students of the retina. Itis the “disc” formation on which
Max Schultze laid so much stress. I now, however, refer
this to a transverse flaking of the internal reticulum, perhaps
a kind of coagulation of the same, as Max Schultze himself
suggested. ‘The transverse flakes are usually deeply coloured
by osmic acid, and often appear exactly like layers of intruded
pigment granules. In preparations not treated with osmic
acid the appearance is not to be found.

The second point is the relation of the phenomena here
detailed to the visual purple. ‘This is said to be produced in
the dark through the interaction of the rods and the pigment
epithelium, i.e. when the epithelium is only in contact with
the tips of the rods, and, further, it is said to be bleached by
the light, i.e. when the rods should, according to my own
observations, be absorbing clear refractive matter from the
epithelial cells, which are then in intimate association with
the rods, inasmuch as tongues of the cells then travel up
between the rods. I am of course aware that it is frequently

STUDIES IN THE RETINA. 46]

maintained that fine protoplasmic processes of the pigment
cells are permanently advanced as far forward as the mem-
brana limitans externa, and are thus always in contact with
the rods. Not in any single one of the retinas of some
twenty-five vertebrates I have yet examined, and their number
must, I think, now amount to fully one hundred, fixed and
stained by all the latest methods, and examined with the
best available microscopic lenses, have I been able to find a
trace of these processes of the epithelial cells permanently
interlocking with the rods. On the contrary, when the pig-
ment is retracted the contour of the pigment cells is per-
fectly straight or rounded as the case may be. Had such
processes existed, I am convinced that at least some evidence
of their presence would have forced itself on my attention
long ago.

I have, therefore, so far no point of connection to offer
between the physiological details here described and the
visual purple, which appears when, according to my own
observations, the rods should be getting rid of the matter
absorbed when last the light forced the pigment cells into
close contact with them, and is bleached when they ought
to be absorbing, and at the same time clarifying, the warm
colouring matter of the pigment. A reconciliation of these
observations will doubtless some day be forthcoming, and
there the matter must be left for the present.

The Ellipsoid.—This somewhat inappropriate name is
usually applied to the body found in the inner limbs of the
Amphibia where these limbs abut against the outer limbs.
Max Schultze regarded it as a plano-convex lens; the name
here adopted was suggested by Krause (‘‘Opticus Ellip-
soid”’). It is here preferred robbed of its prefix “ opticus,”
so aS not necessarily to suggest special functions.’ So far as
the terms describe form alone, “ plano-convex” is prefer-
able to ellipsoid for the Amphibia, for that is the most usual
definitive form assumed in the adult rod, i.e. when the
rod is not very large and thick, as it is in the axolotl, in

1 Krause thought it was the nerve-end organ (‘ Anat. Untersuch.,’ 1860)

462 H. M. BERNARD.

which case the body is usually an irregular flattened disc
(fig. 28).

Asa matter of fact, the body is of very various shapes.
Fig. 15 shows a series of cones and rods (salamander) in
which only in a young cone is the body egg-shaped, in others
it takes the shape of the tip of the swollen inner limb of the
cone: if the latter is large, the body is large; if narrow, the
body is narrow, while in the definitive rod it is uniformly
plano-convex. It thus seems quite plastic in its earlier (cone)
stage, and only assumes a definite form in the full-grown
rod.

Dealing, then, with this body as we have with the other
contents of the rod, we must regard it as an aggregation of
these contents which, for some reason or other, rests perma-
nently against the transverse membrane separating the inner
and outer limbs.

It varies greatly in its staining. It is sometimes intensely
stained, at others it is comparatively clear and refractive.
Tn this latter case a dense stream of staining matter is very
frequently seen descending upon it from the nucleus (see
figs. 10, 28, 27). We cannot be far wrong, then, if we
refer the variation in the intensity with which the body
takes stain to the relative proportions of staining matter and
refractive matter which compose it. For out of these two
substances, which, as we have seen, together constitute the
visible contents of the rods, it must surely consist.

Regarding it for the moment in its definitive plano-convex
form, it seems to me that we have, both in its shape and in
its position, striking confirmation of our conclusion as to the
origins of the contents of the rod. On the one hand, we
have an outwardly streaming reticulum of staining matter
which, so far as we can see, only manages to get further, 1. e.
into the outer limb by way of the outer walls. There cer-
tainly seems to be some condensation of the reticulum against
the blind end of the inner limb (see fig. 27, left-hand figure).
On the other hand, coming into the rods from the opposite
direction, viz. from the pigment epithelium, we have the re-

STUDIES IN THE RETINA, 463

fractive matter. ‘his, as we have seen, is absorbed by the walls
of the rods filling them up till they are turgid. This matter
would thus find its way inevitably up against the transverse
membrane separating inner from outer limb, and, seeing that
it passed through the outer wall into the rod, there is no
apparent reason why it should not pass through this transverse
membrane from: the outer limb into the inner limb. This,
then, I believe, is what takes place, the very form of the
ellipsoid being suggestive of its having been forced through
to form a kind of drop on the proximal side of-the transverse
membrane. Confirmatory evidence will later be adduced from
other retinas, but sufficient to establish the point will be
found in what follows.

When we come to the ellipsoid in the cones (see figs.
15, c—e) it would seem that the explanation we have given
of it in the rod could hardly apply. ‘There appears to be a
transverse membrane (fig. 29, f, 7), but there is no swollen
outer limb filling up with refractive matter. Nevertheless
the explanation of the ellipsoid is practically the same, as
we can gather from the conditions seen in the frog. In the
cones of the frog there is invariably a round refractive globule
at the junction of the basal and the conical portion. In
well-stained specimens a mass of staining matter is generally
seen abutting against this globule, as if they mutually blocked
the way for one another. We thus get practically the same
condition as in the rod, though in this case we do not know
exactly where the transverse membrane is, i.e. whether the
refractive globule is on its inner or outer side.

This parallel assumes (1) that the refractive globule of the
cones of the frog is of the same substance and has the same
source as the refractive matter in the rod, and (2) that this
refractive globule and the adjacent staining matter will later
fuse together to form the definitive ellipsoid.

The former of these assumptions is, | think, fully justi-
fiable. We have seen how readily pigment granules cling
to the thin protoplasmic walls of the cones, and can be seen
fading away on the fine membranous walls of the inner limbs

4.64 H. M. BERNARD.

of the rods, as if in the act of being absorbed. Hence it is
but natural to assume that some of the refractive matter
which later fills these vesicles to overflowing should early find
its way into the tips of the cones and be squeezed out by
the lateral pressure described in Part I as existing in the rod
layer, so as to appear as refractive globules just above the
line where the pressure of the rods ceases, i.e. on a line
between the junctions of the inner and outer limbs. The
secondary thrusting back again of these globules in cones
(c,), described in Part I of this paper, needs no comment.

Then, again, I mentioned in Part I that in young tadpoles
it was possible at times to see these globules actually dis-
appearing in the ellipsoids of young rods (see Pl. 3, fig. 15),
showing clearly that, in this refractive globule of the cone
with its adjoining staining matter, we really have the ele-
ments of the future ellipsoid, though not blended together.
Further, in one of my slides of a young frog tadpole the
refractive matter absorbed by the rod is not always dis-
coloured; globules of bright reddish-brown matter exactly
resembling the pigment in colour occur high up in the rod,
near the transverse membrane, while as a complete confirma-
tion of the argument, globules of exactly the same colour
can here and there be found in the ellipsoids of the same
rods.

The condition found in the cones of the frog thus helps us
to understand the ellipsoid in the cones of the other Am-
phibia here dealt with. It has long been known that the
refractive globule was absent from the cones of the toad, an
absence which was disconcerting to the earlier investigators,
who would attribute to it an important dioptric function. Itis
also absent from the cones of the salamander and the axolotl.
In these cases, from our point of view, it is not so much that
the refractive matter is absent, but that it never really forms
as a distinct globule ; it is mixed with the staining matter to
become the ellipsoid as fast as it collects.

In the case of the newt, all students will remember that

STUDIES IN THE RETINA. 465

Max Schultze, and others after him,! described and figured a
combination of two “lenses,’’ a biconvex and a plano-convex,
as a higher specialisation than the simple plano-convex “lens”
(the ellipsoid) of the frog, toad, salamander, etc. Max
Schultze even claimed that this lens could be isolated. The
body which he figured can be seen frequently enough, but
not by any means always in the shape of a biconvex lens.
It is nothing but a fluid vacuole, more sharply defined than
usual. Fig. 30 shows two rods of a newt side by side; in
one there is a well-defined vacuole resting on the ellipsoid,
and in the other a quite undefined vacuole like that usually
found in other Amphibia. The former is interesting because
its origin from the nucleus can be seen, a second one appear-
ing ready to escape. Most of the nuclei in this preparation
have vacuoles about the same size as shown in fig. 30.
Further, in very many of the outer limbs of the rods rows of
fluid globules of different sizes can be seen. Compare the
views as to the origin of the fluid on p. 453.

Let us sum up the conclusions so far arrived at, forbearing
to enter more fully into the physiological results obtained
till the corroborative evidence yielded by the eyes of verte-
brates other than Amphibia can be prepared for publi-
cation.

The rods in the Amphibia are specialised protrusions of
the retina, consisting of extremely delicate protoplasmic vesi-
cles, each divided by a transverse membrane into an inner
and an outer compartment. The staining reticulum which
traverses these vesicles is especially developed in the outer-
most, into which it finds its way in threads down the walls.
These threads, at short distances, give off other threads from
small nodes into the interiors of these outer vesicles. These
latter further become filled with refractive matter absorbed
from the pigment epithelium, and certainly largely obtained
from the pigment granules. This matter absorbed through
the walls condenses the mass of the reticulum into the axes

1 Cf. ‘Arch. mikr. Anat.,’ Bd. v, 1869, pl. xxii, fig. 2@. See also‘ Bronn’s
Thierreich ’ (Amphibia).

466 H. M. BERNARD.

of the rods. <A portion of this refractive matter exudes
through the transverse membrane, where it mixes with the
staining matter of the inner limb, and forms the body infe-
licitously termed the ellipsoid.

(lo be continued.)

EXPLANATION OF PLATES 30 and 31,

Illustrating Part IT of Mr. H. M. Bernard’s “ Studies in
the Retina: Rods and Cones in the Frog and in some
other Amphibia.”

PLATE 30.

Fic. 1.—The basal portions of two rods (newt) with remains of ellipsoids,
and showing the relations of inner and outer limb as two vesicles separated
by a thin membrane.

Fic. 2.—Part of an adult rod (toad) distorted, and showing its sac-like
character, the contents having ruptured down the middle. Cf. Max Schultze’s
Fig., ‘A.M. A.,’ iii, pl. xiii, fig. 18¢ (Pike).

Fic. 3.—a, 6. Two surface views of rods (frog, Flemming), showing the
dotted appearance of the longitudinal striation. 4. With a pigment granule
to show relative size of dots.

Fic. 4.—a. Rod of young salamander (boiling corrosive sublimate), showing
the longitudinal striation straggling irregularly. 6. Cross-section of a rod of
same eye, showing a simple reticulum attached to the dots in the walls.

Fies. 5 and 6.—Parts of rods and cross-sections from retina of frog (exposed
to osmic vapour); in Fig. 6, a, the rod was abnormally stretched.

Fie. 7.—a, 6. Two cross-sections of rods (frog) according to Hensen. a.
“Optical section of a fresh rod.’ 4, ‘ Optical section of a rod fixed with
osmic acid.”

Fig. 8.—A cross-section of the same according to Max Schultze.

Fie. 9.—Part of a rod (a), with cross-section (4), and a small piece cut off
tangentially (c) from retina of axolotl; note the density of reticulum uear the
tip of the rod. Cf. @ with Hensen’s section, Fig. 7, a.

Fie. 10.—Upper portion of rod of same (Perenyi), partly in optical section,
showing part of nucleus and inner limb; the reticulum of the latter can be
seen (1) coming from the nucleus, (2) condensing on what is called the ellip-

STUDIES IN THE RETINA. 467

soid (cf. Figs. 28 and 27), and (3) passing in fine threads on to the outer
limb ; the striz which continue them are obscured by the internal reticulum.

Fic. 11.—A diagram of the same showing the surface arrangement of the
threads ; the symmetry is, however, seldom so complete on the inner limb.
The threads thicken towards tle tip of the rod.

Fig. 12.—Tangential section through a group of rods of a salamnander after
three hours’ exposure to intense light. The cross-sections compare with
Hensen’s section, Fig. 7, 6. Among the rods is the nucleus of a dislocated
pigment cell, from which all the cytoplasm and pigment have been absorbed.

Fic. 13.—From a toad fixed in boiling corrosive sublimate. a, 4, show
a thick basket-like reticulum (“ Faserkorb”) of inner limb near the ellipsoid ;
c, part of a rod broken away, and showing the longitudinal threads beaded
with dots of staining matter; d, a Schwalbe’s rod with threads apparently
running over the deeply stained ellipsoid and on to the outer limb; e,f, 9,
sections of rods cut at different angles; from cto /what appears to be a
homogeneous rind is seen; 4, a rod 9 pw thick, with a “rind” 1°5 p thick;
part of the same in 2, as interpreted when examined under high magnifica-
tion (1500 times; apoch. 2 mm.; N.A. 1°4, comp. oc. 12); 7, a part of a
cross-section, showing the material of the rind fitting into irregular spaces
of the reticulum; #, cross-section of a narrow rod 5 p thick.

Fic, 14.—Optical section of a rod from the other eye of the same animal ;
the “rind” seen only on one side.

Fic. 15.—Salamander (boiling corrosive sublimate) : @, the upper parts of
a rod, showing dark body in or on ellipsoid, flattened like the latter against
the transverse membrane (see Fig. 1), and from this body two staining
threads descend on the outer wall of the rod; 4, ellipsoid, egg-shaped in
young cone; ¢c, the same with its shape adapted to its position, and with mass
of staining material in contact with it; d, e, the same; f, g, ellipsoids in
rods and in their definitive shapes, with dense staining matter condensing on
them (cf. Fig. 10).

Fre. 16.—Toad (strong alcohol): a, two condensed rod-nuclei protruding
through the mem. lim. externa, and a vesicular cone-nucleus in contact with
the outer reticular layer; 4, the same, with similar vesicular nuclei, one in
the middle, and one in the outer, nuclear layer.

Fic. 17.—The same, condensed rod-nuclei with fluid vacuoles in the inner
limbs, as if exuded by these nuclei.

Fic. 18.—Axolotl (Perenyi) ; a condensed rod-nucleus with a much larger
vesicular nucleus above it (cf. also Fig. 23).
Fic. 19.—The same; a rod forced farther than its neighbours into the pig-

ment, apparently by the intrusion of fluid into the inner limb, thereby dis-
placing the ellipsoid.

468 H. M. BERNARD.

Fic. 20.—Frog (Flemming) ; a nucleus in position of cone-nuclei, but con-
densed, with a vacuole near it and apparently discharged from it.

Fie. 21.—Frog (Flemming); two nuclei passing, with ameeboid changes of
shape, through the outer reticular layer.

Fie. 22.—Axolot], showing the similarity of the cone-nuclei to those of
middle nuclear layer.

PLATE 31.

Fic. 23.—The same, showing a nucleus passing through the outer reticular
layer; a double cone, one with condensed and the other with vesicular
nucleus (cf. Fig. 18, right-hand nuclei).

Fic. 24.—The same, showing a spot where one nucleus, or perhaps two
nuclei, have passed outwards, and apparently dragged the tissue outwards
with them and left a large gap in the middle nuclear layer.

Fie. 25.—Salamander (Perenyi), showing three nuclei, two working out-
wards through the outer reticular layer.

Fic. 26.— Frog (Flemming); a nucleus passing through the outer reticular
layer, preceded by a fluid space and an exquisitely fine staining network.

Fie. 27.—Axolotl (Perenyi); a large nucleus with a similar fine staining
network above it, and the reticulum condensing at the distal end of its inner
limb; on the right is a small condensed rod-nucleus, with the reticulum of
the inner limb condensing on the ellipsoid (cf. also Figs. 10 and 23).

Fic. 28.—Frog (Flemming), showing upper end of a cone with a reticulum
traversing the basal vacuole.

Fic. 29.—a—j, a series of figures from the retina of a young salamander
(seven weeks old, Perenyi), showing in developing cones and in rods the
origin of the staining reticulum of outer limbs from that of the inner limbs,
and sometimes from bright round masses of chromatin. ‘lhe connection of
the reticulum with that of the nucleus is shown in a, 4, 9 (ef. ¢, d, e, 4, g, with
Fig. 15, a).

Fic. 30.—Two rods from the retina of the newt. The very delicate longi-
tudinal striation of the outer limbs is shown, and in connection with the outer-
most rim of the coarsely granular ellipsoid. Above the ellipsoid is a fluid
vesicle very variously developed, but not infrequently slightly flattened (Max
Schultze’s ‘‘ biconvex lens”). Fluid vesicles or vacuoles are also very com-
monly seen in the nuclei and in the axes of the rods.

STAINING WITH BRAZILIN. 469

Staining with Brazilin.
By

Sydney J. Hickson, F.R.S.,
Beyer Professor of Zoology in the Owens College, Manchester.

Brazitin is extracted from the wood of Czxsalpinia
echinata, growing in South America. Formerly, under the
name Brazil-wood, it was used commercially as a dye; but
of recent years it has been superseded by other colouring
substances, and practically driven out of the market.

Chemically it has considerable resemblance to hematoxylin,
which is also extracted from the wood of a tree belonging to
the order Leguminose, but brazilin (C,,H,,0;) differs from
hematoxylin (C,,H,,0,) in having one hydroxy] less. Further
information concerning the chemistry of brazilin may be
found in the papers by Professor Perkin and Dr. Gilbody
published in the ‘Proceedings of the Chemical Society’
during the past four or five years.

Brazilin has been tried as a stain for animal tissues by
Flechsig and Breglia, but apparently the results were not
very satisfactory. Their methods, though mentioned in the
third edition of Lee’s ‘ Microtomist’s Vade-Mecum,’ are
omitted from the fourth edition,

Brazilin is mentioned as a chromatin stain in Rawitz’s
‘ Leitfaden fiir histologische Untersuchungen,’ p. 98, but it
is not mentioned in the elaborate ‘Tabellen zum Gebrauch
bei mikroscopischen Arbeiten,’ by W. Behrens, published in
1898, nor in Bohm and Oppel’s ‘'Taschenbuch der mikro-
scopischen Technik,’ published in 1896.

My first experiment was to attempt to use brazilin instead
of hematoxylin in the method suggested by Heidenhain,

4,70 SYDNEY J. HICKSON.

and now widely used in Eneland and elsewhere under
the name of the “ Tron-Heematoxylin method.” This experi-
ment did not give results that were at all satisfactory.
Another series of experiments which have been made in the
Owens College laboratory by Mr. Wadsworth, under my
direction, have shown that the most satisfactory results are
obtained by the following method :—“ The sections are placed
in a solution of iron-alum (1 per cent. iron-alum in 70 per cent.
alcohol) for one to three hours, and then placed, after slight
washing in 70 per cent. alcohol, in a 3 per cent. solution of
pure brazilin in 70 per cent. spirit.”

Brazilin stains much more slowly than hematoxylin, and
we have found that generally it takes several hours (three to
sixteen) to give a good sharp definition. After staining, the
sections are washed in pure 70 per cent. spirit, passed
through the usual stages, and mounted. There is seldom any
need to wash the sections in iron-alum after staining. It will
thus be seen at the outset that this method possesses two
advantages over iron-hematoxylin,—the sections are never
taken down into water, and the number of washings is con-
siderably reduced.

The results are, to my mind, eminently satisfactory ; for
not only is brazilin a definite chromatin stain, but in nearly
all tissues some parts of the cytoplasm are also stained,
though of a different colour. It isin most cases a double
stain, but with some tissues it is a treble stain.

We have used as a test object a series of sections through
the body of a larval newt. Each of these sections exhibits so
many subjects of histological interest that it is impossible to
describe them briefly. I think it would be admitted, however,
by all observers that nearly, if not all the tissues are well
stained. ‘The chromosomes in the cells that are dividing in
the skin and ovary are very sharply defined and of a deep
purple colour, while the granules in the cytoplasm are brown.
On making a comparison of our preparations of newt’s testis
stained by the iron-brazilin method and by the iron-hema-
toxylin method, it was found that the karyokinetic figures

STAINING WITH BRAZILIN. 471

were equally well defined in the two preparations; but the
iron-brazilin showed an advantage in that the spermatozoa
were triple stained, the head, middle piece, and tail being
clearly of different colours.

A preparation of the sciatic nerve of a dog stained in
brazilin exhibits the neurokeratin remarkably clearly, but
the axis-cylinder is not well stained.

Our method has great advantages in showing the blood-
supply of an organ, the red blood-corpuscles being stained
very deeply. Preparations of ovary of a cat and kidney of a
dog exhibit the capillary plexus excellently.

It is, however, principally in cytological work I think that
the new method will be valuable, as we have found after
many experiments, in our investigations on the very delicate
nuclear structures of the Suctorian Dendrocometes, that
it gives us by far the best general results.

It is obvious that a good deal more must be done before an
opinion can be expressed as to the general value of this
method, but I have thought it right to bring it before the
notice of workers in various branches of microscopic science
at the present time, because I feel that it is a method well
worthy of further trial.

In conclusion I may say that the somewhat severe test of
exposing slides to the action of direct sunlight for several
months has not indicated any appreciable fading effect.

I am indebted to Professor W. H. Perkin, jun., F.R.S., for
much valuable advice and assistance in working out this
method.

eS

Mi A 0 ‘raraeeae? Wie ads hy ies wlk ein
a - He aera s Bah ales Ur Si bead! > Fe, bi) et

any jet ye tines ary pas abet od mal e> en neue)
ot ee ieee co (ipoon Bll. nm
ve J Le & 01 ‘fs ic se ony ee -

zc ae g i ei in| Ok 4 - ast a3 iitieg
y

a Taine eet: ; “wae gta

| . : ‘
at ‘ ees Lek Atel »

With Ten Plates, Royal 4to, 5s.
CONTRIBUTIONS TO THE KNOWLEDGE OF RHABDOPLEURA
AND AMPHIOXUS.
By E. RAY LANKESTER, M.A., LL.D., F.R.S.
London: J. & A. eee 7, Great eee Street.

Science.

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

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

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

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

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

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

TO CORRESPONDEN TS.

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

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

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

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

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

Messrs. ApLARD & Son, 223, Bartholomew Close, E.C., on
the following terms:

For every four pages or less—

25 copies : : : 5/-
0M =; ; 6/-
100 7/-

Plates, 2/- per 25 if uncoloured ; if colour -ed, at the same rate for
every colour.
Prepayment by P.O. Order is requested.
ALL COMMUNICATIONS FOR THE EDITORS TO BE ADDRESSED TO THE CARE
oF Messrs. J. & A. CHURCHILL, 7, GREAT MarLBoROUGH STREET,
Lonpon, W.

THE MARINE BIOLOGICAL ASSOCIATION

OF THE

UNITED KINGDOM.

be 0:
Patron—HIS MAJESTY THE KING.

President—Prof. E. RAY LANKESTER, LL.D., F.R.S.

20:

\

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

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

Naturalists desiring to work at the Laboratory

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

Works published by the Association

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

The Journal of the Marine Biological Association

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

In addition to these publications, the re8ults of work done in the Laboratory
are recorded in the ‘Quarterly Journal of Microscopical Science,’ and in other
scientific journals, British and foreign.

Specimens of Marine Animals and Plants,

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

TERMS ‘OF MEMBERSHIP.

ANNUAL MEMBERS . : ‘ . £1‘ 1 Oper annum:

LIFE MEMBERS . : ; ; . 15 15 0 Composition Fee.
FOUNDERS . 5 é : ; LOO=Ons0 sg 33
Governors (Life Members of Council) 500 O O 5 =

Members have the following rights and privileges:—They elect annually the
Officers and Council; they receive the Journal free by post; they are admitted to
view the Labotatory 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, shouft communicate with—

The DIRECTOR,
The Laboratory,
Plymouth.

New Series, No. 176 (Vol. 44, Part 4). Price 10s.

sa,

AUGUST, 1991.

THE

QUARTERLY JOURNAL

OF

MICROSCOPICAL SCIENCE.

EDITED LY

i. RAY -LANKESTER, M.A., LL.D., F.R.S.,

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

WITH THE CO-OPERATION OF

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

FELLOW AND TUTOR OF TRINITY COLLEGE, CAMBRIDGE};

Wi, Fo B.. WELDON, M.A., FP R:S:.,

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

AND

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

BEYER PROFESSOR OF ZOOLOGY IN THE OWENS COLLEGE, MANCHESTER.

WITH LITHOGRAPHIC PLATES AND ENGRAVINGS ON WOOD.

4, ONDON:
J. & A. CHURCHILL, 7, GREAT MARLBOROUGH STREET.
1901.
|

Adlard and Son,] ; [Bartholomew Close.

CONTENTS OF No. 176.—New Series.

MEMOIRS:
PAGE
On Two New Species of Onychophora from the Siamese Malay
States. By Ricuarp Evans, M.A., B.Sce., of Jesus College, Oxford.
| (With Plates 32—37) . : ; ‘ 5 : : ; . 473

Hoperipatus Butleri (nov.sp.). By Rrcuarp Evans, M.A., B.Sc.,
of Jesus College, Oxford. (With Plate 38) . : ; 539

On Two New British Nemerteans. By R.C. Punnett, B.A. (With

Plates 39, 40) 547
The Ceelomic Fluid in Acanthodrilids. By W. Braxtanp Benuam,
D.Se., M.A., F.Z.S., Professor of Biology in the University of
Otago, New Zealand. (With Plate 41) Ate yf : : . 565
The Crystalline Style of Lamellibrauchia. By 8. B. Mirra, of Cal-
cutta, late of University College, London. (With Plate 42) < c

TITLE, ConTENTS, AND INDEX.

TWO NEW SPECIES OF ONYCHOPHOKA,. 473

On Two New Species! of Onychophora from the
Siamese Malay States.

By

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

With Plates 32—37.

Part I.
ContTeENTS.

PAGE
I. Introduction - ; : : . 474
II. The Classification of the Ghyekooiora : : . 477
ILI. Some Notes on the Type specimen of E. Simatt anus . 484
IV. Description of E. Weldoni and E. Horsti . ‘ . 486
a. Introduction : : ; : . 486
B. External Characters : ‘ : . 487
c. Internal Anatomy . : , . 496
(a) Introductory Rewatke : : : . 496
(6) Ventral Organs : ; ; . 496
(c) Salivary Glands : : ; : . 497
(d) Renal Organs ; : : . 498
(e) Female Reproductive Organs . 500
(/) Male Reproductive Organs. : . 509
(g) Male Accessory glands : : : . 513
(A) Crural Glands : : ‘ 2 > O19
V. Structure of the Ovum. 520

VI. The Relations of the various Genes of the Pevipatide to one
another. ; : ; : : . (525
Addendum... : ; : : : . 530
VIL. Bibliography : ; ; : ; : . 5380
VILLI. Explanation of Plates. : f : : ~ 52

1 The discovery of the two species here deseribed was announced in
‘Nature,’ vol. Ix, p. 591, by E. B. Poulton, Esq., M.A., F.R.S., Hope Pro-
fessor of Zoology in the University of Oxford.

VoL. 44, pART 4,—NEW SERIES. H H

474, RICHARD EVANS.

IT. Inrropuction.

Peripatus, although found in the islands of the Malay
and Melanesian Archipelagos, in Sumatra and New Britain,
has not been hitherto discovered on the mainland of the
Asiatic continent. Dr. R. Horst (8) recorded a specimen
from Sumatra in 1886. The specimen in question was subse-
quently named P. sumatranus by Sedgwick (15). Several
years (1898) afterwards Dr. Arthur Willey described a
species which he himself collected in New Britain (17, 18).

The general characters of the Sumatran species are those
common to all the Neotropical forms of Peripatus. But
Dr. Willey points out that, though New Britain is geographi-
cally an intermediate locality between Sumatra and the Neo-
tropical region, the New Britain species does not possess a
single external character of importance in common with the
Neotropical forms, although by a singular coincidence the
female has the same number of claw-bearing legs—twenty-
four pairs—as the alleged Sumatran specimen. Dr. Willey
further remarks that under the circumstances the evidence
that the latter was found in Sumatra, which Sedgwick re-
gards as inconclusive, must appear more than ever worthy of
suspicion. The material which will be described in the
present account will show that the doubt expressed by Mr.
Sedgwick, and subsequently by Dr. Willey, was not well
founded, though it was, under the circumstances, justifiable.
While the New Britain species possesses no single external
feature in common with the Neotropical forms, the Malay
species described in the following pages, on the contrary,
agree in almost all the most important external characters
with the Neotropical forms as well as with the Sumatran.
This fact of itself goes far to prove that Dr. Horst’s speci-
men really came from Sumatra.

At one time it was supposed that the various species of
Peripatide could be arranged in groups in accordance with
their geographical distribution (15). Mr. Sedgwick described

TWO NEW SPECIES OF ONYCHOPHORA. 475

three groups, namely, the Neotropical, the Australasian, and
the Ethiopian. Dr. Willey, in correspondence with this
nomenclature, states that the New Britain form may be
designated Melanesian (18). M. Bouvier has already shown
that such a grouping of the Peripatidz has no foundation
in fact (4). he discovery in the Malay peninsula of a
number of species which, as above mentioned, agree with
Neotropical forms in almost all their most important charac-
ters, will effectually dispose of the above supposition.

Conditions under which Peripatus was obtained
in the Malay Peninsula.—The material at my disposal
consisted of thirteen specimens, six males and seven females.
I obtained my first specimen on the 5th, and the second on
the 6th, of May, 1899. Wewere then camping on the side
of a mountain 3300 feet high, the position of our camp being
about 2250 feet above sea level. The mountain in question
is designated Bukit besar, in the native tongue, and is
situated on the boundary line between the States of Nawng-
chick and Jalor, a full day’s journey from the town of
Patani. The first specimen was brought me by a Malay, who
carried it between the prongs of a split stick. When I took
hold of the stick and began to handle the animal between
its prongs, it suddenly squirted a whitish, sticky slime to a
distance of fully eighteen inches. My hands were covered
with it, and though it did not stick to the body of the
animal itself, I only succeeded in getting it off with a consider-
able amount of difficulty. This shme, though fluid when
first squirted out, solidifies almost immediately. When in a
semi-solid state it can be drawn out in threads, much in the
same way as the juice of the rubber tree. he natives, when
they saw the slime being squirted, were much frightened.
This whitish secretion is considered to be poisonous, and
accounts for the manner in which they brought me the
animal,

We had, unfortunately, arranged to leave Bukit besar
next morning. It was, however, agreed among the members
of the expedition that I should spend the day on the moun-

4'76 RICHARD EVANS.

tain in search of Peripatus, and make the descent towards
the evening, so as to be ready to leave the neighbourhood on
the following day. I took with me the man who brought me
the first specimen, in the hope that he would be able to find
some more. ‘lhe day’s work resulted in the capture of one
more individual. The Malay who was with me searched all
day on the ground under leaves, but was unable to find any
more specimens; while I myself turned but a few leaves and
spent almost the whole time in splitting and chopping dead
tree trunks. Knowing, as I did, that the natives were ac-
quainted with Peripatus, and had a name for it (Ulat
chelawah), I considered, at the time, that the Malay’s persis-
tence in turning over leaves was probably significant of the
habits of the animal. Later, however, I learnt that this was
not so, for all my specimens were found in dead wood. The
one specimen which I obtained on the 6th of May was found
in a stump of a tree about six feet in height. The stump
in question was almost completely dry, and, situated in its
interior, at a height of four to five feet from the ground, I
came across Peripatus in its usual habitat for the first time.
It was ina torpid state, and endured a considerable amount
of handling before it realised—as it were—what was taking
place. However, it was not long before it began to squirt
with celerity and great force its slimy secretion from the
slime-glands, which open at the tips of the oral papille.

The remaining eleven specimens were obtained more than
three months later at Kuala Aring, in the State of Kelantan,
a distance of at least one hundred and twenty miles in a
southerly direction from Bukit besar. The first batch,
consisting of five specimens, was found on the 18th of August
ina dead tree, which for the most part was exceedingly damp ;
but the exact spot in which they were found was dry. They
were discovered by Mr. Skeat’s Javanese “boy,” who, to-
gether with the other Malay servants, had been carefully
instructed by me several months previously as to the import-
ance of Peripatus. This native happened at the time to be
out collecting with Mr. Laidlaw, who brought the specimens

TWO NEW SPECIES OF ONYCHOPHORA. 477

tome. On the following day the same man found an addi-
tional specimen in the very tree in which he had discovered
the other five on the previous day. On the same morning
another Malay servant came across five more specimens in
another fallen tree, which was almost completely dry and
rotten. On the day in question I myself was with the men
who discovered the specimens, and am able to testify as to
the nature of the places in which they were found. The
kind of place which they seem to prefer for hiding during
the daytime is the interior of dry rotten wood.

Preservation.—Of the thirteen specimens obtained five
were preserved in a mixture composed of ninety-two volumes
of saturated solution of corrosive sublimate in water, and
eight volumes of glacial acetic acid ; three specimens in abso-
lute alcohol; four specimens in Flemming’s weak solution ;
and one specimen in four to five per cent. formaline, that is, in
a mixture of one volume of the commercial formaldehyde with
seven of water. All the specimens save the one preserved in
formaline were cut by a longitudinal incision, either dorsally or
dorso-laterally, so as to preserve the internal organs. Unfor-
tunately, however, it happened—as if by the irony of fate—
that only male specimens were preserved in Flemming’s solu-
tion, all the female specimens having been preserved either
in absolute alcohol or in the corrosive and acetic mixture.
This was a great misfortune, for the material preserved in
Flemming’s fluid was almost perfect, while that treated with
the other solutions was far from reaching the same standard
of excellency.

II. THe CLAssIFICATION OF THE ONYCHOPHORA.

Before I proceed to describe the species here considered,
it seems necessary to give a historical account of the views
which have been put forward of the classification of the
Onychophora.

Mr. Sedgwick, as was mentioned above, pointed out that
the Onychophora fell into three groups, namely, the

4.78 RICHARD EVANS.

African, the Neotropical, and the Australian, though
he did not consider the differences in character between
these groups of sufficient importance to entitle them to the
rank of three distinct genera. Consequently all the species
then known were described by Mr. Sedgwick in his Mono-
graph under the genus Peripatus (15). Later, however,
the genus Peripatus, thus constituted, was broken up by Mr.
Pocock, and arranged in three genera, corresponding to Mr.
Sedgwick’s three groups. Mr. Pocock retained the name
Peripatus for the Neotropical forms, and gave the names
Peripatoides and Peripatopsis to the Australian and
African forms respectively.

In the year 1898 Dr. Willey published a complete descrip-
tion of a species which he had discovered in New Britain,
and gave it the name Paraperipatus Nove-Britannie.
Willey, however, considers the three genera formed by
Pocock, as well as Paraperipatus, worthy only of sub-
generic rank, and prefers along with Mr. Sedgwick to
include all the species in one genus. Prior to the publication
of Willey’s account of P. Nove-Britanniw, M. Bouvier
gave a short account of P. Tholloni, for which Dr. Willey
justly remarks a new sub-genus must be formed if the other
sub-generic names—as he describes them—are to be re-
tained ; in other words, he is of opinion that P. Tholloni,
from a classificatory point of view, is equivalent to any one
of the other four sub-genera, a view which is adopted in the
present paper, with the difference that Willey’s sub-genera
are accorded generic rank. Because the species Tholloni
is intermediate in character between the genus Peripatus
(Pocock) on the one hand, and the genus Peripatopsis
(Pocock) on the other hand, I propose to give it the generic
name Mesoperipatus.

In addition to the above-mentioned genera a new genus,
designated Opisthopatus, has been established by Purcell,
to include those African species which in some respects
approach the Australian forms (18).

In a recent number of this Journal, M. Bouvier published

TWO NEW SPECIES OF ONYCHOPHORA. A479

the results of a thorough and complete study of a great
number of species, and embodied the classificatory results at
the close of his paper in a tabular form, from which the
following is taken (8).

I. Peripatus (1) Genus
Ul (eae anal Pee (3) | eo

"LL Opisthopatus. He E() 2 7S;
III. Peripatopsis (4),
IV. Paraperipatus (a) 0 wat

In the above classification the genus Peripatus includes
the species Tholloni and Sumatranus. In the present
paper the former of these two species is placed in a genus
by itself, and is called by the generic name Mesoperi-
patus; the latter, together with the Malayan species, con-
stitutes a new genus, to which the name Eoperipatus will
be given. Bouvier has shown that M. Tholloni is inter-
mediate in character between the Neotropical and_ the
African species, and that it is more closely related to the
former than to the latter. From the description given in the
present paper it will be seen that the Malay forms are more
nearly akin to the Neotropical genus than to any other, and
that the genus Mesoperipatus is intermediate between the
Malay and African species. By dividing the genus Peri-
patus, as constituted by Bouvier, into three genera, the
differences between the Malayan, African, and Neotropical
forms contained in it will be emphasised. Bouvier has
already united together the genera Peripatoides and
Opisthopatus, this indicating the belief that they are
closely related. In the same way the close relationship
existing between the genera Peripatus, Mesoperipatus,
and Eoperipatus may be indicated by placing them to-
gether in one sub-family.

The following classification seems to be a fair expression
of our present knowledge both of the anatomy and develop-
ment of the Onychophora. It will undoubtedly be improved
upon as our knowledge of the class advances.

4.80 RICHARD EVANS.

Crass ONYCHOPHORA.

Family I. Peripatide.
Sub-family 1. Peripatinee.
Genus 1. Eoperipatus! (gen. nov.).
» 2. Peripatus (Pocock) (12).
» 9& Mesoperipatus (gen. nov.).
Sub-family 2. Peripatoidinee.
Genus 4. Peripatoides (Pocock) (12).
» 9. Opisthopatus (Purcell) (18).
Sub-family 3. Peripatopsine.
Genus 6. Peripatopsis (Pocock) (12).
Sub-family 4. Paraperipatine.
Genus 7. Paraperipatus (Willey) (17, 18).

The above classification follows the same lines as that of
Bouvier. In fact, it amounts to little else than the breaking
up of Bouvier’s genus Peripatus into three genera, and the
formation of a number of sub-families by the grouping to-
gether of the genera.

The feature which has been considered in the formation of
sub-families is the gradual degeneration of the last two pairs
of legs, and the position of the genital openings. These
characters formed the basis of Mr. Sedgwick’s division into
groups, and later of Mr. Pocock’s formation of three genera,
as well as of M. Bouvier’s four divisions.

I shall now proceed to give a formal definition of the
above four sub-families.

1. Peripatine:—The genital orifice is situated between
the penultimate pair of legs. There is a slight reduction in
size, but no actual degeneration in structure of the last two
pairs of legs, further than that they do not possess the full
number of pads.

1 Tam indebted to J. W. Jenkinson, Esq., M.A., of Exeter College, Ox-

ford, for the term Hoperipatus, as expressive of the distribution of the
genus ip the Hast.

TWO NEW SPECIES OF ONYCHOPHORA. A481

2. Peripatoidinze:—The genital orifice is situated be-
tween the last pair of legs which are not reduced in size.
The hindermost pair, found in the Peripatinz, is obsolete.

3. Peripatopsine:—The genital orifice is situated be-
tween the last pair of legs (anal papille), which are much
reduced in size and degenerate in structure. According to
Purcell the anal papillae may be obsolete (18).

4, Paraperipatine:—The genital orifice is situated
behind the last pair of legs. ‘The vestigial pair found im the
Peripatopsine has completely disappeared, and the last
existing pair, which is pregenital in position, is reduced in
size.

Now that the sub-families have been defined, it seems
necessary to sum up the main characters of the genera. It
is impossible to form concise definitions of them, because in
several cases some of the more important characters are
common to more than one genus—a fact which points to the
desirability of arranging the genera on a wider plan than
the one which has been adopted, and of giving the rank of
families to the groups here described as sub-families. It is
highly probable that future investigation and more extensive
knowledge of the class Onychophora will tend towards the
adoption of the course above suggested. However, it may
be useful to give a summary of the leading features of the
several genera here recognised.

Genus 1. Eoperipatus :—The legs have either four or
five pads. The renal papille of the fourth and fifth pairs of
legs are situated either in the middle of the proximal pad,
that is, the fourth pad, or on its proximal side. The legs in
question have only four pads. ‘he feet are provided with
two papille situated on their distal margin, one in front and
one behind. ‘here are four ventral prominences on the feet,
which are provided with one or more spines, similar to those
found on the erect distal papillae, and pointing downwards.
The outer blade of the jaw has two accessory teeth on the
mner side of the main tooth. The inner blade may have
three accessory teeth on the inner side of the main tooth.

482 RICHARD EVANS.

The inner blade has a diastema followed by a row of small
denticles. The male genital pore has the shape of a cross.
The ductus ejaculatorius is very long and forms a loop. The
male accessory glands open to the exterior by a median pore
situated between the last pair of legs. The female opening
is a transverse slit. Receptacula ovorum and seminis are
present. The ovary is large, and spreads out over the other
organs. The ova are large, full of yolk, and exogenous. The
developing embryos have no kind of placental structure, and
measure 20—27 mm. in length at birth.

Genus 2. Peripatus:—The legs have either four or five
pads. The renal papille of the fourth and fifth pairs of legs
are situated either on the proximal side of the third pad or in
the groove between the third and fourth pads. The feet
have a variable number of papille ; the hinder margin having
either one or two, while the anterior margin has from one to
three. ‘The outer blade of the jaw has one or two accessory
teeth, the inner blade has the same number followed by a
diastema and a row of small denticles. The ductus ejacula-
torius forms a very long loop. ‘The male accessory glands
(anal glands) open beside the anus into two pouches, which
have the same structure as the rectum. Receptacula seminis
and ovorum are present. The ovary is small and compact.
The ova are endogenous and devoid of yolk. The young are
provided with a nutritive organ which has been called the
“placenta,” and measures 20—25 mm. in length at birth.

Genus 8. Mesoperipatus :—The legs have only three
pads. ‘The renal papille of the fourth and fifth pairs of legs
make a slight indentation in the proximal pad, that is, the
third pad. The number of legs varies from twenty-two to
twenty-five pairs. The two blades of each jaw have an
accessory tooth on the inner side of the main tooth, and the
inner blade has a long row of denticles. The ridges of the
skin are continuous across the back. ‘The female has a pair
of receptacula seminis ; presumably there are no receptacula
ovorum.

Genus 4. Peripatoides:—The legs have three pads.

TWO NEW SPECIES OF ONYCHOPHORA. 4838

The renal papillae of the fourth and fifth pairs of legs are
continuous with the proximal pad (the third), but do not
divide it in two. The feet carry three papille near the edge,
one in front, one behind, and one dorsally. The outer blade
of the jaw has lost its accessory denticle, the inner blade has
no diastema and the denticles are few. The ductus ejacula-
torius is almost as long as in Hoperipatus. Male accessory
glands open to the exterior by one or two openings situated
between the genital pore and the anus. When there are two
they are widely separated. The genital opening (female) is
round, There is a pair of receptacula seminis, but no recep-
tacula ovcrum. The ova are large, full of yolk, and exo-
genous. Development may take place in the uteri, or ova are
discharged. When development takes place in the uteri the
young at birth measure 5 mm.

Genus 5. Opisthopatus :—The legs have three pads.
The renal papille of the fourth and fifth pairs of legs are
situated in the proximal pad. The feet carry three papille,
one in front, one behind, and one dorsally. The outer blade
of the jaw has a small denticle inside the bigger one; the
inner blade has no diastema, but a row of small denticles.
Male unknown. The female genital opening is a transverse
slit. No receptacula seminis or ovorum. The young at
birth are large in size. They are of different ages in the
uteri.

Genus 6. Peripatopsis:—The legs have three pads, and
the excretory papille of the fourth and fifth pairs of legs are
situated in the proximal pad. Feet have three papillae near
the apex, two on the anterior side, and one on the posterior
side. "he outer blade of the jaws has a small accessory tooth
at the base of the main tooth; the inner blade has no
diastema, but has a row of small denticles. The ductus
ejaculatorius is shorter than in genera 1, 2, and 4. The male
accessory glands open into the ductus ejaculatorius. The
female genital pore is a longitudinal sht. There are no
receptacula ovorum or seminis. ‘The ova are large, devoid of
yolk, and arise exogenously. The young at birth are of

484, RICHARD EVANS.

medium size. All the embryos in the uteri are of the same
age.

Genus 7. Paraperipatus:—The legs have three pads.
The renal papille of the fourth and fifth pairs (sometimes the
sixth as well) are situated in the proximal pad, completely
dividing it into two. The feet carry three papille, one on the
anterior side, one on the posterior side, and one variable in
position. The outer blade of the jaws is simple, without a
small accessory tooth at the base. The inner blade has no
diastema, but has a row of small denticles. ‘he male
genital pore is at the tip of a conical papilla. The ductus
ejaculatorius is short and median. The male accessory
glands (“ pygidial glands’’) debouch into a bulbus, which
opens to the exterior by a median pore above the anus.
There are no receptacula ovorum. There are two recep-
tacula seminis. The ova are small, devoid of yolk, and arise
exogenously. The young at birth are of medium size.
During the early uterine development they are provided with
a trophic sac, which becomes completely enclosed in the later

stages.

III. Some Nores on tHE ‘'l'ypE SPECIMEN OF
K. SUMATRANUS.

In order to compare the two species here described with
E. sumatranus (Horst), at the end of the summer I visited
the Leyden Museum. There I met with every courtesy on
the part of Dr. Horst, who first described the species in
question (8), and under whose care the type specimen is at
present. He kindly gave me every facility in its examination.

The following notes made at the Leyden Museum confirm,
add to, and to some extent correct Dr. Horst’s description.

The colour of the dorsal surface of H. sumatranus is
dark brown, while that of the ventral surface is considerably
lighter. ‘Che white spots found on the dorsal surface are due
almost entirely to the broken condition of the large primary
papille. ‘lhe ventral surface completely lacks the pink

TWO NEW SPECIES OF ONYCHOPHORA. 485

found in the species Horsti. There is a dark line along the
middle of the back, as in the Malay forms. ‘lhe line in
question is due to the greater development of pigment in the
papillz situated on either side of the non-pigmented but
narrow line which occupies the mid-dorsal position, as in the
Malay species. ‘The segmentally arranged areas situated on
the ventral surface in the species from the peninsula are
scarcely visible in the Sumatran form, although, having pre-
viously seen them in the former, I was able to find traces of
them in the latter.

The mouth, genital orifice, and the anus are situated in the
same position, and have the same structure as in H. Horsti.
Between the genital orifice and the anus, i.e. between the
last pair of legs, is found the common opening of the male
accessory glands. ‘This opening was mistaken by Dr. Horst
for the anus, which is a small longitudinal slit situated very
shghtly subterminal.

The antennz have the same shape as in the Malay species,
and very nearly the same number of rings, although some of
these are small and consequently very difficult to count. It
seems that there are in all about fifty or fifty-one rings.

The disposition of the legs, of which there are twenty-four
pairs, is practically the same as in the species Horsti. The
distance between the successive pairs in the anterior part of the
body is not much greater than that between similarly situated
pairs in the posterior part. There are four crescentic pads
on every leg of the first twenty-two pairs, the penultimate
pair having only three, and the last only two on each leg.
The crural grooves are all closed, but are easily made out on
all the legs except the last two pairs, though they are smaller
on the three anterior pairs. On the last two pairs of legs
in the species sumatranus there is no sign of the papilla,
which are supposed to represent the whitish lips of the
reduced crural grooves.

The feet in EK. sumatranus have the two papille found
in Weldoni and Horsti. The two distal papillw, one in
front and one behind, are divided into two parts, the top

486 RICHARD EVANS.

part carrying the usual spine. The spines were not observed
on the four ventral prominences; they were probably rubbed
off.

The opening of the renal organ of the fourth and fifth
pairs of legs is situated on the top of a papilla, which com-
pletely divides the proximal pad in two. In this feature
Sumatranus agrees with the species Weldoni, and not with
Horsti. ‘I'he structure of the skin, however, has no resem-
blance to that of Weldoni, but is very similar to that of
Horsti. The papillz are arranged on the transverse ridges,
and are scarcely ever found in the grooves between them.
The transverse ridges rise up suddenly, and the grooves are
well marked. Owing to the broken condition of the skin
papillae it is almost impossible to distinguish in all cases
between primary and accessory papille, save in so far as this
can be done by noticing the difference in situation and size.
The primary papille consist of a truncated base, with a
cylindrical part on top of it, the latter carrying a spine.
The primary papille occupy the whole width of the ridge on
which they are situated, while the accessory ones are placed
more or less laterally, and are less numerous than in the
species Horsti, to be described in the present memoir.

IV. Descriprion or THE wo SpEctEs,

Korerrpatus Wernponrt AND EKorrripatus Horstt.

A. Introduction.

Of the thirteen specimens at my disposal only two of them
—both females—belong to the species Weldoni; the remain-
ing eleven belong to the species Horsti, and consist of six
males and five females. ‘The former are the specimens
obtained on Bukit besar, the latter are those obtained at
Kuala Aring in Kelantan. The proportion between the
number of males and females of the species Horsti is rather
a commentary on the statement generally made, that the
females are more numerous than the males. In order to

TWO NEW SPECIES OF ONYCHOPHORA. A487

avoid repetition the two species will be described together
as far as possible, and the features in which they differ will
be pointed out as they arise.

B. External Characters.

Colour.—The colour of the two species here described is
shehtly different, especially on the ventral surface. ‘The
dorsal surface of the species Weldoni is coloured dark
brown, while that of the species Horsti is lighter; that is,
it is warmer in tone, as if there were a certain amount of
pink mixed with the brown. The mid-dorsal position in
both species is occupied by a very dark chocolate-coloured
line, of a much deeper hue than the neighbouring parts. It
extends uninterruptedly from close to the anterior end
almost to the terminal anus. The appearance of a dark line
is brought about by the greater development of pigment in
the papillz situated on either side of a completely non-pig-
mented but exceedingly narrow line which occupies the
median position. ‘The dorsal surface, especially im the species
W eldoni, has pale spots, distributed with a certain amount of
regularity all over it. ‘Chis appearance results, on the one hand,
from broken primary papilla, and, on the other hand, from
the primary papillz containing less pigment than the others,
and consequently presenting a yellowish-white appearance.

The ventral surface in the species weldoni is coloured
yellowish-grey with small spots of brown. In the species
Horsti it may be almost as in Weldoni, or it may have a
decidedly pink appearance, which in some cases may even
push between the legs towards the dorsal surface. In both
species there is a row of whitish areas, segmentally arranged
along the mid-ventral line between the legs of each pair.
They correspond to, and lie below, the ventral organs, which
in the genus Hoperipatus do not completely degenerate in
the adult condition,

Dimensions.—The two species here described differ con-
siderably in size, Weldoni being much larger than Horsti.,

488 RICHARD EVANS.

The average length of the specimens belonging to the species
Weldoni is 58 mm., and average breadth is 6°25 mm., while
the average length of the female specimens of Horsti is
only 35:4 mm., and average breadth is 4°6 mm.; and of the
male'specimens the average length is 33°5 mm., and average
breadth 3°8 mm. If the eleven specimens of Horsti be
taken together, the average length is 34 mm. and average
breadth is 4:16 mm. ‘lherefore, in whichever way the com-
parison is made, the size of Horsti is to that of Weldoni
approximately as three is to four. Another fact brought out
by the above measurements is that the females of Horsti are
both longer and broader than the males. However, these
measurements must not be taken as the maximuin length
and breadth of the species Horsti; for one female in my
collection measures 46 mm. in length and 5:5 mm. in
breadth, and one male 40 mm. in length and 4°5 mm. in
breadth.

The last fact mentioned seems to dispose of the specific
difference in size, but such is not really the case, for the
larger specimen of Weldoni is 65 mm. long and 75 mm.
broad, and therefore exceeds the largest specimen of Horsti
in length by 19 mm., and in breadth by 2 mm.

The Skin.—The skin is thrown into folds, of which there
are about twelve between each successive pair of feet in the
middle part of the body. An examination of figs. 10 and
13 will serve better than the best possible description to em-
phasise the most characteristic difference between the two
species here described. When the skin is examined with a
hand lens the folds appear continuous across the back, but
when looked at through the microscope a narrow, non-pig-
mented line is revealed. On these folds are found the papilla
which are one of the constant features of the Peripatide.
In the Malay forms, as well as in the Sumatran species, they
resemble in structure the papille of the Neotropical forms.
Among the dorsal papillae are found both primary and acces-
sory ones, the former consisting of a basal portion which
varies in shape between a cylinder and a cone (figs. 7 and 8),

TWO NEW SPECIES OF ONYCHOPHORA., 4.89

and an apical part which may be either conical, cylindrical,
or spherical in form, and carries on its top a pointed spine ;
the latter consisting merely of a conical elevation of the skin.
The basal portion of the primary papille almost invariably
stretches across the folds of the skin from one groove to the
other, while that of the accessory papille only occupies a
portion of that area. The scales which cover the basal por-
tion of the primary papille, and correspond to the epidermal
cells of the skin, are shorter and broader than those which
cover the apical part.

In the species Weldoni all the papillae have a somewhat
polygonal and distinct basal outline, and the consecutive folds
of the skin come so close together as almost to obliterate
the grooves between them (fig. 10); while in the species
Horsti the papille have a round and indistinct basal outline,
and the grooves between the folds are at least half as wide
as the folds themselves (fig. 13). In Weldoni it is not
unusual to find secondary papillz situated in the grooves;
while in Horsti they scarcely ever occupy that position; in
the former the grooves are narrow and shallow, and the folds
rise up gradually ; in the latter the grooves are broad and
deep, and the folds rise up suddenly.

The Median External Openings :—In the female there
are only three median openings, while in the male there are
four. The latter has an opening which does not occur in the
female, and which is situated between the last pair of legs.
It was described by Horst as the anus, which he said was
subterminal (8).

The Mouth:—The mouth is surrounded by a ring of
yellowish-white papille of large size. Inside the ring, and
in front of the actual mouth-opening, are situated four pairs
of large papillae, which become smaller in size from in front
backwards. Between these internal papille the so-called
tongue appears, and carries a number of complex denticles
(fig. 4 and fig. 14).

The Anus:—The anus is a small, longitudinal slit-like
opening, situated at the terminal end of the animal. In

VOL. 44, PARY 4.—NEW SERIES, 11

4.90 RIGHARD EVANS.

describing the species sumatranus Horst missed it, and
described the above-mentioned opening found in the male as
the anus (fig. 5 aud figs. 15 and 16).

The Genital Orifice:—The genital opening in the
female is situated between the penultimate pair of legs
(fig. 5 and fig. 15), and has the form of a deep, transverse
sht, surrounded by tumid lips, which consist of whitish
papille similar in characters to those which surround the
mouth.

The genital opening of the male of the species Horsti is
cross-shaped, the cross lines lying in the longitudinal and
transverse axes of the animal. The lobes situated in the
angles of the cross are made up of numerous papille, which
are fused together to form four masses, on the top of which
occurs a small circular pit. ‘The position of the male genital
orifice is between the penultimate pair of legs, as in the
female. ‘he characters of the genital openings constitute
an unfailing external difference between the male and female
of the species Horsti (fig. 16).

The Accessory Pore:—The opening of the male acces-
sory glands is situated between the last pair of legs. It is
triangular in shape, but its sides are slightly concave to-
wards the exterior. Papille similar to those which sur-
round the male genital orifice form its boundary. — Its
presence absolutely distinguishes the male from the female
(fig. 16).

The Number of Appendages:—The two specimens
belonging to the species Weldoni have,twenty-four pairs of
legs. The number of legs possessed by the young taken
from the uteri of these two specimens varies between twenty-
three and twenty-five pairs.

Of the eleven specimens belonging to the species Horsti
two females have twenty-five pairs; three females and one
male have twenty-four pairs, and five males have twenty-
three pairs. Out of five females two have twenty-five pairs,
and three twenty-four pairs ; while ont of the six males only
one has as many as twenty-four pairs, the remaining’ five

TWO NEW SPECIES OF ONYCHOPHORA 4.9 |

possessing only twenty-three pairs. As a rule, the male of
the species Horsti has only twenty-three pairs of legs, while
the female has at least twenty-four, and as often as not it has
twenty-five pairs. The net result of the above facts is that
the male has fewer legs than the female by at least one pair.
Another fact brought out by this small collection is the vari-
ability im number of legs possessed by both male and female
—a feature which is more characteristic of the Peripatine
than of any other sub-family.

Antennez:—The antenne taper gradually as far as the
thirty-fourth or thirty-fifth ring, but from that position on-
wards they increase in circumference so as to become club-
shaped. All the rings, as far as the position above men-
tioned, have the same thickness ; but a considerable number
of the last fifteen rings may be either thicker or thinner than
the rest. ‘The presence of the thin ones makes it almost im-
possible to count the number of rings which constitute the
antennee, which are scarcely ever fully extended. On this
account the number that can be counted varies between forty-
six and fifty, or perhaps in some instances fifty-one.

The Jaws:—The jaws are different from those of any
species that has been so far described. he outer blade has
two denticles on the imner side of the main tooth. Asa rule,
in the genera Peripatus and Peripatopsis there is only
one denticle ; while in Paraperipatus and Peripatoides
there are none in this position, Similarly the inner blade has
two denticles on the inner side of the main tooth, which are
followed by a diastema, on the inner side of which is found a
nine or ten in number. There

row of smaller denticles
seems to be no essential difference between the jaws of the
two species Weldoni and Horsti. ‘l'wo vestigial denticles,
however, were noticed at the inner moiety of the diastema of
the inner blade in one specimen of the species Horsti
(fig. 12a). They are not constant, for the inner blade of the
other jaw of the same specimen did not possess them.
Nevertheless, they are interesting in that they point to
the series being at one time continuous, much as it is at

4.92 RICHARD EVANS.

present in P. capensis, in which the reduction in number
seems to have taken place at the inner end of the series and
not in its middle, as in the Peripatinee (figs. lla and 11);
fies. 12a and 125).

The Oral Papille:
sides of the mouth, consist of two rings, which do not carry
skin papilla, and of a knob-like end-piece which is provided
with skin papilla, chiefly on the dorsal aspect. The opening
of the slime-gland is slightly sub-terminal, and is surrounded
by four large papille, similar in character to those forming
the ring round the mouth. ‘The oral papilla seem to be
extremely contractile, and are scarcely ever fully extended
in preserved specimens (fig. 4).

The Legs:—The legs, which vary in number from twenty-

The oral papille, situated at the

three to twenty-five pairs, are short and stumpy. ‘There is a
marked difference between the arrangement of the legs in
the two species. In the species Weldoni they are crowded
towards the posterior end, the distance between the successive
pairs in the anterior moiety being much larger than in the
posterior one. In the species Horsti they are almost evenly
distributed along the whole length of the body, except in the
region of the last two or three pairs.

The Leg-pads:—With the exception of those of the last
two pairs, every leg carries four pads. The last pair has
only two on each leg, and it often happens that they are in
no way well marked from each other. The penultimate pair
has only three, of which the proximal one is often a mere
vestige. Similarly the fourth pad of the antepenultimate
pair of legs may be fully developed.

The Crural Grooves :—Crural grooves may occur on all
the legs, but on the first and last two pairs they are very
feebly developed, and there is often no trace of them. The
grooves on the second and third pairs of legs are much
smaller than on those further back. They extend from the
third row of papille, counted from the proximal pad of the
leg, to a considerable distance on the ventral surface of the

body.

TWO NEW SPECIES OF ONYCHOPHORA. 493

In connection with the crural grooves there are whitish
structures of variable size and shape, according to the part
of the body to which the leg belongs and according to the
species considered.

In the male of the species Horsti there is a whitish
papilla of round shape on every leg of the last two pairs,
where there is hardly a trace of the crural grooves. On the
legs of the genital segment the crural groove is surrounded
by a thin and irregularly-shaped fold, white in colour. It is
specially thickened round the distal angle of the groove.
The same arrangement occurs on seven or eight pairs of
legs in front of the genital segment (fig. 16), but as we
pass forwards the proximal ends of the folds become less
marked. On the next six or seven pairs the white folds are
reduced to two papilla-like structures, situated close to-
gether at the distal angle of the crural grooves, and some-
times fused together to form an U-shaped body. There are no
signs of white papillee or folds on the anterior five pairs of legs.

In one female of the species Horsti whitish sucker-like
structures were observed on all the legs, with the exception
of the five anterior pairs. On the last two pairs of legs they
were smaller in size than on those in front. These structures
in the female present an appearance which is remarkably like
that figured in Mr. Sedgwick’s monograph, and described as
occurring in P. Edwarsdii (compare PI. 28, fig. 12 [15]). In
another female of the same species the legs were so contracted
that the structures described above could not be seen, the
crural grooves appearing as narrow slits. Whether in the
male or in the female the renal pores are situated proximally
to the above structures, even in those legs of the female
where they are only partially developed.

In the female of the species Weldoni the sucker-like
structures, described above in the female Horsti, do not
occur. On the last two pairs of legs there is a round papilla
similar to that found on the male of the species Horsti. On
the legs of the genital segment, as well as on all those
situated in front, with the exception of those of the first pair,

494 RICHARD EVANS.

there is an U-shaped papilla, which surrounds the outer
angle of the crural groove. Such a structure was not found
on all of them, but it was observed on the second, the fourth,
the seventh, and other pairs of legs. Its occurrence on some
of the legs of the anterior five pairs contrasts with the con-
dition found in Horsti, where the whitish papille do not
appear on the legs situated in front of the sixth pair, which
is supplied with them in both male and female. In the ante-
rior half of the female Weldoni the U-shaped bodies are
often divided into two papille, situated close together and
flattened against each other—a condition found to occur on
the sixth to the twelfth pair of legs in the male Horsti.

The Renal Apertures:—The openings of the renal
organs are situated in the crural grooves at the junction of
the legs with the body, except in the fourth and fifth pairs,
in which they are removed to the vicinity of the proximal
pad of the leg.

‘be actual position of the renal papillz of the fourth and
fifth pairs of legs differs in the two species here described.
In the species Weldoni they are situated in the proximal
pad, and therefore divide it into two (fig. 9); but im the
species Horsti they are situated on the proximal side of
the pad (fig. 21). The position of these openings in H.
Sumatranus agrees with that of the species Weldoni,
and not with that of Horsti.

The position of the renal papille of the fourth and fifth
pairs of legs in Hoperipatus is different from that found
in any other genus of the Onychophora. It is more primi-
tive than even that found in the genus Peripatus, which is
its closest ally.

Renal pores do not occur on the penultimate or on the last
pair of legs in the males. ‘They are found, however, on the
last pair of legs in the female of the species Horsti, in which
there is a large renal organ in the segment behind the genital
one. ‘lhey were not seen in the adult female of the species
Weldoni, in which the outer end of the renal duct of the
penultimate segment seems to disappcar in the adult.

{WG NEW SPECIES OF ONYCHOPHORA. 495

The Openings the Crural Glands :—Crural glands
only occur in the males, and open into the crural grooves, so
close to the aperture of the renal organs that they may be
said to debouch into a common pit in the groove. They
occur only in the two pairs of legs situated in front of the
genital orifice, and are placed on the outer side of the renal
pores. ‘There are two openings, corresponding to the couple
of glands found in every leg of the two pairs above men-
tioned. It is just possible that, were a sufficient number of
specimens examined, they would be found still further for-
wards, as is the case in P. Edwardsii (7).

The Feet:—The feet present the same general structure
as in all the Peripatidw. They are provided with a terminal
pair of sickle-shaped claws and a number of papille. Dr.
Horst described them in KH. sumatranus as being provided
with only two papille, and as having their ventral surface
divided by a longitudinal and a transverse groove into four
elevations (8). Mr. Sedgwick pointed out that, if Horst’s
description was correct, H. sumatranus was unique; for in
all other species of Peripatides the foot was provided with at
least three papillee (15).

Renewed examination of KE. Sumatranus has proved the
correctness of the above description as far as it goes. It is
equally applicable to the feet of the species from the Malay
Peninsula, which have only two primary papille situated one
on the anterior and one on the posterior distal margin. The
four elevations on the ventral surface seem to be four papillee
which are but slightly raised above the general surface, and
are pressed against the sides and ventral surface of the foot.
The elevations in question carry at least one spine which was
not mentioned by Horst in the species Sumatranus, but has
been described by Bouvier in some American forms (P.
Geayi, Bouvier). Asa rule the distance between the erect
papillz and the distal one of the above elevations equals that
between the two ventral elevations on each side of the foot ;
but occasionally the distal elevation is displaced, and be-
comes located near the erect papillae which are situated on

496 RICHARD EVANS.

the distal margin. This displacement only takes place on
either the anterior or posterior aspect of the foot, or on both
together. In such cases the foot presents an appearance
which has a certain resemblance to the condition found in
other Peripatidz where it is provided with more than two
papillee.

c. Internal Anatomy.

(a) Introductory Remarks:—The internal anatomy of
all the Peripatide presents a great degree of resemblance ;
consequently it is not necessary to describe at length the
whole anatomy of the species here considered; but as they,
together with KH. sumatranus, constitute a new genus, it is
advisable to describe some systems of organs at greater
length than would be otherwise necessary.

The muscular and vascular systems call for no special
remark. The nervous system resembles that of other
genera to such a degree that there is no need to describe it
further than to mention one or two points which seem to
demand attention. One of the points in question is the
difference between the male and the female nervous systems
immediately in front of the genital orifice. In the male the
cords are provided with specially enlarged swellings or gan-
glia in the above-mentioned position (fig. 35), while in the
female there is no such swelling beyond that which occurs
opposite each pair of legs (fig. 33). The second point
is the well-developed condition, even in the adult, of the
strands of nervous matter which pass from the lateral nerve-
cords to the ventral organs. These strands contrast strongly
with the nerves given off from the nerve-cords to the body-
wall and the legs, in that they consist of cells in the more or
less undifferentiated state in which they are found in the
lateral nerve-cords (fig. 34).

(6) Ventral Organs.—As the ventral organs have been
mentioned already, it is advisable to describe them at the
present juncture. ‘hey occur between each pair of legs in
the mid-ventral line, and correspond to the segmentally

TWO NEW SPECIES OF ONYCHOPHORA. 497

arranged spots mentioned in describing the external cha-
racters. Even in the adult they are not as degenerate as
usually represented in other genera. In section they are
seen to consist of a group of long cells with oval nuclei
situated near their internal end. In a median section the
cells seem to be arranged fan-wise round what appears to be
a cavity which apparently communicates with the exterior.
If this is not the case, the cuticle covering the skin clearly
dips down into the space situated inside the group of cells
constituting the ventral organ.

(c) The Salivary Glands :—The salivary glands present
the same general arrangement as in the other genera of the
Peripatidee.

Their coelomic end-sacs are of enormous size, and spread
themselves, chiefly in a dorso-ventral direction, under the
lateral longitudinal muscles. Their walls are thick, and
consist of cells which present no definite cell outlines, and
which are supplied with large nuclei and highly vacuolated
cytoplasm.

From the postero-dorsal corner of the coelomic end-sacs
a short duct passes and opens into the tubular part of the
slime-gland on its dorsal aspect, on a level with the first pair
of legs. The tubular parts of the glands lie in the lateral
body-cavity immediately above the nerve-cords. ‘Towards
their posterior end they are circular in transverse section,
but anteriorly they are triangular, the apex of the triangle
being wedged in the upper angle of the lateral body-cavity. As
a rule their lining cells are tall and columnar, but in places
they are short, especially on the lower side of the triangular
section and towards the anterior end of the glands. They
open into the anterior outer corner of two rather large sacs
which are situated, one under each of the nerve-cords.
hese sacs have not been seen in any other genus. They
certainly do not exist in P, capensis, or else Balfour and
Sedgwick would have seen them in sections, for in Koperi-
patus they are the most prominent feature of a section
passing through that region. From the inner anterior corner

498 RICHARD KVANS.

of each sac a short duct passes into the posterior outer corner
of the median diverticulum of the buccal cavity.

Shortly before the tubular glands pass into the above-
mentioned sacs there is a sudden, well-marked change in the
characters of the lining cells. The lining of the glandular
part in this region consists of short columnar cells with no
well-defined cell outlines, and with nuclei in the centre, while
the lining of the portion which enters the sacs consists of tall
columnar cells with large nuclei situated at their free ends
and with well-defined cell outlines. In both respects the
lining of the sacs resembles the latter, which is a proof
that they are mere diverticula of the ectodermal ducts
(fig. 32). The lining of the backward diverticulum of the
mouth, that is, the so-called common duct of the salivary
glands, consists of much shorter cells with nuclei in the
centre, but the linings of the two regions pass gradually into
each other.

The sharp distinction between the lining cells shown in
figure 32 probably marks the external limit of the imeso-
dermal portion of the gland.

Mr. Sedgwick describes the elongated glandular portion
as being produced by the backward extension of the duct,
and since the duct is mesodermal in origin the gland must
be so too (14). If we judge from the histology of the parts
in question, there seems to be no doubt but that the tubular
gland, which runs along at least two thirds of the length of
the body, is mesodermal. Kennel would probably consider
it ectodermal, but the histology of the salivary glands, the
renal organs, the genital organs, and finally of the male
accessory glands, all of which appear to be homologous
organs, tends to disprove Kennel’s view of the nature of the
renal and genital ducts, as well as of the salivary duct, from
which the salivary gland is produced by a backward exten-
sion.

(dq) I'he Renal Organs :—Renal organs occur im all the -
leg-bearing segments except the genital one. However, a
complete duct was not found in the last segment provided

{WO NEW SPECIES OF ONYCHOPHORA. 4.99

with legs in either the adult female of the species Weldoni,
or the adult male of the species Horsti; though in the
female of the latter species the duct was very highly de-
veloped (fig. 34).

A typical renal organ is usually said to consist of four
parts: namely, the coelomic end-sac ; the thickened funnel ;
the coiled tube; and the bladder. It appears, however, that
there is a fifth part which is, most probably, as constant as
any other part, without making an exception even of the
ecelomic end-sac, which is always present. The part in ques-
tion is the ectodermal duct which puts the ccelomic portion
of the organ in communication with the exterior. It varies
in length according to the position of the renal organ ; im
that of the ninth segment, which is represented in fig. 26,
this ectodermal duct is short ; but in the one from the fourth
leg-bearing segment which is shown im fig. 25 it is much
longer. Of the other parts, the bladder, the funnel, or
the coiled duct may be absent. In the renal organ of the
fourth and fifth leg-bearing segments the bladder is wanting,
while in those placed in the first, second, third, and the two
pregenital segments, there is neither a differentiated funnel
nor a bladder, and the coiled tube is represented only by a
short, straight duct.

A typical renal organ from the ninth leg-bearing segment,
with the five parts above mentioned, is shown in fig. 26.
The short duct situated externally to the bladder in the
above-mentioned figure is a well-marked structure, and its
distinctive histological characters are clearly seen in fig. 31.
The lining cells of this ectodermal duct resemble exactly
those which cover the external body wall, while the lining
cells of the bladder approach much nearer to those of the
coiled tube. ‘The bladder, it would appear, is nothing more
than a dilated portion of the outer end of the coiled tube.
In the renal organ, shown in fig. 25, which has no dilated
bladder, the short duct represented in fig. 26 appears as a
much elongated tube which stretches from the level of the
nerve-cord to the renal papilla situated on the proximal side

500 RICHARD EVANS.

of the fourth pad. Fig 24 represents the renal organ of the
second pair of legs. ‘lhe only parts represented here are
the coelomic end-sac, and an undifferentiated short duct
which passes from it to the exterior. Figs. 27 and 28 show
a similar reduction of the renal organs of the two pairs of
legs situated in front of the genital orifice. In both cases
they consist of a small coelomic end-sac and a short narrow
duct which passes to the exterior. The examples given
above suffice to show that there is a considerable amount of
difference in the structure of the renal organs in the various
parts of the body. The most prominent feature is the sim-
plification of structure towards either end of the animal.

The coelomic end-sac presents in section an irregular and
collapsed appearance, but is easily demonstrated in the adult.
In the middle part of the body it extends through several
sections, but 1s much smaller towards either end of the
animal. Its walls are thin and seem to consist of two layers
of cells, which are flattened out though not to any great
extent (figs. 29 and 30).

‘The funnel has a rim which projects into the cavity of the
coelomic end-sac, and is never provided with cilia. Its walls
consist of an internal lining of closely packed columnar cells
with deeply staining nuclei, and an external layer of flattened
cells (figs. 29 and 30).

The coiled tube and the bladder have: practically the same
structure as the coelomic end-sac; the lining cells of the
former, however, are less flattened than those of the end-sac,
while those of the bladder are more so, and strongly contrast
with those lining the short duct which passes from the
bladder to the exterior (fig. 31).

(ec) The Female Reproductive Organs :—When an
adult female is opened the first structures visible are either
the stomach or the coils of the uteri, and the ramifications of
the slime-glands, which extend backwards as far as the ovary.
In some cases the coils of the uteri are above the stomach
and completely conceal it, but in others they are below it and
are hidden by it. In either case the branches of the slime-

TWO NEW SPECIES OF ONYCHOPHORA. 501

glands intertwine with the uterine coils. The ramifications
of the slime-glands seem to vary considerably in size, espe-
cially in transverse section—a variation which is probably due
to the condition of the glandular secreting cells which line
the interior, and to the amount of secreted matter in the
lumen of the glands. The length of the lining cells of the
sline-glands is variable, sometimes long, sometimes short,
but always more or less columnar. Their nuclei are large,
granular, and clear, and on this account a section of one of
the finer branches of the slime-glands is distinguished with
ease from other small tubes, such as the finer coils of the
vasa deferentia. he clot of slime, also, which they usually
contain, helps to distinguish them (fig. 41, s. g.).

The Ovary :—The ovaries in the two species here con-
sidered resemble each other in their most essential features
to such an extent that there is no need to describe them
separately. In both cases the ovary is situated dorsally in
the region of the third and fourth pair from the posterior
end of the body. Itis found under the floor of the pericardium,
and is attached toit not by asingle ligament, but by an exten-
sive surface, thus differing from all the generaas yet described.
This feature of itself makes it almost impossible to dissect it
out. Not only is it attached to the pericardium, but it spreads
out over the rectum and uteri like a saddle, and pushes itself
into any space that may be unoccupied, both between as well
as outside the uteri. It becomes closely adherent to the uterine
walls on the one hand, and to the peritoneal lining of the
body-cavity (hemoccele) on the other hand. Thus the fusion
of the ovary wall with two or three other structures makes it
almost impossible to remove it in an unbroken condition.

In the adult the ovary consists of a shapeless sac with an
immense cavity which presents no sign whatever of its double
origin. Its walls are folded and carry follicular outgrowths
which are suspended in the body-cavity, and contain ova in
various stages of development. The ovarian cavity com-
municates by means of a large, irregularly shaped opening
with the oviducts. ‘The opening in question is situated, as a

502 RICHARD EVANS.

rule, at the posterior end of the ovary, the oviducts! passing
forwards from it.

No essential difference could be made out between the
oviduct and the proximal part of the uteri by looking at a
complete preparation; but when sections were examined a
most marked difference was immediately noticed ; but there
seems to be no sharp line of demarcation between tle two
parts (compare fig. 50, ovid., with fig. 51, wt.).

It is difficult to make out that the cells lining the oviducts
are columnar, because the cell outlines are not easily seen,
and the nuclei may take up any position whatever in the
cells; that is, they may be found at the base, in the middle,
or at the free end of the cell (fig. 50). Another feature
of the cell-lining of the oviduct is the sharp and well-
defined limit which the cells present towards the lumen
of the duct, which is in no way narrower than that of the
proximal portions of the uteri.

The cells which line the uteri are quite different i charac-
ter from those which line the oviducts. In those portions
where there are no embryos they are distinctly columnar,
with well-marked cell outlines, with nuclei invariably situated
at their base, with granular cytoplasm, and their free ends
rounded and separate from one another; the result being the
absence of a well-defined limit towards the cavity of the
duct. The cells lining the uteri, whether near the recep-
taculum seminis or towards the exterior, possess glandular
and secreting characters.

Up to the present time sucha difference as has been shown
to occur between the lining of the oviduct and that of the
uteri in Eoperipatus has been found only in Paraperipa-
tus Nove-Britanniez, discovered and described by Dr.
Willey (17, 18).

The oviducts have thick walls, consisting of the same

1 The term oviduct is restricted in the present account to that portion of
the genital duct which is situated between the ovary and the receptaculum
seminis; the term uterus being used to designate the portion of the genital

duct in which the embryos develop, and which is situated between the re-
ceptaculum seminis and the exterior,

TWO NEW SPECIES OF ONYCHOPHORA. 503

layers as the uterine wall, namely a peritoneal investment, a
tunica muscularis, a tunica propria, and a lining epithelium.
‘The first, second, and fourth of these layers are very distinct
in Eoperipatus, but the tunica propria is not well developed
either in the oviducts or in the uteri.

‘The ova in both species are large and full of yolk. In size
and structure they resemble those of the New Zealand forms,
and differ most of all from those of the genus Peripatus;
although as far as external characters are concerned Eoperi-
patus is more closely related to that genus than to any other.
When this feature of the ovum is considered in connection
with the external characters, it is impossible not to accept
Mr. Sedewick’s conclusion (14), and to reject that of Kennel
(9) and Willey (18), who think that the yolk-bearing con-
dition of the ovum is not a primitive but a secondary feature.
‘This question, however, will be discussed further on, as well
as the question as to which is the most primitive genus of the
Peripatidee.

The Receptaculum seminis :—In Koperipatus there
are a couple of well-developed receptacula seminis, such as are
found in the genera Peripatus, Paraperipatus, and
Peripatoides. In the genus Paraperipatus, the “in-
fundibulum,” which corresponds to what has been called the
oviduct in the present memoir, passes directly into the recep-
taculum seminis, and the uterine canal starts from the
opposite side of the same. The “infundibulum ” and the
uterine canal are put into communication with each other by
means of a narrow secondary duct (18). In the genus Peri-
patus, on the other hand, the main duct passes alongside
the receptaculum seminis, and seems to comm unicate with it
only in an indirect way, by means of a couple of narrow ducts
opening into the receptaculum seminis on either side, and
into the main duct by a common aperture. The condition
occurring in Hoperipatus is exactly parallel to that found
in the genus Peripatus. ‘The main canal passes alongside
the receptaculum seminis, and communicates with it only in
an indirect way by means of two diverging narrow ducts,

504, RICHARD EVANS.

which for the greater part of their course are embedded in
the wall of the sac (fig. 51).

The wall of the receptaculum seminis is thin, and consists
of two layers of cells. The outer layer is made up of flattened
cells belonging to the peritoneal epithelium; the inner layer
consists of short cells, with rather large nuclei and clear
cytoplasm. ‘They contrast in the most marked way with the
columnar lining of the adjacent genital ducts (fig. 51).
The two narrow ducts of the receptaculum seminis lie for the
most part in its walls, and are lined with short columnar cells
provided with small and closely set nuclei. The receptaculum
seminis seems to be a storehouse, in which the spermatozoa
are kept until they are wanted. In Eoperipatus copulation
appears to take place, and the spermatozoa must pass up the
uterine canal into the receptaculum seminis. In one speci-
men which was sectionised the lower parts of the uteri were
absolutely full of sperm-cells, enough, it would seem, for a
lifetime. Indeed, it is difficult to understand how fecunda-
tion can take place in EHoperipatus once the ova have
entered the genital ducts, and have started on their journey
down the uteri. HEmbryos in all stages of development are
found in the uteri, and probably go on developing all the
year round, so that the uteri are never empty. The natural
conclusion come to is that fecundation takes place only
once; that is, before the ova have ever entered the uteri.
The receptacula seminis should be considered as a couple
of sacs formed for the purpose of retaining the sperma-
tozoa, which are transmitted into the female uteri during
copulation, which takes place before any ova have ever passed
down the oviducts to the uteri.

The Receptaculum Ovorum :—In Koperipatus there
are two receptacula ovorum with thick walls, exactly like
those figured and described by Gaffron as open funnels
(7), and by von Kennel as closed sacs (9). They occur only
in the genera Peripatus and Hoperipatus. Paraperi-
patus, Mesoperipatus, Opisthopatus, Peripatopsis,
and Peripatoides have no receptacula oyorum. In Eo-

TWO NEW SPECIES OF ONYCHOPHORA. 505

peripatus they are situated close to the ovarian opening of
the oviducts, with which they communicate. Their free end,
which was described by Gaffron in the genus Peripatus as
an open funnel, is closed by a thin membrane. ‘The lining of
the diverticulum in the region situated close to the above-
mentioned membrane consists of columnar cells, which pre-
sent a certain amount of resemblance to the lining cells of
the renal funnels communicating with the coelomic end-sac
of the renal organs ; but the lining of that part which opens
into the oviduct presents the characters possessed by the
lining of the latter (fig. 50).

Willey is of opinion that the presence or absence of recep-
tacula ovorum is correlated with the occurrence of what he
describes as “epithelial ova” and ‘follicular ova” respec-
tively. The presence in Eoperipatus of “follicular ova”
as well as receptacula ovorum, proves Willey’s suggestion to
be unsound. It seems, however, that there is no reason for
supposing that the so-called receptacula ovorum function as
such in Hoperipatus. It is much more probable that here,
as in all forms which possess follicular ova, the stalks of the
follicles, as Willey expresses it, represent so many secondary
ducts discharging into the main ovarian cavity, which in
Koperipatus is of immense size, and which plays the part
of the receptacula ovorum.

There seems to be in the literature of the Peripatide a
certain amount of confusion as to the exact meaning of Mr.
Sedewick’s suggestion of the homology of the parts under
consideration. Having said that Kennel distinctly states that
he does not regard the receptaculum ovorum as homologous
with the funnel of the renal organ, apparently because the
thin-walled vesicle closes its free end, Mr. Sedgwick proceeds
to explain his view of the homology of the parts in question.
He draws the conclusions that the thin-walled vesicle of the
receptaculum ovorum is homologous with the coelomic end-
sac of the renal organs, and that the diverticulam—Ovarian-
trichter of Gaffron—is homologous with the so-called funnel
of the renal organs; two conclusions which derive support

you. 44, PART 4,—NEW SERIES, KK

506 RICHARD EVANS.

from the histological details of the parts under consideration
in Hoperipatus.

The only desirable modification of the above conclusions is
that not the whole, but a portion, of the canal passing from
the thin-walled vesicle to the oviduct is homologous with the
renal funnel—the part situated nearest the oviduct, of which,
though it 1s curved, it seems to be a continuation, being homo-
logous with a portion of the coiled tube of the renal organs.

To recapitulate, the homology of the parts seems to be as
follows :—The thin-walled vesicle or membrane closing the
ovariantrichter of Gaffron is homologous with the coelomic
end-sac of the renal organs; the portion of the diverticulum
which is situated nearest the thin-walled vesicle corresponds
to the renal funnel; the remaining portion of the diverti-
culum, the oviducts, and the uteri, would be homologous with
the coiled tube and the dilated bladder ; and finally, the ecto-
dermal portion of the female genital system would correspond
to the short, ectodermal duct of the renal organ.

On the above view of the homology of the parts here con-
sidered the receptaculum ovorum is the direct continuation
of the oviduct into the terminal end of which it leads. It
follows that the renal funnel is not represented by the pore
leading from the oviduct into the diverticulum. It may be so
represented in other genera of the Peripatide, but certainly
it is not so in Hoperipatus.

It does not follow from the above view that the cavity of
the ovary is a mere continuation of the oviduct. Sedgwick
describes the “germinal nuclei” as appearing in the sixteenth
to the twentieth somite, both inclusive. The twenty-first
pair of somites, in which germinal nuclei do not appear, are
completely used up in the formation of the genital ducts. In
this case the ovary formed from the dorsal ccelom of several
somites becomes grafted—so to speak—on the twenty-first
somite. Consequently it can in no way be described as a
mere continuation of the oviduct which is developed inde-
pendently from a different pair of somites.

Kennel, differmg from Sedgwick, has described the ovary

TWO NEW SPECIES OF ONYCHOPHORA. 507

and oviducts as being developed from the genital somite
alone in the genus Peripatus.

It is quite possible that there is a difference in this respect
in the Peripatidee; but it is necessary to point out that the
position of the ovary in the adult of the genus Peripatus, as
well as of other genera, tends to show that Kennel is in error,
for the ovary is located in the pregeuital seement, just where
we should expect to find it had it been developed from the
dorsal portion of a pregenital somite. If this be so, the
ovary with its oviduct is a composite structure even in the
genus Peripatus, as is certainly the case in the genus Peri-
patopsis. But, though Kennel may be in error in deriving
the ovary from the genital somite alone in the genus Peri-
patus, it is highly improbable that it is formed from as many
somites as in the genus Peripatopsis; for to make a mistake
with regard to one somite is quite different from falling into
error regarding six somites. It would seem that there is a
difference between Peripatopsis and Peripatus in this
respect ; a difference waich points to the possible participa-
tion of a great number of somites in the formation of the
ovary in the ancestral Peripatus. In the genus Hoperi-
patus the ovary is formed from the dorsal portion of the
four pairs of somites situated immediately in front of the
genital segment, which, it would seem, takes no part in the
formation of the ovary.

The Uteri:—tThis term is applied to the main part of the
generative ducts, that part which extends from the receptacula
seminis to the vagina, and contains the developing embryos,
if there are any. The uteri present no constant arrange-
ment, and it is impossible to say whether any particular one
is predominant. In the specimens shown in figs, 22 and 23,
the arrangements are exceedingly different. In both cases
the uteri contain a number of embryos. In the specimen
shown in fig. 22, the oldest embryo in the left uterus is on
the right side, and the one in the right uterus is on the left
side—a result brought about by the crossing of the uteri close

to the vagina.

508 RICHARD EVANS.

In the specimen represented in fig. 23 the uteri do not
cross one another; consequently the embryo in the right
uterus is on the right side, and the one in the left is on the
left side. In the two specimens represented in figs. 22 and
23, there is another marked difference in the topography of
the uteri. In fig. 22 the uteri are packed under a loop of
the posterior end of the stomach, and do not extend forward
on either side of the alimentary system; while in figure 23
they are placed on the right side of the stomach, and extend
forward as far as the first pair of legs. The stomach in this
case has no loop at its posterior end, and this may be the
cause of the forward extension of the uteri.

As has been stated above, the ovary is constant in its
position. The proximal ends, that is, the ovarian ends of
the uteri, always pass forwards to a greater or less distance ;
the middle portions are intricately coiled among them-
selves, and may form one or two loops round the stomach ;
the posterior ends descend and pass on the outer side
of the nerve-cords to the short vagina (figs. 22, 25, and
33).

The thickness of the uterine wall is highly variable. It
may be thick and consist of a well-developed peritoneal
investment, a tunica muscularis, a tunica propria—which is
usually very thin,—and a lining of tall columnar cells
(figs. 51, 52, and 54), or it may be thin, when it seems to
consist of thin layers of cells. ‘This difference, however, is
more apparent than real, for it is due to the expansion of
the uterus brought about by the increasing volume of the
developing embryos, as well as the secretion of nutritive
material by the lining cells, and consequent diminution in
size. hat there is no actual loss or destruction of the lining
cells is amply proved by the way the columnar cells which
line those portions of the uteri situated between the several
embryos gradually pass into the flattened layer which lines
the portions of the uteri in which the embryos are found
(fig. 53).

In mature specimens the uteri contain a variable number

TWO NEW SPECIES OF ONYCHOPHORA. 509

of embryos which represent a number of stages in the
development. It seems that not all the ova which pass into
the uteri develop. Many of them, either for want of room or
for some other reason, fail to advance beyond the segmenta-
tion stages, even if they do segment at all, and as the embryos
which are successful in the struggle for existence elongate,
the unsuccessful ones become wedged in between the con-
tinually growing young.

There seems to be no evidence in favour of the view, which
has been put forward more than once, that accompanying
parturition there is a resorption of the terminal ends of the
uteri. When the pigmented embryo, which measures from
20 to 27 mm. in length, is born, that portion of the uterus
which contained such an embryo must contract; consequently
the wall must thicken and the flattened cells of the lining,
being no longer called upon to secrete nutritive material, once
more assume their normal form. The embryos are always
entirely free in the uterus, that is, there is no organic con-
nection between the developing young and the uterine wall,
though they are in close contact with each other. From
what can be seen in Hoperipatus, it seems much more
reasonable to suppose that the embryos pass gradually down
the uteri than that the uteri are resorbed at their vaginal
ends,

(f) The Male Reproductive Organs:—Though the
female reproductive organs of HKoperipatus differ con-
siderably from those of the genus Peripatus, the male
reproductive organs of the former present a peculiar agree-
ment with those of the latter.

The tubular testis communicates with the seminal vesicle
by means of a short duct which opens into the latter near
its anterior end. Close to the posterior end of the seminal
vesicle the vas deferens’ arises and coils about in the body-
cavity (hemoccel). The vasa deferentia pass backwards, and

1 No distinction is made here between vas efferens and vas deferens,
because there seems to be no satisfactory reason for drawing such a dis-
tinction.

510 RICHARD EVANS.

on reaching the level of the antepenultimate pair of legs,
leave the dorsal aspect of the animal and pass in an oblique
direction towards the ventral surface. The right vas deferens
makes its way under the corresponding nerve-cord and the
ductus ejaculatorius. Asa rule, the left vas deferens passes
under the corresponding nerve-cord, but, in some cases, ib
does not do so. The vasa deferentia unite to form acommon
duct, the point of union being situated either on the inner or
on the outer side of the left nerve-cord, according as the left
vas deferens does or does not make its way under the cord.
When the point of union is on the outer side of the cord, the
right vas deferens passes under it so as to unite with the left.
After the vasa deferentia have united, they pass forwards in
a common sheath to the level of the antepenultimate pair of
legs, where the canals themselves unite.

The common duct runs forward as far as the third pre-
genital pair of legs, and there, turning round, makes its way
backwards to the genital orifice. When the point of union
of the vasa deferentia is on the outer side of the left nerve-
cord, the ductus ejaculatorius in passing to the genital
opening has to make its way under the cord; but when the
point of union is inside the nerve-cord, the ductus ejacula-
torius does not pass under it (figs. 36 and 37).

‘he testes are provided with a peritoneal investment and a
lining of tall columnar cells, with nuclei at their base, and a
sharp cell outline. ‘he nuclei are comparatively large,
and the cytoplasm is clear (fig. 38). The duct which
passes from the testis to the seminal vesicle has the same
two layers as the testis, together with an intervening
muscular layer (fig. 39). Its lining cells are columnar
in form and are provided with basal nuclei, well-defined cell
outlines, and clear cytoplasm. ‘The wall of the seminal
vesicle presents close resemblance to that of the receptaculum
seminis. The vas deferens at its commencement has much
the same appearance in section as the short duct which passes
from the testis to the seminal vesicle, with the difference
that the intermediate or muscular layer is thicker (fig. 40).

TWO NEW SPECIES OF ONYCHOPHORA. Sale

Further along the vasa deferentia, however, the wall becomes
thinner, the muscular coat less developed, and the lining
cells less columnar. This feature becomes more and more
emphasised towards the point of union of the two genital
ducts, as can be seen on comparing figs. 40, 41, and 42.
At the above-mentioned point, however, there is a sudden
change in the characters of the lining cells as well as in the
inuscular coat. In both the ascending and descending limbs
of the common duct the lining cells become columnar and, as
a rule, the cell limits are well marked. The muscular coat of
the ascending limb, however, remains comparatively thin,
and even in the descending limb it only becomes greatly
thickened at the level of the antepenultimate pair of legs.
From that position onwards it is extremely thick, and the
term ductus ejaculatorius should be confined to this thickened
portion of the common duct (figs. 46, 47, and 48).

The most interesting character of the male genital organs
of Koperipatus is the great length of the unpaired duct,
which almost equals in extent that of the genus Peripatus.
The variation that occurs in the length of the unpaired
portion of the male organs in the Peripatide is a very
interesting feature. It is longest in the genera Peripatus
and Hoperipatus (7). In the genus Peripatoides it is not
so long (16), and is still shorter in Peripatopsis (2); but is
shortest of all in Paraperipatus, in which, according to
Willey, the unpaired portion of the male duct is hardly any
longer than the vagina (18). Whether this feature of Para-
peripatus is primitive may well be doubted, when the fact
that there are no spermatophores in this genus is taken
into consideration. As in the genera Peripatus and Peri-
patoides, there is in the unpaired duct of Koperipatus an
enormously long spermatophore, which, however, lacks the
horny coat described by MissSheldon in P. Nove-Zealandiz
(16), and by Gaffronin P. Edwardsii (7). It is nevertheless
provided with a horny cap, which covers its foremost end,
and this seems to be the only advance which it has made, in
the direction of forming a coat, from the condition described

512 RICHARD EVANS.

by Professor Lankester many years ago in two species of the
genus Tubifex (11). In this Annelid genus the tails of the
spermatozoa project freely from the wall of the spermato-
phore, and are found in the living state in continual vibration.
In Eoperipatus the tails were not observed to project freely
from the body of the spermatophore, but it must be remem-
bered that it was not examined in the fresh state. If it were
examined fresh, it is quite possible that the tails of the
spermatozoa would be found to project freely into the cavity
of the duct. In the spermatophore of Hoperipatus as in
that of Tubifex the spermatozoa are arranged radially round
a central core, which is free of them, as was figured in
Tubifex by Professor Lankester (11). It is evident that
the spermatophore of Hoperipatus is but little in advance
of that found in the Annelid ‘l'ubifex. ‘The next stage in
the evolutionary series is found in the genera Peripatus
and Peripatoides, where both the head and the body of
the spermatophore are provided with a horny coat (7) (16).
‘he genus Peripatopsis would supply a still further stage
in the series; a stage in which the long spermatophore
of the other genera has been broken up, and is represented
by a number of small oval bodies each consisting of a thin
case full of spermatozoa (15). The last stage would be
represented by the condition of things found in Para-
peripatus, where the male genital ducts are described by
Willey as “containing abundant loose felted spermatozoa”
(18).

If, on the one hand, the above account represents cor-
rectly the phylogenetic history of the spermatophore in the
Peripatidee, and if, on the other hand, as Willey seems to
think (18), the presence or absence of a long unpaired duct
is in correlation with the formation of a long spermatophore,
it seems that there are good reasons for doubting the con-
clusion that the condition of the genital ducts in the genus
Paraperipatus is primitive, though on a priori grounds it
appears to be so. ‘his explanation of the features presented
to us enables us to place the Malay and Neotropical genera

i’ WO NEW SPECIES OF ONYCHOPHORA. 51S

nearer the base of the Onychophoron branch of the great
Arthropodan phylum than the African, New Zealand,
and Australian genera. ‘This agrees with the arrangement
to which we are forced on other grounds, which will be dis-
cussed later.

(g) The Male Accessory Glands.—These are a pair of
glands which occur only in the male (fig. 35, m. a. q.,
and fig, 36, m. a.g.). Their external opening has already
been described (p. 490). The general course which they take
is the following. They start as straight tubes, occupying a
cavity of their own situated in the dorso-lateral aspect of
the body. Their position is shown in fig. 35, which repre-
sents a transverse section passing just in front of the male
genital pore, from the level of which they pass obliquely
downwards on the outer side of the nerve cords to the
common opening which is situated between the last pair of
legs. In passing round the nerve cords they press against
them in such a way as to become partially embedded in
them.

The inner ends of the male accessory glands have thin
walls and a fairly large lumen, which is circular in shape
(fig. m. a. g.), and which is not lined by a chitinous
intima. ‘The lining cells of this part are short and colum-
nar; the peritoneal investment consists of much-flattened
cells, and the intervening layer is thin. The walls of the
portion situated nearer to the exterior which passes round
the nerve cords are much thicker, and the lumen, which may
be of any shape, and small, is lined with a chitinous intima.
It consists of the same three layers as the thinner portion
situated more internally, but the middle or muscular layer is
much thicker.

It is difficult to say what the homology of these glands
may be, as their development has not been worked out; but
it seems necessary to discuss the possibilities of the question,
were it only to clear the ground. The great difficulty of the
subject lies in the different positions occupied by the ex-
ternal openings of a pair of glands which are found more or

514 RICHARD EVANS.
less closely related to the male genital system in the dif-
ferent genera of the Peripatidee.

In Peripatus Edwardsii the “anal glands” open on
either side of the anus as shown by Gaffron (7). In Peri-
patoides Nove-Zealandiz, according to Miss Sheldon, the
“accessory glands” open outside the nerve-cords near the
posterior end of the body (16). In Peripatoides leuck-
arti, Fletcher describes the “ accessory glands ” as opening
close together between the generative pore and the anus (6).
In Peripatopsis capensis, according to Balfour, the “ male
accessory glands” open into the ductus ejaculatorius (2).
Willey describes the “ pygidial glands” of Paraperipatus
Nove-Britannie as debouching into a much-thickened bul-
bus, which in its turn opens to the exterior above the anus (18).
In Koperipatus the male accessory glands debouch into a
median opening situated between the last pair of legs.

From the above short statement it will be seen that the
glands found at the posterior end of the Peripatide open
to the exterior in half a dozen different positions. As to
their homology two views have been put forward, and obvi-
ously two views are possible. Balfour. thought the male
accessory glands of P. capensis were homologous with the
crural glands (2). Kennel thinks that the anal glands of P.
Kdwardsii are homologous with the renal organs (9), and
Willey seems to accept this view (18). The latter author in
his most brilliant account of the anatomy and development
of Paraperipatus Nove-Britanniz writes as follows of
the “ pygidial glands :”—“ There are a pair of glands... .
homologous with the ‘accessory glands’ of the African
and Australian species, and with the ‘anal glands’ of the
Neotropical species.” On another page he has the following
expression :—“ It is advisable to give separate names to
structures, even though obviously homologous, when they
have such very different anatomical relations.” (In both of
the above quotations the italics are mine.) Dr. Willey has
brought forward few or no reasons in favour of the view
that all these glands are “ obviously homologous.’ At the

TWO NEW SPECIES OF ONYCHOPHORA. 515
present juncture it is proposed to examine the known facts
of anatomy and development in order to see how far such a
statement is justified.

Gaffron’s description of the “anal glands” of P. Ed-
wardsii, and Willey’s account of the “ pygidial glands” of
P. Nove-Britanniz are in agreement in so far as that
they show that the glands consist of two parts, which Gaffron
designated the “ectodermal” and the “ entodermal” por-
tions, and which Willey describes as “ ectodermal ” and
“mesodermal.” The same distinction occurs in Eoperipa-
tus, though it is perhaps not so well marked. So far, then,
there is in these three generaan agreement between the acces-
sory glands—to use one term for all of them—and the renal
organs, which consist of an ectodermal portion, however
short, and a mesodermal or ccelomic portion. It must be
pointed out, however, that too much importance should not
be attached to the difference between the so-called ecto-
dermal and mesodermal portions of these glands, because
the lining cells of the ccelomic portions of the renal organs
differ widely from one part to another; for example, the
lining of the funnel differs more from that of the coiled tube
or of the ccelomic end-sac than the lining of the two parts of
the accessory glands do from one another. It follows that the
argument derived from histology in favour of the view that
the accessory glands consist of ectodermal and mesodermal
parts homologous with the similarly situated parts of the renal
organs is not a very strong one. It seems that the only point
of any importance is the existence of a chitinous intima lining
the so-called ectodermal portions, which are always short.

Kennel described the development of the anal glands in
P. Kdwardsii as taking place from the apodal anal seg-
ment ; that is, the second segment behind the genital one,
and also found a vestigial representative of them in the
young female. The conclusion which naturally follows is
that the anal glands of P. Edwardsii are homologous with
the renal organs and genital ducts.

In Peripatoides Nove-Zealandizw the accessory

516 RICHARD EVANS.

glands open by the sides of the nerve cords, near the
posterior end, that is, they open almost exactly where we
should expect a renal organ to open, and so far this is in
favour of their homology with the renal organs; but we
know nothing of their development. It is not known
whether they belong to the anal segment, as in P, Hd-
wardsii, or to another segment situated between the anal
and the genital segment, which in P. Nove-Zealandie
has lost all traces of its appendages. If they belong to the
former there is a very close relation between them and the
anal glands of P. Edwardsii, but if they belong to the
latter, the most that can be said is that they are thus homolo-
gous in the same sense that the renal organs of the various
segments are homologous with them.

In P. Leuckarti the ‘accessory glands” differ from those
of P. Nove-Zealandiz only as regards their external open-
ings, which are situated close together near the middle line,
between the genital orifice and the anus—a position reached
by a very slight amount of shifting towards the mid-ventral
line. It seems probable, when other genera are taken into
consideration, that in the above two species which belong to
the genus Peripatoides, the “accessory glands” are de-
rived from the anal somite.

The condition existing in P. Leuckarti leads naturally to
that found in Hoperipatus, where the accessory glands open
into a common pit situated in the mid-ventral line between the
genital pore and the anus; that is, the external opening is
placed exactly where we should expect to find it if the acces-
sory glands belonged to the somite of the last leg-bearing
segment. But if the accessory glands of Hoperipatus are
homologous with the renal organs, they cannot belong to the
last leg-bearing segment; for in both male and female there
exists a well-developed renal organ in that segment, though
in the adult male the external portion of the duct is not
found. It follows that the accessory glands of Hoperi-
patus must be either the renal organs of the anal segment
as in P, Edwardsii, or the crural glands of the last leg-

TWO NEW SPECIES OF ONYCHOPHORA. 517

bearing segment. Their position is in favour of the view
that they are crural glands which have passed from the
outer to the inner side of the renal organs. This would be
possible, because the renal openings on this segment have
moved slightly forward, as can be seen from the section
represented in fig. 34, which, though passing through the
renal opening, is taken in front of the last pair of legs.
Their structure, however, is in favour of the view that they
are derived from the apodal anal segment, and contain a
ceelomic element, but that their openings have moved both
forwards and inwards, and have united together in the mid-
ventral line. Of the two views above suggested the latter
seems the more probable, because the accessory glands of
Koperipatus and the anal glands of Peripatus would on
that view belong to the same somite.

The next form to be considered is Paraperipatus, in
which the glands open dorsally above the anus. In Paraperi-
patus the penultimate pair of legs has completely dis-
appeared, and the genital segment is apodal. As _ the
development of the “ pygidial glands” is unknown, it is not
at all certain whether they are crural glands or renal organs
which have been modified. Their structure is certainly in
favour of the latter view. Again, supposing that they are
modified renal organs, it is not known whether they are
derived from the somite next the genital somite or from the
one which follows, namely, the anal somite; the position of
their external opening, however, is in favour of the latter view.

In all the above forms it is probable that the accessory
glands belong to the apodal anal segment—a statement,
however, which is far from having been proved. As regards
their external openings, there seems to have been a gradual
shifting probably in two directions. If these glands represent
renal organs it is only natural that they should have opened
at first in the same position as the renal organs. In fact, the
position of their openings in P. Nove-Zealandiz closely
resembles that of the renal organs, and a slight shifting for-
wards and inwards would produce the condition found in

518 RICHARD EVANS.

P. Leuckarti, and finally in Eoperipatus. If instead of
shifting forward they had shifted backwards, the condition
found in P. Hdwardsii would result, and by shifting up-
wards that found in Paraperipatus would be brought
about. It seems that there are two reasons for this shifting ;
first, the slightly subterminal position of the anus; secondly,
the shortening of the anal cone. Probably the latter was the
more potent element in driving the opening to the mid-dorsal
line in Paraperipatus, the question of space becoming a
determining factor in bringing about the change of position
of the openings. In Paraperipatus, as the anus and
genital orifice approached each other, the glands and their
openings were forced towards the dorsal aspect of the animal.

We may conclude that it is possible to homologise the
anal glands of Peripatus, the pygidial glands of Para-
peripatus, and the accessory glands of Peripatoides and
Koperipatus with the renal organs, and with one another,
without violently twisting the facts known about them,
though in many respects we are still treading on uncertain
ground. ;

There remains to be considered, however, one well-known
form, namely, Peripatopsis capensis. In this form the
male accessory glands open into the terminal portion of the
ductus ejaculatorius. In this feature Peripatopsis is quite
unique. In fact, in this genus the relation of the male
accessory glands to the ductus ejaculatorius is exactly what
it should be, on the view that they are the crural glands of
the genital segment. In Balfour’s posthumous works the
following expression occurs :—“'The accessory gland in the
male is probably a modification of one of these organs ;” i.e.
of the crural glands (1). Mr. Sedgwick in his account of
the development of the species under consideration writes as
follows :—“There are rudiments of two pairs of somites
behind the somites of the anal papille in Stage HK. One
of these is just visible in Stage F. They vanish completely
at the end of Stage F. No appendages or rudiments of such
are developed in connection with them”? (14). From this

TWO NEW SPECIES OF ONYCHOPHORA, 519

passaye it seems absolutely clear that the somites behind the
genital segment do not give rise to the male accessory glands
of P. capensis. It follows from this that the glands in
question cannot in any way be described as modified renal
organs, unless the genital somite is capable of giving rise to
two sets of similar organs, one of which is modified as the
genital ducts, and the other as the male accessory glands—
an absolutely gratuitous supposition, for the body segments
of the Peripatide, as far as is known, never give rise to
more than one renal organ or its modified homologue. It
seems, therefore, that the “obvious homology ” of the male
accessory glands of P. capensis with the anal glands of P.
Kdwardstii, the pygidial glands of P. Nove-Britannia,
and the accessory glands of P. Nove-Zealandiz, P.
Leuckarti,and Hoperipatus, is far from being proved, and
that the possibility of their being merely modified crural
glands must not be lost sight of. It may be that, when we
shall have learnt more about the origin of these glands in
the different genera, they will be shown to be homologous ;
but for the present we must at least suspend judgment, as
the formation of a premature conclusion tends to obscure the
problem which has to be solved, and the view that they
are homologous is still sub judice.

(h) The Crural Glands.—The occurrence of these glands
has already been mentioned in describing the position of their
external apertures. ‘here are four pairs of them; two in
each of the first and second pregenital pairs of legs. They
are tubular glands which almost exactiy equal in size
the renal organs of the segments in which they occur
(figs. 27 and 28). They are lined internally with a thin
chitinous layer, under which are found the short columnar
cells, which are supported by a thin muscular coat and a
peritoneal investment. The occurrence of two crural glands,
in each of the legs in which they occur, is a feature in which
Hoperipatus Horstiresembles P. Edwardsii. It appears,
however, that they occur in a far greater number of legs in
the latter than in the former (7). .

520 RICHARD EVANS.

V. Tse SrructurE OF THE Ovum.

In no group of comparatively small size does the structure
of the ovum vary as in the Onychophora. In Peri-
patoides and Hoperipatus the ovum is large and full of
yolk; in Peripatopsis it is large and devoid of yolk; m
Peripatus and Paraperipatus it is small and free from
yolk.

With regard to the primitive condition of the ovum there
are two views—the one put forward by von Kennel (9), and
recently supported by Dr. Willey; the other held by Mr.
Sedgwick (15), and adopted by Korschelt and Heider in their
‘'Text-book of Embryology.’ Von Kennel’s view, so frequently
adopted by Dr. Willey, is that the ancestral Peripatus had
a small yolkless egg, which it laid in water; Mr. Sedgwick’s
view, adopted by Korschelt and Heider, is that the ancestral
form had a large egg full of yolk. Kennel, with whom the
first view originated, thought that the course of embryonic
development in the Peripatidze had followed two divergent
lines of evolution, the one leading towards the type of
development occurring in the genus Peripatus, the othe:
towards that found in the genus Peripatoides; the ovum
remaining yolkless in the former, but developing yolk in the
latter—a feature which is considered by Willey to be a
secondary one, which culminates in the oviparity of P.
oviparus (Dendy). Sedgwick, disagreeing with Kennel,
sees only one line of evolution; according to him the yolk-
bearing egg of Peripatoides is primitive, the vesicular
ege of Peripatopsis is intermediate, while that of Peri-
patus represents the end result of a series of modifications
(15, p. 463).

Dr. Willey has recently expressed himself very strongly on
these questions. Even with regard to the condition found in P.
capensis, he declares that the opposite view to that of Sedg-
wick could be sustained with equal force, though he admits
“there is no means at present known of deciding between the
two views in this particular case,” and “both of them seem to

TWO NEW SPECIES OF ONYCHOPHORA. SPA

be equally possible.” In the case of the egg of Paraperi-
patus, as well as in that of the Neotropical forms, he goes
even further and says, “There is no reason whatever to
suppose that there has been a secondary loss of yolk
in these cases” (18). He does not stop to argue and con-
sider the point, but gives us his conclusions unsupported by
either fact or argument of any kind. He begins the paragraph
from which the above quotations are taken as follows :—‘ I
will not attempt to discuss this very difficult subject.”

Since Dr. Willey did not consider it advisable to make us
acquainted with the line of thought which led him to adopt
the above conclusions, he can hardly find fault with us for
not accepting them, and perhaps for trying to supply the
reasons which induced him to adopt them.

The view has been held for many years, and is being held
to-day—most probably quite correctly —that of all the forms
of Peripatide hitherto known, the Neotropical ones are
the most primitive. Kenuel, probably thinking that an
animal which is in several respects primitive must be so in
every respect, came to the conclusion that, since the
Neotropical forms are primitive as regards general cha-
racters, they must also possess a primitive ovum; that is, as
he thought, a small yolkless ovum. Dr. Willey may have
unconsciously fallen into the same error, for he mentions
several features which he considers primitive in Paraperi-
patus, for example, the structure of the male genital ducts,
and it is quite possible that for such reasons he has accepted
von Kennel’s view of the original condition of the ovum in
the Peripatidee.

Evidently there are two ways of explaining the structure
of the ovum of P. capensis. ‘The condition of the ovum in
this species may be a vestigial one, the yolk having been
lost, but the spaces in which the yolk was found in the not
very distant past, being still retained, as well as the com-
paratively large size of the ovum; or it may be a case of what
may be described as prophetic adaptation in phylogeny,
the egg preparing itself—so to speak—for the reception of

voL, 44, PART 4,—NEW SERIES. L

ale RICHARD EVANS.

yolk which is to be stored in it in the not very distant future,
such as is now found in the ovum of both Hoperipatus and
Peripatoides. he first explanation is the one put forward
by Sedgwick, and seems to be the only possible one; for the
second alters the sequence of events in such a way that the
effect precedes the cause. ‘The cause of the large size of all
ova seems to be the presence of yolk in one form or another.
But there is no yolk in P. capensis if we are to accept
Sedewick’s account of the structure of the ovum in that
species. From these considerations it seems fairly evident
that it is not equally possible to explain the structure of the
ovum of P. capensis in both ways, and that the only
possible explanation of it is the one put forward by Mr.
Sedgwick. It has been pointed out above that the main
reason for assuming the ovum of the Neotropical genus,
Peripatus, to be primitive as compared with that of the
New Zealand genus, Peripatopsis, is the generally primitive
character of the former, and Dr. Willey might justifiably say
that there is no such difficulty in explaining the structure of
the ovum of Peripatus and Paraperipatus as is raised
by the explanation of the case of Peripatopsis. But
now that the Malay forms have been discovered, forms
which in all respects are as primitive, and in some respects
more primitive, than the Neotropical genus ;—e. g., in the
position of the renal opening of the fourth and fifth pairs of
legs; the presence in the species Horsti of a well-developed
renal organ in the last pair of legs, and in the species
Weldoni of a vestigial one,—a new difficulty crops up, for we
have two groups, to which the dignity of genera has been
accorded in the present memoir,—which are closely related in
all their external and internal characters except in those of
the ovum, ovary, and mode of development. In the genus
Peripatus the ovary is small and compact, and the ova
arise endogenously and are devoid of yolk, but in the
genus Hoperipatus the ovary is large and spreads out, and
the ova arise exogenously and are full of yolk. It becomes
necessary to decide which of these conditions is the more

TWO NEW SPECIES OF ONYCHOPHORA, 523

primitive. If there was no difficulty in accepting Kennel’s
view before Eoperipatus was discovered, there certainly
seems to be a difficulty now that this genus has to be con-
sidered, and that of such a nature as cannot very easily be
put aside. When the genera Peripatus and Peri-
patoides are compared from this point of view, it seems
less difficult to adopt Kennel’s view, because in Peri-
patoides the young are born small in size, or are hatched
from ova outside the body; but when Peripatus and Ho-
peripatus are compared this reason no longer exists. In
both Peripatus and Hoperipatus the young measure from
22 to 27 mm. in length at birth, and are coloured much in
the same way as the adult.

If Kennel’s and Willey’s view be correct, the ovum of EKo-
peripatus must have acquired its yolk since it took to
uterine development, and the condition of the ovum in the
genus Peripatus would in this case be primitive; but if
Sedgwick’s view be the correct one, the structure of the
ovum in Koperipatus has been inherited from an ancestor
which discharged a yolk-bearing egg either in water or on
land, and the condition of the ovum in the genus Peripatus
would be a secondarily acquired one.

The above seems to be a fairly clear statement of the facts
of the problem discussed. If Kennel’s view be correct, it
seems that we have to consider the question, what advan
tage would it be to Hoperipatus, once it had taken to
nourishing its young in the uteri, to produce yolk in its egg
as well? It is difficult enough to explain the presence of
yolk in the egg of Hoperipatus on the supposition that it
has been inherited from a former ancestor which discharged
a yolk-bearing egg; for we should expect to find the yolk
disappearing, as in the genus Peripatopsis, when uterine
development became habitual; but when we are asked
why yolk was produced after the uterine method had become
the habitual mode of development, we find it impossible to
answer, simply because we cannot conceive of any advantage
that would accrue to the animal in the struggle for hfe from

524 RICHARD EVANS,

the adoption of such a course. In fact, it seems a decided
disadvantage for an egg which develops inside the mother,
within easy reach of any amount of food material in the form
of secretion from the tubular walls of the oviducts and uteri,
to have to move along from the ovary to the uteri carrying
with it a mass of yolk of an immense size, of which it has no
need. he improbability of the view held by Kennel and
Willey when applied to Hoperipatus, the young of which
measure 22 to 27 mm. in length at birth, is so great that
we are forced to reject it, and to adopt Sedgwick’s view
of the primitive condition of the ovum in the ancestral
form.

Having reached this position by a method of reasoning
which appears to be perfectly legitimate, the explanation
which Sedgwick gave of the structure of the ovum in P.
capensis follows naturally, and the ovum of Peripatoides
would have to be explained in the same way as that of
Koperipatus. The next conclusion of necessity follows,
which is, that the oviparity of P. oviparus is primitive and
not secondarily acquired.

To sum up, the following seem to be the stages in the
evolutionary changes of structure of the ovum, as its
development became more and more confined to the uterus,
Peripatoides represents a primitive condition, and produces
a yolk-bearing egg, which either develops within the uterus to
a small embryo, or is discharged and develops outside. The
second step is met with in Hoperipatus, which has a very
large yolk-bearing egg, from which is developed an embryo
measuring from 22 to27 mm.in lengthat birth. Peripatopsis
supplies the third step, with a large egg possessing highly
vacuolated cytoplasm, and produces an embryo of medium size
in the uterus. Paraperipatus represents the fourth step,
with a yolkless egg, much reduced in size, and gives rise to
an embryo of medium length. ‘The genus Peripatus seems
to present the culminating point in these changes, for it not
only produces a yolkless egg, but seems to possess what is a
highly modified mode of development.

TWO NEW SPECIES OF ONYCHOPHORA. PA)

It has been mentioned above as a general principle that
animals which are primitive in some respects need not of
necessity be so in all respects. The primitive genus Eo-
peripatus has the most primitive kind of ovum, while the
almost equally primitive genus Peripatus has completely
lost its food-yolk, and in consequence the embryo has assumed
an entirely secondary mode of obtaining nutriment. This
seems to be equally true of Paraperipatus, where one portion
of the ovum is turned into a kind of trophic sac, which comes
in contact with the uterine wall and draws nutritive material
from it, by means of which the young embryo develops and
increases in size.

We now pass on to discuss the relations of the various
genera to one another.

VI. Tue Retations oF THE VARIOUS GENERA OF THE PERIPA-
TIDE TO ONE ANOTHER, TOGETHER WITH SOME PHYLOGENETIC
CONSIDERATIONS.

M. Bouvier, in his recently published memoir, has put
forward the view that the most primitive forms are those
which show the least amount of degeneration of the two
pairs of legs situated posteriorly (5). He is of opinion that
there is a gradual degeneration of the last and penultimate
pairs of legs, and from this point of view the family Peri-
patide naturally falls into four sub-families (see pp. 480 to
484). This view is accepted without reserve in the present me-
moir. It seems so well founded that there is no need to say
anything in its favour. However, it may not be out of place to
point out one embryological fact which supports this view,
namely, the presence in P. capensis of two vestigial somites
behind the somite of the anal papille or last pair of legs (14).
This is the exact number that is wanted in that species to
represent the segments found by Kennel in the Neotropical
forms. They completely disappear in the former, but in the
latter the hinder somite, that is, the anal somite, gives rise to
a vestigial renal organ in the female, and to the anal elands

526 RICHARD EVANS.

in the male (9). This goes far to prove that the genus
Peripatus is more primitive than the genus Peripatopsis,
and that the ancestral form possessed two somites behind the
genital one. This view assumes the position of the genital
orifice to remain constant, that is, between the originally
penultimate pair of legs—an assumption which is sup-
ported by embryology and anatomy. Taking the condition
of the last two pairs of legs as the basis of our classification,
the Peripatide fall into four groups, the first of which
contains three genera which show only a slight diminution
in size of the last pair of legs; in the second there are two
genera in which the last pair has disappeared, but the last
but one is still well developed; in the third there is only one
genus, in which the second pair of legs is very much reduced ;
and in the fourth there is one genus in which there are no
signs whatever of the originally penultimate pair, the genital
pore being situated behind the last existing pair of legs,
which is reduced in size.

In a group so small and so widely distributed it would
hardly be expected that one genus could be derived from
another. The consequence is that any kind of branching
arrangement intended to show their phylogenic relations
seems impossible, for in one genus one primitive character
has been retained while in another genus another such
character is found, and comparison of the several features of
the genera lead to the most divergent results. Hence the
gradation which can be traced among the several genera of
the Onychophora only represents steps which have been
reached by each genus more or less independently. If these
gradations be considered, as has already been done in the
case of the male accessory glands, the male ducts and
spermatophores, the ovum, and finally the genital orifice
and the last two pairs of legs, we arrive at a different result
in almost every case—a fact which justifies the statement that
no branching arrangement can possibly represent the phylo-
genetic relations of the different genera, which appear to
have developed more or less independently from a common

TWO NEW SPECIES OF ONYCHOPHORA. 527

ancestor, which is more closely represented in Hoperipatus
than in any other genus; for this genus combines the most
complete development of the two posterior pairs of legs with
what seems to be the most primitive ovum. Peripatoides
lacks the last pair of legs, though its development is as
primitive as that of Hoperipatus, and, even more so, in that
oviparity occurs. However, the actual development is so
little changed in Eoperipatus, that it may well lay claim to
being, on general grounds, the most primitive genus among
the Peripatide.

Before bringing the present part of my account of Ho-
peripatus to its close, it seems necessary to make a brief
reference to the geographical distribution of the Peri-
patidee.

About two years ago M. Bouvier published a preliminary
note on the geographical distribution and the evolution of
Peripatus (4), in which he came to the following conclu-
sions: first, that the African, American, and Australasian
groups are definitely related to the Continents the names of
which they bear; secondly, that it appears quite certain
that Central America and the Caribbean Region have been
the centre of origin and migration of the species of Peri-
patus. From the above-mentioned regions they are supposed
to have travelled towards the east to Africa, and towards
the west to Australia.

The first conclusion has been adopted in the present
memoir, and the discovery of the Malay forms considerably
strengthens it. But the same discovery seems to weaken the
second conclusion, which, to say the least, has been prema-
turely formed.

The only feature in which the American species seems to
be more primitive than the Malay ones is the possession of a
ereater number of segments in all of them. The number of
denticles in each blade of the jaws of P. tuberculatus
(Bouvier) seems in some respects to show a more primitive
condition than that which occurs in the Malay forms; but im
other respects the jaws of the latter seem to present the more

528 RICHARD EVANS.

primitive structure. ‘The greater number of denticles in the
outer blade of the jaws of P. tuberculatus seems the more
primitive, but of the inner blades those of the Malay species,
containing in all a greater number of denticles, seem to
approach nearer the original arrangement. In the position
of the renal papille of the fourth and fifth pairs of legs, the
arrangement occurring in Eoperipatus is decidedly the
more primitive, and in the structure of the ovary and the ova,
as well as in their method of development, the condition
found in the Malay species is nearer the ancestral arrange-
ment than that found in the genus Peripatus.

In the same communication Bouvier says that it would be
possible to believe the species of Oceania to be derived from
those of Africa; but, seeing that they have some primitive
features, e.g. the position of the sexual orifice and of the
nephridial pores of the fourth and fifth pairs of legs, he says
that it may be supposed that the dispersal of the group took
place in both directions at the same time; towards the east
in the case of Africa, and towards the west as regards
Australia and the adjacent region. In a later publication
Bouvier does not seem so definite in his opinion; for he uses
the word certain in his first paper (4), while in the second
he uses the word probably (8). Curiously enough, the
description given by Bouvier of P. Tholloni in the paper in
which he argues the American origin of the Peripatide is
exactly what was wanted to prove the possibility of the
origin of the Australian forms from Africa ;.for it does away
with any force there may be in the argument based upon the
possession of primitive features by the latter. ‘The prob-
ability of the African origin of the Australian forms is
strengthened by the discovery of the genus Opisthopatus
by Purcell—a genus which Bouvier himself has placed beside
the Australian genus Peripatoides because the genital
orifice is between the last pair of legs (5). When to all this
is added the argument drawn from the primitive structure of
the Malay species, the view that the Peripatidz originated
in Central America and the Caribbean Region seems well-
nigh smpossible.

TWO NEW SPECIES OF ONYCHOPHORA. 529

On looking through the list of sub-families given on page
8, we note that three of them are represented in Africa,
which is a rather significant fact, while only one of them is
found in America. If it were granted that Central America
and the Caribbean Region were the centres from which the
Peripatidze dispersed, the presence in the Malay Region
of such primitive forms would be a problem of great
difficulty, which would have to be solved. Besides, the
Australasian forms are much more closely allied to the
Malay species than to the Neotropical ones

another telling
fact which tends to make Bouvier’s view impossible.

It seems that there are two views to choose between, each
of which seems more satisfactory than the one adopted by
Bouvier. It may be concluded that the centre of origin of
the Peripatide was somewhere in Africa, and that they
travelled in one direction towards America and in another to
the Malay Region, and finally to Australasia; or it may be
concluded that the ancestral Peripatus had an exceedingly
wide distribution, ranging over a tract of land stretching
from South and Central America to South Africa, and across
to the Malay Region. When this tract of land became partly
submerged, the widely distributed Peripatidze became
separated in groups corresponding to the unsubmerged
continents. Hvery group then developed along one or
more lines of evolution. The American forms kept to one
line of evolution; the African species seem to have branched
off in three directions; the Malay forms adhered to one plan
of development; the Australasian species, hkewise, followed
one course, and finally the New Britain form has gone along
a line of its own.

It seems that there are at present no means of deciding
between the above-mentioned views. The adoption of the
latter view would certainly enable us to explain the occur-
rence of one primitive feature in one, and of another in
another genus. It seems that we must wait until more is
known of the specific characters of the Peripatidz before
we can decide with certainty between the two views above

530 RICHARD EVANS.

formulated, though from our present knowledge of them it
appears that Bouvier’s view must be dismissed.

THe DEPARTMENT OF CoMPARATIVE ANATOMY,
Tue Museum, Oxrorp;
November 380th, 1900.

ADDENDUM.

In connection with the present memoir, my sincerest thanks
are due to Professor Weldon for allowing me the free use of
his laboratory, for giving me the benefit of his opinion on
difficult questions bearing on the subject, for procuring for
my use some of the literature on the Peripatidas, and
finally for allowing Mr. Bayzand to help me with the illus-
trations ; to Professor Poulton for announcing the discovery
of the forms here described, for reading over the proof sheets,
as well as for many other kindnesses; to Professor Lankester
for supplying me with copies of some of the most important
papers on the subject; to Dr. R. Horst, of the Leyden
Museum, for his courtesy in allowing me to examine the type
specimen of EH. sumatranus; tothe Principal and Fellows of
Jesus College, Oxford, and to the Royal Society, for grants to
enable me to proceed with my researches.

BIBLIOGRAPHY.

1. Batrour, F. M.—*The Anatomy and Development of Peripatus
capensis,” posthumous memoir edited by H. N. Moseley and A.
Sedewick, ‘Quart. Journ, Mier. Sci.,’ vol. xxii, N.S., pp. 213—259,
Plates 13—20.

2. Bouvier, EB. L.—‘ Sur Organisation du Peripatus Tholloni,” Bouv.,
‘Bull. de la Soc. Ent. de France,’ pp. 197, 198 (1898).

3. Bouvier, lH. L.—“‘ Nouvelles observations sur les Peripatus,”

‘Comptes rendus Ac. des Sce.,’ t. exxvi, pp. 1524, 1525 (1898);
‘Ann. Mag. Nat. Hist.,’ vol. 11 (7), pp. 354, 355.

10.

11.

12.

13.

14.

15.

16.

TWO NEW SPECIES OF ONYCHOPHORA, bell

. Bouvier, K. |..—‘‘ Note preliminaire sur Ja distribution geographique et

Vevolution des Peripatus,” ‘Comptes rendus Ac. des Sc.,’ t. exxvi,
pp. 1858—1361; ‘Ann. Mag. Nat. Hist.,’ vol. ii (7), pp. 351—354.

. Bouvier, BE. L.—‘ Quelques observations sur les Onychophores (Peri-

patus) de la Collection du Musée Britannique,” ‘ Quart. Journ. Mier,
Sci.,’ vol. xliii, N.S., pp. 867—373.

. FLretcner, J. J—“On the Specific identity of the Australian Peri-

patus, usually supposed to be Peripatus Leuckarti, Saenger,”
‘ Proc. Lin. Soc., New South Wales,’ vol. x, p. 172 (1895).

. Garrron, E.—‘ Beitrage zur Anatomie und Histologie von Peripatus,”’

‘ Zoologische Beitrage’? (Schneider), vol. 1, 1895; Theil i, pp. 83—60,
Taf. 6—11; Theil ii, pp. 145—165, Taf. 21—23.

. Horst, R.—‘‘On a specimen of Peripatus, Guild., from Sumatra,”

‘Notes from the Leyden Museum,’ vol. vill, 1886, pp. 37—41, Plate 2.

. Kennet, J.—“ Entwicklungsgeschichte von Peripatus Kdwardsii,

Blanch, und Peripatus torquatus, n. sp.,” Theil i, ‘ Arbeiten a.
d. Zool. Zoot. Inst., Wurzberg,’ vil, pp. 95—229, Taf. 5—11; Theil
ii, ibid, vil, pp. 1—93, Taf. 1—6.

Korscuett, I., und Hetper, K.— Lehrbuch der vergleichenden Ent-
wicklungsgeschichte der Wirbellosen Thiere,’ Jena, 1890-1593. Kne-
lish translation, London, Part ILI, pp. 164—217.

Lankester, HK. R.—‘On the Structure and Origin of the Spermato-
phores or Sperm Ropes of two species of ‘Tubifex,” ‘Quart. Journ.
Mier. Sci.,’ vol ii, 1871, pp. 180—197, Plates 10, 1}.

Pocock, R. [.—* Contributions to our Knowledge of the Arthropod
fauna of the West Indies. Part II. Malocopoda and Protracheata,”
‘Journ. Lin. Soe. Zool.,’ vol. xxiv, 1894, pp. 518— 526.

Purceny, W. F.—* On the South African Species of Peripatidee in the
Collection of the South African Museum,” ‘Ann. So, Afr. Museum,’
vol. 1, pp. 8331—351.

Sepewick, A.—* The Development of the Cape Species of Peripatus,”
Part I, ‘Quart. Journ. Mier. Sci,’ vol. xxv, N.S. pp. 449—468,
Plate 1. Part II, ibid, vol. xxvi, pp. 175—212, Plates 12-15. Part.
ILT, ibid, vol. xxvii, pp. 467—550, Plates 34—37. Part IV, ibid, voi.
xxvii, pp. 373—3896, Plates 26—29.

Sepewick, A.—“ A Monograph on the Species and Distribution of the
genus Peripatus, Guilding,” ‘Quart. Journ. Mier. Sci.,’ vol. xxviii,
N.S., pp. 431 —493, Plates 34—40.

Suetpon, L.—-“‘ Notes on the Anatomy of Peripatus capensis and
Peripatus Nove-Zealandiz,” ‘Quart. Journ. Mier. Sci.,’ vol
xxvill, N.S., pp. 495—499.

532 RICHARD EVANS.

17. Wittey, A.—*‘On Peripatus Nove-Britannie, n. sp.,” ‘Ann. Mag.
Nat. Hist.,’ vol. i (7), 1898, p. 286.

18. Wiitey, A.—‘The Anatomy and Development of Peripatus Nove-
Britannie,” ‘ Zool. Results based on material from New Britain, etc.,’
Part I, pp. 1—52, Plates 1—4, Cambridge, 1898.

EXPLANATION OF PLATES 32—37,

Illustrating Mr. Richard Evans’ paper on “ ‘Two New Species
of Onychophora from the Siamese Malay States.”

Figs. 2, 6, 10, 16, 17, 18, 36, and 37 were drawn by Mr. Bayzand, the
able artist in the Department of Comparative Anatomy, at Oxford; all the
remaining figures were drawn by the author.

Figs. 24, 25, 26, 27, 36, and 37 were drawn from reconstructions by the
author of series of sections.

SIGNIFICANCE OF THE LETTERING.

a.p. Accessory papilla. cav. Cavity of the ovary. c.e.s. Coelomic end-
sac. ¢.g. Crural gland. d.e. Ductus ejaculatorius. jf. Funnel. 9.0.
Genital orifice. 4%. Heart. JZ. Left uterus. /.v.d. Left vas deferens.
m.a.g. Male accessory glands. m.d./. Mid-dorsal line. x. ¢. Nerve-cord.
op. Opening of the oviduct to the ovarian sae. ovid. Oviduct. p.c. Peri-
cardium. p.p. Primary papilla, &. Right uterus. —re.o. Reeeptaculum
ovorum. 7e.s. Receptaculum seminis. 7.0. Renal organ. 7. v.d. Right
vas deferens. s. g. Slime-gland. sp. Spermatozoa. wf. Uterus. v. 0.

Ventral organ.
PLATE 32.
Fic. 1 (x 2).—Dorsal view of Eoperipatus Weldoni.
Fic. 2 (x 2)—Ventral view of Koperipatus Weldoni.
PLATE 33.

Fie. 4.—Ventral view of the anterior end of Hoperipatus Weldoni,

Fie. 5.—Ventral view of the posterior end of Hoperipatus Weldon.

TWO NEW SPECIES OF ONYCHOPHORA. 533

Fic. 6.—Lateral view of one of the legs of Hoperipatus Weldoni.
Note the two distal papilla situated on the distal margin, one in front and
one behind. Hach of these papillee is divided into a basal and an apical piece,
the latter of which carries a spine. Also note the ventral prominences or
ridges which have the appearance of two recumbent papille pressed against
the latero-ventral aspect of the foot. Hvery prominence carries a spine
similar to the one on the distal papille.

Fie. 7.—A primary papilla from the dorsal surface of Eoperipatus
Weldoni. Note the conical basal part.

Fic. 8.—A primary papilla from the dorsal aspect of the leg shown in
Fig. 6. Note the cylindrical basal part, and compare with the conical basal
part of the papilla shown in Fig.7. Intermediate stages between the papille
represented in Figs. 7 and 8 are very common.

Fie. 9.—A ventral view of the fourth and fifth legs of Eoperipatus
Weldoni. Note the position of the renal papilla, on the top of which the
renal pore is situated. Note that by its position it divides the pad into two
pieces.

Fic. 10.—This figure represents a portion of the skin of the dorsal surface
of Koperipatus Weldoni. Note that the primary papille (p.p.) have a
polygonal or perfectly irregular outline, and that they stretch across the
ridges from one groove to the other. ‘The grooves between the ridges are
exceedingly narrow, and the accessory papille are often situated in the
grooves as well as on the ridges. The accessory papillae are very numerous,
and their outlines are quite distinct. At the lower left-hand corner of the
figure there is no special arrangement of the papilla, though the position in
question is situated above the leg. The position marked m. d./. represents
the mid-dorsal line.

Fie. 11.—The inner (lla) and the outer (114) blades of the jaws of
Koperipatus Weldoni. Note that there are two denticles on the inner
side of the large tooth in each blade. ‘The inner blade has a diastema, fol-
lowed by a number of smaller denticles.

PLATE 34.

Kie. 12.—The inner (12a) and the outer (124) blades of the jaws of Eo-
peripatus Horsti. Note that they have almost the same characters as
those figured on the previous plate, though they are different in size.

Fic. 13.—This figure represents a portion of the skin of the dorsal surface
of Eoperipatus Horsti, and should be compared with Fig. 10 on the
previous plate. Note the well-defined transverse ridges and grooves. The
primary papille have not the sharp, irregularly shaped outline found in

554 RICHARD EVANS.

Hoperipatus Weldoni. The accessory papillae, which are here much less
numerous, and are scarcely ever found in the grooves, lose their individuality
in that of the transverse ridges which rise up suddenly from the general sur- ~
face. ‘The middle of the back is occupied by a clear narrow line. At the
lower right-hand corner of the figure is represented the special arrangement
of ridges and grooves which exists in relation to the leg. This should be
compared with the lack of arrangement in the lower left-hand corner of
Fig. 10.

Fie. 14.—Ventral view of the anterior end of Eoperipatus Horsti.

Hie. 15.—Ventral view of the posterior end of Hoperipatus Horsti
(female). .

Fic. 16.—Ventral view of the posterior end of HKoperipatus Horsti
(male).

Fic. 17.—This figure represents an antero-ventral view of one of the feet
of HKoperipatus Horsti. Note the two distal papille and the four ventral
prominences or ridges. ‘The two distal papilla, asin K. Weldoni, are divided
into basal and apical portions, and carry a spine. The ventral prominences
also carry a spine.

l'te. 18.—Postero-dorsal view of the same foot, as shown in Fig. 17.

Fie. 19.—A primary papilla from the dorsal surface of Koperipatus
Horsti.

Fig. 20.—A primary papilla from the dorsal aspect of one of the legs of
Koperipatus Horsti. Note the same difference between the basal por-
tions of the papillae shown in Figs. 19 and 20, as was noticed between those
represented in Figs. 7 and 8.

Fic. 21.—Ventral view of the fourth and fifth legs of Hoperipatus
Horsti. Note the position of the papilla which carry the renal openings
as compared with that which occurs in Fig. 9.

Fres. 22 and 23.—These figures represent two dissected females of Ho-
peripatus Weldoni. Note the difference in the arrangement of the uteri
in the two figures. In Fig, 22 the uteri, full of embryos, are packed near the
posterior end, under the loop of the intestinal canal, while in Fig. 23 they are
arranged on the right side, and extend forward as far as the first pair of legs.

PLATE 35.

Fie. 24.—The renal organ of the second pair of legs of Koperipatus
Horsti. There is neither a terminal bladder nor a differentiated funnel. The
ccelomic end-sac is well developed.

Fic. 25.—The renal organ of the fourth pair of legs of Hoperipatus
Horsti. ‘There is no terminal bladder, but the funnel is well developed. The
renal duct, as a whole, is very long and coiled.

TWO NEW SPECIES OF ONYCHOPHORA. BSN)

Fie. 26.—The renal organ of the ninth pair of legs of Eoperipatus
Horsti. All the parts of a typical renal organ are represented here, namely,
a dilated bladder, a coiled duct, a well-developed funnel, and a ecelomic end-
sac. The dilated bladder above mentioned is not close to the renal pore.

Fie. 27.—The renal organ and the crural glands of the fourth last pair of
legs of Hoperipatus Horsti (male).

Fig. 28.—The renal organ and the crural glands of the third last pair of
legs of Eoperipatus Horsti (male).

Fie. 29.—The funnel and coelomic end-sac of the renal organ of the fourth
pair of legs of Eoperipatus Horsti.

Fie. 30.—A section through the fourth leg of Eoperipatus Horsti.
Note especially the coelomic end-sae (c. ¢. s.), the funnel (/.), and the renal
duct, which passes down the leg to the papilla situated on the proximal side
of the fourth pad.

Fie. 31.—A section through the outer portion of the renal duct shown in
Fig. 26. Note the difference in the arrangement and character of the cells.
In the upper portion, which represents the wall of the biadder, the cells are
Jarge and extended, and their nuclei are far from one another. In the lower
portion the cells are small and columnar, and their nuclei are closely packed
together, as they are in the layer of cells covering the external surface. Note
especially that there is vo gradual transition between the two portions above
mentioned, but that there is a sharp rim.

Fic. 32.—A section of the duct of the salivary gland of Koperipatus
Horsti. Note the difference in the characters of the lining cells. The tall
columnar cells with large nuclei at their free end line the portion nearest the
buccal cavity. The short cells with small nuclei line the portion nearest the
gland.

Fic. 33.—This figure represents a transverse section passing through the
anterior edge of the female genital orifice of Eoperipatus Horsti, the
actual opening being found in the third section from the one drawn. The
uteri (w/.) pass on the outer side of the nerve-cords (w. ¢.), and are full of
spermatozoa (sp.). The nerve-cords (x. ¢.) are widely separate. The heart
(4.) is still a well-developed tube lying in the pericardium. The section
passes through one of the ridges of the skin—a fact which explains the pre-
sence of the large number of papille.

Fie. 34.—This figure represents a transverse section passing immediately
in front of the last pair of legs of the female of Eoperipatus Horsti.
From the nerve-cords (x. ¢.), which are still widely separate, nervous strands
(z.s.) pass to the ventral organ (v.0.). The heart no longer exists. The
most noticeable feature of the section is the large renal organ (r. 0.) with its
coelomic end-sac (ce. é. s.).

536 RICHARD EVANS.

PLATE 36.

Fic. 35.—This figure represents a transverse section passing immediately
in front of the penultimate pair of legs of E. Horsti. The left vas deferens
(J. v. d.) is cut across twice, the right one (r.v.d.) passes under the nerve-
cord (x. c.), and is cut across twice on the right side. The nerve-cords (x. ¢.)
approach each other below the rectum (r¢.), and present a kind of a ganglionic
swelling situated immediately in front of the male genital orifice. In the
dorso-lateral part, in cavities of their own, are seen the male accessory glands
(m.a.g-). The ductus ejaculatorius with its much-thickened wall appears on
ihe left side immediately under the left nerve-cord. In order to understand
the topographical relations of the male genital ducts and the nerve-cords,
Figs. 835—87 should be compared.

Fic. 36.—An enlarged representation of the male genital organs of Ko-
peripatus Horsti. The small numbers (38—48) placed outside the figure,
and at the ends of lines crossing the genital ducts in various positions, refer
approximately to the positions of the sectious drawn in Figs, 88—48. ‘The
rectum (ré.), which passes between the coils of the vasa deferentia, is repre-
sented as having been cut close to the posterior end of the stomach on the
one hand, and near the anus on the other. Note that the ductus ejacula-
torius is drawn out into along loop on the left side, the lines bearing the
numbers 44—48 cross it at various places. ‘The male accessory glands
(m. a.g.) are shown as two small tubes situated between the terminal end of
the ductus ejaculatorius and the posterior end of the rectum.

Fic. 37.—This figure shows the relations of the nerve-cords, which are
represented in black, to the various organs shown in Fig. 36, which in the
present figure are drawn only in outline.

Fic. 38.—A transverse section of the testes of Hoperipatus Horsti.
See number 38 in Fig. 36.

Fie. 39.—A transverse section of the duct passing from the testes to the
seminal vesicle of Hoperipatus Horsti. See number 39 in Fig. 36.

Fic. 40.—A transverse section of the vas deferens, close to the seminal
vesicle, of Hoperipatus Horsti. See number 40 in Fig. 36.

Fic. 41.—A transverse section of the much-coiled vasa deferentia of
Eoperipatus Horsti. The tube marked s. g. represents one of the branches
of the slime-gland. In the tubes on the left of the figure are several loosely-
arranged spermatozoa, but in those on the right they are arranged in a compact
and characteristic way round a common centre. See number 41 in Fig. 36.

Fie. 42.—A transverse section of the vasa deferentia of Hoperipatus
Horsti. Theducts havea common sheath, but their lumen are still separate.
The spermatophores are completely formed. See number 42 in Fig. 36.

Fic. 43.—A transverse section of the common genital duct close to the

TWO NEW SPECIES OF ONYCHOPHORA., 537

union of the vasa deferentia. The spermatophores have not yet fused. See
number 48 in Fig. 36.

Fic. 44.—A transverse section of the common genital duct at some distance
from that shown in Fig. 43. The spermatophores have fused together, but
in the arrangement of the spermatozoa there are signs of a double origin. See
number 44 in Fig. 36,

Fie. 45.—A transverse section of the descending limb of the common
genital duct of Hoperipatus Horsti, The head of the spermatophore has
been cut through twice, and the spermatozoa are arranged in two groups
round two centres. ‘The wall of the duct is as yet comparatively thin. See
number 45 in Fig. 36.

PLATE 37.

Fic. 46.—A transverse section of the ductus ejaculatorius of Eoperi-
patus Horsti. The wall is greatly thickened, but the lining cells are tall
and columnar. See number 46 in Fig. 36.

Fic. 47.—A transverse section of the ductus ejaculatorius of Eoperi-
patus Horsti. The wall is very thick, but the lining cells, though columnar,
are much shorter. See number 47 in Fig. 36.

Fie. 48.—A transverse section of the ductus ejaculatorius of Eoperi-
patus Horsti. The wall is extremely thick, and the lumen is correspond-
ingly small. See number 48 in Fig. 36.

Fic. 49.—This figure represents a dissected-out ovary, ete., of Koperi-
patus Weldoni, It shows the ovary enormously spread out, one recep-
taculum ovorum (re. 0.), and two receptacula seminis (ve. s.), as well as the
commencement of the two uteri. The ova are also seen suspended in thin-
walled sacs, which hang freely in the vascular body-cavity. The lines bearing
the numbers 50, 51, and 52 mark the position of the sections shown in Figs,
50, 51, and 52.

Fig. 50.—This figure represents a section along the line marked 50 in Fig,
49. It passes through the receptaculum ovorum (ve. 0.) and through one
oviduct (ovd.). The other oviduct is not cut across, as the section goes
through the opening passing from the extended cavity (cav.) of the ovary to
the oviduct.

Fie. 51.—This figure represents a section through the receptaculum
seminis (ve. s.) and the proximal end of the uterus (w/.). Note the highly
columnar lining of the latter, and the comparatively thin wall of the former.
One of the ducts passing to the receptaculum seminis is shown embedded in
its wall, and lined by cells with small nuclei.

Fig. 52.—This figure represents a section across the uterus at the position
marked by the line bearing the number 52 in Fig. 49. Note the comparatively
thick wall and columnar lining.

von. 44, pART 4.—NEW SERIES. MM

538 RICHARD EVANS.

Fic. 53.—This figure represents a section across the uterus in the neigh-
bourhood of one of the embryos contained in it. Note that on one side the
lining cells are columnar, while on the other side they are becoming flattened ;
that is, there is a gradual transition from the tall and columnar condition to
the flattened one. There is no destruction or loss of cells in any way.

Fie. 54.—This figure represents a transverse section of one of the uteri of
Eoperipatus Weldoni. The section is from close to the posterior end,
and should be compared with the representation of the uterine walls of Ko-
peripatus Horsti shown in Fig, 33.

EOPERIPATUS BUTLERIT. 539

Eoperipatus Butleri (nov. sp.).!
By

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

With Plate 38.

—E

ConTENTS.
PAGE
I. Introduction - : : F : soad
Il. Description of KE. Butlers ; 3 : : . 540
(a) Colour : é : é : : . 540
(6) Dimensions . : : . 540
(c) ‘The characters of tle skin. : : : . 540
(d) The median external openings , : ; = OAD
(e) Antenne, jaws, oral papille, legs, and feet . : . 542
IIT. Conclusion . : : E ; . 544
Explanation of Plate : : ; ; - . 544

J.—Intropvuction.

THE material described in this paper consists of one female
specimen, which was kindly sent me by Mr. R. I. Pocock, on
the suggestion of Professor Ray Lankester, Director of the
British Museum (Natural History). Mr. Pocock obtained it
from Mr. A. M. Butler, Curator of the Museum at Selangor,

' This short paper is considered as a supplement to the foregoing memoir
**On the Two New Species of Onychophora from the Siamese Malay States,”
rather than as an independent paper, and must be read in connection with
that memoir ; for in the description given here of E. Butleri, especially in
comparing it with the other species, constant allusion is made to the facts
recorded in the above-mentioned paper, in which alone a list of references is
given,

540 RICHARD EVANS.

Straits Settlements, who discovered it on Larut Hills at the
height of 4000 feet. In accordance with a suggestion made
by Mr. Pocock, I have given it the specific name Butleri in
honour of its discoverer.

IJ.—Descriprion oF HKOoPERIPATUS BUTLERI.

(a) Colour.—The dorsal surface is coloured dark brown,
with pale spots scattered about with a certain amount of regu-
larity over the whole of the animal’s back. ‘These spots
represent the large primary papille, which contain less pig-
ment than the other parts of the skin, and which in almost all
cases have lost their apical part. The mid-dorsal position
is occupied by a dark chocolate-coloured line, which extends
from the region of the first pair of legs nearly as far back as
the anus. When the skin is examined with the microscope a
narrow, non-pigmented line is seen to occupy the centre of the
dark line, as in the other Malay forms. ‘he colour of the dorsal
surface is remarkably like that of Hoperipatus Weldoni.
The colour of the ventral surface is shghtly paler than that of
the dorsal, though the difference is in no way well marked. ‘The
seomentally arranged spots, which correspond to the ventral
organs, are very evident, owing to their yellowish white ap-
pearance. Hoperipatus Butleri contrasts strongly with
BH. Weldoni as regards the colour of the ventral surface. In
the latter it is grey with sparsely scattered brownish spots,
while in the former it is not very different from the dorsal
surface. It also contrasts with the greater number of my
specimens of the species Horsti, for it absolutely lacks the
pink found in that species.

(b) Dimensions.—The specimen here described measures
52 mm. in length, 6 mm. in greatest breadth, and 5 mm. in
greatest dorso-ventral diameter.

(c) The Characters of the Skin.—The skin is thrown
into folds, of which there are about fourteen to each sezment
in the middle part of the body. The folds, when examined
with a hand lens, seem continuous across the back, but when
looked at through the microscope they are seen to be divided

EOPERIPATUS BUTLERI. 541

by a narrow, non-pigmented line similar to that found in other
Malay species. The folds of the skin exhibit a varying degree of
continuity on the latero-dorsal aspect, for those which corres-
pond to the intervals between the legs are continuous, while
those which are situated above the legs are discontinuous and
often break up, so that the papille display a diffuse arrange-
ment in patches of varying shape and size. Comparison of
figs. 4 and 5 will probably help the reader to understand the
difference and appreciate the distinction which has been made
above.

The conformation of the ridges and the arrangement of the
papillze on them near the mid-dorsal line resemble those found
in K. Weldoni, in which the folds rise up gradually, and are
covered with primary and accessory papillae. The primary
papillze are few in number and extend over the greater part
of the width of the ridge or fold, while the accessory ones are
very numerous and occupy only a portion of the width. The
grooves between the folds are somewhat wider in EK. Butleri
than they are in KH. Weldoni. On the sides between the
successive pairs of legs the folds display exceedingly different
characters from those which have been described above. In this
position they resemble much more closely those of KE. Horsti
than of EK. Weldoni. They rise up suddenly, and are pro-
vided almost exclusively with large primary papille. The
accessory papille, which are few in number in the region
under consideration, tend to lose their individuality in that
of the ridges, a feature which seems to accompany the forma-
tion of folds of this particular kind, for it also occurs in
HK. Horsti. ‘The primary papille are provided with apical
portions, which are differentiated from the basal ones. In
K. Weldoni and K. Horsti, as a rule, the apical portion is
either oval or subspherical in form, and is comparatively
large in size, but in the species Butleri the apical part is
conicalin shape and small in size. Even on the dorsal aspect
of the leg, a position in which the apical part attains its
greatest development in KH. Weldoni and K. Horsti, it is
almost impossible to see it in H. Butleri.

542 RICHARD EVANS.

(d) The Median External Openings.—There are three
of these openings, namely, the mouth, the genital orifice, and
the anus.

The Mouth:—The mouth is in an exceedingly extended
condition, and consequently all the papillae surrounding it are
clearly seen (PI. 38, fig. 1). The incomplete circle of inner
papille consists of four pairs symmetrically arranged on either
side of the anterior moiety of the mouth. ‘The outer circle,
which is complete, has a pair of papillz symmetrically situated
in front of the mouth, and a median one behind. ‘The latter
seems to be composed of three papille which are united to-
gether. ‘he papillz placed laterally to this median one are
specially enlarged, and, lke those of the incomplete ring
situated in front, form the boundary of the buccal cavity.
Owing to the extruded condition of the mouth it is possible
to make out three pairs of papille in the interior of the
buccal cavity behind the jaws. ‘The tongue, situated in front
of and between the jaws, carries a number of complex
denticles. All the papille surrounding the mouth are pro-
vided with one or more spines.

The Genital Orifice.—The genital orifice is situated
between the penultimate pair of legs. It has the form of a
transverse opening surrounded by tumid lips, made up of
numerous white papille, similar to those which surround the
buceal cavity.

The Anus:—The anus is a slit-like aperture, situated at
the terminal end of the short anal cone. It inclines towards
the dorsal surface rather than towards the ventral, a result
brought about probably by contraction.

(e) Antenne, Jaws, Oral Papille, Legs and Feet.
The Antenne:—The antennze present the same general
characters as in the other species of the genus Hoperipatus.
The club-shaped appearance of the distal extremity is much
less marked. The number of rings that can be counted with
certainty is forty-seven, as in HE. sumatranus and in some
of the specimens of HK. Horsti and EH. Weldoni. It seems,
however, that there are a few small rings intercalated among

EOPERIPATUS BUTLERI. 543

the larger ones in the distal third, which would bring the
number of rings—both large and small—up to fifty or fifty-
one.

The Jaws.—The outer blade of the jaws has the same
structure in E. Butlerias in HK. Horsti and E. Weldoni;
that is, there are two small denticles on the inner side of the
main tooth (fig. 6). The inner blade, however, differs from
that of the above-mentioned species in that it has three small
denticles, instead of two, between the main tooth and the
diastema, as well as fourteen smaller denticles on the inner
side of the diastema instead of the nine or ten found in E.
Weldoni and H, Horsti (fig. 7). The jaw-blades are larger
in K. Butleri than in the other species.

The Oral Papillze.—The oral papille are in no way
peculiar. They consist of two rings which do not carry
papillz, and of an end-knob which is provided with papille
mainly on the dorsal aspect. ‘The opening of the slime-gland
is slightly sub-terminal.

The Legs.—There are twenty-four pairs of legs, which
are arranged as in the species Horsti, with almost the same
distance between the successive pairs of feet along the whole
length of the body, with the exception of the last two or three
pairs; a feature which distinguishes KH. Butleri, even ata
glance, from EH. Weldoni, in addition to the fact that the
legs are shorter and stouter. The legs, with the exception
of the last two pairs, are provided with four pads, and many
of those situated behind the fifth pair have a vestige
of an additional one (fig. 2). ‘The penultimate pair has only
three pads, which are reduced on the last pair to two.
On neither of the last two pairs of legs are the pads well
separated from one another.

Crural grooves occur on all the legs, but are least developed
on the first pair, where they are hardly visible. On the distal
angle of these grooves there is a whitish structure, which may
consist of two papillz lying close together, or of an U-shaped
body formed by the fusion of the two papille on their distal
side. ‘I'hey seem to occur on all the legs, though owing to

544. RICHARD EVANS.

the contracted state of the latter, in many cases they are
drawn into the grooves.

The Feet:—The feet have almost the same structure as
in the other species belonging to the genus Eoperipatus.
They carry only two primary papillz on the distal margin,
one in front and one behind. Hach papilla consists of a basal
and an apical part, the latter being provided with a pointed
spine. The ventral elevations or ridges are not so well
marked as in the species Weldoni and Horsti. The proxi-
mal pair of these elevations agrees with those of the above-
mentioned species in that they carry only one spine, but the
distal pair differs, for they are provided with two spines to
each elevation (fig. 2). Lest the second spine should have
been missed, a renewed examination of the feet of both
Weldoni and Horsti was made, but only to confirm the
conclusion previously reached.

III. Conciusion.

In conclusion, my best thanks are due to Professor Lan-
kester and to Mr. Pocock, who sent me for examination the
first specimen obtained in this country from the Malay
Peninsula, of the species here described, and for allowing me
to dissect it as far as was necessary to determine its specific
characters.

EXPLANATION OF PLATE 388,

Illustrating Mr. Richard Evans’ paper “On Hoperipatus
Butleri” (nov. sp.).

All the figures were traced with the camera lucida.

Fic. 1.—This figure represents the mouth opening and the papillae which
surround it. It is in a beautifully expanded condition, and will serve to show
the general arrangement of the papilla round the mouth in all the Malay

EOPERIPATUS BUTLERI. 045

species which agree with one another to a wonderful extent as regards their
accessory structures. The figure was carefully traced with the camera lucida
by the author, and very carefully finished by Mr. P. Bayzand.

Fic. 2.—This figure represents the latero-ventral aspect of the sixth leg of
the right side. Note the presence, in addition to the usual four crescentic
pads, of a vestige of a fifth pad; the two papille on the distal margin of the
foot, and the four ventral prominences. Note that each prominence of the
distal pair carries two spines, those of the proximal pair being provided with
only one each,

I'ic. 3.—This figure represents the crescentic pads of the fifth leg of the
right side. Note that there is no trace of a fifth pad, and that the renal
papilla is situated in the fourth pad, and divides it into two halves.

Fic, 4.—This figure represents a portion of the skin of the dorsal surface.
Note the narrow, clear line which occupies the mid-dorsal position; the large
primary papille which stretch almost across the ridges from one groove to
another; the numerous accessory papille which are scattered among the
primary ones, and which follow no definite arrangement; and finally, the
narrow fold with small papills which fails to reach the mid-dorsal line.

Fic. 5.—This figure represents a portion of the skin taken from the area
situated immediately above one of the legs, the upper limit of which is shown
at the lower end of the figure, where the papillx are arranged in oblique rows.
Note that the ridges above the leg break up into patches. Compare the two
ridges, one on either side of the figure, with those shown in the previous
figure, and note the gradual diminution in number of the accessory papille.

Fic. 6.—This figure represents the outer blade of the jaw.

Fic. 7.—This figure represents the inner blade of the jaw. Note the three
small denticles situated on the inner side of the main tooth, and the great
number of smailer denticles on the inner side of the diastema. Note the large
size of the jaw-blades as compared with those of E. Weldoni, and especially
of K. Horsti.

ns

etna palais? Geilo dale a Ad Dp er treed yeh
Pieter ty ber

oes Ae ety F,

EOE ab oe y ira tas verve ihe

AL ee ThE 7) gk OTL anya ra.

a ea a hadaw sdvigeateges Gul oar stag Rah ;
RES Hyp rear eats, Hoek Welly Gh. ,2s abonaay! fens tng? alk ds
Renee usenet Hey Lait anny, el rw eet biperyrt i
e aa f se |

sift ie oH aah. : ut JG@evas NGS Stole wee “pa a
Mary 23 bork paul ee ; fea a ta amis eyo! upae her 8 S 4

eerie 228 one | Aye Tyee fed tet Tea paleaitp

[OPS ite Hae Hd ATA eee Pe Asn ; eee Ogee fH
MS WT. PAS OEE “ae ergs fede se, WAR Ning
je 44 & a Te AT Sac inliei ale aay eee

i to@aest hia dit MIG -) Ge ma, Ge winntl deoelt?

wuss 7 t y nv = ‘ipey Teale
4 t 7 ale D
AR pab 4 ? BAAAD Th H Wists ) ‘4 ‘ a
: oe rs ry ; 7
Wri Vico Meh Tinie Dee ip Al aH ; rh
’ a t 5 L 2 Dey
wee, ui Tergteh era abittess od arith es pu ings atl
{
ort ri at i yet 2G Pai ca in pt Ait wid towel eo? WM afte 4
* i Ay _
wives, Pie Be kode Most Liye ~mee wel ee igs is Cyt

; o - . , . ak
etre ek, F Re ager oi Cae Ti. ae ka Pers CP ij pets af ee tomes
‘ea r ‘ =7 ;
sont sila VOGUE vee bay atl lay eagle ee

*
ord sei) a Gh Gh ee Nak s\01G): a
Hy, aoe fie we. ober t , ay og iN
4 1
4 ‘ ni. a Oita, 3 CP) oe et iawn i i i
4 +
Orere Mae . aay FR cok, We } a & Mii ia
j iy
e
ty
oy

TWO NEW BRITISH NEMERTEANS. 547

On Two New British Nemerteans.
By

R. C. Punnett, B.A.

With Plates 39 and 40.

THE two new species of Heteronemerteans described below
each form a new genus. Both were foundat Plymouth. The
first, Micrella rufa, was found by myself whilst digging
in the mud of the river Yealm at low water mark. Amongst
other forms dug up at the same time may be mentioned
Carinella superba, Nemertes echinoderma, together
with species of Synapta, Nephthys, Solen, and Capitel-
lide. For two preserved examples of the second form I am
indebted to the kindness of Mr. W. I. Beaumont, who dis-
covered it among some dredge material from near the
Mewstone, and with whose name I have much pleasure in
associating it. As will appear below, it has seemed advisable
to create anew genus to receive it. Consequently its full name
appears as Oxypolia beaumontiana. A third and larger
example has recently been sent to me alive from the same
locality. By means of it I have been able to confirm the obser-
vations Mr. Beaumont made upon his specimens when alive.

Micrella rufa belongs to the family of the Lineide,
Oxypolia beaumontiana to that of the Kupoliude.

Fam. LINEIDz.

Micrella, nov. gen.

Body elongated, slender and dorso-ventrally flattened pos-
teriorly. No side folds present. Caudal appendage present.

548 R. C. PUNNETT.

Rhynchoccelom to posterior end, and with pockets in the
cesophageal region. Proboscis two-layered and with muscle
crosses. Hxcretory system with long duct and a single pair
of openings at the posterior end. No neurochord cells. Side
organ present just behind excretory pore.

Micrella rufa, n. sp.

Two specimens were obtained, one of which lacked the
anterior end. The perfect specimen was about 18 cm. long
when alive and extended, and about 2—3 mm. broad. In
colour it was of a bright vermilion, shading off into yellow
near the anterior end. Through the orange-coloured head
the brain showed bright red. In the intestinal region the gut
and its pockets showed brown through the body-wall. The
head was somewhat pointed in life, a feature which became
more marked as the animal essayed to burrow through the
bottom of the glass vessel which contained it. The larger
and imperfect specimen was probably, before the severance
of the anterior end, about twice the size of the above.

The epithelium is crowded with unicellular glands,
which stain readily with picric acid. ‘They are wanting only
on the tip of the head, in the head slits and side organs,
on the ventral surface of the caudal appendage, at the junc-
tion of the caudal appendage with the trunk, and in sundry
patches ventrally near the posterior end. ‘The epithelium
rests on a fine basement membrane. Beneath this is an ex-
ceedingly delicate layer of circular muscle-fibrils. he large
cutis glands lie in the outer longitudinal muscle layer, and in
the cesophageal region they reach inwards as far as the
delicate nervous sheath surrounding the circular muscle layer
(fig. 2). Shortly after the cesophageal region they entirely
disappear.

The muscular system in front of the brain consists
mainly of longitudinal fibres. hose directly surrounding
the rhynchodeum and cephalic vascular lacune are sur-
rounded by a thin layer of circular muscle. Of the three
muscle layers in the cesophageal region, the inner longitudinal

TWO NEW BRITISH NEMERTEANS. 549

is the thickest, and this relation obtains throughout the body.
In the posterior cesophageal region the alimentary canal is
completely surrounded by a layer of longitudinal muscles,
separated off from the inner longitudinal layer. The inner
longitudinal layer also separates the circular layer and the
proboscis sheath. Dorso-ventral fibres occur in the posterior
cesophageal region as well as between the intestinal pouches.
The last are, however, but poorly developed. There are a
few horizontal fibres above the mouth. The outer longi-
tudinal layer is feeble in the cesophageal region, owing
possibly to the great development of the cutis glands. It
becomes more strongly marked in the anterior part of the
intestinal region, but entirely disappears towards the pos-
terior end of the animal. The caudal appendage contains
prolongations of the circular and internal longitudinal layers
(fig. 8).

The proboscis sheath possesses an outer circular and an
inner longitudinal layer of muscles. It extends throughout
the length of the animal, though it does not reach into the
caudal appendage. In the cesophageal region occur diver-
ticula (figs. 1 and 11, rhe. p.) from the proboscis sheath
in which the muscle layers are absent. These diverticula
are closely embraced by the lateral vascular lacune in this
region. They are crowded with large rhynchoccelomic cor-
puscles, which are oval in shape and greatly flattened (fig. 9).
Kach contains a nucleus in which the chromatin is arranged
in four small circular masses, all connected by a more or less
circular thread. Similar corpuscles are to be found in the
cavity of rhynchoccelom, and also in the vascular lacune in
this region. The proboscis is not long. In its middle
portion it is composed (when retracted) of an outer longi-
tudinal muscle layer directly beneath the rhynchoccelomic
epithelium, containing two muscle crosses (fig. 4) formed by
fibres from the thinner circular layer directly beneath it.
Beneath this again is the high and glandular proboscis
epithelium. Just inside the circular muscle layer are several
nerves on either side; that is to say, if the proboscis is

550 Rk. CG. PUNNETT.

so orientated that the muscle crosses are dorsal and ventral,
the nerves are then lateral in position. There is no continuous
nervous layer such as occurs in the majority of the group.

The alimentary canal presents no features of special
interest in its structure. ‘The cesophageal epithelium contains
unicellular glands, and unicellular glands also form a layer
round it. In the intestinal region the gut pockets are deep,
and there is a well-marked ventral gutter. The gut pockets
are continued to the anus, which is a comparatively large
opening at the posterior end of the body on the dorsal
surface just in front of the caudal appendage. For the last
millimetre or so the alimentary canal and its pockets are
devoid of gland cells. The whole canal is richly ciliated
throughout.

The vascular system (fig. 11) in the snout consists of a
large lacuna which divides just in front of the brain. At the
level of the brain commissures these unite ventrally, and then
again divide into two lateral and a median dorsal vessel. The
lateral vessels form lacunz round the cerebralorgan. At this
level they again communicate by the buccal commissure,
though no buccal vessels are formed. Behind the cerebral
organ the lateral vessels pass backwards to the cesophagus,
where they form the cesophageal lacune characteristic of the
order. This lacunar network is co-extensive with the excretory
tubules. Assoon as the tubules cease the ceesophageal lacunze
are gathered into a very large lacuna on either side (fig. 1).
This condition lasts until just after the level of the excretory
pore, when the lacunes become constricted and surrounded with
the peculiar parenchymatous tissue found in the rest of the
Heteronemertini. ‘he median dorsal vessel runs in the pro-
boscis sheath until the level of the excretory pore,! when it
emerges and becomes surrounded by parenchymatous tissue
like the lateral vessels. In the intestinal region the lateral

1 Ina previous paper I have already drawn attention to the curious fact
that the dorsal vessel almost invariably leaves the proboscis sheath at the
level of the hind end of the excretory system, whatever may be the extent
of the latter, (‘Quart. Journ. Mier. Sci,’ vol, 44, p. 136.)

TWO NEW BRITISH NEMERTEANS, 551

vessels communicate in the usual way with the median dorsal
vessel. Just in front of the caudal appendage the median
dorsal vessel ends, while the lateral vessels form a ventral
commissure, from which arise two minute vessels. ‘hese soon
fuse, forming a cord of cells which is continued into the caudal
appendage.

The excretory system consists of a duct on each side,
into which run a number of tubules. The tubules lie in close
relation with the cesophageal lacunz, and extend both dorsally
and ventrally to the level of the nervous side stems. After
the tubules come to an end the large duct is continued back-
wards, and opens by a single pore on either side just above the
side stems. It is remarkable that at least half the total ex-
tent of the excretory system is taken up by the large duct
unaccompanied by any excretory tubules.

The gonads were in each case testes alternating with the
intestinal pouches. The ducts are much nearer to the median
dorsal line than to the side stems. No gonidial pouches are
found in the caudal appendage.

The nervous system is formed on the usual Lineid type.
Of the four different kinds of ganglion cells enumerated by
Birger, all are present with the exception of the neurochord
cells. ‘‘hese are also absent from the side stems. The
cerebral organ is not very strongly developed. Its glandular
epithelium reaches forwards dorsally over the hinder part of
the dorsal ganglion (fig. 10). The epithelium of the ciliated
canal is not so highly differentiated as in the rest of the
members of the family in which it has been described. The
large and characteristic cells found on the external side of
canal are not present in Micrella, the whole canal being lined
by epithelium similar to that found on the inner side of the
ciliated canal of other Lineidze. The head slits are not deep,
extending only halfway to the brain. They end abruptly at
the level where the ciliated canal comes off.

With regard to the other sense-organs, both eyes and
frontal organ are absent. There is, however, a lateral sense-
organ on either side (fig. 2) shortly behind the excretory pore.

552 R. C. PUNNETT.

In the preserved animal it is conspicuous as a small longi-
tudinal slit (fig. 5) about ‘75 mm. long on eitherside. It is
lined with characteristic glandular epithelium resembling that
found in the head slits (fig. 6).

The head glands are feebly developed.

The foregoing account shows that Micrella presents several
features which separate it from the rest of the Lineidee, and
it may be profitable to consider them in rather more detail.
In his monograph (5, p. 715) Biirger derives the Hetero-
nemerteans from such Protonemerteans as Carinella. The
Carinellidze are characterised by a side organ in the neigh-
bourhood of the excretory pore, a feature which is shown
only by Zygeupolia (7, p. 151) and Micrella among the
Heteronemerteans. Its position and structure in the last-
named forms would lead us to infer that it is homologous in
both cases with that in Carinella. The excretory system
again in Micrella, whilst typically Heteronemertean in the
arrangement of the tubules closely connected with the
cesophageal lacune, resembles that of Carinella in the size
and length of the main duct, and in the single pair of very
posteriorly situated pores (cf. Birger [5], pl. xxviii, fig. 2).
The proboscis also shows a Protonemertean feature in the
absence of a continuous nervous layer and the presence of
but two muscular layers. It differs, however, from that of a
Carinella in having muscle crosses, a feature hitherto only
found among the Heteronemerteans.

The cutis, again, is not so highly differentiated asis usually
the case in the group where the outer longitudinal muscle
layer is usually separated by connective tissue from a cutis
containing muscle fibrils and glands. It is hardly possible to
speak of a cutis in Micrella, which shows a condition
similar to that described by Biirger! for Lineus lacteus (5,
p. 621, and pl. xxii, fig. 37).

With regard to the vascular system also the cesophageal

! A somewhat similar condition occurs in a fragment christened Cere-

bratulus medullatus by Hubrecht (vide ‘‘ Nemertea,” in ‘ Challenger
Reports,’ vol, xix, p. 39, and pl, xii, fig. 10).

TWO NEW BRITISH NEMERTEANS. 553

lacunee have not nearly so great an extent as in the majority
of the group, where they reach almost or quite to the intestinal
region.

Again, the epithelium lining the ciliated canal shows a
lower degree of specialisation than is the case in any other
Lined.

Though possessing many apparently primitive features
tending to connect the Lineidee with the Carinellide, Micrella
yet shows evidence of specialisation in other organs.

The extent of the rhynchoccelom over the whole length of
the body is a feature common enough in the Lineidze, though
never found in a Protonemertine. The curious rhyncho-
coelomic pockets find their closest parallel in those of certain
Metanemertines, such as! Drepanophorus, though the intimate
connection established between rhynchoccelom and_ lateral
blood lacuna seems to suggest a comparison with the lateral
rhynchoccelom vessels found in Carinella, Carinoma, and
Cerebratulus (cf. Biirger [5], pl. xii, fig. 7; pl. xiv, fig. 4;
pl. xxii, fig. 6). Whilst, however, in all these cases the blood-
vessel projects into the rhynchoccelom, in Micrella the
rhynchoccelom projects into the blood lacuna.

A caudal appendage is a feature characteristic of many
Lineide, though what its significance may be is very doubtful.
Birger regards it as ‘das stark und meist plétzlich verjiingte
hintere K6rperende”’ (5, p. 238), which apparently remains
in a more or less embryonic state, possibly reminiscent of
an ancestral condition in which the body was relatively much
longer. He holds that it contains prolongations of all the
organs and layers found in the intestinal region with the
single exception of the rhynchoccelom. Further, he finds
that the anus opens at its tip. Coe, however (4, p. 493),
found that the anus in Cerebratulus lacteus opened at
the posterior end of the body just below where the caudal

Somewhat similar diverticula have been recently described by Montgomery
for another genus of Metanemerteans—Proneurotes. Here however they arise
as ventral outgrowths of the proboscis sheath. (‘ Zool. Jahr. Abt. f. Syst.,’

1897, p. 4.)
VoL. 44, PART 4.—NEW SERIKS. NN

554 R. C. PUNNETT.

appendage joined it. He gives no account of the structure
of the organ, though one would suppose that it contained no
portion of the alimentary canal.

In Micrella, on the other hand, the appendage joins the
body ventral to the anus, and contains neither gonidial
pouches, intestine, nor outer longitudinal muscle layer, whilst
the vascular system in itis rudimentary. In the light of such
conflicting evidence it can only be conjectured that we are
probably not dealing with homologous structures in each case,
but that the caudal appendage in Heteronemerteans may have
an entirely different morphological significance—unless, indeed,
the anus is not homologous in the different members of the
group. Owing to the fact that the primary subdivisions of
the great fainily of the Lineide are based upon the presence
or absence of a caudal appendage, the study of its structure
in a number of forms would be of the highest importance
for the systematist, whilst at the same time it might be ex-
pected to throw some light upon the morphological signifi-
cance of a very puzzling and enigmatical formation.

One of the most interesting features connected with
Micrella is the light which it throws upon the relations of
the two Heteronemertean families—the Hupoliide and the
Lineide. In his monograph (p. 715) Birger has sketched a
family tree of the group. From it may be seen that he
derives the Lineide directly from a form such as Eupolia.
In the last-named genus we find an excretory system with
many ducts, such as occurs in many Lineide. Micrella
alone in this family presents a condition of this system ap-
proaching that of the Protonemerteans, from which all the
Heteronemerteans are probably to be derived. But Micrella
already possesses the characteristic head slits, consequently
we must suppose that the Lineidee branched off the common
stock before the type of excretory system with many ducts had
been evolved, and that this latter type has arisen independently
in the two families. The family tree given by Biirger must
therefore be amended somewhat in the way which the accom-
panying scheme indicates,

Or
qr

TWO NEW BRITISH NEMERTEANS. De

Rest of

Lineide.

| Fupoliide.
Mierella. |

Ne)

/

Whe
Heteronemertini.

Protonemertini.

Fam. Europ.
Oxypolia, nov. gen.

Short and stout in build, and with pointed head. Proboscis
pore ventral. Circular ciliated groove round kead just in front
of mouth. Rhynchoccelom to end of body. Excretory system
with many ducts. Cerebral organs small and not surrounded
by blood lacunee. Proboseis with three muscle layers, but
without muscle crosses.

Oxypolia beaumontiana.

‘The largest specimen of this worm, which was sent me in the
living state from Plymouth, measured about 12 em. in length
and about 5 mm. in breadth. When extended the whole
worm became greatly flattened. In contraction this was
much less noticeable, the anterior portion of the body
becoming quite cylindrical. The colour was pure white in the
anterior portion, whilst the intestinal region was of a pale
rose-colour. The following notes on the live animal were
made by Mr. Beaumont, who kindly gave them to me together
with the two specimeus which he had procured :—‘‘ Anterior
half milk-white (and this part is rounder in section than the
rest), whilst the remaining portion has a brownish look
about it, and shows more opaque rings, not very regular,

556 Rk. GC. PUNNETY.

about 3—4 mm. apart. Head shaped like a spear-head, but
not quite as wide as the succeeding portion of the body. It
is very flat, as is the animal throughout, especially the
posterior part of the body, which is almost oar-like. When
squeezed the brain was noted as a small yellowish mass in
front of the mouth opening. ‘The gonads were regularly
arranged. The proboscis extended over nearly half the
length.” On preservation the anterior portion becomes
eylindrical, though the posterior half of the body remains
somewhat flattened.

The epithelium is not high. It is crowded with nuclei,
and contains a number of unicellular glands which stain
vividly with picric acid. here is a fine but well-marked
basement membrane in the oesophageal region, beneath which
is a well-developed layer of circular muscle-fibres (fig. 17).
Underneath this again is a thick layer of gelatinous-like con-
nective tissue, which stains deeply with nigrosin, though
faintly with carmine, thionin, or picric acid. It contains a
number of small glands (fig. 15) whose contents stain deeply
with thionin, and whose secretion can, with the help of this
reagent, be traced through the epithelium ‘The layer bears
some resemblance to the gelatinous connective-tissue layer
found in Eupolia, though in that genus the cutis glands are
ageregated nearer the outer surface. In the intestinal region
the basement membrane disappears, the cutis glands become
smaller, and the connective-tissue layer more fibrillated in
appearance. ‘I'he epithelium of the circular head groove is
characterised by the absence of the unicellular glands and the
rich ciliation.

The muscle layers of the body-wall are well developed.
The internal longitudinal layer is thicker than the circular.
In the cesophageal region it forms a well-marked layer dorsal
to the alimentary canal, between the latter and the proboscis
sheath. It is also continued dorsally round the proboscis
sheath, completely separating the latter from the circular
muscle layer. The outer longitudinal muscle layer is con-
siderably thicker than either of the other two. There are no

TWO NEW BRITISH NEMERTEANS. 557

horizontal muscles above the mouth. The dorso-ventral
muscles are feebly developed.

The proboscis sheath extends to within a millimetre of
the posterior end. It is composed of the usual outer circular
and inner longitudinal muscle layers.

The proboscis is very stout and well developed. The
rhynchoccelomic epithelium which covers it externally (in the
retracted state) rests on a well-developed basement membrane
(fig. 14). This basement membrane is succeeded by an outer
longitudinal and a circular muscle layer, both of which are
very thin. Just inside the circular muscles is the nervous
layer, beneath which is the exceedingly thick ‘nner longi-
tudinal muscle layer upon which rests the thin and almost
aglandular epithelium of the proboscis.

As regards the alimentary canal, the mouth is behind
the brain and the anus terminal. In the cesophageal region
the epithelium contains but few glands. In the intestinal
region the epithelium is very granular. In the posterior part
of the intestine the epithelium is but slightly glandular,
whilst the intestinal diverticula become much shallower in
depth and less compressed.

The vascular system in front of the brain shows wide
lacunee dorsal to the rhynchodzeum as in the genus Eupolia.
These reach forwards in front of the proboscis pore (fig. 16).
The lateral vessels give off no diverticula embracing the cere-
bral organs (fig. 19, a—e, and fig. 22). There is a large
buccal commissure behind the commissure whence the median
dorsal vessel arises, but no buccal vessels arise from it. ‘he
rest of the system is on the usual Heteronemertean plan.
The cesophageal lacunar network extends to the beginning of
the intestinal region.

The excretory system closely resembles that described
by Biirger (5, p. 181) for Eupolia. ‘The tubules extend some
way dorsally and ventrally to the level of the side stems.
There are a number of ducts on either side (fig. 22), many
of which are incomplete, not piercing the circular muscle
layer. ‘The number of ducts is not, however, so great as in

558 R. GC. PUNNETT.

most species of EKupolia (cf. 8, pp. 116 and 120, 10, p. 577,
and 2, p. 44). ,

The gonads in both the specimens sectioned contained
minute ova in various stages. hey alternate with the
intestinal pouches, and open to the exterior by well-marked
ducts just above the side stems. The cavities in which the
young ova lie are not lined by any kind of epithelium, but
are merely somewhat indefinite spaces in the gelatinous
mesenchymatous tissue Probably the ova arise from the
mesenchyme cells, as has been suggested by Montgomery in
the case of Cerebratulus lacteus (6, p. 17).

The brain is somewhat high in comparison with its length
(fig. 20). The side stems form a well-marked ventral com-
missure beneath the anus. ‘The cesophageal commissure and
nerves are small. ‘he arrangement of the dorsal nerve
shows a peculiar feature (fig. 18). After rising from the
dorsal commissure it passes backwards for some distance
between the outer longitudinal muscle layer and the cutis.
It is not until the intestinal region is almost reached that it
dips down and joins the median dorsal thickening of the
nervous layer surrounding the circular muscle layer. The
median dorsal thickening just outside the circular muscle
layer is found in all Heteronemerteans springing from the
dorsal brain commissure there (Riickennerv of Birger [5, p.
363]). In Oxypoliathis nerve is well marked, -but does not
reach forwards as far as the brain (fig. 18, nd.). The
arrangement in this genus finds a close parallel in Cari-
noma armandi (95, p. 364, pl. xiv, figs. 4—8). Apparently
the so-called median dorsal nerve of other Heteronemerteans
must be regarded as containing two elements: (1) the true
median dorsal nerve springing from the dorsal commissure,
and (2) a specialised thickened portion of the nervous layer ~
surrounding the circular muscles. In Oxypolia both have
round them afew nuclei of what are apparently ganglion
cells. Just beneath the circular muscles in Oxypolia is
found the “ untere Riickennerv.”

The cerebral organ is small and considerably flattened

TWO NEW BRITISH NEMERTEANS. 559

both dorso-ventrally and in an antero-posterior direction
(fig. 19, b—e, and fig. 20). It is not embraced by a blood
lacuna. The epithelium of the ciliated canal (as is usual
among the Heteronemerteans) contains specialised large cells
externally, seven in number, as seen in transverse section.
The ciliated canal arises dorso-laterally from the ciliated
circular ring surrounding the head just in front of the mouth.

Neither eyes nor frontal organ are present.

The head glands are largely developed, as in Eupolia.
They reach backwards dorsally, and to a less extent ventrally
past the brain, lying in the outer longitudinal muscle layer.
Their substance stains deeply with thionin.

At the time when Biirger’s monograph was published the
Kupoliidee contained but three genera, viz. Eupolia,
Valencinia, and Poliopsis. Since then three other
genera (including the present one) have been added, viz.
Parapolia (Coe [4]), Zygeupolia (Thompson [7]), and
Oxypolia; consequently I have thought it advisable to add
a table showing the main differences presented by the six
genera which now form the family.

From this table it will be seen that Oxypolia holds a
position more or less intermediate between Valencinia and
Kupolia. It is more closely related to the former genus,
though, in addition to characters given above, it may be dis-
tinguished by the following:

(1) The body is shorter and stouter.in build than in Va-
lencinia. Moreover in Oxypolia the posterior portion is
not thicker than the anterior.

(2) There is no circular ciliated head furrow in Valen-
cinia.

(5) Whilst the head glands in Oxy polia exactly resemble
those of KHupolia, those of Valeucinia, according to
Birger (5, p. 186), are slighter in build, recalling those of
many Lineide.

(4) The cephalic vascular lacune in Valencinia forma
broken ring anteriorly (2, pl. i, fig. 53). In Oxypolia

they are quite horizontal as in Kupolia.

k. C. PUNNETT.

560

*£1e}U9 UL
-Svay atam suauteds sit yjoq yyy suorqueut “(61g “a ‘F)
‘OM AdYJO JY} WoaMgod pasodaozut sf akvy [RUOSLIp BeLaYM — LAABMOT, 901) ‘gjoooyouAyt oyg Jo quezxe oyg ur vipodneas
‘(qT “d ‘g) VULMTRASOURLaU “gq UT SEB ‘sdodUl aposnot Oy UBT | - fy, sojquasor vipodvavg yeyy soqeqs (4) UOswoUy, Sst z
atout ssassod ATyeuorsvooo Avur vitpodug Fo stosoqoad oy], ¢ ‘Tua AaUIAUOSOTY B Yq ‘UvazramauolezoH v you ATquqoad st ( 10z
‘(Tat-d 2) efopa| ad *h) YeoqwTOMT JO stUicosouLT MiuToueTeA’ euL eanhuy
-1poyd ‘gq ‘go ‘snuad siyq wr quosoad Ayquriavaur jou are sok , | ~nosnut stosoqord ay} Jo FtMOdoR oY} Ut Ajjwroadsa “(goe *d “g)
‘(9¢g “d “6 ‘Jo) toate we A[quqoad st siya qnq ‘eyvauTTap | UIqnof Fo FvYyZ Woy FVYMOMOS SIo]fIp orm “VIUTOUVT VA
‘3 Ur syonp Fo awd ouo Ajuo st asoy Fvyy SozeJs desing ¢ jo AWozVUR ayy FO JuNODDY S.La5ING PeMolfoF eavy J 1
| | | OINGB]
| pata Avy, OM, posahv-90iU J, d potadu[-90tq],| paradRy-OM JT, ¢ patadvp-OMy, -NOsMUt — sTOSOqoag
| quasq V¥ quasqy yuasatg | yuasqy quasqy » JUasatg ° ; SYNC |
quosalg quesqy (¢) quesqy quesq Vy qresqy| quasqy * " 0BS.10 apig
| BUNde| vunoep = BUNK] poojq
| wunory pooyq Aq = pooyq Aq paovaq pooyq &q paovaq Aq peovaqmia = BunoR] pooyq Aq
paoraqula PUB adABTT -Wa JOU PUR [[RUG aSaerq -we jou pur qpeurg Apard pay Teug peovrquia pue ASIVT* SULSLO [RAqatag
ared aud und aug d | Aur] Auryy ¢ AuvyL* sjyonp [erpiiyda yy
| poaqeponoea | |
qouyL “quas | IY]
quasqy quasqy -aad onssty anprmig quasatg quesq Vy, quasatg synod snourqulay
quarqy quasq ¥ A quasedg quasadg | quasa.tg spurs pro
pua THo]aI0ua
pus pay oy, ¢(¢) pue pury oy, é | pus pury oy, pua pury 03 Apawa ny pury oF Ajiveu gon -udya Jo yuoyxg
| atod stasoq
[eurUtteayqng [B.1yU9 A euinday,| peaqua A 1 [eaqUa {| yeutuntayqngy -o1d Fo uoryIsog
| |
| = ——— ——————- —— — — ——_——_ — — ———<$—_— — a FF — —
‘erpodnas hz, ‘erpoduing *sisdorpog “epodAxQ RIULOUI[R A erpodngy

TWO NEW BRITISH NEMERTEANS. 561

(5) The peculiar arrangement of the median dorsal nerve
noted above only occurs in Oxy polia.

Oxypolia also shows affinities with Parapolia, though
the absence of a head gland and the single excretory pore
readily distinguish this genus from it. With regard to the
relative positions of the six genera in the genealogical tree, it
seems probable that the six genera fall into at least two
groups.

On the one hand, the American genera Zygeupolia and
Parapolia stand together somewhat apart from the rest.
Three of the features which bring them together—i.e. the
absence of a specialised gelatinous connective layer in the
cutis, the absence of peculiar head glands, and the presence
of but a single excretory pore—seem at the same time to point
to their more primitive nature if we accept Biirger’s view
that the group is derived from a Carinella-like form. This
view is emphasised by the presence of a side organ in
Zygeupolia, an organ only found elsewhere in Carinella
and the primitive Lineid Micrella. Again, the three genera
Kupolia, Valencinia, and Oxypolia form a group cha.
racterised by the presence of peculiar head glands, the thick
gelatinous cutis layer (except in Valencinia), and an ex-
cretory system with many ducts. Of the position of Poli-
opsis it is impossible to speak while so many points in its
anatomy remain unknown.

Summary.

1. Description of two new genera of Heteronemerteans
belonging to the families Eupoliidee and Lineide, viz. Oxy-
polia and Micrella.

2. Micrella rufa is the most primitive member of the
Lineidz, showing great resemblance to the Carinellide, more
especially in its excretory system and in the presence of a side
organ.

3. The structure of the caudal appendage in Micrella
rufa differs from that of any other Heteronemertean yet

562 R. CO. PUNNET'S.

described, and throws doubt on the homology of that organ
throughout the group.

4, ‘he presence in a Heteronemertean of rhynchoccelomic

pouches comparable to those of Drepanophorus among the
Metanemerteans.

5. he arrangement of the median dorsal nerve in Oxy-

polia is peculiar, as it runs for some way outside the ex-
ternal longitudinal muscle layer.

6. A comparison of Oxypolia with the other genera of

the Hupoliidee.

BIBLIOGRAPHY.

McIntosu, W. C.—‘ British Annelida: the Nemerteans,’ 1873-74.

2. Oupemans, A. C.—‘* The Circulatory and Nephridial Apparatus of the

10.

Nemertea,” ‘Quart. Journ. Mier. Se.,’ vol. xxv, Suppl., 1885.

. Jounin, L.—‘“* Recherches sur les Turbellaires des Cotes de France,”

‘Archiv. de Zoologie,’ tom. vill, 1890.

. Coz, W.—* On the Anatomy of a Species of Nemertean, etc.,” ‘Trans.

Connecticut Acad.,’ vol. ix, 1895.

. BUrcER, O.— Die Nemertinen,” ‘ Naples Monograph,’ 1895.

. Monteomery, T. H.— On the Connective Tissues and Body-cavities of

the Nemerteans,”’ ‘ Zool. Jalirb.,’ Bd. x, 1897.

. T'nomson, CaroLine B.—“ Preliminary Description of Zygeupolia

littoralis, ete.,” § Zoologischer Anzeiger,’ Bd. xxii, 1900.

. Punnett, R. C.—“ On a Collection of Nemerteans from Singapore,”

‘Quart. Journ. Mier. Sce.,’ vol. 44, 1900.

. Punnett, R. C.—‘‘On some Nemerteans from Torres Straits,” ‘ Proc.
?

Zool. Soce.,’ 1901.

Punnett, R. C.—“ On some South Pacific Nemertines collected by Dr.
Willey,” ‘ Willey’s Zool. Results,’ Pl. v, 1900.

TWO NEW BRITISH NEMERTEANS. 563

EXPLANATION OF PLATES 39 and 40,

Illustrating Mr. Punnett’s paper on ‘ ‘l'wo New British
Nemerteans.”

ABBREVIATIONS IN PLATES.

a. Auus. 6.c. Bucealcommissure. 6.m. Basement membrane. c. Cutis.
c. c. Ciliated canal of cerebral organ. cer. Cerebrum. ¢. org. Cerebral
organ. c. org. gl. Glands of cerebral organ. ¢. ¢. Connective tissue. cu. gl.
Cutis glands. d. 4. . Dorsal blood-vessel. @. ce. Dorsal comniissure of brain.
d. g. Dorsal ganglion. e. p. Epithelium. ea. d. Excretory duct. ea. p.
Exeretory pore. ea. ¢. Exeretory tubules. g. Gonad. 4. s. Head slit.
i. d. Intestinal diverticulum. J/. 6.2. Lateral blood lacuna. J. 4. v. Lateral
blood-vessel. m.¢. Circular muscle layer of body-wall. m. c.¢. Circular
muscles of cutis. m.c. p. Circular muscle layer of proboscis. m. cr. Muscle
cross. m.d.v. Dorso-ventral muscles. m. 2.7. Inner longitudinal layer of
body-wall. m. J. i. p. Inner longitudinal muscles of proboscis. m. /.0. Outer
longitudinal muscles of body-wall. m. Z. 0. p. Outer longitudinal muscles of
proboscis. 2. c. Lateral side stem. x. d. Median dorsal nerve. 2. /.
Nervous layer. 2. p. Proboscis nerve. o. d. 7. Superior median dorsal
nerve (in Oxypolia). @s. Gsophagus. @s. e, Gsophageal epithelium.
es. 1. @sophageal lacune. p. ep. Proboscidial epithelium. 7. Rhyncho-
deum. r. ep. Rhynchocwlomic epithelium. rhe. p. Rhynchocelomic
pocket. s.0. Side organ. v.g. Ventral ganglion. v. vc. Ventral vascular
commissure.

Fic. 1—Micrella rufa. ‘Transverse section through the cesophageal
region just behind the termination of the excretory tubules. x 45.

Fie. 2.—M. rufa. ‘Transverse section through the level of the side organ.
x 45.

Vic. 3.—M. rufa. Transverse section through the level of the anus.
x 80.

Vic. 4.—M. rufa. ‘Transverse section through the proboscis about the
middle. x 160.

Fic. 5.—M. rufa. Side view of anterior end of the animal after preserva-
tion. X 2.

Fic. 6.—M. rufa. Section through the side organ. x 160.

Fie. 7.—M. rufa. Section through outer portion of ventral body-wall in
cesophageal region. xX 160.

564 Rk. C. PUNNETT.

Fic. 8.—M. rufa. Section through caudal appendage. x 80.

Fie. 9.—M. rufa. Corpuscles from rhynchocceelom pocket, two of which
are seen sideways. x 530.

Fig. 10, a—e.—M. rufa. Transverse sections through brain, cerebral
organ, and head slit. taken at intervals of 30 p. x 45.

Fig. 1].—M. rufa. Diagrammatic reconstruction of the anterior end of
the animal. Only a small portion of the proboscis sheath is shown, viz. the
portion which gives off the pockets. x 20.

Fic. 12.—Oxypolia beaumontiana. Sketch of live animal. Some-
what reduced.

Fic. 18.—O. beaumontiana. Sketches of anterior end made by Mr. W.
I. Beaumont from the live animal. Enlarged.

Vie. 14.—O. beaumontiana. Section througha portion of the proboscis.
xelo:

Fig, 15.—O. beaumontiana. Section through a portion of the skin in
the esophageal region. From a thionin preparation to show the cutis glands.
x 160.

Fic. 16.—O. beaumontiana. Section through anterior end near pro-
boscis pore. xX 45.

Fic. 17.—O. beaumontiana. Section through body-wall in esophageal
region. xX 80.

Fic. 18.—O. beaumontiana. Diagram showing arrangement of median
dorsal nerve.

Fie. 19, a—e.—O. beaumontiana. Section through brain and cerebral
organ, Distance between @ and 6 = 69 pw; between the succeeding sections
= 25 p. X 45.

Fic. 20.—O. beaumontiana. Two views of brain from model recon-
structed from sections. (a) Seen from the side. (4) Seen from behind.
x 1b.

Fig. 21.—O. beaumontiana. Longitudinal horizontal section through
intestinal region. xX 45.

Fic. 22.—O. beaumontiana. Diagrammatic reconstruction of anterior
end, showing the arrangement of the various systems. ‘The proboscis sheath
is omitted. x 10.

jor)
ce

THE C@LOMIC FLUID IN ACANTHODRILIDS. 5

The Cclomic Fluid in Acanthodrilids.
By

W. Blaxiand Benham, D.Sc., M.A., F.Z.S.,
Professor of Biology in the University of Otago, New Zealand.

With Plate 41.

Tue corpusculated fluid contained in the coelom of Kuro-
pean earthworms belonging to the family Lumbricide has
been the subject of a memoir by Dr. Rosa—from whose
careful studies we have learnt that, in certain species of the
genus Allolobophora, the corpuscles are of a varied
character—differing somewhat in different species; and,
moreover, that the commonly accepted account of the
formed elements of the fluid is not only very incomplete, but
more or less erroneous. ‘The usual description refers to this
fluid as “a colourless fluid with numerous amoeboid corpus-
cles.” ‘This is an imperfect truth, for, in addition to
“amoebocytes,” there are, in the commoner species of
Allolobophora, numerous “ Hleocyte”’ cells contaiming
refringent oily globules, which are, however, absent in
species of the genus Lumbricus, where, it seems, their
place is taken by ‘‘ vacuolated lymphocytes,” which are not
endowed with amoeboid movement.

Soon after my arrival in New Zealand, in 1898, I was
surprised to note, in the fluid of Octochetus multiporus
Beddard—which served my students as the type of an
earthworm—not only very abundant cells, recalling the
eleocytes of Rosa, but curious “ thread-containing cells,”
similar to those then recently described by Goodrich as a
constituent of the coelomic fluid of Huchytraus hortensis,

566 W. BLAXLAND BENHAM.

This observation led me to commence a detailed study of
the ccelomic fluid in various species of our New Zealand
earthworms ; the work has been interrupted from time to
time, owing to a variety of causes, but it seems at least worth
while to bring together such facts as I have gathered with
regard to this species, and to leave aside, for the present, the
less perfect observations on other species.

Octochetus multiporus has become familiar to zoolo-
gists from the various memoirs of Mr. Beddard, dealing with
the peculiar and exceptional situation of the gonads on the
posterior wall of their segments, the interesting condition of
the nephridia, and so forth.

The worm is one of our largest species and, like the three
other species! of the genus, is pale, indeed almost white,
owing to the absence of pigment in the body wall, which
allows the opaque white fluid contents of the ccelom to show
through. ‘The worm is curiously sluggish and inert; if one
be taken in the hand it makes no attempt to wriggle out of
it, but, by the contraction of the longitudinal muscles the
worm shortens itself, and at the same time the circular
muscles are contracted, so that it becomes quite tense and
firm to the touch. When placed upon a table it remains
quiescent, and is extremely dilatory in making an effort to
escape. In a pie-dish the worm seems to lack the strength
to raise its body up the sides—in the way so familiar to
students of most earthworms.

But when handled roughly « small amount of ccelomic
fluid issues from the dorsal pores. When the worm is placed
in weak alcohol or vapour of acetic acid for the purpose of
killing it the discharge from the dorsal pores is abundant,
and a similar fluid is copiously discharged from the mouth,
which, on examination, is found to be ceelomic fluid, with
all the usual cells. Whether this fluid enters the buccal

' The account of the fluid here given refers especially to O. multiporus,
but I have examined the fluid of two other species of Octochetus and find
a close similarity in the constituent cells, as well as in Acanthodrilus
annectens,

THE C@LOMIC FLULD IN ACANTHODRILIDS. 567

region of the gut by way of the peptonephridia—which are
known to be provided with funnels in the young (Beddard, 2),
though these structures have not been recognised in the adult
—remains to be discovered; the alternative 1s a rupture of
the buccal wall.

If the surface of the worm be touched with corrosive
sublimate, or if an incision be made in the tense wall, the
discharge of the white milky fluid from the dorsal pores at
once becomes active. In the latter case the discharge only
takes place from afew pores in the immediate neighbourhood
of the incision.

This fluid is opaque-white, resembling cream in appear-
ance; it has a varying consistency, but generally that of a
thick gum-mucilage or clotted cream. It does not “ flow ”
from an incision over the surface of the body, but spreads
slowly over it. So, too, when a drop is placed upon a glass

ing

slip, it seems to “set” at once, forming a mass of sufficient

consistency to support the cover-slip, for it does not—as the

fluid of Lumbricus does—flow over the slide to form an even

sheet. ‘his property necessarily makes the passage of
reagents somewhat difficult, but the use of normal salt
solution obviates this.

2‘ 60

The fluid discharged into a dish soon “ coagulates’
form a dirty white sticky and slimy sheet.

In the case of Acanthodrilus annectens (Beddard),
the slightest handling of the worm leads to a very copious
discharge of the coelomic fluid through the dorsal pores.
The fluid is cream coloured when fresh, and of the consist-
ency of.a thick gum solution ; it very soon becomes firm, and
in a few minutes hardens to form a pale yellow chalky mass.
It is so tenacious that it clogs scissors, sticks the fingers
together, forms a cake on the scalpel, and, in fact, is quite
unpleasant to deal with. Its great density renders the
examination of its micro-chemical reactions even more
difficult than in Octochetus, especially as normal salt
solution does not readily mix with it. ‘he plasma coagu-
lates almost at once, and forms a clot, through which fluids

568 W. BLAXLAND BENHAM.

do not pass. The cells die very much more readily than in
the case of Octochetus, the amcebocytes soon forming a
“plasmodium,” which Rosa has shown to be precedent to
death—one stage in degeneration.

The method of observation of the fluid naturally varied
according to the object to be attaimed. For the investiga-
tion of the form and structure of the hving elements a
hanging drop of the fluid was examined, the cover-slip being
supported on a ring or by some other means. In this way no
pressure on the cells occurred ; but the exposure to air, even
as Rosa has noted—certain

for a very brief moment, has
effects upon the cells, and I found it advisable to support one
side only of the cover, so that various depths of fluid could
be observed. The cells in the centre being protected from
air, and at the same time being subjected to a minimum
pressure—if any,—were probably in a normal condition.
This method, too, allowed the use of various reagents.

In fixing the cells I followed Rosa’s suggestion of killmg
them as they issue from the dorsal pore; by touching the
body with a drop of corrosive sublimate the fluid is dis-
charged and fixed.

The Cellular Elements of the Fluid.

The cells in the ccelomic fluid are extremely abundant, the
plasma being, relatively, as little as in the blood of a frog,
for example. ‘Thus the cells are crowded together, and seem
to exert a certain degree of mutual pressure upon one
another.

I recognise four distinct kinds of cells (PI. 41, fig. 1), viz. :

1. Amcebocytes.—Naked, more or less granular cells,
capable of thrusting out psendopodia, in fact typical “ leuco-
cyte-like ” elements.

2. Hleocytes.—Large rounded cells with a distinct
limiting pellicle, and incapable of forming pseudopodia.
‘The cytoplasm is filled with highly refringent oily globules.

THE C@LOMIC FLUID IN ACANTHODRILIDS. 569

3. Lamprocytes.'—Large rounded cells with a distinct
pedicle ; without pseudopodia. The cytoplasm is occupied
by numerous “ vacuoles,” each of which usually contains
a small, highly refringent “ granule.”

4. Linocytes.*—Smaller clear cells, containing one or
more thread-like products.

The relative proportions occupied by these four kinds of
cell varies ; but the “lamprocytes” are the most abundant,
and in a drop of fluid fill almost the entire field (fig. 1). The
“Jinocytes” are usually the fewest in number, but I have
evidence that they are more abundant in some worms than in
others ; but whether this is a seasonal or physiological or an
individual peculiarity I am at present unable to determine.

The amcebocytes are only moderately numerous; and they
also vary in numbers.

While in Octochztus these cells remain distinct and
separate, those of Ac. annectens adhere together to form
clumps, so that when a drop of the fluid is placed on the
slide one sees what at first appear to be enormous cells,
visible to the naked eye. But examined under the micro-
scope—and especially after the use of iodine—these large
cells are seen to be groups of cells adhering closely to one
another,

A variable number of cells occur in each group—usually
from six to twelve, but sometimes even more—and in each
group there is generally one linocyte, one or two eleocytes,
and the rest lamprocytes.

1. Amcebocytes.—These appear under tivo forms, depend-
ing on the character of the granules and of the pseudopodia.
The most usual form has a spherical, rather coarsely
granular body, with few, clear, filamentary

y pseudopods,
chiefly arising from one side (fig. 2). Whether or not this
form of pseudopods is the true and normal condition assumed

during the life of the cell in the body of the worm I am

' Naprpog = shining, in reference to the highly refringent granules,
> Nwwov = thread.

VoL. 44, pART 4,—NEW SERIES. 00

570 W. BLAXLAND BENHAM.

unable to state; certain it is that this is the form presented
by an amoebocyte examined under the ordinary conditions—
either fresh in its own fluid, or after the addition of salt
solution. But Rosa denies that such is the true form of the
corpuscle within the body—as Cattaneo and others have done
in the case of Arthropods and molluscs. According to him
this form, usually described and figured, is only assumed
under abnormal conditions.

He describes the true amcebocyte of the common earth-
worms (Lumbricus and Allolobophora species) as consist-
ing of a small central body, surrounded by more or less

> each of which consists of

numerous ‘ petaloid pseudopods,’
a firmer margin and an extremely transparent central portion,
so that the pseudopods look like a number of loops (see Rosa,
fig. 41).

I have certainly observed similar amcebocytes in Octo-
chetus (Pl. 41, fig. 3); nevertheless, I have—even after
taking the precautions suggested by Rosa, of fixing the cells
with osmic acid or corrosive sublimate as the fluid issues
from the pores—observed corpuscles with a form represented
in fig. 2.

Some of these amcebocytes contain yellow globules of
chlorogogen ; though it does not seem necessary to distin-
guish these as an independent kind of cell.

In Acanthodrilus annectens some of the amcebocytes
contained both these yellow globules or granules of chloro-
gogen, and in addition some clear refringent globules ; these
also have short pseudopods, and few of them.

In addition to these more or less spherical amcebocytes I
have noted in Octochetus—though as rarities in the dis-
charged fluid—some much elongated, spindle-shaped cells,
with clearer cytoplasm, containing finer granules; the pseudo-
pods are few, at the ends of the cell (fig. 4).

These long cells I found in considerable numbers by scrap-
ing gently the inner surface of the body-wall, along which
they appear to be creeping.. 'I'hese appear to correspond with
the “ spindle-shaped cell’? described by Ling Boom Keng in

THE C@LOMIC FLUID IN ACAN'THODRILIDS. 571

his account of the ‘Ccelomic Fluid of the Common English
Karthworm’ (pl. 4, fig. 15). It bears some resemblance to
the “mucocyte” of Allol. mucosa, figured by Rosa (fig. 23),
but that cell is much larger than the eleocyte ; further, his
account of the “mucocyte” shows it to be a very different
cell.

A good deal of work has in the last few years been done
in the staining reaction of the granules of amcebocytes in
various animals; amongst them, Ling Boom Keng has de-
scribed the various forms of amcebocyte in Lumbricus
terrestris. I have not made any observations on those of
Octochetus, and am, therefore, not ina position to confirm
any of his statements; but it is to be regretted that he did
not give a more detailed account of the living cell.

2. Kleocytes (Pl. 41, fig. 5).—These cells are moderately
abundant, of relatively large size, irregularly oval or circular
in outline, and measure, on the average, 40 m.

The cell is more or less spherical, but owing, perhaps, to
mutual pressure the cells assume irregular shapes. I have
not been able to detect any “diffluence” or automatic
change of outline, and it appears to me that cytoplasm so
loaded with endoplastic products could scarcely retain suffi-
cient energy to move so large a mass.

The cytoplasm is transparent (i. e. very finely granular) in
the living cell, and relatively small in amount; it is limited
externally by a very definite pellicle or cell-membrane, which
remains when, by the action of various reagents, the cyto-
plasm has been rendered absolutely transparent or has been
destroyed.

The characteristic feature of this cell is the presence of
numerous clear, colourless globules of oil, which crowd the
cytoplasm; and—unlike those recorded for Allolobophora—
are not limited to the periphery, but occur through the entire
depth of the cell. These globules are highly refringent, and
conceal the nucleus and cytoplasm in the living state.

The nucleus is excentric, circular in outline, and is em-

572 W. BLAXLAND BENHAM.

bedded in a central mass of cytoplasm, the rest of which is
reduced to delicate threads ramifying between the globules.

I found iodine a useful reagent whereby to stain the cyto-
plasm,—which takes on a sherry-brown tint—contrasting
therein with the cytoplasm of the other cells of the fluid; it
presents the appearance of groups of brown granules between
the globules, which are unaffected. This colour disappears
on warming, or rather becomes much lighter; but there is no
reappearance of the dark tint on cooling. I conclude, there-
fore, that these granules are not glycogen.

These eleocytes, however, contain in some cases highly
refringent granules in addition to the oily globules; only
rarely do the latter occur by themselves; but generally the
“oranules” are few in number; in a few cases, however,
they preponderate. I will return to them in describing the
‘“ oranule cells.”

It will be convenient now to describe the reactions of these
oily globules.

Firstly, with respect to stains :

Rosa finds that gentian violet colours the globules in the
eleocytes of Allolobophora blue, the nucleus being violet.
The globules in Octocheetus do not stain in a solution of
gentian violet in normal salty the nucleus, however, stains
violet, but much less readily than do the other cells of the fluid.

But if the cells be first killed in corrosive sublimate, and
the stain run in, the oily globules take on a blue tint.

Cyanin, too, is recommended for fat; and in the fresh
condition I find that the oil globules become coloured blue
with this reagent.

The Action of Acids, Alkalis, ete.—Nitric acid
(strong) causes the cells to swell, thus exhibiting very clearly
the pellicle, which is seen to be folded and creased on its
surface—evidence of a membrane of some toughness. ‘lhe
oil globules swell up, gradually losing their refringency as
they do so; but after the prolonged action of the strong acid,
and even after boiling the acid, they remain undissolved.

Hydrochloric acid gives a similar reaction.

HE CCLOMIC FLUID IN ACANTHODRILIDS. 7D

Sulphuric acid reacts at first lke the preceding; but the
globules are dissolved, leaving a coagulated network (? cyto-
plasmic) pervading the cell, whose pellicle, however, per-
sists.

Acetic acid (glacial) does not dissolve the globules, which
are equally insoluble in oxalic acid.

[The shde is heated over a bunsen till bubbles appear, and
the thin film of fluid boils more or less fiercely.].

When caustic potash (70 per cent.) reaches the cell, the
globules rapidly disappear one after the other; they are, in
fact, instantaneously dissolved.

Kther dissolves the globules.

Absolute alcohol, when poured suddenly and in consider-
able quantity on a cover-slip, leaves many of the globules
undissolved ; but when it is run in below the cover, I have
seen the globules disappear.

3. The Lamprocytes are the most abundant of all the
elements in the fluid (PI. 41, fig. 6). They resemble in size
and general outline the preceding eleocytes, but are, as a
matter of fact, rather flattened—as can be seen as they roll
over in a current of reagent; the oval or roundish
outline is more or less irregular, and I believe the cells are
capable of a certain degree of difluence. At any rate, they
are very readily capable of being compressed, and of again
resuming the normal form.

The cytoplasm—bounded by a definite pellicle—is clear
and transparent ; itis crowded with clear, colourless, circular,
vacuole-like structures, most of which contain a small but
very highly. refringent body. ‘This “granule” differs
chemically as well as physically from the ‘ vacuole,” and
each differs from the globule of the eleocyte.

I have used the expression ‘ vacuole-like,” for I feel doubt
as to whether we are here dealing with true vacuoles in the
cytoplasm ; when the cell is broken these structures are freed
and retain theirform. They seem to be of firmer consistency
than the cytoplasm, but have no definite ‘“‘ membrane ;” each

574 W. BLAXLAND BENHAM.

seems to be a droplet of some fluid, but not of an oily
character; there is no marked refringency, and reagents
point to different substance.

The “ granules,” which are somewhat greenish in colour,
are contained within the vacuoles, as can most certainly be
recognised in crushed cells (see fig. 6, a), and each granule
exhibits “ Brownian movement” therein, whether the vacuole
be still within its cell or isolated. As a rule each vacuole
contains a small granule, and never more than one, but
frequently the vacuole contains none. I have not seen any
granule independent of a vacuole; the two are genetically
related, but whether the granule is formed within the
vacuole, or the fluid of the vacuole arises as a result of solu-
tion of the granule I cannot determine.

These vacuoles are fairly constant in size, but the granules
vary within small limits and in different cells, while the
number in different cells is also subject to considerable
variations. The resemblance in size that a vacuole bears to
a globule naturally suggests some relation between the two,
but, as will be seen below, there is a chemical difference
between these things, though it seems probable that there is
a genetic bond connecting them in a series.

I have already mentioned that most of the eleocytes
contain, also, a few “ granules” in vacuoles ; sometimes the
characteristics of the two cells, which for convenience I refer
to by separate names, are united in a single cell (see fig. 7).
There is therefore little doubt as to the relation of one to the
other, and in some specimens of Octochetus the resem-
blance is still closer, in that some of the “ lamprocytes ”
contain “ vacuoles ” which are without granules,

We thus have, if we regard the eleocytes and lamprocytes
as derivatives one from the other, four conditions :

(a) Cells containing nothing but oily globules.

(b) Cells chiefly with globules, with few or many granules
in vacuoles.

(c) Cells with only one granule in each vacuole.

(d) Cells with vacuoles only.

THE CCLOMIO FLUID IN ACANTHODRILIDS. 575

I am, however, unable to say which of the two cells is
derived, or which is the earlier stage in the history.

In Ac. annectens these lamprocytes contain much
larger “granules,” about twice the size of those in Octo-
chetus; they have a much higher refringency, and are so
abundant that, when viewed by transmitted light, the whole
cell appears opaque and nearly black. The granule nearly
fills the vacuole.

Actions of Reagents.—Stains.—Gentian violet, in the
fresh, stains the nucleus deep violet: the cell membrane is
also stained, though less so than the nucleus. The cytoplasm
is scarcely tinted, while the vacuoles become a very pale
violet, so that in a glycerine mount they show up very dis-
tinctly.

Iodine.—The cytoplasm is stained only a very faint yellow
—quite different from the brown colour exhibited by the
eleocyte. The “vacuoles” also share in this yellow colora-
tion, as is best seen in those isolated and freed from the
cytoplasm, but the granules remain uncoloured.

Acids, Alkalies, etc.—When treated with nitric acid
the “ vacuoles” burst after swelling; the granules are thus
released, but soon dissolve, accompanied by a good deal of
turmoil in the cell. Though I could not detect any actual
bubbles, yet the cell-contents seemed to be ‘on the boil,”
as Goodrich has expressed it in regard to the action of
certain reagents on cells of Vermiculus.

Hydrochloric and sulphuric acids have the same effect.

Acetic acid (glacial) dissolves the granules after first
causing the “vacuoles” to swell and disappear. The
granules are insoluble in oxalic acid.

In potash the entire cell swells, and the contents disappear
instantaneously on the arrival of the reagent.

Neither absolute alcohol nor ether dissolve the granules.

4, Linocytes.—These thread-containing cells are, with-
out doubt, the most interesting and puzzling of the cell-
constituents of the fluid, and, though bearing some resem-

576 W. BLAXLAND BENHAM.

blance to the “thread-containing cells” of Vermiculus,
described by Goodrich, differ in a few details.

The relative number varies somewhat in different worms,
and the appearance of the cell and degree of development of
the thread within are subject to considerable variation ;
though whether this is due to individual or other causes I
am as yet uncertain.

When a drop of the ccelomic fluid is examined under a low
power there are seen, amongst the refringent cells just
described, a few clear, almost transparent, and somewhat
yellowish cells (fig. 1, d.), much larger, as a rule, than the
granular amoebocytes, and sometimes nearly as large as the
eleocytes. If this fluid has been mounted without any
precautions, but merely taken from the ccelom, this clear cell
will show, either immediately or after a short time, a circular
vacuole within—or sometimes more than one vacuole,—which
may be circular, or oval, or irregular. The margin of the
vacuoles is clear and slightly more refringent than the cyto-
plasm, and has the appearance of a ring, which becomes
more evident after the death of the cell. This “ring,” when
carefully examined, appears to be made up of a coiled fine
thread, which is faintly yellow, but prolonged study modified
this conception of a “coiled thread.” Before discussing the
interpretation to be put upon this cell-product, it will be
convenient to describe an average form and some less usual
types.

Usually, the limocyte (fig. 9) is spherical, of about half
the size of an eleocyte. The cytoplasm is finely but regu-
larly granulated, and forms a superficial envelope to the con-
tents, and it is bounded externally by a distinct envelope.
‘he nucleus 1s oval, and les in the peripheral cytoplasmic
coat, which is thicker in its neighbourhood than elsewhere.
The greater part of the cell is occupied by a slightly
refringent, clear, faintly yellowish inclusion—which for
convenience may be termed a ‘coiled thread,’ for even in
the most carefully mounted preparations the refringent
outline of the inclusion, be it oval or circular, soon shows

THE CQ@LOMIC FLUID IN ACANTHODRILIDS. 577

minute concentric fibrils ; and moreover, the refringency is
not limited to the outline of the inclusion, but crosses the
‘“ vacuole” in curved lines, each of which presents the same
appearance of fibrillation, and, as one focusses this strange
inclusion, the whole resembles a coil or tangle of cotton.

The use of certain reagents renders the thread more dis-
tinct, and separates the fibrils from one another, so that the
tangle appears to be unravelled before one’s eyes (see fig. 25) ;
but I have failed, after long search, to detect a free end ;
nor is there any regular spiral arrangement such as both
Goodrich and Eisen have indicated.

Whereas the majority of the cells have but one such “ coil
of thread,” a few cells (or in one specimen, at least, the
majority of the cells) contain several threads, which are of
different sizes and degrees of development. Thus, on July
11th I noted a great variety in the form of the coil, as these
figures well illustrate (see figs. 18, 19, 20). One particularly
curious cell is figured (fig. 21); 1t shows two coils, a circular
and an hour-glass shaped one, which is further represented
enlarged at the sides. But generally, in the case of the
several small ‘ coils,’ each lies in one plane, is simple, and
more or less circular ; whereas, in the case of what I think
may be regarded as normal linocytes, the “coil” is sphe-
roidal, and complicated by crossings and “intertwining ” as
it were.

From a series of observations, made at two different
seasons of the year, in worms of different degrees of sexual
maturity I have been able to trace the development of this
curious cell-product.

The lnocyte is at first a spherical, colourless, non-amce-
boid cell, filled with cytoplasm only, and bounded by a
delicate but distinct membrane (fig. 10).

The cytoplasm is very finely granular, and is occupied by
numerous very small, circular vacuoles, so regularly arranged
as to deserve the descriptive term “honeycomb.” The
nucleus, even in the younger cell noted by me,! is peripherally

1 It is probable that a still earlier phase of this linocyte bears some rela-

578 W. BLAXLAND BENHAM.

situated just below the cell membrane ; moreover, even in
the earlier phase observed, there is a larger more centrally
placed vacuole, the outline of which is at first not specially
distinct. This vacuole gradually increases in size, but I
could not determine whether this results from the union of
the smaller vacuoles with one another, which appears pro-
bable. With this increase of the vacuole the cytoplasmic
envelope becomes thinner and loses its ‘honeycomb ” ap-
pearance, etc. (fig. 11). At first, and for some little time,
the outline of the vacuole, though distinct, exhibits no pecu-
liarity ; but after a time it becomes more definite, and
appears as a gradually thickening wall, which then becomes
refringent. This refringency (fig. 12) commences to be evi-
dent at one side, sometimes on the side next the nucleus,
but as often at any other point. ‘This refringent are gradu-
ally extends so as to become crescentic (fig. 13), the ends
always thinner than the central region; and by a continua-
tion of this procedure the central vacuole becomes completely
surrounded by a circular refringent ring (fig. 14). These
stages were particularly well exhibited in a fully mature
individual examined in November.

This ring now continues to thicken, so that the vacuole
becomes constantly reduced; and as it does so, the ring
seems to become differentiated into irregularly concentric
layers, alternately more and less refringent; in this way the
fibrils of the thread are established (fig. 15). But mean-
while the contents of the vacuole have also become con-
centrated and refringent along transverse lines, giving rise
to curved loops passing from one side of the ring to the
other ; and in this way the “coil” represented in fig. 9 1s
brought about.

In this history there are many points of resemblance to
the development of a nematocyst within a cnidoblast ; and at
an early stage the likeness of the young linocyte to a fat cell
is very evident. The fibrillation of the “ ring ”’ is not evident

tions to an amceboid cell; or at least to some indifferent cell with a centra
nucleus.

THE CGLOMIC FLUID IN ACANTHODRILIDS. 579

in absolutely fresh cells, but when the cell is examined in
salt solutions or even after exposure to air, soon becomes
marked, and is still more intensified by the use of reagents.
The inclusion seems, then, to be a semi-solid spheroidal
structure, which is easily disintegrated into a “ thread.”

Turning now to the less frequent condition where a multi-
plicity of “rings” exists (as observed in certain immature
specimens examined in July). The earlier phases are iden-
tical with the one first described, but in place of a single,
large, central vacuole two, or more, smaller and irregularly
distributed vacuoles make their appearance (fig. 16) ; each of
which, then, becomes surrounded by a refringent ring, which
usually remains simple (fig. 17, ete.).

Action of Reagents.—Stains.—Gentian violet stains the
thread and the fluid in the vacuole very rapidly a bright
blue ; it is the first of the cells of the fluid to take the stain ;
the nucleus is stained violet.

Todine stains the thread a bright yellow, much more deeply
than the cytoplasm.

Acids, Alkalies, etc.—Nitric acid.—At the first contact
the cell membrane shrinks somewhat, and the thread becomes
more evident. But soon the cytoplasmic envelope becomes
coagulated, the transparency being replaced by opaque granu-
lations that conceal the thread within. The cell swells, and
these granules then give place to a series of radial lines (fig.
23) ; while the thread is concealed, and in the first experiments
I believed that it had been reduced to granules which are
more refringent than those of the cytoplasm; finally, the
cell membrane becomes ruptured, and the thread issues in
loops from the margin (fig. 24), sometimes at one point only,
though usually at several points round the circumference.
The radial lines observed in the cytoplasmic envelopes are
probably caused by the “unravelment” of the thread, and
represent the limbs of the loops that ultimately burst out of
the cell.

Boiling nitric acid reduces the whole thread to granules,
which are not dissolved by further action.

580 W. BLAXLAND BENHAM.

The action of hydrochloric and of sulphuric acid is
similar,

In acetic acid the thread cell becomes transparent; the
thread gradually swells and loses its distinctness till it
finally disappears; while oxalic acid reduces the thread to
granules, but does not dissolve it.

In caustic potash (30 per cent.), just at first the thread
becomes more evident; the cell swells and bursts, leaving
the thread behind. This now in its turn swells up, and the
fibrillation becomes more and more definite, the thread, indeed,
has the appearance of becoming gradually unravelled, so
that the ring is replaced by a coil, which becomes looser as
the action of the reagent is prolonged (see fig. 25, a, b, ¢).
By this method it is possible to ascertain the existence or
non-existence of a “free end.’ I have been quite unable
to see anything of the kind; nor is the thread, as represented
by its fibrils, coiled in a spiral.

The process of unravelling continues, but the potash has
no further action upon it. Hven after boiling the prepara-
tions the thread remains undissolved, and appears as a
continuous thread in the form of a chain or a wreath of
loops of varied shapes; but even now retains it refringency
and definiteness (fig. 26).

In earlier experiments, where the potash was weaker, the
cell after swelling simply burst, and the thread issued in
loops, much as in the reaction with mineral acids.

Absolute alcohol does not dissolve the thread, though it
shrinks a good deal, forming a refringent, irregular mass in
the cell.

Kther converts the thread into granules,

Osmic acid has an action similar to absolute alcohol, and
does not sensibly brown the thread.

Remarks on the Celomic Fluid in general.

The fluid, as already stated, is milky to creamy in colour,
owing to the great abundance of the eleocytes and lampro-

THE CQLOMIC FLUID IN ACANTHODRILIDS. 581

cytes. The great consistency, and the absence of ready flow,
which is so noticeable a feature, as compared with the fluid of
Lumbricids, is related no doubt to the abundance of cell
elements and small proportion of liquid plasma. But this
plasma itself must be much less fluid than blood-plasma, for,
as above mentioned, a drop of the fluid is sufficiently firm to
support a cover-slip, and yet the cells are protected from
bursting or injury; indeed, they are but slightly compressed
except at the margins of the drop.
The fluid appears to ‘ coagulate ’
unable to detect any fibrin-like threads except after exposure

> rapidly, but I have been

to air for some time.

The stuff is sticky; adheres to the glass, or dish, or fingers.

The dish in which the worm was kept during examination
of the fluid soon became coated at the bottom with a
tenacious slime, becoming slightly buff coloured after a time.
This slime when lifted seems stringy, and fine threads, fibrin-
like, hold the corpuscles together. But these have no rela-
tion, so faras I could discover, to the threads in the linocytes.
I imagine they are chemically produced in the plasma.

Octochetus multiporus and O. antarcticus are
highly photogenic or phosphorescent, and when handled in
the dark it is at once seen that this light has its seat in the
ecelomic fluid as it issues from the dorsal pores and slowly
spreads over the surface of the worm. The effect is much
more brilliant if the worm be stimulated by a little vapour of
acetic acid; then the abundantly discharged fluid gleams
with considerable brilliance.

It has been suggested (Beddard, ‘ Nature,’ vol. lx, p. 52)
that the photogeny of Microscolex modestus and of
Allolobophora fcetida is due to photogenic bacteria ; but
as I have indicated (loc. cit., p. 591) there is reason to be-
lieve that the phenomenon is connected with the eleocytes
of the fluid. It is well known that in a number of animals
photogeny occurs in direct association with cells containing
fatty matters; and that it is by no means always or neces-
sarily associated with the presence of bacteria (for example,

582 W. BLAXLAND BENHAM.

the glowworm, and firefly, and others). It has long been
regarded as connected with metabolism and rapid oxidation
of fat. Radziszewski has carried out a series of experiments
with various organic chemical substances, such as fats,
ethereal oils, hydrocarbons, and alcohols; and (I quote from
Max Verworn’s ‘General Physiology, p. 256) he “found
that a whole series of organic bodies emit light when they
are slowly combined with oxygen in an alkaline solution.”
Further, Verworn states, “It is in the highest degree prob-
able that the luminosity of living substances depends upon
analogous processes;”? and he comes to the conclusion, as
others have before him, that the photogenic substance is
produced in cell-metabolism.

Now in the eleocytes of the coelomic fluid, it seems to me,
we have just the very conditions for the emission of light,
and we need not summon bacteria to their aid. As a matter
of fact, I have seen no bacteria in this photogenic fluid.

The cells possess considerable vitality and power of resist-
ance to pressure and to drying, for after several hours’ ex-
posure such a slime shows abundant cells of all kinds. In
one instance a living worm was first operated on at about
10.30 in the morning by a small incision; it remained alive
in the dish, being repeatedly incised for fluid, all day, and
the slime, when examined at 4.30 the same afternoon, ex-
hibited all the usual cells apparently alive; at any rate the
amoebocytes were still moving, thrusting out pseudopods as
actively as when freshly extruded from the worm; the cells
presented their normal appearance, and there was no evi-
dence that the linocytes discharge the thread, for the cells
were present in the usual proportion, and nothing resembling
empty cells existed.

Again, the bottom of the tin in which several worms had
been kept for a week or more in grass, was covered with
slime, mixed with earth that had been discharged through
the anus. his dirty slime, when examined, also contained
uninjured cells; but I was unable to detect many of the lino-
cytes. Some I saw, but apparently they were not as abundant

THE CGLOMIC FLUID IN ACANTHODRILIDS. 583

as in the fluid. However, it is not an easy matter to explore
a fluid thick with finely comminuted dirt, and I do not think
any conclusion can be drawn from the apparent fewness of
the linocytes here.

Comparison of the Cells with those of other
Oligochetes.

In his monograph Beddard gives but little information
as to the formed constituents of the coelomic fluid. On p. 26
he states that ‘in the higher Oligocheta” the corpuscles are
“apparently of two kinds: ” viz. amceboid cells, and large
spherical cells loaded with granules; these “are probably
merely stages in growth.”

On p. 27 he says, “among earthworms there is generally
not such a great abundance of corpuscles” (as in the Naiads
and HEnchytreeids), but he mentions the milky-white appear-
ance of certain Kudrilids as being due to the great abundance
of cells.

Beddard himself, in studying the development of Octo-
chetus multiporus, observed a large quantity of cor-
puscles ; for in the later stages the ccelom “ was almost com-
pletely filled with granular corpuscles (i.e. lamprocytes), which
represent a further development of the small non-granular
cells” (i.e. amoebocytes) (2, p. 509). In his monograph he
considers this fact “ related to rapid growth and excretion ;”
but it is curious that in his various studies on the anatomy of
the adult, his attention had not been attracted to the flaky,
white substance that in preserved specimens fills the ccelom ;
but even if it had been examined, it is improbable that much
additional information could have been derived from it.

Since Kukenthal’s account of the “lymphoid cells” of
Annelids, the most detailed description of the ccelomic
corpuscles is contained in Rosa’s memoir, to which reference
has already been made. He finds in most of the species of
Lumbricids three kinds of cells—amcebocytes, eleocytes,
and “ vacuolated lymphocytes.”

584. W. BLAXLAND BENHAM.

I have referred to the amcebocytes above. ‘The eleocytes
of Octocheetus differ but slightly from those described by
Rosa ; and, indeed, different species of Allolobophora
contain eleocytes that differ slightly amongst themselves.
‘hus in some cases (e.g. A. foetida) the oily globules fuse
with one another when the cell is exposed to the air, giving
rise to a peripheral halo of oil round a central nucleus ;
whereas in A. putris this fusion does not occur. I have not
been able to detect the ‘ centrospheres” which, though
absent in some species, seems to be a very conspicuous fea-
ture in other eleocytes.

The ‘‘ vacuolated lymphocytes,” which exist in those species
in which eleocytes are less abundant—as in A. caliginera,
Lumbricus and sp.—differ chiefly from the “lamprocytes”
in the absence of the refringent granules. Rosa notes also
their slow coloration with gentian violet.

It may be mentioned here that Rosa has shown that these
refringent globules in eleocytes—which are yellow in some
species—are easily distinguished from chloragogen globules
by various reactions ; and he shows the error of the idea—due
originally, | believe, to Prof. Ray Lankester,! and later on to
Kiikenthal’s work—that these spherical cells are simply
amcebocytes gorged with chloragogen granules ; a view that
has crept into a number of text-books, from its plausibility
and from the ready way in which the function of the cells
was thereby explained.

Cuénot, too, confused the eleocytes with chloragogen
cells, but explained their history rather differently.

There seems, according to Rosa, to be no doubt as to the
distinction between the oily globules and the chloragogen.
But in the “granules” of the lamprocytes occurring in
Octochetus we have quite a different substance, resembling
some form of chloragogen in its insolubility in absolute

1 The view that the chloragogen cells are metamorphosed into free cells of
the coelomic fluid is due to d’Udekem, who in his monograph on ‘ Tubifex
rivulorum ” (‘Mem. Couron. Acad. Belg.’) develops this view and gives
figures in support of it.—E, R, L,

THE C@LOMIC FLUID IN ACANTHODRILIDS. 585

alcohol, in ether, and in potash; and this opens up another
question—the relation of the two kinds of cells.

To this I am not prepared with any suggestion, but it is
fairly evident that one is derived from the other.

Before leaving these eleocytes, reference may be made to
Mr. Picton’s observations on the corpuscles of the ccelomic
fluid of Amphitrite, of which he figures (figs. 50, 52)
examples, which appear to agree closely with those of Oligo-
cheta. Further, he states that in some of the eleocytes in
which little fat has accumulated there are “ strongly defined
granules of yellow pigment” (p. 290), which are represented
in his fig. 51, where each appears as a highly refringent
granule in the centre of a ‘‘ vacuole,” similar to the condition
noted in the “lamprocytes’”’ of Octochetus; in fact, the
figure of the whole corpuscle, with its vacuolated structure,
might stand for such a cell.

In this paper the author gives an account of the series of
chemical tests applied by him to the cell-contents of heart-
body and other cells, in a series of Polycheta.

With regard to the linocy te or thread-containing cells :—
these seem to have been observed in ccelomic fluid for the
first time by Goodrich—in Enchytrewus hortensis. His
account of the reactions of the thread agrees very closely with
those enumerated above ; but he states, with some apparent
doubt, that the thread is dissolved in boiling potash. ‘he
form of the cell and of the thread itself is different, however.
The “ thread-containing cell” of Enchytreus agrees with
the “granular cells” in containing highly refringent
globules, and he states (p. 58) “that the thread itself
appears to be formed at the expense of the granules,”
though he allows that the appearance upon which he relies
may be deceptive.

In Octochextus it does not appear that there is any

relation between either the “ granules ”

or the globules on
the one hand and the “ thread’”’ on the other; we have seen
that the characters of the cells are entirely different, and
while the ‘eleocyte”’ and the “lamprocyte’”’? may be, and

VoL. 44, PART 4,—NEW SERIES, PP

586 W. BLAXLAND BENHAM.

probably are, related to one another, I am of opinion that
the “linocyte ” is quite an independent cell.

Still more recently ‘‘ thread-containing cells” have been
noted by Hisen in certain Oligocheta. He uses the term
“nematocytes” in describing them—a term which I think is
not altogether suitable, in view of the familiar ‘ nematocyst ”
of the Cnidaria. His account of these ‘ thread-containing
cells” is based upon sections, and this has, I believe, led him
to certain erroneous conclusions as to the form and character
of the cell and thread.

In Ocnerodrilus panamaensis he finds such cells, and
he states (p. 131) that ‘‘the whole cytoplasm is filiform, and
takes the shape of a single, continuous, narrow strand,
wound regularly, like a coil of rope.” . . . ‘The beginning
and end of the thread could be clearly seen” (p. 182).
“There is no cytoplasm visible either inside or outside the
cytoplasmic thread, which does not fill the centre of the cell,
there being always a space there.” I think, from his de-
scription and figures (pl. xi, fig. 116), that the discrepancy
between this account and my own may be explained by the
fact that he examined no fresh material.

But I cannot accept his suggestion that “the rope may
serve to catch bacteria, sperm fragments, or other foreign
substance in the lymph.”’ Had he examined the fluid fresh I
believe that he would not have hazarded the suggestion. At
the same time I am not in a position to form any idea as to
the functions of the thread.

In this paper Hisen also gives some account of other kinds
of ccelomic corpuscles occurring in various earthworms, but,
since this account is founded on observations of preserved
material only, I have no occasion to refer to it more
specially.

With regard to the chemical nature of these cell-products,
I regret that I have not been able to form any definite con-
clusion, owing to my ignorance of micro-chemistry and the
lack of available literature dealing with the subject.

As will be seen in the above notes, I have used the same

587

THE CQ@LOMIC FLUID IN ACANTHODRILIDS.

‘aorearvdoad arg Ul pojou adoM—Spot JOYS AYIT S1dyqO
‘pedeys-puomeip omos—svysdao jpems jo Aqquenb vw plov orsoe YALA guewyvoag 104} W—'9ION

* * x | x *% eqfoouly |
— — — — — + — — — ut peed],
a) f001d mv]
—_ ae a + | om an ae — — Ui = SayNUBA
| |
| (Ayer) a3A40090]9 |
a te = sat S| ies =e ar ste prey Ae Se) LCT AL)
= | = | : = == = : : |
| ‘PPV IW mrepratee 9 |
ae poy poy | ‘pry poy % OF [oyooly
1ayu : Loy : gees UII TY :
| FEM oumydyng mene OLIGIN IYRXO 14009 svqog aynposqy |
‘poaTossipun ynq ‘sopnueso 07 peonpa.l i e S y USIS OT,
‘OUI[IOG UO UAAV “paATOSsIp Jou . e % — USIS OL],

‘paajossip st gonpoad ays qVyy SeqVolpul + USS OTF,

“SLOOGOUd-TTE/) AHL dO SNOLLOVAY TVYOINGH()-OdDT IV

588 W. BLAXLAND BENHAM.

series of reagents as Goodrich employed, under the advice of
Professor Gotch ; and I will tabulate the results obtained by
me for ready comparison with Goodrich’s results.

The observations of Schaeppi, of Picton, and of several
others, on the intra-cellular products of Polycheta, point to
the great variety of these cell-contents. But, in order to
obtain some further light upon this difficult problem, it seems
to me desirable that the whole question should be dealt
with by a competent chemist. It is, at present, somewhat
difficult to correlate the various results, or to form a definite
opinion as to the function performed by these cells—whether
excretory or otherwise,—or the stages by which the varied
chemical substances are built up. ‘lhe whole subject is one
of great interest and importance, and I can only give my
results for what they are worth.

A comparison of this table with that given by Goodrich
shows that these “granules” agree with the “ white gran-
ules” in the lymphocytes of Hnchytreus hortensis,
which are neither mucin nor chitin. They do not appear to

agree with “ Chloragogen ”
tion.

a word of very wide applica-

DUNEDIN ;
November 380th, 1900.

LITERATURE.

=

. Bepparp, F. E.—‘ Monograph of the Oligoch:eta,’ Oxford, 1895.

2. Benparp, I. E.—‘‘ Development of Ac. multiporus,” ‘ Quart. Journ.
Micros. Sci.,’ vol. xxxili, p. 495.

3. Carranro, —.— Sulla morfologia delle cellule amceboidi dei Molluschi
e Artropodi,” ‘ Boli. Sci.,’ Pavia, vol. xi, 1889.

4. Cutnor, L.—“Etude sur le sang et les glandes lymphatiques,” ‘Arch. de
Zool. Expér.,’ 2nd series, vol. ix, 1891.

5. Eisen, G.—“ Researches on American Oligocheta, ete.,” ‘Proce. Cali-

fornian Acad, Sci.,’ 3rd series, vol. ii, 1900, p. 85,

THE CQLOMIG FLUID IN ACANTHODRILIDS. 589

6. Goopricu, E. S.—<*‘ Notes on Oligochetes, ete.,” ‘Quart. Journ. Micros.
Sci.,’ vol. xxxix, 1896, p. 51.
7. KiixentHaLt, W.— Die Lymphoidzellen der Anneliden,” ‘Gen. Zeit.,’
1885.
8. Linc Boom Krene.—“On the Ceelomic Fluid of Lumbricus ter-
restris,” ‘ Phil. Trans.,’ 186, A. 1895, p. 383.
9. Picton, L. J—“On the Heart, Body, and Celomic Fluid of certain
Polycheeta,” ‘Quart. Journ. Micros. Sci.,’ vol. xli, 1898, p. 263.
10. Rosa, D.—‘I lingociti degli Oligocheti,’ ‘Mem. Accad. R. d. Sei.
Torino,’ 2nd series, vol. xlvi, 1896, p. 149.
See also Lankester, BE. Ray, “On some Migrations of Cells,” ‘ Quart.
Journ. Micr. Sci.,’ vol. x, 1870, p. 265.

EXPLANATION OF PLATE 41,

Illustrating Professor W. Blaxland Benham’s paper on “‘ The
Ceelomic Fluid in Acanthodrilids.”

Fie. 1.—A group of coelomic corpuscles of Octochetus multiporus,
fresh, without reagent. (Camera. x 375.) a. Ameebocyte. 4. Eleocyte.
e. Lamprocyte. d. Linocyte.

Fie. 2.—A granular ameebocyte (a of Fig. 1), the appearance described in
ordinary preparations, with short, clear, digitiform pseudopods,

Fie. 3.—An ameebocyte with petaloid processes; seen when the greatest
care is taken to obtain a preparation in a natural condition.

Fie, 4.—Spindle-shaped, faintly granular ameebocyte; found creeping along
the inner surface of the body-wall.

Fic. 5.—An eleocyte, slightly compressed, but without having been acted
upon by any reagent. In addition to the characteristic oily globules, it con-
tains a few refringent granules ; the vacuole in which each lies is not seen.

Fic. 6.—A typical lamprocyte, slightly compressed, fresh; the vacuoles,
with their highly refringent granules, are also shown at 6, a, as seen when a
cell is broken up.

Fie. 7.—A cell intermediate in character, combining the features of both
eleocyte and lamprocyte, slightly compressed, fresh.

590 W. BLAXLAND BENHAM.

Fic. 8.—A lamprocyte stained with gentian violet. (Camera. x 700.)
The vacuoles and granules are represented of their real size. The nucleus is
shown and the creasing of the cell-membrane.

Fic. 9.—A typical linocyte, from discharged fluid, slightly compressed and
dead, showing nucleus and complex “ thread-coil.”

Fie. 10.—A young linoeyte, showing the honeycomb appearance of the
cytoplasm, and the great central vacuole and peripheral nucleus.

Fre. 11.—A slightly older linocyte, with enlarged vacuole ; the houeycomb
appearance of the cytoplasm no longer exists.

Fies. 12—15.—Sketches of four linocytes in different stages of develop-
ment, showing the method of origin of the “thread”? within the vacuole.
These sketches are slightly diagrammatic, but represent faithfully the ap-
pearance of the thread.

Fie. 16.—A young linocyte in which several large vacuoles exist, in each
of which a thread-coil will arise (fresh).

Fic. 17.—A linocyte with several independent rings or thread-coils.

Kies 18—20.—Living linocytes with more than one thread-coil, to illus-
trate the variety of form assumed by this cell product. ‘Ihe details of the
cytoplasm are not fully represented.

Fic. 21.—A livocyte with a single thread-coil and three large vacuoles, in
which probably threads will appear later on.

Fic. 22.—A linocyte (dead) containing two thread-coils, one of which has
an exceptional form, which is further indicated at a; z is the nucleus.

Fie. 23.—A linocyte after the addition of nitric acid to the preparation.
The thread is indistinct, but a series of radiating lines occupy the peripheral
region of the cell, which are possibly parts of the disentangled thread.

Fic. 24.—A linocyte after further action of nitric acid; the cell membrane
has become ruptured, and the unravelling thread is issuing in the form of
loops.

Fie, 25.—The thread-coil after potash has dissolved the cytoplasm. As the
reagent acts the thread undergoes the changes represented at a, 4, ¢, the
thread-coil or ring swells up, and the fibrils separate from one another.

Fie. 26.—Three sketches of threads left after the action of boiling potash,
to illustrate the continuity of the thread.

THE CRYSTALLINE STYLE OF LAMELLIBRANCHIA. 591

The Crystalline Style of Lamellibranchia.
By

s. B. Mitra,
Of Calcutta, late of University College, London.

With Plate 42.

Four hypotheses have been framed with regard to what
the crystalline style of the Lamellibranchia is and does: (1)
Gegenbaur thought it was a secretion from the enteric
epithelium (‘Elements of Comparative Anatomy,’ English
translation, p. 359). (2) Balfour suggested that it was to be
considered as a rudiment of the radular sac of the Glossophora
(see Professor Lankester’s article on ‘‘ Mollusca,” in the
‘Encyclopedia Britannica,’ ninth edition, p. 685). (8)
Claus regarded it as an excretion of the enteric epithelium.
(4) Sedgwick thinks it is to be regarded as a “reserve of
nutriment”’ (‘Student’s ‘l'ext-book of Zoology,’ vol. i, p. 339).
We may say at the outset that of all the hypotheses
Gegenbaur’s was the nearest to the truth, for it is in reality
a secretion, a digestive ferment whose function it is to
digest starch, i.e. to convert starch into a reducible sugar.

The crystalline style cannot be regarded as a rudimentary
structure representing the radular sac of the Glossophora,
for the following reasons :!

(1) Its comparative size is not like that of a rudimentary
structure. It is fully three fourths as long as is the animal

1 A structure apparently identical in nature with the crystalline style of
Lamellibranchs co-exists in some Gastropoda with the radula. If the

identity of the Gastropod’s aud Lamellibranch’s crystalline styles be admitted
there can be no question of relationship to the radula—E. R. L.

592 S. B. MITRA.

itself (Anodon). A structure of such a comparatively great
length cannot be regarded as a vestigial structure unless
there are cogent reasons for taking it in that light; but there
are none.

(2) The cecum (a diverticulum of the alimentary canal
starting from the pyloric end of the stomach), in which in
some Lamellibranchs the style is lodged, and which the
hypothesis under consideration would take to represent the
sac proper of the radula of a glossophorous Mollusc, is lined
with epithelial cells bearmg active cila which are much
longer and better developed than the cilia of the epithelial
cells that line the alimentary canal itself. Here we might
say parenthetically that the alimentary canal of the Lamelli-
branchia is lined throughout with ciliated epithelium. Now,
as cilia always perform an important function in the economy
of the organism that possesses them, we should, if the caecum
were really a vestigial structure, expect to see either that it
was not lined with ciliated epithelium at all, or that if it
was lined with ciliated epithelium at all, the cilia were shorter
and less well developed than the cilia of the cells that line
the alimentary canal itself. The fact that the epithelial
cells lining the ceecum bear such highly active and longer
cilia, suggests forcibly the idea that the caecum performs
some important function, and so cannot be regarded as a
vestigial structure. In fact, there are reasons (as we shall
see later on) for regarding it as a higher stage in the evolu-
tion of the receptacle for the lodgment of the crystalline
style, the more primitive stage being still found in some
species of the Lamellibranchia.

(3) The ceecum in which the crystalline style is lodged in
some Lamellibranchs starts from the pyloric end of the
stomach, whereas the radular sac is formed by a diverticulum
of the wall of the cesophagus. If the former represented the
latter, we should expect to see the caecum start from some
part of the esophagus. We admit that, taken alone, this
reason is not very forcible, as there may be considerable
change in the position of the same fundamental structure in

THE CRYSTALLINE STYLE OF LAMELLIBRANCHIA. 598

the different species of the same group of animals; but
taken in conjunction with the other reasons, it must be
allowed to possess some value.

(4) We invariably found in literally hundreds and hun-
dreds of the fresh-water mussels and other Lamellibranchs
that a considerable quantity of food material surrounds and
is embedded in that end of the crystalline style that projects
into the stomach (see figs. 2, 4, 6). This is very suggestive
to say the least.

(5) The fact that the crystalline style is periodically
renewed (Claus, Sedgwick) does not fit in well with the idea
that 1t is a rudimentary structure. Why should a vestigial
structure be formed and vanish again and again, say once a
day (about that is the frequency of renewal that I found to
obtain amongst the mussels which I kept in an aquarium) ?
It rather suggests the idea that the style 1s somehow con-
nected with some important function that is performed by
some organ of the animal. And we have found that that is
the case. It is connected with the digestive function.
Whenever digestion is going on actively in the animal, as
evidenced by the presence of undigested and half digested
food material in the stomach and first portions of the intes-
tine, and of excrementitious matter in the last portion, one is
sure to find the style. When that function is for any reason
in abeyance, one fails to find it. ‘This is shown by the
following observations. I used to get my supplies of mussels
in batches of about 200 each. After they had been taken
out of a pond they were kept and brought to me ina very
small quantity of water in a pail-like vessel. About ten
hours elapsed from the time they had been taken out of a
pond and the time they reached me. All this time their
position in the vessel was anything but natural. hey lay in
the vessel like a mass of pears in a basket,—some horizon-
tally, some vertically, others obliquely, every one being in
contact with the surrounding ones. Moreover, during the
transit from the country to the town they had to undergo a
considerable amount of jostling and rubbing against one

594. S. B. MITRA,

another. All these circumstances were unfavourable for the
proper functioning of their digestive organs. Now, on their
arrival at my place, I used to open some fifty at a time, and
found the crystalline style in none. I could also at the same
time see that their digestive function was naturally enough
in abeyance. ‘lhe remaining mussels I would then put into
a fresh-water aquarium, where there was plenty of space,
good water, and food material. After two or three hours I
used to open some fifty more of the same batch, taken fresh
from the aquarium. ‘Then I invariably found the crystalline
style in one and all of these fifty mussels, and at the same
time noticed that the digestive function in them had just
begun again. An interesting experiment proving the same
thing—that is, the existence of a connection between digestion
and the crystalline style—got performed over and over again,
one might almost say of itself, in one of my aquaria. ‘This
aquarium had a leak, and the water in it would gradually
drain away in the daytime. During the night there would
not be left sufficient water to enable the mussels to set up
the well-known in-going and out-going currents and so carry
on respiration and digestion actively. On the next morning
at about 7 o’clock there would be left no water at all in the
aquarium. At 8 a.m. every day for two months I examined
some twenty mussels, and found the crystalline style in none,
and the digestive function in abeyance in them. At about 9
a.m. the tap was turned on every day, and the aquarium
filled with water. ‘'T'wo or three hours afterwards, on opening
twenty other mussels taken fresh from the same aquarium
and from the same batch, I invariably found the crystalline
style in them all, and at the same time noticed that digestion
had begun again in all of them. These observations show
conclusively that a functional relationship exists between
digestion and the crystalline style,—that the style either
somehow aids digestion, or else is a product or waste-product
of digestion. And we shall see presently that it aids diges-
tion in a very important way.

(6) But the most convincing, the most irrefragable proof

THE CRYSTALLINE STYLE OF LAMELLIBRANCHIA. 595

that the crystalline style cannot be regarded as a rudimentary
structure, is that it is an active amylolytic ferment, as we
shall see presently. Only if we could regard the ptyalin or
the pepsin of other animals as a rudimentary structure, could
we take the crystalline style as a rudimentary structure.

Let us now describe and state fully the physical, chemical,
and physiological characteristics of the crystalline style,
that is to say, see what it really is and what it really does.
After we have done that we shall dispose of the third and
the fourth hypotheses.

A fully-formed crystalline style that has been shed some
twelve hours is a flexible, solid transparent body that is
thicker at one end than at the other. Broadly speaking its
form is like that of a slender cone (see figs. 1, 2, 4, 5, 6).
Under the microscope it is seen to be longitudinally striated
(see figs. 4, 6). The striation is due to the fact that the style
is composed of concentric or rather co-axial (placed round a
common axis) layers of a colloid substance of greater and
lesser density. A cross section of a fully formed style looks
under the microscope like the cross section of an onion (see
fig. 3). We must state here that very often one notices in
the style of Anodon a central much softer core, which is much
less perfectly striated, and which has embedded in it particles
of food material (see fig. 5). In a freshly formed style in
Anodon, that has been shed a few minutes ago, this central
core forms a very marked feature. Under the microscope
it is seen to occupy three fourths or more of the space
occupied by the whole style, that is to say, to possess a
diameter that is three fourths or more as long as the diameter
of the whole style at the corresponding part; to bea viscous
liquid of a finely bubbly appearance ; and to be surrounded
by a comparatively thin nearly homogeneous sheath-like
layer (fig. 5). In fact, such a freshly formed style may
briefly be said to be formed of a viscous liquid that has got
formed round it, apparently through condensation of its own
substance, a thin sheath-like layer. This observation is very
important, as showing that a fully-formed style is not pro-

596 S. B. MITRA.

duced as a tough, solid substance, but is shed as a viscous,
finely bubbly liquid, which in the receptacle gradually gets
thicker and thicker in consistency, till it becomes a flexible
solid. It should be stated here that occasionally one notices
in a pretty freshly formed style in Anodon, an axial zone
that consists chiefly of particles of food material. Such a
zone is never found in fully formed styles in Anodon, and its
occasional presence in a freshly formed style in that animal
would seem to point to the imperfection of the method of
storing the ferment, and passing on the food material that
obtains in this species of the Lamellibranchia. It is inte-
resting to note in this connection, that such a zone is never
found in the style of a species of Pholas examined by me, in
which the style is lodged in a cecum, and not in the ali-
mentary canal itself. The style, let us add, in this species
of Pholas always possesses a central, liquid, finely bubbly
core (fig. 6), which never contains a particle of food material.
In fact, the main body of the style of this Pholas is always
an assertion that
cannot be made with regard to the style of Anodon. Under

free from a single particle of food material

these circumstances one is surely justified in thinking that
the complete freedom of Pholas’ style from food particles is
due to a superior, more differentiated mechanism for storing
the ferment and passing the food material through the ali-
mentary canal.

As is well known, the style is lodged in some species in
the alimentary canal itself (Anodon), in others (fig. 9) in a
diverticulum (cecum) of the canal, which starts from the
pyloric end of the stomach (Pholas). But the most curious
fact in this connection, that has never before been observed,
is that in Anodon the first portion of the intestine—the
portion that lodges the style—is divided into two longitudinal
compartments by two longitudinal ciliated ridges that project
into the lumen of the canal. A cross section of the canal at
this part, seen from behind, is represented diagrammatically
in the figure 7. One ridge is placed dorsally, the other
ventrally. The two compartments may, therefore, be called

THE CRYSTALLINE STYLE OF LAMELLIBRANCHIA. 597

one the right and the other the left compartment. They
are open at both ends: but the stomach-ends of both are
guarded by a cuticular value. It is in the left compartment
that the style is lodged. Now what is the use of the right
compartment ? It took me some time to find out. It is
through the right compartment that food material is, as a
rule, passed on from the stomach into the rest of the alimen-
tary canal. By careful dissection of the animals when they
are digesting food actively, one can see with the naked eye
particles of food material forming a very slender cord (figs
7, 8), and being hurried along through this compartment by
the action of the cilia. It is easy to see that if there were no
such division of the lumen into compartments there would be
no proper storage of the ferment, and food material might be
passed on without being properly mixed up with the ferment.
It is not out of place to mention here that the cxcal
diverticulum in Pholas that lodges the crystalline style
arises from the same point of the stomach as the intestine,
that it runs parallel with the first portion of the intestine,
and that it is placed to the left of the same portion (fig. 9).
These facts with regard to the point of origin and the position
of the cecum in Pholas, coupled with the undoubted fact
that Pholas is a more highly specialised form than Anodon,
and the fact mentioned above, that the crystalline style when
it is lodged in a cecum is completely free from food particles,
whereas, when it is lodged in the alimentary canal itself
food particles are occasionally found in its substance, force
the idea on one’s mind that this blind cxcum is only a
differentiated part of the first portion of the intestine, and
that it has been evolved from the left compartment of the
first portion of the intestine of Anodon by the permanent
coalescence of the ciliated ridges and shutting up of the
distal end of the compartment. It is to be remembered here
that this phenomenon—viz. the permanent coalescence of two
ciliated surfaces—must have taken place extensively and over
and over again in the evolution of the more complicated
forms of the Lamellibranch gill-plate,

598 S. B. MITRA.

Here also the important fact should be stated definitely,
that one end of the crystalline style, whether it is lodged in
a cecal diverticuleam or in the alimentary canal itself,
invariably projects a little into the stomach, and that this
end is as invariably surrounded by and has embedded in its
substance a very considerable quantity of food material (fig.
8). In fact, one can see that this projecting end is slowly
and gradually dissolved in the stomach, and is there mixed
up with food material, which then is passed on into the
intestine (fig. 8). In Anodon the thicker end projects into
the stomach, in Pholas the thinner end.

One more fact should be mentioned here, and that is that
there is no cellular element in the style, and that it is com-
posed entirely of a colloid substance.

Let us now describe the chemical properties of the style.
The style is soluble in distilled water, but this solubility is
due to the presence in the style of a minute quantity of salts.
The solution is neutral. That it is a proteid substance is
proved by the following colour reactions. It gives the xantho-
proteic reaction with nitric acid and ammonia or caustic
potash ; the Millon’s reaction with Millon’s reagent ; and gives
violet coloration with copper sulphate and caustic potash
(Piotrowski’s reaction). Its solution in distilled water coagu-
lates on heating. (The solution must be concentrated, other-
wise this result cannot be obtained.) The solution in dis-
tilled water is precipitated by nitric acid. The precipitate is not
dissolved by heat. The solution is also precipitated by tannin,
ammonium sulphate, magnesium sulphate, sodium chloride,
and alcohol.

It is important to know to what class of proteids the style
belongs. It belongs to the globulin class—the class to which
fibrin-ferment, which has been proved to bea proteid, belongs.
It is not an albumin, because its solution in water is com-
pletely precipitated by saturation with magnesium sulphate.
It is not an acid or alkali-albumin, because its solution in
water is neutral, and coagulates on heating. ‘There is no
peptone in the style, because when its solution in water is

THE CRYSTALLINE STYLE OF LAMELLIBRANCHIA. 599

saturated with ammonium sulphate and filtered, the filtrate
does not give the characteristic reaction of a peptone.
Nor are there any albumoses in the style, because when its
solution in water is precipitated with alcohol, and the precipi-
tate has been kept under absolute alcohol for six months,
and is then treated with water, the water does not dissolve
the slightest quantity of the precipitate, and does not give
the characteristic reactions of albumoses. he method of
exclusion, therefore, shows that the proteid of the style must
belong to the globulin class, and it shows the characteristics
of that class. It is soluble in dilute saline solutions, and
insoluble in concentrated solutions of NaCl, Mg SO, and
Am, SO,._ Its solution in water, as has been stated above, is
precipitated by heat.

Analysis shows that there is about 88 per cent. of
water in the style, about 12 per cent. of a proteid
(globulin), and about 1 per cent. of salts. As far, therefore,
as the proportion of water to the solids goes, the composition
of the style is not far different from that of the pancreatic
secretion of the dog.

Let us now examine the physiological properties of the
style. If two styles from fresh-water mussels are added to
thirty minims of starch solution, all the starch will be con-
verted into reducible sugar in about three hours. If seven
styles are dissolved in distilled water, and the solution added
to thirty minims of the same starch solution, all the starch
will be transformed into a reducible sugar in about twenty
minutes. ‘lhese very simple experiments show that there is
an amylolytic ferment in the style. An intermediate product
of the nature of dextrin is formed, just as it is formed during
the conversion of starch into sugar by saliva. In fact, all the
stages of this transformation of starch into sugar in salivary
digestion are noticeable during the conversion of starch into
sugar by the crystalline style. It should be stated here that
the style can also transform raw starch into sugar, but more
slowly. It also acts on glycogen as ptyalin does, converting
it slowly into sugar. We could not detect any action of the

600 S. B. MITRA.

crystalline style on a proteid such as egg-albumin, boiled egg
fibrin, muscular fibres, etc.

But granted, it may be said, that there is an amylolytic
ferment in the style, still the proteid in the style may serve
as a reserve of proteid nutriment. And so, it may be said,
the fourth hypothesis may still be regarded to embody, not
the whole truth it is true, but part of the truth. But this
idea is negatived by several considerations and facts. In the
first place a proteid matter (like that of the style) reserved as
a food material in an adult animal is practically unknown in
the animal kingdom, Then, if the style solution in water is
precipitated by alcohol, and the precipitate, consisting of a
globulin, kept under alcohol, it is seen that while the super-
natant alcohol contains no ferment matter, the more insoluble
the precipitate becomes, the less is the ferment-activity of
the precipitate ; until at last, when the precipitate becomes
completely insoluble in water, its ferment-activity is also lost
completely. This points strongly to the conclusion that the
proteid of the style and the ferment are identical. But the
most striking proof that they are identical is furnished by
the fact that the temperature at which the proteid in a
watery solution of the style coagulates, and so loses its dis-
tinctive characteristics, is the same as that at which the
solution loses its ferment-activity completely. Under the
circumstances the proteid of the style and the ferment must
be regarded as identical. This being so, one cannot regard
that proteid as a reserve of nutriment.

After all this, it is hardly necessary to say that the third
hypothesis, that the style is to be regarded as an excretory
matter, is quite untenable. It is not an excretion in the
strict sense of the word, because it is a ferment, and a very
active and important ferment too. Besides, we know that it
performs an important function in the organism by coming into
contact with the food material in the stomach and acting on it.

The style cannot be regarded as a product of digestion,
because there is neither acid-albumin, nor alkali-albumin,
nor any albumose, nor any peptone in it,

THE CRYSTALLINE STYLE OF LAMELLIBRANCHIA. 601

Now the important question arises, Where does the crystal-
line style come from? There are grounds for believing that
it is secreted by the so-called liver. The chief ground is that
there is in the liver an amylolytic ferment exactly like the
ferment of the style. The ferment in the liver behaves
exactly as the style ferment does. On the other hand, we
could hardly detect any amylolytic ferment in the enteric
epithelium. The very small quantity that may occasionally
be detected may be due to the adherence to the epithelium
of a minute quantity of the ferment from the outer surface
of the style. There is also another fact, which must be allowed
to have some force in this connection. It is that yellow pig-
ment cells from the liver are occasionally seen to form the
axial zone of freshly formed styles (fig. 10). If this fact
does not show anything else, it shows at least the possibility
of a product of the liver being carried easily into the recep-
tacle in which the style is lodged. If pigment-fed cells from
the liver may pass into the receptacle, a secretion of the liver
may do so too.

The crystalline style has been found in Mactra, Donax,
Unio, Anodon, in a species of Pholas, in a species of Mytilus,
and in other species. Considering the high importance of
the ferment, one is justified in predicting that future obser-~
vation will reveal its presence in all the species of the
Lamellibranchia. Its presence in the Lipocephala is pos-
sibly connected with the absence in them of specially differ-
entiated salivary glands, just as it is possible to connect this
absence of special salivary glands with the absence of a
buccal mass. But the styles and cords which appear to
represent it in Gastropoda should be now examined afresh,
both as to their origin and chemical properties.

The conclusion that is to be drawn from the foregoing
observations and experiments is, that the crystalline style is
an active amylolytic ferment; that it is secreted as a viscous
liquid, most probably by the liver; that it is stored up as a
flexible solid in the cecum, or in a compartment of the
alimentary canal itself; that the end of it that projects into

VoL. 44, pART 4,—NEW SERIES. QQ

602 S. B. MI'TRA.

the stomach is slowly and gradually dissolved there, and is
mixed there with particles of food material, the starchy
portion of which is transformed by it into a reducible sugar.

DESCRIPTION OF PLATE 42,

Illustrating Mr. 8. B. Mitra’s paper on “The Crystalline
Style of Lamellibranchia.”

Fre. 1.—Crystalline style of Anodon, natural size.
Fic. 2.—Enlarged drawing of another specimen.

Fic. 3.—Transverse section of the crystalline style of Anodon to show the
laminated structure.

Fie. 4.—Greatly enlarged view of a style of Anodon, showing the longi-
tudinal striation and attached food particles.

Fic. 5.—Optical longitudinal section of a similar specimen.

Fic. 6.—Optical longitudinal section of a crystalline style of Pholas.

Fic. 7.—Transverse section, showing the right and left compartments of
that portion of the intestine which lodges the crystalline style in Anodon
(which has no special caecum for the style as has Pholas).

Fie. 8.—Diagram showing the crystalline style in the left compartment of
the intestine of Anodon with the stream of food particles in the right com-
partment.

Fig. 9.—Diagram showing the separate cecum and intestine of Pholas,
formed by completion and fusion of the dividing ridges seen in Fig. 7.

Fic. 10.—Crystalline style of Anodon, showing pigmented liver-cells in
the axis.

N.B.—Details are explained by the lettering on the Plate. All the Figs.
except Fig. 6 and Fig. 9 represent the crystalline style of Anodon,

INDEX TO VOL. 44,

NEW SERIES.

Acanthodrilus, ecelomic fluid of, by |

Benham, 565

Benham on the ccelomic fluid of
Acanthodrilus, 565

Benham on Heteropleuron Hec-
tori, 273

Bernard, studies in the retina, 443

Brazilin, staining with, by Hickson, |

469

Ceelomie fluid of Acanthodrilus, by
Benham, 565

Crystalline style of Lamellibranchia,
by Mitra, 591

Dolichorhynchus indicus, a

new Acraniate, by Arthur Willey, |

269

Drew, Professor Gilman, on_ life-
history of Nucula delphino-
donta, 313

Kchinus esculentus, parasites
from, by Shipley, 281

EKoperipatus, new genus from the
Malay peninsula, by Evans, 539

Ephydatia blembingia, by
Richard Evans, 71
Evans, Richard, on Ephydatia

blembingia, 71

Kvans, Richard, on new species of
Peripatus from the Malay pen-
insula, 473

Frog, rods and cones of retina of the,
by Bernard, 443

Goodrich on Saccocirrus, 413

Harrison on the development and suc-
cession of the teeth in Hatteria
punctata, 161

Hatteria punctata, teeth of, by
Spencer Harrison, 161

Heape, Walter, on the ‘
son’”’ of mammals, 1

Heteropleuron Hectori,
New Zealand lancelet, by Blaxland
Benham, 273

Hickson, staining with brazilin, 469

sexual sea-

the

Lamellibranchia, crystalline style of,
by Mitra, 591

Lancelet, new, by Benham, 273

— by Willey, 269

Malarial diseases, priority of dis-
coveries in regard to, by Nuttall,
429

Mammals, ‘‘sexual season”
Walter Heape, 1

of, by

604

Menstruation and “sexual season”
of mammals, by Heape, 1

Mitra, S. B., on the crystalline style
of Lamellibranchia, 591

Molgula manhattensis, the pro-
tostigmata of, by Arthur Willey,
141

Mylodon Listai, structure of hair
of, by Ridewood, 393

Nemerteans from Singapore, by R.C.
Punnett, 111

Nemerteans, new genera of, by R. C.
Punnett, 547

Nucula delphinodonta, life-his-
tory of, by Gilman Drew, 313

Nuttall on priority with regard to
discoveries as to etiology of ma-
larial diseases, 429

Parasites in Echinus esculentus,
by Shipley, 281

Peripatus, new species from the
Malay peninsula, by Evans, 473

Pleurotomaria Beyricihii, the
anatomy of, by Martin Woodward,
215

Pocock on the Scottish Silurian Scor-
pion, 291

Pro-cestrum and menstruation, by

Heape, 1

INDEX.

; Punnett on Nemerteans from Singa-

pore, 111
Punnett on new genera of Nemer-
teans, 547

Retina, studies in the, by Bernard,
443

Ridewood on the structure of the
hairs of Mylodon Listai, 393

Saccocirrus, structure and affinities
of, by Edwin 8. Goodrich, 413

Scorpion, the Scottish Silurian, by
Pocock, 291

“Sexual season” of mammals, by
Heape, 1

Shipley on parasites in Nchinus
esculentus, 281

Sponge, new freshwater, by Evans, 7]

Staining with Brazilin, by Hickson,
469

Teeth of Hatteria punctata, by
Spencer Harrison, 161

Willey on Dolichorhynchus, 269

Willey on the protostigmata of Mol -
gula manhattensis, 141

Woodward, Martin, on the anatomy
of Pleurotomaria, 215

PRINTED BY ADLARD AND SON,
BARTHOLOMEW CLOSE, E.C., AND 20 HANOVER SQUARE, W.

With Ten Plates, Royal 4to, 5s.
CONTRIBUTIONS TO THE KNOWLEDGE OF RHABDOPLEURA
AND AMPHIOXUS.
By E. RAY LANKESTER, M.A., LL.D., F.R.S.
London: J. & A. CuHurcuity, 7, Great Marlborough Street.

Quarterly Journal of Microscopical

Science.

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

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

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

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

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

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

TO CORRESPONDENTS.

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

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

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

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

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

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

For every four pages or less—

25 copies : g ; : 5/-
Bi) anor . 6/-
“TOYO ase ate : 7/-

Plates, 2/- per 25 if uncoloured; if coloured, at the same rate for
every colour.
Prepayment by P.O. Order is requested.
ALL COMMUNICATIONS FOR THE EDITORS TO BE ADDRESSED TO THE CARE
or Mussrs. J. & A. Cuurcuiny, 7, Great MarLporouGH Street,
Lonpon, W.

THE MARINE BIOLOGICAL ASSOCIATION

OF THE

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

20;

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

The Laboratory at Plymouth

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

Naturalists desiring to work at the Laboratory

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

Works published by the Association

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

The Journal of the Marine Biological Association

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

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

Specimens of Marine Animals and Plants,

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

TERMS OF MEMBERSHIP.

ANNUAL MEMBERS . : é . &1 1 Operannun.

LIFE MEMBERS . : é : . 15 15 O Composition Fee.
FOUNDERS . ; ; é ; 7 LOORTOR AO 5 35
Governors (Life Members of Council) 500 O O By 5

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

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

The DIRECTOR,
The Laboratory,
Plymouth.

i ¥ . A a
tt a :
a» J : :
ke i
mere F
oe j

‘
g “
.
td
-
.
‘
a t
i j
.
a : '
i
i i

bs
ol ‘ eS) : 7 7
“ «
i f-
» .
tal of
+
,
7 4
. = -
>
°
= ”
’ a =
‘ Ae
te ele -
a y
= . * =
— 7 - —
» .
1 5
|
a i z
das -
- ri 8
a
if
= » = ’
e
. = A * —~ 6
. . 7
Ls :
= Sage!
J + iw a
} >
‘
-
at &
q - - >»
; -
- 7 ‘
= 7 ; =
= oe ; i
_ bl i
- i" '
x
> : ~
ya 7 5
: se
. 7 7 s
is <4
‘
~ a S
- 7
: - a
. ‘
7 -
, 3)
Sy
boa aa i .
-
i ‘ ’
F »
®
a a
' -
; - 7 ne : 7
‘

SUT

3 2044 1

COere ee rege pene

~reqercer,
TOR

>),

ELLE IPE IIE LLOEGR LON et

\ ies

F- | BRS:
Ad ‘ef

oS th ‘

SE RRS

\ i
t

SEER Bee